1 2014-04-29 James Greenhalgh <james.greenhalgh@arm.com>
3 * calls.c (initialize_argument_information): Always treat
4 PUSH_ARGS_REVERSED as 1, simplify code accordingly.
5 (expand_call): Likewise.
6 (emit_library_call_calue_1): Likewise.
7 * expr.c (PUSH_ARGS_REVERSED): Do not define.
8 (emit_push_insn): Always treat PUSH_ARGS_REVERSED as 1, simplify
11 2014-04-29 Nick Clifton <nickc@redhat.com>
13 * config/arc/arc.c (arc_select_cc_mode): Fix parentheses.
14 (arc_init_reg_tables): Use a machine_mode enum to iterate over
16 * config/m32r/m32r.c (init_reg_tables): Likewise.
17 * config/m32c/m32c.c (m32c_illegal_subreg_p): Use a machine_mode
18 enum to hold the modes.
20 2014-04-29 Richard Biener <rguenther@suse.de>
22 * dominance.c (free_dominance_info): Add overload with
24 (dom_info_state): Likewise.
25 (dom_info_available_p): Likewise.
26 * basic-block.h (free_dominance_info, dom_info_state,
27 dom_info_available_p): Declare overloads.
28 * passes.c (execute_function_todo): Verify that verifiers
29 don't change dominator info state. Drop dominator info
30 for IPA pass invocations.
31 * cgraph.c (release_function_body): Restore asserts that
32 dominator information is released.
34 2014-04-29 Patrick Palka <patrick@parcs.ath.cx>
36 * doc/invoke.texi: Fix typo.
37 * tree-vrp.c: Fix typos.
38 * gimple.c (infer_nonnull_range): Reorder operands of an &&
41 2014-04-29 Zhenqiang Chen <zhenqiang.chen@linaro.org>
43 * config/aarch64/aarch64.md (mov<mode>cc): New for GPF.
45 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
47 * config/aarch64/aarch64-builtins.c
48 (aarch64_types_storestruct_lane_qualifiers): New.
49 (TYPES_STORESTRUCT_LANE): Likewise.
50 * config/aarch64/aarch64-simd-builtins.def (st2_lane): New.
53 * config/aarch64/aarch64-simd.md (vec_store_lanesoi_lane<mode>): New.
54 (vec_store_lanesci_lane<mode>): Likewise.
55 (vec_store_lanesxi_lane<mode>): Likewise.
56 (aarch64_st2_lane<VQ:mode>): Likewise.
57 (aarch64_st3_lane<VQ:mode>): Likewise.
58 (aarch64_st4_lane<VQ:mode>): Likewise.
59 * config/aarch64/aarch64.md (unspec): Add UNSPEC_ST{2,3,4}_LANE.
60 * config/aarch64/arm_neon.h
61 (__ST2_LANE_FUNC): Rewrite using builtins, update use points to
62 use new macro arguments.
63 (__ST3_LANE_FUNC): Likewise.
64 (__ST4_LANE_FUNC): Likewise.
65 * config/aarch64/iterators.md (V_TWO_ELEM): New.
66 (V_THREE_ELEM): Likewise.
67 (V_FOUR_ELEM): Likewise.
69 2014-04-28 David Malcolm <dmalcolm@redhat.com>
71 * doc/gimple.texi: Replace the description of the now-defunct
72 union gimple_statement_d with a diagram showing the
73 gimple_statement_base class hierarchy and its relationships to
74 the GSS_ and GIMPLE_ enums.
76 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
78 * config/aarch64/aarch64-protos.h (aarch64_modes_tieable_p): New.
79 * config/aarch64/aarch64.c
80 (aarch64_cannot_change_mode_class): Weaken conditions.
81 (aarch64_modes_tieable_p): New.
82 * config/aarch64/aarch64.h (MODES_TIEABLE_P): Use it.
84 2014-04-28 Pat Haugen <pthaugen@us.ibm.com>
86 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
87 (loadsync_<mode>): Change mode.
88 (load_quadpti, store_quadpti): New.
89 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
90 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
92 2014-04-28 Martin Jambor <mjambor@suse.cz>
94 * tree-sra.c (sra_modify_expr): Generate new memory accesses with
95 same alias type as the original statement.
96 (subreplacement_assignment_data): New type.
97 (handle_unscalarized_data_in_subtree): New type of parameter,
98 generate new memory accesses with same alias type as the original
100 (load_assign_lhs_subreplacements): Likewise.
101 (sra_modify_constructor_assign): Generate new memory accesses with
102 same alias type as the original statement.
104 2014-04-28 Richard Biener <rguenther@suse.de>
106 * tree-pass.h (TODO_verify_il): Define.
107 (TODO_verify_all): Complete properly.
108 * passes.c (execute_function_todo): Move existing loop-closed
109 SSA verification under TODO_verify_il.
110 (execute_one_pass): Trigger TODO_verify_il at todo-after time.
111 * graphite-sese-to-poly.c (rewrite_cross_bb_scalar_deps):
112 Fix tree sharing issue.
114 2014-04-28 Richard Biener <rguenther@suse.de>
117 * builtins.def (DEF_C11_BUILTIN): Add.
118 (BUILT_IN_ALIGNED_ALLOC): Likewise.
119 * coretypes.h (enum function_class): Add function_c11_misc.
120 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
121 BUILT_IN_ALIGNED_ALLOC like BUILT_IN_MALLOC.
122 (call_may_clobber_ref_p_1): Likewise.
123 * tree-ssa-dce.c (mark_stmt_if_obviously_necessary): Likewise.
124 (mark_all_reaching_defs_necessary_1): Likewise.
125 (propagate_necessity): Likewise.
126 (eliminate_unnecessary_stmts): Likewise.
127 * tree-ssa-ccp.c (evaluate_stmt): Handle BUILT_IN_ALIGNED_ALLOC.
129 2014-04-28 Richard Biener <rguenther@suse.de>
131 * tree-vrp.c (vrp_var_may_overflow): Remove.
132 (vrp_visit_phi_node): Rather than bumping to +-INF possibly
133 with overflow immediately bump to one before that value and
134 let iteration figure out overflow status.
136 2014-04-28 Richard Biener <rguenther@suse.de>
138 * configure.ac: Do valgrind header checks unconditionally.
139 Add --enable-valgrind-annotations.
140 * system.h: Guard valgrind header inclusion with
141 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING.
142 * alloc-pool.c (pool_alloc, pool_free): Use
143 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING
144 to guard possibly dead code.
145 * config.in: Regenerated.
146 * configure: Likewise.
148 2014-04-28 Jeff Law <law@redhat.com>
150 PR tree-optimization/60902
151 * tree-ssa-threadedge.c
152 (record_temporary_equivalences_from_stmts_at_dest): Only iterate
153 over real defs when invalidating outputs from statements that do not
154 produce useful outputs for threading.
156 2014-04-28 Richard Biener <rguenther@suse.de>
158 PR tree-optimization/60979
159 * graphite-scop-detection.c (scopdet_basic_block_info): Reject
160 SCOPs that end in a block with a successor with abnormal
163 2014-04-28 Richard Biener <rguenther@suse.de>
165 * tree-pass.h (execute_pass_list): Adjust prototype.
166 * passes.c (pass_manager::execute_early_local_passes):
168 (do_per_function): Change callback signature, push all actual
169 work to the callbals.
170 (do_per_function_toporder): Likewise.
171 (execute_function_dump): Adjust.
172 (execute_function_todo): Likewise.
173 (clear_last_verified): Likewise.
174 (verify_curr_properties): Likewise.
175 (update_properties_after_pass): Likewise.
176 (execute_pass_list_1): Split out from ...
177 (execute_pass_list): ... here. Adjust.
178 (execute_ipa_pass_list): Likewise.
179 * cgraphunit.c (cgraph_add_new_function): Adjust.
180 (analyze_function): Likewise.
181 (expand_function): Likewise.
182 * cgraph.c (release_function_body): Free dominance info
183 here instead of asserting it was magically freed elsewhere.
185 2014-04-28 Eric Botcazou <ebotcazou@adacore.com>
187 * configure.ac: Tweak GAS check for LEON instructions on SPARC.
188 * configure: Regenerate.
189 * config/sparc/sparc.opt (muser-mode): New option.
190 * config/sparc/sync.md (atomic_compare_and_swap<mode>_1): Do not enable
192 (atomic_compare_and_swap_leon3_1): New instruction for LEON3.
193 * doc/invoke.texi (SPARC options): Document -muser-mode.
195 2014-04-27 Richard Sandiford <rdsandiford@googlemail.com>
197 * cselib.c (find_slot_memmode): Delete.
198 (cselib_hasher): Change compare_type to a struct.
199 (cselib_hasher::equal): Update accordingly. Don't expect wrapped
201 (preserve_constants_and_equivs): Adjust for new compare_type.
202 (cselib_find_slot): Likewise. Take the mode of the rtx as argument.
203 (wrap_constant): Delete.
204 (cselib_lookup_mem, cselib_lookup_1): Update calls to cselib_find_slot.
206 2014-04-26 Markus Trippelsdorf <markus@trippelsdorf.de>
208 * doc/install.texi (Building with profile feedback): Remove
211 2014-04-26 Tom de Vries <tom@codesourcery.com>
213 * config/i386/i386.md (define_expand "ldexpxf3"): Fix out-of-bounds
216 2014-04-25 Cary Coutant <ccoutant@google.com>
219 * dwarf2out.c (should_move_die_to_comdat): A type definition
220 can contain a subprogram definition, but don't move it to a
222 (clone_as_declaration): Copy DW_AT_abstract_origin attribute.
223 (generate_skeleton_bottom_up): Remove DW_AT_object_pointer attribute
225 (clone_tree_hash): Rename to...
226 (clone_tree_partial): ...this; change callers. Copy
227 DW_TAG_subprogram DIEs as declarations.
228 (copy_decls_walk): Don't copy children of a declaration into a
231 2014-04-25 H.J. Lu <hongjiu.lu@intel.com>
234 * config/i386/i386.md (*movsf_internal): Set MODE to SI for
237 2014-04-25 Jiong Wang <jiong.wang@arm.com>
239 * config/arm/predicates.md (call_insn_operand): Add long_call check.
240 * config/arm/arm.md (sibcall, sibcall_value): Force the address to
242 * config/arm/arm.c (arm_function_ok_for_sibcall): Remove long_call
245 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
247 * config/arm/arm.c (arm_cortex_a8_tune): Initialise
250 2014-04-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
252 PR tree-optimization/60930
253 * gimple-ssa-strength-reduction.c (create_mul_imm_cand): Reject
254 creating a multiply candidate by folding two constant
255 multiplicands when the result overflows.
257 2014-04-25 Jakub Jelinek <jakub@redhat.com>
259 PR tree-optimization/60960
260 * tree-vect-generic.c (expand_vector_operation): Only call
261 expand_vector_divmod if type's mode satisfies VECTOR_MODE_P.
263 2014-04-25 Tom de Vries <tom@codesourcery.com>
265 * expr.c (clobber_reg_mode): New function.
266 * expr.h (clobber_reg): New function.
268 2014-04-25 Tom de Vries <tom@codesourcery.com>
270 * rtlanal.c (find_all_hard_reg_sets): Note INSN_CALL_FUNCTION_USAGE
273 2014-04-25 Radovan Obradovic <robradovic@mips.com>
274 Tom de Vries <tom@codesourcery.com>
276 * rtlanal.c (find_all_hard_reg_sets): Add bool implicit parameter and
278 * rtl.h (find_all_hard_reg_sets): Add bool parameter.
279 * haifa-sched.c (recompute_todo_spec, check_clobbered_conditions): Add
280 new argument to find_all_hard_reg_sets call.
282 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
284 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p):
285 Use HOST_WIDE_INT_C for mask literal.
286 (aarch_rev16_shleft_mask_imm_p): Likewise.
288 2014-04-25 Eric Botcazou <ebotcazou@adacore.com>
291 * config/sparc/sparc.md (ashlsi3_extend): Delete.
293 2014-04-25 Marc Glisse <marc.glisse@inria.fr>
295 PR preprocessor/56540
296 * config/i386/i386-c.c (ix86_target_macros): Define
297 __SIZEOF_FLOAT80__ and __SIZEOF_FLOAT128__.
299 2014-04-25 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
301 * configure.ac (tga_func): Remove.
302 (LIB_TLS_SPEC): Remove.
303 * configure: Regenerate.
304 * config.in: Regenerate.
305 * config/sol2.h (LIB_SPEC): Don't use LIB_TLS_SPEC.
307 2014-04-25 Richard Biener <rguenther@suse.de>
310 * tree-ssa-structalias.c (ipa_pta_execute): Compute direct
311 call stmt use/clobber sets during stmt walk instead of
312 walking the possibly incomplete set of caller edges.
314 2014-04-25 Richard Biener <rguenther@suse.de>
317 * passes.c (apply_ipa_transforms): Inline into only caller ...
318 (execute_one_pass): ... here. Properly bring in function
319 bodies for nodes we want to apply IPA transforms to.
321 2014-04-24 Cong Hou <congh@google.com>
323 PR tree-optimization/60896
324 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Pick up
325 all statements in PATTERN_DEF_SEQ in recognized widen-mult pattern.
326 (vect_mark_pattern_stmts): Set the def type of all statements in
327 PATTERN_DEF_SEQ as vect_internal_def.
329 2014-04-24 Michael Meissner <meissner@linux.vnet.ibm.com>
331 * doc/extend.texi (PowerPC Built-in Functions): Document new
332 powerpc extended divide, bcd, pack/unpack 128-bit, builtin
334 (PowerPC AltiVec/VSX Built-in Functions): Likewise.
336 * config/rs6000/predicates.md (const_0_to_3_operand): New
337 predicate to match 0..3 integer constants.
339 * config/rs6000/rs6000-builtin.def (BU_DFP_MISC_1): Add new macros
340 to support adding miscellaneous builtin functions.
341 (BU_DFP_MISC_2): Likewise.
342 (BU_P7_MISC_1): Likewise.
343 (BU_P7_MISC_2): Likewise.
344 (BU_P8V_MISC_3): Likewise.
345 (BU_MISC_1): Likewise.
346 (BU_MISC_2): Likewise.
347 (DIVWE): Add extended divide builtin functions.
355 (DXEX): Add decimal floating-point builtin functions.
367 (CDTBCD): Add new BCD builtin functions.
371 (BCDADD_LT): Likewise.
372 (BCDADD_EQ): Likewise.
373 (BCDADD_GT): Likewise.
374 (BCDADD_OV): Likewise.
376 (BCDSUB_LT): Likewise.
377 (BCDSUB_EQ): Likewise.
378 (BCDSUB_GT): Likewise.
379 (BCDSUB_OV): Likewise.
380 (PACK_TD): Add new pack/unpack 128-bit type builtin functions.
381 (UNPACK_TD): Likewise.
383 (UNPACK_TF): Likewise.
384 (UNPACK_TF_0): Likewise.
385 (UNPACK_TF_1): Likewise.
386 (PACK_V1TI): Likewise.
387 (UNPACK_V1TI): Likewise.
389 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Add
390 support for decimal floating point builtin functions.
391 (rs6000_expand_ternop_builtin): Add checks for the new builtin
392 functions that take constant arguments.
393 (rs6000_invalid_builtin): Add decimal floating point builtin
395 (rs6000_init_builtins): Setup long double, _Decimal64, and
396 _Decimal128 types for new builtin functions.
397 (builtin_function_type): Set the unsigned flags appropriately for
398 the new builtin functions.
399 (rs6000_opt_masks): Add support for decimal floating point builtin
402 * config/rs6000/rs6000.h (RS6000_BTM_DFP): Add support for decimal
403 floating point builtin functions.
404 (RS6000_BTM_COMMON): Likewise.
405 (RS6000_BTI_long_double): Likewise.
406 (RS6000_BTI_dfloat64): Likewise.
407 (RS6000_BTI_dfloat128): Likewise.
408 (long_double_type_internal_node): Likewise.
409 (dfloat64_type_internal_node): Likewise.
410 (dfloat128_type_internal_node): Likewise.
412 * config/rs6000/altivec.h (UNSPEC_BCDADD): Add support for ISA
413 2.07 bcd arithmetic instructions.
414 (UNSPEC_BCDSUB): Likewise.
415 (UNSPEC_BCD_OVERFLOW): Likewise.
416 (UNSPEC_BCD_ADD_SUB): Likewise.
417 (bcd_add_sub): Likewise.
418 (BCD_TEST): Likewise.
419 (bcd<bcd_add_sub>): Likewise.
420 (bcd<bcd_add_sub>_test): Likewise.
421 (bcd<bcd_add_sub>_test2): Likewise.
422 (bcd<bcd_add_sub>_<code>): Likewise.
423 (peephole2 for combined bcd ops): Likewise.
425 * config/rs6000/dfp.md (UNSPEC_DDEDPD): Add support for new
426 decimal floating point builtin functions.
427 (UNSPEC_DENBCD): Likewise.
428 (UNSPEC_DXEX): Likewise.
429 (UNSPEC_DIEX): Likewise.
430 (UNSPEC_DSCLI): Likewise.
431 (UNSPEC_DSCRI): Likewise.
432 (D64_D128): Likewise.
433 (dfp_suffix): Likewise.
434 (dfp_ddedpd_<mode>): Likewise.
435 (dfp_denbcd_<mode>): Likewise.
436 (dfp_dxex_<mode>): Likewise.
437 (dfp_diex_<mode>): Likewise.
438 (dfp_dscli_<mode>): Likewise.
439 (dfp_dscri_<mode>): Likewise.
441 * config/rs6000/rs6000.md (UNSPEC_ADDG6S): Add support for new BCD
443 (UNSPEC_CDTBCD): Likewise.
444 (UNSPEC_CBCDTD): Likewise.
445 (UNSPEC_DIVE): Add support for new extended divide builtin
447 (UNSPEC_DIVEO): Likewise.
448 (UNSPEC_DIVEU): Likewise.
449 (UNSPEC_DIVEUO): Likewise.
450 (UNSPEC_UNPACK_128BIT): Add support for new builtin functions to
451 pack/unpack 128-bit types.
452 (UNSPEC_PACK_128BIT): Likewise.
453 (idiv_ldiv): New mode attribute to set the 32/64-bit divide type.
454 (udiv<mode>3): Use idiv_ldiv mode attribute.
455 (div<mode>3): Likewise.
456 (addg6s): Add new BCD builtin functions.
459 (UNSPEC_DIV_EXTEND): Add support for new extended divide
461 (div_extend): Likewise.
462 (div<div_extend>_<mode>"): Likewise.
463 (FP128_64): Add support for new builtin functions to pack/unpack
465 (unpack<mode>): Likewise.
466 (unpacktf_0): Likewise.
467 (unpacktf_1): Likewise.
468 (unpack<mode>_dm): Likewise.
469 (unpack<mode>_nodm): Likewise.
470 (pack<mode>): Likewise.
471 (unpackv1ti): Likewise.
472 (packv1ti): Likewise.
474 2014-04-24 Vishnu K S <Vishnu.k_s@atmel.com>
476 * gcc/config/avr/avr.c: Add comment on why -fdelete-null-pointer-checks
479 2014-04-24 Jakub Jelinek <jakub@redhat.com>
481 * tree.h (OMP_CLAUSE_LINEAR_GIMPLE_SEQ): Define.
482 * gimplify.c (omp_is_private): Change last argument's type to int.
483 Only diagnose lastprivate if the simd argument is 1, only diagnose
484 linear if the simd argument is 2.
485 (gimplify_omp_for): Adjust omp_is_private callers. When adding
486 lastprivate or private, add the clause to OMP_FOR_CLAUSES. Pass
487 GOVD_EXPLICIT to omp_add_variable. For simd with collapse == 1
488 create OMP_CLAUSE_LINEAR rather than OMP_CLAUSE_PRIVATE for var.
489 If var != decl and decl is in OMP_CLAUSE_LINEAR, gimplify decl
490 increment to OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
491 * omp-low.c (scan_sharing_clauses, lower_lastprivate_clauses): Handle
492 OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
493 * tree-nested.c (convert_nonlocal_omp_clauses,
494 convert_local_omp_clauses): Handle OMP_CLAUSE_LINEAR.
496 2014-04-24 Segher Boessenkool <segher@kernel.crashing.org>
499 * config/m68k/m68k.md (extendplussidi): Don't allow memory for
502 2014-04-24 Dimitris Papavasiliou <dpapavas@gmail.com>
504 * flag-types.h (enum ivar_visibility): Add.
506 2014-04-24 Trevor Saunders <tsaunders@mozilla.com>
508 * config/sh/sh_treg_combine.c (sh_treg_combine::execute): Take
511 2014-04-24 Alan Lawrence <alan.lawrence@arm.com>
513 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Enable for bigendian.
515 2014-04-24 Radovan Obradovic <robradovic@mips.com>
516 Tom de Vries <tom@codesourcery.com>
518 * reg-notes.def (REG_NOTE (CALL_DECL)): New reg-note REG_CALL_DECL.
519 * calls.c (expand_call, emit_library_call_value_1): Add REG_CALL_DECL
521 * combine.c (distribute_notes): Handle REG_CALL_DECL reg-note.
522 * emit-rtl.c (try_split): Same.
524 2014-04-24 Radovan Obradovic <robradovic@mips.com>
525 Tom de Vries <tom@codesourcery.com>
527 * common.opt (fuse-caller-save): New option.
529 2014-04-24 Tejas Belagod <tejas.belagod@arm.com>
531 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Reverse order of elements
534 2014-04-24 Richard Biener <rguenther@suse.de>
536 * expr.c (expand_expr_real_1): Avoid gimple_assign_rhs_to_tree
537 during TER and instead use the sepops interface for expanding
538 non-GIMPLE_SINGLE_RHS.
540 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
542 * config/i386/sol2.h (ASM_PREFERRED_EH_DATA_FORMAT): Only redefine
543 if not HAVE_AS_IX86_DIFF_SECT_DELTA.
545 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
547 * configure.ac (gcc_cv_as_cfi_directive): Support Solaris/x86
548 assembler 64-bit option.
549 * configure: Regenerate.
551 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
553 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Check
554 TARGET_SIMD rather than TARGET_GENERAL_REGS_ONLY.
555 (TARGET_SIMD): Take AARCH64_ISA_SIMD into account.
556 (TARGET_FLOAT): Take AARCH64_ISA_FP into account.
557 (TARGET_CRYPTO): Take TARGET_SIMD into account.
559 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
561 * config/aarch64/aarch64-builtins.c
562 (aarch64_builtin_vectorized_function): Handle BUILT_IN_BSWAP16,
563 BUILT_IN_BSWAP32, BUILT_IN_BSWAP64.
564 * config/aarch64/aarch64-simd.md (bswap<mode>): New pattern.
565 * config/aarch64/aarch64-simd-builtins.def: Define vector bswap
567 * config/aarch64/iterator.md (VDQHSD): New mode iterator.
568 (Vrevsuff): New mode attribute.
570 2014-04-24 Terry Guo <terry.guo@arm.com>
572 * config/arm/arm.h (machine_function): Define variable
573 after_arm_reorg here.
574 * config/arm/arm.c (after_arm_reorg): Remove the definition.
575 (arm_split_constant): Update the way to access variable
578 (arm_output_function_epilogue): Remove the reset of after_arm_reorg.
580 2014-04-23 Tom de Vries <tom@codesourcery.com>
582 * target-hooks-macros.h: Fix DEFHOOKPOD argument order in comment.
584 2014-04-23 David Malcolm <dmalcolm@redhat.com>
586 * is-a.h: Update comments to reflect the following changes to the
587 "pointerness" of the API, making the template parameter match the
588 return type, allowing use of is-a.h with typedefs of pointers.
589 (is_a_helper::cast): Return a T rather then a pointer to a T, so
590 that the return type matches the parameter to the is_a_helper.
592 (dyn_cast): Likewise.
594 * cgraph.c (cgraph_node_for_asm): Update for removal of implicit
595 pointer from the is-a.h API.
597 * cgraph.h (is_a_helper <cgraph_node>::test): Convert to...
598 (is_a_helper <cgraph_node *>::test): ...this, matching change to
600 (is_a_helper <varpool_node>::test): Likewise, convert to...
601 (is_a_helper <varpool_node *>::test): ...this.
603 (varpool_first_variable): Update for removal of implicit pointer
605 (varpool_next_variable): Likewise.
606 (varpool_first_static_initializer): Likewise.
607 (varpool_next_static_initializer): Likewise.
608 (varpool_first_defined_variable): Likewise.
609 (varpool_next_defined_variable): Likewise.
610 (cgraph_first_defined_function): Likewise.
611 (cgraph_next_defined_function): Likewise.
612 (cgraph_first_function): Likewise.
613 (cgraph_next_function): Likewise.
614 (cgraph_first_function_with_gimple_body): Likewise.
615 (cgraph_next_function_with_gimple_body): Likewise.
616 (cgraph_alias_target): Likewise.
617 (varpool_alias_target): Likewise.
618 (cgraph_function_or_thunk_node): Likewise.
619 (varpool_variable_node): Likewise.
620 (symtab_real_symbol_p): Likewise.
621 * cgraphunit.c (referred_to_p): Likewise.
622 (analyze_functions): Likewise.
623 (handle_alias_pairs): Likewise.
624 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
625 * gimple-ssa.h (gimple_vuse_op): Likewise.
626 (gimple_vdef_op): Likewise.
627 * gimple-streamer-in.c (input_gimple_stmt): Likewise.
628 * gimple.c (gimple_build_asm_1): Likewise.
629 (gimple_build_try): Likewise.
630 (gimple_build_resx): Likewise.
631 (gimple_build_eh_dispatch): Likewise.
632 (gimple_build_omp_for): Likewise.
633 (gimple_omp_for_set_clauses): Likewise.
635 * gimple.h (is_a_helper <gimple_statement_asm>::test): Convert to...
636 (is_a_helper <gimple_statement_asm *>::test): ...this.
637 (is_a_helper <gimple_statement_bind>::test): Convert to...
638 (is_a_helper <gimple_statement_bind *>::test): ...this.
639 (is_a_helper <gimple_statement_call>::test): Convert to...
640 (is_a_helper <gimple_statement_call *>::test): ...this.
641 (is_a_helper <gimple_statement_catch>::test): Convert to...
642 (is_a_helper <gimple_statement_catch *>::test): ...this.
643 (is_a_helper <gimple_statement_resx>::test): Convert to...
644 (is_a_helper <gimple_statement_resx *>::test): ...this.
645 (is_a_helper <gimple_statement_eh_dispatch>::test): Convert to...
646 (is_a_helper <gimple_statement_eh_dispatch *>::test): ...this.
647 (is_a_helper <gimple_statement_eh_else>::test): Convert to...
648 (is_a_helper <gimple_statement_eh_else *>::test): ...this.
649 (is_a_helper <gimple_statement_eh_filter>::test): Convert to...
650 (is_a_helper <gimple_statement_eh_filter *>::test): ...this.
651 (is_a_helper <gimple_statement_eh_mnt>::test): Convert to...
652 (is_a_helper <gimple_statement_eh_mnt *>::test): ...this.
653 (is_a_helper <gimple_statement_omp_atomic_load>::test): Convert to...
654 (is_a_helper <gimple_statement_omp_atomic_load *>::test): ...this.
655 (is_a_helper <gimple_statement_omp_atomic_store>::test): Convert to...
656 (is_a_helper <gimple_statement_omp_atomic_store *>::test): ...this.
657 (is_a_helper <gimple_statement_omp_return>::test): Convert to...
658 (is_a_helper <gimple_statement_omp_return *>::test): ...this.
659 (is_a_helper <gimple_statement_omp_continue>::test): Convert to...
660 (is_a_helper <gimple_statement_omp_continue *>::test): ...this.
661 (is_a_helper <gimple_statement_omp_critical>::test): Convert to...
662 (is_a_helper <gimple_statement_omp_critical *>::test): ...this.
663 (is_a_helper <gimple_statement_omp_for>::test): Convert to...
664 (is_a_helper <gimple_statement_omp_for *>::test): ...this.
665 (is_a_helper <gimple_statement_omp_taskreg>::test): Convert to...
666 (is_a_helper <gimple_statement_omp_taskreg *>::test): ...this.
667 (is_a_helper <gimple_statement_omp_parallel>::test): Convert to...
668 (is_a_helper <gimple_statement_omp_parallel *>::test): ...this.
669 (is_a_helper <gimple_statement_omp_target>::test): Convert to...
670 (is_a_helper <gimple_statement_omp_target *>::test): ...this.
671 (is_a_helper <gimple_statement_omp_sections>::test): Convert to...
672 (is_a_helper <gimple_statement_omp_sections *>::test): ...this.
673 (is_a_helper <gimple_statement_omp_single>::test): Convert to...
674 (is_a_helper <gimple_statement_omp_single *>::test): ...this.
675 (is_a_helper <gimple_statement_omp_teams>::test): Convert to...
676 (is_a_helper <gimple_statement_omp_teams *>::test): ...this.
677 (is_a_helper <gimple_statement_omp_task>::test): Convert to...
678 (is_a_helper <gimple_statement_omp_task *>::test): ...this.
679 (is_a_helper <gimple_statement_phi>::test): Convert to...
680 (is_a_helper <gimple_statement_phi *>::test): ...this.
681 (is_a_helper <gimple_statement_transaction>::test): Convert to...
682 (is_a_helper <gimple_statement_transaction *>::test): ...this.
683 (is_a_helper <gimple_statement_try>::test): Convert to...
684 (is_a_helper <gimple_statement_try *>::test): ...this.
685 (is_a_helper <gimple_statement_wce>::test): Convert to...
686 (is_a_helper <gimple_statement_wce *>::test): ...this.
687 (is_a_helper <const gimple_statement_asm>::test): Convert to...
688 (is_a_helper <const gimple_statement_asm *>::test): ...this.
689 (is_a_helper <const gimple_statement_bind>::test): Convert to...
690 (is_a_helper <const gimple_statement_bind *>::test): ...this.
691 (is_a_helper <const gimple_statement_call>::test): Convert to...
692 (is_a_helper <const gimple_statement_call *>::test): ...this.
693 (is_a_helper <const gimple_statement_catch>::test): Convert to...
694 (is_a_helper <const gimple_statement_catch *>::test): ...this.
695 (is_a_helper <const gimple_statement_resx>::test): Convert to...
696 (is_a_helper <const gimple_statement_resx *>::test): ...this.
697 (is_a_helper <const gimple_statement_eh_dispatch>::test):
699 (is_a_helper <const gimple_statement_eh_dispatch *>::test):
701 (is_a_helper <const gimple_statement_eh_filter>::test): Convert
703 (is_a_helper <const gimple_statement_eh_filter *>::test): ...this.
704 (is_a_helper <const gimple_statement_omp_atomic_load>::test):
706 (is_a_helper <const gimple_statement_omp_atomic_load *>::test):
708 (is_a_helper <const gimple_statement_omp_atomic_store>::test):
710 (is_a_helper <const gimple_statement_omp_atomic_store *>::test):
712 (is_a_helper <const gimple_statement_omp_return>::test): Convert
714 (is_a_helper <const gimple_statement_omp_return *>::test): ...this.
715 (is_a_helper <const gimple_statement_omp_continue>::test): Convert
717 (is_a_helper <const gimple_statement_omp_continue *>::test): ...this.
718 (is_a_helper <const gimple_statement_omp_critical>::test): Convert
720 (is_a_helper <const gimple_statement_omp_critical *>::test): ...this.
721 (is_a_helper <const gimple_statement_omp_for>::test): Convert to...
722 (is_a_helper <const gimple_statement_omp_for *>::test): ...this.
723 (is_a_helper <const gimple_statement_omp_taskreg>::test): Convert
725 (is_a_helper <const gimple_statement_omp_taskreg *>::test): ...this.
726 (is_a_helper <const gimple_statement_omp_parallel>::test): Convert
728 (is_a_helper <const gimple_statement_omp_parallel *>::test): ...this.
729 (is_a_helper <const gimple_statement_omp_target>::test): Convert
731 (is_a_helper <const gimple_statement_omp_target *>::test): ...this.
732 (is_a_helper <const gimple_statement_omp_sections>::test): Convert
734 (is_a_helper <const gimple_statement_omp_sections *>::test): ...this.
735 (is_a_helper <const gimple_statement_omp_single>::test): Convert
737 (is_a_helper <const gimple_statement_omp_single *>::test): ...this.
738 (is_a_helper <const gimple_statement_omp_teams>::test): Convert
740 (is_a_helper <const gimple_statement_omp_teams *>::test): ...this.
741 (is_a_helper <const gimple_statement_omp_task>::test): Convert
743 (is_a_helper <const gimple_statement_omp_task *>::test): ...this.
744 (is_a_helper <const gimple_statement_phi>::test): Convert to...
745 (is_a_helper <const gimple_statement_phi *>::test): ...this.
746 (is_a_helper <const gimple_statement_transaction>::test): Convert
748 (is_a_helper <const gimple_statement_transaction *>::test): ...this.
749 (is_a_helper <const gimple_statement_with_ops>::test): Convert
751 (is_a_helper <const gimple_statement_with_ops *>::test): ...this.
752 (is_a_helper <gimple_statement_with_ops>::test): Convert to...
753 (is_a_helper <gimple_statement_with_ops *>::test): ...this.
754 (is_a_helper <const gimple_statement_with_memory_ops>::test):
756 (is_a_helper <const gimple_statement_with_memory_ops *>::test):
758 (is_a_helper <gimple_statement_with_memory_ops>::test): Convert
760 (is_a_helper <gimple_statement_with_memory_ops *>::test): ...this.
762 (gimple_use_ops): Update for removal of implicit pointer from the
764 (gimple_set_use_ops): Likewise.
765 (gimple_vuse): Likewise.
766 (gimple_vdef): Likewise.
767 (gimple_vuse_ptr): Likewise.
768 (gimple_vdef_ptr): Likewise.
769 (gimple_set_vuse): Likewise.
770 (gimple_set_vdef): Likewise.
771 (gimple_omp_return_set_lhs): Likewise.
772 (gimple_omp_return_lhs): Likewise.
773 (gimple_omp_return_lhs_ptr): Likewise.
774 (gimple_call_fntype): Likewise.
775 (gimple_call_set_fntype): Likewise.
776 (gimple_call_set_internal_fn): Likewise.
777 (gimple_call_use_set): Likewise.
778 (gimple_call_clobber_set): Likewise.
779 (gimple_bind_vars): Likewise.
780 (gimple_bind_set_vars): Likewise.
781 (gimple_bind_body_ptr): Likewise.
782 (gimple_bind_set_body): Likewise.
783 (gimple_bind_add_stmt): Likewise.
784 (gimple_bind_block): Likewise.
785 (gimple_bind_set_block): Likewise.
786 (gimple_asm_ninputs): Likewise.
787 (gimple_asm_noutputs): Likewise.
788 (gimple_asm_nclobbers): Likewise.
789 (gimple_asm_nlabels): Likewise.
790 (gimple_asm_input_op): Likewise.
791 (gimple_asm_input_op_ptr): Likewise.
792 (gimple_asm_output_op): Likewise.
793 (gimple_asm_output_op_ptr): Likewise.
794 (gimple_asm_set_output_op): Likewise.
795 (gimple_asm_clobber_op): Likewise.
796 (gimple_asm_set_clobber_op): Likewise.
797 (gimple_asm_label_op): Likewise.
798 (gimple_asm_set_label_op): Likewise.
799 (gimple_asm_string): Likewise.
800 (gimple_catch_types): Likewise.
801 (gimple_catch_types_ptr): Likewise.
802 (gimple_catch_handler_ptr): Likewise.
803 (gimple_catch_set_types): Likewise.
804 (gimple_catch_set_handler): Likewise.
805 (gimple_eh_filter_types): Likewise.
806 (gimple_eh_filter_types_ptr): Likewise.
807 (gimple_eh_filter_failure_ptr): Likewise.
808 (gimple_eh_filter_set_types): Likewise.
809 (gimple_eh_filter_set_failure): Likewise.
810 (gimple_eh_must_not_throw_fndecl): Likewise.
811 (gimple_eh_must_not_throw_set_fndecl): Likewise.
812 (gimple_eh_else_n_body_ptr): Likewise.
813 (gimple_eh_else_e_body_ptr): Likewise.
814 (gimple_eh_else_set_n_body): Likewise.
815 (gimple_eh_else_set_e_body): Likewise.
816 (gimple_try_eval_ptr): Likewise.
817 (gimple_try_cleanup_ptr): Likewise.
818 (gimple_try_set_eval): Likewise.
819 (gimple_try_set_cleanup): Likewise.
820 (gimple_wce_cleanup_ptr): Likewise.
821 (gimple_wce_set_cleanup): Likewise.
822 (gimple_phi_capacity): Likewise.
823 (gimple_phi_num_args): Likewise.
824 (gimple_phi_result): Likewise.
825 (gimple_phi_result_ptr): Likewise.
826 (gimple_phi_set_result): Likewise.
827 (gimple_phi_arg): Likewise.
828 (gimple_phi_set_arg): Likewise.
829 (gimple_resx_region): Likewise.
830 (gimple_resx_set_region): Likewise.
831 (gimple_eh_dispatch_region): Likewise.
832 (gimple_eh_dispatch_set_region): Likewise.
833 (gimple_omp_critical_name): Likewise.
834 (gimple_omp_critical_name_ptr): Likewise.
835 (gimple_omp_critical_set_name): Likewise.
836 (gimple_omp_for_clauses): Likewise.
837 (gimple_omp_for_clauses_ptr): Likewise.
838 (gimple_omp_for_set_clauses): Likewise.
839 (gimple_omp_for_collapse): Likewise.
840 (gimple_omp_for_index): Likewise.
841 (gimple_omp_for_index_ptr): Likewise.
842 (gimple_omp_for_set_index): Likewise.
843 (gimple_omp_for_initial): Likewise.
844 (gimple_omp_for_initial_ptr): Likewise.
845 (gimple_omp_for_set_initial): Likewise.
846 (gimple_omp_for_final): Likewise.
847 (gimple_omp_for_final_ptr): Likewise.
848 (gimple_omp_for_set_final): Likewise.
849 (gimple_omp_for_incr): Likewise.
850 (gimple_omp_for_incr_ptr): Likewise.
851 (gimple_omp_for_set_incr): Likewise.
852 (gimple_omp_for_pre_body_ptr): Likewise.
853 (gimple_omp_for_set_pre_body): Likewise.
854 (gimple_omp_parallel_clauses): Likewise.
855 (gimple_omp_parallel_clauses_ptr): Likewise.
856 (gimple_omp_parallel_set_clauses): Likewise.
857 (gimple_omp_parallel_child_fn): Likewise.
858 (gimple_omp_parallel_child_fn_ptr): Likewise.
859 (gimple_omp_parallel_set_child_fn): Likewise.
860 (gimple_omp_parallel_data_arg): Likewise.
861 (gimple_omp_parallel_data_arg_ptr): Likewise.
862 (gimple_omp_parallel_set_data_arg): Likewise.
863 (gimple_omp_task_clauses): Likewise.
864 (gimple_omp_task_clauses_ptr): Likewise.
865 (gimple_omp_task_set_clauses): Likewise.
866 (gimple_omp_task_child_fn): Likewise.
867 (gimple_omp_task_child_fn_ptr): Likewise.
868 (gimple_omp_task_set_child_fn): Likewise.
869 (gimple_omp_task_data_arg): Likewise.
870 (gimple_omp_task_data_arg_ptr): Likewise.
871 (gimple_omp_task_set_data_arg): Likewise.
872 (gimple_omp_taskreg_clauses): Likewise.
873 (gimple_omp_taskreg_clauses_ptr): Likewise.
874 (gimple_omp_taskreg_set_clauses): Likewise.
875 (gimple_omp_taskreg_child_fn): Likewise.
876 (gimple_omp_taskreg_child_fn_ptr): Likewise.
877 (gimple_omp_taskreg_set_child_fn): Likewise.
878 (gimple_omp_taskreg_data_arg): Likewise.
879 (gimple_omp_taskreg_data_arg_ptr): Likewise.
880 (gimple_omp_taskreg_set_data_arg): Likewise.
881 (gimple_omp_task_copy_fn): Likewise.
882 (gimple_omp_task_copy_fn_ptr): Likewise.
883 (gimple_omp_task_set_copy_fn): Likewise.
884 (gimple_omp_task_arg_size): Likewise.
885 (gimple_omp_task_arg_size_ptr): Likewise.
886 (gimple_omp_task_set_arg_size): Likewise.
887 (gimple_omp_task_arg_align): Likewise.
888 (gimple_omp_task_arg_align_ptr): Likewise.
889 (gimple_omp_task_set_arg_align): Likewise.
890 (gimple_omp_single_clauses): Likewise.
891 (gimple_omp_single_clauses_ptr): Likewise.
892 (gimple_omp_single_set_clauses): Likewise.
893 (gimple_omp_target_clauses): Likewise.
894 (gimple_omp_target_clauses_ptr): Likewise.
895 (gimple_omp_target_set_clauses): Likewise.
896 (gimple_omp_target_child_fn): Likewise.
897 (gimple_omp_target_child_fn_ptr): Likewise.
898 (gimple_omp_target_set_child_fn): Likewise.
899 (gimple_omp_target_data_arg): Likewise.
900 (gimple_omp_target_data_arg_ptr): Likewise.
901 (gimple_omp_target_set_data_arg): Likewise.
902 (gimple_omp_teams_clauses): Likewise.
903 (gimple_omp_teams_clauses_ptr): Likewise.
904 (gimple_omp_teams_set_clauses): Likewise.
905 (gimple_omp_sections_clauses): Likewise.
906 (gimple_omp_sections_clauses_ptr): Likewise.
907 (gimple_omp_sections_set_clauses): Likewise.
908 (gimple_omp_sections_control): Likewise.
909 (gimple_omp_sections_control_ptr): Likewise.
910 (gimple_omp_sections_set_control): Likewise.
911 (gimple_omp_for_set_cond): Likewise.
912 (gimple_omp_for_cond): Likewise.
913 (gimple_omp_atomic_store_set_val): Likewise.
914 (gimple_omp_atomic_store_val): Likewise.
915 (gimple_omp_atomic_store_val_ptr): Likewise.
916 (gimple_omp_atomic_load_set_lhs): Likewise.
917 (gimple_omp_atomic_load_lhs): Likewise.
918 (gimple_omp_atomic_load_lhs_ptr): Likewise.
919 (gimple_omp_atomic_load_set_rhs): Likewise.
920 (gimple_omp_atomic_load_rhs): Likewise.
921 (gimple_omp_atomic_load_rhs_ptr): Likewise.
922 (gimple_omp_continue_control_def): Likewise.
923 (gimple_omp_continue_control_def_ptr): Likewise.
924 (gimple_omp_continue_set_control_def): Likewise.
925 (gimple_omp_continue_control_use): Likewise.
926 (gimple_omp_continue_control_use_ptr): Likewise.
927 (gimple_omp_continue_set_control_use): Likewise.
928 (gimple_transaction_body_ptr): Likewise.
929 (gimple_transaction_label): Likewise.
930 (gimple_transaction_label_ptr): Likewise.
931 (gimple_transaction_set_body): Likewise.
932 (gimple_transaction_set_label): Likewise.
934 * ipa-devirt.c (build_type_inheritance_graph): Likewise.
935 * ipa-inline-analysis.c (inline_write_summary): Likewise.
936 * ipa-ref.c (ipa_record_reference): Likewise.
937 * ipa-reference.c (analyze_function): Likewise.
938 (ipa_reference_write_optimization_summary): Likewise.
939 * ipa.c (symtab_remove_unreachable_nodes): Likewise.
940 (address_taken_from_non_vtable_p): Likewise.
941 (comdat_can_be_unshared_p_1): Likewise.
942 * lto-cgraph.c (lto_output_ref): Likewise.
943 (add_references): Likewise.
944 (compute_ltrans_boundary): Likewise.
945 (output_symtab): Likewise.
946 (input_ref): Likewise.
947 (input_cgraph_1): Likewise.
948 (output_cgraph_opt_summary): Likewise.
949 * lto-streamer-out.c (lto_output): Likewise.
950 (output_symbol_p): Likewise.
951 * lto-streamer.h (lsei_next_function_in_partition): Likewise.
952 (lsei_start_function_in_partition): Likewise.
953 (lsei_next_variable_in_partition): Likewise.
954 (lsei_start_variable_in_partition): Likewise.
955 * symtab.c (insert_to_assembler_name_hash): Likewise.
956 (unlink_from_assembler_name_hash): Likewise.
957 (symtab_unregister_node): Likewise.
958 (symtab_remove_node): Likewise.
959 (dump_symtab_node): Likewise.
960 (verify_symtab_base): Likewise.
961 (verify_symtab_node): Likewise.
962 (symtab_make_decl_local): Likewise.
963 (symtab_alias_ultimate_target): Likewise.
964 (symtab_resolve_alias): Likewise.
965 (symtab_get_symbol_partitioning_class): Likewise.
966 * tree-phinodes.c (allocate_phi_node): Likewise.
967 (reserve_phi_args_for_new_edge): Likewise.
968 (remove_phi_args): Likewise.
969 * varpool.c (varpool_node_for_asm): Likewise.
970 (varpool_remove_unreferenced_decls): Likewise.
972 2014-04-23 Jeff Law <law@redhat.com>
974 PR tree-optimization/60902
975 * tree-ssa-threadedge.c
976 (record_temporary_equivalences_from_stmts_at_dest): Make sure to
977 invalidate outputs from statements that do not produce useful
978 outputs for threading.
980 2014-04-23 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
982 * config/aarch64/aarch64.md (stack_protect_set, stack_protect_test)
983 (stack_protect_set_<mode>, stack_protect_test_<mode>): Add
984 machine descriptions for Stack Smashing Protector.
986 2014-04-23 Richard Earnshaw <rearnsha@arm.com>
988 * aarch64.md (<optab>_rol<mode>3): New pattern.
989 (<optab>_rolsi3_uxtw): Likewise.
990 * aarch64.c (aarch64_strip_shift): Handle ROTATE and ROTATERT.
992 2014-04-23 James Greenhalgh <james.greenhalgh@arm.com>
994 * config/arm/arm.c (arm_cortex_a57_tune): Initialize all fields.
995 (arm_cortex_a12_tune): Likewise.
997 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
999 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle BSWAP.
1001 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1003 * config/arm/arm.md (arm_rev16si2): New pattern.
1004 (arm_rev16si2_alt): Likewise.
1005 * config/arm/arm.c (arm_new_rtx_costs): Handle rev16 case.
1007 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1009 * config/aarch64/aarch64.md (rev16<mode>2): New pattern.
1010 (rev16<mode>2_alt): Likewise.
1011 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle rev16 case.
1012 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p): New.
1013 (aarch_rev16_shleft_mask_imm_p): Likewise.
1014 (aarch_rev16_p_1): Likewise.
1015 (aarch_rev16_p): Likewise.
1016 * config/arm/aarch-common-protos.h (aarch_rev16_p): Declare extern.
