1 2014-04-29 Alan Lawrence <alan.lawrence@arm.com>
3 * config/aarch64/arm_neon.h (vzip1_f32, vzip1_p8, vzip1_p16, vzip1_s8,
4 vzip1_s16, vzip1_s32, vzip1_u8, vzip1_u16, vzip1_u32, vzip1q_f32,
5 vzip1q_f64, vzip1q_p8, vzip1q_p16, vzip1q_s8, vzip1q_s16, vzip1q_s32,
6 vzip1q_s64, vzip1q_u8, vzip1q_u16, vzip1q_u32, vzip1q_u64, vzip2_f32,
7 vzip2_p8, vzip2_p16, vzip2_s8, vzip2_s16, vzip2_s32, vzip2_u8,
8 vzip2_u16, vzip2_u32, vzip2q_f32, vzip2q_f64, vzip2q_p8, vzip2q_p16,
9 vzip2q_s8, vzip2q_s16, vzip2q_s32, vzip2q_s64, vzip2q_u8, vzip2q_u16,
10 vzip2q_u32, vzip2q_u64): Replace inline __asm__ with __builtin_shuffle.
12 2014-04-29 David Malcolm <dmalcolm@redhat.com>
14 * tree-cfg.c (dump_function_to_file): Dump the return type of
15 functions, in a line to itself before the function body, mimicking
16 the layout of a C function.
18 2014-04-29 Jakub Jelinek <jakub@redhat.com>
20 PR tree-optimization/60971
21 * tree-tailcall.c (process_assignment): Reject conversions which
24 2014-04-29 James Greenhalgh <james.greenhalgh@arm.com>
26 * calls.c (initialize_argument_information): Always treat
27 PUSH_ARGS_REVERSED as 1, simplify code accordingly.
28 (expand_call): Likewise.
29 (emit_library_call_calue_1): Likewise.
30 * expr.c (PUSH_ARGS_REVERSED): Do not define.
31 (emit_push_insn): Always treat PUSH_ARGS_REVERSED as 1, simplify
34 2014-04-29 Nick Clifton <nickc@redhat.com>
36 * config/msp430/msp430.md (umulsidi): Fix typo.
37 (mulhisi3): Enable even inside interrupt handlers.
38 * config/msp430/msp430.c (msp430_print_operand): %O: Allow for the
39 bigger return address pushed in large mode.
41 2014-04-29 Nick Clifton <nickc@redhat.com>
43 * config/arc/arc.c (arc_select_cc_mode): Fix parentheses.
44 (arc_init_reg_tables): Use a machine_mode enum to iterate over
46 * config/m32r/m32r.c (init_reg_tables): Likewise.
47 * config/m32c/m32c.c (m32c_illegal_subreg_p): Use a machine_mode
48 enum to hold the modes.
50 2014-04-29 Richard Biener <rguenther@suse.de>
52 * dominance.c (free_dominance_info): Add overload with
54 (dom_info_state): Likewise.
55 (dom_info_available_p): Likewise.
56 * basic-block.h (free_dominance_info, dom_info_state,
57 dom_info_available_p): Declare overloads.
58 * passes.c (execute_function_todo): Verify that verifiers
59 don't change dominator info state. Drop dominator info
60 for IPA pass invocations.
61 * cgraph.c (release_function_body): Restore asserts that
62 dominator information is released.
64 2014-04-29 Patrick Palka <patrick@parcs.ath.cx>
66 * doc/invoke.texi: Fix typo.
67 * tree-vrp.c: Fix typos.
68 * gimple.c (infer_nonnull_range): Reorder operands of an &&
71 2014-04-29 Zhenqiang Chen <zhenqiang.chen@linaro.org>
73 * config/aarch64/aarch64.md (mov<mode>cc): New for GPF.
75 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
77 * config/aarch64/aarch64-builtins.c
78 (aarch64_types_storestruct_lane_qualifiers): New.
79 (TYPES_STORESTRUCT_LANE): Likewise.
80 * config/aarch64/aarch64-simd-builtins.def (st2_lane): New.
83 * config/aarch64/aarch64-simd.md (vec_store_lanesoi_lane<mode>): New.
84 (vec_store_lanesci_lane<mode>): Likewise.
85 (vec_store_lanesxi_lane<mode>): Likewise.
86 (aarch64_st2_lane<VQ:mode>): Likewise.
87 (aarch64_st3_lane<VQ:mode>): Likewise.
88 (aarch64_st4_lane<VQ:mode>): Likewise.
89 * config/aarch64/aarch64.md (unspec): Add UNSPEC_ST{2,3,4}_LANE.
90 * config/aarch64/arm_neon.h
91 (__ST2_LANE_FUNC): Rewrite using builtins, update use points to
92 use new macro arguments.
93 (__ST3_LANE_FUNC): Likewise.
94 (__ST4_LANE_FUNC): Likewise.
95 * config/aarch64/iterators.md (V_TWO_ELEM): New.
96 (V_THREE_ELEM): Likewise.
97 (V_FOUR_ELEM): Likewise.
99 2014-04-28 David Malcolm <dmalcolm@redhat.com>
101 * doc/gimple.texi: Replace the description of the now-defunct
102 union gimple_statement_d with a diagram showing the
103 gimple_statement_base class hierarchy and its relationships to
104 the GSS_ and GIMPLE_ enums.
106 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
108 * config/aarch64/aarch64-protos.h (aarch64_modes_tieable_p): New.
109 * config/aarch64/aarch64.c
110 (aarch64_cannot_change_mode_class): Weaken conditions.
111 (aarch64_modes_tieable_p): New.
112 * config/aarch64/aarch64.h (MODES_TIEABLE_P): Use it.
114 2014-04-28 Pat Haugen <pthaugen@us.ibm.com>
116 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
117 (loadsync_<mode>): Change mode.
118 (load_quadpti, store_quadpti): New.
119 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
120 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
122 2014-04-28 Martin Jambor <mjambor@suse.cz>
124 * tree-sra.c (sra_modify_expr): Generate new memory accesses with
125 same alias type as the original statement.
126 (subreplacement_assignment_data): New type.
127 (handle_unscalarized_data_in_subtree): New type of parameter,
128 generate new memory accesses with same alias type as the original
130 (load_assign_lhs_subreplacements): Likewise.
131 (sra_modify_constructor_assign): Generate new memory accesses with
132 same alias type as the original statement.
134 2014-04-28 Richard Biener <rguenther@suse.de>
136 * tree-pass.h (TODO_verify_il): Define.
137 (TODO_verify_all): Complete properly.
138 * passes.c (execute_function_todo): Move existing loop-closed
139 SSA verification under TODO_verify_il.
140 (execute_one_pass): Trigger TODO_verify_il at todo-after time.
141 * graphite-sese-to-poly.c (rewrite_cross_bb_scalar_deps):
142 Fix tree sharing issue.
144 2014-04-28 Richard Biener <rguenther@suse.de>
147 * builtins.def (DEF_C11_BUILTIN): Add.
148 (BUILT_IN_ALIGNED_ALLOC): Likewise.
149 * coretypes.h (enum function_class): Add function_c11_misc.
150 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
151 BUILT_IN_ALIGNED_ALLOC like BUILT_IN_MALLOC.
152 (call_may_clobber_ref_p_1): Likewise.
153 * tree-ssa-dce.c (mark_stmt_if_obviously_necessary): Likewise.
154 (mark_all_reaching_defs_necessary_1): Likewise.
155 (propagate_necessity): Likewise.
156 (eliminate_unnecessary_stmts): Likewise.
157 * tree-ssa-ccp.c (evaluate_stmt): Handle BUILT_IN_ALIGNED_ALLOC.
159 2014-04-28 Richard Biener <rguenther@suse.de>
161 * tree-vrp.c (vrp_var_may_overflow): Remove.
162 (vrp_visit_phi_node): Rather than bumping to +-INF possibly
163 with overflow immediately bump to one before that value and
164 let iteration figure out overflow status.
166 2014-04-28 Richard Biener <rguenther@suse.de>
168 * configure.ac: Do valgrind header checks unconditionally.
169 Add --enable-valgrind-annotations.
170 * system.h: Guard valgrind header inclusion with
171 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING.
172 * alloc-pool.c (pool_alloc, pool_free): Use
173 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING
174 to guard possibly dead code.
175 * config.in: Regenerated.
176 * configure: Likewise.
178 2014-04-28 Jeff Law <law@redhat.com>
180 PR tree-optimization/60902
181 * tree-ssa-threadedge.c
182 (record_temporary_equivalences_from_stmts_at_dest): Only iterate
183 over real defs when invalidating outputs from statements that do not
184 produce useful outputs for threading.
186 2014-04-28 Richard Biener <rguenther@suse.de>
188 PR tree-optimization/60979
189 * graphite-scop-detection.c (scopdet_basic_block_info): Reject
190 SCOPs that end in a block with a successor with abnormal
193 2014-04-28 Richard Biener <rguenther@suse.de>
195 * tree-pass.h (execute_pass_list): Adjust prototype.
196 * passes.c (pass_manager::execute_early_local_passes):
198 (do_per_function): Change callback signature, push all actual
199 work to the callbals.
200 (do_per_function_toporder): Likewise.
201 (execute_function_dump): Adjust.
202 (execute_function_todo): Likewise.
203 (clear_last_verified): Likewise.
204 (verify_curr_properties): Likewise.
205 (update_properties_after_pass): Likewise.
206 (execute_pass_list_1): Split out from ...
207 (execute_pass_list): ... here. Adjust.
208 (execute_ipa_pass_list): Likewise.
209 * cgraphunit.c (cgraph_add_new_function): Adjust.
210 (analyze_function): Likewise.
211 (expand_function): Likewise.
212 * cgraph.c (release_function_body): Free dominance info
213 here instead of asserting it was magically freed elsewhere.
215 2014-04-28 Eric Botcazou <ebotcazou@adacore.com>
217 * configure.ac: Tweak GAS check for LEON instructions on SPARC.
218 * configure: Regenerate.
219 * config/sparc/sparc.opt (muser-mode): New option.
220 * config/sparc/sync.md (atomic_compare_and_swap<mode>_1): Do not enable
222 (atomic_compare_and_swap_leon3_1): New instruction for LEON3.
223 * doc/invoke.texi (SPARC options): Document -muser-mode.
225 2014-04-27 Richard Sandiford <rdsandiford@googlemail.com>
227 * cselib.c (find_slot_memmode): Delete.
228 (cselib_hasher): Change compare_type to a struct.
229 (cselib_hasher::equal): Update accordingly. Don't expect wrapped
231 (preserve_constants_and_equivs): Adjust for new compare_type.
232 (cselib_find_slot): Likewise. Take the mode of the rtx as argument.
233 (wrap_constant): Delete.
234 (cselib_lookup_mem, cselib_lookup_1): Update calls to cselib_find_slot.
236 2014-04-26 Markus Trippelsdorf <markus@trippelsdorf.de>
238 * doc/install.texi (Building with profile feedback): Remove
241 2014-04-26 Tom de Vries <tom@codesourcery.com>
243 * config/i386/i386.md (define_expand "ldexpxf3"): Fix out-of-bounds
246 2014-04-25 Cary Coutant <ccoutant@google.com>
249 * dwarf2out.c (should_move_die_to_comdat): A type definition
250 can contain a subprogram definition, but don't move it to a
252 (clone_as_declaration): Copy DW_AT_abstract_origin attribute.
253 (generate_skeleton_bottom_up): Remove DW_AT_object_pointer attribute
255 (clone_tree_hash): Rename to...
256 (clone_tree_partial): ...this; change callers. Copy
257 DW_TAG_subprogram DIEs as declarations.
258 (copy_decls_walk): Don't copy children of a declaration into a
261 2014-04-25 H.J. Lu <hongjiu.lu@intel.com>
264 * config/i386/i386.md (*movsf_internal): Set MODE to SI for
267 2014-04-25 Jiong Wang <jiong.wang@arm.com>
269 * config/arm/predicates.md (call_insn_operand): Add long_call check.
270 * config/arm/arm.md (sibcall, sibcall_value): Force the address to
272 * config/arm/arm.c (arm_function_ok_for_sibcall): Remove long_call
275 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
277 * config/arm/arm.c (arm_cortex_a8_tune): Initialise
280 2014-04-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
282 PR tree-optimization/60930
283 * gimple-ssa-strength-reduction.c (create_mul_imm_cand): Reject
284 creating a multiply candidate by folding two constant
285 multiplicands when the result overflows.
287 2014-04-25 Jakub Jelinek <jakub@redhat.com>
289 PR tree-optimization/60960
290 * tree-vect-generic.c (expand_vector_operation): Only call
291 expand_vector_divmod if type's mode satisfies VECTOR_MODE_P.
293 2014-04-25 Tom de Vries <tom@codesourcery.com>
295 * expr.c (clobber_reg_mode): New function.
296 * expr.h (clobber_reg): New function.
298 2014-04-25 Tom de Vries <tom@codesourcery.com>
300 * rtlanal.c (find_all_hard_reg_sets): Note INSN_CALL_FUNCTION_USAGE
303 2014-04-25 Radovan Obradovic <robradovic@mips.com>
304 Tom de Vries <tom@codesourcery.com>
306 * rtlanal.c (find_all_hard_reg_sets): Add bool implicit parameter and
308 * rtl.h (find_all_hard_reg_sets): Add bool parameter.
309 * haifa-sched.c (recompute_todo_spec, check_clobbered_conditions): Add
310 new argument to find_all_hard_reg_sets call.
312 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
314 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p):
315 Use HOST_WIDE_INT_C for mask literal.
316 (aarch_rev16_shleft_mask_imm_p): Likewise.
318 2014-04-25 Eric Botcazou <ebotcazou@adacore.com>
321 * config/sparc/sparc.md (ashlsi3_extend): Delete.
323 2014-04-25 Marc Glisse <marc.glisse@inria.fr>
325 PR preprocessor/56540
326 * config/i386/i386-c.c (ix86_target_macros): Define
327 __SIZEOF_FLOAT80__ and __SIZEOF_FLOAT128__.
329 2014-04-25 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
331 * configure.ac (tga_func): Remove.
332 (LIB_TLS_SPEC): Remove.
333 * configure: Regenerate.
334 * config.in: Regenerate.
335 * config/sol2.h (LIB_SPEC): Don't use LIB_TLS_SPEC.
337 2014-04-25 Richard Biener <rguenther@suse.de>
340 * tree-ssa-structalias.c (ipa_pta_execute): Compute direct
341 call stmt use/clobber sets during stmt walk instead of
342 walking the possibly incomplete set of caller edges.
344 2014-04-25 Richard Biener <rguenther@suse.de>
347 * passes.c (apply_ipa_transforms): Inline into only caller ...
348 (execute_one_pass): ... here. Properly bring in function
349 bodies for nodes we want to apply IPA transforms to.
351 2014-04-24 Cong Hou <congh@google.com>
353 PR tree-optimization/60896
354 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Pick up
355 all statements in PATTERN_DEF_SEQ in recognized widen-mult pattern.
356 (vect_mark_pattern_stmts): Set the def type of all statements in
357 PATTERN_DEF_SEQ as vect_internal_def.
359 2014-04-24 Michael Meissner <meissner@linux.vnet.ibm.com>
361 * doc/extend.texi (PowerPC Built-in Functions): Document new
362 powerpc extended divide, bcd, pack/unpack 128-bit, builtin
364 (PowerPC AltiVec/VSX Built-in Functions): Likewise.
366 * config/rs6000/predicates.md (const_0_to_3_operand): New
367 predicate to match 0..3 integer constants.
369 * config/rs6000/rs6000-builtin.def (BU_DFP_MISC_1): Add new macros
370 to support adding miscellaneous builtin functions.
371 (BU_DFP_MISC_2): Likewise.
372 (BU_P7_MISC_1): Likewise.
373 (BU_P7_MISC_2): Likewise.
374 (BU_P8V_MISC_3): Likewise.
375 (BU_MISC_1): Likewise.
376 (BU_MISC_2): Likewise.
377 (DIVWE): Add extended divide builtin functions.
385 (DXEX): Add decimal floating-point builtin functions.
397 (CDTBCD): Add new BCD builtin functions.
401 (BCDADD_LT): Likewise.
402 (BCDADD_EQ): Likewise.
403 (BCDADD_GT): Likewise.
404 (BCDADD_OV): Likewise.
406 (BCDSUB_LT): Likewise.
407 (BCDSUB_EQ): Likewise.
408 (BCDSUB_GT): Likewise.
409 (BCDSUB_OV): Likewise.
410 (PACK_TD): Add new pack/unpack 128-bit type builtin functions.
411 (UNPACK_TD): Likewise.
413 (UNPACK_TF): Likewise.
414 (UNPACK_TF_0): Likewise.
415 (UNPACK_TF_1): Likewise.
416 (PACK_V1TI): Likewise.
417 (UNPACK_V1TI): Likewise.
419 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Add
420 support for decimal floating point builtin functions.
421 (rs6000_expand_ternop_builtin): Add checks for the new builtin
422 functions that take constant arguments.
423 (rs6000_invalid_builtin): Add decimal floating point builtin
425 (rs6000_init_builtins): Setup long double, _Decimal64, and
426 _Decimal128 types for new builtin functions.
427 (builtin_function_type): Set the unsigned flags appropriately for
428 the new builtin functions.
429 (rs6000_opt_masks): Add support for decimal floating point builtin
432 * config/rs6000/rs6000.h (RS6000_BTM_DFP): Add support for decimal
433 floating point builtin functions.
434 (RS6000_BTM_COMMON): Likewise.
435 (RS6000_BTI_long_double): Likewise.
436 (RS6000_BTI_dfloat64): Likewise.
437 (RS6000_BTI_dfloat128): Likewise.
438 (long_double_type_internal_node): Likewise.
439 (dfloat64_type_internal_node): Likewise.
440 (dfloat128_type_internal_node): Likewise.
442 * config/rs6000/altivec.h (UNSPEC_BCDADD): Add support for ISA
443 2.07 bcd arithmetic instructions.
444 (UNSPEC_BCDSUB): Likewise.
445 (UNSPEC_BCD_OVERFLOW): Likewise.
446 (UNSPEC_BCD_ADD_SUB): Likewise.
447 (bcd_add_sub): Likewise.
448 (BCD_TEST): Likewise.
449 (bcd<bcd_add_sub>): Likewise.
450 (bcd<bcd_add_sub>_test): Likewise.
451 (bcd<bcd_add_sub>_test2): Likewise.
452 (bcd<bcd_add_sub>_<code>): Likewise.
453 (peephole2 for combined bcd ops): Likewise.
455 * config/rs6000/dfp.md (UNSPEC_DDEDPD): Add support for new
456 decimal floating point builtin functions.
457 (UNSPEC_DENBCD): Likewise.
458 (UNSPEC_DXEX): Likewise.
459 (UNSPEC_DIEX): Likewise.
460 (UNSPEC_DSCLI): Likewise.
461 (UNSPEC_DSCRI): Likewise.
462 (D64_D128): Likewise.
463 (dfp_suffix): Likewise.
464 (dfp_ddedpd_<mode>): Likewise.
465 (dfp_denbcd_<mode>): Likewise.
466 (dfp_dxex_<mode>): Likewise.
467 (dfp_diex_<mode>): Likewise.
468 (dfp_dscli_<mode>): Likewise.
469 (dfp_dscri_<mode>): Likewise.
471 * config/rs6000/rs6000.md (UNSPEC_ADDG6S): Add support for new BCD
473 (UNSPEC_CDTBCD): Likewise.
474 (UNSPEC_CBCDTD): Likewise.
475 (UNSPEC_DIVE): Add support for new extended divide builtin
477 (UNSPEC_DIVEO): Likewise.
478 (UNSPEC_DIVEU): Likewise.
479 (UNSPEC_DIVEUO): Likewise.
480 (UNSPEC_UNPACK_128BIT): Add support for new builtin functions to
481 pack/unpack 128-bit types.
482 (UNSPEC_PACK_128BIT): Likewise.
483 (idiv_ldiv): New mode attribute to set the 32/64-bit divide type.
484 (udiv<mode>3): Use idiv_ldiv mode attribute.
485 (div<mode>3): Likewise.
486 (addg6s): Add new BCD builtin functions.
489 (UNSPEC_DIV_EXTEND): Add support for new extended divide
491 (div_extend): Likewise.
492 (div<div_extend>_<mode>"): Likewise.
493 (FP128_64): Add support for new builtin functions to pack/unpack
495 (unpack<mode>): Likewise.
496 (unpacktf_0): Likewise.
497 (unpacktf_1): Likewise.
498 (unpack<mode>_dm): Likewise.
499 (unpack<mode>_nodm): Likewise.
500 (pack<mode>): Likewise.
501 (unpackv1ti): Likewise.
502 (packv1ti): Likewise.
504 2014-04-24 Vishnu K S <Vishnu.k_s@atmel.com>
506 * gcc/config/avr/avr.c: Add comment on why -fdelete-null-pointer-checks
509 2014-04-24 Jakub Jelinek <jakub@redhat.com>
511 * tree.h (OMP_CLAUSE_LINEAR_GIMPLE_SEQ): Define.
512 * gimplify.c (omp_is_private): Change last argument's type to int.
513 Only diagnose lastprivate if the simd argument is 1, only diagnose
514 linear if the simd argument is 2.
515 (gimplify_omp_for): Adjust omp_is_private callers. When adding
516 lastprivate or private, add the clause to OMP_FOR_CLAUSES. Pass
517 GOVD_EXPLICIT to omp_add_variable. For simd with collapse == 1
518 create OMP_CLAUSE_LINEAR rather than OMP_CLAUSE_PRIVATE for var.
519 If var != decl and decl is in OMP_CLAUSE_LINEAR, gimplify decl
520 increment to OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
521 * omp-low.c (scan_sharing_clauses, lower_lastprivate_clauses): Handle
522 OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
523 * tree-nested.c (convert_nonlocal_omp_clauses,
524 convert_local_omp_clauses): Handle OMP_CLAUSE_LINEAR.
526 2014-04-24 Segher Boessenkool <segher@kernel.crashing.org>
529 * config/m68k/m68k.md (extendplussidi): Don't allow memory for
532 2014-04-24 Dimitris Papavasiliou <dpapavas@gmail.com>
534 * flag-types.h (enum ivar_visibility): Add.
536 2014-04-24 Trevor Saunders <tsaunders@mozilla.com>
538 * config/sh/sh_treg_combine.c (sh_treg_combine::execute): Take
541 2014-04-24 Alan Lawrence <alan.lawrence@arm.com>
543 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Enable for bigendian.
545 2014-04-24 Radovan Obradovic <robradovic@mips.com>
546 Tom de Vries <tom@codesourcery.com>
548 * reg-notes.def (REG_NOTE (CALL_DECL)): New reg-note REG_CALL_DECL.
549 * calls.c (expand_call, emit_library_call_value_1): Add REG_CALL_DECL
551 * combine.c (distribute_notes): Handle REG_CALL_DECL reg-note.
552 * emit-rtl.c (try_split): Same.
554 2014-04-24 Radovan Obradovic <robradovic@mips.com>
555 Tom de Vries <tom@codesourcery.com>
557 * common.opt (fuse-caller-save): New option.
559 2014-04-24 Tejas Belagod <tejas.belagod@arm.com>
561 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Reverse order of elements
564 2014-04-24 Richard Biener <rguenther@suse.de>
566 * expr.c (expand_expr_real_1): Avoid gimple_assign_rhs_to_tree
567 during TER and instead use the sepops interface for expanding
568 non-GIMPLE_SINGLE_RHS.
570 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
572 * config/i386/sol2.h (ASM_PREFERRED_EH_DATA_FORMAT): Only redefine
573 if not HAVE_AS_IX86_DIFF_SECT_DELTA.
575 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
577 * configure.ac (gcc_cv_as_cfi_directive): Support Solaris/x86
578 assembler 64-bit option.
579 * configure: Regenerate.
581 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
583 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Check
584 TARGET_SIMD rather than TARGET_GENERAL_REGS_ONLY.
585 (TARGET_SIMD): Take AARCH64_ISA_SIMD into account.
586 (TARGET_FLOAT): Take AARCH64_ISA_FP into account.
587 (TARGET_CRYPTO): Take TARGET_SIMD into account.
589 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
591 * config/aarch64/aarch64-builtins.c
592 (aarch64_builtin_vectorized_function): Handle BUILT_IN_BSWAP16,
593 BUILT_IN_BSWAP32, BUILT_IN_BSWAP64.
594 * config/aarch64/aarch64-simd.md (bswap<mode>): New pattern.
595 * config/aarch64/aarch64-simd-builtins.def: Define vector bswap
597 * config/aarch64/iterator.md (VDQHSD): New mode iterator.
598 (Vrevsuff): New mode attribute.
600 2014-04-24 Terry Guo <terry.guo@arm.com>
602 * config/arm/arm.h (machine_function): Define variable
603 after_arm_reorg here.
604 * config/arm/arm.c (after_arm_reorg): Remove the definition.
605 (arm_split_constant): Update the way to access variable
608 (arm_output_function_epilogue): Remove the reset of after_arm_reorg.
610 2014-04-23 Tom de Vries <tom@codesourcery.com>
612 * target-hooks-macros.h: Fix DEFHOOKPOD argument order in comment.
614 2014-04-23 David Malcolm <dmalcolm@redhat.com>
616 * is-a.h: Update comments to reflect the following changes to the
617 "pointerness" of the API, making the template parameter match the
618 return type, allowing use of is-a.h with typedefs of pointers.
619 (is_a_helper::cast): Return a T rather then a pointer to a T, so
620 that the return type matches the parameter to the is_a_helper.
622 (dyn_cast): Likewise.
624 * cgraph.c (cgraph_node_for_asm): Update for removal of implicit
625 pointer from the is-a.h API.
627 * cgraph.h (is_a_helper <cgraph_node>::test): Convert to...
628 (is_a_helper <cgraph_node *>::test): ...this, matching change to
630 (is_a_helper <varpool_node>::test): Likewise, convert to...
631 (is_a_helper <varpool_node *>::test): ...this.
633 (varpool_first_variable): Update for removal of implicit pointer
635 (varpool_next_variable): Likewise.
636 (varpool_first_static_initializer): Likewise.
637 (varpool_next_static_initializer): Likewise.
638 (varpool_first_defined_variable): Likewise.
639 (varpool_next_defined_variable): Likewise.
640 (cgraph_first_defined_function): Likewise.
641 (cgraph_next_defined_function): Likewise.
642 (cgraph_first_function): Likewise.
643 (cgraph_next_function): Likewise.
644 (cgraph_first_function_with_gimple_body): Likewise.
645 (cgraph_next_function_with_gimple_body): Likewise.
646 (cgraph_alias_target): Likewise.
647 (varpool_alias_target): Likewise.
648 (cgraph_function_or_thunk_node): Likewise.
649 (varpool_variable_node): Likewise.
650 (symtab_real_symbol_p): Likewise.
651 * cgraphunit.c (referred_to_p): Likewise.
652 (analyze_functions): Likewise.
653 (handle_alias_pairs): Likewise.
654 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
655 * gimple-ssa.h (gimple_vuse_op): Likewise.
656 (gimple_vdef_op): Likewise.
657 * gimple-streamer-in.c (input_gimple_stmt): Likewise.
658 * gimple.c (gimple_build_asm_1): Likewise.
659 (gimple_build_try): Likewise.
660 (gimple_build_resx): Likewise.
661 (gimple_build_eh_dispatch): Likewise.
662 (gimple_build_omp_for): Likewise.
663 (gimple_omp_for_set_clauses): Likewise.
665 * gimple.h (is_a_helper <gimple_statement_asm>::test): Convert to...
666 (is_a_helper <gimple_statement_asm *>::test): ...this.
667 (is_a_helper <gimple_statement_bind>::test): Convert to...
668 (is_a_helper <gimple_statement_bind *>::test): ...this.
669 (is_a_helper <gimple_statement_call>::test): Convert to...
670 (is_a_helper <gimple_statement_call *>::test): ...this.
671 (is_a_helper <gimple_statement_catch>::test): Convert to...
672 (is_a_helper <gimple_statement_catch *>::test): ...this.
673 (is_a_helper <gimple_statement_resx>::test): Convert to...
674 (is_a_helper <gimple_statement_resx *>::test): ...this.
675 (is_a_helper <gimple_statement_eh_dispatch>::test): Convert to...
676 (is_a_helper <gimple_statement_eh_dispatch *>::test): ...this.
677 (is_a_helper <gimple_statement_eh_else>::test): Convert to...
678 (is_a_helper <gimple_statement_eh_else *>::test): ...this.
679 (is_a_helper <gimple_statement_eh_filter>::test): Convert to...
680 (is_a_helper <gimple_statement_eh_filter *>::test): ...this.
681 (is_a_helper <gimple_statement_eh_mnt>::test): Convert to...
682 (is_a_helper <gimple_statement_eh_mnt *>::test): ...this.
683 (is_a_helper <gimple_statement_omp_atomic_load>::test): Convert to...
684 (is_a_helper <gimple_statement_omp_atomic_load *>::test): ...this.
685 (is_a_helper <gimple_statement_omp_atomic_store>::test): Convert to...
686 (is_a_helper <gimple_statement_omp_atomic_store *>::test): ...this.
687 (is_a_helper <gimple_statement_omp_return>::test): Convert to...
688 (is_a_helper <gimple_statement_omp_return *>::test): ...this.
689 (is_a_helper <gimple_statement_omp_continue>::test): Convert to...
690 (is_a_helper <gimple_statement_omp_continue *>::test): ...this.
691 (is_a_helper <gimple_statement_omp_critical>::test): Convert to...
692 (is_a_helper <gimple_statement_omp_critical *>::test): ...this.
693 (is_a_helper <gimple_statement_omp_for>::test): Convert to...
694 (is_a_helper <gimple_statement_omp_for *>::test): ...this.
695 (is_a_helper <gimple_statement_omp_taskreg>::test): Convert to...
696 (is_a_helper <gimple_statement_omp_taskreg *>::test): ...this.
697 (is_a_helper <gimple_statement_omp_parallel>::test): Convert to...
698 (is_a_helper <gimple_statement_omp_parallel *>::test): ...this.
699 (is_a_helper <gimple_statement_omp_target>::test): Convert to...
700 (is_a_helper <gimple_statement_omp_target *>::test): ...this.
701 (is_a_helper <gimple_statement_omp_sections>::test): Convert to...
702 (is_a_helper <gimple_statement_omp_sections *>::test): ...this.
703 (is_a_helper <gimple_statement_omp_single>::test): Convert to...
704 (is_a_helper <gimple_statement_omp_single *>::test): ...this.
705 (is_a_helper <gimple_statement_omp_teams>::test): Convert to...
706 (is_a_helper <gimple_statement_omp_teams *>::test): ...this.
707 (is_a_helper <gimple_statement_omp_task>::test): Convert to...
708 (is_a_helper <gimple_statement_omp_task *>::test): ...this.
709 (is_a_helper <gimple_statement_phi>::test): Convert to...
710 (is_a_helper <gimple_statement_phi *>::test): ...this.
711 (is_a_helper <gimple_statement_transaction>::test): Convert to...
712 (is_a_helper <gimple_statement_transaction *>::test): ...this.
713 (is_a_helper <gimple_statement_try>::test): Convert to...
714 (is_a_helper <gimple_statement_try *>::test): ...this.
715 (is_a_helper <gimple_statement_wce>::test): Convert to...
716 (is_a_helper <gimple_statement_wce *>::test): ...this.
717 (is_a_helper <const gimple_statement_asm>::test): Convert to...
718 (is_a_helper <const gimple_statement_asm *>::test): ...this.
719 (is_a_helper <const gimple_statement_bind>::test): Convert to...
720 (is_a_helper <const gimple_statement_bind *>::test): ...this.
721 (is_a_helper <const gimple_statement_call>::test): Convert to...
722 (is_a_helper <const gimple_statement_call *>::test): ...this.
723 (is_a_helper <const gimple_statement_catch>::test): Convert to...
724 (is_a_helper <const gimple_statement_catch *>::test): ...this.
725 (is_a_helper <const gimple_statement_resx>::test): Convert to...
726 (is_a_helper <const gimple_statement_resx *>::test): ...this.
727 (is_a_helper <const gimple_statement_eh_dispatch>::test):
729 (is_a_helper <const gimple_statement_eh_dispatch *>::test):
731 (is_a_helper <const gimple_statement_eh_filter>::test): Convert
733 (is_a_helper <const gimple_statement_eh_filter *>::test): ...this.
734 (is_a_helper <const gimple_statement_omp_atomic_load>::test):
736 (is_a_helper <const gimple_statement_omp_atomic_load *>::test):
738 (is_a_helper <const gimple_statement_omp_atomic_store>::test):
740 (is_a_helper <const gimple_statement_omp_atomic_store *>::test):
742 (is_a_helper <const gimple_statement_omp_return>::test): Convert
744 (is_a_helper <const gimple_statement_omp_return *>::test): ...this.
745 (is_a_helper <const gimple_statement_omp_continue>::test): Convert
747 (is_a_helper <const gimple_statement_omp_continue *>::test): ...this.
748 (is_a_helper <const gimple_statement_omp_critical>::test): Convert
750 (is_a_helper <const gimple_statement_omp_critical *>::test): ...this.
751 (is_a_helper <const gimple_statement_omp_for>::test): Convert to...
752 (is_a_helper <const gimple_statement_omp_for *>::test): ...this.
753 (is_a_helper <const gimple_statement_omp_taskreg>::test): Convert
755 (is_a_helper <const gimple_statement_omp_taskreg *>::test): ...this.
756 (is_a_helper <const gimple_statement_omp_parallel>::test): Convert
758 (is_a_helper <const gimple_statement_omp_parallel *>::test): ...this.
759 (is_a_helper <const gimple_statement_omp_target>::test): Convert
761 (is_a_helper <const gimple_statement_omp_target *>::test): ...this.
762 (is_a_helper <const gimple_statement_omp_sections>::test): Convert
764 (is_a_helper <const gimple_statement_omp_sections *>::test): ...this.
765 (is_a_helper <const gimple_statement_omp_single>::test): Convert
767 (is_a_helper <const gimple_statement_omp_single *>::test): ...this.
768 (is_a_helper <const gimple_statement_omp_teams>::test): Convert
770 (is_a_helper <const gimple_statement_omp_teams *>::test): ...this.
771 (is_a_helper <const gimple_statement_omp_task>::test): Convert
773 (is_a_helper <const gimple_statement_omp_task *>::test): ...this.
774 (is_a_helper <const gimple_statement_phi>::test): Convert to...
775 (is_a_helper <const gimple_statement_phi *>::test): ...this.
776 (is_a_helper <const gimple_statement_transaction>::test): Convert
778 (is_a_helper <const gimple_statement_transaction *>::test): ...this.
779 (is_a_helper <const gimple_statement_with_ops>::test): Convert
781 (is_a_helper <const gimple_statement_with_ops *>::test): ...this.
782 (is_a_helper <gimple_statement_with_ops>::test): Convert to...
783 (is_a_helper <gimple_statement_with_ops *>::test): ...this.
784 (is_a_helper <const gimple_statement_with_memory_ops>::test):
786 (is_a_helper <const gimple_statement_with_memory_ops *>::test):
788 (is_a_helper <gimple_statement_with_memory_ops>::test): Convert
790 (is_a_helper <gimple_statement_with_memory_ops *>::test): ...this.
792 (gimple_use_ops): Update for removal of implicit pointer from the
794 (gimple_set_use_ops): Likewise.
795 (gimple_vuse): Likewise.
796 (gimple_vdef): Likewise.
797 (gimple_vuse_ptr): Likewise.
798 (gimple_vdef_ptr): Likewise.
799 (gimple_set_vuse): Likewise.
800 (gimple_set_vdef): Likewise.
801 (gimple_omp_return_set_lhs): Likewise.
802 (gimple_omp_return_lhs): Likewise.
803 (gimple_omp_return_lhs_ptr): Likewise.
804 (gimple_call_fntype): Likewise.
805 (gimple_call_set_fntype): Likewise.
806 (gimple_call_set_internal_fn): Likewise.
807 (gimple_call_use_set): Likewise.
808 (gimple_call_clobber_set): Likewise.
809 (gimple_bind_vars): Likewise.
810 (gimple_bind_set_vars): Likewise.
811 (gimple_bind_body_ptr): Likewise.
812 (gimple_bind_set_body): Likewise.
813 (gimple_bind_add_stmt): Likewise.
814 (gimple_bind_block): Likewise.
815 (gimple_bind_set_block): Likewise.
816 (gimple_asm_ninputs): Likewise.
817 (gimple_asm_noutputs): Likewise.
818 (gimple_asm_nclobbers): Likewise.
819 (gimple_asm_nlabels): Likewise.
820 (gimple_asm_input_op): Likewise.
821 (gimple_asm_input_op_ptr): Likewise.
822 (gimple_asm_output_op): Likewise.
823 (gimple_asm_output_op_ptr): Likewise.
824 (gimple_asm_set_output_op): Likewise.
825 (gimple_asm_clobber_op): Likewise.
826 (gimple_asm_set_clobber_op): Likewise.
827 (gimple_asm_label_op): Likewise.
828 (gimple_asm_set_label_op): Likewise.
829 (gimple_asm_string): Likewise.
830 (gimple_catch_types): Likewise.
831 (gimple_catch_types_ptr): Likewise.
832 (gimple_catch_handler_ptr): Likewise.
833 (gimple_catch_set_types): Likewise.
834 (gimple_catch_set_handler): Likewise.
835 (gimple_eh_filter_types): Likewise.
836 (gimple_eh_filter_types_ptr): Likewise.
837 (gimple_eh_filter_failure_ptr): Likewise.
838 (gimple_eh_filter_set_types): Likewise.
839 (gimple_eh_filter_set_failure): Likewise.
840 (gimple_eh_must_not_throw_fndecl): Likewise.
841 (gimple_eh_must_not_throw_set_fndecl): Likewise.
842 (gimple_eh_else_n_body_ptr): Likewise.
843 (gimple_eh_else_e_body_ptr): Likewise.
844 (gimple_eh_else_set_n_body): Likewise.
845 (gimple_eh_else_set_e_body): Likewise.
846 (gimple_try_eval_ptr): Likewise.
847 (gimple_try_cleanup_ptr): Likewise.
848 (gimple_try_set_eval): Likewise.
849 (gimple_try_set_cleanup): Likewise.
850 (gimple_wce_cleanup_ptr): Likewise.
851 (gimple_wce_set_cleanup): Likewise.
852 (gimple_phi_capacity): Likewise.
853 (gimple_phi_num_args): Likewise.
854 (gimple_phi_result): Likewise.
855 (gimple_phi_result_ptr): Likewise.
856 (gimple_phi_set_result): Likewise.
857 (gimple_phi_arg): Likewise.
858 (gimple_phi_set_arg): Likewise.
859 (gimple_resx_region): Likewise.
860 (gimple_resx_set_region): Likewise.
861 (gimple_eh_dispatch_region): Likewise.
862 (gimple_eh_dispatch_set_region): Likewise.
863 (gimple_omp_critical_name): Likewise.
864 (gimple_omp_critical_name_ptr): Likewise.
865 (gimple_omp_critical_set_name): Likewise.
866 (gimple_omp_for_clauses): Likewise.
867 (gimple_omp_for_clauses_ptr): Likewise.
868 (gimple_omp_for_set_clauses): Likewise.
869 (gimple_omp_for_collapse): Likewise.
870 (gimple_omp_for_index): Likewise.
871 (gimple_omp_for_index_ptr): Likewise.
872 (gimple_omp_for_set_index): Likewise.
873 (gimple_omp_for_initial): Likewise.
874 (gimple_omp_for_initial_ptr): Likewise.
875 (gimple_omp_for_set_initial): Likewise.
876 (gimple_omp_for_final): Likewise.
877 (gimple_omp_for_final_ptr): Likewise.
878 (gimple_omp_for_set_final): Likewise.
879 (gimple_omp_for_incr): Likewise.
880 (gimple_omp_for_incr_ptr): Likewise.
881 (gimple_omp_for_set_incr): Likewise.
882 (gimple_omp_for_pre_body_ptr): Likewise.
883 (gimple_omp_for_set_pre_body): Likewise.
884 (gimple_omp_parallel_clauses): Likewise.
885 (gimple_omp_parallel_clauses_ptr): Likewise.
886 (gimple_omp_parallel_set_clauses): Likewise.
887 (gimple_omp_parallel_child_fn): Likewise.
888 (gimple_omp_parallel_child_fn_ptr): Likewise.
889 (gimple_omp_parallel_set_child_fn): Likewise.
890 (gimple_omp_parallel_data_arg): Likewise.
891 (gimple_omp_parallel_data_arg_ptr): Likewise.
892 (gimple_omp_parallel_set_data_arg): Likewise.
893 (gimple_omp_task_clauses): Likewise.
894 (gimple_omp_task_clauses_ptr): Likewise.
895 (gimple_omp_task_set_clauses): Likewise.
896 (gimple_omp_task_child_fn): Likewise.
897 (gimple_omp_task_child_fn_ptr): Likewise.
898 (gimple_omp_task_set_child_fn): Likewise.
899 (gimple_omp_task_data_arg): Likewise.
900 (gimple_omp_task_data_arg_ptr): Likewise.
901 (gimple_omp_task_set_data_arg): Likewise.
902 (gimple_omp_taskreg_clauses): Likewise.
903 (gimple_omp_taskreg_clauses_ptr): Likewise.
904 (gimple_omp_taskreg_set_clauses): Likewise.
905 (gimple_omp_taskreg_child_fn): Likewise.
906 (gimple_omp_taskreg_child_fn_ptr): Likewise.
907 (gimple_omp_taskreg_set_child_fn): Likewise.
908 (gimple_omp_taskreg_data_arg): Likewise.
909 (gimple_omp_taskreg_data_arg_ptr): Likewise.
910 (gimple_omp_taskreg_set_data_arg): Likewise.
911 (gimple_omp_task_copy_fn): Likewise.
912 (gimple_omp_task_copy_fn_ptr): Likewise.
913 (gimple_omp_task_set_copy_fn): Likewise.
914 (gimple_omp_task_arg_size): Likewise.
915 (gimple_omp_task_arg_size_ptr): Likewise.
916 (gimple_omp_task_set_arg_size): Likewise.
917 (gimple_omp_task_arg_align): Likewise.
918 (gimple_omp_task_arg_align_ptr): Likewise.
919 (gimple_omp_task_set_arg_align): Likewise.
920 (gimple_omp_single_clauses): Likewise.
921 (gimple_omp_single_clauses_ptr): Likewise.
922 (gimple_omp_single_set_clauses): Likewise.
923 (gimple_omp_target_clauses): Likewise.
924 (gimple_omp_target_clauses_ptr): Likewise.
925 (gimple_omp_target_set_clauses): Likewise.
926 (gimple_omp_target_child_fn): Likewise.
927 (gimple_omp_target_child_fn_ptr): Likewise.
928 (gimple_omp_target_set_child_fn): Likewise.
929 (gimple_omp_target_data_arg): Likewise.
930 (gimple_omp_target_data_arg_ptr): Likewise.
931 (gimple_omp_target_set_data_arg): Likewise.
932 (gimple_omp_teams_clauses): Likewise.
933 (gimple_omp_teams_clauses_ptr): Likewise.
934 (gimple_omp_teams_set_clauses): Likewise.
935 (gimple_omp_sections_clauses): Likewise.
936 (gimple_omp_sections_clauses_ptr): Likewise.
937 (gimple_omp_sections_set_clauses): Likewise.
938 (gimple_omp_sections_control): Likewise.
939 (gimple_omp_sections_control_ptr): Likewise.
940 (gimple_omp_sections_set_control): Likewise.
941 (gimple_omp_for_set_cond): Likewise.
942 (gimple_omp_for_cond): Likewise.
943 (gimple_omp_atomic_store_set_val): Likewise.
944 (gimple_omp_atomic_store_val): Likewise.
945 (gimple_omp_atomic_store_val_ptr): Likewise.
946 (gimple_omp_atomic_load_set_lhs): Likewise.
947 (gimple_omp_atomic_load_lhs): Likewise.
948 (gimple_omp_atomic_load_lhs_ptr): Likewise.
949 (gimple_omp_atomic_load_set_rhs): Likewise.
950 (gimple_omp_atomic_load_rhs): Likewise.
951 (gimple_omp_atomic_load_rhs_ptr): Likewise.
952 (gimple_omp_continue_control_def): Likewise.
953 (gimple_omp_continue_control_def_ptr): Likewise.
954 (gimple_omp_continue_set_control_def): Likewise.
955 (gimple_omp_continue_control_use): Likewise.
956 (gimple_omp_continue_control_use_ptr): Likewise.
957 (gimple_omp_continue_set_control_use): Likewise.
958 (gimple_transaction_body_ptr): Likewise.
959 (gimple_transaction_label): Likewise.
960 (gimple_transaction_label_ptr): Likewise.
961 (gimple_transaction_set_body): Likewise.
962 (gimple_transaction_set_label): Likewise.
964 * ipa-devirt.c (build_type_inheritance_graph): Likewise.
965 * ipa-inline-analysis.c (inline_write_summary): Likewise.
966 * ipa-ref.c (ipa_record_reference): Likewise.
967 * ipa-reference.c (analyze_function): Likewise.
968 (ipa_reference_write_optimization_summary): Likewise.
969 * ipa.c (symtab_remove_unreachable_nodes): Likewise.
970 (address_taken_from_non_vtable_p): Likewise.
971 (comdat_can_be_unshared_p_1): Likewise.
972 * lto-cgraph.c (lto_output_ref): Likewise.
973 (add_references): Likewise.
974 (compute_ltrans_boundary): Likewise.
975 (output_symtab): Likewise.
976 (input_ref): Likewise.
977 (input_cgraph_1): Likewise.
978 (output_cgraph_opt_summary): Likewise.
979 * lto-streamer-out.c (lto_output): Likewise.
980 (output_symbol_p): Likewise.
981 * lto-streamer.h (lsei_next_function_in_partition): Likewise.
982 (lsei_start_function_in_partition): Likewise.
983 (lsei_next_variable_in_partition): Likewise.
984 (lsei_start_variable_in_partition): Likewise.
985 * symtab.c (insert_to_assembler_name_hash): Likewise.
986 (unlink_from_assembler_name_hash): Likewise.
987 (symtab_unregister_node): Likewise.
988 (symtab_remove_node): Likewise.
989 (dump_symtab_node): Likewise.
990 (verify_symtab_base): Likewise.
991 (verify_symtab_node): Likewise.
992 (symtab_make_decl_local): Likewise.
993 (symtab_alias_ultimate_target): Likewise.
994 (symtab_resolve_alias): Likewise.
995 (symtab_get_symbol_partitioning_class): Likewise.
996 * tree-phinodes.c (allocate_phi_node): Likewise.
997 (reserve_phi_args_for_new_edge): Likewise.
998 (remove_phi_args): Likewise.
999 * varpool.c (varpool_node_for_asm): Likewise.
1000 (varpool_remove_unreferenced_decls): Likewise.
1002 2014-04-23 Jeff Law <law@redhat.com>
1004 PR tree-optimization/60902
1005 * tree-ssa-threadedge.c
1006 (record_temporary_equivalences_from_stmts_at_dest): Make sure to
1007 invalidate outputs from statements that do not produce useful
1008 outputs for threading.
