1 2014-04-30 Marek Polacek <polacek@redhat.com>
3 * gcc.c (sanitize_spec_function): Handle SANITIZE_FLOAT_DIVIDE.
4 * builtins.def: Initialize builtins even for SANITIZE_FLOAT_DIVIDE.
5 * flag-types.h (enum sanitize_code): Add SANITIZE_FLOAT_DIVIDE.
6 * opts.c (common_handle_option): Add -fsanitize=float-divide-by-zero.
8 2014-04-29 Alan Lawrence <alan.lawrence@arm.com>
10 * config/aarch64/arm_neon.h (vzip1_f32, vzip1_p8, vzip1_p16, vzip1_s8,
11 vzip1_s16, vzip1_s32, vzip1_u8, vzip1_u16, vzip1_u32, vzip1q_f32,
12 vzip1q_f64, vzip1q_p8, vzip1q_p16, vzip1q_s8, vzip1q_s16, vzip1q_s32,
13 vzip1q_s64, vzip1q_u8, vzip1q_u16, vzip1q_u32, vzip1q_u64, vzip2_f32,
14 vzip2_p8, vzip2_p16, vzip2_s8, vzip2_s16, vzip2_s32, vzip2_u8,
15 vzip2_u16, vzip2_u32, vzip2q_f32, vzip2q_f64, vzip2q_p8, vzip2q_p16,
16 vzip2q_s8, vzip2q_s16, vzip2q_s32, vzip2q_s64, vzip2q_u8, vzip2q_u16,
17 vzip2q_u32, vzip2q_u64): Replace inline __asm__ with __builtin_shuffle.
19 2014-04-29 David Malcolm <dmalcolm@redhat.com>
21 * tree-cfg.c (dump_function_to_file): Dump the return type of
22 functions, in a line to itself before the function body, mimicking
23 the layout of a C function.
25 2014-04-29 Jakub Jelinek <jakub@redhat.com>
27 PR tree-optimization/60971
28 * tree-tailcall.c (process_assignment): Reject conversions which
31 2014-04-29 James Greenhalgh <james.greenhalgh@arm.com>
33 * calls.c (initialize_argument_information): Always treat
34 PUSH_ARGS_REVERSED as 1, simplify code accordingly.
35 (expand_call): Likewise.
36 (emit_library_call_calue_1): Likewise.
37 * expr.c (PUSH_ARGS_REVERSED): Do not define.
38 (emit_push_insn): Always treat PUSH_ARGS_REVERSED as 1, simplify
41 2014-04-29 Nick Clifton <nickc@redhat.com>
43 * config/msp430/msp430.md (umulsidi): Fix typo.
44 (mulhisi3): Enable even inside interrupt handlers.
45 * config/msp430/msp430.c (msp430_print_operand): %O: Allow for the
46 bigger return address pushed in large mode.
48 2014-04-29 Nick Clifton <nickc@redhat.com>
50 * config/arc/arc.c (arc_select_cc_mode): Fix parentheses.
51 (arc_init_reg_tables): Use a machine_mode enum to iterate over
53 * config/m32r/m32r.c (init_reg_tables): Likewise.
54 * config/m32c/m32c.c (m32c_illegal_subreg_p): Use a machine_mode
55 enum to hold the modes.
57 2014-04-29 Richard Biener <rguenther@suse.de>
59 * dominance.c (free_dominance_info): Add overload with
61 (dom_info_state): Likewise.
62 (dom_info_available_p): Likewise.
63 * basic-block.h (free_dominance_info, dom_info_state,
64 dom_info_available_p): Declare overloads.
65 * passes.c (execute_function_todo): Verify that verifiers
66 don't change dominator info state. Drop dominator info
67 for IPA pass invocations.
68 * cgraph.c (release_function_body): Restore asserts that
69 dominator information is released.
71 2014-04-29 Patrick Palka <patrick@parcs.ath.cx>
73 * doc/invoke.texi: Fix typo.
74 * tree-vrp.c: Fix typos.
75 * gimple.c (infer_nonnull_range): Reorder operands of an &&
78 2014-04-29 Zhenqiang Chen <zhenqiang.chen@linaro.org>
80 * config/aarch64/aarch64.md (mov<mode>cc): New for GPF.
82 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
84 * config/aarch64/aarch64-builtins.c
85 (aarch64_types_storestruct_lane_qualifiers): New.
86 (TYPES_STORESTRUCT_LANE): Likewise.
87 * config/aarch64/aarch64-simd-builtins.def (st2_lane): New.
90 * config/aarch64/aarch64-simd.md (vec_store_lanesoi_lane<mode>): New.
91 (vec_store_lanesci_lane<mode>): Likewise.
92 (vec_store_lanesxi_lane<mode>): Likewise.
93 (aarch64_st2_lane<VQ:mode>): Likewise.
94 (aarch64_st3_lane<VQ:mode>): Likewise.
95 (aarch64_st4_lane<VQ:mode>): Likewise.
96 * config/aarch64/aarch64.md (unspec): Add UNSPEC_ST{2,3,4}_LANE.
97 * config/aarch64/arm_neon.h
98 (__ST2_LANE_FUNC): Rewrite using builtins, update use points to
99 use new macro arguments.
100 (__ST3_LANE_FUNC): Likewise.
101 (__ST4_LANE_FUNC): Likewise.
102 * config/aarch64/iterators.md (V_TWO_ELEM): New.
103 (V_THREE_ELEM): Likewise.
104 (V_FOUR_ELEM): Likewise.
106 2014-04-28 David Malcolm <dmalcolm@redhat.com>
108 * doc/gimple.texi: Replace the description of the now-defunct
109 union gimple_statement_d with a diagram showing the
110 gimple_statement_base class hierarchy and its relationships to
111 the GSS_ and GIMPLE_ enums.
113 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
115 * config/aarch64/aarch64-protos.h (aarch64_modes_tieable_p): New.
116 * config/aarch64/aarch64.c
117 (aarch64_cannot_change_mode_class): Weaken conditions.
118 (aarch64_modes_tieable_p): New.
119 * config/aarch64/aarch64.h (MODES_TIEABLE_P): Use it.
121 2014-04-28 Pat Haugen <pthaugen@us.ibm.com>
123 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
124 (loadsync_<mode>): Change mode.
125 (load_quadpti, store_quadpti): New.
126 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
127 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
129 2014-04-28 Martin Jambor <mjambor@suse.cz>
131 * tree-sra.c (sra_modify_expr): Generate new memory accesses with
132 same alias type as the original statement.
133 (subreplacement_assignment_data): New type.
134 (handle_unscalarized_data_in_subtree): New type of parameter,
135 generate new memory accesses with same alias type as the original
137 (load_assign_lhs_subreplacements): Likewise.
138 (sra_modify_constructor_assign): Generate new memory accesses with
139 same alias type as the original statement.
141 2014-04-28 Richard Biener <rguenther@suse.de>
143 * tree-pass.h (TODO_verify_il): Define.
144 (TODO_verify_all): Complete properly.
145 * passes.c (execute_function_todo): Move existing loop-closed
146 SSA verification under TODO_verify_il.
147 (execute_one_pass): Trigger TODO_verify_il at todo-after time.
148 * graphite-sese-to-poly.c (rewrite_cross_bb_scalar_deps):
149 Fix tree sharing issue.
151 2014-04-28 Richard Biener <rguenther@suse.de>
154 * builtins.def (DEF_C11_BUILTIN): Add.
155 (BUILT_IN_ALIGNED_ALLOC): Likewise.
156 * coretypes.h (enum function_class): Add function_c11_misc.
157 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
158 BUILT_IN_ALIGNED_ALLOC like BUILT_IN_MALLOC.
159 (call_may_clobber_ref_p_1): Likewise.
160 * tree-ssa-dce.c (mark_stmt_if_obviously_necessary): Likewise.
161 (mark_all_reaching_defs_necessary_1): Likewise.
162 (propagate_necessity): Likewise.
163 (eliminate_unnecessary_stmts): Likewise.
164 * tree-ssa-ccp.c (evaluate_stmt): Handle BUILT_IN_ALIGNED_ALLOC.
166 2014-04-28 Richard Biener <rguenther@suse.de>
168 * tree-vrp.c (vrp_var_may_overflow): Remove.
169 (vrp_visit_phi_node): Rather than bumping to +-INF possibly
170 with overflow immediately bump to one before that value and
171 let iteration figure out overflow status.
173 2014-04-28 Richard Biener <rguenther@suse.de>
175 * configure.ac: Do valgrind header checks unconditionally.
176 Add --enable-valgrind-annotations.
177 * system.h: Guard valgrind header inclusion with
178 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING.
179 * alloc-pool.c (pool_alloc, pool_free): Use
180 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING
181 to guard possibly dead code.
182 * config.in: Regenerated.
183 * configure: Likewise.
185 2014-04-28 Jeff Law <law@redhat.com>
187 PR tree-optimization/60902
188 * tree-ssa-threadedge.c
189 (record_temporary_equivalences_from_stmts_at_dest): Only iterate
190 over real defs when invalidating outputs from statements that do not
191 produce useful outputs for threading.
193 2014-04-28 Richard Biener <rguenther@suse.de>
195 PR tree-optimization/60979
196 * graphite-scop-detection.c (scopdet_basic_block_info): Reject
197 SCOPs that end in a block with a successor with abnormal
200 2014-04-28 Richard Biener <rguenther@suse.de>
202 * tree-pass.h (execute_pass_list): Adjust prototype.
203 * passes.c (pass_manager::execute_early_local_passes):
205 (do_per_function): Change callback signature, push all actual
206 work to the callbals.
207 (do_per_function_toporder): Likewise.
208 (execute_function_dump): Adjust.
209 (execute_function_todo): Likewise.
210 (clear_last_verified): Likewise.
211 (verify_curr_properties): Likewise.
212 (update_properties_after_pass): Likewise.
213 (execute_pass_list_1): Split out from ...
214 (execute_pass_list): ... here. Adjust.
215 (execute_ipa_pass_list): Likewise.
216 * cgraphunit.c (cgraph_add_new_function): Adjust.
217 (analyze_function): Likewise.
218 (expand_function): Likewise.
219 * cgraph.c (release_function_body): Free dominance info
220 here instead of asserting it was magically freed elsewhere.
222 2014-04-28 Eric Botcazou <ebotcazou@adacore.com>
224 * configure.ac: Tweak GAS check for LEON instructions on SPARC.
225 * configure: Regenerate.
226 * config/sparc/sparc.opt (muser-mode): New option.
227 * config/sparc/sync.md (atomic_compare_and_swap<mode>_1): Do not enable
229 (atomic_compare_and_swap_leon3_1): New instruction for LEON3.
230 * doc/invoke.texi (SPARC options): Document -muser-mode.
232 2014-04-27 Richard Sandiford <rdsandiford@googlemail.com>
234 * cselib.c (find_slot_memmode): Delete.
235 (cselib_hasher): Change compare_type to a struct.
236 (cselib_hasher::equal): Update accordingly. Don't expect wrapped
238 (preserve_constants_and_equivs): Adjust for new compare_type.
239 (cselib_find_slot): Likewise. Take the mode of the rtx as argument.
240 (wrap_constant): Delete.
241 (cselib_lookup_mem, cselib_lookup_1): Update calls to cselib_find_slot.
243 2014-04-26 Markus Trippelsdorf <markus@trippelsdorf.de>
245 * doc/install.texi (Building with profile feedback): Remove
248 2014-04-26 Tom de Vries <tom@codesourcery.com>
250 * config/i386/i386.md (define_expand "ldexpxf3"): Fix out-of-bounds
253 2014-04-25 Cary Coutant <ccoutant@google.com>
256 * dwarf2out.c (should_move_die_to_comdat): A type definition
257 can contain a subprogram definition, but don't move it to a
259 (clone_as_declaration): Copy DW_AT_abstract_origin attribute.
260 (generate_skeleton_bottom_up): Remove DW_AT_object_pointer attribute
262 (clone_tree_hash): Rename to...
263 (clone_tree_partial): ...this; change callers. Copy
264 DW_TAG_subprogram DIEs as declarations.
265 (copy_decls_walk): Don't copy children of a declaration into a
268 2014-04-25 H.J. Lu <hongjiu.lu@intel.com>
271 * config/i386/i386.md (*movsf_internal): Set MODE to SI for
274 2014-04-25 Jiong Wang <jiong.wang@arm.com>
276 * config/arm/predicates.md (call_insn_operand): Add long_call check.
277 * config/arm/arm.md (sibcall, sibcall_value): Force the address to
279 * config/arm/arm.c (arm_function_ok_for_sibcall): Remove long_call
282 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
284 * config/arm/arm.c (arm_cortex_a8_tune): Initialise
287 2014-04-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
289 PR tree-optimization/60930
290 * gimple-ssa-strength-reduction.c (create_mul_imm_cand): Reject
291 creating a multiply candidate by folding two constant
292 multiplicands when the result overflows.
294 2014-04-25 Jakub Jelinek <jakub@redhat.com>
296 PR tree-optimization/60960
297 * tree-vect-generic.c (expand_vector_operation): Only call
298 expand_vector_divmod if type's mode satisfies VECTOR_MODE_P.
300 2014-04-25 Tom de Vries <tom@codesourcery.com>
302 * expr.c (clobber_reg_mode): New function.
303 * expr.h (clobber_reg): New function.
305 2014-04-25 Tom de Vries <tom@codesourcery.com>
307 * rtlanal.c (find_all_hard_reg_sets): Note INSN_CALL_FUNCTION_USAGE
310 2014-04-25 Radovan Obradovic <robradovic@mips.com>
311 Tom de Vries <tom@codesourcery.com>
313 * rtlanal.c (find_all_hard_reg_sets): Add bool implicit parameter and
315 * rtl.h (find_all_hard_reg_sets): Add bool parameter.
316 * haifa-sched.c (recompute_todo_spec, check_clobbered_conditions): Add
317 new argument to find_all_hard_reg_sets call.
319 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
321 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p):
322 Use HOST_WIDE_INT_C for mask literal.
323 (aarch_rev16_shleft_mask_imm_p): Likewise.
325 2014-04-25 Eric Botcazou <ebotcazou@adacore.com>
328 * config/sparc/sparc.md (ashlsi3_extend): Delete.
330 2014-04-25 Marc Glisse <marc.glisse@inria.fr>
332 PR preprocessor/56540
333 * config/i386/i386-c.c (ix86_target_macros): Define
334 __SIZEOF_FLOAT80__ and __SIZEOF_FLOAT128__.
336 2014-04-25 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
338 * configure.ac (tga_func): Remove.
339 (LIB_TLS_SPEC): Remove.
340 * configure: Regenerate.
341 * config.in: Regenerate.
342 * config/sol2.h (LIB_SPEC): Don't use LIB_TLS_SPEC.
344 2014-04-25 Richard Biener <rguenther@suse.de>
347 * tree-ssa-structalias.c (ipa_pta_execute): Compute direct
348 call stmt use/clobber sets during stmt walk instead of
349 walking the possibly incomplete set of caller edges.
351 2014-04-25 Richard Biener <rguenther@suse.de>
354 * passes.c (apply_ipa_transforms): Inline into only caller ...
355 (execute_one_pass): ... here. Properly bring in function
356 bodies for nodes we want to apply IPA transforms to.
358 2014-04-24 Cong Hou <congh@google.com>
360 PR tree-optimization/60896
361 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Pick up
362 all statements in PATTERN_DEF_SEQ in recognized widen-mult pattern.
363 (vect_mark_pattern_stmts): Set the def type of all statements in
364 PATTERN_DEF_SEQ as vect_internal_def.
366 2014-04-24 Michael Meissner <meissner@linux.vnet.ibm.com>
368 * doc/extend.texi (PowerPC Built-in Functions): Document new
369 powerpc extended divide, bcd, pack/unpack 128-bit, builtin
371 (PowerPC AltiVec/VSX Built-in Functions): Likewise.
373 * config/rs6000/predicates.md (const_0_to_3_operand): New
374 predicate to match 0..3 integer constants.
376 * config/rs6000/rs6000-builtin.def (BU_DFP_MISC_1): Add new macros
377 to support adding miscellaneous builtin functions.
378 (BU_DFP_MISC_2): Likewise.
379 (BU_P7_MISC_1): Likewise.
380 (BU_P7_MISC_2): Likewise.
381 (BU_P8V_MISC_3): Likewise.
382 (BU_MISC_1): Likewise.
383 (BU_MISC_2): Likewise.
384 (DIVWE): Add extended divide builtin functions.
392 (DXEX): Add decimal floating-point builtin functions.
404 (CDTBCD): Add new BCD builtin functions.
408 (BCDADD_LT): Likewise.
409 (BCDADD_EQ): Likewise.
410 (BCDADD_GT): Likewise.
411 (BCDADD_OV): Likewise.
413 (BCDSUB_LT): Likewise.
414 (BCDSUB_EQ): Likewise.
415 (BCDSUB_GT): Likewise.
416 (BCDSUB_OV): Likewise.
417 (PACK_TD): Add new pack/unpack 128-bit type builtin functions.
418 (UNPACK_TD): Likewise.
420 (UNPACK_TF): Likewise.
421 (UNPACK_TF_0): Likewise.
422 (UNPACK_TF_1): Likewise.
423 (PACK_V1TI): Likewise.
424 (UNPACK_V1TI): Likewise.
426 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Add
427 support for decimal floating point builtin functions.
428 (rs6000_expand_ternop_builtin): Add checks for the new builtin
429 functions that take constant arguments.
430 (rs6000_invalid_builtin): Add decimal floating point builtin
432 (rs6000_init_builtins): Setup long double, _Decimal64, and
433 _Decimal128 types for new builtin functions.
434 (builtin_function_type): Set the unsigned flags appropriately for
435 the new builtin functions.
436 (rs6000_opt_masks): Add support for decimal floating point builtin
439 * config/rs6000/rs6000.h (RS6000_BTM_DFP): Add support for decimal
440 floating point builtin functions.
441 (RS6000_BTM_COMMON): Likewise.
442 (RS6000_BTI_long_double): Likewise.
443 (RS6000_BTI_dfloat64): Likewise.
444 (RS6000_BTI_dfloat128): Likewise.
445 (long_double_type_internal_node): Likewise.
446 (dfloat64_type_internal_node): Likewise.
447 (dfloat128_type_internal_node): Likewise.
449 * config/rs6000/altivec.h (UNSPEC_BCDADD): Add support for ISA
450 2.07 bcd arithmetic instructions.
451 (UNSPEC_BCDSUB): Likewise.
452 (UNSPEC_BCD_OVERFLOW): Likewise.
453 (UNSPEC_BCD_ADD_SUB): Likewise.
454 (bcd_add_sub): Likewise.
455 (BCD_TEST): Likewise.
456 (bcd<bcd_add_sub>): Likewise.
457 (bcd<bcd_add_sub>_test): Likewise.
458 (bcd<bcd_add_sub>_test2): Likewise.
459 (bcd<bcd_add_sub>_<code>): Likewise.
460 (peephole2 for combined bcd ops): Likewise.
462 * config/rs6000/dfp.md (UNSPEC_DDEDPD): Add support for new
463 decimal floating point builtin functions.
464 (UNSPEC_DENBCD): Likewise.
465 (UNSPEC_DXEX): Likewise.
466 (UNSPEC_DIEX): Likewise.
467 (UNSPEC_DSCLI): Likewise.
468 (UNSPEC_DSCRI): Likewise.
469 (D64_D128): Likewise.
470 (dfp_suffix): Likewise.
471 (dfp_ddedpd_<mode>): Likewise.
472 (dfp_denbcd_<mode>): Likewise.
473 (dfp_dxex_<mode>): Likewise.
474 (dfp_diex_<mode>): Likewise.
475 (dfp_dscli_<mode>): Likewise.
476 (dfp_dscri_<mode>): Likewise.
478 * config/rs6000/rs6000.md (UNSPEC_ADDG6S): Add support for new BCD
480 (UNSPEC_CDTBCD): Likewise.
481 (UNSPEC_CBCDTD): Likewise.
482 (UNSPEC_DIVE): Add support for new extended divide builtin
484 (UNSPEC_DIVEO): Likewise.
485 (UNSPEC_DIVEU): Likewise.
486 (UNSPEC_DIVEUO): Likewise.
487 (UNSPEC_UNPACK_128BIT): Add support for new builtin functions to
488 pack/unpack 128-bit types.
489 (UNSPEC_PACK_128BIT): Likewise.
490 (idiv_ldiv): New mode attribute to set the 32/64-bit divide type.
491 (udiv<mode>3): Use idiv_ldiv mode attribute.
492 (div<mode>3): Likewise.
493 (addg6s): Add new BCD builtin functions.
496 (UNSPEC_DIV_EXTEND): Add support for new extended divide
498 (div_extend): Likewise.
499 (div<div_extend>_<mode>"): Likewise.
500 (FP128_64): Add support for new builtin functions to pack/unpack
502 (unpack<mode>): Likewise.
503 (unpacktf_0): Likewise.
504 (unpacktf_1): Likewise.
505 (unpack<mode>_dm): Likewise.
506 (unpack<mode>_nodm): Likewise.
507 (pack<mode>): Likewise.
508 (unpackv1ti): Likewise.
509 (packv1ti): Likewise.
511 2014-04-24 Vishnu K S <Vishnu.k_s@atmel.com>
513 * gcc/config/avr/avr.c: Add comment on why -fdelete-null-pointer-checks
516 2014-04-24 Jakub Jelinek <jakub@redhat.com>
518 * tree.h (OMP_CLAUSE_LINEAR_GIMPLE_SEQ): Define.
519 * gimplify.c (omp_is_private): Change last argument's type to int.
520 Only diagnose lastprivate if the simd argument is 1, only diagnose
521 linear if the simd argument is 2.
522 (gimplify_omp_for): Adjust omp_is_private callers. When adding
523 lastprivate or private, add the clause to OMP_FOR_CLAUSES. Pass
524 GOVD_EXPLICIT to omp_add_variable. For simd with collapse == 1
525 create OMP_CLAUSE_LINEAR rather than OMP_CLAUSE_PRIVATE for var.
526 If var != decl and decl is in OMP_CLAUSE_LINEAR, gimplify decl
527 increment to OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
528 * omp-low.c (scan_sharing_clauses, lower_lastprivate_clauses): Handle
529 OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
530 * tree-nested.c (convert_nonlocal_omp_clauses,
531 convert_local_omp_clauses): Handle OMP_CLAUSE_LINEAR.
533 2014-04-24 Segher Boessenkool <segher@kernel.crashing.org>
536 * config/m68k/m68k.md (extendplussidi): Don't allow memory for
539 2014-04-24 Dimitris Papavasiliou <dpapavas@gmail.com>
541 * flag-types.h (enum ivar_visibility): Add.
543 2014-04-24 Trevor Saunders <tsaunders@mozilla.com>
545 * config/sh/sh_treg_combine.c (sh_treg_combine::execute): Take
548 2014-04-24 Alan Lawrence <alan.lawrence@arm.com>
550 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Enable for bigendian.
552 2014-04-24 Radovan Obradovic <robradovic@mips.com>
553 Tom de Vries <tom@codesourcery.com>
555 * reg-notes.def (REG_NOTE (CALL_DECL)): New reg-note REG_CALL_DECL.
556 * calls.c (expand_call, emit_library_call_value_1): Add REG_CALL_DECL
558 * combine.c (distribute_notes): Handle REG_CALL_DECL reg-note.
559 * emit-rtl.c (try_split): Same.
561 2014-04-24 Radovan Obradovic <robradovic@mips.com>
562 Tom de Vries <tom@codesourcery.com>
564 * common.opt (fuse-caller-save): New option.
566 2014-04-24 Tejas Belagod <tejas.belagod@arm.com>
568 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Reverse order of elements
571 2014-04-24 Richard Biener <rguenther@suse.de>
573 * expr.c (expand_expr_real_1): Avoid gimple_assign_rhs_to_tree
574 during TER and instead use the sepops interface for expanding
575 non-GIMPLE_SINGLE_RHS.
577 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
579 * config/i386/sol2.h (ASM_PREFERRED_EH_DATA_FORMAT): Only redefine
580 if not HAVE_AS_IX86_DIFF_SECT_DELTA.
582 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
584 * configure.ac (gcc_cv_as_cfi_directive): Support Solaris/x86
585 assembler 64-bit option.
586 * configure: Regenerate.
588 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
590 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Check
591 TARGET_SIMD rather than TARGET_GENERAL_REGS_ONLY.
592 (TARGET_SIMD): Take AARCH64_ISA_SIMD into account.
593 (TARGET_FLOAT): Take AARCH64_ISA_FP into account.
594 (TARGET_CRYPTO): Take TARGET_SIMD into account.
596 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
598 * config/aarch64/aarch64-builtins.c
599 (aarch64_builtin_vectorized_function): Handle BUILT_IN_BSWAP16,
600 BUILT_IN_BSWAP32, BUILT_IN_BSWAP64.
601 * config/aarch64/aarch64-simd.md (bswap<mode>): New pattern.
602 * config/aarch64/aarch64-simd-builtins.def: Define vector bswap
604 * config/aarch64/iterator.md (VDQHSD): New mode iterator.
605 (Vrevsuff): New mode attribute.
607 2014-04-24 Terry Guo <terry.guo@arm.com>
609 * config/arm/arm.h (machine_function): Define variable
610 after_arm_reorg here.
611 * config/arm/arm.c (after_arm_reorg): Remove the definition.
612 (arm_split_constant): Update the way to access variable
615 (arm_output_function_epilogue): Remove the reset of after_arm_reorg.
617 2014-04-23 Tom de Vries <tom@codesourcery.com>
619 * target-hooks-macros.h: Fix DEFHOOKPOD argument order in comment.
621 2014-04-23 David Malcolm <dmalcolm@redhat.com>
623 * is-a.h: Update comments to reflect the following changes to the
624 "pointerness" of the API, making the template parameter match the
625 return type, allowing use of is-a.h with typedefs of pointers.
626 (is_a_helper::cast): Return a T rather then a pointer to a T, so
627 that the return type matches the parameter to the is_a_helper.
629 (dyn_cast): Likewise.
631 * cgraph.c (cgraph_node_for_asm): Update for removal of implicit
632 pointer from the is-a.h API.
634 * cgraph.h (is_a_helper <cgraph_node>::test): Convert to...
635 (is_a_helper <cgraph_node *>::test): ...this, matching change to
637 (is_a_helper <varpool_node>::test): Likewise, convert to...
638 (is_a_helper <varpool_node *>::test): ...this.
640 (varpool_first_variable): Update for removal of implicit pointer
642 (varpool_next_variable): Likewise.
643 (varpool_first_static_initializer): Likewise.
644 (varpool_next_static_initializer): Likewise.
645 (varpool_first_defined_variable): Likewise.
646 (varpool_next_defined_variable): Likewise.
647 (cgraph_first_defined_function): Likewise.
648 (cgraph_next_defined_function): Likewise.
649 (cgraph_first_function): Likewise.
650 (cgraph_next_function): Likewise.
651 (cgraph_first_function_with_gimple_body): Likewise.
652 (cgraph_next_function_with_gimple_body): Likewise.
653 (cgraph_alias_target): Likewise.
654 (varpool_alias_target): Likewise.
655 (cgraph_function_or_thunk_node): Likewise.
656 (varpool_variable_node): Likewise.
657 (symtab_real_symbol_p): Likewise.
658 * cgraphunit.c (referred_to_p): Likewise.
659 (analyze_functions): Likewise.
660 (handle_alias_pairs): Likewise.
661 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
662 * gimple-ssa.h (gimple_vuse_op): Likewise.
663 (gimple_vdef_op): Likewise.
664 * gimple-streamer-in.c (input_gimple_stmt): Likewise.
665 * gimple.c (gimple_build_asm_1): Likewise.
666 (gimple_build_try): Likewise.
667 (gimple_build_resx): Likewise.
668 (gimple_build_eh_dispatch): Likewise.
669 (gimple_build_omp_for): Likewise.
670 (gimple_omp_for_set_clauses): Likewise.
672 * gimple.h (is_a_helper <gimple_statement_asm>::test): Convert to...
673 (is_a_helper <gimple_statement_asm *>::test): ...this.
674 (is_a_helper <gimple_statement_bind>::test): Convert to...
675 (is_a_helper <gimple_statement_bind *>::test): ...this.
676 (is_a_helper <gimple_statement_call>::test): Convert to...
677 (is_a_helper <gimple_statement_call *>::test): ...this.
678 (is_a_helper <gimple_statement_catch>::test): Convert to...
679 (is_a_helper <gimple_statement_catch *>::test): ...this.
680 (is_a_helper <gimple_statement_resx>::test): Convert to...
681 (is_a_helper <gimple_statement_resx *>::test): ...this.
682 (is_a_helper <gimple_statement_eh_dispatch>::test): Convert to...
683 (is_a_helper <gimple_statement_eh_dispatch *>::test): ...this.
684 (is_a_helper <gimple_statement_eh_else>::test): Convert to...
685 (is_a_helper <gimple_statement_eh_else *>::test): ...this.
686 (is_a_helper <gimple_statement_eh_filter>::test): Convert to...
687 (is_a_helper <gimple_statement_eh_filter *>::test): ...this.
688 (is_a_helper <gimple_statement_eh_mnt>::test): Convert to...
689 (is_a_helper <gimple_statement_eh_mnt *>::test): ...this.
690 (is_a_helper <gimple_statement_omp_atomic_load>::test): Convert to...
691 (is_a_helper <gimple_statement_omp_atomic_load *>::test): ...this.
692 (is_a_helper <gimple_statement_omp_atomic_store>::test): Convert to...
693 (is_a_helper <gimple_statement_omp_atomic_store *>::test): ...this.
694 (is_a_helper <gimple_statement_omp_return>::test): Convert to...
695 (is_a_helper <gimple_statement_omp_return *>::test): ...this.
696 (is_a_helper <gimple_statement_omp_continue>::test): Convert to...
697 (is_a_helper <gimple_statement_omp_continue *>::test): ...this.
698 (is_a_helper <gimple_statement_omp_critical>::test): Convert to...
699 (is_a_helper <gimple_statement_omp_critical *>::test): ...this.
700 (is_a_helper <gimple_statement_omp_for>::test): Convert to...
701 (is_a_helper <gimple_statement_omp_for *>::test): ...this.
702 (is_a_helper <gimple_statement_omp_taskreg>::test): Convert to...
703 (is_a_helper <gimple_statement_omp_taskreg *>::test): ...this.
704 (is_a_helper <gimple_statement_omp_parallel>::test): Convert to...
705 (is_a_helper <gimple_statement_omp_parallel *>::test): ...this.
706 (is_a_helper <gimple_statement_omp_target>::test): Convert to...
707 (is_a_helper <gimple_statement_omp_target *>::test): ...this.
708 (is_a_helper <gimple_statement_omp_sections>::test): Convert to...
709 (is_a_helper <gimple_statement_omp_sections *>::test): ...this.
710 (is_a_helper <gimple_statement_omp_single>::test): Convert to...
711 (is_a_helper <gimple_statement_omp_single *>::test): ...this.
712 (is_a_helper <gimple_statement_omp_teams>::test): Convert to...
713 (is_a_helper <gimple_statement_omp_teams *>::test): ...this.
714 (is_a_helper <gimple_statement_omp_task>::test): Convert to...
715 (is_a_helper <gimple_statement_omp_task *>::test): ...this.
716 (is_a_helper <gimple_statement_phi>::test): Convert to...
717 (is_a_helper <gimple_statement_phi *>::test): ...this.
718 (is_a_helper <gimple_statement_transaction>::test): Convert to...
719 (is_a_helper <gimple_statement_transaction *>::test): ...this.
720 (is_a_helper <gimple_statement_try>::test): Convert to...
721 (is_a_helper <gimple_statement_try *>::test): ...this.
722 (is_a_helper <gimple_statement_wce>::test): Convert to...
723 (is_a_helper <gimple_statement_wce *>::test): ...this.
724 (is_a_helper <const gimple_statement_asm>::test): Convert to...
725 (is_a_helper <const gimple_statement_asm *>::test): ...this.
726 (is_a_helper <const gimple_statement_bind>::test): Convert to...
727 (is_a_helper <const gimple_statement_bind *>::test): ...this.
728 (is_a_helper <const gimple_statement_call>::test): Convert to...
729 (is_a_helper <const gimple_statement_call *>::test): ...this.
730 (is_a_helper <const gimple_statement_catch>::test): Convert to...
731 (is_a_helper <const gimple_statement_catch *>::test): ...this.
732 (is_a_helper <const gimple_statement_resx>::test): Convert to...
733 (is_a_helper <const gimple_statement_resx *>::test): ...this.
734 (is_a_helper <const gimple_statement_eh_dispatch>::test):
736 (is_a_helper <const gimple_statement_eh_dispatch *>::test):
738 (is_a_helper <const gimple_statement_eh_filter>::test): Convert
740 (is_a_helper <const gimple_statement_eh_filter *>::test): ...this.
741 (is_a_helper <const gimple_statement_omp_atomic_load>::test):
743 (is_a_helper <const gimple_statement_omp_atomic_load *>::test):
745 (is_a_helper <const gimple_statement_omp_atomic_store>::test):
747 (is_a_helper <const gimple_statement_omp_atomic_store *>::test):
749 (is_a_helper <const gimple_statement_omp_return>::test): Convert
751 (is_a_helper <const gimple_statement_omp_return *>::test): ...this.
752 (is_a_helper <const gimple_statement_omp_continue>::test): Convert
754 (is_a_helper <const gimple_statement_omp_continue *>::test): ...this.
755 (is_a_helper <const gimple_statement_omp_critical>::test): Convert
757 (is_a_helper <const gimple_statement_omp_critical *>::test): ...this.
758 (is_a_helper <const gimple_statement_omp_for>::test): Convert to...
759 (is_a_helper <const gimple_statement_omp_for *>::test): ...this.
760 (is_a_helper <const gimple_statement_omp_taskreg>::test): Convert
762 (is_a_helper <const gimple_statement_omp_taskreg *>::test): ...this.
763 (is_a_helper <const gimple_statement_omp_parallel>::test): Convert
765 (is_a_helper <const gimple_statement_omp_parallel *>::test): ...this.
766 (is_a_helper <const gimple_statement_omp_target>::test): Convert
768 (is_a_helper <const gimple_statement_omp_target *>::test): ...this.
769 (is_a_helper <const gimple_statement_omp_sections>::test): Convert
771 (is_a_helper <const gimple_statement_omp_sections *>::test): ...this.
772 (is_a_helper <const gimple_statement_omp_single>::test): Convert
774 (is_a_helper <const gimple_statement_omp_single *>::test): ...this.
775 (is_a_helper <const gimple_statement_omp_teams>::test): Convert
777 (is_a_helper <const gimple_statement_omp_teams *>::test): ...this.
778 (is_a_helper <const gimple_statement_omp_task>::test): Convert
780 (is_a_helper <const gimple_statement_omp_task *>::test): ...this.
781 (is_a_helper <const gimple_statement_phi>::test): Convert to...
782 (is_a_helper <const gimple_statement_phi *>::test): ...this.
783 (is_a_helper <const gimple_statement_transaction>::test): Convert
785 (is_a_helper <const gimple_statement_transaction *>::test): ...this.
786 (is_a_helper <const gimple_statement_with_ops>::test): Convert
788 (is_a_helper <const gimple_statement_with_ops *>::test): ...this.
789 (is_a_helper <gimple_statement_with_ops>::test): Convert to...
790 (is_a_helper <gimple_statement_with_ops *>::test): ...this.
791 (is_a_helper <const gimple_statement_with_memory_ops>::test):
793 (is_a_helper <const gimple_statement_with_memory_ops *>::test):
795 (is_a_helper <gimple_statement_with_memory_ops>::test): Convert
797 (is_a_helper <gimple_statement_with_memory_ops *>::test): ...this.
799 (gimple_use_ops): Update for removal of implicit pointer from the
801 (gimple_set_use_ops): Likewise.
802 (gimple_vuse): Likewise.
803 (gimple_vdef): Likewise.
804 (gimple_vuse_ptr): Likewise.
805 (gimple_vdef_ptr): Likewise.
806 (gimple_set_vuse): Likewise.
807 (gimple_set_vdef): Likewise.
808 (gimple_omp_return_set_lhs): Likewise.
809 (gimple_omp_return_lhs): Likewise.
810 (gimple_omp_return_lhs_ptr): Likewise.
811 (gimple_call_fntype): Likewise.
812 (gimple_call_set_fntype): Likewise.
813 (gimple_call_set_internal_fn): Likewise.
814 (gimple_call_use_set): Likewise.
815 (gimple_call_clobber_set): Likewise.
816 (gimple_bind_vars): Likewise.
817 (gimple_bind_set_vars): Likewise.
818 (gimple_bind_body_ptr): Likewise.
819 (gimple_bind_set_body): Likewise.
820 (gimple_bind_add_stmt): Likewise.
821 (gimple_bind_block): Likewise.
822 (gimple_bind_set_block): Likewise.
823 (gimple_asm_ninputs): Likewise.
824 (gimple_asm_noutputs): Likewise.
825 (gimple_asm_nclobbers): Likewise.
826 (gimple_asm_nlabels): Likewise.
827 (gimple_asm_input_op): Likewise.
828 (gimple_asm_input_op_ptr): Likewise.
829 (gimple_asm_output_op): Likewise.
830 (gimple_asm_output_op_ptr): Likewise.
831 (gimple_asm_set_output_op): Likewise.
832 (gimple_asm_clobber_op): Likewise.
833 (gimple_asm_set_clobber_op): Likewise.
834 (gimple_asm_label_op): Likewise.
835 (gimple_asm_set_label_op): Likewise.
836 (gimple_asm_string): Likewise.
837 (gimple_catch_types): Likewise.
838 (gimple_catch_types_ptr): Likewise.
839 (gimple_catch_handler_ptr): Likewise.
840 (gimple_catch_set_types): Likewise.
841 (gimple_catch_set_handler): Likewise.
842 (gimple_eh_filter_types): Likewise.
843 (gimple_eh_filter_types_ptr): Likewise.
844 (gimple_eh_filter_failure_ptr): Likewise.
845 (gimple_eh_filter_set_types): Likewise.
846 (gimple_eh_filter_set_failure): Likewise.
847 (gimple_eh_must_not_throw_fndecl): Likewise.
848 (gimple_eh_must_not_throw_set_fndecl): Likewise.
849 (gimple_eh_else_n_body_ptr): Likewise.
850 (gimple_eh_else_e_body_ptr): Likewise.
851 (gimple_eh_else_set_n_body): Likewise.
852 (gimple_eh_else_set_e_body): Likewise.
853 (gimple_try_eval_ptr): Likewise.
854 (gimple_try_cleanup_ptr): Likewise.
855 (gimple_try_set_eval): Likewise.
856 (gimple_try_set_cleanup): Likewise.
857 (gimple_wce_cleanup_ptr): Likewise.
858 (gimple_wce_set_cleanup): Likewise.
859 (gimple_phi_capacity): Likewise.
860 (gimple_phi_num_args): Likewise.
861 (gimple_phi_result): Likewise.
862 (gimple_phi_result_ptr): Likewise.
863 (gimple_phi_set_result): Likewise.
864 (gimple_phi_arg): Likewise.
865 (gimple_phi_set_arg): Likewise.
866 (gimple_resx_region): Likewise.
867 (gimple_resx_set_region): Likewise.
868 (gimple_eh_dispatch_region): Likewise.
869 (gimple_eh_dispatch_set_region): Likewise.
870 (gimple_omp_critical_name): Likewise.
871 (gimple_omp_critical_name_ptr): Likewise.
872 (gimple_omp_critical_set_name): Likewise.
873 (gimple_omp_for_clauses): Likewise.
874 (gimple_omp_for_clauses_ptr): Likewise.
875 (gimple_omp_for_set_clauses): Likewise.
876 (gimple_omp_for_collapse): Likewise.
877 (gimple_omp_for_index): Likewise.
878 (gimple_omp_for_index_ptr): Likewise.
879 (gimple_omp_for_set_index): Likewise.
880 (gimple_omp_for_initial): Likewise.
881 (gimple_omp_for_initial_ptr): Likewise.
882 (gimple_omp_for_set_initial): Likewise.
883 (gimple_omp_for_final): Likewise.
884 (gimple_omp_for_final_ptr): Likewise.
885 (gimple_omp_for_set_final): Likewise.
886 (gimple_omp_for_incr): Likewise.
887 (gimple_omp_for_incr_ptr): Likewise.
888 (gimple_omp_for_set_incr): Likewise.
889 (gimple_omp_for_pre_body_ptr): Likewise.
890 (gimple_omp_for_set_pre_body): Likewise.
891 (gimple_omp_parallel_clauses): Likewise.
892 (gimple_omp_parallel_clauses_ptr): Likewise.
893 (gimple_omp_parallel_set_clauses): Likewise.
894 (gimple_omp_parallel_child_fn): Likewise.
895 (gimple_omp_parallel_child_fn_ptr): Likewise.
896 (gimple_omp_parallel_set_child_fn): Likewise.
897 (gimple_omp_parallel_data_arg): Likewise.
898 (gimple_omp_parallel_data_arg_ptr): Likewise.
899 (gimple_omp_parallel_set_data_arg): Likewise.
900 (gimple_omp_task_clauses): Likewise.
901 (gimple_omp_task_clauses_ptr): Likewise.
902 (gimple_omp_task_set_clauses): Likewise.
903 (gimple_omp_task_child_fn): Likewise.
904 (gimple_omp_task_child_fn_ptr): Likewise.
905 (gimple_omp_task_set_child_fn): Likewise.
906 (gimple_omp_task_data_arg): Likewise.
907 (gimple_omp_task_data_arg_ptr): Likewise.
908 (gimple_omp_task_set_data_arg): Likewise.
909 (gimple_omp_taskreg_clauses): Likewise.
910 (gimple_omp_taskreg_clauses_ptr): Likewise.
911 (gimple_omp_taskreg_set_clauses): Likewise.
912 (gimple_omp_taskreg_child_fn): Likewise.
913 (gimple_omp_taskreg_child_fn_ptr): Likewise.
914 (gimple_omp_taskreg_set_child_fn): Likewise.
915 (gimple_omp_taskreg_data_arg): Likewise.
916 (gimple_omp_taskreg_data_arg_ptr): Likewise.
917 (gimple_omp_taskreg_set_data_arg): Likewise.
918 (gimple_omp_task_copy_fn): Likewise.
919 (gimple_omp_task_copy_fn_ptr): Likewise.
920 (gimple_omp_task_set_copy_fn): Likewise.
921 (gimple_omp_task_arg_size): Likewise.
922 (gimple_omp_task_arg_size_ptr): Likewise.
923 (gimple_omp_task_set_arg_size): Likewise.
924 (gimple_omp_task_arg_align): Likewise.
925 (gimple_omp_task_arg_align_ptr): Likewise.
926 (gimple_omp_task_set_arg_align): Likewise.
927 (gimple_omp_single_clauses): Likewise.
928 (gimple_omp_single_clauses_ptr): Likewise.
929 (gimple_omp_single_set_clauses): Likewise.
930 (gimple_omp_target_clauses): Likewise.
931 (gimple_omp_target_clauses_ptr): Likewise.
932 (gimple_omp_target_set_clauses): Likewise.
933 (gimple_omp_target_child_fn): Likewise.
934 (gimple_omp_target_child_fn_ptr): Likewise.
935 (gimple_omp_target_set_child_fn): Likewise.
936 (gimple_omp_target_data_arg): Likewise.
937 (gimple_omp_target_data_arg_ptr): Likewise.
938 (gimple_omp_target_set_data_arg): Likewise.
939 (gimple_omp_teams_clauses): Likewise.
940 (gimple_omp_teams_clauses_ptr): Likewise.
941 (gimple_omp_teams_set_clauses): Likewise.
942 (gimple_omp_sections_clauses): Likewise.
943 (gimple_omp_sections_clauses_ptr): Likewise.
944 (gimple_omp_sections_set_clauses): Likewise.
945 (gimple_omp_sections_control): Likewise.
946 (gimple_omp_sections_control_ptr): Likewise.
947 (gimple_omp_sections_set_control): Likewise.
948 (gimple_omp_for_set_cond): Likewise.
949 (gimple_omp_for_cond): Likewise.
950 (gimple_omp_atomic_store_set_val): Likewise.
951 (gimple_omp_atomic_store_val): Likewise.
952 (gimple_omp_atomic_store_val_ptr): Likewise.
953 (gimple_omp_atomic_load_set_lhs): Likewise.
954 (gimple_omp_atomic_load_lhs): Likewise.
955 (gimple_omp_atomic_load_lhs_ptr): Likewise.
956 (gimple_omp_atomic_load_set_rhs): Likewise.
957 (gimple_omp_atomic_load_rhs): Likewise.
958 (gimple_omp_atomic_load_rhs_ptr): Likewise.
959 (gimple_omp_continue_control_def): Likewise.
960 (gimple_omp_continue_control_def_ptr): Likewise.
961 (gimple_omp_continue_set_control_def): Likewise.
962 (gimple_omp_continue_control_use): Likewise.
963 (gimple_omp_continue_control_use_ptr): Likewise.
964 (gimple_omp_continue_set_control_use): Likewise.
965 (gimple_transaction_body_ptr): Likewise.
966 (gimple_transaction_label): Likewise.
967 (gimple_transaction_label_ptr): Likewise.
968 (gimple_transaction_set_body): Likewise.
969 (gimple_transaction_set_label): Likewise.
971 * ipa-devirt.c (build_type_inheritance_graph): Likewise.
972 * ipa-inline-analysis.c (inline_write_summary): Likewise.
973 * ipa-ref.c (ipa_record_reference): Likewise.
974 * ipa-reference.c (analyze_function): Likewise.
975 (ipa_reference_write_optimization_summary): Likewise.
976 * ipa.c (symtab_remove_unreachable_nodes): Likewise.
977 (address_taken_from_non_vtable_p): Likewise.
978 (comdat_can_be_unshared_p_1): Likewise.
979 * lto-cgraph.c (lto_output_ref): Likewise.
980 (add_references): Likewise.
981 (compute_ltrans_boundary): Likewise.
982 (output_symtab): Likewise.
983 (input_ref): Likewise.
984 (input_cgraph_1): Likewise.
985 (output_cgraph_opt_summary): Likewise.
986 * lto-streamer-out.c (lto_output): Likewise.
987 (output_symbol_p): Likewise.
988 * lto-streamer.h (lsei_next_function_in_partition): Likewise.
989 (lsei_start_function_in_partition): Likewise.
990 (lsei_next_variable_in_partition): Likewise.
991 (lsei_start_variable_in_partition): Likewise.
992 * symtab.c (insert_to_assembler_name_hash): Likewise.
993 (unlink_from_assembler_name_hash): Likewise.
994 (symtab_unregister_node): Likewise.
995 (symtab_remove_node): Likewise.
996 (dump_symtab_node): Likewise.
997 (verify_symtab_base): Likewise.
998 (verify_symtab_node): Likewise.
999 (symtab_make_decl_local): Likewise.
1000 (symtab_alias_ultimate_target): Likewise.
1001 (symtab_resolve_alias): Likewise.
1002 (symtab_get_symbol_partitioning_class): Likewise.
1003 * tree-phinodes.c (allocate_phi_node): Likewise.
