1 2014-04-29 Nick Clifton <nickc@redhat.com>
3 * config/arc/arc.c (arc_select_cc_mode): Fix parentheses.
4 (arc_init_reg_tables): Use a machine_mode enum to iterate over
6 * config/m32r/m32r.c (init_reg_tables): Likewise.
7 * config/m32c/m32c.c (m32c_illegal_subreg_p): Use a machine_mode
8 enum to hold the modes.
10 2014-04-29 Richard Biener <rguenther@suse.de>
12 * dominance.c (free_dominance_info): Add overload with
14 (dom_info_state): Likewise.
15 (dom_info_available_p): Likewise.
16 * basic-block.h (free_dominance_info, dom_info_state,
17 dom_info_available_p): Declare overloads.
18 * passes.c (execute_function_todo): Verify that verifiers
19 don't change dominator info state. Drop dominator info
20 for IPA pass invocations.
21 * cgraph.c (release_function_body): Restore asserts that
22 dominator information is released.
24 2014-04-29 Patrick Palka <patrick@parcs.ath.cx>
26 * doc/invoke.texi: Fix typo.
27 * tree-vrp.c: Fix typos.
28 * gimple.c (infer_nonnull_range): Reorder operands of an &&
31 2014-04-29 Zhenqiang Chen <zhenqiang.chen@linaro.org>
33 * config/aarch64/aarch64.md (mov<mode>cc): New for GPF.
35 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
37 * config/aarch64/aarch64-builtins.c
38 (aarch64_types_storestruct_lane_qualifiers): New.
39 (TYPES_STORESTRUCT_LANE): Likewise.
40 * config/aarch64/aarch64-simd-builtins.def (st2_lane): New.
43 * config/aarch64/aarch64-simd.md (vec_store_lanesoi_lane<mode>): New.
44 (vec_store_lanesci_lane<mode>): Likewise.
45 (vec_store_lanesxi_lane<mode>): Likewise.
46 (aarch64_st2_lane<VQ:mode>): Likewise.
47 (aarch64_st3_lane<VQ:mode>): Likewise.
48 (aarch64_st4_lane<VQ:mode>): Likewise.
49 * config/aarch64/aarch64.md (unspec): Add UNSPEC_ST{2,3,4}_LANE.
50 * config/aarch64/arm_neon.h
51 (__ST2_LANE_FUNC): Rewrite using builtins, update use points to
52 use new macro arguments.
53 (__ST3_LANE_FUNC): Likewise.
54 (__ST4_LANE_FUNC): Likewise.
55 * config/aarch64/iterators.md (V_TWO_ELEM): New.
56 (V_THREE_ELEM): Likewise.
57 (V_FOUR_ELEM): Likewise.
59 2014-04-28 David Malcolm <dmalcolm@redhat.com>
61 * doc/gimple.texi: Replace the description of the now-defunct
62 union gimple_statement_d with a diagram showing the
63 gimple_statement_base class hierarchy and its relationships to
64 the GSS_ and GIMPLE_ enums.
66 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
68 * config/aarch64/aarch64-protos.h (aarch64_modes_tieable_p): New.
69 * config/aarch64/aarch64.c
70 (aarch64_cannot_change_mode_class): Weaken conditions.
71 (aarch64_modes_tieable_p): New.
72 * config/aarch64/aarch64.h (MODES_TIEABLE_P): Use it.
74 2014-04-28 Pat Haugen <pthaugen@us.ibm.com>
76 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
77 (loadsync_<mode>): Change mode.
78 (load_quadpti, store_quadpti): New.
79 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
80 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
82 2014-04-28 Martin Jambor <mjambor@suse.cz>
84 * tree-sra.c (sra_modify_expr): Generate new memory accesses with
85 same alias type as the original statement.
86 (subreplacement_assignment_data): New type.
87 (handle_unscalarized_data_in_subtree): New type of parameter,
88 generate new memory accesses with same alias type as the original
90 (load_assign_lhs_subreplacements): Likewise.
91 (sra_modify_constructor_assign): Generate new memory accesses with
92 same alias type as the original statement.
94 2014-04-28 Richard Biener <rguenther@suse.de>
96 * tree-pass.h (TODO_verify_il): Define.
97 (TODO_verify_all): Complete properly.
98 * passes.c (execute_function_todo): Move existing loop-closed
99 SSA verification under TODO_verify_il.
100 (execute_one_pass): Trigger TODO_verify_il at todo-after time.
101 * graphite-sese-to-poly.c (rewrite_cross_bb_scalar_deps):
102 Fix tree sharing issue.
104 2014-04-28 Richard Biener <rguenther@suse.de>
107 * builtins.def (DEF_C11_BUILTIN): Add.
108 (BUILT_IN_ALIGNED_ALLOC): Likewise.
109 * coretypes.h (enum function_class): Add function_c11_misc.
110 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
111 BUILT_IN_ALIGNED_ALLOC like BUILT_IN_MALLOC.
112 (call_may_clobber_ref_p_1): Likewise.
113 * tree-ssa-dce.c (mark_stmt_if_obviously_necessary): Likewise.
114 (mark_all_reaching_defs_necessary_1): Likewise.
115 (propagate_necessity): Likewise.
116 (eliminate_unnecessary_stmts): Likewise.
117 * tree-ssa-ccp.c (evaluate_stmt): Handle BUILT_IN_ALIGNED_ALLOC.
119 2014-04-28 Richard Biener <rguenther@suse.de>
121 * tree-vrp.c (vrp_var_may_overflow): Remove.
122 (vrp_visit_phi_node): Rather than bumping to +-INF possibly
123 with overflow immediately bump to one before that value and
124 let iteration figure out overflow status.
126 2014-04-28 Richard Biener <rguenther@suse.de>
128 * configure.ac: Do valgrind header checks unconditionally.
129 Add --enable-valgrind-annotations.
130 * system.h: Guard valgrind header inclusion with
131 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING.
132 * alloc-pool.c (pool_alloc, pool_free): Use
133 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING
134 to guard possibly dead code.
135 * config.in: Regenerated.
136 * configure: Likewise.
138 2014-04-28 Jeff Law <law@redhat.com>
140 PR tree-optimization/60902
141 * tree-ssa-threadedge.c
142 (record_temporary_equivalences_from_stmts_at_dest): Only iterate
143 over real defs when invalidating outputs from statements that do not
144 produce useful outputs for threading.
146 2014-04-28 Richard Biener <rguenther@suse.de>
148 PR tree-optimization/60979
149 * graphite-scop-detection.c (scopdet_basic_block_info): Reject
150 SCOPs that end in a block with a successor with abnormal
153 2014-04-28 Richard Biener <rguenther@suse.de>
155 * tree-pass.h (execute_pass_list): Adjust prototype.
156 * passes.c (pass_manager::execute_early_local_passes):
158 (do_per_function): Change callback signature, push all actual
159 work to the callbals.
160 (do_per_function_toporder): Likewise.
161 (execute_function_dump): Adjust.
162 (execute_function_todo): Likewise.
163 (clear_last_verified): Likewise.
164 (verify_curr_properties): Likewise.
165 (update_properties_after_pass): Likewise.
166 (execute_pass_list_1): Split out from ...
167 (execute_pass_list): ... here. Adjust.
168 (execute_ipa_pass_list): Likewise.
169 * cgraphunit.c (cgraph_add_new_function): Adjust.
170 (analyze_function): Likewise.
171 (expand_function): Likewise.
172 * cgraph.c (release_function_body): Free dominance info
173 here instead of asserting it was magically freed elsewhere.
175 2014-04-28 Eric Botcazou <ebotcazou@adacore.com>
177 * configure.ac: Tweak GAS check for LEON instructions on SPARC.
178 * configure: Regenerate.
179 * config/sparc/sparc.opt (muser-mode): New option.
180 * config/sparc/sync.md (atomic_compare_and_swap<mode>_1): Do not enable
182 (atomic_compare_and_swap_leon3_1): New instruction for LEON3.
183 * doc/invoke.texi (SPARC options): Document -muser-mode.
185 2014-04-27 Richard Sandiford <rdsandiford@googlemail.com>
187 * cselib.c (find_slot_memmode): Delete.
188 (cselib_hasher): Change compare_type to a struct.
189 (cselib_hasher::equal): Update accordingly. Don't expect wrapped
191 (preserve_constants_and_equivs): Adjust for new compare_type.
192 (cselib_find_slot): Likewise. Take the mode of the rtx as argument.
193 (wrap_constant): Delete.
194 (cselib_lookup_mem, cselib_lookup_1): Update calls to cselib_find_slot.
196 2014-04-26 Markus Trippelsdorf <markus@trippelsdorf.de>
198 * doc/install.texi (Building with profile feedback): Remove
201 2014-04-26 Tom de Vries <tom@codesourcery.com>
203 * config/i386/i386.md (define_expand "ldexpxf3"): Fix out-of-bounds
206 2014-04-25 Cary Coutant <ccoutant@google.com>
209 * dwarf2out.c (should_move_die_to_comdat): A type definition
210 can contain a subprogram definition, but don't move it to a
212 (clone_as_declaration): Copy DW_AT_abstract_origin attribute.
213 (generate_skeleton_bottom_up): Remove DW_AT_object_pointer attribute
215 (clone_tree_hash): Rename to...
216 (clone_tree_partial): ...this; change callers. Copy
217 DW_TAG_subprogram DIEs as declarations.
218 (copy_decls_walk): Don't copy children of a declaration into a
221 2014-04-25 H.J. Lu <hongjiu.lu@intel.com>
224 * config/i386/i386.md (*movsf_internal): Set MODE to SI for
227 2014-04-25 Jiong Wang <jiong.wang@arm.com>
229 * config/arm/predicates.md (call_insn_operand): Add long_call check.
230 * config/arm/arm.md (sibcall, sibcall_value): Force the address to
232 * config/arm/arm.c (arm_function_ok_for_sibcall): Remove long_call
235 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
237 * config/arm/arm.c (arm_cortex_a8_tune): Initialise
240 2014-04-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
242 PR tree-optimization/60930
243 * gimple-ssa-strength-reduction.c (create_mul_imm_cand): Reject
244 creating a multiply candidate by folding two constant
245 multiplicands when the result overflows.
247 2014-04-25 Jakub Jelinek <jakub@redhat.com>
249 PR tree-optimization/60960
250 * tree-vect-generic.c (expand_vector_operation): Only call
251 expand_vector_divmod if type's mode satisfies VECTOR_MODE_P.
253 2014-04-25 Tom de Vries <tom@codesourcery.com>
255 * expr.c (clobber_reg_mode): New function.
256 * expr.h (clobber_reg): New function.
258 2014-04-25 Tom de Vries <tom@codesourcery.com>
260 * rtlanal.c (find_all_hard_reg_sets): Note INSN_CALL_FUNCTION_USAGE
263 2014-04-25 Radovan Obradovic <robradovic@mips.com>
264 Tom de Vries <tom@codesourcery.com>
266 * rtlanal.c (find_all_hard_reg_sets): Add bool implicit parameter and
268 * rtl.h (find_all_hard_reg_sets): Add bool parameter.
269 * haifa-sched.c (recompute_todo_spec, check_clobbered_conditions): Add
270 new argument to find_all_hard_reg_sets call.
272 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
274 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p):
275 Use HOST_WIDE_INT_C for mask literal.
276 (aarch_rev16_shleft_mask_imm_p): Likewise.
278 2014-04-25 Eric Botcazou <ebotcazou@adacore.com>
281 * config/sparc/sparc.md (ashlsi3_extend): Delete.