1017 (aarch_rev16_shright_mask_imm_p): Likewise.
1018 (aarch_rev16_shleft_mask_imm_p): Likewise.
1020 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1022 * config/arm/aarch-common-protos.h (alu_cost_table): Add rev field.
1023 * config/arm/aarch-cost-tables.h (generic_extra_costs): Specify
1025 (cortex_a53_extra_costs): Likewise.
1026 (cortex_a57_extra_costs): Likewise.
1027 * config/arm/arm.c (cortexa9_extra_costs): Likewise.
1028 (cortexa7_extra_costs): Likewise.
1029 (cortexa8_extra_costs): Likewise.
1030 (cortexa12_extra_costs): Likewise.
1031 (cortexa15_extra_costs): Likewise.
1032 (v7m_extra_costs): Likewise.
1033 (arm_new_rtx_costs): Handle BSWAP.
1035 2013-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1037 * config/arm/arm.c (cortexa8_extra_costs): New table.
1038 (arm_cortex_a8_tune): New tuning struct.
1039 * config/arm/arm-cores.def (cortex-a8): Use cortex_a8 tuning struct.
1041 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1043 * config/arm/arm.c (arm_new_rtx_costs): Handle FMA.
1045 2014-04-23 Richard Biener <rguenther@suse.de>
1047 * Makefile.in (OBJS): Remove loop-unswitch.o.
1048 * loop-unswitch.c: Delete.
1049 * tree-pass.h (make_pass_rtl_unswitch): Remove.
1050 * passes.def (pass_rtl_unswitch): Likewise.
1051 * loop-init.c (gate_rtl_unswitch): Likewise.
1052 (rtl_unswitch): Likewise.
1053 (pass_data_rtl_unswitch): Likewise.
1054 (pass_rtl_unswitch): Likewise.
1055 (make_pass_rtl_unswitch): Likewise.
1056 * rtl.h (reversed_condition): Likewise.
1057 (compare_and_jump_seq): Likewise.
1058 * loop-iv.c (reversed_condition): Move here from loop-unswitch.c
1060 * loop-unroll.c (compare_and_jump_seq): Likewise.
1062 2014-04-23 Richard Biener <rguenther@suse.de>
1064 PR tree-optimization/60903
1065 * tree-ssa-loop-im.c (analyze_memory_references): Remove
1066 commented code block.
1067 (execute_sm_if_changed): Properly apply IRREDUCIBLE_LOOP
1068 loop flags to newly created BBs and edges.
1070 2014-04-23 Nick Clifton <nickc@redhat.com>
1072 * config/msp430/msp430.c (msp430_handle_option): Move function
1074 (msp430_option_override): Simplify mcu and mcpu option handling.
1075 (msp430_is_f5_mcu): Rename to msp430_use_f5_series_hwmult. Add
1076 support for -mhwmult command line option.
1077 (has_32bit_hwmult): Rename to use_32bit_hwmult. Add support for
1078 -mhwmult command line option.
1079 (msp430_hwmult_enabled): Delete.
1080 (msp43o_output_labelref): Add support for -mhwmult command line
1082 * config/msp430/msp430.md (mulhisi3, umulhisi3, mulsidi3)
1083 (umulsidi3): Likewise.
1084 * config/msp430/msp430.opt (mmcu): Add Report attribute.
1085 (mcpu, mlarge, msmall): Likewise.
1086 (mhwmult): New option.
1087 * config/msp430/msp430-protos.h (msp430_hwmult_enabled): Remove
1089 (msp430_is_f5_mcu): Remove prototype.
1090 (msp430_use_f5_series_hwmult): Add prototype.
1091 * config/msp430/msp430-opts.h: New file.
1092 * common/config/msp430: New directory.
1093 * common/config/msp430/msp430-common.c: New file.
1094 * config.gcc (msp430): Remove target_has_targetm_common.
1095 * doc/invoke.texi: Document -mhwmult command line option.
1097 2014-04-23 Nick Clifton <nickc@redhat.com>
1099 * config/i386/cygwin.h (ENDFILE_SPEC): Include
1100 default-manifest.o if it can be found in the search path.
1101 * config/i386/mingw32.h (ENDFILE_SPEC): Likewise.
1103 2014-04-23 Terry Guo <terry.guo@arm.com>
1105 * config/arm/arm.h (ASM_APP_OFF): Re-define it in a cleaner way.
1107 2014-04-23 Richard Biener <rguenther@suse.de>
1110 * tree-inline.c (declare_return_variable): Use mark_addressable.
1112 2014-04-23 Richard Biener <rguenther@suse.de>
1115 * loop-init.c (loop_optimizer_init): Make sure to apply
1116 LOOPS_MAY_HAVE_MULTIPLE_LATCHES before fixing up loops.
1118 2014-04-22 Jakub Jelinek <jakub@redhat.com>
1121 * common.opt (fsanitize-recover, fsanitize-undefined-trap-on-error):
1123 * gcc.c (sanitize_spec_function): Don't return "" for "undefined"
1124 if flag_sanitize_undefined_trap_on_error.
1125 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW_ABORT,
1126 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS_ABORT,
1127 BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE_ABORT,
1128 BUILT_IN_UBSAN_HANDLE_TYPE_MISMATCH_ABORT,
1129 BUILT_IN_UBSAN_HANDLE_ADD_OVERFLOW_ABORT,
1130 BUILT_IN_UBSAN_HANDLE_SUB_OVERFLOW_ABORT,
1131 BUILT_IN_UBSAN_HANDLE_MUL_OVERFLOW_ABORT,
1132 BUILT_IN_UBSAN_HANDLE_NEGATE_OVERFLOW_ABORT,
1133 BUILT_IN_UBSAN_HANDLE_LOAD_INVALID_VALUE_ABORT): New builtins.
1134 * ubsan.c (ubsan_instrument_unreachable): Return
1135 __builtin_trap () if flag_sanitize_undefined_trap_on_error.
1136 (ubsan_expand_null_ifn): Emit __builtin_trap ()
1137 if flag_sanitize_undefined_trap_on_error and
1138 __ubsan_handle_type_mismatch_abort if !flag_sanitize_recover.
1139 (ubsan_expand_null_ifn, ubsan_build_overflow_builtin,
1140 instrument_bool_enum_load): Emit __builtin_trap () if
1141 flag_sanitize_undefined_trap_on_error and
1142 __builtin_handle_*_abort () if !flag_sanitize_recover.
1143 * doc/invoke.texi (-fsanitize-recover,
1144 -fsanitize-undefined-trap-on-error): Document.
1146 2014-04-22 Christian Bruel <christian.bruel@st.com>
1148 * config/sh/sh.md (mov<mode>): Replace movQIHI.
1149 Force immediates to SImode.
1151 2014-04-22 Sandra Loosemore <sandra@codesourcery.com>
1153 * config/nios2/nios2.md (UNSPEC_ROUND): New.
1155 * config/nios2/nios2.opt (mno-custom-round, mcustom-round=): New.
1156 * config/nios2/nios2-opts.h (N2FPU_ALL_CODES): Add round.
1157 * config/nios2/nios2.c (N2F_NO_ERRNO): Define.
1158 (nios2_fpu_insn): Add entry for round.
1159 (N2FPU_NO_ERRNO_P): Define.
1160 (nios2_custom_check_insns): Add check for N2F_NO_ERRNO and
1162 * doc/invoke.texi (Nios II Options): Document -mcustom-round.
1164 2014-04-22 Richard Henderson <rth@redhat.com>
1166 * config/aarch64/aarch64 (addti3, subti3): New expanders.
1167 (add<GPI>3_compare0): Remove leading * from name.
1168 (add<GPI>3_carryin): Likewise.
1169 (sub<GPI>3_compare0): Likewise.
1170 (sub<GPI>3_carryin): Likewise.
1171 (<su_optab>mulditi3): New expander.
1172 (multi3): New expander.
1173 (madd<GPI>): Remove leading * from name.
1175 2014-04-22 Martin Jambor <mjambor@suse.cz>
1177 * cgraphclones.c (cgraph_function_versioning): Copy
1178 ipa_transforms_to_apply instead of asserting it is empty.
1180 2014-04-22 H.J. Lu <hongjiu.lu@intel.com>
1183 * config/i386/i386.c (ix86_expand_set_or_movmem): Call counter_mode
1184 on count_exp to get mode.
1186 2014-04-22 Andrew Pinski <apinski@cavium.com>
1188 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
1189 Handle TLS for ILP32.
1190 * config/aarch64/aarch64.md (tlsie_small): Rename to ...
1191 (tlsie_small_<mode>): this and handle PTR.
1192 (tlsie_small_sidi): New pattern.
1193 (tlsle_small): Change to an expand to handle ILP32.
1194 (tlsle_small_<mode>): New pattern.
1195 (tlsdesc_small): Rename to ...
1196 (tlsdesc_small_<mode>): this and handle PTR.
1198 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1200 * config/aarch64/aarch64.c (TARGET_FLAGS_REGNUM): Define.
1202 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1204 * config/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
1205 (aarch64_types_signed_unsigned_qualifiers): Qualifier added.
1206 (aarch64_types_signed_poly_qualifiers): Likewise.
1207 (aarch64_types_unsigned_signed_qualifiers): Likewise.
1208 (aarch64_types_poly_signed_qualifiers): Likewise.
1209 (TYPES_REINTERP_SS): Type macro added.
1210 (TYPES_REINTERP_SU): Likewise.
1211 (TYPES_REINTERP_SP): Likewise.
1212 (TYPES_REINTERP_US): Likewise.
1213 (TYPES_REINTERP_PS): Likewise.
1214 (aarch64_fold_builtin): New expression folding added.
1215 * config/aarch64/aarch64-simd-builtins.def (REINTERP):
1216 Declarations removed.
1217 (REINTERP_SS): Declarations added.
1218 (REINTERP_US): Likewise.
1219 (REINTERP_PS): Likewise.
1220 (REINTERP_SU): Likewise.
1221 (REINTERP_SP): Likewise.
1222 * config/aarch64/arm_neon.h (vreinterpret_p8_f64): Implemented.
1223 (vreinterpretq_p8_f64): Likewise.
1224 (vreinterpret_p16_f64): Likewise.
1225 (vreinterpretq_p16_f64): Likewise.
1226 (vreinterpret_f32_f64): Likewise.
1227 (vreinterpretq_f32_f64): Likewise.
1228 (vreinterpret_f64_f32): Likewise.
1229 (vreinterpret_f64_p8): Likewise.
1230 (vreinterpret_f64_p16): Likewise.
1231 (vreinterpret_f64_s8): Likewise.
1232 (vreinterpret_f64_s16): Likewise.
1233 (vreinterpret_f64_s32): Likewise.
1234 (vreinterpret_f64_s64): Likewise.
1235 (vreinterpret_f64_u8): Likewise.
1236 (vreinterpret_f64_u16): Likewise.
1237 (vreinterpret_f64_u32): Likewise.
1238 (vreinterpret_f64_u64): Likewise.
1239 (vreinterpretq_f64_f32): Likewise.
1240 (vreinterpretq_f64_p8): Likewise.
1241 (vreinterpretq_f64_p16): Likewise.
1242 (vreinterpretq_f64_s8): Likewise.
1243 (vreinterpretq_f64_s16): Likewise.
1244 (vreinterpretq_f64_s32): Likewise.
1245 (vreinterpretq_f64_s64): Likewise.
1246 (vreinterpretq_f64_u8): Likewise.
1247 (vreinterpretq_f64_u16): Likewise.
1248 (vreinterpretq_f64_u32): Likewise.
1249 (vreinterpretq_f64_u64): Likewise.
1250 (vreinterpret_s64_f64): Likewise.
1251 (vreinterpretq_s64_f64): Likewise.
1252 (vreinterpret_u64_f64): Likewise.
1253 (vreinterpretq_u64_f64): Likewise.
1254 (vreinterpret_s8_f64): Likewise.
1255 (vreinterpretq_s8_f64): Likewise.
1256 (vreinterpret_s16_f64): Likewise.
1257 (vreinterpretq_s16_f64): Likewise.
1258 (vreinterpret_s32_f64): Likewise.
1259 (vreinterpretq_s32_f64): Likewise.
1260 (vreinterpret_u8_f64): Likewise.
1261 (vreinterpretq_u8_f64): Likewise.
1262 (vreinterpret_u16_f64): Likewise.
1263 (vreinterpretq_u16_f64): Likewise.
1264 (vreinterpret_u32_f64): Likewise.
1265 (vreinterpretq_u32_f64): Likewise.
1267 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1269 * config/aarch64/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
1270 * config/aarch64/aarch64/aarch64-simd-builtins.def (REINTERP): Removed.
1271 (vreinterpret_p8_s8): Likewise.
1272 * config/aarch64/aarch64/arm_neon.h (vreinterpret_p8_s8): Uses cast.
1273 (vreinterpret_p8_s16): Likewise.
1274 (vreinterpret_p8_s32): Likewise.
1275 (vreinterpret_p8_s64): Likewise.
1276 (vreinterpret_p8_f32): Likewise.
1277 (vreinterpret_p8_u8): Likewise.
1278 (vreinterpret_p8_u16): Likewise.
1279 (vreinterpret_p8_u32): Likewise.
1280 (vreinterpret_p8_u64): Likewise.
1281 (vreinterpret_p8_p16): Likewise.
1282 (vreinterpretq_p8_s8): Likewise.
1283 (vreinterpretq_p8_s16): Likewise.
1284 (vreinterpretq_p8_s32): Likewise.
1285 (vreinterpretq_p8_s64): Likewise.
1286 (vreinterpretq_p8_f32): Likewise.
1287 (vreinterpretq_p8_u8): Likewise.
1288 (vreinterpretq_p8_u16): Likewise.
1289 (vreinterpretq_p8_u32): Likewise.
1290 (vreinterpretq_p8_u64): Likewise.
1291 (vreinterpretq_p8_p16): Likewise.
1292 (vreinterpret_p16_s8): Likewise.
1293 (vreinterpret_p16_s16): Likewise.
1294 (vreinterpret_p16_s32): Likewise.
1295 (vreinterpret_p16_s64): Likewise.
1296 (vreinterpret_p16_f32): Likewise.
1297 (vreinterpret_p16_u8): Likewise.
1298 (vreinterpret_p16_u16): Likewise.
1299 (vreinterpret_p16_u32): Likewise.
1300 (vreinterpret_p16_u64): Likewise.
1301 (vreinterpret_p16_p8): Likewise.
1302 (vreinterpretq_p16_s8): Likewise.
1303 (vreinterpretq_p16_s16): Likewise.
1304 (vreinterpretq_p16_s32): Likewise.
1305 (vreinterpretq_p16_s64): Likewise.
1306 (vreinterpretq_p16_f32): Likewise.
1307 (vreinterpretq_p16_u8): Likewise.
1308 (vreinterpretq_p16_u16): Likewise.
1309 (vreinterpretq_p16_u32): Likewise.
1310 (vreinterpretq_p16_u64): Likewise.
1311 (vreinterpretq_p16_p8): Likewise.
1312 (vreinterpret_f32_s8): Likewise.
1313 (vreinterpret_f32_s16): Likewise.
1314 (vreinterpret_f32_s32): Likewise.
1315 (vreinterpret_f32_s64): Likewise.
1316 (vreinterpret_f32_u8): Likewise.
1317 (vreinterpret_f32_u16): Likewise.
1318 (vreinterpret_f32_u32): Likewise.
1319 (vreinterpret_f32_u64): Likewise.
1320 (vreinterpret_f32_p8): Likewise.
1321 (vreinterpret_f32_p16): Likewise.
1322 (vreinterpretq_f32_s8): Likewise.
1323 (vreinterpretq_f32_s16): Likewise.
1324 (vreinterpretq_f32_s32): Likewise.
1325 (vreinterpretq_f32_s64): Likewise.
1326 (vreinterpretq_f32_u8): Likewise.
1327 (vreinterpretq_f32_u16): Likewise.
1328 (vreinterpretq_f32_u32): Likewise.
1329 (vreinterpretq_f32_u64): Likewise.
1330 (vreinterpretq_f32_p8): Likewise.
1331 (vreinterpretq_f32_p16): Likewise.
1332 (vreinterpret_s64_s8): Likewise.
1333 (vreinterpret_s64_s16): Likewise.
1334 (vreinterpret_s64_s32): Likewise.
1335 (vreinterpret_s64_f32): Likewise.
1336 (vreinterpret_s64_u8): Likewise.
1337 (vreinterpret_s64_u16): Likewise.
1338 (vreinterpret_s64_u32): Likewise.
1339 (vreinterpret_s64_u64): Likewise.
1340 (vreinterpret_s64_p8): Likewise.
1341 (vreinterpret_s64_p16): Likewise.
1342 (vreinterpretq_s64_s8): Likewise.
1343 (vreinterpretq_s64_s16): Likewise.
1344 (vreinterpretq_s64_s32): Likewise.
1345 (vreinterpretq_s64_f32): Likewise.
1346 (vreinterpretq_s64_u8): Likewise.
1347 (vreinterpretq_s64_u16): Likewise.
1348 (vreinterpretq_s64_u32): Likewise.
1349 (vreinterpretq_s64_u64): Likewise.
1350 (vreinterpretq_s64_p8): Likewise.
1351 (vreinterpretq_s64_p16): Likewise.
1352 (vreinterpret_u64_s8): Likewise.
1353 (vreinterpret_u64_s16): Likewise.
1354 (vreinterpret_u64_s32): Likewise.
1355 (vreinterpret_u64_s64): Likewise.
1356 (vreinterpret_u64_f32): Likewise.
1357 (vreinterpret_u64_u8): Likewise.
1358 (vreinterpret_u64_u16): Likewise.
1359 (vreinterpret_u64_u32): Likewise.
1360 (vreinterpret_u64_p8): Likewise.
1361 (vreinterpret_u64_p16): Likewise.
1362 (vreinterpretq_u64_s8): Likewise.
1363 (vreinterpretq_u64_s16): Likewise.
1364 (vreinterpretq_u64_s32): Likewise.
1365 (vreinterpretq_u64_s64): Likewise.
1366 (vreinterpretq_u64_f32): Likewise.
1367 (vreinterpretq_u64_u8): Likewise.
1368 (vreinterpretq_u64_u16): Likewise.
1369 (vreinterpretq_u64_u32): Likewise.
1370 (vreinterpretq_u64_p8): Likewise.
1371 (vreinterpretq_u64_p16): Likewise.
1372 (vreinterpret_s8_s16): Likewise.
1373 (vreinterpret_s8_s32): Likewise.
1374 (vreinterpret_s8_s64): Likewise.
1375 (vreinterpret_s8_f32): Likewise.
1376 (vreinterpret_s8_u8): Likewise.
1377 (vreinterpret_s8_u16): Likewise.
1378 (vreinterpret_s8_u32): Likewise.
1379 (vreinterpret_s8_u64): Likewise.
1380 (vreinterpret_s8_p8): Likewise.
1381 (vreinterpret_s8_p16): Likewise.
1382 (vreinterpretq_s8_s16): Likewise.
1383 (vreinterpretq_s8_s32): Likewise.
1384 (vreinterpretq_s8_s64): Likewise.
1385 (vreinterpretq_s8_f32): Likewise.
1386 (vreinterpretq_s8_u8): Likewise.
1387 (vreinterpretq_s8_u16): Likewise.
1388 (vreinterpretq_s8_u32): Likewise.
1389 (vreinterpretq_s8_u64): Likewise.
1390 (vreinterpretq_s8_p8): Likewise.
1391 (vreinterpretq_s8_p16): Likewise.
1392 (vreinterpret_s16_s8): Likewise.
1393 (vreinterpret_s16_s32): Likewise.
1394 (vreinterpret_s16_s64): Likewise.
1395 (vreinterpret_s16_f32): Likewise.
1396 (vreinterpret_s16_u8): Likewise.
1397 (vreinterpret_s16_u16): Likewise.
1398 (vreinterpret_s16_u32): Likewise.
1399 (vreinterpret_s16_u64): Likewise.
1400 (vreinterpret_s16_p8): Likewise.
1401 (vreinterpret_s16_p16): Likewise.
1402 (vreinterpretq_s16_s8): Likewise.
1403 (vreinterpretq_s16_s32): Likewise.
1404 (vreinterpretq_s16_s64): Likewise.
1405 (vreinterpretq_s16_f32): Likewise.
1406 (vreinterpretq_s16_u8): Likewise.
1407 (vreinterpretq_s16_u16): Likewise.
1408 (vreinterpretq_s16_u32): Likewise.
1409 (vreinterpretq_s16_u64): Likewise.
1410 (vreinterpretq_s16_p8): Likewise.
1411 (vreinterpretq_s16_p16): Likewise.
1412 (vreinterpret_s32_s8): Likewise.
1413 (vreinterpret_s32_s16): Likewise.
1414 (vreinterpret_s32_s64): Likewise.
1415 (vreinterpret_s32_f32): Likewise.
1416 (vreinterpret_s32_u8): Likewise.
1417 (vreinterpret_s32_u16): Likewise.
1418 (vreinterpret_s32_u32): Likewise.
1419 (vreinterpret_s32_u64): Likewise.
1420 (vreinterpret_s32_p8): Likewise.
1421 (vreinterpret_s32_p16): Likewise.
1422 (vreinterpretq_s32_s8): Likewise.
1423 (vreinterpretq_s32_s16): Likewise.
1424 (vreinterpretq_s32_s64): Likewise.
1425 (vreinterpretq_s32_f32): Likewise.
1426 (vreinterpretq_s32_u8): Likewise.
1427 (vreinterpretq_s32_u16): Likewise.
1428 (vreinterpretq_s32_u32): Likewise.
1429 (vreinterpretq_s32_u64): Likewise.
1430 (vreinterpretq_s32_p8): Likewise.
1431 (vreinterpretq_s32_p16): Likewise.
1432 (vreinterpret_u8_s8): Likewise.
1433 (vreinterpret_u8_s16): Likewise.
1434 (vreinterpret_u8_s32): Likewise.
1435 (vreinterpret_u8_s64): Likewise.
1436 (vreinterpret_u8_f32): Likewise.
1437 (vreinterpret_u8_u16): Likewise.
1438 (vreinterpret_u8_u32): Likewise.
1439 (vreinterpret_u8_u64): Likewise.
1440 (vreinterpret_u8_p8): Likewise.
1441 (vreinterpret_u8_p16): Likewise.
1442 (vreinterpretq_u8_s8): Likewise.
1443 (vreinterpretq_u8_s16): Likewise.
1444 (vreinterpretq_u8_s32): Likewise.
1445 (vreinterpretq_u8_s64): Likewise.
1446 (vreinterpretq_u8_f32): Likewise.
1447 (vreinterpretq_u8_u16): Likewise.
1448 (vreinterpretq_u8_u32): Likewise.
1449 (vreinterpretq_u8_u64): Likewise.
1450 (vreinterpretq_u8_p8): Likewise.
1451 (vreinterpretq_u8_p16): Likewise.
1452 (vreinterpret_u16_s8): Likewise.
1453 (vreinterpret_u16_s16): Likewise.
1454 (vreinterpret_u16_s32): Likewise.
1455 (vreinterpret_u16_s64): Likewise.
1456 (vreinterpret_u16_f32): Likewise.
1457 (vreinterpret_u16_u8): Likewise.
1458 (vreinterpret_u16_u32): Likewise.
1459 (vreinterpret_u16_u64): Likewise.
1460 (vreinterpret_u16_p8): Likewise.
1461 (vreinterpret_u16_p16): Likewise.
1462 (vreinterpretq_u16_s8): Likewise.
1463 (vreinterpretq_u16_s16): Likewise.
1464 (vreinterpretq_u16_s32): Likewise.
1465 (vreinterpretq_u16_s64): Likewise.
1466 (vreinterpretq_u16_f32): Likewise.
1467 (vreinterpretq_u16_u8): Likewise.
1468 (vreinterpretq_u16_u32): Likewise.
1469 (vreinterpretq_u16_u64): Likewise.
1470 (vreinterpretq_u16_p8): Likewise.
1471 (vreinterpretq_u16_p16): Likewise.
1472 (vreinterpret_u32_s8): Likewise.
1473 (vreinterpret_u32_s16): Likewise.
1474 (vreinterpret_u32_s32): Likewise.
1475 (vreinterpret_u32_s64): Likewise.
1476 (vreinterpret_u32_f32): Likewise.
1477 (vreinterpret_u32_u8): Likewise.
1478 (vreinterpret_u32_u16): Likewise.
1479 (vreinterpret_u32_u64): Likewise.
1480 (vreinterpret_u32_p8): Likewise.
1481 (vreinterpret_u32_p16): Likewise.
1482 (vreinterpretq_u32_s8): Likewise.
1483 (vreinterpretq_u32_s16): Likewise.
1484 (vreinterpretq_u32_s32): Likewise.
1485 (vreinterpretq_u32_s64): Likewise.
1486 (vreinterpretq_u32_f32): Likewise.
1487 (vreinterpretq_u32_u8): Likewise.
1488 (vreinterpretq_u32_u16): Likewise.
1489 (vreinterpretq_u32_u64): Likewise.
1490 (vreinterpretq_u32_p8): Likewise.
1491 (vreinterpretq_u32_p16): Likewise.
1493 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1495 * gcc/config/aarch64/aarch64-simd.md (aarch64_s<optab><mode>):
1497 * config/aarch64/aarch64-simd-builtins.def (sqneg): Iterator
1500 * config/aarch64/arm_neon.h (vqneg_s64): New intrinsic.
1501 (vqnegd_s64): Likewise.
1502 (vqabs_s64): Likewise.
1503 (vqabsd_s64): Likewise.
1505 2014-04-22 Richard Henderson <rth@redhat.com>
1507 * config/sparc/sparc.c (sparc_init_modes): Hoist GET_MODE_SIZE
1508 computation to the top of the loop.
1510 2014-04-22 Renlin <renlin.li@arm.com>
1511 Jiong Wang <jiong.wang@arm.com>
1513 * config/aarch64/aarch64.h (aarch64_frame): Delete "fp_lr_offset".
1514 * config/aarch64/aarch64.c (aarch64_layout_frame)
1515 (aarch64_initial_elimination_offset): Likewise.
1517 2014-04-22 Marcus Shawcroft <marcus.shawcroft@arm.com>
1519 * config/aarch64/aarch64.c (aarch64_initial_elimination_offset):
1522 2014-04-22 Richard Sandiford <rdsandiford@googlemail.com>
1524 * machmode.h (bitwise_mode_for_mode): Declare.
1525 * stor-layout.h (bitwise_type_for_mode): Likewise.
1526 * stor-layout.c (bitwise_mode_for_mode): New function.
1527 (bitwise_type_for_mode): Likewise.
1528 * builtins.c (fold_builtin_memory_op): Use it instead of
1529 int_mode_for_mode and build_nonstandard_integer_type.
1531 2014-04-22 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1533 * config.gcc (enable_obsolete): Remove *-*-solaris2.9*.
1534 (*-*-solaris2.[0-9] | *-*-solaris2.[0-9].*): Mark unsupported.
1535 (*-*-solaris2*): Simplify.
1536 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Likewise.
1537 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Remove
1538 *-*-solaris2.9* handling.
1540 * configure.ac (gcc_cv_as_hidden): Remove test for Solaris 9/x86
1542 (gcc_cv_ld_hidden): Remove *-*-solaris2.9* handling.
1543 (ld_tls_support): Remove i?86-*-solaris2.9, sparc*-*-solaris2.9
1545 (gcc_cv_as_gstabs_flag): Remove workaround for Solaris 9/x86 as bug.
1546 * configure: Regenerate.
1548 * config/i386/sol2-9.h: Remove.
1550 * doc/install.texi (Specific, i?86-*-solaris2.9): Remove.
1551 (Specific, *-*-solaris2*): Mention Solaris 9 support removal.
1552 Remove Solaris 9 references.
1554 2014-04-22 Vidya Praveen <vidyapraveen@arm.com>
1556 * aarch64.md (float<GPI:mode><GPF:mode>2): Remove.
1557 (floatuns<GPI:mode><GPF:mode>2): Remove.
1558 (<optab><fcvt_target><GPF:mode>2): New pattern for equal width float
1559 and floatuns conversions.
1560 (<optab><fcvt_iesize><GPF:mode>2): New pattern for inequal width float
1561 and floatuns conversions.
1562 * iterators.md (fcvt_target, FCVT_TARGET): Support SF and DF modes.
1563 (w1,w2): New mode attributes for inequal width conversions.
1565 2014-04-22 Renlin Li <Renlin.Li@arm.com>
1567 * config/aarch64/aarch64.c (aarch64_print_operand_address): Adjust
1568 the output asm format.
1570 2014-04-22 James Greenhalgh <james.greenhalgh@arm.com>
1572 * config/aarch64/aarch64-simd.md
1573 (aarch64_cm<optab>di): Always split.
1574 (*aarch64_cm<optab>di): New.
1575 (aarch64_cmtstdi): Always split.
1576 (*aarch64_cmtstdi): New.
1578 2014-04-22 Jakub Jelinek <jakub@redhat.com>
1580 PR tree-optimization/60823
1581 * omp-low.c (ipa_simd_modify_function_body): Go through
1582 all SSA_NAMEs and for those refering to vector arguments
1583 which are going to be replaced adjust SSA_NAME_VAR and,
1584 if it is a default definition, change it into a non-default
1585 definition assigned at the beginning of function from new_decl.
1586 (ipa_simd_modify_stmt_ops): Rewritten.
1587 * tree-dfa.c (set_ssa_default_def): When removing default def,
1588 check for NULL loc instead of NULL *loc.
1590 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1592 * config/arm/arm.c (arm_hard_regno_mode_ok): Loosen
1593 restrictions on core registers for DImode values in Thumb2.
1595 2014-04-22 Ian Bolton <ian.bolton@arm.com>
1597 * config/arm/arm.md (*anddi_notdi_zesidi): New pattern.
1598 * config/arm/thumb2.md (*iordi_notdi_zesidi): New pattern.
1600 2014-04-22 Ian Bolton <ian.bolton@arm.com>
1602 * config/arm/thumb2.md (*iordi_notdi_di): New pattern.
1603 (*iordi_notzesidi_di): Likewise.
1604 (*iordi_notsesidi_di): Likewise.
1606 2014-04-22 Ian Bolton <ian.bolton@arm.com>
1608 * config/arm/arm-protos.h (tune_params): New struct members.
1609 * config/arm/arm.c: Initialise tune_params per processor.
1610 (thumb2_reorg): Suppress conversion from t32 to t16 when optimizing
1611 for speed, based on new tune_params.
1613 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1615 * config/aarch64/aarch64-builtins.c (BUILTIN_VDQF_DF): Macro
1617 * config/aarch64/aarch64-simd-builtins.def (frintn): Use added
1619 * config/aarch64/aarch64-simd.md (<frint_pattern>): Comment
1621 * config/aarch64/aarch64.md (<frint_pattern>): Likewise.
1622 * config/aarch64/arm_neon.h (vrnd_f64): Added.
1623 (vrnda_f64): Likewise.
1624 (vrndi_f64): Likewise.
1625 (vrndm_f64): Likewise.
1626 (vrndn_f64): Likewise.
1627 (vrndp_f64): Likewise.
1628 (vrndx_f64): Likewise.
1630 2014-04-22 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1632 * config/arm/arm.c (arm_print_operand, thumb_exit): Make sure
1633 GET_MODE_SIZE argument is enum machine_mode.
1635 2014-04-22 Jakub Jelinek <jakub@redhat.com>
1638 * config/sparc/sparc.c (sparc_init_modes): Pass enum machine_mode
1639 value instead of int to GET_MODE_CLASS and GET_MODE_SIZE macros.
1641 2014-04-22 Lin Zuojian <manjian2006@gmail.com>
1644 * asan.c (asan_emit_stack_protection): Force the base to align to
1645 appropriate bits if STRICT_ALIGNMENT. Set shadow_mem align to
1646 appropriate bits if STRICT_ALIGNMENT.
1647 * cfgexpand.c (expand_stack_vars): Set base_align appropriately
1649 (expand_used_vars): Leave a space in the stack frame for alignment
1650 if STRICT_ALIGNMENT.
1652 2014-04-21 David Malcolm <dmalcolm@redhat.com>
1654 * gimple.h (gimple_assign_single_p): Accept a const_gimple rather
1656 (gimple_store_p): Likewise.
1657 (gimple_assign_load_p): Likewise.
1658 (gimple_assign_cast_p): Likewise.
1659 (gimple_clobber_p): Likewise.
1661 * doc/gimple.texi (gimple_assign_cast_p): Accept a const_gimple
1662 rather than a gimple.
1663 (gimple_assign_cast_p): Likewise.
1665 2014-04-21 Michael Meissner <meissner@linux.vnet.ibm.com>
1668 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64 case):
1669 If mode is DDmode and TARGET_E500_DOUBLE allow move.
1671 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print some
1672 more debug information for E500 if -mdebug=reg.
1674 2014-04-21 Uros Bizjak <ubizjak@gmail.com>
1677 * config/i386/i386.c (ix86_expand_builtin)
1678 <case IX86_BUILTIN_RDRAND{16,32,64}_STEP>: Use temporary
1679 register for target RTX.
1680 <case IX86_BUILTIN_RDSEED{16,32,64}_STEP>: Ditto.
1682 2014-04-18 Cong Hou <congh@google.com>
1684 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): Enhance
1685 the widen-mult pattern by handling two operands with different sizes,
1686 and operands whose size is smaller than half of the result type.
1688 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
1690 * ipa-inline.h (INLINE_HINT_known_hot): New hint.
1691 * ipa-inline-analysis.c (dump_inline_hints): Dump it.
1692 (do_estimate_edge_time): Compute it.
1693 * ipa-inline.c (want_inline_small_function_p): Bypass
1694 INLINE_INSNS_AUTO/SINGLE limits for calls that are known to be hot.
1696 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
1698 * ipa-inline.c (spec_rem): New static variable.
1699 (dump_overall_stats): New function.
1700 (dump_inline_stats): New function.
1702 2014-04-18 Richard Henderson <rth@redhat.com>
1704 * config/aarch64/aarch64.c (aarch64_register_move_cost): Pass a mode
1705 to GET_MODE_SIZE, not a reg_class_t.
1707 2014-04-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1709 * config/rs6000/vsx.md (vsx_xxmrghw_<mode>): Adjust for little-endian.
1710 (vsx_xxmrglw_<mode>): Likewise.
1712 2014-04-17 Michael Meissner <meissner@linux.vnet.ibm.com>
1715 * config/rs6000/rs6000.c (rs6000_setup_reg_addr_masks): Make sure
1716 GET_MODE_SIZE gets passed an enum machine_mode type and not integer.
1717 (rs6000_init_hard_regno_mode_ok): Likewise.
1719 2014-04-17 Jan Hubicka <hubicka@ucw.cz>
1721 * ipa-inline.c (inline_small_functions): Account only non-cold
1723 * doc/invoke.texi (inline-unit-growth): Update documentation.
1725 2014-04-17 Pat Haugen <pthaugen@us.ibm.com>
1727 * config/rs6000/rs6000.md (addti3, subti3): New.
1729 2014-04-17 H.J. Lu <hongjiu.lu@intel.com>
1732 * config/i386/i386.c (ix86_expand_clear): Remove outdated
1733 comment. Check optimize_insn_for_size_p instead of
1734 optimize_insn_for_speed_p.
1736 2014-04-17 Martin Jambor <mjambor@suse.cz>
1738 * gimple-iterator.c (gsi_start_edge): New function.
1739 * gimple-iterator.h (gsi_start_edge): Declare.
1740 * tree-sra.c (single_non_eh_succ): New function.
1741 (disqualify_ops_if_throwing_stmt): Renamed to
1742 disqualify_if_bad_bb_terminating_stmt. Allow throwing statements
1743 having one non-EH successor BB.
1744 (sra_modify_expr): If stmt ends bb, use single non-EH successor to
1745 generate loads into replacements.
1746 (sra_modify_assign): Likewise and and also use the simple path for
1748 (sra_modify_function_body): Commit statements on edges.
1750 2014-04-17 Richard Biener <rguenther@suse.de>
1753 * tree-ssa-propagate.c (valid_gimple_rhs_p): Allow vector
1754 comparison results and add clarifying comment.
1756 2014-04-17 Jakub Jelinek <jakub@redhat.com>
1758 * genmodes.c (struct mode_data): Add need_bytesize_adj field.
1759 (blank_mode): Initialize it.
1760 (emit_mode_size_inline, emit_mode_nunits_inline,
1761 emit_mode_inner_inline): New functions.
1762 (emit_insn_modes_h): Call them and surround their output with
1763 #if GCC_VERSION >= 4001 ... #endif.
1764 * machmode.h (GET_MODE_SIZE, GET_MODE_NUNITS, GET_MODE_INNER):
1765 For GCC_VERSION >= 4001 use mode_*_inline routines instead of
1766 mode_* arrays if the argument is __builtin_constant_p.
1767 * lower-subreg.c (dump_choices): Make sure GET_MODE_SIZE argument
1768 is enum machine_mode.
1770 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1772 * passes.c (opt_pass::execute): Adjust.
1773 (pass_manager::execute_pass_mode_switching): Likewise.
1774 (early_local_passes::execute): Likewise.
1775 (execute_one_pass): Pass cfun to the pass's execute method.
1776 * tree-pass.h (opt_pass::execute): Add function * argument.
1777 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
1778 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
1779 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
1780 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
1781 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
1782 config/sparc/sparc.c, cprop.c, dce.c, df-core.c, dse.c, dwarf2cfi.c,
1783 except.c, final.c, function.c, fwprop.c, gcse.c, gimple-low.c,
1784 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
1785 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
1786 ipa-inline-analysis.c, ipa-inline.c, ipa-profile.c, ipa-pure-const.c,
1787 ipa-reference.c, ipa-split.c, ipa.c, ira.c, jump.c, loop-init.c,
1788 lower-subreg.c, mode-switching.c, omp-low.c, postreload-gcse.c,
1789 postreload.c, predict.c, recog.c, ree.c, reg-stack.c, regcprop.c,
1790 reginfo.c, regrename.c, reorg.c, sched-rgn.c, stack-ptr-mod.c,
1791 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
1792 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
1793 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
1794 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-ssa-ccp.c,
1795 tree-ssa-copy.c, tree-ssa-copyrename.c, tree-ssa-dce.c,
1796 tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
1797 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
1798 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
1799 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
1800 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
1801 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
1802 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
1803 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
1804 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
1805 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
1808 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1810 * passes.c (opt_pass::gate): Take function * argument.
1811 (gate_all_early_local_passes): Merge into
1812 (early_local_passes::gate): this.
1813 (gate_all_early_optimizations): Merge into
1814 (all_early_optimizations::gate): this.
1815 (gate_all_optimizations): Mege into
1816 (all_optimizations::gate): this.
1817 (gate_all_optimizations_g): Merge into
1818 (all_optimizations_g::gate): this.
1819 (gate_rest_of_compilation): Mege into
1820 (rest_of_compilation::gate): this.
1821 (gate_postreload): Merge into
1822 (postreload::gate): this.
1823 (dump_one_pass): Pass cfun to the pass's gate method.
1824 (execute_ipa_summary_passes): Likewise.
1825 (execute_one_pass): Likewise.
1826 (ipa_write_summaries_2): Likewise.
1827 (ipa_write_optimization_summaries_1): Likewise.
1828 (ipa_read_summaries_1): Likewise.
1829 (ipa_read_optimization_summaries_1): Likewise.
1830 (execute_ipa_stmt_fixups): Likewise.
1831 * tree-pass.h (opt_pass::gate): Add function * argument.
1832 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c,
1833 combine-stack-adj.c, combine.c, compare-elim.c,
1834 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
1835 config/rl78/rl78.c, config/sh/sh_optimize_sett_clrt.cc,
1836 config/sh/sh_treg_combine.cc, config/sparc/sparc.c, cprop.c, cse.c,
1837 dce.c, df-core.c, dse.c, dwarf2cfi.c, except.c, fwprop.c, gcse.c,
1838 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
1839 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
1840 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
1841 loop-init.c, lower-subreg.c, mode-switching.c, modulo-sched.c,
1842 omp-low.c, postreload-gcse.c, postreload.c, predict.c, recog.c, ree.c,
1843 reg-stack.c, regcprop.c, regrename.c, reorg.c, sched-rgn.c,
1844 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
1845 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
1846 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c,
1847 tree-nrv.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
1848 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
1849 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
1850 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
1851 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
1852 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
1853 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
1854 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
1855 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
1856 tree-ssa.c, tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
1857 tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c, tsan.c, ubsan.c,
1858 var-tracking.c, vtable-verify.c, web.c: Adjust.
1860 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1862 * configure.ac: Check for -Woverloaded-virtual and enable it if found.
1863 * configure: Regenerate.
1865 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1867 * passes.c (dump_one_pass): don't check pass->has_gate.
1868 (execute_ipa_summary_passes): Likewise.
1869 (execute_one_pass): Likewise.
1870 (ipa_write_summaries_2): Likewise.
1871 (ipa_write_optimization_summaries_1): Likewise.
1872 (ipa_read_optimization_summaries_1): Likewise.
1873 (execute_ipa_stmt_fixups): Likewise.
1874 * tree-pass.h (pass_data::has_gate): Remove.
1875 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
1876 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
1877 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
1878 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
1879 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
1880 config/sh/sh_optimize_sett_clrt.cc, config/sh/sh_treg_combine.cc,
1881 config/sparc/sparc.c, cprop.c, cse.c, dce.c, df-core.c, dse.c,
1882 dwarf2cfi.c, except.c, final.c, function.c, fwprop.c, gcse.c,
1883 gimple-low.c, gimple-ssa-isolate-paths.c,
1884 gimple-ssa-strength-reduction.c, graphite.c, ifcvt.c, init-regs.c,
1885 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
1886 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
1887 ira.c, jump.c, loop-init.c, lower-subreg.c, mode-switching.c,
1888 modulo-sched.c, omp-low.c, postreload-gcse.c, postreload.c, predict.c,
1889 recog.c, ree.c, reg-stack.c, regcprop.c, reginfo.c, regrename.c,
1890 reorg.c, sched-rgn.c, stack-ptr-mod.c, store-motion.c, tracer.c,
1891 trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
1892 tree-complex.c, tree-eh.c, tree-emutls.c, tree-if-conv.c,
1893 tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
1894 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
1895 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
1896 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
1897 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
1898 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
1899 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
1900 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
1901 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
1902 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
1903 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
1904 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
1905 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
1908 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1910 * pass_manager.h (pass_manager::register_dump_files_1): Remove
1912 * passes.c (pass_manager::register_dump_files_1): Merge into
1913 (pass_manager::register_dump_files): this, and remove its handling of
1914 properties since the pass always has the properties anyway.
1915 (pass_manager::pass_manager): Adjust.
1917 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1919 * pass_manager.h (pass_manager::register_dump_files_1): Adjust.
1920 * passes.c (pass_manager::register_dump_files_1): Remove dead code
1921 dealing with properties.
1922 (pass_manager::register_dump_files): Adjust.
1924 2014-03-20 Mark Wielaard <mjw@redhat.com>
1926 * dwarf2out.c (add_bound_info): If HOST_WIDE_INT is big enough,
1927 then represent the bound as normal constant value.
1929 2014-04-17 Jakub Jelinek <jakub@redhat.com>
1932 Forward port from 4.8 branch
1933 2013-07-19 Kirill Yukhin <kirill.yukhin@intel.com>
1935 * config/i386/bmiintrin.h (_blsi_u32): New.
1939 (_blsmsk_u32): Ditto.
1940 (_blsmsk_u64): Ditto.
1941 (_tzcnt_u32): Ditto.
1942 (_tzcnt_u64): Ditto.
1944 2014-04-17 Kito Cheng <kito@0xlab.org>
1946 * gcc.c (used_arg): Prevent out of bound access for multilib_options.
1948 2014-04-17 Richard Biener <rguenther@suse.de>
1951 * tree-ssa-propagate.c (valid_gimple_rhs_p): Only allow effective
1952 boolean results for comparisons.
1954 2014-04-17 Richard Biener <rguenther@suse.de>
1956 PR tree-optimization/60836
1957 * tree-vect-loop.c (vect_create_epilog_for_reduction): Force
1958 initial PHI args to be gimple values.
1960 2014-04-17 Richard Biener <rguenther@suse.de>
1962 PR tree-optimization/60841
1963 * tree-vect-data-refs.c (vect_analyze_data_refs): Count stmts.
1964 * tree-vect-loop.c (vect_analyze_loop_2): Pass down number
1965 of stmts to SLP build.
1966 * tree-vect-slp.c (vect_slp_analyze_bb_1): Likewise.
1967 (vect_analyze_slp): Likewise.
1968 (vect_analyze_slp_instance): Likewise.
1969 (vect_build_slp_tree): Limit overall SLP tree growth.
1970 * tree-vectorizer.h (vect_analyze_data_refs,
1971 vect_analyze_slp): Adjust prototypes.
1973 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
1975 * config/i386/i386.c (x86_add_stmt_cost): Fix vector cost model for
1978 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
1980 * config/i386/x86-tune.def (TARGET_SLOW_PSHUFB): New tune definition.
1981 * config/i386/i386.h (TARGET_SLOW_PSHUFB): New tune flag.
1982 * config/i386/i386.c (expand_vec_perm_even_odd_1): Avoid byte shuffles
1983 for TARGET_SLOW_PSHUFB
1985 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
1987 * config/i386/i386.c (slm_cost): Adjust vec_to_scalar_cost.
1988 * config/i386/i386.c (intel_cost): Ditto.
1990 2014-04-17 Joey Ye <joey.ye@arm.com>
1992 * opts.c (OPT_fif_conversion, OPT_fif_conversion2): Disable for Og.
1994 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
1996 * opts.c (common_handle_option): Disable -fipa-reference coorectly
1999 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
2001 * ipa-devirt.c (odr_type_d): Add field all_derivations_known.
2002 (type_all_derivations_known_p): New predicate.
2003 (type_all_ctors_visible_p): New predicate.
2004 (type_possibly_instantiated_p): New predicate.
2005 (get_odr_type): Compute all_derivations_known.
2006 (dump_odr_type): Dump the flag.
2007 (maybe_record_type): Cleanup.
2008 (record_target_from_binfo): Add bases_to_consider array;
2009 record bases for types w/o instances and skip CXX destructor.
2010 (possible_polymorphic_call_targets_1): Add bases_to_consider
2011 and consider_construction parameters; check if type may have instance.
2012 (get_polymorphic_call_info): Set maybe_in_construction to true
2013 when we know nothing.
2014 (record_targets_from_bases): Skip CXX destructors; they are
2015 never called for types in construction.
2016 (possible_polymorphic_call_targets): Do not record target when
2017 type may not have instance.
2019 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
2022 * ipa.c (symtab_remove_unreachable_nodes): Mark targets of
2023 external aliases alive, too.
2025 2014-04-16 Andrew Pinski <apinski@cavium.com>
2027 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Change aarch64 ilp32
2030 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
2032 * final.c (compute_alignments): Do not apply loop alignment to a block
2033 falling through to the exit.