1004 (reserve_phi_args_for_new_edge): Likewise.
1005 (remove_phi_args): Likewise.
1006 * varpool.c (varpool_node_for_asm): Likewise.
1007 (varpool_remove_unreferenced_decls): Likewise.
1009 2014-04-23 Jeff Law <law@redhat.com>
1011 PR tree-optimization/60902
1012 * tree-ssa-threadedge.c
1013 (record_temporary_equivalences_from_stmts_at_dest): Make sure to
1014 invalidate outputs from statements that do not produce useful
1015 outputs for threading.
1017 2014-04-23 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
1019 * config/aarch64/aarch64.md (stack_protect_set, stack_protect_test)
1020 (stack_protect_set_<mode>, stack_protect_test_<mode>): Add
1021 machine descriptions for Stack Smashing Protector.
1023 2014-04-23 Richard Earnshaw <rearnsha@arm.com>
1025 * aarch64.md (<optab>_rol<mode>3): New pattern.
1026 (<optab>_rolsi3_uxtw): Likewise.
1027 * aarch64.c (aarch64_strip_shift): Handle ROTATE and ROTATERT.
1029 2014-04-23 James Greenhalgh <james.greenhalgh@arm.com>
1031 * config/arm/arm.c (arm_cortex_a57_tune): Initialize all fields.
1032 (arm_cortex_a12_tune): Likewise.
1034 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1036 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle BSWAP.
1038 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1040 * config/arm/arm.md (arm_rev16si2): New pattern.
1041 (arm_rev16si2_alt): Likewise.
1042 * config/arm/arm.c (arm_new_rtx_costs): Handle rev16 case.
1044 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1046 * config/aarch64/aarch64.md (rev16<mode>2): New pattern.
1047 (rev16<mode>2_alt): Likewise.
1048 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle rev16 case.
1049 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p): New.
1050 (aarch_rev16_shleft_mask_imm_p): Likewise.
1051 (aarch_rev16_p_1): Likewise.
1052 (aarch_rev16_p): Likewise.
1053 * config/arm/aarch-common-protos.h (aarch_rev16_p): Declare extern.
1054 (aarch_rev16_shright_mask_imm_p): Likewise.
1055 (aarch_rev16_shleft_mask_imm_p): Likewise.
1057 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1059 * config/arm/aarch-common-protos.h (alu_cost_table): Add rev field.
1060 * config/arm/aarch-cost-tables.h (generic_extra_costs): Specify
1062 (cortex_a53_extra_costs): Likewise.
1063 (cortex_a57_extra_costs): Likewise.
1064 * config/arm/arm.c (cortexa9_extra_costs): Likewise.
1065 (cortexa7_extra_costs): Likewise.
1066 (cortexa8_extra_costs): Likewise.
1067 (cortexa12_extra_costs): Likewise.
1068 (cortexa15_extra_costs): Likewise.
1069 (v7m_extra_costs): Likewise.
1070 (arm_new_rtx_costs): Handle BSWAP.
1072 2013-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1074 * config/arm/arm.c (cortexa8_extra_costs): New table.
1075 (arm_cortex_a8_tune): New tuning struct.
1076 * config/arm/arm-cores.def (cortex-a8): Use cortex_a8 tuning struct.
1078 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1080 * config/arm/arm.c (arm_new_rtx_costs): Handle FMA.
1082 2014-04-23 Richard Biener <rguenther@suse.de>
1084 * Makefile.in (OBJS): Remove loop-unswitch.o.
1085 * loop-unswitch.c: Delete.
1086 * tree-pass.h (make_pass_rtl_unswitch): Remove.
1087 * passes.def (pass_rtl_unswitch): Likewise.
1088 * loop-init.c (gate_rtl_unswitch): Likewise.
1089 (rtl_unswitch): Likewise.
1090 (pass_data_rtl_unswitch): Likewise.
1091 (pass_rtl_unswitch): Likewise.
1092 (make_pass_rtl_unswitch): Likewise.
1093 * rtl.h (reversed_condition): Likewise.
1094 (compare_and_jump_seq): Likewise.
1095 * loop-iv.c (reversed_condition): Move here from loop-unswitch.c
1097 * loop-unroll.c (compare_and_jump_seq): Likewise.
1099 2014-04-23 Richard Biener <rguenther@suse.de>
1101 PR tree-optimization/60903
1102 * tree-ssa-loop-im.c (analyze_memory_references): Remove
1103 commented code block.
1104 (execute_sm_if_changed): Properly apply IRREDUCIBLE_LOOP
1105 loop flags to newly created BBs and edges.
1107 2014-04-23 Nick Clifton <nickc@redhat.com>
1109 * config/msp430/msp430.c (msp430_handle_option): Move function
1111 (msp430_option_override): Simplify mcu and mcpu option handling.
1112 (msp430_is_f5_mcu): Rename to msp430_use_f5_series_hwmult. Add
1113 support for -mhwmult command line option.
1114 (has_32bit_hwmult): Rename to use_32bit_hwmult. Add support for
1115 -mhwmult command line option.
1116 (msp430_hwmult_enabled): Delete.
1117 (msp43o_output_labelref): Add support for -mhwmult command line
1119 * config/msp430/msp430.md (mulhisi3, umulhisi3, mulsidi3)
1120 (umulsidi3): Likewise.
1121 * config/msp430/msp430.opt (mmcu): Add Report attribute.
1122 (mcpu, mlarge, msmall): Likewise.
1123 (mhwmult): New option.
1124 * config/msp430/msp430-protos.h (msp430_hwmult_enabled): Remove
1126 (msp430_is_f5_mcu): Remove prototype.
1127 (msp430_use_f5_series_hwmult): Add prototype.
1128 * config/msp430/msp430-opts.h: New file.
1129 * common/config/msp430: New directory.
1130 * common/config/msp430/msp430-common.c: New file.
1131 * config.gcc (msp430): Remove target_has_targetm_common.
1132 * doc/invoke.texi: Document -mhwmult command line option.
1134 2014-04-23 Nick Clifton <nickc@redhat.com>
1136 * config/i386/cygwin.h (ENDFILE_SPEC): Include
1137 default-manifest.o if it can be found in the search path.
1138 * config/i386/mingw32.h (ENDFILE_SPEC): Likewise.
1140 2014-04-23 Terry Guo <terry.guo@arm.com>
1142 * config/arm/arm.h (ASM_APP_OFF): Re-define it in a cleaner way.
1144 2014-04-23 Richard Biener <rguenther@suse.de>
1147 * tree-inline.c (declare_return_variable): Use mark_addressable.
1149 2014-04-23 Richard Biener <rguenther@suse.de>
1152 * loop-init.c (loop_optimizer_init): Make sure to apply
1153 LOOPS_MAY_HAVE_MULTIPLE_LATCHES before fixing up loops.
1155 2014-04-22 Jakub Jelinek <jakub@redhat.com>
1158 * common.opt (fsanitize-recover, fsanitize-undefined-trap-on-error):
1160 * gcc.c (sanitize_spec_function): Don't return "" for "undefined"
1161 if flag_sanitize_undefined_trap_on_error.
1162 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW_ABORT,
1163 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS_ABORT,
1164 BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE_ABORT,
1165 BUILT_IN_UBSAN_HANDLE_TYPE_MISMATCH_ABORT,
1166 BUILT_IN_UBSAN_HANDLE_ADD_OVERFLOW_ABORT,
1167 BUILT_IN_UBSAN_HANDLE_SUB_OVERFLOW_ABORT,
1168 BUILT_IN_UBSAN_HANDLE_MUL_OVERFLOW_ABORT,
1169 BUILT_IN_UBSAN_HANDLE_NEGATE_OVERFLOW_ABORT,
1170 BUILT_IN_UBSAN_HANDLE_LOAD_INVALID_VALUE_ABORT): New builtins.
1171 * ubsan.c (ubsan_instrument_unreachable): Return
1172 __builtin_trap () if flag_sanitize_undefined_trap_on_error.
1173 (ubsan_expand_null_ifn): Emit __builtin_trap ()
1174 if flag_sanitize_undefined_trap_on_error and
1175 __ubsan_handle_type_mismatch_abort if !flag_sanitize_recover.
1176 (ubsan_expand_null_ifn, ubsan_build_overflow_builtin,
1177 instrument_bool_enum_load): Emit __builtin_trap () if
1178 flag_sanitize_undefined_trap_on_error and
1179 __builtin_handle_*_abort () if !flag_sanitize_recover.
1180 * doc/invoke.texi (-fsanitize-recover,
1181 -fsanitize-undefined-trap-on-error): Document.
1183 2014-04-22 Christian Bruel <christian.bruel@st.com>
1185 * config/sh/sh.md (mov<mode>): Replace movQIHI.
1186 Force immediates to SImode.
1188 2014-04-22 Sandra Loosemore <sandra@codesourcery.com>
1190 * config/nios2/nios2.md (UNSPEC_ROUND): New.
1192 * config/nios2/nios2.opt (mno-custom-round, mcustom-round=): New.
1193 * config/nios2/nios2-opts.h (N2FPU_ALL_CODES): Add round.
1194 * config/nios2/nios2.c (N2F_NO_ERRNO): Define.
1195 (nios2_fpu_insn): Add entry for round.
1196 (N2FPU_NO_ERRNO_P): Define.
1197 (nios2_custom_check_insns): Add check for N2F_NO_ERRNO and
1199 * doc/invoke.texi (Nios II Options): Document -mcustom-round.
1201 2014-04-22 Richard Henderson <rth@redhat.com>
1203 * config/aarch64/aarch64 (addti3, subti3): New expanders.
1204 (add<GPI>3_compare0): Remove leading * from name.
1205 (add<GPI>3_carryin): Likewise.
1206 (sub<GPI>3_compare0): Likewise.
1207 (sub<GPI>3_carryin): Likewise.
1208 (<su_optab>mulditi3): New expander.
1209 (multi3): New expander.
1210 (madd<GPI>): Remove leading * from name.
1212 2014-04-22 Martin Jambor <mjambor@suse.cz>
1214 * cgraphclones.c (cgraph_function_versioning): Copy
1215 ipa_transforms_to_apply instead of asserting it is empty.
1217 2014-04-22 H.J. Lu <hongjiu.lu@intel.com>
1220 * config/i386/i386.c (ix86_expand_set_or_movmem): Call counter_mode
1221 on count_exp to get mode.
1223 2014-04-22 Andrew Pinski <apinski@cavium.com>
1225 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
1226 Handle TLS for ILP32.
1227 * config/aarch64/aarch64.md (tlsie_small): Rename to ...
1228 (tlsie_small_<mode>): this and handle PTR.
1229 (tlsie_small_sidi): New pattern.
1230 (tlsle_small): Change to an expand to handle ILP32.
1231 (tlsle_small_<mode>): New pattern.
1232 (tlsdesc_small): Rename to ...
1233 (tlsdesc_small_<mode>): this and handle PTR.
1235 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1237 * config/aarch64/aarch64.c (TARGET_FLAGS_REGNUM): Define.
1239 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1241 * config/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
1242 (aarch64_types_signed_unsigned_qualifiers): Qualifier added.
1243 (aarch64_types_signed_poly_qualifiers): Likewise.
1244 (aarch64_types_unsigned_signed_qualifiers): Likewise.
1245 (aarch64_types_poly_signed_qualifiers): Likewise.
1246 (TYPES_REINTERP_SS): Type macro added.
1247 (TYPES_REINTERP_SU): Likewise.
1248 (TYPES_REINTERP_SP): Likewise.
1249 (TYPES_REINTERP_US): Likewise.
1250 (TYPES_REINTERP_PS): Likewise.
1251 (aarch64_fold_builtin): New expression folding added.
1252 * config/aarch64/aarch64-simd-builtins.def (REINTERP):
1253 Declarations removed.
1254 (REINTERP_SS): Declarations added.
1255 (REINTERP_US): Likewise.
1256 (REINTERP_PS): Likewise.
1257 (REINTERP_SU): Likewise.
1258 (REINTERP_SP): Likewise.
1259 * config/aarch64/arm_neon.h (vreinterpret_p8_f64): Implemented.
1260 (vreinterpretq_p8_f64): Likewise.
1261 (vreinterpret_p16_f64): Likewise.
1262 (vreinterpretq_p16_f64): Likewise.
1263 (vreinterpret_f32_f64): Likewise.
1264 (vreinterpretq_f32_f64): Likewise.
1265 (vreinterpret_f64_f32): Likewise.
1266 (vreinterpret_f64_p8): Likewise.
1267 (vreinterpret_f64_p16): Likewise.
1268 (vreinterpret_f64_s8): Likewise.
1269 (vreinterpret_f64_s16): Likewise.
1270 (vreinterpret_f64_s32): Likewise.
1271 (vreinterpret_f64_s64): Likewise.
1272 (vreinterpret_f64_u8): Likewise.
1273 (vreinterpret_f64_u16): Likewise.
1274 (vreinterpret_f64_u32): Likewise.
1275 (vreinterpret_f64_u64): Likewise.
1276 (vreinterpretq_f64_f32): Likewise.
1277 (vreinterpretq_f64_p8): Likewise.
1278 (vreinterpretq_f64_p16): Likewise.
1279 (vreinterpretq_f64_s8): Likewise.
1280 (vreinterpretq_f64_s16): Likewise.
1281 (vreinterpretq_f64_s32): Likewise.
1282 (vreinterpretq_f64_s64): Likewise.
1283 (vreinterpretq_f64_u8): Likewise.
1284 (vreinterpretq_f64_u16): Likewise.
1285 (vreinterpretq_f64_u32): Likewise.
1286 (vreinterpretq_f64_u64): Likewise.
1287 (vreinterpret_s64_f64): Likewise.
1288 (vreinterpretq_s64_f64): Likewise.
1289 (vreinterpret_u64_f64): Likewise.
1290 (vreinterpretq_u64_f64): Likewise.
1291 (vreinterpret_s8_f64): Likewise.
1292 (vreinterpretq_s8_f64): Likewise.
1293 (vreinterpret_s16_f64): Likewise.
1294 (vreinterpretq_s16_f64): Likewise.
1295 (vreinterpret_s32_f64): Likewise.
1296 (vreinterpretq_s32_f64): Likewise.
1297 (vreinterpret_u8_f64): Likewise.
1298 (vreinterpretq_u8_f64): Likewise.
1299 (vreinterpret_u16_f64): Likewise.
1300 (vreinterpretq_u16_f64): Likewise.
1301 (vreinterpret_u32_f64): Likewise.
1302 (vreinterpretq_u32_f64): Likewise.
1304 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1306 * config/aarch64/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
1307 * config/aarch64/aarch64/aarch64-simd-builtins.def (REINTERP): Removed.
1308 (vreinterpret_p8_s8): Likewise.
1309 * config/aarch64/aarch64/arm_neon.h (vreinterpret_p8_s8): Uses cast.
1310 (vreinterpret_p8_s16): Likewise.
1311 (vreinterpret_p8_s32): Likewise.
1312 (vreinterpret_p8_s64): Likewise.
1313 (vreinterpret_p8_f32): Likewise.
1314 (vreinterpret_p8_u8): Likewise.
1315 (vreinterpret_p8_u16): Likewise.
1316 (vreinterpret_p8_u32): Likewise.
1317 (vreinterpret_p8_u64): Likewise.
1318 (vreinterpret_p8_p16): Likewise.
1319 (vreinterpretq_p8_s8): Likewise.
1320 (vreinterpretq_p8_s16): Likewise.
1321 (vreinterpretq_p8_s32): Likewise.
1322 (vreinterpretq_p8_s64): Likewise.
1323 (vreinterpretq_p8_f32): Likewise.
1324 (vreinterpretq_p8_u8): Likewise.
1325 (vreinterpretq_p8_u16): Likewise.
1326 (vreinterpretq_p8_u32): Likewise.
1327 (vreinterpretq_p8_u64): Likewise.
1328 (vreinterpretq_p8_p16): Likewise.
1329 (vreinterpret_p16_s8): Likewise.
1330 (vreinterpret_p16_s16): Likewise.
1331 (vreinterpret_p16_s32): Likewise.
1332 (vreinterpret_p16_s64): Likewise.
1333 (vreinterpret_p16_f32): Likewise.
1334 (vreinterpret_p16_u8): Likewise.
1335 (vreinterpret_p16_u16): Likewise.
1336 (vreinterpret_p16_u32): Likewise.
1337 (vreinterpret_p16_u64): Likewise.
1338 (vreinterpret_p16_p8): Likewise.
1339 (vreinterpretq_p16_s8): Likewise.
1340 (vreinterpretq_p16_s16): Likewise.
1341 (vreinterpretq_p16_s32): Likewise.
1342 (vreinterpretq_p16_s64): Likewise.
1343 (vreinterpretq_p16_f32): Likewise.
1344 (vreinterpretq_p16_u8): Likewise.
1345 (vreinterpretq_p16_u16): Likewise.
1346 (vreinterpretq_p16_u32): Likewise.
1347 (vreinterpretq_p16_u64): Likewise.
1348 (vreinterpretq_p16_p8): Likewise.
1349 (vreinterpret_f32_s8): Likewise.
1350 (vreinterpret_f32_s16): Likewise.
1351 (vreinterpret_f32_s32): Likewise.
1352 (vreinterpret_f32_s64): Likewise.
1353 (vreinterpret_f32_u8): Likewise.
1354 (vreinterpret_f32_u16): Likewise.
1355 (vreinterpret_f32_u32): Likewise.
1356 (vreinterpret_f32_u64): Likewise.
1357 (vreinterpret_f32_p8): Likewise.
1358 (vreinterpret_f32_p16): Likewise.
1359 (vreinterpretq_f32_s8): Likewise.
1360 (vreinterpretq_f32_s16): Likewise.
1361 (vreinterpretq_f32_s32): Likewise.
1362 (vreinterpretq_f32_s64): Likewise.
1363 (vreinterpretq_f32_u8): Likewise.
1364 (vreinterpretq_f32_u16): Likewise.
1365 (vreinterpretq_f32_u32): Likewise.
1366 (vreinterpretq_f32_u64): Likewise.
1367 (vreinterpretq_f32_p8): Likewise.
1368 (vreinterpretq_f32_p16): Likewise.
1369 (vreinterpret_s64_s8): Likewise.
1370 (vreinterpret_s64_s16): Likewise.
1371 (vreinterpret_s64_s32): Likewise.
1372 (vreinterpret_s64_f32): Likewise.
1373 (vreinterpret_s64_u8): Likewise.
1374 (vreinterpret_s64_u16): Likewise.
1375 (vreinterpret_s64_u32): Likewise.
1376 (vreinterpret_s64_u64): Likewise.
1377 (vreinterpret_s64_p8): Likewise.
1378 (vreinterpret_s64_p16): Likewise.
1379 (vreinterpretq_s64_s8): Likewise.
1380 (vreinterpretq_s64_s16): Likewise.
1381 (vreinterpretq_s64_s32): Likewise.
1382 (vreinterpretq_s64_f32): Likewise.
1383 (vreinterpretq_s64_u8): Likewise.
1384 (vreinterpretq_s64_u16): Likewise.
1385 (vreinterpretq_s64_u32): Likewise.
1386 (vreinterpretq_s64_u64): Likewise.
1387 (vreinterpretq_s64_p8): Likewise.
1388 (vreinterpretq_s64_p16): Likewise.
1389 (vreinterpret_u64_s8): Likewise.
1390 (vreinterpret_u64_s16): Likewise.
1391 (vreinterpret_u64_s32): Likewise.
1392 (vreinterpret_u64_s64): Likewise.
1393 (vreinterpret_u64_f32): Likewise.
1394 (vreinterpret_u64_u8): Likewise.
1395 (vreinterpret_u64_u16): Likewise.
1396 (vreinterpret_u64_u32): Likewise.
1397 (vreinterpret_u64_p8): Likewise.
1398 (vreinterpret_u64_p16): Likewise.
1399 (vreinterpretq_u64_s8): Likewise.
1400 (vreinterpretq_u64_s16): Likewise.
1401 (vreinterpretq_u64_s32): Likewise.
1402 (vreinterpretq_u64_s64): Likewise.
1403 (vreinterpretq_u64_f32): Likewise.
1404 (vreinterpretq_u64_u8): Likewise.
1405 (vreinterpretq_u64_u16): Likewise.
1406 (vreinterpretq_u64_u32): Likewise.
1407 (vreinterpretq_u64_p8): Likewise.
1408 (vreinterpretq_u64_p16): Likewise.
1409 (vreinterpret_s8_s16): Likewise.
1410 (vreinterpret_s8_s32): Likewise.
1411 (vreinterpret_s8_s64): Likewise.
1412 (vreinterpret_s8_f32): Likewise.
1413 (vreinterpret_s8_u8): Likewise.
1414 (vreinterpret_s8_u16): Likewise.
1415 (vreinterpret_s8_u32): Likewise.
1416 (vreinterpret_s8_u64): Likewise.
1417 (vreinterpret_s8_p8): Likewise.
1418 (vreinterpret_s8_p16): Likewise.
1419 (vreinterpretq_s8_s16): Likewise.
1420 (vreinterpretq_s8_s32): Likewise.
1421 (vreinterpretq_s8_s64): Likewise.
1422 (vreinterpretq_s8_f32): Likewise.
1423 (vreinterpretq_s8_u8): Likewise.
1424 (vreinterpretq_s8_u16): Likewise.
1425 (vreinterpretq_s8_u32): Likewise.
1426 (vreinterpretq_s8_u64): Likewise.
1427 (vreinterpretq_s8_p8): Likewise.
1428 (vreinterpretq_s8_p16): Likewise.
1429 (vreinterpret_s16_s8): Likewise.
1430 (vreinterpret_s16_s32): Likewise.
1431 (vreinterpret_s16_s64): Likewise.
1432 (vreinterpret_s16_f32): Likewise.
1433 (vreinterpret_s16_u8): Likewise.
1434 (vreinterpret_s16_u16): Likewise.
1435 (vreinterpret_s16_u32): Likewise.
1436 (vreinterpret_s16_u64): Likewise.
1437 (vreinterpret_s16_p8): Likewise.
1438 (vreinterpret_s16_p16): Likewise.
1439 (vreinterpretq_s16_s8): Likewise.
1440 (vreinterpretq_s16_s32): Likewise.
1441 (vreinterpretq_s16_s64): Likewise.
1442 (vreinterpretq_s16_f32): Likewise.
1443 (vreinterpretq_s16_u8): Likewise.
1444 (vreinterpretq_s16_u16): Likewise.
1445 (vreinterpretq_s16_u32): Likewise.
1446 (vreinterpretq_s16_u64): Likewise.
1447 (vreinterpretq_s16_p8): Likewise.
1448 (vreinterpretq_s16_p16): Likewise.
1449 (vreinterpret_s32_s8): Likewise.
1450 (vreinterpret_s32_s16): Likewise.
1451 (vreinterpret_s32_s64): Likewise.
1452 (vreinterpret_s32_f32): Likewise.
1453 (vreinterpret_s32_u8): Likewise.
1454 (vreinterpret_s32_u16): Likewise.
1455 (vreinterpret_s32_u32): Likewise.
1456 (vreinterpret_s32_u64): Likewise.
1457 (vreinterpret_s32_p8): Likewise.
1458 (vreinterpret_s32_p16): Likewise.
1459 (vreinterpretq_s32_s8): Likewise.
1460 (vreinterpretq_s32_s16): Likewise.
1461 (vreinterpretq_s32_s64): Likewise.
1462 (vreinterpretq_s32_f32): Likewise.
1463 (vreinterpretq_s32_u8): Likewise.
1464 (vreinterpretq_s32_u16): Likewise.
1465 (vreinterpretq_s32_u32): Likewise.
1466 (vreinterpretq_s32_u64): Likewise.
1467 (vreinterpretq_s32_p8): Likewise.
1468 (vreinterpretq_s32_p16): Likewise.
1469 (vreinterpret_u8_s8): Likewise.
1470 (vreinterpret_u8_s16): Likewise.
1471 (vreinterpret_u8_s32): Likewise.
1472 (vreinterpret_u8_s64): Likewise.
1473 (vreinterpret_u8_f32): Likewise.
1474 (vreinterpret_u8_u16): Likewise.
1475 (vreinterpret_u8_u32): Likewise.
1476 (vreinterpret_u8_u64): Likewise.
1477 (vreinterpret_u8_p8): Likewise.
1478 (vreinterpret_u8_p16): Likewise.
1479 (vreinterpretq_u8_s8): Likewise.
1480 (vreinterpretq_u8_s16): Likewise.
1481 (vreinterpretq_u8_s32): Likewise.
1482 (vreinterpretq_u8_s64): Likewise.
1483 (vreinterpretq_u8_f32): Likewise.
1484 (vreinterpretq_u8_u16): Likewise.
1485 (vreinterpretq_u8_u32): Likewise.
1486 (vreinterpretq_u8_u64): Likewise.
1487 (vreinterpretq_u8_p8): Likewise.
1488 (vreinterpretq_u8_p16): Likewise.
1489 (vreinterpret_u16_s8): Likewise.
1490 (vreinterpret_u16_s16): Likewise.
1491 (vreinterpret_u16_s32): Likewise.
1492 (vreinterpret_u16_s64): Likewise.
1493 (vreinterpret_u16_f32): Likewise.
1494 (vreinterpret_u16_u8): Likewise.
1495 (vreinterpret_u16_u32): Likewise.
1496 (vreinterpret_u16_u64): Likewise.
1497 (vreinterpret_u16_p8): Likewise.
1498 (vreinterpret_u16_p16): Likewise.
1499 (vreinterpretq_u16_s8): Likewise.
1500 (vreinterpretq_u16_s16): Likewise.
1501 (vreinterpretq_u16_s32): Likewise.
1502 (vreinterpretq_u16_s64): Likewise.
1503 (vreinterpretq_u16_f32): Likewise.
1504 (vreinterpretq_u16_u8): Likewise.
1505 (vreinterpretq_u16_u32): Likewise.
1506 (vreinterpretq_u16_u64): Likewise.
1507 (vreinterpretq_u16_p8): Likewise.
1508 (vreinterpretq_u16_p16): Likewise.
1509 (vreinterpret_u32_s8): Likewise.
1510 (vreinterpret_u32_s16): Likewise.
1511 (vreinterpret_u32_s32): Likewise.
1512 (vreinterpret_u32_s64): Likewise.
1513 (vreinterpret_u32_f32): Likewise.
1514 (vreinterpret_u32_u8): Likewise.
1515 (vreinterpret_u32_u16): Likewise.
1516 (vreinterpret_u32_u64): Likewise.
1517 (vreinterpret_u32_p8): Likewise.
1518 (vreinterpret_u32_p16): Likewise.
1519 (vreinterpretq_u32_s8): Likewise.
1520 (vreinterpretq_u32_s16): Likewise.
1521 (vreinterpretq_u32_s32): Likewise.
1522 (vreinterpretq_u32_s64): Likewise.
1523 (vreinterpretq_u32_f32): Likewise.
1524 (vreinterpretq_u32_u8): Likewise.
1525 (vreinterpretq_u32_u16): Likewise.
1526 (vreinterpretq_u32_u64): Likewise.
1527 (vreinterpretq_u32_p8): Likewise.
1528 (vreinterpretq_u32_p16): Likewise.
1530 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1532 * gcc/config/aarch64/aarch64-simd.md (aarch64_s<optab><mode>):
1534 * config/aarch64/aarch64-simd-builtins.def (sqneg): Iterator
1537 * config/aarch64/arm_neon.h (vqneg_s64): New intrinsic.
1538 (vqnegd_s64): Likewise.
1539 (vqabs_s64): Likewise.
1540 (vqabsd_s64): Likewise.
1542 2014-04-22 Richard Henderson <rth@redhat.com>
1544 * config/sparc/sparc.c (sparc_init_modes): Hoist GET_MODE_SIZE
1545 computation to the top of the loop.
1547 2014-04-22 Renlin <renlin.li@arm.com>
1548 Jiong Wang <jiong.wang@arm.com>
1550 * config/aarch64/aarch64.h (aarch64_frame): Delete "fp_lr_offset".
1551 * config/aarch64/aarch64.c (aarch64_layout_frame)
1552 (aarch64_initial_elimination_offset): Likewise.
1554 2014-04-22 Marcus Shawcroft <marcus.shawcroft@arm.com>
1556 * config/aarch64/aarch64.c (aarch64_initial_elimination_offset):
1559 2014-04-22 Richard Sandiford <rdsandiford@googlemail.com>
1561 * machmode.h (bitwise_mode_for_mode): Declare.
1562 * stor-layout.h (bitwise_type_for_mode): Likewise.
1563 * stor-layout.c (bitwise_mode_for_mode): New function.
1564 (bitwise_type_for_mode): Likewise.
1565 * builtins.c (fold_builtin_memory_op): Use it instead of
1566 int_mode_for_mode and build_nonstandard_integer_type.
1568 2014-04-22 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1570 * config.gcc (enable_obsolete): Remove *-*-solaris2.9*.
1571 (*-*-solaris2.[0-9] | *-*-solaris2.[0-9].*): Mark unsupported.
1572 (*-*-solaris2*): Simplify.
1573 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Likewise.
1574 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Remove
1575 *-*-solaris2.9* handling.
1577 * configure.ac (gcc_cv_as_hidden): Remove test for Solaris 9/x86
1579 (gcc_cv_ld_hidden): Remove *-*-solaris2.9* handling.
1580 (ld_tls_support): Remove i?86-*-solaris2.9, sparc*-*-solaris2.9
1582 (gcc_cv_as_gstabs_flag): Remove workaround for Solaris 9/x86 as bug.
1583 * configure: Regenerate.
1585 * config/i386/sol2-9.h: Remove.
1587 * doc/install.texi (Specific, i?86-*-solaris2.9): Remove.
1588 (Specific, *-*-solaris2*): Mention Solaris 9 support removal.
1589 Remove Solaris 9 references.
1591 2014-04-22 Vidya Praveen <vidyapraveen@arm.com>
1593 * aarch64.md (float<GPI:mode><GPF:mode>2): Remove.
1594 (floatuns<GPI:mode><GPF:mode>2): Remove.
1595 (<optab><fcvt_target><GPF:mode>2): New pattern for equal width float
1596 and floatuns conversions.
1597 (<optab><fcvt_iesize><GPF:mode>2): New pattern for inequal width float
1598 and floatuns conversions.
1599 * iterators.md (fcvt_target, FCVT_TARGET): Support SF and DF modes.
1600 (w1,w2): New mode attributes for inequal width conversions.
1602 2014-04-22 Renlin Li <Renlin.Li@arm.com>
1604 * config/aarch64/aarch64.c (aarch64_print_operand_address): Adjust
1605 the output asm format.
1607 2014-04-22 James Greenhalgh <james.greenhalgh@arm.com>
1609 * config/aarch64/aarch64-simd.md
1610 (aarch64_cm<optab>di): Always split.
1611 (*aarch64_cm<optab>di): New.
1612 (aarch64_cmtstdi): Always split.
1613 (*aarch64_cmtstdi): New.
1615 2014-04-22 Jakub Jelinek <jakub@redhat.com>
1617 PR tree-optimization/60823
1618 * omp-low.c (ipa_simd_modify_function_body): Go through
1619 all SSA_NAMEs and for those refering to vector arguments
1620 which are going to be replaced adjust SSA_NAME_VAR and,
1621 if it is a default definition, change it into a non-default
1622 definition assigned at the beginning of function from new_decl.
1623 (ipa_simd_modify_stmt_ops): Rewritten.
1624 * tree-dfa.c (set_ssa_default_def): When removing default def,
1625 check for NULL loc instead of NULL *loc.
1627 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1629 * config/arm/arm.c (arm_hard_regno_mode_ok): Loosen
1630 restrictions on core registers for DImode values in Thumb2.
1632 2014-04-22 Ian Bolton <ian.bolton@arm.com>
1634 * config/arm/arm.md (*anddi_notdi_zesidi): New pattern.
1635 * config/arm/thumb2.md (*iordi_notdi_zesidi): New pattern.
1637 2014-04-22 Ian Bolton <ian.bolton@arm.com>
1639 * config/arm/thumb2.md (*iordi_notdi_di): New pattern.
1640 (*iordi_notzesidi_di): Likewise.
1641 (*iordi_notsesidi_di): Likewise.
1643 2014-04-22 Ian Bolton <ian.bolton@arm.com>
1645 * config/arm/arm-protos.h (tune_params): New struct members.
1646 * config/arm/arm.c: Initialise tune_params per processor.
1647 (thumb2_reorg): Suppress conversion from t32 to t16 when optimizing
1648 for speed, based on new tune_params.
1650 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1652 * config/aarch64/aarch64-builtins.c (BUILTIN_VDQF_DF): Macro
1654 * config/aarch64/aarch64-simd-builtins.def (frintn): Use added
1656 * config/aarch64/aarch64-simd.md (<frint_pattern>): Comment
1658 * config/aarch64/aarch64.md (<frint_pattern>): Likewise.
1659 * config/aarch64/arm_neon.h (vrnd_f64): Added.
1660 (vrnda_f64): Likewise.
1661 (vrndi_f64): Likewise.
1662 (vrndm_f64): Likewise.
1663 (vrndn_f64): Likewise.
1664 (vrndp_f64): Likewise.
1665 (vrndx_f64): Likewise.
1667 2014-04-22 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1669 * config/arm/arm.c (arm_print_operand, thumb_exit): Make sure
1670 GET_MODE_SIZE argument is enum machine_mode.
1672 2014-04-22 Jakub Jelinek <jakub@redhat.com>
1675 * config/sparc/sparc.c (sparc_init_modes): Pass enum machine_mode
1676 value instead of int to GET_MODE_CLASS and GET_MODE_SIZE macros.
1678 2014-04-22 Lin Zuojian <manjian2006@gmail.com>
1681 * asan.c (asan_emit_stack_protection): Force the base to align to
1682 appropriate bits if STRICT_ALIGNMENT. Set shadow_mem align to
1683 appropriate bits if STRICT_ALIGNMENT.
1684 * cfgexpand.c (expand_stack_vars): Set base_align appropriately
1686 (expand_used_vars): Leave a space in the stack frame for alignment
1687 if STRICT_ALIGNMENT.
1689 2014-04-21 David Malcolm <dmalcolm@redhat.com>
1691 * gimple.h (gimple_assign_single_p): Accept a const_gimple rather
1693 (gimple_store_p): Likewise.
1694 (gimple_assign_load_p): Likewise.
1695 (gimple_assign_cast_p): Likewise.
1696 (gimple_clobber_p): Likewise.
1698 * doc/gimple.texi (gimple_assign_cast_p): Accept a const_gimple
1699 rather than a gimple.
1700 (gimple_assign_cast_p): Likewise.
1702 2014-04-21 Michael Meissner <meissner@linux.vnet.ibm.com>
1705 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64 case):
1706 If mode is DDmode and TARGET_E500_DOUBLE allow move.
1708 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print some
1709 more debug information for E500 if -mdebug=reg.
1711 2014-04-21 Uros Bizjak <ubizjak@gmail.com>
1714 * config/i386/i386.c (ix86_expand_builtin)
1715 <case IX86_BUILTIN_RDRAND{16,32,64}_STEP>: Use temporary
1716 register for target RTX.
1717 <case IX86_BUILTIN_RDSEED{16,32,64}_STEP>: Ditto.
1719 2014-04-18 Cong Hou <congh@google.com>
1721 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): Enhance
1722 the widen-mult pattern by handling two operands with different sizes,
1723 and operands whose size is smaller than half of the result type.
1725 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
1727 * ipa-inline.h (INLINE_HINT_known_hot): New hint.
1728 * ipa-inline-analysis.c (dump_inline_hints): Dump it.
1729 (do_estimate_edge_time): Compute it.
1730 * ipa-inline.c (want_inline_small_function_p): Bypass
1731 INLINE_INSNS_AUTO/SINGLE limits for calls that are known to be hot.
1733 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
1735 * ipa-inline.c (spec_rem): New static variable.
1736 (dump_overall_stats): New function.
1737 (dump_inline_stats): New function.
1739 2014-04-18 Richard Henderson <rth@redhat.com>
1741 * config/aarch64/aarch64.c (aarch64_register_move_cost): Pass a mode
1742 to GET_MODE_SIZE, not a reg_class_t.
1744 2014-04-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1746 * config/rs6000/vsx.md (vsx_xxmrghw_<mode>): Adjust for little-endian.
1747 (vsx_xxmrglw_<mode>): Likewise.
1749 2014-04-17 Michael Meissner <meissner@linux.vnet.ibm.com>
1752 * config/rs6000/rs6000.c (rs6000_setup_reg_addr_masks): Make sure
1753 GET_MODE_SIZE gets passed an enum machine_mode type and not integer.
1754 (rs6000_init_hard_regno_mode_ok): Likewise.
1756 2014-04-17 Jan Hubicka <hubicka@ucw.cz>
1758 * ipa-inline.c (inline_small_functions): Account only non-cold
1760 * doc/invoke.texi (inline-unit-growth): Update documentation.
1762 2014-04-17 Pat Haugen <pthaugen@us.ibm.com>
1764 * config/rs6000/rs6000.md (addti3, subti3): New.
1766 2014-04-17 H.J. Lu <hongjiu.lu@intel.com>
1769 * config/i386/i386.c (ix86_expand_clear): Remove outdated
1770 comment. Check optimize_insn_for_size_p instead of
1771 optimize_insn_for_speed_p.
1773 2014-04-17 Martin Jambor <mjambor@suse.cz>
1775 * gimple-iterator.c (gsi_start_edge): New function.
1776 * gimple-iterator.h (gsi_start_edge): Declare.
1777 * tree-sra.c (single_non_eh_succ): New function.
1778 (disqualify_ops_if_throwing_stmt): Renamed to
1779 disqualify_if_bad_bb_terminating_stmt. Allow throwing statements
1780 having one non-EH successor BB.
1781 (sra_modify_expr): If stmt ends bb, use single non-EH successor to
1782 generate loads into replacements.
1783 (sra_modify_assign): Likewise and and also use the simple path for
1785 (sra_modify_function_body): Commit statements on edges.
1787 2014-04-17 Richard Biener <rguenther@suse.de>
1790 * tree-ssa-propagate.c (valid_gimple_rhs_p): Allow vector
1791 comparison results and add clarifying comment.
1793 2014-04-17 Jakub Jelinek <jakub@redhat.com>
1795 * genmodes.c (struct mode_data): Add need_bytesize_adj field.
1796 (blank_mode): Initialize it.
1797 (emit_mode_size_inline, emit_mode_nunits_inline,
1798 emit_mode_inner_inline): New functions.
1799 (emit_insn_modes_h): Call them and surround their output with
1800 #if GCC_VERSION >= 4001 ... #endif.
1801 * machmode.h (GET_MODE_SIZE, GET_MODE_NUNITS, GET_MODE_INNER):
1802 For GCC_VERSION >= 4001 use mode_*_inline routines instead of
1803 mode_* arrays if the argument is __builtin_constant_p.
1804 * lower-subreg.c (dump_choices): Make sure GET_MODE_SIZE argument
1805 is enum machine_mode.
1807 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1809 * passes.c (opt_pass::execute): Adjust.
1810 (pass_manager::execute_pass_mode_switching): Likewise.
1811 (early_local_passes::execute): Likewise.
1812 (execute_one_pass): Pass cfun to the pass's execute method.
1813 * tree-pass.h (opt_pass::execute): Add function * argument.
1814 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
1815 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
1816 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
1817 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
1818 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
1819 config/sparc/sparc.c, cprop.c, dce.c, df-core.c, dse.c, dwarf2cfi.c,
1820 except.c, final.c, function.c, fwprop.c, gcse.c, gimple-low.c,
1821 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
1822 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
1823 ipa-inline-analysis.c, ipa-inline.c, ipa-profile.c, ipa-pure-const.c,
1824 ipa-reference.c, ipa-split.c, ipa.c, ira.c, jump.c, loop-init.c,
1825 lower-subreg.c, mode-switching.c, omp-low.c, postreload-gcse.c,
1826 postreload.c, predict.c, recog.c, ree.c, reg-stack.c, regcprop.c,
1827 reginfo.c, regrename.c, reorg.c, sched-rgn.c, stack-ptr-mod.c,
1828 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
1829 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
1830 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
1831 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-ssa-ccp.c,
1832 tree-ssa-copy.c, tree-ssa-copyrename.c, tree-ssa-dce.c,
1833 tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
1834 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
1835 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
1836 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
1837 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
1838 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
1839 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
1840 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
1841 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
1842 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
1845 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1847 * passes.c (opt_pass::gate): Take function * argument.
1848 (gate_all_early_local_passes): Merge into
1849 (early_local_passes::gate): this.
1850 (gate_all_early_optimizations): Merge into
1851 (all_early_optimizations::gate): this.
1852 (gate_all_optimizations): Mege into
1853 (all_optimizations::gate): this.
1854 (gate_all_optimizations_g): Merge into
1855 (all_optimizations_g::gate): this.
1856 (gate_rest_of_compilation): Mege into
1857 (rest_of_compilation::gate): this.
1858 (gate_postreload): Merge into
1859 (postreload::gate): this.
1860 (dump_one_pass): Pass cfun to the pass's gate method.
1861 (execute_ipa_summary_passes): Likewise.
1862 (execute_one_pass): Likewise.
1863 (ipa_write_summaries_2): Likewise.
1864 (ipa_write_optimization_summaries_1): Likewise.
1865 (ipa_read_summaries_1): Likewise.
1866 (ipa_read_optimization_summaries_1): Likewise.
1867 (execute_ipa_stmt_fixups): Likewise.
1868 * tree-pass.h (opt_pass::gate): Add function * argument.
1869 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c,
1870 combine-stack-adj.c, combine.c, compare-elim.c,
1871 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
1872 config/rl78/rl78.c, config/sh/sh_optimize_sett_clrt.cc,
1873 config/sh/sh_treg_combine.cc, config/sparc/sparc.c, cprop.c, cse.c,
1874 dce.c, df-core.c, dse.c, dwarf2cfi.c, except.c, fwprop.c, gcse.c,
1875 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
1876 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
1877 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
1878 loop-init.c, lower-subreg.c, mode-switching.c, modulo-sched.c,
1879 omp-low.c, postreload-gcse.c, postreload.c, predict.c, recog.c, ree.c,
1880 reg-stack.c, regcprop.c, regrename.c, reorg.c, sched-rgn.c,
1881 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
1882 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
1883 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c,
1884 tree-nrv.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
1885 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
1886 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
1887 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
1888 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
1889 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
1890 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
1891 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
1892 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
1893 tree-ssa.c, tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
1894 tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c, tsan.c, ubsan.c,
1895 var-tracking.c, vtable-verify.c, web.c: Adjust.
1897 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1899 * configure.ac: Check for -Woverloaded-virtual and enable it if found.
1900 * configure: Regenerate.
1902 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1904 * passes.c (dump_one_pass): don't check pass->has_gate.
1905 (execute_ipa_summary_passes): Likewise.
1906 (execute_one_pass): Likewise.
1907 (ipa_write_summaries_2): Likewise.
1908 (ipa_write_optimization_summaries_1): Likewise.
1909 (ipa_read_optimization_summaries_1): Likewise.
1910 (execute_ipa_stmt_fixups): Likewise.
1911 * tree-pass.h (pass_data::has_gate): Remove.
1912 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
1913 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
1914 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
1915 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
1916 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
1917 config/sh/sh_optimize_sett_clrt.cc, config/sh/sh_treg_combine.cc,
1918 config/sparc/sparc.c, cprop.c, cse.c, dce.c, df-core.c, dse.c,
1919 dwarf2cfi.c, except.c, final.c, function.c, fwprop.c, gcse.c,
1920 gimple-low.c, gimple-ssa-isolate-paths.c,
1921 gimple-ssa-strength-reduction.c, graphite.c, ifcvt.c, init-regs.c,
1922 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
1923 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
1924 ira.c, jump.c, loop-init.c, lower-subreg.c, mode-switching.c,
1925 modulo-sched.c, omp-low.c, postreload-gcse.c, postreload.c, predict.c,
1926 recog.c, ree.c, reg-stack.c, regcprop.c, reginfo.c, regrename.c,
1927 reorg.c, sched-rgn.c, stack-ptr-mod.c, store-motion.c, tracer.c,
1928 trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
1929 tree-complex.c, tree-eh.c, tree-emutls.c, tree-if-conv.c,
1930 tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
1931 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
1932 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
1933 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
1934 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
1935 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
1936 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
1937 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
1938 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
1939 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
1940 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
1941 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
1942 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
1945 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1947 * pass_manager.h (pass_manager::register_dump_files_1): Remove
1949 * passes.c (pass_manager::register_dump_files_1): Merge into
1950 (pass_manager::register_dump_files): this, and remove its handling of
1951 properties since the pass always has the properties anyway.
1952 (pass_manager::pass_manager): Adjust.
1954 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1956 * pass_manager.h (pass_manager::register_dump_files_1): Adjust.
1957 * passes.c (pass_manager::register_dump_files_1): Remove dead code
1958 dealing with properties.
1959 (pass_manager::register_dump_files): Adjust.
1961 2014-03-20 Mark Wielaard <mjw@redhat.com>
1963 * dwarf2out.c (add_bound_info): If HOST_WIDE_INT is big enough,
1964 then represent the bound as normal constant value.
1966 2014-04-17 Jakub Jelinek <jakub@redhat.com>
1969 Forward port from 4.8 branch
1970 2013-07-19 Kirill Yukhin <kirill.yukhin@intel.com>
1972 * config/i386/bmiintrin.h (_blsi_u32): New.
1976 (_blsmsk_u32): Ditto.
1977 (_blsmsk_u64): Ditto.
1978 (_tzcnt_u32): Ditto.
1979 (_tzcnt_u64): Ditto.
1981 2014-04-17 Kito Cheng <kito@0xlab.org>
1983 * gcc.c (used_arg): Prevent out of bound access for multilib_options.
1985 2014-04-17 Richard Biener <rguenther@suse.de>
1988 * tree-ssa-propagate.c (valid_gimple_rhs_p): Only allow effective
1989 boolean results for comparisons.
1991 2014-04-17 Richard Biener <rguenther@suse.de>
1993 PR tree-optimization/60836
1994 * tree-vect-loop.c (vect_create_epilog_for_reduction): Force
1995 initial PHI args to be gimple values.
1997 2014-04-17 Richard Biener <rguenther@suse.de>
1999 PR tree-optimization/60841
2000 * tree-vect-data-refs.c (vect_analyze_data_refs): Count stmts.
2001 * tree-vect-loop.c (vect_analyze_loop_2): Pass down number
2002 of stmts to SLP build.
2003 * tree-vect-slp.c (vect_slp_analyze_bb_1): Likewise.
2004 (vect_analyze_slp): Likewise.
2005 (vect_analyze_slp_instance): Likewise.
2006 (vect_build_slp_tree): Limit overall SLP tree growth.
2007 * tree-vectorizer.h (vect_analyze_data_refs,
2008 vect_analyze_slp): Adjust prototypes.
2010 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
2012 * config/i386/i386.c (x86_add_stmt_cost): Fix vector cost model for
2015 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
2017 * config/i386/x86-tune.def (TARGET_SLOW_PSHUFB): New tune definition.
2018 * config/i386/i386.h (TARGET_SLOW_PSHUFB): New tune flag.