283 2014-04-25 Marc Glisse <marc.glisse@inria.fr>
285 PR preprocessor/56540
286 * config/i386/i386-c.c (ix86_target_macros): Define
287 __SIZEOF_FLOAT80__ and __SIZEOF_FLOAT128__.
289 2014-04-25 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
291 * configure.ac (tga_func): Remove.
292 (LIB_TLS_SPEC): Remove.
293 * configure: Regenerate.
294 * config.in: Regenerate.
295 * config/sol2.h (LIB_SPEC): Don't use LIB_TLS_SPEC.
297 2014-04-25 Richard Biener <rguenther@suse.de>
300 * tree-ssa-structalias.c (ipa_pta_execute): Compute direct
301 call stmt use/clobber sets during stmt walk instead of
302 walking the possibly incomplete set of caller edges.
304 2014-04-25 Richard Biener <rguenther@suse.de>
307 * passes.c (apply_ipa_transforms): Inline into only caller ...
308 (execute_one_pass): ... here. Properly bring in function
309 bodies for nodes we want to apply IPA transforms to.
311 2014-04-24 Cong Hou <congh@google.com>
313 PR tree-optimization/60896
314 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Pick up
315 all statements in PATTERN_DEF_SEQ in recognized widen-mult pattern.
316 (vect_mark_pattern_stmts): Set the def type of all statements in
317 PATTERN_DEF_SEQ as vect_internal_def.
319 2014-04-24 Michael Meissner <meissner@linux.vnet.ibm.com>
321 * doc/extend.texi (PowerPC Built-in Functions): Document new
322 powerpc extended divide, bcd, pack/unpack 128-bit, builtin
324 (PowerPC AltiVec/VSX Built-in Functions): Likewise.
326 * config/rs6000/predicates.md (const_0_to_3_operand): New
327 predicate to match 0..3 integer constants.
329 * config/rs6000/rs6000-builtin.def (BU_DFP_MISC_1): Add new macros
330 to support adding miscellaneous builtin functions.
331 (BU_DFP_MISC_2): Likewise.
332 (BU_P7_MISC_1): Likewise.
333 (BU_P7_MISC_2): Likewise.
334 (BU_P8V_MISC_3): Likewise.
335 (BU_MISC_1): Likewise.
336 (BU_MISC_2): Likewise.
337 (DIVWE): Add extended divide builtin functions.
345 (DXEX): Add decimal floating-point builtin functions.
357 (CDTBCD): Add new BCD builtin functions.
361 (BCDADD_LT): Likewise.
362 (BCDADD_EQ): Likewise.
363 (BCDADD_GT): Likewise.
364 (BCDADD_OV): Likewise.
366 (BCDSUB_LT): Likewise.
367 (BCDSUB_EQ): Likewise.
368 (BCDSUB_GT): Likewise.
369 (BCDSUB_OV): Likewise.
370 (PACK_TD): Add new pack/unpack 128-bit type builtin functions.
371 (UNPACK_TD): Likewise.
373 (UNPACK_TF): Likewise.
374 (UNPACK_TF_0): Likewise.
375 (UNPACK_TF_1): Likewise.
376 (PACK_V1TI): Likewise.
377 (UNPACK_V1TI): Likewise.
379 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Add
380 support for decimal floating point builtin functions.
381 (rs6000_expand_ternop_builtin): Add checks for the new builtin
382 functions that take constant arguments.
383 (rs6000_invalid_builtin): Add decimal floating point builtin
385 (rs6000_init_builtins): Setup long double, _Decimal64, and
386 _Decimal128 types for new builtin functions.
387 (builtin_function_type): Set the unsigned flags appropriately for
388 the new builtin functions.
389 (rs6000_opt_masks): Add support for decimal floating point builtin
392 * config/rs6000/rs6000.h (RS6000_BTM_DFP): Add support for decimal
393 floating point builtin functions.
394 (RS6000_BTM_COMMON): Likewise.
395 (RS6000_BTI_long_double): Likewise.
396 (RS6000_BTI_dfloat64): Likewise.
397 (RS6000_BTI_dfloat128): Likewise.
398 (long_double_type_internal_node): Likewise.
399 (dfloat64_type_internal_node): Likewise.
400 (dfloat128_type_internal_node): Likewise.
402 * config/rs6000/altivec.h (UNSPEC_BCDADD): Add support for ISA
403 2.07 bcd arithmetic instructions.
404 (UNSPEC_BCDSUB): Likewise.
405 (UNSPEC_BCD_OVERFLOW): Likewise.
406 (UNSPEC_BCD_ADD_SUB): Likewise.
407 (bcd_add_sub): Likewise.
408 (BCD_TEST): Likewise.
409 (bcd<bcd_add_sub>): Likewise.
410 (bcd<bcd_add_sub>_test): Likewise.
411 (bcd<bcd_add_sub>_test2): Likewise.
412 (bcd<bcd_add_sub>_<code>): Likewise.
413 (peephole2 for combined bcd ops): Likewise.
415 * config/rs6000/dfp.md (UNSPEC_DDEDPD): Add support for new
416 decimal floating point builtin functions.
417 (UNSPEC_DENBCD): Likewise.
418 (UNSPEC_DXEX): Likewise.
419 (UNSPEC_DIEX): Likewise.
420 (UNSPEC_DSCLI): Likewise.
421 (UNSPEC_DSCRI): Likewise.
422 (D64_D128): Likewise.
423 (dfp_suffix): Likewise.
424 (dfp_ddedpd_<mode>): Likewise.
425 (dfp_denbcd_<mode>): Likewise.
426 (dfp_dxex_<mode>): Likewise.
427 (dfp_diex_<mode>): Likewise.
428 (dfp_dscli_<mode>): Likewise.
429 (dfp_dscri_<mode>): Likewise.
431 * config/rs6000/rs6000.md (UNSPEC_ADDG6S): Add support for new BCD
433 (UNSPEC_CDTBCD): Likewise.
434 (UNSPEC_CBCDTD): Likewise.
435 (UNSPEC_DIVE): Add support for new extended divide builtin
437 (UNSPEC_DIVEO): Likewise.
438 (UNSPEC_DIVEU): Likewise.
439 (UNSPEC_DIVEUO): Likewise.
440 (UNSPEC_UNPACK_128BIT): Add support for new builtin functions to
441 pack/unpack 128-bit types.
442 (UNSPEC_PACK_128BIT): Likewise.
443 (idiv_ldiv): New mode attribute to set the 32/64-bit divide type.
444 (udiv<mode>3): Use idiv_ldiv mode attribute.
445 (div<mode>3): Likewise.
446 (addg6s): Add new BCD builtin functions.
449 (UNSPEC_DIV_EXTEND): Add support for new extended divide
451 (div_extend): Likewise.
452 (div<div_extend>_<mode>"): Likewise.
453 (FP128_64): Add support for new builtin functions to pack/unpack
455 (unpack<mode>): Likewise.
456 (unpacktf_0): Likewise.
457 (unpacktf_1): Likewise.
458 (unpack<mode>_dm): Likewise.
459 (unpack<mode>_nodm): Likewise.
460 (pack<mode>): Likewise.
461 (unpackv1ti): Likewise.
462 (packv1ti): Likewise.
464 2014-04-24 Vishnu K S <Vishnu.k_s@atmel.com>
466 * gcc/config/avr/avr.c: Add comment on why -fdelete-null-pointer-checks
469 2014-04-24 Jakub Jelinek <jakub@redhat.com>
471 * tree.h (OMP_CLAUSE_LINEAR_GIMPLE_SEQ): Define.
472 * gimplify.c (omp_is_private): Change last argument's type to int.
473 Only diagnose lastprivate if the simd argument is 1, only diagnose
474 linear if the simd argument is 2.
475 (gimplify_omp_for): Adjust omp_is_private callers. When adding
476 lastprivate or private, add the clause to OMP_FOR_CLAUSES. Pass
477 GOVD_EXPLICIT to omp_add_variable. For simd with collapse == 1
478 create OMP_CLAUSE_LINEAR rather than OMP_CLAUSE_PRIVATE for var.
479 If var != decl and decl is in OMP_CLAUSE_LINEAR, gimplify decl
480 increment to OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
481 * omp-low.c (scan_sharing_clauses, lower_lastprivate_clauses): Handle
482 OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
483 * tree-nested.c (convert_nonlocal_omp_clauses,
484 convert_local_omp_clauses): Handle OMP_CLAUSE_LINEAR.
486 2014-04-24 Segher Boessenkool <segher@kernel.crashing.org>
489 * config/m68k/m68k.md (extendplussidi): Don't allow memory for
492 2014-04-24 Dimitris Papavasiliou <dpapavas@gmail.com>
494 * flag-types.h (enum ivar_visibility): Add.
496 2014-04-24 Trevor Saunders <tsaunders@mozilla.com>
498 * config/sh/sh_treg_combine.c (sh_treg_combine::execute): Take
501 2014-04-24 Alan Lawrence <alan.lawrence@arm.com>
503 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Enable for bigendian.
505 2014-04-24 Radovan Obradovic <robradovic@mips.com>
506 Tom de Vries <tom@codesourcery.com>
508 * reg-notes.def (REG_NOTE (CALL_DECL)): New reg-note REG_CALL_DECL.
509 * calls.c (expand_call, emit_library_call_value_1): Add REG_CALL_DECL
511 * combine.c (distribute_notes): Handle REG_CALL_DECL reg-note.
512 * emit-rtl.c (try_split): Same.
514 2014-04-24 Radovan Obradovic <robradovic@mips.com>
515 Tom de Vries <tom@codesourcery.com>
517 * common.opt (fuse-caller-save): New option.
519 2014-04-24 Tejas Belagod <tejas.belagod@arm.com>
521 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Reverse order of elements
524 2014-04-24 Richard Biener <rguenther@suse.de>
526 * expr.c (expand_expr_real_1): Avoid gimple_assign_rhs_to_tree
527 during TER and instead use the sepops interface for expanding
528 non-GIMPLE_SINGLE_RHS.
530 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
532 * config/i386/sol2.h (ASM_PREFERRED_EH_DATA_FORMAT): Only redefine
533 if not HAVE_AS_IX86_DIFF_SECT_DELTA.
535 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
537 * configure.ac (gcc_cv_as_cfi_directive): Support Solaris/x86
538 assembler 64-bit option.
539 * configure: Regenerate.
541 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
543 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Check
544 TARGET_SIMD rather than TARGET_GENERAL_REGS_ONLY.
545 (TARGET_SIMD): Take AARCH64_ISA_SIMD into account.
546 (TARGET_FLOAT): Take AARCH64_ISA_FP into account.
547 (TARGET_CRYPTO): Take TARGET_SIMD into account.
549 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
551 * config/aarch64/aarch64-builtins.c
552 (aarch64_builtin_vectorized_function): Handle BUILT_IN_BSWAP16,
553 BUILT_IN_BSWAP32, BUILT_IN_BSWAP64.
554 * config/aarch64/aarch64-simd.md (bswap<mode>): New pattern.
555 * config/aarch64/aarch64-simd-builtins.def: Define vector bswap
557 * config/aarch64/iterator.md (VDQHSD): New mode iterator.
558 (Vrevsuff): New mode attribute.
560 2014-04-24 Terry Guo <terry.guo@arm.com>
562 * config/arm/arm.h (machine_function): Define variable
563 after_arm_reorg here.
564 * config/arm/arm.c (after_arm_reorg): Remove the definition.
565 (arm_split_constant): Update the way to access variable
568 (arm_output_function_epilogue): Remove the reset of after_arm_reorg.
570 2014-04-23 Tom de Vries <tom@codesourcery.com>
572 * target-hooks-macros.h: Fix DEFHOOKPOD argument order in comment.