2035 2014-04-16 Catherine Moore <clm@codesourcery.com>
2037 * mips.md (*mov<mode>_internal, *movhi_internal, *movqi_internal):
2038 Adjust constraints for microMIPS store patterns.
2040 2014-04-16 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
2042 * config/avr/avr-mcus.def: Correct typo for atxmega256a3bu macro.
2044 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
2046 * tree-ssa-operands.c (create_vop_var): Set DECL_IGNORED_P.
2047 (append_use): Run at -O0.
2048 (append_vdef): Likewise.
2049 * tree-ssa-ter.c (ter_is_replaceable_p): Do not special-case -O0.
2050 * tree-ssa-uninit.c (warn_uninitialized_vars): Remove obsolete comment.
2052 2014-04-16 Jakub Jelinek <jakub@redhat.com>
2054 PR tree-optimization/60844
2055 * tree-ssa-reassoc.c (reassoc_remove_stmt): New function.
2056 (propagate_op_to_single_use, remove_visited_stmt_chain,
2057 linearize_expr, repropagate_negates, reassociate_bb): Use it
2058 instead of gsi_remove.
2060 2014-04-16 Martin Jambor <mjambor@suse.cz>
2062 * cgraphclones.c (cgraph_create_virtual_clone): Duplicate
2063 ipa_transforms_to_apply.
2064 (cgraph_function_versioning): Assert that old_node has empty
2065 ipa_transforms_to_apply.
2066 * trans-mem.c (ipa_tm_create_version): Likewise.
2067 * tree-inline.c (tree_function_versioning): Do not duplicate
2068 ipa_transforms_to_apply.
2070 2014-04-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2073 * configure.ac (set_have_as_tls): Merge i[34567]86-*-* and
2075 Pass necessary as flags on 64-bit Solaris/x86.
2076 Use lowercase relocs for x86_64-*-*.
2077 * configure: Regenerate.
2079 2014-04-15 Jan Hubicka <jh@suse.cz>
2081 * ipa-devirt.c (referenced_from_vtable_p): New predicate.
2082 (maybe_record_node, likely_target_p): Use it.
2084 2014-04-15 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2087 Revert following patch
2089 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
2092 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
2093 software floating point or no floating point registers, do not
2094 allow any type in the FPRs. Eliminate a test for SPE SIMD types
2095 in GPRs that occurs after we tested for GPRs that would never be
2098 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
2099 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
2100 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
2101 specifically allow DDmode, since that does not use the SPE SIMD
2104 2014-03-21 Mark Wielaard <mjw@redhat.com>
2106 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_const_value
2107 as unsigned or int depending on type and value used.
2109 2014-04-15 Richard Biener <rguenther@suse.de>
2111 PR rtl-optimization/56965
2112 * alias.c (ncr_compar, nonoverlapping_component_refs_p): Move ...
2113 * tree-ssa-alias.c (ncr_compar, nonoverlapping_component_refs_p):
2115 * alias.c (true_dependence_1): Do not call
2116 nonoverlapping_component_refs_p.
2117 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Call
2118 nonoverlapping_component_refs_p.
2119 (indirect_refs_may_alias_p): Likewise.
2121 2014-04-15 Teresa Johnson <tejohnson@google.com>
2123 * cfg.c (dump_bb_info): Fix flags check.
2124 * tree-cfg.c (remove_bb): Only dump TDF_BLOCKS when removing.
2126 2014-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2128 PR rtl-optimization/60663
2129 * config/arm/arm.c (arm_new_rtx_costs): Improve ASM_OPERANDS case,
2132 2014-04-15 Richard Biener <rguenther@suse.de>
2134 * lto-streamer.h (LTO_major_version): Bump to 4.
2136 2014-04-15 Richard Biener <rguenther@suse.de>
2138 * common.opt (lto_partition_model): New enum.
2139 (flto-partition=): Merge separate options with a single with argument,
2140 add -flto-partition=one support.
2141 * flag-types.h (enum lto_partition_model): Declare.
2142 * opts.c (finish_options): Remove duplicate -flto-partition=
2144 * lto-wrapper.c (run_gcc): Adjust.
2146 2014-04-15 Richard Biener <rguenther@suse.de>
2148 * alias.c (ncr_compar): New function.
2149 (nonoverlapping_component_refs_p): Re-implement in O (n log n).
2151 2014-04-15 Richard Biener <rguenther@suse.de>
2153 * alias.c (record_component_aliases): Do not walk BINFOs.
2155 2014-04-15 Richard Biener <rguenther@suse.de>
2157 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
2158 Add struct function argument and adjust.
2159 (find_func_aliases_for_call): Likewise.
2160 (find_func_aliases): Likewise.
2161 (find_func_clobbers): Likewise.
2162 (intra_create_variable_infos): Likewise.
2163 (compute_points_to_sets): Likewise.
2164 (ipa_pta_execute): Adjust. Do not push/pop cfun.
2166 2014-04-15 Richard Biener <rguenther@suse.de>
2168 * tree.c (iterative_hash_expr): Use enum tree_code_class
2169 to store TREE_CODE_CLASS.
2170 (tree_block): Likewise.
2171 (tree_set_block): Likewise.
2172 * tree.h (fold_build_pointer_plus_loc): Use
2173 convert_to_ptrofftype_loc.
2175 2014-04-15 Jakub Jelinek <jakub@redhat.com>
2178 * Makefile.in (PLUGIN_HEADERS): Add various headers that have been
2181 2014-04-15 Eric Botcazou <ebotcazou@adacore.com>
2183 * cfgloop.h (struct loop): Move force_vectorize down.
2184 * gimplify.c (gimple_boolify) <ANNOTATE_EXPR>: Handle new kinds.
2185 (gimplify_expr) <ANNOTATE_EXPR>: Minor tweak.
2186 * lto-streamer-in.c (input_cfg): Read dont_vectorize field.
2187 * lto-streamer-out.c (output_cfg): Write dont_vectorize field.
2188 * tree-cfg.c (replace_loop_annotate): Revamp and handle new kinds.
2189 * tree-core.h (enum annot_expr_kind): Add new kind values.
2190 * tree-inline.c (copy_loops): Copy dont_vectorize field and reorder.
2191 * tree-pretty-print.c (dump_generic_node) <ANNOTATE_EXPR>: Handle new
2193 * tree.def (ANNOTATE_EXPR): Tweak comment.
2195 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
2197 * ipa-devirt.c (maybe_record_node): Ignore all non-methods (including
2200 2014-04-14 Paolo Carlini <paolo.carlini@oracle.com>
2202 * tree.h (TYPE_IDENTIFIER): Declare.
2203 * tree.c (subrange_type_for_debug_p): Use it.
2204 * godump.c (go_format_type): Likewise.
2205 * dwarf2out.c (is_cxx_auto, modified_type_die,
2206 gen_type_die_with_usage, gen_type_die_with_usage): Likewise.
2207 * dbxout.c (dbxout_type, dbxout_symbol): Likewise.
2209 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
2212 * varpool.c (varpool_remove_node): Do not alter decls when streaming.
2214 2014-04-14 Uros Bizjak <ubizjak@gmail.com>
2216 * config/i386/i386.c (examine_argument): Return bool. Return true if
2217 parameter should be passed in memory.
2218 <case X86_64_COMPLEX_X87_CLASS>: Adjust.
2219 (construct_container): Update calls to examine_argument.
2220 (function_arg_advance_64): Ditto.
2221 (return_in_memory_32): Merge with ix86_return_in_memory.
2222 (return_in_memory_64): Ditto.
2223 (return_in_memory_ms_64): Ditto.
2225 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
2227 * ipa-utils.c (ipa_merge_profiles): Merge profile_id.
2228 * coverage.c (coverage_compute_profile_id): Handle externally visible
2231 2014-04-14 Martin Jambor <mjambor@suse.cz>
2233 * tree-sra.c (ipa_sra_preliminary_function_checks): Skip
2234 DECL_DISREGARD_INLINE_LIMITS functions.
2236 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
2239 * config/i386/i386.md (*fixuns_trunc<mode>_1): Revert the last change.
2241 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
2244 * config/i386/i386.md (*fixuns_trunc<mode>_1): Check
2245 optimize_insn_for_speed_p instead of
2246 optimize_function_for_speed_p.
2248 2014-04-14 Yufeng Zhang <yufeng.zhang@arm.com>
2250 * doc/invoke.texi (free): Document AArch64.
2252 2014-04-14 Richard Biener <rguenther@suse.de>
2254 PR tree-optimization/60042
2255 * tree-ssa-pre.c (inhibit_phi_insertion): Remove.
2256 (insert_into_preds_of_block): Do not prevent PHI insertion
2257 for REFERENCE exprs here ...
2258 (eliminate_dom_walker::before_dom_children): ... but prevent
2259 their use here under similar conditions when applied to the
2260 IL after PRE optimizations.
2262 2014-04-14 Richard Biener <rguenther@suse.de>
2264 * passes.def: Move early points-to after early SRA.
2266 2014-04-14 Richard Biener <rguenther@suse.de>
2268 * tree-ssa-forwprop.c (simplify_gimple_switch): Enhance
2269 check for which sign-changes we allow when forwarding
2270 a converted value into a switch.
2272 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
2274 * stor-layout.c (place_field): Finalize non-constant offset for the
2277 2014-04-14 Richard Biener <rguenther@suse.de>
2279 * tree-switch-conversion.c (lshift_cheap_p): Get speed_p
2281 (expand_switch_using_bit_tests_p): Likewise.
2282 (process_switch): Compute and pass on speed_p based on the
2284 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Use
2285 optimize_bb_for_speed_p.
2287 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
2289 * cfgloop.h (struct loop): Rename force_vect into force_vectorize.
2290 * function.h (struct function): Rename has_force_vect_loops into
2291 has_force_vectorize_loops.
2292 * lto-streamer-in.c (input_cfg): Adjust for renaming.
2293 (input_struct_function_base): Likewise.
2294 * lto-streamer-out.c (output_cfg): Likewise.
2295 (output_struct_function_base): Likewise.
2296 * omp-low.c (expand_omp_simd): Likewise.
2297 * tree-cfg.c (move_sese_region_to_fn): Likewise.
2298 * tree-if-conv.c (ifcvt_can_use_mask_load_store): Likewise.
2299 (version_loop_for_if_conversion): Likewise.
2300 (tree_if_conversion): Likewise.
2301 (main_tree_if_conversion): Likewise.
2302 (gate_tree_if_conversion): Likewise.
2303 * tree-inline.c (copy_loops): Likewise.
2304 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Likewise.
2305 * tree-ssa-loop.c (tree_loop_vectorize): Likewise.
2306 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Likewise.
2307 * tree-vect-loop.c (vect_estimate_min_profitable_iters): Likewise.
2308 * tree-vectorizer.c (vectorize_loops): Likewise.
2309 * tree-vectorizer.h (unlimited_cost_model): Likewise.
2311 2014-04-14 Richard Biener <rguenther@suse.de>
2314 * lto-streamer-out.c (wrap_refs): New function.
2315 (lto_output): Wrap symbol references in global initializes in
2316 type-preserving MEM_REFs.
2318 2014-04-14 Christian Bruel <christian.bruel@st.com>
2320 * config/sh/sh-mem.cc (sh_expand_strlen): Unroll last word.
2322 2014-04-14 Christian Bruel <christian.bruel@st.com>
2324 * config/sh/sh.md (setmemqi): New expand pattern.
2325 * config/sh/sh.h (CLEAR_RATIO): Define.
2326 * config/sh/sh-mem.cc (sh_expand_setmem): Define.
2327 * config/sh/sh-protos.h (sh_expand_setmem): Declare.
2329 2014-04-14 Richard Biener <rguenther@suse.de>
2332 * fold-const.c (negate_expr_p): Don't negate directional rounding
2334 (fold_negate_expr): Likewise.
2336 2014-04-14 Richard Biener <rguenther@suse.de>
2338 PR tree-optimization/59817
2339 PR tree-optimization/60453
2340 * graphite-scop-detection.c (graphite_can_represent_scev): Complete
2341 recursion to catch all CHRECs in the scalar evolution and restrict
2342 the predicate for the remains appropriately.
2344 2014-04-12 Catherine Moore <clm@codesourcery.com>
2346 * config/mips/constraints.md: Add new register constraint "kb".
2347 * config/mips/mips.md (*mov<mode>_internal): Use constraint "kb".
2348 (*movhi_internal): Likewise.
2349 (*movqi_internal): Likewise.
2350 * config/mips/mips.h (M16_STORE_REGS): New register class.
2351 (REG_CLASS_NAMES): Add M16_STORE_REGS.
2352 (REG_CLASS_CONTENTS): Likewise.
2353 * config/mips/mips.c (mips_regno_to_class): Add M16_STORE_REGS.
2355 2014-04-11 Tobias Burnus <burnus@net-b.de>
2358 * doc/invoke.texi (-Wformat-signedness): Document it.
2359 (Wformat=2): Mention that this enables -Wformat-signedness.
2361 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
2363 * common/config/epiphany/epiphany-common.c
2364 (epiphany_option_optimization_table): Enable section anchors by
2365 default at -O1 or higher.
2366 * config/epiphany/epiphany.c (TARGET_MAX_ANCHOR_OFFSET): Define.
2367 (TARGET_MIN_ANCHOR_OFFSET): Likewise.
2368 (epiphany_rtx_costs) <SET>: For binary operators, the set as such
2369 carries no extra cost.
2370 (epiphany_legitimate_address_p): For BLKmode, apply SImode check.
2371 * config/epiphany/epiphany.h (ASM_OUTPUT_DEF): Define.
2372 * config/epiphany/predicates.md (memclob_operand): New predicate.
2373 * config/epiphany/epiphany.md (stack_adjust_add, stack_adjust_str):
2374 Use memclob_operand predicate and X constraint for operand 3.
2376 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
2378 * config/epiphany/epiphany.c (epiphany_rtx_cost): Compare
2379 with CC_N_NE / CC_C_LTU / CC_C_GTU carries no extra cost for
2382 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
2384 PR rtl-optimization/60651
2385 * mode-switching.c (optimize_mode_switching): Make sure to emit
2386 sets of a lower numbered entity before sets of a higher numbered
2387 entity to a mode of the same or lower priority.
2388 When creating a seginfo for a basic block that starts with a code
2389 label, move the insertion point past the code label.
2390 (new_seginfo): Document and enforce requirement that
2391 NOTE_INSN_BASIC_BLOCK only appears for empty blocks.
2392 * doc/tm.texi.in: Document ordering constraint for emitted mode sets.
2393 * doc/tm.texi: Regenerate.
2395 2014-01-11 Joern Rennecke <joern.rennecke@embecosm.com>
2398 * config/arc/arc.c (arc_save_restore): Fix assert typo.
2400 2013-04-11 Jakub Jelinek <jakub@redhat.com>
2402 * BASE-VER: Set to 4.10.0.
2404 2014-04-11 Tobias Burnus <burnus@net-b.de>
2407 * doc/bugreport.texi (Bugs): Remove nodes pointing to the nirvana.
2408 * doc/gcc.texi (Service): Update description in the @menu
2409 * doc/invoke.texi (Option Summary): Remove misplaced and
2412 2014-04-11 Steve Ellcey <sellcey@mips.com>
2413 Jakub Jelinek <jakub@redhat.com>
2416 * expr.c (convert_move): Use emit_store_flag_force instead of
2417 emit_store_flag. Pass lowpart_mode instead of VOIDmode as 5th
2420 2014-04-11 Richard Biener <rguenther@suse.de>
2423 * varasm.c (assemble_alias): Avoid endless error reporting
2424 recursion by setting TREE_ASM_WRITTEN.
2426 2014-04-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2428 * config/s390/s390.md: Add a splitter for NOT rtx.
2430 2014-04-11 Jakub Jelinek <jakub@redhat.com>
2432 PR rtl-optimization/60663
2433 * cse.c (cse_insn): Set src_volatile on ASM_OPERANDS in PARALLEL.
2435 2014-04-10 Jan Hubicka <hubicka@ucw.cz>
2436 Jakub Jelinek <jakub@redhat.com>
2439 * ipa.c (function_and_variable_visibility): Copy forced_by_abi
2440 flag from decl_node to node.
2442 2014-04-10 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2445 * config/arm/arm.c (TARGET_CONST_NOT_OK_FOR_DEBUG_P): Define
2446 (arm_const_not_ok_for_debug_p): Reject MINUS with SYM_REF's
2447 ameliorating the cases where it can be.
2449 2014-04-09 David Edelsohn <dje.gcc@gmail.com>
2452 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
2454 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
2455 (loadsync_<mode>): Change mode.
2456 (load_quadpti, store_quadpti): New.
2457 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
2458 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
2459 * config/rs6000/predicates.md (quad_memory_operand): !TARGET_SYNC_TI.
2461 2014-04-09 Cong Hou <congh@google.com>
2464 * doc/sourcebuild.texi (vect_widen_mult_si_to_di_pattern): Add
2467 2014-04-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2469 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Use vnand
2470 instead of vnor to exploit possible fusion opportunity in the
2472 (altivec_expand_vec_perm_const_le): Likewise.
2474 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
2476 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
2477 (loadsync_<mode>): Change mode.
2478 (load_quadpti, store_quadpti): New.
2479 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
2480 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
2482 2014-04-08 Richard Sandiford <rdsandiford@googlemail.com>
2485 * config/rs6000/vsx.md (vsx_xscvdpspn_scalar): Change input to DImode.
2486 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Update accordingly.
2487 Use gen_rtx_REG rather than simplify_gen_subreg for op0_di.
2489 2014-04-08 Richard Biener <rguenther@suse.de>
2492 * tree-pretty-print.c (pp_double_int): For HWI32 hosts with
2493 a 64bit widest int print double-int similar to on HWI64 hosts.
2495 2014-04-08 Richard Biener <rguenther@suse.de>
2497 PR tree-optimization/60785
2498 * graphite-sese-to-poly.c (rewrite_phi_out_of_ssa): Treat
2499 default defs properly.
2501 2014-04-08 Nathan Sidwell <nathan@codesourcery.com>
2503 * doc/invoke (Wnon-virtual-dtor): Update to match implementation.
2504 (Weffc++): Likewise.
2506 2014-04-07 Jan Hubicka <hubcika@ucw.cz>
2508 * ipa-devirt.c (maybe_record_node): When node is not recorded,
2509 set completep to false rather than true.
2511 2014-04-07 Douglas B Rupp <rupp@adacore.com>
2514 * config/arm/arm.h (ASM_PREFERRED_EH_DATA_FORMAT): Expose from
2515 ARM_TARGET2_DWARF_FORMAT.
2517 2014-04-07 Charles Baylis <charles.baylis@linaro.org>
2520 * config/arm/arm.h (ASM_OUTPUT_CASE_END): Remove.
2521 (LABEL_ALIGN_AFTER_BARRIER): Align barriers which occur after
2524 2014-04-07 Richard Biener <rguenther@suse.de>
2526 PR tree-optimization/60766
2527 * tree-ssa-loop-ivopts.c (cand_value_at): Compute in an unsigned type.
2528 (may_eliminate_iv): Convert cand_value_at result to desired type.
2530 2014-04-07 Jason Merrill <jason@redhat.com>
2533 * common.opt (-fno-gnu-unique): Add.
2534 * config/elfos.h (USE_GNU_UNIQUE_OBJECT): Check it.
2536 2014-04-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2538 * haifa-sched.c: Fix outdated function reference and minor
2539 grammar errors in introductory comment.
2541 2014-04-07 Richard Biener <rguenther@suse.de>
2544 * tree-ssa-operands.c (maybe_add_call_vops): Also add VDEFs
2546 * tree-cfgcleanup.c (fixup_noreturn_call): Do not remove VDEFs.
2548 2014-04-06 John David Anglin <danglin@gcc.gnu.org>
2551 * config/pa/pa.c (pa_output_function_epilogue): Skip address and code
2552 size accounting for thunks.
2553 (pa_asm_output_mi_thunk): Use final_start_function() and
2554 final_end_function() to output function start and end directives.
2556 2014-04-05 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
2558 * config/avr/avr-arch.h (avr_mcu_t): Add dev_attribute field to have
2559 device specific ISA/ feature information. Remove short_sp and
2560 errata_skip ds. Add avr_device_specific_features enum to have device
2562 * config/avr/avr-c.c (avr_cpu_cpp_builtins): use dev_attribute to check
2563 errata_skip. Add __AVR_ISA_RMW__ builtin macro if RMW ISA available.
2564 * config/avr/avr-devices.c (avr_mcu_types): Update AVR_MCU macro for
2565 updated device specific info.
2566 * config/avr/avr-mcus.def: Merge device specific details to
2567 dev_attribute field.
2568 * config/avr/avr.c (avr_2word_insn_p): use dev_attribute field to check
2570 * config/avr/avr.h (AVR_HAVE_8BIT_SP): same for short sp info.
2571 * config/avr/driver-avr.c (avr_device_to_as): Pass -mrmw option to
2572 assembler if RMW isa supported by current device.
2573 * config/avr/genmultilib.awk: Update as device info structure changed.
2574 * doc/invoke.texi: Add info for __AVR_ISA_RMW__ builtin macro
2576 2014-04-04 Cong Hou <congh@google.com>
2578 PR tree-optimization/60656
2579 * tree-vect-stmts.c (supportable_widening_operation):
2580 Fix a bug that elements in a vector with vect_used_by_reduction
2581 property are incorrectly reordered when the operation on it is not
2582 consistant with the one in reduction operation.
2584 2014-04-04 John David Anglin <danglin@gcc.gnu.org>
2586 PR rtl-optimization/60155
2587 * gcse.c (record_set_data): New function.
2588 (single_set_gcse): New function.
2589 (gcse_emit_move_after): Use single_set_gcse instead of single_set.
2590 (hoist_code): Likewise.
2591 (get_pressure_class_and_nregs): Likewise.
2593 2014-04-04 Eric Botcazou <ebotcazou@adacore.com>
2595 * explow.c (probe_stack_range): Emit a final optimization blockage.
2597 2014-04-04 Anthony Green <green@moxielogic.com>
2599 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2): Fix
2602 2014-04-04 Jan Hubicka <hubicka@ucw.cz>
2605 * lto-cgraph.c (input_overwrite_node): Check that partitioning
2606 flags are set only during streaming.
2607 * ipa.c (process_references, walk_polymorphic_call_targets,
2608 symtab_remove_unreachable_nodes): Drop bodies of always inline
2609 after early inlining.
2610 (symtab_remove_unreachable_nodes): Remove always_inline attribute.
2612 2014-04-04 Jakub Jelinek <jakub@redhat.com>
2613 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2616 * dwarf2out.c (const_ok_for_output_1): Reject expressions
2619 2014-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2622 * config/arm/cortex-a53.md (cortex_a53_fdivs): Reduce reservation
2624 (cortex_a53_fdivd): Likewise.
2626 2014-04-04 Martin Jambor <mjambor@suse.cz>
2629 * cgraph.h (cgraph_clone_node): New parameter added to declaration.
2631 * cgraph.c (clone_of_p): Also return true if thunks match.
2632 (verify_edge_corresponds_to_fndecl): Removed extraneous call to
2633 cgraph_function_or_thunk_node and an obsolete comment.
2634 * cgraphclones.c (build_function_type_skip_args): Moved upwards in the
2636 (build_function_decl_skip_args): Likewise.
2637 (set_new_clone_decl_and_node_flags): New function.
2638 (duplicate_thunk_for_node): Likewise.
2639 (redirect_edge_duplicating_thunks): Likewise.
2640 (cgraph_clone_node): New parameter args_to_skip, pass it to
2641 redirect_edge_duplicating_thunks which is called instead of
2642 cgraph_redirect_edge_callee.
2643 (cgraph_create_virtual_clone): Pass args_to_skip to cgraph_clone_node,
2644 moved setting of a lot of flags to set_new_clone_decl_and_node_flags.
2646 2014-04-04 Jeff Law <law@redhat.com>
2649 * config/arm/predicates.md (const_int_I_operand): New predicate.
2650 (const_int_M_operand): Similarly.
2651 * config/arm/arm.md (insv_zero): Use const_int_M_operand instead of
2653 (insv_t2, extv_reg, extzv_t2): Likewise.
2654 (load_multiple_with_writeback): Similarly for const_int_I_operand.
2655 (pop_multiple_with_writeback_and_return): Likewise.
2656 (vfp_pop_multiple_with_writeback): Likewise
2658 2014-04-04 Richard Biener <rguenther@suse.de>
2661 * tree-ssanames.c (make_ssa_name_fn): Fix assert.
2662 * gimple.c (gimple_set_bb): Avoid ICEing for NULL cfun for
2664 * gimplify.h (gimple_add_tmp_var_fn): Declare.
2665 * gimplify.c (gimple_add_tmp_var_fn): New function.
2666 * gimple-expr.h (create_tmp_reg_fn): Declare.
2667 * gimple-expr.c (create_tmp_reg_fn): New function.
2668 * gimple-low.c (record_vars_into): Don't change cfun.
2669 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix
2670 code generation without cfun.
2672 2014-04-04 Thomas Schwinge <thomas@codesourcery.com>
2675 * Makefile.in (install-driver): Fix shell scripting.
2677 2014-04-03 Cong Hou <congh@google.com>
2679 PR tree-optimization/60505
2680 * tree-vectorizer.h (struct _stmt_vec_info): Add th field as the
2681 threshold of number of iterations below which no vectorization
2683 * tree-vect-loop.c (new_loop_vec_info):
2684 Initialize LOOP_VINFO_COST_MODEL_THRESHOLD.
2685 * tree-vect-loop.c (vect_analyze_loop_operations):
2686 Set LOOP_VINFO_COST_MODEL_THRESHOLD.
2687 * tree-vect-loop.c (vect_transform_loop):
2688 Use LOOP_VINFO_COST_MODEL_THRESHOLD.
2689 * tree-vect-loop.c (vect_analyze_loop_2): Check the maximum number
2690 of iterations of the loop and see if we should build the epilogue.
2692 2014-04-03 Richard Biener <rguenther@suse.de>
2694 * tree-streamer.h (struct streamer_tree_cache_d): Add next_idx member.
2695 (streamer_tree_cache_create): Adjust.
2696 * tree-streamer.c (streamer_tree_cache_add_to_node_array): Adjust
2697 to allow optional nodes array.
2698 (streamer_tree_cache_insert_1): Use next_idx to assign idx.
2699 (streamer_tree_cache_append): Likewise.
2700 (streamer_tree_cache_create): Create nodes array optionally
2701 as specified by parameter.
2702 * lto-streamer-out.c (create_output_block): Avoid maintaining
2703 the node array in the writer cache.
2704 (DFS_write_tree): Remove assertion.
2705 (produce_asm_for_decls): Free the out decl state hash table early.
2706 * lto-streamer-in.c (lto_data_in_create): Adjust for
2707 streamer_tree_cache_create prototype change.
2709 2014-04-03 Richard Biener <rguenther@suse.de>
2711 * tree-streamer-out.c (streamer_write_chain): Do not temporarily
2712 set TREE_CHAIN to NULL_TREE.
2714 2014-04-03 Richard Biener <rguenther@suse.de>
2716 PR tree-optimization/60740
2717 * graphite-scop-detection.c (stmt_simple_for_scop_p): Iterate
2718 over all GIMPLE_COND operands.
2720 2014-04-03 Nathan Sidwell <nathan@codesourcery.com>
2722 * doc/invoke.texi (Wnon-virtual-dtor): Adjust documentation.
2723 (Weffc++): Remove Scott's numbering, merge lists and reference
2726 2014-04-03 Nick Clifton <nickc@redhat.com>
2728 * config/rl78/rl78-expand.md (movqi): Handle (SUBREG (SYMBOL_REF))
2731 2014-04-03 Martin Jambor <mjambor@suse.cz>
2733 * ipa-cp.c (ipcp_verify_propagated_values): Also dump symtab and
2734 mention gcc_unreachable before failing.
2735 * ipa.c (symtab_remove_unreachable_nodes): Also print order of
2738 2014-04-02 Jan Hubicka <hubicka@ucw.cz>
2741 * ipa-devirt.c (get_polymorphic_call_info): Do not ICE on type
2742 inconsistent code and instead mark the context inconsistent.
2743 (possible_polymorphic_call_targets): For inconsistent contexts
2744 return empty complete list.
2746 2014-04-02 Anthony Green <green@moxielogic.com>
2748 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2)
2749 (extendqisi2, extendhisi2): Define.
2750 * config/moxie/moxie.h (DEFAULT_SIGNED_CHAR): Change to 0.
2751 (WCHAR_TYPE): Change to unsigned int.
2753 2014-04-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2755 PR tree-optimization/60733
2756 * gimple-ssa-strength-reduction.c (ncd_with_phi): Change required
2757 insertion point for PHI candidates to be the end of the feeding
2758 block for the PHI argument.
2760 2014-04-02 Vladimir Makarov <vmakarov@redhat.com>
2762 PR rtl-optimization/60650
2763 * lra-constraints.c (process_alt_operands): Decrease reject for
2764 earlyclobber matching.
2766 2014-04-02 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2768 * config/s390/s390.c (s390_expand_insv): Use GET_MODE_BITSIZE.
2770 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
2772 * config/spu/spu.c (pad_bb): Do not crash when the last
2773 insn is CODE_FOR_blockage.
2775 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
2777 * config/spu/spu.md ("insv"): Fail if bitoffset+bitsize
2778 lies outside the target mode.
2780 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
2783 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
2784 software floating point or no floating point registers, do not
2785 allow any type in the FPRs. Eliminate a test for SPE SIMD types
2786 in GPRs that occurs after we tested for GPRs that would never be
2789 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
2790 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
2791 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
2792 specifically allow DDmode, since that does not use the SPE SIMD
2795 2014-04-02 Richard Biener <rguenther@suse.de>
2798 * optabs.c (expand_abs_nojump): Honor flag_trapv only for
2799 MODE_INTs. Properly use negv_optab.
2800 (expand_abs): Likewise.
2802 2014-04-02 Richard Biener <rguenther@suse.de>
2805 * Makefile.in (install-driver): Guard extra installs with special
2808 2014-04-01 Michael Meissner <meissner@linux.vnet.ibm.com>
2810 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
2813 2014-04-01 Richard Henderson <rth@redhat.com>
2816 * config/i386/i386.md (*float<SWI48><MODEF>2_sse): Leave the second
2817 alternative enabled before register allocation.
2819 2014-04-01 Chung-Lin Tang <cltang@codesourcery.com>
2821 * config/nios2/nios2.md (unspec): Remove UNSPEC_TLS, UNSPEC_TLS_LDM.
2822 * config/nios2/nios2.c (nios2_function_profiler): Fix addi operand
2824 (nios2_large_got_address): Remove unneeded 'sym' parameter.
2825 (nios2_got_address): Update nios2_large_got_address call site.
2826 (nios2_delegitimize_address): New function.
2827 (TARGET_DELEGITIMIZE_ADDRESS): Define to nios2_delegitimize_address.
2828 * config/nios2/linux.h (GLIBC_DYNAMIC_LINKER): Define.
2829 (LINK_SPEC): Specify dynamic linker using GNU_USER_DYNAMIC_LINKER.
2831 2014-04-01 Martin Husemann <martin@duskware.de>
2833 * config/mips/netbsd.h (TARGET_OS_CPP_BUILTINS): Define __mips_o32
2836 2014-04-01 Richard Sandiford <rdsandiford@googlemail.com>
2838 PR rtl-optimization/60604
2839 * recog.c (general_operand): Incorporate REG_CANNOT_CHANGE_MODE_P
2840 check from register_operand.
2841 (register_operand): Redefine in terms of general_operand.
2842 (nonmemory_operand): Use register_operand for the non-constant cases.
2844 2014-04-01 Richard Biener <rguenther@suse.de>
2846 * gimple.h (struct gimple_statement_base): Align subcode to 16 bits.
2848 2014-04-01 Sebastian Huber <sebastian.huber@embedded-brains.de>
2850 * doc/invoke.texi (mapp-regs): Clarify.
2852 2014-03-31 Ulrich Drepper <drepper@gmail.com>
2854 * config/i386/avx512fintrin.h (__v32hi): Define type.
2855 (__v64qi): Likewise.
2856 (_mm512_set1_epi8): Define.
2857 (_mm512_set1_epi16): Define.
2858 (_mm512_set4_epi32): Define.
2859 (_mm512_set4_epi64): Define.
2860 (_mm512_set4_pd): Define.
2861 (_mm512_set4_ps): Define.
2862 (_mm512_setr4_epi64): Define.
2863 (_mm512_setr4_epi32): Define.
2864 (_mm512_setr4_pd): Define.
2865 (_mm512_setr4_ps): Define.
2866 (_mm512_setzero_epi32): Define.
2868 2014-03-31 Martin Jambor <mjambor@suse.cz>
2871 * tree-sra.c (callsite_has_enough_arguments_p): Renamed to
2872 callsite_arguments_match_p. Updated all callers. Also check types of
2873 corresponding formal parameters and actual arguments.
2874 (not_all_callers_have_enough_arguments_p) Renamed to
2875 some_callers_have_mismatched_arguments_p.
2877 2014-03-31 Yuri Rumyantsev <ysrumyan@gmail.com>
2879 * tree-inline.c (copy_loops): Add missed copy of 'safelen'.
2881 2014-03-31 Kugan Vivekanandarajah <kuganv@linaro.org>
2884 * aarch64/aarch64.c (aarch64_classify_address): Fix alignment for
2887 2014-03-30 Uros Bizjak <ubizjak@gmail.com>
2889 * config/i386/sse.md (FMAMODE_NOVF512): New mode iterator.
2890 (<sd_mask_codefor>fma_fmadd_<mode><sd_maskz_name><round_name>):
2892 <sd_mask_codefor>fma_fmadd_<VF_512:mode><sd_maskz_name><round_name>.
2893 Use FMAMODE_NOVF512 mode iterator.
2894 (<sd_mask_codefor>fma_fmsub_<mode><sd_maskz_name><round_name>): Ditto.
2895 (<sd_mask_codefor>fma_fnmadd_<mode><sd_maskz_name><round_name>): Ditto.
2896 (<sd_mask_codefor>fma_fnmsub_<mode><sd_maskz_name><round_name>): Ditto.
2897 (<sd_mask_codefor>fma_fmaddsub_<mode><sd_maskz_name><round_name>):
2899 <sd_mask_codefor>fma_fmaddsub_<VF_512:mode><sd_maskz_name><round_name>.
2900 Use VF_128_256 mode iterator.
2901 (<sd_mask_codefor>fma_fmsubadd_<mode><sd_maskz_name><round_name>):
2904 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
2906 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Clear
2907 static chain if needed.
2909 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
2912 * lra-constraints.c (index_part_to_reg): New.
2913 (process_address): Use it.
2915 2014-03-27 Jeff Law <law@redhat.com>
2916 Jakub Jelinek <jakub@redhat.com>
2919 * expr.c (do_tablejump): Use simplify_gen_binary rather than
2920 gen_rtx_{PLUS,MULT} to build up the address expression.
2922 * i386/i386.c (ix86_legitimize_address): Use copy_addr_to_reg to avoid
2923 creating non-canonical RTL.
2925 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
2928 * ipa-inline.c (want_inline_small_function_p): Short circuit large
2929 functions; reorganize to make cheap checks first.
2930 (inline_small_functions): Do not estimate growth when dumping;
2932 * ipa-inline.h (inline_summary): Add min_size.
2933 (growth_likely_positive): New function.
2934 * ipa-inline-analysis.c (dump_inline_summary): Add min_size.
2935 (set_cond_stmt_execution_predicate): Cleanup.
2936 (estimate_edge_size_and_time): Compute min_size.
2937 (estimate_calls_size_and_time): Likewise.
2938 (estimate_node_size_and_time): Likewise.
2939 (inline_update_overall_summary): Update min_size.
2940 (do_estimate_edge_time): Likewise.
2941 (do_estimate_edge_size): Update.
2942 (do_estimate_edge_hints): Update.
2943 (growth_likely_positive): New function.
2945 2014-03-28 Jakub Jelinek <jakub@redhat.com>
2948 * config/i386/i386.c (ix86_copy_addr_to_reg): Call copy_addr_to_reg
2949 also if addr has VOIDmode.
2951 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2953 * config/arm/aarch-common.c (aarch_crypto_can_dual_issue): New.
2954 * config/arm/aarch-common-protos.h (aarch_crypto_can_dual_issue):
2956 * config/arm/cortex-a53.md: Add reservations and bypass for crypto
2957 instructions as well as AdvancedSIMD loads.
2959 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2961 * config/aarch64/aarch64-simd.md (aarch64_crypto_aes<aes_op>v16qi):
2962 Use crypto_aese type.
2963 (aarch64_crypto_aes<aesmc_op>v16qi): Use crypto_aesmc type.
2964 * config/arm/arm.md (is_neon_type): Replace crypto_aes with
2965 crypto_aese, crypto_aesmc. Move to types.md.
2966 * config/arm/types.md (crypto_aes): Split into crypto_aese,
2968 * config/arm/iterators.md (crypto_type): Likewise.
2970 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
2972 * cgraph.c: Include expr.h and tree-dfa.h.
2973 (cgraph_redirect_edge_call_stmt_to_callee): If call in noreturn;
2976 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
2979 * lra-assigns.c (find_hard_regno_for): Remove unavailable hard
2980 regs from checking multi-reg pseudos.
2982 2014-03-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2984 * config/arm/t-aprofile (MULTILIB_MATCHES): Correct A12 rule.
2986 2014-03-28 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
2988 * config/rs6000/rs6000.c (fusion_gpr_load_p): Refuse optimization
2989 if it would clobber the stack pointer, even temporarily.
2991 2014-03-28 Eric Botcazou <ebotcazou@adacore.com>
2993 * mode-switching.c: Make small adjustments to the top comment.
2995 2014-03-27 Michael Meissner <meissner@linux.vnet.ibm.com>
2997 * config/rs6000/constraints.md (wD constraint): New constraint to
2998 match the constant integer to get the top DImode/DFmode out of a
2999 vector in a VSX register.
3001 * config/rs6000/predicates.md (vsx_scalar_64bit): New predicate to
3002 match the constant integer to get the top DImode/DFmode out of a
3003 vector in a VSX register.
3005 * config/rs6000/rs6000-builtins.def (VBPERMQ): Add vbpermq builtin
3008 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
3011 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
3012 -mdebug=reg, print value of VECTOR_ELEMENT_SCALAR_64BIT.
3014 * config/rs6000/vsx.md (vsx_extract_<mode>, V2DI/V2DF modes):
3015 Optimize vec_extract of 64-bit values, where the value being
3016 extracted is in the top word, where we can use scalar
3017 instructions. Add direct move and store support. Combine the big
3018 endian/little endian vector select load support into a single insn.
3019 (vsx_extract_<mode>_internal1): Likewise.
3020 (vsx_extract_<mode>_internal2): Likewise.
3021 (vsx_extract_<mode>_load): Likewise.
3022 (vsx_extract_<mode>_store): Likewise.
3023 (vsx_extract_<mode>_zero): Delete, big and little endian insns are
3024 combined into vsx_extract_<mode>_load.
3025 (vsx_extract_<mode>_one_le): Likewise.
3027 * config/rs6000/rs6000.h (VECTOR_ELEMENT_SCALAR_64BIT): Macro to
3028 define the top 64-bit vector element.
3030 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wD
3033 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
3034 Document vec_vbpermq builtin.
3037 * config/rs6000/altivec.h (vec_xxsldwi): Add missing define to
3038 enable use of xxsldwi and xxpermdi builtin functions.
3039 (vec_xxpermdi): Likewise.
3041 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
3042 Document use of vec_xxsldwi and vec_xxpermdi builtins.
3044 2014-03-27 Vladimir Makarov <vmakarov@redhat.com>
3046 PR rtl-optimization/60650
3047 * lra-assign.c (find_hard_regno_for, spill_for): Add parameter
3049 (find_spills_for): New.
3050 (assign_by_spills): Pass the new parameter to find_hard_regno_for.
3051 Spill all pseudos on the second iteration.
3053 2014-03-27 Marek Polacek <polacek@redhat.com>
3056 * doc/extend.texi (ffs Builtins): Change unsigned types to signed
3059 2014-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3061 * config/s390/s390.c (s390_can_use_return_insn): Check for
3062 call-saved FPRs on 31 bit.
3064 2014-03-27 Jakub Jelinek <jakub@redhat.com>
3067 * omp-low.c (lower_omp_1): For gimple_clobber_p stmts,
3068 if they need regimplification, just drop them instead of
3069 calling gimple_regimplify_operands on them.
3071 2014-03-27 Marcus Shawcroft <marcus.shawcroft@arm.com>
3074 * config/aarch64/aarch64.c (faked_omit_frame_pointer): Remove.
3075 (aarch64_frame_pointer_required): Adjust logic.
3076 (aarch64_can_eliminate): Adjust logic.
3077 (aarch64_override_options_after_change): Adjust logic.
3079 2014-03-27 Dehao Chen <dehao@google.com>
3081 * ipa-inline.c (early_inliner): Update node's inline info.
3083 2014-03-26 Dehao Chen <dehao@google.com>
3085 * dojump.c (do_compare_rtx_and_jump): Sets correct probability for
3086 compiler inserted conditional jumps for NAN float check.
3088 2014-03-26 Jakub Jelinek <jakub@redhat.com>
3090 * ubsan.h (ubsan_create_data): Change second argument's type
3091 to const location_t *.
3092 * ubsan.c (ubsan_source_location): If xloc.file is NULL, set it to
3094 (ubsan_create_data): Change second argument to const location_t *PLOC.
3095 Create Loc field whenever PLOC is non-NULL.
3096 (ubsan_instrument_unreachable, ubsan_expand_null_ifn,
3097 ubsan_build_overflow_builtin, instrument_bool_enum_load): Adjust
3101 * real.c (real_to_integer2): Change type of low to UHWI.
3103 2014-03-26 Tobias Burnus <burnus@net-b.de>
3105 * gcc.c (LINK_COMMAND_SPEC): Use libcilkrts.spec for -fcilkplus.
3106 (CILK_SELF_SPECS): New define.
3107 (driver_self_specs): Use it.
3109 2014-03-26 Richard Biener <rguenther@suse.de>
3111 * tree-pretty-print.c (percent_K_format): Implement special
3112 case for LTO and its stripped down BLOCK tree.
3114 2014-03-26 Jakub Jelinek <jakub@redhat.com>
3117 * ubsan.c (instrument_si_overflow): Instrument ABS_EXPR.
3119 * tree-vrp.c (simplify_internal_call_using_ranges): If only
3120 one range is range_int_cst_p, but not both, at least optimize
3121 addition/subtraction of 0 and multiplication by 0 or 1.
3122 * gimple-fold.c (gimple_fold_call): Fold
3123 IFN_UBSAN_CHECK_{ADD,SUB,MUL}.
3124 (gimple_fold_stmt_to_constant_1): If both op0 and op1 aren't
3125 INTEGER_CSTs, try to fold at least x * 0 and y - y.
3127 2014-03-26 Eric Botcazou <ebotcazou@adacore.com>
3129 PR rtl-optimization/60452
3130 * rtlanal.c (rtx_addr_can_trap_p_1): Fix head comment.
3131 <case REG>: Return 1 for invalid offsets from the frame pointer.
3133 2014-03-26 Marek Polacek <polacek@redhat.com>
3136 * doc/extend.texi (C Extensions): Mention variable-length arrays in
3139 2014-03-26 Marek Polacek <polacek@redhat.com>
3142 * doc/extend.texi (Designated Inits): Describe what happens to omitted
3145 2014-03-26 Marek Polacek <polacek@redhat.com>
3148 * ira-color.c (update_conflict_hard_regno_costs): Perform the
3149 multiplication in unsigned type.
3151 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
3153 * doc/install.texi: Document nds32le-*-elf and nds32be-*-elf.
3155 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
3157 * doc/contrib.texi: Add myself as Andes nds32 port contributor.
3159 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
3162 * cif-code.def (UNREACHABLE) New code.
3163 * ipa-inline.c (inline_small_functions): Skip edges to
3164 __builtlin_unreachable.
3165 (estimate_edge_growth): Allow edges to __builtlin_unreachable.
3166 * ipa-inline-analysis.c (edge_set_predicate): Redirect edges with false
3167 predicate to __bulitin_unreachable.
3168 (set_cond_stmt_execution_predicate): Fix issue when
3169 invert_tree_comparison returns ERROR_MARK.
3170 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow): Do not
3171 propagate to inline clones.
3172 * cgraph.c (verify_edge_corresponds_to_fndecl): Allow redirection
3174 * ipa-cp.c (create_specialized_node): Be ready for new node to appear.
3175 * cgraphclones.c (cgraph_clone_node): If call destination is already
3176 ureachable, do not redirect it back.
3177 * tree-inline.c (fold_marked_statements): Hanlde calls becoming
3180 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
3182 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow):
3183 Do not modify inline clones.
3185 2014-03-25 Jakub Jelinek <jakub@redhat.com>
3187 * config/i386/i386.md (general_sext_operand): New mode attr.
3188 (addv<mode>4, subv<mode>4, mulv<mode>4): If operands[2] is CONST_INT,
3189 don't generate (sign_extend (const_int)).
3190 (*addv<mode>4, *subv<mode>4, *mulv<mode>4): Disallow CONST_INT_P
3191 operands[2]. Use We constraint instead of <i> and
3192 <general_sext_operand> predicate instead of <general_operand>.
3193 (*addv<mode>4_1, *subv<mode>4_1, *mulv<mode>4_1): New insns.
3194 * config/i386/constraints.md (We): New constraint.
3195 * config/i386/predicates.md (x86_64_sext_operand,
3196 sext_operand): New predicates.
3198 2014-03-25 Martin Jambor <mjambor@suse.cz>
3201 * ipa-cp.c (ipa_get_indirect_edge_target_1): Redirect type
3202 inconsistent devirtualizations to __builtin_unreachable.
3204 2014-03-25 Marek Polacek <polacek@redhat.com>
3207 * doc/extend.texi (Example of asm with clobbered asm reg): Fix typo.
3209 2014-03-25 Alan Lawrence <alan.lawrence@arm.com>
3211 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Reverse
3212 order of elements for big-endian.
3214 2014-03-25 Richard Biener <rguenther@suse.de>
3217 * gimplify-me.c (gimple_regimplify_operands): Update the
3220 2014-03-25 Martin Jambor <mjambor@suse.cz>
3223 * lto-cgraph.c (lto_output_node): Stream body_removed flag.
3224 (lto_output_varpool_node): Likewise.
3225 (input_overwrite_node): Likewise.
3226 (input_varpool_node): Likewise.
3228 2014-03-25 Richard Biener <rguenther@suse.de>
3230 * lto-wrapper.c (merge_and_complain): Handle OPT_fPIE like OPT_fpie.
3231 (run_gcc): Likewise.
3233 2014-03-25 Jakub Jelinek <jakub@redhat.com>
3235 * combine.c (simplify_compare_const): Add MODE argument.