2019 * config/i386/i386.c (expand_vec_perm_even_odd_1): Avoid byte shuffles
2020 for TARGET_SLOW_PSHUFB
2022 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
2024 * config/i386/i386.c (slm_cost): Adjust vec_to_scalar_cost.
2025 * config/i386/i386.c (intel_cost): Ditto.
2027 2014-04-17 Joey Ye <joey.ye@arm.com>
2029 * opts.c (OPT_fif_conversion, OPT_fif_conversion2): Disable for Og.
2031 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
2033 * opts.c (common_handle_option): Disable -fipa-reference coorectly
2036 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
2038 * ipa-devirt.c (odr_type_d): Add field all_derivations_known.
2039 (type_all_derivations_known_p): New predicate.
2040 (type_all_ctors_visible_p): New predicate.
2041 (type_possibly_instantiated_p): New predicate.
2042 (get_odr_type): Compute all_derivations_known.
2043 (dump_odr_type): Dump the flag.
2044 (maybe_record_type): Cleanup.
2045 (record_target_from_binfo): Add bases_to_consider array;
2046 record bases for types w/o instances and skip CXX destructor.
2047 (possible_polymorphic_call_targets_1): Add bases_to_consider
2048 and consider_construction parameters; check if type may have instance.
2049 (get_polymorphic_call_info): Set maybe_in_construction to true
2050 when we know nothing.
2051 (record_targets_from_bases): Skip CXX destructors; they are
2052 never called for types in construction.
2053 (possible_polymorphic_call_targets): Do not record target when
2054 type may not have instance.
2056 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
2059 * ipa.c (symtab_remove_unreachable_nodes): Mark targets of
2060 external aliases alive, too.
2062 2014-04-16 Andrew Pinski <apinski@cavium.com>
2064 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Change aarch64 ilp32
2067 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
2069 * final.c (compute_alignments): Do not apply loop alignment to a block
2070 falling through to the exit.
2072 2014-04-16 Catherine Moore <clm@codesourcery.com>
2074 * mips.md (*mov<mode>_internal, *movhi_internal, *movqi_internal):
2075 Adjust constraints for microMIPS store patterns.
2077 2014-04-16 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
2079 * config/avr/avr-mcus.def: Correct typo for atxmega256a3bu macro.
2081 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
2083 * tree-ssa-operands.c (create_vop_var): Set DECL_IGNORED_P.
2084 (append_use): Run at -O0.
2085 (append_vdef): Likewise.
2086 * tree-ssa-ter.c (ter_is_replaceable_p): Do not special-case -O0.
2087 * tree-ssa-uninit.c (warn_uninitialized_vars): Remove obsolete comment.
2089 2014-04-16 Jakub Jelinek <jakub@redhat.com>
2091 PR tree-optimization/60844
2092 * tree-ssa-reassoc.c (reassoc_remove_stmt): New function.
2093 (propagate_op_to_single_use, remove_visited_stmt_chain,
2094 linearize_expr, repropagate_negates, reassociate_bb): Use it
2095 instead of gsi_remove.
2097 2014-04-16 Martin Jambor <mjambor@suse.cz>
2099 * cgraphclones.c (cgraph_create_virtual_clone): Duplicate
2100 ipa_transforms_to_apply.
2101 (cgraph_function_versioning): Assert that old_node has empty
2102 ipa_transforms_to_apply.
2103 * trans-mem.c (ipa_tm_create_version): Likewise.
2104 * tree-inline.c (tree_function_versioning): Do not duplicate
2105 ipa_transforms_to_apply.
2107 2014-04-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2110 * configure.ac (set_have_as_tls): Merge i[34567]86-*-* and
2112 Pass necessary as flags on 64-bit Solaris/x86.
2113 Use lowercase relocs for x86_64-*-*.
2114 * configure: Regenerate.
2116 2014-04-15 Jan Hubicka <jh@suse.cz>
2118 * ipa-devirt.c (referenced_from_vtable_p): New predicate.
2119 (maybe_record_node, likely_target_p): Use it.
2121 2014-04-15 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2124 Revert following patch
2126 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
2129 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
2130 software floating point or no floating point registers, do not
2131 allow any type in the FPRs. Eliminate a test for SPE SIMD types
2132 in GPRs that occurs after we tested for GPRs that would never be
2135 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
2136 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
2137 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
2138 specifically allow DDmode, since that does not use the SPE SIMD
2141 2014-03-21 Mark Wielaard <mjw@redhat.com>
2143 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_const_value
2144 as unsigned or int depending on type and value used.
2146 2014-04-15 Richard Biener <rguenther@suse.de>
2148 PR rtl-optimization/56965
2149 * alias.c (ncr_compar, nonoverlapping_component_refs_p): Move ...
2150 * tree-ssa-alias.c (ncr_compar, nonoverlapping_component_refs_p):
2152 * alias.c (true_dependence_1): Do not call
2153 nonoverlapping_component_refs_p.
2154 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Call
2155 nonoverlapping_component_refs_p.
2156 (indirect_refs_may_alias_p): Likewise.
2158 2014-04-15 Teresa Johnson <tejohnson@google.com>
2160 * cfg.c (dump_bb_info): Fix flags check.
2161 * tree-cfg.c (remove_bb): Only dump TDF_BLOCKS when removing.
2163 2014-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2165 PR rtl-optimization/60663
2166 * config/arm/arm.c (arm_new_rtx_costs): Improve ASM_OPERANDS case,
2169 2014-04-15 Richard Biener <rguenther@suse.de>
2171 * lto-streamer.h (LTO_major_version): Bump to 4.
2173 2014-04-15 Richard Biener <rguenther@suse.de>
2175 * common.opt (lto_partition_model): New enum.
2176 (flto-partition=): Merge separate options with a single with argument,
2177 add -flto-partition=one support.
2178 * flag-types.h (enum lto_partition_model): Declare.
2179 * opts.c (finish_options): Remove duplicate -flto-partition=
2181 * lto-wrapper.c (run_gcc): Adjust.
2183 2014-04-15 Richard Biener <rguenther@suse.de>
2185 * alias.c (ncr_compar): New function.
2186 (nonoverlapping_component_refs_p): Re-implement in O (n log n).
2188 2014-04-15 Richard Biener <rguenther@suse.de>
2190 * alias.c (record_component_aliases): Do not walk BINFOs.
2192 2014-04-15 Richard Biener <rguenther@suse.de>
2194 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
2195 Add struct function argument and adjust.
2196 (find_func_aliases_for_call): Likewise.
2197 (find_func_aliases): Likewise.
2198 (find_func_clobbers): Likewise.
2199 (intra_create_variable_infos): Likewise.
2200 (compute_points_to_sets): Likewise.
2201 (ipa_pta_execute): Adjust. Do not push/pop cfun.
2203 2014-04-15 Richard Biener <rguenther@suse.de>
2205 * tree.c (iterative_hash_expr): Use enum tree_code_class
2206 to store TREE_CODE_CLASS.
2207 (tree_block): Likewise.
2208 (tree_set_block): Likewise.
2209 * tree.h (fold_build_pointer_plus_loc): Use
2210 convert_to_ptrofftype_loc.
2212 2014-04-15 Jakub Jelinek <jakub@redhat.com>
2215 * Makefile.in (PLUGIN_HEADERS): Add various headers that have been
2218 2014-04-15 Eric Botcazou <ebotcazou@adacore.com>
2220 * cfgloop.h (struct loop): Move force_vectorize down.
2221 * gimplify.c (gimple_boolify) <ANNOTATE_EXPR>: Handle new kinds.
2222 (gimplify_expr) <ANNOTATE_EXPR>: Minor tweak.
2223 * lto-streamer-in.c (input_cfg): Read dont_vectorize field.
2224 * lto-streamer-out.c (output_cfg): Write dont_vectorize field.
2225 * tree-cfg.c (replace_loop_annotate): Revamp and handle new kinds.
2226 * tree-core.h (enum annot_expr_kind): Add new kind values.
2227 * tree-inline.c (copy_loops): Copy dont_vectorize field and reorder.
2228 * tree-pretty-print.c (dump_generic_node) <ANNOTATE_EXPR>: Handle new
2230 * tree.def (ANNOTATE_EXPR): Tweak comment.
2232 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
2234 * ipa-devirt.c (maybe_record_node): Ignore all non-methods (including
2237 2014-04-14 Paolo Carlini <paolo.carlini@oracle.com>
2239 * tree.h (TYPE_IDENTIFIER): Declare.
2240 * tree.c (subrange_type_for_debug_p): Use it.
2241 * godump.c (go_format_type): Likewise.
2242 * dwarf2out.c (is_cxx_auto, modified_type_die,
2243 gen_type_die_with_usage, gen_type_die_with_usage): Likewise.
2244 * dbxout.c (dbxout_type, dbxout_symbol): Likewise.
2246 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
2249 * varpool.c (varpool_remove_node): Do not alter decls when streaming.
2251 2014-04-14 Uros Bizjak <ubizjak@gmail.com>
2253 * config/i386/i386.c (examine_argument): Return bool. Return true if
2254 parameter should be passed in memory.
2255 <case X86_64_COMPLEX_X87_CLASS>: Adjust.
2256 (construct_container): Update calls to examine_argument.
2257 (function_arg_advance_64): Ditto.
2258 (return_in_memory_32): Merge with ix86_return_in_memory.
2259 (return_in_memory_64): Ditto.
2260 (return_in_memory_ms_64): Ditto.
2262 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
2264 * ipa-utils.c (ipa_merge_profiles): Merge profile_id.
2265 * coverage.c (coverage_compute_profile_id): Handle externally visible
2268 2014-04-14 Martin Jambor <mjambor@suse.cz>
2270 * tree-sra.c (ipa_sra_preliminary_function_checks): Skip
2271 DECL_DISREGARD_INLINE_LIMITS functions.
2273 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
2276 * config/i386/i386.md (*fixuns_trunc<mode>_1): Revert the last change.
2278 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
2281 * config/i386/i386.md (*fixuns_trunc<mode>_1): Check
2282 optimize_insn_for_speed_p instead of
2283 optimize_function_for_speed_p.
2285 2014-04-14 Yufeng Zhang <yufeng.zhang@arm.com>
2287 * doc/invoke.texi (free): Document AArch64.
2289 2014-04-14 Richard Biener <rguenther@suse.de>
2291 PR tree-optimization/60042
2292 * tree-ssa-pre.c (inhibit_phi_insertion): Remove.
2293 (insert_into_preds_of_block): Do not prevent PHI insertion
2294 for REFERENCE exprs here ...
2295 (eliminate_dom_walker::before_dom_children): ... but prevent
2296 their use here under similar conditions when applied to the
2297 IL after PRE optimizations.
2299 2014-04-14 Richard Biener <rguenther@suse.de>
2301 * passes.def: Move early points-to after early SRA.
2303 2014-04-14 Richard Biener <rguenther@suse.de>
2305 * tree-ssa-forwprop.c (simplify_gimple_switch): Enhance
2306 check for which sign-changes we allow when forwarding
2307 a converted value into a switch.
2309 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
2311 * stor-layout.c (place_field): Finalize non-constant offset for the
2314 2014-04-14 Richard Biener <rguenther@suse.de>
2316 * tree-switch-conversion.c (lshift_cheap_p): Get speed_p
2318 (expand_switch_using_bit_tests_p): Likewise.
2319 (process_switch): Compute and pass on speed_p based on the
2321 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Use
2322 optimize_bb_for_speed_p.
2324 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
2326 * cfgloop.h (struct loop): Rename force_vect into force_vectorize.
2327 * function.h (struct function): Rename has_force_vect_loops into
2328 has_force_vectorize_loops.
2329 * lto-streamer-in.c (input_cfg): Adjust for renaming.
2330 (input_struct_function_base): Likewise.
2331 * lto-streamer-out.c (output_cfg): Likewise.
2332 (output_struct_function_base): Likewise.
2333 * omp-low.c (expand_omp_simd): Likewise.
2334 * tree-cfg.c (move_sese_region_to_fn): Likewise.
2335 * tree-if-conv.c (ifcvt_can_use_mask_load_store): Likewise.
2336 (version_loop_for_if_conversion): Likewise.
2337 (tree_if_conversion): Likewise.
2338 (main_tree_if_conversion): Likewise.
2339 (gate_tree_if_conversion): Likewise.
2340 * tree-inline.c (copy_loops): Likewise.
2341 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Likewise.
2342 * tree-ssa-loop.c (tree_loop_vectorize): Likewise.
2343 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Likewise.
2344 * tree-vect-loop.c (vect_estimate_min_profitable_iters): Likewise.
2345 * tree-vectorizer.c (vectorize_loops): Likewise.
2346 * tree-vectorizer.h (unlimited_cost_model): Likewise.
2348 2014-04-14 Richard Biener <rguenther@suse.de>
2351 * lto-streamer-out.c (wrap_refs): New function.
2352 (lto_output): Wrap symbol references in global initializes in
2353 type-preserving MEM_REFs.
2355 2014-04-14 Christian Bruel <christian.bruel@st.com>
2357 * config/sh/sh-mem.cc (sh_expand_strlen): Unroll last word.
2359 2014-04-14 Christian Bruel <christian.bruel@st.com>
2361 * config/sh/sh.md (setmemqi): New expand pattern.
2362 * config/sh/sh.h (CLEAR_RATIO): Define.
2363 * config/sh/sh-mem.cc (sh_expand_setmem): Define.
2364 * config/sh/sh-protos.h (sh_expand_setmem): Declare.
2366 2014-04-14 Richard Biener <rguenther@suse.de>
2369 * fold-const.c (negate_expr_p): Don't negate directional rounding
2371 (fold_negate_expr): Likewise.
2373 2014-04-14 Richard Biener <rguenther@suse.de>
2375 PR tree-optimization/59817
2376 PR tree-optimization/60453
2377 * graphite-scop-detection.c (graphite_can_represent_scev): Complete
2378 recursion to catch all CHRECs in the scalar evolution and restrict
2379 the predicate for the remains appropriately.
2381 2014-04-12 Catherine Moore <clm@codesourcery.com>
2383 * config/mips/constraints.md: Add new register constraint "kb".
2384 * config/mips/mips.md (*mov<mode>_internal): Use constraint "kb".
2385 (*movhi_internal): Likewise.
2386 (*movqi_internal): Likewise.
2387 * config/mips/mips.h (M16_STORE_REGS): New register class.
2388 (REG_CLASS_NAMES): Add M16_STORE_REGS.
2389 (REG_CLASS_CONTENTS): Likewise.
2390 * config/mips/mips.c (mips_regno_to_class): Add M16_STORE_REGS.
2392 2014-04-11 Tobias Burnus <burnus@net-b.de>
2395 * doc/invoke.texi (-Wformat-signedness): Document it.
2396 (Wformat=2): Mention that this enables -Wformat-signedness.
2398 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
2400 * common/config/epiphany/epiphany-common.c
2401 (epiphany_option_optimization_table): Enable section anchors by
2402 default at -O1 or higher.
2403 * config/epiphany/epiphany.c (TARGET_MAX_ANCHOR_OFFSET): Define.
2404 (TARGET_MIN_ANCHOR_OFFSET): Likewise.
2405 (epiphany_rtx_costs) <SET>: For binary operators, the set as such
2406 carries no extra cost.
2407 (epiphany_legitimate_address_p): For BLKmode, apply SImode check.
2408 * config/epiphany/epiphany.h (ASM_OUTPUT_DEF): Define.
2409 * config/epiphany/predicates.md (memclob_operand): New predicate.
2410 * config/epiphany/epiphany.md (stack_adjust_add, stack_adjust_str):
2411 Use memclob_operand predicate and X constraint for operand 3.
2413 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
2415 * config/epiphany/epiphany.c (epiphany_rtx_cost): Compare
2416 with CC_N_NE / CC_C_LTU / CC_C_GTU carries no extra cost for
2419 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
2421 PR rtl-optimization/60651
2422 * mode-switching.c (optimize_mode_switching): Make sure to emit
2423 sets of a lower numbered entity before sets of a higher numbered
2424 entity to a mode of the same or lower priority.
2425 When creating a seginfo for a basic block that starts with a code
2426 label, move the insertion point past the code label.
2427 (new_seginfo): Document and enforce requirement that
2428 NOTE_INSN_BASIC_BLOCK only appears for empty blocks.
2429 * doc/tm.texi.in: Document ordering constraint for emitted mode sets.
2430 * doc/tm.texi: Regenerate.
2432 2014-01-11 Joern Rennecke <joern.rennecke@embecosm.com>
2435 * config/arc/arc.c (arc_save_restore): Fix assert typo.
2437 2013-04-11 Jakub Jelinek <jakub@redhat.com>
2439 * BASE-VER: Set to 4.10.0.
2441 2014-04-11 Tobias Burnus <burnus@net-b.de>
2444 * doc/bugreport.texi (Bugs): Remove nodes pointing to the nirvana.
2445 * doc/gcc.texi (Service): Update description in the @menu
2446 * doc/invoke.texi (Option Summary): Remove misplaced and
2449 2014-04-11 Steve Ellcey <sellcey@mips.com>
2450 Jakub Jelinek <jakub@redhat.com>
2453 * expr.c (convert_move): Use emit_store_flag_force instead of
2454 emit_store_flag. Pass lowpart_mode instead of VOIDmode as 5th
2457 2014-04-11 Richard Biener <rguenther@suse.de>
2460 * varasm.c (assemble_alias): Avoid endless error reporting
2461 recursion by setting TREE_ASM_WRITTEN.
2463 2014-04-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2465 * config/s390/s390.md: Add a splitter for NOT rtx.
2467 2014-04-11 Jakub Jelinek <jakub@redhat.com>
2469 PR rtl-optimization/60663
2470 * cse.c (cse_insn): Set src_volatile on ASM_OPERANDS in PARALLEL.
2472 2014-04-10 Jan Hubicka <hubicka@ucw.cz>
2473 Jakub Jelinek <jakub@redhat.com>
2476 * ipa.c (function_and_variable_visibility): Copy forced_by_abi
2477 flag from decl_node to node.
2479 2014-04-10 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2482 * config/arm/arm.c (TARGET_CONST_NOT_OK_FOR_DEBUG_P): Define
2483 (arm_const_not_ok_for_debug_p): Reject MINUS with SYM_REF's
2484 ameliorating the cases where it can be.
2486 2014-04-09 David Edelsohn <dje.gcc@gmail.com>
2489 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
2491 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
2492 (loadsync_<mode>): Change mode.
2493 (load_quadpti, store_quadpti): New.
2494 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
2495 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
2496 * config/rs6000/predicates.md (quad_memory_operand): !TARGET_SYNC_TI.
2498 2014-04-09 Cong Hou <congh@google.com>
2501 * doc/sourcebuild.texi (vect_widen_mult_si_to_di_pattern): Add
2504 2014-04-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2506 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Use vnand
2507 instead of vnor to exploit possible fusion opportunity in the
2509 (altivec_expand_vec_perm_const_le): Likewise.
2511 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
2513 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
2514 (loadsync_<mode>): Change mode.
2515 (load_quadpti, store_quadpti): New.
2516 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
2517 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
2519 2014-04-08 Richard Sandiford <rdsandiford@googlemail.com>
2522 * config/rs6000/vsx.md (vsx_xscvdpspn_scalar): Change input to DImode.
2523 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Update accordingly.
2524 Use gen_rtx_REG rather than simplify_gen_subreg for op0_di.
2526 2014-04-08 Richard Biener <rguenther@suse.de>
2529 * tree-pretty-print.c (pp_double_int): For HWI32 hosts with
2530 a 64bit widest int print double-int similar to on HWI64 hosts.
2532 2014-04-08 Richard Biener <rguenther@suse.de>
2534 PR tree-optimization/60785
2535 * graphite-sese-to-poly.c (rewrite_phi_out_of_ssa): Treat
2536 default defs properly.
2538 2014-04-08 Nathan Sidwell <nathan@codesourcery.com>
2540 * doc/invoke (Wnon-virtual-dtor): Update to match implementation.
2541 (Weffc++): Likewise.
2543 2014-04-07 Jan Hubicka <hubcika@ucw.cz>
2545 * ipa-devirt.c (maybe_record_node): When node is not recorded,
2546 set completep to false rather than true.
2548 2014-04-07 Douglas B Rupp <rupp@adacore.com>
2551 * config/arm/arm.h (ASM_PREFERRED_EH_DATA_FORMAT): Expose from
2552 ARM_TARGET2_DWARF_FORMAT.
2554 2014-04-07 Charles Baylis <charles.baylis@linaro.org>
2557 * config/arm/arm.h (ASM_OUTPUT_CASE_END): Remove.
2558 (LABEL_ALIGN_AFTER_BARRIER): Align barriers which occur after
2561 2014-04-07 Richard Biener <rguenther@suse.de>
2563 PR tree-optimization/60766
2564 * tree-ssa-loop-ivopts.c (cand_value_at): Compute in an unsigned type.
2565 (may_eliminate_iv): Convert cand_value_at result to desired type.
2567 2014-04-07 Jason Merrill <jason@redhat.com>
2570 * common.opt (-fno-gnu-unique): Add.
2571 * config/elfos.h (USE_GNU_UNIQUE_OBJECT): Check it.
2573 2014-04-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2575 * haifa-sched.c: Fix outdated function reference and minor
2576 grammar errors in introductory comment.
2578 2014-04-07 Richard Biener <rguenther@suse.de>
2581 * tree-ssa-operands.c (maybe_add_call_vops): Also add VDEFs
2583 * tree-cfgcleanup.c (fixup_noreturn_call): Do not remove VDEFs.
2585 2014-04-06 John David Anglin <danglin@gcc.gnu.org>
2588 * config/pa/pa.c (pa_output_function_epilogue): Skip address and code
2589 size accounting for thunks.
2590 (pa_asm_output_mi_thunk): Use final_start_function() and
2591 final_end_function() to output function start and end directives.
2593 2014-04-05 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
2595 * config/avr/avr-arch.h (avr_mcu_t): Add dev_attribute field to have
2596 device specific ISA/ feature information. Remove short_sp and
2597 errata_skip ds. Add avr_device_specific_features enum to have device
2599 * config/avr/avr-c.c (avr_cpu_cpp_builtins): use dev_attribute to check
2600 errata_skip. Add __AVR_ISA_RMW__ builtin macro if RMW ISA available.
2601 * config/avr/avr-devices.c (avr_mcu_types): Update AVR_MCU macro for
2602 updated device specific info.
2603 * config/avr/avr-mcus.def: Merge device specific details to
2604 dev_attribute field.
2605 * config/avr/avr.c (avr_2word_insn_p): use dev_attribute field to check
2607 * config/avr/avr.h (AVR_HAVE_8BIT_SP): same for short sp info.
2608 * config/avr/driver-avr.c (avr_device_to_as): Pass -mrmw option to
2609 assembler if RMW isa supported by current device.
2610 * config/avr/genmultilib.awk: Update as device info structure changed.
2611 * doc/invoke.texi: Add info for __AVR_ISA_RMW__ builtin macro
2613 2014-04-04 Cong Hou <congh@google.com>
2615 PR tree-optimization/60656
2616 * tree-vect-stmts.c (supportable_widening_operation):
2617 Fix a bug that elements in a vector with vect_used_by_reduction
2618 property are incorrectly reordered when the operation on it is not
2619 consistant with the one in reduction operation.
2621 2014-04-04 John David Anglin <danglin@gcc.gnu.org>
2623 PR rtl-optimization/60155
2624 * gcse.c (record_set_data): New function.
2625 (single_set_gcse): New function.
2626 (gcse_emit_move_after): Use single_set_gcse instead of single_set.
2627 (hoist_code): Likewise.
2628 (get_pressure_class_and_nregs): Likewise.
2630 2014-04-04 Eric Botcazou <ebotcazou@adacore.com>
2632 * explow.c (probe_stack_range): Emit a final optimization blockage.
2634 2014-04-04 Anthony Green <green@moxielogic.com>
2636 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2): Fix
2639 2014-04-04 Jan Hubicka <hubicka@ucw.cz>
2642 * lto-cgraph.c (input_overwrite_node): Check that partitioning
2643 flags are set only during streaming.
2644 * ipa.c (process_references, walk_polymorphic_call_targets,
2645 symtab_remove_unreachable_nodes): Drop bodies of always inline
2646 after early inlining.
2647 (symtab_remove_unreachable_nodes): Remove always_inline attribute.
2649 2014-04-04 Jakub Jelinek <jakub@redhat.com>
2650 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2653 * dwarf2out.c (const_ok_for_output_1): Reject expressions
2656 2014-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2659 * config/arm/cortex-a53.md (cortex_a53_fdivs): Reduce reservation
2661 (cortex_a53_fdivd): Likewise.
2663 2014-04-04 Martin Jambor <mjambor@suse.cz>
2666 * cgraph.h (cgraph_clone_node): New parameter added to declaration.
2668 * cgraph.c (clone_of_p): Also return true if thunks match.
2669 (verify_edge_corresponds_to_fndecl): Removed extraneous call to
2670 cgraph_function_or_thunk_node and an obsolete comment.
2671 * cgraphclones.c (build_function_type_skip_args): Moved upwards in the
2673 (build_function_decl_skip_args): Likewise.
2674 (set_new_clone_decl_and_node_flags): New function.
2675 (duplicate_thunk_for_node): Likewise.
2676 (redirect_edge_duplicating_thunks): Likewise.
2677 (cgraph_clone_node): New parameter args_to_skip, pass it to
2678 redirect_edge_duplicating_thunks which is called instead of
2679 cgraph_redirect_edge_callee.
2680 (cgraph_create_virtual_clone): Pass args_to_skip to cgraph_clone_node,
2681 moved setting of a lot of flags to set_new_clone_decl_and_node_flags.
2683 2014-04-04 Jeff Law <law@redhat.com>
2686 * config/arm/predicates.md (const_int_I_operand): New predicate.
2687 (const_int_M_operand): Similarly.
2688 * config/arm/arm.md (insv_zero): Use const_int_M_operand instead of
2690 (insv_t2, extv_reg, extzv_t2): Likewise.
2691 (load_multiple_with_writeback): Similarly for const_int_I_operand.
2692 (pop_multiple_with_writeback_and_return): Likewise.
2693 (vfp_pop_multiple_with_writeback): Likewise
2695 2014-04-04 Richard Biener <rguenther@suse.de>
2698 * tree-ssanames.c (make_ssa_name_fn): Fix assert.
2699 * gimple.c (gimple_set_bb): Avoid ICEing for NULL cfun for
2701 * gimplify.h (gimple_add_tmp_var_fn): Declare.
2702 * gimplify.c (gimple_add_tmp_var_fn): New function.
2703 * gimple-expr.h (create_tmp_reg_fn): Declare.
2704 * gimple-expr.c (create_tmp_reg_fn): New function.
2705 * gimple-low.c (record_vars_into): Don't change cfun.
2706 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix
2707 code generation without cfun.
2709 2014-04-04 Thomas Schwinge <thomas@codesourcery.com>
2712 * Makefile.in (install-driver): Fix shell scripting.
2714 2014-04-03 Cong Hou <congh@google.com>
2716 PR tree-optimization/60505
2717 * tree-vectorizer.h (struct _stmt_vec_info): Add th field as the
2718 threshold of number of iterations below which no vectorization
2720 * tree-vect-loop.c (new_loop_vec_info):
2721 Initialize LOOP_VINFO_COST_MODEL_THRESHOLD.
2722 * tree-vect-loop.c (vect_analyze_loop_operations):
2723 Set LOOP_VINFO_COST_MODEL_THRESHOLD.
2724 * tree-vect-loop.c (vect_transform_loop):
2725 Use LOOP_VINFO_COST_MODEL_THRESHOLD.
2726 * tree-vect-loop.c (vect_analyze_loop_2): Check the maximum number
2727 of iterations of the loop and see if we should build the epilogue.
2729 2014-04-03 Richard Biener <rguenther@suse.de>
2731 * tree-streamer.h (struct streamer_tree_cache_d): Add next_idx member.
2732 (streamer_tree_cache_create): Adjust.
2733 * tree-streamer.c (streamer_tree_cache_add_to_node_array): Adjust
2734 to allow optional nodes array.
2735 (streamer_tree_cache_insert_1): Use next_idx to assign idx.
2736 (streamer_tree_cache_append): Likewise.
2737 (streamer_tree_cache_create): Create nodes array optionally
2738 as specified by parameter.
2739 * lto-streamer-out.c (create_output_block): Avoid maintaining
2740 the node array in the writer cache.
2741 (DFS_write_tree): Remove assertion.
2742 (produce_asm_for_decls): Free the out decl state hash table early.
2743 * lto-streamer-in.c (lto_data_in_create): Adjust for
2744 streamer_tree_cache_create prototype change.
2746 2014-04-03 Richard Biener <rguenther@suse.de>
2748 * tree-streamer-out.c (streamer_write_chain): Do not temporarily
2749 set TREE_CHAIN to NULL_TREE.
2751 2014-04-03 Richard Biener <rguenther@suse.de>
2753 PR tree-optimization/60740
2754 * graphite-scop-detection.c (stmt_simple_for_scop_p): Iterate
2755 over all GIMPLE_COND operands.
2757 2014-04-03 Nathan Sidwell <nathan@codesourcery.com>
2759 * doc/invoke.texi (Wnon-virtual-dtor): Adjust documentation.
2760 (Weffc++): Remove Scott's numbering, merge lists and reference
2763 2014-04-03 Nick Clifton <nickc@redhat.com>
2765 * config/rl78/rl78-expand.md (movqi): Handle (SUBREG (SYMBOL_REF))
2768 2014-04-03 Martin Jambor <mjambor@suse.cz>
2770 * ipa-cp.c (ipcp_verify_propagated_values): Also dump symtab and
2771 mention gcc_unreachable before failing.
2772 * ipa.c (symtab_remove_unreachable_nodes): Also print order of
2775 2014-04-02 Jan Hubicka <hubicka@ucw.cz>
2778 * ipa-devirt.c (get_polymorphic_call_info): Do not ICE on type
2779 inconsistent code and instead mark the context inconsistent.
2780 (possible_polymorphic_call_targets): For inconsistent contexts
2781 return empty complete list.
2783 2014-04-02 Anthony Green <green@moxielogic.com>
2785 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2)
2786 (extendqisi2, extendhisi2): Define.
2787 * config/moxie/moxie.h (DEFAULT_SIGNED_CHAR): Change to 0.
2788 (WCHAR_TYPE): Change to unsigned int.
2790 2014-04-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2792 PR tree-optimization/60733
2793 * gimple-ssa-strength-reduction.c (ncd_with_phi): Change required
2794 insertion point for PHI candidates to be the end of the feeding
2795 block for the PHI argument.
2797 2014-04-02 Vladimir Makarov <vmakarov@redhat.com>
2799 PR rtl-optimization/60650
2800 * lra-constraints.c (process_alt_operands): Decrease reject for
2801 earlyclobber matching.
2803 2014-04-02 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2805 * config/s390/s390.c (s390_expand_insv): Use GET_MODE_BITSIZE.
2807 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
2809 * config/spu/spu.c (pad_bb): Do not crash when the last
2810 insn is CODE_FOR_blockage.
2812 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
2814 * config/spu/spu.md ("insv"): Fail if bitoffset+bitsize
2815 lies outside the target mode.
2817 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
2820 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
2821 software floating point or no floating point registers, do not
2822 allow any type in the FPRs. Eliminate a test for SPE SIMD types
2823 in GPRs that occurs after we tested for GPRs that would never be
2826 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
2827 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
2828 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
2829 specifically allow DDmode, since that does not use the SPE SIMD
2832 2014-04-02 Richard Biener <rguenther@suse.de>
2835 * optabs.c (expand_abs_nojump): Honor flag_trapv only for
2836 MODE_INTs. Properly use negv_optab.
2837 (expand_abs): Likewise.
2839 2014-04-02 Richard Biener <rguenther@suse.de>
2842 * Makefile.in (install-driver): Guard extra installs with special
2845 2014-04-01 Michael Meissner <meissner@linux.vnet.ibm.com>
2847 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
2850 2014-04-01 Richard Henderson <rth@redhat.com>
2853 * config/i386/i386.md (*float<SWI48><MODEF>2_sse): Leave the second
2854 alternative enabled before register allocation.
2856 2014-04-01 Chung-Lin Tang <cltang@codesourcery.com>
2858 * config/nios2/nios2.md (unspec): Remove UNSPEC_TLS, UNSPEC_TLS_LDM.
2859 * config/nios2/nios2.c (nios2_function_profiler): Fix addi operand
2861 (nios2_large_got_address): Remove unneeded 'sym' parameter.
2862 (nios2_got_address): Update nios2_large_got_address call site.
2863 (nios2_delegitimize_address): New function.
2864 (TARGET_DELEGITIMIZE_ADDRESS): Define to nios2_delegitimize_address.
2865 * config/nios2/linux.h (GLIBC_DYNAMIC_LINKER): Define.
2866 (LINK_SPEC): Specify dynamic linker using GNU_USER_DYNAMIC_LINKER.
2868 2014-04-01 Martin Husemann <martin@duskware.de>
2870 * config/mips/netbsd.h (TARGET_OS_CPP_BUILTINS): Define __mips_o32
2873 2014-04-01 Richard Sandiford <rdsandiford@googlemail.com>
2875 PR rtl-optimization/60604
2876 * recog.c (general_operand): Incorporate REG_CANNOT_CHANGE_MODE_P
2877 check from register_operand.
2878 (register_operand): Redefine in terms of general_operand.
2879 (nonmemory_operand): Use register_operand for the non-constant cases.
2881 2014-04-01 Richard Biener <rguenther@suse.de>
2883 * gimple.h (struct gimple_statement_base): Align subcode to 16 bits.
2885 2014-04-01 Sebastian Huber <sebastian.huber@embedded-brains.de>
2887 * doc/invoke.texi (mapp-regs): Clarify.
2889 2014-03-31 Ulrich Drepper <drepper@gmail.com>
2891 * config/i386/avx512fintrin.h (__v32hi): Define type.
2892 (__v64qi): Likewise.
2893 (_mm512_set1_epi8): Define.
2894 (_mm512_set1_epi16): Define.
2895 (_mm512_set4_epi32): Define.
2896 (_mm512_set4_epi64): Define.
2897 (_mm512_set4_pd): Define.
2898 (_mm512_set4_ps): Define.
2899 (_mm512_setr4_epi64): Define.
2900 (_mm512_setr4_epi32): Define.
2901 (_mm512_setr4_pd): Define.
2902 (_mm512_setr4_ps): Define.
2903 (_mm512_setzero_epi32): Define.
2905 2014-03-31 Martin Jambor <mjambor@suse.cz>
2908 * tree-sra.c (callsite_has_enough_arguments_p): Renamed to
2909 callsite_arguments_match_p. Updated all callers. Also check types of
2910 corresponding formal parameters and actual arguments.
2911 (not_all_callers_have_enough_arguments_p) Renamed to
2912 some_callers_have_mismatched_arguments_p.
2914 2014-03-31 Yuri Rumyantsev <ysrumyan@gmail.com>
2916 * tree-inline.c (copy_loops): Add missed copy of 'safelen'.
2918 2014-03-31 Kugan Vivekanandarajah <kuganv@linaro.org>
2921 * aarch64/aarch64.c (aarch64_classify_address): Fix alignment for
2924 2014-03-30 Uros Bizjak <ubizjak@gmail.com>
2926 * config/i386/sse.md (FMAMODE_NOVF512): New mode iterator.
2927 (<sd_mask_codefor>fma_fmadd_<mode><sd_maskz_name><round_name>):
2929 <sd_mask_codefor>fma_fmadd_<VF_512:mode><sd_maskz_name><round_name>.
2930 Use FMAMODE_NOVF512 mode iterator.
2931 (<sd_mask_codefor>fma_fmsub_<mode><sd_maskz_name><round_name>): Ditto.
2932 (<sd_mask_codefor>fma_fnmadd_<mode><sd_maskz_name><round_name>): Ditto.
2933 (<sd_mask_codefor>fma_fnmsub_<mode><sd_maskz_name><round_name>): Ditto.
2934 (<sd_mask_codefor>fma_fmaddsub_<mode><sd_maskz_name><round_name>):
2936 <sd_mask_codefor>fma_fmaddsub_<VF_512:mode><sd_maskz_name><round_name>.
2937 Use VF_128_256 mode iterator.
2938 (<sd_mask_codefor>fma_fmsubadd_<mode><sd_maskz_name><round_name>):
2941 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
2943 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Clear
2944 static chain if needed.
2946 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
2949 * lra-constraints.c (index_part_to_reg): New.
2950 (process_address): Use it.
2952 2014-03-27 Jeff Law <law@redhat.com>
2953 Jakub Jelinek <jakub@redhat.com>
2956 * expr.c (do_tablejump): Use simplify_gen_binary rather than
2957 gen_rtx_{PLUS,MULT} to build up the address expression.
2959 * i386/i386.c (ix86_legitimize_address): Use copy_addr_to_reg to avoid
2960 creating non-canonical RTL.
2962 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
2965 * ipa-inline.c (want_inline_small_function_p): Short circuit large
2966 functions; reorganize to make cheap checks first.
2967 (inline_small_functions): Do not estimate growth when dumping;
2969 * ipa-inline.h (inline_summary): Add min_size.
2970 (growth_likely_positive): New function.
2971 * ipa-inline-analysis.c (dump_inline_summary): Add min_size.
2972 (set_cond_stmt_execution_predicate): Cleanup.
2973 (estimate_edge_size_and_time): Compute min_size.
2974 (estimate_calls_size_and_time): Likewise.
2975 (estimate_node_size_and_time): Likewise.
2976 (inline_update_overall_summary): Update min_size.
2977 (do_estimate_edge_time): Likewise.
2978 (do_estimate_edge_size): Update.
2979 (do_estimate_edge_hints): Update.
2980 (growth_likely_positive): New function.
2982 2014-03-28 Jakub Jelinek <jakub@redhat.com>
2985 * config/i386/i386.c (ix86_copy_addr_to_reg): Call copy_addr_to_reg
2986 also if addr has VOIDmode.
2988 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2990 * config/arm/aarch-common.c (aarch_crypto_can_dual_issue): New.
2991 * config/arm/aarch-common-protos.h (aarch_crypto_can_dual_issue):
2993 * config/arm/cortex-a53.md: Add reservations and bypass for crypto
2994 instructions as well as AdvancedSIMD loads.
2996 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2998 * config/aarch64/aarch64-simd.md (aarch64_crypto_aes<aes_op>v16qi):
2999 Use crypto_aese type.
3000 (aarch64_crypto_aes<aesmc_op>v16qi): Use crypto_aesmc type.
3001 * config/arm/arm.md (is_neon_type): Replace crypto_aes with
3002 crypto_aese, crypto_aesmc. Move to types.md.
3003 * config/arm/types.md (crypto_aes): Split into crypto_aese,
3005 * config/arm/iterators.md (crypto_type): Likewise.
3007 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
3009 * cgraph.c: Include expr.h and tree-dfa.h.
3010 (cgraph_redirect_edge_call_stmt_to_callee): If call in noreturn;
3013 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
3016 * lra-assigns.c (find_hard_regno_for): Remove unavailable hard
3017 regs from checking multi-reg pseudos.
3019 2014-03-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3021 * config/arm/t-aprofile (MULTILIB_MATCHES): Correct A12 rule.
3023 2014-03-28 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
3025 * config/rs6000/rs6000.c (fusion_gpr_load_p): Refuse optimization
3026 if it would clobber the stack pointer, even temporarily.
3028 2014-03-28 Eric Botcazou <ebotcazou@adacore.com>
3030 * mode-switching.c: Make small adjustments to the top comment.
3032 2014-03-27 Michael Meissner <meissner@linux.vnet.ibm.com>
3034 * config/rs6000/constraints.md (wD constraint): New constraint to
3035 match the constant integer to get the top DImode/DFmode out of a
3036 vector in a VSX register.
3038 * config/rs6000/predicates.md (vsx_scalar_64bit): New predicate to
3039 match the constant integer to get the top DImode/DFmode out of a
3040 vector in a VSX register.
3042 * config/rs6000/rs6000-builtins.def (VBPERMQ): Add vbpermq builtin
3045 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
3048 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
3049 -mdebug=reg, print value of VECTOR_ELEMENT_SCALAR_64BIT.
3051 * config/rs6000/vsx.md (vsx_extract_<mode>, V2DI/V2DF modes):
3052 Optimize vec_extract of 64-bit values, where the value being
3053 extracted is in the top word, where we can use scalar
3054 instructions. Add direct move and store support. Combine the big
3055 endian/little endian vector select load support into a single insn.
3056 (vsx_extract_<mode>_internal1): Likewise.
3057 (vsx_extract_<mode>_internal2): Likewise.
3058 (vsx_extract_<mode>_load): Likewise.
3059 (vsx_extract_<mode>_store): Likewise.
3060 (vsx_extract_<mode>_zero): Delete, big and little endian insns are
3061 combined into vsx_extract_<mode>_load.
3062 (vsx_extract_<mode>_one_le): Likewise.
3064 * config/rs6000/rs6000.h (VECTOR_ELEMENT_SCALAR_64BIT): Macro to
3065 define the top 64-bit vector element.
3067 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wD
3070 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
3071 Document vec_vbpermq builtin.
3074 * config/rs6000/altivec.h (vec_xxsldwi): Add missing define to
3075 enable use of xxsldwi and xxpermdi builtin functions.
3076 (vec_xxpermdi): Likewise.
3078 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
3079 Document use of vec_xxsldwi and vec_xxpermdi builtins.
3081 2014-03-27 Vladimir Makarov <vmakarov@redhat.com>
3083 PR rtl-optimization/60650
3084 * lra-assign.c (find_hard_regno_for, spill_for): Add parameter
3086 (find_spills_for): New.
3087 (assign_by_spills): Pass the new parameter to find_hard_regno_for.
3088 Spill all pseudos on the second iteration.
3090 2014-03-27 Marek Polacek <polacek@redhat.com>
3093 * doc/extend.texi (ffs Builtins): Change unsigned types to signed
3096 2014-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3098 * config/s390/s390.c (s390_can_use_return_insn): Check for
3099 call-saved FPRs on 31 bit.
3101 2014-03-27 Jakub Jelinek <jakub@redhat.com>
3104 * omp-low.c (lower_omp_1): For gimple_clobber_p stmts,
3105 if they need regimplification, just drop them instead of
3106 calling gimple_regimplify_operands on them.
3108 2014-03-27 Marcus Shawcroft <marcus.shawcroft@arm.com>
3111 * config/aarch64/aarch64.c (faked_omit_frame_pointer): Remove.
3112 (aarch64_frame_pointer_required): Adjust logic.
3113 (aarch64_can_eliminate): Adjust logic.
3114 (aarch64_override_options_after_change): Adjust logic.
3116 2014-03-27 Dehao Chen <dehao@google.com>
3118 * ipa-inline.c (early_inliner): Update node's inline info.
3120 2014-03-26 Dehao Chen <dehao@google.com>
3122 * dojump.c (do_compare_rtx_and_jump): Sets correct probability for
3123 compiler inserted conditional jumps for NAN float check.
3125 2014-03-26 Jakub Jelinek <jakub@redhat.com>
3127 * ubsan.h (ubsan_create_data): Change second argument's type
3128 to const location_t *.
3129 * ubsan.c (ubsan_source_location): If xloc.file is NULL, set it to
3131 (ubsan_create_data): Change second argument to const location_t *PLOC.
3132 Create Loc field whenever PLOC is non-NULL.
3133 (ubsan_instrument_unreachable, ubsan_expand_null_ifn,
3134 ubsan_build_overflow_builtin, instrument_bool_enum_load): Adjust
3138 * real.c (real_to_integer2): Change type of low to UHWI.
3140 2014-03-26 Tobias Burnus <burnus@net-b.de>
3142 * gcc.c (LINK_COMMAND_SPEC): Use libcilkrts.spec for -fcilkplus.
3143 (CILK_SELF_SPECS): New define.
3144 (driver_self_specs): Use it.
3146 2014-03-26 Richard Biener <rguenther@suse.de>
3148 * tree-pretty-print.c (percent_K_format): Implement special
3149 case for LTO and its stripped down BLOCK tree.
3151 2014-03-26 Jakub Jelinek <jakub@redhat.com>
3154 * ubsan.c (instrument_si_overflow): Instrument ABS_EXPR.
3156 * tree-vrp.c (simplify_internal_call_using_ranges): If only
3157 one range is range_int_cst_p, but not both, at least optimize
3158 addition/subtraction of 0 and multiplication by 0 or 1.
3159 * gimple-fold.c (gimple_fold_call): Fold
3160 IFN_UBSAN_CHECK_{ADD,SUB,MUL}.
3161 (gimple_fold_stmt_to_constant_1): If both op0 and op1 aren't
3162 INTEGER_CSTs, try to fold at least x * 0 and y - y.
3164 2014-03-26 Eric Botcazou <ebotcazou@adacore.com>
3166 PR rtl-optimization/60452
3167 * rtlanal.c (rtx_addr_can_trap_p_1): Fix head comment.
3168 <case REG>: Return 1 for invalid offsets from the frame pointer.
3170 2014-03-26 Marek Polacek <polacek@redhat.com>
3173 * doc/extend.texi (C Extensions): Mention variable-length arrays in
3176 2014-03-26 Marek Polacek <polacek@redhat.com>
3179 * doc/extend.texi (Designated Inits): Describe what happens to omitted
3182 2014-03-26 Marek Polacek <polacek@redhat.com>
3185 * ira-color.c (update_conflict_hard_regno_costs): Perform the
3186 multiplication in unsigned type.
3188 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
3190 * doc/install.texi: Document nds32le-*-elf and nds32be-*-elf.
3192 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
3194 * doc/contrib.texi: Add myself as Andes nds32 port contributor.
3196 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
3199 * cif-code.def (UNREACHABLE) New code.
3200 * ipa-inline.c (inline_small_functions): Skip edges to
3201 __builtlin_unreachable.
3202 (estimate_edge_growth): Allow edges to __builtlin_unreachable.
3203 * ipa-inline-analysis.c (edge_set_predicate): Redirect edges with false
3204 predicate to __bulitin_unreachable.
3205 (set_cond_stmt_execution_predicate): Fix issue when
3206 invert_tree_comparison returns ERROR_MARK.
3207 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow): Do not
3208 propagate to inline clones.
3209 * cgraph.c (verify_edge_corresponds_to_fndecl): Allow redirection
3211 * ipa-cp.c (create_specialized_node): Be ready for new node to appear.
3212 * cgraphclones.c (cgraph_clone_node): If call destination is already
3213 ureachable, do not redirect it back.
3214 * tree-inline.c (fold_marked_statements): Hanlde calls becoming
3217 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
3219 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow):
3220 Do not modify inline clones.
3222 2014-03-25 Jakub Jelinek <jakub@redhat.com>
3224 * config/i386/i386.md (general_sext_operand): New mode attr.
3225 (addv<mode>4, subv<mode>4, mulv<mode>4): If operands[2] is CONST_INT,
3226 don't generate (sign_extend (const_int)).
3227 (*addv<mode>4, *subv<mode>4, *mulv<mode>4): Disallow CONST_INT_P
3228 operands[2]. Use We constraint instead of <i> and
3229 <general_sext_operand> predicate instead of <general_operand>.