574 2014-04-23 David Malcolm <dmalcolm@redhat.com>
576 * is-a.h: Update comments to reflect the following changes to the
577 "pointerness" of the API, making the template parameter match the
578 return type, allowing use of is-a.h with typedefs of pointers.
579 (is_a_helper::cast): Return a T rather then a pointer to a T, so
580 that the return type matches the parameter to the is_a_helper.
582 (dyn_cast): Likewise.
584 * cgraph.c (cgraph_node_for_asm): Update for removal of implicit
585 pointer from the is-a.h API.
587 * cgraph.h (is_a_helper <cgraph_node>::test): Convert to...
588 (is_a_helper <cgraph_node *>::test): ...this, matching change to
590 (is_a_helper <varpool_node>::test): Likewise, convert to...
591 (is_a_helper <varpool_node *>::test): ...this.
593 (varpool_first_variable): Update for removal of implicit pointer
595 (varpool_next_variable): Likewise.
596 (varpool_first_static_initializer): Likewise.
597 (varpool_next_static_initializer): Likewise.
598 (varpool_first_defined_variable): Likewise.
599 (varpool_next_defined_variable): Likewise.
600 (cgraph_first_defined_function): Likewise.
601 (cgraph_next_defined_function): Likewise.
602 (cgraph_first_function): Likewise.
603 (cgraph_next_function): Likewise.
604 (cgraph_first_function_with_gimple_body): Likewise.
605 (cgraph_next_function_with_gimple_body): Likewise.
606 (cgraph_alias_target): Likewise.
607 (varpool_alias_target): Likewise.
608 (cgraph_function_or_thunk_node): Likewise.
609 (varpool_variable_node): Likewise.
610 (symtab_real_symbol_p): Likewise.
611 * cgraphunit.c (referred_to_p): Likewise.
612 (analyze_functions): Likewise.
613 (handle_alias_pairs): Likewise.
614 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
615 * gimple-ssa.h (gimple_vuse_op): Likewise.
616 (gimple_vdef_op): Likewise.
617 * gimple-streamer-in.c (input_gimple_stmt): Likewise.
618 * gimple.c (gimple_build_asm_1): Likewise.
619 (gimple_build_try): Likewise.
620 (gimple_build_resx): Likewise.
621 (gimple_build_eh_dispatch): Likewise.
622 (gimple_build_omp_for): Likewise.
623 (gimple_omp_for_set_clauses): Likewise.
625 * gimple.h (is_a_helper <gimple_statement_asm>::test): Convert to...
626 (is_a_helper <gimple_statement_asm *>::test): ...this.
627 (is_a_helper <gimple_statement_bind>::test): Convert to...
628 (is_a_helper <gimple_statement_bind *>::test): ...this.
629 (is_a_helper <gimple_statement_call>::test): Convert to...
630 (is_a_helper <gimple_statement_call *>::test): ...this.
631 (is_a_helper <gimple_statement_catch>::test): Convert to...
632 (is_a_helper <gimple_statement_catch *>::test): ...this.
633 (is_a_helper <gimple_statement_resx>::test): Convert to...
634 (is_a_helper <gimple_statement_resx *>::test): ...this.
635 (is_a_helper <gimple_statement_eh_dispatch>::test): Convert to...
636 (is_a_helper <gimple_statement_eh_dispatch *>::test): ...this.
637 (is_a_helper <gimple_statement_eh_else>::test): Convert to...
638 (is_a_helper <gimple_statement_eh_else *>::test): ...this.
639 (is_a_helper <gimple_statement_eh_filter>::test): Convert to...
640 (is_a_helper <gimple_statement_eh_filter *>::test): ...this.
641 (is_a_helper <gimple_statement_eh_mnt>::test): Convert to...
642 (is_a_helper <gimple_statement_eh_mnt *>::test): ...this.
643 (is_a_helper <gimple_statement_omp_atomic_load>::test): Convert to...
644 (is_a_helper <gimple_statement_omp_atomic_load *>::test): ...this.
645 (is_a_helper <gimple_statement_omp_atomic_store>::test): Convert to...
646 (is_a_helper <gimple_statement_omp_atomic_store *>::test): ...this.
647 (is_a_helper <gimple_statement_omp_return>::test): Convert to...
648 (is_a_helper <gimple_statement_omp_return *>::test): ...this.
649 (is_a_helper <gimple_statement_omp_continue>::test): Convert to...
650 (is_a_helper <gimple_statement_omp_continue *>::test): ...this.
651 (is_a_helper <gimple_statement_omp_critical>::test): Convert to...
652 (is_a_helper <gimple_statement_omp_critical *>::test): ...this.
653 (is_a_helper <gimple_statement_omp_for>::test): Convert to...
654 (is_a_helper <gimple_statement_omp_for *>::test): ...this.
655 (is_a_helper <gimple_statement_omp_taskreg>::test): Convert to...
656 (is_a_helper <gimple_statement_omp_taskreg *>::test): ...this.
657 (is_a_helper <gimple_statement_omp_parallel>::test): Convert to...
658 (is_a_helper <gimple_statement_omp_parallel *>::test): ...this.
659 (is_a_helper <gimple_statement_omp_target>::test): Convert to...
660 (is_a_helper <gimple_statement_omp_target *>::test): ...this.
661 (is_a_helper <gimple_statement_omp_sections>::test): Convert to...
662 (is_a_helper <gimple_statement_omp_sections *>::test): ...this.
663 (is_a_helper <gimple_statement_omp_single>::test): Convert to...
664 (is_a_helper <gimple_statement_omp_single *>::test): ...this.
665 (is_a_helper <gimple_statement_omp_teams>::test): Convert to...
666 (is_a_helper <gimple_statement_omp_teams *>::test): ...this.
667 (is_a_helper <gimple_statement_omp_task>::test): Convert to...
668 (is_a_helper <gimple_statement_omp_task *>::test): ...this.
669 (is_a_helper <gimple_statement_phi>::test): Convert to...
670 (is_a_helper <gimple_statement_phi *>::test): ...this.
671 (is_a_helper <gimple_statement_transaction>::test): Convert to...
672 (is_a_helper <gimple_statement_transaction *>::test): ...this.
673 (is_a_helper <gimple_statement_try>::test): Convert to...
674 (is_a_helper <gimple_statement_try *>::test): ...this.
675 (is_a_helper <gimple_statement_wce>::test): Convert to...
676 (is_a_helper <gimple_statement_wce *>::test): ...this.
677 (is_a_helper <const gimple_statement_asm>::test): Convert to...
678 (is_a_helper <const gimple_statement_asm *>::test): ...this.
679 (is_a_helper <const gimple_statement_bind>::test): Convert to...
680 (is_a_helper <const gimple_statement_bind *>::test): ...this.
681 (is_a_helper <const gimple_statement_call>::test): Convert to...
682 (is_a_helper <const gimple_statement_call *>::test): ...this.
683 (is_a_helper <const gimple_statement_catch>::test): Convert to...
684 (is_a_helper <const gimple_statement_catch *>::test): ...this.
685 (is_a_helper <const gimple_statement_resx>::test): Convert to...
686 (is_a_helper <const gimple_statement_resx *>::test): ...this.
687 (is_a_helper <const gimple_statement_eh_dispatch>::test):
689 (is_a_helper <const gimple_statement_eh_dispatch *>::test):
691 (is_a_helper <const gimple_statement_eh_filter>::test): Convert
693 (is_a_helper <const gimple_statement_eh_filter *>::test): ...this.
694 (is_a_helper <const gimple_statement_omp_atomic_load>::test):
696 (is_a_helper <const gimple_statement_omp_atomic_load *>::test):
698 (is_a_helper <const gimple_statement_omp_atomic_store>::test):
700 (is_a_helper <const gimple_statement_omp_atomic_store *>::test):
702 (is_a_helper <const gimple_statement_omp_return>::test): Convert
704 (is_a_helper <const gimple_statement_omp_return *>::test): ...this.
705 (is_a_helper <const gimple_statement_omp_continue>::test): Convert
707 (is_a_helper <const gimple_statement_omp_continue *>::test): ...this.
708 (is_a_helper <const gimple_statement_omp_critical>::test): Convert
710 (is_a_helper <const gimple_statement_omp_critical *>::test): ...this.
711 (is_a_helper <const gimple_statement_omp_for>::test): Convert to...
712 (is_a_helper <const gimple_statement_omp_for *>::test): ...this.
713 (is_a_helper <const gimple_statement_omp_taskreg>::test): Convert
715 (is_a_helper <const gimple_statement_omp_taskreg *>::test): ...this.
716 (is_a_helper <const gimple_statement_omp_parallel>::test): Convert
718 (is_a_helper <const gimple_statement_omp_parallel *>::test): ...this.
719 (is_a_helper <const gimple_statement_omp_target>::test): Convert
721 (is_a_helper <const gimple_statement_omp_target *>::test): ...this.
722 (is_a_helper <const gimple_statement_omp_sections>::test): Convert
724 (is_a_helper <const gimple_statement_omp_sections *>::test): ...this.
725 (is_a_helper <const gimple_statement_omp_single>::test): Convert
727 (is_a_helper <const gimple_statement_omp_single *>::test): ...this.
728 (is_a_helper <const gimple_statement_omp_teams>::test): Convert
730 (is_a_helper <const gimple_statement_omp_teams *>::test): ...this.
731 (is_a_helper <const gimple_statement_omp_task>::test): Convert
733 (is_a_helper <const gimple_statement_omp_task *>::test): ...this.
734 (is_a_helper <const gimple_statement_phi>::test): Convert to...
735 (is_a_helper <const gimple_statement_phi *>::test): ...this.
736 (is_a_helper <const gimple_statement_transaction>::test): Convert
738 (is_a_helper <const gimple_statement_transaction *>::test): ...this.
739 (is_a_helper <const gimple_statement_with_ops>::test): Convert
741 (is_a_helper <const gimple_statement_with_ops *>::test): ...this.
742 (is_a_helper <gimple_statement_with_ops>::test): Convert to...
743 (is_a_helper <gimple_statement_with_ops *>::test): ...this.
744 (is_a_helper <const gimple_statement_with_memory_ops>::test):
746 (is_a_helper <const gimple_statement_with_memory_ops *>::test):
748 (is_a_helper <gimple_statement_with_memory_ops>::test): Convert
750 (is_a_helper <gimple_statement_with_memory_ops *>::test): ...this.
752 (gimple_use_ops): Update for removal of implicit pointer from the
754 (gimple_set_use_ops): Likewise.
755 (gimple_vuse): Likewise.
756 (gimple_vdef): Likewise.
757 (gimple_vuse_ptr): Likewise.
758 (gimple_vdef_ptr): Likewise.
759 (gimple_set_vuse): Likewise.
760 (gimple_set_vdef): Likewise.
761 (gimple_omp_return_set_lhs): Likewise.
762 (gimple_omp_return_lhs): Likewise.
763 (gimple_omp_return_lhs_ptr): Likewise.
764 (gimple_call_fntype): Likewise.
765 (gimple_call_set_fntype): Likewise.
766 (gimple_call_set_internal_fn): Likewise.
767 (gimple_call_use_set): Likewise.
768 (gimple_call_clobber_set): Likewise.
769 (gimple_bind_vars): Likewise.
770 (gimple_bind_set_vars): Likewise.
771 (gimple_bind_body_ptr): Likewise.
772 (gimple_bind_set_body): Likewise.
773 (gimple_bind_add_stmt): Likewise.
774 (gimple_bind_block): Likewise.
775 (gimple_bind_set_block): Likewise.
776 (gimple_asm_ninputs): Likewise.
777 (gimple_asm_noutputs): Likewise.
778 (gimple_asm_nclobbers): Likewise.
779 (gimple_asm_nlabels): Likewise.