3236 Handle mode_width 0 as very large mode_width.
3237 (try_combine, simplify_comparison): Adjust callers.
3239 * cselib.c (cselib_hash_rtx): Perform addition in unsigned
3240 type to avoid signed integer overflow.
3241 * explow.c (plus_constant): Likewise.
3243 2014-03-25 Dominik Vogt <vogt@linux.vnet.ibm.com>
3245 * doc/generic.texi: Correct typos.
3247 2014-03-24 Tobias Burnus <burnus@net-b.de>
3249 * doc/invoke.texi (-flto): Expand section about
3250 using static libraries with LTO.
3252 2014-03-24 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3254 PR rtl-optimization/60501
3255 * optabs.def (addptr3_optab): New optab.
3256 * optabs.c (gen_addptr3_insn, have_addptr3_insn): New function.
3257 * doc/md.texi ("addptrm3"): Document new RTL standard expander.
3258 * expr.h (gen_addptr3_insn, have_addptr3_insn): Add prototypes.
3260 * lra.c (emit_add3_insn): Use the addptr pattern if available.
3262 * config/s390/s390.md ("addptrdi3", "addptrsi3"): New expanders.
3264 2014-03-24 Ulrich Drepper <drepper@gmail.com>
3266 * config/i386/avx512fintrin.h: Define _mm512_set1_ps and
3269 * config/i386/avxintrin.h (_mm256_undefined_si256): Define.
3270 (_mm256_undefined_ps): Define.
3271 (_mm256_undefined_pd): Define.
3272 * config/i386/emmintrin.h (_mm_undefined_si128): Define.
3273 (_mm_undefined_pd): Define.
3274 * config/i386/xmmintrin.h (_mm_undefined_ps): Define.
3275 * config/i386/avx512fintrin.h (_mm512_undefined_si512): Define.
3276 (_mm512_undefined_ps): Define.
3277 (_mm512_undefined_pd): Define.
3278 Use _mm*_undefined_*.
3279 * config/i386/avx2intrin.h: Use _mm*_undefined_*.
3281 2014-03-24 Alex Velenko <Alex.Velenko@arm.com>
3283 * config/aarch64/aarch64-simd-builtins.def (lshr): DI mode excluded.
3284 (lshr_simd): DI mode added.
3285 * config/aarch64/aarch64-simd.md (aarch64_lshr_simddi): New pattern.
3286 (aarch64_ushr_simddi): Likewise.
3287 * config/aarch64/aarch64.md (UNSPEC_USHR64): New unspec.
3288 * config/aarch64/arm_neon.h (vshr_n_u64): Intrinsic fixed.
3289 (vshrd_n_u64): Likewise.
3291 2014-03-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3293 * Makefile.in (s-macro_list): Depend on cc1.
3295 2014-03-23 Teresa Johnson <tejohnson@google.com>
3297 * ipa-utils.c (ipa_print_order): Use specified dump file.
3299 2014-03-23 Eric Botcazou <ebotcazou@adacore.com>
3301 PR rtl-optimization/60601
3302 * bb-reorder.c (fix_up_fall_thru_edges): Test EDGE_FALLTHRU everywhere.
3304 * gcc.c (eval_spec_function): Initialize save_growing_value.
3306 2014-03-22 Jakub Jelinek <jakub@redhat.com>
3309 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): For
3310 code == MINUS_EXPR, never swap op0 with op1.
3312 * toplev.c (init_local_tick): Avoid signed integer multiplication
3314 * genautomata.c (reserv_sets_hash_value): Fix rotate idiom, avoid
3315 shift by first operand's bitsize.
3317 2014-03-21 Jakub Jelinek <jakub@redhat.com>
3320 * config/i386/i386.h (TARGET_64BIT_P): If not TARGET_BI_ARCH,
3322 * config/i386/darwin.h (TARGET_64BIT_P): Redefine to
3323 TARGET_ISA_64BIT_P(x).
3325 2014-03-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3327 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Generate a
3328 pattern for vector nor instead of subtract from splat(-1).
3329 (altivec_expand_vec_perm_const_le): Likewise.
3331 2014-03-21 Richard Henderson <rth@twiddle.net>
3334 * ifcvt.c (dead_or_predicable): Return FALSE if there are any frame
3335 related insns after epilogue_completed.
3337 2014-03-21 Martin Jambor <mjambor@suse.cz>
3340 * cgraph.h (symtab_node): New flag body_removed.
3341 * ipa.c (symtab_remove_unreachable_nodes): Set body_removed flag
3342 when removing bodies.
3343 * symtab.c (dump_symtab_base): Dump body_removed flag.
3344 * cgraph.c (verify_edge_corresponds_to_fndecl): Skip nodes which
3345 had their bodies removed.
3347 2014-03-21 Martin Jambor <mjambor@suse.cz>
3350 * ipa.c (symtab_remove_unreachable_nodes): Clear thunk flag of nodes
3353 2014-03-21 Richard Biener <rguenther@suse.de>
3355 PR tree-optimization/60577
3356 * tree-core.h (struct tree_base): Document nothrow_flag use
3357 in VAR_DECL_NONALIASED.
3358 * tree.h (VAR_DECL_NONALIASED): New.
3359 (may_be_aliased): Adjust.
3360 * coverage.c (build_var): Set VAR_DECL_NONALIASED.
3362 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
3364 * expr.c (expand_expr_real_1): Remove outdated comment.
3366 2014-03-20 Jakub Jelinek <jakub@redhat.com>
3369 * ira.c (adjust_cleared_regs): Call copy_rtx on
3370 *reg_equiv[REGNO (loc)].src_p before passing it to
3371 simplify_replace_fn_rtx.
3374 * config/i386/i386.c (x86_output_mi_thunk): Surround UNSPEC_GOT
3375 into CONST, put pic register as first operand of PLUS. Use
3376 gen_const_mem for both 32-bit and 64-bit PIC got loads.
3378 2014-03-20 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3380 * config/aarch64/aarch64.c (MEMORY_MOVE_COST): Delete.
3382 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
3384 * config/sparc/sparc.c (sparc_do_work_around_errata): Implement work
3385 around for store forwarding issue in the FPU on the UT699.
3386 * config/sparc/sparc.md (in_branch_delay): Return false for single FP
3387 loads and operations if -mfix-ut699 is specified.
3388 (divtf3_hq): Tweak attribute.
3389 (sqrttf2_hq): Likewise.
3391 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
3393 * calls.c (store_one_arg): Remove incorrect const qualification on the
3394 type of the temporary.
3395 * cfgexpand.c (expand_return): Likewise.
3396 * expr.c (expand_constructor): Likewise.
3397 (expand_expr_real_1): Likewise.
3399 2014-03-20 Zhenqiang Chen <zhenqiang.chen@linaro.org>
3401 * config/arm/arm.c (arm_dwarf_register_span): Update the element number
3404 2014-03-19 Kaz Kojima <kkojima@gcc.gnu.org>
3407 * config/sh/sh.md (udivsi3_i1): Clobber R1 register.
3409 2014-03-19 James Greenhalgh <james.greenhalgh@arm.com>
3411 * config/arm/aarch-common-protos.h
3412 (alu_cost_table): Fix spelling of "extend".
3413 * config/arm/arm.c (arm_new_rtx_costs): Fix spelling of "extend".
3415 2014-03-19 Richard Biener <rguenther@suse.de>
3418 * tree-core.h (tree_type_common): Re-order pointer members
3419 to reduce recursion depth during GC walks.
3421 2014-03-19 Marek Polacek <polacek@redhat.com>
3424 * ubsan.c (ubsan_type_descriptor): Check that DECL_NAME is nonnull
3425 before accessing it.
3427 2014-03-19 Richard Biener <rguenther@suse.de>
3430 * lto-streamer-in.c (input_function): In WPA stage do not drop
3433 2014-03-19 Jakub Jelinek <jakub@redhat.com>
3435 PR tree-optimization/60559
3436 * vectorizable_mask_load_store): Replace scalar MASK_LOAD
3437 with build_zero_cst assignment.
3439 2014-03-18 Kai Tietz <ktietz@redhat.com>
3441 PR rtl-optimization/56356
3442 * sdbout.c (sdbout_parms): Verify that parms'
3443 incoming argument is valid.
3444 (sdbout_reg_parms): Likewise.
3446 2014-03-18 Richard Henderson <rth@redhat.com>
3449 * config/i386/i386.md (*float<SWI48x><MODEF>2_i387): Move down to
3450 be shadowed by *float<SWI48><MODEF>2_sse. Test X87_ENABLE_FLOAT.
3451 (*float<SWI48><MODEF>2_sse): Check X87_ENABLE_FLOAT for alternative 0.
3453 2014-03-18 Basile Starynkevitch <basile@starynkevitch.net>
3455 * plugin.def: Improve comment for PLUGIN_INCLUDE_FILE.
3456 * doc/plugins.texi (Plugin callbacks): Mention PLUGIN_INCLUDE_FILE.
3457 Italicize plugin event names in description. Explain that
3458 PLUGIN_PRAGMAS has no sense for lto1. Explain PLUGIN_INCLUDE_FILE.
3459 Remind that no GCC functions should be called after PLUGIN_FINISH.
3460 Explain what pragmas with expansion are.
3462 2014-03-18 Martin Liska <mliska@suse.cz>
3464 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Added case when
3465 gimple call statement is update.
3466 * gimple-fold.c (gimple_fold_call): Changed order for GIMPLE_ASSIGN and
3467 GIMPLE_CALL, where gsi iterator still points to GIMPLE CALL.
3469 2014-03-18 Jakub Jelinek <jakub@redhat.com>
3472 * ubsan.c (ubsan_instrument_unreachable): Call
3473 initialize_sanitizer_builtins.
3474 (ubsan_pass): Likewise.
3477 * ubsan.c (ubsan_type_descriptor, ubsan_create_data): Call
3478 varpool_finalize_decl instead of rest_of_decl_compilation.
3480 2014-03-18 Richard Biener <rguenther@suse.de>
3482 * df-problems.c (df_rd_confluence_n): Avoid bitmap_copy
3483 by using bitmap_and_compl instead of bitmap_and_compl_into.
3484 (df_rd_transfer_function): Likewise.
3486 2014-03-18 Richard Biener <rguenther@suse.de>
3488 * doc/lto.texi (fresolution): Fix typo.
3490 2014-03-18 Richard Biener <rguenther@suse.de>
3492 * doc/invoke.texi (flto): Update for changes in 4.9.
3494 2014-03-18 Richard Biener <rguenther@suse.de>
3496 * doc/loop.texi: Remove section on the removed lambda framework.
3497 Update loop docs with recent changes in preserving loop structure.
3499 2014-03-18 Richard Biener <rguenther@suse.de>
3501 * doc/lto.texi (-fresolution): Document.
3503 2014-03-18 Richard Biener <rguenther@suse.de>
3505 * doc/contrib.texi: Adjust my name.
3507 2014-03-18 Jakub Jelinek <jakub@redhat.com>
3510 * internal-fn.c: Include diagnostic-core.h.
3511 (expand_BUILTIN_EXPECT): New function.
3512 * gimplify.c (gimplify_call_expr): Use false instead of FALSE.
3513 (gimplify_modify_expr): Gimplify 3 argument __builtin_expect into
3514 IFN_BUILTIN_EXPECT call instead of __builtin_expect builtin call.
3515 * ipa-inline-analysis.c (find_foldable_builtin_expect): Handle
3517 * predict.c (expr_expected_value_1): Handle IFN_BUILTIN_EXPECT.
3518 Revert 3 argument __builtin_expect code.
3519 (strip_predict_hints): Handle IFN_BUILTIN_EXPECT.
3520 * gimple-fold.c (gimple_fold_call): Likewise.
3521 * tree.h (fold_builtin_expect): New prototype.
3522 * builtins.c (build_builtin_expect_predicate): Add predictor
3523 argument, if non-NULL, create 3 argument __builtin_expect.
3524 (fold_builtin_expect): No longer static. Add ARG2 argument,
3525 pass it through to build_builtin_expect_predicate.
3526 (fold_builtin_2): Adjust caller.
3527 (fold_builtin_3): Handle BUILT_IN_EXPECT.
3528 * internal-fn.def (BUILTIN_EXPECT): New.
3530 2014-03-18 Tobias Burnus <burnus@net-b.de>
3533 * predict.def (PRED_FORTRAN_OVERFLOW, PRED_FORTRAN_FAIL_ALLOC,
3534 PRED_FORTRAN_FAIL_IO, PRED_FORTRAN_WARN_ONCE, PRED_FORTRAN_SIZE_ZERO,
3535 PRED_FORTRAN_INVALID_BOUND, PRED_FORTRAN_ABSENT_DUMMY): Add.
3537 2014-03-18 Jan Hubicka <hubicka@ucw.cz>
3540 * predict.c (combine_predictions_for_bb): Fix up formatting.
3541 (expr_expected_value_1, expr_expected_value): Add predictor argument,
3542 fill what it points to if non-NULL.
3543 (tree_predict_by_opcode): Adjust caller, use the predictor.
3544 * predict.def (PRED_COMPARE_AND_SWAP): Add.
3546 2014-03-18 Eric Botcazou <ebotcazou@adacore.com>
3548 * config/sparc/sparc.c (sparc_do_work_around_errata): Speed up and use
3549 proper constant for the store mode.
3551 2014-03-18 Ilya Enkovich <ilya.enkovich@intel.com>
3553 * symtab.c (change_decl_assembler_name): Fix transparent alias
3556 2014-03-16 Renlin Li <Renlin.Li@arm.com>
3558 * config/aarch64/aarch64.c: Correct the comments about the
3559 aarch64 stack layout.
3561 2014-03-18 Thomas Schwinge <thomas@codesourcery.com>
3563 * omp-low.c (lower_rec_input_clauses) <build_omp_barrier>: Restore
3564 check for GF_OMP_FOR_KIND_FOR.
3566 2013-03-18 Kirill Yukhin <kirill.yukhin@intel.com>
3568 * config/i386/i386.h (ADDITIONAL_REGISTER_NAMES): Add
3569 ymm and zmm register names.
3571 2014-03-17 Jakub Jelinek <jakub@redhat.com>
3574 * config/i386/i386.c (ix86_expand_epilogue): Adjust REG_CFA_ADJUST_CFA
3575 note creation for the 2010-08-31 changes.
3577 2014-03-17 Marek Polacek <polacek@redhat.com>
3580 * omp-low.c (omp_max_vf): Treat -fno-tree-loop-optimize the same
3581 as -fno-tree-loop-vectorize.
3582 (expand_omp_simd): Likewise.
3584 2014-03-15 Eric Botcazou <ebotcazou@adacore.com>
3586 * config/sparc/sparc-protos.h (tls_call_delay): Delete.
3587 (eligible_for_call_delay): New prototype.
3588 * config/sparc/sparc.c (tls_call_delay): Rename into...
3589 (eligible_for_call_delay): ...this. Return false if the instruction
3590 cannot be put in the delay slot of a branch.
3591 (eligible_for_restore_insn): Simplify.
3592 (eligible_for_return_delay): Return false if the instruction cannot be
3593 put in the delay slot of a branch and simplify.
3594 (eligible_for_sibcall_delay): Return false if the instruction cannot be
3595 put in the delay slot of a branch.
3596 * config/sparc/sparc.md (fix_ut699): New attribute.
3597 (tls_call_delay): Delete.
3598 (in_call_delay): Reimplement.
3599 (eligible_for_sibcall_delay): Rename into...
3600 (in_sibcall_delay): ...this.
3601 (eligible_for_return_delay): Rename into...
3602 (in_return_delay): ...this.
3603 (in_branch_delay): Reimplement.
3604 (in_uncond_branch_delay): Delete.
3605 (in_annul_branch_delay): Delete.
3607 2014-03-14 Richard Henderson <rth@redhat.com>
3610 * config/i386/i386.md (floathi<X87MODEF>2): Delete expander; rename
3611 define_insn from *floathi<X87MODEF>2_i387; allow nonimmediate_operand.
3612 (*floathi<X87MODEF>2_i387_with_temp): Remove.
3613 (floathi splitters): Remove.
3614 (float<SWI48x>xf2): New pattern.
3615 (float<SWI48><MODEF>2): Rename from float<SWI48x><X87MODEF>2. Drop
3616 code that tried to handle DImode for 32-bit, but which was excluded
3617 by the pattern's condition. Drop allocation of stack temporary.
3618 (*floatsi<MODEF>2_vector_mixed_with_temp): Remove.
3619 (*float<SWI48><MODEF>2_mixed_with_temp): Remove.
3620 (*float<SWI48><MODEF>2_mixed_interunit): Remove.
3621 (*float<SWI48><MODEF>2_mixed_nointerunit): Remove.
3622 (*floatsi<MODEF>2_vector_sse_with_temp): Remove.
3623 (*float<SWI48><MODEF>2_sse_with_temp): Remove.
3624 (*float<SWI48><MODEF>2_sse_interunit): Remove.
3625 (*float<SWI48><MODEF>2_sse_nointerunit): Remove.
3626 (*float<SWI48x><X87MODEF>2_i387_with_temp): Remove.
3627 (*float<SWI48x><X87MODEF>2_i387): Remove.
3628 (all float _with_temp splitters): Remove.
3629 (*float<SWI48x><MODEF>2_i387): New pattern.
3630 (*float<SWI48><MODEF>2_sse): New pattern.
3631 (float TARGET_USE_VECTOR_CONVERTS splitters): Merge them.
3632 (float TARGET_SSE_PARTIAL_REG_DEPENDENCY splitters): Merge them.
3634 2014-03-14 Jakub Jelinek <jakub@redhat.com>
3635 Marek Polacek <polacek@redhat.com>
3638 * common.opt (dump_base_name_prefixed): New Variable.
3639 * opts.c (finish_options): Don't prepend directory to x_dump_base_name
3640 if x_dump_base_name_prefixed is already set, set it at the end.
3642 2014-03-14 Vladimir Makarov <vmakarov@redhat.com>
3644 PR rtl-optimization/60508
3645 * lra-constraints.c (get_reload_reg): Add new parameter
3647 (process_addr_reg, simplify_operand_subreg, curr_insn_transform):
3648 Pass the new parameter values.
3650 2014-03-14 Richard Biener <rguenther@suse.de>
3652 * common.opt: Revert unintented changes from r205065.
3655 2014-03-14 Richard Biener <rguenther@suse.de>
3658 * cfghooks.c (split_block): Properly adjust all loops the
3659 block was a latch of.
3661 2014-03-14 Martin Jambor <mjambor@suse.cz>
3664 * ipa-prop.c (ipa_modify_call_arguments): Fix iteration condition
3667 2014-03-14 Georg-Johann Lay <avr@gjlay.de>
3670 * config/avr/avr.c (avr_set_current_function): Pass function name
3671 through default_strip_name_encoding before sanity checking instead
3672 of skipping the first char of the assembler name.
3674 2014-03-13 Richard Henderson <rth@redhat.com>
3677 * config/i386/i386.c (ix86_split_fp_branch): Remove pushed argument.
3678 (ix86_force_to_memory, ix86_free_from_memory): Remove.
3679 * config/i386/i386-protos.h: Likewise.
3680 * config/i386/i386.md (floathi<X87MODEF>2): Use assign_386_stack_local
3681 in the expander instead of a splitter.
3682 (float<SWI48x><X87MODEF>2): Use assign_386_stack_local if there is
3683 any possibility of requiring a memory.
3684 (*floatsi<MODEF>2_vector_mixed): Remove, and the splitters.
3685 (*floatsi<MODEF>2_vector_sse): Remove, and the splitters.
3686 (fp branch splitters): Update for ix86_split_fp_branch.
3687 (*jcc<X87MODEF>_<SWI24>_i387): Remove r/f alternative.
3688 (*jcc<X87MODEF>_<SWI24>_r_i387): Likewise.
3689 (splitter for jcc<X87MODEF>_<SWI24>_i387 r/f): Remove.
3690 (*fop_<MODEF>_2_i387): Remove f/r alternative.
3691 (*fop_<MODEF>_3_i387): Likewise.
3692 (*fop_xf_2_i387, *fop_xf_3_i387): Likewise.
3693 (splitters for the fop_* register patterns): Remove.
3694 (fscalexf4_i387): Rename from *fscalexf4_i387.
3695 (ldexpxf3): Use gen_floatsixf2 and gen_fscalexf4_i387.
3697 2014-03-13 Jakub Jelinek <jakub@redhat.com>
3699 PR tree-optimization/59779
3700 * tree-dfa.c (get_ref_base_and_extent): Use double_int
3701 type for bitsize and maxsize instead of HOST_WIDE_INT.
3703 2014-03-13 Steven Bosscher <steven@gcc.gnu.org>
3705 PR rtl-optimization/57320
3706 * function.c (rest_of_handle_thread_prologue_and_epilogue): Cleanup
3707 the CFG after thread_prologue_and_epilogue_insns.
3709 2014-03-13 Vladimir Makarov <vmakarov@redhat.com>
3711 PR rtl-optimization/57189
3712 * lra-constraints.c (process_alt_operands): Disfavor spilling
3715 2014-03-13 Cesar Philippidis <cesar@codesourcery.com>
3717 * lto-wrapper.c (maybe_unlink_file): Suppress diagnostic messages.
3719 2014-03-13 Jakub Jelinek <jakub@redhat.com>
3721 PR tree-optimization/59025
3723 * tree-ssa-reassoc.c (sort_by_operand_rank): For SSA_NAMEs with the
3724 same rank, sort by bb_rank and gimple_uid of SSA_NAME_DEF_STMT first.
3726 2014-03-13 Georg-Johann Lay <avr@gjlay.de>
3729 * config/avr/avr.c (avr_out_plus): Swap cc_plus and cc_minus in
3730 calls of avr_out_plus_1.
3732 2014-03-13 Bin Cheng <bin.cheng@arm.com>
3734 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Record
3735 BB's single pred and update the father loop's latch info later.
3737 2014-03-12 Michael Meissner <meissner@linux.vnet.ibm.com>
3739 * config/rs6000/vector.md (VEC_L): Add V1TI mode to vector types.
3743 (VEC_base): Likewise.
3744 (mov<MODE>, VEC_M modes): If we are loading TImode into VSX
3745 registers, we need to swap double words in little endian mode.
3747 * config/rs6000/rs6000-modes.def (V1TImode): Add new vector mode
3748 to be a container mode for 128-bit integer operations added in ISA
3749 2.07. Unlike TImode and PTImode, the preferred register set is
3750 the Altivec/VMX registers for the 128-bit operations.
3752 * config/rs6000/rs6000-protos.h (rs6000_move_128bit_ok_p): Add
3754 (rs6000_split_128bit_ok_p): Likewise.
3756 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_3): Add new support
3757 macros for creating ISA 2.07 normal and overloaded builtin
3758 functions with 3 arguments.
3759 (BU_P8V_OVERLOAD_3): Likewise.
3760 (VPERM_1T): Add support for V1TImode in 128-bit vector operations
3761 for use as overloaded functions.
3762 (VPERM_1TI_UNS): Likewise.
3763 (VSEL_1TI): Likewise.
3764 (VSEL_1TI_UNS): Likewise.
3765 (ST_INTERNAL_1ti): Likewise.
3766 (LD_INTERNAL_1ti): Likewise.
3767 (XXSEL_1TI): Likewise.
3768 (XXSEL_1TI_UNS): Likewise.
3769 (VPERM_1TI): Likewise.
3770 (VPERM_1TI_UNS): Likewise.
3771 (XXPERMDI_1TI): Likewise.
3772 (SET_1TI): Likewise.
3773 (LXVD2X_V1TI): Likewise.
3774 (STXVD2X_V1TI): Likewise.
3775 (VEC_INIT_V1TI): Likewise.
3776 (VEC_SET_V1TI): Likewise.
3777 (VEC_EXT_V1TI): Likewise.
3778 (EQV_V1TI): Likewise.
3779 (NAND_V1TI): Likewise.
3780 (ORC_V1TI): Likewise.
3781 (VADDCUQ): Add support for 128-bit integer arithmetic instructions
3782 added in ISA 2.07. Add both normal 'altivec' builtins, and the
3784 (VADDUQM): Likewise.
3785 (VSUBCUQ): Likewise.
3786 (VADDEUQM): Likewise.
3787 (VADDECUQ): Likewise.
3788 (VSUBEUQM): Likewise.
3789 (VSUBECUQ): Likewise.
3791 * config/rs6000/rs6000-c.c (__int128_type): New static to hold
3792 __int128_t and __uint128_t types.
3793 (__uint128_type): Likewise.
3794 (altivec_categorize_keyword): Add support for vector __int128_t,
3795 vector __uint128_t, vector __int128, and vector unsigned __int128
3796 as a container type for TImode operations that need to be done in
3797 VSX/Altivec registers.
3798 (rs6000_macro_to_expand): Likewise.
3799 (altivec_overloaded_builtins): Add ISA 2.07 overloaded functions
3800 to support 128-bit integer instructions vaddcuq, vadduqm,
3801 vaddecuq, vaddeuqm, vsubcuq, vsubuqm, vsubecuq, vsubeuqm.
3802 (altivec_resolve_overloaded_builtin): Add support for V1TImode.
3804 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Add support
3805 for V1TImode, and set up preferences to use VSX/Altivec registers.
3806 Setup VSX reload handlers.
3807 (rs6000_debug_reg_global): Likewise.
3808 (rs6000_init_hard_regno_mode_ok): Likewise.
3809 (rs6000_preferred_simd_mode): Likewise.
3810 (vspltis_constant): Do not allow V1TImode as easy altivec constants.
3811 (easy_altivec_constant): Likewise.
3812 (output_vec_const_move): Likewise.
3813 (rs6000_expand_vector_set): Convert V1TImode set and extract to
3815 (rs6000_expand_vector_extract): Likewise.
3816 (reg_offset_addressing_ok_p): Setup V1TImode to use VSX reg+reg
3818 (rs6000_const_vec): Add support for V1TImode.
3819 (rs6000_emit_le_vsx_load): Swap double words when loading or
3820 storing TImode/V1TImode.
3821 (rs6000_emit_le_vsx_store): Likewise.
3822 (rs6000_emit_le_vsx_move): Likewise.
3823 (rs6000_emit_move): Add support for V1TImode.
3824 (altivec_expand_ld_builtin): Likewise.
3825 (altivec_expand_st_builtin): Likewise.
3826 (altivec_expand_vec_init_builtin): Likewise.
3827 (altivec_expand_builtin): Likewise.
3828 (rs6000_init_builtins): Add support for V1TImode type. Add
3829 support for ISA 2.07 128-bit integer builtins. Define type names
3830 for the VSX/Altivec vector types.
3831 (altivec_init_builtins): Add support for overloaded vector
3832 functions with V1TImode type.
3833 (rs6000_preferred_reload_class): Prefer Altivec registers for V1TImode.
3834 (rs6000_move_128bit_ok_p): Move 128-bit move/split validation to
3836 (rs6000_split_128bit_ok_p): Likewise.
3837 (rs6000_handle_altivec_attribute): Create V1TImode from vector
3838 __int128_t and vector __uint128_t.
3840 * config/rs6000/vsx.md (VSX_L): Add V1TImode to vector iterators
3841 and mode attributes.
3848 (VS_scalar): Likewise.
3849 (VS_double): Likewise.
3850 (vsx_set_v1ti): New builtin function to create V1TImode from TImode.
3852 * config/rs6000/rs6000.h (TARGET_VADDUQM): New macro to say whether
3853 we support the ISA 2.07 128-bit integer arithmetic instructions.
3854 (ALTIVEC_OR_VSX_VECTOR_MODE): Add V1TImode.
3855 (enum rs6000_builtin_type_index): Add fields to hold V1TImode
3856 and TImode types for use with the builtin functions.
3857 (V1TI_type_node): Likewise.
3858 (unsigned_V1TI_type_node): Likewise.
3859 (intTI_type_internal_node): Likewise.
3860 (uintTI_type_internal_node): Likewise.
3862 * config/rs6000/altivec.md (UNSPEC_VADDCUQ): New unspecs for ISA 2.07
3863 128-bit builtin functions.
3864 (UNSPEC_VADDEUQM): Likewise.
3865 (UNSPEC_VADDECUQ): Likewise.
3866 (UNSPEC_VSUBCUQ): Likewise.
3867 (UNSPEC_VSUBEUQM): Likewise.
3868 (UNSPEC_VSUBECUQ): Likewise.
3869 (VM): Add V1TImode to vector mode iterators.
3871 (VI_unit): Likewise.
3872 (altivec_vadduqm): Add ISA 2.07 128-bit binary builtins.
3873 (altivec_vaddcuq): Likewise.
3874 (altivec_vsubuqm): Likewise.
3875 (altivec_vsubcuq): Likewise.
3876 (altivec_vaddeuqm): Likewise.
3877 (altivec_vaddecuq): Likewise.
3878 (altivec_vsubeuqm): Likewise.
3879 (altivec_vsubecuq): Likewise.
3881 * config/rs6000/rs6000.md (FMOVE128_GPR): Add V1TImode to vector
3883 (BOOL_128): Likewise.
3884 (BOOL_REGS_OUTPUT): Likewise.
3885 (BOOL_REGS_OP1): Likewise.
3886 (BOOL_REGS_OP2): Likewise.
3887 (BOOL_REGS_UNARY): Likewise.
3888 (BOOL_REGS_AND_CR0): Likewise.
3890 * config/rs6000/altivec.h (vec_vaddcuq): Add support for ISA 2.07
3891 128-bit integer builtin support.
3892 (vec_vadduqm): Likewise.
3893 (vec_vaddecuq): Likewise.
3894 (vec_vaddeuqm): Likewise.
3895 (vec_vsubecuq): Likewise.
3896 (vec_vsubeuqm): Likewise.
3897 (vec_vsubcuq): Likewise.
3898 (vec_vsubuqm): Likewise.
3900 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
3901 Document vec_vaddcuq, vec_vadduqm, vec_vaddecuq, vec_vaddeuqm,
3902 vec_subecuq, vec_subeuqm, vec_vsubcuq, vec_vsubeqm builtins adding
3903 128-bit integer add/subtract to ISA 2.07.
3905 2014-03-12 Joern Rennecke <joern.rennecke@embecosm.com>
3907 * config/arc/arc.c (arc_predicate_delay_insns):
3908 Fix third argument passed to conditionalize_nonjump.
3910 2014-03-12 Yufeng Zhang <yufeng.zhang@arm.com>
3912 * config/aarch64/aarch64-builtins.c
3913 (aarch64_builtin_vectorized_function): Add BUILT_IN_LFLOORF,
3914 BUILT_IN_LLFLOOR, BUILT_IN_LCEILF and BUILT_IN_LLCEIL.
3915 * config/aarch64/arm_neon.h (vcvtaq_u64_f64): Call __builtin_llfloor
3916 instead of __builtin_lfloor.
3917 (vcvtnq_u64_f64): Call __builtin_llceil instead of __builtin_lceil.
3919 2014-03-12 Jakub Jelinek <jakub@redhat.com>
3921 * tree-ssa-ifcombine.c (forwarder_block_to): New function.
3922 (tree_ssa_ifcombine_bb_1): New function.
3923 (tree_ssa_ifcombine_bb): Use it. Handle also cases where else_bb
3924 is an empty forwarder block to then_bb or vice versa and then_bb
3925 and else_bb are effectively swapped.
3927 2014-03-12 Christian Bruel <christian.bruel@st.com>
3930 * config/arm/arm.c (arm_emit_vfp_multi_reg_pop): Emit a
3931 REG_CFA_DEF_CFA note.
3932 (arm_expand_epilogue_apcs_frame): call arm_add_cfa_adjust_cfa_note.
3933 (arm_unwind_emit): Allow REG_CFA_DEF_CFA.
3935 2014-03-12 Thomas Preud'homme <thomas.preudhomme@arm.com>
3937 PR tree-optimization/60454
3938 * tree-ssa-math-opts.c (find_bswap_1): Fix bswap detection.
3940 2014-03-12 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3942 * config.gcc (aarch64*-*-*): Use ISA flags from aarch64-arches.def.
3943 Do not define target_cpu_default2 to generic.
3944 * config/aarch64/aarch64.h (TARGET_CPU_DEFAULT): Use generic cpu.
3945 * config/aarch64/aarch64.c (aarch64_override_options): Update comment.
3946 * config/aarch64/aarch64-arches.def (armv8-a): Use generic cpu.
3948 2014-03-12 Jakub Jelinek <jakub@redhat.com>
3949 Marc Glisse <marc.glisse@inria.fr>
3951 PR tree-optimization/60502
3952 * tree-ssa-reassoc.c (eliminate_not_pairs): Use build_all_ones_cst
3953 instead of build_low_bits_mask.
3955 2014-03-12 Jakub Jelinek <jakub@redhat.com>
3958 * tree-vrp.c (register_edge_assert_for_1): Don't add assert
3959 if there are multiple uses, but op doesn't live on E edge.
3960 * tree-cfg.c (assert_unreachable_fallthru_edge_p): Also ignore
3961 clobber stmts before __builtin_unreachable.
3963 2014-03-11 Richard Sandiford <rdsandiford@googlemail.com>
3965 * builtins.c (expand_builtin_setjmp_receiver): Use and clobber
3966 hard_frame_pointer_rtx.
3967 * cse.c (cse_insn): Remove volatile check.
3968 * cselib.c (cselib_process_insn): Likewise.
3969 * dse.c (scan_insn): Likewise.
3971 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
3973 * config/arc/arc.c (conditionalize_nonjump): New function,
3975 (arc_ifcvt): ... this.
3976 (arc_predicate_delay_insns): Use it.
3978 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
3980 * config/arc/predicates.md (extend_operand): During/after reload,
3981 allow const_int_operand.
3982 * config/arc/arc.md (mulsidi3_700): Use extend_operand predicate.
3983 (umulsidi3_700): Likewise. Change operand 2 constraint back to "cL".
3984 (mulsi3_highpart): Change operand 2 constraint alternatives 2 and 3
3986 (umulsi3_highpart_i): Likewise.
3988 2014-03-11 Richard Biener <rguenther@suse.de>
3990 * tree-ssa-structalias.c (get_constraint_for_ptr_offset):
3991 Add asserts to guard possible wrong-code bugs.
3993 2014-03-11 Richard Biener <rguenther@suse.de>
3995 PR tree-optimization/60429
3996 PR tree-optimization/60485
3997 * tree-ssa-structalias.c (set_union_with_increment): Properly
3998 take into account all fields that overlap the shifted vars.
3999 (do_sd_constraint): Likewise.
4000 (do_ds_constraint): Likewise.
4001 (get_constraint_for_ptr_offset): Likewise.
4003 2014-03-11 Chung-Lin Tang <cltang@codesourcery.com>
4005 * config/nios2/nios2.c (machine_function): Add fp_save_offset field.
4006 (nios2_compute_frame_layout):
4007 Add calculation of cfun->machine->fp_save_offset.
4008 (nios2_expand_prologue): Correct setting of frame pointer register
4010 (nios2_expand_epilogue): Update recovery of stack pointer from
4011 frame pointer accordingly.
4012 (nios2_initial_elimination_offset): Update calculation of offset
4013 for eliminating to HARD_FRAME_POINTER_REGNUM.
4015 2014-03-10 Jakub Jelinek <jakub@redhat.com>
4018 * ipa.c (symtab_remove_unreachable_nodes): Don't call
4019 cgraph_get_create_node on VAR_DECLs.
4021 2014-03-10 Richard Biener <rguenther@suse.de>
4024 * tree.c (signed_or_unsigned_type_for): Handle OFFSET_TYPEs.
4026 2014-03-08 Douglas B Rupp <rupp@gnat.com>
4028 * config/vms/vms.opt (vms_float_format): New variable.
4030 2014-03-08 Tobias Burnus <burnus@net-b.de>
4032 * doc/invoke.texi (-fcilkplus): Update implementation status.
4034 2014-03-08 Paulo Matos <paulo@matos-sorge.com>
4035 Richard Biener <rguenther@suse.de>
4037 * lto-wrapper.c (merge_and_complain): Ensure -fshort-double is used
4038 consistently accross all TUs.
4039 (run_gcc): Enable -fshort-double automatically at link at link-time
4040 and disallow override.
4042 2014-03-08 Richard Sandiford <rdsandiford@googlemail.com>
4045 * config/mips/mips.c (mips_option_override): Promote -mpaired-single
4046 warning to an error. Disable TARGET_PAIRED_SINGLE and TARGET_MIPS3D
4047 if they can't be used.
4049 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4051 * configure.ac (HAVE_AS_IX86_TLSLDMPLT): Improve test
4052 for Solaris 11/x86 ld.
4053 * configure: Regenerate.
4055 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4057 * configure.ac (TLS_SECTION_ASM_FLAG): Save as tls_section_flag.
4058 (LIB_TLS_SPEC): Save as ld_tls_libs.
4059 (HAVE_AS_IX86_TLSLDMPLT): Define as 1/0.
4060 (HAVE_AS_IX86_TLSLDM): New test.
4061 * configure, config.in: Regenerate.
4062 * config/i386/i386.c (legitimize_tls_address): Fall back to
4063 TLS_MODEL_GLOBAL_DYNAMIC on 32-bit Solaris/x86 if tool chain
4064 cannot support TLS_MODEL_LOCAL_DYNAMIC.
4065 * config/i386/i386.md (*tls_local_dynamic_base_32_gnu): Use if
4066 instead of #ifdef in HAVE_AS_IX86_TLSLDMPLT test.
4068 2014-03-07 Paulo Matos <paulo@matos-sorge.com>
4070 * common.opt (fira-loop-pressure): Mark as optimization.
4072 2014-03-07 Thomas Schwinge <thomas@codesourcery.com>
4074 * langhooks.c (lhd_omp_mappable_type): The error_mark_node is not
4075 an OpenMP mappable type.
4077 2014-03-06 Matthias Klose <doko@ubuntu.com>
4079 * Makefile.in (s-mlib): Only pass MULTIARCH_DIRNAME if
4080 MULTILIB_OSDIRNAMES is not defined.
4082 2014-03-06 Jakub Jelinek <jakub@redhat.com>
4083 Meador Inge <meadori@codesourcery.com>
4086 * config/arm/arm.c (arm_tls_symbol_p): Remove.
4087 (arm_legitimize_address): Call legitimize_tls_address for any
4088 arm_tls_referenced_p expression, handle constant addend. Call it
4089 before testing for !TARGET_ARM.
4090 (thumb_legitimize_address): Don't handle arm_tls_symbol_p here.
4092 2014-03-06 Richard Biener <rguenther@suse.de>
4098 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
4100 * tree-streamer.c (record_common_node): Assert we don't record
4101 nodes with type double.
4102 (preload_common_node): Skip type double, complex double and double
4103 pointer since it is now frontend dependent due to fshort-double option.
4105 2014-03-06 Richard Biener <rguenther@suse.de>
4107 * gcc.c (PLUGIN_COND): Always enable unless -fno-use-linker-plugin
4108 or -fno-lto is specified and the linker has full plugin support.
4109 * collect2.c (lto_mode): Default to LTO_MODE_WHOPR if LTO is enabled.
4110 (main): Remove -flto processing, adjust lto_mode using use_plugin late.
4111 * lto-wrapper.c (merge_and_complain): Merge compile-time
4112 optimization levels.
4113 (run_gcc): And pass it through to the link options.
4115 2014-03-06 Alexandre Oliva <aoliva@redhat.com>
4119 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4121 * cselib.c (remove_useless_values): Skip to avoid quadratic
4122 behavior if the condition moved from...
4123 (cselib_process_insn): ... here holds.
4125 2014-03-05 Jakub Jelinek <jakub@redhat.com>
4128 * Makefile.in (PLUGIN_HEADERS): Add tree-phinodes.h, stor-layout.h,
4129 ssa-iterators.h, $(RESOURCE_H) and tree-cfgcleanup.h.
4132 * config/i386/t-i386 (OPTIONS_H_EXTRA): Add stringop.def.
4133 (TM_H): Add x86-tune.def.
4135 2014-03-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4137 * config/aarch64/aarch64.c (generic_tunings):
4138 Use cortexa57_extra_costs.
4140 2014-03-05 Jakub Jelinek <jakub@redhat.com>
4143 * cfgexpand.c (expand_used_vars): Do not assume all SSA_NAMEs
4144 of PARM/RESULT_DECLs must be coalesced with optimize && in_lto_p.
4145 * tree-ssa-coalesce.c (coalesce_ssa_name): Use MUST_COALESCE_COST - 1
4148 2014-03-04 Heiher <r@hev.cc>
4150 * config/mips/mips-cpus.def (loongson3a): Mark as a MIPS64r2 processor.
4151 * config/mips/mips.h (MIPS_ISA_LEVEL_SPEC): Adjust accordingly.
4153 2014-03-04 Uros Bizjak <ubizjak@gmail.com>
4155 * config/i386/predicates.md (const2356_operand): Change to ...
4156 (const2367_operand): ... this.
4157 * config/i386/sse.md (avx512pf_scatterpf<mode>sf): Use
4159 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
4160 (*avx512pf_scatterpf<mode>sf): Ditto.
4161 (avx512pf_scatterpf<mode>df): Ditto.
4162 (*avx512pf_scatterpf<mode>df_mask): Ditto.
4163 (*avx512pf_scatterpf<mode>df): Ditto.
4164 * config/i386/i386.c (ix86_expand_builtin): Update
4165 incorrect hint operand error message.
4167 2014-03-04 Richard Biener <rguenther@suse.de>
4169 * lto-section-in.c (lto_get_section_data): Fix const cast.
4171 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
4173 * tree-streamer.c (record_common_node): Assert we don't record
4174 nodes with type double.
4175 (preload_common_node): Skip type double, complex double and double
4176 pointer since it is now frontend dependent due to fshort-double option.
4178 2014-03-04 Richard Biener <rguenther@suse.de>
4181 * lto-streamer-in.c (lto_read_body): Remove LTO bytecode version check.
4182 (lto_input_toplevel_asms): Likewise.
4183 * lto-section-in.c (lto_get_section_data): Instead do it here
4186 2014-03-04 Richard Biener <rguenther@suse.de>
4188 PR tree-optimization/60382
4189 * tree-vect-loop.c (vect_is_simple_reduction_1): Do not consider
4190 dead PHIs a reduction.
4192 2014-03-03 Uros Bizjak <ubizjak@gmail.com>
4194 * config/i386/xmmintrin.h (enum _mm_hint) <_MM_HINT_ET0>: Correct
4196 (_mm_prefetch): Move out of GCC target("sse") pragma.
4197 * config/i386/prfchwintrin.h (_m_prefetchw): Move out of
4198 GCC target("prfchw") pragma.
4199 * config/i386/i386.md (prefetch): Emit prefetchwt1 only
4201 * config/i386/i386.c (ix86_option_override_internal): Enable
4202 -mprfchw with -mprefetchwt1.
4204 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
4206 * config/arc/arc.md (casesi_load) <length attribute alternative 0>:
4209 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
4211 * opts.h (CL_PCH_IGNORE): Define.
4212 * targhooks.c (option_affects_pch_p):
4213 Return false for options that have CL_PCH_IGNORE set.
4214 * opt-functions.awk: Process PchIgnore.
4215 * doc/options.texi: Document PchIgnore.
4217 * config/arc/arc.opt (misize): Add PchIgnore property.
4219 2014-03-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4221 * config/rs6000/rs6000.c (rs6000_preferred_reload_class): Disallow
4222 reload of PLUS rtx's outside of GENERAL_REGS or BASE_REGS; relax
4223 constraint on constants to permit them being loaded into
4224 GENERAL_REGS or BASE_REGS.
4226 2014-03-03 Nick Clifton <nickc@redhat.com>
4228 * config/rl78/rl78-real.md (cbranchsi4_real_signed): Add
4229 anti-cacnonical alternatives.
4230 (negandhi3_real): New pattern.
4231 * config/rl78/rl78-virt.md (negandhi3_virt): New pattern.
4233 2014-03-03 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
4235 * config/avr/avr-mcus.def: Remove atxmega16x1.
4236 * config/avr/avr-tables.opt: Regenerate.
4237 * config/avr/t-multilib: Regenerate.
4238 * doc/avr-mmcu.texi: Regenerate.
4240 2014-03-03 Tobias Grosser <tobias@grosser.es>
4241 Mircea Namolaru <mircea.namolaru@inria.fr>
4243 PR tree-optimization/58028
4244 * graphite-clast-to-gimple.c (set_cloog_options): Don't remove
4247 2014-03-03 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4249 * config/arm/neon.md (*movmisalign<mode>): Legitimize addresses
4250 not handled by recognizers.
4252 2014-03-03 Jakub Jelinek <jakub@redhat.com>
4255 * function.c (expand_function_end): Don't emit
4256 clobber_return_register sequence if clobber_after is a BARRIER.
4257 * cfgexpand.c (construct_exit_block): Append instructions before
4258 return_label to prev_bb.
4260 2014-03-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4262 * config/rs6000/constraints.md: Document reserved use of "wc".
4264 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
4267 * ipa.c (function_and_variable_visibility): When dissolving comdat
4268 group, also set all symbols to local.
4270 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
4275 2013-12-14 Jan Hubicka <jh@suse.cz>
4277 * ipa-prop.c (stmt_may_be_vtbl_ptr_store): Skip clobbers.
4279 2014-03-02 Jon Beniston <jon@beniston.com>
4285 * config/lm32/lm32.c (lm32_legitimate_constant_p): Remove, as incorrect.
4286 (TARGET_LEGITIMATE_CONSTANT_P): Undefine, as not needed.
4287 * config/lm32/lm32.md (movsi_insn): Add 32-bit immediate support.
4288 (simple_return, *simple_return): New patterns
4289 * config/lm32/predicates.md (movsi_rhs_operand): Remove as obsolete.
4290 * configure.ac (force_sjlj_exceptions): Force sjlj exceptions for lm32.
4292 2014-03-01 Paolo Carlini <paolo.carlini@oracle.com>
4294 * dwarf2out.c (gen_subprogram_die): Tidy.
4296 2014-03-01 Oleg Endo <olegendo@gcc.gnu.org>
4299 * config/sh/sh.md (*mov_t_msb_neg): Split into ...
4300 (*mov_t_msb_neg_negc): ... this new insn.
4302 2014-02-28 Jason Merrill <jason@redhat.com>
4305 * ipa-devirt.c (ipa_devirt): Don't choose an implicitly-declared
4308 2014-02-28 Paolo Carlini <paolo.carlini@oracle.com>
4311 * dwarf2out.c (decltype_auto_die): New static.
4312 (gen_subprogram_die): Handle 'decltype(auto)' like 'auto'.
4313 (gen_type_die_with_usage): Handle 'decltype(auto)'.
4314 (is_cxx_auto): Likewise.
4316 2014-02-28 Ian Bolton <ian.bolton@arm.com>
4318 * config/aarch64/aarch64.h: Define __ARM_NEON by default if
4319 we are not using general regs only.