1010 2014-04-23 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
1012 * config/aarch64/aarch64.md (stack_protect_set, stack_protect_test)
1013 (stack_protect_set_<mode>, stack_protect_test_<mode>): Add
1014 machine descriptions for Stack Smashing Protector.
1016 2014-04-23 Richard Earnshaw <rearnsha@arm.com>
1018 * aarch64.md (<optab>_rol<mode>3): New pattern.
1019 (<optab>_rolsi3_uxtw): Likewise.
1020 * aarch64.c (aarch64_strip_shift): Handle ROTATE and ROTATERT.
1022 2014-04-23 James Greenhalgh <james.greenhalgh@arm.com>
1024 * config/arm/arm.c (arm_cortex_a57_tune): Initialize all fields.
1025 (arm_cortex_a12_tune): Likewise.
1027 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1029 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle BSWAP.
1031 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1033 * config/arm/arm.md (arm_rev16si2): New pattern.
1034 (arm_rev16si2_alt): Likewise.
1035 * config/arm/arm.c (arm_new_rtx_costs): Handle rev16 case.
1037 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1039 * config/aarch64/aarch64.md (rev16<mode>2): New pattern.
1040 (rev16<mode>2_alt): Likewise.
1041 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle rev16 case.
1042 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p): New.
1043 (aarch_rev16_shleft_mask_imm_p): Likewise.
1044 (aarch_rev16_p_1): Likewise.
1045 (aarch_rev16_p): Likewise.
1046 * config/arm/aarch-common-protos.h (aarch_rev16_p): Declare extern.
1047 (aarch_rev16_shright_mask_imm_p): Likewise.
1048 (aarch_rev16_shleft_mask_imm_p): Likewise.
1050 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1052 * config/arm/aarch-common-protos.h (alu_cost_table): Add rev field.
1053 * config/arm/aarch-cost-tables.h (generic_extra_costs): Specify
1055 (cortex_a53_extra_costs): Likewise.
1056 (cortex_a57_extra_costs): Likewise.
1057 * config/arm/arm.c (cortexa9_extra_costs): Likewise.
1058 (cortexa7_extra_costs): Likewise.
1059 (cortexa8_extra_costs): Likewise.
1060 (cortexa12_extra_costs): Likewise.
1061 (cortexa15_extra_costs): Likewise.
1062 (v7m_extra_costs): Likewise.
1063 (arm_new_rtx_costs): Handle BSWAP.
1065 2013-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1067 * config/arm/arm.c (cortexa8_extra_costs): New table.
1068 (arm_cortex_a8_tune): New tuning struct.
1069 * config/arm/arm-cores.def (cortex-a8): Use cortex_a8 tuning struct.
1071 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1073 * config/arm/arm.c (arm_new_rtx_costs): Handle FMA.
1075 2014-04-23 Richard Biener <rguenther@suse.de>
1077 * Makefile.in (OBJS): Remove loop-unswitch.o.
1078 * loop-unswitch.c: Delete.
1079 * tree-pass.h (make_pass_rtl_unswitch): Remove.
1080 * passes.def (pass_rtl_unswitch): Likewise.
1081 * loop-init.c (gate_rtl_unswitch): Likewise.
1082 (rtl_unswitch): Likewise.
1083 (pass_data_rtl_unswitch): Likewise.
1084 (pass_rtl_unswitch): Likewise.
1085 (make_pass_rtl_unswitch): Likewise.
1086 * rtl.h (reversed_condition): Likewise.
1087 (compare_and_jump_seq): Likewise.
1088 * loop-iv.c (reversed_condition): Move here from loop-unswitch.c
1090 * loop-unroll.c (compare_and_jump_seq): Likewise.
1092 2014-04-23 Richard Biener <rguenther@suse.de>
1094 PR tree-optimization/60903
1095 * tree-ssa-loop-im.c (analyze_memory_references): Remove
1096 commented code block.
1097 (execute_sm_if_changed): Properly apply IRREDUCIBLE_LOOP
1098 loop flags to newly created BBs and edges.
1100 2014-04-23 Nick Clifton <nickc@redhat.com>
1102 * config/msp430/msp430.c (msp430_handle_option): Move function
1104 (msp430_option_override): Simplify mcu and mcpu option handling.
1105 (msp430_is_f5_mcu): Rename to msp430_use_f5_series_hwmult. Add
1106 support for -mhwmult command line option.
1107 (has_32bit_hwmult): Rename to use_32bit_hwmult. Add support for
1108 -mhwmult command line option.
1109 (msp430_hwmult_enabled): Delete.
1110 (msp43o_output_labelref): Add support for -mhwmult command line
1112 * config/msp430/msp430.md (mulhisi3, umulhisi3, mulsidi3)
1113 (umulsidi3): Likewise.
1114 * config/msp430/msp430.opt (mmcu): Add Report attribute.
1115 (mcpu, mlarge, msmall): Likewise.
1116 (mhwmult): New option.
1117 * config/msp430/msp430-protos.h (msp430_hwmult_enabled): Remove
1119 (msp430_is_f5_mcu): Remove prototype.
1120 (msp430_use_f5_series_hwmult): Add prototype.
1121 * config/msp430/msp430-opts.h: New file.
1122 * common/config/msp430: New directory.
1123 * common/config/msp430/msp430-common.c: New file.
1124 * config.gcc (msp430): Remove target_has_targetm_common.
1125 * doc/invoke.texi: Document -mhwmult command line option.
1127 2014-04-23 Nick Clifton <nickc@redhat.com>
1129 * config/i386/cygwin.h (ENDFILE_SPEC): Include
1130 default-manifest.o if it can be found in the search path.
1131 * config/i386/mingw32.h (ENDFILE_SPEC): Likewise.
1133 2014-04-23 Terry Guo <terry.guo@arm.com>
1135 * config/arm/arm.h (ASM_APP_OFF): Re-define it in a cleaner way.
1137 2014-04-23 Richard Biener <rguenther@suse.de>
1140 * tree-inline.c (declare_return_variable): Use mark_addressable.
1142 2014-04-23 Richard Biener <rguenther@suse.de>
1145 * loop-init.c (loop_optimizer_init): Make sure to apply
1146 LOOPS_MAY_HAVE_MULTIPLE_LATCHES before fixing up loops.
1148 2014-04-22 Jakub Jelinek <jakub@redhat.com>
1151 * common.opt (fsanitize-recover, fsanitize-undefined-trap-on-error):
1153 * gcc.c (sanitize_spec_function): Don't return "" for "undefined"
1154 if flag_sanitize_undefined_trap_on_error.
1155 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW_ABORT,
1156 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS_ABORT,
1157 BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE_ABORT,
1158 BUILT_IN_UBSAN_HANDLE_TYPE_MISMATCH_ABORT,
1159 BUILT_IN_UBSAN_HANDLE_ADD_OVERFLOW_ABORT,
1160 BUILT_IN_UBSAN_HANDLE_SUB_OVERFLOW_ABORT,
1161 BUILT_IN_UBSAN_HANDLE_MUL_OVERFLOW_ABORT,
1162 BUILT_IN_UBSAN_HANDLE_NEGATE_OVERFLOW_ABORT,
1163 BUILT_IN_UBSAN_HANDLE_LOAD_INVALID_VALUE_ABORT): New builtins.
1164 * ubsan.c (ubsan_instrument_unreachable): Return
1165 __builtin_trap () if flag_sanitize_undefined_trap_on_error.
1166 (ubsan_expand_null_ifn): Emit __builtin_trap ()
1167 if flag_sanitize_undefined_trap_on_error and
1168 __ubsan_handle_type_mismatch_abort if !flag_sanitize_recover.
1169 (ubsan_expand_null_ifn, ubsan_build_overflow_builtin,
1170 instrument_bool_enum_load): Emit __builtin_trap () if
1171 flag_sanitize_undefined_trap_on_error and
1172 __builtin_handle_*_abort () if !flag_sanitize_recover.
1173 * doc/invoke.texi (-fsanitize-recover,
1174 -fsanitize-undefined-trap-on-error): Document.
1176 2014-04-22 Christian Bruel <christian.bruel@st.com>
1178 * config/sh/sh.md (mov<mode>): Replace movQIHI.
1179 Force immediates to SImode.
1181 2014-04-22 Sandra Loosemore <sandra@codesourcery.com>
1183 * config/nios2/nios2.md (UNSPEC_ROUND): New.
1185 * config/nios2/nios2.opt (mno-custom-round, mcustom-round=): New.
1186 * config/nios2/nios2-opts.h (N2FPU_ALL_CODES): Add round.
1187 * config/nios2/nios2.c (N2F_NO_ERRNO): Define.
1188 (nios2_fpu_insn): Add entry for round.
1189 (N2FPU_NO_ERRNO_P): Define.
1190 (nios2_custom_check_insns): Add check for N2F_NO_ERRNO and
1192 * doc/invoke.texi (Nios II Options): Document -mcustom-round.
1194 2014-04-22 Richard Henderson <rth@redhat.com>
1196 * config/aarch64/aarch64 (addti3, subti3): New expanders.
1197 (add<GPI>3_compare0): Remove leading * from name.
1198 (add<GPI>3_carryin): Likewise.
1199 (sub<GPI>3_compare0): Likewise.
1200 (sub<GPI>3_carryin): Likewise.
1201 (<su_optab>mulditi3): New expander.
1202 (multi3): New expander.
1203 (madd<GPI>): Remove leading * from name.
1205 2014-04-22 Martin Jambor <mjambor@suse.cz>
1207 * cgraphclones.c (cgraph_function_versioning): Copy
1208 ipa_transforms_to_apply instead of asserting it is empty.
1210 2014-04-22 H.J. Lu <hongjiu.lu@intel.com>
1213 * config/i386/i386.c (ix86_expand_set_or_movmem): Call counter_mode
1214 on count_exp to get mode.
1216 2014-04-22 Andrew Pinski <apinski@cavium.com>
1218 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
1219 Handle TLS for ILP32.
1220 * config/aarch64/aarch64.md (tlsie_small): Rename to ...
1221 (tlsie_small_<mode>): this and handle PTR.
1222 (tlsie_small_sidi): New pattern.
1223 (tlsle_small): Change to an expand to handle ILP32.
1224 (tlsle_small_<mode>): New pattern.
1225 (tlsdesc_small): Rename to ...
1226 (tlsdesc_small_<mode>): this and handle PTR.
1228 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1230 * config/aarch64/aarch64.c (TARGET_FLAGS_REGNUM): Define.
1232 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1234 * config/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
1235 (aarch64_types_signed_unsigned_qualifiers): Qualifier added.
1236 (aarch64_types_signed_poly_qualifiers): Likewise.
1237 (aarch64_types_unsigned_signed_qualifiers): Likewise.
1238 (aarch64_types_poly_signed_qualifiers): Likewise.
1239 (TYPES_REINTERP_SS): Type macro added.
1240 (TYPES_REINTERP_SU): Likewise.
1241 (TYPES_REINTERP_SP): Likewise.
1242 (TYPES_REINTERP_US): Likewise.
1243 (TYPES_REINTERP_PS): Likewise.
1244 (aarch64_fold_builtin): New expression folding added.
1245 * config/aarch64/aarch64-simd-builtins.def (REINTERP):
1246 Declarations removed.
1247 (REINTERP_SS): Declarations added.
1248 (REINTERP_US): Likewise.
1249 (REINTERP_PS): Likewise.
1250 (REINTERP_SU): Likewise.
1251 (REINTERP_SP): Likewise.
1252 * config/aarch64/arm_neon.h (vreinterpret_p8_f64): Implemented.
1253 (vreinterpretq_p8_f64): Likewise.
1254 (vreinterpret_p16_f64): Likewise.
1255 (vreinterpretq_p16_f64): Likewise.
1256 (vreinterpret_f32_f64): Likewise.
1257 (vreinterpretq_f32_f64): Likewise.
1258 (vreinterpret_f64_f32): Likewise.
1259 (vreinterpret_f64_p8): Likewise.
1260 (vreinterpret_f64_p16): Likewise.
1261 (vreinterpret_f64_s8): Likewise.
1262 (vreinterpret_f64_s16): Likewise.
1263 (vreinterpret_f64_s32): Likewise.
1264 (vreinterpret_f64_s64): Likewise.
1265 (vreinterpret_f64_u8): Likewise.
1266 (vreinterpret_f64_u16): Likewise.
1267 (vreinterpret_f64_u32): Likewise.
1268 (vreinterpret_f64_u64): Likewise.
1269 (vreinterpretq_f64_f32): Likewise.
1270 (vreinterpretq_f64_p8): Likewise.
1271 (vreinterpretq_f64_p16): Likewise.
1272 (vreinterpretq_f64_s8): Likewise.
1273 (vreinterpretq_f64_s16): Likewise.
1274 (vreinterpretq_f64_s32): Likewise.
1275 (vreinterpretq_f64_s64): Likewise.
1276 (vreinterpretq_f64_u8): Likewise.
1277 (vreinterpretq_f64_u16): Likewise.
1278 (vreinterpretq_f64_u32): Likewise.
1279 (vreinterpretq_f64_u64): Likewise.
1280 (vreinterpret_s64_f64): Likewise.
1281 (vreinterpretq_s64_f64): Likewise.
1282 (vreinterpret_u64_f64): Likewise.
1283 (vreinterpretq_u64_f64): Likewise.
1284 (vreinterpret_s8_f64): Likewise.
1285 (vreinterpretq_s8_f64): Likewise.
1286 (vreinterpret_s16_f64): Likewise.
1287 (vreinterpretq_s16_f64): Likewise.
1288 (vreinterpret_s32_f64): Likewise.
1289 (vreinterpretq_s32_f64): Likewise.
1290 (vreinterpret_u8_f64): Likewise.
1291 (vreinterpretq_u8_f64): Likewise.
1292 (vreinterpret_u16_f64): Likewise.
1293 (vreinterpretq_u16_f64): Likewise.
1294 (vreinterpret_u32_f64): Likewise.
1295 (vreinterpretq_u32_f64): Likewise.
1297 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1299 * config/aarch64/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
1300 * config/aarch64/aarch64/aarch64-simd-builtins.def (REINTERP): Removed.
1301 (vreinterpret_p8_s8): Likewise.
1302 * config/aarch64/aarch64/arm_neon.h (vreinterpret_p8_s8): Uses cast.
1303 (vreinterpret_p8_s16): Likewise.
1304 (vreinterpret_p8_s32): Likewise.
1305 (vreinterpret_p8_s64): Likewise.
1306 (vreinterpret_p8_f32): Likewise.
1307 (vreinterpret_p8_u8): Likewise.
1308 (vreinterpret_p8_u16): Likewise.
1309 (vreinterpret_p8_u32): Likewise.
1310 (vreinterpret_p8_u64): Likewise.
1311 (vreinterpret_p8_p16): Likewise.
1312 (vreinterpretq_p8_s8): Likewise.
1313 (vreinterpretq_p8_s16): Likewise.
1314 (vreinterpretq_p8_s32): Likewise.
1315 (vreinterpretq_p8_s64): Likewise.
1316 (vreinterpretq_p8_f32): Likewise.
1317 (vreinterpretq_p8_u8): Likewise.
1318 (vreinterpretq_p8_u16): Likewise.
1319 (vreinterpretq_p8_u32): Likewise.
1320 (vreinterpretq_p8_u64): Likewise.
1321 (vreinterpretq_p8_p16): Likewise.
1322 (vreinterpret_p16_s8): Likewise.
1323 (vreinterpret_p16_s16): Likewise.
1324 (vreinterpret_p16_s32): Likewise.
1325 (vreinterpret_p16_s64): Likewise.
1326 (vreinterpret_p16_f32): Likewise.
1327 (vreinterpret_p16_u8): Likewise.
1328 (vreinterpret_p16_u16): Likewise.
1329 (vreinterpret_p16_u32): Likewise.
1330 (vreinterpret_p16_u64): Likewise.
1331 (vreinterpret_p16_p8): Likewise.
1332 (vreinterpretq_p16_s8): Likewise.
1333 (vreinterpretq_p16_s16): Likewise.
1334 (vreinterpretq_p16_s32): Likewise.
1335 (vreinterpretq_p16_s64): Likewise.
1336 (vreinterpretq_p16_f32): Likewise.
1337 (vreinterpretq_p16_u8): Likewise.
1338 (vreinterpretq_p16_u16): Likewise.
1339 (vreinterpretq_p16_u32): Likewise.
1340 (vreinterpretq_p16_u64): Likewise.
1341 (vreinterpretq_p16_p8): Likewise.
1342 (vreinterpret_f32_s8): Likewise.
1343 (vreinterpret_f32_s16): Likewise.
1344 (vreinterpret_f32_s32): Likewise.
1345 (vreinterpret_f32_s64): Likewise.
1346 (vreinterpret_f32_u8): Likewise.
1347 (vreinterpret_f32_u16): Likewise.
1348 (vreinterpret_f32_u32): Likewise.
1349 (vreinterpret_f32_u64): Likewise.
1350 (vreinterpret_f32_p8): Likewise.
1351 (vreinterpret_f32_p16): Likewise.
1352 (vreinterpretq_f32_s8): Likewise.
1353 (vreinterpretq_f32_s16): Likewise.
1354 (vreinterpretq_f32_s32): Likewise.
1355 (vreinterpretq_f32_s64): Likewise.
1356 (vreinterpretq_f32_u8): Likewise.
1357 (vreinterpretq_f32_u16): Likewise.
1358 (vreinterpretq_f32_u32): Likewise.
1359 (vreinterpretq_f32_u64): Likewise.
1360 (vreinterpretq_f32_p8): Likewise.
1361 (vreinterpretq_f32_p16): Likewise.
1362 (vreinterpret_s64_s8): Likewise.
1363 (vreinterpret_s64_s16): Likewise.
1364 (vreinterpret_s64_s32): Likewise.
1365 (vreinterpret_s64_f32): Likewise.
1366 (vreinterpret_s64_u8): Likewise.
1367 (vreinterpret_s64_u16): Likewise.
1368 (vreinterpret_s64_u32): Likewise.
1369 (vreinterpret_s64_u64): Likewise.
1370 (vreinterpret_s64_p8): Likewise.
1371 (vreinterpret_s64_p16): Likewise.
1372 (vreinterpretq_s64_s8): Likewise.
1373 (vreinterpretq_s64_s16): Likewise.
1374 (vreinterpretq_s64_s32): Likewise.
1375 (vreinterpretq_s64_f32): Likewise.
1376 (vreinterpretq_s64_u8): Likewise.
1377 (vreinterpretq_s64_u16): Likewise.
1378 (vreinterpretq_s64_u32): Likewise.
1379 (vreinterpretq_s64_u64): Likewise.
1380 (vreinterpretq_s64_p8): Likewise.
1381 (vreinterpretq_s64_p16): Likewise.
1382 (vreinterpret_u64_s8): Likewise.
1383 (vreinterpret_u64_s16): Likewise.
1384 (vreinterpret_u64_s32): Likewise.
1385 (vreinterpret_u64_s64): Likewise.
1386 (vreinterpret_u64_f32): Likewise.
1387 (vreinterpret_u64_u8): Likewise.
1388 (vreinterpret_u64_u16): Likewise.
1389 (vreinterpret_u64_u32): Likewise.
1390 (vreinterpret_u64_p8): Likewise.
1391 (vreinterpret_u64_p16): Likewise.
1392 (vreinterpretq_u64_s8): Likewise.
1393 (vreinterpretq_u64_s16): Likewise.
1394 (vreinterpretq_u64_s32): Likewise.
1395 (vreinterpretq_u64_s64): Likewise.
1396 (vreinterpretq_u64_f32): Likewise.
1397 (vreinterpretq_u64_u8): Likewise.
1398 (vreinterpretq_u64_u16): Likewise.
1399 (vreinterpretq_u64_u32): Likewise.
1400 (vreinterpretq_u64_p8): Likewise.
1401 (vreinterpretq_u64_p16): Likewise.
1402 (vreinterpret_s8_s16): Likewise.
1403 (vreinterpret_s8_s32): Likewise.
1404 (vreinterpret_s8_s64): Likewise.
1405 (vreinterpret_s8_f32): Likewise.
1406 (vreinterpret_s8_u8): Likewise.
1407 (vreinterpret_s8_u16): Likewise.
1408 (vreinterpret_s8_u32): Likewise.
1409 (vreinterpret_s8_u64): Likewise.
1410 (vreinterpret_s8_p8): Likewise.
1411 (vreinterpret_s8_p16): Likewise.
1412 (vreinterpretq_s8_s16): Likewise.
1413 (vreinterpretq_s8_s32): Likewise.
1414 (vreinterpretq_s8_s64): Likewise.
1415 (vreinterpretq_s8_f32): Likewise.
1416 (vreinterpretq_s8_u8): Likewise.
1417 (vreinterpretq_s8_u16): Likewise.
1418 (vreinterpretq_s8_u32): Likewise.
1419 (vreinterpretq_s8_u64): Likewise.
1420 (vreinterpretq_s8_p8): Likewise.
1421 (vreinterpretq_s8_p16): Likewise.
1422 (vreinterpret_s16_s8): Likewise.
1423 (vreinterpret_s16_s32): Likewise.
1424 (vreinterpret_s16_s64): Likewise.
1425 (vreinterpret_s16_f32): Likewise.
1426 (vreinterpret_s16_u8): Likewise.
1427 (vreinterpret_s16_u16): Likewise.
1428 (vreinterpret_s16_u32): Likewise.
1429 (vreinterpret_s16_u64): Likewise.
1430 (vreinterpret_s16_p8): Likewise.
1431 (vreinterpret_s16_p16): Likewise.
1432 (vreinterpretq_s16_s8): Likewise.
1433 (vreinterpretq_s16_s32): Likewise.
1434 (vreinterpretq_s16_s64): Likewise.
1435 (vreinterpretq_s16_f32): Likewise.
1436 (vreinterpretq_s16_u8): Likewise.
1437 (vreinterpretq_s16_u16): Likewise.
1438 (vreinterpretq_s16_u32): Likewise.
1439 (vreinterpretq_s16_u64): Likewise.
1440 (vreinterpretq_s16_p8): Likewise.
1441 (vreinterpretq_s16_p16): Likewise.
1442 (vreinterpret_s32_s8): Likewise.
1443 (vreinterpret_s32_s16): Likewise.
1444 (vreinterpret_s32_s64): Likewise.
1445 (vreinterpret_s32_f32): Likewise.
1446 (vreinterpret_s32_u8): Likewise.
1447 (vreinterpret_s32_u16): Likewise.
1448 (vreinterpret_s32_u32): Likewise.
1449 (vreinterpret_s32_u64): Likewise.
1450 (vreinterpret_s32_p8): Likewise.
1451 (vreinterpret_s32_p16): Likewise.
1452 (vreinterpretq_s32_s8): Likewise.
1453 (vreinterpretq_s32_s16): Likewise.
1454 (vreinterpretq_s32_s64): Likewise.
1455 (vreinterpretq_s32_f32): Likewise.
1456 (vreinterpretq_s32_u8): Likewise.
1457 (vreinterpretq_s32_u16): Likewise.
1458 (vreinterpretq_s32_u32): Likewise.
1459 (vreinterpretq_s32_u64): Likewise.
1460 (vreinterpretq_s32_p8): Likewise.
1461 (vreinterpretq_s32_p16): Likewise.
1462 (vreinterpret_u8_s8): Likewise.
1463 (vreinterpret_u8_s16): Likewise.
1464 (vreinterpret_u8_s32): Likewise.
1465 (vreinterpret_u8_s64): Likewise.
1466 (vreinterpret_u8_f32): Likewise.
1467 (vreinterpret_u8_u16): Likewise.
1468 (vreinterpret_u8_u32): Likewise.
1469 (vreinterpret_u8_u64): Likewise.
1470 (vreinterpret_u8_p8): Likewise.
1471 (vreinterpret_u8_p16): Likewise.
1472 (vreinterpretq_u8_s8): Likewise.
1473 (vreinterpretq_u8_s16): Likewise.
1474 (vreinterpretq_u8_s32): Likewise.
1475 (vreinterpretq_u8_s64): Likewise.
1476 (vreinterpretq_u8_f32): Likewise.
1477 (vreinterpretq_u8_u16): Likewise.
1478 (vreinterpretq_u8_u32): Likewise.
1479 (vreinterpretq_u8_u64): Likewise.
1480 (vreinterpretq_u8_p8): Likewise.
1481 (vreinterpretq_u8_p16): Likewise.
1482 (vreinterpret_u16_s8): Likewise.
1483 (vreinterpret_u16_s16): Likewise.
1484 (vreinterpret_u16_s32): Likewise.
1485 (vreinterpret_u16_s64): Likewise.
1486 (vreinterpret_u16_f32): Likewise.
1487 (vreinterpret_u16_u8): Likewise.
1488 (vreinterpret_u16_u32): Likewise.
1489 (vreinterpret_u16_u64): Likewise.
1490 (vreinterpret_u16_p8): Likewise.
1491 (vreinterpret_u16_p16): Likewise.
1492 (vreinterpretq_u16_s8): Likewise.
1493 (vreinterpretq_u16_s16): Likewise.
1494 (vreinterpretq_u16_s32): Likewise.
1495 (vreinterpretq_u16_s64): Likewise.
1496 (vreinterpretq_u16_f32): Likewise.
1497 (vreinterpretq_u16_u8): Likewise.
1498 (vreinterpretq_u16_u32): Likewise.
1499 (vreinterpretq_u16_u64): Likewise.
1500 (vreinterpretq_u16_p8): Likewise.
1501 (vreinterpretq_u16_p16): Likewise.
1502 (vreinterpret_u32_s8): Likewise.
1503 (vreinterpret_u32_s16): Likewise.
1504 (vreinterpret_u32_s32): Likewise.
1505 (vreinterpret_u32_s64): Likewise.
1506 (vreinterpret_u32_f32): Likewise.
1507 (vreinterpret_u32_u8): Likewise.
1508 (vreinterpret_u32_u16): Likewise.
1509 (vreinterpret_u32_u64): Likewise.
1510 (vreinterpret_u32_p8): Likewise.
1511 (vreinterpret_u32_p16): Likewise.
1512 (vreinterpretq_u32_s8): Likewise.
1513 (vreinterpretq_u32_s16): Likewise.
1514 (vreinterpretq_u32_s32): Likewise.
1515 (vreinterpretq_u32_s64): Likewise.
1516 (vreinterpretq_u32_f32): Likewise.
1517 (vreinterpretq_u32_u8): Likewise.
1518 (vreinterpretq_u32_u16): Likewise.
1519 (vreinterpretq_u32_u64): Likewise.
1520 (vreinterpretq_u32_p8): Likewise.
1521 (vreinterpretq_u32_p16): Likewise.
1523 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1525 * gcc/config/aarch64/aarch64-simd.md (aarch64_s<optab><mode>):
1527 * config/aarch64/aarch64-simd-builtins.def (sqneg): Iterator
1530 * config/aarch64/arm_neon.h (vqneg_s64): New intrinsic.
1531 (vqnegd_s64): Likewise.
1532 (vqabs_s64): Likewise.
1533 (vqabsd_s64): Likewise.
1535 2014-04-22 Richard Henderson <rth@redhat.com>
1537 * config/sparc/sparc.c (sparc_init_modes): Hoist GET_MODE_SIZE
1538 computation to the top of the loop.
1540 2014-04-22 Renlin <renlin.li@arm.com>
1541 Jiong Wang <jiong.wang@arm.com>
1543 * config/aarch64/aarch64.h (aarch64_frame): Delete "fp_lr_offset".
1544 * config/aarch64/aarch64.c (aarch64_layout_frame)
1545 (aarch64_initial_elimination_offset): Likewise.
1547 2014-04-22 Marcus Shawcroft <marcus.shawcroft@arm.com>
1549 * config/aarch64/aarch64.c (aarch64_initial_elimination_offset):
1552 2014-04-22 Richard Sandiford <rdsandiford@googlemail.com>
1554 * machmode.h (bitwise_mode_for_mode): Declare.
1555 * stor-layout.h (bitwise_type_for_mode): Likewise.
1556 * stor-layout.c (bitwise_mode_for_mode): New function.
1557 (bitwise_type_for_mode): Likewise.
1558 * builtins.c (fold_builtin_memory_op): Use it instead of
1559 int_mode_for_mode and build_nonstandard_integer_type.
1561 2014-04-22 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1563 * config.gcc (enable_obsolete): Remove *-*-solaris2.9*.
1564 (*-*-solaris2.[0-9] | *-*-solaris2.[0-9].*): Mark unsupported.
1565 (*-*-solaris2*): Simplify.
1566 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Likewise.
1567 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Remove
1568 *-*-solaris2.9* handling.
1570 * configure.ac (gcc_cv_as_hidden): Remove test for Solaris 9/x86
1572 (gcc_cv_ld_hidden): Remove *-*-solaris2.9* handling.
1573 (ld_tls_support): Remove i?86-*-solaris2.9, sparc*-*-solaris2.9
1575 (gcc_cv_as_gstabs_flag): Remove workaround for Solaris 9/x86 as bug.
1576 * configure: Regenerate.
1578 * config/i386/sol2-9.h: Remove.
1580 * doc/install.texi (Specific, i?86-*-solaris2.9): Remove.
1581 (Specific, *-*-solaris2*): Mention Solaris 9 support removal.
1582 Remove Solaris 9 references.
1584 2014-04-22 Vidya Praveen <vidyapraveen@arm.com>
1586 * aarch64.md (float<GPI:mode><GPF:mode>2): Remove.
1587 (floatuns<GPI:mode><GPF:mode>2): Remove.
1588 (<optab><fcvt_target><GPF:mode>2): New pattern for equal width float
1589 and floatuns conversions.
1590 (<optab><fcvt_iesize><GPF:mode>2): New pattern for inequal width float
1591 and floatuns conversions.
1592 * iterators.md (fcvt_target, FCVT_TARGET): Support SF and DF modes.
1593 (w1,w2): New mode attributes for inequal width conversions.
1595 2014-04-22 Renlin Li <Renlin.Li@arm.com>
1597 * config/aarch64/aarch64.c (aarch64_print_operand_address): Adjust
1598 the output asm format.
1600 2014-04-22 James Greenhalgh <james.greenhalgh@arm.com>
1602 * config/aarch64/aarch64-simd.md
1603 (aarch64_cm<optab>di): Always split.
1604 (*aarch64_cm<optab>di): New.
1605 (aarch64_cmtstdi): Always split.
1606 (*aarch64_cmtstdi): New.
1608 2014-04-22 Jakub Jelinek <jakub@redhat.com>
1610 PR tree-optimization/60823
1611 * omp-low.c (ipa_simd_modify_function_body): Go through
1612 all SSA_NAMEs and for those refering to vector arguments
1613 which are going to be replaced adjust SSA_NAME_VAR and,
1614 if it is a default definition, change it into a non-default
1615 definition assigned at the beginning of function from new_decl.
1616 (ipa_simd_modify_stmt_ops): Rewritten.
1617 * tree-dfa.c (set_ssa_default_def): When removing default def,
1618 check for NULL loc instead of NULL *loc.
1620 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1622 * config/arm/arm.c (arm_hard_regno_mode_ok): Loosen
1623 restrictions on core registers for DImode values in Thumb2.
1625 2014-04-22 Ian Bolton <ian.bolton@arm.com>
1627 * config/arm/arm.md (*anddi_notdi_zesidi): New pattern.
1628 * config/arm/thumb2.md (*iordi_notdi_zesidi): New pattern.
1630 2014-04-22 Ian Bolton <ian.bolton@arm.com>
1632 * config/arm/thumb2.md (*iordi_notdi_di): New pattern.
1633 (*iordi_notzesidi_di): Likewise.
1634 (*iordi_notsesidi_di): Likewise.
1636 2014-04-22 Ian Bolton <ian.bolton@arm.com>
1638 * config/arm/arm-protos.h (tune_params): New struct members.
1639 * config/arm/arm.c: Initialise tune_params per processor.
1640 (thumb2_reorg): Suppress conversion from t32 to t16 when optimizing
1641 for speed, based on new tune_params.
1643 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1645 * config/aarch64/aarch64-builtins.c (BUILTIN_VDQF_DF): Macro
1647 * config/aarch64/aarch64-simd-builtins.def (frintn): Use added
1649 * config/aarch64/aarch64-simd.md (<frint_pattern>): Comment
1651 * config/aarch64/aarch64.md (<frint_pattern>): Likewise.
1652 * config/aarch64/arm_neon.h (vrnd_f64): Added.
1653 (vrnda_f64): Likewise.
1654 (vrndi_f64): Likewise.
1655 (vrndm_f64): Likewise.
1656 (vrndn_f64): Likewise.
1657 (vrndp_f64): Likewise.
1658 (vrndx_f64): Likewise.
1660 2014-04-22 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1662 * config/arm/arm.c (arm_print_operand, thumb_exit): Make sure
1663 GET_MODE_SIZE argument is enum machine_mode.
1665 2014-04-22 Jakub Jelinek <jakub@redhat.com>
1668 * config/sparc/sparc.c (sparc_init_modes): Pass enum machine_mode
1669 value instead of int to GET_MODE_CLASS and GET_MODE_SIZE macros.
1671 2014-04-22 Lin Zuojian <manjian2006@gmail.com>
1674 * asan.c (asan_emit_stack_protection): Force the base to align to
1675 appropriate bits if STRICT_ALIGNMENT. Set shadow_mem align to
1676 appropriate bits if STRICT_ALIGNMENT.
1677 * cfgexpand.c (expand_stack_vars): Set base_align appropriately
1679 (expand_used_vars): Leave a space in the stack frame for alignment
1680 if STRICT_ALIGNMENT.
1682 2014-04-21 David Malcolm <dmalcolm@redhat.com>
1684 * gimple.h (gimple_assign_single_p): Accept a const_gimple rather
1686 (gimple_store_p): Likewise.
1687 (gimple_assign_load_p): Likewise.
1688 (gimple_assign_cast_p): Likewise.
1689 (gimple_clobber_p): Likewise.
1691 * doc/gimple.texi (gimple_assign_cast_p): Accept a const_gimple
1692 rather than a gimple.
1693 (gimple_assign_cast_p): Likewise.
1695 2014-04-21 Michael Meissner <meissner@linux.vnet.ibm.com>
1698 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64 case):
1699 If mode is DDmode and TARGET_E500_DOUBLE allow move.
1701 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print some
1702 more debug information for E500 if -mdebug=reg.
1704 2014-04-21 Uros Bizjak <ubizjak@gmail.com>
1707 * config/i386/i386.c (ix86_expand_builtin)
1708 <case IX86_BUILTIN_RDRAND{16,32,64}_STEP>: Use temporary
1709 register for target RTX.
1710 <case IX86_BUILTIN_RDSEED{16,32,64}_STEP>: Ditto.
1712 2014-04-18 Cong Hou <congh@google.com>
1714 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): Enhance
1715 the widen-mult pattern by handling two operands with different sizes,
1716 and operands whose size is smaller than half of the result type.
1718 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
1720 * ipa-inline.h (INLINE_HINT_known_hot): New hint.
1721 * ipa-inline-analysis.c (dump_inline_hints): Dump it.
1722 (do_estimate_edge_time): Compute it.
1723 * ipa-inline.c (want_inline_small_function_p): Bypass
1724 INLINE_INSNS_AUTO/SINGLE limits for calls that are known to be hot.
1726 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
1728 * ipa-inline.c (spec_rem): New static variable.
1729 (dump_overall_stats): New function.
1730 (dump_inline_stats): New function.
1732 2014-04-18 Richard Henderson <rth@redhat.com>
1734 * config/aarch64/aarch64.c (aarch64_register_move_cost): Pass a mode
1735 to GET_MODE_SIZE, not a reg_class_t.
1737 2014-04-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1739 * config/rs6000/vsx.md (vsx_xxmrghw_<mode>): Adjust for little-endian.
1740 (vsx_xxmrglw_<mode>): Likewise.
1742 2014-04-17 Michael Meissner <meissner@linux.vnet.ibm.com>
1745 * config/rs6000/rs6000.c (rs6000_setup_reg_addr_masks): Make sure
1746 GET_MODE_SIZE gets passed an enum machine_mode type and not integer.
1747 (rs6000_init_hard_regno_mode_ok): Likewise.
1749 2014-04-17 Jan Hubicka <hubicka@ucw.cz>
1751 * ipa-inline.c (inline_small_functions): Account only non-cold
1753 * doc/invoke.texi (inline-unit-growth): Update documentation.
1755 2014-04-17 Pat Haugen <pthaugen@us.ibm.com>
1757 * config/rs6000/rs6000.md (addti3, subti3): New.
1759 2014-04-17 H.J. Lu <hongjiu.lu@intel.com>
1762 * config/i386/i386.c (ix86_expand_clear): Remove outdated
1763 comment. Check optimize_insn_for_size_p instead of
1764 optimize_insn_for_speed_p.
1766 2014-04-17 Martin Jambor <mjambor@suse.cz>
1768 * gimple-iterator.c (gsi_start_edge): New function.
1769 * gimple-iterator.h (gsi_start_edge): Declare.
1770 * tree-sra.c (single_non_eh_succ): New function.
1771 (disqualify_ops_if_throwing_stmt): Renamed to
1772 disqualify_if_bad_bb_terminating_stmt. Allow throwing statements
1773 having one non-EH successor BB.
1774 (sra_modify_expr): If stmt ends bb, use single non-EH successor to
1775 generate loads into replacements.
1776 (sra_modify_assign): Likewise and and also use the simple path for
1778 (sra_modify_function_body): Commit statements on edges.
1780 2014-04-17 Richard Biener <rguenther@suse.de>
1783 * tree-ssa-propagate.c (valid_gimple_rhs_p): Allow vector
1784 comparison results and add clarifying comment.
1786 2014-04-17 Jakub Jelinek <jakub@redhat.com>
1788 * genmodes.c (struct mode_data): Add need_bytesize_adj field.
1789 (blank_mode): Initialize it.
1790 (emit_mode_size_inline, emit_mode_nunits_inline,
1791 emit_mode_inner_inline): New functions.
1792 (emit_insn_modes_h): Call them and surround their output with
1793 #if GCC_VERSION >= 4001 ... #endif.
1794 * machmode.h (GET_MODE_SIZE, GET_MODE_NUNITS, GET_MODE_INNER):
1795 For GCC_VERSION >= 4001 use mode_*_inline routines instead of
1796 mode_* arrays if the argument is __builtin_constant_p.
1797 * lower-subreg.c (dump_choices): Make sure GET_MODE_SIZE argument
1798 is enum machine_mode.
1800 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1802 * passes.c (opt_pass::execute): Adjust.
1803 (pass_manager::execute_pass_mode_switching): Likewise.
1804 (early_local_passes::execute): Likewise.
1805 (execute_one_pass): Pass cfun to the pass's execute method.
1806 * tree-pass.h (opt_pass::execute): Add function * argument.
1807 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
1808 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
1809 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
1810 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
1811 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
1812 config/sparc/sparc.c, cprop.c, dce.c, df-core.c, dse.c, dwarf2cfi.c,
1813 except.c, final.c, function.c, fwprop.c, gcse.c, gimple-low.c,
1814 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
1815 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
1816 ipa-inline-analysis.c, ipa-inline.c, ipa-profile.c, ipa-pure-const.c,
1817 ipa-reference.c, ipa-split.c, ipa.c, ira.c, jump.c, loop-init.c,
1818 lower-subreg.c, mode-switching.c, omp-low.c, postreload-gcse.c,
1819 postreload.c, predict.c, recog.c, ree.c, reg-stack.c, regcprop.c,
1820 reginfo.c, regrename.c, reorg.c, sched-rgn.c, stack-ptr-mod.c,
1821 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
1822 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
1823 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
1824 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-ssa-ccp.c,
1825 tree-ssa-copy.c, tree-ssa-copyrename.c, tree-ssa-dce.c,
1826 tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
1827 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
1828 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
1829 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
1830 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
1831 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
1832 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
1833 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
1834 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
1835 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
1838 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1840 * passes.c (opt_pass::gate): Take function * argument.
1841 (gate_all_early_local_passes): Merge into
1842 (early_local_passes::gate): this.
1843 (gate_all_early_optimizations): Merge into
1844 (all_early_optimizations::gate): this.
1845 (gate_all_optimizations): Mege into
1846 (all_optimizations::gate): this.
1847 (gate_all_optimizations_g): Merge into
1848 (all_optimizations_g::gate): this.
1849 (gate_rest_of_compilation): Mege into
1850 (rest_of_compilation::gate): this.
1851 (gate_postreload): Merge into
1852 (postreload::gate): this.
1853 (dump_one_pass): Pass cfun to the pass's gate method.
1854 (execute_ipa_summary_passes): Likewise.
1855 (execute_one_pass): Likewise.
1856 (ipa_write_summaries_2): Likewise.
1857 (ipa_write_optimization_summaries_1): Likewise.
1858 (ipa_read_summaries_1): Likewise.
1859 (ipa_read_optimization_summaries_1): Likewise.
1860 (execute_ipa_stmt_fixups): Likewise.
1861 * tree-pass.h (opt_pass::gate): Add function * argument.
1862 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c,
1863 combine-stack-adj.c, combine.c, compare-elim.c,
1864 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
1865 config/rl78/rl78.c, config/sh/sh_optimize_sett_clrt.cc,
1866 config/sh/sh_treg_combine.cc, config/sparc/sparc.c, cprop.c, cse.c,
1867 dce.c, df-core.c, dse.c, dwarf2cfi.c, except.c, fwprop.c, gcse.c,
1868 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
1869 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
1870 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
1871 loop-init.c, lower-subreg.c, mode-switching.c, modulo-sched.c,
1872 omp-low.c, postreload-gcse.c, postreload.c, predict.c, recog.c, ree.c,
1873 reg-stack.c, regcprop.c, regrename.c, reorg.c, sched-rgn.c,
1874 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
1875 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
1876 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c,
1877 tree-nrv.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
1878 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
1879 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
1880 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
1881 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
1882 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
1883 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
1884 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
1885 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
1886 tree-ssa.c, tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
1887 tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c, tsan.c, ubsan.c,
1888 var-tracking.c, vtable-verify.c, web.c: Adjust.
1890 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1892 * configure.ac: Check for -Woverloaded-virtual and enable it if found.
1893 * configure: Regenerate.
1895 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1897 * passes.c (dump_one_pass): don't check pass->has_gate.
1898 (execute_ipa_summary_passes): Likewise.
1899 (execute_one_pass): Likewise.
1900 (ipa_write_summaries_2): Likewise.
1901 (ipa_write_optimization_summaries_1): Likewise.
1902 (ipa_read_optimization_summaries_1): Likewise.
1903 (execute_ipa_stmt_fixups): Likewise.
1904 * tree-pass.h (pass_data::has_gate): Remove.
1905 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
1906 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
1907 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
1908 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
1909 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
1910 config/sh/sh_optimize_sett_clrt.cc, config/sh/sh_treg_combine.cc,
1911 config/sparc/sparc.c, cprop.c, cse.c, dce.c, df-core.c, dse.c,
1912 dwarf2cfi.c, except.c, final.c, function.c, fwprop.c, gcse.c,
1913 gimple-low.c, gimple-ssa-isolate-paths.c,
1914 gimple-ssa-strength-reduction.c, graphite.c, ifcvt.c, init-regs.c,
1915 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
1916 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
1917 ira.c, jump.c, loop-init.c, lower-subreg.c, mode-switching.c,
1918 modulo-sched.c, omp-low.c, postreload-gcse.c, postreload.c, predict.c,
1919 recog.c, ree.c, reg-stack.c, regcprop.c, reginfo.c, regrename.c,
1920 reorg.c, sched-rgn.c, stack-ptr-mod.c, store-motion.c, tracer.c,
1921 trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
1922 tree-complex.c, tree-eh.c, tree-emutls.c, tree-if-conv.c,
1923 tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
1924 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
1925 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
1926 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
1927 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
1928 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
1929 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
1930 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
1931 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
1932 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
1933 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
1934 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
1935 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
1938 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1940 * pass_manager.h (pass_manager::register_dump_files_1): Remove
1942 * passes.c (pass_manager::register_dump_files_1): Merge into
1943 (pass_manager::register_dump_files): this, and remove its handling of
1944 properties since the pass always has the properties anyway.
1945 (pass_manager::pass_manager): Adjust.
1947 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1949 * pass_manager.h (pass_manager::register_dump_files_1): Adjust.
1950 * passes.c (pass_manager::register_dump_files_1): Remove dead code
1951 dealing with properties.
1952 (pass_manager::register_dump_files): Adjust.
1954 2014-03-20 Mark Wielaard <mjw@redhat.com>
1956 * dwarf2out.c (add_bound_info): If HOST_WIDE_INT is big enough,
1957 then represent the bound as normal constant value.
1959 2014-04-17 Jakub Jelinek <jakub@redhat.com>
1962 Forward port from 4.8 branch
1963 2013-07-19 Kirill Yukhin <kirill.yukhin@intel.com>
1965 * config/i386/bmiintrin.h (_blsi_u32): New.
1969 (_blsmsk_u32): Ditto.
1970 (_blsmsk_u64): Ditto.
1971 (_tzcnt_u32): Ditto.
1972 (_tzcnt_u64): Ditto.
1974 2014-04-17 Kito Cheng <kito@0xlab.org>
1976 * gcc.c (used_arg): Prevent out of bound access for multilib_options.
1978 2014-04-17 Richard Biener <rguenther@suse.de>
1981 * tree-ssa-propagate.c (valid_gimple_rhs_p): Only allow effective
1982 boolean results for comparisons.
1984 2014-04-17 Richard Biener <rguenther@suse.de>
1986 PR tree-optimization/60836
1987 * tree-vect-loop.c (vect_create_epilog_for_reduction): Force
1988 initial PHI args to be gimple values.
1990 2014-04-17 Richard Biener <rguenther@suse.de>
1992 PR tree-optimization/60841
1993 * tree-vect-data-refs.c (vect_analyze_data_refs): Count stmts.
1994 * tree-vect-loop.c (vect_analyze_loop_2): Pass down number
1995 of stmts to SLP build.
1996 * tree-vect-slp.c (vect_slp_analyze_bb_1): Likewise.
1997 (vect_analyze_slp): Likewise.
1998 (vect_analyze_slp_instance): Likewise.
1999 (vect_build_slp_tree): Limit overall SLP tree growth.
2000 * tree-vectorizer.h (vect_analyze_data_refs,
2001 vect_analyze_slp): Adjust prototypes.
2003 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
2005 * config/i386/i386.c (x86_add_stmt_cost): Fix vector cost model for
2008 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
2010 * config/i386/x86-tune.def (TARGET_SLOW_PSHUFB): New tune definition.
2011 * config/i386/i386.h (TARGET_SLOW_PSHUFB): New tune flag.
2012 * config/i386/i386.c (expand_vec_perm_even_odd_1): Avoid byte shuffles
2013 for TARGET_SLOW_PSHUFB
2015 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
2017 * config/i386/i386.c (slm_cost): Adjust vec_to_scalar_cost.
2018 * config/i386/i386.c (intel_cost): Ditto.