780 (gimple_asm_input_op): Likewise.
781 (gimple_asm_input_op_ptr): Likewise.
782 (gimple_asm_output_op): Likewise.
783 (gimple_asm_output_op_ptr): Likewise.
784 (gimple_asm_set_output_op): Likewise.
785 (gimple_asm_clobber_op): Likewise.
786 (gimple_asm_set_clobber_op): Likewise.
787 (gimple_asm_label_op): Likewise.
788 (gimple_asm_set_label_op): Likewise.
789 (gimple_asm_string): Likewise.
790 (gimple_catch_types): Likewise.
791 (gimple_catch_types_ptr): Likewise.
792 (gimple_catch_handler_ptr): Likewise.
793 (gimple_catch_set_types): Likewise.
794 (gimple_catch_set_handler): Likewise.
795 (gimple_eh_filter_types): Likewise.
796 (gimple_eh_filter_types_ptr): Likewise.
797 (gimple_eh_filter_failure_ptr): Likewise.
798 (gimple_eh_filter_set_types): Likewise.
799 (gimple_eh_filter_set_failure): Likewise.
800 (gimple_eh_must_not_throw_fndecl): Likewise.
801 (gimple_eh_must_not_throw_set_fndecl): Likewise.
802 (gimple_eh_else_n_body_ptr): Likewise.
803 (gimple_eh_else_e_body_ptr): Likewise.
804 (gimple_eh_else_set_n_body): Likewise.
805 (gimple_eh_else_set_e_body): Likewise.
806 (gimple_try_eval_ptr): Likewise.
807 (gimple_try_cleanup_ptr): Likewise.
808 (gimple_try_set_eval): Likewise.
809 (gimple_try_set_cleanup): Likewise.
810 (gimple_wce_cleanup_ptr): Likewise.
811 (gimple_wce_set_cleanup): Likewise.
812 (gimple_phi_capacity): Likewise.
813 (gimple_phi_num_args): Likewise.
814 (gimple_phi_result): Likewise.
815 (gimple_phi_result_ptr): Likewise.
816 (gimple_phi_set_result): Likewise.
817 (gimple_phi_arg): Likewise.
818 (gimple_phi_set_arg): Likewise.
819 (gimple_resx_region): Likewise.
820 (gimple_resx_set_region): Likewise.
821 (gimple_eh_dispatch_region): Likewise.
822 (gimple_eh_dispatch_set_region): Likewise.
823 (gimple_omp_critical_name): Likewise.
824 (gimple_omp_critical_name_ptr): Likewise.
825 (gimple_omp_critical_set_name): Likewise.
826 (gimple_omp_for_clauses): Likewise.
827 (gimple_omp_for_clauses_ptr): Likewise.
828 (gimple_omp_for_set_clauses): Likewise.
829 (gimple_omp_for_collapse): Likewise.
830 (gimple_omp_for_index): Likewise.
831 (gimple_omp_for_index_ptr): Likewise.
832 (gimple_omp_for_set_index): Likewise.
833 (gimple_omp_for_initial): Likewise.
834 (gimple_omp_for_initial_ptr): Likewise.
835 (gimple_omp_for_set_initial): Likewise.
836 (gimple_omp_for_final): Likewise.
837 (gimple_omp_for_final_ptr): Likewise.
838 (gimple_omp_for_set_final): Likewise.
839 (gimple_omp_for_incr): Likewise.
840 (gimple_omp_for_incr_ptr): Likewise.
841 (gimple_omp_for_set_incr): Likewise.
842 (gimple_omp_for_pre_body_ptr): Likewise.
843 (gimple_omp_for_set_pre_body): Likewise.
844 (gimple_omp_parallel_clauses): Likewise.
845 (gimple_omp_parallel_clauses_ptr): Likewise.
846 (gimple_omp_parallel_set_clauses): Likewise.
847 (gimple_omp_parallel_child_fn): Likewise.
848 (gimple_omp_parallel_child_fn_ptr): Likewise.
849 (gimple_omp_parallel_set_child_fn): Likewise.
850 (gimple_omp_parallel_data_arg): Likewise.
851 (gimple_omp_parallel_data_arg_ptr): Likewise.
852 (gimple_omp_parallel_set_data_arg): Likewise.
853 (gimple_omp_task_clauses): Likewise.
854 (gimple_omp_task_clauses_ptr): Likewise.
855 (gimple_omp_task_set_clauses): Likewise.
856 (gimple_omp_task_child_fn): Likewise.
857 (gimple_omp_task_child_fn_ptr): Likewise.
858 (gimple_omp_task_set_child_fn): Likewise.
859 (gimple_omp_task_data_arg): Likewise.
860 (gimple_omp_task_data_arg_ptr): Likewise.
861 (gimple_omp_task_set_data_arg): Likewise.
862 (gimple_omp_taskreg_clauses): Likewise.
863 (gimple_omp_taskreg_clauses_ptr): Likewise.
864 (gimple_omp_taskreg_set_clauses): Likewise.
865 (gimple_omp_taskreg_child_fn): Likewise.
866 (gimple_omp_taskreg_child_fn_ptr): Likewise.
867 (gimple_omp_taskreg_set_child_fn): Likewise.
868 (gimple_omp_taskreg_data_arg): Likewise.
869 (gimple_omp_taskreg_data_arg_ptr): Likewise.
870 (gimple_omp_taskreg_set_data_arg): Likewise.
871 (gimple_omp_task_copy_fn): Likewise.
872 (gimple_omp_task_copy_fn_ptr): Likewise.
873 (gimple_omp_task_set_copy_fn): Likewise.
874 (gimple_omp_task_arg_size): Likewise.
875 (gimple_omp_task_arg_size_ptr): Likewise.
876 (gimple_omp_task_set_arg_size): Likewise.
877 (gimple_omp_task_arg_align): Likewise.
878 (gimple_omp_task_arg_align_ptr): Likewise.
879 (gimple_omp_task_set_arg_align): Likewise.
880 (gimple_omp_single_clauses): Likewise.
881 (gimple_omp_single_clauses_ptr): Likewise.
882 (gimple_omp_single_set_clauses): Likewise.
883 (gimple_omp_target_clauses): Likewise.
884 (gimple_omp_target_clauses_ptr): Likewise.
885 (gimple_omp_target_set_clauses): Likewise.
886 (gimple_omp_target_child_fn): Likewise.
887 (gimple_omp_target_child_fn_ptr): Likewise.
888 (gimple_omp_target_set_child_fn): Likewise.
889 (gimple_omp_target_data_arg): Likewise.
890 (gimple_omp_target_data_arg_ptr): Likewise.
891 (gimple_omp_target_set_data_arg): Likewise.
892 (gimple_omp_teams_clauses): Likewise.
893 (gimple_omp_teams_clauses_ptr): Likewise.
894 (gimple_omp_teams_set_clauses): Likewise.
895 (gimple_omp_sections_clauses): Likewise.
896 (gimple_omp_sections_clauses_ptr): Likewise.
897 (gimple_omp_sections_set_clauses): Likewise.
898 (gimple_omp_sections_control): Likewise.
899 (gimple_omp_sections_control_ptr): Likewise.
900 (gimple_omp_sections_set_control): Likewise.
901 (gimple_omp_for_set_cond): Likewise.
902 (gimple_omp_for_cond): Likewise.
903 (gimple_omp_atomic_store_set_val): Likewise.
904 (gimple_omp_atomic_store_val): Likewise.
905 (gimple_omp_atomic_store_val_ptr): Likewise.
906 (gimple_omp_atomic_load_set_lhs): Likewise.
907 (gimple_omp_atomic_load_lhs): Likewise.
908 (gimple_omp_atomic_load_lhs_ptr): Likewise.
909 (gimple_omp_atomic_load_set_rhs): Likewise.
910 (gimple_omp_atomic_load_rhs): Likewise.
911 (gimple_omp_atomic_load_rhs_ptr): Likewise.
912 (gimple_omp_continue_control_def): Likewise.
913 (gimple_omp_continue_control_def_ptr): Likewise.
914 (gimple_omp_continue_set_control_def): Likewise.
915 (gimple_omp_continue_control_use): Likewise.
916 (gimple_omp_continue_control_use_ptr): Likewise.
917 (gimple_omp_continue_set_control_use): Likewise.
918 (gimple_transaction_body_ptr): Likewise.
919 (gimple_transaction_label): Likewise.
920 (gimple_transaction_label_ptr): Likewise.
921 (gimple_transaction_set_body): Likewise.
922 (gimple_transaction_set_label): Likewise.
924 * ipa-devirt.c (build_type_inheritance_graph): Likewise.
925 * ipa-inline-analysis.c (inline_write_summary): Likewise.
926 * ipa-ref.c (ipa_record_reference): Likewise.
927 * ipa-reference.c (analyze_function): Likewise.
928 (ipa_reference_write_optimization_summary): Likewise.
929 * ipa.c (symtab_remove_unreachable_nodes): Likewise.
930 (address_taken_from_non_vtable_p): Likewise.
931 (comdat_can_be_unshared_p_1): Likewise.
932 * lto-cgraph.c (lto_output_ref): Likewise.
933 (add_references): Likewise.
934 (compute_ltrans_boundary): Likewise.
935 (output_symtab): Likewise.
936 (input_ref): Likewise.
937 (input_cgraph_1): Likewise.
938 (output_cgraph_opt_summary): Likewise.
939 * lto-streamer-out.c (lto_output): Likewise.
940 (output_symbol_p): Likewise.
941 * lto-streamer.h (lsei_next_function_in_partition): Likewise.
942 (lsei_start_function_in_partition): Likewise.
943 (lsei_next_variable_in_partition): Likewise.
944 (lsei_start_variable_in_partition): Likewise.
945 * symtab.c (insert_to_assembler_name_hash): Likewise.
946 (unlink_from_assembler_name_hash): Likewise.
947 (symtab_unregister_node): Likewise.
948 (symtab_remove_node): Likewise.
949 (dump_symtab_node): Likewise.
950 (verify_symtab_base): Likewise.
951 (verify_symtab_node): Likewise.
952 (symtab_make_decl_local): Likewise.
953 (symtab_alias_ultimate_target): Likewise.
954 (symtab_resolve_alias): Likewise.
955 (symtab_get_symbol_partitioning_class): Likewise.
956 * tree-phinodes.c (allocate_phi_node): Likewise.
957 (reserve_phi_args_for_new_edge): Likewise.
958 (remove_phi_args): Likewise.
959 * varpool.c (varpool_node_for_asm): Likewise.
960 (varpool_remove_unreferenced_decls): Likewise.
962 2014-04-23 Jeff Law <law@redhat.com>
964 PR tree-optimization/60902
965 * tree-ssa-threadedge.c
966 (record_temporary_equivalences_from_stmts_at_dest): Make sure to
967 invalidate outputs from statements that do not produce useful
968 outputs for threading.
970 2014-04-23 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
972 * config/aarch64/aarch64.md (stack_protect_set, stack_protect_test)
973 (stack_protect_set_<mode>, stack_protect_test_<mode>): Add
974 machine descriptions for Stack Smashing Protector.
976 2014-04-23 Richard Earnshaw <rearnsha@arm.com>
978 * aarch64.md (<optab>_rol<mode>3): New pattern.
979 (<optab>_rolsi3_uxtw): Likewise.
980 * aarch64.c (aarch64_strip_shift): Handle ROTATE and ROTATERT.
982 2014-04-23 James Greenhalgh <james.greenhalgh@arm.com>
984 * config/arm/arm.c (arm_cortex_a57_tune): Initialize all fields.
985 (arm_cortex_a12_tune): Likewise.
987 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
989 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle BSWAP.
991 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
993 * config/arm/arm.md (arm_rev16si2): New pattern.