4321 2014-02-28 Richard Biener <rguenther@suse.de>
4324 * tree-cfgcleanup.c (tree_forwarder_block_p): Restrict
4325 previous fix and only allow to remove trivial pre-headers
4326 and latches. Also honor LOOPS_MAY_HAVE_MULTIPLE_LATCHES.
4327 (remove_forwarder_block): Properly update the latch of a loop.
4329 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4332 * cselib.c (cselib_hasher::equal): Special-case VALUE lookup.
4333 (cselib_preserved_hash_table): New.
4334 (preserve_constants_and_equivs): Move preserved vals to it.
4335 (cselib_find_slot): Look it up first.
4336 (cselib_init): Initialize it.
4337 (cselib_finish): Release it.
4338 (dump_cselib_table): Dump it.
4340 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4343 * cselib.c (remove_useless_values): Skip to avoid quadratic
4344 behavior if the condition moved from...
4345 (cselib_process_insn): ... here holds.
4347 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4350 * var-tracking.c (vt_initialize): Apply the same condition to
4351 preserve the CFA base value.
4353 2014-02-28 Joey Ye <joey.ye@arm.com>
4356 * config/arm/arm.c (thumb_far_jump_used_p): Don't change
4357 if reload in progress or completed.
4359 2014-02-28 Tobias Burnus <burnus@net-b.de>
4362 * tree-pretty-print.c (dump_generic_node, print_declaration): Handle
4365 2014-02-27 H.J. Lu <hongjiu.lu@intel.com>
4367 * doc/tm.texi.in (Condition Code Status): Update documention for
4368 relative locations of cc0-setter and cc0-user.
4370 2014-02-27 Jeff Law <law@redhat.com>
4372 PR rtl-optimization/52714
4373 * combine.c (try_combine): When splitting an unrecognized PARALLEL
4374 into two independent simple sets, if I3 is a jump, ensure the
4375 pattern we place into I3 is a (set (pc) ...).
4377 2014-02-27 Mikael Pettersson <mikpe@it.uu.se>
4378 Jeff Law <law@redhat.com>
4380 PR rtl-optimization/49847
4381 * cse.c (fold_rtx) Handle case where cc0 setter and cc0 user
4382 are in different blocks.
4383 * doc/tm.texi (Condition Code Status): Update documention for
4384 relative locations of cc0-setter and cc0-user.
4386 2014-02-27 Vladimir Makarov <vmakarov@redhat.com>
4389 * lra.c (lra_emit_add): Check SUBREG too.
4391 2014-02-27 Andreas Schwab <schwab@suse.de>
4393 * config/m68k/m68k.c (m68k_option_override): Disable
4394 -flive-range-shrinkage for classic m68k.
4395 (m68k_override_options_after_change): Likewise.
4397 2014-02-27 Marek Polacek <polacek@redhat.com>
4400 * tree-ssa-uninit.c (gate_warn_uninitialized): Run the pass even for
4401 -Wmaybe-uninitialized.
4403 2014-02-27 Alan Modra <amodra@gmail.com>
4406 * reload1.c (emit_input_reload_insns): When reload_override_in,
4407 set old to rl->in_reg when rl->in_reg is a subreg.
4409 2014-02-26 Richard Biener <rguenther@suse.de>
4412 * lra-assigns.c (spill_for): Avoid mixed-sign comparison.
4414 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
4416 * common/config/i386/predicates.md (const1256_operand): Remove.
4417 (const2356_operand): New.
4418 (const_1_to_2_operand): Remove.
4419 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
4420 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
4421 (*avx512pf_gatherpf<mode>sf): Ditto.
4422 (avx512pf_gatherpf<mode>df): Ditto.
4423 (*avx512pf_gatherpf<mode>df_mask): Ditto.
4424 (*avx512pf_gatherpf<mode>df): Ditto.
4425 (avx512pf_scatterpf<mode>sf): Ditto.
4426 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
4427 (*avx512pf_scatterpf<mode>sf): Ditto.
4428 (avx512pf_scatterpf<mode>df): Ditto.
4429 (*avx512pf_scatterpf<mode>df_mask): Ditto.
4430 (*avx512pf_scatterpf<mode>df): Ditto.
4431 * common/config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET0.
4433 2014-02-26 Ilya Tocar <ilya.tocar@intel.com>
4435 * config/i386/avx512fintrin.h (_mm512_testn_epi32_mask),
4436 (_mm512_mask_testn_epi32_mask), (_mm512_testn_epi64_mask),
4437 (_mm512_mask_testn_epi64_mask): Move to ...
4438 * config/i386/avx512cdintrin.h: Here.
4439 * config/i386/i386.c (bdesc_args): Change MASK_ISA for testnm.
4440 * config/i386/sse.md (avx512f_vmscalef<mode><round_name>): Remove %.
4441 (avx512f_scalef<mode><mask_name><round_name>): Ditto.
4442 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Change conditon to
4443 TARGET_AVX512F from TARGET_AVX512CD.
4445 2014-02-26 Richard Biener <rguenther@suse.de>
4448 * ipa.c (walk_polymorphic_call_targets): Properly guard
4449 call to inline_update_overall_summary.
4451 2014-02-26 Bin Cheng <bin.cheng@arm.com>
4454 * tree-cfgcleanup.c (tree_forwarder_block_p): Protect loop preheaders
4455 and latches only if requested. Fix latch if it is removed.
4456 * tree-ssa-dom.c (tree_ssa_dominator_optimize): Set
4457 LOOPS_HAVE_PREHEADERS.
4459 2014-02-25 Andrew Pinski <apinski@cavium.com>
4461 * builtins.c (expand_builtin_thread_pointer): Create a new target
4462 when the target is NULL.
4464 2014-02-25 Vladimir Makarov <vmakarov@redhat.com>
4466 PR rtl-optimization/60317
4467 * params.def (PARAM_LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
4468 * params.h (LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
4469 * lra-assigns.c: Include params.h.
4470 (spill_for): Use LRA_MAX_CONSIDERED_RELOAD_PSEUDOS as guard for
4471 other reload pseudos considerations.
4473 2014-02-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4475 * config/rs6000/vector.md (*vector_unordered<mode>): Change split
4476 to use canonical form for nor<mode>3.
4478 2014-02-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4481 * config/arm/arm.h (CANNOT_CHANGE_MODE_CLASS): Allow 128 to 64-bit
4484 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
4486 * common/config/i386/i386-common.c (OPTION_MASK_ISA_PREFETCHWT1_SET),
4487 (OPTION_MASK_ISA_PREFETCHWT1_UNSET): New.
4488 (ix86_handle_option): Handle OPT_mprefetchwt1.
4489 * config/i386/cpuid.h (bit_PREFETCHWT1): New.
4490 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
4492 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
4493 OPTION_MASK_ISA_PREFETCHWT1.
4494 * config/i386/i386.c (ix86_target_string): Handle mprefetchwt1.
4495 (PTA_PREFETCHWT1): New.
4496 (ix86_option_override_internal): Handle PTA_PREFETCHWT1.
4497 (ix86_valid_target_attribute_inner_p): Handle OPT_mprefetchwt1.
4498 * config/i386/i386.h (TARGET_PREFETCHWT1, TARGET_PREFETCHWT1_P): New.
4499 * config/i386/i386.md (prefetch): Check TARGET_PREFETCHWT1
4500 (*prefetch_avx512pf_<mode>_: Change into ...
4501 (*prefetch_prefetchwt1_<mode>: This.
4502 * config/i386/i386.opt (mprefetchwt1): New.
4503 * config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET1.
4504 (_mm_prefetch): Handle intent to write.
4505 * doc/invoke.texi (mprefetchwt1), (mno-prefetchwt1): Doccument.
4507 2014-02-25 Richard Biener <rguenther@suse.de>
4510 * emit-rtl.c (mem_attrs_htab): Remove.
4511 (mem_attrs_htab_hash): Likewise.
4512 (mem_attrs_htab_eq): Likewise.
4513 (set_mem_attrs): Always allocate new mem-attrs when something changed.
4514 (init_emit_once): Do not allocate mem_attrs_htab.
4516 2014-02-25 Richard Biener <rguenther@suse.de>
4519 * lto-opts.c (lto_write_options): Output non-explicit conservative
4520 -fwrapv, -fno-trapv and -fno-strict-overflow.
4521 * lto-wrapper.c (merge_and_complain): Handle merging those options.
4522 (run_gcc): And pass them through.
4524 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
4526 * sel-sched.c (calculate_new_fences): New parameter ptime.
4527 Calculate it as a maximum over all fence cycles.
4528 (sel_sched_region_2): Adjust the call to calculate_new_fences.
4529 Print the final schedule timing when sched_verbose.
4531 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
4533 PR rtl-optimization/60292
4534 * sel-sched.c (fill_vec_av_set): Do not reset target availability
4535 bit fot the fence instruction.
4537 2014-02-24 Alangi Derick <alangiderick@gmail.com>
4539 * calls.h: Fix typo in comment.
4541 2014-02-24 John David Anglin <danglin@gcc.gnu.org>
4543 * config/pa/pa.c (pa_output_move_double): Don't valididate when
4544 adjusting offsetable addresses.
4546 2014-02-24 Guozhi Wei <carrot@google.com>
4548 * sparseset.h (sparseset_pop): Fix the wrong index.
4550 2014-02-24 Walter Lee <walt@tilera.com>
4552 * config.gcc (tilepro-*-*): Change to tilepro*-*-*.
4553 (tilegx-*-linux*): Change to tilegx*-*-linux*; Support tilegxbe
4555 * common/config/tilegx/tilegx-common.c
4556 (TARGET_DEFAULT_TARGET_FLAGS): Define.
4557 * config/tilegx/linux.h (ASM_SPEC): Add endian_spec.
4559 * config/tilegx/sync.md (atomic_test_and_set): Handle big endian.
4560 * config/tilegx/tilegx.c (tilegx_return_in_msb): New.
4561 (tilegx_gimplify_va_arg_expr): Handle big endian.
4562 (tilegx_expand_unaligned_load): Ditto.
4563 (tilegx_expand_unaligned_store): Ditto.
4564 (TARGET_RETURN_IN_MSB): New.
4565 * config/tilegx/tilegx.h (TARGET_DEFAULT): New.
4566 (TARGET_ENDIAN_DEFAULT): New.
4567 (TARGET_BIG_ENDIAN): Handle big endian.
4568 (BYTES_BIG_ENDIAN): Ditto.
4569 (WORDS_BIG_ENDIAN): Ditto.
4570 (FLOAT_WORDS_BIG_ENDIAN): Ditto.
4573 * config/tilegx/tilegx.md (extv): Handle big endian.
4575 (insn_st<n>): Ditto.
4576 (insn_st<n>_add<bitsuffix>): Ditto.
4577 (insn_stnt<n>): Ditto.
4578 (insn_stnt<n>_add<bitsuffix>):Ditto.
4579 (vec_interleave_highv8qi): Handle big endian.
4580 (vec_interleave_highv8qi_be): New.
4581 (vec_interleave_highv8qi_le): New.
4582 (insn_v1int_h): Handle big endian.
4583 (vec_interleave_lowv8qi): Handle big endian.
4584 (vec_interleave_lowv8qi_be): New.
4585 (vec_interleave_lowv8qi_le): New.
4586 (insn_v1int_l): Handle big endian.
4587 (vec_interleave_highv4hi): Handle big endian.
4588 (vec_interleave_highv4hi_be): New.
4589 (vec_interleave_highv4hi_le): New.
4590 (insn_v2int_h): Handle big endian.
4591 (vec_interleave_lowv4hi): Handle big endian.
4592 (vec_interleave_lowv4hi_be): New.
4593 (vec_interleave_lowv4hi_le): New.
4594 (insn_v2int_l): Handle big endian.
4595 (vec_interleave_highv2si): Handle big endian.
4596 (vec_interleave_highv2si_be): New.
4597 (vec_interleave_highv2si_le): New.
4598 (insn_v4int_h): Handle big endian.
4599 (vec_interleave_lowv2si): Handle big endian.
4600 (vec_interleave_lowv2si_be): New.
4601 (vec_interleave_lowv2si_le): New.
4602 (insn_v4int_l): Handle big endian.
4603 * config/tilegx/tilegx.opt (mbig-endian): New option.
4604 (mlittle-endian): New option.
4605 * doc/install.texi: Document tilegxbe-linux.
4606 * doc/invoke.texi: Document -mbig-endian and -mlittle-endian.
4608 2014-02-24 Martin Jambor <mjambor@suse.cz>
4611 * ipa-cp.c (propagate_constants_accross_call): Bail out early if
4612 there are no parameter descriptors.
4614 2014-02-24 Andrey Belevantsev <abel@ispras.ru>
4616 PR rtl-optimization/60268
4617 * sched-rgn.c (haifa_find_rgns): Move the nr_regions_initial variable
4618 initialization to ...
4619 (sched_rgn_init): ... here.
4620 (schedule_region): Check for SCHED_PRESSURE_NONE earlier.
4622 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
4624 * config/microblaze/microblaze.md: Correct ashrsi_reg / lshrsi_reg
4627 2014-02-23 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
4629 * config/microblaze/microblaze.h: Remove SECONDARY_MEMORY_NEEDED
4632 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
4634 * /config/microblaze/microblaze.c: Add microblaze_asm_output_mi_thunk,
4635 define TARGET_ASM_OUTPUT_MI_THUNK and TARGET_ASM_CAN_OUTPUT_MI_THUNK.
4637 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
4639 * config/microblaze/predicates.md: Add cmp_op predicate.
4640 * config/microblaze/microblaze.md: Add branch_compare instruction
4641 which uses cmp_op predicate and emits cmp insn before branch.
4642 * config/microblaze/microblaze.c (microblaze_emit_compare): Rename
4643 to microblaze_expand_conditional_branch and consolidate logic.
4644 (microblaze_expand_conditional_branch): emit branch_compare
4645 insn instead of handling cmp op separate from branch insn.
4647 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4649 * config/rs6000/rs6000.c (rs6000_emit_le_vsx_move): Relax assert
4652 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4654 * config/rs6000/altivec.md (altivec_lve<VI_char>x): Replace
4655 define_insn with define_expand and new define_insn
4656 *altivec_lve<VI_char>x_internal.
4657 (altivec_stve<VI_char>x): Replace define_insn with define_expand
4658 and new define_insn *altivec_stve<VI_char>x_internal.
4659 * config/rs6000/rs6000-protos.h (altivec_expand_stvex_be): New
4661 * config/rs6000/rs6000.c (altivec_expand_lvx_be): Document use by
4663 (altivec_expand_stvex_be): New function.
4665 2014-02-22 Joern Rennecke <joern.rennecke@embecosm.com>
4667 * config/avr/avr.c (avr_can_eliminate): Allow elimination from
4668 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM if !frame_pointer_needed.
4669 * config/avr/avr.c (ELIMINABLE_REGS): Add elimination from
4670 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM.
4672 2014-02-21 Vladimir Makarov <vmakarov@redhat.com>
4675 * lra-constraints.c (inherit_reload_reg): Use lra_emit_move
4676 instead of emit_move_insn.
4678 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4680 * config/rs6000/altivec.md (altivec_vsumsws): Replace second
4682 (reduc_uplus_v16qi): Use gen_altivec_vsumsws_direct instead of
4683 gen_altivec_vsumsws.
4685 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4687 * config/rs6000/altivec.md (altivec_lvxl): Rename as
4688 *altivec_lvxl_<mode>_internal and use VM2 iterator instead of V4SI.
4689 (altivec_lvxl_<mode>): New define_expand incorporating
4690 -maltivec=be semantics where needed.
4691 (altivec_lvx): Rename as *altivec_lvx_<mode>_internal.
4692 (altivec_lvx_<mode>): New define_expand incorporating -maltivec=be
4693 semantics where needed.
4694 (altivec_stvx): Rename as *altivec_stvx_<mode>_internal.
4695 (altivec_stvx_<mode>): New define_expand incorporating
4696 -maltivec=be semantics where needed.
4697 (altivec_stvxl): Rename as *altivec_stvxl_<mode>_internal and use
4698 VM2 iterator instead of V4SI.
4699 (altivec_stvxl_<mode>): New define_expand incorporating
4700 -maltivec=be semantics where needed.
4701 * config/rs6000/rs6000-builtin.def: Add new built-in definitions
4702 LVXL_V2DF, LVXL_V2DI, LVXL_V4SF, LVXL_V4SI, LVXL_V8HI, LVXL_V16QI,
4703 LVX_V2DF, LVX_V2DI, LVX_V4SF, LVX_V4SI, LVX_V8HI, LVX_V16QI, STVX_V2DF,
4704 STVX_V2DI, STVX_V4SF, STVX_V4SI, STVX_V8HI, STVX_V16QI, STVXL_V2DF,
4705 STVXL_V2DI, STVXL_V4SF, STVXL_V4SI, STVXL_V8HI, STVXL_V16QI.
4706 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Replace
4707 ALTIVEC_BUILTIN_LVX with ALTIVEC_BUILTIN_LVX_<MODE> throughout;
4708 similarly for ALTIVEC_BUILTIN_LVXL, ALTIVEC_BUILTIN_STVX, and
4709 ALTIVEC_BUILTIN_STVXL.
4710 * config/rs6000/rs6000-protos.h (altivec_expand_lvx_be): New prototype.
4711 (altivec_expand_stvx_be): Likewise.
4712 * config/rs6000/rs6000.c (swap_selector_for_mode): New function.
4713 (altivec_expand_lvx_be): Likewise.
4714 (altivec_expand_stvx_be): Likewise.
4715 (altivec_expand_builtin): Add cases for
4716 ALTIVEC_BUILTIN_STVX_<MODE>, ALTIVEC_BUILTIN_STVXL_<MODE>,
4717 ALTIVEC_BUILTIN_LVXL_<MODE>, and ALTIVEC_BUILTIN_LVX_<MODE>.
4718 (altivec_init_builtins): Add definitions for
4719 __builtin_altivec_lvxl_<mode>, __builtin_altivec_lvx_<mode>,
4720 __builtin_altivec_stvx_<mode>, and __builtin_altivec_stvxl_<mode>.
4722 2014-02-21 Catherine Moore <clm@codesourcery.com>
4724 * doc/invoke.texi (mvirt, mno-virt): Document.
4725 * config/mips/mips.opt (mvirt): New option.
4726 * config/mips/mips.h (ASM_SPEC): Pass mvirt to the assembler.
4728 2014-02-21 Richard Biener <rguenther@suse.de>
4730 PR tree-optimization/60276
4731 * tree-vectorizer.h (struct _stmt_vec_info): Add min_neg_dist field.
4732 (STMT_VINFO_MIN_NEG_DIST): New macro.
4733 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Record
4734 STMT_VINFO_MIN_NEG_DIST.
4735 * tree-vect-stmts.c (vectorizable_load): Verify if assumptions
4736 made for negative dependence distances still hold.
4738 2014-02-21 Richard Biener <rguenther@suse.de>
4741 * tree-ssa-live.c (mark_all_vars_used_1): Do not walk
4742 DECL_INITIAL for globals not in the current function context.
4744 2014-02-21 Jakub Jelinek <jakub@redhat.com>
4746 PR tree-optimization/56490
4747 * params.def (PARAM_UNINIT_CONTROL_DEP_ATTEMPTS): New param.
4748 * tree-ssa-uninit.c: Include params.h.
4749 (compute_control_dep_chain): Add num_calls argument, return false
4750 if it exceed PARAM_UNINIT_CONTROL_DEP_ATTEMPTS param, pass
4751 num_calls to recursive call.
4752 (find_predicates): Change dep_chain into normal array,
4753 cur_chain into auto_vec<edge, MAX_CHAIN_LEN + 1>, add num_calls
4754 variable and adjust compute_control_dep_chain caller.
4755 (find_def_preds): Likewise.
4757 2014-02-21 Thomas Schwinge <thomas@codesourcery.com>
4759 * gimple-pretty-print.c (dump_gimple_omp_for) [flags & TDF_RAW]
4760 <case GF_OMP_FOR_KIND_CILKSIMD>: Add missing break statement.
4762 2014-02-21 Nick Clifton <nickc@redhat.com>
4764 * config/stormy16/stormy16.md (pushdqi1): Add mode to post_inc.
4765 (pushhi1): Likewise.
4766 (popqi1): Add mode to pre_dec.
4769 2014-02-21 Jakub Jelinek <jakub@redhat.com>
4771 * config/i386/i386.c (ix86_expand_vec_perm): Use V8SImode
4772 mode for mask of V8SFmode permutation.
4774 2014-02-20 Richard Henderson <rth@redhat.com>
4777 * builtins.c (expand_builtin_atomic_compare_exchange): Always make
4778 a new pseudo for OLDVAL.
4780 2014-02-20 Jakub Jelinek <jakub@redhat.com>
4783 * config/i386/i386.c (expand_vec_perm_interleave2): Don't call
4784 gen_reg_rtx if d->testing_p.
4785 (expand_vec_perm_pshufb2, expand_vec_perm_broadcast_1): Return early
4786 if d->testing_p and we will certainly return true.
4787 (expand_vec_perm_even_odd_1): Likewise. Don't call gen_reg_rtx
4790 2014-02-20 Uros Bizjak <ubizjak@gmail.com>
4792 * emit-rtl.c (gen_reg_rtx): Assert that
4793 crtl->emit.regno_pointer_align_length is non-zero.
4795 2014-02-20 Richard Henderson <rth@redhat.com>
4798 * builtins.c (expand_builtin_atomic_compare_exchange): Conditionalize
4799 on failure the store back into EXPECT.
4801 2014-02-20 Chung-Lin Tang <cltang@codesourcery.com>
4802 Sandra Loosemore <sandra@codesourcery.com>
4804 * config/nios2/nios2.md (unspec): Add UNSPEC_PIC_GOTOFF_SYM enum.
4805 * config/nios2/nios2.c (nios2_function_profiler): Add
4806 -fPIC (flag_pic == 2) support.
4807 (nios2_handle_custom_fpu_cfg): Fix warning parameter.
4808 (nios2_large_offset_p): New function.
4809 (nios2_unspec_reloc_p): Move up position, update to use
4810 nios2_large_offset_p.
4811 (nios2_unspec_address): Remove function.
4812 (nios2_unspec_offset): New function.
4813 (nios2_large_got_address): New function.
4814 (nios2_got_address): Add large offset support.
4815 (nios2_legitimize_tls_address): Update usage of removed and new
4817 (nios2_symbol_binds_local_p): New function.
4818 (nios2_load_pic_address): Add -fPIC (flag_pic == 2) support.
4819 (nios2_legitimize_address): Update to use nios2_large_offset_p.
4820 (nios2_emit_move_sequence): Avoid legitimizing (const (unspec ...)).
4821 (nios2_print_operand): Merge H/L processing, add hiadj/lo
4822 processing for (const (unspec ...)).
4823 (nios2_unspec_reloc_name): Add UNSPEC_PIC_GOTOFF_SYM case.
4825 2014-02-20 Richard Biener <rguenther@suse.de>
4827 * tree-cfg.c (replace_uses_by): Mark altered BBs before
4828 doing the substitution.
4829 (verify_gimple_assign_single): Also verify bare MEM_REFs on the lhs.
4831 2014-02-20 Martin Jambor <mjambor@suse.cz>
4834 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Uce correct
4835 info when checking whether lattices are bottom.
4837 2014-02-20 Richard Biener <rguenther@suse.de>
4840 * tree-eh.c (execute_cleanup_eh_1): Also cleanup empty EH
4843 2014-02-20 Jan Hubicka <hubicka@ucw.cz>
4846 * ipa-inline-transform.c (clone_inlined_nodes): Add freq_scale
4847 parameter specifying the scaling.
4848 (inline_call): Update.
4849 (want_inline_recursively): Guard division by zero.
4850 (recursive_inlining): Update.
4851 * ipa-inline.h (clone_inlined_nodes): Update.
4853 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
4856 * config/i386/i386.c (classify_argument): Pass structures of size
4857 64 bytes or less in register.
4859 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
4860 Kirill Yukhin <kirill.yukhin@intel.com>
4862 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): Swap operands.
4863 (_mm_rcp28_round_ss): Ditto.
4864 (_mm_rsqrt28_round_sd): Ditto.
4865 (_mm_rsqrt28_round_ss): Ditto.
4866 * config/i386/avx512erintrin.h (_mm_rcp14_round_sd): Ditto.
4867 (_mm_rcp14_round_ss): Ditto.
4868 (_mm_rsqrt14_round_sd): Ditto.
4869 (_mm_rsqrt14_round_ss): Ditto.
4870 * config/i386/sse.md (rsqrt14<mode>): Put nonimmediate operand as
4871 the first input operand, get rid of match_dup.
4872 (avx512er_exp2<mode><mask_name><round_saeonly_name>): Set type
4874 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
4876 (avx512er_vmrcp28<mode><round_saeonly_name>): Put nonimmediate
4877 operand as the first input operand, set type attribute.
4878 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
4880 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Put nonimmediate
4881 operand as the first input operand, set type attribute.
4883 2014-02-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4885 * config/rs6000/rs6000.c (vspltis_constant): Fix most significant
4888 2014-02-19 H.J. Lu <hongjiu.lu@intel.com>
4891 * config/i386/i386.c (construct_container): Remove TFmode check
4892 for X86_64_INTEGER_CLASS.
4894 2014-02-19 Uros Bizjak <ubizjak@gmail.com>
4897 * config/i386/i386.c (type_natural_mode): Warn for ABI changes
4898 only when -Wpsabi is enabled.
4900 2014-02-19 Michael Hudson-Doyle <michael.hudson@linaro.org>
4903 * config/aarch64/aarch64.c (aarch64_pass_by_reference): The rules for
4904 passing arrays in registers are the same as for structs, so remove the
4905 special case for them.
4907 2014-02-19 Eric Botcazou <ebotcazou@adacore.com>
4909 * expr.c (expand_expr_real_1) <case VIEW_CONVERT_EXPR>: For a bit-field
4910 destination type, extract only the valid bits if the source type is not
4911 integral and has a different mode.
4913 2014-02-19 Richard Biener <rguenther@suse.de>
4916 * tree-inline.c (estimate_num_insns): Avoid calling cgraph_get_node
4919 2014-02-19 Richard Biener <rguenther@suse.de>
4922 * ipa-prop.c: Include stringpool.h and tree-ssanames.h.
4923 (ipa_modify_call_arguments): Emit an argument load explicitely and
4924 preserve virtual SSA form there and for the replacement call.
4925 Do not update SSA form nor free dominance info.
4927 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
4929 * ipa.c (function_and_variable_visibility): Also clear WEAK
4930 flag when disolving COMDAT_GROUP.
4932 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
4934 * ipa-prop.h (ipa_ancestor_jf_data): Update ocmment.
4935 * ipa-prop.c (ipa_set_jf_known_type): Return early when
4937 (ipa_set_ancestor_jf): Set type to NULL hwen it is not preserved;
4938 do more sanity checks.
4939 (detect_type_change): Return true when giving up early.
4940 (compute_complex_assign_jump_func): Fix type parameter of
4941 ipa_set_ancestor_jf.
4942 (compute_complex_ancestor_jump_func): Likewise.
4943 (update_jump_functions_after_inlining): Fix updating of
4945 * ipa-cp.c (ipa_get_jf_ancestor_result): Be ready for type to be NULL.
4947 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
4949 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Also remove
4950 inline clones when edge disappears.
4952 2014-02-18 Michael Meissner <meissner@linux.vnet.ibm.com>
4955 * config/rs6000/rs6000.md (mov<mode>_64bit, TF/TDmode moves):
4956 Split 64-bit moves into 2 patterns. Do not allow the use of
4957 direct move for TDmode in little endian, since the decimal value
4958 has little endian bytes within a word, but the 64-bit pieces are
4959 ordered in a big endian fashion, and normal subreg's of TDmode are
4961 (mov<mode>_64bit_dm): Likewise.
4962 (movtd_64bit_nodm): Likewise.
4964 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
4966 PR tree-optimization/60174
4967 * tree-ssa-reassoc.c (init_range_entry): Do not look into the defining
4968 statement of an SSA_NAME that occurs in an abnormal PHI node.
4970 2014-02-18 Jakub Jelinek <jakub@redhat.com>
4973 * final.c (SEEN_BB): Remove.
4974 (SEEN_NOTE, SEEN_EMITTED): Renumber.
4975 (final_scan_insn): Don't force_source_line on second
4976 NOTE_INSN_BASIC_BLOCK.
4978 2014-02-18 Uros Bizjak <ubizjak@gmail.com>
4981 * config/i386/i386.h (struct ix86_args): Add warn_avx512f.
4982 * config/i386/i386.c (init_cumulative_args): Initialize warn_avx512f.
4983 (type_natural_mode): Warn ABI change when %zmm register is not
4984 available for AVX512F vector value passing.
4986 2014-02-18 Kai Tietz <ktietz@redhat.com>
4989 * config/i386/i386.c (ix86_expand_prologue): Use value in
4990 rax register as displacement when restoring %r10 or %rax.
4991 Fix wrong offset when restoring both registers.
4993 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
4995 * ipa-prop.c (compute_complex_ancestor_jump_func): Replace overzealous
4996 assertion with conditional return.
4998 2014-02-18 Jakub Jelinek <jakub@redhat.com>
4999 Uros Bizjak <ubizjak@gmail.com>
5002 * config/i386/driver-i386.c (host_detect_local_cpu): If
5003 YMM state is not saved by the OS, also clear has_f16c. Move
5004 CPUID 0x80000001 handling before YMM state saving checking.
5006 2014-02-18 Andrey Belevantsev <abel@ispras.ru>
5008 PR rtl-optimization/58960
5009 * haifa-sched.c (alloc_global_sched_pressure_data): New,
5010 factored out from ...
5011 (sched_init): ... here.
5012 (free_global_sched_pressure_data): New, factored out from ...
5013 (sched_finish): ... here.
5014 * sched-int.h (free_global_sched_pressure_data): Declare.
5015 * sched-rgn.c (nr_regions_initial): New static global.
5016 (haifa_find_rgns): Initialize it.
5017 (schedule_region): Disable sched-pressure for the newly
5020 2014-02-17 Richard Biener <rguenther@suse.de>
5022 * tree-vect-stmts.c (free_stmt_vec_info): Clear BB and
5023 release SSA defs of pattern stmts.
5025 2014-02-17 Richard Biener <rguenther@suse.de>
5027 * tree-inline.c (expand_call_inline): Release the virtual
5028 operand defined by the call we are about to inline.
5030 2014-02-17 Richard Biener <rguenther@suse.de>
5032 * tree-ssa.c (verify_ssa): If verify_def found an error, ICE.
5034 2014-02-17 Kirill Yukhin <kirill.yukhin@intel.com>
5035 Ilya Tocar <ilya.tocar@intel.com>
5037 * config/i386/avx512fintrin.h (_mm512_maskz_permutexvar_epi64): Swap
5038 arguments order in builtin.
5039 (_mm512_permutexvar_epi64): Ditto.
5040 (_mm512_mask_permutexvar_epi64): Ditto
5041 (_mm512_maskz_permutexvar_epi32): Ditto
5042 (_mm512_permutexvar_epi32): Ditto
5043 (_mm512_mask_permutexvar_epi32): Ditto
5045 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5047 * config/rs6000/altivec.md (p8_vmrgew): Handle little endian targets.
5048 (p8_vmrgow): Likewise.
5050 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5052 * config/rs6000/vsx.md (vsx_xxpermdi_<mode>): Handle little
5055 2014-02-15 Michael Meissner <meissner@linux.vnet.ibm.com>
5058 * config/rs6000/rs6000.md (rreg): Add TFmode, TDmode constraints.
5059 (mov<mode>_internal, TFmode/TDmode): Split TFmode/TDmode moves
5060 into 64-bit and 32-bit moves. On 64-bit moves, add support for
5061 using direct move instructions on ISA 2.07. Also adjust
5062 instruction length for 64-bit.
5063 (mov<mode>_64bit, TFmode/TDmode): Likewise.
5064 (mov<mode>_32bit, TFmode/TDmode): Likewise.
5066 2014-02-15 Alan Modra <amodra@gmail.com>
5070 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
5071 find_replacement on parts of insn rtl that might be reloaded.
5073 2014-02-15 Richard Biener <rguenther@suse.de>
5075 PR tree-optimization/60183
5076 * tree-ssa-phiprop.c (propagate_with_phi): Avoid speculating loads.
5077 (tree_ssa_phiprop): Calculate and free post-dominators.
5079 2014-02-14 Jeff Law <law@redhat.com>
5081 PR rtl-optimization/60131
5082 * ree.c (get_extended_src_reg): New function.
5083 (combine_reaching_defs): Use it rather than assuming location of REG.
5084 (find_and_remove_re): Verify first operand of extension is
5085 a REG before adding the insns to the copy list.
5087 2014-02-14 Roland McGrath <mcgrathr@google.com>
5089 * configure.ac (HAVE_AS_IX86_UD2): New test for 'ud2' mnemonic.
5090 * configure: Regenerated.
5091 * config.in: Regenerated.
5092 * config/i386/i386.md (trap) [HAVE_AS_IX86_UD2]: Use the mnemonic
5093 instead of ASM_SHORT.
5095 2014-02-14 Vladimir Makarov <vmakarov@redhat.com>
5096 Richard Earnshaw <rearnsha@arm.com>
5098 PR rtl-optimization/59535
5099 * lra-constraints.c (process_alt_operands): Encourage alternative
5100 when unassigned pseudo class is superset of the alternative class.
5101 (inherit_reload_reg): Don't inherit when optimizing for code size.
5102 * config/arm/arm.h (MODE_BASE_REG_CLASS): Add version for LRA
5103 returning CORE_REGS for anything but Thumb1 and BASE_REGS for
5104 modes not less than 4 for Thumb1.
5106 2014-02-14 Kyle McMartin <kyle@redhat.com>
5109 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Define for AArch64.
5111 2014-02-14 Richard Biener <rguenther@suse.de>
5113 * cilk-common.c (cilk_arrow): Build a MEM_REF, not an INDIRECT_REF.
5114 (get_frame_arg): Drop the assert with langhook types_compatible_p.
5115 Do not strip INDIRECT_REFs.
5117 2014-02-14 Richard Biener <rguenther@suse.de>
5120 * lto-streamer-out.c (DFS_write_tree_body): Do not follow
5121 DECL_FUNCTION_SPECIFIC_TARGET.
5122 (hash_tree): Do not hash DECL_FUNCTION_SPECIFIC_TARGET.
5123 * tree-streamer-out.c (pack_ts_target_option): Remove.
5124 (streamer_pack_tree_bitfields): Do not stream TS_TARGET_OPTION.
5125 (write_ts_function_decl_tree_pointers): Do not stream
5126 DECL_FUNCTION_SPECIFIC_TARGET.
5127 * tree-streamer-in.c (unpack_ts_target_option): Remove.
5128 (unpack_value_fields): Do not stream TS_TARGET_OPTION.
5129 (lto_input_ts_function_decl_tree_pointers): Do not stream
5130 DECL_FUNCTION_SPECIFIC_TARGET.
5132 2014-02-14 Jakub Jelinek <jakub@redhat.com>
5134 * tree-vect-loop.c (vect_is_slp_reduction): Don't set use_stmt twice.
5135 (get_initial_def_for_induction, vectorizable_induction): Ignore
5136 debug stmts when looking for exit_phi.
5137 (vectorizable_live_operation): Fix up condition.
5139 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
5141 * config/nds32/nds32.c (nds32_asm_function_prologue): Do not use
5142 nreverse() because it changes the content of original tree list.
5144 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
5146 * config/nds32/t-mlibs (MULTILIB_OPTIONS): Fix typo in comment.
5147 * config/nds32/nds32.c (nds32_merge_decl_attributes): Likewise.
5149 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
5151 * config/nds32/nds32.c (nds32_naked_function_p): Follow the
5152 GNU coding standards.
5154 2014-02-13 Jakub Jelinek <jakub@redhat.com>
5157 * dwarf2out.c (gen_subprogram_die): Don't call
5158 add_calling_convention_attribute if subr_die is old_die.
5160 2014-02-13 Sharad Singhai <singhai@google.com>
5162 * doc/optinfo.texi: Fix order of nodes.
5164 2014-02-13 Uros Bizjak <ubizjak@gmail.com>
5166 * config/i386/sse.md (xop_vmfrcz<mode>2): Generate const0 in
5167 operands[2], not operands[3].
5169 2014-02-13 Richard Biener <rguenther@suse.de>
5172 * doc/install.texi (ISL): Update recommended version to 0.12.2,
5173 mention the possibility of an in-tree build.
5174 (CLooG): Update recommended version to 0.18.1, mention the
5175 possibility of an in-tree build and clarify that the ISL
5176 bundled with CLooG does not work.
5178 2014-02-13 Jakub Jelinek <jakub@redhat.com>
5181 * expr.c (compress_float_constant): If x is a hard register,
5182 extend into a pseudo and then move to x.
5184 2014-02-13 Dominik Vogt <vogt@linux.vnet.ibm.com>
5186 * config/s390/s390.c (s390_asm_output_function_label): Fix crash
5187 caused by bad second argument to warning_at() with -mhotpatch and
5188 nested functions (e.g. with gfortran).
5190 2014-02-13 Richard Sandiford <rdsandiford@googlemail.com>
5192 * opts.c (option_name): Remove "enabled by default" rider.
5194 2014-02-12 John David Anglin <danglin@gcc.gnu.org>
5196 * config/pa/pa.c (pa_option_override): Remove auto increment FIXME.
5198 2014-02-12 H.J. Lu <hongjiu.lu@intel.com>
5199 Uros Bizjak <ubizjak@gmail.com>
5202 * configure.ac (HAVE_AS_GOTOFF_IN_DATA): Pass --32 to GNU assembler.
5203 * configure: Regenerated.
5205 2014-02-12 Richard Biener <rguenther@suse.de>
5207 * vec.c (vec_prefix::calculate_allocation): Move as
5208 inline variant to vec.h.
5209 (vec_prefix::calculate_allocation_1): New out-of-line version.
5210 * vec.h (vec_prefix::calculate_allocation_1): Declare.
5211 (vec_prefix::m_has_auto_buf): Rename to ...
5212 (vec_prefix::m_using_auto_storage): ... this.
5213 (vec_prefix::calculate_allocation): Inline the easy cases
5214 and dispatch to calculate_allocation_1 which doesn't need the
5216 (va_heap::reserve): Use gcc_checking_assert.
5217 (vec<T, A, vl_embed>::embedded_init): Add argument to initialize
5218 m_using_auto_storage.
5219 (auto_vec): Change m_vecpfx member to a vec<T, va_heap, vl_embed>
5221 (vec<T, va_heap, vl_ptr>::reserve): Remove redundant check.
5222 (vec<T, va_heap, vl_ptr>::release): Avoid casting.
5223 (vec<T, va_heap, vl_ptr>::using_auto_storage): Simplify.
5225 2014-02-12 Richard Biener <rguenther@suse.de>
5227 * gcse.c (compute_transp): break from loop over canon_modify_mem_list
5228 when we found a dependence.
5230 2014-02-12 Thomas Schwinge <thomas@codesourcery.com>
5232 * gimplify.c (gimplify_call_expr, gimplify_modify_expr): Move
5234 (maybe_fold_stmt): ... into this new function.
5235 * omp-low.c (lower_omp): Update comment.
5237 * omp-low.c (lower_omp_target): Add clobber for sizes array, after
5240 * omp-low.c (diagnose_sb_0): Make sure label_ctx is valid to
5243 2014-02-12 James Greenhalgh <james.greenhalgh@arm.com>
5245 * config/arm/aarch-cost-tables.h (generic_extra_costs): Fix
5246 identifiers in comments.
5247 (cortexa53_extra_costs): Likewise.
5248 * config/arm/arm.c (cortexa9_extra_costs): Fix identifiers in comments.
5249 (cortexa7_extra_costs): Likewise.
5250 (cortexa12_extra_costs): Likewise.
5251 (cortexa15_extra_costs): Likewise.
5252 (v7m_extra_costs): Likewise.
5254 2014-02-12 Richard Biener <rguenther@suse.de>
5257 * gimple-low.c (lower_builtin_posix_memalign): Lower conditional
5258 of posix_memalign being successful.
5259 (lower_stmt): Restrict lowering of posix_memalign to when
5260 -ftree-bit-ccp is enabled.
5262 2014-02-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
5264 * config/avr/avr-c.c (avr_resolve_overloaded_builtin): Pass vNULL for
5266 * config/spu/spu-c.c (spu_resolve_overloaded_builtin): Likewise.
5268 2014-02-12 Eric Botcazou <ebotcazou@adacore.com>
5270 PR rtl-optimization/60116
5271 * combine.c (try_combine): Also remove dangling REG_DEAD notes on the
5272 other_insn once the combination has been validated.
5274 2014-02-11 Jan Hubicka <hubicka@ucw.cz>
5277 * ipa-utils.h (possible_polymorphic_call_targets): Update prototype
5279 * ipa-devirt.c: Include demangle.h
5280 (odr_violation_reported): New static variable.
5281 (add_type_duplicate): Update odr_violations.
5282 (maybe_record_node): Add completep parameter; update it.
5283 (record_target_from_binfo): Add COMPLETEP parameter;
5284 update it as needed.
5285 (possible_polymorphic_call_targets_1): Likewise.
5286 (struct polymorphic_call_target_d): Add nonconstruction_targets;
5287 rename FINAL to COMPLETE.
5288 (record_targets_from_bases): Sanity check we found the binfo;
5289 fix COMPLETEP updating.
5290 (possible_polymorphic_call_targets): Add NONCONSTRUTION_TARGETSP
5291 parameter, fix computing of COMPLETEP.
5292 (dump_possible_polymorphic_call_targets): Imrove readability of dump;
5293 at LTO time do demangling.
5294 (ipa_devirt): Use nonconstruction_targets; Improve dumps.
5295 * gimple-fold.c (gimple_get_virt_method_for_vtable): Add can_refer
5297 (gimple_get_virt_method_for_binfo): Likewise.
5298 * gimple-fold.h (gimple_get_virt_method_for_binfo,
5299 gimple_get_virt_method_for_vtable): Update prototypes.
5301 2014-02-11 Vladimir Makarov <vmakarov@redhat.com>
5304 * genautomata.c (add_presence_absence): Fix typo with
5305 {final_}presence_list.
5307 2014-02-11 Michael Meissner <meissner@linux.vnet.ibm.com>
5310 * config/rs6000/rs6000.md (128-bit GPR splitter): Add a splitter
5311 for VSX/Altivec vectors that land in GPR registers.
5313 2014-02-11 Richard Henderson <rth@redhat.com>
5314 Jakub Jelinek <jakub@redhat.com>
5317 * tree-sra.c (load_assign_lhs_subreplacements): Add VIEW_CONVERT_EXPR
5318 around drhs if type conversion to lacc->type is not useless.
5320 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5322 * config/aarch64/aarch64-cores.def (cortex-a57): Use cortexa57
5324 (cortex-a57.cortex-a53): Likewise.
5325 * config/aarch64/aarch64.c (cortexa57_tunings): New tuning struct.
5327 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5329 * config/arm/thumb2.md (*thumb2_movhi_insn): Add alternatives for
5332 2014-02-11 Renlin Li <Renlin.Li@arm.com>
5334 * doc/sourcebuild.texi: Document check_effective_target_arm_vfp3_ok and
5335 add_options_for_arm_vfp3.
5337 2014-02-11 Jeff Law <law@redhat.com>
5340 * expr.c (expand_expr_addr_expr_1): Handle expand_expr returning an
5341 object with an undesirable mode.
5343 2014-02-11 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5346 * config/i386/sol2-9.h: New file.
5347 * config.gcc (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*,
5348 *-*-solaris2.9*): Use it.
5350 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
5352 * config/microblaze/microblaze.md: Add movsi4_rev insn pattern.
5353 * config/microblaze/predicates.md: Add reg_or_mem_operand predicate.
5355 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
5357 * config/microblaze/microblaze.c: Extend mcpu version format
5359 2014-02-10 David Holsgrove <david.holsgrove@xilinx.com>
5361 * config/microblaze/microblaze.h: Define SIZE_TYPE and PTRDIFF_TYPE.
5363 2014-02-10 Richard Henderson <rth@redhat.com>
5366 * calls.c (expand_call): Don't double-push for reg_parm_stack_space.
5367 * config/i386/i386.c (init_cumulative_args): Remove sorry for 64-bit
5368 ms-abi vs -mno-accumulate-outgoing-args.
5369 (ix86_expand_prologue): Unconditionally call ix86_eax_live_at_start_p.
5370 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Fix comment with
5373 2014-02-10 Bernd Edlinger <bernd.edlinger@hotmail.de>
5376 * cfgexpand.c (expand_asm_operands): Attach source location to
5377 ASM_INPUT rtx objects.
5378 * print-rtl.c (print_rtx): Check for UNKNOWN_LOCATION.
5380 2014-02-10 Nick Clifton <nickc@redhat.com>
5382 * config/mn10300/mn10300.c (popcount): New function.
5383 (mn10300_expand_prologue): Include saved registers in stack usage
5386 2014-02-10 Jeff Law <law@redhat.com>
5389 * reload1.c (emit_input_reload_insns): Do not create invalid RTL
5390 when changing the SET_DEST of a prior insn to avoid an input reload.
5392 2014-02-10 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
5394 * config/rs6000/sysv4.h (ENDIAN_SELECT): Do not attempt to enforce
5395 big-endian mode for -mcall-aixdesc, -mcall-freebsd, -mcall-netbsd,
5396 -mcall-openbsd, or -mcall-linux.
5397 (CC1_ENDIAN_BIG_SPEC): Remove.
5398 (CC1_ENDIAN_LITTLE_SPEC): Remove.
5399 (CC1_ENDIAN_DEFAULT_SPEC): Remove.
5400 (CC1_SPEC): Remove (always empty) %cc1_endian_... spec.
5401 (SUBTARGET_EXTRA_SPECS): Remove %cc1_endian_big, %cc1_endian_little,
5402 and %cc1_endian_default.
5403 * config/rs6000/sysv4le.h (CC1_ENDIAN_DEFAULT_SPEC): Remove.
5405 2014-02-10 Richard Biener <rguenther@suse.de>
5407 PR tree-optimization/60115
5408 * tree-eh.c (tree_could_trap_p): Unify TARGET_MEM_REF and
5409 MEM_REF handling. Properly verify that the accesses are not
5410 out of the objects bound.
5412 2014-02-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5414 * config/aarch64/aarch64.c (aarch64_override_options): Fix typo from
5417 2014-02-10 Eric Botcazou <ebotcazou@adacore.com>
5419 * ipa-devirt.c (get_polymorphic_call_info_from_invariant): Return
5420 proper constants and fix formatting.
5421 (possible_polymorphic_call_targets): Fix formatting.
5423 2014-02-10 Kirill Yukhin <kirill.yukhin@intel.com>
5424 Ilya Tocar <ilya.tocar@intel.com>
5426 * config/i386/avx512fintrin.h (_mm512_storeu_epi64): Removed.
5427 (_mm512_loadu_epi32): Renamed into...