2020 2014-04-17 Joey Ye <joey.ye@arm.com>
2022 * opts.c (OPT_fif_conversion, OPT_fif_conversion2): Disable for Og.
2024 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
2026 * opts.c (common_handle_option): Disable -fipa-reference coorectly
2029 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
2031 * ipa-devirt.c (odr_type_d): Add field all_derivations_known.
2032 (type_all_derivations_known_p): New predicate.
2033 (type_all_ctors_visible_p): New predicate.
2034 (type_possibly_instantiated_p): New predicate.
2035 (get_odr_type): Compute all_derivations_known.
2036 (dump_odr_type): Dump the flag.
2037 (maybe_record_type): Cleanup.
2038 (record_target_from_binfo): Add bases_to_consider array;
2039 record bases for types w/o instances and skip CXX destructor.
2040 (possible_polymorphic_call_targets_1): Add bases_to_consider
2041 and consider_construction parameters; check if type may have instance.
2042 (get_polymorphic_call_info): Set maybe_in_construction to true
2043 when we know nothing.
2044 (record_targets_from_bases): Skip CXX destructors; they are
2045 never called for types in construction.
2046 (possible_polymorphic_call_targets): Do not record target when
2047 type may not have instance.
2049 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
2052 * ipa.c (symtab_remove_unreachable_nodes): Mark targets of
2053 external aliases alive, too.
2055 2014-04-16 Andrew Pinski <apinski@cavium.com>
2057 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Change aarch64 ilp32
2060 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
2062 * final.c (compute_alignments): Do not apply loop alignment to a block
2063 falling through to the exit.
2065 2014-04-16 Catherine Moore <clm@codesourcery.com>
2067 * mips.md (*mov<mode>_internal, *movhi_internal, *movqi_internal):
2068 Adjust constraints for microMIPS store patterns.
2070 2014-04-16 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
2072 * config/avr/avr-mcus.def: Correct typo for atxmega256a3bu macro.
2074 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
2076 * tree-ssa-operands.c (create_vop_var): Set DECL_IGNORED_P.
2077 (append_use): Run at -O0.
2078 (append_vdef): Likewise.
2079 * tree-ssa-ter.c (ter_is_replaceable_p): Do not special-case -O0.
2080 * tree-ssa-uninit.c (warn_uninitialized_vars): Remove obsolete comment.
2082 2014-04-16 Jakub Jelinek <jakub@redhat.com>
2084 PR tree-optimization/60844
2085 * tree-ssa-reassoc.c (reassoc_remove_stmt): New function.
2086 (propagate_op_to_single_use, remove_visited_stmt_chain,
2087 linearize_expr, repropagate_negates, reassociate_bb): Use it
2088 instead of gsi_remove.
2090 2014-04-16 Martin Jambor <mjambor@suse.cz>
2092 * cgraphclones.c (cgraph_create_virtual_clone): Duplicate
2093 ipa_transforms_to_apply.
2094 (cgraph_function_versioning): Assert that old_node has empty
2095 ipa_transforms_to_apply.
2096 * trans-mem.c (ipa_tm_create_version): Likewise.
2097 * tree-inline.c (tree_function_versioning): Do not duplicate
2098 ipa_transforms_to_apply.
2100 2014-04-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2103 * configure.ac (set_have_as_tls): Merge i[34567]86-*-* and
2105 Pass necessary as flags on 64-bit Solaris/x86.
2106 Use lowercase relocs for x86_64-*-*.
2107 * configure: Regenerate.
2109 2014-04-15 Jan Hubicka <jh@suse.cz>
2111 * ipa-devirt.c (referenced_from_vtable_p): New predicate.
2112 (maybe_record_node, likely_target_p): Use it.
2114 2014-04-15 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2117 Revert following patch
2119 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
2122 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
2123 software floating point or no floating point registers, do not
2124 allow any type in the FPRs. Eliminate a test for SPE SIMD types
2125 in GPRs that occurs after we tested for GPRs that would never be
2128 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
2129 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
2130 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
2131 specifically allow DDmode, since that does not use the SPE SIMD
2134 2014-03-21 Mark Wielaard <mjw@redhat.com>
2136 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_const_value
2137 as unsigned or int depending on type and value used.
2139 2014-04-15 Richard Biener <rguenther@suse.de>
2141 PR rtl-optimization/56965
2142 * alias.c (ncr_compar, nonoverlapping_component_refs_p): Move ...
2143 * tree-ssa-alias.c (ncr_compar, nonoverlapping_component_refs_p):
2145 * alias.c (true_dependence_1): Do not call
2146 nonoverlapping_component_refs_p.
2147 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Call
2148 nonoverlapping_component_refs_p.
2149 (indirect_refs_may_alias_p): Likewise.
2151 2014-04-15 Teresa Johnson <tejohnson@google.com>
2153 * cfg.c (dump_bb_info): Fix flags check.
2154 * tree-cfg.c (remove_bb): Only dump TDF_BLOCKS when removing.
2156 2014-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2158 PR rtl-optimization/60663
2159 * config/arm/arm.c (arm_new_rtx_costs): Improve ASM_OPERANDS case,
2162 2014-04-15 Richard Biener <rguenther@suse.de>
2164 * lto-streamer.h (LTO_major_version): Bump to 4.
2166 2014-04-15 Richard Biener <rguenther@suse.de>
2168 * common.opt (lto_partition_model): New enum.
2169 (flto-partition=): Merge separate options with a single with argument,
2170 add -flto-partition=one support.
2171 * flag-types.h (enum lto_partition_model): Declare.
2172 * opts.c (finish_options): Remove duplicate -flto-partition=
2174 * lto-wrapper.c (run_gcc): Adjust.
2176 2014-04-15 Richard Biener <rguenther@suse.de>
2178 * alias.c (ncr_compar): New function.
2179 (nonoverlapping_component_refs_p): Re-implement in O (n log n).
2181 2014-04-15 Richard Biener <rguenther@suse.de>
2183 * alias.c (record_component_aliases): Do not walk BINFOs.
2185 2014-04-15 Richard Biener <rguenther@suse.de>
2187 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
2188 Add struct function argument and adjust.
2189 (find_func_aliases_for_call): Likewise.
2190 (find_func_aliases): Likewise.
2191 (find_func_clobbers): Likewise.
2192 (intra_create_variable_infos): Likewise.
2193 (compute_points_to_sets): Likewise.
2194 (ipa_pta_execute): Adjust. Do not push/pop cfun.
2196 2014-04-15 Richard Biener <rguenther@suse.de>
2198 * tree.c (iterative_hash_expr): Use enum tree_code_class
2199 to store TREE_CODE_CLASS.
2200 (tree_block): Likewise.
2201 (tree_set_block): Likewise.
2202 * tree.h (fold_build_pointer_plus_loc): Use
2203 convert_to_ptrofftype_loc.
2205 2014-04-15 Jakub Jelinek <jakub@redhat.com>
2208 * Makefile.in (PLUGIN_HEADERS): Add various headers that have been
2211 2014-04-15 Eric Botcazou <ebotcazou@adacore.com>
2213 * cfgloop.h (struct loop): Move force_vectorize down.
2214 * gimplify.c (gimple_boolify) <ANNOTATE_EXPR>: Handle new kinds.
2215 (gimplify_expr) <ANNOTATE_EXPR>: Minor tweak.
2216 * lto-streamer-in.c (input_cfg): Read dont_vectorize field.
2217 * lto-streamer-out.c (output_cfg): Write dont_vectorize field.
2218 * tree-cfg.c (replace_loop_annotate): Revamp and handle new kinds.
2219 * tree-core.h (enum annot_expr_kind): Add new kind values.
2220 * tree-inline.c (copy_loops): Copy dont_vectorize field and reorder.
2221 * tree-pretty-print.c (dump_generic_node) <ANNOTATE_EXPR>: Handle new
2223 * tree.def (ANNOTATE_EXPR): Tweak comment.
2225 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
2227 * ipa-devirt.c (maybe_record_node): Ignore all non-methods (including
2230 2014-04-14 Paolo Carlini <paolo.carlini@oracle.com>
2232 * tree.h (TYPE_IDENTIFIER): Declare.
2233 * tree.c (subrange_type_for_debug_p): Use it.
2234 * godump.c (go_format_type): Likewise.
2235 * dwarf2out.c (is_cxx_auto, modified_type_die,
2236 gen_type_die_with_usage, gen_type_die_with_usage): Likewise.
2237 * dbxout.c (dbxout_type, dbxout_symbol): Likewise.
2239 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
2242 * varpool.c (varpool_remove_node): Do not alter decls when streaming.
2244 2014-04-14 Uros Bizjak <ubizjak@gmail.com>
2246 * config/i386/i386.c (examine_argument): Return bool. Return true if
2247 parameter should be passed in memory.
2248 <case X86_64_COMPLEX_X87_CLASS>: Adjust.
2249 (construct_container): Update calls to examine_argument.
2250 (function_arg_advance_64): Ditto.
2251 (return_in_memory_32): Merge with ix86_return_in_memory.
2252 (return_in_memory_64): Ditto.
2253 (return_in_memory_ms_64): Ditto.
2255 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
2257 * ipa-utils.c (ipa_merge_profiles): Merge profile_id.
2258 * coverage.c (coverage_compute_profile_id): Handle externally visible
2261 2014-04-14 Martin Jambor <mjambor@suse.cz>
2263 * tree-sra.c (ipa_sra_preliminary_function_checks): Skip
2264 DECL_DISREGARD_INLINE_LIMITS functions.
2266 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
2269 * config/i386/i386.md (*fixuns_trunc<mode>_1): Revert the last change.
2271 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
2274 * config/i386/i386.md (*fixuns_trunc<mode>_1): Check
2275 optimize_insn_for_speed_p instead of
2276 optimize_function_for_speed_p.
2278 2014-04-14 Yufeng Zhang <yufeng.zhang@arm.com>
2280 * doc/invoke.texi (free): Document AArch64.
2282 2014-04-14 Richard Biener <rguenther@suse.de>
2284 PR tree-optimization/60042
2285 * tree-ssa-pre.c (inhibit_phi_insertion): Remove.
2286 (insert_into_preds_of_block): Do not prevent PHI insertion
2287 for REFERENCE exprs here ...
2288 (eliminate_dom_walker::before_dom_children): ... but prevent
2289 their use here under similar conditions when applied to the
2290 IL after PRE optimizations.
2292 2014-04-14 Richard Biener <rguenther@suse.de>
2294 * passes.def: Move early points-to after early SRA.
2296 2014-04-14 Richard Biener <rguenther@suse.de>
2298 * tree-ssa-forwprop.c (simplify_gimple_switch): Enhance
2299 check for which sign-changes we allow when forwarding
2300 a converted value into a switch.
2302 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
2304 * stor-layout.c (place_field): Finalize non-constant offset for the
2307 2014-04-14 Richard Biener <rguenther@suse.de>
2309 * tree-switch-conversion.c (lshift_cheap_p): Get speed_p
2311 (expand_switch_using_bit_tests_p): Likewise.
2312 (process_switch): Compute and pass on speed_p based on the
2314 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Use
2315 optimize_bb_for_speed_p.
2317 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
2319 * cfgloop.h (struct loop): Rename force_vect into force_vectorize.
2320 * function.h (struct function): Rename has_force_vect_loops into
2321 has_force_vectorize_loops.
2322 * lto-streamer-in.c (input_cfg): Adjust for renaming.
2323 (input_struct_function_base): Likewise.
2324 * lto-streamer-out.c (output_cfg): Likewise.
2325 (output_struct_function_base): Likewise.
2326 * omp-low.c (expand_omp_simd): Likewise.
2327 * tree-cfg.c (move_sese_region_to_fn): Likewise.
2328 * tree-if-conv.c (ifcvt_can_use_mask_load_store): Likewise.
2329 (version_loop_for_if_conversion): Likewise.
2330 (tree_if_conversion): Likewise.
2331 (main_tree_if_conversion): Likewise.
2332 (gate_tree_if_conversion): Likewise.
2333 * tree-inline.c (copy_loops): Likewise.
2334 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Likewise.
2335 * tree-ssa-loop.c (tree_loop_vectorize): Likewise.
2336 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Likewise.
2337 * tree-vect-loop.c (vect_estimate_min_profitable_iters): Likewise.
2338 * tree-vectorizer.c (vectorize_loops): Likewise.
2339 * tree-vectorizer.h (unlimited_cost_model): Likewise.
2341 2014-04-14 Richard Biener <rguenther@suse.de>
2344 * lto-streamer-out.c (wrap_refs): New function.
2345 (lto_output): Wrap symbol references in global initializes in
2346 type-preserving MEM_REFs.
2348 2014-04-14 Christian Bruel <christian.bruel@st.com>
2350 * config/sh/sh-mem.cc (sh_expand_strlen): Unroll last word.
2352 2014-04-14 Christian Bruel <christian.bruel@st.com>
2354 * config/sh/sh.md (setmemqi): New expand pattern.
2355 * config/sh/sh.h (CLEAR_RATIO): Define.
2356 * config/sh/sh-mem.cc (sh_expand_setmem): Define.
2357 * config/sh/sh-protos.h (sh_expand_setmem): Declare.
2359 2014-04-14 Richard Biener <rguenther@suse.de>
2362 * fold-const.c (negate_expr_p): Don't negate directional rounding
2364 (fold_negate_expr): Likewise.
2366 2014-04-14 Richard Biener <rguenther@suse.de>
2368 PR tree-optimization/59817
2369 PR tree-optimization/60453
2370 * graphite-scop-detection.c (graphite_can_represent_scev): Complete
2371 recursion to catch all CHRECs in the scalar evolution and restrict
2372 the predicate for the remains appropriately.
2374 2014-04-12 Catherine Moore <clm@codesourcery.com>
2376 * config/mips/constraints.md: Add new register constraint "kb".
2377 * config/mips/mips.md (*mov<mode>_internal): Use constraint "kb".
2378 (*movhi_internal): Likewise.
2379 (*movqi_internal): Likewise.
2380 * config/mips/mips.h (M16_STORE_REGS): New register class.
2381 (REG_CLASS_NAMES): Add M16_STORE_REGS.
2382 (REG_CLASS_CONTENTS): Likewise.
2383 * config/mips/mips.c (mips_regno_to_class): Add M16_STORE_REGS.
2385 2014-04-11 Tobias Burnus <burnus@net-b.de>
2388 * doc/invoke.texi (-Wformat-signedness): Document it.
2389 (Wformat=2): Mention that this enables -Wformat-signedness.
2391 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
2393 * common/config/epiphany/epiphany-common.c
2394 (epiphany_option_optimization_table): Enable section anchors by
2395 default at -O1 or higher.
2396 * config/epiphany/epiphany.c (TARGET_MAX_ANCHOR_OFFSET): Define.
2397 (TARGET_MIN_ANCHOR_OFFSET): Likewise.
2398 (epiphany_rtx_costs) <SET>: For binary operators, the set as such
2399 carries no extra cost.
2400 (epiphany_legitimate_address_p): For BLKmode, apply SImode check.
2401 * config/epiphany/epiphany.h (ASM_OUTPUT_DEF): Define.
2402 * config/epiphany/predicates.md (memclob_operand): New predicate.
2403 * config/epiphany/epiphany.md (stack_adjust_add, stack_adjust_str):
2404 Use memclob_operand predicate and X constraint for operand 3.
2406 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
2408 * config/epiphany/epiphany.c (epiphany_rtx_cost): Compare
2409 with CC_N_NE / CC_C_LTU / CC_C_GTU carries no extra cost for
2412 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
2414 PR rtl-optimization/60651
2415 * mode-switching.c (optimize_mode_switching): Make sure to emit
2416 sets of a lower numbered entity before sets of a higher numbered
2417 entity to a mode of the same or lower priority.
2418 When creating a seginfo for a basic block that starts with a code
2419 label, move the insertion point past the code label.
2420 (new_seginfo): Document and enforce requirement that
2421 NOTE_INSN_BASIC_BLOCK only appears for empty blocks.
2422 * doc/tm.texi.in: Document ordering constraint for emitted mode sets.
2423 * doc/tm.texi: Regenerate.
2425 2014-01-11 Joern Rennecke <joern.rennecke@embecosm.com>
2428 * config/arc/arc.c (arc_save_restore): Fix assert typo.
2430 2013-04-11 Jakub Jelinek <jakub@redhat.com>
2432 * BASE-VER: Set to 4.10.0.
2434 2014-04-11 Tobias Burnus <burnus@net-b.de>
2437 * doc/bugreport.texi (Bugs): Remove nodes pointing to the nirvana.
2438 * doc/gcc.texi (Service): Update description in the @menu
2439 * doc/invoke.texi (Option Summary): Remove misplaced and
2442 2014-04-11 Steve Ellcey <sellcey@mips.com>
2443 Jakub Jelinek <jakub@redhat.com>
2446 * expr.c (convert_move): Use emit_store_flag_force instead of
2447 emit_store_flag. Pass lowpart_mode instead of VOIDmode as 5th
2450 2014-04-11 Richard Biener <rguenther@suse.de>
2453 * varasm.c (assemble_alias): Avoid endless error reporting
2454 recursion by setting TREE_ASM_WRITTEN.
2456 2014-04-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2458 * config/s390/s390.md: Add a splitter for NOT rtx.
2460 2014-04-11 Jakub Jelinek <jakub@redhat.com>
2462 PR rtl-optimization/60663
2463 * cse.c (cse_insn): Set src_volatile on ASM_OPERANDS in PARALLEL.
2465 2014-04-10 Jan Hubicka <hubicka@ucw.cz>
2466 Jakub Jelinek <jakub@redhat.com>
2469 * ipa.c (function_and_variable_visibility): Copy forced_by_abi
2470 flag from decl_node to node.
2472 2014-04-10 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2475 * config/arm/arm.c (TARGET_CONST_NOT_OK_FOR_DEBUG_P): Define
2476 (arm_const_not_ok_for_debug_p): Reject MINUS with SYM_REF's
2477 ameliorating the cases where it can be.
2479 2014-04-09 David Edelsohn <dje.gcc@gmail.com>
2482 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
2484 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
2485 (loadsync_<mode>): Change mode.
2486 (load_quadpti, store_quadpti): New.
2487 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
2488 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
2489 * config/rs6000/predicates.md (quad_memory_operand): !TARGET_SYNC_TI.
2491 2014-04-09 Cong Hou <congh@google.com>
2494 * doc/sourcebuild.texi (vect_widen_mult_si_to_di_pattern): Add
2497 2014-04-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2499 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Use vnand
2500 instead of vnor to exploit possible fusion opportunity in the
2502 (altivec_expand_vec_perm_const_le): Likewise.
2504 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
2506 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
2507 (loadsync_<mode>): Change mode.
2508 (load_quadpti, store_quadpti): New.
2509 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
2510 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
2512 2014-04-08 Richard Sandiford <rdsandiford@googlemail.com>
2515 * config/rs6000/vsx.md (vsx_xscvdpspn_scalar): Change input to DImode.
2516 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Update accordingly.
2517 Use gen_rtx_REG rather than simplify_gen_subreg for op0_di.
2519 2014-04-08 Richard Biener <rguenther@suse.de>
2522 * tree-pretty-print.c (pp_double_int): For HWI32 hosts with
2523 a 64bit widest int print double-int similar to on HWI64 hosts.
2525 2014-04-08 Richard Biener <rguenther@suse.de>
2527 PR tree-optimization/60785
2528 * graphite-sese-to-poly.c (rewrite_phi_out_of_ssa): Treat
2529 default defs properly.
2531 2014-04-08 Nathan Sidwell <nathan@codesourcery.com>
2533 * doc/invoke (Wnon-virtual-dtor): Update to match implementation.
2534 (Weffc++): Likewise.
2536 2014-04-07 Jan Hubicka <hubcika@ucw.cz>
2538 * ipa-devirt.c (maybe_record_node): When node is not recorded,
2539 set completep to false rather than true.
2541 2014-04-07 Douglas B Rupp <rupp@adacore.com>
2544 * config/arm/arm.h (ASM_PREFERRED_EH_DATA_FORMAT): Expose from
2545 ARM_TARGET2_DWARF_FORMAT.
2547 2014-04-07 Charles Baylis <charles.baylis@linaro.org>
2550 * config/arm/arm.h (ASM_OUTPUT_CASE_END): Remove.
2551 (LABEL_ALIGN_AFTER_BARRIER): Align barriers which occur after
2554 2014-04-07 Richard Biener <rguenther@suse.de>
2556 PR tree-optimization/60766
2557 * tree-ssa-loop-ivopts.c (cand_value_at): Compute in an unsigned type.
2558 (may_eliminate_iv): Convert cand_value_at result to desired type.
2560 2014-04-07 Jason Merrill <jason@redhat.com>
2563 * common.opt (-fno-gnu-unique): Add.
2564 * config/elfos.h (USE_GNU_UNIQUE_OBJECT): Check it.
2566 2014-04-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2568 * haifa-sched.c: Fix outdated function reference and minor
2569 grammar errors in introductory comment.
2571 2014-04-07 Richard Biener <rguenther@suse.de>
2574 * tree-ssa-operands.c (maybe_add_call_vops): Also add VDEFs
2576 * tree-cfgcleanup.c (fixup_noreturn_call): Do not remove VDEFs.
2578 2014-04-06 John David Anglin <danglin@gcc.gnu.org>
2581 * config/pa/pa.c (pa_output_function_epilogue): Skip address and code
2582 size accounting for thunks.
2583 (pa_asm_output_mi_thunk): Use final_start_function() and
2584 final_end_function() to output function start and end directives.
2586 2014-04-05 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
2588 * config/avr/avr-arch.h (avr_mcu_t): Add dev_attribute field to have
2589 device specific ISA/ feature information. Remove short_sp and
2590 errata_skip ds. Add avr_device_specific_features enum to have device
2592 * config/avr/avr-c.c (avr_cpu_cpp_builtins): use dev_attribute to check
2593 errata_skip. Add __AVR_ISA_RMW__ builtin macro if RMW ISA available.
2594 * config/avr/avr-devices.c (avr_mcu_types): Update AVR_MCU macro for
2595 updated device specific info.
2596 * config/avr/avr-mcus.def: Merge device specific details to
2597 dev_attribute field.
2598 * config/avr/avr.c (avr_2word_insn_p): use dev_attribute field to check
2600 * config/avr/avr.h (AVR_HAVE_8BIT_SP): same for short sp info.
2601 * config/avr/driver-avr.c (avr_device_to_as): Pass -mrmw option to
2602 assembler if RMW isa supported by current device.
2603 * config/avr/genmultilib.awk: Update as device info structure changed.
2604 * doc/invoke.texi: Add info for __AVR_ISA_RMW__ builtin macro
2606 2014-04-04 Cong Hou <congh@google.com>
2608 PR tree-optimization/60656
2609 * tree-vect-stmts.c (supportable_widening_operation):
2610 Fix a bug that elements in a vector with vect_used_by_reduction
2611 property are incorrectly reordered when the operation on it is not
2612 consistant with the one in reduction operation.
2614 2014-04-04 John David Anglin <danglin@gcc.gnu.org>
2616 PR rtl-optimization/60155
2617 * gcse.c (record_set_data): New function.
2618 (single_set_gcse): New function.
2619 (gcse_emit_move_after): Use single_set_gcse instead of single_set.
2620 (hoist_code): Likewise.
2621 (get_pressure_class_and_nregs): Likewise.
2623 2014-04-04 Eric Botcazou <ebotcazou@adacore.com>
2625 * explow.c (probe_stack_range): Emit a final optimization blockage.
2627 2014-04-04 Anthony Green <green@moxielogic.com>
2629 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2): Fix
2632 2014-04-04 Jan Hubicka <hubicka@ucw.cz>
2635 * lto-cgraph.c (input_overwrite_node): Check that partitioning
2636 flags are set only during streaming.
2637 * ipa.c (process_references, walk_polymorphic_call_targets,
2638 symtab_remove_unreachable_nodes): Drop bodies of always inline
2639 after early inlining.
2640 (symtab_remove_unreachable_nodes): Remove always_inline attribute.
2642 2014-04-04 Jakub Jelinek <jakub@redhat.com>
2643 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2646 * dwarf2out.c (const_ok_for_output_1): Reject expressions
2649 2014-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2652 * config/arm/cortex-a53.md (cortex_a53_fdivs): Reduce reservation
2654 (cortex_a53_fdivd): Likewise.
2656 2014-04-04 Martin Jambor <mjambor@suse.cz>
2659 * cgraph.h (cgraph_clone_node): New parameter added to declaration.
2661 * cgraph.c (clone_of_p): Also return true if thunks match.
2662 (verify_edge_corresponds_to_fndecl): Removed extraneous call to
2663 cgraph_function_or_thunk_node and an obsolete comment.
2664 * cgraphclones.c (build_function_type_skip_args): Moved upwards in the
2666 (build_function_decl_skip_args): Likewise.
2667 (set_new_clone_decl_and_node_flags): New function.
2668 (duplicate_thunk_for_node): Likewise.
2669 (redirect_edge_duplicating_thunks): Likewise.
2670 (cgraph_clone_node): New parameter args_to_skip, pass it to
2671 redirect_edge_duplicating_thunks which is called instead of
2672 cgraph_redirect_edge_callee.
2673 (cgraph_create_virtual_clone): Pass args_to_skip to cgraph_clone_node,
2674 moved setting of a lot of flags to set_new_clone_decl_and_node_flags.
2676 2014-04-04 Jeff Law <law@redhat.com>
2679 * config/arm/predicates.md (const_int_I_operand): New predicate.
2680 (const_int_M_operand): Similarly.
2681 * config/arm/arm.md (insv_zero): Use const_int_M_operand instead of
2683 (insv_t2, extv_reg, extzv_t2): Likewise.
2684 (load_multiple_with_writeback): Similarly for const_int_I_operand.
2685 (pop_multiple_with_writeback_and_return): Likewise.
2686 (vfp_pop_multiple_with_writeback): Likewise
2688 2014-04-04 Richard Biener <rguenther@suse.de>
2691 * tree-ssanames.c (make_ssa_name_fn): Fix assert.
2692 * gimple.c (gimple_set_bb): Avoid ICEing for NULL cfun for
2694 * gimplify.h (gimple_add_tmp_var_fn): Declare.
2695 * gimplify.c (gimple_add_tmp_var_fn): New function.
2696 * gimple-expr.h (create_tmp_reg_fn): Declare.
2697 * gimple-expr.c (create_tmp_reg_fn): New function.
2698 * gimple-low.c (record_vars_into): Don't change cfun.
2699 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix
2700 code generation without cfun.
2702 2014-04-04 Thomas Schwinge <thomas@codesourcery.com>
2705 * Makefile.in (install-driver): Fix shell scripting.
2707 2014-04-03 Cong Hou <congh@google.com>
2709 PR tree-optimization/60505
2710 * tree-vectorizer.h (struct _stmt_vec_info): Add th field as the
2711 threshold of number of iterations below which no vectorization
2713 * tree-vect-loop.c (new_loop_vec_info):
2714 Initialize LOOP_VINFO_COST_MODEL_THRESHOLD.
2715 * tree-vect-loop.c (vect_analyze_loop_operations):
2716 Set LOOP_VINFO_COST_MODEL_THRESHOLD.
2717 * tree-vect-loop.c (vect_transform_loop):
2718 Use LOOP_VINFO_COST_MODEL_THRESHOLD.
2719 * tree-vect-loop.c (vect_analyze_loop_2): Check the maximum number
2720 of iterations of the loop and see if we should build the epilogue.
2722 2014-04-03 Richard Biener <rguenther@suse.de>
2724 * tree-streamer.h (struct streamer_tree_cache_d): Add next_idx member.
2725 (streamer_tree_cache_create): Adjust.
2726 * tree-streamer.c (streamer_tree_cache_add_to_node_array): Adjust
2727 to allow optional nodes array.
2728 (streamer_tree_cache_insert_1): Use next_idx to assign idx.
2729 (streamer_tree_cache_append): Likewise.
2730 (streamer_tree_cache_create): Create nodes array optionally
2731 as specified by parameter.
2732 * lto-streamer-out.c (create_output_block): Avoid maintaining
2733 the node array in the writer cache.
2734 (DFS_write_tree): Remove assertion.
2735 (produce_asm_for_decls): Free the out decl state hash table early.
2736 * lto-streamer-in.c (lto_data_in_create): Adjust for
2737 streamer_tree_cache_create prototype change.
2739 2014-04-03 Richard Biener <rguenther@suse.de>
2741 * tree-streamer-out.c (streamer_write_chain): Do not temporarily
2742 set TREE_CHAIN to NULL_TREE.
2744 2014-04-03 Richard Biener <rguenther@suse.de>
2746 PR tree-optimization/60740
2747 * graphite-scop-detection.c (stmt_simple_for_scop_p): Iterate
2748 over all GIMPLE_COND operands.
2750 2014-04-03 Nathan Sidwell <nathan@codesourcery.com>
2752 * doc/invoke.texi (Wnon-virtual-dtor): Adjust documentation.
2753 (Weffc++): Remove Scott's numbering, merge lists and reference
2756 2014-04-03 Nick Clifton <nickc@redhat.com>
2758 * config/rl78/rl78-expand.md (movqi): Handle (SUBREG (SYMBOL_REF))
2761 2014-04-03 Martin Jambor <mjambor@suse.cz>
2763 * ipa-cp.c (ipcp_verify_propagated_values): Also dump symtab and
2764 mention gcc_unreachable before failing.
2765 * ipa.c (symtab_remove_unreachable_nodes): Also print order of
2768 2014-04-02 Jan Hubicka <hubicka@ucw.cz>
2771 * ipa-devirt.c (get_polymorphic_call_info): Do not ICE on type
2772 inconsistent code and instead mark the context inconsistent.
2773 (possible_polymorphic_call_targets): For inconsistent contexts
2774 return empty complete list.
2776 2014-04-02 Anthony Green <green@moxielogic.com>
2778 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2)
2779 (extendqisi2, extendhisi2): Define.
2780 * config/moxie/moxie.h (DEFAULT_SIGNED_CHAR): Change to 0.
2781 (WCHAR_TYPE): Change to unsigned int.
2783 2014-04-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2785 PR tree-optimization/60733
2786 * gimple-ssa-strength-reduction.c (ncd_with_phi): Change required
2787 insertion point for PHI candidates to be the end of the feeding
2788 block for the PHI argument.
2790 2014-04-02 Vladimir Makarov <vmakarov@redhat.com>
2792 PR rtl-optimization/60650
2793 * lra-constraints.c (process_alt_operands): Decrease reject for
2794 earlyclobber matching.
2796 2014-04-02 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2798 * config/s390/s390.c (s390_expand_insv): Use GET_MODE_BITSIZE.
2800 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
2802 * config/spu/spu.c (pad_bb): Do not crash when the last
2803 insn is CODE_FOR_blockage.
2805 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
2807 * config/spu/spu.md ("insv"): Fail if bitoffset+bitsize
2808 lies outside the target mode.
2810 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
2813 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
2814 software floating point or no floating point registers, do not
2815 allow any type in the FPRs. Eliminate a test for SPE SIMD types
2816 in GPRs that occurs after we tested for GPRs that would never be
2819 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
2820 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
2821 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
2822 specifically allow DDmode, since that does not use the SPE SIMD
2825 2014-04-02 Richard Biener <rguenther@suse.de>
2828 * optabs.c (expand_abs_nojump): Honor flag_trapv only for
2829 MODE_INTs. Properly use negv_optab.
2830 (expand_abs): Likewise.
2832 2014-04-02 Richard Biener <rguenther@suse.de>
2835 * Makefile.in (install-driver): Guard extra installs with special
2838 2014-04-01 Michael Meissner <meissner@linux.vnet.ibm.com>
2840 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
2843 2014-04-01 Richard Henderson <rth@redhat.com>
2846 * config/i386/i386.md (*float<SWI48><MODEF>2_sse): Leave the second
2847 alternative enabled before register allocation.
2849 2014-04-01 Chung-Lin Tang <cltang@codesourcery.com>
2851 * config/nios2/nios2.md (unspec): Remove UNSPEC_TLS, UNSPEC_TLS_LDM.
2852 * config/nios2/nios2.c (nios2_function_profiler): Fix addi operand
2854 (nios2_large_got_address): Remove unneeded 'sym' parameter.
2855 (nios2_got_address): Update nios2_large_got_address call site.
2856 (nios2_delegitimize_address): New function.
2857 (TARGET_DELEGITIMIZE_ADDRESS): Define to nios2_delegitimize_address.
2858 * config/nios2/linux.h (GLIBC_DYNAMIC_LINKER): Define.
2859 (LINK_SPEC): Specify dynamic linker using GNU_USER_DYNAMIC_LINKER.
2861 2014-04-01 Martin Husemann <martin@duskware.de>
2863 * config/mips/netbsd.h (TARGET_OS_CPP_BUILTINS): Define __mips_o32
2866 2014-04-01 Richard Sandiford <rdsandiford@googlemail.com>
2868 PR rtl-optimization/60604
2869 * recog.c (general_operand): Incorporate REG_CANNOT_CHANGE_MODE_P
2870 check from register_operand.
2871 (register_operand): Redefine in terms of general_operand.
2872 (nonmemory_operand): Use register_operand for the non-constant cases.
2874 2014-04-01 Richard Biener <rguenther@suse.de>
2876 * gimple.h (struct gimple_statement_base): Align subcode to 16 bits.
2878 2014-04-01 Sebastian Huber <sebastian.huber@embedded-brains.de>
2880 * doc/invoke.texi (mapp-regs): Clarify.
2882 2014-03-31 Ulrich Drepper <drepper@gmail.com>
2884 * config/i386/avx512fintrin.h (__v32hi): Define type.
2885 (__v64qi): Likewise.
2886 (_mm512_set1_epi8): Define.
2887 (_mm512_set1_epi16): Define.
2888 (_mm512_set4_epi32): Define.
2889 (_mm512_set4_epi64): Define.
2890 (_mm512_set4_pd): Define.
2891 (_mm512_set4_ps): Define.
2892 (_mm512_setr4_epi64): Define.
2893 (_mm512_setr4_epi32): Define.
2894 (_mm512_setr4_pd): Define.
2895 (_mm512_setr4_ps): Define.
2896 (_mm512_setzero_epi32): Define.
2898 2014-03-31 Martin Jambor <mjambor@suse.cz>
2901 * tree-sra.c (callsite_has_enough_arguments_p): Renamed to
2902 callsite_arguments_match_p. Updated all callers. Also check types of
2903 corresponding formal parameters and actual arguments.
2904 (not_all_callers_have_enough_arguments_p) Renamed to
2905 some_callers_have_mismatched_arguments_p.
2907 2014-03-31 Yuri Rumyantsev <ysrumyan@gmail.com>
2909 * tree-inline.c (copy_loops): Add missed copy of 'safelen'.
2911 2014-03-31 Kugan Vivekanandarajah <kuganv@linaro.org>
2914 * aarch64/aarch64.c (aarch64_classify_address): Fix alignment for
2917 2014-03-30 Uros Bizjak <ubizjak@gmail.com>
2919 * config/i386/sse.md (FMAMODE_NOVF512): New mode iterator.
2920 (<sd_mask_codefor>fma_fmadd_<mode><sd_maskz_name><round_name>):
2922 <sd_mask_codefor>fma_fmadd_<VF_512:mode><sd_maskz_name><round_name>.
2923 Use FMAMODE_NOVF512 mode iterator.
2924 (<sd_mask_codefor>fma_fmsub_<mode><sd_maskz_name><round_name>): Ditto.
2925 (<sd_mask_codefor>fma_fnmadd_<mode><sd_maskz_name><round_name>): Ditto.
2926 (<sd_mask_codefor>fma_fnmsub_<mode><sd_maskz_name><round_name>): Ditto.
2927 (<sd_mask_codefor>fma_fmaddsub_<mode><sd_maskz_name><round_name>):
2929 <sd_mask_codefor>fma_fmaddsub_<VF_512:mode><sd_maskz_name><round_name>.
2930 Use VF_128_256 mode iterator.
2931 (<sd_mask_codefor>fma_fmsubadd_<mode><sd_maskz_name><round_name>):
2934 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
2936 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Clear
2937 static chain if needed.
2939 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
2942 * lra-constraints.c (index_part_to_reg): New.
2943 (process_address): Use it.
2945 2014-03-27 Jeff Law <law@redhat.com>
2946 Jakub Jelinek <jakub@redhat.com>
2949 * expr.c (do_tablejump): Use simplify_gen_binary rather than
2950 gen_rtx_{PLUS,MULT} to build up the address expression.
2952 * i386/i386.c (ix86_legitimize_address): Use copy_addr_to_reg to avoid
2953 creating non-canonical RTL.
2955 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
2958 * ipa-inline.c (want_inline_small_function_p): Short circuit large
2959 functions; reorganize to make cheap checks first.
2960 (inline_small_functions): Do not estimate growth when dumping;
2962 * ipa-inline.h (inline_summary): Add min_size.
2963 (growth_likely_positive): New function.
2964 * ipa-inline-analysis.c (dump_inline_summary): Add min_size.
2965 (set_cond_stmt_execution_predicate): Cleanup.
2966 (estimate_edge_size_and_time): Compute min_size.
2967 (estimate_calls_size_and_time): Likewise.
2968 (estimate_node_size_and_time): Likewise.
2969 (inline_update_overall_summary): Update min_size.
2970 (do_estimate_edge_time): Likewise.
2971 (do_estimate_edge_size): Update.
2972 (do_estimate_edge_hints): Update.
2973 (growth_likely_positive): New function.
2975 2014-03-28 Jakub Jelinek <jakub@redhat.com>
2978 * config/i386/i386.c (ix86_copy_addr_to_reg): Call copy_addr_to_reg
2979 also if addr has VOIDmode.
2981 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2983 * config/arm/aarch-common.c (aarch_crypto_can_dual_issue): New.
2984 * config/arm/aarch-common-protos.h (aarch_crypto_can_dual_issue):
2986 * config/arm/cortex-a53.md: Add reservations and bypass for crypto
2987 instructions as well as AdvancedSIMD loads.
2989 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2991 * config/aarch64/aarch64-simd.md (aarch64_crypto_aes<aes_op>v16qi):
2992 Use crypto_aese type.
2993 (aarch64_crypto_aes<aesmc_op>v16qi): Use crypto_aesmc type.
2994 * config/arm/arm.md (is_neon_type): Replace crypto_aes with
2995 crypto_aese, crypto_aesmc. Move to types.md.
2996 * config/arm/types.md (crypto_aes): Split into crypto_aese,
2998 * config/arm/iterators.md (crypto_type): Likewise.
3000 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
3002 * cgraph.c: Include expr.h and tree-dfa.h.
3003 (cgraph_redirect_edge_call_stmt_to_callee): If call in noreturn;
3006 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
3009 * lra-assigns.c (find_hard_regno_for): Remove unavailable hard
3010 regs from checking multi-reg pseudos.
3012 2014-03-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3014 * config/arm/t-aprofile (MULTILIB_MATCHES): Correct A12 rule.
3016 2014-03-28 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
3018 * config/rs6000/rs6000.c (fusion_gpr_load_p): Refuse optimization
3019 if it would clobber the stack pointer, even temporarily.
3021 2014-03-28 Eric Botcazou <ebotcazou@adacore.com>
3023 * mode-switching.c: Make small adjustments to the top comment.
3025 2014-03-27 Michael Meissner <meissner@linux.vnet.ibm.com>
3027 * config/rs6000/constraints.md (wD constraint): New constraint to
3028 match the constant integer to get the top DImode/DFmode out of a
3029 vector in a VSX register.
3031 * config/rs6000/predicates.md (vsx_scalar_64bit): New predicate to
3032 match the constant integer to get the top DImode/DFmode out of a
3033 vector in a VSX register.
3035 * config/rs6000/rs6000-builtins.def (VBPERMQ): Add vbpermq builtin
3038 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
3041 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
3042 -mdebug=reg, print value of VECTOR_ELEMENT_SCALAR_64BIT.
3044 * config/rs6000/vsx.md (vsx_extract_<mode>, V2DI/V2DF modes):
3045 Optimize vec_extract of 64-bit values, where the value being
3046 extracted is in the top word, where we can use scalar
3047 instructions. Add direct move and store support. Combine the big
3048 endian/little endian vector select load support into a single insn.
3049 (vsx_extract_<mode>_internal1): Likewise.
3050 (vsx_extract_<mode>_internal2): Likewise.
3051 (vsx_extract_<mode>_load): Likewise.
3052 (vsx_extract_<mode>_store): Likewise.
3053 (vsx_extract_<mode>_zero): Delete, big and little endian insns are
3054 combined into vsx_extract_<mode>_load.
3055 (vsx_extract_<mode>_one_le): Likewise.
3057 * config/rs6000/rs6000.h (VECTOR_ELEMENT_SCALAR_64BIT): Macro to
3058 define the top 64-bit vector element.
3060 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wD
3063 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
3064 Document vec_vbpermq builtin.
3067 * config/rs6000/altivec.h (vec_xxsldwi): Add missing define to
3068 enable use of xxsldwi and xxpermdi builtin functions.
3069 (vec_xxpermdi): Likewise.
3071 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
3072 Document use of vec_xxsldwi and vec_xxpermdi builtins.
3074 2014-03-27 Vladimir Makarov <vmakarov@redhat.com>
3076 PR rtl-optimization/60650
3077 * lra-assign.c (find_hard_regno_for, spill_for): Add parameter
3079 (find_spills_for): New.
3080 (assign_by_spills): Pass the new parameter to find_hard_regno_for.
3081 Spill all pseudos on the second iteration.
3083 2014-03-27 Marek Polacek <polacek@redhat.com>
3086 * doc/extend.texi (ffs Builtins): Change unsigned types to signed
3089 2014-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3091 * config/s390/s390.c (s390_can_use_return_insn): Check for
3092 call-saved FPRs on 31 bit.
3094 2014-03-27 Jakub Jelinek <jakub@redhat.com>
3097 * omp-low.c (lower_omp_1): For gimple_clobber_p stmts,
3098 if they need regimplification, just drop them instead of
3099 calling gimple_regimplify_operands on them.
3101 2014-03-27 Marcus Shawcroft <marcus.shawcroft@arm.com>
3104 * config/aarch64/aarch64.c (faked_omit_frame_pointer): Remove.
3105 (aarch64_frame_pointer_required): Adjust logic.
3106 (aarch64_can_eliminate): Adjust logic.
3107 (aarch64_override_options_after_change): Adjust logic.
3109 2014-03-27 Dehao Chen <dehao@google.com>
3111 * ipa-inline.c (early_inliner): Update node's inline info.
3113 2014-03-26 Dehao Chen <dehao@google.com>
3115 * dojump.c (do_compare_rtx_and_jump): Sets correct probability for
3116 compiler inserted conditional jumps for NAN float check.
3118 2014-03-26 Jakub Jelinek <jakub@redhat.com>
3120 * ubsan.h (ubsan_create_data): Change second argument's type
3121 to const location_t *.
3122 * ubsan.c (ubsan_source_location): If xloc.file is NULL, set it to
3124 (ubsan_create_data): Change second argument to const location_t *PLOC.
3125 Create Loc field whenever PLOC is non-NULL.
3126 (ubsan_instrument_unreachable, ubsan_expand_null_ifn,
3127 ubsan_build_overflow_builtin, instrument_bool_enum_load): Adjust
3131 * real.c (real_to_integer2): Change type of low to UHWI.
3133 2014-03-26 Tobias Burnus <burnus@net-b.de>
3135 * gcc.c (LINK_COMMAND_SPEC): Use libcilkrts.spec for -fcilkplus.
3136 (CILK_SELF_SPECS): New define.
3137 (driver_self_specs): Use it.
3139 2014-03-26 Richard Biener <rguenther@suse.de>
3141 * tree-pretty-print.c (percent_K_format): Implement special
3142 case for LTO and its stripped down BLOCK tree.
3144 2014-03-26 Jakub Jelinek <jakub@redhat.com>
3147 * ubsan.c (instrument_si_overflow): Instrument ABS_EXPR.
3149 * tree-vrp.c (simplify_internal_call_using_ranges): If only
3150 one range is range_int_cst_p, but not both, at least optimize
3151 addition/subtraction of 0 and multiplication by 0 or 1.
3152 * gimple-fold.c (gimple_fold_call): Fold
3153 IFN_UBSAN_CHECK_{ADD,SUB,MUL}.
3154 (gimple_fold_stmt_to_constant_1): If both op0 and op1 aren't
3155 INTEGER_CSTs, try to fold at least x * 0 and y - y.
3157 2014-03-26 Eric Botcazou <ebotcazou@adacore.com>
3159 PR rtl-optimization/60452
3160 * rtlanal.c (rtx_addr_can_trap_p_1): Fix head comment.
3161 <case REG>: Return 1 for invalid offsets from the frame pointer.
3163 2014-03-26 Marek Polacek <polacek@redhat.com>
3166 * doc/extend.texi (C Extensions): Mention variable-length arrays in
3169 2014-03-26 Marek Polacek <polacek@redhat.com>
3172 * doc/extend.texi (Designated Inits): Describe what happens to omitted
3175 2014-03-26 Marek Polacek <polacek@redhat.com>
3178 * ira-color.c (update_conflict_hard_regno_costs): Perform the
3179 multiplication in unsigned type.
3181 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
3183 * doc/install.texi: Document nds32le-*-elf and nds32be-*-elf.
3185 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
3187 * doc/contrib.texi: Add myself as Andes nds32 port contributor.
3189 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
3192 * cif-code.def (UNREACHABLE) New code.
3193 * ipa-inline.c (inline_small_functions): Skip edges to
3194 __builtlin_unreachable.
3195 (estimate_edge_growth): Allow edges to __builtlin_unreachable.
3196 * ipa-inline-analysis.c (edge_set_predicate): Redirect edges with false
3197 predicate to __bulitin_unreachable.
3198 (set_cond_stmt_execution_predicate): Fix issue when
3199 invert_tree_comparison returns ERROR_MARK.
3200 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow): Do not
3201 propagate to inline clones.
3202 * cgraph.c (verify_edge_corresponds_to_fndecl): Allow redirection
3204 * ipa-cp.c (create_specialized_node): Be ready for new node to appear.
3205 * cgraphclones.c (cgraph_clone_node): If call destination is already
3206 ureachable, do not redirect it back.
3207 * tree-inline.c (fold_marked_statements): Hanlde calls becoming
3210 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
3212 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow):
3213 Do not modify inline clones.
3215 2014-03-25 Jakub Jelinek <jakub@redhat.com>
3217 * config/i386/i386.md (general_sext_operand): New mode attr.
3218 (addv<mode>4, subv<mode>4, mulv<mode>4): If operands[2] is CONST_INT,
3219 don't generate (sign_extend (const_int)).
3220 (*addv<mode>4, *subv<mode>4, *mulv<mode>4): Disallow CONST_INT_P
3221 operands[2]. Use We constraint instead of <i> and
3222 <general_sext_operand> predicate instead of <general_operand>.
3223 (*addv<mode>4_1, *subv<mode>4_1, *mulv<mode>4_1): New insns.
3224 * config/i386/constraints.md (We): New constraint.
3225 * config/i386/predicates.md (x86_64_sext_operand,
3226 sext_operand): New predicates.