994 (arm_rev16si2_alt): Likewise.
995 * config/arm/arm.c (arm_new_rtx_costs): Handle rev16 case.
997 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
999 * config/aarch64/aarch64.md (rev16<mode>2): New pattern.
1000 (rev16<mode>2_alt): Likewise.
1001 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle rev16 case.
1002 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p): New.
1003 (aarch_rev16_shleft_mask_imm_p): Likewise.
1004 (aarch_rev16_p_1): Likewise.
1005 (aarch_rev16_p): Likewise.
1006 * config/arm/aarch-common-protos.h (aarch_rev16_p): Declare extern.
1007 (aarch_rev16_shright_mask_imm_p): Likewise.
1008 (aarch_rev16_shleft_mask_imm_p): Likewise.
1010 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1012 * config/arm/aarch-common-protos.h (alu_cost_table): Add rev field.
1013 * config/arm/aarch-cost-tables.h (generic_extra_costs): Specify
1015 (cortex_a53_extra_costs): Likewise.
1016 (cortex_a57_extra_costs): Likewise.
1017 * config/arm/arm.c (cortexa9_extra_costs): Likewise.
1018 (cortexa7_extra_costs): Likewise.
1019 (cortexa8_extra_costs): Likewise.
1020 (cortexa12_extra_costs): Likewise.
1021 (cortexa15_extra_costs): Likewise.
1022 (v7m_extra_costs): Likewise.
1023 (arm_new_rtx_costs): Handle BSWAP.
1025 2013-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1027 * config/arm/arm.c (cortexa8_extra_costs): New table.
1028 (arm_cortex_a8_tune): New tuning struct.
1029 * config/arm/arm-cores.def (cortex-a8): Use cortex_a8 tuning struct.
1031 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1033 * config/arm/arm.c (arm_new_rtx_costs): Handle FMA.
1035 2014-04-23 Richard Biener <rguenther@suse.de>
1037 * Makefile.in (OBJS): Remove loop-unswitch.o.
1038 * loop-unswitch.c: Delete.
1039 * tree-pass.h (make_pass_rtl_unswitch): Remove.
1040 * passes.def (pass_rtl_unswitch): Likewise.
1041 * loop-init.c (gate_rtl_unswitch): Likewise.
1042 (rtl_unswitch): Likewise.
1043 (pass_data_rtl_unswitch): Likewise.
1044 (pass_rtl_unswitch): Likewise.
1045 (make_pass_rtl_unswitch): Likewise.
1046 * rtl.h (reversed_condition): Likewise.
1047 (compare_and_jump_seq): Likewise.
1048 * loop-iv.c (reversed_condition): Move here from loop-unswitch.c
1050 * loop-unroll.c (compare_and_jump_seq): Likewise.
1052 2014-04-23 Richard Biener <rguenther@suse.de>
1054 PR tree-optimization/60903
1055 * tree-ssa-loop-im.c (analyze_memory_references): Remove
1056 commented code block.
1057 (execute_sm_if_changed): Properly apply IRREDUCIBLE_LOOP
1058 loop flags to newly created BBs and edges.
1060 2014-04-23 Nick Clifton <nickc@redhat.com>
1062 * config/msp430/msp430.c (msp430_handle_option): Move function
1064 (msp430_option_override): Simplify mcu and mcpu option handling.
1065 (msp430_is_f5_mcu): Rename to msp430_use_f5_series_hwmult. Add
1066 support for -mhwmult command line option.
1067 (has_32bit_hwmult): Rename to use_32bit_hwmult. Add support for
1068 -mhwmult command line option.
1069 (msp430_hwmult_enabled): Delete.
1070 (msp43o_output_labelref): Add support for -mhwmult command line
1072 * config/msp430/msp430.md (mulhisi3, umulhisi3, mulsidi3)
1073 (umulsidi3): Likewise.
1074 * config/msp430/msp430.opt (mmcu): Add Report attribute.
1075 (mcpu, mlarge, msmall): Likewise.
1076 (mhwmult): New option.
1077 * config/msp430/msp430-protos.h (msp430_hwmult_enabled): Remove
1079 (msp430_is_f5_mcu): Remove prototype.
1080 (msp430_use_f5_series_hwmult): Add prototype.
1081 * config/msp430/msp430-opts.h: New file.
1082 * common/config/msp430: New directory.
1083 * common/config/msp430/msp430-common.c: New file.
1084 * config.gcc (msp430): Remove target_has_targetm_common.
1085 * doc/invoke.texi: Document -mhwmult command line option.
1087 2014-04-23 Nick Clifton <nickc@redhat.com>
1089 * config/i386/cygwin.h (ENDFILE_SPEC): Include
1090 default-manifest.o if it can be found in the search path.
1091 * config/i386/mingw32.h (ENDFILE_SPEC): Likewise.
1093 2014-04-23 Terry Guo <terry.guo@arm.com>
1095 * config/arm/arm.h (ASM_APP_OFF): Re-define it in a cleaner way.
1097 2014-04-23 Richard Biener <rguenther@suse.de>
1100 * tree-inline.c (declare_return_variable): Use mark_addressable.
1102 2014-04-23 Richard Biener <rguenther@suse.de>
1105 * loop-init.c (loop_optimizer_init): Make sure to apply
1106 LOOPS_MAY_HAVE_MULTIPLE_LATCHES before fixing up loops.
1108 2014-04-22 Jakub Jelinek <jakub@redhat.com>
1111 * common.opt (fsanitize-recover, fsanitize-undefined-trap-on-error):
1113 * gcc.c (sanitize_spec_function): Don't return "" for "undefined"
1114 if flag_sanitize_undefined_trap_on_error.
1115 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW_ABORT,
1116 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS_ABORT,
1117 BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE_ABORT,
1118 BUILT_IN_UBSAN_HANDLE_TYPE_MISMATCH_ABORT,
1119 BUILT_IN_UBSAN_HANDLE_ADD_OVERFLOW_ABORT,
1120 BUILT_IN_UBSAN_HANDLE_SUB_OVERFLOW_ABORT,
1121 BUILT_IN_UBSAN_HANDLE_MUL_OVERFLOW_ABORT,
1122 BUILT_IN_UBSAN_HANDLE_NEGATE_OVERFLOW_ABORT,
1123 BUILT_IN_UBSAN_HANDLE_LOAD_INVALID_VALUE_ABORT): New builtins.
1124 * ubsan.c (ubsan_instrument_unreachable): Return
1125 __builtin_trap () if flag_sanitize_undefined_trap_on_error.
1126 (ubsan_expand_null_ifn): Emit __builtin_trap ()
1127 if flag_sanitize_undefined_trap_on_error and
1128 __ubsan_handle_type_mismatch_abort if !flag_sanitize_recover.
1129 (ubsan_expand_null_ifn, ubsan_build_overflow_builtin,
1130 instrument_bool_enum_load): Emit __builtin_trap () if
1131 flag_sanitize_undefined_trap_on_error and
1132 __builtin_handle_*_abort () if !flag_sanitize_recover.
1133 * doc/invoke.texi (-fsanitize-recover,
1134 -fsanitize-undefined-trap-on-error): Document.
1136 2014-04-22 Christian Bruel <christian.bruel@st.com>
1138 * config/sh/sh.md (mov<mode>): Replace movQIHI.
1139 Force immediates to SImode.
1141 2014-04-22 Sandra Loosemore <sandra@codesourcery.com>
1143 * config/nios2/nios2.md (UNSPEC_ROUND): New.
1145 * config/nios2/nios2.opt (mno-custom-round, mcustom-round=): New.
1146 * config/nios2/nios2-opts.h (N2FPU_ALL_CODES): Add round.
1147 * config/nios2/nios2.c (N2F_NO_ERRNO): Define.
1148 (nios2_fpu_insn): Add entry for round.
1149 (N2FPU_NO_ERRNO_P): Define.
1150 (nios2_custom_check_insns): Add check for N2F_NO_ERRNO and
1152 * doc/invoke.texi (Nios II Options): Document -mcustom-round.
1154 2014-04-22 Richard Henderson <rth@redhat.com>
1156 * config/aarch64/aarch64 (addti3, subti3): New expanders.
1157 (add<GPI>3_compare0): Remove leading * from name.
1158 (add<GPI>3_carryin): Likewise.
1159 (sub<GPI>3_compare0): Likewise.
1160 (sub<GPI>3_carryin): Likewise.
1161 (<su_optab>mulditi3): New expander.
1162 (multi3): New expander.
1163 (madd<GPI>): Remove leading * from name.
1165 2014-04-22 Martin Jambor <mjambor@suse.cz>
1167 * cgraphclones.c (cgraph_function_versioning): Copy
1168 ipa_transforms_to_apply instead of asserting it is empty.
1170 2014-04-22 H.J. Lu <hongjiu.lu@intel.com>
1173 * config/i386/i386.c (ix86_expand_set_or_movmem): Call counter_mode
1174 on count_exp to get mode.
1176 2014-04-22 Andrew Pinski <apinski@cavium.com>
1178 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
1179 Handle TLS for ILP32.
1180 * config/aarch64/aarch64.md (tlsie_small): Rename to ...
1181 (tlsie_small_<mode>): this and handle PTR.
1182 (tlsie_small_sidi): New pattern.
1183 (tlsle_small): Change to an expand to handle ILP32.
1184 (tlsle_small_<mode>): New pattern.
1185 (tlsdesc_small): Rename to ...
1186 (tlsdesc_small_<mode>): this and handle PTR.
1188 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1190 * config/aarch64/aarch64.c (TARGET_FLAGS_REGNUM): Define.
1192 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1194 * config/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
1195 (aarch64_types_signed_unsigned_qualifiers): Qualifier added.
1196 (aarch64_types_signed_poly_qualifiers): Likewise.
1197 (aarch64_types_unsigned_signed_qualifiers): Likewise.
1198 (aarch64_types_poly_signed_qualifiers): Likewise.
1199 (TYPES_REINTERP_SS): Type macro added.
1200 (TYPES_REINTERP_SU): Likewise.
1201 (TYPES_REINTERP_SP): Likewise.
1202 (TYPES_REINTERP_US): Likewise.
1203 (TYPES_REINTERP_PS): Likewise.
1204 (aarch64_fold_builtin): New expression folding added.
1205 * config/aarch64/aarch64-simd-builtins.def (REINTERP):
1206 Declarations removed.
1207 (REINTERP_SS): Declarations added.
1208 (REINTERP_US): Likewise.
1209 (REINTERP_PS): Likewise.
1210 (REINTERP_SU): Likewise.
1211 (REINTERP_SP): Likewise.
1212 * config/aarch64/arm_neon.h (vreinterpret_p8_f64): Implemented.
1213 (vreinterpretq_p8_f64): Likewise.
1214 (vreinterpret_p16_f64): Likewise.
1215 (vreinterpretq_p16_f64): Likewise.
1216 (vreinterpret_f32_f64): Likewise.
1217 (vreinterpretq_f32_f64): Likewise.
1218 (vreinterpret_f64_f32): Likewise.
1219 (vreinterpret_f64_p8): Likewise.
1220 (vreinterpret_f64_p16): Likewise.
1221 (vreinterpret_f64_s8): Likewise.
1222 (vreinterpret_f64_s16): Likewise.
1223 (vreinterpret_f64_s32): Likewise.
1224 (vreinterpret_f64_s64): Likewise.
1225 (vreinterpret_f64_u8): Likewise.
1226 (vreinterpret_f64_u16): Likewise.
1227 (vreinterpret_f64_u32): Likewise.
1228 (vreinterpret_f64_u64): Likewise.
1229 (vreinterpretq_f64_f32): Likewise.