5428 (_mm512_loadu_si512): This.
5429 (_mm512_storeu_epi32): Renamed into...
5430 (_mm512_storeu_si512): This.
5431 (_mm512_maskz_ceil_ps): Removed.
5432 (_mm512_maskz_ceil_pd): Ditto.
5433 (_mm512_maskz_floor_ps): Ditto.
5434 (_mm512_maskz_floor_pd): Ditto.
5435 (_mm512_floor_round_ps): Ditto.
5436 (_mm512_floor_round_pd): Ditto.
5437 (_mm512_ceil_round_ps): Ditto.
5438 (_mm512_ceil_round_pd): Ditto.
5439 (_mm512_mask_floor_round_ps): Ditto.
5440 (_mm512_mask_floor_round_pd): Ditto.
5441 (_mm512_mask_ceil_round_ps): Ditto.
5442 (_mm512_mask_ceil_round_pd): Ditto.
5443 (_mm512_maskz_floor_round_ps): Ditto.
5444 (_mm512_maskz_floor_round_pd): Ditto.
5445 (_mm512_maskz_ceil_round_ps): Ditto.
5446 (_mm512_maskz_ceil_round_pd): Ditto.
5447 (_mm512_expand_pd): Ditto.
5448 (_mm512_expand_ps): Ditto.
5449 * config/i386/i386.c (ix86_builtins): Remove
5450 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK.
5451 (bdesc_args): Ditto.
5452 * config/i386/predicates.md (const1256_operand): New.
5453 (const_1_to_2_operand): Ditto.
5454 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
5455 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
5456 (*avx512pf_gatherpf<mode>sf): Ditto.
5457 (avx512pf_gatherpf<mode>df): Ditto.
5458 (*avx512pf_gatherpf<mode>df_mask): Ditto.
5459 (*avx512pf_gatherpf<mode>df): Ditto.
5460 (avx512pf_scatterpf<mode>sf): Ditto.
5461 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
5462 (*avx512pf_scatterpf<mode>sf): Ditto.
5463 (avx512pf_scatterpf<mode>df): Ditto.
5464 (*avx512pf_scatterpf<mode>df_mask): Ditto.
5465 (*avx512pf_scatterpf<mode>df): Ditto.
5466 (avx512f_expand<mode>): Removed.
5467 (<shift_insn><mode>3<mask_name>): Change predicate type.
5469 2014-02-08 Jakub Jelinek <jakub@redhat.com>
5471 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers
5472 not at the end of datarefs vector use ordered_remove to avoid
5473 reordering datarefs vector.
5476 * gimplify.c (gimplify_bind_expr): In ORT_SIMD region
5477 mark local addressable non-static vars as GOVD_PRIVATE
5478 instead of GOVD_LOCAL.
5479 * omp-low.c (lower_omp_for): Move gimple_bind_vars
5480 and BLOCK_VARS of gimple_bind_block to new_stmt rather
5484 * tree-ssa-ccp.c (surely_varying_stmt_p): Don't return true
5485 if TYPE_ATTRIBUTES (gimple_call_fntype ()) contain
5486 assume_aligned or alloc_align attributes.
5487 (bit_value_assume_aligned): Add ATTR, PTRVAL and ALLOC_ALIGN
5488 arguments. Handle also assume_aligned and alloc_align attributes.
5489 (evaluate_stmt): Adjust bit_value_assume_aligned caller. Handle
5490 calls to functions with assume_aligned or alloc_align attributes.
5491 * doc/extend.texi: Document assume_aligned and alloc_align attributes.
5493 2014-02-08 Terry Guo <terry.guo@arm.com>
5495 * doc/invoke.texi: Document ARM -march=armv7e-m.
5497 2014-02-08 Jakub Jelinek <jakub@redhat.com>
5499 * cilk-common.c (cilk_init_builtins): Clear TREE_NOTHROW
5500 flag on __cilkrts_rethrow builtin.
5503 * ipa-cp.c (determine_versionability): Fail at -O0
5504 or __attribute__((optimize (0))) or -fno-ipa-cp functions.
5505 * tree-sra.c (ipa_sra_preliminary_function_checks): Similarly.
5508 2014-02-04 Jakub Jelinek <jakub@redhat.com>
5511 * tree-inline.c (copy_forbidden): Fail for
5512 __attribute__((optimize (0))) functions.
5514 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
5516 * varpool.c: Include pointer-set.h.
5517 (varpool_remove_unreferenced_decls): Variables in other partitions
5518 will not be output; be however careful to not lose information
5521 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
5523 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do O(1)
5524 lookup in the vtable constructor.
5526 2014-02-07 Jeff Law <law@redhat.com>
5529 * config/m68k/m68k.md (ashldi_extsi): Turn into a
5530 define_insn_and_split.
5532 * ipa-inline.c (inline_small_functions): Fix typos.
5534 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5536 * config/s390/s390-protos.h (s390_can_use_simple_return_insn)
5537 (s390_can_use_return_insn): Declare.
5538 * config/s390/s390.h (EPILOGUE_USES): Define.
5539 * config/s390/s390.c (s390_mainpool_start): Allow two main_pool
5541 (s390_chunkify_start): Handle return JUMP_LABELs.
5542 (s390_early_mach): Emit a main_pool instruction on the entry edge.
5543 (s300_set_up_by_prologue, s390_can_use_simple_return_insn)
5544 (s390_can_use_return_insn): New functions.
5545 (s390_fix_long_loop_prediction): Handle conditional returns.
5546 (TARGET_SET_UP_BY_PROLOGUE): Define.
5547 * config/s390/s390.md (ANY_RETURN): New code iterator.
5548 (*creturn, *csimple_return, return, simple_return): New patterns.
5550 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5552 * config/s390/s390.c (s390_restore_gprs_from_fprs): Add REG_CFA_RESTORE
5553 notes to each restore. Also add REG_CFA_DEF_CFA when restoring %r15.
5554 (s390_optimize_prologue): Don't clear RTX_FRAME_RELATED_P. Update the
5555 REG_CFA_RESTORE list when deciding not to restore a register.
5557 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5559 * config/s390/s390.c: Include tree-pass.h and context.h.
5560 (s390_early_mach): New function, split out from...
5561 (s390_emit_prologue): ...here.
5562 (pass_data_s390_early_mach): New pass structure.
5563 (pass_s390_early_mach): New class.
5564 (s390_option_override): Create and register early_mach pass.
5565 Move to end of file.
5567 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5569 * var-tracking.c (vt_stack_adjustments): Don't require stack_adjusts
5570 to match for the exit block.
5572 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5574 * config/s390/s390.md ("atomic_load<mode>", "atomic_store<mode>")
5575 ("atomic_compare_and_swap<mode>", "atomic_fetch_<atomic><mode>"):
5576 Reject misaligned operands.
5578 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5580 * optabs.c (expand_atomic_compare_and_swap): Allow expander to fail.
5582 2014-02-07 Richard Biener <rguenther@suse.de>
5585 * gimple-low.c (lower_builtin_posix_memalign): New function.
5586 (lower_stmt): Call it to lower posix_memalign in a way
5587 to make alignment info accessible.
5589 2014-02-07 Jakub Jelinek <jakub@redhat.com>
5592 * tree.c (build_common_builtin_nodes): Set ECF_LEAF for
5593 __builtin_setjmp_receiver.
5595 2014-02-07 Richard Biener <rguenther@suse.de>
5598 * builtin-types.def (BT_FN_INT_PTRPTR_SIZE_SIZE): Add.
5599 * builtins.def (BUILT_IN_POSIX_MEMALIGN): Likewise.
5600 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
5601 Handle BUILT_IN_POSIX_MEMALIGN.
5602 (find_func_clobbers): Likewise.
5603 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
5604 (call_may_clobber_ref_p_1): Likewise.
5606 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5609 * ipa-devirt.c (record_target_from_binfo): Remove overactive
5612 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5615 * lto-cgraph.c (lto_output_node): Use
5616 symtab_get_symbol_partitioning_class.
5617 (lto_output_varpool_node): likewise.
5618 (symtab_get_symbol_partitioning_class): Move here from
5620 * cgraph.h (symbol_partitioning_class): Likewise.
5621 (symtab_get_symbol_partitioning_class): Declare.
5623 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5625 * ggc.h (ggc_internal_cleared_alloc): New macro.
5626 * vec.h (vec_safe_copy): Handle memory stats.
5627 * omp-low.c (simd_clone_struct_alloc): Use ggc_internal_cleared_alloc.
5628 * target-globals.c (save_target_globals): Likewise.
5630 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5633 * expr.c (emit_move_resolve_push): Export; be bit more selective
5634 on when to clear alias set.
5635 * expr.h (emit_move_resolve_push): Declare.
5636 * function.h (struct function): Add tail_call_marked.
5637 * tree-tailcall.c (optimize_tail_call): Set tail_call_marked.
5638 * config/i386/i386-protos.h (ix86_expand_push): Remove.
5639 * config/i386/i386.md (TImode move expander): De not call
5641 (FP push expanders): Preserve memory attributes.
5642 * config/i386/sse.md (push<mode>1): Remove.
5643 * config/i386/i386.c (ix86_expand_vector_move): Handle push operation.
5644 (ix86_expand_push): Remove.
5645 * config/i386/mmx.md (push<mode>1): Remove.
5647 2014-02-06 Jakub Jelinek <jakub@redhat.com>
5649 PR rtl-optimization/60030
5650 * internal-fn.c (ubsan_expand_si_overflow_mul_check): Surround
5651 lopart with paradoxical subreg before shifting it up by hprec.
5653 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5655 * config/arm/aarch-cost-tables.h (cortexa57_extra_costs): New table.
5656 Remove extra newline at end of file.
5657 * config/arm/arm.c (arm_cortex_a57_tune): New tuning struct.
5658 (arm_issue_rate): Handle cortexa57.
5659 * config/arm/arm-cores.def (cortex-a57): Use cortex_a57 tuning.
5660 (cortex-a57.cortex-a53): Likewise.
5662 2014-02-06 Jakub Jelinek <jakub@redhat.com>
5665 * config/arm/arm.c (emit_multi_reg_push): Add dwarf_regs_mask argument,
5666 don't record in REG_FRAME_RELATED_EXPR registers not set in that
5668 (arm_expand_prologue): Adjust all callers.
5669 (arm_unwind_emit_sequence): Allow saved, but not important for unwind
5670 info, registers also at the lowest numbered registers side. Use
5671 gcc_assert instead of abort, and SET_SRC/SET_DEST macros instead of
5675 * var-tracking.c (adjust_mems): Before adding a SET to
5676 amd->side_effects, adjust it's SET_SRC using simplify_replace_fn_rtx.
5678 2014-02-06 Alan Modra <amodra@gmail.com>
5681 * config/rs6000/rs6000.c (rs6000_secondary_memory_needed_mode): Only
5682 change SDmode to DDmode when lra_in_progress.
5684 2014-02-06 Jakub Jelinek <jakub@redhat.com>
5687 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers, call
5688 free_data_ref on the dr first, and before goto again also set dr
5689 to the next dr. For simd_lane_access, free old datarefs[i] before
5690 overwriting it. For get_vectype_for_scalar_type failure, don't
5691 free_data_ref if simd_lane_access.
5693 * Makefile.in (prefix.o, cppbuiltin.o): Depend on $(BASEVER).
5696 * tree.h (opts_for_fn): New inline function.
5697 (opt_for_fn): Define.
5698 * config/i386/i386.c (ix86_function_regparm): Use
5699 opt_for_fn (decl, optimize) instead of optimize.
5701 2014-02-06 Marcus Shawcroft <marcus.shawcroft@arm.com>
5703 * config/aarch64/aarch64.c (aarch64_classify_symbol): Fix logic
5704 for SYMBOL_REF in large memory model.
5706 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5708 * config/aarch64/aarch64-cores.def (cortex-a53): Specify CRC32
5710 (cortex-a57): Likewise.
5711 (cortex-a57.cortex-a53): Likewise.
5713 2014-02-06 Yury Gribov <y.gribov@samsung.com>
5714 Kugan Vivekanandarajah <kuganv@linaro.org>
5716 * config/arm/arm.c (arm_vector_alignment_reachable): Check
5718 * config/arm/arm.c (arm_builtin_support_vector_misalignment): Likewise.
5720 2014-02-06 Richard Biener <rguenther@suse.de>
5722 * tree-cfg.c (gimple_duplicate_sese_region): Fix ordering of
5723 set_loop_copy and initialize_original_copy_tables.
5725 2014-02-06 Alex Velenko <Alex.Velenko@arm.com>
5727 * config/aarch64/aarch64-simd.md
5728 (aarch64_ashr_simddi): Change QI to SI.
5730 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
5731 Jakub Jelinek <jakub@redhat.com>
5734 * ipa-inline-analysis.c (compute_bb_predicates): Ensure monotonicity
5737 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5739 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Change
5740 CODE_FOR_altivec_vpku[hw]um to
5741 CODE_FOR_altivec_vpku[hw]um_direct.
5742 * config/rs6000/altivec.md (vec_unpacks_hi_<VP_small_lc>): Change
5743 UNSPEC_VUNPACK_HI_SIGN to UNSPEC_VUNPACK_HI_SIGN_DIRECT.
5744 (vec_unpacks_lo_<VP_small_lc>): Change UNSPEC_VUNPACK_LO_SIGN to
5745 UNSPEC_VUNPACK_LO_SIGN_DIRECT.
5747 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5749 * config/rs6000/altivec.md (altivec_vsum2sws): Adjust code
5750 generation for -maltivec=be.
5751 (altivec_vsumsws): Simplify redundant test.
5753 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5755 * altivec.md (UNSPEC_VPACK_UNS_UNS_MOD_DIRECT): New unspec.
5756 (UNSPEC_VUNPACK_HI_SIGN_DIRECT): Likewise.
5757 (UNSPEC_VUNPACK_LO_SIGN_DIRECT): Likewise.
5758 (mulv8hi3): Use gen_altivec_vpkuwum_direct instead of
5759 gen_altivec_vpkuwum.
5760 (altivec_vpkpx): Test for VECTOR_ELT_ORDER_BIG instead of for
5762 (altivec_vpks<VI_char>ss): Likewise.
5763 (altivec_vpks<VI_char>us): Likewise.
5764 (altivec_vpku<VI_char>us): Likewise.
5765 (altivec_vpku<VI_char>um): Likewise.
5766 (altivec_vpku<VI_char>um_direct): New (copy of
5767 altivec_vpku<VI_char>um that still relies on BYTES_BIG_ENDIAN, for
5769 (altivec_vupkhs<VU_char>): Emit vupkls* instead of vupkhs* when
5770 target is little endian and -maltivec=be is not specified.
5771 (*altivec_vupkhs<VU_char>_direct): New (copy of
5772 altivec_vupkhs<VU_char> that always emits vupkhs*, for internal use).
5773 (altivec_vupkls<VU_char>): Emit vupkhs* instead of vupkls* when
5774 target is little endian and -maltivec=be is not specified.
5775 (*altivec_vupkls<VU_char>_direct): New (copy of
5776 altivec_vupkls<VU_char> that always emits vupkls*, for internal use).
5777 (altivec_vupkhpx): Emit vupklpx instead of vupkhpx when target is
5778 little endian and -maltivec=be is not specified.
5779 (altivec_vupklpx): Emit vupkhpx instead of vupklpx when target is
5780 little endian and -maltivec=be is not specified.
5782 2014-02-05 Richard Henderson <rth@redhat.com>
5785 * combine-stack-adj.c: Revert r206943.
5786 * sched-int.h (struct deps_desc): Add last_args_size.
5787 * sched-deps.c (init_deps): Initialize it.
5788 (sched_analyze_insn): Add OUTPUT dependencies between insns that
5789 contain REG_ARGS_SIZE notes.
5791 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
5793 * lto-cgraph.c (asm_nodes_output): Make global.
5794 * lto-wrapper.c (run_gcc): Pass down paralelizm to WPA.
5795 * gcc.c (AS_NEEDS_DASH_FOR_PIPED_INPUT): Allow WPA parameter
5796 (driver_handle_option): Handle OPT_fwpa.
5798 2014-02-05 Jakub Jelinek <jakub@redhat.com>
5801 * ipa-devirt.c (possible_polymorphic_call_targets): Fix
5802 a comment typo and formatting issue. If odr_hash hasn't been
5803 created, return vNULL and set *completep to false.
5806 * tree-eh.c (cleanup_empty_eh): Bail out on totally empty
5807 bb with no successors.
5809 2014-02-05 James Greenhalgh <james.greenhalgh@arm.com>
5812 * doc/invoke.texi (-march): Clarify documentation for ARM.
5816 2014-02-05 Richard Biener <rguenther@suse.de>
5818 * tree-vect-loop.c (vect_analyze_loop_2): Be more informative
5819 when not vectorizing because of too many alias checks.
5820 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
5821 Add more verboseness, avoid duplicate MSG_MISSED_OPTIMIZATION.
5823 2014-02-05 Nick Clifton <nickc@redhat.com>
5825 * config/mn10300/mn10300.c (mn10300_hard_regno_mode_ok): Do not
5826 accept extended registers in any mode when compiling for the MN10300.
5828 2014-02-05 Yury Gribov <y.gribov@samsung.com>
5830 * cif-code.def (ATTRIBUTE_MISMATCH): New CIF code.
5831 * ipa-inline.c (report_inline_failed_reason): Handle mismatched
5832 sanitization attributes.
5833 (can_inline_edge_p): Likewise.
5834 (sanitize_attrs_match_for_inline_p): New function.
5836 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
5838 * ipa-prop.c (detect_type_change): Shor circuit testing of
5839 type changes on THIS pointer.
5841 2014-02-04 John David Anglin <danglin@gcc.gnu.org>
5844 * config/pa/pa.c (legitimize_tls_address): Return original address
5845 if not passed a SYMBOL_REF rtx.
5846 (hppa_legitimize_address): Call legitimize_tls_address for all TLS
5848 (pa_emit_move_sequence): Simplify TLS source operands.
5849 (pa_legitimate_constant_p): Reject all TLS constants.
5850 * config/pa/pa.h (PA_SYMBOL_REF_TLS_P): Correct comment.
5851 (CONSTANT_ADDRESS_P): Reject TLS CONST addresses.
5853 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
5855 * ipa.c (function_and_variable_visibility): Decompose DECL_ONE_ONLY
5856 groups when we know they are controlled by LTO.
5857 * varasm.c (default_binds_local_p_1): If object is in other partition,
5858 it will be resolved locally.
5860 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
5862 * config/host-linux.c (linux_gt_pch_use_address): Don't
5863 use SSIZE_MAX because it is not always defined.
5865 2014-02-04 Vladimir Makarov <vmakarov@redhat.com>
5868 * lra-constraints.c (need_for_split_p): Use more 3 reloads as
5869 threshold for pseudo splitting.
5870 (update_ebb_live_info): Process call argument hard registers and
5871 hard registers from insn definition too.
5872 (max_small_class_regs_num): New constant.
5873 (inherit_in_ebb): Update live hard regs through EBBs. Update
5874 reloads_num only for small register classes. Don't split for
5877 2014-02-04 Markus Trippelsdorf <markus@trippelsdorf.de>
5880 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that target
5883 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
5885 * gimple-fold.c (can_refer_decl_in_current_unit_p): Default
5888 2014-02-04 Marek Polacek <polacek@redhat.com>
5890 * gdbinit.in (pel): Define.
5892 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
5894 * doc/invoke.texi (fstrict-volatile-bitfields): Clarify current
5897 2014-02-04 Richard Biener <rguenther@suse.de>
5900 * lto-streamer-out.c (tree_is_indexable): Force NAMELIST_DECLs
5901 in function context local.
5902 (lto_output_tree_ref): Do not write trees from lto_output_tree_ref.
5903 * lto-streamer-in.c (lto_input_tree_ref): Handle LTO_namelist_decl_ref
5904 similar to LTO_imported_decl_ref.
5906 2014-02-04 Jakub Jelinek <jakub@redhat.com>
5908 PR tree-optimization/60002
5909 * cgraphclones.c (build_function_decl_skip_args): Clear
5912 PR tree-optimization/60023
5913 * tree-if-conv.c (predicate_mem_writes): Pass true instead of
5914 false to gsi_replace.
5915 * tree-vect-stmts.c (vect_finish_stmt_generation): If stmt
5916 has been in some EH region and vec_stmt could throw, add
5917 vec_stmt into the same EH region.
5918 * tree-data-ref.c (get_references_in_stmt): If IFN_MASK_LOAD
5919 has no lhs, ignore it.
5920 * internal-fn.c (expand_MASK_LOAD): Likewise.
5923 * tree-inline.c (copy_forbidden): Fail for
5924 __attribute__((optimize (0))) functions.
5927 * omp-low.c (simd_clone_struct_copy): If from->inbranch
5928 is set, copy one less argument.
5929 (expand_simd_clones): Don't subtract clone_info->inbranch
5930 from simd_clone_struct_alloc argument.
5932 PR rtl-optimization/57915
5933 * recog.c (simplify_while_replacing): If all unary/binary/relational
5934 operation arguments are constant, attempt to simplify those.
5937 * expmed.c (expand_mult): For MODE_VECTOR_INT multiplication
5938 if there is no vashl<mode>3 or ashl<mode>3 insn, skip_synth.
5940 2014-02-04 Richard Biener <rguenther@suse.de>
5942 PR tree-optimization/60012
5943 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Apply
5944 TBAA disambiguation to all DDRs.
5946 2014-02-04 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5949 * config/sol2.h (LINK_LIBGCC_MAPFILE_SPEC): Define.
5950 (LINK_SPEC): Use it for -shared, -shared-libgcc.
5952 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
5955 * tree.c (get_binfo_at_offset): Do not get confused by empty classes;
5957 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
5959 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Remove.
5960 * gimple-fold.h (gimple_extract_devirt_binfo_from_cst): Remove.
5962 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
5965 * ipa-cp.c (ipa_get_indirect_edge_target_1): Use ipa-devirt
5966 to figure out targets of polymorphic calls with known decl.
5967 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
5968 * ipa-utils.h (get_polymorphic_call_info_from_invariant): Declare.
5969 * ipa-devirt.c (get_polymorphic_call_info_for_decl): Break out from ...
5970 (get_polymorphic_call_info): ... here.
5971 (get_polymorphic_call_info_from_invariant): New function.
5973 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
5975 * ipa-cp.c (ipa_get_indirect_edge_target_1): Do direct
5976 lookup via vtable pointer; check for type consistency
5977 and turn inconsitent facts into UNREACHABLE.
5978 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
5979 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do not ICE on
5980 type inconsistent querries; return UNREACHABLE instead.
5982 2014-02-03 Richard Henderson <rth@twiddle.net>
5985 * tree-ssa-uninit.c (push_to_worklist): Don't re-push if we've
5986 already processed this node.
5987 (normalize_one_pred_1): Pass along mark_set.
5988 (normalize_one_pred): Create and destroy a pointer_set_t.
5989 (normalize_one_pred_chain): Likewise.
5991 2014-02-03 Laurent Aflonsi <laurent.alfonsi@st.com>
5993 PR gcov-profile/58602
5994 * gcc/gcov-io.c (gcov_open): Open with truncation when mode < 0.
5996 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
5999 * ipa-cp.c (ipa_get_indirect_edge_target_1): Give up on
6000 -fno-devirtualize; try to devirtualize by the knowledge of
6001 virtual table pointer given by aggregate propagation.
6002 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
6003 (ipa_print_node_jump_functions): Dump also offset that
6004 is relevant for polymorphic calls.
6005 (determine_known_aggregate_parts): Add arg_type parameter; use it
6006 instead of determining the type from pointer type.
6007 (ipa_compute_jump_functions_for_edge): Update call of
6008 determine_known_aggregate_parts.
6009 * gimple-fold.c (gimple_get_virt_method_for_vtable): Break out from ...
6010 (gimple_get_virt_method_for_binfo): ... here; simplify using
6011 vtable_pointer_value_to_vtable.
6012 * gimple-fold.h (gimple_get_virt_method_for_vtable): Declare.
6013 * ipa-devirt.c (subbinfo_with_vtable_at_offset): Turn OFFSET parameter
6014 to unsigned HOST_WIDE_INT; use vtable_pointer_value_to_vtable.
6015 (vtable_pointer_value_to_vtable): Break out from ...; handle also
6017 (vtable_pointer_value_to_binfo): ... here.
6018 * ipa-utils.h (vtable_pointer_value_to_vtable): Declare.
6020 2014-02-03 Teresa Johnson <tejohnson@google.com>
6022 * tree-vect-slp.c (vect_supported_load_permutation_p): Avoid
6023 redef of outer loop index variable.
6025 2014-02-03 Marc Glisse <marc.glisse@inria.fr>
6029 * doc/extend.texi (Function Attributes): Typo.
6031 2014-02-03 Cong Hou <congh@google.com>
6033 PR tree-optimization/60000
6034 * tree-vect-loop.c (vect_transform_loop): Set pattern_def_seq to NULL
6035 if the vectorized statement is a store. A store statement can only
6036 appear at the end of pattern statements.
6038 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
6040 * config/i386/i386.c (flag_opts): Add -mlong-double-128.
6041 (ix86_option_override_internal): Default long double to 64-bit for
6042 32-bit Bionic and to 128-bit for 64-bit Bionic.
6044 * config/i386/i386.h (LONG_DOUBLE_TYPE_SIZE): Use 128 if
6045 TARGET_LONG_DOUBLE_128 is true.
6046 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Likewise.
6048 * config/i386/i386.opt (mlong-double-80): Negate -mlong-double-64.
6049 (mlong-double-64): Negate -mlong-double-128.
6050 (mlong-double-128): New option.
6052 * config/i386/i386-c.c (ix86_target_macros): Define
6053 __LONG_DOUBLE_128__ for TARGET_LONG_DOUBLE_128.
6055 * doc/invoke.texi: Document -mlong-double-128.
6057 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
6059 PR rtl-optimization/60024
6060 * sel-sched.c (init_regs_for_mode): Check if mode is OK first.
6062 2014-02-03 Markus Trippelsdorf <markus@trippelsdorf.de>
6064 * doc/invoke.texi (fprofile-reorder-functions): Fix typo.
6066 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
6068 PR rtl-optimization/57662
6069 * sel-sched.c (code_motion_path_driver): Do not mark already not
6070 existing blocks in the visiting bitmap.
6072 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
6074 * sel-sched-ir.c (sel_gen_insn_from_expr_after): Reset INSN_DELETED_P
6075 on the insn being emitted.
6077 2014-02-03 James Greenhalgh <james.greenhalgh@arm.com>
6078 Will Deacon <will.deacon@arm.com>
6080 * doc/gimple.texi (gimple_asm_clear_volatile): Remove.
6082 2014-02-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6084 * config/arm/arm-tables.opt: Regenerate.
6086 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6088 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): Generalize
6089 for vector types other than V16QImode.
6090 * config/rs6000/altivec.md (altivec_vperm_<mode>): Change to a
6091 define_expand, and call altivec_expand_vec_perm_le when producing
6092 code with little endian element order.
6093 (*altivec_vperm_<mode>_internal): New insn having previous
6094 behavior of altivec_vperm_<mode>.
6095 (altivec_vperm_<mode>_uns): Change to a define_expand, and call
6096 altivec_expand_vec_perm_le when producing code with little endian
6098 (*altivec_vperm_<mode>_uns_internal): New insn having previous
6099 behavior of altivec_vperm_<mode>_uns.
6101 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6103 * config/rs6000/altivec.md (UNSPEC_VSUMSWS_DIRECT): New unspec.
6104 (altivec_vsumsws): Add handling for -maltivec=be with a little
6106 (altivec_vsumsws_direct): New.
6107 (reduc_splus_<mode>): Call gen_altivec_vsumsws_direct instead of
6108 gen_altivec_vsumsws.
6110 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
6112 * ipa-devirt.c (subbinfo_with_vtable_at_offset,
6113 vtable_pointer_value_to_binfo): New functions.
6114 * ipa-utils.h (vtable_pointer_value_to_binfo): Declare.
6115 * ipa-prop.c (extr_type_from_vtbl_ptr_store): Use it.
6117 2014-02-02 Sandra Loosemore <sandra@codesourcery.com>
6119 * config/nios2/nios2.md (load_got_register): Initialize GOT
6120 pointer from _gp_got instead of _GLOBAL_OFFSET_TABLE_.
6121 * config/nios2/nios2.c (nios2_function_profiler): Likewise.
6123 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
6125 * ipa-prop.c (update_jump_functions_after_inlining): When type is not
6126 preserverd by passthrough, do not propagate the type.
6128 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
6130 * config/mips/mips.c (MIPS_GET_FCSR, MIPS_SET_FCSR): New macros.
6131 (mips_atomic_assign_expand_fenv): New function.
6132 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): Define.
6134 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
6136 * doc/extend.texi (__builtin_mips_get_fcsr): Document.
6137 (__builtin_mips_set_fcsr): Likewise.
6138 * config/mips/mips-ftypes.def: Add MIPS_VOID_FTYPE_USI and
6139 MIPS_USI_FTYPE_VOID.
6140 * config/mips/mips-protos.h (mips16_expand_get_fcsr): Declare
6141 (mips16_expand_set_fcsr): Likewise.
6142 * config/mips/mips.c (mips16_get_fcsr_stub): New variable.
6143 (mips16_set_fcsr_stub): Likewise.
6144 (mips16_get_fcsr_one_only_stub): New class.
6145 (mips16_set_fcsr_one_only_stub): Likewise.
6146 (mips16_expand_get_fcsr, mips16_expand_set_fcsr): New functions.
6147 (mips_code_end): Output the get_fcsr and set_fcsr stubs, if needed.
6148 (BUILTIN_AVAIL_MIPS16, AVAIL_ALL): New macros.
6149 (hard_float): New availability predicate.
6150 (mips_builtins): Add get_fcsr and set_fcsr.
6151 (mips_expand_builtin): Check BUILTIN_AVAIL_MIPS16.
6152 * config/mips/mips.md (UNSPEC_GET_FCSR, UNSPEC_SET_FCSR): New unspecs.
6153 (GET_FCSR_REGNUM, SET_FCSR_REGNUM): New constants.
6154 (mips_get_fcsr, *mips_get_fcsr, mips_get_fcsr_mips16_<mode>)
6155 (mips_set_fcsr, *mips_set_fcsr, mips_set_fcsr_mips16_<mode>): New
6158 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
6160 * config/mips/mips.c (mips_one_only_stub): New class.
6161 (mips_need_mips16_rdhwr_p): Replace with...
6162 (mips16_rdhwr_stub): ...this new variable.
6163 (mips16_stub_call_address): New function.
6164 (mips16_rdhwr_one_only_stub): New class.
6165 (mips_expand_thread_pointer): Use mips16_stub_call_address.
6166 (mips_output_mips16_rdhwr): Delete.
6167 (mips_finish_stub): New function.
6168 (mips_code_end): Use it to handle rdhwr stubs.
6170 2014-02-02 Uros Bizjak <ubizjak@gmail.com>
6173 * config/i386/i386.c (classify_argument): Fix handling of bit_offset
6174 when calculating size of integer atomic types.
6176 2014-02-02 H.J. Lu <hongjiu.lu@intel.com>
6178 * ipa-inline-analysis.c (true_predicate_p): Fix a typo in comments.
6180 2014-02-01 Jakub Jelinek <jakub@redhat.com>
6182 PR tree-optimization/60003
6183 * gimple-low.c (lower_builtin_setjmp): Set cfun->has_nonlocal_label.
6184 * profile.c (branch_prob): Use gimple_call_builtin_p
6185 to check for BUILT_IN_SETJMP_RECEIVER.
6186 * tree-inline.c (copy_bb): Call notice_special_calls.
6188 2014-01-31 Vladimir Makarov <vmakarov@redhat.com>
6191 * lra-constraints.c (process_alt_operands): Update reload_sum only
6194 2014-01-31 Richard Henderson <rth@redhat.com>
6197 * tree-eh.c (lower_try_finally_switch): Delay lowering finally block
6198 until after else_eh is processed.
6200 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
6202 * config/i386/avx512fintrin.h (_MM_FROUND_TO_NEAREST_INT),
6203 (_MM_FROUND_TO_NEG_INF), (_MM_FROUND_TO_POS_INF),
6204 (_MM_FROUND_TO_ZERO), (_MM_FROUND_CUR_DIRECTION): Are already defined
6205 in smmintrin.h, remove them.
6206 (_MM_FROUND_NO_EXC): Same as above, bit also wrong value.
6207 * config/i386/i386.c (ix86_print_operand): Split sae and rounding.
6208 * config/i386/i386.md (ROUND_SAE): Fix value.
6209 * config/i386/predicates.md (const_4_or_8_to_11_operand): New.
6210 (const48_operand): New.
6211 * config/i386/subst.md (round), (round_expand): Use
6212 const_4_or_8_to_11_operand.
6213 (round_saeonly), (round_saeonly_expand): Use const48_operand.
6215 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
6217 * config/i386/constraints.md (Yk): Swap meaning with k.
6218 * config/i386/i386.md (movhi_internal): Change Yk to k.
6219 (movqi_internal): Ditto.
6220 (*k<logic><mode>): Ditto.
6223 (kandn<mode>): Ditto.
6224 (*<code>hi_1): Ditto.
6225 (*<code>qi_1): Ditto.
6226 (kxnor<mode>): Ditto.
6227 (kortestzhi): Ditto.
6228 (kortestchi): Ditto.
6230 (*one_cmplhi2_1): Ditto.
6231 (*one_cmplqi2_1): Ditto.
6232 * config/i386/sse.md (): Change k to Yk.
6233 (avx512f_load<mode>_mask): Ditto.
6234 (avx512f_blendm<mode>): Ditto.
6235 (avx512f_store<mode>_mask): Ditto.
6236 (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
6237 (avx512f_storedqu<mode>_mask): Ditto.
6238 (avx512f_cmp<mode>3<mask_scalar_merge_name><round_saeonly_name>):
6240 (avx512f_ucmp<mode>3<mask_scalar_merge_name>): Ditto.
6241 (avx512f_vmcmp<mode>3<round_saeonly_name>): Ditto.
6242 (avx512f_vmcmp<mode>3_mask<round_saeonly_name>): Ditto.
6243 (avx512f_maskcmp<mode>3): Ditto.
6244 (avx512f_fmadd_<mode>_mask<round_name>): Ditto.
6245 (avx512f_fmadd_<mode>_mask3<round_name>): Ditto.
6246 (avx512f_fmsub_<mode>_mask<round_name>): Ditto.
6247 (avx512f_fmsub_<mode>_mask3<round_name>): Ditto.
6248 (avx512f_fnmadd_<mode>_mask<round_name>): Ditto.
6249 (avx512f_fnmadd_<mode>_mask3<round_name>): Ditto.
6250 (avx512f_fnmsub_<mode>_mask<round_name>): Ditto.
6251 (avx512f_fnmsub_<mode>_mask3<round_name>): Ditto.
6252 (avx512f_fmaddsub_<mode>_mask<round_name>): Ditto.
6253 (avx512f_fmaddsub_<mode>_mask3<round_name>): Ditto.
6254 (avx512f_fmsubadd_<mode>_mask<round_name>): Ditto.
6255 (avx512f_fmsubadd_<mode>_mask3<round_name>): Ditto.
6256 (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
6257 (vec_extract_lo_<mode>_maskm): Ditto.
6258 (vec_extract_hi_<mode>_maskm): Ditto.
6259 (avx512f_vternlog<mode>_mask): Ditto.
6260 (avx512f_fixupimm<mode>_mask<round_saeonly_name>): Ditto.
6261 (avx512f_sfixupimm<mode>_mask<round_saeonly_name>): Ditto.
6262 (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
6263 (avx512f_<code>v8div16qi2_mask): Ditto.
6264 (avx512f_<code>v8div16qi2_mask_store): Ditto.
6265 (avx512f_eq<mode>3<mask_scalar_merge_name>_1): Ditto.
6266 (avx512f_gt<mode>3<mask_scalar_merge_name>): Ditto.
6267 (avx512f_testm<mode>3<mask_scalar_merge_name>): Ditto.
6268 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Ditto.
6269 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
6270 (*avx512pf_gatherpf<mode>df_mask): Ditto.
6271 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
6272 (*avx512pf_scatterpf<mode>df_mask): Ditto.
6273 (avx512cd_maskb_vec_dupv8di): Ditto.
6274 (avx512cd_maskw_vec_dupv16si): Ditto.
6275 (avx512f_vpermi2var<mode>3_maskz): Ditto.
6276 (avx512f_vpermi2var<mode>3_mask): Ditto.
6277 (avx512f_vpermi2var<mode>3_mask): Ditto.
6278 (avx512f_vpermt2var<mode>3_maskz): Ditto.
6279 (*avx512f_gathersi<mode>): Ditto.
6280 (*avx512f_gathersi<mode>_2): Ditto.
6281 (*avx512f_gatherdi<mode>): Ditto.
6282 (*avx512f_gatherdi<mode>_2): Ditto.
6283 (*avx512f_scattersi<mode>): Ditto.
6284 (*avx512f_scatterdi<mode>): Ditto.
6285 (avx512f_compress<mode>_mask): Ditto.
6286 (avx512f_compressstore<mode>_mask): Ditto.
6287 (avx512f_expand<mode>_mask): Ditto.
6288 * config/i386/subst.md (mask): Change k to Yk.
6289 (mask_scalar_merge): Ditto.
6292 2014-01-31 Marc Glisse <marc.glisse@inria.fr>
6294 * doc/extend.texi (Vector Extensions): Document ?: in C++.
6296 2014-01-31 Richard Biener <rguenther@suse.de>
6299 * builtins.c (fold_builtin_memory_op): Make sure to not
6300 use a floating-point mode or a boolean or enumeral type for
6303 2014-01-30 DJ Delorie <dj@redhat.com>
6305 * config/msp430/msp430.h (LIB_SPEC): Add -lcrt
6306 * config/msp430/msp430.md (msp430_refsym_need_exit): New.
6307 * config/msp430/msp430.c (msp430_expand_epilogue): Call it
6308 whenever main() has an epilogue.
6310 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6312 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Remove
6313 unused variable "field".
6314 * config/rs6000/vsx.md (vsx_mergel_<mode>): Add missing DONE.
6315 (vsx_mergeh_<mode>): Likewise.
6316 * config/rs6000/altivec.md (altivec_vmrghb): Likewise.
6317 (altivec_vmrghh): Likewise.
6318 (altivec_vmrghw): Likewise.
6319 (altivec_vmrglb): Likewise.
6320 (altivec_vmrglh): Likewise.
6321 (altivec_vmrglw): Likewise.
6322 (altivec_vspltb): Add missing uses.
6323 (altivec_vsplth): Likewise.
6324 (altivec_vspltw): Likewise.
6325 (altivec_vspltsf): Likewise.
6327 2014-01-30 Jakub Jelinek <jakub@redhat.com>
6330 * ifcvt.c (cond_exec_process_insns): Don't conditionalize
6331 frame related instructions.
6333 2014-01-30 Vladimir Makarov <vmakarov@redhat.com>
6335 PR rtl-optimization/59959
6336 * lra-constrains.c (simplify_operand_subreg): Assign NO_REGS to
6337 any reload of register whose subreg is invalid.
6339 2014-01-30 Jakub Jelinek <jakub@redhat.com>
6341 * config/i386/f16cintrin.h (_cvtsh_ss): Avoid -Wnarrowing warning.
6342 * config/i386/avx512fintrin.h (_mm512_mask_cvtusepi64_storeu_epi32):
6343 Add missing return type - void.
6345 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6347 * gcc/config/rs6000/rs6000.c (rs6000_expand_vector_init): Use
6348 gen_vsx_xxspltw_v4sf_direct instead of gen_vsx_xxspltw_v4sf;
6349 remove element index adjustment for endian (now handled in vsx.md
6351 (altivec_expand_vec_perm_const): Use
6352 gen_altivec_vsplt[bhw]_direct instead of gen_altivec_vsplt[bhw].
6353 * gcc/config/rs6000/vsx.md (UNSPEC_VSX_XXSPLTW): New unspec.
6354 (vsx_xxspltw_<mode>): Adjust element index for little endian.
6355 * gcc/config/rs6000/altivec.md (altivec_vspltb): Divide into a
6356 define_expand and a new define_insn *altivec_vspltb_internal;
6357 adjust for -maltivec=be on a little endian target.
6358 (altivec_vspltb_direct): New.
6359 (altivec_vsplth): Divide into a define_expand and a new
6360 define_insn *altivec_vsplth_internal; adjust for -maltivec=be on a
6361 little endian target.
6362 (altivec_vsplth_direct): New.
6363 (altivec_vspltw): Divide into a define_expand and a new
6364 define_insn *altivec_vspltw_internal; adjust for -maltivec=be on a
6365 little endian target.
6366 (altivec_vspltw_direct): New.
6367 (altivec_vspltsf): Divide into a define_expand and a new
6368 define_insn *altivec_vspltsf_internal; adjust for -maltivec=be on
6369 a little endian target.
6371 2014-01-30 Richard Biener <rguenther@suse.de>
6373 PR tree-optimization/59993
6374 * tree-ssa-forwprop.c (associate_pointerplus): Check we
6375 can propagate form the earlier stmt and avoid the transform
6376 when the intermediate result is needed.
6378 2014-01-30 Alangi Derick <alangiderick@gmail.com>
6380 * README.Portability: Fix typo.
6382 2014-01-30 David Holsgrove <david.holsgrove@xilinx.com>
6384 * config/microblaze/microblaze.md(cstoresf4, cbranchsf4): Replace
6385 comparison_operator with ordered_comparison_operator.
6387 2014-01-30 Nick Clifton <nickc@redhat.com>
6389 * config/mn10300/mn10300-protos.h (mn10300_store_multiple_operation_p):
6390 Rename to mn10300_store_multiple_regs.
6391 * config/mn10300/mn10300.c: Likewise.
6392 * config/mn10300/mn10300.md (store_movm): Fix typo: call
6393 store_multiple_regs.
6394 * config/mn10300/predicates.md (mn10300_store_multiple_operation):
6395 Call mn10300_store_multiple_regs.
6397 2014-01-30 Nick Clifton <nickc@redhat.com>
6398 DJ Delorie <dj@redhat.com>
6400 * config/rl78/rl78.c (register_sizes): Make the "upper half" of
6401 %fp 2 to keep registers after it properly word-aligned.
6402 (rl78_alloc_physical_registers_umul): Handle the case where both
6403 input operands are the same.
6405 2014-01-30 Richard Biener <rguenther@suse.de>
6407 PR tree-optimization/59903
6408 * tree-vect-loop.c (vect_transform_loop): Guard multiple-types
6411 2014-01-30 Jason Merrill <jason@redhat.com>
6414 * tree.c (walk_type_fields): Handle VECTOR_TYPE.
6417 * cgraphunit.c (expand_thunk): Copy volatile arg to a temporary.
6419 2014-01-30 Richard Biener <rguenther@suse.de>
6421 PR tree-optimization/59951
6422 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Skip uses in debug insns.
6424 2014-01-30 Savin Zlobec <savin.zlobec@gmail.com>
6427 * config/nios2/nios2.c (nios2_fpu_insn_asm): Fix asm output of
6428 SFmode to DFmode case.
6430 2014-01-29 DJ Delorie <dj@redhat.com>
6432 * config/msp430/msp430.opt (-minrt): New.
6433 * config/msp430/msp430.h (STARTFILE_SPEC): Link alternate runtime
6435 (ENDFILE_SPEC): Likewise.
6437 2014-01-29 Jan Hubicka <hubicka@ucw.cz>
6439 * ipa-inline-analysis.c (clobber_only_eh_bb_p): New function.
6440 (estimate_function_body_sizes): Use it.
6442 2014-01-29 Paolo Carlini <paolo.carlini@oracle.com>
6445 * dwarf2out.c (is_cxx_auto): New.
6446 (is_base_type): Use it.
6447 (gen_type_die_with_usage): Likewise.
6449 2014-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6451 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Use
6452 CODE_FOR_altivec_vmrg*_direct rather than CODE_FOR_altivec_vmrg*.
6453 * config/rs6000/vsx.md (vsx_mergel_<mode>): Adjust for
6454 -maltivec=be with LE targets.
6455 (vsx_mergeh_<mode>): Likewise.
6456 * config/rs6000/altivec.md (UNSPEC_VMRG[HL]_DIRECT): New unspecs.
6457 (mulv8hi3): Use gen_altivec_vmrg[hl]w_direct.
6458 (altivec_vmrghb): Replace with define_expand and new
6459 *altivec_vmrghb_internal insn; adjust for -maltivec=be with LE targets.
6460 (altivec_vmrghb_direct): New define_insn.
6461 (altivec_vmrghh): Replace with define_expand and new
6462 *altivec_vmrghh_internal insn; adjust for -maltivec=be with LE targets.
6463 (altivec_vmrghh_direct): New define_insn.
6464 (altivec_vmrghw): Replace with define_expand and new
6465 *altivec_vmrghw_internal insn; adjust for -maltivec=be with LE targets.
6466 (altivec_vmrghw_direct): New define_insn.
6467 (*altivec_vmrghsf): Adjust for endianness.
6468 (altivec_vmrglb): Replace with define_expand and new
6469 *altivec_vmrglb_internal insn; adjust for -maltivec=be with LE targets.
6470 (altivec_vmrglb_direct): New define_insn.
6471 (altivec_vmrglh): Replace with define_expand and new
6472 *altivec_vmrglh_internal insn; adjust for -maltivec=be with LE targets.
6473 (altivec_vmrglh_direct): New define_insn.
6474 (altivec_vmrglw): Replace with define_expand and new
6475 *altivec_vmrglw_internal insn; adjust for -maltivec=be with LE targets.
6476 (altivec_vmrglw_direct): New define_insn.
6477 (*altivec_vmrglsf): Adjust for endianness.
6478 (vec_widen_umult_hi_v16qi): Use gen_altivec_vmrghh_direct.
6479 (vec_widen_umult_lo_v16qi): Use gen_altivec_vmrglh_direct.
6480 (vec_widen_smult_hi_v16qi): Use gen_altivec_vmrghh_direct.
6481 (vec_widen_smult_lo_v16qi): Use gen_altivec_vmrglh_direct.
6482 (vec_widen_umult_hi_v8hi): Use gen_altivec_vmrghw_direct.
6483 (vec_widen_umult_lo_v8hi): Use gen_altivec_vmrglw_direct.
6484 (vec_widen_smult_hi_v8hi): Use gen_altivec_vmrghw_direct.
6485 (vec_widen_smult_lo_v8hi): Use gen_altivec_vmrglw_direct.
6487 2014-01-29 Marcus Shawcroft <marcus.shawcroft@arm.com>
6489 * config/aarch64/aarch64.c (aarch64_expand_mov_immediate)
6490 (aarch64_legitimate_address_p, aarch64_class_max_nregs): Adjust
6493 2014-01-29 Richard Biener <rguenther@suse.de>
6495 PR tree-optimization/58742
6496 * tree-ssa-forwprop.c (associate_pointerplus): Rename to
6497 associate_pointerplus_align.