3230 (*addv<mode>4_1, *subv<mode>4_1, *mulv<mode>4_1): New insns.
3231 * config/i386/constraints.md (We): New constraint.
3232 * config/i386/predicates.md (x86_64_sext_operand,
3233 sext_operand): New predicates.
3235 2014-03-25 Martin Jambor <mjambor@suse.cz>
3238 * ipa-cp.c (ipa_get_indirect_edge_target_1): Redirect type
3239 inconsistent devirtualizations to __builtin_unreachable.
3241 2014-03-25 Marek Polacek <polacek@redhat.com>
3244 * doc/extend.texi (Example of asm with clobbered asm reg): Fix typo.
3246 2014-03-25 Alan Lawrence <alan.lawrence@arm.com>
3248 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Reverse
3249 order of elements for big-endian.
3251 2014-03-25 Richard Biener <rguenther@suse.de>
3254 * gimplify-me.c (gimple_regimplify_operands): Update the
3257 2014-03-25 Martin Jambor <mjambor@suse.cz>
3260 * lto-cgraph.c (lto_output_node): Stream body_removed flag.
3261 (lto_output_varpool_node): Likewise.
3262 (input_overwrite_node): Likewise.
3263 (input_varpool_node): Likewise.
3265 2014-03-25 Richard Biener <rguenther@suse.de>
3267 * lto-wrapper.c (merge_and_complain): Handle OPT_fPIE like OPT_fpie.
3268 (run_gcc): Likewise.
3270 2014-03-25 Jakub Jelinek <jakub@redhat.com>
3272 * combine.c (simplify_compare_const): Add MODE argument.
3273 Handle mode_width 0 as very large mode_width.
3274 (try_combine, simplify_comparison): Adjust callers.
3276 * cselib.c (cselib_hash_rtx): Perform addition in unsigned
3277 type to avoid signed integer overflow.
3278 * explow.c (plus_constant): Likewise.
3280 2014-03-25 Dominik Vogt <vogt@linux.vnet.ibm.com>
3282 * doc/generic.texi: Correct typos.
3284 2014-03-24 Tobias Burnus <burnus@net-b.de>
3286 * doc/invoke.texi (-flto): Expand section about
3287 using static libraries with LTO.
3289 2014-03-24 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3291 PR rtl-optimization/60501
3292 * optabs.def (addptr3_optab): New optab.
3293 * optabs.c (gen_addptr3_insn, have_addptr3_insn): New function.
3294 * doc/md.texi ("addptrm3"): Document new RTL standard expander.
3295 * expr.h (gen_addptr3_insn, have_addptr3_insn): Add prototypes.
3297 * lra.c (emit_add3_insn): Use the addptr pattern if available.
3299 * config/s390/s390.md ("addptrdi3", "addptrsi3"): New expanders.
3301 2014-03-24 Ulrich Drepper <drepper@gmail.com>
3303 * config/i386/avx512fintrin.h: Define _mm512_set1_ps and
3306 * config/i386/avxintrin.h (_mm256_undefined_si256): Define.
3307 (_mm256_undefined_ps): Define.
3308 (_mm256_undefined_pd): Define.
3309 * config/i386/emmintrin.h (_mm_undefined_si128): Define.
3310 (_mm_undefined_pd): Define.
3311 * config/i386/xmmintrin.h (_mm_undefined_ps): Define.
3312 * config/i386/avx512fintrin.h (_mm512_undefined_si512): Define.
3313 (_mm512_undefined_ps): Define.
3314 (_mm512_undefined_pd): Define.
3315 Use _mm*_undefined_*.
3316 * config/i386/avx2intrin.h: Use _mm*_undefined_*.
3318 2014-03-24 Alex Velenko <Alex.Velenko@arm.com>
3320 * config/aarch64/aarch64-simd-builtins.def (lshr): DI mode excluded.
3321 (lshr_simd): DI mode added.
3322 * config/aarch64/aarch64-simd.md (aarch64_lshr_simddi): New pattern.
3323 (aarch64_ushr_simddi): Likewise.
3324 * config/aarch64/aarch64.md (UNSPEC_USHR64): New unspec.
3325 * config/aarch64/arm_neon.h (vshr_n_u64): Intrinsic fixed.
3326 (vshrd_n_u64): Likewise.
3328 2014-03-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3330 * Makefile.in (s-macro_list): Depend on cc1.
3332 2014-03-23 Teresa Johnson <tejohnson@google.com>
3334 * ipa-utils.c (ipa_print_order): Use specified dump file.
3336 2014-03-23 Eric Botcazou <ebotcazou@adacore.com>
3338 PR rtl-optimization/60601
3339 * bb-reorder.c (fix_up_fall_thru_edges): Test EDGE_FALLTHRU everywhere.
3341 * gcc.c (eval_spec_function): Initialize save_growing_value.
3343 2014-03-22 Jakub Jelinek <jakub@redhat.com>
3346 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): For
3347 code == MINUS_EXPR, never swap op0 with op1.
3349 * toplev.c (init_local_tick): Avoid signed integer multiplication
3351 * genautomata.c (reserv_sets_hash_value): Fix rotate idiom, avoid
3352 shift by first operand's bitsize.
3354 2014-03-21 Jakub Jelinek <jakub@redhat.com>
3357 * config/i386/i386.h (TARGET_64BIT_P): If not TARGET_BI_ARCH,
3359 * config/i386/darwin.h (TARGET_64BIT_P): Redefine to
3360 TARGET_ISA_64BIT_P(x).
3362 2014-03-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3364 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Generate a
3365 pattern for vector nor instead of subtract from splat(-1).
3366 (altivec_expand_vec_perm_const_le): Likewise.
3368 2014-03-21 Richard Henderson <rth@twiddle.net>
3371 * ifcvt.c (dead_or_predicable): Return FALSE if there are any frame
3372 related insns after epilogue_completed.
3374 2014-03-21 Martin Jambor <mjambor@suse.cz>
3377 * cgraph.h (symtab_node): New flag body_removed.
3378 * ipa.c (symtab_remove_unreachable_nodes): Set body_removed flag
3379 when removing bodies.
3380 * symtab.c (dump_symtab_base): Dump body_removed flag.
3381 * cgraph.c (verify_edge_corresponds_to_fndecl): Skip nodes which
3382 had their bodies removed.
3384 2014-03-21 Martin Jambor <mjambor@suse.cz>
3387 * ipa.c (symtab_remove_unreachable_nodes): Clear thunk flag of nodes
3390 2014-03-21 Richard Biener <rguenther@suse.de>
3392 PR tree-optimization/60577
3393 * tree-core.h (struct tree_base): Document nothrow_flag use
3394 in VAR_DECL_NONALIASED.
3395 * tree.h (VAR_DECL_NONALIASED): New.
3396 (may_be_aliased): Adjust.
3397 * coverage.c (build_var): Set VAR_DECL_NONALIASED.
3399 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
3401 * expr.c (expand_expr_real_1): Remove outdated comment.
3403 2014-03-20 Jakub Jelinek <jakub@redhat.com>
3406 * ira.c (adjust_cleared_regs): Call copy_rtx on
3407 *reg_equiv[REGNO (loc)].src_p before passing it to
3408 simplify_replace_fn_rtx.
3411 * config/i386/i386.c (x86_output_mi_thunk): Surround UNSPEC_GOT
3412 into CONST, put pic register as first operand of PLUS. Use
3413 gen_const_mem for both 32-bit and 64-bit PIC got loads.
3415 2014-03-20 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3417 * config/aarch64/aarch64.c (MEMORY_MOVE_COST): Delete.
3419 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
3421 * config/sparc/sparc.c (sparc_do_work_around_errata): Implement work
3422 around for store forwarding issue in the FPU on the UT699.
3423 * config/sparc/sparc.md (in_branch_delay): Return false for single FP
3424 loads and operations if -mfix-ut699 is specified.
3425 (divtf3_hq): Tweak attribute.
3426 (sqrttf2_hq): Likewise.
3428 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
3430 * calls.c (store_one_arg): Remove incorrect const qualification on the
3431 type of the temporary.
3432 * cfgexpand.c (expand_return): Likewise.
3433 * expr.c (expand_constructor): Likewise.
3434 (expand_expr_real_1): Likewise.
3436 2014-03-20 Zhenqiang Chen <zhenqiang.chen@linaro.org>
3438 * config/arm/arm.c (arm_dwarf_register_span): Update the element number
3441 2014-03-19 Kaz Kojima <kkojima@gcc.gnu.org>
3444 * config/sh/sh.md (udivsi3_i1): Clobber R1 register.
3446 2014-03-19 James Greenhalgh <james.greenhalgh@arm.com>
3448 * config/arm/aarch-common-protos.h
3449 (alu_cost_table): Fix spelling of "extend".
3450 * config/arm/arm.c (arm_new_rtx_costs): Fix spelling of "extend".
3452 2014-03-19 Richard Biener <rguenther@suse.de>
3455 * tree-core.h (tree_type_common): Re-order pointer members
3456 to reduce recursion depth during GC walks.
3458 2014-03-19 Marek Polacek <polacek@redhat.com>
3461 * ubsan.c (ubsan_type_descriptor): Check that DECL_NAME is nonnull
3462 before accessing it.
3464 2014-03-19 Richard Biener <rguenther@suse.de>
3467 * lto-streamer-in.c (input_function): In WPA stage do not drop
3470 2014-03-19 Jakub Jelinek <jakub@redhat.com>
3472 PR tree-optimization/60559
3473 * vectorizable_mask_load_store): Replace scalar MASK_LOAD
3474 with build_zero_cst assignment.
3476 2014-03-18 Kai Tietz <ktietz@redhat.com>
3478 PR rtl-optimization/56356
3479 * sdbout.c (sdbout_parms): Verify that parms'
3480 incoming argument is valid.
3481 (sdbout_reg_parms): Likewise.
3483 2014-03-18 Richard Henderson <rth@redhat.com>
3486 * config/i386/i386.md (*float<SWI48x><MODEF>2_i387): Move down to
3487 be shadowed by *float<SWI48><MODEF>2_sse. Test X87_ENABLE_FLOAT.
3488 (*float<SWI48><MODEF>2_sse): Check X87_ENABLE_FLOAT for alternative 0.
3490 2014-03-18 Basile Starynkevitch <basile@starynkevitch.net>
3492 * plugin.def: Improve comment for PLUGIN_INCLUDE_FILE.
3493 * doc/plugins.texi (Plugin callbacks): Mention PLUGIN_INCLUDE_FILE.
3494 Italicize plugin event names in description. Explain that
3495 PLUGIN_PRAGMAS has no sense for lto1. Explain PLUGIN_INCLUDE_FILE.
3496 Remind that no GCC functions should be called after PLUGIN_FINISH.
3497 Explain what pragmas with expansion are.
3499 2014-03-18 Martin Liska <mliska@suse.cz>
3501 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Added case when
3502 gimple call statement is update.
3503 * gimple-fold.c (gimple_fold_call): Changed order for GIMPLE_ASSIGN and
3504 GIMPLE_CALL, where gsi iterator still points to GIMPLE CALL.
3506 2014-03-18 Jakub Jelinek <jakub@redhat.com>
3509 * ubsan.c (ubsan_instrument_unreachable): Call
3510 initialize_sanitizer_builtins.
3511 (ubsan_pass): Likewise.
3514 * ubsan.c (ubsan_type_descriptor, ubsan_create_data): Call
3515 varpool_finalize_decl instead of rest_of_decl_compilation.
3517 2014-03-18 Richard Biener <rguenther@suse.de>
3519 * df-problems.c (df_rd_confluence_n): Avoid bitmap_copy
3520 by using bitmap_and_compl instead of bitmap_and_compl_into.
3521 (df_rd_transfer_function): Likewise.
3523 2014-03-18 Richard Biener <rguenther@suse.de>
3525 * doc/lto.texi (fresolution): Fix typo.
3527 2014-03-18 Richard Biener <rguenther@suse.de>
3529 * doc/invoke.texi (flto): Update for changes in 4.9.
3531 2014-03-18 Richard Biener <rguenther@suse.de>
3533 * doc/loop.texi: Remove section on the removed lambda framework.
3534 Update loop docs with recent changes in preserving loop structure.
3536 2014-03-18 Richard Biener <rguenther@suse.de>
3538 * doc/lto.texi (-fresolution): Document.
3540 2014-03-18 Richard Biener <rguenther@suse.de>
3542 * doc/contrib.texi: Adjust my name.
3544 2014-03-18 Jakub Jelinek <jakub@redhat.com>
3547 * internal-fn.c: Include diagnostic-core.h.
3548 (expand_BUILTIN_EXPECT): New function.
3549 * gimplify.c (gimplify_call_expr): Use false instead of FALSE.
3550 (gimplify_modify_expr): Gimplify 3 argument __builtin_expect into
3551 IFN_BUILTIN_EXPECT call instead of __builtin_expect builtin call.
3552 * ipa-inline-analysis.c (find_foldable_builtin_expect): Handle
3554 * predict.c (expr_expected_value_1): Handle IFN_BUILTIN_EXPECT.
3555 Revert 3 argument __builtin_expect code.
3556 (strip_predict_hints): Handle IFN_BUILTIN_EXPECT.
3557 * gimple-fold.c (gimple_fold_call): Likewise.
3558 * tree.h (fold_builtin_expect): New prototype.
3559 * builtins.c (build_builtin_expect_predicate): Add predictor
3560 argument, if non-NULL, create 3 argument __builtin_expect.
3561 (fold_builtin_expect): No longer static. Add ARG2 argument,
3562 pass it through to build_builtin_expect_predicate.
3563 (fold_builtin_2): Adjust caller.
3564 (fold_builtin_3): Handle BUILT_IN_EXPECT.
3565 * internal-fn.def (BUILTIN_EXPECT): New.
3567 2014-03-18 Tobias Burnus <burnus@net-b.de>
3570 * predict.def (PRED_FORTRAN_OVERFLOW, PRED_FORTRAN_FAIL_ALLOC,
3571 PRED_FORTRAN_FAIL_IO, PRED_FORTRAN_WARN_ONCE, PRED_FORTRAN_SIZE_ZERO,
3572 PRED_FORTRAN_INVALID_BOUND, PRED_FORTRAN_ABSENT_DUMMY): Add.
3574 2014-03-18 Jan Hubicka <hubicka@ucw.cz>
3577 * predict.c (combine_predictions_for_bb): Fix up formatting.
3578 (expr_expected_value_1, expr_expected_value): Add predictor argument,
3579 fill what it points to if non-NULL.
3580 (tree_predict_by_opcode): Adjust caller, use the predictor.
3581 * predict.def (PRED_COMPARE_AND_SWAP): Add.
3583 2014-03-18 Eric Botcazou <ebotcazou@adacore.com>
3585 * config/sparc/sparc.c (sparc_do_work_around_errata): Speed up and use
3586 proper constant for the store mode.
3588 2014-03-18 Ilya Enkovich <ilya.enkovich@intel.com>
3590 * symtab.c (change_decl_assembler_name): Fix transparent alias
3593 2014-03-16 Renlin Li <Renlin.Li@arm.com>
3595 * config/aarch64/aarch64.c: Correct the comments about the
3596 aarch64 stack layout.
3598 2014-03-18 Thomas Schwinge <thomas@codesourcery.com>
3600 * omp-low.c (lower_rec_input_clauses) <build_omp_barrier>: Restore
3601 check for GF_OMP_FOR_KIND_FOR.
3603 2013-03-18 Kirill Yukhin <kirill.yukhin@intel.com>
3605 * config/i386/i386.h (ADDITIONAL_REGISTER_NAMES): Add
3606 ymm and zmm register names.
3608 2014-03-17 Jakub Jelinek <jakub@redhat.com>
3611 * config/i386/i386.c (ix86_expand_epilogue): Adjust REG_CFA_ADJUST_CFA
3612 note creation for the 2010-08-31 changes.
3614 2014-03-17 Marek Polacek <polacek@redhat.com>
3617 * omp-low.c (omp_max_vf): Treat -fno-tree-loop-optimize the same
3618 as -fno-tree-loop-vectorize.
3619 (expand_omp_simd): Likewise.
3621 2014-03-15 Eric Botcazou <ebotcazou@adacore.com>
3623 * config/sparc/sparc-protos.h (tls_call_delay): Delete.
3624 (eligible_for_call_delay): New prototype.
3625 * config/sparc/sparc.c (tls_call_delay): Rename into...
3626 (eligible_for_call_delay): ...this. Return false if the instruction
3627 cannot be put in the delay slot of a branch.
3628 (eligible_for_restore_insn): Simplify.
3629 (eligible_for_return_delay): Return false if the instruction cannot be
3630 put in the delay slot of a branch and simplify.
3631 (eligible_for_sibcall_delay): Return false if the instruction cannot be
3632 put in the delay slot of a branch.
3633 * config/sparc/sparc.md (fix_ut699): New attribute.
3634 (tls_call_delay): Delete.
3635 (in_call_delay): Reimplement.
3636 (eligible_for_sibcall_delay): Rename into...
3637 (in_sibcall_delay): ...this.
3638 (eligible_for_return_delay): Rename into...
3639 (in_return_delay): ...this.
3640 (in_branch_delay): Reimplement.
3641 (in_uncond_branch_delay): Delete.
3642 (in_annul_branch_delay): Delete.
3644 2014-03-14 Richard Henderson <rth@redhat.com>
3647 * config/i386/i386.md (floathi<X87MODEF>2): Delete expander; rename
3648 define_insn from *floathi<X87MODEF>2_i387; allow nonimmediate_operand.
3649 (*floathi<X87MODEF>2_i387_with_temp): Remove.
3650 (floathi splitters): Remove.
3651 (float<SWI48x>xf2): New pattern.
3652 (float<SWI48><MODEF>2): Rename from float<SWI48x><X87MODEF>2. Drop
3653 code that tried to handle DImode for 32-bit, but which was excluded
3654 by the pattern's condition. Drop allocation of stack temporary.
3655 (*floatsi<MODEF>2_vector_mixed_with_temp): Remove.
3656 (*float<SWI48><MODEF>2_mixed_with_temp): Remove.
3657 (*float<SWI48><MODEF>2_mixed_interunit): Remove.
3658 (*float<SWI48><MODEF>2_mixed_nointerunit): Remove.
3659 (*floatsi<MODEF>2_vector_sse_with_temp): Remove.
3660 (*float<SWI48><MODEF>2_sse_with_temp): Remove.
3661 (*float<SWI48><MODEF>2_sse_interunit): Remove.
3662 (*float<SWI48><MODEF>2_sse_nointerunit): Remove.
3663 (*float<SWI48x><X87MODEF>2_i387_with_temp): Remove.
3664 (*float<SWI48x><X87MODEF>2_i387): Remove.
3665 (all float _with_temp splitters): Remove.
3666 (*float<SWI48x><MODEF>2_i387): New pattern.
3667 (*float<SWI48><MODEF>2_sse): New pattern.
3668 (float TARGET_USE_VECTOR_CONVERTS splitters): Merge them.
3669 (float TARGET_SSE_PARTIAL_REG_DEPENDENCY splitters): Merge them.
3671 2014-03-14 Jakub Jelinek <jakub@redhat.com>
3672 Marek Polacek <polacek@redhat.com>
3675 * common.opt (dump_base_name_prefixed): New Variable.
3676 * opts.c (finish_options): Don't prepend directory to x_dump_base_name
3677 if x_dump_base_name_prefixed is already set, set it at the end.
3679 2014-03-14 Vladimir Makarov <vmakarov@redhat.com>
3681 PR rtl-optimization/60508
3682 * lra-constraints.c (get_reload_reg): Add new parameter
3684 (process_addr_reg, simplify_operand_subreg, curr_insn_transform):
3685 Pass the new parameter values.
3687 2014-03-14 Richard Biener <rguenther@suse.de>
3689 * common.opt: Revert unintented changes from r205065.
3692 2014-03-14 Richard Biener <rguenther@suse.de>
3695 * cfghooks.c (split_block): Properly adjust all loops the
3696 block was a latch of.
3698 2014-03-14 Martin Jambor <mjambor@suse.cz>
3701 * ipa-prop.c (ipa_modify_call_arguments): Fix iteration condition
3704 2014-03-14 Georg-Johann Lay <avr@gjlay.de>
3707 * config/avr/avr.c (avr_set_current_function): Pass function name
3708 through default_strip_name_encoding before sanity checking instead
3709 of skipping the first char of the assembler name.
3711 2014-03-13 Richard Henderson <rth@redhat.com>
3714 * config/i386/i386.c (ix86_split_fp_branch): Remove pushed argument.
3715 (ix86_force_to_memory, ix86_free_from_memory): Remove.
3716 * config/i386/i386-protos.h: Likewise.
3717 * config/i386/i386.md (floathi<X87MODEF>2): Use assign_386_stack_local
3718 in the expander instead of a splitter.
3719 (float<SWI48x><X87MODEF>2): Use assign_386_stack_local if there is
3720 any possibility of requiring a memory.
3721 (*floatsi<MODEF>2_vector_mixed): Remove, and the splitters.
3722 (*floatsi<MODEF>2_vector_sse): Remove, and the splitters.
3723 (fp branch splitters): Update for ix86_split_fp_branch.
3724 (*jcc<X87MODEF>_<SWI24>_i387): Remove r/f alternative.
3725 (*jcc<X87MODEF>_<SWI24>_r_i387): Likewise.
3726 (splitter for jcc<X87MODEF>_<SWI24>_i387 r/f): Remove.
3727 (*fop_<MODEF>_2_i387): Remove f/r alternative.
3728 (*fop_<MODEF>_3_i387): Likewise.
3729 (*fop_xf_2_i387, *fop_xf_3_i387): Likewise.
3730 (splitters for the fop_* register patterns): Remove.
3731 (fscalexf4_i387): Rename from *fscalexf4_i387.
3732 (ldexpxf3): Use gen_floatsixf2 and gen_fscalexf4_i387.
3734 2014-03-13 Jakub Jelinek <jakub@redhat.com>
3736 PR tree-optimization/59779
3737 * tree-dfa.c (get_ref_base_and_extent): Use double_int
3738 type for bitsize and maxsize instead of HOST_WIDE_INT.
3740 2014-03-13 Steven Bosscher <steven@gcc.gnu.org>
3742 PR rtl-optimization/57320
3743 * function.c (rest_of_handle_thread_prologue_and_epilogue): Cleanup
3744 the CFG after thread_prologue_and_epilogue_insns.
3746 2014-03-13 Vladimir Makarov <vmakarov@redhat.com>
3748 PR rtl-optimization/57189
3749 * lra-constraints.c (process_alt_operands): Disfavor spilling
3752 2014-03-13 Cesar Philippidis <cesar@codesourcery.com>
3754 * lto-wrapper.c (maybe_unlink_file): Suppress diagnostic messages.
3756 2014-03-13 Jakub Jelinek <jakub@redhat.com>
3758 PR tree-optimization/59025
3760 * tree-ssa-reassoc.c (sort_by_operand_rank): For SSA_NAMEs with the
3761 same rank, sort by bb_rank and gimple_uid of SSA_NAME_DEF_STMT first.
3763 2014-03-13 Georg-Johann Lay <avr@gjlay.de>
3766 * config/avr/avr.c (avr_out_plus): Swap cc_plus and cc_minus in
3767 calls of avr_out_plus_1.
3769 2014-03-13 Bin Cheng <bin.cheng@arm.com>
3771 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Record
3772 BB's single pred and update the father loop's latch info later.
3774 2014-03-12 Michael Meissner <meissner@linux.vnet.ibm.com>
3776 * config/rs6000/vector.md (VEC_L): Add V1TI mode to vector types.
3780 (VEC_base): Likewise.
3781 (mov<MODE>, VEC_M modes): If we are loading TImode into VSX
3782 registers, we need to swap double words in little endian mode.
3784 * config/rs6000/rs6000-modes.def (V1TImode): Add new vector mode
3785 to be a container mode for 128-bit integer operations added in ISA
3786 2.07. Unlike TImode and PTImode, the preferred register set is
3787 the Altivec/VMX registers for the 128-bit operations.
3789 * config/rs6000/rs6000-protos.h (rs6000_move_128bit_ok_p): Add
3791 (rs6000_split_128bit_ok_p): Likewise.
3793 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_3): Add new support
3794 macros for creating ISA 2.07 normal and overloaded builtin
3795 functions with 3 arguments.
3796 (BU_P8V_OVERLOAD_3): Likewise.
3797 (VPERM_1T): Add support for V1TImode in 128-bit vector operations
3798 for use as overloaded functions.
3799 (VPERM_1TI_UNS): Likewise.
3800 (VSEL_1TI): Likewise.
3801 (VSEL_1TI_UNS): Likewise.
3802 (ST_INTERNAL_1ti): Likewise.
3803 (LD_INTERNAL_1ti): Likewise.
3804 (XXSEL_1TI): Likewise.
3805 (XXSEL_1TI_UNS): Likewise.
3806 (VPERM_1TI): Likewise.
3807 (VPERM_1TI_UNS): Likewise.
3808 (XXPERMDI_1TI): Likewise.
3809 (SET_1TI): Likewise.
3810 (LXVD2X_V1TI): Likewise.
3811 (STXVD2X_V1TI): Likewise.
3812 (VEC_INIT_V1TI): Likewise.
3813 (VEC_SET_V1TI): Likewise.
3814 (VEC_EXT_V1TI): Likewise.
3815 (EQV_V1TI): Likewise.
3816 (NAND_V1TI): Likewise.
3817 (ORC_V1TI): Likewise.
3818 (VADDCUQ): Add support for 128-bit integer arithmetic instructions
3819 added in ISA 2.07. Add both normal 'altivec' builtins, and the
3821 (VADDUQM): Likewise.
3822 (VSUBCUQ): Likewise.
3823 (VADDEUQM): Likewise.
3824 (VADDECUQ): Likewise.
3825 (VSUBEUQM): Likewise.
3826 (VSUBECUQ): Likewise.
3828 * config/rs6000/rs6000-c.c (__int128_type): New static to hold
3829 __int128_t and __uint128_t types.
3830 (__uint128_type): Likewise.
3831 (altivec_categorize_keyword): Add support for vector __int128_t,
3832 vector __uint128_t, vector __int128, and vector unsigned __int128
3833 as a container type for TImode operations that need to be done in
3834 VSX/Altivec registers.
3835 (rs6000_macro_to_expand): Likewise.
3836 (altivec_overloaded_builtins): Add ISA 2.07 overloaded functions
3837 to support 128-bit integer instructions vaddcuq, vadduqm,
3838 vaddecuq, vaddeuqm, vsubcuq, vsubuqm, vsubecuq, vsubeuqm.
3839 (altivec_resolve_overloaded_builtin): Add support for V1TImode.
3841 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Add support
3842 for V1TImode, and set up preferences to use VSX/Altivec registers.
3843 Setup VSX reload handlers.
3844 (rs6000_debug_reg_global): Likewise.
3845 (rs6000_init_hard_regno_mode_ok): Likewise.
3846 (rs6000_preferred_simd_mode): Likewise.
3847 (vspltis_constant): Do not allow V1TImode as easy altivec constants.
3848 (easy_altivec_constant): Likewise.
3849 (output_vec_const_move): Likewise.
3850 (rs6000_expand_vector_set): Convert V1TImode set and extract to
3852 (rs6000_expand_vector_extract): Likewise.
3853 (reg_offset_addressing_ok_p): Setup V1TImode to use VSX reg+reg
3855 (rs6000_const_vec): Add support for V1TImode.
3856 (rs6000_emit_le_vsx_load): Swap double words when loading or
3857 storing TImode/V1TImode.
3858 (rs6000_emit_le_vsx_store): Likewise.
3859 (rs6000_emit_le_vsx_move): Likewise.
3860 (rs6000_emit_move): Add support for V1TImode.
3861 (altivec_expand_ld_builtin): Likewise.
3862 (altivec_expand_st_builtin): Likewise.
3863 (altivec_expand_vec_init_builtin): Likewise.
3864 (altivec_expand_builtin): Likewise.
3865 (rs6000_init_builtins): Add support for V1TImode type. Add
3866 support for ISA 2.07 128-bit integer builtins. Define type names
3867 for the VSX/Altivec vector types.
3868 (altivec_init_builtins): Add support for overloaded vector
3869 functions with V1TImode type.
3870 (rs6000_preferred_reload_class): Prefer Altivec registers for V1TImode.
3871 (rs6000_move_128bit_ok_p): Move 128-bit move/split validation to
3873 (rs6000_split_128bit_ok_p): Likewise.
3874 (rs6000_handle_altivec_attribute): Create V1TImode from vector
3875 __int128_t and vector __uint128_t.
3877 * config/rs6000/vsx.md (VSX_L): Add V1TImode to vector iterators
3878 and mode attributes.
3885 (VS_scalar): Likewise.
3886 (VS_double): Likewise.
3887 (vsx_set_v1ti): New builtin function to create V1TImode from TImode.
3889 * config/rs6000/rs6000.h (TARGET_VADDUQM): New macro to say whether
3890 we support the ISA 2.07 128-bit integer arithmetic instructions.
3891 (ALTIVEC_OR_VSX_VECTOR_MODE): Add V1TImode.
3892 (enum rs6000_builtin_type_index): Add fields to hold V1TImode
3893 and TImode types for use with the builtin functions.
3894 (V1TI_type_node): Likewise.
3895 (unsigned_V1TI_type_node): Likewise.
3896 (intTI_type_internal_node): Likewise.
3897 (uintTI_type_internal_node): Likewise.
3899 * config/rs6000/altivec.md (UNSPEC_VADDCUQ): New unspecs for ISA 2.07
3900 128-bit builtin functions.
3901 (UNSPEC_VADDEUQM): Likewise.
3902 (UNSPEC_VADDECUQ): Likewise.
3903 (UNSPEC_VSUBCUQ): Likewise.
3904 (UNSPEC_VSUBEUQM): Likewise.
3905 (UNSPEC_VSUBECUQ): Likewise.
3906 (VM): Add V1TImode to vector mode iterators.
3908 (VI_unit): Likewise.
3909 (altivec_vadduqm): Add ISA 2.07 128-bit binary builtins.
3910 (altivec_vaddcuq): Likewise.
3911 (altivec_vsubuqm): Likewise.
3912 (altivec_vsubcuq): Likewise.
3913 (altivec_vaddeuqm): Likewise.
3914 (altivec_vaddecuq): Likewise.
3915 (altivec_vsubeuqm): Likewise.
3916 (altivec_vsubecuq): Likewise.
3918 * config/rs6000/rs6000.md (FMOVE128_GPR): Add V1TImode to vector
3920 (BOOL_128): Likewise.
3921 (BOOL_REGS_OUTPUT): Likewise.
3922 (BOOL_REGS_OP1): Likewise.
3923 (BOOL_REGS_OP2): Likewise.
3924 (BOOL_REGS_UNARY): Likewise.
3925 (BOOL_REGS_AND_CR0): Likewise.
3927 * config/rs6000/altivec.h (vec_vaddcuq): Add support for ISA 2.07
3928 128-bit integer builtin support.
3929 (vec_vadduqm): Likewise.
3930 (vec_vaddecuq): Likewise.
3931 (vec_vaddeuqm): Likewise.
3932 (vec_vsubecuq): Likewise.
3933 (vec_vsubeuqm): Likewise.
3934 (vec_vsubcuq): Likewise.
3935 (vec_vsubuqm): Likewise.
3937 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
3938 Document vec_vaddcuq, vec_vadduqm, vec_vaddecuq, vec_vaddeuqm,
3939 vec_subecuq, vec_subeuqm, vec_vsubcuq, vec_vsubeqm builtins adding
3940 128-bit integer add/subtract to ISA 2.07.
3942 2014-03-12 Joern Rennecke <joern.rennecke@embecosm.com>
3944 * config/arc/arc.c (arc_predicate_delay_insns):
3945 Fix third argument passed to conditionalize_nonjump.
3947 2014-03-12 Yufeng Zhang <yufeng.zhang@arm.com>
3949 * config/aarch64/aarch64-builtins.c
3950 (aarch64_builtin_vectorized_function): Add BUILT_IN_LFLOORF,
3951 BUILT_IN_LLFLOOR, BUILT_IN_LCEILF and BUILT_IN_LLCEIL.
3952 * config/aarch64/arm_neon.h (vcvtaq_u64_f64): Call __builtin_llfloor
3953 instead of __builtin_lfloor.
3954 (vcvtnq_u64_f64): Call __builtin_llceil instead of __builtin_lceil.
3956 2014-03-12 Jakub Jelinek <jakub@redhat.com>
3958 * tree-ssa-ifcombine.c (forwarder_block_to): New function.
3959 (tree_ssa_ifcombine_bb_1): New function.
3960 (tree_ssa_ifcombine_bb): Use it. Handle also cases where else_bb
3961 is an empty forwarder block to then_bb or vice versa and then_bb
3962 and else_bb are effectively swapped.
3964 2014-03-12 Christian Bruel <christian.bruel@st.com>
3967 * config/arm/arm.c (arm_emit_vfp_multi_reg_pop): Emit a
3968 REG_CFA_DEF_CFA note.
3969 (arm_expand_epilogue_apcs_frame): call arm_add_cfa_adjust_cfa_note.
3970 (arm_unwind_emit): Allow REG_CFA_DEF_CFA.
3972 2014-03-12 Thomas Preud'homme <thomas.preudhomme@arm.com>
3974 PR tree-optimization/60454
3975 * tree-ssa-math-opts.c (find_bswap_1): Fix bswap detection.
3977 2014-03-12 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3979 * config.gcc (aarch64*-*-*): Use ISA flags from aarch64-arches.def.
3980 Do not define target_cpu_default2 to generic.
3981 * config/aarch64/aarch64.h (TARGET_CPU_DEFAULT): Use generic cpu.
3982 * config/aarch64/aarch64.c (aarch64_override_options): Update comment.
3983 * config/aarch64/aarch64-arches.def (armv8-a): Use generic cpu.
3985 2014-03-12 Jakub Jelinek <jakub@redhat.com>
3986 Marc Glisse <marc.glisse@inria.fr>
3988 PR tree-optimization/60502
3989 * tree-ssa-reassoc.c (eliminate_not_pairs): Use build_all_ones_cst
3990 instead of build_low_bits_mask.
3992 2014-03-12 Jakub Jelinek <jakub@redhat.com>
3995 * tree-vrp.c (register_edge_assert_for_1): Don't add assert
3996 if there are multiple uses, but op doesn't live on E edge.
3997 * tree-cfg.c (assert_unreachable_fallthru_edge_p): Also ignore
3998 clobber stmts before __builtin_unreachable.
4000 2014-03-11 Richard Sandiford <rdsandiford@googlemail.com>
4002 * builtins.c (expand_builtin_setjmp_receiver): Use and clobber
4003 hard_frame_pointer_rtx.
4004 * cse.c (cse_insn): Remove volatile check.
4005 * cselib.c (cselib_process_insn): Likewise.
4006 * dse.c (scan_insn): Likewise.
4008 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
4010 * config/arc/arc.c (conditionalize_nonjump): New function,
4012 (arc_ifcvt): ... this.
4013 (arc_predicate_delay_insns): Use it.
4015 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
4017 * config/arc/predicates.md (extend_operand): During/after reload,
4018 allow const_int_operand.
4019 * config/arc/arc.md (mulsidi3_700): Use extend_operand predicate.
4020 (umulsidi3_700): Likewise. Change operand 2 constraint back to "cL".
4021 (mulsi3_highpart): Change operand 2 constraint alternatives 2 and 3
4023 (umulsi3_highpart_i): Likewise.
4025 2014-03-11 Richard Biener <rguenther@suse.de>
4027 * tree-ssa-structalias.c (get_constraint_for_ptr_offset):
4028 Add asserts to guard possible wrong-code bugs.
4030 2014-03-11 Richard Biener <rguenther@suse.de>
4032 PR tree-optimization/60429
4033 PR tree-optimization/60485
4034 * tree-ssa-structalias.c (set_union_with_increment): Properly
4035 take into account all fields that overlap the shifted vars.
4036 (do_sd_constraint): Likewise.
4037 (do_ds_constraint): Likewise.
4038 (get_constraint_for_ptr_offset): Likewise.
4040 2014-03-11 Chung-Lin Tang <cltang@codesourcery.com>
4042 * config/nios2/nios2.c (machine_function): Add fp_save_offset field.
4043 (nios2_compute_frame_layout):
4044 Add calculation of cfun->machine->fp_save_offset.
4045 (nios2_expand_prologue): Correct setting of frame pointer register
4047 (nios2_expand_epilogue): Update recovery of stack pointer from
4048 frame pointer accordingly.
4049 (nios2_initial_elimination_offset): Update calculation of offset
4050 for eliminating to HARD_FRAME_POINTER_REGNUM.
4052 2014-03-10 Jakub Jelinek <jakub@redhat.com>
4055 * ipa.c (symtab_remove_unreachable_nodes): Don't call
4056 cgraph_get_create_node on VAR_DECLs.
4058 2014-03-10 Richard Biener <rguenther@suse.de>
4061 * tree.c (signed_or_unsigned_type_for): Handle OFFSET_TYPEs.
4063 2014-03-08 Douglas B Rupp <rupp@gnat.com>
4065 * config/vms/vms.opt (vms_float_format): New variable.
4067 2014-03-08 Tobias Burnus <burnus@net-b.de>
4069 * doc/invoke.texi (-fcilkplus): Update implementation status.
4071 2014-03-08 Paulo Matos <paulo@matos-sorge.com>
4072 Richard Biener <rguenther@suse.de>
4074 * lto-wrapper.c (merge_and_complain): Ensure -fshort-double is used
4075 consistently accross all TUs.
4076 (run_gcc): Enable -fshort-double automatically at link at link-time
4077 and disallow override.
4079 2014-03-08 Richard Sandiford <rdsandiford@googlemail.com>
4082 * config/mips/mips.c (mips_option_override): Promote -mpaired-single
4083 warning to an error. Disable TARGET_PAIRED_SINGLE and TARGET_MIPS3D
4084 if they can't be used.
4086 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4088 * configure.ac (HAVE_AS_IX86_TLSLDMPLT): Improve test
4089 for Solaris 11/x86 ld.
4090 * configure: Regenerate.
4092 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4094 * configure.ac (TLS_SECTION_ASM_FLAG): Save as tls_section_flag.
4095 (LIB_TLS_SPEC): Save as ld_tls_libs.
4096 (HAVE_AS_IX86_TLSLDMPLT): Define as 1/0.
4097 (HAVE_AS_IX86_TLSLDM): New test.
4098 * configure, config.in: Regenerate.
4099 * config/i386/i386.c (legitimize_tls_address): Fall back to
4100 TLS_MODEL_GLOBAL_DYNAMIC on 32-bit Solaris/x86 if tool chain
4101 cannot support TLS_MODEL_LOCAL_DYNAMIC.
4102 * config/i386/i386.md (*tls_local_dynamic_base_32_gnu): Use if
4103 instead of #ifdef in HAVE_AS_IX86_TLSLDMPLT test.
4105 2014-03-07 Paulo Matos <paulo@matos-sorge.com>
4107 * common.opt (fira-loop-pressure): Mark as optimization.
4109 2014-03-07 Thomas Schwinge <thomas@codesourcery.com>
4111 * langhooks.c (lhd_omp_mappable_type): The error_mark_node is not
4112 an OpenMP mappable type.
4114 2014-03-06 Matthias Klose <doko@ubuntu.com>
4116 * Makefile.in (s-mlib): Only pass MULTIARCH_DIRNAME if
4117 MULTILIB_OSDIRNAMES is not defined.
4119 2014-03-06 Jakub Jelinek <jakub@redhat.com>
4120 Meador Inge <meadori@codesourcery.com>
4123 * config/arm/arm.c (arm_tls_symbol_p): Remove.
4124 (arm_legitimize_address): Call legitimize_tls_address for any
4125 arm_tls_referenced_p expression, handle constant addend. Call it
4126 before testing for !TARGET_ARM.
4127 (thumb_legitimize_address): Don't handle arm_tls_symbol_p here.
4129 2014-03-06 Richard Biener <rguenther@suse.de>
4135 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
4137 * tree-streamer.c (record_common_node): Assert we don't record
4138 nodes with type double.
4139 (preload_common_node): Skip type double, complex double and double
4140 pointer since it is now frontend dependent due to fshort-double option.
4142 2014-03-06 Richard Biener <rguenther@suse.de>
4144 * gcc.c (PLUGIN_COND): Always enable unless -fno-use-linker-plugin
4145 or -fno-lto is specified and the linker has full plugin support.
4146 * collect2.c (lto_mode): Default to LTO_MODE_WHOPR if LTO is enabled.
4147 (main): Remove -flto processing, adjust lto_mode using use_plugin late.
4148 * lto-wrapper.c (merge_and_complain): Merge compile-time
4149 optimization levels.
4150 (run_gcc): And pass it through to the link options.
4152 2014-03-06 Alexandre Oliva <aoliva@redhat.com>
4156 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4158 * cselib.c (remove_useless_values): Skip to avoid quadratic
4159 behavior if the condition moved from...
4160 (cselib_process_insn): ... here holds.
4162 2014-03-05 Jakub Jelinek <jakub@redhat.com>
4165 * Makefile.in (PLUGIN_HEADERS): Add tree-phinodes.h, stor-layout.h,
4166 ssa-iterators.h, $(RESOURCE_H) and tree-cfgcleanup.h.
4169 * config/i386/t-i386 (OPTIONS_H_EXTRA): Add stringop.def.
4170 (TM_H): Add x86-tune.def.
4172 2014-03-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4174 * config/aarch64/aarch64.c (generic_tunings):
4175 Use cortexa57_extra_costs.
4177 2014-03-05 Jakub Jelinek <jakub@redhat.com>
4180 * cfgexpand.c (expand_used_vars): Do not assume all SSA_NAMEs
4181 of PARM/RESULT_DECLs must be coalesced with optimize && in_lto_p.
4182 * tree-ssa-coalesce.c (coalesce_ssa_name): Use MUST_COALESCE_COST - 1
4185 2014-03-04 Heiher <r@hev.cc>
4187 * config/mips/mips-cpus.def (loongson3a): Mark as a MIPS64r2 processor.
4188 * config/mips/mips.h (MIPS_ISA_LEVEL_SPEC): Adjust accordingly.
4190 2014-03-04 Uros Bizjak <ubizjak@gmail.com>
4192 * config/i386/predicates.md (const2356_operand): Change to ...
4193 (const2367_operand): ... this.
4194 * config/i386/sse.md (avx512pf_scatterpf<mode>sf): Use
4196 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
4197 (*avx512pf_scatterpf<mode>sf): Ditto.
4198 (avx512pf_scatterpf<mode>df): Ditto.
4199 (*avx512pf_scatterpf<mode>df_mask): Ditto.
4200 (*avx512pf_scatterpf<mode>df): Ditto.
4201 * config/i386/i386.c (ix86_expand_builtin): Update
4202 incorrect hint operand error message.
4204 2014-03-04 Richard Biener <rguenther@suse.de>
4206 * lto-section-in.c (lto_get_section_data): Fix const cast.
4208 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
4210 * tree-streamer.c (record_common_node): Assert we don't record
4211 nodes with type double.
4212 (preload_common_node): Skip type double, complex double and double
4213 pointer since it is now frontend dependent due to fshort-double option.
4215 2014-03-04 Richard Biener <rguenther@suse.de>
4218 * lto-streamer-in.c (lto_read_body): Remove LTO bytecode version check.
4219 (lto_input_toplevel_asms): Likewise.
4220 * lto-section-in.c (lto_get_section_data): Instead do it here
4223 2014-03-04 Richard Biener <rguenther@suse.de>
4225 PR tree-optimization/60382
4226 * tree-vect-loop.c (vect_is_simple_reduction_1): Do not consider
4227 dead PHIs a reduction.
4229 2014-03-03 Uros Bizjak <ubizjak@gmail.com>
4231 * config/i386/xmmintrin.h (enum _mm_hint) <_MM_HINT_ET0>: Correct
4233 (_mm_prefetch): Move out of GCC target("sse") pragma.
4234 * config/i386/prfchwintrin.h (_m_prefetchw): Move out of
4235 GCC target("prfchw") pragma.
4236 * config/i386/i386.md (prefetch): Emit prefetchwt1 only
4238 * config/i386/i386.c (ix86_option_override_internal): Enable
4239 -mprfchw with -mprefetchwt1.
4241 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
4243 * config/arc/arc.md (casesi_load) <length attribute alternative 0>:
4246 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
4248 * opts.h (CL_PCH_IGNORE): Define.
4249 * targhooks.c (option_affects_pch_p):
4250 Return false for options that have CL_PCH_IGNORE set.
4251 * opt-functions.awk: Process PchIgnore.
4252 * doc/options.texi: Document PchIgnore.
4254 * config/arc/arc.opt (misize): Add PchIgnore property.
4256 2014-03-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4258 * config/rs6000/rs6000.c (rs6000_preferred_reload_class): Disallow
4259 reload of PLUS rtx's outside of GENERAL_REGS or BASE_REGS; relax
4260 constraint on constants to permit them being loaded into
4261 GENERAL_REGS or BASE_REGS.
4263 2014-03-03 Nick Clifton <nickc@redhat.com>
4265 * config/rl78/rl78-real.md (cbranchsi4_real_signed): Add
4266 anti-cacnonical alternatives.
4267 (negandhi3_real): New pattern.
4268 * config/rl78/rl78-virt.md (negandhi3_virt): New pattern.
4270 2014-03-03 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
4272 * config/avr/avr-mcus.def: Remove atxmega16x1.
4273 * config/avr/avr-tables.opt: Regenerate.
4274 * config/avr/t-multilib: Regenerate.
4275 * doc/avr-mmcu.texi: Regenerate.
4277 2014-03-03 Tobias Grosser <tobias@grosser.es>
4278 Mircea Namolaru <mircea.namolaru@inria.fr>
4280 PR tree-optimization/58028
4281 * graphite-clast-to-gimple.c (set_cloog_options): Don't remove
4284 2014-03-03 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4286 * config/arm/neon.md (*movmisalign<mode>): Legitimize addresses
4287 not handled by recognizers.
4289 2014-03-03 Jakub Jelinek <jakub@redhat.com>
4292 * function.c (expand_function_end): Don't emit
4293 clobber_return_register sequence if clobber_after is a BARRIER.
4294 * cfgexpand.c (construct_exit_block): Append instructions before
4295 return_label to prev_bb.
4297 2014-03-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4299 * config/rs6000/constraints.md: Document reserved use of "wc".
4301 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
4304 * ipa.c (function_and_variable_visibility): When dissolving comdat
4305 group, also set all symbols to local.
4307 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
4312 2013-12-14 Jan Hubicka <jh@suse.cz>
4314 * ipa-prop.c (stmt_may_be_vtbl_ptr_store): Skip clobbers.
4316 2014-03-02 Jon Beniston <jon@beniston.com>
4322 * config/lm32/lm32.c (lm32_legitimate_constant_p): Remove, as incorrect.
4323 (TARGET_LEGITIMATE_CONSTANT_P): Undefine, as not needed.
4324 * config/lm32/lm32.md (movsi_insn): Add 32-bit immediate support.
4325 (simple_return, *simple_return): New patterns
4326 * config/lm32/predicates.md (movsi_rhs_operand): Remove as obsolete.
4327 * configure.ac (force_sjlj_exceptions): Force sjlj exceptions for lm32.