1230 (vreinterpretq_f64_p8): Likewise.
1231 (vreinterpretq_f64_p16): Likewise.
1232 (vreinterpretq_f64_s8): Likewise.
1233 (vreinterpretq_f64_s16): Likewise.
1234 (vreinterpretq_f64_s32): Likewise.
1235 (vreinterpretq_f64_s64): Likewise.
1236 (vreinterpretq_f64_u8): Likewise.
1237 (vreinterpretq_f64_u16): Likewise.
1238 (vreinterpretq_f64_u32): Likewise.
1239 (vreinterpretq_f64_u64): Likewise.
1240 (vreinterpret_s64_f64): Likewise.
1241 (vreinterpretq_s64_f64): Likewise.
1242 (vreinterpret_u64_f64): Likewise.
1243 (vreinterpretq_u64_f64): Likewise.
1244 (vreinterpret_s8_f64): Likewise.
1245 (vreinterpretq_s8_f64): Likewise.
1246 (vreinterpret_s16_f64): Likewise.
1247 (vreinterpretq_s16_f64): Likewise.
1248 (vreinterpret_s32_f64): Likewise.
1249 (vreinterpretq_s32_f64): Likewise.
1250 (vreinterpret_u8_f64): Likewise.
1251 (vreinterpretq_u8_f64): Likewise.
1252 (vreinterpret_u16_f64): Likewise.
1253 (vreinterpretq_u16_f64): Likewise.
1254 (vreinterpret_u32_f64): Likewise.
1255 (vreinterpretq_u32_f64): Likewise.
1257 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1259 * config/aarch64/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
1260 * config/aarch64/aarch64/aarch64-simd-builtins.def (REINTERP): Removed.
1261 (vreinterpret_p8_s8): Likewise.
1262 * config/aarch64/aarch64/arm_neon.h (vreinterpret_p8_s8): Uses cast.
1263 (vreinterpret_p8_s16): Likewise.
1264 (vreinterpret_p8_s32): Likewise.
1265 (vreinterpret_p8_s64): Likewise.
1266 (vreinterpret_p8_f32): Likewise.
1267 (vreinterpret_p8_u8): Likewise.
1268 (vreinterpret_p8_u16): Likewise.
1269 (vreinterpret_p8_u32): Likewise.
1270 (vreinterpret_p8_u64): Likewise.
1271 (vreinterpret_p8_p16): Likewise.
1272 (vreinterpretq_p8_s8): Likewise.
1273 (vreinterpretq_p8_s16): Likewise.
1274 (vreinterpretq_p8_s32): Likewise.
1275 (vreinterpretq_p8_s64): Likewise.
1276 (vreinterpretq_p8_f32): Likewise.
1277 (vreinterpretq_p8_u8): Likewise.
1278 (vreinterpretq_p8_u16): Likewise.
1279 (vreinterpretq_p8_u32): Likewise.
1280 (vreinterpretq_p8_u64): Likewise.
1281 (vreinterpretq_p8_p16): Likewise.
1282 (vreinterpret_p16_s8): Likewise.
1283 (vreinterpret_p16_s16): Likewise.
1284 (vreinterpret_p16_s32): Likewise.
1285 (vreinterpret_p16_s64): Likewise.
1286 (vreinterpret_p16_f32): Likewise.
1287 (vreinterpret_p16_u8): Likewise.
1288 (vreinterpret_p16_u16): Likewise.
1289 (vreinterpret_p16_u32): Likewise.
1290 (vreinterpret_p16_u64): Likewise.
1291 (vreinterpret_p16_p8): Likewise.
1292 (vreinterpretq_p16_s8): Likewise.
1293 (vreinterpretq_p16_s16): Likewise.
1294 (vreinterpretq_p16_s32): Likewise.
1295 (vreinterpretq_p16_s64): Likewise.
1296 (vreinterpretq_p16_f32): Likewise.
1297 (vreinterpretq_p16_u8): Likewise.
1298 (vreinterpretq_p16_u16): Likewise.
1299 (vreinterpretq_p16_u32): Likewise.
1300 (vreinterpretq_p16_u64): Likewise.
1301 (vreinterpretq_p16_p8): Likewise.
1302 (vreinterpret_f32_s8): Likewise.
1303 (vreinterpret_f32_s16): Likewise.
1304 (vreinterpret_f32_s32): Likewise.
1305 (vreinterpret_f32_s64): Likewise.
1306 (vreinterpret_f32_u8): Likewise.
1307 (vreinterpret_f32_u16): Likewise.
1308 (vreinterpret_f32_u32): Likewise.
1309 (vreinterpret_f32_u64): Likewise.
1310 (vreinterpret_f32_p8): Likewise.
1311 (vreinterpret_f32_p16): Likewise.
1312 (vreinterpretq_f32_s8): Likewise.
1313 (vreinterpretq_f32_s16): Likewise.
1314 (vreinterpretq_f32_s32): Likewise.
1315 (vreinterpretq_f32_s64): Likewise.
1316 (vreinterpretq_f32_u8): Likewise.
1317 (vreinterpretq_f32_u16): Likewise.
1318 (vreinterpretq_f32_u32): Likewise.
1319 (vreinterpretq_f32_u64): Likewise.
1320 (vreinterpretq_f32_p8): Likewise.
1321 (vreinterpretq_f32_p16): Likewise.
1322 (vreinterpret_s64_s8): Likewise.
1323 (vreinterpret_s64_s16): Likewise.
1324 (vreinterpret_s64_s32): Likewise.
1325 (vreinterpret_s64_f32): Likewise.
1326 (vreinterpret_s64_u8): Likewise.
1327 (vreinterpret_s64_u16): Likewise.
1328 (vreinterpret_s64_u32): Likewise.
1329 (vreinterpret_s64_u64): Likewise.
1330 (vreinterpret_s64_p8): Likewise.
1331 (vreinterpret_s64_p16): Likewise.
1332 (vreinterpretq_s64_s8): Likewise.
1333 (vreinterpretq_s64_s16): Likewise.
1334 (vreinterpretq_s64_s32): Likewise.
1335 (vreinterpretq_s64_f32): Likewise.
1336 (vreinterpretq_s64_u8): Likewise.
1337 (vreinterpretq_s64_u16): Likewise.
1338 (vreinterpretq_s64_u32): Likewise.
1339 (vreinterpretq_s64_u64): Likewise.
1340 (vreinterpretq_s64_p8): Likewise.
1341 (vreinterpretq_s64_p16): Likewise.
1342 (vreinterpret_u64_s8): Likewise.
1343 (vreinterpret_u64_s16): Likewise.
1344 (vreinterpret_u64_s32): Likewise.
1345 (vreinterpret_u64_s64): Likewise.
1346 (vreinterpret_u64_f32): Likewise.
1347 (vreinterpret_u64_u8): Likewise.
1348 (vreinterpret_u64_u16): Likewise.
1349 (vreinterpret_u64_u32): Likewise.
1350 (vreinterpret_u64_p8): Likewise.
1351 (vreinterpret_u64_p16): Likewise.
1352 (vreinterpretq_u64_s8): Likewise.
1353 (vreinterpretq_u64_s16): Likewise.
1354 (vreinterpretq_u64_s32): Likewise.
1355 (vreinterpretq_u64_s64): Likewise.
1356 (vreinterpretq_u64_f32): Likewise.
1357 (vreinterpretq_u64_u8): Likewise.
1358 (vreinterpretq_u64_u16): Likewise.
1359 (vreinterpretq_u64_u32): Likewise.
1360 (vreinterpretq_u64_p8): Likewise.
1361 (vreinterpretq_u64_p16): Likewise.
1362 (vreinterpret_s8_s16): Likewise.
1363 (vreinterpret_s8_s32): Likewise.
1364 (vreinterpret_s8_s64): Likewise.
1365 (vreinterpret_s8_f32): Likewise.
1366 (vreinterpret_s8_u8): Likewise.
1367 (vreinterpret_s8_u16): Likewise.
1368 (vreinterpret_s8_u32): Likewise.
1369 (vreinterpret_s8_u64): Likewise.
1370 (vreinterpret_s8_p8): Likewise.
1371 (vreinterpret_s8_p16): Likewise.
1372 (vreinterpretq_s8_s16): Likewise.
1373 (vreinterpretq_s8_s32): Likewise.
1374 (vreinterpretq_s8_s64): Likewise.
1375 (vreinterpretq_s8_f32): Likewise.
1376 (vreinterpretq_s8_u8): Likewise.
1377 (vreinterpretq_s8_u16): Likewise.
1378 (vreinterpretq_s8_u32): Likewise.
1379 (vreinterpretq_s8_u64): Likewise.
1380 (vreinterpretq_s8_p8): Likewise.
1381 (vreinterpretq_s8_p16): Likewise.
1382 (vreinterpret_s16_s8): Likewise.
1383 (vreinterpret_s16_s32): Likewise.
1384 (vreinterpret_s16_s64): Likewise.
1385 (vreinterpret_s16_f32): Likewise.
1386 (vreinterpret_s16_u8): Likewise.
1387 (vreinterpret_s16_u16): Likewise.
1388 (vreinterpret_s16_u32): Likewise.
1389 (vreinterpret_s16_u64): Likewise.
1390 (vreinterpret_s16_p8): Likewise.
1391 (vreinterpret_s16_p16): Likewise.
1392 (vreinterpretq_s16_s8): Likewise.
1393 (vreinterpretq_s16_s32): Likewise.
1394 (vreinterpretq_s16_s64): Likewise.
1395 (vreinterpretq_s16_f32): Likewise.
1396 (vreinterpretq_s16_u8): Likewise.
1397 (vreinterpretq_s16_u16): Likewise.
1398 (vreinterpretq_s16_u32): Likewise.
1399 (vreinterpretq_s16_u64): Likewise.
1400 (vreinterpretq_s16_p8): Likewise.
1401 (vreinterpretq_s16_p16): Likewise.
1402 (vreinterpret_s32_s8): Likewise.
1403 (vreinterpret_s32_s16): Likewise.
1404 (vreinterpret_s32_s64): Likewise.
1405 (vreinterpret_s32_f32): Likewise.
1406 (vreinterpret_s32_u8): Likewise.
1407 (vreinterpret_s32_u16): Likewise.
1408 (vreinterpret_s32_u32): Likewise.
1409 (vreinterpret_s32_u64): Likewise.
1410 (vreinterpret_s32_p8): Likewise.
1411 (vreinterpret_s32_p16): Likewise.
1412 (vreinterpretq_s32_s8): Likewise.
1413 (vreinterpretq_s32_s16): Likewise.
1414 (vreinterpretq_s32_s64): Likewise.
1415 (vreinterpretq_s32_f32): Likewise.
1416 (vreinterpretq_s32_u8): Likewise.
1417 (vreinterpretq_s32_u16): Likewise.
1418 (vreinterpretq_s32_u32): Likewise.
1419 (vreinterpretq_s32_u64): Likewise.
1420 (vreinterpretq_s32_p8): Likewise.
1421 (vreinterpretq_s32_p16): Likewise.
1422 (vreinterpret_u8_s8): Likewise.
1423 (vreinterpret_u8_s16): Likewise.
1424 (vreinterpret_u8_s32): Likewise.
1425 (vreinterpret_u8_s64): Likewise.
1426 (vreinterpret_u8_f32): Likewise.
1427 (vreinterpret_u8_u16): Likewise.
1428 (vreinterpret_u8_u32): Likewise.
1429 (vreinterpret_u8_u64): Likewise.
1430 (vreinterpret_u8_p8): Likewise.
1431 (vreinterpret_u8_p16): Likewise.
1432 (vreinterpretq_u8_s8): Likewise.
1433 (vreinterpretq_u8_s16): Likewise.