6498 (associate_pointerplus_diff): New function.
6499 (associate_pointerplus): Likewise. Call associate_pointerplus_align
6500 and associate_pointerplus_diff.
6502 2014-01-29 Richard Biener <rguenther@suse.de>
6504 * lto-streamer.h (LTO_major_version): Bump to 3.
6505 (LTO_minor_version): Reset to 0.
6507 2014-01-29 Renlin Li <Renlin.Li@arm.com>
6509 * config/arm/arm-arches.def (ARM_ARCH): Add armv7ve arch.
6510 * config/arm/arm.c (FL_FOR_ARCH7VE): New.
6511 (arm_file_start): Generate correct asm header for armv7ve.
6512 * config/arm/bpabi.h: Add multilib support for armv7ve.
6513 * config/arm/driver-arm.c: Change the architectures of cortex-a7
6514 and cortex-a15 to armv7ve.
6515 * config/arm/t-aprofile: Add multilib support for armv7ve.
6516 * doc/invoke.texi: Document -march=armv7ve.
6518 2014-01-29 Richard Biener <rguenther@suse.de>
6520 PR tree-optimization/58742
6521 * tree-ssa-forwprop.c (associate_plusminus): Return true
6522 if we changed sth, defer EH cleanup to ...
6523 (ssa_forward_propagate_and_combine): ... here. Call simplify_mult.
6524 (simplify_mult): New function.
6526 2014-01-29 Jakub Jelinek <jakub@redhat.com>
6529 PR tree-optimization/59920
6530 * tree.c (build_common_builtin_nodes): Remove
6531 __builtin_setjmp_dispatcher initialization.
6532 * omp-low.h (make_gimple_omp_edges): Add a new int * argument.
6533 * profile.c (branch_prob): Use gsi_start_nondebug_after_labels_bb
6534 instead of gsi_after_labels + manually skipping debug stmts.
6535 Don't ignore bbs with BUILT_IN_SETJMP_DISPATCHER, instead
6536 ignore bbs with IFN_ABNORMAL_DISPATCHER.
6537 * tree-inline.c (copy_edges_for_bb): Remove
6538 can_make_abnormal_goto argument, instead add abnormal_goto_dest
6539 argument. Ignore computed_goto_p stmts. Don't call
6540 make_abnormal_goto_edges. If a call might need abnormal edges
6541 for non-local gotos, see if it already has an edge to
6542 IFN_ABNORMAL_DISPATCHER or if it is IFN_ABNORMAL_DISPATCHER
6543 with true argument, don't do anything then, otherwise add
6544 EDGE_ABNORMAL from the call's bb to abnormal_goto_dest.
6545 (copy_cfg_body): Compute abnormal_goto_dest, adjust copy_edges_for_bb
6547 * gimple-low.c (struct lower_data): Remove calls_builtin_setjmp.
6548 (lower_function_body): Don't emit __builtin_setjmp_dispatcher.
6549 (lower_stmt): Don't set data->calls_builtin_setjmp.
6550 (lower_builtin_setjmp): Adjust comment.
6551 * builtins.def (BUILT_IN_SETJMP_DISPATCHER): Remove.
6552 * tree-cfg.c (found_computed_goto): Remove.
6553 (factor_computed_gotos): Remove.
6554 (make_goto_expr_edges): Return bool, true for computed gotos.
6555 Don't call make_abnormal_goto_edges.
6556 (build_gimple_cfg): Don't set found_computed_goto, don't call
6557 factor_computed_gotos.
6558 (computed_goto_p): No longer static.
6559 (make_blocks): Don't set found_computed_goto.
6560 (get_abnormal_succ_dispatcher, handle_abnormal_edges): New functions.
6561 (make_edges): If make_goto_expr_edges returns true, push bb
6562 into ab_edge_goto vector, for stmt_can_make_abnormal_goto calls
6563 instead of calling make_abnormal_goto_edges push bb into ab_edge_call
6564 vector. Record mapping between bbs and OpenMP regions if there
6565 are any, adjust make_gimple_omp_edges caller. Call
6566 handle_abnormal_edges.
6567 (make_abnormal_goto_edges): Remove.
6568 * tree-cfg.h (make_abnormal_goto_edges): Remove.
6569 (computed_goto_p, get_abnormal_succ_dispatcher): New prototypes.
6570 * internal-fn.c (expand_ABNORMAL_DISPATCHER): New function.
6571 * builtins.c (expand_builtin): Don't handle BUILT_IN_SETJMP_DISPATCHER.
6572 * internal-fn.def (ABNORMAL_DISPATCHER): New.
6573 * omp-low.c (make_gimple_omp_edges): Add region_idx argument, when
6574 filling *region also set *region_idx to (*region)->entry->index.
6577 * read-rtl.c (read_rtx_code): Clear all of RTX_CODE_SIZE (code).
6578 For REGs set ORIGINAL_REGNO.
6580 2014-01-29 Bingfeng Mei <bmei@broadcom.com>
6582 * doc/md.texi: Mention that a target shouldn't implement
6583 vec_widen_(s|u)mul_even/odd pair if it is less efficient
6586 2014-01-29 Jakub Jelinek <jakub@redhat.com>
6588 PR tree-optimization/59594
6589 * tree-vect-data-refs.c (vect_analyze_data_ref_accesses): Sort
6590 a copy of the datarefs vector rather than the vector itself.
6592 2014-01-28 Jason Merrill <jason@redhat.com>
6595 * dwarf2out.c (auto_die): New static.
6596 (gen_type_die_with_usage): Handle C++1y 'auto'.
6597 (gen_subprogram_die): If in-class DIE had 'auto', emit type again
6600 2014-01-28 H.J. Lu <hongjiu.lu@intel.com>
6603 * config/i386/gnu-user64.h (SPEC_32): Add "m16|" to "m32".
6604 (SPEC_X32): Likewise.
6605 (SPEC_64): Likewise.
6606 * config/i386/i386.c (ix86_option_override_internal): Turn off
6607 OPTION_MASK_ISA_64BIT, OPTION_MASK_ABI_X32 and OPTION_MASK_ABI_64
6609 (x86_file_start): Output .code16gcc for TARGET_16BIT.
6610 * config/i386/i386.h (TARGET_16BIT): New macro.
6611 (TARGET_16BIT_P): Likewise.
6612 * config/i386/i386.opt: Add m16.
6613 * doc/invoke.texi: Document -m16.
6615 2014-01-28 Jakub Jelinek <jakub@redhat.com>
6617 PR preprocessor/59935
6618 * input.c (location_get_source_line): Bail out on when line number
6619 is zero, and test the return value of lookup_or_add_file_to_cache_tab.
6621 2014-01-28 Richard Biener <rguenther@suse.de>
6623 PR tree-optimization/58742
6624 * tree-ssa-forwprop.c (associate_plusminus): Handle
6625 pointer subtraction of the form (T)(P + A) - (T)P.
6627 2014-01-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6629 * config/arm/arm.c (arm_new_rtx_costs): Remove useless statement
6632 2014-01-28 Richard Biener <rguenther@suse.de>
6635 2014-01-28 Richard Biener <rguenther@suse.de>
6637 PR rtl-optimization/45364
6638 PR rtl-optimization/59890
6639 * var-tracking.c (local_get_addr_clear_given_value): Handle
6640 already cleared slot.
6641 (val_reset): Handle not allocated local_get_addr_cache.
6642 (vt_find_locations): Use post-order on the inverted CFG.
6644 2014-01-28 Richard Biener <rguenther@suse.de>
6646 * tree-data-ref.h (ddr_is_anti_dependent, ddrs_have_anti_deps): Remove.
6648 2014-01-28 Richard Biener <rguenther@suse.de>
6650 PR rtl-optimization/45364
6651 PR rtl-optimization/59890
6652 * var-tracking.c (local_get_addr_clear_given_value): Handle
6653 already cleared slot.
6654 (val_reset): Handle not allocated local_get_addr_cache.
6655 (vt_find_locations): Use post-order on the inverted CFG.
6657 2014-01-28 Alan Modra <amodra@gmail.com>
6659 * Makefile.in (BUILD_CPPFLAGS): Do not use ALL_CPPFLAGS.
6660 * configure.ac <recursive call for build != host>: Define
6661 GENERATOR_FILE. Comment. Use CXX_FOR_BUILD, CXXFLAGS_FOR_BUILD
6662 and LD_FOR_BUILD too.
6663 * configure: Regenerate.
6665 2014-01-27 Allan Sandfeld Jensen <sandfeld@kde.org>
6667 * config/i386/i386.c (get_builtin_code_for_version): Separate
6668 Westmere from Nehalem, Ivy Bridge from Sandy Bridge and
6669 Broadwell from Haswell.
6671 2014-01-27 Steve Ellcey <sellcey@mips.com>
6673 * common/config/mips/mips-common.c (TARGET_DEFAULT_TARGET_FLAGS):
6674 Remove TARGET_FP_EXCEPTIONS_DEFAULT and MASK_FUSED_MADD.
6675 * config/mips/mips.c (mips_option_override): Change setting
6677 * config/mips/mips.h (TARGET_FP_EXCEPTIONS_DEFAULT): Remove.
6678 * config/mips/mips.opt (DSP, DSPR2, FP_EXCEPTIONS, FUSED_MADD, MIPS3D):
6679 Change from Mask to Var.
6681 2014-01-27 Jeff Law <law@redhat.com>
6683 * ipa-inline.c (inline_small_functions): Fix typo.
6685 2014-01-27 Ilya Tocar <ilya.tocar@intel.com>
6687 * config/i386/avx512fintrin.h (_mm512_mask_cvtepi32_storeu_epi8): New.
6688 (_mm512_mask_cvtsepi32_storeu_epi8): Ditto.
6689 (_mm512_mask_cvtusepi32_storeu_epi8): Ditto.
6690 (_mm512_mask_cvtepi32_storeu_epi16): Ditto.
6691 (_mm512_mask_cvtsepi32_storeu_epi16): Ditto.
6692 (_mm512_mask_cvtusepi32_storeu_epi16): Ditto.
6693 (_mm512_mask_cvtepi64_storeu_epi32): Ditto.
6694 (_mm512_mask_cvtsepi64_storeu_epi32): Ditto.
6695 (_mm512_mask_cvtusepi64_storeu_epi32): Ditto.
6696 (_mm512_mask_cvtepi64_storeu_epi16): Ditto.
6697 (_mm512_mask_cvtsepi64_storeu_epi16): Ditto.
6698 (_mm512_mask_cvtusepi64_storeu_epi16): Ditto.
6699 (_mm512_mask_cvtepi64_storeu_epi8): Ditto.
6700 (_mm512_mask_cvtsepi64_storeu_epi8): Ditto.
6701 (_mm512_mask_cvtusepi64_storeu_epi8): Ditto.
6702 (_mm512_storeu_epi64): Ditto.
6703 (_mm512_cmpge_epi32_mask): Ditto.
6704 (_mm512_cmpge_epu32_mask): Ditto.
6705 (_mm512_cmpge_epi64_mask): Ditto.
6706 (_mm512_cmpge_epu64_mask): Ditto.
6707 (_mm512_cmple_epi32_mask): Ditto.
6708 (_mm512_cmple_epu32_mask): Ditto.
6709 (_mm512_cmple_epi64_mask): Ditto.
6710 (_mm512_cmple_epu64_mask): Ditto.
6711 (_mm512_cmplt_epi32_mask): Ditto.
6712 (_mm512_cmplt_epu32_mask): Ditto.
6713 (_mm512_cmplt_epi64_mask): Ditto.
6714 (_mm512_cmplt_epu64_mask): Ditto.
6715 (_mm512_cmpneq_epi32_mask): Ditto.
6716 (_mm512_cmpneq_epu32_mask): Ditto.
6717 (_mm512_cmpneq_epi64_mask): Ditto.
6718 (_mm512_cmpneq_epu64_mask): Ditto.
6719 (_mm512_expand_pd): Ditto.
6720 (_mm512_expand_ps): Ditto.
6721 * config/i386/i386-builtin-types.def: Add PV16QI, PV16QI, PV16HI,
6722 VOID_PV8SI_V8DI_QI, VOID_PV8HI_V8DI_QI, VOID_PV16QI_V8DI_QI,
6723 VOID_PV16QI_V16SI_HI, VOID_PV16HI_V16SI_HI.
6724 * config/i386/i386.c (ix86_builtins): Add
6725 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK,
6726 IX86_BUILTIN_PMOVDB512_MEM, IX86_BUILTIN_PMOVDW512_MEM,
6727 IX86_BUILTIN_PMOVQB512_MEM, IX86_BUILTIN_PMOVQD512_MEM,
6728 IX86_BUILTIN_PMOVQW512_MEM, IX86_BUILTIN_PMOVSDB512_MEM,
6729 IX86_BUILTIN_PMOVSDW512_MEM, IX86_BUILTIN_PMOVSQB512_MEM,
6730 IX86_BUILTIN_PMOVSQD512_MEM, IX86_BUILTIN_PMOVSQW512_MEM,
6731 IX86_BUILTIN_PMOVUSDB512_MEM, IX86_BUILTIN_PMOVUSDW512_MEM,
6732 IX86_BUILTIN_PMOVUSQB512_MEM, IX86_BUILTIN_PMOVUSQD512_MEM,
6733 IX86_BUILTIN_PMOVUSQW512_MEM.
6734 (bdesc_special_args): Add __builtin_ia32_pmovusqd512mem_mask,
6735 __builtin_ia32_pmovsqd512mem_mask,
6736 __builtin_ia32_pmovqd512mem_mask,
6737 __builtin_ia32_pmovusqw512mem_mask,
6738 __builtin_ia32_pmovsqw512mem_mask,
6739 __builtin_ia32_pmovqw512mem_mask,
6740 __builtin_ia32_pmovusdw512mem_mask,
6741 __builtin_ia32_pmovsdw512mem_mask,
6742 __builtin_ia32_pmovdw512mem_mask,
6743 __builtin_ia32_pmovqb512mem_mask,
6744 __builtin_ia32_pmovusqb512mem_mask,
6745 __builtin_ia32_pmovsqb512mem_mask,
6746 __builtin_ia32_pmovusdb512mem_mask,
6747 __builtin_ia32_pmovsdb512mem_mask,
6748 __builtin_ia32_pmovdb512mem_mask.
6749 (bdesc_args): Add __builtin_ia32_expanddf512,
6750 __builtin_ia32_expandsf512.
6751 (ix86_expand_special_args_builtin): Handle VOID_FTYPE_PV8SI_V8DI_QI,
6752 VOID_FTYPE_PV8HI_V8DI_QI, VOID_FTYPE_PV16HI_V16SI_HI,
6753 VOID_FTYPE_PV16QI_V8DI_QI, VOID_FTYPE_PV16QI_V16SI_HI.
6754 * config/i386/sse.md (unspec): Add UNSPEC_EXPAND_NOMASK.
6755 (avx512f_<code><pmov_src_lower><mode>2_mask_store): New.
6756 (*avx512f_<code>v8div16qi2_store_mask): Renamed to ...
6757 (avx512f_<code>v8div16qi2_mask_store): This.
6758 (avx512f_expand<mode>): New.
6760 2014-01-27 Kirill Yukhin <kirill.yukhin@intel.com>
6762 * config/i386/avx512pfintrin.h (_mm512_mask_prefetch_i32gather_pd):
6764 (_mm512_mask_prefetch_i64gather_pd): Ditto.
6765 (_mm512_prefetch_i32scatter_pd): Ditto.
6766 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
6767 (_mm512_prefetch_i64scatter_pd): Ditto.
6768 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
6769 (_mm512_mask_prefetch_i32gather_ps): Fix operand type.
6770 (_mm512_mask_prefetch_i64gather_ps): Ditto.
6771 (_mm512_prefetch_i32scatter_ps): Ditto.
6772 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
6773 (_mm512_prefetch_i64scatter_ps): Ditto.
6774 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
6775 * config/i386/i386-builtin-types.def: Define
6776 VOID_FTYPE_QI_V8SI_PCINT64_INT_INT
6777 and VOID_FTYPE_QI_V8DI_PCINT64_INT_INT.
6778 * config/i386/i386.c (ix86_builtins): Define IX86_BUILTIN_GATHERPFQPD,
6779 IX86_BUILTIN_GATHERPFDPD, IX86_BUILTIN_SCATTERPFDPD,
6780 IX86_BUILTIN_SCATTERPFQPD.
6781 (ix86_init_mmx_sse_builtins): Define __builtin_ia32_gatherpfdpd,
6782 __builtin_ia32_gatherpfdps, __builtin_ia32_gatherpfqpd,
6783 __builtin_ia32_gatherpfqps, __builtin_ia32_scatterpfdpd,
6784 __builtin_ia32_scatterpfdps, __builtin_ia32_scatterpfqpd,
6785 __builtin_ia32_scatterpfqps.
6786 (ix86_expand_builtin): Expand new built-ins.
6787 * config/i386/sse.md (avx512pf_gatherpf<mode>): Add SF suffix,
6788 fix memory access data type.
6789 (*avx512pf_gatherpf<mode>_mask): Ditto.
6790 (*avx512pf_gatherpf<mode>): Ditto.
6791 (avx512pf_scatterpf<mode>): Ditto.
6792 (*avx512pf_scatterpf<mode>_mask): Ditto.
6793 (*avx512pf_scatterpf<mode>): Ditto.
6794 (GATHER_SCATTER_SF_MEM_MODE): New.
6795 (avx512pf_gatherpf<mode>df): Ditto.
6796 (*avx512pf_gatherpf<mode>df_mask): Ditto.
6797 (*avx512pf_scatterpf<mode>df): Ditto.
6799 2014-01-27 Jakub Jelinek <jakub@redhat.com>
6802 * expmed.h (expmed_mode_index): Rework so that analysis and optimziers
6803 know when the MODE_PARTIAL_INT and MODE_VECTOR_INT cases can never be
6806 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
6808 * common/config/arm/arm-common.c
6809 (arm_rewrite_mcpu): Handle multiple names.
6811 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
6813 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
6815 * gimple-builder.h (create_gimple_tmp): Delete.
6817 2014-01-27 Christian Bruel <christian.bruel@st.com>
6819 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Fix remaining bytes after
6822 2014-01-26 John David Anglin <danglin@gcc.gnu.org>
6824 * config/pa/pa.md (call): Generate indirect long calls to non-local
6825 functions when outputing 32-bit code.
6826 (call_value): Likewise except for special call to buggy powf function.
6828 * config/pa/pa.c (pa_attr_length_indirect_call): Adjust length of
6829 portable runtime and PIC indirect calls.
6830 (pa_output_indirect_call): Remove unnecessary nop from portable runtime
6831 and PIC call sequences. Use ldo instead of blr to set return register
6832 in PIC call sequence.
6834 2014-01-25 Walter Lee <walt@tilera.com>
6836 * config/tilegx/sync.md (atomic_fetch_sub): Fix negation and
6837 avoid clobbering a live register.
6839 2014-01-25 Walter Lee <walt@tilera.com>
6841 * config/tilegx/tilegx-c.c (tilegx_cpu_cpp_builtins):
6842 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2}.
6843 * config/tilegx/tilepro-c.c (tilepro_cpu_cpp_builtins):
6844 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2,4,8}.
6846 2014-01-25 Walter Lee <walt@tilera.com>
6848 * config/tilegx/tilegx.c (tilegx_function_arg): Start 16-byte
6849 arguments on even registers.
6850 (tilegx_gimplify_va_arg_expr): Align 16-byte var args to
6852 * config/tilegx/tilegx.h (STACK_BOUNDARY): Change to 16 bytes.
6853 (BIGGEST_ALIGNMENT): Ditto.
6854 (BIGGEST_FIELD_ALIGNMENT): Ditto.
6856 2014-01-25 Walter Lee <walt@tilera.com>
6858 * config/tilegx/tilegx.c (tilegx_gen_bundles): Delete barrier
6859 insns before bundling.
6860 * config/tilegx/tilegx.md (tile_network_barrier): Update comment.
6862 2014-01-25 Walter Lee <walt@tilera.com>
6864 * config/tilegx/tilegx.c (tilegx_expand_builtin): Set
6865 PREFETCH_SCHEDULE_BARRIER_P to true for prefetches.
6866 * config/tilepro/tilepro.c (tilepro_expand_builtin): Ditto.
6868 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
6870 * config/mips/constraints.md (kl): Delete.
6871 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Turn into
6872 define expands, using...
6873 (divmod<mode>4_mips16, udivmod<mode>4_mips16): ...these new
6874 instructions for MIPS16.
6875 (*divmod<mode>4, *udivmod<mode>4): New patterns, taken from the
6876 non-MIPS16 version of the old divmod<mode>4 and udivmod<mode>4.
6878 2014-01-25 Walter Lee <walt@tilera.com>
6880 * config/tilepro/tilepro.md (ctzdi2): Use register_operand predicate.
6884 2014-01-25 Walter Lee <walt@tilera.com>
6886 * config/tilegx/tilegx.c (tilegx_expand_to_rtl_hook): New.
6887 (TARGET_EXPAND_TO_RTL_HOOK): Define.
6889 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
6891 * rtlanal.c (canonicalize_condition): Split out duplicated mode check.
6894 2014-01-25 Jakub Jelinek <jakub@redhat.com>
6896 * print-rtl.c (in_call_function_usage): New var.
6897 (print_rtx): When in CALL_INSN_FUNCTION_USAGE, always print
6898 EXPR_LIST mode as mode and not as reg note name.
6901 * cfgloopmanip.c (copy_loop_info): If
6902 loop->warned_aggressive_loop_optimizations, make sure
6903 the flag is set in target loop too.
6905 2014-01-24 Balaji V. Iyer <balaji.v.iyer@intel.com>
6907 * builtins.c (is_builtin_name): Renamed flag_enable_cilkplus to
6909 * builtins.def: Likewise.
6910 * cilk.h (fn_contains_cilk_spawn_p): Likewise.
6911 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Likewise.
6912 * ira.c (ira_setup_eliminable_regset): Likewise.
6913 * omp-low.c (gate_expand_omp): Likewise.
6914 (execute_lower_omp): Likewise.
6915 (diagnose_sb_0): Likewise.
6916 (gate_diagnose_omp_blocks): Likewise.
6917 (simd_clone_clauses_extract): Likewise.
6920 2014-01-24 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6922 * config/rs6000/rs6000.c (rs6000_expand_vec_perm_const_1): Remove
6923 correction for little endian...
6924 * config/rs6000/vsx.md (vsx_xxpermdi2_<mode>_1): ...and move it to
6927 2014-01-24 Jeff Law <law@redhat.com>
6929 PR tree-optimization/59919
6930 * tree-vrp.c (find_assert_locations_1): Do not register asserts
6931 for non-returning calls.
6933 2014-01-24 James Greenhalgh <james.greenhalgh@arm.com>
6935 * common/config/aarch64/aarch64-common.c
6936 (aarch64_rewrite_mcpu): Handle multiple names.
6937 * config/aarch64/aarch64.h
6938 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
6940 2014-01-24 Dodji Seketeli <dodji@redhat.com>
6942 * input.c (add_file_to_cache_tab): Handle the case where fopen
6945 2014-01-23 H.J. Lu <hongjiu.lu@intel.com>
6948 * config/i386/i386.md (pushsf splitter): Get stack adjustment
6949 from push operand if code of push isn't PRE_DEC.
6951 2014-01-23 Michael Meissner <meissner@linux.vnet.ibm.com>
6954 * doc/invoke.texi (RS/6000 and PowerPC Options): Document
6955 -mquad-memory-atomic. Update -mquad-memory documentation to say
6956 it is only used for non-atomic loads/stores.
6958 * config/rs6000/predicates.md (quad_int_reg_operand): Allow either
6959 -mquad-memory or -mquad-memory-atomic switches.
6961 * config/rs6000/rs6000-cpus.def (ISA_2_7_MASKS_SERVER): Add
6962 -mquad-memory-atomic to ISA 2.07 support.
6964 * config/rs6000/rs6000.opt (-mquad-memory-atomic): Add new switch
6965 to separate support of normal quad word memory operations (ldq, stq)
6966 from the atomic quad word memory operations.
6968 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
6969 support to separate non-atomic quad word operations from atomic
6970 quad word operations. Disable non-atomic quad word operations in
6971 little endian mode so that we don't have to swap words after the
6972 load and before the store.
6973 (quad_load_store_p): Add comment about atomic quad word support.
6974 (rs6000_opt_masks): Add -mquad-memory-atomic to the list of
6975 options printed with -mdebug=reg.
6977 * config/rs6000/rs6000.h (TARGET_SYNC_TI): Use
6978 -mquad-memory-atomic as the test for whether we have quad word
6979 atomic instructions.
6980 (TARGET_SYNC_HI_QI): If either -mquad-memory-atomic, -mquad-memory,
6981 or -mp8-vector are used, allow byte/half-word atomic operations.
6983 * config/rs6000/sync.md (load_lockedti): Insure that the address
6984 is a proper indexed or indirect address for the lqarx instruction.
6985 On little endian systems, swap the hi/lo registers after the lqarx
6987 (load_lockedpti): Use indexed_or_indirect_operand predicate to
6988 insure the address is valid for the lqarx instruction.
6989 (store_conditionalti): Insure that the address is a proper indexed
6990 or indirect address for the stqcrx. instruction. On little endian
6991 systems, swap the hi/lo registers before doing the stqcrx.
6993 (store_conditionalpti): Use indexed_or_indirect_operand predicate to
6994 insure the address is valid for the stqcrx. instruction.
6996 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
6997 Define __QUAD_MEMORY__ and __QUAD_MEMORY_ATOMIC__ based on what
6998 type of quad memory support is available.
7000 2014-01-23 Vladimir Makarov <vmakarov@redhat.com>
7003 * lra-constraints.c (simplify_operand_subreg): Spill pseudo if
7004 there is a danger of looping.
7006 2014-01-23 Pat Haugen <pthaugen@us.ibm.com>
7008 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
7009 force flag_ira_loop_pressure if set via command line.
7011 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
7013 * config/aarch64/aarch64-simd-builtins.def (ashr): DI mode removed.
7014 (ashr_simd): New builtin handling DI mode.
7015 * config/aarch64/aarch64-simd.md (aarch64_ashr_simddi): New pattern.
7016 (aarch64_sshr_simddi): New match pattern.
7017 * config/aarch64/arm_neon.h (vshr_n_s32): Builtin call modified.
7018 (vshrd_n_s64): Likewise.
7019 * config/aarch64/predicates.md (aarch64_shift_imm64_di): New predicate.
7021 2014-01-23 Nick Clifton <nickc@redhat.com>
7023 * config/msp430/msp430.h (ASM_SPEC): Pass the -mcpu as -mcpu.
7024 (LIB_SPEC): Drop use of memory.ld and peripherals.ld scripts in
7025 favour of mcu specific scripts.
7026 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more matches for
7029 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
7030 Alex Velenko <Alex.Velenko@arm.com>
7032 * config/aarch64/arm_neon.h (vaddv_s8): __LANE0 cleanup.
7033 (vaddv_s16): Likewise.
7034 (vaddv_s32): Likewise.
7035 (vaddv_u8): Likewise.
7036 (vaddv_u16): Likewise.
7037 (vaddv_u32): Likewise.
7038 (vaddvq_s8): Likewise.
7039 (vaddvq_s16): Likewise.
7040 (vaddvq_s32): Likewise.
7041 (vaddvq_s64): Likewise.
7042 (vaddvq_u8): Likewise.
7043 (vaddvq_u16): Likewise.
7044 (vaddvq_u32): Likewise.
7045 (vaddvq_u64): Likewise.
7046 (vaddv_f32): Likewise.
7047 (vaddvq_f32): Likewise.
7048 (vaddvq_f64): Likewise.
7049 (vmaxv_f32): Likewise.
7050 (vmaxv_s8): Likewise.
7051 (vmaxv_s16): Likewise.
7052 (vmaxv_s32): Likewise.
7053 (vmaxv_u8): Likewise.
7054 (vmaxv_u16): Likewise.
7055 (vmaxv_u32): Likewise.
7056 (vmaxvq_f32): Likewise.
7057 (vmaxvq_f64): Likewise.
7058 (vmaxvq_s8): Likewise.
7059 (vmaxvq_s16): Likewise.
7060 (vmaxvq_s32): Likewise.
7061 (vmaxvq_u8): Likewise.
7062 (vmaxvq_u16): Likewise.
7063 (vmaxvq_u32): Likewise.
7064 (vmaxnmv_f32): Likewise.
7065 (vmaxnmvq_f32): Likewise.
7066 (vmaxnmvq_f64): Likewise.
7067 (vminv_f32): Likewise.
7068 (vminv_s8): Likewise.
7069 (vminv_s16): Likewise.
7070 (vminv_s32): Likewise.
7071 (vminv_u8): Likewise.
7072 (vminv_u16): Likewise.
7073 (vminv_u32): Likewise.
7074 (vminvq_f32): Likewise.
7075 (vminvq_f64): Likewise.
7076 (vminvq_s8): Likewise.
7077 (vminvq_s16): Likewise.
7078 (vminvq_s32): Likewise.
7079 (vminvq_u8): Likewise.
7080 (vminvq_u16): Likewise.
7081 (vminvq_u32): Likewise.
7082 (vminnmv_f32): Likewise.
7083 (vminnmvq_f32): Likewise.
7084 (vminnmvq_f64): Likewise.
7086 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
7088 * config/aarch64/aarch64-simd.md
7089 (aarch64_dup_lane<mode>): Correct lane number on big-endian.
7090 (aarch64_dup_lane_<vswap_widthi_name><mode>): Likewise.
7091 (*aarch64_mul3_elt<mode>): Likewise.
7092 (*aarch64_mul3_elt<vswap_width_name><mode>): Likewise.
7093 (*aarch64_mul3_elt_to_64v2df): Likewise.
7094 (*aarch64_mla_elt<mode>): Likewise.
7095 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
7096 (*aarch64_mls_elt<mode>): Likewise.
7097 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
7098 (*aarch64_fma4_elt<mode>): Likewise.
7099 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
7100 (*aarch64_fma4_elt_to_64v2df): Likewise.
7101 (*aarch64_fnma4_elt<mode>): Likewise.
7102 (*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
7103 (*aarch64_fnma4_elt_to_64v2df): Likewise.
7104 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
7105 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
7106 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
7107 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
7108 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
7109 (aarch64_sqdmull_lane<mode>_internal): Likewise.
7110 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
7112 2013-01-23 Alex Velenko <Alex.Velenko@arm.com>
7114 * config/aarch64/aarch64-simd.md
7115 (aarch64_be_checked_get_lane<mode>): New define_expand.
7116 * config/aarch64/aarch64-simd-builtins.def
7117 (BUILTIN_VALL (GETLANE, be_checked_get_lane, 0)):
7118 New builtin definition.
7119 * config/aarch64/arm_neon.h: (__aarch64_vget_lane_any):
7120 Use new safe be builtin.
7122 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
7124 * config/aarch64/aarch64-simd.md (aarch64_be_ld1<mode>):
7126 (aarch64_be_st1<mode>): Likewise.
7127 (aarch_ld1<VALL:mode>): Define_expand modified.
7128 (aarch_st1<VALL:mode>): Likewise.
7129 * config/aarch64/aarch64.md (UNSPEC_LD1): New unspec definition.
7130 (UNSPEC_ST1): Likewise.
7132 2014-01-23 David Holsgrove <david.holsgrove@xilinx.com>
7134 * config/microblaze/microblaze.md: Add trap insn and attribute
7136 2014-01-23 Dodji Seketeli <dodji@redhat.com>
7138 PR preprocessor/58580
7139 * input.h (location_get_source_line): Take an additional line_size
7141 (void diagnostics_file_cache_fini): Declare new function.
7142 * input.c (struct fcache): New type.
7143 (fcache_tab_size, fcache_buffer_size, fcache_line_record_size):
7144 New static constants.
7145 (diagnostic_file_cache_init, total_lines_num)
7146 (lookup_file_in_cache_tab, evicted_cache_tab_entry)
7147 (add_file_to_cache_tab, lookup_or_add_file_to_cache_tab)
7148 (needs_read, needs_grow, maybe_grow, read_data, maybe_read_data)
7149 (get_next_line, read_next_line, goto_next_line, read_line_num):
7150 New static function definitions.
7151 (diagnostic_file_cache_fini): New function.
7152 (location_get_source_line): Take an additional output line_len
7153 parameter. Re-write using lookup_or_add_file_to_cache_tab and
7155 * diagnostic.c (diagnostic_finish): Call
7156 diagnostic_file_cache_fini.
7157 (adjust_line): Take an additional input parameter for the length
7158 of the line, rather than calculating it with strlen.
7159 (diagnostic_show_locus): Adjust the use of
7160 location_get_source_line and adjust_line with respect to their new
7161 signature. While displaying a line now, do not stop at the first
7162 null byte. Rather, display the zero byte as a space and keep
7163 going until we reach the size of the line.
7164 * Makefile.in: Add vec.o to OBJS-libcommon
7166 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
7167 Ilya Tocar <ilya.tocar@intel.com>
7169 * config/i386/avx512fintrin.h (_mm512_kmov): New.
7170 * config/i386/i386.c (IX86_BUILTIN_KMOV16): Ditto.
7171 (__builtin_ia32_kmov16): Ditto.
7172 * config/i386/i386.md (UNSPEC_KMOV): New.
7175 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
7177 * config/i386/avx512fintrin.h (_mm512_loadu_si512): Rename.
7178 (_mm512_storeu_si512): Ditto.
7180 2014-01-23 Richard Sandiford <rdsandiford@googlemail.com>
7183 * rtl.h (get_referenced_operands): Declare.
7184 * recog.c (get_referenced_operands): New function.
7185 * config/mips/mips.c (mips_reorg_process_insns): Check which asm
7186 operands have been referenced when recording LO_SUM references.
7188 2014-01-22 David Holsgrove <david.holsgrove@xilinx.com>
7190 * config/microblaze/microblaze.md: Correct bswaphi2 insn.
7192 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
7194 * config/i386/x86-tune.def (X86_TUNE_ACCUMULATE_OUTGOING_ARGS):
7195 Enable for generic and recent AMD targets.
7197 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
7199 * combine-stack-adj.c (combine_stack_adjustments_for_block): Remove
7200 ARG_SIZE note when adjustment was eliminated.
7202 2014-01-22 Jeff Law <law@redhat.com>
7204 PR tree-optimization/59597
7205 * tree-ssa-threadupdate.c (dump_jump_thread_path): Move to earlier
7206 in file. Accept new argument REGISTERING and use it to modify
7207 dump output appropriately.
7208 (register_jump_thread): Corresponding changes.
7209 (mark_threaded_blocks): Reinstate code to cancel unprofitable
7210 thread paths involving joiner blocks. Add code to dump cancelled
7211 jump threading paths.
7213 2014-01-22 Vladimir Makarov <vmakarov@redhat.com>
7215 PR rtl-optimization/59477
7216 * lra-constraints.c (inherit_in_ebb): Process call for living hard
7217 regs. Update reloads_num and potential_reload_hard_regs for all insns.
7219 2014-01-22 Tom Tromey <tromey@redhat.com>
7221 * config/i386/i386-interix.h (i386_pe_unique_section): Don't use
7223 * config/cr16/cr16-protos.h (notice_update_cc): Don't use PARAMS.
7225 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
7227 PR rtl-optimization/59896
7228 * lra-constraints.c (process_alt_operands): Check unused note for
7229 matched operands of insn with no output reloads.
7231 2014-01-21 Richard Sandiford <rdsandiford@googlemail.com>
7233 * config/mips/mips.c (mips_move_to_gpr_cost): Add M16_REGS case.
7234 (mips_move_from_gpr_cost): Likewise.
7236 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
7238 PR rtl-optimization/59858
7239 * lra-constraints.c (SMALL_REGISTER_CLASS_P): Use
7240 ira_class_hard_regs_num.
7241 (process_alt_operands): Increase reject for dying matched operand.
7243 2014-01-21 Jakub Jelinek <jakub@redhat.com>
7246 * config/i386/i386.c (expand_small_movmem_or_setmem): If mode is
7247 smaller than size, perform several stores or loads and stores
7248 at dst + count - size to store or copy all of size bytes, rather
7249 than just last modesize bytes.
7251 2014-01-20 DJ Delorie <dj@redhat.com>
7253 * config/rl78/rl78.c (rl78_propogate_register_origins): Verify
7254 that CLOBBERs are REGs before propogating their values.
7256 2014-01-20 H.J. Lu <hongjiu.lu@intel.com>
7259 * cgraph.c (cgraph_inline_failed_string): Add type to DEFCIFCODE.
7260 (cgraph_inline_failed_type): New function.
7261 * cgraph.h (DEFCIFCODE): Add type.
7262 (cgraph_inline_failed_type_t): New enum.
7263 (cgraph_inline_failed_type): New prototype.
7264 * cif-code.def: Add CIF_FINAL_NORMAL to OK, FUNCTION_NOT_CONSIDERED,
7265 FUNCTION_NOT_OPTIMIZED, REDEFINED_EXTERN_INLINE,
7266 FUNCTION_NOT_INLINE_CANDIDATE, LARGE_FUNCTION_GROWTH_LIMIT,
7267 LARGE_STACK_FRAME_GROWTH_LIMIT, MAX_INLINE_INSNS_SINGLE_LIMIT,
7268 MAX_INLINE_INSNS_AUTO_LIMIT, INLINE_UNIT_GROWTH_LIMIT,
7269 RECURSIVE_INLINING, UNLIKELY_CALL, NOT_DECLARED_INLINED,
7270 OPTIMIZING_FOR_SIZE, ORIGINALLY_INDIRECT_CALL,
7271 INDIRECT_UNKNOWN_CALL, USES_COMDAT_LOCAL.
7272 Add CIF_FINAL_ERROR to UNSPECIFIED, BODY_NOT_AVAILABLE,
7273 FUNCTION_NOT_INLINABLE, OVERWRITABLE, MISMATCHED_ARGUMENTS,
7274 EH_PERSONALITY, NON_CALL_EXCEPTIONS, TARGET_OPTION_MISMATCH,
7275 OPTIMIZATION_MISMATCH.
7276 * tree-inline.c (expand_call_inline): Emit errors during
7277 early_inlining if cgraph_inline_failed_type returns CIF_FINAL_ERROR.
7279 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
7282 * config/i386/sse.md (*andnot<mode>3<mask_name>): Handle MODE_V16SF
7283 mode attribute in insn output.
7285 2014-01-20 Eric Botcazou <ebotcazou@adacore.com>
7287 * output.h (output_constant): Delete.
7288 * varasm.c (output_constant): Make private.
7290 2014-01-20 Alex Velenko <Alex.Velenko@arm.com>
7292 * config/aarch64/aarch64-simd.md (vec_perm<mode>): Add BE check.
7294 2014-01-20 Jakub Jelinek <jakub@redhat.com>
7297 * tree.h (fold_builtin_strcat): New prototype.
7298 * builtins.c (fold_builtin_strcat): No longer static. Add len
7299 argument, if non-NULL, don't call c_strlen. Optimize
7300 directly into __builtin_memcpy instead of __builtin_strcpy.
7301 (fold_builtin_2): Adjust fold_builtin_strcat caller.
7302 * gimple-fold.c (gimple_fold_builtin): Handle BUILT_IN_STRCAT.
7304 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
7306 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
7307 for SImode_address_operand operands, having only a REG argument.
7309 2014-01-20 Marcus Shawcroft <marcus.shawcroft@arm.com>
7311 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER): Expand
7312 loader name using mbig-endian.
7313 (LINUX_TARGET_LINK_SPEC): Pass linker -m flag.
7315 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
7317 * doc/invoke.texi (-march): Clarify documentation for AArch64.
7321 2014-01-20 Tejas Belagod <tejas.belagod@arm.com>
7323 * config/aarch64/aarch64-protos.h
7324 (aarch64_cannot_change_mode_class_ptr): Declare.
7325 * config/aarch64/aarch64.c (aarch64_cannot_change_mode_class,
7326 aarch64_cannot_change_mode_class_ptr): New.
7327 * config/aarch64/aarch64.h (CANNOT_CHANGE_MODE_CLASS): Change to call
7328 backend hook aarch64_cannot_change_mode_class.
7330 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
7332 * common/config/aarch64/aarch64-common.c
7333 (aarch64_handle_option): Don't handle any option order logic here.
7334 * config/aarch64/aarch64.c (aarch64_parse_arch): Do not override
7335 selected_cpu, warn on architecture version mismatch.
7336 (aarch64_override_options): Fix parsing order for option strings.
7338 2014-01-20 Jan-Benedict Glaw <jbglaw@lug-owl.de>
7339 Iain Sandoe <iain@codesourcery.com>
7342 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Fix unused variable
7343 warning. Amend comment to reflect current functionality.
7345 2014-01-20 Richard Biener <rguenther@suse.de>
7348 * builtins.c (fold_builtin_strcat): Remove case better handled
7349 by tree-ssa-strlen.c.
7351 2014-01-20 Alan Lawrence <alan.lawrence@arm.com>
7353 * config/aarch64/aarch64.opt
7354 (mcpu, march, mtune): Make case-insensitive.
7356 2014-01-20 Jakub Jelinek <jakub@redhat.com>
7359 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
7360 if operands[1] is a REG or ZERO_EXTEND of a REG.
7362 2014-01-19 Jan Hubicka <hubicka@ucw.cz>
7364 * varasm.c (compute_reloc_for_constant): Use targetm.binds_local_p.
7366 2014-01-19 John David Anglin <danglin@gcc.gnu.org>
7368 * config/pa/pa.c (pa_attr_length_millicode_call): Correct length of
7369 long non-pic millicode calls.
7371 2014-01-19 Jan-Benedict Glaw <jbglaw@lug-owl.de>
7373 * config/vax/vax.h (FUNCTION_ARG_REGNO_P): Fix unused variable warning.
7375 2014-01-19 Kito Cheng <kito@0xlab.org>
7377 * builtins.c (expand_movstr): Check movstr expand done or fail.
7379 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
7380 H.J. Lu <hongjiu.lu@intel.com>
7383 * config/i386/i386.md (*lea<mode>): Zero-extend return register
7384 to DImode for zero-extended addresses.
7386 2014-01-19 Jakub Jelinek <jakub@redhat.com>
7388 PR rtl-optimization/57763
7389 * bb-reorder.c (fix_crossing_unconditional_branches): Set JUMP_LABEL
7390 on the new indirect jump_insn and increment LABEL_NUSES (label).
7392 2014-01-18 H.J. Lu <hongjiu.lu@intel.com>
7396 * config.gcc (x86_archs): New variable.
7397 (x86_64_archs): Likewise.
7398 (x86_cpus): Likewise.
7399 Use $x86_archs, $x86_64_archs and $x86_cpus to check valid
7400 --with-arch/--with-cpu= options.
7401 Support --with-arch=/--with-cpu={nehalem,westmere,
7402 sandybridge,ivybridge,haswell,broadwell,bonnell,silvermont}.
7404 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
7406 * config/i386/i386.c (ix86_adjust_cost): Reorder PROCESSOR_K8
7407 and PROCESSOR_ATHLON to simplify code. Move "memory" calculation.
7409 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
7411 * config/i386/i386.md (*swap<mode>): Rename from swap<mode>.
7413 2014-01-18 Jakub Jelinek <jakub@redhat.com>
7416 * config/i386/i386-c.c (ix86_pragma_target_parse): Temporarily
7417 clear cpp_get_options (parse_in)->warn_unused_macros for
7418 ix86_target_macros_internal with cpp_define.
7420 2014-01-18 Richard Sandiford <rdsandiford@googlemail.com>
7422 * jump.c (delete_related_insns): Keep (use (insn))s.
7423 * reorg.c (redundant_insn): Check for barriers too.
7425 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
7427 * config/i386/i386.c (ix86_split_lea_for_addr): Fix a comment typo.
7429 2014-01-17 John David Anglin <danglin@gcc.gnu.org>
7431 * config/pa/pa.c (pa_attr_length_indirect_call): Don't output a short
7432 call to $$dyncall when TARGET_LONG_CALLS is true.
7434 2014-01-17 Jeff Law <law@redhat.com>
7436 * ree.c (combine_set_extension): Temporarily disable test for
7437 changing number of hard registers.
7439 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
7442 * ipa-inline-analysis.c (inline_free_summary):
7443 Do not free summary of aliases.
7445 2014-01-17 Jakub Jelinek <jakub@redhat.com>
7448 * gimplify.c (gimplify_expr): Use create_tmp_var
7449 instead of create_tmp_var_raw. If cond doesn't have
7450 integral type, don't add the IFN_ANNOTATE builtin at all.
7452 2014-01-17 Martin Jambor <mjambor@suse.cz>
7455 * ipa-cp.c (prev_edge_clone): New variable.
7456 (grow_next_edge_clone_vector): Renamed to grow_edge_clone_vectors.
7457 Also resize prev_edge_clone vector.
7458 (ipcp_edge_duplication_hook): Also update prev_edge_clone.
7459 (ipcp_edge_removal_hook): New function.
7460 (ipcp_driver): Register ipcp_edge_removal_hook.
7462 2014-01-17 Andrew Pinski <apinski@cavium.com>
7463 Steve Ellcey <sellcey@mips.com>
7466 * config/mips/mips.c (mips_print_operand): Check operand mode instead
7469 2014-01-17 Jeff Law <law@redhat.com>
7472 * passes.def: Reorder pass_copy_prop, pass_unrolli, pass_ccp sequence
7473 so that pass_ccp runs first.
7475 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
7477 * config/i386/i386.c (ix86_lea_outperforms): Use TARGET_XXX.
7478 (ix86_adjust_cost): Use !TARGET_XXX.
7479 (do_reorder_for_imul): Likewise.
7480 (swap_top_of_ready_list): Likewise.
7481 (ix86_sched_reorder): Likewise.
7483 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
7485 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
7486 PROCESSOR_INTEL. Treat like PROCESSOR_GENERIC.
7487 * config/i386/i386.c (intel_memcpy): New. Duplicate slm_memcpy.
7488 (intel_memset): New. Duplicate slm_memset.
7489 (intel_cost): New. Duplicate slm_cost.
7490 (m_INTEL): New macro.
7491 (processor_target_table): Add "intel".
7492 (ix86_option_override_internal): Replace PROCESSOR_SILVERMONT
7493 with PROCESSOR_INTEL for "intel".
7494 (ix86_lea_outperforms): Support PROCESSOR_INTEL. Duplicate
7495 PROCESSOR_SILVERMONT.
7496 (ix86_issue_rate): Likewise.
7497 (ix86_adjust_cost): Likewise.
7498 (ia32_multipass_dfa_lookahead): Likewise.
7499 (swap_top_of_ready_list): Likewise.
7500 (ix86_sched_reorder): Likewise.
7501 (ix86_avoid_lea_for_addr): Check TARGET_AVOID_LEA_FOR_ADDR
7502 instead of TARGET_OPT_AGU.