3228 2014-03-25 Martin Jambor <mjambor@suse.cz>
3231 * ipa-cp.c (ipa_get_indirect_edge_target_1): Redirect type
3232 inconsistent devirtualizations to __builtin_unreachable.
3234 2014-03-25 Marek Polacek <polacek@redhat.com>
3237 * doc/extend.texi (Example of asm with clobbered asm reg): Fix typo.
3239 2014-03-25 Alan Lawrence <alan.lawrence@arm.com>
3241 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Reverse
3242 order of elements for big-endian.
3244 2014-03-25 Richard Biener <rguenther@suse.de>
3247 * gimplify-me.c (gimple_regimplify_operands): Update the
3250 2014-03-25 Martin Jambor <mjambor@suse.cz>
3253 * lto-cgraph.c (lto_output_node): Stream body_removed flag.
3254 (lto_output_varpool_node): Likewise.
3255 (input_overwrite_node): Likewise.
3256 (input_varpool_node): Likewise.
3258 2014-03-25 Richard Biener <rguenther@suse.de>
3260 * lto-wrapper.c (merge_and_complain): Handle OPT_fPIE like OPT_fpie.
3261 (run_gcc): Likewise.
3263 2014-03-25 Jakub Jelinek <jakub@redhat.com>
3265 * combine.c (simplify_compare_const): Add MODE argument.
3266 Handle mode_width 0 as very large mode_width.
3267 (try_combine, simplify_comparison): Adjust callers.
3269 * cselib.c (cselib_hash_rtx): Perform addition in unsigned
3270 type to avoid signed integer overflow.
3271 * explow.c (plus_constant): Likewise.
3273 2014-03-25 Dominik Vogt <vogt@linux.vnet.ibm.com>
3275 * doc/generic.texi: Correct typos.
3277 2014-03-24 Tobias Burnus <burnus@net-b.de>
3279 * doc/invoke.texi (-flto): Expand section about
3280 using static libraries with LTO.
3282 2014-03-24 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3284 PR rtl-optimization/60501
3285 * optabs.def (addptr3_optab): New optab.
3286 * optabs.c (gen_addptr3_insn, have_addptr3_insn): New function.
3287 * doc/md.texi ("addptrm3"): Document new RTL standard expander.
3288 * expr.h (gen_addptr3_insn, have_addptr3_insn): Add prototypes.
3290 * lra.c (emit_add3_insn): Use the addptr pattern if available.
3292 * config/s390/s390.md ("addptrdi3", "addptrsi3"): New expanders.
3294 2014-03-24 Ulrich Drepper <drepper@gmail.com>
3296 * config/i386/avx512fintrin.h: Define _mm512_set1_ps and
3299 * config/i386/avxintrin.h (_mm256_undefined_si256): Define.
3300 (_mm256_undefined_ps): Define.
3301 (_mm256_undefined_pd): Define.
3302 * config/i386/emmintrin.h (_mm_undefined_si128): Define.
3303 (_mm_undefined_pd): Define.
3304 * config/i386/xmmintrin.h (_mm_undefined_ps): Define.
3305 * config/i386/avx512fintrin.h (_mm512_undefined_si512): Define.
3306 (_mm512_undefined_ps): Define.
3307 (_mm512_undefined_pd): Define.
3308 Use _mm*_undefined_*.
3309 * config/i386/avx2intrin.h: Use _mm*_undefined_*.
3311 2014-03-24 Alex Velenko <Alex.Velenko@arm.com>
3313 * config/aarch64/aarch64-simd-builtins.def (lshr): DI mode excluded.
3314 (lshr_simd): DI mode added.
3315 * config/aarch64/aarch64-simd.md (aarch64_lshr_simddi): New pattern.
3316 (aarch64_ushr_simddi): Likewise.
3317 * config/aarch64/aarch64.md (UNSPEC_USHR64): New unspec.
3318 * config/aarch64/arm_neon.h (vshr_n_u64): Intrinsic fixed.
3319 (vshrd_n_u64): Likewise.
3321 2014-03-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3323 * Makefile.in (s-macro_list): Depend on cc1.
3325 2014-03-23 Teresa Johnson <tejohnson@google.com>
3327 * ipa-utils.c (ipa_print_order): Use specified dump file.
3329 2014-03-23 Eric Botcazou <ebotcazou@adacore.com>
3331 PR rtl-optimization/60601
3332 * bb-reorder.c (fix_up_fall_thru_edges): Test EDGE_FALLTHRU everywhere.
3334 * gcc.c (eval_spec_function): Initialize save_growing_value.
3336 2014-03-22 Jakub Jelinek <jakub@redhat.com>
3339 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): For
3340 code == MINUS_EXPR, never swap op0 with op1.
3342 * toplev.c (init_local_tick): Avoid signed integer multiplication
3344 * genautomata.c (reserv_sets_hash_value): Fix rotate idiom, avoid
3345 shift by first operand's bitsize.
3347 2014-03-21 Jakub Jelinek <jakub@redhat.com>
3350 * config/i386/i386.h (TARGET_64BIT_P): If not TARGET_BI_ARCH,
3352 * config/i386/darwin.h (TARGET_64BIT_P): Redefine to
3353 TARGET_ISA_64BIT_P(x).
3355 2014-03-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3357 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Generate a
3358 pattern for vector nor instead of subtract from splat(-1).
3359 (altivec_expand_vec_perm_const_le): Likewise.
3361 2014-03-21 Richard Henderson <rth@twiddle.net>
3364 * ifcvt.c (dead_or_predicable): Return FALSE if there are any frame
3365 related insns after epilogue_completed.
3367 2014-03-21 Martin Jambor <mjambor@suse.cz>
3370 * cgraph.h (symtab_node): New flag body_removed.
3371 * ipa.c (symtab_remove_unreachable_nodes): Set body_removed flag
3372 when removing bodies.
3373 * symtab.c (dump_symtab_base): Dump body_removed flag.
3374 * cgraph.c (verify_edge_corresponds_to_fndecl): Skip nodes which
3375 had their bodies removed.
3377 2014-03-21 Martin Jambor <mjambor@suse.cz>
3380 * ipa.c (symtab_remove_unreachable_nodes): Clear thunk flag of nodes
3383 2014-03-21 Richard Biener <rguenther@suse.de>
3385 PR tree-optimization/60577
3386 * tree-core.h (struct tree_base): Document nothrow_flag use
3387 in VAR_DECL_NONALIASED.
3388 * tree.h (VAR_DECL_NONALIASED): New.
3389 (may_be_aliased): Adjust.
3390 * coverage.c (build_var): Set VAR_DECL_NONALIASED.
3392 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
3394 * expr.c (expand_expr_real_1): Remove outdated comment.
3396 2014-03-20 Jakub Jelinek <jakub@redhat.com>
3399 * ira.c (adjust_cleared_regs): Call copy_rtx on
3400 *reg_equiv[REGNO (loc)].src_p before passing it to
3401 simplify_replace_fn_rtx.
3404 * config/i386/i386.c (x86_output_mi_thunk): Surround UNSPEC_GOT
3405 into CONST, put pic register as first operand of PLUS. Use
3406 gen_const_mem for both 32-bit and 64-bit PIC got loads.
3408 2014-03-20 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3410 * config/aarch64/aarch64.c (MEMORY_MOVE_COST): Delete.
3412 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
3414 * config/sparc/sparc.c (sparc_do_work_around_errata): Implement work
3415 around for store forwarding issue in the FPU on the UT699.
3416 * config/sparc/sparc.md (in_branch_delay): Return false for single FP
3417 loads and operations if -mfix-ut699 is specified.
3418 (divtf3_hq): Tweak attribute.
3419 (sqrttf2_hq): Likewise.
3421 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
3423 * calls.c (store_one_arg): Remove incorrect const qualification on the
3424 type of the temporary.
3425 * cfgexpand.c (expand_return): Likewise.
3426 * expr.c (expand_constructor): Likewise.
3427 (expand_expr_real_1): Likewise.
3429 2014-03-20 Zhenqiang Chen <zhenqiang.chen@linaro.org>
3431 * config/arm/arm.c (arm_dwarf_register_span): Update the element number
3434 2014-03-19 Kaz Kojima <kkojima@gcc.gnu.org>
3437 * config/sh/sh.md (udivsi3_i1): Clobber R1 register.
3439 2014-03-19 James Greenhalgh <james.greenhalgh@arm.com>
3441 * config/arm/aarch-common-protos.h
3442 (alu_cost_table): Fix spelling of "extend".
3443 * config/arm/arm.c (arm_new_rtx_costs): Fix spelling of "extend".
3445 2014-03-19 Richard Biener <rguenther@suse.de>
3448 * tree-core.h (tree_type_common): Re-order pointer members
3449 to reduce recursion depth during GC walks.
3451 2014-03-19 Marek Polacek <polacek@redhat.com>
3454 * ubsan.c (ubsan_type_descriptor): Check that DECL_NAME is nonnull
3455 before accessing it.
3457 2014-03-19 Richard Biener <rguenther@suse.de>
3460 * lto-streamer-in.c (input_function): In WPA stage do not drop
3463 2014-03-19 Jakub Jelinek <jakub@redhat.com>
3465 PR tree-optimization/60559
3466 * vectorizable_mask_load_store): Replace scalar MASK_LOAD
3467 with build_zero_cst assignment.
3469 2014-03-18 Kai Tietz <ktietz@redhat.com>
3471 PR rtl-optimization/56356
3472 * sdbout.c (sdbout_parms): Verify that parms'
3473 incoming argument is valid.
3474 (sdbout_reg_parms): Likewise.
3476 2014-03-18 Richard Henderson <rth@redhat.com>
3479 * config/i386/i386.md (*float<SWI48x><MODEF>2_i387): Move down to
3480 be shadowed by *float<SWI48><MODEF>2_sse. Test X87_ENABLE_FLOAT.
3481 (*float<SWI48><MODEF>2_sse): Check X87_ENABLE_FLOAT for alternative 0.
3483 2014-03-18 Basile Starynkevitch <basile@starynkevitch.net>
3485 * plugin.def: Improve comment for PLUGIN_INCLUDE_FILE.
3486 * doc/plugins.texi (Plugin callbacks): Mention PLUGIN_INCLUDE_FILE.
3487 Italicize plugin event names in description. Explain that
3488 PLUGIN_PRAGMAS has no sense for lto1. Explain PLUGIN_INCLUDE_FILE.
3489 Remind that no GCC functions should be called after PLUGIN_FINISH.
3490 Explain what pragmas with expansion are.
3492 2014-03-18 Martin Liska <mliska@suse.cz>
3494 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Added case when
3495 gimple call statement is update.
3496 * gimple-fold.c (gimple_fold_call): Changed order for GIMPLE_ASSIGN and
3497 GIMPLE_CALL, where gsi iterator still points to GIMPLE CALL.
3499 2014-03-18 Jakub Jelinek <jakub@redhat.com>
3502 * ubsan.c (ubsan_instrument_unreachable): Call
3503 initialize_sanitizer_builtins.
3504 (ubsan_pass): Likewise.
3507 * ubsan.c (ubsan_type_descriptor, ubsan_create_data): Call
3508 varpool_finalize_decl instead of rest_of_decl_compilation.
3510 2014-03-18 Richard Biener <rguenther@suse.de>
3512 * df-problems.c (df_rd_confluence_n): Avoid bitmap_copy
3513 by using bitmap_and_compl instead of bitmap_and_compl_into.
3514 (df_rd_transfer_function): Likewise.
3516 2014-03-18 Richard Biener <rguenther@suse.de>
3518 * doc/lto.texi (fresolution): Fix typo.
3520 2014-03-18 Richard Biener <rguenther@suse.de>
3522 * doc/invoke.texi (flto): Update for changes in 4.9.
3524 2014-03-18 Richard Biener <rguenther@suse.de>
3526 * doc/loop.texi: Remove section on the removed lambda framework.
3527 Update loop docs with recent changes in preserving loop structure.
3529 2014-03-18 Richard Biener <rguenther@suse.de>
3531 * doc/lto.texi (-fresolution): Document.
3533 2014-03-18 Richard Biener <rguenther@suse.de>
3535 * doc/contrib.texi: Adjust my name.
3537 2014-03-18 Jakub Jelinek <jakub@redhat.com>
3540 * internal-fn.c: Include diagnostic-core.h.
3541 (expand_BUILTIN_EXPECT): New function.
3542 * gimplify.c (gimplify_call_expr): Use false instead of FALSE.
3543 (gimplify_modify_expr): Gimplify 3 argument __builtin_expect into
3544 IFN_BUILTIN_EXPECT call instead of __builtin_expect builtin call.
3545 * ipa-inline-analysis.c (find_foldable_builtin_expect): Handle
3547 * predict.c (expr_expected_value_1): Handle IFN_BUILTIN_EXPECT.
3548 Revert 3 argument __builtin_expect code.
3549 (strip_predict_hints): Handle IFN_BUILTIN_EXPECT.
3550 * gimple-fold.c (gimple_fold_call): Likewise.
3551 * tree.h (fold_builtin_expect): New prototype.
3552 * builtins.c (build_builtin_expect_predicate): Add predictor
3553 argument, if non-NULL, create 3 argument __builtin_expect.
3554 (fold_builtin_expect): No longer static. Add ARG2 argument,
3555 pass it through to build_builtin_expect_predicate.
3556 (fold_builtin_2): Adjust caller.
3557 (fold_builtin_3): Handle BUILT_IN_EXPECT.
3558 * internal-fn.def (BUILTIN_EXPECT): New.
3560 2014-03-18 Tobias Burnus <burnus@net-b.de>
3563 * predict.def (PRED_FORTRAN_OVERFLOW, PRED_FORTRAN_FAIL_ALLOC,
3564 PRED_FORTRAN_FAIL_IO, PRED_FORTRAN_WARN_ONCE, PRED_FORTRAN_SIZE_ZERO,
3565 PRED_FORTRAN_INVALID_BOUND, PRED_FORTRAN_ABSENT_DUMMY): Add.
3567 2014-03-18 Jan Hubicka <hubicka@ucw.cz>
3570 * predict.c (combine_predictions_for_bb): Fix up formatting.
3571 (expr_expected_value_1, expr_expected_value): Add predictor argument,
3572 fill what it points to if non-NULL.
3573 (tree_predict_by_opcode): Adjust caller, use the predictor.
3574 * predict.def (PRED_COMPARE_AND_SWAP): Add.
3576 2014-03-18 Eric Botcazou <ebotcazou@adacore.com>
3578 * config/sparc/sparc.c (sparc_do_work_around_errata): Speed up and use
3579 proper constant for the store mode.
3581 2014-03-18 Ilya Enkovich <ilya.enkovich@intel.com>
3583 * symtab.c (change_decl_assembler_name): Fix transparent alias
3586 2014-03-16 Renlin Li <Renlin.Li@arm.com>
3588 * config/aarch64/aarch64.c: Correct the comments about the
3589 aarch64 stack layout.
3591 2014-03-18 Thomas Schwinge <thomas@codesourcery.com>
3593 * omp-low.c (lower_rec_input_clauses) <build_omp_barrier>: Restore
3594 check for GF_OMP_FOR_KIND_FOR.
3596 2013-03-18 Kirill Yukhin <kirill.yukhin@intel.com>
3598 * config/i386/i386.h (ADDITIONAL_REGISTER_NAMES): Add
3599 ymm and zmm register names.
3601 2014-03-17 Jakub Jelinek <jakub@redhat.com>
3604 * config/i386/i386.c (ix86_expand_epilogue): Adjust REG_CFA_ADJUST_CFA
3605 note creation for the 2010-08-31 changes.
3607 2014-03-17 Marek Polacek <polacek@redhat.com>
3610 * omp-low.c (omp_max_vf): Treat -fno-tree-loop-optimize the same
3611 as -fno-tree-loop-vectorize.
3612 (expand_omp_simd): Likewise.
3614 2014-03-15 Eric Botcazou <ebotcazou@adacore.com>
3616 * config/sparc/sparc-protos.h (tls_call_delay): Delete.
3617 (eligible_for_call_delay): New prototype.
3618 * config/sparc/sparc.c (tls_call_delay): Rename into...
3619 (eligible_for_call_delay): ...this. Return false if the instruction
3620 cannot be put in the delay slot of a branch.
3621 (eligible_for_restore_insn): Simplify.
3622 (eligible_for_return_delay): Return false if the instruction cannot be
3623 put in the delay slot of a branch and simplify.
3624 (eligible_for_sibcall_delay): Return false if the instruction cannot be
3625 put in the delay slot of a branch.
3626 * config/sparc/sparc.md (fix_ut699): New attribute.
3627 (tls_call_delay): Delete.
3628 (in_call_delay): Reimplement.
3629 (eligible_for_sibcall_delay): Rename into...
3630 (in_sibcall_delay): ...this.
3631 (eligible_for_return_delay): Rename into...
3632 (in_return_delay): ...this.
3633 (in_branch_delay): Reimplement.
3634 (in_uncond_branch_delay): Delete.
3635 (in_annul_branch_delay): Delete.
3637 2014-03-14 Richard Henderson <rth@redhat.com>
3640 * config/i386/i386.md (floathi<X87MODEF>2): Delete expander; rename
3641 define_insn from *floathi<X87MODEF>2_i387; allow nonimmediate_operand.
3642 (*floathi<X87MODEF>2_i387_with_temp): Remove.
3643 (floathi splitters): Remove.
3644 (float<SWI48x>xf2): New pattern.
3645 (float<SWI48><MODEF>2): Rename from float<SWI48x><X87MODEF>2. Drop
3646 code that tried to handle DImode for 32-bit, but which was excluded
3647 by the pattern's condition. Drop allocation of stack temporary.
3648 (*floatsi<MODEF>2_vector_mixed_with_temp): Remove.
3649 (*float<SWI48><MODEF>2_mixed_with_temp): Remove.
3650 (*float<SWI48><MODEF>2_mixed_interunit): Remove.
3651 (*float<SWI48><MODEF>2_mixed_nointerunit): Remove.
3652 (*floatsi<MODEF>2_vector_sse_with_temp): Remove.
3653 (*float<SWI48><MODEF>2_sse_with_temp): Remove.
3654 (*float<SWI48><MODEF>2_sse_interunit): Remove.
3655 (*float<SWI48><MODEF>2_sse_nointerunit): Remove.
3656 (*float<SWI48x><X87MODEF>2_i387_with_temp): Remove.
3657 (*float<SWI48x><X87MODEF>2_i387): Remove.
3658 (all float _with_temp splitters): Remove.
3659 (*float<SWI48x><MODEF>2_i387): New pattern.
3660 (*float<SWI48><MODEF>2_sse): New pattern.
3661 (float TARGET_USE_VECTOR_CONVERTS splitters): Merge them.
3662 (float TARGET_SSE_PARTIAL_REG_DEPENDENCY splitters): Merge them.
3664 2014-03-14 Jakub Jelinek <jakub@redhat.com>
3665 Marek Polacek <polacek@redhat.com>
3668 * common.opt (dump_base_name_prefixed): New Variable.
3669 * opts.c (finish_options): Don't prepend directory to x_dump_base_name
3670 if x_dump_base_name_prefixed is already set, set it at the end.
3672 2014-03-14 Vladimir Makarov <vmakarov@redhat.com>
3674 PR rtl-optimization/60508
3675 * lra-constraints.c (get_reload_reg): Add new parameter
3677 (process_addr_reg, simplify_operand_subreg, curr_insn_transform):
3678 Pass the new parameter values.
3680 2014-03-14 Richard Biener <rguenther@suse.de>
3682 * common.opt: Revert unintented changes from r205065.
3685 2014-03-14 Richard Biener <rguenther@suse.de>
3688 * cfghooks.c (split_block): Properly adjust all loops the
3689 block was a latch of.
3691 2014-03-14 Martin Jambor <mjambor@suse.cz>
3694 * ipa-prop.c (ipa_modify_call_arguments): Fix iteration condition
3697 2014-03-14 Georg-Johann Lay <avr@gjlay.de>
3700 * config/avr/avr.c (avr_set_current_function): Pass function name
3701 through default_strip_name_encoding before sanity checking instead
3702 of skipping the first char of the assembler name.
3704 2014-03-13 Richard Henderson <rth@redhat.com>
3707 * config/i386/i386.c (ix86_split_fp_branch): Remove pushed argument.
3708 (ix86_force_to_memory, ix86_free_from_memory): Remove.
3709 * config/i386/i386-protos.h: Likewise.
3710 * config/i386/i386.md (floathi<X87MODEF>2): Use assign_386_stack_local
3711 in the expander instead of a splitter.
3712 (float<SWI48x><X87MODEF>2): Use assign_386_stack_local if there is
3713 any possibility of requiring a memory.
3714 (*floatsi<MODEF>2_vector_mixed): Remove, and the splitters.
3715 (*floatsi<MODEF>2_vector_sse): Remove, and the splitters.
3716 (fp branch splitters): Update for ix86_split_fp_branch.
3717 (*jcc<X87MODEF>_<SWI24>_i387): Remove r/f alternative.
3718 (*jcc<X87MODEF>_<SWI24>_r_i387): Likewise.
3719 (splitter for jcc<X87MODEF>_<SWI24>_i387 r/f): Remove.
3720 (*fop_<MODEF>_2_i387): Remove f/r alternative.
3721 (*fop_<MODEF>_3_i387): Likewise.
3722 (*fop_xf_2_i387, *fop_xf_3_i387): Likewise.
3723 (splitters for the fop_* register patterns): Remove.
3724 (fscalexf4_i387): Rename from *fscalexf4_i387.
3725 (ldexpxf3): Use gen_floatsixf2 and gen_fscalexf4_i387.
3727 2014-03-13 Jakub Jelinek <jakub@redhat.com>
3729 PR tree-optimization/59779
3730 * tree-dfa.c (get_ref_base_and_extent): Use double_int
3731 type for bitsize and maxsize instead of HOST_WIDE_INT.
3733 2014-03-13 Steven Bosscher <steven@gcc.gnu.org>
3735 PR rtl-optimization/57320
3736 * function.c (rest_of_handle_thread_prologue_and_epilogue): Cleanup
3737 the CFG after thread_prologue_and_epilogue_insns.
3739 2014-03-13 Vladimir Makarov <vmakarov@redhat.com>
3741 PR rtl-optimization/57189
3742 * lra-constraints.c (process_alt_operands): Disfavor spilling
3745 2014-03-13 Cesar Philippidis <cesar@codesourcery.com>
3747 * lto-wrapper.c (maybe_unlink_file): Suppress diagnostic messages.
3749 2014-03-13 Jakub Jelinek <jakub@redhat.com>
3751 PR tree-optimization/59025
3753 * tree-ssa-reassoc.c (sort_by_operand_rank): For SSA_NAMEs with the
3754 same rank, sort by bb_rank and gimple_uid of SSA_NAME_DEF_STMT first.
3756 2014-03-13 Georg-Johann Lay <avr@gjlay.de>
3759 * config/avr/avr.c (avr_out_plus): Swap cc_plus and cc_minus in
3760 calls of avr_out_plus_1.
3762 2014-03-13 Bin Cheng <bin.cheng@arm.com>
3764 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Record
3765 BB's single pred and update the father loop's latch info later.
3767 2014-03-12 Michael Meissner <meissner@linux.vnet.ibm.com>
3769 * config/rs6000/vector.md (VEC_L): Add V1TI mode to vector types.
3773 (VEC_base): Likewise.
3774 (mov<MODE>, VEC_M modes): If we are loading TImode into VSX
3775 registers, we need to swap double words in little endian mode.
3777 * config/rs6000/rs6000-modes.def (V1TImode): Add new vector mode
3778 to be a container mode for 128-bit integer operations added in ISA
3779 2.07. Unlike TImode and PTImode, the preferred register set is
3780 the Altivec/VMX registers for the 128-bit operations.
3782 * config/rs6000/rs6000-protos.h (rs6000_move_128bit_ok_p): Add
3784 (rs6000_split_128bit_ok_p): Likewise.
3786 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_3): Add new support
3787 macros for creating ISA 2.07 normal and overloaded builtin
3788 functions with 3 arguments.
3789 (BU_P8V_OVERLOAD_3): Likewise.
3790 (VPERM_1T): Add support for V1TImode in 128-bit vector operations
3791 for use as overloaded functions.
3792 (VPERM_1TI_UNS): Likewise.
3793 (VSEL_1TI): Likewise.
3794 (VSEL_1TI_UNS): Likewise.
3795 (ST_INTERNAL_1ti): Likewise.
3796 (LD_INTERNAL_1ti): Likewise.
3797 (XXSEL_1TI): Likewise.
3798 (XXSEL_1TI_UNS): Likewise.
3799 (VPERM_1TI): Likewise.
3800 (VPERM_1TI_UNS): Likewise.
3801 (XXPERMDI_1TI): Likewise.
3802 (SET_1TI): Likewise.
3803 (LXVD2X_V1TI): Likewise.
3804 (STXVD2X_V1TI): Likewise.
3805 (VEC_INIT_V1TI): Likewise.
3806 (VEC_SET_V1TI): Likewise.
3807 (VEC_EXT_V1TI): Likewise.
3808 (EQV_V1TI): Likewise.
3809 (NAND_V1TI): Likewise.
3810 (ORC_V1TI): Likewise.
3811 (VADDCUQ): Add support for 128-bit integer arithmetic instructions
3812 added in ISA 2.07. Add both normal 'altivec' builtins, and the
3814 (VADDUQM): Likewise.
3815 (VSUBCUQ): Likewise.
3816 (VADDEUQM): Likewise.
3817 (VADDECUQ): Likewise.
3818 (VSUBEUQM): Likewise.
3819 (VSUBECUQ): Likewise.
3821 * config/rs6000/rs6000-c.c (__int128_type): New static to hold
3822 __int128_t and __uint128_t types.
3823 (__uint128_type): Likewise.
3824 (altivec_categorize_keyword): Add support for vector __int128_t,
3825 vector __uint128_t, vector __int128, and vector unsigned __int128
3826 as a container type for TImode operations that need to be done in
3827 VSX/Altivec registers.
3828 (rs6000_macro_to_expand): Likewise.
3829 (altivec_overloaded_builtins): Add ISA 2.07 overloaded functions
3830 to support 128-bit integer instructions vaddcuq, vadduqm,
3831 vaddecuq, vaddeuqm, vsubcuq, vsubuqm, vsubecuq, vsubeuqm.
3832 (altivec_resolve_overloaded_builtin): Add support for V1TImode.
3834 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Add support
3835 for V1TImode, and set up preferences to use VSX/Altivec registers.
3836 Setup VSX reload handlers.
3837 (rs6000_debug_reg_global): Likewise.
3838 (rs6000_init_hard_regno_mode_ok): Likewise.
3839 (rs6000_preferred_simd_mode): Likewise.
3840 (vspltis_constant): Do not allow V1TImode as easy altivec constants.
3841 (easy_altivec_constant): Likewise.
3842 (output_vec_const_move): Likewise.
3843 (rs6000_expand_vector_set): Convert V1TImode set and extract to
3845 (rs6000_expand_vector_extract): Likewise.
3846 (reg_offset_addressing_ok_p): Setup V1TImode to use VSX reg+reg
3848 (rs6000_const_vec): Add support for V1TImode.
3849 (rs6000_emit_le_vsx_load): Swap double words when loading or
3850 storing TImode/V1TImode.
3851 (rs6000_emit_le_vsx_store): Likewise.
3852 (rs6000_emit_le_vsx_move): Likewise.
3853 (rs6000_emit_move): Add support for V1TImode.
3854 (altivec_expand_ld_builtin): Likewise.
3855 (altivec_expand_st_builtin): Likewise.
3856 (altivec_expand_vec_init_builtin): Likewise.
3857 (altivec_expand_builtin): Likewise.
3858 (rs6000_init_builtins): Add support for V1TImode type. Add
3859 support for ISA 2.07 128-bit integer builtins. Define type names
3860 for the VSX/Altivec vector types.
3861 (altivec_init_builtins): Add support for overloaded vector
3862 functions with V1TImode type.
3863 (rs6000_preferred_reload_class): Prefer Altivec registers for V1TImode.
3864 (rs6000_move_128bit_ok_p): Move 128-bit move/split validation to
3866 (rs6000_split_128bit_ok_p): Likewise.
3867 (rs6000_handle_altivec_attribute): Create V1TImode from vector
3868 __int128_t and vector __uint128_t.
3870 * config/rs6000/vsx.md (VSX_L): Add V1TImode to vector iterators
3871 and mode attributes.
3878 (VS_scalar): Likewise.
3879 (VS_double): Likewise.
3880 (vsx_set_v1ti): New builtin function to create V1TImode from TImode.
3882 * config/rs6000/rs6000.h (TARGET_VADDUQM): New macro to say whether
3883 we support the ISA 2.07 128-bit integer arithmetic instructions.
3884 (ALTIVEC_OR_VSX_VECTOR_MODE): Add V1TImode.
3885 (enum rs6000_builtin_type_index): Add fields to hold V1TImode
3886 and TImode types for use with the builtin functions.
3887 (V1TI_type_node): Likewise.
3888 (unsigned_V1TI_type_node): Likewise.
3889 (intTI_type_internal_node): Likewise.
3890 (uintTI_type_internal_node): Likewise.
3892 * config/rs6000/altivec.md (UNSPEC_VADDCUQ): New unspecs for ISA 2.07
3893 128-bit builtin functions.
3894 (UNSPEC_VADDEUQM): Likewise.
3895 (UNSPEC_VADDECUQ): Likewise.
3896 (UNSPEC_VSUBCUQ): Likewise.
3897 (UNSPEC_VSUBEUQM): Likewise.
3898 (UNSPEC_VSUBECUQ): Likewise.
3899 (VM): Add V1TImode to vector mode iterators.
3901 (VI_unit): Likewise.
3902 (altivec_vadduqm): Add ISA 2.07 128-bit binary builtins.
3903 (altivec_vaddcuq): Likewise.
3904 (altivec_vsubuqm): Likewise.
3905 (altivec_vsubcuq): Likewise.
3906 (altivec_vaddeuqm): Likewise.
3907 (altivec_vaddecuq): Likewise.
3908 (altivec_vsubeuqm): Likewise.
3909 (altivec_vsubecuq): Likewise.
3911 * config/rs6000/rs6000.md (FMOVE128_GPR): Add V1TImode to vector
3913 (BOOL_128): Likewise.
3914 (BOOL_REGS_OUTPUT): Likewise.
3915 (BOOL_REGS_OP1): Likewise.
3916 (BOOL_REGS_OP2): Likewise.
3917 (BOOL_REGS_UNARY): Likewise.
3918 (BOOL_REGS_AND_CR0): Likewise.
3920 * config/rs6000/altivec.h (vec_vaddcuq): Add support for ISA 2.07
3921 128-bit integer builtin support.
3922 (vec_vadduqm): Likewise.
3923 (vec_vaddecuq): Likewise.
3924 (vec_vaddeuqm): Likewise.
3925 (vec_vsubecuq): Likewise.
3926 (vec_vsubeuqm): Likewise.
3927 (vec_vsubcuq): Likewise.
3928 (vec_vsubuqm): Likewise.
3930 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
3931 Document vec_vaddcuq, vec_vadduqm, vec_vaddecuq, vec_vaddeuqm,
3932 vec_subecuq, vec_subeuqm, vec_vsubcuq, vec_vsubeqm builtins adding
3933 128-bit integer add/subtract to ISA 2.07.
3935 2014-03-12 Joern Rennecke <joern.rennecke@embecosm.com>
3937 * config/arc/arc.c (arc_predicate_delay_insns):
3938 Fix third argument passed to conditionalize_nonjump.
3940 2014-03-12 Yufeng Zhang <yufeng.zhang@arm.com>
3942 * config/aarch64/aarch64-builtins.c
3943 (aarch64_builtin_vectorized_function): Add BUILT_IN_LFLOORF,
3944 BUILT_IN_LLFLOOR, BUILT_IN_LCEILF and BUILT_IN_LLCEIL.
3945 * config/aarch64/arm_neon.h (vcvtaq_u64_f64): Call __builtin_llfloor
3946 instead of __builtin_lfloor.
3947 (vcvtnq_u64_f64): Call __builtin_llceil instead of __builtin_lceil.
3949 2014-03-12 Jakub Jelinek <jakub@redhat.com>
3951 * tree-ssa-ifcombine.c (forwarder_block_to): New function.
3952 (tree_ssa_ifcombine_bb_1): New function.
3953 (tree_ssa_ifcombine_bb): Use it. Handle also cases where else_bb
3954 is an empty forwarder block to then_bb or vice versa and then_bb
3955 and else_bb are effectively swapped.
3957 2014-03-12 Christian Bruel <christian.bruel@st.com>
3960 * config/arm/arm.c (arm_emit_vfp_multi_reg_pop): Emit a
3961 REG_CFA_DEF_CFA note.
3962 (arm_expand_epilogue_apcs_frame): call arm_add_cfa_adjust_cfa_note.
3963 (arm_unwind_emit): Allow REG_CFA_DEF_CFA.
3965 2014-03-12 Thomas Preud'homme <thomas.preudhomme@arm.com>
3967 PR tree-optimization/60454
3968 * tree-ssa-math-opts.c (find_bswap_1): Fix bswap detection.
3970 2014-03-12 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3972 * config.gcc (aarch64*-*-*): Use ISA flags from aarch64-arches.def.
3973 Do not define target_cpu_default2 to generic.
3974 * config/aarch64/aarch64.h (TARGET_CPU_DEFAULT): Use generic cpu.
3975 * config/aarch64/aarch64.c (aarch64_override_options): Update comment.
3976 * config/aarch64/aarch64-arches.def (armv8-a): Use generic cpu.
3978 2014-03-12 Jakub Jelinek <jakub@redhat.com>
3979 Marc Glisse <marc.glisse@inria.fr>
3981 PR tree-optimization/60502
3982 * tree-ssa-reassoc.c (eliminate_not_pairs): Use build_all_ones_cst
3983 instead of build_low_bits_mask.
3985 2014-03-12 Jakub Jelinek <jakub@redhat.com>
3988 * tree-vrp.c (register_edge_assert_for_1): Don't add assert
3989 if there are multiple uses, but op doesn't live on E edge.
3990 * tree-cfg.c (assert_unreachable_fallthru_edge_p): Also ignore
3991 clobber stmts before __builtin_unreachable.
3993 2014-03-11 Richard Sandiford <rdsandiford@googlemail.com>
3995 * builtins.c (expand_builtin_setjmp_receiver): Use and clobber
3996 hard_frame_pointer_rtx.
3997 * cse.c (cse_insn): Remove volatile check.
3998 * cselib.c (cselib_process_insn): Likewise.
3999 * dse.c (scan_insn): Likewise.
4001 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
4003 * config/arc/arc.c (conditionalize_nonjump): New function,
4005 (arc_ifcvt): ... this.
4006 (arc_predicate_delay_insns): Use it.
4008 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
4010 * config/arc/predicates.md (extend_operand): During/after reload,
4011 allow const_int_operand.
4012 * config/arc/arc.md (mulsidi3_700): Use extend_operand predicate.
4013 (umulsidi3_700): Likewise. Change operand 2 constraint back to "cL".
4014 (mulsi3_highpart): Change operand 2 constraint alternatives 2 and 3
4016 (umulsi3_highpart_i): Likewise.
4018 2014-03-11 Richard Biener <rguenther@suse.de>
4020 * tree-ssa-structalias.c (get_constraint_for_ptr_offset):
4021 Add asserts to guard possible wrong-code bugs.
4023 2014-03-11 Richard Biener <rguenther@suse.de>
4025 PR tree-optimization/60429
4026 PR tree-optimization/60485
4027 * tree-ssa-structalias.c (set_union_with_increment): Properly
4028 take into account all fields that overlap the shifted vars.
4029 (do_sd_constraint): Likewise.
4030 (do_ds_constraint): Likewise.
4031 (get_constraint_for_ptr_offset): Likewise.
4033 2014-03-11 Chung-Lin Tang <cltang@codesourcery.com>
4035 * config/nios2/nios2.c (machine_function): Add fp_save_offset field.
4036 (nios2_compute_frame_layout):
4037 Add calculation of cfun->machine->fp_save_offset.
4038 (nios2_expand_prologue): Correct setting of frame pointer register
4040 (nios2_expand_epilogue): Update recovery of stack pointer from
4041 frame pointer accordingly.
4042 (nios2_initial_elimination_offset): Update calculation of offset
4043 for eliminating to HARD_FRAME_POINTER_REGNUM.
4045 2014-03-10 Jakub Jelinek <jakub@redhat.com>
4048 * ipa.c (symtab_remove_unreachable_nodes): Don't call
4049 cgraph_get_create_node on VAR_DECLs.
4051 2014-03-10 Richard Biener <rguenther@suse.de>
4054 * tree.c (signed_or_unsigned_type_for): Handle OFFSET_TYPEs.
4056 2014-03-08 Douglas B Rupp <rupp@gnat.com>
4058 * config/vms/vms.opt (vms_float_format): New variable.
4060 2014-03-08 Tobias Burnus <burnus@net-b.de>
4062 * doc/invoke.texi (-fcilkplus): Update implementation status.
4064 2014-03-08 Paulo Matos <paulo@matos-sorge.com>
4065 Richard Biener <rguenther@suse.de>
4067 * lto-wrapper.c (merge_and_complain): Ensure -fshort-double is used
4068 consistently accross all TUs.
4069 (run_gcc): Enable -fshort-double automatically at link at link-time
4070 and disallow override.
4072 2014-03-08 Richard Sandiford <rdsandiford@googlemail.com>
4075 * config/mips/mips.c (mips_option_override): Promote -mpaired-single
4076 warning to an error. Disable TARGET_PAIRED_SINGLE and TARGET_MIPS3D
4077 if they can't be used.
4079 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4081 * configure.ac (HAVE_AS_IX86_TLSLDMPLT): Improve test
4082 for Solaris 11/x86 ld.
4083 * configure: Regenerate.
4085 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4087 * configure.ac (TLS_SECTION_ASM_FLAG): Save as tls_section_flag.
4088 (LIB_TLS_SPEC): Save as ld_tls_libs.
4089 (HAVE_AS_IX86_TLSLDMPLT): Define as 1/0.
4090 (HAVE_AS_IX86_TLSLDM): New test.
4091 * configure, config.in: Regenerate.
4092 * config/i386/i386.c (legitimize_tls_address): Fall back to
4093 TLS_MODEL_GLOBAL_DYNAMIC on 32-bit Solaris/x86 if tool chain
4094 cannot support TLS_MODEL_LOCAL_DYNAMIC.
4095 * config/i386/i386.md (*tls_local_dynamic_base_32_gnu): Use if
4096 instead of #ifdef in HAVE_AS_IX86_TLSLDMPLT test.
4098 2014-03-07 Paulo Matos <paulo@matos-sorge.com>
4100 * common.opt (fira-loop-pressure): Mark as optimization.
4102 2014-03-07 Thomas Schwinge <thomas@codesourcery.com>
4104 * langhooks.c (lhd_omp_mappable_type): The error_mark_node is not
4105 an OpenMP mappable type.
4107 2014-03-06 Matthias Klose <doko@ubuntu.com>
4109 * Makefile.in (s-mlib): Only pass MULTIARCH_DIRNAME if
4110 MULTILIB_OSDIRNAMES is not defined.
4112 2014-03-06 Jakub Jelinek <jakub@redhat.com>
4113 Meador Inge <meadori@codesourcery.com>
4116 * config/arm/arm.c (arm_tls_symbol_p): Remove.
4117 (arm_legitimize_address): Call legitimize_tls_address for any
4118 arm_tls_referenced_p expression, handle constant addend. Call it
4119 before testing for !TARGET_ARM.
4120 (thumb_legitimize_address): Don't handle arm_tls_symbol_p here.
4122 2014-03-06 Richard Biener <rguenther@suse.de>
4128 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
4130 * tree-streamer.c (record_common_node): Assert we don't record
4131 nodes with type double.
4132 (preload_common_node): Skip type double, complex double and double
4133 pointer since it is now frontend dependent due to fshort-double option.
4135 2014-03-06 Richard Biener <rguenther@suse.de>
4137 * gcc.c (PLUGIN_COND): Always enable unless -fno-use-linker-plugin
4138 or -fno-lto is specified and the linker has full plugin support.
4139 * collect2.c (lto_mode): Default to LTO_MODE_WHOPR if LTO is enabled.
4140 (main): Remove -flto processing, adjust lto_mode using use_plugin late.
4141 * lto-wrapper.c (merge_and_complain): Merge compile-time
4142 optimization levels.
4143 (run_gcc): And pass it through to the link options.
4145 2014-03-06 Alexandre Oliva <aoliva@redhat.com>
4149 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4151 * cselib.c (remove_useless_values): Skip to avoid quadratic
4152 behavior if the condition moved from...
4153 (cselib_process_insn): ... here holds.
4155 2014-03-05 Jakub Jelinek <jakub@redhat.com>
4158 * Makefile.in (PLUGIN_HEADERS): Add tree-phinodes.h, stor-layout.h,
4159 ssa-iterators.h, $(RESOURCE_H) and tree-cfgcleanup.h.
4162 * config/i386/t-i386 (OPTIONS_H_EXTRA): Add stringop.def.
4163 (TM_H): Add x86-tune.def.
4165 2014-03-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4167 * config/aarch64/aarch64.c (generic_tunings):
4168 Use cortexa57_extra_costs.
4170 2014-03-05 Jakub Jelinek <jakub@redhat.com>
4173 * cfgexpand.c (expand_used_vars): Do not assume all SSA_NAMEs
4174 of PARM/RESULT_DECLs must be coalesced with optimize && in_lto_p.
4175 * tree-ssa-coalesce.c (coalesce_ssa_name): Use MUST_COALESCE_COST - 1
4178 2014-03-04 Heiher <r@hev.cc>
4180 * config/mips/mips-cpus.def (loongson3a): Mark as a MIPS64r2 processor.
4181 * config/mips/mips.h (MIPS_ISA_LEVEL_SPEC): Adjust accordingly.
4183 2014-03-04 Uros Bizjak <ubizjak@gmail.com>
4185 * config/i386/predicates.md (const2356_operand): Change to ...
4186 (const2367_operand): ... this.
4187 * config/i386/sse.md (avx512pf_scatterpf<mode>sf): Use
4189 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
4190 (*avx512pf_scatterpf<mode>sf): Ditto.
4191 (avx512pf_scatterpf<mode>df): Ditto.
4192 (*avx512pf_scatterpf<mode>df_mask): Ditto.
4193 (*avx512pf_scatterpf<mode>df): Ditto.
4194 * config/i386/i386.c (ix86_expand_builtin): Update
4195 incorrect hint operand error message.
4197 2014-03-04 Richard Biener <rguenther@suse.de>
4199 * lto-section-in.c (lto_get_section_data): Fix const cast.
4201 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
4203 * tree-streamer.c (record_common_node): Assert we don't record
4204 nodes with type double.
4205 (preload_common_node): Skip type double, complex double and double
4206 pointer since it is now frontend dependent due to fshort-double option.
4208 2014-03-04 Richard Biener <rguenther@suse.de>
4211 * lto-streamer-in.c (lto_read_body): Remove LTO bytecode version check.
4212 (lto_input_toplevel_asms): Likewise.
4213 * lto-section-in.c (lto_get_section_data): Instead do it here
4216 2014-03-04 Richard Biener <rguenther@suse.de>
4218 PR tree-optimization/60382
4219 * tree-vect-loop.c (vect_is_simple_reduction_1): Do not consider
4220 dead PHIs a reduction.
4222 2014-03-03 Uros Bizjak <ubizjak@gmail.com>
4224 * config/i386/xmmintrin.h (enum _mm_hint) <_MM_HINT_ET0>: Correct
4226 (_mm_prefetch): Move out of GCC target("sse") pragma.
4227 * config/i386/prfchwintrin.h (_m_prefetchw): Move out of
4228 GCC target("prfchw") pragma.
4229 * config/i386/i386.md (prefetch): Emit prefetchwt1 only
4231 * config/i386/i386.c (ix86_option_override_internal): Enable
4232 -mprfchw with -mprefetchwt1.
4234 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
4236 * config/arc/arc.md (casesi_load) <length attribute alternative 0>:
4239 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
4241 * opts.h (CL_PCH_IGNORE): Define.
4242 * targhooks.c (option_affects_pch_p):
4243 Return false for options that have CL_PCH_IGNORE set.
4244 * opt-functions.awk: Process PchIgnore.
4245 * doc/options.texi: Document PchIgnore.
4247 * config/arc/arc.opt (misize): Add PchIgnore property.
4249 2014-03-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4251 * config/rs6000/rs6000.c (rs6000_preferred_reload_class): Disallow
4252 reload of PLUS rtx's outside of GENERAL_REGS or BASE_REGS; relax
4253 constraint on constants to permit them being loaded into
4254 GENERAL_REGS or BASE_REGS.
4256 2014-03-03 Nick Clifton <nickc@redhat.com>
4258 * config/rl78/rl78-real.md (cbranchsi4_real_signed): Add
4259 anti-cacnonical alternatives.
4260 (negandhi3_real): New pattern.
4261 * config/rl78/rl78-virt.md (negandhi3_virt): New pattern.
4263 2014-03-03 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
4265 * config/avr/avr-mcus.def: Remove atxmega16x1.
4266 * config/avr/avr-tables.opt: Regenerate.
4267 * config/avr/t-multilib: Regenerate.
4268 * doc/avr-mmcu.texi: Regenerate.
4270 2014-03-03 Tobias Grosser <tobias@grosser.es>
4271 Mircea Namolaru <mircea.namolaru@inria.fr>
4273 PR tree-optimization/58028
4274 * graphite-clast-to-gimple.c (set_cloog_options): Don't remove
4277 2014-03-03 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4279 * config/arm/neon.md (*movmisalign<mode>): Legitimize addresses
4280 not handled by recognizers.
4282 2014-03-03 Jakub Jelinek <jakub@redhat.com>
4285 * function.c (expand_function_end): Don't emit
4286 clobber_return_register sequence if clobber_after is a BARRIER.
4287 * cfgexpand.c (construct_exit_block): Append instructions before
4288 return_label to prev_bb.
4290 2014-03-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4292 * config/rs6000/constraints.md: Document reserved use of "wc".
4294 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
4297 * ipa.c (function_and_variable_visibility): When dissolving comdat
4298 group, also set all symbols to local.
4300 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
4305 2013-12-14 Jan Hubicka <jh@suse.cz>
4307 * ipa-prop.c (stmt_may_be_vtbl_ptr_store): Skip clobbers.
4309 2014-03-02 Jon Beniston <jon@beniston.com>
4315 * config/lm32/lm32.c (lm32_legitimate_constant_p): Remove, as incorrect.
4316 (TARGET_LEGITIMATE_CONSTANT_P): Undefine, as not needed.
4317 * config/lm32/lm32.md (movsi_insn): Add 32-bit immediate support.
4318 (simple_return, *simple_return): New patterns
4319 * config/lm32/predicates.md (movsi_rhs_operand): Remove as obsolete.
4320 * configure.ac (force_sjlj_exceptions): Force sjlj exceptions for lm32.
4322 2014-03-01 Paolo Carlini <paolo.carlini@oracle.com>
4324 * dwarf2out.c (gen_subprogram_die): Tidy.