1434 (vreinterpretq_u8_s32): Likewise.
1435 (vreinterpretq_u8_s64): Likewise.
1436 (vreinterpretq_u8_f32): Likewise.
1437 (vreinterpretq_u8_u16): Likewise.
1438 (vreinterpretq_u8_u32): Likewise.
1439 (vreinterpretq_u8_u64): Likewise.
1440 (vreinterpretq_u8_p8): Likewise.
1441 (vreinterpretq_u8_p16): Likewise.
1442 (vreinterpret_u16_s8): Likewise.
1443 (vreinterpret_u16_s16): Likewise.
1444 (vreinterpret_u16_s32): Likewise.
1445 (vreinterpret_u16_s64): Likewise.
1446 (vreinterpret_u16_f32): Likewise.
1447 (vreinterpret_u16_u8): Likewise.
1448 (vreinterpret_u16_u32): Likewise.
1449 (vreinterpret_u16_u64): Likewise.
1450 (vreinterpret_u16_p8): Likewise.
1451 (vreinterpret_u16_p16): Likewise.
1452 (vreinterpretq_u16_s8): Likewise.
1453 (vreinterpretq_u16_s16): Likewise.
1454 (vreinterpretq_u16_s32): Likewise.
1455 (vreinterpretq_u16_s64): Likewise.
1456 (vreinterpretq_u16_f32): Likewise.
1457 (vreinterpretq_u16_u8): Likewise.
1458 (vreinterpretq_u16_u32): Likewise.
1459 (vreinterpretq_u16_u64): Likewise.
1460 (vreinterpretq_u16_p8): Likewise.
1461 (vreinterpretq_u16_p16): Likewise.
1462 (vreinterpret_u32_s8): Likewise.
1463 (vreinterpret_u32_s16): Likewise.
1464 (vreinterpret_u32_s32): Likewise.
1465 (vreinterpret_u32_s64): Likewise.
1466 (vreinterpret_u32_f32): Likewise.
1467 (vreinterpret_u32_u8): Likewise.
1468 (vreinterpret_u32_u16): Likewise.
1469 (vreinterpret_u32_u64): Likewise.
1470 (vreinterpret_u32_p8): Likewise.
1471 (vreinterpret_u32_p16): Likewise.
1472 (vreinterpretq_u32_s8): Likewise.
1473 (vreinterpretq_u32_s16): Likewise.
1474 (vreinterpretq_u32_s32): Likewise.
1475 (vreinterpretq_u32_s64): Likewise.
1476 (vreinterpretq_u32_f32): Likewise.
1477 (vreinterpretq_u32_u8): Likewise.
1478 (vreinterpretq_u32_u16): Likewise.
1479 (vreinterpretq_u32_u64): Likewise.
1480 (vreinterpretq_u32_p8): Likewise.
1481 (vreinterpretq_u32_p16): Likewise.
1483 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1485 * gcc/config/aarch64/aarch64-simd.md (aarch64_s<optab><mode>):
1487 * config/aarch64/aarch64-simd-builtins.def (sqneg): Iterator
1490 * config/aarch64/arm_neon.h (vqneg_s64): New intrinsic.
1491 (vqnegd_s64): Likewise.
1492 (vqabs_s64): Likewise.
1493 (vqabsd_s64): Likewise.
1495 2014-04-22 Richard Henderson <rth@redhat.com>
1497 * config/sparc/sparc.c (sparc_init_modes): Hoist GET_MODE_SIZE
1498 computation to the top of the loop.
1500 2014-04-22 Renlin <renlin.li@arm.com>
1501 Jiong Wang <jiong.wang@arm.com>
1503 * config/aarch64/aarch64.h (aarch64_frame): Delete "fp_lr_offset".
1504 * config/aarch64/aarch64.c (aarch64_layout_frame)
1505 (aarch64_initial_elimination_offset): Likewise.
1507 2014-04-22 Marcus Shawcroft <marcus.shawcroft@arm.com>
1509 * config/aarch64/aarch64.c (aarch64_initial_elimination_offset):
1512 2014-04-22 Richard Sandiford <rdsandiford@googlemail.com>
1514 * machmode.h (bitwise_mode_for_mode): Declare.
1515 * stor-layout.h (bitwise_type_for_mode): Likewise.
1516 * stor-layout.c (bitwise_mode_for_mode): New function.
1517 (bitwise_type_for_mode): Likewise.
1518 * builtins.c (fold_builtin_memory_op): Use it instead of
1519 int_mode_for_mode and build_nonstandard_integer_type.
1521 2014-04-22 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1523 * config.gcc (enable_obsolete): Remove *-*-solaris2.9*.
1524 (*-*-solaris2.[0-9] | *-*-solaris2.[0-9].*): Mark unsupported.
1525 (*-*-solaris2*): Simplify.
1526 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Likewise.
1527 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Remove
1528 *-*-solaris2.9* handling.
1530 * configure.ac (gcc_cv_as_hidden): Remove test for Solaris 9/x86
1532 (gcc_cv_ld_hidden): Remove *-*-solaris2.9* handling.
1533 (ld_tls_support): Remove i?86-*-solaris2.9, sparc*-*-solaris2.9
1535 (gcc_cv_as_gstabs_flag): Remove workaround for Solaris 9/x86 as bug.
1536 * configure: Regenerate.
1538 * config/i386/sol2-9.h: Remove.
1540 * doc/install.texi (Specific, i?86-*-solaris2.9): Remove.
1541 (Specific, *-*-solaris2*): Mention Solaris 9 support removal.
1542 Remove Solaris 9 references.
1544 2014-04-22 Vidya Praveen <vidyapraveen@arm.com>
1546 * aarch64.md (float<GPI:mode><GPF:mode>2): Remove.
1547 (floatuns<GPI:mode><GPF:mode>2): Remove.
1548 (<optab><fcvt_target><GPF:mode>2): New pattern for equal width float
1549 and floatuns conversions.
1550 (<optab><fcvt_iesize><GPF:mode>2): New pattern for inequal width float
1551 and floatuns conversions.
1552 * iterators.md (fcvt_target, FCVT_TARGET): Support SF and DF modes.
1553 (w1,w2): New mode attributes for inequal width conversions.
1555 2014-04-22 Renlin Li <Renlin.Li@arm.com>
1557 * config/aarch64/aarch64.c (aarch64_print_operand_address): Adjust
1558 the output asm format.
1560 2014-04-22 James Greenhalgh <james.greenhalgh@arm.com>
1562 * config/aarch64/aarch64-simd.md
1563 (aarch64_cm<optab>di): Always split.
1564 (*aarch64_cm<optab>di): New.
1565 (aarch64_cmtstdi): Always split.
1566 (*aarch64_cmtstdi): New.
1568 2014-04-22 Jakub Jelinek <jakub@redhat.com>
1570 PR tree-optimization/60823
1571 * omp-low.c (ipa_simd_modify_function_body): Go through
1572 all SSA_NAMEs and for those refering to vector arguments
1573 which are going to be replaced adjust SSA_NAME_VAR and,
1574 if it is a default definition, change it into a non-default
1575 definition assigned at the beginning of function from new_decl.
1576 (ipa_simd_modify_stmt_ops): Rewritten.
1577 * tree-dfa.c (set_ssa_default_def): When removing default def,
1578 check for NULL loc instead of NULL *loc.
1580 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1582 * config/arm/arm.c (arm_hard_regno_mode_ok): Loosen
1583 restrictions on core registers for DImode values in Thumb2.
1585 2014-04-22 Ian Bolton <ian.bolton@arm.com>
1587 * config/arm/arm.md (*anddi_notdi_zesidi): New pattern.
1588 * config/arm/thumb2.md (*iordi_notdi_zesidi): New pattern.
1590 2014-04-22 Ian Bolton <ian.bolton@arm.com>
1592 * config/arm/thumb2.md (*iordi_notdi_di): New pattern.
1593 (*iordi_notzesidi_di): Likewise.
1594 (*iordi_notsesidi_di): Likewise.
1596 2014-04-22 Ian Bolton <ian.bolton@arm.com>
1598 * config/arm/arm-protos.h (tune_params): New struct members.
1599 * config/arm/arm.c: Initialise tune_params per processor.
1600 (thumb2_reorg): Suppress conversion from t32 to t16 when optimizing
1601 for speed, based on new tune_params.
1603 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1605 * config/aarch64/aarch64-builtins.c (BUILTIN_VDQF_DF): Macro
1607 * config/aarch64/aarch64-simd-builtins.def (frintn): Use added
1609 * config/aarch64/aarch64-simd.md (<frint_pattern>): Comment
1611 * config/aarch64/aarch64.md (<frint_pattern>): Likewise.
1612 * config/aarch64/arm_neon.h (vrnd_f64): Added.
1613 (vrnda_f64): Likewise.
1614 (vrndi_f64): Likewise.
1615 (vrndm_f64): Likewise.
1616 (vrndn_f64): Likewise.
1617 (vrndp_f64): Likewise.
1618 (vrndx_f64): Likewise.
1620 2014-04-22 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1622 * config/arm/arm.c (arm_print_operand, thumb_exit): Make sure
1623 GET_MODE_SIZE argument is enum machine_mode.
1625 2014-04-22 Jakub Jelinek <jakub@redhat.com>
1628 * config/sparc/sparc.c (sparc_init_modes): Pass enum machine_mode
1629 value instead of int to GET_MODE_CLASS and GET_MODE_SIZE macros.
1631 2014-04-22 Lin Zuojian <manjian2006@gmail.com>
1634 * asan.c (asan_emit_stack_protection): Force the base to align to
1635 appropriate bits if STRICT_ALIGNMENT. Set shadow_mem align to
1636 appropriate bits if STRICT_ALIGNMENT.
1637 * cfgexpand.c (expand_stack_vars): Set base_align appropriately
1639 (expand_used_vars): Leave a space in the stack frame for alignment
1640 if STRICT_ALIGNMENT.
1642 2014-04-21 David Malcolm <dmalcolm@redhat.com>
1644 * gimple.h (gimple_assign_single_p): Accept a const_gimple rather
1646 (gimple_store_p): Likewise.
1647 (gimple_assign_load_p): Likewise.
1648 (gimple_assign_cast_p): Likewise.
1649 (gimple_clobber_p): Likewise.
1651 * doc/gimple.texi (gimple_assign_cast_p): Accept a const_gimple
1652 rather than a gimple.
1653 (gimple_assign_cast_p): Likewise.
1655 2014-04-21 Michael Meissner <meissner@linux.vnet.ibm.com>
1658 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64 case):
1659 If mode is DDmode and TARGET_E500_DOUBLE allow move.
1661 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print some
1662 more debug information for E500 if -mdebug=reg.
1664 2014-04-21 Uros Bizjak <ubizjak@gmail.com>
1667 * config/i386/i386.c (ix86_expand_builtin)
1668 <case IX86_BUILTIN_RDRAND{16,32,64}_STEP>: Use temporary
1669 register for target RTX.
1670 <case IX86_BUILTIN_RDSEED{16,32,64}_STEP>: Ditto.
1672 2014-04-18 Cong Hou <congh@google.com>
1674 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): Enhance
1675 the widen-mult pattern by handling two operands with different sizes,
1676 and operands whose size is smaller than half of the result type.
1678 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
1680 * ipa-inline.h (INLINE_HINT_known_hot): New hint.
1681 * ipa-inline-analysis.c (dump_inline_hints): Dump it.
1682 (do_estimate_edge_time): Compute it.
1683 * ipa-inline.c (want_inline_small_function_p): Bypass
1684 INLINE_INSNS_AUTO/SINGLE limits for calls that are known to be hot.
1686 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
1688 * ipa-inline.c (spec_rem): New static variable.