4329 2014-03-01 Paolo Carlini <paolo.carlini@oracle.com>
4331 * dwarf2out.c (gen_subprogram_die): Tidy.
4333 2014-03-01 Oleg Endo <olegendo@gcc.gnu.org>
4336 * config/sh/sh.md (*mov_t_msb_neg): Split into ...
4337 (*mov_t_msb_neg_negc): ... this new insn.
4339 2014-02-28 Jason Merrill <jason@redhat.com>
4342 * ipa-devirt.c (ipa_devirt): Don't choose an implicitly-declared
4345 2014-02-28 Paolo Carlini <paolo.carlini@oracle.com>
4348 * dwarf2out.c (decltype_auto_die): New static.
4349 (gen_subprogram_die): Handle 'decltype(auto)' like 'auto'.
4350 (gen_type_die_with_usage): Handle 'decltype(auto)'.
4351 (is_cxx_auto): Likewise.
4353 2014-02-28 Ian Bolton <ian.bolton@arm.com>
4355 * config/aarch64/aarch64.h: Define __ARM_NEON by default if
4356 we are not using general regs only.
4358 2014-02-28 Richard Biener <rguenther@suse.de>
4361 * tree-cfgcleanup.c (tree_forwarder_block_p): Restrict
4362 previous fix and only allow to remove trivial pre-headers
4363 and latches. Also honor LOOPS_MAY_HAVE_MULTIPLE_LATCHES.
4364 (remove_forwarder_block): Properly update the latch of a loop.
4366 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4369 * cselib.c (cselib_hasher::equal): Special-case VALUE lookup.
4370 (cselib_preserved_hash_table): New.
4371 (preserve_constants_and_equivs): Move preserved vals to it.
4372 (cselib_find_slot): Look it up first.
4373 (cselib_init): Initialize it.
4374 (cselib_finish): Release it.
4375 (dump_cselib_table): Dump it.
4377 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4380 * cselib.c (remove_useless_values): Skip to avoid quadratic
4381 behavior if the condition moved from...
4382 (cselib_process_insn): ... here holds.
4384 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4387 * var-tracking.c (vt_initialize): Apply the same condition to
4388 preserve the CFA base value.
4390 2014-02-28 Joey Ye <joey.ye@arm.com>
4393 * config/arm/arm.c (thumb_far_jump_used_p): Don't change
4394 if reload in progress or completed.
4396 2014-02-28 Tobias Burnus <burnus@net-b.de>
4399 * tree-pretty-print.c (dump_generic_node, print_declaration): Handle
4402 2014-02-27 H.J. Lu <hongjiu.lu@intel.com>
4404 * doc/tm.texi.in (Condition Code Status): Update documention for
4405 relative locations of cc0-setter and cc0-user.
4407 2014-02-27 Jeff Law <law@redhat.com>
4409 PR rtl-optimization/52714
4410 * combine.c (try_combine): When splitting an unrecognized PARALLEL
4411 into two independent simple sets, if I3 is a jump, ensure the
4412 pattern we place into I3 is a (set (pc) ...).
4414 2014-02-27 Mikael Pettersson <mikpe@it.uu.se>
4415 Jeff Law <law@redhat.com>
4417 PR rtl-optimization/49847
4418 * cse.c (fold_rtx) Handle case where cc0 setter and cc0 user
4419 are in different blocks.
4420 * doc/tm.texi (Condition Code Status): Update documention for
4421 relative locations of cc0-setter and cc0-user.
4423 2014-02-27 Vladimir Makarov <vmakarov@redhat.com>
4426 * lra.c (lra_emit_add): Check SUBREG too.
4428 2014-02-27 Andreas Schwab <schwab@suse.de>
4430 * config/m68k/m68k.c (m68k_option_override): Disable
4431 -flive-range-shrinkage for classic m68k.
4432 (m68k_override_options_after_change): Likewise.
4434 2014-02-27 Marek Polacek <polacek@redhat.com>
4437 * tree-ssa-uninit.c (gate_warn_uninitialized): Run the pass even for
4438 -Wmaybe-uninitialized.
4440 2014-02-27 Alan Modra <amodra@gmail.com>
4443 * reload1.c (emit_input_reload_insns): When reload_override_in,
4444 set old to rl->in_reg when rl->in_reg is a subreg.
4446 2014-02-26 Richard Biener <rguenther@suse.de>
4449 * lra-assigns.c (spill_for): Avoid mixed-sign comparison.
4451 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
4453 * common/config/i386/predicates.md (const1256_operand): Remove.
4454 (const2356_operand): New.
4455 (const_1_to_2_operand): Remove.
4456 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
4457 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
4458 (*avx512pf_gatherpf<mode>sf): Ditto.
4459 (avx512pf_gatherpf<mode>df): Ditto.
4460 (*avx512pf_gatherpf<mode>df_mask): Ditto.
4461 (*avx512pf_gatherpf<mode>df): Ditto.
4462 (avx512pf_scatterpf<mode>sf): Ditto.
4463 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
4464 (*avx512pf_scatterpf<mode>sf): Ditto.
4465 (avx512pf_scatterpf<mode>df): Ditto.
4466 (*avx512pf_scatterpf<mode>df_mask): Ditto.
4467 (*avx512pf_scatterpf<mode>df): Ditto.
4468 * common/config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET0.
4470 2014-02-26 Ilya Tocar <ilya.tocar@intel.com>
4472 * config/i386/avx512fintrin.h (_mm512_testn_epi32_mask),
4473 (_mm512_mask_testn_epi32_mask), (_mm512_testn_epi64_mask),
4474 (_mm512_mask_testn_epi64_mask): Move to ...
4475 * config/i386/avx512cdintrin.h: Here.
4476 * config/i386/i386.c (bdesc_args): Change MASK_ISA for testnm.
4477 * config/i386/sse.md (avx512f_vmscalef<mode><round_name>): Remove %.
4478 (avx512f_scalef<mode><mask_name><round_name>): Ditto.
4479 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Change conditon to
4480 TARGET_AVX512F from TARGET_AVX512CD.
4482 2014-02-26 Richard Biener <rguenther@suse.de>
4485 * ipa.c (walk_polymorphic_call_targets): Properly guard
4486 call to inline_update_overall_summary.
4488 2014-02-26 Bin Cheng <bin.cheng@arm.com>
4491 * tree-cfgcleanup.c (tree_forwarder_block_p): Protect loop preheaders
4492 and latches only if requested. Fix latch if it is removed.
4493 * tree-ssa-dom.c (tree_ssa_dominator_optimize): Set
4494 LOOPS_HAVE_PREHEADERS.
4496 2014-02-25 Andrew Pinski <apinski@cavium.com>
4498 * builtins.c (expand_builtin_thread_pointer): Create a new target
4499 when the target is NULL.
4501 2014-02-25 Vladimir Makarov <vmakarov@redhat.com>
4503 PR rtl-optimization/60317
4504 * params.def (PARAM_LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
4505 * params.h (LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
4506 * lra-assigns.c: Include params.h.
4507 (spill_for): Use LRA_MAX_CONSIDERED_RELOAD_PSEUDOS as guard for
4508 other reload pseudos considerations.
4510 2014-02-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4512 * config/rs6000/vector.md (*vector_unordered<mode>): Change split
4513 to use canonical form for nor<mode>3.
4515 2014-02-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4518 * config/arm/arm.h (CANNOT_CHANGE_MODE_CLASS): Allow 128 to 64-bit
4521 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
4523 * common/config/i386/i386-common.c (OPTION_MASK_ISA_PREFETCHWT1_SET),
4524 (OPTION_MASK_ISA_PREFETCHWT1_UNSET): New.
4525 (ix86_handle_option): Handle OPT_mprefetchwt1.
4526 * config/i386/cpuid.h (bit_PREFETCHWT1): New.
4527 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
4529 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
4530 OPTION_MASK_ISA_PREFETCHWT1.
4531 * config/i386/i386.c (ix86_target_string): Handle mprefetchwt1.
4532 (PTA_PREFETCHWT1): New.
4533 (ix86_option_override_internal): Handle PTA_PREFETCHWT1.
4534 (ix86_valid_target_attribute_inner_p): Handle OPT_mprefetchwt1.
4535 * config/i386/i386.h (TARGET_PREFETCHWT1, TARGET_PREFETCHWT1_P): New.
4536 * config/i386/i386.md (prefetch): Check TARGET_PREFETCHWT1
4537 (*prefetch_avx512pf_<mode>_: Change into ...
4538 (*prefetch_prefetchwt1_<mode>: This.
4539 * config/i386/i386.opt (mprefetchwt1): New.
4540 * config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET1.
4541 (_mm_prefetch): Handle intent to write.
4542 * doc/invoke.texi (mprefetchwt1), (mno-prefetchwt1): Doccument.
4544 2014-02-25 Richard Biener <rguenther@suse.de>
4547 * emit-rtl.c (mem_attrs_htab): Remove.
4548 (mem_attrs_htab_hash): Likewise.
4549 (mem_attrs_htab_eq): Likewise.
4550 (set_mem_attrs): Always allocate new mem-attrs when something changed.
4551 (init_emit_once): Do not allocate mem_attrs_htab.
4553 2014-02-25 Richard Biener <rguenther@suse.de>
4556 * lto-opts.c (lto_write_options): Output non-explicit conservative
4557 -fwrapv, -fno-trapv and -fno-strict-overflow.
4558 * lto-wrapper.c (merge_and_complain): Handle merging those options.
4559 (run_gcc): And pass them through.
4561 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
4563 * sel-sched.c (calculate_new_fences): New parameter ptime.
4564 Calculate it as a maximum over all fence cycles.
4565 (sel_sched_region_2): Adjust the call to calculate_new_fences.
4566 Print the final schedule timing when sched_verbose.
4568 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
4570 PR rtl-optimization/60292
4571 * sel-sched.c (fill_vec_av_set): Do not reset target availability
4572 bit fot the fence instruction.
4574 2014-02-24 Alangi Derick <alangiderick@gmail.com>
4576 * calls.h: Fix typo in comment.
4578 2014-02-24 John David Anglin <danglin@gcc.gnu.org>
4580 * config/pa/pa.c (pa_output_move_double): Don't valididate when
4581 adjusting offsetable addresses.
4583 2014-02-24 Guozhi Wei <carrot@google.com>
4585 * sparseset.h (sparseset_pop): Fix the wrong index.
4587 2014-02-24 Walter Lee <walt@tilera.com>
4589 * config.gcc (tilepro-*-*): Change to tilepro*-*-*.
4590 (tilegx-*-linux*): Change to tilegx*-*-linux*; Support tilegxbe
4592 * common/config/tilegx/tilegx-common.c
4593 (TARGET_DEFAULT_TARGET_FLAGS): Define.
4594 * config/tilegx/linux.h (ASM_SPEC): Add endian_spec.
4596 * config/tilegx/sync.md (atomic_test_and_set): Handle big endian.
4597 * config/tilegx/tilegx.c (tilegx_return_in_msb): New.
4598 (tilegx_gimplify_va_arg_expr): Handle big endian.
4599 (tilegx_expand_unaligned_load): Ditto.
4600 (tilegx_expand_unaligned_store): Ditto.
4601 (TARGET_RETURN_IN_MSB): New.
4602 * config/tilegx/tilegx.h (TARGET_DEFAULT): New.
4603 (TARGET_ENDIAN_DEFAULT): New.
4604 (TARGET_BIG_ENDIAN): Handle big endian.
4605 (BYTES_BIG_ENDIAN): Ditto.
4606 (WORDS_BIG_ENDIAN): Ditto.
4607 (FLOAT_WORDS_BIG_ENDIAN): Ditto.
4610 * config/tilegx/tilegx.md (extv): Handle big endian.
4612 (insn_st<n>): Ditto.
4613 (insn_st<n>_add<bitsuffix>): Ditto.
4614 (insn_stnt<n>): Ditto.
4615 (insn_stnt<n>_add<bitsuffix>):Ditto.
4616 (vec_interleave_highv8qi): Handle big endian.
4617 (vec_interleave_highv8qi_be): New.
4618 (vec_interleave_highv8qi_le): New.
4619 (insn_v1int_h): Handle big endian.
4620 (vec_interleave_lowv8qi): Handle big endian.
4621 (vec_interleave_lowv8qi_be): New.
4622 (vec_interleave_lowv8qi_le): New.
4623 (insn_v1int_l): Handle big endian.
4624 (vec_interleave_highv4hi): Handle big endian.
4625 (vec_interleave_highv4hi_be): New.
4626 (vec_interleave_highv4hi_le): New.
4627 (insn_v2int_h): Handle big endian.
4628 (vec_interleave_lowv4hi): Handle big endian.
4629 (vec_interleave_lowv4hi_be): New.
4630 (vec_interleave_lowv4hi_le): New.
4631 (insn_v2int_l): Handle big endian.
4632 (vec_interleave_highv2si): Handle big endian.
4633 (vec_interleave_highv2si_be): New.
4634 (vec_interleave_highv2si_le): New.
4635 (insn_v4int_h): Handle big endian.
4636 (vec_interleave_lowv2si): Handle big endian.
4637 (vec_interleave_lowv2si_be): New.
4638 (vec_interleave_lowv2si_le): New.
4639 (insn_v4int_l): Handle big endian.
4640 * config/tilegx/tilegx.opt (mbig-endian): New option.
4641 (mlittle-endian): New option.
4642 * doc/install.texi: Document tilegxbe-linux.
4643 * doc/invoke.texi: Document -mbig-endian and -mlittle-endian.
4645 2014-02-24 Martin Jambor <mjambor@suse.cz>
4648 * ipa-cp.c (propagate_constants_accross_call): Bail out early if
4649 there are no parameter descriptors.
4651 2014-02-24 Andrey Belevantsev <abel@ispras.ru>
4653 PR rtl-optimization/60268
4654 * sched-rgn.c (haifa_find_rgns): Move the nr_regions_initial variable
4655 initialization to ...
4656 (sched_rgn_init): ... here.
4657 (schedule_region): Check for SCHED_PRESSURE_NONE earlier.
4659 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
4661 * config/microblaze/microblaze.md: Correct ashrsi_reg / lshrsi_reg
4664 2014-02-23 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
4666 * config/microblaze/microblaze.h: Remove SECONDARY_MEMORY_NEEDED
4669 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
4671 * /config/microblaze/microblaze.c: Add microblaze_asm_output_mi_thunk,
4672 define TARGET_ASM_OUTPUT_MI_THUNK and TARGET_ASM_CAN_OUTPUT_MI_THUNK.
4674 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
4676 * config/microblaze/predicates.md: Add cmp_op predicate.
4677 * config/microblaze/microblaze.md: Add branch_compare instruction
4678 which uses cmp_op predicate and emits cmp insn before branch.
4679 * config/microblaze/microblaze.c (microblaze_emit_compare): Rename
4680 to microblaze_expand_conditional_branch and consolidate logic.
4681 (microblaze_expand_conditional_branch): emit branch_compare
4682 insn instead of handling cmp op separate from branch insn.
4684 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4686 * config/rs6000/rs6000.c (rs6000_emit_le_vsx_move): Relax assert
4689 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4691 * config/rs6000/altivec.md (altivec_lve<VI_char>x): Replace
4692 define_insn with define_expand and new define_insn
4693 *altivec_lve<VI_char>x_internal.
4694 (altivec_stve<VI_char>x): Replace define_insn with define_expand
4695 and new define_insn *altivec_stve<VI_char>x_internal.
4696 * config/rs6000/rs6000-protos.h (altivec_expand_stvex_be): New
4698 * config/rs6000/rs6000.c (altivec_expand_lvx_be): Document use by
4700 (altivec_expand_stvex_be): New function.
4702 2014-02-22 Joern Rennecke <joern.rennecke@embecosm.com>
4704 * config/avr/avr.c (avr_can_eliminate): Allow elimination from
4705 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM if !frame_pointer_needed.
4706 * config/avr/avr.c (ELIMINABLE_REGS): Add elimination from
4707 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM.
4709 2014-02-21 Vladimir Makarov <vmakarov@redhat.com>
4712 * lra-constraints.c (inherit_reload_reg): Use lra_emit_move
4713 instead of emit_move_insn.
4715 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4717 * config/rs6000/altivec.md (altivec_vsumsws): Replace second
4719 (reduc_uplus_v16qi): Use gen_altivec_vsumsws_direct instead of
4720 gen_altivec_vsumsws.
4722 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4724 * config/rs6000/altivec.md (altivec_lvxl): Rename as
4725 *altivec_lvxl_<mode>_internal and use VM2 iterator instead of V4SI.
4726 (altivec_lvxl_<mode>): New define_expand incorporating
4727 -maltivec=be semantics where needed.
4728 (altivec_lvx): Rename as *altivec_lvx_<mode>_internal.
4729 (altivec_lvx_<mode>): New define_expand incorporating -maltivec=be
4730 semantics where needed.
4731 (altivec_stvx): Rename as *altivec_stvx_<mode>_internal.
4732 (altivec_stvx_<mode>): New define_expand incorporating
4733 -maltivec=be semantics where needed.
4734 (altivec_stvxl): Rename as *altivec_stvxl_<mode>_internal and use
4735 VM2 iterator instead of V4SI.
4736 (altivec_stvxl_<mode>): New define_expand incorporating
4737 -maltivec=be semantics where needed.
4738 * config/rs6000/rs6000-builtin.def: Add new built-in definitions
4739 LVXL_V2DF, LVXL_V2DI, LVXL_V4SF, LVXL_V4SI, LVXL_V8HI, LVXL_V16QI,
4740 LVX_V2DF, LVX_V2DI, LVX_V4SF, LVX_V4SI, LVX_V8HI, LVX_V16QI, STVX_V2DF,
4741 STVX_V2DI, STVX_V4SF, STVX_V4SI, STVX_V8HI, STVX_V16QI, STVXL_V2DF,
4742 STVXL_V2DI, STVXL_V4SF, STVXL_V4SI, STVXL_V8HI, STVXL_V16QI.
4743 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Replace
4744 ALTIVEC_BUILTIN_LVX with ALTIVEC_BUILTIN_LVX_<MODE> throughout;
4745 similarly for ALTIVEC_BUILTIN_LVXL, ALTIVEC_BUILTIN_STVX, and
4746 ALTIVEC_BUILTIN_STVXL.
4747 * config/rs6000/rs6000-protos.h (altivec_expand_lvx_be): New prototype.
4748 (altivec_expand_stvx_be): Likewise.
4749 * config/rs6000/rs6000.c (swap_selector_for_mode): New function.
4750 (altivec_expand_lvx_be): Likewise.
4751 (altivec_expand_stvx_be): Likewise.
4752 (altivec_expand_builtin): Add cases for
4753 ALTIVEC_BUILTIN_STVX_<MODE>, ALTIVEC_BUILTIN_STVXL_<MODE>,
4754 ALTIVEC_BUILTIN_LVXL_<MODE>, and ALTIVEC_BUILTIN_LVX_<MODE>.
4755 (altivec_init_builtins): Add definitions for
4756 __builtin_altivec_lvxl_<mode>, __builtin_altivec_lvx_<mode>,
4757 __builtin_altivec_stvx_<mode>, and __builtin_altivec_stvxl_<mode>.
4759 2014-02-21 Catherine Moore <clm@codesourcery.com>
4761 * doc/invoke.texi (mvirt, mno-virt): Document.
4762 * config/mips/mips.opt (mvirt): New option.
4763 * config/mips/mips.h (ASM_SPEC): Pass mvirt to the assembler.
4765 2014-02-21 Richard Biener <rguenther@suse.de>
4767 PR tree-optimization/60276
4768 * tree-vectorizer.h (struct _stmt_vec_info): Add min_neg_dist field.
4769 (STMT_VINFO_MIN_NEG_DIST): New macro.
4770 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Record
4771 STMT_VINFO_MIN_NEG_DIST.
4772 * tree-vect-stmts.c (vectorizable_load): Verify if assumptions
4773 made for negative dependence distances still hold.
4775 2014-02-21 Richard Biener <rguenther@suse.de>
4778 * tree-ssa-live.c (mark_all_vars_used_1): Do not walk
4779 DECL_INITIAL for globals not in the current function context.
4781 2014-02-21 Jakub Jelinek <jakub@redhat.com>
4783 PR tree-optimization/56490
4784 * params.def (PARAM_UNINIT_CONTROL_DEP_ATTEMPTS): New param.
4785 * tree-ssa-uninit.c: Include params.h.
4786 (compute_control_dep_chain): Add num_calls argument, return false
4787 if it exceed PARAM_UNINIT_CONTROL_DEP_ATTEMPTS param, pass
4788 num_calls to recursive call.
4789 (find_predicates): Change dep_chain into normal array,
4790 cur_chain into auto_vec<edge, MAX_CHAIN_LEN + 1>, add num_calls
4791 variable and adjust compute_control_dep_chain caller.
4792 (find_def_preds): Likewise.
4794 2014-02-21 Thomas Schwinge <thomas@codesourcery.com>
4796 * gimple-pretty-print.c (dump_gimple_omp_for) [flags & TDF_RAW]
4797 <case GF_OMP_FOR_KIND_CILKSIMD>: Add missing break statement.
4799 2014-02-21 Nick Clifton <nickc@redhat.com>
4801 * config/stormy16/stormy16.md (pushdqi1): Add mode to post_inc.
4802 (pushhi1): Likewise.
4803 (popqi1): Add mode to pre_dec.
4806 2014-02-21 Jakub Jelinek <jakub@redhat.com>
4808 * config/i386/i386.c (ix86_expand_vec_perm): Use V8SImode
4809 mode for mask of V8SFmode permutation.
4811 2014-02-20 Richard Henderson <rth@redhat.com>
4814 * builtins.c (expand_builtin_atomic_compare_exchange): Always make
4815 a new pseudo for OLDVAL.
4817 2014-02-20 Jakub Jelinek <jakub@redhat.com>
4820 * config/i386/i386.c (expand_vec_perm_interleave2): Don't call
4821 gen_reg_rtx if d->testing_p.
4822 (expand_vec_perm_pshufb2, expand_vec_perm_broadcast_1): Return early
4823 if d->testing_p and we will certainly return true.
4824 (expand_vec_perm_even_odd_1): Likewise. Don't call gen_reg_rtx
4827 2014-02-20 Uros Bizjak <ubizjak@gmail.com>
4829 * emit-rtl.c (gen_reg_rtx): Assert that
4830 crtl->emit.regno_pointer_align_length is non-zero.
4832 2014-02-20 Richard Henderson <rth@redhat.com>
4835 * builtins.c (expand_builtin_atomic_compare_exchange): Conditionalize
4836 on failure the store back into EXPECT.
4838 2014-02-20 Chung-Lin Tang <cltang@codesourcery.com>
4839 Sandra Loosemore <sandra@codesourcery.com>
4841 * config/nios2/nios2.md (unspec): Add UNSPEC_PIC_GOTOFF_SYM enum.
4842 * config/nios2/nios2.c (nios2_function_profiler): Add
4843 -fPIC (flag_pic == 2) support.
4844 (nios2_handle_custom_fpu_cfg): Fix warning parameter.
4845 (nios2_large_offset_p): New function.
4846 (nios2_unspec_reloc_p): Move up position, update to use
4847 nios2_large_offset_p.
4848 (nios2_unspec_address): Remove function.
4849 (nios2_unspec_offset): New function.
4850 (nios2_large_got_address): New function.
4851 (nios2_got_address): Add large offset support.
4852 (nios2_legitimize_tls_address): Update usage of removed and new
4854 (nios2_symbol_binds_local_p): New function.
4855 (nios2_load_pic_address): Add -fPIC (flag_pic == 2) support.
4856 (nios2_legitimize_address): Update to use nios2_large_offset_p.
4857 (nios2_emit_move_sequence): Avoid legitimizing (const (unspec ...)).
4858 (nios2_print_operand): Merge H/L processing, add hiadj/lo
4859 processing for (const (unspec ...)).
4860 (nios2_unspec_reloc_name): Add UNSPEC_PIC_GOTOFF_SYM case.
4862 2014-02-20 Richard Biener <rguenther@suse.de>
4864 * tree-cfg.c (replace_uses_by): Mark altered BBs before
4865 doing the substitution.
4866 (verify_gimple_assign_single): Also verify bare MEM_REFs on the lhs.
4868 2014-02-20 Martin Jambor <mjambor@suse.cz>
4871 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Uce correct
4872 info when checking whether lattices are bottom.
4874 2014-02-20 Richard Biener <rguenther@suse.de>
4877 * tree-eh.c (execute_cleanup_eh_1): Also cleanup empty EH
4880 2014-02-20 Jan Hubicka <hubicka@ucw.cz>
4883 * ipa-inline-transform.c (clone_inlined_nodes): Add freq_scale
4884 parameter specifying the scaling.
4885 (inline_call): Update.
4886 (want_inline_recursively): Guard division by zero.
4887 (recursive_inlining): Update.
4888 * ipa-inline.h (clone_inlined_nodes): Update.
4890 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
4893 * config/i386/i386.c (classify_argument): Pass structures of size
4894 64 bytes or less in register.
4896 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
4897 Kirill Yukhin <kirill.yukhin@intel.com>
4899 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): Swap operands.
4900 (_mm_rcp28_round_ss): Ditto.
4901 (_mm_rsqrt28_round_sd): Ditto.
4902 (_mm_rsqrt28_round_ss): Ditto.
4903 * config/i386/avx512erintrin.h (_mm_rcp14_round_sd): Ditto.
4904 (_mm_rcp14_round_ss): Ditto.
4905 (_mm_rsqrt14_round_sd): Ditto.
4906 (_mm_rsqrt14_round_ss): Ditto.
4907 * config/i386/sse.md (rsqrt14<mode>): Put nonimmediate operand as
4908 the first input operand, get rid of match_dup.
4909 (avx512er_exp2<mode><mask_name><round_saeonly_name>): Set type
4911 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
4913 (avx512er_vmrcp28<mode><round_saeonly_name>): Put nonimmediate
4914 operand as the first input operand, set type attribute.
4915 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
4917 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Put nonimmediate
4918 operand as the first input operand, set type attribute.
4920 2014-02-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4922 * config/rs6000/rs6000.c (vspltis_constant): Fix most significant
4925 2014-02-19 H.J. Lu <hongjiu.lu@intel.com>
4928 * config/i386/i386.c (construct_container): Remove TFmode check
4929 for X86_64_INTEGER_CLASS.
4931 2014-02-19 Uros Bizjak <ubizjak@gmail.com>
4934 * config/i386/i386.c (type_natural_mode): Warn for ABI changes
4935 only when -Wpsabi is enabled.
4937 2014-02-19 Michael Hudson-Doyle <michael.hudson@linaro.org>
4940 * config/aarch64/aarch64.c (aarch64_pass_by_reference): The rules for
4941 passing arrays in registers are the same as for structs, so remove the
4942 special case for them.
4944 2014-02-19 Eric Botcazou <ebotcazou@adacore.com>
4946 * expr.c (expand_expr_real_1) <case VIEW_CONVERT_EXPR>: For a bit-field
4947 destination type, extract only the valid bits if the source type is not
4948 integral and has a different mode.
4950 2014-02-19 Richard Biener <rguenther@suse.de>
4953 * tree-inline.c (estimate_num_insns): Avoid calling cgraph_get_node
4956 2014-02-19 Richard Biener <rguenther@suse.de>
4959 * ipa-prop.c: Include stringpool.h and tree-ssanames.h.
4960 (ipa_modify_call_arguments): Emit an argument load explicitely and
4961 preserve virtual SSA form there and for the replacement call.
4962 Do not update SSA form nor free dominance info.
4964 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
4966 * ipa.c (function_and_variable_visibility): Also clear WEAK
4967 flag when disolving COMDAT_GROUP.
4969 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
4971 * ipa-prop.h (ipa_ancestor_jf_data): Update ocmment.
4972 * ipa-prop.c (ipa_set_jf_known_type): Return early when
4974 (ipa_set_ancestor_jf): Set type to NULL hwen it is not preserved;
4975 do more sanity checks.
4976 (detect_type_change): Return true when giving up early.
4977 (compute_complex_assign_jump_func): Fix type parameter of
4978 ipa_set_ancestor_jf.
4979 (compute_complex_ancestor_jump_func): Likewise.
4980 (update_jump_functions_after_inlining): Fix updating of
4982 * ipa-cp.c (ipa_get_jf_ancestor_result): Be ready for type to be NULL.
4984 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
4986 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Also remove
4987 inline clones when edge disappears.
4989 2014-02-18 Michael Meissner <meissner@linux.vnet.ibm.com>
4992 * config/rs6000/rs6000.md (mov<mode>_64bit, TF/TDmode moves):
4993 Split 64-bit moves into 2 patterns. Do not allow the use of
4994 direct move for TDmode in little endian, since the decimal value
4995 has little endian bytes within a word, but the 64-bit pieces are
4996 ordered in a big endian fashion, and normal subreg's of TDmode are
4998 (mov<mode>_64bit_dm): Likewise.
4999 (movtd_64bit_nodm): Likewise.
5001 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
5003 PR tree-optimization/60174
5004 * tree-ssa-reassoc.c (init_range_entry): Do not look into the defining
5005 statement of an SSA_NAME that occurs in an abnormal PHI node.
5007 2014-02-18 Jakub Jelinek <jakub@redhat.com>
5010 * final.c (SEEN_BB): Remove.
5011 (SEEN_NOTE, SEEN_EMITTED): Renumber.
5012 (final_scan_insn): Don't force_source_line on second
5013 NOTE_INSN_BASIC_BLOCK.
5015 2014-02-18 Uros Bizjak <ubizjak@gmail.com>
5018 * config/i386/i386.h (struct ix86_args): Add warn_avx512f.
5019 * config/i386/i386.c (init_cumulative_args): Initialize warn_avx512f.
5020 (type_natural_mode): Warn ABI change when %zmm register is not
5021 available for AVX512F vector value passing.
5023 2014-02-18 Kai Tietz <ktietz@redhat.com>
5026 * config/i386/i386.c (ix86_expand_prologue): Use value in
5027 rax register as displacement when restoring %r10 or %rax.
5028 Fix wrong offset when restoring both registers.
5030 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
5032 * ipa-prop.c (compute_complex_ancestor_jump_func): Replace overzealous
5033 assertion with conditional return.
5035 2014-02-18 Jakub Jelinek <jakub@redhat.com>
5036 Uros Bizjak <ubizjak@gmail.com>
5039 * config/i386/driver-i386.c (host_detect_local_cpu): If
5040 YMM state is not saved by the OS, also clear has_f16c. Move
5041 CPUID 0x80000001 handling before YMM state saving checking.
5043 2014-02-18 Andrey Belevantsev <abel@ispras.ru>
5045 PR rtl-optimization/58960
5046 * haifa-sched.c (alloc_global_sched_pressure_data): New,
5047 factored out from ...
5048 (sched_init): ... here.
5049 (free_global_sched_pressure_data): New, factored out from ...
5050 (sched_finish): ... here.
5051 * sched-int.h (free_global_sched_pressure_data): Declare.
5052 * sched-rgn.c (nr_regions_initial): New static global.
5053 (haifa_find_rgns): Initialize it.
5054 (schedule_region): Disable sched-pressure for the newly
5057 2014-02-17 Richard Biener <rguenther@suse.de>
5059 * tree-vect-stmts.c (free_stmt_vec_info): Clear BB and
5060 release SSA defs of pattern stmts.
5062 2014-02-17 Richard Biener <rguenther@suse.de>
5064 * tree-inline.c (expand_call_inline): Release the virtual
5065 operand defined by the call we are about to inline.
5067 2014-02-17 Richard Biener <rguenther@suse.de>
5069 * tree-ssa.c (verify_ssa): If verify_def found an error, ICE.
5071 2014-02-17 Kirill Yukhin <kirill.yukhin@intel.com>
5072 Ilya Tocar <ilya.tocar@intel.com>
5074 * config/i386/avx512fintrin.h (_mm512_maskz_permutexvar_epi64): Swap
5075 arguments order in builtin.
5076 (_mm512_permutexvar_epi64): Ditto.
5077 (_mm512_mask_permutexvar_epi64): Ditto
5078 (_mm512_maskz_permutexvar_epi32): Ditto
5079 (_mm512_permutexvar_epi32): Ditto
5080 (_mm512_mask_permutexvar_epi32): Ditto
5082 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5084 * config/rs6000/altivec.md (p8_vmrgew): Handle little endian targets.
5085 (p8_vmrgow): Likewise.
5087 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5089 * config/rs6000/vsx.md (vsx_xxpermdi_<mode>): Handle little
5092 2014-02-15 Michael Meissner <meissner@linux.vnet.ibm.com>
5095 * config/rs6000/rs6000.md (rreg): Add TFmode, TDmode constraints.
5096 (mov<mode>_internal, TFmode/TDmode): Split TFmode/TDmode moves
5097 into 64-bit and 32-bit moves. On 64-bit moves, add support for
5098 using direct move instructions on ISA 2.07. Also adjust
5099 instruction length for 64-bit.
5100 (mov<mode>_64bit, TFmode/TDmode): Likewise.
5101 (mov<mode>_32bit, TFmode/TDmode): Likewise.
5103 2014-02-15 Alan Modra <amodra@gmail.com>
5107 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
5108 find_replacement on parts of insn rtl that might be reloaded.
5110 2014-02-15 Richard Biener <rguenther@suse.de>
5112 PR tree-optimization/60183
5113 * tree-ssa-phiprop.c (propagate_with_phi): Avoid speculating loads.
5114 (tree_ssa_phiprop): Calculate and free post-dominators.
5116 2014-02-14 Jeff Law <law@redhat.com>
5118 PR rtl-optimization/60131
5119 * ree.c (get_extended_src_reg): New function.
5120 (combine_reaching_defs): Use it rather than assuming location of REG.
5121 (find_and_remove_re): Verify first operand of extension is
5122 a REG before adding the insns to the copy list.
5124 2014-02-14 Roland McGrath <mcgrathr@google.com>
5126 * configure.ac (HAVE_AS_IX86_UD2): New test for 'ud2' mnemonic.
5127 * configure: Regenerated.
5128 * config.in: Regenerated.
5129 * config/i386/i386.md (trap) [HAVE_AS_IX86_UD2]: Use the mnemonic
5130 instead of ASM_SHORT.
5132 2014-02-14 Vladimir Makarov <vmakarov@redhat.com>
5133 Richard Earnshaw <rearnsha@arm.com>
5135 PR rtl-optimization/59535
5136 * lra-constraints.c (process_alt_operands): Encourage alternative
5137 when unassigned pseudo class is superset of the alternative class.
5138 (inherit_reload_reg): Don't inherit when optimizing for code size.
5139 * config/arm/arm.h (MODE_BASE_REG_CLASS): Add version for LRA
5140 returning CORE_REGS for anything but Thumb1 and BASE_REGS for
5141 modes not less than 4 for Thumb1.
5143 2014-02-14 Kyle McMartin <kyle@redhat.com>
5146 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Define for AArch64.
5148 2014-02-14 Richard Biener <rguenther@suse.de>
5150 * cilk-common.c (cilk_arrow): Build a MEM_REF, not an INDIRECT_REF.
5151 (get_frame_arg): Drop the assert with langhook types_compatible_p.
5152 Do not strip INDIRECT_REFs.
5154 2014-02-14 Richard Biener <rguenther@suse.de>
5157 * lto-streamer-out.c (DFS_write_tree_body): Do not follow
5158 DECL_FUNCTION_SPECIFIC_TARGET.
5159 (hash_tree): Do not hash DECL_FUNCTION_SPECIFIC_TARGET.
5160 * tree-streamer-out.c (pack_ts_target_option): Remove.
5161 (streamer_pack_tree_bitfields): Do not stream TS_TARGET_OPTION.
5162 (write_ts_function_decl_tree_pointers): Do not stream
5163 DECL_FUNCTION_SPECIFIC_TARGET.
5164 * tree-streamer-in.c (unpack_ts_target_option): Remove.
5165 (unpack_value_fields): Do not stream TS_TARGET_OPTION.
5166 (lto_input_ts_function_decl_tree_pointers): Do not stream
5167 DECL_FUNCTION_SPECIFIC_TARGET.
5169 2014-02-14 Jakub Jelinek <jakub@redhat.com>
5171 * tree-vect-loop.c (vect_is_slp_reduction): Don't set use_stmt twice.
5172 (get_initial_def_for_induction, vectorizable_induction): Ignore
5173 debug stmts when looking for exit_phi.
5174 (vectorizable_live_operation): Fix up condition.
5176 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
5178 * config/nds32/nds32.c (nds32_asm_function_prologue): Do not use
5179 nreverse() because it changes the content of original tree list.
5181 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
5183 * config/nds32/t-mlibs (MULTILIB_OPTIONS): Fix typo in comment.
5184 * config/nds32/nds32.c (nds32_merge_decl_attributes): Likewise.
5186 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
5188 * config/nds32/nds32.c (nds32_naked_function_p): Follow the
5189 GNU coding standards.
5191 2014-02-13 Jakub Jelinek <jakub@redhat.com>
5194 * dwarf2out.c (gen_subprogram_die): Don't call
5195 add_calling_convention_attribute if subr_die is old_die.
5197 2014-02-13 Sharad Singhai <singhai@google.com>
5199 * doc/optinfo.texi: Fix order of nodes.
5201 2014-02-13 Uros Bizjak <ubizjak@gmail.com>
5203 * config/i386/sse.md (xop_vmfrcz<mode>2): Generate const0 in
5204 operands[2], not operands[3].
5206 2014-02-13 Richard Biener <rguenther@suse.de>
5209 * doc/install.texi (ISL): Update recommended version to 0.12.2,
5210 mention the possibility of an in-tree build.
5211 (CLooG): Update recommended version to 0.18.1, mention the
5212 possibility of an in-tree build and clarify that the ISL
5213 bundled with CLooG does not work.
5215 2014-02-13 Jakub Jelinek <jakub@redhat.com>
5218 * expr.c (compress_float_constant): If x is a hard register,
5219 extend into a pseudo and then move to x.
5221 2014-02-13 Dominik Vogt <vogt@linux.vnet.ibm.com>
5223 * config/s390/s390.c (s390_asm_output_function_label): Fix crash
5224 caused by bad second argument to warning_at() with -mhotpatch and
5225 nested functions (e.g. with gfortran).
5227 2014-02-13 Richard Sandiford <rdsandiford@googlemail.com>
5229 * opts.c (option_name): Remove "enabled by default" rider.
5231 2014-02-12 John David Anglin <danglin@gcc.gnu.org>
5233 * config/pa/pa.c (pa_option_override): Remove auto increment FIXME.
5235 2014-02-12 H.J. Lu <hongjiu.lu@intel.com>
5236 Uros Bizjak <ubizjak@gmail.com>
5239 * configure.ac (HAVE_AS_GOTOFF_IN_DATA): Pass --32 to GNU assembler.
5240 * configure: Regenerated.
5242 2014-02-12 Richard Biener <rguenther@suse.de>
5244 * vec.c (vec_prefix::calculate_allocation): Move as
5245 inline variant to vec.h.
5246 (vec_prefix::calculate_allocation_1): New out-of-line version.
5247 * vec.h (vec_prefix::calculate_allocation_1): Declare.
5248 (vec_prefix::m_has_auto_buf): Rename to ...
5249 (vec_prefix::m_using_auto_storage): ... this.
5250 (vec_prefix::calculate_allocation): Inline the easy cases
5251 and dispatch to calculate_allocation_1 which doesn't need the
5253 (va_heap::reserve): Use gcc_checking_assert.
5254 (vec<T, A, vl_embed>::embedded_init): Add argument to initialize
5255 m_using_auto_storage.
5256 (auto_vec): Change m_vecpfx member to a vec<T, va_heap, vl_embed>
5258 (vec<T, va_heap, vl_ptr>::reserve): Remove redundant check.
5259 (vec<T, va_heap, vl_ptr>::release): Avoid casting.
5260 (vec<T, va_heap, vl_ptr>::using_auto_storage): Simplify.
5262 2014-02-12 Richard Biener <rguenther@suse.de>
5264 * gcse.c (compute_transp): break from loop over canon_modify_mem_list
5265 when we found a dependence.
5267 2014-02-12 Thomas Schwinge <thomas@codesourcery.com>
5269 * gimplify.c (gimplify_call_expr, gimplify_modify_expr): Move
5271 (maybe_fold_stmt): ... into this new function.
5272 * omp-low.c (lower_omp): Update comment.
5274 * omp-low.c (lower_omp_target): Add clobber for sizes array, after
5277 * omp-low.c (diagnose_sb_0): Make sure label_ctx is valid to
5280 2014-02-12 James Greenhalgh <james.greenhalgh@arm.com>
5282 * config/arm/aarch-cost-tables.h (generic_extra_costs): Fix
5283 identifiers in comments.
5284 (cortexa53_extra_costs): Likewise.
5285 * config/arm/arm.c (cortexa9_extra_costs): Fix identifiers in comments.
5286 (cortexa7_extra_costs): Likewise.
5287 (cortexa12_extra_costs): Likewise.
5288 (cortexa15_extra_costs): Likewise.
5289 (v7m_extra_costs): Likewise.
5291 2014-02-12 Richard Biener <rguenther@suse.de>
5294 * gimple-low.c (lower_builtin_posix_memalign): Lower conditional
5295 of posix_memalign being successful.
5296 (lower_stmt): Restrict lowering of posix_memalign to when
5297 -ftree-bit-ccp is enabled.
5299 2014-02-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
5301 * config/avr/avr-c.c (avr_resolve_overloaded_builtin): Pass vNULL for
5303 * config/spu/spu-c.c (spu_resolve_overloaded_builtin): Likewise.
5305 2014-02-12 Eric Botcazou <ebotcazou@adacore.com>
5307 PR rtl-optimization/60116
5308 * combine.c (try_combine): Also remove dangling REG_DEAD notes on the
5309 other_insn once the combination has been validated.
5311 2014-02-11 Jan Hubicka <hubicka@ucw.cz>
5314 * ipa-utils.h (possible_polymorphic_call_targets): Update prototype
5316 * ipa-devirt.c: Include demangle.h
5317 (odr_violation_reported): New static variable.
5318 (add_type_duplicate): Update odr_violations.
5319 (maybe_record_node): Add completep parameter; update it.
5320 (record_target_from_binfo): Add COMPLETEP parameter;
5321 update it as needed.
5322 (possible_polymorphic_call_targets_1): Likewise.
5323 (struct polymorphic_call_target_d): Add nonconstruction_targets;
5324 rename FINAL to COMPLETE.
5325 (record_targets_from_bases): Sanity check we found the binfo;
5326 fix COMPLETEP updating.
5327 (possible_polymorphic_call_targets): Add NONCONSTRUTION_TARGETSP
5328 parameter, fix computing of COMPLETEP.
5329 (dump_possible_polymorphic_call_targets): Imrove readability of dump;
5330 at LTO time do demangling.
5331 (ipa_devirt): Use nonconstruction_targets; Improve dumps.
5332 * gimple-fold.c (gimple_get_virt_method_for_vtable): Add can_refer
5334 (gimple_get_virt_method_for_binfo): Likewise.
5335 * gimple-fold.h (gimple_get_virt_method_for_binfo,
5336 gimple_get_virt_method_for_vtable): Update prototypes.
5338 2014-02-11 Vladimir Makarov <vmakarov@redhat.com>
5341 * genautomata.c (add_presence_absence): Fix typo with
5342 {final_}presence_list.
5344 2014-02-11 Michael Meissner <meissner@linux.vnet.ibm.com>
5347 * config/rs6000/rs6000.md (128-bit GPR splitter): Add a splitter
5348 for VSX/Altivec vectors that land in GPR registers.
5350 2014-02-11 Richard Henderson <rth@redhat.com>
5351 Jakub Jelinek <jakub@redhat.com>
5354 * tree-sra.c (load_assign_lhs_subreplacements): Add VIEW_CONVERT_EXPR
5355 around drhs if type conversion to lacc->type is not useless.
5357 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5359 * config/aarch64/aarch64-cores.def (cortex-a57): Use cortexa57
5361 (cortex-a57.cortex-a53): Likewise.
5362 * config/aarch64/aarch64.c (cortexa57_tunings): New tuning struct.
5364 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5366 * config/arm/thumb2.md (*thumb2_movhi_insn): Add alternatives for
5369 2014-02-11 Renlin Li <Renlin.Li@arm.com>
5371 * doc/sourcebuild.texi: Document check_effective_target_arm_vfp3_ok and
5372 add_options_for_arm_vfp3.
5374 2014-02-11 Jeff Law <law@redhat.com>
5377 * expr.c (expand_expr_addr_expr_1): Handle expand_expr returning an
5378 object with an undesirable mode.
5380 2014-02-11 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5383 * config/i386/sol2-9.h: New file.
5384 * config.gcc (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*,
5385 *-*-solaris2.9*): Use it.
5387 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
5389 * config/microblaze/microblaze.md: Add movsi4_rev insn pattern.
5390 * config/microblaze/predicates.md: Add reg_or_mem_operand predicate.
5392 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
5394 * config/microblaze/microblaze.c: Extend mcpu version format
5396 2014-02-10 David Holsgrove <david.holsgrove@xilinx.com>
5398 * config/microblaze/microblaze.h: Define SIZE_TYPE and PTRDIFF_TYPE.
5400 2014-02-10 Richard Henderson <rth@redhat.com>
5403 * calls.c (expand_call): Don't double-push for reg_parm_stack_space.
5404 * config/i386/i386.c (init_cumulative_args): Remove sorry for 64-bit
5405 ms-abi vs -mno-accumulate-outgoing-args.
5406 (ix86_expand_prologue): Unconditionally call ix86_eax_live_at_start_p.
5407 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Fix comment with
5410 2014-02-10 Bernd Edlinger <bernd.edlinger@hotmail.de>
5413 * cfgexpand.c (expand_asm_operands): Attach source location to
5414 ASM_INPUT rtx objects.
5415 * print-rtl.c (print_rtx): Check for UNKNOWN_LOCATION.
5417 2014-02-10 Nick Clifton <nickc@redhat.com>
5419 * config/mn10300/mn10300.c (popcount): New function.
5420 (mn10300_expand_prologue): Include saved registers in stack usage
5423 2014-02-10 Jeff Law <law@redhat.com>
5426 * reload1.c (emit_input_reload_insns): Do not create invalid RTL
5427 when changing the SET_DEST of a prior insn to avoid an input reload.
5429 2014-02-10 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
5431 * config/rs6000/sysv4.h (ENDIAN_SELECT): Do not attempt to enforce
5432 big-endian mode for -mcall-aixdesc, -mcall-freebsd, -mcall-netbsd,
5433 -mcall-openbsd, or -mcall-linux.
5434 (CC1_ENDIAN_BIG_SPEC): Remove.
5435 (CC1_ENDIAN_LITTLE_SPEC): Remove.
5436 (CC1_ENDIAN_DEFAULT_SPEC): Remove.
5437 (CC1_SPEC): Remove (always empty) %cc1_endian_... spec.
5438 (SUBTARGET_EXTRA_SPECS): Remove %cc1_endian_big, %cc1_endian_little,
5439 and %cc1_endian_default.
5440 * config/rs6000/sysv4le.h (CC1_ENDIAN_DEFAULT_SPEC): Remove.