4326 2014-03-01 Oleg Endo <olegendo@gcc.gnu.org>
4329 * config/sh/sh.md (*mov_t_msb_neg): Split into ...
4330 (*mov_t_msb_neg_negc): ... this new insn.
4332 2014-02-28 Jason Merrill <jason@redhat.com>
4335 * ipa-devirt.c (ipa_devirt): Don't choose an implicitly-declared
4338 2014-02-28 Paolo Carlini <paolo.carlini@oracle.com>
4341 * dwarf2out.c (decltype_auto_die): New static.
4342 (gen_subprogram_die): Handle 'decltype(auto)' like 'auto'.
4343 (gen_type_die_with_usage): Handle 'decltype(auto)'.
4344 (is_cxx_auto): Likewise.
4346 2014-02-28 Ian Bolton <ian.bolton@arm.com>
4348 * config/aarch64/aarch64.h: Define __ARM_NEON by default if
4349 we are not using general regs only.
4351 2014-02-28 Richard Biener <rguenther@suse.de>
4354 * tree-cfgcleanup.c (tree_forwarder_block_p): Restrict
4355 previous fix and only allow to remove trivial pre-headers
4356 and latches. Also honor LOOPS_MAY_HAVE_MULTIPLE_LATCHES.
4357 (remove_forwarder_block): Properly update the latch of a loop.
4359 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4362 * cselib.c (cselib_hasher::equal): Special-case VALUE lookup.
4363 (cselib_preserved_hash_table): New.
4364 (preserve_constants_and_equivs): Move preserved vals to it.
4365 (cselib_find_slot): Look it up first.
4366 (cselib_init): Initialize it.
4367 (cselib_finish): Release it.
4368 (dump_cselib_table): Dump it.
4370 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4373 * cselib.c (remove_useless_values): Skip to avoid quadratic
4374 behavior if the condition moved from...
4375 (cselib_process_insn): ... here holds.
4377 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4380 * var-tracking.c (vt_initialize): Apply the same condition to
4381 preserve the CFA base value.
4383 2014-02-28 Joey Ye <joey.ye@arm.com>
4386 * config/arm/arm.c (thumb_far_jump_used_p): Don't change
4387 if reload in progress or completed.
4389 2014-02-28 Tobias Burnus <burnus@net-b.de>
4392 * tree-pretty-print.c (dump_generic_node, print_declaration): Handle
4395 2014-02-27 H.J. Lu <hongjiu.lu@intel.com>
4397 * doc/tm.texi.in (Condition Code Status): Update documention for
4398 relative locations of cc0-setter and cc0-user.
4400 2014-02-27 Jeff Law <law@redhat.com>
4402 PR rtl-optimization/52714
4403 * combine.c (try_combine): When splitting an unrecognized PARALLEL
4404 into two independent simple sets, if I3 is a jump, ensure the
4405 pattern we place into I3 is a (set (pc) ...).
4407 2014-02-27 Mikael Pettersson <mikpe@it.uu.se>
4408 Jeff Law <law@redhat.com>
4410 PR rtl-optimization/49847
4411 * cse.c (fold_rtx) Handle case where cc0 setter and cc0 user
4412 are in different blocks.
4413 * doc/tm.texi (Condition Code Status): Update documention for
4414 relative locations of cc0-setter and cc0-user.
4416 2014-02-27 Vladimir Makarov <vmakarov@redhat.com>
4419 * lra.c (lra_emit_add): Check SUBREG too.
4421 2014-02-27 Andreas Schwab <schwab@suse.de>
4423 * config/m68k/m68k.c (m68k_option_override): Disable
4424 -flive-range-shrinkage for classic m68k.
4425 (m68k_override_options_after_change): Likewise.
4427 2014-02-27 Marek Polacek <polacek@redhat.com>
4430 * tree-ssa-uninit.c (gate_warn_uninitialized): Run the pass even for
4431 -Wmaybe-uninitialized.
4433 2014-02-27 Alan Modra <amodra@gmail.com>
4436 * reload1.c (emit_input_reload_insns): When reload_override_in,
4437 set old to rl->in_reg when rl->in_reg is a subreg.
4439 2014-02-26 Richard Biener <rguenther@suse.de>
4442 * lra-assigns.c (spill_for): Avoid mixed-sign comparison.
4444 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
4446 * common/config/i386/predicates.md (const1256_operand): Remove.
4447 (const2356_operand): New.
4448 (const_1_to_2_operand): Remove.
4449 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
4450 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
4451 (*avx512pf_gatherpf<mode>sf): Ditto.
4452 (avx512pf_gatherpf<mode>df): Ditto.
4453 (*avx512pf_gatherpf<mode>df_mask): Ditto.
4454 (*avx512pf_gatherpf<mode>df): Ditto.
4455 (avx512pf_scatterpf<mode>sf): Ditto.
4456 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
4457 (*avx512pf_scatterpf<mode>sf): Ditto.
4458 (avx512pf_scatterpf<mode>df): Ditto.
4459 (*avx512pf_scatterpf<mode>df_mask): Ditto.
4460 (*avx512pf_scatterpf<mode>df): Ditto.
4461 * common/config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET0.
4463 2014-02-26 Ilya Tocar <ilya.tocar@intel.com>
4465 * config/i386/avx512fintrin.h (_mm512_testn_epi32_mask),
4466 (_mm512_mask_testn_epi32_mask), (_mm512_testn_epi64_mask),
4467 (_mm512_mask_testn_epi64_mask): Move to ...
4468 * config/i386/avx512cdintrin.h: Here.
4469 * config/i386/i386.c (bdesc_args): Change MASK_ISA for testnm.
4470 * config/i386/sse.md (avx512f_vmscalef<mode><round_name>): Remove %.
4471 (avx512f_scalef<mode><mask_name><round_name>): Ditto.
4472 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Change conditon to
4473 TARGET_AVX512F from TARGET_AVX512CD.
4475 2014-02-26 Richard Biener <rguenther@suse.de>
4478 * ipa.c (walk_polymorphic_call_targets): Properly guard
4479 call to inline_update_overall_summary.
4481 2014-02-26 Bin Cheng <bin.cheng@arm.com>
4484 * tree-cfgcleanup.c (tree_forwarder_block_p): Protect loop preheaders
4485 and latches only if requested. Fix latch if it is removed.
4486 * tree-ssa-dom.c (tree_ssa_dominator_optimize): Set
4487 LOOPS_HAVE_PREHEADERS.
4489 2014-02-25 Andrew Pinski <apinski@cavium.com>
4491 * builtins.c (expand_builtin_thread_pointer): Create a new target
4492 when the target is NULL.
4494 2014-02-25 Vladimir Makarov <vmakarov@redhat.com>
4496 PR rtl-optimization/60317
4497 * params.def (PARAM_LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
4498 * params.h (LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
4499 * lra-assigns.c: Include params.h.
4500 (spill_for): Use LRA_MAX_CONSIDERED_RELOAD_PSEUDOS as guard for
4501 other reload pseudos considerations.
4503 2014-02-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4505 * config/rs6000/vector.md (*vector_unordered<mode>): Change split
4506 to use canonical form for nor<mode>3.
4508 2014-02-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4511 * config/arm/arm.h (CANNOT_CHANGE_MODE_CLASS): Allow 128 to 64-bit
4514 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
4516 * common/config/i386/i386-common.c (OPTION_MASK_ISA_PREFETCHWT1_SET),
4517 (OPTION_MASK_ISA_PREFETCHWT1_UNSET): New.
4518 (ix86_handle_option): Handle OPT_mprefetchwt1.
4519 * config/i386/cpuid.h (bit_PREFETCHWT1): New.
4520 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
4522 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
4523 OPTION_MASK_ISA_PREFETCHWT1.
4524 * config/i386/i386.c (ix86_target_string): Handle mprefetchwt1.
4525 (PTA_PREFETCHWT1): New.
4526 (ix86_option_override_internal): Handle PTA_PREFETCHWT1.
4527 (ix86_valid_target_attribute_inner_p): Handle OPT_mprefetchwt1.
4528 * config/i386/i386.h (TARGET_PREFETCHWT1, TARGET_PREFETCHWT1_P): New.
4529 * config/i386/i386.md (prefetch): Check TARGET_PREFETCHWT1
4530 (*prefetch_avx512pf_<mode>_: Change into ...
4531 (*prefetch_prefetchwt1_<mode>: This.
4532 * config/i386/i386.opt (mprefetchwt1): New.
4533 * config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET1.
4534 (_mm_prefetch): Handle intent to write.
4535 * doc/invoke.texi (mprefetchwt1), (mno-prefetchwt1): Doccument.
4537 2014-02-25 Richard Biener <rguenther@suse.de>
4540 * emit-rtl.c (mem_attrs_htab): Remove.
4541 (mem_attrs_htab_hash): Likewise.
4542 (mem_attrs_htab_eq): Likewise.
4543 (set_mem_attrs): Always allocate new mem-attrs when something changed.
4544 (init_emit_once): Do not allocate mem_attrs_htab.
4546 2014-02-25 Richard Biener <rguenther@suse.de>
4549 * lto-opts.c (lto_write_options): Output non-explicit conservative
4550 -fwrapv, -fno-trapv and -fno-strict-overflow.
4551 * lto-wrapper.c (merge_and_complain): Handle merging those options.
4552 (run_gcc): And pass them through.
4554 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
4556 * sel-sched.c (calculate_new_fences): New parameter ptime.
4557 Calculate it as a maximum over all fence cycles.
4558 (sel_sched_region_2): Adjust the call to calculate_new_fences.
4559 Print the final schedule timing when sched_verbose.
4561 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
4563 PR rtl-optimization/60292
4564 * sel-sched.c (fill_vec_av_set): Do not reset target availability
4565 bit fot the fence instruction.
4567 2014-02-24 Alangi Derick <alangiderick@gmail.com>
4569 * calls.h: Fix typo in comment.
4571 2014-02-24 John David Anglin <danglin@gcc.gnu.org>
4573 * config/pa/pa.c (pa_output_move_double): Don't valididate when
4574 adjusting offsetable addresses.
4576 2014-02-24 Guozhi Wei <carrot@google.com>
4578 * sparseset.h (sparseset_pop): Fix the wrong index.
4580 2014-02-24 Walter Lee <walt@tilera.com>
4582 * config.gcc (tilepro-*-*): Change to tilepro*-*-*.
4583 (tilegx-*-linux*): Change to tilegx*-*-linux*; Support tilegxbe
4585 * common/config/tilegx/tilegx-common.c
4586 (TARGET_DEFAULT_TARGET_FLAGS): Define.
4587 * config/tilegx/linux.h (ASM_SPEC): Add endian_spec.
4589 * config/tilegx/sync.md (atomic_test_and_set): Handle big endian.
4590 * config/tilegx/tilegx.c (tilegx_return_in_msb): New.
4591 (tilegx_gimplify_va_arg_expr): Handle big endian.
4592 (tilegx_expand_unaligned_load): Ditto.
4593 (tilegx_expand_unaligned_store): Ditto.
4594 (TARGET_RETURN_IN_MSB): New.
4595 * config/tilegx/tilegx.h (TARGET_DEFAULT): New.
4596 (TARGET_ENDIAN_DEFAULT): New.
4597 (TARGET_BIG_ENDIAN): Handle big endian.
4598 (BYTES_BIG_ENDIAN): Ditto.
4599 (WORDS_BIG_ENDIAN): Ditto.
4600 (FLOAT_WORDS_BIG_ENDIAN): Ditto.
4603 * config/tilegx/tilegx.md (extv): Handle big endian.
4605 (insn_st<n>): Ditto.
4606 (insn_st<n>_add<bitsuffix>): Ditto.
4607 (insn_stnt<n>): Ditto.
4608 (insn_stnt<n>_add<bitsuffix>):Ditto.
4609 (vec_interleave_highv8qi): Handle big endian.
4610 (vec_interleave_highv8qi_be): New.
4611 (vec_interleave_highv8qi_le): New.
4612 (insn_v1int_h): Handle big endian.
4613 (vec_interleave_lowv8qi): Handle big endian.
4614 (vec_interleave_lowv8qi_be): New.
4615 (vec_interleave_lowv8qi_le): New.
4616 (insn_v1int_l): Handle big endian.
4617 (vec_interleave_highv4hi): Handle big endian.
4618 (vec_interleave_highv4hi_be): New.
4619 (vec_interleave_highv4hi_le): New.
4620 (insn_v2int_h): Handle big endian.
4621 (vec_interleave_lowv4hi): Handle big endian.
4622 (vec_interleave_lowv4hi_be): New.
4623 (vec_interleave_lowv4hi_le): New.
4624 (insn_v2int_l): Handle big endian.
4625 (vec_interleave_highv2si): Handle big endian.
4626 (vec_interleave_highv2si_be): New.
4627 (vec_interleave_highv2si_le): New.
4628 (insn_v4int_h): Handle big endian.
4629 (vec_interleave_lowv2si): Handle big endian.
4630 (vec_interleave_lowv2si_be): New.
4631 (vec_interleave_lowv2si_le): New.
4632 (insn_v4int_l): Handle big endian.
4633 * config/tilegx/tilegx.opt (mbig-endian): New option.
4634 (mlittle-endian): New option.
4635 * doc/install.texi: Document tilegxbe-linux.
4636 * doc/invoke.texi: Document -mbig-endian and -mlittle-endian.
4638 2014-02-24 Martin Jambor <mjambor@suse.cz>
4641 * ipa-cp.c (propagate_constants_accross_call): Bail out early if
4642 there are no parameter descriptors.
4644 2014-02-24 Andrey Belevantsev <abel@ispras.ru>
4646 PR rtl-optimization/60268
4647 * sched-rgn.c (haifa_find_rgns): Move the nr_regions_initial variable
4648 initialization to ...
4649 (sched_rgn_init): ... here.
4650 (schedule_region): Check for SCHED_PRESSURE_NONE earlier.
4652 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
4654 * config/microblaze/microblaze.md: Correct ashrsi_reg / lshrsi_reg
4657 2014-02-23 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
4659 * config/microblaze/microblaze.h: Remove SECONDARY_MEMORY_NEEDED
4662 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
4664 * /config/microblaze/microblaze.c: Add microblaze_asm_output_mi_thunk,
4665 define TARGET_ASM_OUTPUT_MI_THUNK and TARGET_ASM_CAN_OUTPUT_MI_THUNK.
4667 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
4669 * config/microblaze/predicates.md: Add cmp_op predicate.
4670 * config/microblaze/microblaze.md: Add branch_compare instruction
4671 which uses cmp_op predicate and emits cmp insn before branch.
4672 * config/microblaze/microblaze.c (microblaze_emit_compare): Rename
4673 to microblaze_expand_conditional_branch and consolidate logic.
4674 (microblaze_expand_conditional_branch): emit branch_compare
4675 insn instead of handling cmp op separate from branch insn.
4677 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4679 * config/rs6000/rs6000.c (rs6000_emit_le_vsx_move): Relax assert
4682 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4684 * config/rs6000/altivec.md (altivec_lve<VI_char>x): Replace
4685 define_insn with define_expand and new define_insn
4686 *altivec_lve<VI_char>x_internal.
4687 (altivec_stve<VI_char>x): Replace define_insn with define_expand
4688 and new define_insn *altivec_stve<VI_char>x_internal.
4689 * config/rs6000/rs6000-protos.h (altivec_expand_stvex_be): New
4691 * config/rs6000/rs6000.c (altivec_expand_lvx_be): Document use by
4693 (altivec_expand_stvex_be): New function.
4695 2014-02-22 Joern Rennecke <joern.rennecke@embecosm.com>
4697 * config/avr/avr.c (avr_can_eliminate): Allow elimination from
4698 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM if !frame_pointer_needed.
4699 * config/avr/avr.c (ELIMINABLE_REGS): Add elimination from
4700 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM.
4702 2014-02-21 Vladimir Makarov <vmakarov@redhat.com>
4705 * lra-constraints.c (inherit_reload_reg): Use lra_emit_move
4706 instead of emit_move_insn.
4708 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4710 * config/rs6000/altivec.md (altivec_vsumsws): Replace second
4712 (reduc_uplus_v16qi): Use gen_altivec_vsumsws_direct instead of
4713 gen_altivec_vsumsws.
4715 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4717 * config/rs6000/altivec.md (altivec_lvxl): Rename as
4718 *altivec_lvxl_<mode>_internal and use VM2 iterator instead of V4SI.
4719 (altivec_lvxl_<mode>): New define_expand incorporating
4720 -maltivec=be semantics where needed.
4721 (altivec_lvx): Rename as *altivec_lvx_<mode>_internal.
4722 (altivec_lvx_<mode>): New define_expand incorporating -maltivec=be
4723 semantics where needed.
4724 (altivec_stvx): Rename as *altivec_stvx_<mode>_internal.
4725 (altivec_stvx_<mode>): New define_expand incorporating
4726 -maltivec=be semantics where needed.
4727 (altivec_stvxl): Rename as *altivec_stvxl_<mode>_internal and use
4728 VM2 iterator instead of V4SI.
4729 (altivec_stvxl_<mode>): New define_expand incorporating
4730 -maltivec=be semantics where needed.
4731 * config/rs6000/rs6000-builtin.def: Add new built-in definitions
4732 LVXL_V2DF, LVXL_V2DI, LVXL_V4SF, LVXL_V4SI, LVXL_V8HI, LVXL_V16QI,
4733 LVX_V2DF, LVX_V2DI, LVX_V4SF, LVX_V4SI, LVX_V8HI, LVX_V16QI, STVX_V2DF,
4734 STVX_V2DI, STVX_V4SF, STVX_V4SI, STVX_V8HI, STVX_V16QI, STVXL_V2DF,
4735 STVXL_V2DI, STVXL_V4SF, STVXL_V4SI, STVXL_V8HI, STVXL_V16QI.
4736 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Replace
4737 ALTIVEC_BUILTIN_LVX with ALTIVEC_BUILTIN_LVX_<MODE> throughout;
4738 similarly for ALTIVEC_BUILTIN_LVXL, ALTIVEC_BUILTIN_STVX, and
4739 ALTIVEC_BUILTIN_STVXL.
4740 * config/rs6000/rs6000-protos.h (altivec_expand_lvx_be): New prototype.
4741 (altivec_expand_stvx_be): Likewise.
4742 * config/rs6000/rs6000.c (swap_selector_for_mode): New function.
4743 (altivec_expand_lvx_be): Likewise.
4744 (altivec_expand_stvx_be): Likewise.
4745 (altivec_expand_builtin): Add cases for
4746 ALTIVEC_BUILTIN_STVX_<MODE>, ALTIVEC_BUILTIN_STVXL_<MODE>,
4747 ALTIVEC_BUILTIN_LVXL_<MODE>, and ALTIVEC_BUILTIN_LVX_<MODE>.
4748 (altivec_init_builtins): Add definitions for
4749 __builtin_altivec_lvxl_<mode>, __builtin_altivec_lvx_<mode>,
4750 __builtin_altivec_stvx_<mode>, and __builtin_altivec_stvxl_<mode>.
4752 2014-02-21 Catherine Moore <clm@codesourcery.com>
4754 * doc/invoke.texi (mvirt, mno-virt): Document.
4755 * config/mips/mips.opt (mvirt): New option.
4756 * config/mips/mips.h (ASM_SPEC): Pass mvirt to the assembler.
4758 2014-02-21 Richard Biener <rguenther@suse.de>
4760 PR tree-optimization/60276
4761 * tree-vectorizer.h (struct _stmt_vec_info): Add min_neg_dist field.
4762 (STMT_VINFO_MIN_NEG_DIST): New macro.
4763 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Record
4764 STMT_VINFO_MIN_NEG_DIST.
4765 * tree-vect-stmts.c (vectorizable_load): Verify if assumptions
4766 made for negative dependence distances still hold.
4768 2014-02-21 Richard Biener <rguenther@suse.de>
4771 * tree-ssa-live.c (mark_all_vars_used_1): Do not walk
4772 DECL_INITIAL for globals not in the current function context.
4774 2014-02-21 Jakub Jelinek <jakub@redhat.com>
4776 PR tree-optimization/56490
4777 * params.def (PARAM_UNINIT_CONTROL_DEP_ATTEMPTS): New param.
4778 * tree-ssa-uninit.c: Include params.h.
4779 (compute_control_dep_chain): Add num_calls argument, return false
4780 if it exceed PARAM_UNINIT_CONTROL_DEP_ATTEMPTS param, pass
4781 num_calls to recursive call.
4782 (find_predicates): Change dep_chain into normal array,
4783 cur_chain into auto_vec<edge, MAX_CHAIN_LEN + 1>, add num_calls
4784 variable and adjust compute_control_dep_chain caller.
4785 (find_def_preds): Likewise.
4787 2014-02-21 Thomas Schwinge <thomas@codesourcery.com>
4789 * gimple-pretty-print.c (dump_gimple_omp_for) [flags & TDF_RAW]
4790 <case GF_OMP_FOR_KIND_CILKSIMD>: Add missing break statement.
4792 2014-02-21 Nick Clifton <nickc@redhat.com>
4794 * config/stormy16/stormy16.md (pushdqi1): Add mode to post_inc.
4795 (pushhi1): Likewise.
4796 (popqi1): Add mode to pre_dec.
4799 2014-02-21 Jakub Jelinek <jakub@redhat.com>
4801 * config/i386/i386.c (ix86_expand_vec_perm): Use V8SImode
4802 mode for mask of V8SFmode permutation.
4804 2014-02-20 Richard Henderson <rth@redhat.com>
4807 * builtins.c (expand_builtin_atomic_compare_exchange): Always make
4808 a new pseudo for OLDVAL.
4810 2014-02-20 Jakub Jelinek <jakub@redhat.com>
4813 * config/i386/i386.c (expand_vec_perm_interleave2): Don't call
4814 gen_reg_rtx if d->testing_p.
4815 (expand_vec_perm_pshufb2, expand_vec_perm_broadcast_1): Return early
4816 if d->testing_p and we will certainly return true.
4817 (expand_vec_perm_even_odd_1): Likewise. Don't call gen_reg_rtx
4820 2014-02-20 Uros Bizjak <ubizjak@gmail.com>
4822 * emit-rtl.c (gen_reg_rtx): Assert that
4823 crtl->emit.regno_pointer_align_length is non-zero.
4825 2014-02-20 Richard Henderson <rth@redhat.com>
4828 * builtins.c (expand_builtin_atomic_compare_exchange): Conditionalize
4829 on failure the store back into EXPECT.
4831 2014-02-20 Chung-Lin Tang <cltang@codesourcery.com>
4832 Sandra Loosemore <sandra@codesourcery.com>
4834 * config/nios2/nios2.md (unspec): Add UNSPEC_PIC_GOTOFF_SYM enum.
4835 * config/nios2/nios2.c (nios2_function_profiler): Add
4836 -fPIC (flag_pic == 2) support.
4837 (nios2_handle_custom_fpu_cfg): Fix warning parameter.
4838 (nios2_large_offset_p): New function.
4839 (nios2_unspec_reloc_p): Move up position, update to use
4840 nios2_large_offset_p.
4841 (nios2_unspec_address): Remove function.
4842 (nios2_unspec_offset): New function.
4843 (nios2_large_got_address): New function.
4844 (nios2_got_address): Add large offset support.
4845 (nios2_legitimize_tls_address): Update usage of removed and new
4847 (nios2_symbol_binds_local_p): New function.
4848 (nios2_load_pic_address): Add -fPIC (flag_pic == 2) support.
4849 (nios2_legitimize_address): Update to use nios2_large_offset_p.
4850 (nios2_emit_move_sequence): Avoid legitimizing (const (unspec ...)).
4851 (nios2_print_operand): Merge H/L processing, add hiadj/lo
4852 processing for (const (unspec ...)).
4853 (nios2_unspec_reloc_name): Add UNSPEC_PIC_GOTOFF_SYM case.
4855 2014-02-20 Richard Biener <rguenther@suse.de>
4857 * tree-cfg.c (replace_uses_by): Mark altered BBs before
4858 doing the substitution.
4859 (verify_gimple_assign_single): Also verify bare MEM_REFs on the lhs.
4861 2014-02-20 Martin Jambor <mjambor@suse.cz>
4864 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Uce correct
4865 info when checking whether lattices are bottom.
4867 2014-02-20 Richard Biener <rguenther@suse.de>
4870 * tree-eh.c (execute_cleanup_eh_1): Also cleanup empty EH
4873 2014-02-20 Jan Hubicka <hubicka@ucw.cz>
4876 * ipa-inline-transform.c (clone_inlined_nodes): Add freq_scale
4877 parameter specifying the scaling.
4878 (inline_call): Update.
4879 (want_inline_recursively): Guard division by zero.
4880 (recursive_inlining): Update.
4881 * ipa-inline.h (clone_inlined_nodes): Update.
4883 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
4886 * config/i386/i386.c (classify_argument): Pass structures of size
4887 64 bytes or less in register.
4889 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
4890 Kirill Yukhin <kirill.yukhin@intel.com>
4892 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): Swap operands.
4893 (_mm_rcp28_round_ss): Ditto.
4894 (_mm_rsqrt28_round_sd): Ditto.
4895 (_mm_rsqrt28_round_ss): Ditto.
4896 * config/i386/avx512erintrin.h (_mm_rcp14_round_sd): Ditto.
4897 (_mm_rcp14_round_ss): Ditto.
4898 (_mm_rsqrt14_round_sd): Ditto.
4899 (_mm_rsqrt14_round_ss): Ditto.
4900 * config/i386/sse.md (rsqrt14<mode>): Put nonimmediate operand as
4901 the first input operand, get rid of match_dup.
4902 (avx512er_exp2<mode><mask_name><round_saeonly_name>): Set type
4904 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
4906 (avx512er_vmrcp28<mode><round_saeonly_name>): Put nonimmediate
4907 operand as the first input operand, set type attribute.
4908 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
4910 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Put nonimmediate
4911 operand as the first input operand, set type attribute.
4913 2014-02-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4915 * config/rs6000/rs6000.c (vspltis_constant): Fix most significant
4918 2014-02-19 H.J. Lu <hongjiu.lu@intel.com>
4921 * config/i386/i386.c (construct_container): Remove TFmode check
4922 for X86_64_INTEGER_CLASS.
4924 2014-02-19 Uros Bizjak <ubizjak@gmail.com>
4927 * config/i386/i386.c (type_natural_mode): Warn for ABI changes
4928 only when -Wpsabi is enabled.
4930 2014-02-19 Michael Hudson-Doyle <michael.hudson@linaro.org>
4933 * config/aarch64/aarch64.c (aarch64_pass_by_reference): The rules for
4934 passing arrays in registers are the same as for structs, so remove the
4935 special case for them.
4937 2014-02-19 Eric Botcazou <ebotcazou@adacore.com>
4939 * expr.c (expand_expr_real_1) <case VIEW_CONVERT_EXPR>: For a bit-field
4940 destination type, extract only the valid bits if the source type is not
4941 integral and has a different mode.
4943 2014-02-19 Richard Biener <rguenther@suse.de>
4946 * tree-inline.c (estimate_num_insns): Avoid calling cgraph_get_node
4949 2014-02-19 Richard Biener <rguenther@suse.de>
4952 * ipa-prop.c: Include stringpool.h and tree-ssanames.h.
4953 (ipa_modify_call_arguments): Emit an argument load explicitely and
4954 preserve virtual SSA form there and for the replacement call.
4955 Do not update SSA form nor free dominance info.
4957 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
4959 * ipa.c (function_and_variable_visibility): Also clear WEAK
4960 flag when disolving COMDAT_GROUP.
4962 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
4964 * ipa-prop.h (ipa_ancestor_jf_data): Update ocmment.
4965 * ipa-prop.c (ipa_set_jf_known_type): Return early when
4967 (ipa_set_ancestor_jf): Set type to NULL hwen it is not preserved;
4968 do more sanity checks.
4969 (detect_type_change): Return true when giving up early.
4970 (compute_complex_assign_jump_func): Fix type parameter of
4971 ipa_set_ancestor_jf.
4972 (compute_complex_ancestor_jump_func): Likewise.
4973 (update_jump_functions_after_inlining): Fix updating of
4975 * ipa-cp.c (ipa_get_jf_ancestor_result): Be ready for type to be NULL.
4977 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
4979 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Also remove
4980 inline clones when edge disappears.
4982 2014-02-18 Michael Meissner <meissner@linux.vnet.ibm.com>
4985 * config/rs6000/rs6000.md (mov<mode>_64bit, TF/TDmode moves):
4986 Split 64-bit moves into 2 patterns. Do not allow the use of
4987 direct move for TDmode in little endian, since the decimal value
4988 has little endian bytes within a word, but the 64-bit pieces are
4989 ordered in a big endian fashion, and normal subreg's of TDmode are
4991 (mov<mode>_64bit_dm): Likewise.
4992 (movtd_64bit_nodm): Likewise.
4994 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
4996 PR tree-optimization/60174
4997 * tree-ssa-reassoc.c (init_range_entry): Do not look into the defining
4998 statement of an SSA_NAME that occurs in an abnormal PHI node.
5000 2014-02-18 Jakub Jelinek <jakub@redhat.com>
5003 * final.c (SEEN_BB): Remove.
5004 (SEEN_NOTE, SEEN_EMITTED): Renumber.
5005 (final_scan_insn): Don't force_source_line on second
5006 NOTE_INSN_BASIC_BLOCK.
5008 2014-02-18 Uros Bizjak <ubizjak@gmail.com>
5011 * config/i386/i386.h (struct ix86_args): Add warn_avx512f.
5012 * config/i386/i386.c (init_cumulative_args): Initialize warn_avx512f.
5013 (type_natural_mode): Warn ABI change when %zmm register is not
5014 available for AVX512F vector value passing.
5016 2014-02-18 Kai Tietz <ktietz@redhat.com>
5019 * config/i386/i386.c (ix86_expand_prologue): Use value in
5020 rax register as displacement when restoring %r10 or %rax.
5021 Fix wrong offset when restoring both registers.
5023 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
5025 * ipa-prop.c (compute_complex_ancestor_jump_func): Replace overzealous
5026 assertion with conditional return.
5028 2014-02-18 Jakub Jelinek <jakub@redhat.com>
5029 Uros Bizjak <ubizjak@gmail.com>
5032 * config/i386/driver-i386.c (host_detect_local_cpu): If
5033 YMM state is not saved by the OS, also clear has_f16c. Move
5034 CPUID 0x80000001 handling before YMM state saving checking.
5036 2014-02-18 Andrey Belevantsev <abel@ispras.ru>
5038 PR rtl-optimization/58960
5039 * haifa-sched.c (alloc_global_sched_pressure_data): New,
5040 factored out from ...
5041 (sched_init): ... here.
5042 (free_global_sched_pressure_data): New, factored out from ...
5043 (sched_finish): ... here.
5044 * sched-int.h (free_global_sched_pressure_data): Declare.
5045 * sched-rgn.c (nr_regions_initial): New static global.
5046 (haifa_find_rgns): Initialize it.
5047 (schedule_region): Disable sched-pressure for the newly
5050 2014-02-17 Richard Biener <rguenther@suse.de>
5052 * tree-vect-stmts.c (free_stmt_vec_info): Clear BB and
5053 release SSA defs of pattern stmts.
5055 2014-02-17 Richard Biener <rguenther@suse.de>
5057 * tree-inline.c (expand_call_inline): Release the virtual
5058 operand defined by the call we are about to inline.
5060 2014-02-17 Richard Biener <rguenther@suse.de>
5062 * tree-ssa.c (verify_ssa): If verify_def found an error, ICE.
5064 2014-02-17 Kirill Yukhin <kirill.yukhin@intel.com>
5065 Ilya Tocar <ilya.tocar@intel.com>
5067 * config/i386/avx512fintrin.h (_mm512_maskz_permutexvar_epi64): Swap
5068 arguments order in builtin.
5069 (_mm512_permutexvar_epi64): Ditto.
5070 (_mm512_mask_permutexvar_epi64): Ditto
5071 (_mm512_maskz_permutexvar_epi32): Ditto
5072 (_mm512_permutexvar_epi32): Ditto
5073 (_mm512_mask_permutexvar_epi32): Ditto
5075 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5077 * config/rs6000/altivec.md (p8_vmrgew): Handle little endian targets.
5078 (p8_vmrgow): Likewise.
5080 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5082 * config/rs6000/vsx.md (vsx_xxpermdi_<mode>): Handle little
5085 2014-02-15 Michael Meissner <meissner@linux.vnet.ibm.com>
5088 * config/rs6000/rs6000.md (rreg): Add TFmode, TDmode constraints.
5089 (mov<mode>_internal, TFmode/TDmode): Split TFmode/TDmode moves
5090 into 64-bit and 32-bit moves. On 64-bit moves, add support for
5091 using direct move instructions on ISA 2.07. Also adjust
5092 instruction length for 64-bit.
5093 (mov<mode>_64bit, TFmode/TDmode): Likewise.
5094 (mov<mode>_32bit, TFmode/TDmode): Likewise.
5096 2014-02-15 Alan Modra <amodra@gmail.com>
5100 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
5101 find_replacement on parts of insn rtl that might be reloaded.
5103 2014-02-15 Richard Biener <rguenther@suse.de>
5105 PR tree-optimization/60183
5106 * tree-ssa-phiprop.c (propagate_with_phi): Avoid speculating loads.
5107 (tree_ssa_phiprop): Calculate and free post-dominators.
5109 2014-02-14 Jeff Law <law@redhat.com>
5111 PR rtl-optimization/60131
5112 * ree.c (get_extended_src_reg): New function.
5113 (combine_reaching_defs): Use it rather than assuming location of REG.
5114 (find_and_remove_re): Verify first operand of extension is
5115 a REG before adding the insns to the copy list.
5117 2014-02-14 Roland McGrath <mcgrathr@google.com>
5119 * configure.ac (HAVE_AS_IX86_UD2): New test for 'ud2' mnemonic.
5120 * configure: Regenerated.
5121 * config.in: Regenerated.
5122 * config/i386/i386.md (trap) [HAVE_AS_IX86_UD2]: Use the mnemonic
5123 instead of ASM_SHORT.
5125 2014-02-14 Vladimir Makarov <vmakarov@redhat.com>
5126 Richard Earnshaw <rearnsha@arm.com>
5128 PR rtl-optimization/59535
5129 * lra-constraints.c (process_alt_operands): Encourage alternative
5130 when unassigned pseudo class is superset of the alternative class.
5131 (inherit_reload_reg): Don't inherit when optimizing for code size.
5132 * config/arm/arm.h (MODE_BASE_REG_CLASS): Add version for LRA
5133 returning CORE_REGS for anything but Thumb1 and BASE_REGS for
5134 modes not less than 4 for Thumb1.
5136 2014-02-14 Kyle McMartin <kyle@redhat.com>
5139 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Define for AArch64.
5141 2014-02-14 Richard Biener <rguenther@suse.de>
5143 * cilk-common.c (cilk_arrow): Build a MEM_REF, not an INDIRECT_REF.
5144 (get_frame_arg): Drop the assert with langhook types_compatible_p.
5145 Do not strip INDIRECT_REFs.
5147 2014-02-14 Richard Biener <rguenther@suse.de>
5150 * lto-streamer-out.c (DFS_write_tree_body): Do not follow
5151 DECL_FUNCTION_SPECIFIC_TARGET.
5152 (hash_tree): Do not hash DECL_FUNCTION_SPECIFIC_TARGET.
5153 * tree-streamer-out.c (pack_ts_target_option): Remove.
5154 (streamer_pack_tree_bitfields): Do not stream TS_TARGET_OPTION.
5155 (write_ts_function_decl_tree_pointers): Do not stream
5156 DECL_FUNCTION_SPECIFIC_TARGET.
5157 * tree-streamer-in.c (unpack_ts_target_option): Remove.
5158 (unpack_value_fields): Do not stream TS_TARGET_OPTION.
5159 (lto_input_ts_function_decl_tree_pointers): Do not stream
5160 DECL_FUNCTION_SPECIFIC_TARGET.
5162 2014-02-14 Jakub Jelinek <jakub@redhat.com>
5164 * tree-vect-loop.c (vect_is_slp_reduction): Don't set use_stmt twice.
5165 (get_initial_def_for_induction, vectorizable_induction): Ignore
5166 debug stmts when looking for exit_phi.
5167 (vectorizable_live_operation): Fix up condition.
5169 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
5171 * config/nds32/nds32.c (nds32_asm_function_prologue): Do not use
5172 nreverse() because it changes the content of original tree list.
5174 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
5176 * config/nds32/t-mlibs (MULTILIB_OPTIONS): Fix typo in comment.
5177 * config/nds32/nds32.c (nds32_merge_decl_attributes): Likewise.
5179 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
5181 * config/nds32/nds32.c (nds32_naked_function_p): Follow the
5182 GNU coding standards.
5184 2014-02-13 Jakub Jelinek <jakub@redhat.com>
5187 * dwarf2out.c (gen_subprogram_die): Don't call
5188 add_calling_convention_attribute if subr_die is old_die.
5190 2014-02-13 Sharad Singhai <singhai@google.com>
5192 * doc/optinfo.texi: Fix order of nodes.
5194 2014-02-13 Uros Bizjak <ubizjak@gmail.com>
5196 * config/i386/sse.md (xop_vmfrcz<mode>2): Generate const0 in
5197 operands[2], not operands[3].
5199 2014-02-13 Richard Biener <rguenther@suse.de>
5202 * doc/install.texi (ISL): Update recommended version to 0.12.2,
5203 mention the possibility of an in-tree build.
5204 (CLooG): Update recommended version to 0.18.1, mention the
5205 possibility of an in-tree build and clarify that the ISL
5206 bundled with CLooG does not work.
5208 2014-02-13 Jakub Jelinek <jakub@redhat.com>
5211 * expr.c (compress_float_constant): If x is a hard register,
5212 extend into a pseudo and then move to x.
5214 2014-02-13 Dominik Vogt <vogt@linux.vnet.ibm.com>
5216 * config/s390/s390.c (s390_asm_output_function_label): Fix crash
5217 caused by bad second argument to warning_at() with -mhotpatch and
5218 nested functions (e.g. with gfortran).
5220 2014-02-13 Richard Sandiford <rdsandiford@googlemail.com>
5222 * opts.c (option_name): Remove "enabled by default" rider.
5224 2014-02-12 John David Anglin <danglin@gcc.gnu.org>
5226 * config/pa/pa.c (pa_option_override): Remove auto increment FIXME.
5228 2014-02-12 H.J. Lu <hongjiu.lu@intel.com>
5229 Uros Bizjak <ubizjak@gmail.com>
5232 * configure.ac (HAVE_AS_GOTOFF_IN_DATA): Pass --32 to GNU assembler.
5233 * configure: Regenerated.
5235 2014-02-12 Richard Biener <rguenther@suse.de>
5237 * vec.c (vec_prefix::calculate_allocation): Move as
5238 inline variant to vec.h.
5239 (vec_prefix::calculate_allocation_1): New out-of-line version.
5240 * vec.h (vec_prefix::calculate_allocation_1): Declare.
5241 (vec_prefix::m_has_auto_buf): Rename to ...
5242 (vec_prefix::m_using_auto_storage): ... this.
5243 (vec_prefix::calculate_allocation): Inline the easy cases
5244 and dispatch to calculate_allocation_1 which doesn't need the
5246 (va_heap::reserve): Use gcc_checking_assert.
5247 (vec<T, A, vl_embed>::embedded_init): Add argument to initialize
5248 m_using_auto_storage.
5249 (auto_vec): Change m_vecpfx member to a vec<T, va_heap, vl_embed>
5251 (vec<T, va_heap, vl_ptr>::reserve): Remove redundant check.
5252 (vec<T, va_heap, vl_ptr>::release): Avoid casting.
5253 (vec<T, va_heap, vl_ptr>::using_auto_storage): Simplify.
5255 2014-02-12 Richard Biener <rguenther@suse.de>
5257 * gcse.c (compute_transp): break from loop over canon_modify_mem_list
5258 when we found a dependence.
5260 2014-02-12 Thomas Schwinge <thomas@codesourcery.com>
5262 * gimplify.c (gimplify_call_expr, gimplify_modify_expr): Move
5264 (maybe_fold_stmt): ... into this new function.
5265 * omp-low.c (lower_omp): Update comment.
5267 * omp-low.c (lower_omp_target): Add clobber for sizes array, after
5270 * omp-low.c (diagnose_sb_0): Make sure label_ctx is valid to
5273 2014-02-12 James Greenhalgh <james.greenhalgh@arm.com>
5275 * config/arm/aarch-cost-tables.h (generic_extra_costs): Fix
5276 identifiers in comments.
5277 (cortexa53_extra_costs): Likewise.
5278 * config/arm/arm.c (cortexa9_extra_costs): Fix identifiers in comments.
5279 (cortexa7_extra_costs): Likewise.
5280 (cortexa12_extra_costs): Likewise.
5281 (cortexa15_extra_costs): Likewise.
5282 (v7m_extra_costs): Likewise.
5284 2014-02-12 Richard Biener <rguenther@suse.de>
5287 * gimple-low.c (lower_builtin_posix_memalign): Lower conditional
5288 of posix_memalign being successful.
5289 (lower_stmt): Restrict lowering of posix_memalign to when
5290 -ftree-bit-ccp is enabled.
5292 2014-02-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
5294 * config/avr/avr-c.c (avr_resolve_overloaded_builtin): Pass vNULL for
5296 * config/spu/spu-c.c (spu_resolve_overloaded_builtin): Likewise.
5298 2014-02-12 Eric Botcazou <ebotcazou@adacore.com>
5300 PR rtl-optimization/60116
5301 * combine.c (try_combine): Also remove dangling REG_DEAD notes on the
5302 other_insn once the combination has been validated.
5304 2014-02-11 Jan Hubicka <hubicka@ucw.cz>
5307 * ipa-utils.h (possible_polymorphic_call_targets): Update prototype
5309 * ipa-devirt.c: Include demangle.h
5310 (odr_violation_reported): New static variable.
5311 (add_type_duplicate): Update odr_violations.
5312 (maybe_record_node): Add completep parameter; update it.
5313 (record_target_from_binfo): Add COMPLETEP parameter;
5314 update it as needed.
5315 (possible_polymorphic_call_targets_1): Likewise.
5316 (struct polymorphic_call_target_d): Add nonconstruction_targets;
5317 rename FINAL to COMPLETE.
5318 (record_targets_from_bases): Sanity check we found the binfo;
5319 fix COMPLETEP updating.
5320 (possible_polymorphic_call_targets): Add NONCONSTRUTION_TARGETSP
5321 parameter, fix computing of COMPLETEP.
5322 (dump_possible_polymorphic_call_targets): Imrove readability of dump;
5323 at LTO time do demangling.
5324 (ipa_devirt): Use nonconstruction_targets; Improve dumps.
5325 * gimple-fold.c (gimple_get_virt_method_for_vtable): Add can_refer
5327 (gimple_get_virt_method_for_binfo): Likewise.
5328 * gimple-fold.h (gimple_get_virt_method_for_binfo,
5329 gimple_get_virt_method_for_vtable): Update prototypes.
5331 2014-02-11 Vladimir Makarov <vmakarov@redhat.com>
5334 * genautomata.c (add_presence_absence): Fix typo with
5335 {final_}presence_list.
5337 2014-02-11 Michael Meissner <meissner@linux.vnet.ibm.com>
5340 * config/rs6000/rs6000.md (128-bit GPR splitter): Add a splitter
5341 for VSX/Altivec vectors that land in GPR registers.
5343 2014-02-11 Richard Henderson <rth@redhat.com>
5344 Jakub Jelinek <jakub@redhat.com>
5347 * tree-sra.c (load_assign_lhs_subreplacements): Add VIEW_CONVERT_EXPR
5348 around drhs if type conversion to lacc->type is not useless.
5350 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5352 * config/aarch64/aarch64-cores.def (cortex-a57): Use cortexa57
5354 (cortex-a57.cortex-a53): Likewise.
5355 * config/aarch64/aarch64.c (cortexa57_tunings): New tuning struct.
5357 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5359 * config/arm/thumb2.md (*thumb2_movhi_insn): Add alternatives for
5362 2014-02-11 Renlin Li <Renlin.Li@arm.com>
5364 * doc/sourcebuild.texi: Document check_effective_target_arm_vfp3_ok and
5365 add_options_for_arm_vfp3.
5367 2014-02-11 Jeff Law <law@redhat.com>
5370 * expr.c (expand_expr_addr_expr_1): Handle expand_expr returning an
5371 object with an undesirable mode.
5373 2014-02-11 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5376 * config/i386/sol2-9.h: New file.
5377 * config.gcc (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*,
5378 *-*-solaris2.9*): Use it.
5380 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
5382 * config/microblaze/microblaze.md: Add movsi4_rev insn pattern.
5383 * config/microblaze/predicates.md: Add reg_or_mem_operand predicate.
5385 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
5387 * config/microblaze/microblaze.c: Extend mcpu version format
5389 2014-02-10 David Holsgrove <david.holsgrove@xilinx.com>
5391 * config/microblaze/microblaze.h: Define SIZE_TYPE and PTRDIFF_TYPE.
5393 2014-02-10 Richard Henderson <rth@redhat.com>
5396 * calls.c (expand_call): Don't double-push for reg_parm_stack_space.
5397 * config/i386/i386.c (init_cumulative_args): Remove sorry for 64-bit
5398 ms-abi vs -mno-accumulate-outgoing-args.
5399 (ix86_expand_prologue): Unconditionally call ix86_eax_live_at_start_p.
5400 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Fix comment with
5403 2014-02-10 Bernd Edlinger <bernd.edlinger@hotmail.de>
5406 * cfgexpand.c (expand_asm_operands): Attach source location to
5407 ASM_INPUT rtx objects.
5408 * print-rtl.c (print_rtx): Check for UNKNOWN_LOCATION.
5410 2014-02-10 Nick Clifton <nickc@redhat.com>
5412 * config/mn10300/mn10300.c (popcount): New function.
5413 (mn10300_expand_prologue): Include saved registers in stack usage
5416 2014-02-10 Jeff Law <law@redhat.com>
5419 * reload1.c (emit_input_reload_insns): Do not create invalid RTL
5420 when changing the SET_DEST of a prior insn to avoid an input reload.
5422 2014-02-10 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
5424 * config/rs6000/sysv4.h (ENDIAN_SELECT): Do not attempt to enforce
5425 big-endian mode for -mcall-aixdesc, -mcall-freebsd, -mcall-netbsd,
5426 -mcall-openbsd, or -mcall-linux.
5427 (CC1_ENDIAN_BIG_SPEC): Remove.
5428 (CC1_ENDIAN_LITTLE_SPEC): Remove.
5429 (CC1_ENDIAN_DEFAULT_SPEC): Remove.
5430 (CC1_SPEC): Remove (always empty) %cc1_endian_... spec.
5431 (SUBTARGET_EXTRA_SPECS): Remove %cc1_endian_big, %cc1_endian_little,
5432 and %cc1_endian_default.
5433 * config/rs6000/sysv4le.h (CC1_ENDIAN_DEFAULT_SPEC): Remove.
5435 2014-02-10 Richard Biener <rguenther@suse.de>
5437 PR tree-optimization/60115
5438 * tree-eh.c (tree_could_trap_p): Unify TARGET_MEM_REF and
5439 MEM_REF handling. Properly verify that the accesses are not
5440 out of the objects bound.