7503 * config/i386/i386.h (TARGET_INTEL): New.
7504 (TARGET_AVOID_LEA_FOR_ADDR): Likewise.
7505 (processor_type): Add PROCESSOR_INTEL.
7506 * config/i386/x86-tune.def: Support m_INTEL. Duplicate m_SILVERMONT.
7507 Add X86_TUNE_AVOID_LEA_FOR_ADDR.
7509 2014-01-17 Marek Polacek <polacek@redhat.com>
7512 * gimple-fold.c (fold_array_ctor_reference): Don't fold if element
7515 2014-01-17 Richard Biener <rguenther@suse.de>
7517 PR tree-optimization/46590
7518 * opts.c (default_options_table): Add entries for
7519 OPT_fbranch_count_reg, OPT_fmove_loop_invariants and OPT_ftree_pta,
7520 all enabled at -O1 but not for -Og.
7521 * common.opt (fbranch-count-reg): Remove Init(1).
7522 (fmove-loop-invariants): Likewise.
7523 (ftree-pta): Likewise.
7525 2014-01-17 Jakub Jelinek <jakub@redhat.com>
7527 * config/i386/i386.c (ix86_data_alignment): For compatibility with
7528 (incorrect) GCC 4.8 and earlier alignment assumptions ensure we align
7529 decls to at least the GCC 4.8 used alignments.
7532 * tree-nested.c (convert_nonlocal_reference_stmt,
7533 convert_local_reference_stmt): For NAMELIST_DECLs in gimple_bind_vars
7534 of GIMPLE_BIND stmts, adjust associated decls.
7536 2014-01-17 Richard Biener <rguenther@suse.de>
7538 PR tree-optimization/46590
7539 * vec.h (vec<>::bseach): New member function implementing
7540 binary search according to C89 bsearch.
7541 (vec<>::qsort): Avoid calling ::qsort for vectors with sizes 0 or 1.
7542 * tree-ssa-loop-im.c (struct mem_ref): Make stored member a
7543 bitmap pointer again. Make accesses_in_loop a flat array.
7544 (mem_ref_obstack): New global.
7545 (outermost_indep_loop): Adjust for mem_ref->stored changes.
7546 (mark_ref_stored): Likewise.
7547 (ref_indep_loop_p_2): Likewise.
7548 (set_ref_stored_in_loop): New helper function.
7549 (mem_ref_alloc): Allocate mem_refs on the mem_ref_obstack obstack.
7550 (memref_free): Adjust.
7551 (record_mem_ref_loc): Simplify.
7552 (gather_mem_refs_stmt): Adjust.
7553 (sort_locs_in_loop_postorder_cmp): New function.
7554 (analyze_memory_references): Sort accesses_in_loop after
7555 loop postorder number.
7556 (find_ref_loc_in_loop_cmp): New function.
7557 (for_all_locs_in_loop): Find relevant cluster of locs in
7558 accesses_in_loop and iterate without recursion.
7559 (execute_sm): Avoid uninit warning.
7560 (struct ref_always_accessed): Simplify.
7561 (ref_always_accessed::operator ()): Likewise.
7562 (ref_always_accessed_p): Likewise.
7563 (tree_ssa_lim_initialize): Initialize mem_ref_obstack, compute
7564 loop postorder numbers here.
7565 (tree_ssa_lim_finalize): Free mem_ref_obstack and loop postorder
7568 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
7571 * passes.c (rest_of_decl_compilation): Don't call varpool_finalize_decl
7572 on decls for which assemble_alias has been called.
7574 2014-01-17 Nick Clifton <nickc@redhat.com>
7576 * config/msp430/msp430.opt: (mcpu): New option.
7577 * config/msp430/msp430.c (msp430_mcu_name): Use target_mcu.
7578 (msp430_option_override): Parse target_cpu. If the MCU name
7579 matches a generic string, clear target_mcu.
7580 (msp430_attr): Allow numeric interrupt values up to 63.
7581 (msp430_expand_epilogue): No longer invert operand 1 of gen_popm.
7582 * config/msp430/msp430.h (ASM_SPEC): Convert -mcpu into a -mmcu
7584 * config/msp430/t-msp430: (MULTILIB_MATCHES): Remove mcu matches.
7586 * config/msp430/msp430.md (popm): Use %J rather than %I.
7587 (addsi3): Use msp430_nonimmediate_operand for operand 2.
7588 (addhi_cy_i): Use immediate_operand for operand 2.
7589 * doc/invoke.texi: Document -mcpu option.
7591 2014-01-17 Richard Biener <rguenther@suse.de>
7593 PR rtl-optimization/38518
7594 * df.h (df_analyze_loop): Declare.
7595 * df-core.c: Include cfgloop.h.
7596 (df_analyze_1): Split out main part of df_analyze.
7597 (df_analyze): Adjust.
7598 (loop_inverted_post_order_compute): New function.
7599 (loop_post_order_compute): Likewise.
7600 (df_analyze_loop): New function avoiding whole-function
7602 * loop-invariant.c (find_defs): Use df_analyze_loop.
7603 (find_invariants): Adjust.
7604 * loop-iv.c (iv_analysis_loop_init): Use df_analyze_loop.
7606 2014-01-17 Zhenqiang Chen <zhenqiang.chen@arm.com>
7608 * config/arm/arm.c (arm_v7m_tune): Set max_insns_skipped to 2.
7609 (thumb2_final_prescan_insn): Set max to MAX_INSN_PER_IT_BLOCK.
7611 2014-01-16 Ilya Enkovich <ilya.enkovich@intel.com>
7613 * ipa-ref.c (ipa_remove_stmt_references): Fix references
7614 traversal when removing references.
7616 2014-01-16 Jan Hubicka <hubicka@ucw.cz>
7619 * tree.c (get_binfo_at_offset): Look harder for virtual bases.
7621 2014-01-16 Bernd Schmidt <bernds@codesourcery.com>
7624 * reload.c (find_reloads_address_1): Do not use RELOAD_OTHER when
7625 pushing a reload for an autoinc when we had previously reloaded an
7626 inner part of the address.
7628 2014-01-16 Jakub Jelinek <jakub@redhat.com>
7630 * tree-vectorizer.h (struct _loop_vec_info): Add no_data_dependencies
7632 (LOOP_VINFO_NO_DATA_DEPENDENCIES): Define.
7633 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Clear it
7634 when not giving up or versioning for alias only because of
7636 (vect_analyze_data_ref_dependences): Set to true.
7637 * tree-vect-stmts.c (hoist_defs_of_uses): Return false if def_stmt
7639 (vectorizable_load): Use LOOP_VINFO_NO_DATA_DEPENDENCIES instead of
7640 LOOP_REQUIRES_VERSIONING_FOR_ALIAS, add && !nested_in_vect_loop
7644 * expr.c (expand_expr_real_1): Handle init == NULL_TREE.
7647 * config/i386/i386.c (ix86_expand_builtin): If target doesn't satisfy
7648 operand 0 predicate for gathers, use a new pseudo as subtarget.
7650 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
7653 * lra-constraints.c (process_alt_operands): Add printing debug info.
7654 Check absence of input/output reloads for matched operands too.
7656 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
7658 PR rtl-optimization/59835
7659 * ira.c (ira_init_register_move_cost): Increase cost for
7662 2014-01-16 Alan Lawrence <alan.lawrence@arm.com>
7664 * config/arm/arm.opt (mcpu, march, mtune): Make case-insensitive.
7666 2014-01-16 Richard Earnshaw <rearnsha@arm.com>
7669 * aarch64.c (aarch64_split_128bit_move): Don't lookup REGNO on
7670 non-register objects. Use gen_(high/low)part more consistently.
7673 2014-01-16 Michael Meissner <meissner@linux.vnet.ibm.com>
7676 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Add little
7677 endian support, remove tests for WORDS_BIG_ENDIAN.
7678 (p8_mfvsrd_3_<mode>): Likewise.
7679 (reload_gpr_from_vsx<mode>): Likewise.
7680 (reload_gpr_from_vsxsf): Likewise.
7681 (p8_mfvsrd_4_disf): Likewise.
7683 2014-01-16 Richard Biener <rguenther@suse.de>
7685 PR rtl-optimization/46590
7686 * lcm.c (compute_antinout_edge): Use postorder iteration.
7687 (compute_laterin): Use inverted postorder iteration.
7689 2014-01-16 Nick Clifton <nickc@redhat.com>
7692 * varasm.c (output_constant): Return the number of bytes actually
7694 (output_constructor_array_range): Update the field size with the
7695 number of bytes emitted by output_constant.
7696 (output_constructor_regular_field): Likewise. Also do not
7697 complain if the total number of bytes emitted is now greater
7698 than the expected fieldpos.
7699 * output.h (output_constant): Update prototype and descriptive comment.
7701 2014-01-16 Marek Polacek <polacek@redhat.com>
7704 * cgraph.c (gimple_check_call_args): Don't use DECL_ARG_TYPE if
7705 it is error_mark_node.
7707 2014-01-15 Uros Bizjak <ubizjak@gmail.com>
7709 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use
7710 VALID_AVX256_REG_OR_OI_MODE.
7712 2014-01-15 Pat Haugen <pthaugen@us.ibm.com>
7714 * config/rs6000/rs6000.c (rs6000_output_function_prologue): Check if
7715 current procedure should be profiled.
7717 2014-01-15 Andrew Pinski <apinski@cavium.com>
7719 * config/aarch64/aarch64.c (aarch64_register_move_cost): Correct cost
7720 of moving from/to the STACK_REG register class.
7722 2014-01-15 Richard Henderson <rth@redhat.com>
7725 * reginfo.c (global_regs_decl): Globalize.
7726 * rtl.h (global_regs_decl): Declare.
7727 * ira.c (do_reload): Diagnose frame_pointer_needed and it
7728 reserved via global_regs.
7730 2014-01-15 Teresa Johnson <tejohnson@google.com>
7732 * tree-ssa-sccvn.c (visit_reference_op_call): Handle NULL vdef.
7734 2014-01-15 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
7736 * config/rs6000/altivec.md (mulv8hi3): Explicitly generate vmulesh
7737 and vmulosh rather than call gen_vec_widen_smult_*.
7738 (vec_widen_umult_even_v16qi): Test VECTOR_ELT_ORDER_BIG rather
7739 than BYTES_BIG_ENDIAN to determine use of even or odd instruction.
7740 (vec_widen_smult_even_v16qi): Likewise.
7741 (vec_widen_umult_even_v8hi): Likewise.
7742 (vec_widen_smult_even_v8hi): Likewise.
7743 (vec_widen_umult_odd_v16qi): Likewise.
7744 (vec_widen_smult_odd_v16qi): Likewise.
7745 (vec_widen_umult_odd_v8hi): Likewise.
7746 (vec_widen_smult_odd_v8hi): Likewise.
7747 (vec_widen_umult_hi_v16qi): Explicitly generate vmuleub and
7748 vmuloub rather than call gen_vec_widen_umult_*.
7749 (vec_widen_umult_lo_v16qi): Likewise.
7750 (vec_widen_smult_hi_v16qi): Explicitly generate vmulesb and
7751 vmulosb rather than call gen_vec_widen_smult_*.
7752 (vec_widen_smult_lo_v16qi): Likewise.
7753 (vec_widen_umult_hi_v8hi): Explicitly generate vmuleuh and vmulouh
7754 rather than call gen_vec_widen_umult_*.
7755 (vec_widen_umult_lo_v8hi): Likewise.
7756 (vec_widen_smult_hi_v8hi): Explicitly gnerate vmulesh and vmulosh
7757 rather than call gen_vec_widen_smult_*.
7758 (vec_widen_smult_lo_v8hi): Likewise.
7760 2014-01-15 Jeff Law <law@redhat.com>
7762 PR tree-optimization/59747
7763 * ree.c (find_and_remove_re): Properly handle case where a second
7764 eliminated extension requires widening a copy created for elimination
7765 of a prior extension.
7766 (combine_set_extension): Ensure that the number of hard regs needed
7767 for a destination register does not change when we widen it.
7769 2014-01-15 Sebastian Huber <sebastian.huber@embedded-brains.de>
7771 * config.gcc (*-*-rtems*): Add t-rtems to tmake_file.
7772 (arm*-*-uclinux*eabi*): Do not override an existing tmake_file.
7773 (arm*-*-eabi* | arm*-*-symbianelf* | arm*-*-rtems*): Likwise.
7774 (arm*-*-rtems*): Use t-rtems from existing tmake_file.
7775 (avr-*-rtems*): Likewise.
7776 (bfin*-rtems*): Likewise.
7777 (moxie-*-rtems*): Likewise.
7778 (h8300-*-rtems*): Likewise.
7779 (i[34567]86-*-rtems*): Likewise.
7780 (lm32-*-rtems*): Likewise.
7781 (m32r-*-rtems*): Likewise.
7782 (m68k-*-rtems*): Likewise.
7783 (microblaze*-*-rtems*): Likewise.
7784 (mips*-*-rtems*): Likewise.
7785 (powerpc-*-rtems*): Likewise.
7786 (sh-*-rtems*): Likewise.
7787 (sparc-*-rtems*): Likewise.
7788 (sparc64-*-rtems*): Likewise.
7789 (v850-*-rtems*): Likewise.
7790 (m32c-*-rtems*): Likewise.
7792 2014-01-15 Vladimir Makarov <vmakarov@redhat.com>
7794 PR rtl-optimization/59511
7795 * ira.c (ira_init_register_move_cost): Use memory costs for some
7796 cases of register move cost calculations.
7797 * lra-constraints.c (lra_constraints): Use REG_FREQ_FROM_BB
7798 instead of BB frequency.
7799 * lra-coalesce.c (move_freq_compare_func, lra_coalesce): Ditto.
7800 * lra-assigns.c (find_hard_regno_for): Ditto.
7802 2014-01-15 Richard Biener <rguenther@suse.de>
7804 PR tree-optimization/59822
7805 * tree-vect-stmts.c (hoist_defs_of_uses): New function.
7806 (vectorizable_load): Use it to hoist defs of uses of invariant
7807 loads out of the loop.
7809 2014-01-15 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
7810 Kugan Vivekanandarajah <kuganv@linaro.org>
7813 * config/aarch64/aarch64.c (aarch64_build_constant): Fix incorrect
7816 2014-01-15 Richard Biener <rguenther@suse.de>
7818 PR rtl-optimization/59802
7819 * lcm.c (compute_available): Use inverted postorder to seed
7820 the initial worklist.
7822 2014-01-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7825 * config/s390/s390.c (s390_preferred_reload_class): Don't return
7826 ADDR_REGS for invalid symrefs in non-PIC code.
7828 2014-01-15 Jakub Jelinek <jakub@redhat.com>
7831 * builtins.c (determine_block_size): Initialize *probable_max_size
7832 even if len_rtx is CONST_INT.
7834 2014-01-14 Andrew Pinski <apinski@cavium.com>
7836 * config/aarch64/aarch64-protos.h (tune_params): Add issue_rate.
7837 * config/aarch64/aarch64.c (generic_tunings): Add issue rate of 2.
7838 (cortexa53_tunings): Likewise.
7839 (aarch64_sched_issue_rate): New function.
7840 (TARGET_SCHED_ISSUE_RATE): Define.
7842 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
7844 * ira-costs.c (find_costs_and_classes): Add missed
7845 ira_init_register_move_cost_if_necessary.
7847 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
7850 * config/arm/arm.c (arm_coproc_mem_operand): Add lra_in_progress.
7852 2014-01-14 H.J. Lu <hongjiu.lu@intel.com>
7855 * config/i386/i386.c (type_natural_mode): Add a bool parameter
7856 to indicate if type is used for function return value. Warn ABI
7857 change if the vector mode isn't available for function return value.
7858 (ix86_function_arg_advance): Pass false to type_natural_mode.
7859 (ix86_function_arg): Likewise.
7860 (ix86_gimplify_va_arg): Likewise.
7861 (function_arg_32): Don't warn ABI change.
7862 (ix86_function_value): Pass true to type_natural_mode.
7863 (ix86_return_in_memory): Likewise.
7864 (ix86_struct_value_rtx): Removed.
7865 (TARGET_STRUCT_VALUE_RTX): Likewise.
7867 2014-01-14 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7869 * jump.c (redirect_jump_2): Remove REG_CROSSING_JUMP notes when
7870 converting a conditional jump into a conditional return.
7872 2014-01-14 Richard Biener <rguenther@suse.de>
7874 PR tree-optimization/58921
7875 PR tree-optimization/59006
7876 * tree-vect-loop-manip.c (vect_loop_versioning): Remove code
7877 hoisting invariant stmts.
7878 * tree-vect-stmts.c (vectorizable_load): Insert the splat of
7879 invariant loads on the preheader edge if possible.
7881 2014-01-14 Joey Ye <joey.ye@arm.com>
7883 * doc/plugin.texi (Building GCC plugins): Update to C++.
7885 2014-01-14 Kirill Yukhin <kirill.yukhin@intel.com>
7887 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): New.
7888 (_mm_rcp28_round_ss): Ditto.
7889 (_mm_rsqrt28_round_sd): Ditto.
7890 (_mm_rsqrt28_round_ss): Ditto.
7891 (_mm_rcp28_sd): Ditto.
7892 (_mm_rcp28_ss): Ditto.
7893 (_mm_rsqrt28_sd): Ditto.
7894 (_mm_rsqrt28_ss): Ditto.
7895 * config/i386/avx512fintrin.h (_mm512_stream_load_si512): Ditto.
7896 * config/i386/i386-builtin-types.def (V8DI_FTYPE_PV8DI): Ditto.
7897 * config/i386/i386.c (IX86_BUILTIN_MOVNTDQA512): Ditto.
7898 (IX86_BUILTIN_RCP28SD): Ditto.
7899 (IX86_BUILTIN_RCP28SS): Ditto.
7900 (IX86_BUILTIN_RSQRT28SD): Ditto.
7901 (IX86_BUILTIN_RSQRT28SS): Ditto.
7902 (bdesc_special_args): Define __builtin_ia32_movntdqa512,
7903 __builtin_ia32_rcp28sd_round, __builtin_ia32_rcp28ss_round,
7904 __builtin_ia32_rsqrt28sd_round, __builtin_ia32_rsqrt28ss_round.
7905 (ix86_expand_special_args_builtin): Expand new FTYPE.
7906 * config/i386/sse.md (define_mode_attr "sse4_1_avx2"): Expand to V8DI.
7907 (srcp14<mode>): Make insn unary.
7908 (avx512f_vmscalef<mode><round_name>): Use substed predicate.
7909 (avx512f_sgetexp<mode><round_saeonly_name>): Ditto.
7910 (avx512f_rndscale<mode><round_saeonly_name>): Ditto.
7911 (<sse4_1_avx2>_movntdqa): Extend to 512 bits.
7912 (avx512er_exp2<mode><mask_name><round_saeonly_name>):
7913 Fix rounding: make it SAE only.
7914 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
7916 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
7918 (avx512er_vmrcp28<mode><round_saeonly_name>): Ditto.
7919 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Ditto.
7920 (avx512f_getmant<mode><mask_name><round_saeonly_name>): Ditto.
7921 * config/i386/subst.md (round_saeonly_mask_scalar_operand3): Remove.
7922 (round_saeonly_mask_scalar_operand4): Ditto.
7923 (round_saeonly_mask_scalar_op3): Ditto.
7924 (round_saeonly_mask_scalar_op4): Ditto.
7926 2014-01-13 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7928 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
7929 Implement -maltivec=be for vec_insert and vec_extract.
7931 2014-01-10 DJ Delorie <dj@redhat.com>
7933 * config/msp430/msp430.md (call_internal): Don't allow memory
7934 references with SP as the base register.
7935 (call_value_internal): Likewise.
7936 * config/msp430/constraints.md (Yc): New. For memory references
7937 that don't use SP as a base register.
7939 * config/msp430/msp430.c (msp430_print_operand): Add 'J' to mean
7940 "an integer without a # prefix"
7941 * config/msp430/msp430.md (epilogue_helper): Use it.
7943 2014-01-13 Jakub Jelinek <jakub@redhat.com>
7946 * config/i386/i386.c (ix86_vectorize_builtin_gather): Uncomment
7947 AVX512F gather builtins.
7948 * tree-vect-stmts.c (vectorizable_mask_load_store): For now punt
7949 on gather decls with INTEGER_TYPE masktype.
7950 (vectorizable_load): For INTEGER_TYPE masktype, put the INTEGER_CST
7951 directly into the builtin rather than hoisting it before loop.
7953 PR tree-optimization/59387
7954 * tree-scalar-evolution.c: Include gimple-fold.h and gimplify-me.h.
7955 (scev_const_prop): If folded_casts and type has undefined overflow,
7956 use force_gimple_operand instead of force_gimple_operand_gsi and
7957 for each added stmt if it is assign with
7958 arith_code_with_undefined_signed_overflow, call
7959 rewrite_to_defined_overflow.
7960 * tree-ssa-loop-im.c: Don't include gimplify-me.h, include
7961 gimple-fold.h instead.
7962 (arith_code_with_undefined_signed_overflow,
7963 rewrite_to_defined_overflow): Moved to ...
7964 * gimple-fold.c (arith_code_with_undefined_signed_overflow,
7965 rewrite_to_defined_overflow): ... here. No longer static.
7966 Include gimplify-me.h.
7967 * gimple-fold.h (arith_code_with_undefined_signed_overflow,
7968 rewrite_to_defined_overflow): New prototypes.
7970 2014-01-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7972 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Fix typo in description.
7974 2014-01-13 Eric Botcazou <ebotcazou@adacore.com>
7976 * builtins.c (get_object_alignment_2): Minor tweak.
7977 * tree-ssa-loop-ivopts.c (may_be_unaligned_p): Rewrite.
7979 2014-01-13 Christian Bruel <christian.bruel@st.com>
7981 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Unroll small sizes and
7982 optimized non constant lengths.
7984 2014-01-13 Jakub Jelinek <jakub@redhat.com>
7987 * omp-low.c (expand_omp_atomic_pipeline): Expand the initial
7988 load as __atomic_load_N if possible.
7990 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
7992 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): Remove
7994 (rs6000_expand_builtin): Adjust call.
7996 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
7999 * config/rs6000/rs6000.h (SWITCHABLE_TARGET): Define.
8000 * config/rs6000/rs6000.c: Include target-globals.h.
8001 (rs6000_set_current_function): Instead of doing target_reinit
8002 unconditionally, use save_target_globals_default_opts and
8003 restore_target_globals.
8005 * config/rs6000/rs6000-builtin.def (mffs, mtfsf): Add builtins for
8007 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): New.
8008 (rs6000_expand_builtin): Handle mffs and mtfsf.
8009 (rs6000_init_builtins): Define mffs and mtfsf.
8010 * config/rs6000/rs6000.md (UNSPECV_MFFS, UNSPECV_MTFSF): New constants.
8011 (rs6000_mffs): New pattern.
8012 (rs6000_mtfsf): New pattern.
8014 2014-01-11 Bin Cheng <bin.cheng@arm.com>
8016 * tree-ssa-loop-ivopts.c (iv_ca_narrow): New parameter.
8017 Start narrowing with START. Apply candidate-use pair
8018 and check overall cost in narrowing.
8019 (iv_ca_prune): Pass new argument.
8021 2014-01-10 Jeff Law <law@redhat.com>
8024 * ree.c (combine_reaching_defs): Ensure the defining statement
8025 occurs before the extension when optimizing extensions with
8026 different source and destination hard registers.
8028 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
8031 * ipa-devirt.c (build_type_inheritance_graph): Also add types of
8032 vtables into the type inheritance graph.
8034 2014-01-10 Jakub Jelinek <jakub@redhat.com>
8036 PR rtl-optimization/59754
8037 * ree.c (combine_reaching_defs): Disallow !SCALAR_INT_MODE_P
8038 modes in the REGNO != REGNO case.
8040 2014-01-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8042 * config/rs6000/rs6000-builtin.def: Fix pasto for VPKSDUS.
8044 2014-01-10 Jakub Jelinek <jakub@redhat.com>
8046 PR tree-optimization/59745
8047 * tree-predcom.c (tree_predictive_commoning_loop): Call
8048 free_affine_expand_cache if giving up because components is NULL.
8050 * target-globals.c (save_target_globals): Allocate < 4KB structs using
8051 GC in payload of target_globals struct instead of allocating them on
8052 the heap and the larger structs separately using GC.
8053 * target-globals.h (struct target_globals): Make regs, hard_regs,
8054 reload, expmed, ira, ira_int and lra_fields GTY((atomic)) instead
8055 of GTY((skip)) and change type to void *.
8056 (reset_target_globals): Cast loads from those fields to corresponding
8059 2014-01-10 Steve Ellcey <sellcey@mips.com>
8062 * Makefile.in (PLUGIN_HEADERS): Add gimplify.h, gimple-iterator.h,
8063 gimple-ssa.h, fold-const.h, tree-cfg.h, tree-into-ssa.h,
8064 tree-ssanames.h, print-tree.h, varasm.h, and context.h.
8066 2014-01-10 Richard Earnshaw <rearnsha@arm.com>
8069 * aarch64-modes.def (CC_Zmode): New flags mode.
8070 * aarch64.c (aarch64_select_cc_mode): Only allow NEG when the condition
8071 represents an equality.
8072 (aarch64_get_condition_code): Handle CC_Zmode.
8073 * aarch64.md (compare_neg<mode>): Restrict to equality operations.
8075 2014-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8077 * config/s390/s390.c (s390_expand_tbegin): Remove jump over CC
8078 extraction in good case.
8080 2014-01-10 Richard Biener <rguenther@suse.de>
8082 PR tree-optimization/59374
8083 * tree-vect-slp.c (vect_slp_analyze_bb_1): Move dependence
8084 checking after SLP discovery. Mark stmts not participating
8085 in any SLP instance properly.
8087 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8089 * config/arm/arm.c (arm_new_rtx_costs): Use destination mode
8090 when handling a SET rtx.
8092 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8094 * config/arm/arm-cores.def (cortex-a53): Specify FL_CRC32.
8095 (cortex-a57): Likewise.
8096 (cortex-a57.cortex-a53): Likewise. Remove redundant flags.
8098 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8100 * config/arm/arm.c (arm_init_iwmmxt_builtins): Skip
8101 non-iwmmxt builtins.
8103 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
8107 * ipa-devirt.c record_target_from_binfo): Take as argument
8108 stack of binfos and lookup matching one for virtual inheritance.
8109 (possible_polymorphic_call_targets_1): Update.
8111 2014-01-10 Huacai Chen <chenhc@lemote.com>
8113 * config/mips/driver-native.c (host_detect_local_cpu): Handle new
8114 kernel strings for Loongson-2E/2F/3A.
8116 2014-01-10 Jakub Jelinek <jakub@redhat.com>
8119 * tree-vect-data-refs.c (vect_analyze_data_refs): Check
8120 is_gimple_call before calling gimple_call_internal_p.
8122 2014-01-09 Steve Ellcey <sellcey@mips.com>
8124 * Makefile.in (TREE_FLOW_H): Remove.
8125 (TREE_SSA_H): Add file names from tree-flow.h.
8126 * doc/tree-ssa.texi (Annotations): Remove reference to tree-flow.h
8127 * tree.h: Remove tree-flow.h reference.
8128 * hash-table.h: Remove tree-flow.h reference.
8129 * tree-ssa-loop-niter.c (dump_affine_iv): Replace tree-flow.h
8130 reference with tree-ssa-loop.h.
8132 2014-01-09 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8134 * doc/invoke.texi: Add -maltivec={be,le} options, and document
8135 default element-order behavior for -maltivec.
8136 * config/rs6000/rs6000.opt: Add -maltivec={be,le} options.
8137 * config/rs6000/rs6000.c (rs6000_option_override_internal): Ensure
8138 that -maltivec={le,be} implies -maltivec; disallow -maltivec=le
8139 when targeting big endian, at least for now.
8140 * config/rs6000/rs6000.h: Add #define of VECTOR_ELT_ORDER_BIG.
8142 2014-01-09 Jakub Jelinek <jakub@redhat.com>
8145 * cfgexpand.c (expand_one_var): For SSA_NAMEs, if the underlying
8146 var satisfies use_register_for_decl, just take into account type
8147 alignment, rather than decl alignment.
8149 PR tree-optimization/59622
8150 * gimple-fold.c (gimple_fold_call): Fix a typo in message. For
8151 __builtin_unreachable replace the OBJ_TYPE_REF call with a call to
8152 __builtin_unreachable and add if needed a setter of the lhs SSA_NAME.
8153 Don't devirtualize for inplace at all. For targets.length () == 1,
8154 if the call is noreturn and cfun isn't in SSA form yet, clear lhs.
8156 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
8158 * config/i386/i386.md (cpu): Remove the unused btver1.
8160 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
8162 * gdbasan.in: Put a breakpoint on __sanitizer::Report.
8164 2014-01-09 Jakub Jelinek <jakub@redhat.com>
8167 * tree-core.h (struct target_globals): New forward declaration.
8168 (struct tree_target_option): Add globals field.
8169 * tree.h (TREE_TARGET_GLOBALS): Define.
8170 (prepare_target_option_nodes_for_pch): New prototype.
8171 * target-globals.h (struct target_globals): Define even if
8173 * tree.c (prepare_target_option_node_for_pch,
8174 prepare_target_option_nodes_for_pch): New functions.
8175 * config/i386/i386.h (SWITCHABLE_TARGET): Define.
8176 * config/i386/i386.c: Include target-globals.h.
8177 (ix86_set_current_function): Instead of doing target_reinit
8178 unconditionally, use save_target_globals_default_opts and
8179 restore_target_globals.
8181 2014-01-09 Richard Biener <rguenther@suse.de>
8183 PR tree-optimization/59715
8184 * tree-cfg.h (split_critical_edges): Declare.
8185 * tree-cfg.c (split_critical_edges): Export.
8186 * tree-ssa-sink.c (execute_sink_code): Split critical edges.
8188 2014-01-09 Max Ostapenko <m.ostapenko@partner.samsung.com>
8190 * cfgexpand.c (expand_stack_vars): Optionally disable
8191 asan stack protection.
8192 (expand_used_vars): Likewise.
8193 (partition_stack_vars): Likewise.
8194 * asan.c (asan_emit_stack_protection): Optionally disable
8195 after return stack usage.
8196 (instrument_derefs): Optionally disable memory access instrumentation.
8197 (instrument_builtin_call): Likewise.
8198 (instrument_strlen_call): Likewise.
8199 (asan_protect_global): Optionally disable global variables protection.
8200 * doc/invoke.texi: Added doc for new options.
8201 * params.def: Added new options.
8202 * params.h: Likewise.
8204 2014-01-09 Jakub Jelinek <jakub@redhat.com>
8206 PR rtl-optimization/59724
8207 * ifcvt.c (cond_exec_process_if_block): Don't call
8208 flow_find_head_matching_sequence with 0 longest_match.
8209 * cfgcleanup.c (flow_find_head_matching_sequence): Count even
8210 non-active insns if !stop_after.
8211 (try_head_merge_bb): Revert 2014-01-07 changes.
8213 2014-01-08 Jeff Law <law@redhat.com>
8215 * ree.c (get_sub_rtx): New function, extracted from...
8216 (merge_def_and_ext): Here.
8217 (combine_reaching_defs): Use get_sub_rtx.
8219 2014-01-08 Eric Botcazou <ebotcazou@adacore.com>
8221 * cgraph.h (varpool_variable_node): Do not choke on null node.
8223 2014-01-08 Catherine Moore <clm@codesourcery.com>
8225 * config/mips/mips.md (simple_return): Attempt to use JRC
8227 * config/mips/mips.h (MIPS_CALL): Attempt to use JALS for microMIPS.
8229 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
8231 PR rtl-optimization/59137
8232 * reorg.c (steal_delay_list_from_target): Call update_block for
8234 (steal_delay_list_from_fallthrough, relax_delay_slots): Likewise.
8236 2014-01-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8238 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Remove
8239 two duplicate entries.
8241 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
8244 2012-10-07 Richard Sandiford <rdsandiford@googlemail.com>
8246 * config/mips/mips.c (mips_truncated_op_cost): New function.
8247 (mips_rtx_costs): Adjust test for BADDU.
8248 * config/mips/mips.md (*baddu_di<mode>): Push truncates to operands.
8250 2012-10-02 Richard Sandiford <rdsandiford@googlemail.com>
8252 * config/mips/mips.md (*baddu_si_eb, *baddu_si_el): Merge into...
8253 (*baddu_si): ...this new pattern.
8255 2014-01-08 Jakub Jelinek <jakub@redhat.com>
8258 * ipa-prop.c (ipa_analyze_params_uses): Ignore uses in debug stmts.
8260 2014-01-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
8263 * expr.h (expand_expr_real, expand_expr_real_1): Add new parameter
8265 (expand_expr, expand_normal): Adjust.
8266 * expr.c (expand_expr_real, expand_expr_real_1): Add new parameter
8267 inner_reference_p. Use inner_reference_p to expand inner references.
8268 (store_expr): Adjust.
8269 * cfgexpand.c (expand_call_stmt): Adjust.
8271 2014-01-08 Rong Xu <xur@google.com>
8273 * gcov-io.c (gcov_var): Move from gcov-io.h.
8274 (gcov_position): Ditto.
8275 (gcov_is_error): Ditto.
8276 (gcov_rewrite): Ditto.
8277 * gcov-io.h: Refactor. Move gcov_var to gcov-io.h, and libgcov
8278 only part to libgcc/libgcov.h.
8280 2014-01-08 Marek Polacek <polacek@redhat.com>
8283 * omp-low.c (simd_clone_adjust): Don't crash if def is NULL.
8285 2014-01-08 Marek Polacek <polacek@redhat.com>
8288 * ubsan.c (ubsan_type_descriptor): Call strip_array_types on type2.
8290 2014-01-08 Jakub Jelinek <jakub@redhat.com>
8292 PR rtl-optimization/59649
8293 * stor-layout.c (get_mode_bounds): For BImode return
8294 0 and STORE_FLAG_VALUE.
8296 2014-01-08 Richard Biener <rguenther@suse.de>
8299 * gimple.h (is_gimple_builtin_call): Remove.
8300 (gimple_builtin_call_types_compatible_p): New.
8301 (gimple_call_builtin_p): New overload.
8302 * gimple.c (is_gimple_builtin_call): Remove.
8303 (validate_call): Rename to ...
8304 (gimple_builtin_call_types_compatible_p): ... this and export. Also
8306 (validate_type): New static function.
8307 (gimple_call_builtin_p): New overload and adjust.
8308 * gimple-fold.c (gimple_fold_builtin): Fold the return value.
8309 (gimple_fold_call): Likewise. Use gimple_call_builtin_p.
8310 (gimple_fold_stmt_to_constant_1): Likewise.
8311 * tsan.c (instrument_gimple): Use gimple_call_builtin_p.
8313 2014-01-08 Richard Biener <rguenther@suse.de>
8316 * gimplify.c (gimplify_expr): Gimplify register-register type
8317 VIEW_CONVERT_EXPRs to separate stmts.
8319 2014-01-07 Jeff Law <law@redhat.com>
8322 * ree.c (combine_set_extension): Handle case where source
8323 and destination registers in an extension insn are different.
8324 (combine_reaching_defs): Allow source and destination registers
8325 in extension to be different under limited circumstances.
8326 (add_removable_extension): Remove restriction that the
8327 source and destination registers in the extension are the same.
8328 (find_and_remove_re): Emit a copy from the extension's
8329 destination to its source after the defining insn if
8330 the source and destination registers are different.
8333 * ifcvt.c (merge_if_block): If we are merging a block with more than
8334 one successor with a block with no successors, remove any BARRIER
8335 after the second block.
8337 2014-01-07 Dan Xio Qiang <ziyan01@163.com>
8339 * hw-doloop.c (reorg_loops): Release the bitmap obstack.
8341 2014-01-07 John David Anglin <danglin@gcc.gnu.org>
8344 * config/pa/pa.c (pa_legitimate_address_p): Return false before reload
8345 for 14-bit register offsets when INT14_OK_STRICT is false.
8347 2014-01-07 Roland Stigge <stigge@antcom.de>
8348 Michael Meissner <meissner@linux.vnet.ibm.com>
8351 * config/rs6000/rs6000.c (rs6000_legitimate_offset_address_p):
8352 Only check TFmode for SPE constants. Don't check TImode or TDmode.
8354 2014-01-07 James Greenhalgh <james.greenhalgh@arm.com>
8356 * config/aarch64/aarch64-elf.h (ASM_SPEC): Remove identity spec for
8359 2014-01-07 Yufeng Zhang <yufeng.zhang@arm.com>
8361 * config/arm/arm.c (arm_expand_neon_args): Call expand_expr
8362 with EXPAND_MEMORY for NEON_ARG_MEMORY; check if the returned
8363 rtx is const0_rtx or not.
8365 2014-01-07 Richard Sandiford <rdsandiford@googlemail.com>
8368 * target-globals.c (save_target_globals): Remove this_fn_optab
8370 * toplev.c: Include optabs.h.
8371 (target_reinit): Temporarily restore the global options if another
8372 set of options are in force.
8374 2014-01-07 Jakub Jelinek <jakub@redhat.com>
8376 PR rtl-optimization/58668
8377 * cfgcleanup.c (flow_find_cross_jump): Don't count
8378 any jumps if dir_p is NULL. Remove p1 variable, use active_insn_p
8379 to determine what is counted.
8380 (flow_find_head_matching_sequence): Use active_insn_p to determine
8382 (try_head_merge_bb): Adjust for the flow_find_head_matching_sequence
8384 * ifcvt.c (count_bb_insns): Use active_insn_p && !JUMP_P to
8385 determine what is counted.
8387 PR tree-optimization/59643
8388 * tree-predcom.c (split_data_refs_to_components): If one dr is
8389 read and one write, determine_offset fails and the write isn't
8390 in the bad component, just put the read into the bad component.
8392 2014-01-07 Mike Stump <mikestump@comcast.net>
8393 Jakub Jelinek <jakub@redhat.com>
8396 * tree-core.h (struct tree_optimization_option): Change optabs
8397 type from unsigned char * to void *.
8398 * optabs.c (init_tree_optimization_optabs): Adjust
8399 TREE_OPTIMIZATION_OPTABS initialization.
8401 2014-01-06 Jakub Jelinek <jakub@redhat.com>
8404 * config/i386/i386.h (struct machine_function): Add
8405 no_drap_save_restore field.
8406 * config/i386/i386.c (ix86_save_reg): Use
8407 !cfun->machine->no_drap_save_restore instead of
8408 crtl->stack_realign_needed.
8409 (ix86_finalize_stack_realign_flags): Don't clear drap_reg unless
8410 this function clears frame_pointer_needed. Set
8411 cfun->machine->no_drap_save_restore if clearing frame_pointer_needed
8412 and DRAP reg is needed.
8414 2014-01-06 Marek Polacek <polacek@redhat.com>
8417 * doc/implement-c.texi: Mention that other integer types are
8418 permitted as bit-field types in strictly conforming mode.
8420 2014-01-06 Felix Yang <fei.yang0953@gmail.com>
8422 * modulo-sched.c (schedule_reg_moves): Clear distance1_uses if it
8425 2014-01-06 Richard Earnshaw <rearnsha@arm.com>
8427 * aarch64.c (aarch64_rtx_costs): Fix cost calculation for MADD.
8429 2014-01-06 Martin Jambor <mjambor@suse.cz>
8432 * ipa-cp.c (ipcp_discover_new_direct_edges): Changed param_index type
8434 * ipa-prop.c (ipa_print_node_params): Fix indentation.
8436 2014-01-06 Eric Botcazou <ebotcazou@adacore.com>
8440 * var-tracking.c (add_stores): Preserve the value of the source even if
8441 we don't record the store.
8443 2014-01-06 Terry Guo <terry.guo@arm.com>
8445 * config.gcc (arm*-*-*): Check --with-arch against arm-arches.def.
8447 2014-01-05 Iain Sandoe <iain@codesourcery.com>
8450 * config/darwin.c (darwin_function_section): Adjust return values to
8451 correspond to optimisation changes made in r206070.
8453 2014-01-05 Uros Bizjak <ubizjak@gmail.com>
8455 * config/i386/i386.c (ix86_data_alignment): Calculate max_align
8456 from prefetch_block tune setting.
8457 (nocona_cost): Correct size of prefetch block to 64.
8459 2014-01-04 Eric Botcazou <ebotcazou@adacore.com>
8461 * config/arm/arm.c (arm_get_frame_offsets): Revamp long lines.
8462 (arm_expand_epilogue_apcs_frame): Take into account the number of bytes
8463 used to save the static chain register in the computation of the offset
8464 from which the FP registers need to be restored.
8466 2014-01-04 Jakub Jelinek <jakub@redhat.com>
8468 PR tree-optimization/59519
8469 * tree-vect-loop-manip.c (slpeel_update_phi_nodes_for_guard1): Don't
8470 ICE if get_current_def (current_new_name) is already non-NULL, as long
8471 as it is a phi result of some other phi in *new_exit_bb that has
8474 * config/i386/sse.md (avx512f_load<mode>_mask): Emit vmovup{s,d}
8475 or vmovdqu* for misaligned_operand.
8476 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>,
8477 <sse2_avx_avx512f>_loaddqu<mode><mask_name>): Handle <mask_applied>.
8478 * config/i386/i386.c (ix86_expand_special_args_builtin): Set
8479 aligned_mem for AVX512F masked aligned load and store builtins and for
8482 2014-01-03 Bingfeng Mei <bmei@broadcom.com>
8484 PR tree-optimization/59651
8485 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
8486 Address range for negative step should be added by TYPE_SIZE_UNIT.
8488 2014-01-03 Andreas Schwab <schwab@linux-m68k.org>
8490 * config/m68k/m68k.c (handle_move_double): Handle pushes with
8491 overlapping registers also for registers other than the stack pointer.
8493 2014-01-03 Marek Polacek <polacek@redhat.com>
8496 * doc/extend.texi: Fix the return value of __builtin_FUNCTION and
8499 2014-01-03 Jakub Jelinek <jakub@redhat.com>
8502 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Don't consider
8505 * config/i386/i386.md (MODE_SIZE): New mode attribute.
8506 (push splitter): Use <P:MODE_SIZE> instead of
8507 GET_MODE_SIZE (<P:MODE>mode).
8508 (lea splitter): Use <MODE_SIZE> instead of GET_MODE_SIZE (<MODE>mode).
8509 (mov -1, reg peephole2): Likewise.
8510 * config/i386/sse.md (*mov<mode>_internal,
8511 <sse>_storeu<ssemodesuffix><avxsizesuffix>,
8512 <sse2_avx_avx512f>_storedqu<mode>, <sse>_andnot<mode>3,
8513 *<code><mode>3, *andnot<mode>3<mask_name>,
8514 <mask_codefor><code><mode>3<mask_name>): Likewise.
8515 * config/i386/subst.md (mask_mode512bit_condition,
8516 sd_mask_mode512bit_condition): Likewise.
8518 2014-01-02 Xinliang David Li <davidxl@google.com>
8520 PR tree-optimization/59303
8521 * tree-ssa-uninit.c (is_use_properly_guarded): Main cleanup.
8522 (dump_predicates): Better output format.
8523 (pred_equal_p): New function.
8524 (is_neq_relop_p): Ditto.
8525 (is_neq_zero_form_p): Ditto.
8526 (pred_expr_equal_p): Ditto.
8527 (pred_neg_p): Ditto.
8528 (simplify_pred): Ditto.
8529 (simplify_preds_2): Ditto.
8530 (simplify_preds_3): Ditto.
8531 (simplify_preds_4): Ditto.
8532 (simplify_preds): Ditto.
8534 (push_to_worklist): Ditto.
8535 (get_pred_info_from_cmp): Ditto.
8536 (is_degenerated_phi): Ditto.
8537 (normalize_one_pred_1): Ditto.
8538 (normalize_one_pred): Ditto.
8539 (normalize_one_pred_chain): Ditto.
8540 (normalize_preds): Ditto.
8541 (normalize_cond_1): Remove function.
8542 (normalize_cond): Ditto.
8543 (is_gcond_subset_of): Ditto.
8544 (is_subset_of_any): Ditto.
8545 (is_or_set_subset_of): Ditto.
8546 (is_and_set_subset_of): Ditto.
8547 (is_norm_cond_subset_of): Ditto.
8548 (pred_chain_length_cmp): Ditto.
8549 (convert_control_dep_chain_into_preds): Type change.
8550 (find_predicates): Ditto.
8551 (find_def_preds): Ditto.
8552 (destroy_predicates_vecs): Ditto.
8553 (find_matching_predicates_in_rest_chains): Ditto.
8554 (use_pred_not_overlap_with_undef_path_pred): Ditto.
8555 (is_pred_expr_subset): Ditto.
8556 (is_pred_chain_subset_of): Ditto.
8557 (is_included_in): Ditto.
8558 (is_superset_of): Ditto.
8560 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
8562 Update copyright years.
8564 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
8566 * common/config/arc/arc-common.c, config/arc/arc-modes.def,
8567 config/arc/arc-protos.h, config/arc/arc.c, config/arc/arc.h,
8568 config/arc/arc.md, config/arc/arc.opt,
8569 config/arm/arm_neon_builtins.def, config/arm/crypto.def,
8570 config/i386/avx512cdintrin.h, config/i386/avx512erintrin.h,
8571 config/i386/avx512fintrin.h, config/i386/avx512pfintrin.h,
8572 config/i386/btver2.md, config/i386/shaintrin.h, config/i386/slm.md,
8573 config/linux-protos.h, config/linux.c, config/winnt-c.c,
8574 diagnostic-color.c, diagnostic-color.h, gimple-ssa-isolate-paths.c,
8575 vtable-verify.c, vtable-verify.h: Use the standard form for the
8578 2014-01-02 Tobias Burnus <burnus@net-b.de>
8580 * gcc.c (process_command): Update copyright notice dates.
8581 * gcov-dump.c: Ditto.
8583 * doc/cpp.texi: Bump @copying's copyright year.
8584 * doc/cppinternals.texi: Ditto.
8585 * doc/gcc.texi: Ditto.
8586 * doc/gccint.texi: Ditto.
8587 * doc/gcov.texi: Ditto.
8588 * doc/install.texi: Ditto.
8589 * doc/invoke.texi: Ditto.
8591 2014-01-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8593 * config/nios2/nios2.h (BITS_PER_UNIT): Don't define it.
8595 2014-01-01 Jakub Jelinek <jakub@redhat.com>
8597 * config/i386/sse.md (*mov<mode>_internal): Guard
8598 EXT_REX_SSE_REGNO_P (REGNO ()) uses with REG_P.
8600 PR rtl-optimization/59647
8601 * cse.c (cse_process_notes_1): Don't substitute negative VOIDmode
8602 new_rtx into UNSIGNED_FLOAT rtxes.
8604 Copyright (C) 2014 Free Software Foundation, Inc.
8606 Copying and distribution of this file, with or without modification,
8607 are permitted in any medium without royalty provided the copyright
8608 notice and this notice are preserved.