5442 2014-02-10 Richard Biener <rguenther@suse.de>
5444 PR tree-optimization/60115
5445 * tree-eh.c (tree_could_trap_p): Unify TARGET_MEM_REF and
5446 MEM_REF handling. Properly verify that the accesses are not
5447 out of the objects bound.
5449 2014-02-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5451 * config/aarch64/aarch64.c (aarch64_override_options): Fix typo from
5454 2014-02-10 Eric Botcazou <ebotcazou@adacore.com>
5456 * ipa-devirt.c (get_polymorphic_call_info_from_invariant): Return
5457 proper constants and fix formatting.
5458 (possible_polymorphic_call_targets): Fix formatting.
5460 2014-02-10 Kirill Yukhin <kirill.yukhin@intel.com>
5461 Ilya Tocar <ilya.tocar@intel.com>
5463 * config/i386/avx512fintrin.h (_mm512_storeu_epi64): Removed.
5464 (_mm512_loadu_epi32): Renamed into...
5465 (_mm512_loadu_si512): This.
5466 (_mm512_storeu_epi32): Renamed into...
5467 (_mm512_storeu_si512): This.
5468 (_mm512_maskz_ceil_ps): Removed.
5469 (_mm512_maskz_ceil_pd): Ditto.
5470 (_mm512_maskz_floor_ps): Ditto.
5471 (_mm512_maskz_floor_pd): Ditto.
5472 (_mm512_floor_round_ps): Ditto.
5473 (_mm512_floor_round_pd): Ditto.
5474 (_mm512_ceil_round_ps): Ditto.
5475 (_mm512_ceil_round_pd): Ditto.
5476 (_mm512_mask_floor_round_ps): Ditto.
5477 (_mm512_mask_floor_round_pd): Ditto.
5478 (_mm512_mask_ceil_round_ps): Ditto.
5479 (_mm512_mask_ceil_round_pd): Ditto.
5480 (_mm512_maskz_floor_round_ps): Ditto.
5481 (_mm512_maskz_floor_round_pd): Ditto.
5482 (_mm512_maskz_ceil_round_ps): Ditto.
5483 (_mm512_maskz_ceil_round_pd): Ditto.
5484 (_mm512_expand_pd): Ditto.
5485 (_mm512_expand_ps): Ditto.
5486 * config/i386/i386.c (ix86_builtins): Remove
5487 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK.
5488 (bdesc_args): Ditto.
5489 * config/i386/predicates.md (const1256_operand): New.
5490 (const_1_to_2_operand): Ditto.
5491 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
5492 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
5493 (*avx512pf_gatherpf<mode>sf): Ditto.
5494 (avx512pf_gatherpf<mode>df): Ditto.
5495 (*avx512pf_gatherpf<mode>df_mask): Ditto.
5496 (*avx512pf_gatherpf<mode>df): Ditto.
5497 (avx512pf_scatterpf<mode>sf): Ditto.
5498 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
5499 (*avx512pf_scatterpf<mode>sf): Ditto.
5500 (avx512pf_scatterpf<mode>df): Ditto.
5501 (*avx512pf_scatterpf<mode>df_mask): Ditto.
5502 (*avx512pf_scatterpf<mode>df): Ditto.
5503 (avx512f_expand<mode>): Removed.
5504 (<shift_insn><mode>3<mask_name>): Change predicate type.
5506 2014-02-08 Jakub Jelinek <jakub@redhat.com>
5508 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers
5509 not at the end of datarefs vector use ordered_remove to avoid
5510 reordering datarefs vector.
5513 * gimplify.c (gimplify_bind_expr): In ORT_SIMD region
5514 mark local addressable non-static vars as GOVD_PRIVATE
5515 instead of GOVD_LOCAL.
5516 * omp-low.c (lower_omp_for): Move gimple_bind_vars
5517 and BLOCK_VARS of gimple_bind_block to new_stmt rather
5521 * tree-ssa-ccp.c (surely_varying_stmt_p): Don't return true
5522 if TYPE_ATTRIBUTES (gimple_call_fntype ()) contain
5523 assume_aligned or alloc_align attributes.
5524 (bit_value_assume_aligned): Add ATTR, PTRVAL and ALLOC_ALIGN
5525 arguments. Handle also assume_aligned and alloc_align attributes.
5526 (evaluate_stmt): Adjust bit_value_assume_aligned caller. Handle
5527 calls to functions with assume_aligned or alloc_align attributes.
5528 * doc/extend.texi: Document assume_aligned and alloc_align attributes.
5530 2014-02-08 Terry Guo <terry.guo@arm.com>
5532 * doc/invoke.texi: Document ARM -march=armv7e-m.
5534 2014-02-08 Jakub Jelinek <jakub@redhat.com>
5536 * cilk-common.c (cilk_init_builtins): Clear TREE_NOTHROW
5537 flag on __cilkrts_rethrow builtin.
5540 * ipa-cp.c (determine_versionability): Fail at -O0
5541 or __attribute__((optimize (0))) or -fno-ipa-cp functions.
5542 * tree-sra.c (ipa_sra_preliminary_function_checks): Similarly.
5545 2014-02-04 Jakub Jelinek <jakub@redhat.com>
5548 * tree-inline.c (copy_forbidden): Fail for
5549 __attribute__((optimize (0))) functions.
5551 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
5553 * varpool.c: Include pointer-set.h.
5554 (varpool_remove_unreferenced_decls): Variables in other partitions
5555 will not be output; be however careful to not lose information
5558 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
5560 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do O(1)
5561 lookup in the vtable constructor.
5563 2014-02-07 Jeff Law <law@redhat.com>
5566 * config/m68k/m68k.md (ashldi_extsi): Turn into a
5567 define_insn_and_split.
5569 * ipa-inline.c (inline_small_functions): Fix typos.
5571 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5573 * config/s390/s390-protos.h (s390_can_use_simple_return_insn)
5574 (s390_can_use_return_insn): Declare.
5575 * config/s390/s390.h (EPILOGUE_USES): Define.
5576 * config/s390/s390.c (s390_mainpool_start): Allow two main_pool
5578 (s390_chunkify_start): Handle return JUMP_LABELs.
5579 (s390_early_mach): Emit a main_pool instruction on the entry edge.
5580 (s300_set_up_by_prologue, s390_can_use_simple_return_insn)
5581 (s390_can_use_return_insn): New functions.
5582 (s390_fix_long_loop_prediction): Handle conditional returns.
5583 (TARGET_SET_UP_BY_PROLOGUE): Define.
5584 * config/s390/s390.md (ANY_RETURN): New code iterator.
5585 (*creturn, *csimple_return, return, simple_return): New patterns.
5587 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5589 * config/s390/s390.c (s390_restore_gprs_from_fprs): Add REG_CFA_RESTORE
5590 notes to each restore. Also add REG_CFA_DEF_CFA when restoring %r15.
5591 (s390_optimize_prologue): Don't clear RTX_FRAME_RELATED_P. Update the
5592 REG_CFA_RESTORE list when deciding not to restore a register.
5594 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5596 * config/s390/s390.c: Include tree-pass.h and context.h.
5597 (s390_early_mach): New function, split out from...
5598 (s390_emit_prologue): ...here.
5599 (pass_data_s390_early_mach): New pass structure.
5600 (pass_s390_early_mach): New class.
5601 (s390_option_override): Create and register early_mach pass.
5602 Move to end of file.
5604 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5606 * var-tracking.c (vt_stack_adjustments): Don't require stack_adjusts
5607 to match for the exit block.
5609 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5611 * config/s390/s390.md ("atomic_load<mode>", "atomic_store<mode>")
5612 ("atomic_compare_and_swap<mode>", "atomic_fetch_<atomic><mode>"):
5613 Reject misaligned operands.
5615 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5617 * optabs.c (expand_atomic_compare_and_swap): Allow expander to fail.
5619 2014-02-07 Richard Biener <rguenther@suse.de>
5622 * gimple-low.c (lower_builtin_posix_memalign): New function.
5623 (lower_stmt): Call it to lower posix_memalign in a way
5624 to make alignment info accessible.
5626 2014-02-07 Jakub Jelinek <jakub@redhat.com>
5629 * tree.c (build_common_builtin_nodes): Set ECF_LEAF for
5630 __builtin_setjmp_receiver.
5632 2014-02-07 Richard Biener <rguenther@suse.de>
5635 * builtin-types.def (BT_FN_INT_PTRPTR_SIZE_SIZE): Add.
5636 * builtins.def (BUILT_IN_POSIX_MEMALIGN): Likewise.
5637 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
5638 Handle BUILT_IN_POSIX_MEMALIGN.
5639 (find_func_clobbers): Likewise.
5640 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
5641 (call_may_clobber_ref_p_1): Likewise.
5643 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5646 * ipa-devirt.c (record_target_from_binfo): Remove overactive
5649 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5652 * lto-cgraph.c (lto_output_node): Use
5653 symtab_get_symbol_partitioning_class.
5654 (lto_output_varpool_node): likewise.
5655 (symtab_get_symbol_partitioning_class): Move here from
5657 * cgraph.h (symbol_partitioning_class): Likewise.
5658 (symtab_get_symbol_partitioning_class): Declare.
5660 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5662 * ggc.h (ggc_internal_cleared_alloc): New macro.
5663 * vec.h (vec_safe_copy): Handle memory stats.
5664 * omp-low.c (simd_clone_struct_alloc): Use ggc_internal_cleared_alloc.
5665 * target-globals.c (save_target_globals): Likewise.
5667 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5670 * expr.c (emit_move_resolve_push): Export; be bit more selective
5671 on when to clear alias set.
5672 * expr.h (emit_move_resolve_push): Declare.
5673 * function.h (struct function): Add tail_call_marked.
5674 * tree-tailcall.c (optimize_tail_call): Set tail_call_marked.
5675 * config/i386/i386-protos.h (ix86_expand_push): Remove.
5676 * config/i386/i386.md (TImode move expander): De not call
5678 (FP push expanders): Preserve memory attributes.
5679 * config/i386/sse.md (push<mode>1): Remove.
5680 * config/i386/i386.c (ix86_expand_vector_move): Handle push operation.
5681 (ix86_expand_push): Remove.
5682 * config/i386/mmx.md (push<mode>1): Remove.
5684 2014-02-06 Jakub Jelinek <jakub@redhat.com>
5686 PR rtl-optimization/60030
5687 * internal-fn.c (ubsan_expand_si_overflow_mul_check): Surround
5688 lopart with paradoxical subreg before shifting it up by hprec.
5690 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5692 * config/arm/aarch-cost-tables.h (cortexa57_extra_costs): New table.
5693 Remove extra newline at end of file.
5694 * config/arm/arm.c (arm_cortex_a57_tune): New tuning struct.
5695 (arm_issue_rate): Handle cortexa57.
5696 * config/arm/arm-cores.def (cortex-a57): Use cortex_a57 tuning.
5697 (cortex-a57.cortex-a53): Likewise.
5699 2014-02-06 Jakub Jelinek <jakub@redhat.com>
5702 * config/arm/arm.c (emit_multi_reg_push): Add dwarf_regs_mask argument,
5703 don't record in REG_FRAME_RELATED_EXPR registers not set in that
5705 (arm_expand_prologue): Adjust all callers.
5706 (arm_unwind_emit_sequence): Allow saved, but not important for unwind
5707 info, registers also at the lowest numbered registers side. Use
5708 gcc_assert instead of abort, and SET_SRC/SET_DEST macros instead of
5712 * var-tracking.c (adjust_mems): Before adding a SET to
5713 amd->side_effects, adjust it's SET_SRC using simplify_replace_fn_rtx.
5715 2014-02-06 Alan Modra <amodra@gmail.com>
5718 * config/rs6000/rs6000.c (rs6000_secondary_memory_needed_mode): Only
5719 change SDmode to DDmode when lra_in_progress.
5721 2014-02-06 Jakub Jelinek <jakub@redhat.com>
5724 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers, call
5725 free_data_ref on the dr first, and before goto again also set dr
5726 to the next dr. For simd_lane_access, free old datarefs[i] before
5727 overwriting it. For get_vectype_for_scalar_type failure, don't
5728 free_data_ref if simd_lane_access.
5730 * Makefile.in (prefix.o, cppbuiltin.o): Depend on $(BASEVER).
5733 * tree.h (opts_for_fn): New inline function.
5734 (opt_for_fn): Define.
5735 * config/i386/i386.c (ix86_function_regparm): Use
5736 opt_for_fn (decl, optimize) instead of optimize.
5738 2014-02-06 Marcus Shawcroft <marcus.shawcroft@arm.com>
5740 * config/aarch64/aarch64.c (aarch64_classify_symbol): Fix logic
5741 for SYMBOL_REF in large memory model.
5743 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5745 * config/aarch64/aarch64-cores.def (cortex-a53): Specify CRC32
5747 (cortex-a57): Likewise.
5748 (cortex-a57.cortex-a53): Likewise.
5750 2014-02-06 Yury Gribov <y.gribov@samsung.com>
5751 Kugan Vivekanandarajah <kuganv@linaro.org>
5753 * config/arm/arm.c (arm_vector_alignment_reachable): Check
5755 * config/arm/arm.c (arm_builtin_support_vector_misalignment): Likewise.
5757 2014-02-06 Richard Biener <rguenther@suse.de>
5759 * tree-cfg.c (gimple_duplicate_sese_region): Fix ordering of
5760 set_loop_copy and initialize_original_copy_tables.
5762 2014-02-06 Alex Velenko <Alex.Velenko@arm.com>
5764 * config/aarch64/aarch64-simd.md
5765 (aarch64_ashr_simddi): Change QI to SI.
5767 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
5768 Jakub Jelinek <jakub@redhat.com>
5771 * ipa-inline-analysis.c (compute_bb_predicates): Ensure monotonicity
5774 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5776 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Change
5777 CODE_FOR_altivec_vpku[hw]um to
5778 CODE_FOR_altivec_vpku[hw]um_direct.
5779 * config/rs6000/altivec.md (vec_unpacks_hi_<VP_small_lc>): Change
5780 UNSPEC_VUNPACK_HI_SIGN to UNSPEC_VUNPACK_HI_SIGN_DIRECT.
5781 (vec_unpacks_lo_<VP_small_lc>): Change UNSPEC_VUNPACK_LO_SIGN to
5782 UNSPEC_VUNPACK_LO_SIGN_DIRECT.
5784 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5786 * config/rs6000/altivec.md (altivec_vsum2sws): Adjust code
5787 generation for -maltivec=be.
5788 (altivec_vsumsws): Simplify redundant test.
5790 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5792 * altivec.md (UNSPEC_VPACK_UNS_UNS_MOD_DIRECT): New unspec.
5793 (UNSPEC_VUNPACK_HI_SIGN_DIRECT): Likewise.
5794 (UNSPEC_VUNPACK_LO_SIGN_DIRECT): Likewise.
5795 (mulv8hi3): Use gen_altivec_vpkuwum_direct instead of
5796 gen_altivec_vpkuwum.
5797 (altivec_vpkpx): Test for VECTOR_ELT_ORDER_BIG instead of for
5799 (altivec_vpks<VI_char>ss): Likewise.
5800 (altivec_vpks<VI_char>us): Likewise.
5801 (altivec_vpku<VI_char>us): Likewise.
5802 (altivec_vpku<VI_char>um): Likewise.
5803 (altivec_vpku<VI_char>um_direct): New (copy of
5804 altivec_vpku<VI_char>um that still relies on BYTES_BIG_ENDIAN, for
5806 (altivec_vupkhs<VU_char>): Emit vupkls* instead of vupkhs* when
5807 target is little endian and -maltivec=be is not specified.
5808 (*altivec_vupkhs<VU_char>_direct): New (copy of
5809 altivec_vupkhs<VU_char> that always emits vupkhs*, for internal use).
5810 (altivec_vupkls<VU_char>): Emit vupkhs* instead of vupkls* when
5811 target is little endian and -maltivec=be is not specified.
5812 (*altivec_vupkls<VU_char>_direct): New (copy of
5813 altivec_vupkls<VU_char> that always emits vupkls*, for internal use).
5814 (altivec_vupkhpx): Emit vupklpx instead of vupkhpx when target is
5815 little endian and -maltivec=be is not specified.
5816 (altivec_vupklpx): Emit vupkhpx instead of vupklpx when target is
5817 little endian and -maltivec=be is not specified.
5819 2014-02-05 Richard Henderson <rth@redhat.com>
5822 * combine-stack-adj.c: Revert r206943.
5823 * sched-int.h (struct deps_desc): Add last_args_size.
5824 * sched-deps.c (init_deps): Initialize it.
5825 (sched_analyze_insn): Add OUTPUT dependencies between insns that
5826 contain REG_ARGS_SIZE notes.
5828 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
5830 * lto-cgraph.c (asm_nodes_output): Make global.
5831 * lto-wrapper.c (run_gcc): Pass down paralelizm to WPA.
5832 * gcc.c (AS_NEEDS_DASH_FOR_PIPED_INPUT): Allow WPA parameter
5833 (driver_handle_option): Handle OPT_fwpa.
5835 2014-02-05 Jakub Jelinek <jakub@redhat.com>
5838 * ipa-devirt.c (possible_polymorphic_call_targets): Fix
5839 a comment typo and formatting issue. If odr_hash hasn't been
5840 created, return vNULL and set *completep to false.
5843 * tree-eh.c (cleanup_empty_eh): Bail out on totally empty
5844 bb with no successors.
5846 2014-02-05 James Greenhalgh <james.greenhalgh@arm.com>
5849 * doc/invoke.texi (-march): Clarify documentation for ARM.
5853 2014-02-05 Richard Biener <rguenther@suse.de>
5855 * tree-vect-loop.c (vect_analyze_loop_2): Be more informative
5856 when not vectorizing because of too many alias checks.
5857 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
5858 Add more verboseness, avoid duplicate MSG_MISSED_OPTIMIZATION.
5860 2014-02-05 Nick Clifton <nickc@redhat.com>
5862 * config/mn10300/mn10300.c (mn10300_hard_regno_mode_ok): Do not
5863 accept extended registers in any mode when compiling for the MN10300.
5865 2014-02-05 Yury Gribov <y.gribov@samsung.com>
5867 * cif-code.def (ATTRIBUTE_MISMATCH): New CIF code.
5868 * ipa-inline.c (report_inline_failed_reason): Handle mismatched
5869 sanitization attributes.
5870 (can_inline_edge_p): Likewise.
5871 (sanitize_attrs_match_for_inline_p): New function.
5873 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
5875 * ipa-prop.c (detect_type_change): Shor circuit testing of
5876 type changes on THIS pointer.
5878 2014-02-04 John David Anglin <danglin@gcc.gnu.org>
5881 * config/pa/pa.c (legitimize_tls_address): Return original address
5882 if not passed a SYMBOL_REF rtx.
5883 (hppa_legitimize_address): Call legitimize_tls_address for all TLS
5885 (pa_emit_move_sequence): Simplify TLS source operands.
5886 (pa_legitimate_constant_p): Reject all TLS constants.
5887 * config/pa/pa.h (PA_SYMBOL_REF_TLS_P): Correct comment.
5888 (CONSTANT_ADDRESS_P): Reject TLS CONST addresses.
5890 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
5892 * ipa.c (function_and_variable_visibility): Decompose DECL_ONE_ONLY
5893 groups when we know they are controlled by LTO.
5894 * varasm.c (default_binds_local_p_1): If object is in other partition,
5895 it will be resolved locally.
5897 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
5899 * config/host-linux.c (linux_gt_pch_use_address): Don't
5900 use SSIZE_MAX because it is not always defined.
5902 2014-02-04 Vladimir Makarov <vmakarov@redhat.com>
5905 * lra-constraints.c (need_for_split_p): Use more 3 reloads as
5906 threshold for pseudo splitting.
5907 (update_ebb_live_info): Process call argument hard registers and
5908 hard registers from insn definition too.
5909 (max_small_class_regs_num): New constant.
5910 (inherit_in_ebb): Update live hard regs through EBBs. Update
5911 reloads_num only for small register classes. Don't split for
5914 2014-02-04 Markus Trippelsdorf <markus@trippelsdorf.de>
5917 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that target
5920 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
5922 * gimple-fold.c (can_refer_decl_in_current_unit_p): Default
5925 2014-02-04 Marek Polacek <polacek@redhat.com>
5927 * gdbinit.in (pel): Define.
5929 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
5931 * doc/invoke.texi (fstrict-volatile-bitfields): Clarify current
5934 2014-02-04 Richard Biener <rguenther@suse.de>
5937 * lto-streamer-out.c (tree_is_indexable): Force NAMELIST_DECLs
5938 in function context local.
5939 (lto_output_tree_ref): Do not write trees from lto_output_tree_ref.
5940 * lto-streamer-in.c (lto_input_tree_ref): Handle LTO_namelist_decl_ref
5941 similar to LTO_imported_decl_ref.
5943 2014-02-04 Jakub Jelinek <jakub@redhat.com>
5945 PR tree-optimization/60002
5946 * cgraphclones.c (build_function_decl_skip_args): Clear
5949 PR tree-optimization/60023
5950 * tree-if-conv.c (predicate_mem_writes): Pass true instead of
5951 false to gsi_replace.
5952 * tree-vect-stmts.c (vect_finish_stmt_generation): If stmt
5953 has been in some EH region and vec_stmt could throw, add
5954 vec_stmt into the same EH region.
5955 * tree-data-ref.c (get_references_in_stmt): If IFN_MASK_LOAD
5956 has no lhs, ignore it.
5957 * internal-fn.c (expand_MASK_LOAD): Likewise.
5960 * tree-inline.c (copy_forbidden): Fail for
5961 __attribute__((optimize (0))) functions.
5964 * omp-low.c (simd_clone_struct_copy): If from->inbranch
5965 is set, copy one less argument.
5966 (expand_simd_clones): Don't subtract clone_info->inbranch
5967 from simd_clone_struct_alloc argument.
5969 PR rtl-optimization/57915
5970 * recog.c (simplify_while_replacing): If all unary/binary/relational
5971 operation arguments are constant, attempt to simplify those.
5974 * expmed.c (expand_mult): For MODE_VECTOR_INT multiplication
5975 if there is no vashl<mode>3 or ashl<mode>3 insn, skip_synth.
5977 2014-02-04 Richard Biener <rguenther@suse.de>
5979 PR tree-optimization/60012
5980 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Apply
5981 TBAA disambiguation to all DDRs.
5983 2014-02-04 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5986 * config/sol2.h (LINK_LIBGCC_MAPFILE_SPEC): Define.
5987 (LINK_SPEC): Use it for -shared, -shared-libgcc.
5989 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
5992 * tree.c (get_binfo_at_offset): Do not get confused by empty classes;
5994 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
5996 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Remove.
5997 * gimple-fold.h (gimple_extract_devirt_binfo_from_cst): Remove.
5999 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
6002 * ipa-cp.c (ipa_get_indirect_edge_target_1): Use ipa-devirt
6003 to figure out targets of polymorphic calls with known decl.
6004 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
6005 * ipa-utils.h (get_polymorphic_call_info_from_invariant): Declare.
6006 * ipa-devirt.c (get_polymorphic_call_info_for_decl): Break out from ...
6007 (get_polymorphic_call_info): ... here.
6008 (get_polymorphic_call_info_from_invariant): New function.
6010 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
6012 * ipa-cp.c (ipa_get_indirect_edge_target_1): Do direct
6013 lookup via vtable pointer; check for type consistency
6014 and turn inconsitent facts into UNREACHABLE.
6015 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
6016 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do not ICE on
6017 type inconsistent querries; return UNREACHABLE instead.
6019 2014-02-03 Richard Henderson <rth@twiddle.net>
6022 * tree-ssa-uninit.c (push_to_worklist): Don't re-push if we've
6023 already processed this node.
6024 (normalize_one_pred_1): Pass along mark_set.
6025 (normalize_one_pred): Create and destroy a pointer_set_t.
6026 (normalize_one_pred_chain): Likewise.
6028 2014-02-03 Laurent Aflonsi <laurent.alfonsi@st.com>
6030 PR gcov-profile/58602
6031 * gcc/gcov-io.c (gcov_open): Open with truncation when mode < 0.
6033 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
6036 * ipa-cp.c (ipa_get_indirect_edge_target_1): Give up on
6037 -fno-devirtualize; try to devirtualize by the knowledge of
6038 virtual table pointer given by aggregate propagation.
6039 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
6040 (ipa_print_node_jump_functions): Dump also offset that
6041 is relevant for polymorphic calls.
6042 (determine_known_aggregate_parts): Add arg_type parameter; use it
6043 instead of determining the type from pointer type.
6044 (ipa_compute_jump_functions_for_edge): Update call of
6045 determine_known_aggregate_parts.
6046 * gimple-fold.c (gimple_get_virt_method_for_vtable): Break out from ...
6047 (gimple_get_virt_method_for_binfo): ... here; simplify using
6048 vtable_pointer_value_to_vtable.
6049 * gimple-fold.h (gimple_get_virt_method_for_vtable): Declare.
6050 * ipa-devirt.c (subbinfo_with_vtable_at_offset): Turn OFFSET parameter
6051 to unsigned HOST_WIDE_INT; use vtable_pointer_value_to_vtable.
6052 (vtable_pointer_value_to_vtable): Break out from ...; handle also
6054 (vtable_pointer_value_to_binfo): ... here.
6055 * ipa-utils.h (vtable_pointer_value_to_vtable): Declare.
6057 2014-02-03 Teresa Johnson <tejohnson@google.com>
6059 * tree-vect-slp.c (vect_supported_load_permutation_p): Avoid
6060 redef of outer loop index variable.
6062 2014-02-03 Marc Glisse <marc.glisse@inria.fr>
6066 * doc/extend.texi (Function Attributes): Typo.
6068 2014-02-03 Cong Hou <congh@google.com>
6070 PR tree-optimization/60000
6071 * tree-vect-loop.c (vect_transform_loop): Set pattern_def_seq to NULL
6072 if the vectorized statement is a store. A store statement can only
6073 appear at the end of pattern statements.
6075 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
6077 * config/i386/i386.c (flag_opts): Add -mlong-double-128.
6078 (ix86_option_override_internal): Default long double to 64-bit for
6079 32-bit Bionic and to 128-bit for 64-bit Bionic.
6081 * config/i386/i386.h (LONG_DOUBLE_TYPE_SIZE): Use 128 if
6082 TARGET_LONG_DOUBLE_128 is true.
6083 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Likewise.
6085 * config/i386/i386.opt (mlong-double-80): Negate -mlong-double-64.
6086 (mlong-double-64): Negate -mlong-double-128.
6087 (mlong-double-128): New option.
6089 * config/i386/i386-c.c (ix86_target_macros): Define
6090 __LONG_DOUBLE_128__ for TARGET_LONG_DOUBLE_128.
6092 * doc/invoke.texi: Document -mlong-double-128.
6094 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
6096 PR rtl-optimization/60024
6097 * sel-sched.c (init_regs_for_mode): Check if mode is OK first.
6099 2014-02-03 Markus Trippelsdorf <markus@trippelsdorf.de>
6101 * doc/invoke.texi (fprofile-reorder-functions): Fix typo.
6103 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
6105 PR rtl-optimization/57662
6106 * sel-sched.c (code_motion_path_driver): Do not mark already not
6107 existing blocks in the visiting bitmap.
6109 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
6111 * sel-sched-ir.c (sel_gen_insn_from_expr_after): Reset INSN_DELETED_P
6112 on the insn being emitted.
6114 2014-02-03 James Greenhalgh <james.greenhalgh@arm.com>
6115 Will Deacon <will.deacon@arm.com>
6117 * doc/gimple.texi (gimple_asm_clear_volatile): Remove.
6119 2014-02-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6121 * config/arm/arm-tables.opt: Regenerate.
6123 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6125 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): Generalize
6126 for vector types other than V16QImode.
6127 * config/rs6000/altivec.md (altivec_vperm_<mode>): Change to a
6128 define_expand, and call altivec_expand_vec_perm_le when producing
6129 code with little endian element order.
6130 (*altivec_vperm_<mode>_internal): New insn having previous
6131 behavior of altivec_vperm_<mode>.
6132 (altivec_vperm_<mode>_uns): Change to a define_expand, and call
6133 altivec_expand_vec_perm_le when producing code with little endian
6135 (*altivec_vperm_<mode>_uns_internal): New insn having previous
6136 behavior of altivec_vperm_<mode>_uns.
6138 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6140 * config/rs6000/altivec.md (UNSPEC_VSUMSWS_DIRECT): New unspec.
6141 (altivec_vsumsws): Add handling for -maltivec=be with a little
6143 (altivec_vsumsws_direct): New.
6144 (reduc_splus_<mode>): Call gen_altivec_vsumsws_direct instead of
6145 gen_altivec_vsumsws.
6147 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
6149 * ipa-devirt.c (subbinfo_with_vtable_at_offset,
6150 vtable_pointer_value_to_binfo): New functions.
6151 * ipa-utils.h (vtable_pointer_value_to_binfo): Declare.
6152 * ipa-prop.c (extr_type_from_vtbl_ptr_store): Use it.
6154 2014-02-02 Sandra Loosemore <sandra@codesourcery.com>
6156 * config/nios2/nios2.md (load_got_register): Initialize GOT
6157 pointer from _gp_got instead of _GLOBAL_OFFSET_TABLE_.
6158 * config/nios2/nios2.c (nios2_function_profiler): Likewise.
6160 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
6162 * ipa-prop.c (update_jump_functions_after_inlining): When type is not
6163 preserverd by passthrough, do not propagate the type.
6165 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
6167 * config/mips/mips.c (MIPS_GET_FCSR, MIPS_SET_FCSR): New macros.
6168 (mips_atomic_assign_expand_fenv): New function.
6169 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): Define.
6171 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
6173 * doc/extend.texi (__builtin_mips_get_fcsr): Document.
6174 (__builtin_mips_set_fcsr): Likewise.
6175 * config/mips/mips-ftypes.def: Add MIPS_VOID_FTYPE_USI and
6176 MIPS_USI_FTYPE_VOID.
6177 * config/mips/mips-protos.h (mips16_expand_get_fcsr): Declare
6178 (mips16_expand_set_fcsr): Likewise.
6179 * config/mips/mips.c (mips16_get_fcsr_stub): New variable.
6180 (mips16_set_fcsr_stub): Likewise.
6181 (mips16_get_fcsr_one_only_stub): New class.
6182 (mips16_set_fcsr_one_only_stub): Likewise.
6183 (mips16_expand_get_fcsr, mips16_expand_set_fcsr): New functions.
6184 (mips_code_end): Output the get_fcsr and set_fcsr stubs, if needed.
6185 (BUILTIN_AVAIL_MIPS16, AVAIL_ALL): New macros.
6186 (hard_float): New availability predicate.
6187 (mips_builtins): Add get_fcsr and set_fcsr.
6188 (mips_expand_builtin): Check BUILTIN_AVAIL_MIPS16.
6189 * config/mips/mips.md (UNSPEC_GET_FCSR, UNSPEC_SET_FCSR): New unspecs.
6190 (GET_FCSR_REGNUM, SET_FCSR_REGNUM): New constants.
6191 (mips_get_fcsr, *mips_get_fcsr, mips_get_fcsr_mips16_<mode>)
6192 (mips_set_fcsr, *mips_set_fcsr, mips_set_fcsr_mips16_<mode>): New
6195 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
6197 * config/mips/mips.c (mips_one_only_stub): New class.
6198 (mips_need_mips16_rdhwr_p): Replace with...
6199 (mips16_rdhwr_stub): ...this new variable.
6200 (mips16_stub_call_address): New function.
6201 (mips16_rdhwr_one_only_stub): New class.
6202 (mips_expand_thread_pointer): Use mips16_stub_call_address.
6203 (mips_output_mips16_rdhwr): Delete.
6204 (mips_finish_stub): New function.
6205 (mips_code_end): Use it to handle rdhwr stubs.
6207 2014-02-02 Uros Bizjak <ubizjak@gmail.com>
6210 * config/i386/i386.c (classify_argument): Fix handling of bit_offset
6211 when calculating size of integer atomic types.
6213 2014-02-02 H.J. Lu <hongjiu.lu@intel.com>
6215 * ipa-inline-analysis.c (true_predicate_p): Fix a typo in comments.
6217 2014-02-01 Jakub Jelinek <jakub@redhat.com>
6219 PR tree-optimization/60003
6220 * gimple-low.c (lower_builtin_setjmp): Set cfun->has_nonlocal_label.
6221 * profile.c (branch_prob): Use gimple_call_builtin_p
6222 to check for BUILT_IN_SETJMP_RECEIVER.
6223 * tree-inline.c (copy_bb): Call notice_special_calls.
6225 2014-01-31 Vladimir Makarov <vmakarov@redhat.com>
6228 * lra-constraints.c (process_alt_operands): Update reload_sum only
6231 2014-01-31 Richard Henderson <rth@redhat.com>
6234 * tree-eh.c (lower_try_finally_switch): Delay lowering finally block
6235 until after else_eh is processed.
6237 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
6239 * config/i386/avx512fintrin.h (_MM_FROUND_TO_NEAREST_INT),
6240 (_MM_FROUND_TO_NEG_INF), (_MM_FROUND_TO_POS_INF),
6241 (_MM_FROUND_TO_ZERO), (_MM_FROUND_CUR_DIRECTION): Are already defined
6242 in smmintrin.h, remove them.
6243 (_MM_FROUND_NO_EXC): Same as above, bit also wrong value.
6244 * config/i386/i386.c (ix86_print_operand): Split sae and rounding.
6245 * config/i386/i386.md (ROUND_SAE): Fix value.
6246 * config/i386/predicates.md (const_4_or_8_to_11_operand): New.
6247 (const48_operand): New.
6248 * config/i386/subst.md (round), (round_expand): Use
6249 const_4_or_8_to_11_operand.
6250 (round_saeonly), (round_saeonly_expand): Use const48_operand.
6252 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
6254 * config/i386/constraints.md (Yk): Swap meaning with k.
6255 * config/i386/i386.md (movhi_internal): Change Yk to k.
6256 (movqi_internal): Ditto.
6257 (*k<logic><mode>): Ditto.
6260 (kandn<mode>): Ditto.
6261 (*<code>hi_1): Ditto.
6262 (*<code>qi_1): Ditto.
6263 (kxnor<mode>): Ditto.
6264 (kortestzhi): Ditto.
6265 (kortestchi): Ditto.
6267 (*one_cmplhi2_1): Ditto.
6268 (*one_cmplqi2_1): Ditto.
6269 * config/i386/sse.md (): Change k to Yk.
6270 (avx512f_load<mode>_mask): Ditto.
6271 (avx512f_blendm<mode>): Ditto.
6272 (avx512f_store<mode>_mask): Ditto.
6273 (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
6274 (avx512f_storedqu<mode>_mask): Ditto.
6275 (avx512f_cmp<mode>3<mask_scalar_merge_name><round_saeonly_name>):
6277 (avx512f_ucmp<mode>3<mask_scalar_merge_name>): Ditto.
6278 (avx512f_vmcmp<mode>3<round_saeonly_name>): Ditto.
6279 (avx512f_vmcmp<mode>3_mask<round_saeonly_name>): Ditto.
6280 (avx512f_maskcmp<mode>3): Ditto.
6281 (avx512f_fmadd_<mode>_mask<round_name>): Ditto.
6282 (avx512f_fmadd_<mode>_mask3<round_name>): Ditto.
6283 (avx512f_fmsub_<mode>_mask<round_name>): Ditto.
6284 (avx512f_fmsub_<mode>_mask3<round_name>): Ditto.
6285 (avx512f_fnmadd_<mode>_mask<round_name>): Ditto.
6286 (avx512f_fnmadd_<mode>_mask3<round_name>): Ditto.
6287 (avx512f_fnmsub_<mode>_mask<round_name>): Ditto.
6288 (avx512f_fnmsub_<mode>_mask3<round_name>): Ditto.
6289 (avx512f_fmaddsub_<mode>_mask<round_name>): Ditto.
6290 (avx512f_fmaddsub_<mode>_mask3<round_name>): Ditto.
6291 (avx512f_fmsubadd_<mode>_mask<round_name>): Ditto.
6292 (avx512f_fmsubadd_<mode>_mask3<round_name>): Ditto.
6293 (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
6294 (vec_extract_lo_<mode>_maskm): Ditto.
6295 (vec_extract_hi_<mode>_maskm): Ditto.
6296 (avx512f_vternlog<mode>_mask): Ditto.
6297 (avx512f_fixupimm<mode>_mask<round_saeonly_name>): Ditto.
6298 (avx512f_sfixupimm<mode>_mask<round_saeonly_name>): Ditto.
6299 (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
6300 (avx512f_<code>v8div16qi2_mask): Ditto.
6301 (avx512f_<code>v8div16qi2_mask_store): Ditto.
6302 (avx512f_eq<mode>3<mask_scalar_merge_name>_1): Ditto.
6303 (avx512f_gt<mode>3<mask_scalar_merge_name>): Ditto.
6304 (avx512f_testm<mode>3<mask_scalar_merge_name>): Ditto.
6305 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Ditto.
6306 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
6307 (*avx512pf_gatherpf<mode>df_mask): Ditto.
6308 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
6309 (*avx512pf_scatterpf<mode>df_mask): Ditto.
6310 (avx512cd_maskb_vec_dupv8di): Ditto.
6311 (avx512cd_maskw_vec_dupv16si): Ditto.
6312 (avx512f_vpermi2var<mode>3_maskz): Ditto.
6313 (avx512f_vpermi2var<mode>3_mask): Ditto.
6314 (avx512f_vpermi2var<mode>3_mask): Ditto.
6315 (avx512f_vpermt2var<mode>3_maskz): Ditto.
6316 (*avx512f_gathersi<mode>): Ditto.
6317 (*avx512f_gathersi<mode>_2): Ditto.
6318 (*avx512f_gatherdi<mode>): Ditto.
6319 (*avx512f_gatherdi<mode>_2): Ditto.
6320 (*avx512f_scattersi<mode>): Ditto.
6321 (*avx512f_scatterdi<mode>): Ditto.
6322 (avx512f_compress<mode>_mask): Ditto.
6323 (avx512f_compressstore<mode>_mask): Ditto.
6324 (avx512f_expand<mode>_mask): Ditto.
6325 * config/i386/subst.md (mask): Change k to Yk.
6326 (mask_scalar_merge): Ditto.
6329 2014-01-31 Marc Glisse <marc.glisse@inria.fr>
6331 * doc/extend.texi (Vector Extensions): Document ?: in C++.
6333 2014-01-31 Richard Biener <rguenther@suse.de>
6336 * builtins.c (fold_builtin_memory_op): Make sure to not
6337 use a floating-point mode or a boolean or enumeral type for
6340 2014-01-30 DJ Delorie <dj@redhat.com>
6342 * config/msp430/msp430.h (LIB_SPEC): Add -lcrt
6343 * config/msp430/msp430.md (msp430_refsym_need_exit): New.
6344 * config/msp430/msp430.c (msp430_expand_epilogue): Call it
6345 whenever main() has an epilogue.
6347 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6349 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Remove
6350 unused variable "field".
6351 * config/rs6000/vsx.md (vsx_mergel_<mode>): Add missing DONE.
6352 (vsx_mergeh_<mode>): Likewise.
6353 * config/rs6000/altivec.md (altivec_vmrghb): Likewise.
6354 (altivec_vmrghh): Likewise.
6355 (altivec_vmrghw): Likewise.
6356 (altivec_vmrglb): Likewise.
6357 (altivec_vmrglh): Likewise.
6358 (altivec_vmrglw): Likewise.
6359 (altivec_vspltb): Add missing uses.
6360 (altivec_vsplth): Likewise.
6361 (altivec_vspltw): Likewise.
6362 (altivec_vspltsf): Likewise.
6364 2014-01-30 Jakub Jelinek <jakub@redhat.com>
6367 * ifcvt.c (cond_exec_process_insns): Don't conditionalize
6368 frame related instructions.
6370 2014-01-30 Vladimir Makarov <vmakarov@redhat.com>
6372 PR rtl-optimization/59959
6373 * lra-constrains.c (simplify_operand_subreg): Assign NO_REGS to
6374 any reload of register whose subreg is invalid.
6376 2014-01-30 Jakub Jelinek <jakub@redhat.com>
6378 * config/i386/f16cintrin.h (_cvtsh_ss): Avoid -Wnarrowing warning.
6379 * config/i386/avx512fintrin.h (_mm512_mask_cvtusepi64_storeu_epi32):
6380 Add missing return type - void.
6382 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6384 * gcc/config/rs6000/rs6000.c (rs6000_expand_vector_init): Use
6385 gen_vsx_xxspltw_v4sf_direct instead of gen_vsx_xxspltw_v4sf;
6386 remove element index adjustment for endian (now handled in vsx.md
6388 (altivec_expand_vec_perm_const): Use
6389 gen_altivec_vsplt[bhw]_direct instead of gen_altivec_vsplt[bhw].
6390 * gcc/config/rs6000/vsx.md (UNSPEC_VSX_XXSPLTW): New unspec.
6391 (vsx_xxspltw_<mode>): Adjust element index for little endian.
6392 * gcc/config/rs6000/altivec.md (altivec_vspltb): Divide into a
6393 define_expand and a new define_insn *altivec_vspltb_internal;
6394 adjust for -maltivec=be on a little endian target.
6395 (altivec_vspltb_direct): New.
6396 (altivec_vsplth): Divide into a define_expand and a new
6397 define_insn *altivec_vsplth_internal; adjust for -maltivec=be on a
6398 little endian target.
6399 (altivec_vsplth_direct): New.
6400 (altivec_vspltw): Divide into a define_expand and a new
6401 define_insn *altivec_vspltw_internal; adjust for -maltivec=be on a
6402 little endian target.
6403 (altivec_vspltw_direct): New.
6404 (altivec_vspltsf): Divide into a define_expand and a new
6405 define_insn *altivec_vspltsf_internal; adjust for -maltivec=be on
6406 a little endian target.
6408 2014-01-30 Richard Biener <rguenther@suse.de>
6410 PR tree-optimization/59993
6411 * tree-ssa-forwprop.c (associate_pointerplus): Check we
6412 can propagate form the earlier stmt and avoid the transform
6413 when the intermediate result is needed.
6415 2014-01-30 Alangi Derick <alangiderick@gmail.com>
6417 * README.Portability: Fix typo.
6419 2014-01-30 David Holsgrove <david.holsgrove@xilinx.com>
6421 * config/microblaze/microblaze.md(cstoresf4, cbranchsf4): Replace
6422 comparison_operator with ordered_comparison_operator.
6424 2014-01-30 Nick Clifton <nickc@redhat.com>
6426 * config/mn10300/mn10300-protos.h (mn10300_store_multiple_operation_p):
6427 Rename to mn10300_store_multiple_regs.
6428 * config/mn10300/mn10300.c: Likewise.
6429 * config/mn10300/mn10300.md (store_movm): Fix typo: call
6430 store_multiple_regs.
6431 * config/mn10300/predicates.md (mn10300_store_multiple_operation):
6432 Call mn10300_store_multiple_regs.
6434 2014-01-30 Nick Clifton <nickc@redhat.com>
6435 DJ Delorie <dj@redhat.com>
6437 * config/rl78/rl78.c (register_sizes): Make the "upper half" of
6438 %fp 2 to keep registers after it properly word-aligned.
6439 (rl78_alloc_physical_registers_umul): Handle the case where both
6440 input operands are the same.
6442 2014-01-30 Richard Biener <rguenther@suse.de>
6444 PR tree-optimization/59903
6445 * tree-vect-loop.c (vect_transform_loop): Guard multiple-types
6448 2014-01-30 Jason Merrill <jason@redhat.com>
6451 * tree.c (walk_type_fields): Handle VECTOR_TYPE.
6454 * cgraphunit.c (expand_thunk): Copy volatile arg to a temporary.
6456 2014-01-30 Richard Biener <rguenther@suse.de>
6458 PR tree-optimization/59951
6459 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Skip uses in debug insns.
6461 2014-01-30 Savin Zlobec <savin.zlobec@gmail.com>
6464 * config/nios2/nios2.c (nios2_fpu_insn_asm): Fix asm output of
6465 SFmode to DFmode case.
6467 2014-01-29 DJ Delorie <dj@redhat.com>
6469 * config/msp430/msp430.opt (-minrt): New.
6470 * config/msp430/msp430.h (STARTFILE_SPEC): Link alternate runtime
6472 (ENDFILE_SPEC): Likewise.
6474 2014-01-29 Jan Hubicka <hubicka@ucw.cz>
6476 * ipa-inline-analysis.c (clobber_only_eh_bb_p): New function.
6477 (estimate_function_body_sizes): Use it.
6479 2014-01-29 Paolo Carlini <paolo.carlini@oracle.com>
6482 * dwarf2out.c (is_cxx_auto): New.
6483 (is_base_type): Use it.
6484 (gen_type_die_with_usage): Likewise.
6486 2014-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6488 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Use
6489 CODE_FOR_altivec_vmrg*_direct rather than CODE_FOR_altivec_vmrg*.
6490 * config/rs6000/vsx.md (vsx_mergel_<mode>): Adjust for
6491 -maltivec=be with LE targets.
6492 (vsx_mergeh_<mode>): Likewise.
6493 * config/rs6000/altivec.md (UNSPEC_VMRG[HL]_DIRECT): New unspecs.
6494 (mulv8hi3): Use gen_altivec_vmrg[hl]w_direct.
6495 (altivec_vmrghb): Replace with define_expand and new
6496 *altivec_vmrghb_internal insn; adjust for -maltivec=be with LE targets.
6497 (altivec_vmrghb_direct): New define_insn.
6498 (altivec_vmrghh): Replace with define_expand and new
6499 *altivec_vmrghh_internal insn; adjust for -maltivec=be with LE targets.
6500 (altivec_vmrghh_direct): New define_insn.
6501 (altivec_vmrghw): Replace with define_expand and new
6502 *altivec_vmrghw_internal insn; adjust for -maltivec=be with LE targets.
6503 (altivec_vmrghw_direct): New define_insn.
6504 (*altivec_vmrghsf): Adjust for endianness.
6505 (altivec_vmrglb): Replace with define_expand and new
6506 *altivec_vmrglb_internal insn; adjust for -maltivec=be with LE targets.
6507 (altivec_vmrglb_direct): New define_insn.
6508 (altivec_vmrglh): Replace with define_expand and new
6509 *altivec_vmrglh_internal insn; adjust for -maltivec=be with LE targets.
6510 (altivec_vmrglh_direct): New define_insn.
6511 (altivec_vmrglw): Replace with define_expand and new
6512 *altivec_vmrglw_internal insn; adjust for -maltivec=be with LE targets.
6513 (altivec_vmrglw_direct): New define_insn.
6514 (*altivec_vmrglsf): Adjust for endianness.
6515 (vec_widen_umult_hi_v16qi): Use gen_altivec_vmrghh_direct.
6516 (vec_widen_umult_lo_v16qi): Use gen_altivec_vmrglh_direct.
6517 (vec_widen_smult_hi_v16qi): Use gen_altivec_vmrghh_direct.
6518 (vec_widen_smult_lo_v16qi): Use gen_altivec_vmrglh_direct.
6519 (vec_widen_umult_hi_v8hi): Use gen_altivec_vmrghw_direct.
6520 (vec_widen_umult_lo_v8hi): Use gen_altivec_vmrglw_direct.