5442 2014-02-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5444 * config/aarch64/aarch64.c (aarch64_override_options): Fix typo from
5447 2014-02-10 Eric Botcazou <ebotcazou@adacore.com>
5449 * ipa-devirt.c (get_polymorphic_call_info_from_invariant): Return
5450 proper constants and fix formatting.
5451 (possible_polymorphic_call_targets): Fix formatting.
5453 2014-02-10 Kirill Yukhin <kirill.yukhin@intel.com>
5454 Ilya Tocar <ilya.tocar@intel.com>
5456 * config/i386/avx512fintrin.h (_mm512_storeu_epi64): Removed.
5457 (_mm512_loadu_epi32): Renamed into...
5458 (_mm512_loadu_si512): This.
5459 (_mm512_storeu_epi32): Renamed into...
5460 (_mm512_storeu_si512): This.
5461 (_mm512_maskz_ceil_ps): Removed.
5462 (_mm512_maskz_ceil_pd): Ditto.
5463 (_mm512_maskz_floor_ps): Ditto.
5464 (_mm512_maskz_floor_pd): Ditto.
5465 (_mm512_floor_round_ps): Ditto.
5466 (_mm512_floor_round_pd): Ditto.
5467 (_mm512_ceil_round_ps): Ditto.
5468 (_mm512_ceil_round_pd): Ditto.
5469 (_mm512_mask_floor_round_ps): Ditto.
5470 (_mm512_mask_floor_round_pd): Ditto.
5471 (_mm512_mask_ceil_round_ps): Ditto.
5472 (_mm512_mask_ceil_round_pd): Ditto.
5473 (_mm512_maskz_floor_round_ps): Ditto.
5474 (_mm512_maskz_floor_round_pd): Ditto.
5475 (_mm512_maskz_ceil_round_ps): Ditto.
5476 (_mm512_maskz_ceil_round_pd): Ditto.
5477 (_mm512_expand_pd): Ditto.
5478 (_mm512_expand_ps): Ditto.
5479 * config/i386/i386.c (ix86_builtins): Remove
5480 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK.
5481 (bdesc_args): Ditto.
5482 * config/i386/predicates.md (const1256_operand): New.
5483 (const_1_to_2_operand): Ditto.
5484 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
5485 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
5486 (*avx512pf_gatherpf<mode>sf): Ditto.
5487 (avx512pf_gatherpf<mode>df): Ditto.
5488 (*avx512pf_gatherpf<mode>df_mask): Ditto.
5489 (*avx512pf_gatherpf<mode>df): Ditto.
5490 (avx512pf_scatterpf<mode>sf): Ditto.
5491 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
5492 (*avx512pf_scatterpf<mode>sf): Ditto.
5493 (avx512pf_scatterpf<mode>df): Ditto.
5494 (*avx512pf_scatterpf<mode>df_mask): Ditto.
5495 (*avx512pf_scatterpf<mode>df): Ditto.
5496 (avx512f_expand<mode>): Removed.
5497 (<shift_insn><mode>3<mask_name>): Change predicate type.
5499 2014-02-08 Jakub Jelinek <jakub@redhat.com>
5501 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers
5502 not at the end of datarefs vector use ordered_remove to avoid
5503 reordering datarefs vector.
5506 * gimplify.c (gimplify_bind_expr): In ORT_SIMD region
5507 mark local addressable non-static vars as GOVD_PRIVATE
5508 instead of GOVD_LOCAL.
5509 * omp-low.c (lower_omp_for): Move gimple_bind_vars
5510 and BLOCK_VARS of gimple_bind_block to new_stmt rather
5514 * tree-ssa-ccp.c (surely_varying_stmt_p): Don't return true
5515 if TYPE_ATTRIBUTES (gimple_call_fntype ()) contain
5516 assume_aligned or alloc_align attributes.
5517 (bit_value_assume_aligned): Add ATTR, PTRVAL and ALLOC_ALIGN
5518 arguments. Handle also assume_aligned and alloc_align attributes.
5519 (evaluate_stmt): Adjust bit_value_assume_aligned caller. Handle
5520 calls to functions with assume_aligned or alloc_align attributes.
5521 * doc/extend.texi: Document assume_aligned and alloc_align attributes.
5523 2014-02-08 Terry Guo <terry.guo@arm.com>
5525 * doc/invoke.texi: Document ARM -march=armv7e-m.
5527 2014-02-08 Jakub Jelinek <jakub@redhat.com>
5529 * cilk-common.c (cilk_init_builtins): Clear TREE_NOTHROW
5530 flag on __cilkrts_rethrow builtin.
5533 * ipa-cp.c (determine_versionability): Fail at -O0
5534 or __attribute__((optimize (0))) or -fno-ipa-cp functions.
5535 * tree-sra.c (ipa_sra_preliminary_function_checks): Similarly.
5538 2014-02-04 Jakub Jelinek <jakub@redhat.com>
5541 * tree-inline.c (copy_forbidden): Fail for
5542 __attribute__((optimize (0))) functions.
5544 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
5546 * varpool.c: Include pointer-set.h.
5547 (varpool_remove_unreferenced_decls): Variables in other partitions
5548 will not be output; be however careful to not lose information
5551 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
5553 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do O(1)
5554 lookup in the vtable constructor.
5556 2014-02-07 Jeff Law <law@redhat.com>
5559 * config/m68k/m68k.md (ashldi_extsi): Turn into a
5560 define_insn_and_split.
5562 * ipa-inline.c (inline_small_functions): Fix typos.
5564 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5566 * config/s390/s390-protos.h (s390_can_use_simple_return_insn)
5567 (s390_can_use_return_insn): Declare.
5568 * config/s390/s390.h (EPILOGUE_USES): Define.
5569 * config/s390/s390.c (s390_mainpool_start): Allow two main_pool
5571 (s390_chunkify_start): Handle return JUMP_LABELs.
5572 (s390_early_mach): Emit a main_pool instruction on the entry edge.
5573 (s300_set_up_by_prologue, s390_can_use_simple_return_insn)
5574 (s390_can_use_return_insn): New functions.
5575 (s390_fix_long_loop_prediction): Handle conditional returns.
5576 (TARGET_SET_UP_BY_PROLOGUE): Define.
5577 * config/s390/s390.md (ANY_RETURN): New code iterator.
5578 (*creturn, *csimple_return, return, simple_return): New patterns.
5580 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5582 * config/s390/s390.c (s390_restore_gprs_from_fprs): Add REG_CFA_RESTORE
5583 notes to each restore. Also add REG_CFA_DEF_CFA when restoring %r15.
5584 (s390_optimize_prologue): Don't clear RTX_FRAME_RELATED_P. Update the
5585 REG_CFA_RESTORE list when deciding not to restore a register.
5587 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5589 * config/s390/s390.c: Include tree-pass.h and context.h.
5590 (s390_early_mach): New function, split out from...
5591 (s390_emit_prologue): ...here.
5592 (pass_data_s390_early_mach): New pass structure.
5593 (pass_s390_early_mach): New class.
5594 (s390_option_override): Create and register early_mach pass.
5595 Move to end of file.
5597 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5599 * var-tracking.c (vt_stack_adjustments): Don't require stack_adjusts
5600 to match for the exit block.
5602 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5604 * config/s390/s390.md ("atomic_load<mode>", "atomic_store<mode>")
5605 ("atomic_compare_and_swap<mode>", "atomic_fetch_<atomic><mode>"):
5606 Reject misaligned operands.
5608 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5610 * optabs.c (expand_atomic_compare_and_swap): Allow expander to fail.
5612 2014-02-07 Richard Biener <rguenther@suse.de>
5615 * gimple-low.c (lower_builtin_posix_memalign): New function.
5616 (lower_stmt): Call it to lower posix_memalign in a way
5617 to make alignment info accessible.
5619 2014-02-07 Jakub Jelinek <jakub@redhat.com>
5622 * tree.c (build_common_builtin_nodes): Set ECF_LEAF for
5623 __builtin_setjmp_receiver.
5625 2014-02-07 Richard Biener <rguenther@suse.de>
5628 * builtin-types.def (BT_FN_INT_PTRPTR_SIZE_SIZE): Add.
5629 * builtins.def (BUILT_IN_POSIX_MEMALIGN): Likewise.
5630 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
5631 Handle BUILT_IN_POSIX_MEMALIGN.
5632 (find_func_clobbers): Likewise.
5633 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
5634 (call_may_clobber_ref_p_1): Likewise.
5636 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5639 * ipa-devirt.c (record_target_from_binfo): Remove overactive
5642 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5645 * lto-cgraph.c (lto_output_node): Use
5646 symtab_get_symbol_partitioning_class.
5647 (lto_output_varpool_node): likewise.
5648 (symtab_get_symbol_partitioning_class): Move here from
5650 * cgraph.h (symbol_partitioning_class): Likewise.
5651 (symtab_get_symbol_partitioning_class): Declare.
5653 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5655 * ggc.h (ggc_internal_cleared_alloc): New macro.
5656 * vec.h (vec_safe_copy): Handle memory stats.
5657 * omp-low.c (simd_clone_struct_alloc): Use ggc_internal_cleared_alloc.
5658 * target-globals.c (save_target_globals): Likewise.
5660 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5663 * expr.c (emit_move_resolve_push): Export; be bit more selective
5664 on when to clear alias set.
5665 * expr.h (emit_move_resolve_push): Declare.
5666 * function.h (struct function): Add tail_call_marked.
5667 * tree-tailcall.c (optimize_tail_call): Set tail_call_marked.
5668 * config/i386/i386-protos.h (ix86_expand_push): Remove.
5669 * config/i386/i386.md (TImode move expander): De not call
5671 (FP push expanders): Preserve memory attributes.
5672 * config/i386/sse.md (push<mode>1): Remove.
5673 * config/i386/i386.c (ix86_expand_vector_move): Handle push operation.
5674 (ix86_expand_push): Remove.
5675 * config/i386/mmx.md (push<mode>1): Remove.
5677 2014-02-06 Jakub Jelinek <jakub@redhat.com>
5679 PR rtl-optimization/60030
5680 * internal-fn.c (ubsan_expand_si_overflow_mul_check): Surround
5681 lopart with paradoxical subreg before shifting it up by hprec.
5683 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5685 * config/arm/aarch-cost-tables.h (cortexa57_extra_costs): New table.
5686 Remove extra newline at end of file.
5687 * config/arm/arm.c (arm_cortex_a57_tune): New tuning struct.
5688 (arm_issue_rate): Handle cortexa57.
5689 * config/arm/arm-cores.def (cortex-a57): Use cortex_a57 tuning.
5690 (cortex-a57.cortex-a53): Likewise.
5692 2014-02-06 Jakub Jelinek <jakub@redhat.com>
5695 * config/arm/arm.c (emit_multi_reg_push): Add dwarf_regs_mask argument,
5696 don't record in REG_FRAME_RELATED_EXPR registers not set in that
5698 (arm_expand_prologue): Adjust all callers.
5699 (arm_unwind_emit_sequence): Allow saved, but not important for unwind
5700 info, registers also at the lowest numbered registers side. Use
5701 gcc_assert instead of abort, and SET_SRC/SET_DEST macros instead of
5705 * var-tracking.c (adjust_mems): Before adding a SET to
5706 amd->side_effects, adjust it's SET_SRC using simplify_replace_fn_rtx.
5708 2014-02-06 Alan Modra <amodra@gmail.com>
5711 * config/rs6000/rs6000.c (rs6000_secondary_memory_needed_mode): Only
5712 change SDmode to DDmode when lra_in_progress.
5714 2014-02-06 Jakub Jelinek <jakub@redhat.com>
5717 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers, call
5718 free_data_ref on the dr first, and before goto again also set dr
5719 to the next dr. For simd_lane_access, free old datarefs[i] before
5720 overwriting it. For get_vectype_for_scalar_type failure, don't
5721 free_data_ref if simd_lane_access.
5723 * Makefile.in (prefix.o, cppbuiltin.o): Depend on $(BASEVER).
5726 * tree.h (opts_for_fn): New inline function.
5727 (opt_for_fn): Define.
5728 * config/i386/i386.c (ix86_function_regparm): Use
5729 opt_for_fn (decl, optimize) instead of optimize.
5731 2014-02-06 Marcus Shawcroft <marcus.shawcroft@arm.com>
5733 * config/aarch64/aarch64.c (aarch64_classify_symbol): Fix logic
5734 for SYMBOL_REF in large memory model.
5736 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5738 * config/aarch64/aarch64-cores.def (cortex-a53): Specify CRC32
5740 (cortex-a57): Likewise.
5741 (cortex-a57.cortex-a53): Likewise.
5743 2014-02-06 Yury Gribov <y.gribov@samsung.com>
5744 Kugan Vivekanandarajah <kuganv@linaro.org>
5746 * config/arm/arm.c (arm_vector_alignment_reachable): Check
5748 * config/arm/arm.c (arm_builtin_support_vector_misalignment): Likewise.
5750 2014-02-06 Richard Biener <rguenther@suse.de>
5752 * tree-cfg.c (gimple_duplicate_sese_region): Fix ordering of
5753 set_loop_copy and initialize_original_copy_tables.
5755 2014-02-06 Alex Velenko <Alex.Velenko@arm.com>
5757 * config/aarch64/aarch64-simd.md
5758 (aarch64_ashr_simddi): Change QI to SI.
5760 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
5761 Jakub Jelinek <jakub@redhat.com>
5764 * ipa-inline-analysis.c (compute_bb_predicates): Ensure monotonicity
5767 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5769 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Change
5770 CODE_FOR_altivec_vpku[hw]um to
5771 CODE_FOR_altivec_vpku[hw]um_direct.
5772 * config/rs6000/altivec.md (vec_unpacks_hi_<VP_small_lc>): Change
5773 UNSPEC_VUNPACK_HI_SIGN to UNSPEC_VUNPACK_HI_SIGN_DIRECT.
5774 (vec_unpacks_lo_<VP_small_lc>): Change UNSPEC_VUNPACK_LO_SIGN to
5775 UNSPEC_VUNPACK_LO_SIGN_DIRECT.
5777 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5779 * config/rs6000/altivec.md (altivec_vsum2sws): Adjust code
5780 generation for -maltivec=be.
5781 (altivec_vsumsws): Simplify redundant test.
5783 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5785 * altivec.md (UNSPEC_VPACK_UNS_UNS_MOD_DIRECT): New unspec.
5786 (UNSPEC_VUNPACK_HI_SIGN_DIRECT): Likewise.
5787 (UNSPEC_VUNPACK_LO_SIGN_DIRECT): Likewise.
5788 (mulv8hi3): Use gen_altivec_vpkuwum_direct instead of
5789 gen_altivec_vpkuwum.
5790 (altivec_vpkpx): Test for VECTOR_ELT_ORDER_BIG instead of for
5792 (altivec_vpks<VI_char>ss): Likewise.
5793 (altivec_vpks<VI_char>us): Likewise.
5794 (altivec_vpku<VI_char>us): Likewise.
5795 (altivec_vpku<VI_char>um): Likewise.
5796 (altivec_vpku<VI_char>um_direct): New (copy of
5797 altivec_vpku<VI_char>um that still relies on BYTES_BIG_ENDIAN, for
5799 (altivec_vupkhs<VU_char>): Emit vupkls* instead of vupkhs* when
5800 target is little endian and -maltivec=be is not specified.
5801 (*altivec_vupkhs<VU_char>_direct): New (copy of
5802 altivec_vupkhs<VU_char> that always emits vupkhs*, for internal use).
5803 (altivec_vupkls<VU_char>): Emit vupkhs* instead of vupkls* when
5804 target is little endian and -maltivec=be is not specified.
5805 (*altivec_vupkls<VU_char>_direct): New (copy of
5806 altivec_vupkls<VU_char> that always emits vupkls*, for internal use).
5807 (altivec_vupkhpx): Emit vupklpx instead of vupkhpx when target is
5808 little endian and -maltivec=be is not specified.
5809 (altivec_vupklpx): Emit vupkhpx instead of vupklpx when target is
5810 little endian and -maltivec=be is not specified.
5812 2014-02-05 Richard Henderson <rth@redhat.com>
5815 * combine-stack-adj.c: Revert r206943.
5816 * sched-int.h (struct deps_desc): Add last_args_size.
5817 * sched-deps.c (init_deps): Initialize it.
5818 (sched_analyze_insn): Add OUTPUT dependencies between insns that
5819 contain REG_ARGS_SIZE notes.
5821 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
5823 * lto-cgraph.c (asm_nodes_output): Make global.
5824 * lto-wrapper.c (run_gcc): Pass down paralelizm to WPA.
5825 * gcc.c (AS_NEEDS_DASH_FOR_PIPED_INPUT): Allow WPA parameter
5826 (driver_handle_option): Handle OPT_fwpa.
5828 2014-02-05 Jakub Jelinek <jakub@redhat.com>
5831 * ipa-devirt.c (possible_polymorphic_call_targets): Fix
5832 a comment typo and formatting issue. If odr_hash hasn't been
5833 created, return vNULL and set *completep to false.
5836 * tree-eh.c (cleanup_empty_eh): Bail out on totally empty
5837 bb with no successors.
5839 2014-02-05 James Greenhalgh <james.greenhalgh@arm.com>
5842 * doc/invoke.texi (-march): Clarify documentation for ARM.
5846 2014-02-05 Richard Biener <rguenther@suse.de>
5848 * tree-vect-loop.c (vect_analyze_loop_2): Be more informative
5849 when not vectorizing because of too many alias checks.
5850 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
5851 Add more verboseness, avoid duplicate MSG_MISSED_OPTIMIZATION.
5853 2014-02-05 Nick Clifton <nickc@redhat.com>
5855 * config/mn10300/mn10300.c (mn10300_hard_regno_mode_ok): Do not
5856 accept extended registers in any mode when compiling for the MN10300.
5858 2014-02-05 Yury Gribov <y.gribov@samsung.com>
5860 * cif-code.def (ATTRIBUTE_MISMATCH): New CIF code.
5861 * ipa-inline.c (report_inline_failed_reason): Handle mismatched
5862 sanitization attributes.
5863 (can_inline_edge_p): Likewise.
5864 (sanitize_attrs_match_for_inline_p): New function.
5866 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
5868 * ipa-prop.c (detect_type_change): Shor circuit testing of
5869 type changes on THIS pointer.
5871 2014-02-04 John David Anglin <danglin@gcc.gnu.org>
5874 * config/pa/pa.c (legitimize_tls_address): Return original address
5875 if not passed a SYMBOL_REF rtx.
5876 (hppa_legitimize_address): Call legitimize_tls_address for all TLS
5878 (pa_emit_move_sequence): Simplify TLS source operands.
5879 (pa_legitimate_constant_p): Reject all TLS constants.
5880 * config/pa/pa.h (PA_SYMBOL_REF_TLS_P): Correct comment.
5881 (CONSTANT_ADDRESS_P): Reject TLS CONST addresses.
5883 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
5885 * ipa.c (function_and_variable_visibility): Decompose DECL_ONE_ONLY
5886 groups when we know they are controlled by LTO.
5887 * varasm.c (default_binds_local_p_1): If object is in other partition,
5888 it will be resolved locally.
5890 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
5892 * config/host-linux.c (linux_gt_pch_use_address): Don't
5893 use SSIZE_MAX because it is not always defined.
5895 2014-02-04 Vladimir Makarov <vmakarov@redhat.com>
5898 * lra-constraints.c (need_for_split_p): Use more 3 reloads as
5899 threshold for pseudo splitting.
5900 (update_ebb_live_info): Process call argument hard registers and
5901 hard registers from insn definition too.
5902 (max_small_class_regs_num): New constant.
5903 (inherit_in_ebb): Update live hard regs through EBBs. Update
5904 reloads_num only for small register classes. Don't split for
5907 2014-02-04 Markus Trippelsdorf <markus@trippelsdorf.de>
5910 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that target
5913 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
5915 * gimple-fold.c (can_refer_decl_in_current_unit_p): Default
5918 2014-02-04 Marek Polacek <polacek@redhat.com>
5920 * gdbinit.in (pel): Define.
5922 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
5924 * doc/invoke.texi (fstrict-volatile-bitfields): Clarify current
5927 2014-02-04 Richard Biener <rguenther@suse.de>
5930 * lto-streamer-out.c (tree_is_indexable): Force NAMELIST_DECLs
5931 in function context local.
5932 (lto_output_tree_ref): Do not write trees from lto_output_tree_ref.
5933 * lto-streamer-in.c (lto_input_tree_ref): Handle LTO_namelist_decl_ref
5934 similar to LTO_imported_decl_ref.
5936 2014-02-04 Jakub Jelinek <jakub@redhat.com>
5938 PR tree-optimization/60002
5939 * cgraphclones.c (build_function_decl_skip_args): Clear
5942 PR tree-optimization/60023
5943 * tree-if-conv.c (predicate_mem_writes): Pass true instead of
5944 false to gsi_replace.
5945 * tree-vect-stmts.c (vect_finish_stmt_generation): If stmt
5946 has been in some EH region and vec_stmt could throw, add
5947 vec_stmt into the same EH region.
5948 * tree-data-ref.c (get_references_in_stmt): If IFN_MASK_LOAD
5949 has no lhs, ignore it.
5950 * internal-fn.c (expand_MASK_LOAD): Likewise.
5953 * tree-inline.c (copy_forbidden): Fail for
5954 __attribute__((optimize (0))) functions.
5957 * omp-low.c (simd_clone_struct_copy): If from->inbranch
5958 is set, copy one less argument.
5959 (expand_simd_clones): Don't subtract clone_info->inbranch
5960 from simd_clone_struct_alloc argument.
5962 PR rtl-optimization/57915
5963 * recog.c (simplify_while_replacing): If all unary/binary/relational
5964 operation arguments are constant, attempt to simplify those.
5967 * expmed.c (expand_mult): For MODE_VECTOR_INT multiplication
5968 if there is no vashl<mode>3 or ashl<mode>3 insn, skip_synth.
5970 2014-02-04 Richard Biener <rguenther@suse.de>
5972 PR tree-optimization/60012
5973 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Apply
5974 TBAA disambiguation to all DDRs.
5976 2014-02-04 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5979 * config/sol2.h (LINK_LIBGCC_MAPFILE_SPEC): Define.
5980 (LINK_SPEC): Use it for -shared, -shared-libgcc.
5982 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
5985 * tree.c (get_binfo_at_offset): Do not get confused by empty classes;
5987 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
5989 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Remove.
5990 * gimple-fold.h (gimple_extract_devirt_binfo_from_cst): Remove.
5992 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
5995 * ipa-cp.c (ipa_get_indirect_edge_target_1): Use ipa-devirt
5996 to figure out targets of polymorphic calls with known decl.
5997 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
5998 * ipa-utils.h (get_polymorphic_call_info_from_invariant): Declare.
5999 * ipa-devirt.c (get_polymorphic_call_info_for_decl): Break out from ...
6000 (get_polymorphic_call_info): ... here.
6001 (get_polymorphic_call_info_from_invariant): New function.
6003 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
6005 * ipa-cp.c (ipa_get_indirect_edge_target_1): Do direct
6006 lookup via vtable pointer; check for type consistency
6007 and turn inconsitent facts into UNREACHABLE.
6008 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
6009 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do not ICE on
6010 type inconsistent querries; return UNREACHABLE instead.
6012 2014-02-03 Richard Henderson <rth@twiddle.net>
6015 * tree-ssa-uninit.c (push_to_worklist): Don't re-push if we've
6016 already processed this node.
6017 (normalize_one_pred_1): Pass along mark_set.
6018 (normalize_one_pred): Create and destroy a pointer_set_t.
6019 (normalize_one_pred_chain): Likewise.
6021 2014-02-03 Laurent Aflonsi <laurent.alfonsi@st.com>
6023 PR gcov-profile/58602
6024 * gcc/gcov-io.c (gcov_open): Open with truncation when mode < 0.
6026 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
6029 * ipa-cp.c (ipa_get_indirect_edge_target_1): Give up on
6030 -fno-devirtualize; try to devirtualize by the knowledge of
6031 virtual table pointer given by aggregate propagation.
6032 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
6033 (ipa_print_node_jump_functions): Dump also offset that
6034 is relevant for polymorphic calls.
6035 (determine_known_aggregate_parts): Add arg_type parameter; use it
6036 instead of determining the type from pointer type.
6037 (ipa_compute_jump_functions_for_edge): Update call of
6038 determine_known_aggregate_parts.
6039 * gimple-fold.c (gimple_get_virt_method_for_vtable): Break out from ...
6040 (gimple_get_virt_method_for_binfo): ... here; simplify using
6041 vtable_pointer_value_to_vtable.
6042 * gimple-fold.h (gimple_get_virt_method_for_vtable): Declare.
6043 * ipa-devirt.c (subbinfo_with_vtable_at_offset): Turn OFFSET parameter
6044 to unsigned HOST_WIDE_INT; use vtable_pointer_value_to_vtable.
6045 (vtable_pointer_value_to_vtable): Break out from ...; handle also
6047 (vtable_pointer_value_to_binfo): ... here.
6048 * ipa-utils.h (vtable_pointer_value_to_vtable): Declare.
6050 2014-02-03 Teresa Johnson <tejohnson@google.com>
6052 * tree-vect-slp.c (vect_supported_load_permutation_p): Avoid
6053 redef of outer loop index variable.
6055 2014-02-03 Marc Glisse <marc.glisse@inria.fr>
6059 * doc/extend.texi (Function Attributes): Typo.
6061 2014-02-03 Cong Hou <congh@google.com>
6063 PR tree-optimization/60000
6064 * tree-vect-loop.c (vect_transform_loop): Set pattern_def_seq to NULL
6065 if the vectorized statement is a store. A store statement can only
6066 appear at the end of pattern statements.
6068 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
6070 * config/i386/i386.c (flag_opts): Add -mlong-double-128.
6071 (ix86_option_override_internal): Default long double to 64-bit for
6072 32-bit Bionic and to 128-bit for 64-bit Bionic.
6074 * config/i386/i386.h (LONG_DOUBLE_TYPE_SIZE): Use 128 if
6075 TARGET_LONG_DOUBLE_128 is true.
6076 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Likewise.
6078 * config/i386/i386.opt (mlong-double-80): Negate -mlong-double-64.
6079 (mlong-double-64): Negate -mlong-double-128.
6080 (mlong-double-128): New option.
6082 * config/i386/i386-c.c (ix86_target_macros): Define
6083 __LONG_DOUBLE_128__ for TARGET_LONG_DOUBLE_128.
6085 * doc/invoke.texi: Document -mlong-double-128.
6087 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
6089 PR rtl-optimization/60024
6090 * sel-sched.c (init_regs_for_mode): Check if mode is OK first.
6092 2014-02-03 Markus Trippelsdorf <markus@trippelsdorf.de>
6094 * doc/invoke.texi (fprofile-reorder-functions): Fix typo.
6096 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
6098 PR rtl-optimization/57662
6099 * sel-sched.c (code_motion_path_driver): Do not mark already not
6100 existing blocks in the visiting bitmap.
6102 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
6104 * sel-sched-ir.c (sel_gen_insn_from_expr_after): Reset INSN_DELETED_P
6105 on the insn being emitted.
6107 2014-02-03 James Greenhalgh <james.greenhalgh@arm.com>
6108 Will Deacon <will.deacon@arm.com>
6110 * doc/gimple.texi (gimple_asm_clear_volatile): Remove.
6112 2014-02-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6114 * config/arm/arm-tables.opt: Regenerate.
6116 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6118 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): Generalize
6119 for vector types other than V16QImode.
6120 * config/rs6000/altivec.md (altivec_vperm_<mode>): Change to a
6121 define_expand, and call altivec_expand_vec_perm_le when producing
6122 code with little endian element order.
6123 (*altivec_vperm_<mode>_internal): New insn having previous
6124 behavior of altivec_vperm_<mode>.
6125 (altivec_vperm_<mode>_uns): Change to a define_expand, and call
6126 altivec_expand_vec_perm_le when producing code with little endian
6128 (*altivec_vperm_<mode>_uns_internal): New insn having previous
6129 behavior of altivec_vperm_<mode>_uns.
6131 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6133 * config/rs6000/altivec.md (UNSPEC_VSUMSWS_DIRECT): New unspec.
6134 (altivec_vsumsws): Add handling for -maltivec=be with a little
6136 (altivec_vsumsws_direct): New.
6137 (reduc_splus_<mode>): Call gen_altivec_vsumsws_direct instead of
6138 gen_altivec_vsumsws.
6140 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
6142 * ipa-devirt.c (subbinfo_with_vtable_at_offset,
6143 vtable_pointer_value_to_binfo): New functions.
6144 * ipa-utils.h (vtable_pointer_value_to_binfo): Declare.
6145 * ipa-prop.c (extr_type_from_vtbl_ptr_store): Use it.
6147 2014-02-02 Sandra Loosemore <sandra@codesourcery.com>
6149 * config/nios2/nios2.md (load_got_register): Initialize GOT
6150 pointer from _gp_got instead of _GLOBAL_OFFSET_TABLE_.
6151 * config/nios2/nios2.c (nios2_function_profiler): Likewise.
6153 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
6155 * ipa-prop.c (update_jump_functions_after_inlining): When type is not
6156 preserverd by passthrough, do not propagate the type.
6158 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
6160 * config/mips/mips.c (MIPS_GET_FCSR, MIPS_SET_FCSR): New macros.
6161 (mips_atomic_assign_expand_fenv): New function.
6162 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): Define.
6164 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
6166 * doc/extend.texi (__builtin_mips_get_fcsr): Document.
6167 (__builtin_mips_set_fcsr): Likewise.
6168 * config/mips/mips-ftypes.def: Add MIPS_VOID_FTYPE_USI and
6169 MIPS_USI_FTYPE_VOID.
6170 * config/mips/mips-protos.h (mips16_expand_get_fcsr): Declare
6171 (mips16_expand_set_fcsr): Likewise.
6172 * config/mips/mips.c (mips16_get_fcsr_stub): New variable.
6173 (mips16_set_fcsr_stub): Likewise.
6174 (mips16_get_fcsr_one_only_stub): New class.
6175 (mips16_set_fcsr_one_only_stub): Likewise.
6176 (mips16_expand_get_fcsr, mips16_expand_set_fcsr): New functions.
6177 (mips_code_end): Output the get_fcsr and set_fcsr stubs, if needed.
6178 (BUILTIN_AVAIL_MIPS16, AVAIL_ALL): New macros.
6179 (hard_float): New availability predicate.
6180 (mips_builtins): Add get_fcsr and set_fcsr.
6181 (mips_expand_builtin): Check BUILTIN_AVAIL_MIPS16.
6182 * config/mips/mips.md (UNSPEC_GET_FCSR, UNSPEC_SET_FCSR): New unspecs.
6183 (GET_FCSR_REGNUM, SET_FCSR_REGNUM): New constants.
6184 (mips_get_fcsr, *mips_get_fcsr, mips_get_fcsr_mips16_<mode>)
6185 (mips_set_fcsr, *mips_set_fcsr, mips_set_fcsr_mips16_<mode>): New
6188 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
6190 * config/mips/mips.c (mips_one_only_stub): New class.
6191 (mips_need_mips16_rdhwr_p): Replace with...
6192 (mips16_rdhwr_stub): ...this new variable.
6193 (mips16_stub_call_address): New function.
6194 (mips16_rdhwr_one_only_stub): New class.
6195 (mips_expand_thread_pointer): Use mips16_stub_call_address.
6196 (mips_output_mips16_rdhwr): Delete.
6197 (mips_finish_stub): New function.
6198 (mips_code_end): Use it to handle rdhwr stubs.
6200 2014-02-02 Uros Bizjak <ubizjak@gmail.com>
6203 * config/i386/i386.c (classify_argument): Fix handling of bit_offset
6204 when calculating size of integer atomic types.
6206 2014-02-02 H.J. Lu <hongjiu.lu@intel.com>
6208 * ipa-inline-analysis.c (true_predicate_p): Fix a typo in comments.
6210 2014-02-01 Jakub Jelinek <jakub@redhat.com>
6212 PR tree-optimization/60003
6213 * gimple-low.c (lower_builtin_setjmp): Set cfun->has_nonlocal_label.
6214 * profile.c (branch_prob): Use gimple_call_builtin_p
6215 to check for BUILT_IN_SETJMP_RECEIVER.
6216 * tree-inline.c (copy_bb): Call notice_special_calls.
6218 2014-01-31 Vladimir Makarov <vmakarov@redhat.com>
6221 * lra-constraints.c (process_alt_operands): Update reload_sum only
6224 2014-01-31 Richard Henderson <rth@redhat.com>
6227 * tree-eh.c (lower_try_finally_switch): Delay lowering finally block
6228 until after else_eh is processed.
6230 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
6232 * config/i386/avx512fintrin.h (_MM_FROUND_TO_NEAREST_INT),
6233 (_MM_FROUND_TO_NEG_INF), (_MM_FROUND_TO_POS_INF),
6234 (_MM_FROUND_TO_ZERO), (_MM_FROUND_CUR_DIRECTION): Are already defined
6235 in smmintrin.h, remove them.
6236 (_MM_FROUND_NO_EXC): Same as above, bit also wrong value.
6237 * config/i386/i386.c (ix86_print_operand): Split sae and rounding.
6238 * config/i386/i386.md (ROUND_SAE): Fix value.
6239 * config/i386/predicates.md (const_4_or_8_to_11_operand): New.
6240 (const48_operand): New.
6241 * config/i386/subst.md (round), (round_expand): Use
6242 const_4_or_8_to_11_operand.
6243 (round_saeonly), (round_saeonly_expand): Use const48_operand.
6245 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
6247 * config/i386/constraints.md (Yk): Swap meaning with k.
6248 * config/i386/i386.md (movhi_internal): Change Yk to k.
6249 (movqi_internal): Ditto.
6250 (*k<logic><mode>): Ditto.
6253 (kandn<mode>): Ditto.
6254 (*<code>hi_1): Ditto.
6255 (*<code>qi_1): Ditto.
6256 (kxnor<mode>): Ditto.
6257 (kortestzhi): Ditto.
6258 (kortestchi): Ditto.
6260 (*one_cmplhi2_1): Ditto.
6261 (*one_cmplqi2_1): Ditto.
6262 * config/i386/sse.md (): Change k to Yk.
6263 (avx512f_load<mode>_mask): Ditto.
6264 (avx512f_blendm<mode>): Ditto.
6265 (avx512f_store<mode>_mask): Ditto.
6266 (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
6267 (avx512f_storedqu<mode>_mask): Ditto.
6268 (avx512f_cmp<mode>3<mask_scalar_merge_name><round_saeonly_name>):
6270 (avx512f_ucmp<mode>3<mask_scalar_merge_name>): Ditto.
6271 (avx512f_vmcmp<mode>3<round_saeonly_name>): Ditto.
6272 (avx512f_vmcmp<mode>3_mask<round_saeonly_name>): Ditto.
6273 (avx512f_maskcmp<mode>3): Ditto.
6274 (avx512f_fmadd_<mode>_mask<round_name>): Ditto.
6275 (avx512f_fmadd_<mode>_mask3<round_name>): Ditto.
6276 (avx512f_fmsub_<mode>_mask<round_name>): Ditto.
6277 (avx512f_fmsub_<mode>_mask3<round_name>): Ditto.
6278 (avx512f_fnmadd_<mode>_mask<round_name>): Ditto.
6279 (avx512f_fnmadd_<mode>_mask3<round_name>): Ditto.
6280 (avx512f_fnmsub_<mode>_mask<round_name>): Ditto.
6281 (avx512f_fnmsub_<mode>_mask3<round_name>): Ditto.
6282 (avx512f_fmaddsub_<mode>_mask<round_name>): Ditto.
6283 (avx512f_fmaddsub_<mode>_mask3<round_name>): Ditto.
6284 (avx512f_fmsubadd_<mode>_mask<round_name>): Ditto.
6285 (avx512f_fmsubadd_<mode>_mask3<round_name>): Ditto.
6286 (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
6287 (vec_extract_lo_<mode>_maskm): Ditto.
6288 (vec_extract_hi_<mode>_maskm): Ditto.
6289 (avx512f_vternlog<mode>_mask): Ditto.
6290 (avx512f_fixupimm<mode>_mask<round_saeonly_name>): Ditto.
6291 (avx512f_sfixupimm<mode>_mask<round_saeonly_name>): Ditto.
6292 (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
6293 (avx512f_<code>v8div16qi2_mask): Ditto.
6294 (avx512f_<code>v8div16qi2_mask_store): Ditto.
6295 (avx512f_eq<mode>3<mask_scalar_merge_name>_1): Ditto.
6296 (avx512f_gt<mode>3<mask_scalar_merge_name>): Ditto.
6297 (avx512f_testm<mode>3<mask_scalar_merge_name>): Ditto.
6298 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Ditto.
6299 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
6300 (*avx512pf_gatherpf<mode>df_mask): Ditto.
6301 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
6302 (*avx512pf_scatterpf<mode>df_mask): Ditto.
6303 (avx512cd_maskb_vec_dupv8di): Ditto.
6304 (avx512cd_maskw_vec_dupv16si): Ditto.
6305 (avx512f_vpermi2var<mode>3_maskz): Ditto.
6306 (avx512f_vpermi2var<mode>3_mask): Ditto.
6307 (avx512f_vpermi2var<mode>3_mask): Ditto.
6308 (avx512f_vpermt2var<mode>3_maskz): Ditto.
6309 (*avx512f_gathersi<mode>): Ditto.
6310 (*avx512f_gathersi<mode>_2): Ditto.
6311 (*avx512f_gatherdi<mode>): Ditto.
6312 (*avx512f_gatherdi<mode>_2): Ditto.
6313 (*avx512f_scattersi<mode>): Ditto.
6314 (*avx512f_scatterdi<mode>): Ditto.
6315 (avx512f_compress<mode>_mask): Ditto.
6316 (avx512f_compressstore<mode>_mask): Ditto.
6317 (avx512f_expand<mode>_mask): Ditto.
6318 * config/i386/subst.md (mask): Change k to Yk.
6319 (mask_scalar_merge): Ditto.
6322 2014-01-31 Marc Glisse <marc.glisse@inria.fr>
6324 * doc/extend.texi (Vector Extensions): Document ?: in C++.
6326 2014-01-31 Richard Biener <rguenther@suse.de>
6329 * builtins.c (fold_builtin_memory_op): Make sure to not
6330 use a floating-point mode or a boolean or enumeral type for
6333 2014-01-30 DJ Delorie <dj@redhat.com>
6335 * config/msp430/msp430.h (LIB_SPEC): Add -lcrt
6336 * config/msp430/msp430.md (msp430_refsym_need_exit): New.
6337 * config/msp430/msp430.c (msp430_expand_epilogue): Call it
6338 whenever main() has an epilogue.
6340 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6342 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Remove
6343 unused variable "field".
6344 * config/rs6000/vsx.md (vsx_mergel_<mode>): Add missing DONE.
6345 (vsx_mergeh_<mode>): Likewise.
6346 * config/rs6000/altivec.md (altivec_vmrghb): Likewise.
6347 (altivec_vmrghh): Likewise.
6348 (altivec_vmrghw): Likewise.
6349 (altivec_vmrglb): Likewise.
6350 (altivec_vmrglh): Likewise.
6351 (altivec_vmrglw): Likewise.
6352 (altivec_vspltb): Add missing uses.
6353 (altivec_vsplth): Likewise.
6354 (altivec_vspltw): Likewise.
6355 (altivec_vspltsf): Likewise.
6357 2014-01-30 Jakub Jelinek <jakub@redhat.com>
6360 * ifcvt.c (cond_exec_process_insns): Don't conditionalize
6361 frame related instructions.
6363 2014-01-30 Vladimir Makarov <vmakarov@redhat.com>
6365 PR rtl-optimization/59959
6366 * lra-constrains.c (simplify_operand_subreg): Assign NO_REGS to
6367 any reload of register whose subreg is invalid.
6369 2014-01-30 Jakub Jelinek <jakub@redhat.com>
6371 * config/i386/f16cintrin.h (_cvtsh_ss): Avoid -Wnarrowing warning.
6372 * config/i386/avx512fintrin.h (_mm512_mask_cvtusepi64_storeu_epi32):
6373 Add missing return type - void.
6375 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6377 * gcc/config/rs6000/rs6000.c (rs6000_expand_vector_init): Use
6378 gen_vsx_xxspltw_v4sf_direct instead of gen_vsx_xxspltw_v4sf;
6379 remove element index adjustment for endian (now handled in vsx.md
6381 (altivec_expand_vec_perm_const): Use
6382 gen_altivec_vsplt[bhw]_direct instead of gen_altivec_vsplt[bhw].
6383 * gcc/config/rs6000/vsx.md (UNSPEC_VSX_XXSPLTW): New unspec.
6384 (vsx_xxspltw_<mode>): Adjust element index for little endian.
6385 * gcc/config/rs6000/altivec.md (altivec_vspltb): Divide into a
6386 define_expand and a new define_insn *altivec_vspltb_internal;
6387 adjust for -maltivec=be on a little endian target.
6388 (altivec_vspltb_direct): New.
6389 (altivec_vsplth): Divide into a define_expand and a new
6390 define_insn *altivec_vsplth_internal; adjust for -maltivec=be on a
6391 little endian target.
6392 (altivec_vsplth_direct): New.
6393 (altivec_vspltw): Divide into a define_expand and a new
6394 define_insn *altivec_vspltw_internal; adjust for -maltivec=be on a
6395 little endian target.
6396 (altivec_vspltw_direct): New.
6397 (altivec_vspltsf): Divide into a define_expand and a new
6398 define_insn *altivec_vspltsf_internal; adjust for -maltivec=be on
6399 a little endian target.
6401 2014-01-30 Richard Biener <rguenther@suse.de>
6403 PR tree-optimization/59993
6404 * tree-ssa-forwprop.c (associate_pointerplus): Check we
6405 can propagate form the earlier stmt and avoid the transform
6406 when the intermediate result is needed.
6408 2014-01-30 Alangi Derick <alangiderick@gmail.com>
6410 * README.Portability: Fix typo.
6412 2014-01-30 David Holsgrove <david.holsgrove@xilinx.com>
6414 * config/microblaze/microblaze.md(cstoresf4, cbranchsf4): Replace
6415 comparison_operator with ordered_comparison_operator.
6417 2014-01-30 Nick Clifton <nickc@redhat.com>
6419 * config/mn10300/mn10300-protos.h (mn10300_store_multiple_operation_p):
6420 Rename to mn10300_store_multiple_regs.
6421 * config/mn10300/mn10300.c: Likewise.
6422 * config/mn10300/mn10300.md (store_movm): Fix typo: call
6423 store_multiple_regs.
6424 * config/mn10300/predicates.md (mn10300_store_multiple_operation):
6425 Call mn10300_store_multiple_regs.
6427 2014-01-30 Nick Clifton <nickc@redhat.com>
6428 DJ Delorie <dj@redhat.com>
6430 * config/rl78/rl78.c (register_sizes): Make the "upper half" of
6431 %fp 2 to keep registers after it properly word-aligned.
6432 (rl78_alloc_physical_registers_umul): Handle the case where both
6433 input operands are the same.
6435 2014-01-30 Richard Biener <rguenther@suse.de>
6437 PR tree-optimization/59903
6438 * tree-vect-loop.c (vect_transform_loop): Guard multiple-types
6441 2014-01-30 Jason Merrill <jason@redhat.com>
6444 * tree.c (walk_type_fields): Handle VECTOR_TYPE.
6447 * cgraphunit.c (expand_thunk): Copy volatile arg to a temporary.
6449 2014-01-30 Richard Biener <rguenther@suse.de>
6451 PR tree-optimization/59951
6452 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Skip uses in debug insns.
6454 2014-01-30 Savin Zlobec <savin.zlobec@gmail.com>
6457 * config/nios2/nios2.c (nios2_fpu_insn_asm): Fix asm output of
6458 SFmode to DFmode case.
6460 2014-01-29 DJ Delorie <dj@redhat.com>
6462 * config/msp430/msp430.opt (-minrt): New.
6463 * config/msp430/msp430.h (STARTFILE_SPEC): Link alternate runtime
6465 (ENDFILE_SPEC): Likewise.
6467 2014-01-29 Jan Hubicka <hubicka@ucw.cz>
6469 * ipa-inline-analysis.c (clobber_only_eh_bb_p): New function.
6470 (estimate_function_body_sizes): Use it.
6472 2014-01-29 Paolo Carlini <paolo.carlini@oracle.com>
6475 * dwarf2out.c (is_cxx_auto): New.
6476 (is_base_type): Use it.
6477 (gen_type_die_with_usage): Likewise.
6479 2014-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6481 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Use
6482 CODE_FOR_altivec_vmrg*_direct rather than CODE_FOR_altivec_vmrg*.
6483 * config/rs6000/vsx.md (vsx_mergel_<mode>): Adjust for
6484 -maltivec=be with LE targets.
6485 (vsx_mergeh_<mode>): Likewise.
6486 * config/rs6000/altivec.md (UNSPEC_VMRG[HL]_DIRECT): New unspecs.
6487 (mulv8hi3): Use gen_altivec_vmrg[hl]w_direct.
6488 (altivec_vmrghb): Replace with define_expand and new
6489 *altivec_vmrghb_internal insn; adjust for -maltivec=be with LE targets.
6490 (altivec_vmrghb_direct): New define_insn.
6491 (altivec_vmrghh): Replace with define_expand and new
6492 *altivec_vmrghh_internal insn; adjust for -maltivec=be with LE targets.
6493 (altivec_vmrghh_direct): New define_insn.
6494 (altivec_vmrghw): Replace with define_expand and new
6495 *altivec_vmrghw_internal insn; adjust for -maltivec=be with LE targets.
6496 (altivec_vmrghw_direct): New define_insn.
6497 (*altivec_vmrghsf): Adjust for endianness.
6498 (altivec_vmrglb): Replace with define_expand and new
6499 *altivec_vmrglb_internal insn; adjust for -maltivec=be with LE targets.
6500 (altivec_vmrglb_direct): New define_insn.
6501 (altivec_vmrglh): Replace with define_expand and new
6502 *altivec_vmrglh_internal insn; adjust for -maltivec=be with LE targets.
6503 (altivec_vmrglh_direct): New define_insn.
6504 (altivec_vmrglw): Replace with define_expand and new
6505 *altivec_vmrglw_internal insn; adjust for -maltivec=be with LE targets.
6506 (altivec_vmrglw_direct): New define_insn.
6507 (*altivec_vmrglsf): Adjust for endianness.
6508 (vec_widen_umult_hi_v16qi): Use gen_altivec_vmrghh_direct.
6509 (vec_widen_umult_lo_v16qi): Use gen_altivec_vmrglh_direct.
6510 (vec_widen_smult_hi_v16qi): Use gen_altivec_vmrghh_direct.
6511 (vec_widen_smult_lo_v16qi): Use gen_altivec_vmrglh_direct.
6512 (vec_widen_umult_hi_v8hi): Use gen_altivec_vmrghw_direct.
6513 (vec_widen_umult_lo_v8hi): Use gen_altivec_vmrglw_direct.
6514 (vec_widen_smult_hi_v8hi): Use gen_altivec_vmrghw_direct.
6515 (vec_widen_smult_lo_v8hi): Use gen_altivec_vmrglw_direct.
6517 2014-01-29 Marcus Shawcroft <marcus.shawcroft@arm.com>
6519 * config/aarch64/aarch64.c (aarch64_expand_mov_immediate)
6520 (aarch64_legitimate_address_p, aarch64_class_max_nregs): Adjust
6523 2014-01-29 Richard Biener <rguenther@suse.de>
6525 PR tree-optimization/58742
6526 * tree-ssa-forwprop.c (associate_pointerplus): Rename to
6527 associate_pointerplus_align.
6528 (associate_pointerplus_diff): New function.
6529 (associate_pointerplus): Likewise. Call associate_pointerplus_align
6530 and associate_pointerplus_diff.
6532 2014-01-29 Richard Biener <rguenther@suse.de>
6534 * lto-streamer.h (LTO_major_version): Bump to 3.
6535 (LTO_minor_version): Reset to 0.
6537 2014-01-29 Renlin Li <Renlin.Li@arm.com>
6539 * config/arm/arm-arches.def (ARM_ARCH): Add armv7ve arch.
6540 * config/arm/arm.c (FL_FOR_ARCH7VE): New.
6541 (arm_file_start): Generate correct asm header for armv7ve.
6542 * config/arm/bpabi.h: Add multilib support for armv7ve.
6543 * config/arm/driver-arm.c: Change the architectures of cortex-a7
6544 and cortex-a15 to armv7ve.
6545 * config/arm/t-aprofile: Add multilib support for armv7ve.
6546 * doc/invoke.texi: Document -march=armv7ve.
6548 2014-01-29 Richard Biener <rguenther@suse.de>
6550 PR tree-optimization/58742
6551 * tree-ssa-forwprop.c (associate_plusminus): Return true
6552 if we changed sth, defer EH cleanup to ...
6553 (ssa_forward_propagate_and_combine): ... here. Call simplify_mult.
6554 (simplify_mult): New function.
6556 2014-01-29 Jakub Jelinek <jakub@redhat.com>
6559 PR tree-optimization/59920
6560 * tree.c (build_common_builtin_nodes): Remove
6561 __builtin_setjmp_dispatcher initialization.
6562 * omp-low.h (make_gimple_omp_edges): Add a new int * argument.
6563 * profile.c (branch_prob): Use gsi_start_nondebug_after_labels_bb
6564 instead of gsi_after_labels + manually skipping debug stmts.
6565 Don't ignore bbs with BUILT_IN_SETJMP_DISPATCHER, instead
6566 ignore bbs with IFN_ABNORMAL_DISPATCHER.
6567 * tree-inline.c (copy_edges_for_bb): Remove
6568 can_make_abnormal_goto argument, instead add abnormal_goto_dest
6569 argument. Ignore computed_goto_p stmts. Don't call
6570 make_abnormal_goto_edges. If a call might need abnormal edges
6571 for non-local gotos, see if it already has an edge to
6572 IFN_ABNORMAL_DISPATCHER or if it is IFN_ABNORMAL_DISPATCHER
6573 with true argument, don't do anything then, otherwise add
6574 EDGE_ABNORMAL from the call's bb to abnormal_goto_dest.
6575 (copy_cfg_body): Compute abnormal_goto_dest, adjust copy_edges_for_bb
6577 * gimple-low.c (struct lower_data): Remove calls_builtin_setjmp.
6578 (lower_function_body): Don't emit __builtin_setjmp_dispatcher.
6579 (lower_stmt): Don't set data->calls_builtin_setjmp.
6580 (lower_builtin_setjmp): Adjust comment.
6581 * builtins.def (BUILT_IN_SETJMP_DISPATCHER): Remove.
6582 * tree-cfg.c (found_computed_goto): Remove.
6583 (factor_computed_gotos): Remove.
6584 (make_goto_expr_edges): Return bool, true for computed gotos.
6585 Don't call make_abnormal_goto_edges.
6586 (build_gimple_cfg): Don't set found_computed_goto, don't call
6587 factor_computed_gotos.
6588 (computed_goto_p): No longer static.
6589 (make_blocks): Don't set found_computed_goto.
6590 (get_abnormal_succ_dispatcher, handle_abnormal_edges): New functions.
6591 (make_edges): If make_goto_expr_edges returns true, push bb
6592 into ab_edge_goto vector, for stmt_can_make_abnormal_goto calls
6593 instead of calling make_abnormal_goto_edges push bb into ab_edge_call
6594 vector. Record mapping between bbs and OpenMP regions if there
6595 are any, adjust make_gimple_omp_edges caller. Call
6596 handle_abnormal_edges.
6597 (make_abnormal_goto_edges): Remove.
6598 * tree-cfg.h (make_abnormal_goto_edges): Remove.
6599 (computed_goto_p, get_abnormal_succ_dispatcher): New prototypes.
6600 * internal-fn.c (expand_ABNORMAL_DISPATCHER): New function.
6601 * builtins.c (expand_builtin): Don't handle BUILT_IN_SETJMP_DISPATCHER.
6602 * internal-fn.def (ABNORMAL_DISPATCHER): New.
6603 * omp-low.c (make_gimple_omp_edges): Add region_idx argument, when
6604 filling *region also set *region_idx to (*region)->entry->index.
6607 * read-rtl.c (read_rtx_code): Clear all of RTX_CODE_SIZE (code).
6608 For REGs set ORIGINAL_REGNO.
6610 2014-01-29 Bingfeng Mei <bmei@broadcom.com>
6612 * doc/md.texi: Mention that a target shouldn't implement
6613 vec_widen_(s|u)mul_even/odd pair if it is less efficient
6616 2014-01-29 Jakub Jelinek <jakub@redhat.com>
6618 PR tree-optimization/59594
6619 * tree-vect-data-refs.c (vect_analyze_data_ref_accesses): Sort
6620 a copy of the datarefs vector rather than the vector itself.
6622 2014-01-28 Jason Merrill <jason@redhat.com>
6625 * dwarf2out.c (auto_die): New static.
6626 (gen_type_die_with_usage): Handle C++1y 'auto'.
6627 (gen_subprogram_die): If in-class DIE had 'auto', emit type again
6630 2014-01-28 H.J. Lu <hongjiu.lu@intel.com>
6633 * config/i386/gnu-user64.h (SPEC_32): Add "m16|" to "m32".
6634 (SPEC_X32): Likewise.
6635 (SPEC_64): Likewise.
6636 * config/i386/i386.c (ix86_option_override_internal): Turn off
6637 OPTION_MASK_ISA_64BIT, OPTION_MASK_ABI_X32 and OPTION_MASK_ABI_64
6639 (x86_file_start): Output .code16gcc for TARGET_16BIT.
6640 * config/i386/i386.h (TARGET_16BIT): New macro.
6641 (TARGET_16BIT_P): Likewise.
6642 * config/i386/i386.opt: Add m16.
6643 * doc/invoke.texi: Document -m16.
6645 2014-01-28 Jakub Jelinek <jakub@redhat.com>
6647 PR preprocessor/59935
6648 * input.c (location_get_source_line): Bail out on when line number
6649 is zero, and test the return value of lookup_or_add_file_to_cache_tab.
6651 2014-01-28 Richard Biener <rguenther@suse.de>
6653 PR tree-optimization/58742
6654 * tree-ssa-forwprop.c (associate_plusminus): Handle
6655 pointer subtraction of the form (T)(P + A) - (T)P.
6657 2014-01-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6659 * config/arm/arm.c (arm_new_rtx_costs): Remove useless statement
6662 2014-01-28 Richard Biener <rguenther@suse.de>
6665 2014-01-28 Richard Biener <rguenther@suse.de>
6667 PR rtl-optimization/45364
6668 PR rtl-optimization/59890
6669 * var-tracking.c (local_get_addr_clear_given_value): Handle
6670 already cleared slot.
6671 (val_reset): Handle not allocated local_get_addr_cache.
6672 (vt_find_locations): Use post-order on the inverted CFG.
6674 2014-01-28 Richard Biener <rguenther@suse.de>
6676 * tree-data-ref.h (ddr_is_anti_dependent, ddrs_have_anti_deps): Remove.
6678 2014-01-28 Richard Biener <rguenther@suse.de>
6680 PR rtl-optimization/45364
6681 PR rtl-optimization/59890
6682 * var-tracking.c (local_get_addr_clear_given_value): Handle
6683 already cleared slot.
6684 (val_reset): Handle not allocated local_get_addr_cache.
6685 (vt_find_locations): Use post-order on the inverted CFG.
6687 2014-01-28 Alan Modra <amodra@gmail.com>
6689 * Makefile.in (BUILD_CPPFLAGS): Do not use ALL_CPPFLAGS.
6690 * configure.ac <recursive call for build != host>: Define
6691 GENERATOR_FILE. Comment. Use CXX_FOR_BUILD, CXXFLAGS_FOR_BUILD
6692 and LD_FOR_BUILD too.
6693 * configure: Regenerate.
6695 2014-01-27 Allan Sandfeld Jensen <sandfeld@kde.org>
6697 * config/i386/i386.c (get_builtin_code_for_version): Separate
6698 Westmere from Nehalem, Ivy Bridge from Sandy Bridge and
6699 Broadwell from Haswell.
6701 2014-01-27 Steve Ellcey <sellcey@mips.com>
6703 * common/config/mips/mips-common.c (TARGET_DEFAULT_TARGET_FLAGS):
6704 Remove TARGET_FP_EXCEPTIONS_DEFAULT and MASK_FUSED_MADD.
6705 * config/mips/mips.c (mips_option_override): Change setting
6707 * config/mips/mips.h (TARGET_FP_EXCEPTIONS_DEFAULT): Remove.
6708 * config/mips/mips.opt (DSP, DSPR2, FP_EXCEPTIONS, FUSED_MADD, MIPS3D):
6709 Change from Mask to Var.
6711 2014-01-27 Jeff Law <law@redhat.com>
6713 * ipa-inline.c (inline_small_functions): Fix typo.
6715 2014-01-27 Ilya Tocar <ilya.tocar@intel.com>
6717 * config/i386/avx512fintrin.h (_mm512_mask_cvtepi32_storeu_epi8): New.
6718 (_mm512_mask_cvtsepi32_storeu_epi8): Ditto.
6719 (_mm512_mask_cvtusepi32_storeu_epi8): Ditto.
6720 (_mm512_mask_cvtepi32_storeu_epi16): Ditto.
6721 (_mm512_mask_cvtsepi32_storeu_epi16): Ditto.
6722 (_mm512_mask_cvtusepi32_storeu_epi16): Ditto.
6723 (_mm512_mask_cvtepi64_storeu_epi32): Ditto.
6724 (_mm512_mask_cvtsepi64_storeu_epi32): Ditto.
6725 (_mm512_mask_cvtusepi64_storeu_epi32): Ditto.
6726 (_mm512_mask_cvtepi64_storeu_epi16): Ditto.
6727 (_mm512_mask_cvtsepi64_storeu_epi16): Ditto.
6728 (_mm512_mask_cvtusepi64_storeu_epi16): Ditto.
6729 (_mm512_mask_cvtepi64_storeu_epi8): Ditto.
6730 (_mm512_mask_cvtsepi64_storeu_epi8): Ditto.
6731 (_mm512_mask_cvtusepi64_storeu_epi8): Ditto.
6732 (_mm512_storeu_epi64): Ditto.
6733 (_mm512_cmpge_epi32_mask): Ditto.
6734 (_mm512_cmpge_epu32_mask): Ditto.
6735 (_mm512_cmpge_epi64_mask): Ditto.
6736 (_mm512_cmpge_epu64_mask): Ditto.
6737 (_mm512_cmple_epi32_mask): Ditto.
6738 (_mm512_cmple_epu32_mask): Ditto.
6739 (_mm512_cmple_epi64_mask): Ditto.
6740 (_mm512_cmple_epu64_mask): Ditto.
6741 (_mm512_cmplt_epi32_mask): Ditto.
6742 (_mm512_cmplt_epu32_mask): Ditto.
6743 (_mm512_cmplt_epi64_mask): Ditto.
6744 (_mm512_cmplt_epu64_mask): Ditto.
6745 (_mm512_cmpneq_epi32_mask): Ditto.
6746 (_mm512_cmpneq_epu32_mask): Ditto.
6747 (_mm512_cmpneq_epi64_mask): Ditto.
6748 (_mm512_cmpneq_epu64_mask): Ditto.
6749 (_mm512_expand_pd): Ditto.
6750 (_mm512_expand_ps): Ditto.
6751 * config/i386/i386-builtin-types.def: Add PV16QI, PV16QI, PV16HI,
6752 VOID_PV8SI_V8DI_QI, VOID_PV8HI_V8DI_QI, VOID_PV16QI_V8DI_QI,
6753 VOID_PV16QI_V16SI_HI, VOID_PV16HI_V16SI_HI.
6754 * config/i386/i386.c (ix86_builtins): Add
6755 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK,
6756 IX86_BUILTIN_PMOVDB512_MEM, IX86_BUILTIN_PMOVDW512_MEM,
6757 IX86_BUILTIN_PMOVQB512_MEM, IX86_BUILTIN_PMOVQD512_MEM,
6758 IX86_BUILTIN_PMOVQW512_MEM, IX86_BUILTIN_PMOVSDB512_MEM,
6759 IX86_BUILTIN_PMOVSDW512_MEM, IX86_BUILTIN_PMOVSQB512_MEM,
6760 IX86_BUILTIN_PMOVSQD512_MEM, IX86_BUILTIN_PMOVSQW512_MEM,
6761 IX86_BUILTIN_PMOVUSDB512_MEM, IX86_BUILTIN_PMOVUSDW512_MEM,
6762 IX86_BUILTIN_PMOVUSQB512_MEM, IX86_BUILTIN_PMOVUSQD512_MEM,
6763 IX86_BUILTIN_PMOVUSQW512_MEM.
6764 (bdesc_special_args): Add __builtin_ia32_pmovusqd512mem_mask,
6765 __builtin_ia32_pmovsqd512mem_mask,
6766 __builtin_ia32_pmovqd512mem_mask,
6767 __builtin_ia32_pmovusqw512mem_mask,
6768 __builtin_ia32_pmovsqw512mem_mask,
6769 __builtin_ia32_pmovqw512mem_mask,
6770 __builtin_ia32_pmovusdw512mem_mask,
6771 __builtin_ia32_pmovsdw512mem_mask,
6772 __builtin_ia32_pmovdw512mem_mask,
6773 __builtin_ia32_pmovqb512mem_mask,
6774 __builtin_ia32_pmovusqb512mem_mask,
6775 __builtin_ia32_pmovsqb512mem_mask,
6776 __builtin_ia32_pmovusdb512mem_mask,
6777 __builtin_ia32_pmovsdb512mem_mask,
6778 __builtin_ia32_pmovdb512mem_mask.
6779 (bdesc_args): Add __builtin_ia32_expanddf512,
6780 __builtin_ia32_expandsf512.
6781 (ix86_expand_special_args_builtin): Handle VOID_FTYPE_PV8SI_V8DI_QI,
6782 VOID_FTYPE_PV8HI_V8DI_QI, VOID_FTYPE_PV16HI_V16SI_HI,
6783 VOID_FTYPE_PV16QI_V8DI_QI, VOID_FTYPE_PV16QI_V16SI_HI.
6784 * config/i386/sse.md (unspec): Add UNSPEC_EXPAND_NOMASK.
6785 (avx512f_<code><pmov_src_lower><mode>2_mask_store): New.
6786 (*avx512f_<code>v8div16qi2_store_mask): Renamed to ...
6787 (avx512f_<code>v8div16qi2_mask_store): This.
6788 (avx512f_expand<mode>): New.
6790 2014-01-27 Kirill Yukhin <kirill.yukhin@intel.com>
6792 * config/i386/avx512pfintrin.h (_mm512_mask_prefetch_i32gather_pd):
6794 (_mm512_mask_prefetch_i64gather_pd): Ditto.
6795 (_mm512_prefetch_i32scatter_pd): Ditto.
6796 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
6797 (_mm512_prefetch_i64scatter_pd): Ditto.
6798 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
6799 (_mm512_mask_prefetch_i32gather_ps): Fix operand type.
6800 (_mm512_mask_prefetch_i64gather_ps): Ditto.
6801 (_mm512_prefetch_i32scatter_ps): Ditto.
6802 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
6803 (_mm512_prefetch_i64scatter_ps): Ditto.
6804 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
6805 * config/i386/i386-builtin-types.def: Define
6806 VOID_FTYPE_QI_V8SI_PCINT64_INT_INT
6807 and VOID_FTYPE_QI_V8DI_PCINT64_INT_INT.
6808 * config/i386/i386.c (ix86_builtins): Define IX86_BUILTIN_GATHERPFQPD,
6809 IX86_BUILTIN_GATHERPFDPD, IX86_BUILTIN_SCATTERPFDPD,
6810 IX86_BUILTIN_SCATTERPFQPD.
6811 (ix86_init_mmx_sse_builtins): Define __builtin_ia32_gatherpfdpd,
6812 __builtin_ia32_gatherpfdps, __builtin_ia32_gatherpfqpd,
6813 __builtin_ia32_gatherpfqps, __builtin_ia32_scatterpfdpd,
6814 __builtin_ia32_scatterpfdps, __builtin_ia32_scatterpfqpd,
6815 __builtin_ia32_scatterpfqps.
6816 (ix86_expand_builtin): Expand new built-ins.
6817 * config/i386/sse.md (avx512pf_gatherpf<mode>): Add SF suffix,
6818 fix memory access data type.
6819 (*avx512pf_gatherpf<mode>_mask): Ditto.
6820 (*avx512pf_gatherpf<mode>): Ditto.
6821 (avx512pf_scatterpf<mode>): Ditto.
6822 (*avx512pf_scatterpf<mode>_mask): Ditto.
6823 (*avx512pf_scatterpf<mode>): Ditto.
6824 (GATHER_SCATTER_SF_MEM_MODE): New.
6825 (avx512pf_gatherpf<mode>df): Ditto.
6826 (*avx512pf_gatherpf<mode>df_mask): Ditto.
6827 (*avx512pf_scatterpf<mode>df): Ditto.
6829 2014-01-27 Jakub Jelinek <jakub@redhat.com>
6832 * expmed.h (expmed_mode_index): Rework so that analysis and optimziers
6833 know when the MODE_PARTIAL_INT and MODE_VECTOR_INT cases can never be
6836 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
6838 * common/config/arm/arm-common.c
6839 (arm_rewrite_mcpu): Handle multiple names.
6841 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
6843 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
6845 * gimple-builder.h (create_gimple_tmp): Delete.
6847 2014-01-27 Christian Bruel <christian.bruel@st.com>
6849 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Fix remaining bytes after
6852 2014-01-26 John David Anglin <danglin@gcc.gnu.org>
6854 * config/pa/pa.md (call): Generate indirect long calls to non-local
6855 functions when outputing 32-bit code.
6856 (call_value): Likewise except for special call to buggy powf function.
6858 * config/pa/pa.c (pa_attr_length_indirect_call): Adjust length of
6859 portable runtime and PIC indirect calls.
6860 (pa_output_indirect_call): Remove unnecessary nop from portable runtime
6861 and PIC call sequences. Use ldo instead of blr to set return register
6862 in PIC call sequence.
6864 2014-01-25 Walter Lee <walt@tilera.com>
6866 * config/tilegx/sync.md (atomic_fetch_sub): Fix negation and
6867 avoid clobbering a live register.
6869 2014-01-25 Walter Lee <walt@tilera.com>
6871 * config/tilegx/tilegx-c.c (tilegx_cpu_cpp_builtins):
6872 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2}.
6873 * config/tilegx/tilepro-c.c (tilepro_cpu_cpp_builtins):
6874 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2,4,8}.
6876 2014-01-25 Walter Lee <walt@tilera.com>
6878 * config/tilegx/tilegx.c (tilegx_function_arg): Start 16-byte
6879 arguments on even registers.
6880 (tilegx_gimplify_va_arg_expr): Align 16-byte var args to
6882 * config/tilegx/tilegx.h (STACK_BOUNDARY): Change to 16 bytes.
6883 (BIGGEST_ALIGNMENT): Ditto.
6884 (BIGGEST_FIELD_ALIGNMENT): Ditto.
6886 2014-01-25 Walter Lee <walt@tilera.com>
6888 * config/tilegx/tilegx.c (tilegx_gen_bundles): Delete barrier
6889 insns before bundling.
6890 * config/tilegx/tilegx.md (tile_network_barrier): Update comment.
6892 2014-01-25 Walter Lee <walt@tilera.com>
6894 * config/tilegx/tilegx.c (tilegx_expand_builtin): Set
6895 PREFETCH_SCHEDULE_BARRIER_P to true for prefetches.
6896 * config/tilepro/tilepro.c (tilepro_expand_builtin): Ditto.
6898 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
6900 * config/mips/constraints.md (kl): Delete.
6901 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Turn into
6902 define expands, using...
6903 (divmod<mode>4_mips16, udivmod<mode>4_mips16): ...these new
6904 instructions for MIPS16.
6905 (*divmod<mode>4, *udivmod<mode>4): New patterns, taken from the
6906 non-MIPS16 version of the old divmod<mode>4 and udivmod<mode>4.
6908 2014-01-25 Walter Lee <walt@tilera.com>
6910 * config/tilepro/tilepro.md (ctzdi2): Use register_operand predicate.
6914 2014-01-25 Walter Lee <walt@tilera.com>
6916 * config/tilegx/tilegx.c (tilegx_expand_to_rtl_hook): New.
6917 (TARGET_EXPAND_TO_RTL_HOOK): Define.
6919 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
6921 * rtlanal.c (canonicalize_condition): Split out duplicated mode check.
6924 2014-01-25 Jakub Jelinek <jakub@redhat.com>
6926 * print-rtl.c (in_call_function_usage): New var.
6927 (print_rtx): When in CALL_INSN_FUNCTION_USAGE, always print
6928 EXPR_LIST mode as mode and not as reg note name.
6931 * cfgloopmanip.c (copy_loop_info): If
6932 loop->warned_aggressive_loop_optimizations, make sure
6933 the flag is set in target loop too.
6935 2014-01-24 Balaji V. Iyer <balaji.v.iyer@intel.com>
6937 * builtins.c (is_builtin_name): Renamed flag_enable_cilkplus to
6939 * builtins.def: Likewise.
6940 * cilk.h (fn_contains_cilk_spawn_p): Likewise.
6941 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Likewise.
6942 * ira.c (ira_setup_eliminable_regset): Likewise.
6943 * omp-low.c (gate_expand_omp): Likewise.
6944 (execute_lower_omp): Likewise.
6945 (diagnose_sb_0): Likewise.
6946 (gate_diagnose_omp_blocks): Likewise.
6947 (simd_clone_clauses_extract): Likewise.
6950 2014-01-24 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6952 * config/rs6000/rs6000.c (rs6000_expand_vec_perm_const_1): Remove
6953 correction for little endian...
6954 * config/rs6000/vsx.md (vsx_xxpermdi2_<mode>_1): ...and move it to
6957 2014-01-24 Jeff Law <law@redhat.com>
6959 PR tree-optimization/59919
6960 * tree-vrp.c (find_assert_locations_1): Do not register asserts
6961 for non-returning calls.
6963 2014-01-24 James Greenhalgh <james.greenhalgh@arm.com>
6965 * common/config/aarch64/aarch64-common.c
6966 (aarch64_rewrite_mcpu): Handle multiple names.
6967 * config/aarch64/aarch64.h
6968 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
6970 2014-01-24 Dodji Seketeli <dodji@redhat.com>
6972 * input.c (add_file_to_cache_tab): Handle the case where fopen
6975 2014-01-23 H.J. Lu <hongjiu.lu@intel.com>
6978 * config/i386/i386.md (pushsf splitter): Get stack adjustment
6979 from push operand if code of push isn't PRE_DEC.
6981 2014-01-23 Michael Meissner <meissner@linux.vnet.ibm.com>
6984 * doc/invoke.texi (RS/6000 and PowerPC Options): Document
6985 -mquad-memory-atomic. Update -mquad-memory documentation to say
6986 it is only used for non-atomic loads/stores.
6988 * config/rs6000/predicates.md (quad_int_reg_operand): Allow either
6989 -mquad-memory or -mquad-memory-atomic switches.
6991 * config/rs6000/rs6000-cpus.def (ISA_2_7_MASKS_SERVER): Add
6992 -mquad-memory-atomic to ISA 2.07 support.
6994 * config/rs6000/rs6000.opt (-mquad-memory-atomic): Add new switch
6995 to separate support of normal quad word memory operations (ldq, stq)
6996 from the atomic quad word memory operations.
6998 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
6999 support to separate non-atomic quad word operations from atomic
7000 quad word operations. Disable non-atomic quad word operations in
7001 little endian mode so that we don't have to swap words after the
7002 load and before the store.
7003 (quad_load_store_p): Add comment about atomic quad word support.
7004 (rs6000_opt_masks): Add -mquad-memory-atomic to the list of
7005 options printed with -mdebug=reg.
7007 * config/rs6000/rs6000.h (TARGET_SYNC_TI): Use
7008 -mquad-memory-atomic as the test for whether we have quad word
7009 atomic instructions.
7010 (TARGET_SYNC_HI_QI): If either -mquad-memory-atomic, -mquad-memory,
7011 or -mp8-vector are used, allow byte/half-word atomic operations.
7013 * config/rs6000/sync.md (load_lockedti): Insure that the address
7014 is a proper indexed or indirect address for the lqarx instruction.
7015 On little endian systems, swap the hi/lo registers after the lqarx
7017 (load_lockedpti): Use indexed_or_indirect_operand predicate to
7018 insure the address is valid for the lqarx instruction.
7019 (store_conditionalti): Insure that the address is a proper indexed
7020 or indirect address for the stqcrx. instruction. On little endian
7021 systems, swap the hi/lo registers before doing the stqcrx.
7023 (store_conditionalpti): Use indexed_or_indirect_operand predicate to
7024 insure the address is valid for the stqcrx. instruction.
7026 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
7027 Define __QUAD_MEMORY__ and __QUAD_MEMORY_ATOMIC__ based on what
7028 type of quad memory support is available.
7030 2014-01-23 Vladimir Makarov <vmakarov@redhat.com>
7033 * lra-constraints.c (simplify_operand_subreg): Spill pseudo if
7034 there is a danger of looping.
7036 2014-01-23 Pat Haugen <pthaugen@us.ibm.com>
7038 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
7039 force flag_ira_loop_pressure if set via command line.
7041 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
7043 * config/aarch64/aarch64-simd-builtins.def (ashr): DI mode removed.
7044 (ashr_simd): New builtin handling DI mode.
7045 * config/aarch64/aarch64-simd.md (aarch64_ashr_simddi): New pattern.
7046 (aarch64_sshr_simddi): New match pattern.
7047 * config/aarch64/arm_neon.h (vshr_n_s32): Builtin call modified.
7048 (vshrd_n_s64): Likewise.
7049 * config/aarch64/predicates.md (aarch64_shift_imm64_di): New predicate.
7051 2014-01-23 Nick Clifton <nickc@redhat.com>
7053 * config/msp430/msp430.h (ASM_SPEC): Pass the -mcpu as -mcpu.
7054 (LIB_SPEC): Drop use of memory.ld and peripherals.ld scripts in
7055 favour of mcu specific scripts.
7056 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more matches for
7059 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
7060 Alex Velenko <Alex.Velenko@arm.com>
7062 * config/aarch64/arm_neon.h (vaddv_s8): __LANE0 cleanup.
7063 (vaddv_s16): Likewise.
7064 (vaddv_s32): Likewise.
7065 (vaddv_u8): Likewise.
7066 (vaddv_u16): Likewise.
7067 (vaddv_u32): Likewise.
7068 (vaddvq_s8): Likewise.
7069 (vaddvq_s16): Likewise.
7070 (vaddvq_s32): Likewise.
7071 (vaddvq_s64): Likewise.
7072 (vaddvq_u8): Likewise.
7073 (vaddvq_u16): Likewise.
7074 (vaddvq_u32): Likewise.
7075 (vaddvq_u64): Likewise.
7076 (vaddv_f32): Likewise.
7077 (vaddvq_f32): Likewise.
7078 (vaddvq_f64): Likewise.
7079 (vmaxv_f32): Likewise.
7080 (vmaxv_s8): Likewise.
7081 (vmaxv_s16): Likewise.
7082 (vmaxv_s32): Likewise.
7083 (vmaxv_u8): Likewise.
7084 (vmaxv_u16): Likewise.
7085 (vmaxv_u32): Likewise.
7086 (vmaxvq_f32): Likewise.
7087 (vmaxvq_f64): Likewise.
7088 (vmaxvq_s8): Likewise.
7089 (vmaxvq_s16): Likewise.
7090 (vmaxvq_s32): Likewise.
7091 (vmaxvq_u8): Likewise.
7092 (vmaxvq_u16): Likewise.
7093 (vmaxvq_u32): Likewise.
7094 (vmaxnmv_f32): Likewise.
7095 (vmaxnmvq_f32): Likewise.
7096 (vmaxnmvq_f64): Likewise.
7097 (vminv_f32): Likewise.
7098 (vminv_s8): Likewise.
7099 (vminv_s16): Likewise.
7100 (vminv_s32): Likewise.
7101 (vminv_u8): Likewise.
7102 (vminv_u16): Likewise.
7103 (vminv_u32): Likewise.
7104 (vminvq_f32): Likewise.
7105 (vminvq_f64): Likewise.
7106 (vminvq_s8): Likewise.
7107 (vminvq_s16): Likewise.
7108 (vminvq_s32): Likewise.
7109 (vminvq_u8): Likewise.
7110 (vminvq_u16): Likewise.
7111 (vminvq_u32): Likewise.
7112 (vminnmv_f32): Likewise.
7113 (vminnmvq_f32): Likewise.
7114 (vminnmvq_f64): Likewise.
7116 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
7118 * config/aarch64/aarch64-simd.md
7119 (aarch64_dup_lane<mode>): Correct lane number on big-endian.
7120 (aarch64_dup_lane_<vswap_widthi_name><mode>): Likewise.
7121 (*aarch64_mul3_elt<mode>): Likewise.
7122 (*aarch64_mul3_elt<vswap_width_name><mode>): Likewise.
7123 (*aarch64_mul3_elt_to_64v2df): Likewise.
7124 (*aarch64_mla_elt<mode>): Likewise.
7125 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
7126 (*aarch64_mls_elt<mode>): Likewise.
7127 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
7128 (*aarch64_fma4_elt<mode>): Likewise.
7129 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
7130 (*aarch64_fma4_elt_to_64v2df): Likewise.
7131 (*aarch64_fnma4_elt<mode>): Likewise.
7132 (*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
7133 (*aarch64_fnma4_elt_to_64v2df): Likewise.
7134 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
7135 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
7136 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
7137 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
7138 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
7139 (aarch64_sqdmull_lane<mode>_internal): Likewise.
7140 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
7142 2013-01-23 Alex Velenko <Alex.Velenko@arm.com>
7144 * config/aarch64/aarch64-simd.md
7145 (aarch64_be_checked_get_lane<mode>): New define_expand.
7146 * config/aarch64/aarch64-simd-builtins.def
7147 (BUILTIN_VALL (GETLANE, be_checked_get_lane, 0)):
7148 New builtin definition.
7149 * config/aarch64/arm_neon.h: (__aarch64_vget_lane_any):
7150 Use new safe be builtin.
7152 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
7154 * config/aarch64/aarch64-simd.md (aarch64_be_ld1<mode>):
7156 (aarch64_be_st1<mode>): Likewise.
7157 (aarch_ld1<VALL:mode>): Define_expand modified.
7158 (aarch_st1<VALL:mode>): Likewise.
7159 * config/aarch64/aarch64.md (UNSPEC_LD1): New unspec definition.
7160 (UNSPEC_ST1): Likewise.
7162 2014-01-23 David Holsgrove <david.holsgrove@xilinx.com>
7164 * config/microblaze/microblaze.md: Add trap insn and attribute
7166 2014-01-23 Dodji Seketeli <dodji@redhat.com>
7168 PR preprocessor/58580
7169 * input.h (location_get_source_line): Take an additional line_size
7171 (void diagnostics_file_cache_fini): Declare new function.
7172 * input.c (struct fcache): New type.
7173 (fcache_tab_size, fcache_buffer_size, fcache_line_record_size):
7174 New static constants.
7175 (diagnostic_file_cache_init, total_lines_num)
7176 (lookup_file_in_cache_tab, evicted_cache_tab_entry)
7177 (add_file_to_cache_tab, lookup_or_add_file_to_cache_tab)
7178 (needs_read, needs_grow, maybe_grow, read_data, maybe_read_data)
7179 (get_next_line, read_next_line, goto_next_line, read_line_num):
7180 New static function definitions.
7181 (diagnostic_file_cache_fini): New function.
7182 (location_get_source_line): Take an additional output line_len
7183 parameter. Re-write using lookup_or_add_file_to_cache_tab and
7185 * diagnostic.c (diagnostic_finish): Call
7186 diagnostic_file_cache_fini.
7187 (adjust_line): Take an additional input parameter for the length
7188 of the line, rather than calculating it with strlen.
7189 (diagnostic_show_locus): Adjust the use of
7190 location_get_source_line and adjust_line with respect to their new
7191 signature. While displaying a line now, do not stop at the first
7192 null byte. Rather, display the zero byte as a space and keep
7193 going until we reach the size of the line.
7194 * Makefile.in: Add vec.o to OBJS-libcommon
7196 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
7197 Ilya Tocar <ilya.tocar@intel.com>
7199 * config/i386/avx512fintrin.h (_mm512_kmov): New.
7200 * config/i386/i386.c (IX86_BUILTIN_KMOV16): Ditto.
7201 (__builtin_ia32_kmov16): Ditto.
7202 * config/i386/i386.md (UNSPEC_KMOV): New.
7205 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
7207 * config/i386/avx512fintrin.h (_mm512_loadu_si512): Rename.
7208 (_mm512_storeu_si512): Ditto.
7210 2014-01-23 Richard Sandiford <rdsandiford@googlemail.com>
7213 * rtl.h (get_referenced_operands): Declare.
7214 * recog.c (get_referenced_operands): New function.
7215 * config/mips/mips.c (mips_reorg_process_insns): Check which asm
7216 operands have been referenced when recording LO_SUM references.
7218 2014-01-22 David Holsgrove <david.holsgrove@xilinx.com>
7220 * config/microblaze/microblaze.md: Correct bswaphi2 insn.
7222 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
7224 * config/i386/x86-tune.def (X86_TUNE_ACCUMULATE_OUTGOING_ARGS):
7225 Enable for generic and recent AMD targets.
7227 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
7229 * combine-stack-adj.c (combine_stack_adjustments_for_block): Remove
7230 ARG_SIZE note when adjustment was eliminated.
7232 2014-01-22 Jeff Law <law@redhat.com>
7234 PR tree-optimization/59597
7235 * tree-ssa-threadupdate.c (dump_jump_thread_path): Move to earlier
7236 in file. Accept new argument REGISTERING and use it to modify
7237 dump output appropriately.
7238 (register_jump_thread): Corresponding changes.
7239 (mark_threaded_blocks): Reinstate code to cancel unprofitable
7240 thread paths involving joiner blocks. Add code to dump cancelled
7241 jump threading paths.
7243 2014-01-22 Vladimir Makarov <vmakarov@redhat.com>
7245 PR rtl-optimization/59477
7246 * lra-constraints.c (inherit_in_ebb): Process call for living hard
7247 regs. Update reloads_num and potential_reload_hard_regs for all insns.
7249 2014-01-22 Tom Tromey <tromey@redhat.com>
7251 * config/i386/i386-interix.h (i386_pe_unique_section): Don't use
7253 * config/cr16/cr16-protos.h (notice_update_cc): Don't use PARAMS.
7255 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
7257 PR rtl-optimization/59896
7258 * lra-constraints.c (process_alt_operands): Check unused note for
7259 matched operands of insn with no output reloads.
7261 2014-01-21 Richard Sandiford <rdsandiford@googlemail.com>
7263 * config/mips/mips.c (mips_move_to_gpr_cost): Add M16_REGS case.
7264 (mips_move_from_gpr_cost): Likewise.
7266 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
7268 PR rtl-optimization/59858
7269 * lra-constraints.c (SMALL_REGISTER_CLASS_P): Use
7270 ira_class_hard_regs_num.
7271 (process_alt_operands): Increase reject for dying matched operand.
7273 2014-01-21 Jakub Jelinek <jakub@redhat.com>
7276 * config/i386/i386.c (expand_small_movmem_or_setmem): If mode is
7277 smaller than size, perform several stores or loads and stores
7278 at dst + count - size to store or copy all of size bytes, rather
7279 than just last modesize bytes.
7281 2014-01-20 DJ Delorie <dj@redhat.com>
7283 * config/rl78/rl78.c (rl78_propogate_register_origins): Verify
7284 that CLOBBERs are REGs before propogating their values.
7286 2014-01-20 H.J. Lu <hongjiu.lu@intel.com>
7289 * cgraph.c (cgraph_inline_failed_string): Add type to DEFCIFCODE.
7290 (cgraph_inline_failed_type): New function.
7291 * cgraph.h (DEFCIFCODE): Add type.
7292 (cgraph_inline_failed_type_t): New enum.
7293 (cgraph_inline_failed_type): New prototype.
7294 * cif-code.def: Add CIF_FINAL_NORMAL to OK, FUNCTION_NOT_CONSIDERED,
7295 FUNCTION_NOT_OPTIMIZED, REDEFINED_EXTERN_INLINE,
7296 FUNCTION_NOT_INLINE_CANDIDATE, LARGE_FUNCTION_GROWTH_LIMIT,
7297 LARGE_STACK_FRAME_GROWTH_LIMIT, MAX_INLINE_INSNS_SINGLE_LIMIT,
7298 MAX_INLINE_INSNS_AUTO_LIMIT, INLINE_UNIT_GROWTH_LIMIT,
7299 RECURSIVE_INLINING, UNLIKELY_CALL, NOT_DECLARED_INLINED,
7300 OPTIMIZING_FOR_SIZE, ORIGINALLY_INDIRECT_CALL,
7301 INDIRECT_UNKNOWN_CALL, USES_COMDAT_LOCAL.
7302 Add CIF_FINAL_ERROR to UNSPECIFIED, BODY_NOT_AVAILABLE,
7303 FUNCTION_NOT_INLINABLE, OVERWRITABLE, MISMATCHED_ARGUMENTS,
7304 EH_PERSONALITY, NON_CALL_EXCEPTIONS, TARGET_OPTION_MISMATCH,
7305 OPTIMIZATION_MISMATCH.
7306 * tree-inline.c (expand_call_inline): Emit errors during
7307 early_inlining if cgraph_inline_failed_type returns CIF_FINAL_ERROR.
7309 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
7312 * config/i386/sse.md (*andnot<mode>3<mask_name>): Handle MODE_V16SF
7313 mode attribute in insn output.
7315 2014-01-20 Eric Botcazou <ebotcazou@adacore.com>
7317 * output.h (output_constant): Delete.
7318 * varasm.c (output_constant): Make private.
7320 2014-01-20 Alex Velenko <Alex.Velenko@arm.com>
7322 * config/aarch64/aarch64-simd.md (vec_perm<mode>): Add BE check.
7324 2014-01-20 Jakub Jelinek <jakub@redhat.com>
7327 * tree.h (fold_builtin_strcat): New prototype.
7328 * builtins.c (fold_builtin_strcat): No longer static. Add len
7329 argument, if non-NULL, don't call c_strlen. Optimize
7330 directly into __builtin_memcpy instead of __builtin_strcpy.
7331 (fold_builtin_2): Adjust fold_builtin_strcat caller.
7332 * gimple-fold.c (gimple_fold_builtin): Handle BUILT_IN_STRCAT.
7334 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
7336 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
7337 for SImode_address_operand operands, having only a REG argument.
7339 2014-01-20 Marcus Shawcroft <marcus.shawcroft@arm.com>
7341 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER): Expand
7342 loader name using mbig-endian.
7343 (LINUX_TARGET_LINK_SPEC): Pass linker -m flag.
7345 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
7347 * doc/invoke.texi (-march): Clarify documentation for AArch64.
7351 2014-01-20 Tejas Belagod <tejas.belagod@arm.com>
7353 * config/aarch64/aarch64-protos.h
7354 (aarch64_cannot_change_mode_class_ptr): Declare.
7355 * config/aarch64/aarch64.c (aarch64_cannot_change_mode_class,
7356 aarch64_cannot_change_mode_class_ptr): New.
7357 * config/aarch64/aarch64.h (CANNOT_CHANGE_MODE_CLASS): Change to call
7358 backend hook aarch64_cannot_change_mode_class.
7360 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
7362 * common/config/aarch64/aarch64-common.c
7363 (aarch64_handle_option): Don't handle any option order logic here.
7364 * config/aarch64/aarch64.c (aarch64_parse_arch): Do not override
7365 selected_cpu, warn on architecture version mismatch.
7366 (aarch64_override_options): Fix parsing order for option strings.
7368 2014-01-20 Jan-Benedict Glaw <jbglaw@lug-owl.de>
7369 Iain Sandoe <iain@codesourcery.com>
7372 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Fix unused variable
7373 warning. Amend comment to reflect current functionality.
7375 2014-01-20 Richard Biener <rguenther@suse.de>
7378 * builtins.c (fold_builtin_strcat): Remove case better handled
7379 by tree-ssa-strlen.c.
7381 2014-01-20 Alan Lawrence <alan.lawrence@arm.com>
7383 * config/aarch64/aarch64.opt
7384 (mcpu, march, mtune): Make case-insensitive.
7386 2014-01-20 Jakub Jelinek <jakub@redhat.com>
7389 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
7390 if operands[1] is a REG or ZERO_EXTEND of a REG.
7392 2014-01-19 Jan Hubicka <hubicka@ucw.cz>
7394 * varasm.c (compute_reloc_for_constant): Use targetm.binds_local_p.
7396 2014-01-19 John David Anglin <danglin@gcc.gnu.org>
7398 * config/pa/pa.c (pa_attr_length_millicode_call): Correct length of
7399 long non-pic millicode calls.
7401 2014-01-19 Jan-Benedict Glaw <jbglaw@lug-owl.de>
7403 * config/vax/vax.h (FUNCTION_ARG_REGNO_P): Fix unused variable warning.
7405 2014-01-19 Kito Cheng <kito@0xlab.org>
7407 * builtins.c (expand_movstr): Check movstr expand done or fail.
7409 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
7410 H.J. Lu <hongjiu.lu@intel.com>
7413 * config/i386/i386.md (*lea<mode>): Zero-extend return register
7414 to DImode for zero-extended addresses.
7416 2014-01-19 Jakub Jelinek <jakub@redhat.com>
7418 PR rtl-optimization/57763
7419 * bb-reorder.c (fix_crossing_unconditional_branches): Set JUMP_LABEL
7420 on the new indirect jump_insn and increment LABEL_NUSES (label).
7422 2014-01-18 H.J. Lu <hongjiu.lu@intel.com>
7426 * config.gcc (x86_archs): New variable.
7427 (x86_64_archs): Likewise.
7428 (x86_cpus): Likewise.
7429 Use $x86_archs, $x86_64_archs and $x86_cpus to check valid
7430 --with-arch/--with-cpu= options.
7431 Support --with-arch=/--with-cpu={nehalem,westmere,
7432 sandybridge,ivybridge,haswell,broadwell,bonnell,silvermont}.
7434 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
7436 * config/i386/i386.c (ix86_adjust_cost): Reorder PROCESSOR_K8
7437 and PROCESSOR_ATHLON to simplify code. Move "memory" calculation.
7439 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
7441 * config/i386/i386.md (*swap<mode>): Rename from swap<mode>.
7443 2014-01-18 Jakub Jelinek <jakub@redhat.com>
7446 * config/i386/i386-c.c (ix86_pragma_target_parse): Temporarily
7447 clear cpp_get_options (parse_in)->warn_unused_macros for
7448 ix86_target_macros_internal with cpp_define.
7450 2014-01-18 Richard Sandiford <rdsandiford@googlemail.com>
7452 * jump.c (delete_related_insns): Keep (use (insn))s.
7453 * reorg.c (redundant_insn): Check for barriers too.
7455 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
7457 * config/i386/i386.c (ix86_split_lea_for_addr): Fix a comment typo.
7459 2014-01-17 John David Anglin <danglin@gcc.gnu.org>
7461 * config/pa/pa.c (pa_attr_length_indirect_call): Don't output a short
7462 call to $$dyncall when TARGET_LONG_CALLS is true.
7464 2014-01-17 Jeff Law <law@redhat.com>
7466 * ree.c (combine_set_extension): Temporarily disable test for
7467 changing number of hard registers.
7469 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
7472 * ipa-inline-analysis.c (inline_free_summary):
7473 Do not free summary of aliases.
7475 2014-01-17 Jakub Jelinek <jakub@redhat.com>
7478 * gimplify.c (gimplify_expr): Use create_tmp_var
7479 instead of create_tmp_var_raw. If cond doesn't have
7480 integral type, don't add the IFN_ANNOTATE builtin at all.
7482 2014-01-17 Martin Jambor <mjambor@suse.cz>
7485 * ipa-cp.c (prev_edge_clone): New variable.
7486 (grow_next_edge_clone_vector): Renamed to grow_edge_clone_vectors.
7487 Also resize prev_edge_clone vector.
7488 (ipcp_edge_duplication_hook): Also update prev_edge_clone.
7489 (ipcp_edge_removal_hook): New function.
7490 (ipcp_driver): Register ipcp_edge_removal_hook.
7492 2014-01-17 Andrew Pinski <apinski@cavium.com>
7493 Steve Ellcey <sellcey@mips.com>
7496 * config/mips/mips.c (mips_print_operand): Check operand mode instead
7499 2014-01-17 Jeff Law <law@redhat.com>
7502 * passes.def: Reorder pass_copy_prop, pass_unrolli, pass_ccp sequence
7503 so that pass_ccp runs first.
7505 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
7507 * config/i386/i386.c (ix86_lea_outperforms): Use TARGET_XXX.
7508 (ix86_adjust_cost): Use !TARGET_XXX.
7509 (do_reorder_for_imul): Likewise.
7510 (swap_top_of_ready_list): Likewise.
7511 (ix86_sched_reorder): Likewise.
7513 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
7515 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
7516 PROCESSOR_INTEL. Treat like PROCESSOR_GENERIC.
7517 * config/i386/i386.c (intel_memcpy): New. Duplicate slm_memcpy.
7518 (intel_memset): New. Duplicate slm_memset.
7519 (intel_cost): New. Duplicate slm_cost.
7520 (m_INTEL): New macro.
7521 (processor_target_table): Add "intel".
7522 (ix86_option_override_internal): Replace PROCESSOR_SILVERMONT
7523 with PROCESSOR_INTEL for "intel".
7524 (ix86_lea_outperforms): Support PROCESSOR_INTEL. Duplicate
7525 PROCESSOR_SILVERMONT.
7526 (ix86_issue_rate): Likewise.
7527 (ix86_adjust_cost): Likewise.
7528 (ia32_multipass_dfa_lookahead): Likewise.
7529 (swap_top_of_ready_list): Likewise.