6521 (vec_widen_smult_hi_v8hi): Use gen_altivec_vmrghw_direct.
6522 (vec_widen_smult_lo_v8hi): Use gen_altivec_vmrglw_direct.
6524 2014-01-29 Marcus Shawcroft <marcus.shawcroft@arm.com>
6526 * config/aarch64/aarch64.c (aarch64_expand_mov_immediate)
6527 (aarch64_legitimate_address_p, aarch64_class_max_nregs): Adjust
6530 2014-01-29 Richard Biener <rguenther@suse.de>
6532 PR tree-optimization/58742
6533 * tree-ssa-forwprop.c (associate_pointerplus): Rename to
6534 associate_pointerplus_align.
6535 (associate_pointerplus_diff): New function.
6536 (associate_pointerplus): Likewise. Call associate_pointerplus_align
6537 and associate_pointerplus_diff.
6539 2014-01-29 Richard Biener <rguenther@suse.de>
6541 * lto-streamer.h (LTO_major_version): Bump to 3.
6542 (LTO_minor_version): Reset to 0.
6544 2014-01-29 Renlin Li <Renlin.Li@arm.com>
6546 * config/arm/arm-arches.def (ARM_ARCH): Add armv7ve arch.
6547 * config/arm/arm.c (FL_FOR_ARCH7VE): New.
6548 (arm_file_start): Generate correct asm header for armv7ve.
6549 * config/arm/bpabi.h: Add multilib support for armv7ve.
6550 * config/arm/driver-arm.c: Change the architectures of cortex-a7
6551 and cortex-a15 to armv7ve.
6552 * config/arm/t-aprofile: Add multilib support for armv7ve.
6553 * doc/invoke.texi: Document -march=armv7ve.
6555 2014-01-29 Richard Biener <rguenther@suse.de>
6557 PR tree-optimization/58742
6558 * tree-ssa-forwprop.c (associate_plusminus): Return true
6559 if we changed sth, defer EH cleanup to ...
6560 (ssa_forward_propagate_and_combine): ... here. Call simplify_mult.
6561 (simplify_mult): New function.
6563 2014-01-29 Jakub Jelinek <jakub@redhat.com>
6566 PR tree-optimization/59920
6567 * tree.c (build_common_builtin_nodes): Remove
6568 __builtin_setjmp_dispatcher initialization.
6569 * omp-low.h (make_gimple_omp_edges): Add a new int * argument.
6570 * profile.c (branch_prob): Use gsi_start_nondebug_after_labels_bb
6571 instead of gsi_after_labels + manually skipping debug stmts.
6572 Don't ignore bbs with BUILT_IN_SETJMP_DISPATCHER, instead
6573 ignore bbs with IFN_ABNORMAL_DISPATCHER.
6574 * tree-inline.c (copy_edges_for_bb): Remove
6575 can_make_abnormal_goto argument, instead add abnormal_goto_dest
6576 argument. Ignore computed_goto_p stmts. Don't call
6577 make_abnormal_goto_edges. If a call might need abnormal edges
6578 for non-local gotos, see if it already has an edge to
6579 IFN_ABNORMAL_DISPATCHER or if it is IFN_ABNORMAL_DISPATCHER
6580 with true argument, don't do anything then, otherwise add
6581 EDGE_ABNORMAL from the call's bb to abnormal_goto_dest.
6582 (copy_cfg_body): Compute abnormal_goto_dest, adjust copy_edges_for_bb
6584 * gimple-low.c (struct lower_data): Remove calls_builtin_setjmp.
6585 (lower_function_body): Don't emit __builtin_setjmp_dispatcher.
6586 (lower_stmt): Don't set data->calls_builtin_setjmp.
6587 (lower_builtin_setjmp): Adjust comment.
6588 * builtins.def (BUILT_IN_SETJMP_DISPATCHER): Remove.
6589 * tree-cfg.c (found_computed_goto): Remove.
6590 (factor_computed_gotos): Remove.
6591 (make_goto_expr_edges): Return bool, true for computed gotos.
6592 Don't call make_abnormal_goto_edges.
6593 (build_gimple_cfg): Don't set found_computed_goto, don't call
6594 factor_computed_gotos.
6595 (computed_goto_p): No longer static.
6596 (make_blocks): Don't set found_computed_goto.
6597 (get_abnormal_succ_dispatcher, handle_abnormal_edges): New functions.
6598 (make_edges): If make_goto_expr_edges returns true, push bb
6599 into ab_edge_goto vector, for stmt_can_make_abnormal_goto calls
6600 instead of calling make_abnormal_goto_edges push bb into ab_edge_call
6601 vector. Record mapping between bbs and OpenMP regions if there
6602 are any, adjust make_gimple_omp_edges caller. Call
6603 handle_abnormal_edges.
6604 (make_abnormal_goto_edges): Remove.
6605 * tree-cfg.h (make_abnormal_goto_edges): Remove.
6606 (computed_goto_p, get_abnormal_succ_dispatcher): New prototypes.
6607 * internal-fn.c (expand_ABNORMAL_DISPATCHER): New function.
6608 * builtins.c (expand_builtin): Don't handle BUILT_IN_SETJMP_DISPATCHER.
6609 * internal-fn.def (ABNORMAL_DISPATCHER): New.
6610 * omp-low.c (make_gimple_omp_edges): Add region_idx argument, when
6611 filling *region also set *region_idx to (*region)->entry->index.
6614 * read-rtl.c (read_rtx_code): Clear all of RTX_CODE_SIZE (code).
6615 For REGs set ORIGINAL_REGNO.
6617 2014-01-29 Bingfeng Mei <bmei@broadcom.com>
6619 * doc/md.texi: Mention that a target shouldn't implement
6620 vec_widen_(s|u)mul_even/odd pair if it is less efficient
6623 2014-01-29 Jakub Jelinek <jakub@redhat.com>
6625 PR tree-optimization/59594
6626 * tree-vect-data-refs.c (vect_analyze_data_ref_accesses): Sort
6627 a copy of the datarefs vector rather than the vector itself.
6629 2014-01-28 Jason Merrill <jason@redhat.com>
6632 * dwarf2out.c (auto_die): New static.
6633 (gen_type_die_with_usage): Handle C++1y 'auto'.
6634 (gen_subprogram_die): If in-class DIE had 'auto', emit type again
6637 2014-01-28 H.J. Lu <hongjiu.lu@intel.com>
6640 * config/i386/gnu-user64.h (SPEC_32): Add "m16|" to "m32".
6641 (SPEC_X32): Likewise.
6642 (SPEC_64): Likewise.
6643 * config/i386/i386.c (ix86_option_override_internal): Turn off
6644 OPTION_MASK_ISA_64BIT, OPTION_MASK_ABI_X32 and OPTION_MASK_ABI_64
6646 (x86_file_start): Output .code16gcc for TARGET_16BIT.
6647 * config/i386/i386.h (TARGET_16BIT): New macro.
6648 (TARGET_16BIT_P): Likewise.
6649 * config/i386/i386.opt: Add m16.
6650 * doc/invoke.texi: Document -m16.
6652 2014-01-28 Jakub Jelinek <jakub@redhat.com>
6654 PR preprocessor/59935
6655 * input.c (location_get_source_line): Bail out on when line number
6656 is zero, and test the return value of lookup_or_add_file_to_cache_tab.
6658 2014-01-28 Richard Biener <rguenther@suse.de>
6660 PR tree-optimization/58742
6661 * tree-ssa-forwprop.c (associate_plusminus): Handle
6662 pointer subtraction of the form (T)(P + A) - (T)P.
6664 2014-01-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6666 * config/arm/arm.c (arm_new_rtx_costs): Remove useless statement
6669 2014-01-28 Richard Biener <rguenther@suse.de>
6672 2014-01-28 Richard Biener <rguenther@suse.de>
6674 PR rtl-optimization/45364
6675 PR rtl-optimization/59890
6676 * var-tracking.c (local_get_addr_clear_given_value): Handle
6677 already cleared slot.
6678 (val_reset): Handle not allocated local_get_addr_cache.
6679 (vt_find_locations): Use post-order on the inverted CFG.
6681 2014-01-28 Richard Biener <rguenther@suse.de>
6683 * tree-data-ref.h (ddr_is_anti_dependent, ddrs_have_anti_deps): Remove.
6685 2014-01-28 Richard Biener <rguenther@suse.de>
6687 PR rtl-optimization/45364
6688 PR rtl-optimization/59890
6689 * var-tracking.c (local_get_addr_clear_given_value): Handle
6690 already cleared slot.
6691 (val_reset): Handle not allocated local_get_addr_cache.
6692 (vt_find_locations): Use post-order on the inverted CFG.
6694 2014-01-28 Alan Modra <amodra@gmail.com>
6696 * Makefile.in (BUILD_CPPFLAGS): Do not use ALL_CPPFLAGS.
6697 * configure.ac <recursive call for build != host>: Define
6698 GENERATOR_FILE. Comment. Use CXX_FOR_BUILD, CXXFLAGS_FOR_BUILD
6699 and LD_FOR_BUILD too.
6700 * configure: Regenerate.
6702 2014-01-27 Allan Sandfeld Jensen <sandfeld@kde.org>
6704 * config/i386/i386.c (get_builtin_code_for_version): Separate
6705 Westmere from Nehalem, Ivy Bridge from Sandy Bridge and
6706 Broadwell from Haswell.
6708 2014-01-27 Steve Ellcey <sellcey@mips.com>
6710 * common/config/mips/mips-common.c (TARGET_DEFAULT_TARGET_FLAGS):
6711 Remove TARGET_FP_EXCEPTIONS_DEFAULT and MASK_FUSED_MADD.
6712 * config/mips/mips.c (mips_option_override): Change setting
6714 * config/mips/mips.h (TARGET_FP_EXCEPTIONS_DEFAULT): Remove.
6715 * config/mips/mips.opt (DSP, DSPR2, FP_EXCEPTIONS, FUSED_MADD, MIPS3D):
6716 Change from Mask to Var.
6718 2014-01-27 Jeff Law <law@redhat.com>
6720 * ipa-inline.c (inline_small_functions): Fix typo.
6722 2014-01-27 Ilya Tocar <ilya.tocar@intel.com>
6724 * config/i386/avx512fintrin.h (_mm512_mask_cvtepi32_storeu_epi8): New.
6725 (_mm512_mask_cvtsepi32_storeu_epi8): Ditto.
6726 (_mm512_mask_cvtusepi32_storeu_epi8): Ditto.
6727 (_mm512_mask_cvtepi32_storeu_epi16): Ditto.
6728 (_mm512_mask_cvtsepi32_storeu_epi16): Ditto.
6729 (_mm512_mask_cvtusepi32_storeu_epi16): Ditto.
6730 (_mm512_mask_cvtepi64_storeu_epi32): Ditto.
6731 (_mm512_mask_cvtsepi64_storeu_epi32): Ditto.
6732 (_mm512_mask_cvtusepi64_storeu_epi32): Ditto.
6733 (_mm512_mask_cvtepi64_storeu_epi16): Ditto.
6734 (_mm512_mask_cvtsepi64_storeu_epi16): Ditto.
6735 (_mm512_mask_cvtusepi64_storeu_epi16): Ditto.
6736 (_mm512_mask_cvtepi64_storeu_epi8): Ditto.
6737 (_mm512_mask_cvtsepi64_storeu_epi8): Ditto.
6738 (_mm512_mask_cvtusepi64_storeu_epi8): Ditto.
6739 (_mm512_storeu_epi64): Ditto.
6740 (_mm512_cmpge_epi32_mask): Ditto.
6741 (_mm512_cmpge_epu32_mask): Ditto.
6742 (_mm512_cmpge_epi64_mask): Ditto.
6743 (_mm512_cmpge_epu64_mask): Ditto.
6744 (_mm512_cmple_epi32_mask): Ditto.
6745 (_mm512_cmple_epu32_mask): Ditto.
6746 (_mm512_cmple_epi64_mask): Ditto.
6747 (_mm512_cmple_epu64_mask): Ditto.
6748 (_mm512_cmplt_epi32_mask): Ditto.
6749 (_mm512_cmplt_epu32_mask): Ditto.
6750 (_mm512_cmplt_epi64_mask): Ditto.
6751 (_mm512_cmplt_epu64_mask): Ditto.
6752 (_mm512_cmpneq_epi32_mask): Ditto.
6753 (_mm512_cmpneq_epu32_mask): Ditto.
6754 (_mm512_cmpneq_epi64_mask): Ditto.
6755 (_mm512_cmpneq_epu64_mask): Ditto.
6756 (_mm512_expand_pd): Ditto.
6757 (_mm512_expand_ps): Ditto.
6758 * config/i386/i386-builtin-types.def: Add PV16QI, PV16QI, PV16HI,
6759 VOID_PV8SI_V8DI_QI, VOID_PV8HI_V8DI_QI, VOID_PV16QI_V8DI_QI,
6760 VOID_PV16QI_V16SI_HI, VOID_PV16HI_V16SI_HI.
6761 * config/i386/i386.c (ix86_builtins): Add
6762 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK,
6763 IX86_BUILTIN_PMOVDB512_MEM, IX86_BUILTIN_PMOVDW512_MEM,
6764 IX86_BUILTIN_PMOVQB512_MEM, IX86_BUILTIN_PMOVQD512_MEM,
6765 IX86_BUILTIN_PMOVQW512_MEM, IX86_BUILTIN_PMOVSDB512_MEM,
6766 IX86_BUILTIN_PMOVSDW512_MEM, IX86_BUILTIN_PMOVSQB512_MEM,
6767 IX86_BUILTIN_PMOVSQD512_MEM, IX86_BUILTIN_PMOVSQW512_MEM,
6768 IX86_BUILTIN_PMOVUSDB512_MEM, IX86_BUILTIN_PMOVUSDW512_MEM,
6769 IX86_BUILTIN_PMOVUSQB512_MEM, IX86_BUILTIN_PMOVUSQD512_MEM,
6770 IX86_BUILTIN_PMOVUSQW512_MEM.
6771 (bdesc_special_args): Add __builtin_ia32_pmovusqd512mem_mask,
6772 __builtin_ia32_pmovsqd512mem_mask,
6773 __builtin_ia32_pmovqd512mem_mask,
6774 __builtin_ia32_pmovusqw512mem_mask,
6775 __builtin_ia32_pmovsqw512mem_mask,
6776 __builtin_ia32_pmovqw512mem_mask,
6777 __builtin_ia32_pmovusdw512mem_mask,
6778 __builtin_ia32_pmovsdw512mem_mask,
6779 __builtin_ia32_pmovdw512mem_mask,
6780 __builtin_ia32_pmovqb512mem_mask,
6781 __builtin_ia32_pmovusqb512mem_mask,
6782 __builtin_ia32_pmovsqb512mem_mask,
6783 __builtin_ia32_pmovusdb512mem_mask,
6784 __builtin_ia32_pmovsdb512mem_mask,
6785 __builtin_ia32_pmovdb512mem_mask.
6786 (bdesc_args): Add __builtin_ia32_expanddf512,
6787 __builtin_ia32_expandsf512.
6788 (ix86_expand_special_args_builtin): Handle VOID_FTYPE_PV8SI_V8DI_QI,
6789 VOID_FTYPE_PV8HI_V8DI_QI, VOID_FTYPE_PV16HI_V16SI_HI,
6790 VOID_FTYPE_PV16QI_V8DI_QI, VOID_FTYPE_PV16QI_V16SI_HI.
6791 * config/i386/sse.md (unspec): Add UNSPEC_EXPAND_NOMASK.
6792 (avx512f_<code><pmov_src_lower><mode>2_mask_store): New.
6793 (*avx512f_<code>v8div16qi2_store_mask): Renamed to ...
6794 (avx512f_<code>v8div16qi2_mask_store): This.
6795 (avx512f_expand<mode>): New.
6797 2014-01-27 Kirill Yukhin <kirill.yukhin@intel.com>
6799 * config/i386/avx512pfintrin.h (_mm512_mask_prefetch_i32gather_pd):
6801 (_mm512_mask_prefetch_i64gather_pd): Ditto.
6802 (_mm512_prefetch_i32scatter_pd): Ditto.
6803 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
6804 (_mm512_prefetch_i64scatter_pd): Ditto.
6805 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
6806 (_mm512_mask_prefetch_i32gather_ps): Fix operand type.
6807 (_mm512_mask_prefetch_i64gather_ps): Ditto.
6808 (_mm512_prefetch_i32scatter_ps): Ditto.
6809 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
6810 (_mm512_prefetch_i64scatter_ps): Ditto.
6811 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
6812 * config/i386/i386-builtin-types.def: Define
6813 VOID_FTYPE_QI_V8SI_PCINT64_INT_INT
6814 and VOID_FTYPE_QI_V8DI_PCINT64_INT_INT.
6815 * config/i386/i386.c (ix86_builtins): Define IX86_BUILTIN_GATHERPFQPD,
6816 IX86_BUILTIN_GATHERPFDPD, IX86_BUILTIN_SCATTERPFDPD,
6817 IX86_BUILTIN_SCATTERPFQPD.
6818 (ix86_init_mmx_sse_builtins): Define __builtin_ia32_gatherpfdpd,
6819 __builtin_ia32_gatherpfdps, __builtin_ia32_gatherpfqpd,
6820 __builtin_ia32_gatherpfqps, __builtin_ia32_scatterpfdpd,
6821 __builtin_ia32_scatterpfdps, __builtin_ia32_scatterpfqpd,
6822 __builtin_ia32_scatterpfqps.
6823 (ix86_expand_builtin): Expand new built-ins.
6824 * config/i386/sse.md (avx512pf_gatherpf<mode>): Add SF suffix,
6825 fix memory access data type.
6826 (*avx512pf_gatherpf<mode>_mask): Ditto.
6827 (*avx512pf_gatherpf<mode>): Ditto.
6828 (avx512pf_scatterpf<mode>): Ditto.
6829 (*avx512pf_scatterpf<mode>_mask): Ditto.
6830 (*avx512pf_scatterpf<mode>): Ditto.
6831 (GATHER_SCATTER_SF_MEM_MODE): New.
6832 (avx512pf_gatherpf<mode>df): Ditto.
6833 (*avx512pf_gatherpf<mode>df_mask): Ditto.
6834 (*avx512pf_scatterpf<mode>df): Ditto.
6836 2014-01-27 Jakub Jelinek <jakub@redhat.com>
6839 * expmed.h (expmed_mode_index): Rework so that analysis and optimziers
6840 know when the MODE_PARTIAL_INT and MODE_VECTOR_INT cases can never be
6843 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
6845 * common/config/arm/arm-common.c
6846 (arm_rewrite_mcpu): Handle multiple names.
6848 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
6850 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
6852 * gimple-builder.h (create_gimple_tmp): Delete.
6854 2014-01-27 Christian Bruel <christian.bruel@st.com>
6856 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Fix remaining bytes after
6859 2014-01-26 John David Anglin <danglin@gcc.gnu.org>
6861 * config/pa/pa.md (call): Generate indirect long calls to non-local
6862 functions when outputing 32-bit code.
6863 (call_value): Likewise except for special call to buggy powf function.
6865 * config/pa/pa.c (pa_attr_length_indirect_call): Adjust length of
6866 portable runtime and PIC indirect calls.
6867 (pa_output_indirect_call): Remove unnecessary nop from portable runtime
6868 and PIC call sequences. Use ldo instead of blr to set return register
6869 in PIC call sequence.
6871 2014-01-25 Walter Lee <walt@tilera.com>
6873 * config/tilegx/sync.md (atomic_fetch_sub): Fix negation and
6874 avoid clobbering a live register.
6876 2014-01-25 Walter Lee <walt@tilera.com>
6878 * config/tilegx/tilegx-c.c (tilegx_cpu_cpp_builtins):
6879 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2}.
6880 * config/tilegx/tilepro-c.c (tilepro_cpu_cpp_builtins):
6881 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2,4,8}.
6883 2014-01-25 Walter Lee <walt@tilera.com>
6885 * config/tilegx/tilegx.c (tilegx_function_arg): Start 16-byte
6886 arguments on even registers.
6887 (tilegx_gimplify_va_arg_expr): Align 16-byte var args to
6889 * config/tilegx/tilegx.h (STACK_BOUNDARY): Change to 16 bytes.
6890 (BIGGEST_ALIGNMENT): Ditto.
6891 (BIGGEST_FIELD_ALIGNMENT): Ditto.
6893 2014-01-25 Walter Lee <walt@tilera.com>
6895 * config/tilegx/tilegx.c (tilegx_gen_bundles): Delete barrier
6896 insns before bundling.
6897 * config/tilegx/tilegx.md (tile_network_barrier): Update comment.
6899 2014-01-25 Walter Lee <walt@tilera.com>
6901 * config/tilegx/tilegx.c (tilegx_expand_builtin): Set
6902 PREFETCH_SCHEDULE_BARRIER_P to true for prefetches.
6903 * config/tilepro/tilepro.c (tilepro_expand_builtin): Ditto.
6905 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
6907 * config/mips/constraints.md (kl): Delete.
6908 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Turn into
6909 define expands, using...
6910 (divmod<mode>4_mips16, udivmod<mode>4_mips16): ...these new
6911 instructions for MIPS16.
6912 (*divmod<mode>4, *udivmod<mode>4): New patterns, taken from the
6913 non-MIPS16 version of the old divmod<mode>4 and udivmod<mode>4.
6915 2014-01-25 Walter Lee <walt@tilera.com>
6917 * config/tilepro/tilepro.md (ctzdi2): Use register_operand predicate.
6921 2014-01-25 Walter Lee <walt@tilera.com>
6923 * config/tilegx/tilegx.c (tilegx_expand_to_rtl_hook): New.
6924 (TARGET_EXPAND_TO_RTL_HOOK): Define.
6926 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
6928 * rtlanal.c (canonicalize_condition): Split out duplicated mode check.
6931 2014-01-25 Jakub Jelinek <jakub@redhat.com>
6933 * print-rtl.c (in_call_function_usage): New var.
6934 (print_rtx): When in CALL_INSN_FUNCTION_USAGE, always print
6935 EXPR_LIST mode as mode and not as reg note name.
6938 * cfgloopmanip.c (copy_loop_info): If
6939 loop->warned_aggressive_loop_optimizations, make sure
6940 the flag is set in target loop too.
6942 2014-01-24 Balaji V. Iyer <balaji.v.iyer@intel.com>
6944 * builtins.c (is_builtin_name): Renamed flag_enable_cilkplus to
6946 * builtins.def: Likewise.
6947 * cilk.h (fn_contains_cilk_spawn_p): Likewise.
6948 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Likewise.
6949 * ira.c (ira_setup_eliminable_regset): Likewise.
6950 * omp-low.c (gate_expand_omp): Likewise.
6951 (execute_lower_omp): Likewise.
6952 (diagnose_sb_0): Likewise.
6953 (gate_diagnose_omp_blocks): Likewise.
6954 (simd_clone_clauses_extract): Likewise.
6957 2014-01-24 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6959 * config/rs6000/rs6000.c (rs6000_expand_vec_perm_const_1): Remove
6960 correction for little endian...
6961 * config/rs6000/vsx.md (vsx_xxpermdi2_<mode>_1): ...and move it to
6964 2014-01-24 Jeff Law <law@redhat.com>
6966 PR tree-optimization/59919
6967 * tree-vrp.c (find_assert_locations_1): Do not register asserts
6968 for non-returning calls.
6970 2014-01-24 James Greenhalgh <james.greenhalgh@arm.com>
6972 * common/config/aarch64/aarch64-common.c
6973 (aarch64_rewrite_mcpu): Handle multiple names.
6974 * config/aarch64/aarch64.h
6975 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
6977 2014-01-24 Dodji Seketeli <dodji@redhat.com>
6979 * input.c (add_file_to_cache_tab): Handle the case where fopen
6982 2014-01-23 H.J. Lu <hongjiu.lu@intel.com>
6985 * config/i386/i386.md (pushsf splitter): Get stack adjustment
6986 from push operand if code of push isn't PRE_DEC.
6988 2014-01-23 Michael Meissner <meissner@linux.vnet.ibm.com>
6991 * doc/invoke.texi (RS/6000 and PowerPC Options): Document
6992 -mquad-memory-atomic. Update -mquad-memory documentation to say
6993 it is only used for non-atomic loads/stores.
6995 * config/rs6000/predicates.md (quad_int_reg_operand): Allow either
6996 -mquad-memory or -mquad-memory-atomic switches.
6998 * config/rs6000/rs6000-cpus.def (ISA_2_7_MASKS_SERVER): Add
6999 -mquad-memory-atomic to ISA 2.07 support.
7001 * config/rs6000/rs6000.opt (-mquad-memory-atomic): Add new switch
7002 to separate support of normal quad word memory operations (ldq, stq)
7003 from the atomic quad word memory operations.
7005 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
7006 support to separate non-atomic quad word operations from atomic
7007 quad word operations. Disable non-atomic quad word operations in
7008 little endian mode so that we don't have to swap words after the
7009 load and before the store.
7010 (quad_load_store_p): Add comment about atomic quad word support.
7011 (rs6000_opt_masks): Add -mquad-memory-atomic to the list of
7012 options printed with -mdebug=reg.
7014 * config/rs6000/rs6000.h (TARGET_SYNC_TI): Use
7015 -mquad-memory-atomic as the test for whether we have quad word
7016 atomic instructions.
7017 (TARGET_SYNC_HI_QI): If either -mquad-memory-atomic, -mquad-memory,
7018 or -mp8-vector are used, allow byte/half-word atomic operations.
7020 * config/rs6000/sync.md (load_lockedti): Insure that the address
7021 is a proper indexed or indirect address for the lqarx instruction.
7022 On little endian systems, swap the hi/lo registers after the lqarx
7024 (load_lockedpti): Use indexed_or_indirect_operand predicate to
7025 insure the address is valid for the lqarx instruction.
7026 (store_conditionalti): Insure that the address is a proper indexed
7027 or indirect address for the stqcrx. instruction. On little endian
7028 systems, swap the hi/lo registers before doing the stqcrx.
7030 (store_conditionalpti): Use indexed_or_indirect_operand predicate to
7031 insure the address is valid for the stqcrx. instruction.
7033 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
7034 Define __QUAD_MEMORY__ and __QUAD_MEMORY_ATOMIC__ based on what
7035 type of quad memory support is available.
7037 2014-01-23 Vladimir Makarov <vmakarov@redhat.com>
7040 * lra-constraints.c (simplify_operand_subreg): Spill pseudo if
7041 there is a danger of looping.
7043 2014-01-23 Pat Haugen <pthaugen@us.ibm.com>
7045 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
7046 force flag_ira_loop_pressure if set via command line.
7048 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
7050 * config/aarch64/aarch64-simd-builtins.def (ashr): DI mode removed.
7051 (ashr_simd): New builtin handling DI mode.
7052 * config/aarch64/aarch64-simd.md (aarch64_ashr_simddi): New pattern.
7053 (aarch64_sshr_simddi): New match pattern.
7054 * config/aarch64/arm_neon.h (vshr_n_s32): Builtin call modified.
7055 (vshrd_n_s64): Likewise.
7056 * config/aarch64/predicates.md (aarch64_shift_imm64_di): New predicate.
7058 2014-01-23 Nick Clifton <nickc@redhat.com>
7060 * config/msp430/msp430.h (ASM_SPEC): Pass the -mcpu as -mcpu.
7061 (LIB_SPEC): Drop use of memory.ld and peripherals.ld scripts in
7062 favour of mcu specific scripts.
7063 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more matches for
7066 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
7067 Alex Velenko <Alex.Velenko@arm.com>
7069 * config/aarch64/arm_neon.h (vaddv_s8): __LANE0 cleanup.
7070 (vaddv_s16): Likewise.
7071 (vaddv_s32): Likewise.
7072 (vaddv_u8): Likewise.
7073 (vaddv_u16): Likewise.
7074 (vaddv_u32): Likewise.
7075 (vaddvq_s8): Likewise.
7076 (vaddvq_s16): Likewise.
7077 (vaddvq_s32): Likewise.
7078 (vaddvq_s64): Likewise.
7079 (vaddvq_u8): Likewise.
7080 (vaddvq_u16): Likewise.
7081 (vaddvq_u32): Likewise.
7082 (vaddvq_u64): Likewise.
7083 (vaddv_f32): Likewise.
7084 (vaddvq_f32): Likewise.
7085 (vaddvq_f64): Likewise.
7086 (vmaxv_f32): Likewise.
7087 (vmaxv_s8): Likewise.
7088 (vmaxv_s16): Likewise.
7089 (vmaxv_s32): Likewise.
7090 (vmaxv_u8): Likewise.
7091 (vmaxv_u16): Likewise.
7092 (vmaxv_u32): Likewise.
7093 (vmaxvq_f32): Likewise.
7094 (vmaxvq_f64): Likewise.
7095 (vmaxvq_s8): Likewise.
7096 (vmaxvq_s16): Likewise.
7097 (vmaxvq_s32): Likewise.
7098 (vmaxvq_u8): Likewise.
7099 (vmaxvq_u16): Likewise.
7100 (vmaxvq_u32): Likewise.
7101 (vmaxnmv_f32): Likewise.
7102 (vmaxnmvq_f32): Likewise.
7103 (vmaxnmvq_f64): Likewise.
7104 (vminv_f32): Likewise.
7105 (vminv_s8): Likewise.
7106 (vminv_s16): Likewise.
7107 (vminv_s32): Likewise.
7108 (vminv_u8): Likewise.
7109 (vminv_u16): Likewise.
7110 (vminv_u32): Likewise.
7111 (vminvq_f32): Likewise.
7112 (vminvq_f64): Likewise.
7113 (vminvq_s8): Likewise.
7114 (vminvq_s16): Likewise.
7115 (vminvq_s32): Likewise.
7116 (vminvq_u8): Likewise.
7117 (vminvq_u16): Likewise.
7118 (vminvq_u32): Likewise.
7119 (vminnmv_f32): Likewise.
7120 (vminnmvq_f32): Likewise.
7121 (vminnmvq_f64): Likewise.
7123 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
7125 * config/aarch64/aarch64-simd.md
7126 (aarch64_dup_lane<mode>): Correct lane number on big-endian.
7127 (aarch64_dup_lane_<vswap_widthi_name><mode>): Likewise.
7128 (*aarch64_mul3_elt<mode>): Likewise.
7129 (*aarch64_mul3_elt<vswap_width_name><mode>): Likewise.
7130 (*aarch64_mul3_elt_to_64v2df): Likewise.
7131 (*aarch64_mla_elt<mode>): Likewise.
7132 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
7133 (*aarch64_mls_elt<mode>): Likewise.
7134 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
7135 (*aarch64_fma4_elt<mode>): Likewise.
7136 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
7137 (*aarch64_fma4_elt_to_64v2df): Likewise.
7138 (*aarch64_fnma4_elt<mode>): Likewise.
7139 (*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
7140 (*aarch64_fnma4_elt_to_64v2df): Likewise.
7141 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
7142 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
7143 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
7144 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
7145 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
7146 (aarch64_sqdmull_lane<mode>_internal): Likewise.
7147 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
7149 2013-01-23 Alex Velenko <Alex.Velenko@arm.com>
7151 * config/aarch64/aarch64-simd.md
7152 (aarch64_be_checked_get_lane<mode>): New define_expand.
7153 * config/aarch64/aarch64-simd-builtins.def
7154 (BUILTIN_VALL (GETLANE, be_checked_get_lane, 0)):
7155 New builtin definition.
7156 * config/aarch64/arm_neon.h: (__aarch64_vget_lane_any):
7157 Use new safe be builtin.
7159 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
7161 * config/aarch64/aarch64-simd.md (aarch64_be_ld1<mode>):
7163 (aarch64_be_st1<mode>): Likewise.
7164 (aarch_ld1<VALL:mode>): Define_expand modified.
7165 (aarch_st1<VALL:mode>): Likewise.
7166 * config/aarch64/aarch64.md (UNSPEC_LD1): New unspec definition.
7167 (UNSPEC_ST1): Likewise.
7169 2014-01-23 David Holsgrove <david.holsgrove@xilinx.com>
7171 * config/microblaze/microblaze.md: Add trap insn and attribute
7173 2014-01-23 Dodji Seketeli <dodji@redhat.com>
7175 PR preprocessor/58580
7176 * input.h (location_get_source_line): Take an additional line_size
7178 (void diagnostics_file_cache_fini): Declare new function.
7179 * input.c (struct fcache): New type.
7180 (fcache_tab_size, fcache_buffer_size, fcache_line_record_size):
7181 New static constants.
7182 (diagnostic_file_cache_init, total_lines_num)
7183 (lookup_file_in_cache_tab, evicted_cache_tab_entry)
7184 (add_file_to_cache_tab, lookup_or_add_file_to_cache_tab)
7185 (needs_read, needs_grow, maybe_grow, read_data, maybe_read_data)
7186 (get_next_line, read_next_line, goto_next_line, read_line_num):
7187 New static function definitions.
7188 (diagnostic_file_cache_fini): New function.
7189 (location_get_source_line): Take an additional output line_len
7190 parameter. Re-write using lookup_or_add_file_to_cache_tab and
7192 * diagnostic.c (diagnostic_finish): Call
7193 diagnostic_file_cache_fini.
7194 (adjust_line): Take an additional input parameter for the length
7195 of the line, rather than calculating it with strlen.
7196 (diagnostic_show_locus): Adjust the use of
7197 location_get_source_line and adjust_line with respect to their new
7198 signature. While displaying a line now, do not stop at the first
7199 null byte. Rather, display the zero byte as a space and keep
7200 going until we reach the size of the line.
7201 * Makefile.in: Add vec.o to OBJS-libcommon
7203 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
7204 Ilya Tocar <ilya.tocar@intel.com>
7206 * config/i386/avx512fintrin.h (_mm512_kmov): New.
7207 * config/i386/i386.c (IX86_BUILTIN_KMOV16): Ditto.
7208 (__builtin_ia32_kmov16): Ditto.
7209 * config/i386/i386.md (UNSPEC_KMOV): New.
7212 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
7214 * config/i386/avx512fintrin.h (_mm512_loadu_si512): Rename.
7215 (_mm512_storeu_si512): Ditto.
7217 2014-01-23 Richard Sandiford <rdsandiford@googlemail.com>
7220 * rtl.h (get_referenced_operands): Declare.
7221 * recog.c (get_referenced_operands): New function.
7222 * config/mips/mips.c (mips_reorg_process_insns): Check which asm
7223 operands have been referenced when recording LO_SUM references.
7225 2014-01-22 David Holsgrove <david.holsgrove@xilinx.com>
7227 * config/microblaze/microblaze.md: Correct bswaphi2 insn.
7229 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
7231 * config/i386/x86-tune.def (X86_TUNE_ACCUMULATE_OUTGOING_ARGS):
7232 Enable for generic and recent AMD targets.
7234 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
7236 * combine-stack-adj.c (combine_stack_adjustments_for_block): Remove
7237 ARG_SIZE note when adjustment was eliminated.
7239 2014-01-22 Jeff Law <law@redhat.com>
7241 PR tree-optimization/59597
7242 * tree-ssa-threadupdate.c (dump_jump_thread_path): Move to earlier
7243 in file. Accept new argument REGISTERING and use it to modify
7244 dump output appropriately.
7245 (register_jump_thread): Corresponding changes.
7246 (mark_threaded_blocks): Reinstate code to cancel unprofitable
7247 thread paths involving joiner blocks. Add code to dump cancelled
7248 jump threading paths.
7250 2014-01-22 Vladimir Makarov <vmakarov@redhat.com>
7252 PR rtl-optimization/59477
7253 * lra-constraints.c (inherit_in_ebb): Process call for living hard
7254 regs. Update reloads_num and potential_reload_hard_regs for all insns.
7256 2014-01-22 Tom Tromey <tromey@redhat.com>
7258 * config/i386/i386-interix.h (i386_pe_unique_section): Don't use
7260 * config/cr16/cr16-protos.h (notice_update_cc): Don't use PARAMS.
7262 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
7264 PR rtl-optimization/59896
7265 * lra-constraints.c (process_alt_operands): Check unused note for
7266 matched operands of insn with no output reloads.
7268 2014-01-21 Richard Sandiford <rdsandiford@googlemail.com>
7270 * config/mips/mips.c (mips_move_to_gpr_cost): Add M16_REGS case.
7271 (mips_move_from_gpr_cost): Likewise.
7273 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
7275 PR rtl-optimization/59858
7276 * lra-constraints.c (SMALL_REGISTER_CLASS_P): Use
7277 ira_class_hard_regs_num.
7278 (process_alt_operands): Increase reject for dying matched operand.
7280 2014-01-21 Jakub Jelinek <jakub@redhat.com>
7283 * config/i386/i386.c (expand_small_movmem_or_setmem): If mode is
7284 smaller than size, perform several stores or loads and stores
7285 at dst + count - size to store or copy all of size bytes, rather
7286 than just last modesize bytes.
7288 2014-01-20 DJ Delorie <dj@redhat.com>
7290 * config/rl78/rl78.c (rl78_propogate_register_origins): Verify
7291 that CLOBBERs are REGs before propogating their values.
7293 2014-01-20 H.J. Lu <hongjiu.lu@intel.com>
7296 * cgraph.c (cgraph_inline_failed_string): Add type to DEFCIFCODE.
7297 (cgraph_inline_failed_type): New function.
7298 * cgraph.h (DEFCIFCODE): Add type.
7299 (cgraph_inline_failed_type_t): New enum.
7300 (cgraph_inline_failed_type): New prototype.
7301 * cif-code.def: Add CIF_FINAL_NORMAL to OK, FUNCTION_NOT_CONSIDERED,
7302 FUNCTION_NOT_OPTIMIZED, REDEFINED_EXTERN_INLINE,
7303 FUNCTION_NOT_INLINE_CANDIDATE, LARGE_FUNCTION_GROWTH_LIMIT,
7304 LARGE_STACK_FRAME_GROWTH_LIMIT, MAX_INLINE_INSNS_SINGLE_LIMIT,
7305 MAX_INLINE_INSNS_AUTO_LIMIT, INLINE_UNIT_GROWTH_LIMIT,
7306 RECURSIVE_INLINING, UNLIKELY_CALL, NOT_DECLARED_INLINED,
7307 OPTIMIZING_FOR_SIZE, ORIGINALLY_INDIRECT_CALL,
7308 INDIRECT_UNKNOWN_CALL, USES_COMDAT_LOCAL.
7309 Add CIF_FINAL_ERROR to UNSPECIFIED, BODY_NOT_AVAILABLE,
7310 FUNCTION_NOT_INLINABLE, OVERWRITABLE, MISMATCHED_ARGUMENTS,
7311 EH_PERSONALITY, NON_CALL_EXCEPTIONS, TARGET_OPTION_MISMATCH,
7312 OPTIMIZATION_MISMATCH.
7313 * tree-inline.c (expand_call_inline): Emit errors during
7314 early_inlining if cgraph_inline_failed_type returns CIF_FINAL_ERROR.
7316 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
7319 * config/i386/sse.md (*andnot<mode>3<mask_name>): Handle MODE_V16SF
7320 mode attribute in insn output.
7322 2014-01-20 Eric Botcazou <ebotcazou@adacore.com>
7324 * output.h (output_constant): Delete.
7325 * varasm.c (output_constant): Make private.
7327 2014-01-20 Alex Velenko <Alex.Velenko@arm.com>
7329 * config/aarch64/aarch64-simd.md (vec_perm<mode>): Add BE check.
7331 2014-01-20 Jakub Jelinek <jakub@redhat.com>
7334 * tree.h (fold_builtin_strcat): New prototype.
7335 * builtins.c (fold_builtin_strcat): No longer static. Add len
7336 argument, if non-NULL, don't call c_strlen. Optimize
7337 directly into __builtin_memcpy instead of __builtin_strcpy.
7338 (fold_builtin_2): Adjust fold_builtin_strcat caller.
7339 * gimple-fold.c (gimple_fold_builtin): Handle BUILT_IN_STRCAT.
7341 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
7343 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
7344 for SImode_address_operand operands, having only a REG argument.
7346 2014-01-20 Marcus Shawcroft <marcus.shawcroft@arm.com>
7348 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER): Expand
7349 loader name using mbig-endian.
7350 (LINUX_TARGET_LINK_SPEC): Pass linker -m flag.
7352 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
7354 * doc/invoke.texi (-march): Clarify documentation for AArch64.
7358 2014-01-20 Tejas Belagod <tejas.belagod@arm.com>
7360 * config/aarch64/aarch64-protos.h
7361 (aarch64_cannot_change_mode_class_ptr): Declare.
7362 * config/aarch64/aarch64.c (aarch64_cannot_change_mode_class,
7363 aarch64_cannot_change_mode_class_ptr): New.
7364 * config/aarch64/aarch64.h (CANNOT_CHANGE_MODE_CLASS): Change to call
7365 backend hook aarch64_cannot_change_mode_class.
7367 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
7369 * common/config/aarch64/aarch64-common.c
7370 (aarch64_handle_option): Don't handle any option order logic here.
7371 * config/aarch64/aarch64.c (aarch64_parse_arch): Do not override
7372 selected_cpu, warn on architecture version mismatch.
7373 (aarch64_override_options): Fix parsing order for option strings.
7375 2014-01-20 Jan-Benedict Glaw <jbglaw@lug-owl.de>
7376 Iain Sandoe <iain@codesourcery.com>
7379 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Fix unused variable
7380 warning. Amend comment to reflect current functionality.
7382 2014-01-20 Richard Biener <rguenther@suse.de>
7385 * builtins.c (fold_builtin_strcat): Remove case better handled
7386 by tree-ssa-strlen.c.
7388 2014-01-20 Alan Lawrence <alan.lawrence@arm.com>
7390 * config/aarch64/aarch64.opt
7391 (mcpu, march, mtune): Make case-insensitive.
7393 2014-01-20 Jakub Jelinek <jakub@redhat.com>
7396 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
7397 if operands[1] is a REG or ZERO_EXTEND of a REG.
7399 2014-01-19 Jan Hubicka <hubicka@ucw.cz>
7401 * varasm.c (compute_reloc_for_constant): Use targetm.binds_local_p.
7403 2014-01-19 John David Anglin <danglin@gcc.gnu.org>
7405 * config/pa/pa.c (pa_attr_length_millicode_call): Correct length of
7406 long non-pic millicode calls.
7408 2014-01-19 Jan-Benedict Glaw <jbglaw@lug-owl.de>
7410 * config/vax/vax.h (FUNCTION_ARG_REGNO_P): Fix unused variable warning.
7412 2014-01-19 Kito Cheng <kito@0xlab.org>
7414 * builtins.c (expand_movstr): Check movstr expand done or fail.
7416 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
7417 H.J. Lu <hongjiu.lu@intel.com>
7420 * config/i386/i386.md (*lea<mode>): Zero-extend return register
7421 to DImode for zero-extended addresses.
7423 2014-01-19 Jakub Jelinek <jakub@redhat.com>
7425 PR rtl-optimization/57763
7426 * bb-reorder.c (fix_crossing_unconditional_branches): Set JUMP_LABEL
7427 on the new indirect jump_insn and increment LABEL_NUSES (label).
7429 2014-01-18 H.J. Lu <hongjiu.lu@intel.com>
7433 * config.gcc (x86_archs): New variable.
7434 (x86_64_archs): Likewise.
7435 (x86_cpus): Likewise.
7436 Use $x86_archs, $x86_64_archs and $x86_cpus to check valid
7437 --with-arch/--with-cpu= options.
7438 Support --with-arch=/--with-cpu={nehalem,westmere,
7439 sandybridge,ivybridge,haswell,broadwell,bonnell,silvermont}.
7441 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
7443 * config/i386/i386.c (ix86_adjust_cost): Reorder PROCESSOR_K8
7444 and PROCESSOR_ATHLON to simplify code. Move "memory" calculation.
7446 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
7448 * config/i386/i386.md (*swap<mode>): Rename from swap<mode>.
7450 2014-01-18 Jakub Jelinek <jakub@redhat.com>
7453 * config/i386/i386-c.c (ix86_pragma_target_parse): Temporarily
7454 clear cpp_get_options (parse_in)->warn_unused_macros for
7455 ix86_target_macros_internal with cpp_define.
7457 2014-01-18 Richard Sandiford <rdsandiford@googlemail.com>
7459 * jump.c (delete_related_insns): Keep (use (insn))s.
7460 * reorg.c (redundant_insn): Check for barriers too.
7462 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
7464 * config/i386/i386.c (ix86_split_lea_for_addr): Fix a comment typo.
7466 2014-01-17 John David Anglin <danglin@gcc.gnu.org>
7468 * config/pa/pa.c (pa_attr_length_indirect_call): Don't output a short
7469 call to $$dyncall when TARGET_LONG_CALLS is true.
7471 2014-01-17 Jeff Law <law@redhat.com>
7473 * ree.c (combine_set_extension): Temporarily disable test for
7474 changing number of hard registers.
7476 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
7479 * ipa-inline-analysis.c (inline_free_summary):
7480 Do not free summary of aliases.
7482 2014-01-17 Jakub Jelinek <jakub@redhat.com>
7485 * gimplify.c (gimplify_expr): Use create_tmp_var
7486 instead of create_tmp_var_raw. If cond doesn't have
7487 integral type, don't add the IFN_ANNOTATE builtin at all.
7489 2014-01-17 Martin Jambor <mjambor@suse.cz>
7492 * ipa-cp.c (prev_edge_clone): New variable.
7493 (grow_next_edge_clone_vector): Renamed to grow_edge_clone_vectors.
7494 Also resize prev_edge_clone vector.
7495 (ipcp_edge_duplication_hook): Also update prev_edge_clone.
7496 (ipcp_edge_removal_hook): New function.
7497 (ipcp_driver): Register ipcp_edge_removal_hook.
7499 2014-01-17 Andrew Pinski <apinski@cavium.com>
7500 Steve Ellcey <sellcey@mips.com>
7503 * config/mips/mips.c (mips_print_operand): Check operand mode instead
7506 2014-01-17 Jeff Law <law@redhat.com>
7509 * passes.def: Reorder pass_copy_prop, pass_unrolli, pass_ccp sequence
7510 so that pass_ccp runs first.
7512 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
7514 * config/i386/i386.c (ix86_lea_outperforms): Use TARGET_XXX.
7515 (ix86_adjust_cost): Use !TARGET_XXX.
7516 (do_reorder_for_imul): Likewise.
7517 (swap_top_of_ready_list): Likewise.
7518 (ix86_sched_reorder): Likewise.
7520 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
7522 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
7523 PROCESSOR_INTEL. Treat like PROCESSOR_GENERIC.
7524 * config/i386/i386.c (intel_memcpy): New. Duplicate slm_memcpy.
7525 (intel_memset): New. Duplicate slm_memset.
7526 (intel_cost): New. Duplicate slm_cost.
7527 (m_INTEL): New macro.
7528 (processor_target_table): Add "intel".
7529 (ix86_option_override_internal): Replace PROCESSOR_SILVERMONT
7530 with PROCESSOR_INTEL for "intel".
7531 (ix86_lea_outperforms): Support PROCESSOR_INTEL. Duplicate
7532 PROCESSOR_SILVERMONT.
7533 (ix86_issue_rate): Likewise.
7534 (ix86_adjust_cost): Likewise.
7535 (ia32_multipass_dfa_lookahead): Likewise.