7530 (ix86_sched_reorder): Likewise.
7531 (ix86_avoid_lea_for_addr): Check TARGET_AVOID_LEA_FOR_ADDR
7532 instead of TARGET_OPT_AGU.
7533 * config/i386/i386.h (TARGET_INTEL): New.
7534 (TARGET_AVOID_LEA_FOR_ADDR): Likewise.
7535 (processor_type): Add PROCESSOR_INTEL.
7536 * config/i386/x86-tune.def: Support m_INTEL. Duplicate m_SILVERMONT.
7537 Add X86_TUNE_AVOID_LEA_FOR_ADDR.
7539 2014-01-17 Marek Polacek <polacek@redhat.com>
7542 * gimple-fold.c (fold_array_ctor_reference): Don't fold if element
7545 2014-01-17 Richard Biener <rguenther@suse.de>
7547 PR tree-optimization/46590
7548 * opts.c (default_options_table): Add entries for
7549 OPT_fbranch_count_reg, OPT_fmove_loop_invariants and OPT_ftree_pta,
7550 all enabled at -O1 but not for -Og.
7551 * common.opt (fbranch-count-reg): Remove Init(1).
7552 (fmove-loop-invariants): Likewise.
7553 (ftree-pta): Likewise.
7555 2014-01-17 Jakub Jelinek <jakub@redhat.com>
7557 * config/i386/i386.c (ix86_data_alignment): For compatibility with
7558 (incorrect) GCC 4.8 and earlier alignment assumptions ensure we align
7559 decls to at least the GCC 4.8 used alignments.
7562 * tree-nested.c (convert_nonlocal_reference_stmt,
7563 convert_local_reference_stmt): For NAMELIST_DECLs in gimple_bind_vars
7564 of GIMPLE_BIND stmts, adjust associated decls.
7566 2014-01-17 Richard Biener <rguenther@suse.de>
7568 PR tree-optimization/46590
7569 * vec.h (vec<>::bseach): New member function implementing
7570 binary search according to C89 bsearch.
7571 (vec<>::qsort): Avoid calling ::qsort for vectors with sizes 0 or 1.
7572 * tree-ssa-loop-im.c (struct mem_ref): Make stored member a
7573 bitmap pointer again. Make accesses_in_loop a flat array.
7574 (mem_ref_obstack): New global.
7575 (outermost_indep_loop): Adjust for mem_ref->stored changes.
7576 (mark_ref_stored): Likewise.
7577 (ref_indep_loop_p_2): Likewise.
7578 (set_ref_stored_in_loop): New helper function.
7579 (mem_ref_alloc): Allocate mem_refs on the mem_ref_obstack obstack.
7580 (memref_free): Adjust.
7581 (record_mem_ref_loc): Simplify.
7582 (gather_mem_refs_stmt): Adjust.
7583 (sort_locs_in_loop_postorder_cmp): New function.
7584 (analyze_memory_references): Sort accesses_in_loop after
7585 loop postorder number.
7586 (find_ref_loc_in_loop_cmp): New function.
7587 (for_all_locs_in_loop): Find relevant cluster of locs in
7588 accesses_in_loop and iterate without recursion.
7589 (execute_sm): Avoid uninit warning.
7590 (struct ref_always_accessed): Simplify.
7591 (ref_always_accessed::operator ()): Likewise.
7592 (ref_always_accessed_p): Likewise.
7593 (tree_ssa_lim_initialize): Initialize mem_ref_obstack, compute
7594 loop postorder numbers here.
7595 (tree_ssa_lim_finalize): Free mem_ref_obstack and loop postorder
7598 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
7601 * passes.c (rest_of_decl_compilation): Don't call varpool_finalize_decl
7602 on decls for which assemble_alias has been called.
7604 2014-01-17 Nick Clifton <nickc@redhat.com>
7606 * config/msp430/msp430.opt: (mcpu): New option.
7607 * config/msp430/msp430.c (msp430_mcu_name): Use target_mcu.
7608 (msp430_option_override): Parse target_cpu. If the MCU name
7609 matches a generic string, clear target_mcu.
7610 (msp430_attr): Allow numeric interrupt values up to 63.
7611 (msp430_expand_epilogue): No longer invert operand 1 of gen_popm.
7612 * config/msp430/msp430.h (ASM_SPEC): Convert -mcpu into a -mmcu
7614 * config/msp430/t-msp430: (MULTILIB_MATCHES): Remove mcu matches.
7616 * config/msp430/msp430.md (popm): Use %J rather than %I.
7617 (addsi3): Use msp430_nonimmediate_operand for operand 2.
7618 (addhi_cy_i): Use immediate_operand for operand 2.
7619 * doc/invoke.texi: Document -mcpu option.
7621 2014-01-17 Richard Biener <rguenther@suse.de>
7623 PR rtl-optimization/38518
7624 * df.h (df_analyze_loop): Declare.
7625 * df-core.c: Include cfgloop.h.
7626 (df_analyze_1): Split out main part of df_analyze.
7627 (df_analyze): Adjust.
7628 (loop_inverted_post_order_compute): New function.
7629 (loop_post_order_compute): Likewise.
7630 (df_analyze_loop): New function avoiding whole-function
7632 * loop-invariant.c (find_defs): Use df_analyze_loop.
7633 (find_invariants): Adjust.
7634 * loop-iv.c (iv_analysis_loop_init): Use df_analyze_loop.
7636 2014-01-17 Zhenqiang Chen <zhenqiang.chen@arm.com>
7638 * config/arm/arm.c (arm_v7m_tune): Set max_insns_skipped to 2.
7639 (thumb2_final_prescan_insn): Set max to MAX_INSN_PER_IT_BLOCK.
7641 2014-01-16 Ilya Enkovich <ilya.enkovich@intel.com>
7643 * ipa-ref.c (ipa_remove_stmt_references): Fix references
7644 traversal when removing references.
7646 2014-01-16 Jan Hubicka <hubicka@ucw.cz>
7649 * tree.c (get_binfo_at_offset): Look harder for virtual bases.
7651 2014-01-16 Bernd Schmidt <bernds@codesourcery.com>
7654 * reload.c (find_reloads_address_1): Do not use RELOAD_OTHER when
7655 pushing a reload for an autoinc when we had previously reloaded an
7656 inner part of the address.
7658 2014-01-16 Jakub Jelinek <jakub@redhat.com>
7660 * tree-vectorizer.h (struct _loop_vec_info): Add no_data_dependencies
7662 (LOOP_VINFO_NO_DATA_DEPENDENCIES): Define.
7663 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Clear it
7664 when not giving up or versioning for alias only because of
7666 (vect_analyze_data_ref_dependences): Set to true.
7667 * tree-vect-stmts.c (hoist_defs_of_uses): Return false if def_stmt
7669 (vectorizable_load): Use LOOP_VINFO_NO_DATA_DEPENDENCIES instead of
7670 LOOP_REQUIRES_VERSIONING_FOR_ALIAS, add && !nested_in_vect_loop
7674 * expr.c (expand_expr_real_1): Handle init == NULL_TREE.
7677 * config/i386/i386.c (ix86_expand_builtin): If target doesn't satisfy
7678 operand 0 predicate for gathers, use a new pseudo as subtarget.
7680 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
7683 * lra-constraints.c (process_alt_operands): Add printing debug info.
7684 Check absence of input/output reloads for matched operands too.
7686 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
7688 PR rtl-optimization/59835
7689 * ira.c (ira_init_register_move_cost): Increase cost for
7692 2014-01-16 Alan Lawrence <alan.lawrence@arm.com>
7694 * config/arm/arm.opt (mcpu, march, mtune): Make case-insensitive.
7696 2014-01-16 Richard Earnshaw <rearnsha@arm.com>
7699 * aarch64.c (aarch64_split_128bit_move): Don't lookup REGNO on
7700 non-register objects. Use gen_(high/low)part more consistently.
7703 2014-01-16 Michael Meissner <meissner@linux.vnet.ibm.com>
7706 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Add little
7707 endian support, remove tests for WORDS_BIG_ENDIAN.
7708 (p8_mfvsrd_3_<mode>): Likewise.
7709 (reload_gpr_from_vsx<mode>): Likewise.
7710 (reload_gpr_from_vsxsf): Likewise.
7711 (p8_mfvsrd_4_disf): Likewise.
7713 2014-01-16 Richard Biener <rguenther@suse.de>
7715 PR rtl-optimization/46590
7716 * lcm.c (compute_antinout_edge): Use postorder iteration.
7717 (compute_laterin): Use inverted postorder iteration.
7719 2014-01-16 Nick Clifton <nickc@redhat.com>
7722 * varasm.c (output_constant): Return the number of bytes actually
7724 (output_constructor_array_range): Update the field size with the
7725 number of bytes emitted by output_constant.
7726 (output_constructor_regular_field): Likewise. Also do not
7727 complain if the total number of bytes emitted is now greater
7728 than the expected fieldpos.
7729 * output.h (output_constant): Update prototype and descriptive comment.
7731 2014-01-16 Marek Polacek <polacek@redhat.com>
7734 * cgraph.c (gimple_check_call_args): Don't use DECL_ARG_TYPE if
7735 it is error_mark_node.
7737 2014-01-15 Uros Bizjak <ubizjak@gmail.com>
7739 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use
7740 VALID_AVX256_REG_OR_OI_MODE.
7742 2014-01-15 Pat Haugen <pthaugen@us.ibm.com>
7744 * config/rs6000/rs6000.c (rs6000_output_function_prologue): Check if
7745 current procedure should be profiled.
7747 2014-01-15 Andrew Pinski <apinski@cavium.com>
7749 * config/aarch64/aarch64.c (aarch64_register_move_cost): Correct cost
7750 of moving from/to the STACK_REG register class.
7752 2014-01-15 Richard Henderson <rth@redhat.com>
7755 * reginfo.c (global_regs_decl): Globalize.
7756 * rtl.h (global_regs_decl): Declare.
7757 * ira.c (do_reload): Diagnose frame_pointer_needed and it
7758 reserved via global_regs.
7760 2014-01-15 Teresa Johnson <tejohnson@google.com>
7762 * tree-ssa-sccvn.c (visit_reference_op_call): Handle NULL vdef.
7764 2014-01-15 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
7766 * config/rs6000/altivec.md (mulv8hi3): Explicitly generate vmulesh
7767 and vmulosh rather than call gen_vec_widen_smult_*.
7768 (vec_widen_umult_even_v16qi): Test VECTOR_ELT_ORDER_BIG rather
7769 than BYTES_BIG_ENDIAN to determine use of even or odd instruction.
7770 (vec_widen_smult_even_v16qi): Likewise.
7771 (vec_widen_umult_even_v8hi): Likewise.
7772 (vec_widen_smult_even_v8hi): Likewise.
7773 (vec_widen_umult_odd_v16qi): Likewise.
7774 (vec_widen_smult_odd_v16qi): Likewise.
7775 (vec_widen_umult_odd_v8hi): Likewise.
7776 (vec_widen_smult_odd_v8hi): Likewise.
7777 (vec_widen_umult_hi_v16qi): Explicitly generate vmuleub and
7778 vmuloub rather than call gen_vec_widen_umult_*.
7779 (vec_widen_umult_lo_v16qi): Likewise.
7780 (vec_widen_smult_hi_v16qi): Explicitly generate vmulesb and
7781 vmulosb rather than call gen_vec_widen_smult_*.
7782 (vec_widen_smult_lo_v16qi): Likewise.
7783 (vec_widen_umult_hi_v8hi): Explicitly generate vmuleuh and vmulouh
7784 rather than call gen_vec_widen_umult_*.
7785 (vec_widen_umult_lo_v8hi): Likewise.
7786 (vec_widen_smult_hi_v8hi): Explicitly gnerate vmulesh and vmulosh
7787 rather than call gen_vec_widen_smult_*.
7788 (vec_widen_smult_lo_v8hi): Likewise.
7790 2014-01-15 Jeff Law <law@redhat.com>
7792 PR tree-optimization/59747
7793 * ree.c (find_and_remove_re): Properly handle case where a second
7794 eliminated extension requires widening a copy created for elimination
7795 of a prior extension.
7796 (combine_set_extension): Ensure that the number of hard regs needed
7797 for a destination register does not change when we widen it.
7799 2014-01-15 Sebastian Huber <sebastian.huber@embedded-brains.de>
7801 * config.gcc (*-*-rtems*): Add t-rtems to tmake_file.
7802 (arm*-*-uclinux*eabi*): Do not override an existing tmake_file.
7803 (arm*-*-eabi* | arm*-*-symbianelf* | arm*-*-rtems*): Likwise.
7804 (arm*-*-rtems*): Use t-rtems from existing tmake_file.
7805 (avr-*-rtems*): Likewise.
7806 (bfin*-rtems*): Likewise.
7807 (moxie-*-rtems*): Likewise.
7808 (h8300-*-rtems*): Likewise.
7809 (i[34567]86-*-rtems*): Likewise.
7810 (lm32-*-rtems*): Likewise.
7811 (m32r-*-rtems*): Likewise.
7812 (m68k-*-rtems*): Likewise.
7813 (microblaze*-*-rtems*): Likewise.
7814 (mips*-*-rtems*): Likewise.
7815 (powerpc-*-rtems*): Likewise.
7816 (sh-*-rtems*): Likewise.
7817 (sparc-*-rtems*): Likewise.
7818 (sparc64-*-rtems*): Likewise.
7819 (v850-*-rtems*): Likewise.
7820 (m32c-*-rtems*): Likewise.
7822 2014-01-15 Vladimir Makarov <vmakarov@redhat.com>
7824 PR rtl-optimization/59511
7825 * ira.c (ira_init_register_move_cost): Use memory costs for some
7826 cases of register move cost calculations.
7827 * lra-constraints.c (lra_constraints): Use REG_FREQ_FROM_BB
7828 instead of BB frequency.
7829 * lra-coalesce.c (move_freq_compare_func, lra_coalesce): Ditto.
7830 * lra-assigns.c (find_hard_regno_for): Ditto.
7832 2014-01-15 Richard Biener <rguenther@suse.de>
7834 PR tree-optimization/59822
7835 * tree-vect-stmts.c (hoist_defs_of_uses): New function.
7836 (vectorizable_load): Use it to hoist defs of uses of invariant
7837 loads out of the loop.
7839 2014-01-15 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
7840 Kugan Vivekanandarajah <kuganv@linaro.org>
7843 * config/aarch64/aarch64.c (aarch64_build_constant): Fix incorrect
7846 2014-01-15 Richard Biener <rguenther@suse.de>
7848 PR rtl-optimization/59802
7849 * lcm.c (compute_available): Use inverted postorder to seed
7850 the initial worklist.
7852 2014-01-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7855 * config/s390/s390.c (s390_preferred_reload_class): Don't return
7856 ADDR_REGS for invalid symrefs in non-PIC code.
7858 2014-01-15 Jakub Jelinek <jakub@redhat.com>
7861 * builtins.c (determine_block_size): Initialize *probable_max_size
7862 even if len_rtx is CONST_INT.
7864 2014-01-14 Andrew Pinski <apinski@cavium.com>
7866 * config/aarch64/aarch64-protos.h (tune_params): Add issue_rate.
7867 * config/aarch64/aarch64.c (generic_tunings): Add issue rate of 2.
7868 (cortexa53_tunings): Likewise.
7869 (aarch64_sched_issue_rate): New function.
7870 (TARGET_SCHED_ISSUE_RATE): Define.
7872 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
7874 * ira-costs.c (find_costs_and_classes): Add missed
7875 ira_init_register_move_cost_if_necessary.
7877 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
7880 * config/arm/arm.c (arm_coproc_mem_operand): Add lra_in_progress.
7882 2014-01-14 H.J. Lu <hongjiu.lu@intel.com>
7885 * config/i386/i386.c (type_natural_mode): Add a bool parameter
7886 to indicate if type is used for function return value. Warn ABI
7887 change if the vector mode isn't available for function return value.
7888 (ix86_function_arg_advance): Pass false to type_natural_mode.
7889 (ix86_function_arg): Likewise.
7890 (ix86_gimplify_va_arg): Likewise.
7891 (function_arg_32): Don't warn ABI change.
7892 (ix86_function_value): Pass true to type_natural_mode.
7893 (ix86_return_in_memory): Likewise.
7894 (ix86_struct_value_rtx): Removed.
7895 (TARGET_STRUCT_VALUE_RTX): Likewise.
7897 2014-01-14 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7899 * jump.c (redirect_jump_2): Remove REG_CROSSING_JUMP notes when
7900 converting a conditional jump into a conditional return.
7902 2014-01-14 Richard Biener <rguenther@suse.de>
7904 PR tree-optimization/58921
7905 PR tree-optimization/59006
7906 * tree-vect-loop-manip.c (vect_loop_versioning): Remove code
7907 hoisting invariant stmts.
7908 * tree-vect-stmts.c (vectorizable_load): Insert the splat of
7909 invariant loads on the preheader edge if possible.
7911 2014-01-14 Joey Ye <joey.ye@arm.com>
7913 * doc/plugin.texi (Building GCC plugins): Update to C++.
7915 2014-01-14 Kirill Yukhin <kirill.yukhin@intel.com>
7917 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): New.
7918 (_mm_rcp28_round_ss): Ditto.
7919 (_mm_rsqrt28_round_sd): Ditto.
7920 (_mm_rsqrt28_round_ss): Ditto.
7921 (_mm_rcp28_sd): Ditto.
7922 (_mm_rcp28_ss): Ditto.
7923 (_mm_rsqrt28_sd): Ditto.
7924 (_mm_rsqrt28_ss): Ditto.
7925 * config/i386/avx512fintrin.h (_mm512_stream_load_si512): Ditto.
7926 * config/i386/i386-builtin-types.def (V8DI_FTYPE_PV8DI): Ditto.
7927 * config/i386/i386.c (IX86_BUILTIN_MOVNTDQA512): Ditto.
7928 (IX86_BUILTIN_RCP28SD): Ditto.
7929 (IX86_BUILTIN_RCP28SS): Ditto.
7930 (IX86_BUILTIN_RSQRT28SD): Ditto.
7931 (IX86_BUILTIN_RSQRT28SS): Ditto.
7932 (bdesc_special_args): Define __builtin_ia32_movntdqa512,
7933 __builtin_ia32_rcp28sd_round, __builtin_ia32_rcp28ss_round,
7934 __builtin_ia32_rsqrt28sd_round, __builtin_ia32_rsqrt28ss_round.
7935 (ix86_expand_special_args_builtin): Expand new FTYPE.
7936 * config/i386/sse.md (define_mode_attr "sse4_1_avx2"): Expand to V8DI.
7937 (srcp14<mode>): Make insn unary.
7938 (avx512f_vmscalef<mode><round_name>): Use substed predicate.
7939 (avx512f_sgetexp<mode><round_saeonly_name>): Ditto.
7940 (avx512f_rndscale<mode><round_saeonly_name>): Ditto.
7941 (<sse4_1_avx2>_movntdqa): Extend to 512 bits.
7942 (avx512er_exp2<mode><mask_name><round_saeonly_name>):
7943 Fix rounding: make it SAE only.
7944 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
7946 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
7948 (avx512er_vmrcp28<mode><round_saeonly_name>): Ditto.
7949 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Ditto.
7950 (avx512f_getmant<mode><mask_name><round_saeonly_name>): Ditto.
7951 * config/i386/subst.md (round_saeonly_mask_scalar_operand3): Remove.
7952 (round_saeonly_mask_scalar_operand4): Ditto.
7953 (round_saeonly_mask_scalar_op3): Ditto.
7954 (round_saeonly_mask_scalar_op4): Ditto.
7956 2014-01-13 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7958 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
7959 Implement -maltivec=be for vec_insert and vec_extract.
7961 2014-01-10 DJ Delorie <dj@redhat.com>
7963 * config/msp430/msp430.md (call_internal): Don't allow memory
7964 references with SP as the base register.
7965 (call_value_internal): Likewise.
7966 * config/msp430/constraints.md (Yc): New. For memory references
7967 that don't use SP as a base register.
7969 * config/msp430/msp430.c (msp430_print_operand): Add 'J' to mean
7970 "an integer without a # prefix"
7971 * config/msp430/msp430.md (epilogue_helper): Use it.
7973 2014-01-13 Jakub Jelinek <jakub@redhat.com>
7976 * config/i386/i386.c (ix86_vectorize_builtin_gather): Uncomment
7977 AVX512F gather builtins.
7978 * tree-vect-stmts.c (vectorizable_mask_load_store): For now punt
7979 on gather decls with INTEGER_TYPE masktype.
7980 (vectorizable_load): For INTEGER_TYPE masktype, put the INTEGER_CST
7981 directly into the builtin rather than hoisting it before loop.
7983 PR tree-optimization/59387
7984 * tree-scalar-evolution.c: Include gimple-fold.h and gimplify-me.h.
7985 (scev_const_prop): If folded_casts and type has undefined overflow,
7986 use force_gimple_operand instead of force_gimple_operand_gsi and
7987 for each added stmt if it is assign with
7988 arith_code_with_undefined_signed_overflow, call
7989 rewrite_to_defined_overflow.
7990 * tree-ssa-loop-im.c: Don't include gimplify-me.h, include
7991 gimple-fold.h instead.
7992 (arith_code_with_undefined_signed_overflow,
7993 rewrite_to_defined_overflow): Moved to ...
7994 * gimple-fold.c (arith_code_with_undefined_signed_overflow,
7995 rewrite_to_defined_overflow): ... here. No longer static.
7996 Include gimplify-me.h.
7997 * gimple-fold.h (arith_code_with_undefined_signed_overflow,
7998 rewrite_to_defined_overflow): New prototypes.
8000 2014-01-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8002 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Fix typo in description.
8004 2014-01-13 Eric Botcazou <ebotcazou@adacore.com>
8006 * builtins.c (get_object_alignment_2): Minor tweak.
8007 * tree-ssa-loop-ivopts.c (may_be_unaligned_p): Rewrite.
8009 2014-01-13 Christian Bruel <christian.bruel@st.com>
8011 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Unroll small sizes and
8012 optimized non constant lengths.
8014 2014-01-13 Jakub Jelinek <jakub@redhat.com>
8017 * omp-low.c (expand_omp_atomic_pipeline): Expand the initial
8018 load as __atomic_load_N if possible.
8020 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
8022 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): Remove
8024 (rs6000_expand_builtin): Adjust call.
8026 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
8029 * config/rs6000/rs6000.h (SWITCHABLE_TARGET): Define.
8030 * config/rs6000/rs6000.c: Include target-globals.h.
8031 (rs6000_set_current_function): Instead of doing target_reinit
8032 unconditionally, use save_target_globals_default_opts and
8033 restore_target_globals.
8035 * config/rs6000/rs6000-builtin.def (mffs, mtfsf): Add builtins for
8037 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): New.
8038 (rs6000_expand_builtin): Handle mffs and mtfsf.
8039 (rs6000_init_builtins): Define mffs and mtfsf.
8040 * config/rs6000/rs6000.md (UNSPECV_MFFS, UNSPECV_MTFSF): New constants.
8041 (rs6000_mffs): New pattern.
8042 (rs6000_mtfsf): New pattern.
8044 2014-01-11 Bin Cheng <bin.cheng@arm.com>
8046 * tree-ssa-loop-ivopts.c (iv_ca_narrow): New parameter.
8047 Start narrowing with START. Apply candidate-use pair
8048 and check overall cost in narrowing.
8049 (iv_ca_prune): Pass new argument.
8051 2014-01-10 Jeff Law <law@redhat.com>
8054 * ree.c (combine_reaching_defs): Ensure the defining statement
8055 occurs before the extension when optimizing extensions with
8056 different source and destination hard registers.
8058 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
8061 * ipa-devirt.c (build_type_inheritance_graph): Also add types of
8062 vtables into the type inheritance graph.
8064 2014-01-10 Jakub Jelinek <jakub@redhat.com>
8066 PR rtl-optimization/59754
8067 * ree.c (combine_reaching_defs): Disallow !SCALAR_INT_MODE_P
8068 modes in the REGNO != REGNO case.
8070 2014-01-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8072 * config/rs6000/rs6000-builtin.def: Fix pasto for VPKSDUS.
8074 2014-01-10 Jakub Jelinek <jakub@redhat.com>
8076 PR tree-optimization/59745
8077 * tree-predcom.c (tree_predictive_commoning_loop): Call
8078 free_affine_expand_cache if giving up because components is NULL.
8080 * target-globals.c (save_target_globals): Allocate < 4KB structs using
8081 GC in payload of target_globals struct instead of allocating them on
8082 the heap and the larger structs separately using GC.
8083 * target-globals.h (struct target_globals): Make regs, hard_regs,
8084 reload, expmed, ira, ira_int and lra_fields GTY((atomic)) instead
8085 of GTY((skip)) and change type to void *.
8086 (reset_target_globals): Cast loads from those fields to corresponding
8089 2014-01-10 Steve Ellcey <sellcey@mips.com>
8092 * Makefile.in (PLUGIN_HEADERS): Add gimplify.h, gimple-iterator.h,
8093 gimple-ssa.h, fold-const.h, tree-cfg.h, tree-into-ssa.h,
8094 tree-ssanames.h, print-tree.h, varasm.h, and context.h.
8096 2014-01-10 Richard Earnshaw <rearnsha@arm.com>
8099 * aarch64-modes.def (CC_Zmode): New flags mode.
8100 * aarch64.c (aarch64_select_cc_mode): Only allow NEG when the condition
8101 represents an equality.
8102 (aarch64_get_condition_code): Handle CC_Zmode.
8103 * aarch64.md (compare_neg<mode>): Restrict to equality operations.
8105 2014-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8107 * config/s390/s390.c (s390_expand_tbegin): Remove jump over CC
8108 extraction in good case.
8110 2014-01-10 Richard Biener <rguenther@suse.de>
8112 PR tree-optimization/59374
8113 * tree-vect-slp.c (vect_slp_analyze_bb_1): Move dependence
8114 checking after SLP discovery. Mark stmts not participating
8115 in any SLP instance properly.
8117 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8119 * config/arm/arm.c (arm_new_rtx_costs): Use destination mode
8120 when handling a SET rtx.
8122 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8124 * config/arm/arm-cores.def (cortex-a53): Specify FL_CRC32.
8125 (cortex-a57): Likewise.
8126 (cortex-a57.cortex-a53): Likewise. Remove redundant flags.
8128 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8130 * config/arm/arm.c (arm_init_iwmmxt_builtins): Skip
8131 non-iwmmxt builtins.
8133 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
8137 * ipa-devirt.c record_target_from_binfo): Take as argument
8138 stack of binfos and lookup matching one for virtual inheritance.
8139 (possible_polymorphic_call_targets_1): Update.
8141 2014-01-10 Huacai Chen <chenhc@lemote.com>
8143 * config/mips/driver-native.c (host_detect_local_cpu): Handle new
8144 kernel strings for Loongson-2E/2F/3A.
8146 2014-01-10 Jakub Jelinek <jakub@redhat.com>
8149 * tree-vect-data-refs.c (vect_analyze_data_refs): Check
8150 is_gimple_call before calling gimple_call_internal_p.
8152 2014-01-09 Steve Ellcey <sellcey@mips.com>
8154 * Makefile.in (TREE_FLOW_H): Remove.
8155 (TREE_SSA_H): Add file names from tree-flow.h.
8156 * doc/tree-ssa.texi (Annotations): Remove reference to tree-flow.h
8157 * tree.h: Remove tree-flow.h reference.
8158 * hash-table.h: Remove tree-flow.h reference.
8159 * tree-ssa-loop-niter.c (dump_affine_iv): Replace tree-flow.h
8160 reference with tree-ssa-loop.h.
8162 2014-01-09 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8164 * doc/invoke.texi: Add -maltivec={be,le} options, and document
8165 default element-order behavior for -maltivec.
8166 * config/rs6000/rs6000.opt: Add -maltivec={be,le} options.
8167 * config/rs6000/rs6000.c (rs6000_option_override_internal): Ensure
8168 that -maltivec={le,be} implies -maltivec; disallow -maltivec=le
8169 when targeting big endian, at least for now.
8170 * config/rs6000/rs6000.h: Add #define of VECTOR_ELT_ORDER_BIG.
8172 2014-01-09 Jakub Jelinek <jakub@redhat.com>
8175 * cfgexpand.c (expand_one_var): For SSA_NAMEs, if the underlying
8176 var satisfies use_register_for_decl, just take into account type
8177 alignment, rather than decl alignment.
8179 PR tree-optimization/59622
8180 * gimple-fold.c (gimple_fold_call): Fix a typo in message. For
8181 __builtin_unreachable replace the OBJ_TYPE_REF call with a call to
8182 __builtin_unreachable and add if needed a setter of the lhs SSA_NAME.
8183 Don't devirtualize for inplace at all. For targets.length () == 1,
8184 if the call is noreturn and cfun isn't in SSA form yet, clear lhs.
8186 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
8188 * config/i386/i386.md (cpu): Remove the unused btver1.
8190 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
8192 * gdbasan.in: Put a breakpoint on __sanitizer::Report.
8194 2014-01-09 Jakub Jelinek <jakub@redhat.com>
8197 * tree-core.h (struct target_globals): New forward declaration.
8198 (struct tree_target_option): Add globals field.
8199 * tree.h (TREE_TARGET_GLOBALS): Define.
8200 (prepare_target_option_nodes_for_pch): New prototype.
8201 * target-globals.h (struct target_globals): Define even if
8203 * tree.c (prepare_target_option_node_for_pch,
8204 prepare_target_option_nodes_for_pch): New functions.
8205 * config/i386/i386.h (SWITCHABLE_TARGET): Define.
8206 * config/i386/i386.c: Include target-globals.h.
8207 (ix86_set_current_function): Instead of doing target_reinit
8208 unconditionally, use save_target_globals_default_opts and
8209 restore_target_globals.
8211 2014-01-09 Richard Biener <rguenther@suse.de>
8213 PR tree-optimization/59715
8214 * tree-cfg.h (split_critical_edges): Declare.
8215 * tree-cfg.c (split_critical_edges): Export.
8216 * tree-ssa-sink.c (execute_sink_code): Split critical edges.
8218 2014-01-09 Max Ostapenko <m.ostapenko@partner.samsung.com>
8220 * cfgexpand.c (expand_stack_vars): Optionally disable
8221 asan stack protection.
8222 (expand_used_vars): Likewise.
8223 (partition_stack_vars): Likewise.
8224 * asan.c (asan_emit_stack_protection): Optionally disable
8225 after return stack usage.
8226 (instrument_derefs): Optionally disable memory access instrumentation.
8227 (instrument_builtin_call): Likewise.
8228 (instrument_strlen_call): Likewise.
8229 (asan_protect_global): Optionally disable global variables protection.
8230 * doc/invoke.texi: Added doc for new options.
8231 * params.def: Added new options.
8232 * params.h: Likewise.
8234 2014-01-09 Jakub Jelinek <jakub@redhat.com>
8236 PR rtl-optimization/59724
8237 * ifcvt.c (cond_exec_process_if_block): Don't call
8238 flow_find_head_matching_sequence with 0 longest_match.
8239 * cfgcleanup.c (flow_find_head_matching_sequence): Count even
8240 non-active insns if !stop_after.
8241 (try_head_merge_bb): Revert 2014-01-07 changes.
8243 2014-01-08 Jeff Law <law@redhat.com>
8245 * ree.c (get_sub_rtx): New function, extracted from...
8246 (merge_def_and_ext): Here.
8247 (combine_reaching_defs): Use get_sub_rtx.
8249 2014-01-08 Eric Botcazou <ebotcazou@adacore.com>
8251 * cgraph.h (varpool_variable_node): Do not choke on null node.
8253 2014-01-08 Catherine Moore <clm@codesourcery.com>
8255 * config/mips/mips.md (simple_return): Attempt to use JRC
8257 * config/mips/mips.h (MIPS_CALL): Attempt to use JALS for microMIPS.
8259 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
8261 PR rtl-optimization/59137
8262 * reorg.c (steal_delay_list_from_target): Call update_block for
8264 (steal_delay_list_from_fallthrough, relax_delay_slots): Likewise.
8266 2014-01-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8268 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Remove
8269 two duplicate entries.
8271 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
8274 2012-10-07 Richard Sandiford <rdsandiford@googlemail.com>
8276 * config/mips/mips.c (mips_truncated_op_cost): New function.
8277 (mips_rtx_costs): Adjust test for BADDU.
8278 * config/mips/mips.md (*baddu_di<mode>): Push truncates to operands.
8280 2012-10-02 Richard Sandiford <rdsandiford@googlemail.com>
8282 * config/mips/mips.md (*baddu_si_eb, *baddu_si_el): Merge into...
8283 (*baddu_si): ...this new pattern.
8285 2014-01-08 Jakub Jelinek <jakub@redhat.com>
8288 * ipa-prop.c (ipa_analyze_params_uses): Ignore uses in debug stmts.
8290 2014-01-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
8293 * expr.h (expand_expr_real, expand_expr_real_1): Add new parameter
8295 (expand_expr, expand_normal): Adjust.
8296 * expr.c (expand_expr_real, expand_expr_real_1): Add new parameter
8297 inner_reference_p. Use inner_reference_p to expand inner references.
8298 (store_expr): Adjust.
8299 * cfgexpand.c (expand_call_stmt): Adjust.
8301 2014-01-08 Rong Xu <xur@google.com>
8303 * gcov-io.c (gcov_var): Move from gcov-io.h.
8304 (gcov_position): Ditto.
8305 (gcov_is_error): Ditto.
8306 (gcov_rewrite): Ditto.
8307 * gcov-io.h: Refactor. Move gcov_var to gcov-io.h, and libgcov
8308 only part to libgcc/libgcov.h.
8310 2014-01-08 Marek Polacek <polacek@redhat.com>
8313 * omp-low.c (simd_clone_adjust): Don't crash if def is NULL.
8315 2014-01-08 Marek Polacek <polacek@redhat.com>
8318 * ubsan.c (ubsan_type_descriptor): Call strip_array_types on type2.
8320 2014-01-08 Jakub Jelinek <jakub@redhat.com>
8322 PR rtl-optimization/59649
8323 * stor-layout.c (get_mode_bounds): For BImode return
8324 0 and STORE_FLAG_VALUE.
8326 2014-01-08 Richard Biener <rguenther@suse.de>
8329 * gimple.h (is_gimple_builtin_call): Remove.
8330 (gimple_builtin_call_types_compatible_p): New.
8331 (gimple_call_builtin_p): New overload.
8332 * gimple.c (is_gimple_builtin_call): Remove.
8333 (validate_call): Rename to ...
8334 (gimple_builtin_call_types_compatible_p): ... this and export. Also
8336 (validate_type): New static function.
8337 (gimple_call_builtin_p): New overload and adjust.
8338 * gimple-fold.c (gimple_fold_builtin): Fold the return value.
8339 (gimple_fold_call): Likewise. Use gimple_call_builtin_p.
8340 (gimple_fold_stmt_to_constant_1): Likewise.
8341 * tsan.c (instrument_gimple): Use gimple_call_builtin_p.
8343 2014-01-08 Richard Biener <rguenther@suse.de>
8346 * gimplify.c (gimplify_expr): Gimplify register-register type
8347 VIEW_CONVERT_EXPRs to separate stmts.
8349 2014-01-07 Jeff Law <law@redhat.com>
8352 * ree.c (combine_set_extension): Handle case where source
8353 and destination registers in an extension insn are different.
8354 (combine_reaching_defs): Allow source and destination registers
8355 in extension to be different under limited circumstances.
8356 (add_removable_extension): Remove restriction that the
8357 source and destination registers in the extension are the same.
8358 (find_and_remove_re): Emit a copy from the extension's
8359 destination to its source after the defining insn if
8360 the source and destination registers are different.
8363 * ifcvt.c (merge_if_block): If we are merging a block with more than
8364 one successor with a block with no successors, remove any BARRIER
8365 after the second block.
8367 2014-01-07 Dan Xio Qiang <ziyan01@163.com>
8369 * hw-doloop.c (reorg_loops): Release the bitmap obstack.
8371 2014-01-07 John David Anglin <danglin@gcc.gnu.org>
8374 * config/pa/pa.c (pa_legitimate_address_p): Return false before reload
8375 for 14-bit register offsets when INT14_OK_STRICT is false.
8377 2014-01-07 Roland Stigge <stigge@antcom.de>
8378 Michael Meissner <meissner@linux.vnet.ibm.com>
8381 * config/rs6000/rs6000.c (rs6000_legitimate_offset_address_p):
8382 Only check TFmode for SPE constants. Don't check TImode or TDmode.
8384 2014-01-07 James Greenhalgh <james.greenhalgh@arm.com>
8386 * config/aarch64/aarch64-elf.h (ASM_SPEC): Remove identity spec for
8389 2014-01-07 Yufeng Zhang <yufeng.zhang@arm.com>
8391 * config/arm/arm.c (arm_expand_neon_args): Call expand_expr
8392 with EXPAND_MEMORY for NEON_ARG_MEMORY; check if the returned
8393 rtx is const0_rtx or not.
8395 2014-01-07 Richard Sandiford <rdsandiford@googlemail.com>
8398 * target-globals.c (save_target_globals): Remove this_fn_optab
8400 * toplev.c: Include optabs.h.
8401 (target_reinit): Temporarily restore the global options if another
8402 set of options are in force.
8404 2014-01-07 Jakub Jelinek <jakub@redhat.com>
8406 PR rtl-optimization/58668
8407 * cfgcleanup.c (flow_find_cross_jump): Don't count
8408 any jumps if dir_p is NULL. Remove p1 variable, use active_insn_p
8409 to determine what is counted.
8410 (flow_find_head_matching_sequence): Use active_insn_p to determine
8412 (try_head_merge_bb): Adjust for the flow_find_head_matching_sequence
8414 * ifcvt.c (count_bb_insns): Use active_insn_p && !JUMP_P to
8415 determine what is counted.
8417 PR tree-optimization/59643
8418 * tree-predcom.c (split_data_refs_to_components): If one dr is
8419 read and one write, determine_offset fails and the write isn't
8420 in the bad component, just put the read into the bad component.
8422 2014-01-07 Mike Stump <mikestump@comcast.net>
8423 Jakub Jelinek <jakub@redhat.com>
8426 * tree-core.h (struct tree_optimization_option): Change optabs
8427 type from unsigned char * to void *.
8428 * optabs.c (init_tree_optimization_optabs): Adjust
8429 TREE_OPTIMIZATION_OPTABS initialization.
8431 2014-01-06 Jakub Jelinek <jakub@redhat.com>
8434 * config/i386/i386.h (struct machine_function): Add
8435 no_drap_save_restore field.
8436 * config/i386/i386.c (ix86_save_reg): Use
8437 !cfun->machine->no_drap_save_restore instead of
8438 crtl->stack_realign_needed.
8439 (ix86_finalize_stack_realign_flags): Don't clear drap_reg unless
8440 this function clears frame_pointer_needed. Set
8441 cfun->machine->no_drap_save_restore if clearing frame_pointer_needed
8442 and DRAP reg is needed.
8444 2014-01-06 Marek Polacek <polacek@redhat.com>
8447 * doc/implement-c.texi: Mention that other integer types are
8448 permitted as bit-field types in strictly conforming mode.
8450 2014-01-06 Felix Yang <fei.yang0953@gmail.com>
8452 * modulo-sched.c (schedule_reg_moves): Clear distance1_uses if it
8455 2014-01-06 Richard Earnshaw <rearnsha@arm.com>
8457 * aarch64.c (aarch64_rtx_costs): Fix cost calculation for MADD.
8459 2014-01-06 Martin Jambor <mjambor@suse.cz>
8462 * ipa-cp.c (ipcp_discover_new_direct_edges): Changed param_index type
8464 * ipa-prop.c (ipa_print_node_params): Fix indentation.
8466 2014-01-06 Eric Botcazou <ebotcazou@adacore.com>
8470 * var-tracking.c (add_stores): Preserve the value of the source even if
8471 we don't record the store.
8473 2014-01-06 Terry Guo <terry.guo@arm.com>
8475 * config.gcc (arm*-*-*): Check --with-arch against arm-arches.def.
8477 2014-01-05 Iain Sandoe <iain@codesourcery.com>
8480 * config/darwin.c (darwin_function_section): Adjust return values to
8481 correspond to optimisation changes made in r206070.
8483 2014-01-05 Uros Bizjak <ubizjak@gmail.com>
8485 * config/i386/i386.c (ix86_data_alignment): Calculate max_align
8486 from prefetch_block tune setting.
8487 (nocona_cost): Correct size of prefetch block to 64.
8489 2014-01-04 Eric Botcazou <ebotcazou@adacore.com>
8491 * config/arm/arm.c (arm_get_frame_offsets): Revamp long lines.
8492 (arm_expand_epilogue_apcs_frame): Take into account the number of bytes
8493 used to save the static chain register in the computation of the offset
8494 from which the FP registers need to be restored.
8496 2014-01-04 Jakub Jelinek <jakub@redhat.com>
8498 PR tree-optimization/59519
8499 * tree-vect-loop-manip.c (slpeel_update_phi_nodes_for_guard1): Don't
8500 ICE if get_current_def (current_new_name) is already non-NULL, as long
8501 as it is a phi result of some other phi in *new_exit_bb that has
8504 * config/i386/sse.md (avx512f_load<mode>_mask): Emit vmovup{s,d}
8505 or vmovdqu* for misaligned_operand.
8506 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>,
8507 <sse2_avx_avx512f>_loaddqu<mode><mask_name>): Handle <mask_applied>.
8508 * config/i386/i386.c (ix86_expand_special_args_builtin): Set
8509 aligned_mem for AVX512F masked aligned load and store builtins and for
8512 2014-01-03 Bingfeng Mei <bmei@broadcom.com>
8514 PR tree-optimization/59651
8515 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
8516 Address range for negative step should be added by TYPE_SIZE_UNIT.
8518 2014-01-03 Andreas Schwab <schwab@linux-m68k.org>
8520 * config/m68k/m68k.c (handle_move_double): Handle pushes with
8521 overlapping registers also for registers other than the stack pointer.
8523 2014-01-03 Marek Polacek <polacek@redhat.com>
8526 * doc/extend.texi: Fix the return value of __builtin_FUNCTION and
8529 2014-01-03 Jakub Jelinek <jakub@redhat.com>
8532 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Don't consider
8535 * config/i386/i386.md (MODE_SIZE): New mode attribute.
8536 (push splitter): Use <P:MODE_SIZE> instead of
8537 GET_MODE_SIZE (<P:MODE>mode).
8538 (lea splitter): Use <MODE_SIZE> instead of GET_MODE_SIZE (<MODE>mode).
8539 (mov -1, reg peephole2): Likewise.
8540 * config/i386/sse.md (*mov<mode>_internal,
8541 <sse>_storeu<ssemodesuffix><avxsizesuffix>,
8542 <sse2_avx_avx512f>_storedqu<mode>, <sse>_andnot<mode>3,
8543 *<code><mode>3, *andnot<mode>3<mask_name>,
8544 <mask_codefor><code><mode>3<mask_name>): Likewise.
8545 * config/i386/subst.md (mask_mode512bit_condition,
8546 sd_mask_mode512bit_condition): Likewise.
8548 2014-01-02 Xinliang David Li <davidxl@google.com>
8550 PR tree-optimization/59303
8551 * tree-ssa-uninit.c (is_use_properly_guarded): Main cleanup.
8552 (dump_predicates): Better output format.
8553 (pred_equal_p): New function.
8554 (is_neq_relop_p): Ditto.
8555 (is_neq_zero_form_p): Ditto.
8556 (pred_expr_equal_p): Ditto.
8557 (pred_neg_p): Ditto.
8558 (simplify_pred): Ditto.
8559 (simplify_preds_2): Ditto.
8560 (simplify_preds_3): Ditto.
8561 (simplify_preds_4): Ditto.
8562 (simplify_preds): Ditto.
8564 (push_to_worklist): Ditto.
8565 (get_pred_info_from_cmp): Ditto.
8566 (is_degenerated_phi): Ditto.
8567 (normalize_one_pred_1): Ditto.
8568 (normalize_one_pred): Ditto.
8569 (normalize_one_pred_chain): Ditto.
8570 (normalize_preds): Ditto.
8571 (normalize_cond_1): Remove function.
8572 (normalize_cond): Ditto.
8573 (is_gcond_subset_of): Ditto.
8574 (is_subset_of_any): Ditto.
8575 (is_or_set_subset_of): Ditto.
8576 (is_and_set_subset_of): Ditto.
8577 (is_norm_cond_subset_of): Ditto.
8578 (pred_chain_length_cmp): Ditto.
8579 (convert_control_dep_chain_into_preds): Type change.
8580 (find_predicates): Ditto.
8581 (find_def_preds): Ditto.
8582 (destroy_predicates_vecs): Ditto.
8583 (find_matching_predicates_in_rest_chains): Ditto.
8584 (use_pred_not_overlap_with_undef_path_pred): Ditto.
8585 (is_pred_expr_subset): Ditto.
8586 (is_pred_chain_subset_of): Ditto.
8587 (is_included_in): Ditto.
8588 (is_superset_of): Ditto.
8590 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
8592 Update copyright years.
8594 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
8596 * common/config/arc/arc-common.c, config/arc/arc-modes.def,
8597 config/arc/arc-protos.h, config/arc/arc.c, config/arc/arc.h,
8598 config/arc/arc.md, config/arc/arc.opt,
8599 config/arm/arm_neon_builtins.def, config/arm/crypto.def,
8600 config/i386/avx512cdintrin.h, config/i386/avx512erintrin.h,
8601 config/i386/avx512fintrin.h, config/i386/avx512pfintrin.h,
8602 config/i386/btver2.md, config/i386/shaintrin.h, config/i386/slm.md,
8603 config/linux-protos.h, config/linux.c, config/winnt-c.c,
8604 diagnostic-color.c, diagnostic-color.h, gimple-ssa-isolate-paths.c,
8605 vtable-verify.c, vtable-verify.h: Use the standard form for the
8608 2014-01-02 Tobias Burnus <burnus@net-b.de>
8610 * gcc.c (process_command): Update copyright notice dates.
8611 * gcov-dump.c: Ditto.
8613 * doc/cpp.texi: Bump @copying's copyright year.
8614 * doc/cppinternals.texi: Ditto.
8615 * doc/gcc.texi: Ditto.
8616 * doc/gccint.texi: Ditto.
8617 * doc/gcov.texi: Ditto.
8618 * doc/install.texi: Ditto.
8619 * doc/invoke.texi: Ditto.
8621 2014-01-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8623 * config/nios2/nios2.h (BITS_PER_UNIT): Don't define it.
8625 2014-01-01 Jakub Jelinek <jakub@redhat.com>
8627 * config/i386/sse.md (*mov<mode>_internal): Guard
8628 EXT_REX_SSE_REGNO_P (REGNO ()) uses with REG_P.
8630 PR rtl-optimization/59647
8631 * cse.c (cse_process_notes_1): Don't substitute negative VOIDmode
8632 new_rtx into UNSIGNED_FLOAT rtxes.
8634 Copyright (C) 2014 Free Software Foundation, Inc.
8636 Copying and distribution of this file, with or without modification,
8637 are permitted in any medium without royalty provided the copyright
8638 notice and this notice are preserved.