7536 (swap_top_of_ready_list): Likewise.
7537 (ix86_sched_reorder): Likewise.
7538 (ix86_avoid_lea_for_addr): Check TARGET_AVOID_LEA_FOR_ADDR
7539 instead of TARGET_OPT_AGU.
7540 * config/i386/i386.h (TARGET_INTEL): New.
7541 (TARGET_AVOID_LEA_FOR_ADDR): Likewise.
7542 (processor_type): Add PROCESSOR_INTEL.
7543 * config/i386/x86-tune.def: Support m_INTEL. Duplicate m_SILVERMONT.
7544 Add X86_TUNE_AVOID_LEA_FOR_ADDR.
7546 2014-01-17 Marek Polacek <polacek@redhat.com>
7549 * gimple-fold.c (fold_array_ctor_reference): Don't fold if element
7552 2014-01-17 Richard Biener <rguenther@suse.de>
7554 PR tree-optimization/46590
7555 * opts.c (default_options_table): Add entries for
7556 OPT_fbranch_count_reg, OPT_fmove_loop_invariants and OPT_ftree_pta,
7557 all enabled at -O1 but not for -Og.
7558 * common.opt (fbranch-count-reg): Remove Init(1).
7559 (fmove-loop-invariants): Likewise.
7560 (ftree-pta): Likewise.
7562 2014-01-17 Jakub Jelinek <jakub@redhat.com>
7564 * config/i386/i386.c (ix86_data_alignment): For compatibility with
7565 (incorrect) GCC 4.8 and earlier alignment assumptions ensure we align
7566 decls to at least the GCC 4.8 used alignments.
7569 * tree-nested.c (convert_nonlocal_reference_stmt,
7570 convert_local_reference_stmt): For NAMELIST_DECLs in gimple_bind_vars
7571 of GIMPLE_BIND stmts, adjust associated decls.
7573 2014-01-17 Richard Biener <rguenther@suse.de>
7575 PR tree-optimization/46590
7576 * vec.h (vec<>::bseach): New member function implementing
7577 binary search according to C89 bsearch.
7578 (vec<>::qsort): Avoid calling ::qsort for vectors with sizes 0 or 1.
7579 * tree-ssa-loop-im.c (struct mem_ref): Make stored member a
7580 bitmap pointer again. Make accesses_in_loop a flat array.
7581 (mem_ref_obstack): New global.
7582 (outermost_indep_loop): Adjust for mem_ref->stored changes.
7583 (mark_ref_stored): Likewise.
7584 (ref_indep_loop_p_2): Likewise.
7585 (set_ref_stored_in_loop): New helper function.
7586 (mem_ref_alloc): Allocate mem_refs on the mem_ref_obstack obstack.
7587 (memref_free): Adjust.
7588 (record_mem_ref_loc): Simplify.
7589 (gather_mem_refs_stmt): Adjust.
7590 (sort_locs_in_loop_postorder_cmp): New function.
7591 (analyze_memory_references): Sort accesses_in_loop after
7592 loop postorder number.
7593 (find_ref_loc_in_loop_cmp): New function.
7594 (for_all_locs_in_loop): Find relevant cluster of locs in
7595 accesses_in_loop and iterate without recursion.
7596 (execute_sm): Avoid uninit warning.
7597 (struct ref_always_accessed): Simplify.
7598 (ref_always_accessed::operator ()): Likewise.
7599 (ref_always_accessed_p): Likewise.
7600 (tree_ssa_lim_initialize): Initialize mem_ref_obstack, compute
7601 loop postorder numbers here.
7602 (tree_ssa_lim_finalize): Free mem_ref_obstack and loop postorder
7605 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
7608 * passes.c (rest_of_decl_compilation): Don't call varpool_finalize_decl
7609 on decls for which assemble_alias has been called.
7611 2014-01-17 Nick Clifton <nickc@redhat.com>
7613 * config/msp430/msp430.opt: (mcpu): New option.
7614 * config/msp430/msp430.c (msp430_mcu_name): Use target_mcu.
7615 (msp430_option_override): Parse target_cpu. If the MCU name
7616 matches a generic string, clear target_mcu.
7617 (msp430_attr): Allow numeric interrupt values up to 63.
7618 (msp430_expand_epilogue): No longer invert operand 1 of gen_popm.
7619 * config/msp430/msp430.h (ASM_SPEC): Convert -mcpu into a -mmcu
7621 * config/msp430/t-msp430: (MULTILIB_MATCHES): Remove mcu matches.
7623 * config/msp430/msp430.md (popm): Use %J rather than %I.
7624 (addsi3): Use msp430_nonimmediate_operand for operand 2.
7625 (addhi_cy_i): Use immediate_operand for operand 2.
7626 * doc/invoke.texi: Document -mcpu option.
7628 2014-01-17 Richard Biener <rguenther@suse.de>
7630 PR rtl-optimization/38518
7631 * df.h (df_analyze_loop): Declare.
7632 * df-core.c: Include cfgloop.h.
7633 (df_analyze_1): Split out main part of df_analyze.
7634 (df_analyze): Adjust.
7635 (loop_inverted_post_order_compute): New function.
7636 (loop_post_order_compute): Likewise.
7637 (df_analyze_loop): New function avoiding whole-function
7639 * loop-invariant.c (find_defs): Use df_analyze_loop.
7640 (find_invariants): Adjust.
7641 * loop-iv.c (iv_analysis_loop_init): Use df_analyze_loop.
7643 2014-01-17 Zhenqiang Chen <zhenqiang.chen@arm.com>
7645 * config/arm/arm.c (arm_v7m_tune): Set max_insns_skipped to 2.
7646 (thumb2_final_prescan_insn): Set max to MAX_INSN_PER_IT_BLOCK.
7648 2014-01-16 Ilya Enkovich <ilya.enkovich@intel.com>
7650 * ipa-ref.c (ipa_remove_stmt_references): Fix references
7651 traversal when removing references.
7653 2014-01-16 Jan Hubicka <hubicka@ucw.cz>
7656 * tree.c (get_binfo_at_offset): Look harder for virtual bases.
7658 2014-01-16 Bernd Schmidt <bernds@codesourcery.com>
7661 * reload.c (find_reloads_address_1): Do not use RELOAD_OTHER when
7662 pushing a reload for an autoinc when we had previously reloaded an
7663 inner part of the address.
7665 2014-01-16 Jakub Jelinek <jakub@redhat.com>
7667 * tree-vectorizer.h (struct _loop_vec_info): Add no_data_dependencies
7669 (LOOP_VINFO_NO_DATA_DEPENDENCIES): Define.
7670 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Clear it
7671 when not giving up or versioning for alias only because of
7673 (vect_analyze_data_ref_dependences): Set to true.
7674 * tree-vect-stmts.c (hoist_defs_of_uses): Return false if def_stmt
7676 (vectorizable_load): Use LOOP_VINFO_NO_DATA_DEPENDENCIES instead of
7677 LOOP_REQUIRES_VERSIONING_FOR_ALIAS, add && !nested_in_vect_loop
7681 * expr.c (expand_expr_real_1): Handle init == NULL_TREE.
7684 * config/i386/i386.c (ix86_expand_builtin): If target doesn't satisfy
7685 operand 0 predicate for gathers, use a new pseudo as subtarget.
7687 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
7690 * lra-constraints.c (process_alt_operands): Add printing debug info.
7691 Check absence of input/output reloads for matched operands too.
7693 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
7695 PR rtl-optimization/59835
7696 * ira.c (ira_init_register_move_cost): Increase cost for
7699 2014-01-16 Alan Lawrence <alan.lawrence@arm.com>
7701 * config/arm/arm.opt (mcpu, march, mtune): Make case-insensitive.
7703 2014-01-16 Richard Earnshaw <rearnsha@arm.com>
7706 * aarch64.c (aarch64_split_128bit_move): Don't lookup REGNO on
7707 non-register objects. Use gen_(high/low)part more consistently.
7710 2014-01-16 Michael Meissner <meissner@linux.vnet.ibm.com>
7713 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Add little
7714 endian support, remove tests for WORDS_BIG_ENDIAN.
7715 (p8_mfvsrd_3_<mode>): Likewise.
7716 (reload_gpr_from_vsx<mode>): Likewise.
7717 (reload_gpr_from_vsxsf): Likewise.
7718 (p8_mfvsrd_4_disf): Likewise.
7720 2014-01-16 Richard Biener <rguenther@suse.de>
7722 PR rtl-optimization/46590
7723 * lcm.c (compute_antinout_edge): Use postorder iteration.
7724 (compute_laterin): Use inverted postorder iteration.
7726 2014-01-16 Nick Clifton <nickc@redhat.com>
7729 * varasm.c (output_constant): Return the number of bytes actually
7731 (output_constructor_array_range): Update the field size with the
7732 number of bytes emitted by output_constant.
7733 (output_constructor_regular_field): Likewise. Also do not
7734 complain if the total number of bytes emitted is now greater
7735 than the expected fieldpos.
7736 * output.h (output_constant): Update prototype and descriptive comment.
7738 2014-01-16 Marek Polacek <polacek@redhat.com>
7741 * cgraph.c (gimple_check_call_args): Don't use DECL_ARG_TYPE if
7742 it is error_mark_node.
7744 2014-01-15 Uros Bizjak <ubizjak@gmail.com>
7746 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use
7747 VALID_AVX256_REG_OR_OI_MODE.
7749 2014-01-15 Pat Haugen <pthaugen@us.ibm.com>
7751 * config/rs6000/rs6000.c (rs6000_output_function_prologue): Check if
7752 current procedure should be profiled.
7754 2014-01-15 Andrew Pinski <apinski@cavium.com>
7756 * config/aarch64/aarch64.c (aarch64_register_move_cost): Correct cost
7757 of moving from/to the STACK_REG register class.
7759 2014-01-15 Richard Henderson <rth@redhat.com>
7762 * reginfo.c (global_regs_decl): Globalize.
7763 * rtl.h (global_regs_decl): Declare.
7764 * ira.c (do_reload): Diagnose frame_pointer_needed and it
7765 reserved via global_regs.
7767 2014-01-15 Teresa Johnson <tejohnson@google.com>
7769 * tree-ssa-sccvn.c (visit_reference_op_call): Handle NULL vdef.
7771 2014-01-15 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
7773 * config/rs6000/altivec.md (mulv8hi3): Explicitly generate vmulesh
7774 and vmulosh rather than call gen_vec_widen_smult_*.
7775 (vec_widen_umult_even_v16qi): Test VECTOR_ELT_ORDER_BIG rather
7776 than BYTES_BIG_ENDIAN to determine use of even or odd instruction.
7777 (vec_widen_smult_even_v16qi): Likewise.
7778 (vec_widen_umult_even_v8hi): Likewise.
7779 (vec_widen_smult_even_v8hi): Likewise.
7780 (vec_widen_umult_odd_v16qi): Likewise.
7781 (vec_widen_smult_odd_v16qi): Likewise.
7782 (vec_widen_umult_odd_v8hi): Likewise.
7783 (vec_widen_smult_odd_v8hi): Likewise.
7784 (vec_widen_umult_hi_v16qi): Explicitly generate vmuleub and
7785 vmuloub rather than call gen_vec_widen_umult_*.
7786 (vec_widen_umult_lo_v16qi): Likewise.
7787 (vec_widen_smult_hi_v16qi): Explicitly generate vmulesb and
7788 vmulosb rather than call gen_vec_widen_smult_*.
7789 (vec_widen_smult_lo_v16qi): Likewise.
7790 (vec_widen_umult_hi_v8hi): Explicitly generate vmuleuh and vmulouh
7791 rather than call gen_vec_widen_umult_*.
7792 (vec_widen_umult_lo_v8hi): Likewise.
7793 (vec_widen_smult_hi_v8hi): Explicitly gnerate vmulesh and vmulosh
7794 rather than call gen_vec_widen_smult_*.
7795 (vec_widen_smult_lo_v8hi): Likewise.
7797 2014-01-15 Jeff Law <law@redhat.com>
7799 PR tree-optimization/59747
7800 * ree.c (find_and_remove_re): Properly handle case where a second
7801 eliminated extension requires widening a copy created for elimination
7802 of a prior extension.
7803 (combine_set_extension): Ensure that the number of hard regs needed
7804 for a destination register does not change when we widen it.
7806 2014-01-15 Sebastian Huber <sebastian.huber@embedded-brains.de>
7808 * config.gcc (*-*-rtems*): Add t-rtems to tmake_file.
7809 (arm*-*-uclinux*eabi*): Do not override an existing tmake_file.
7810 (arm*-*-eabi* | arm*-*-symbianelf* | arm*-*-rtems*): Likwise.
7811 (arm*-*-rtems*): Use t-rtems from existing tmake_file.
7812 (avr-*-rtems*): Likewise.
7813 (bfin*-rtems*): Likewise.
7814 (moxie-*-rtems*): Likewise.
7815 (h8300-*-rtems*): Likewise.
7816 (i[34567]86-*-rtems*): Likewise.
7817 (lm32-*-rtems*): Likewise.
7818 (m32r-*-rtems*): Likewise.
7819 (m68k-*-rtems*): Likewise.
7820 (microblaze*-*-rtems*): Likewise.
7821 (mips*-*-rtems*): Likewise.
7822 (powerpc-*-rtems*): Likewise.
7823 (sh-*-rtems*): Likewise.
7824 (sparc-*-rtems*): Likewise.
7825 (sparc64-*-rtems*): Likewise.
7826 (v850-*-rtems*): Likewise.
7827 (m32c-*-rtems*): Likewise.
7829 2014-01-15 Vladimir Makarov <vmakarov@redhat.com>
7831 PR rtl-optimization/59511
7832 * ira.c (ira_init_register_move_cost): Use memory costs for some
7833 cases of register move cost calculations.
7834 * lra-constraints.c (lra_constraints): Use REG_FREQ_FROM_BB
7835 instead of BB frequency.
7836 * lra-coalesce.c (move_freq_compare_func, lra_coalesce): Ditto.
7837 * lra-assigns.c (find_hard_regno_for): Ditto.
7839 2014-01-15 Richard Biener <rguenther@suse.de>
7841 PR tree-optimization/59822
7842 * tree-vect-stmts.c (hoist_defs_of_uses): New function.
7843 (vectorizable_load): Use it to hoist defs of uses of invariant
7844 loads out of the loop.
7846 2014-01-15 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
7847 Kugan Vivekanandarajah <kuganv@linaro.org>
7850 * config/aarch64/aarch64.c (aarch64_build_constant): Fix incorrect
7853 2014-01-15 Richard Biener <rguenther@suse.de>
7855 PR rtl-optimization/59802
7856 * lcm.c (compute_available): Use inverted postorder to seed
7857 the initial worklist.
7859 2014-01-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7862 * config/s390/s390.c (s390_preferred_reload_class): Don't return
7863 ADDR_REGS for invalid symrefs in non-PIC code.
7865 2014-01-15 Jakub Jelinek <jakub@redhat.com>
7868 * builtins.c (determine_block_size): Initialize *probable_max_size
7869 even if len_rtx is CONST_INT.
7871 2014-01-14 Andrew Pinski <apinski@cavium.com>
7873 * config/aarch64/aarch64-protos.h (tune_params): Add issue_rate.
7874 * config/aarch64/aarch64.c (generic_tunings): Add issue rate of 2.
7875 (cortexa53_tunings): Likewise.
7876 (aarch64_sched_issue_rate): New function.
7877 (TARGET_SCHED_ISSUE_RATE): Define.
7879 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
7881 * ira-costs.c (find_costs_and_classes): Add missed
7882 ira_init_register_move_cost_if_necessary.
7884 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
7887 * config/arm/arm.c (arm_coproc_mem_operand): Add lra_in_progress.
7889 2014-01-14 H.J. Lu <hongjiu.lu@intel.com>
7892 * config/i386/i386.c (type_natural_mode): Add a bool parameter
7893 to indicate if type is used for function return value. Warn ABI
7894 change if the vector mode isn't available for function return value.
7895 (ix86_function_arg_advance): Pass false to type_natural_mode.
7896 (ix86_function_arg): Likewise.
7897 (ix86_gimplify_va_arg): Likewise.
7898 (function_arg_32): Don't warn ABI change.
7899 (ix86_function_value): Pass true to type_natural_mode.
7900 (ix86_return_in_memory): Likewise.
7901 (ix86_struct_value_rtx): Removed.
7902 (TARGET_STRUCT_VALUE_RTX): Likewise.
7904 2014-01-14 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7906 * jump.c (redirect_jump_2): Remove REG_CROSSING_JUMP notes when
7907 converting a conditional jump into a conditional return.
7909 2014-01-14 Richard Biener <rguenther@suse.de>
7911 PR tree-optimization/58921
7912 PR tree-optimization/59006
7913 * tree-vect-loop-manip.c (vect_loop_versioning): Remove code
7914 hoisting invariant stmts.
7915 * tree-vect-stmts.c (vectorizable_load): Insert the splat of
7916 invariant loads on the preheader edge if possible.
7918 2014-01-14 Joey Ye <joey.ye@arm.com>
7920 * doc/plugin.texi (Building GCC plugins): Update to C++.
7922 2014-01-14 Kirill Yukhin <kirill.yukhin@intel.com>
7924 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): New.
7925 (_mm_rcp28_round_ss): Ditto.
7926 (_mm_rsqrt28_round_sd): Ditto.
7927 (_mm_rsqrt28_round_ss): Ditto.
7928 (_mm_rcp28_sd): Ditto.
7929 (_mm_rcp28_ss): Ditto.
7930 (_mm_rsqrt28_sd): Ditto.
7931 (_mm_rsqrt28_ss): Ditto.
7932 * config/i386/avx512fintrin.h (_mm512_stream_load_si512): Ditto.
7933 * config/i386/i386-builtin-types.def (V8DI_FTYPE_PV8DI): Ditto.
7934 * config/i386/i386.c (IX86_BUILTIN_MOVNTDQA512): Ditto.
7935 (IX86_BUILTIN_RCP28SD): Ditto.
7936 (IX86_BUILTIN_RCP28SS): Ditto.
7937 (IX86_BUILTIN_RSQRT28SD): Ditto.
7938 (IX86_BUILTIN_RSQRT28SS): Ditto.
7939 (bdesc_special_args): Define __builtin_ia32_movntdqa512,
7940 __builtin_ia32_rcp28sd_round, __builtin_ia32_rcp28ss_round,
7941 __builtin_ia32_rsqrt28sd_round, __builtin_ia32_rsqrt28ss_round.
7942 (ix86_expand_special_args_builtin): Expand new FTYPE.
7943 * config/i386/sse.md (define_mode_attr "sse4_1_avx2"): Expand to V8DI.
7944 (srcp14<mode>): Make insn unary.
7945 (avx512f_vmscalef<mode><round_name>): Use substed predicate.
7946 (avx512f_sgetexp<mode><round_saeonly_name>): Ditto.
7947 (avx512f_rndscale<mode><round_saeonly_name>): Ditto.
7948 (<sse4_1_avx2>_movntdqa): Extend to 512 bits.
7949 (avx512er_exp2<mode><mask_name><round_saeonly_name>):
7950 Fix rounding: make it SAE only.
7951 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
7953 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
7955 (avx512er_vmrcp28<mode><round_saeonly_name>): Ditto.
7956 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Ditto.
7957 (avx512f_getmant<mode><mask_name><round_saeonly_name>): Ditto.
7958 * config/i386/subst.md (round_saeonly_mask_scalar_operand3): Remove.
7959 (round_saeonly_mask_scalar_operand4): Ditto.
7960 (round_saeonly_mask_scalar_op3): Ditto.
7961 (round_saeonly_mask_scalar_op4): Ditto.
7963 2014-01-13 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7965 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
7966 Implement -maltivec=be for vec_insert and vec_extract.
7968 2014-01-10 DJ Delorie <dj@redhat.com>
7970 * config/msp430/msp430.md (call_internal): Don't allow memory
7971 references with SP as the base register.
7972 (call_value_internal): Likewise.
7973 * config/msp430/constraints.md (Yc): New. For memory references
7974 that don't use SP as a base register.
7976 * config/msp430/msp430.c (msp430_print_operand): Add 'J' to mean
7977 "an integer without a # prefix"
7978 * config/msp430/msp430.md (epilogue_helper): Use it.
7980 2014-01-13 Jakub Jelinek <jakub@redhat.com>
7983 * config/i386/i386.c (ix86_vectorize_builtin_gather): Uncomment
7984 AVX512F gather builtins.
7985 * tree-vect-stmts.c (vectorizable_mask_load_store): For now punt
7986 on gather decls with INTEGER_TYPE masktype.
7987 (vectorizable_load): For INTEGER_TYPE masktype, put the INTEGER_CST
7988 directly into the builtin rather than hoisting it before loop.
7990 PR tree-optimization/59387
7991 * tree-scalar-evolution.c: Include gimple-fold.h and gimplify-me.h.
7992 (scev_const_prop): If folded_casts and type has undefined overflow,
7993 use force_gimple_operand instead of force_gimple_operand_gsi and
7994 for each added stmt if it is assign with
7995 arith_code_with_undefined_signed_overflow, call
7996 rewrite_to_defined_overflow.
7997 * tree-ssa-loop-im.c: Don't include gimplify-me.h, include
7998 gimple-fold.h instead.
7999 (arith_code_with_undefined_signed_overflow,
8000 rewrite_to_defined_overflow): Moved to ...
8001 * gimple-fold.c (arith_code_with_undefined_signed_overflow,
8002 rewrite_to_defined_overflow): ... here. No longer static.
8003 Include gimplify-me.h.
8004 * gimple-fold.h (arith_code_with_undefined_signed_overflow,
8005 rewrite_to_defined_overflow): New prototypes.
8007 2014-01-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8009 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Fix typo in description.
8011 2014-01-13 Eric Botcazou <ebotcazou@adacore.com>
8013 * builtins.c (get_object_alignment_2): Minor tweak.
8014 * tree-ssa-loop-ivopts.c (may_be_unaligned_p): Rewrite.
8016 2014-01-13 Christian Bruel <christian.bruel@st.com>
8018 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Unroll small sizes and
8019 optimized non constant lengths.
8021 2014-01-13 Jakub Jelinek <jakub@redhat.com>
8024 * omp-low.c (expand_omp_atomic_pipeline): Expand the initial
8025 load as __atomic_load_N if possible.
8027 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
8029 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): Remove
8031 (rs6000_expand_builtin): Adjust call.
8033 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
8036 * config/rs6000/rs6000.h (SWITCHABLE_TARGET): Define.
8037 * config/rs6000/rs6000.c: Include target-globals.h.
8038 (rs6000_set_current_function): Instead of doing target_reinit
8039 unconditionally, use save_target_globals_default_opts and
8040 restore_target_globals.
8042 * config/rs6000/rs6000-builtin.def (mffs, mtfsf): Add builtins for
8044 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): New.
8045 (rs6000_expand_builtin): Handle mffs and mtfsf.
8046 (rs6000_init_builtins): Define mffs and mtfsf.
8047 * config/rs6000/rs6000.md (UNSPECV_MFFS, UNSPECV_MTFSF): New constants.
8048 (rs6000_mffs): New pattern.
8049 (rs6000_mtfsf): New pattern.
8051 2014-01-11 Bin Cheng <bin.cheng@arm.com>
8053 * tree-ssa-loop-ivopts.c (iv_ca_narrow): New parameter.
8054 Start narrowing with START. Apply candidate-use pair
8055 and check overall cost in narrowing.
8056 (iv_ca_prune): Pass new argument.
8058 2014-01-10 Jeff Law <law@redhat.com>
8061 * ree.c (combine_reaching_defs): Ensure the defining statement
8062 occurs before the extension when optimizing extensions with
8063 different source and destination hard registers.
8065 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
8068 * ipa-devirt.c (build_type_inheritance_graph): Also add types of
8069 vtables into the type inheritance graph.
8071 2014-01-10 Jakub Jelinek <jakub@redhat.com>
8073 PR rtl-optimization/59754
8074 * ree.c (combine_reaching_defs): Disallow !SCALAR_INT_MODE_P
8075 modes in the REGNO != REGNO case.
8077 2014-01-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8079 * config/rs6000/rs6000-builtin.def: Fix pasto for VPKSDUS.
8081 2014-01-10 Jakub Jelinek <jakub@redhat.com>
8083 PR tree-optimization/59745
8084 * tree-predcom.c (tree_predictive_commoning_loop): Call
8085 free_affine_expand_cache if giving up because components is NULL.
8087 * target-globals.c (save_target_globals): Allocate < 4KB structs using
8088 GC in payload of target_globals struct instead of allocating them on
8089 the heap and the larger structs separately using GC.
8090 * target-globals.h (struct target_globals): Make regs, hard_regs,
8091 reload, expmed, ira, ira_int and lra_fields GTY((atomic)) instead
8092 of GTY((skip)) and change type to void *.
8093 (reset_target_globals): Cast loads from those fields to corresponding
8096 2014-01-10 Steve Ellcey <sellcey@mips.com>
8099 * Makefile.in (PLUGIN_HEADERS): Add gimplify.h, gimple-iterator.h,
8100 gimple-ssa.h, fold-const.h, tree-cfg.h, tree-into-ssa.h,
8101 tree-ssanames.h, print-tree.h, varasm.h, and context.h.
8103 2014-01-10 Richard Earnshaw <rearnsha@arm.com>
8106 * aarch64-modes.def (CC_Zmode): New flags mode.
8107 * aarch64.c (aarch64_select_cc_mode): Only allow NEG when the condition
8108 represents an equality.
8109 (aarch64_get_condition_code): Handle CC_Zmode.
8110 * aarch64.md (compare_neg<mode>): Restrict to equality operations.
8112 2014-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8114 * config/s390/s390.c (s390_expand_tbegin): Remove jump over CC
8115 extraction in good case.
8117 2014-01-10 Richard Biener <rguenther@suse.de>
8119 PR tree-optimization/59374
8120 * tree-vect-slp.c (vect_slp_analyze_bb_1): Move dependence
8121 checking after SLP discovery. Mark stmts not participating
8122 in any SLP instance properly.
8124 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8126 * config/arm/arm.c (arm_new_rtx_costs): Use destination mode
8127 when handling a SET rtx.
8129 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8131 * config/arm/arm-cores.def (cortex-a53): Specify FL_CRC32.
8132 (cortex-a57): Likewise.
8133 (cortex-a57.cortex-a53): Likewise. Remove redundant flags.
8135 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8137 * config/arm/arm.c (arm_init_iwmmxt_builtins): Skip
8138 non-iwmmxt builtins.
8140 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
8144 * ipa-devirt.c record_target_from_binfo): Take as argument
8145 stack of binfos and lookup matching one for virtual inheritance.
8146 (possible_polymorphic_call_targets_1): Update.
8148 2014-01-10 Huacai Chen <chenhc@lemote.com>
8150 * config/mips/driver-native.c (host_detect_local_cpu): Handle new
8151 kernel strings for Loongson-2E/2F/3A.
8153 2014-01-10 Jakub Jelinek <jakub@redhat.com>
8156 * tree-vect-data-refs.c (vect_analyze_data_refs): Check
8157 is_gimple_call before calling gimple_call_internal_p.
8159 2014-01-09 Steve Ellcey <sellcey@mips.com>
8161 * Makefile.in (TREE_FLOW_H): Remove.
8162 (TREE_SSA_H): Add file names from tree-flow.h.
8163 * doc/tree-ssa.texi (Annotations): Remove reference to tree-flow.h
8164 * tree.h: Remove tree-flow.h reference.
8165 * hash-table.h: Remove tree-flow.h reference.
8166 * tree-ssa-loop-niter.c (dump_affine_iv): Replace tree-flow.h
8167 reference with tree-ssa-loop.h.
8169 2014-01-09 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8171 * doc/invoke.texi: Add -maltivec={be,le} options, and document
8172 default element-order behavior for -maltivec.
8173 * config/rs6000/rs6000.opt: Add -maltivec={be,le} options.
8174 * config/rs6000/rs6000.c (rs6000_option_override_internal): Ensure
8175 that -maltivec={le,be} implies -maltivec; disallow -maltivec=le
8176 when targeting big endian, at least for now.
8177 * config/rs6000/rs6000.h: Add #define of VECTOR_ELT_ORDER_BIG.
8179 2014-01-09 Jakub Jelinek <jakub@redhat.com>
8182 * cfgexpand.c (expand_one_var): For SSA_NAMEs, if the underlying
8183 var satisfies use_register_for_decl, just take into account type
8184 alignment, rather than decl alignment.
8186 PR tree-optimization/59622
8187 * gimple-fold.c (gimple_fold_call): Fix a typo in message. For
8188 __builtin_unreachable replace the OBJ_TYPE_REF call with a call to
8189 __builtin_unreachable and add if needed a setter of the lhs SSA_NAME.
8190 Don't devirtualize for inplace at all. For targets.length () == 1,
8191 if the call is noreturn and cfun isn't in SSA form yet, clear lhs.
8193 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
8195 * config/i386/i386.md (cpu): Remove the unused btver1.
8197 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
8199 * gdbasan.in: Put a breakpoint on __sanitizer::Report.
8201 2014-01-09 Jakub Jelinek <jakub@redhat.com>
8204 * tree-core.h (struct target_globals): New forward declaration.
8205 (struct tree_target_option): Add globals field.
8206 * tree.h (TREE_TARGET_GLOBALS): Define.
8207 (prepare_target_option_nodes_for_pch): New prototype.
8208 * target-globals.h (struct target_globals): Define even if
8210 * tree.c (prepare_target_option_node_for_pch,
8211 prepare_target_option_nodes_for_pch): New functions.
8212 * config/i386/i386.h (SWITCHABLE_TARGET): Define.
8213 * config/i386/i386.c: Include target-globals.h.
8214 (ix86_set_current_function): Instead of doing target_reinit
8215 unconditionally, use save_target_globals_default_opts and
8216 restore_target_globals.
8218 2014-01-09 Richard Biener <rguenther@suse.de>
8220 PR tree-optimization/59715
8221 * tree-cfg.h (split_critical_edges): Declare.
8222 * tree-cfg.c (split_critical_edges): Export.
8223 * tree-ssa-sink.c (execute_sink_code): Split critical edges.
8225 2014-01-09 Max Ostapenko <m.ostapenko@partner.samsung.com>
8227 * cfgexpand.c (expand_stack_vars): Optionally disable
8228 asan stack protection.
8229 (expand_used_vars): Likewise.
8230 (partition_stack_vars): Likewise.
8231 * asan.c (asan_emit_stack_protection): Optionally disable
8232 after return stack usage.
8233 (instrument_derefs): Optionally disable memory access instrumentation.
8234 (instrument_builtin_call): Likewise.
8235 (instrument_strlen_call): Likewise.
8236 (asan_protect_global): Optionally disable global variables protection.
8237 * doc/invoke.texi: Added doc for new options.
8238 * params.def: Added new options.
8239 * params.h: Likewise.
8241 2014-01-09 Jakub Jelinek <jakub@redhat.com>
8243 PR rtl-optimization/59724
8244 * ifcvt.c (cond_exec_process_if_block): Don't call
8245 flow_find_head_matching_sequence with 0 longest_match.
8246 * cfgcleanup.c (flow_find_head_matching_sequence): Count even
8247 non-active insns if !stop_after.
8248 (try_head_merge_bb): Revert 2014-01-07 changes.
8250 2014-01-08 Jeff Law <law@redhat.com>
8252 * ree.c (get_sub_rtx): New function, extracted from...
8253 (merge_def_and_ext): Here.
8254 (combine_reaching_defs): Use get_sub_rtx.
8256 2014-01-08 Eric Botcazou <ebotcazou@adacore.com>
8258 * cgraph.h (varpool_variable_node): Do not choke on null node.
8260 2014-01-08 Catherine Moore <clm@codesourcery.com>
8262 * config/mips/mips.md (simple_return): Attempt to use JRC
8264 * config/mips/mips.h (MIPS_CALL): Attempt to use JALS for microMIPS.
8266 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
8268 PR rtl-optimization/59137
8269 * reorg.c (steal_delay_list_from_target): Call update_block for
8271 (steal_delay_list_from_fallthrough, relax_delay_slots): Likewise.
8273 2014-01-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8275 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Remove
8276 two duplicate entries.
8278 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
8281 2012-10-07 Richard Sandiford <rdsandiford@googlemail.com>
8283 * config/mips/mips.c (mips_truncated_op_cost): New function.
8284 (mips_rtx_costs): Adjust test for BADDU.
8285 * config/mips/mips.md (*baddu_di<mode>): Push truncates to operands.
8287 2012-10-02 Richard Sandiford <rdsandiford@googlemail.com>
8289 * config/mips/mips.md (*baddu_si_eb, *baddu_si_el): Merge into...
8290 (*baddu_si): ...this new pattern.
8292 2014-01-08 Jakub Jelinek <jakub@redhat.com>
8295 * ipa-prop.c (ipa_analyze_params_uses): Ignore uses in debug stmts.
8297 2014-01-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
8300 * expr.h (expand_expr_real, expand_expr_real_1): Add new parameter
8302 (expand_expr, expand_normal): Adjust.
8303 * expr.c (expand_expr_real, expand_expr_real_1): Add new parameter
8304 inner_reference_p. Use inner_reference_p to expand inner references.
8305 (store_expr): Adjust.
8306 * cfgexpand.c (expand_call_stmt): Adjust.
8308 2014-01-08 Rong Xu <xur@google.com>
8310 * gcov-io.c (gcov_var): Move from gcov-io.h.
8311 (gcov_position): Ditto.
8312 (gcov_is_error): Ditto.
8313 (gcov_rewrite): Ditto.
8314 * gcov-io.h: Refactor. Move gcov_var to gcov-io.h, and libgcov
8315 only part to libgcc/libgcov.h.
8317 2014-01-08 Marek Polacek <polacek@redhat.com>
8320 * omp-low.c (simd_clone_adjust): Don't crash if def is NULL.
8322 2014-01-08 Marek Polacek <polacek@redhat.com>
8325 * ubsan.c (ubsan_type_descriptor): Call strip_array_types on type2.
8327 2014-01-08 Jakub Jelinek <jakub@redhat.com>
8329 PR rtl-optimization/59649
8330 * stor-layout.c (get_mode_bounds): For BImode return
8331 0 and STORE_FLAG_VALUE.
8333 2014-01-08 Richard Biener <rguenther@suse.de>
8336 * gimple.h (is_gimple_builtin_call): Remove.
8337 (gimple_builtin_call_types_compatible_p): New.
8338 (gimple_call_builtin_p): New overload.
8339 * gimple.c (is_gimple_builtin_call): Remove.
8340 (validate_call): Rename to ...
8341 (gimple_builtin_call_types_compatible_p): ... this and export. Also
8343 (validate_type): New static function.
8344 (gimple_call_builtin_p): New overload and adjust.
8345 * gimple-fold.c (gimple_fold_builtin): Fold the return value.
8346 (gimple_fold_call): Likewise. Use gimple_call_builtin_p.
8347 (gimple_fold_stmt_to_constant_1): Likewise.
8348 * tsan.c (instrument_gimple): Use gimple_call_builtin_p.
8350 2014-01-08 Richard Biener <rguenther@suse.de>
8353 * gimplify.c (gimplify_expr): Gimplify register-register type
8354 VIEW_CONVERT_EXPRs to separate stmts.
8356 2014-01-07 Jeff Law <law@redhat.com>
8359 * ree.c (combine_set_extension): Handle case where source
8360 and destination registers in an extension insn are different.
8361 (combine_reaching_defs): Allow source and destination registers
8362 in extension to be different under limited circumstances.
8363 (add_removable_extension): Remove restriction that the
8364 source and destination registers in the extension are the same.
8365 (find_and_remove_re): Emit a copy from the extension's
8366 destination to its source after the defining insn if
8367 the source and destination registers are different.
8370 * ifcvt.c (merge_if_block): If we are merging a block with more than
8371 one successor with a block with no successors, remove any BARRIER
8372 after the second block.
8374 2014-01-07 Dan Xio Qiang <ziyan01@163.com>
8376 * hw-doloop.c (reorg_loops): Release the bitmap obstack.
8378 2014-01-07 John David Anglin <danglin@gcc.gnu.org>
8381 * config/pa/pa.c (pa_legitimate_address_p): Return false before reload
8382 for 14-bit register offsets when INT14_OK_STRICT is false.
8384 2014-01-07 Roland Stigge <stigge@antcom.de>
8385 Michael Meissner <meissner@linux.vnet.ibm.com>
8388 * config/rs6000/rs6000.c (rs6000_legitimate_offset_address_p):
8389 Only check TFmode for SPE constants. Don't check TImode or TDmode.
8391 2014-01-07 James Greenhalgh <james.greenhalgh@arm.com>
8393 * config/aarch64/aarch64-elf.h (ASM_SPEC): Remove identity spec for
8396 2014-01-07 Yufeng Zhang <yufeng.zhang@arm.com>
8398 * config/arm/arm.c (arm_expand_neon_args): Call expand_expr
8399 with EXPAND_MEMORY for NEON_ARG_MEMORY; check if the returned
8400 rtx is const0_rtx or not.
8402 2014-01-07 Richard Sandiford <rdsandiford@googlemail.com>
8405 * target-globals.c (save_target_globals): Remove this_fn_optab
8407 * toplev.c: Include optabs.h.
8408 (target_reinit): Temporarily restore the global options if another
8409 set of options are in force.
8411 2014-01-07 Jakub Jelinek <jakub@redhat.com>
8413 PR rtl-optimization/58668
8414 * cfgcleanup.c (flow_find_cross_jump): Don't count
8415 any jumps if dir_p is NULL. Remove p1 variable, use active_insn_p
8416 to determine what is counted.
8417 (flow_find_head_matching_sequence): Use active_insn_p to determine
8419 (try_head_merge_bb): Adjust for the flow_find_head_matching_sequence
8421 * ifcvt.c (count_bb_insns): Use active_insn_p && !JUMP_P to
8422 determine what is counted.
8424 PR tree-optimization/59643
8425 * tree-predcom.c (split_data_refs_to_components): If one dr is
8426 read and one write, determine_offset fails and the write isn't
8427 in the bad component, just put the read into the bad component.
8429 2014-01-07 Mike Stump <mikestump@comcast.net>
8430 Jakub Jelinek <jakub@redhat.com>
8433 * tree-core.h (struct tree_optimization_option): Change optabs
8434 type from unsigned char * to void *.
8435 * optabs.c (init_tree_optimization_optabs): Adjust
8436 TREE_OPTIMIZATION_OPTABS initialization.
8438 2014-01-06 Jakub Jelinek <jakub@redhat.com>
8441 * config/i386/i386.h (struct machine_function): Add
8442 no_drap_save_restore field.
8443 * config/i386/i386.c (ix86_save_reg): Use
8444 !cfun->machine->no_drap_save_restore instead of
8445 crtl->stack_realign_needed.
8446 (ix86_finalize_stack_realign_flags): Don't clear drap_reg unless
8447 this function clears frame_pointer_needed. Set
8448 cfun->machine->no_drap_save_restore if clearing frame_pointer_needed
8449 and DRAP reg is needed.
8451 2014-01-06 Marek Polacek <polacek@redhat.com>
8454 * doc/implement-c.texi: Mention that other integer types are
8455 permitted as bit-field types in strictly conforming mode.
8457 2014-01-06 Felix Yang <fei.yang0953@gmail.com>
8459 * modulo-sched.c (schedule_reg_moves): Clear distance1_uses if it
8462 2014-01-06 Richard Earnshaw <rearnsha@arm.com>
8464 * aarch64.c (aarch64_rtx_costs): Fix cost calculation for MADD.
8466 2014-01-06 Martin Jambor <mjambor@suse.cz>
8469 * ipa-cp.c (ipcp_discover_new_direct_edges): Changed param_index type
8471 * ipa-prop.c (ipa_print_node_params): Fix indentation.
8473 2014-01-06 Eric Botcazou <ebotcazou@adacore.com>
8477 * var-tracking.c (add_stores): Preserve the value of the source even if
8478 we don't record the store.
8480 2014-01-06 Terry Guo <terry.guo@arm.com>
8482 * config.gcc (arm*-*-*): Check --with-arch against arm-arches.def.
8484 2014-01-05 Iain Sandoe <iain@codesourcery.com>
8487 * config/darwin.c (darwin_function_section): Adjust return values to
8488 correspond to optimisation changes made in r206070.
8490 2014-01-05 Uros Bizjak <ubizjak@gmail.com>
8492 * config/i386/i386.c (ix86_data_alignment): Calculate max_align
8493 from prefetch_block tune setting.
8494 (nocona_cost): Correct size of prefetch block to 64.
8496 2014-01-04 Eric Botcazou <ebotcazou@adacore.com>
8498 * config/arm/arm.c (arm_get_frame_offsets): Revamp long lines.
8499 (arm_expand_epilogue_apcs_frame): Take into account the number of bytes
8500 used to save the static chain register in the computation of the offset
8501 from which the FP registers need to be restored.
8503 2014-01-04 Jakub Jelinek <jakub@redhat.com>
8505 PR tree-optimization/59519
8506 * tree-vect-loop-manip.c (slpeel_update_phi_nodes_for_guard1): Don't
8507 ICE if get_current_def (current_new_name) is already non-NULL, as long
8508 as it is a phi result of some other phi in *new_exit_bb that has
8511 * config/i386/sse.md (avx512f_load<mode>_mask): Emit vmovup{s,d}
8512 or vmovdqu* for misaligned_operand.
8513 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>,
8514 <sse2_avx_avx512f>_loaddqu<mode><mask_name>): Handle <mask_applied>.
8515 * config/i386/i386.c (ix86_expand_special_args_builtin): Set
8516 aligned_mem for AVX512F masked aligned load and store builtins and for
8519 2014-01-03 Bingfeng Mei <bmei@broadcom.com>
8521 PR tree-optimization/59651
8522 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
8523 Address range for negative step should be added by TYPE_SIZE_UNIT.
8525 2014-01-03 Andreas Schwab <schwab@linux-m68k.org>
8527 * config/m68k/m68k.c (handle_move_double): Handle pushes with
8528 overlapping registers also for registers other than the stack pointer.
8530 2014-01-03 Marek Polacek <polacek@redhat.com>
8533 * doc/extend.texi: Fix the return value of __builtin_FUNCTION and
8536 2014-01-03 Jakub Jelinek <jakub@redhat.com>
8539 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Don't consider
8542 * config/i386/i386.md (MODE_SIZE): New mode attribute.
8543 (push splitter): Use <P:MODE_SIZE> instead of
8544 GET_MODE_SIZE (<P:MODE>mode).
8545 (lea splitter): Use <MODE_SIZE> instead of GET_MODE_SIZE (<MODE>mode).
8546 (mov -1, reg peephole2): Likewise.
8547 * config/i386/sse.md (*mov<mode>_internal,
8548 <sse>_storeu<ssemodesuffix><avxsizesuffix>,
8549 <sse2_avx_avx512f>_storedqu<mode>, <sse>_andnot<mode>3,
8550 *<code><mode>3, *andnot<mode>3<mask_name>,
8551 <mask_codefor><code><mode>3<mask_name>): Likewise.
8552 * config/i386/subst.md (mask_mode512bit_condition,
8553 sd_mask_mode512bit_condition): Likewise.
8555 2014-01-02 Xinliang David Li <davidxl@google.com>
8557 PR tree-optimization/59303
8558 * tree-ssa-uninit.c (is_use_properly_guarded): Main cleanup.
8559 (dump_predicates): Better output format.
8560 (pred_equal_p): New function.
8561 (is_neq_relop_p): Ditto.
8562 (is_neq_zero_form_p): Ditto.
8563 (pred_expr_equal_p): Ditto.
8564 (pred_neg_p): Ditto.
8565 (simplify_pred): Ditto.
8566 (simplify_preds_2): Ditto.
8567 (simplify_preds_3): Ditto.
8568 (simplify_preds_4): Ditto.
8569 (simplify_preds): Ditto.
8571 (push_to_worklist): Ditto.
8572 (get_pred_info_from_cmp): Ditto.
8573 (is_degenerated_phi): Ditto.
8574 (normalize_one_pred_1): Ditto.
8575 (normalize_one_pred): Ditto.
8576 (normalize_one_pred_chain): Ditto.
8577 (normalize_preds): Ditto.
8578 (normalize_cond_1): Remove function.
8579 (normalize_cond): Ditto.
8580 (is_gcond_subset_of): Ditto.
8581 (is_subset_of_any): Ditto.
8582 (is_or_set_subset_of): Ditto.
8583 (is_and_set_subset_of): Ditto.
8584 (is_norm_cond_subset_of): Ditto.
8585 (pred_chain_length_cmp): Ditto.
8586 (convert_control_dep_chain_into_preds): Type change.
8587 (find_predicates): Ditto.
8588 (find_def_preds): Ditto.
8589 (destroy_predicates_vecs): Ditto.
8590 (find_matching_predicates_in_rest_chains): Ditto.
8591 (use_pred_not_overlap_with_undef_path_pred): Ditto.
8592 (is_pred_expr_subset): Ditto.
8593 (is_pred_chain_subset_of): Ditto.
8594 (is_included_in): Ditto.
8595 (is_superset_of): Ditto.
8597 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
8599 Update copyright years.
8601 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
8603 * common/config/arc/arc-common.c, config/arc/arc-modes.def,
8604 config/arc/arc-protos.h, config/arc/arc.c, config/arc/arc.h,
8605 config/arc/arc.md, config/arc/arc.opt,
8606 config/arm/arm_neon_builtins.def, config/arm/crypto.def,
8607 config/i386/avx512cdintrin.h, config/i386/avx512erintrin.h,
8608 config/i386/avx512fintrin.h, config/i386/avx512pfintrin.h,
8609 config/i386/btver2.md, config/i386/shaintrin.h, config/i386/slm.md,
8610 config/linux-protos.h, config/linux.c, config/winnt-c.c,
8611 diagnostic-color.c, diagnostic-color.h, gimple-ssa-isolate-paths.c,
8612 vtable-verify.c, vtable-verify.h: Use the standard form for the
8615 2014-01-02 Tobias Burnus <burnus@net-b.de>
8617 * gcc.c (process_command): Update copyright notice dates.
8618 * gcov-dump.c: Ditto.
8620 * doc/cpp.texi: Bump @copying's copyright year.
8621 * doc/cppinternals.texi: Ditto.
8622 * doc/gcc.texi: Ditto.
8623 * doc/gccint.texi: Ditto.
8624 * doc/gcov.texi: Ditto.
8625 * doc/install.texi: Ditto.
8626 * doc/invoke.texi: Ditto.
8628 2014-01-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8630 * config/nios2/nios2.h (BITS_PER_UNIT): Don't define it.
8632 2014-01-01 Jakub Jelinek <jakub@redhat.com>
8634 * config/i386/sse.md (*mov<mode>_internal): Guard
8635 EXT_REX_SSE_REGNO_P (REGNO ()) uses with REG_P.
8637 PR rtl-optimization/59647
8638 * cse.c (cse_process_notes_1): Don't substitute negative VOIDmode
8639 new_rtx into UNSIGNED_FLOAT rtxes.
8641 Copyright (C) 2014 Free Software Foundation, Inc.
8643 Copying and distribution of this file, with or without modification,
8644 are permitted in any medium without royalty provided the copyright
8645 notice and this notice are preserved.