1689 (dump_overall_stats): New function.
1690 (dump_inline_stats): New function.
1692 2014-04-18 Richard Henderson <rth@redhat.com>
1694 * config/aarch64/aarch64.c (aarch64_register_move_cost): Pass a mode
1695 to GET_MODE_SIZE, not a reg_class_t.
1697 2014-04-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1699 * config/rs6000/vsx.md (vsx_xxmrghw_<mode>): Adjust for little-endian.
1700 (vsx_xxmrglw_<mode>): Likewise.
1702 2014-04-17 Michael Meissner <meissner@linux.vnet.ibm.com>
1705 * config/rs6000/rs6000.c (rs6000_setup_reg_addr_masks): Make sure
1706 GET_MODE_SIZE gets passed an enum machine_mode type and not integer.
1707 (rs6000_init_hard_regno_mode_ok): Likewise.
1709 2014-04-17 Jan Hubicka <hubicka@ucw.cz>
1711 * ipa-inline.c (inline_small_functions): Account only non-cold
1713 * doc/invoke.texi (inline-unit-growth): Update documentation.
1715 2014-04-17 Pat Haugen <pthaugen@us.ibm.com>
1717 * config/rs6000/rs6000.md (addti3, subti3): New.
1719 2014-04-17 H.J. Lu <hongjiu.lu@intel.com>
1722 * config/i386/i386.c (ix86_expand_clear): Remove outdated
1723 comment. Check optimize_insn_for_size_p instead of
1724 optimize_insn_for_speed_p.
1726 2014-04-17 Martin Jambor <mjambor@suse.cz>
1728 * gimple-iterator.c (gsi_start_edge): New function.
1729 * gimple-iterator.h (gsi_start_edge): Declare.
1730 * tree-sra.c (single_non_eh_succ): New function.
1731 (disqualify_ops_if_throwing_stmt): Renamed to
1732 disqualify_if_bad_bb_terminating_stmt. Allow throwing statements
1733 having one non-EH successor BB.
1734 (sra_modify_expr): If stmt ends bb, use single non-EH successor to
1735 generate loads into replacements.
1736 (sra_modify_assign): Likewise and and also use the simple path for
1738 (sra_modify_function_body): Commit statements on edges.
1740 2014-04-17 Richard Biener <rguenther@suse.de>
1743 * tree-ssa-propagate.c (valid_gimple_rhs_p): Allow vector
1744 comparison results and add clarifying comment.
1746 2014-04-17 Jakub Jelinek <jakub@redhat.com>
1748 * genmodes.c (struct mode_data): Add need_bytesize_adj field.
1749 (blank_mode): Initialize it.
1750 (emit_mode_size_inline, emit_mode_nunits_inline,
1751 emit_mode_inner_inline): New functions.
1752 (emit_insn_modes_h): Call them and surround their output with
1753 #if GCC_VERSION >= 4001 ... #endif.
1754 * machmode.h (GET_MODE_SIZE, GET_MODE_NUNITS, GET_MODE_INNER):
1755 For GCC_VERSION >= 4001 use mode_*_inline routines instead of
1756 mode_* arrays if the argument is __builtin_constant_p.
1757 * lower-subreg.c (dump_choices): Make sure GET_MODE_SIZE argument
1758 is enum machine_mode.
1760 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1762 * passes.c (opt_pass::execute): Adjust.
1763 (pass_manager::execute_pass_mode_switching): Likewise.
1764 (early_local_passes::execute): Likewise.
1765 (execute_one_pass): Pass cfun to the pass's execute method.
1766 * tree-pass.h (opt_pass::execute): Add function * argument.
1767 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
1768 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
1769 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
1770 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
1771 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
1772 config/sparc/sparc.c, cprop.c, dce.c, df-core.c, dse.c, dwarf2cfi.c,
1773 except.c, final.c, function.c, fwprop.c, gcse.c, gimple-low.c,
1774 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
1775 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
1776 ipa-inline-analysis.c, ipa-inline.c, ipa-profile.c, ipa-pure-const.c,
1777 ipa-reference.c, ipa-split.c, ipa.c, ira.c, jump.c, loop-init.c,
1778 lower-subreg.c, mode-switching.c, omp-low.c, postreload-gcse.c,
1779 postreload.c, predict.c, recog.c, ree.c, reg-stack.c, regcprop.c,
1780 reginfo.c, regrename.c, reorg.c, sched-rgn.c, stack-ptr-mod.c,
1781 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
1782 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
1783 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
1784 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-ssa-ccp.c,
1785 tree-ssa-copy.c, tree-ssa-copyrename.c, tree-ssa-dce.c,
1786 tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
1787 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
1788 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
1789 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
1790 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
1791 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
1792 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
1793 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
1794 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
1795 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
1798 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1800 * passes.c (opt_pass::gate): Take function * argument.
1801 (gate_all_early_local_passes): Merge into
1802 (early_local_passes::gate): this.
1803 (gate_all_early_optimizations): Merge into
1804 (all_early_optimizations::gate): this.
1805 (gate_all_optimizations): Mege into
1806 (all_optimizations::gate): this.
1807 (gate_all_optimizations_g): Merge into
1808 (all_optimizations_g::gate): this.
1809 (gate_rest_of_compilation): Mege into
1810 (rest_of_compilation::gate): this.
1811 (gate_postreload): Merge into
1812 (postreload::gate): this.
1813 (dump_one_pass): Pass cfun to the pass's gate method.
1814 (execute_ipa_summary_passes): Likewise.
1815 (execute_one_pass): Likewise.
1816 (ipa_write_summaries_2): Likewise.
1817 (ipa_write_optimization_summaries_1): Likewise.
1818 (ipa_read_summaries_1): Likewise.
1819 (ipa_read_optimization_summaries_1): Likewise.
1820 (execute_ipa_stmt_fixups): Likewise.
1821 * tree-pass.h (opt_pass::gate): Add function * argument.
1822 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c,
1823 combine-stack-adj.c, combine.c, compare-elim.c,
1824 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
1825 config/rl78/rl78.c, config/sh/sh_optimize_sett_clrt.cc,
1826 config/sh/sh_treg_combine.cc, config/sparc/sparc.c, cprop.c, cse.c,
1827 dce.c, df-core.c, dse.c, dwarf2cfi.c, except.c, fwprop.c, gcse.c,
1828 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
1829 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
1830 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
1831 loop-init.c, lower-subreg.c, mode-switching.c, modulo-sched.c,
1832 omp-low.c, postreload-gcse.c, postreload.c, predict.c, recog.c, ree.c,
1833 reg-stack.c, regcprop.c, regrename.c, reorg.c, sched-rgn.c,
1834 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
1835 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
1836 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c,
1837 tree-nrv.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
1838 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
1839 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
1840 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
1841 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
1842 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
1843 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
1844 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
1845 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
1846 tree-ssa.c, tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
1847 tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c, tsan.c, ubsan.c,
1848 var-tracking.c, vtable-verify.c, web.c: Adjust.
1850 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1852 * configure.ac: Check for -Woverloaded-virtual and enable it if found.
1853 * configure: Regenerate.
1855 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1857 * passes.c (dump_one_pass): don't check pass->has_gate.
1858 (execute_ipa_summary_passes): Likewise.
1859 (execute_one_pass): Likewise.
1860 (ipa_write_summaries_2): Likewise.
1861 (ipa_write_optimization_summaries_1): Likewise.
1862 (ipa_read_optimization_summaries_1): Likewise.
1863 (execute_ipa_stmt_fixups): Likewise.
1864 * tree-pass.h (pass_data::has_gate): Remove.
1865 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
1866 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
1867 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
1868 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
1869 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
1870 config/sh/sh_optimize_sett_clrt.cc, config/sh/sh_treg_combine.cc,
1871 config/sparc/sparc.c, cprop.c, cse.c, dce.c, df-core.c, dse.c,
1872 dwarf2cfi.c, except.c, final.c, function.c, fwprop.c, gcse.c,
1873 gimple-low.c, gimple-ssa-isolate-paths.c,
1874 gimple-ssa-strength-reduction.c, graphite.c, ifcvt.c, init-regs.c,
1875 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
1876 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
1877 ira.c, jump.c, loop-init.c, lower-subreg.c, mode-switching.c,
1878 modulo-sched.c, omp-low.c, postreload-gcse.c, postreload.c, predict.c,
1879 recog.c, ree.c, reg-stack.c, regcprop.c, reginfo.c, regrename.c,
1880 reorg.c, sched-rgn.c, stack-ptr-mod.c, store-motion.c, tracer.c,
1881 trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
1882 tree-complex.c, tree-eh.c, tree-emutls.c, tree-if-conv.c,
1883 tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
1884 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
1885 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
1886 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
1887 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
1888 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
1889 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
1890 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
1891 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
1892 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
1893 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
1894 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
1895 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
1898 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1900 * pass_manager.h (pass_manager::register_dump_files_1): Remove
1902 * passes.c (pass_manager::register_dump_files_1): Merge into
1903 (pass_manager::register_dump_files): this, and remove its handling of
1904 properties since the pass always has the properties anyway.
1905 (pass_manager::pass_manager): Adjust.
1907 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1909 * pass_manager.h (pass_manager::register_dump_files_1): Adjust.
1910 * passes.c (pass_manager::register_dump_files_1): Remove dead code
1911 dealing with properties.
1912 (pass_manager::register_dump_files): Adjust.
1914 2014-03-20 Mark Wielaard <mjw@redhat.com>
1916 * dwarf2out.c (add_bound_info): If HOST_WIDE_INT is big enough,
1917 then represent the bound as normal constant value.
1919 2014-04-17 Jakub Jelinek <jakub@redhat.com>
1922 Forward port from 4.8 branch
1923 2013-07-19 Kirill Yukhin <kirill.yukhin@intel.com>
1925 * config/i386/bmiintrin.h (_blsi_u32): New.
1929 (_blsmsk_u32): Ditto.
1930 (_blsmsk_u64): Ditto.
1931 (_tzcnt_u32): Ditto.
1932 (_tzcnt_u64): Ditto.
1934 2014-04-17 Kito Cheng <kito@0xlab.org>
1936 * gcc.c (used_arg): Prevent out of bound access for multilib_options.
1938 2014-04-17 Richard Biener <rguenther@suse.de>
1941 * tree-ssa-propagate.c (valid_gimple_rhs_p): Only allow effective
1942 boolean results for comparisons.
1944 2014-04-17 Richard Biener <rguenther@suse.de>
1946 PR tree-optimization/60836
1947 * tree-vect-loop.c (vect_create_epilog_for_reduction): Force
1948 initial PHI args to be gimple values.
1950 2014-04-17 Richard Biener <rguenther@suse.de>
1952 PR tree-optimization/60841
1953 * tree-vect-data-refs.c (vect_analyze_data_refs): Count stmts.
1954 * tree-vect-loop.c (vect_analyze_loop_2): Pass down number
1955 of stmts to SLP build.
1956 * tree-vect-slp.c (vect_slp_analyze_bb_1): Likewise.
1957 (vect_analyze_slp): Likewise.
1958 (vect_analyze_slp_instance): Likewise.
1959 (vect_build_slp_tree): Limit overall SLP tree growth.
1960 * tree-vectorizer.h (vect_analyze_data_refs,
1961 vect_analyze_slp): Adjust prototypes.
1963 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
1965 * config/i386/i386.c (x86_add_stmt_cost): Fix vector cost model for
1968 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
1970 * config/i386/x86-tune.def (TARGET_SLOW_PSHUFB): New tune definition.
1971 * config/i386/i386.h (TARGET_SLOW_PSHUFB): New tune flag.
1972 * config/i386/i386.c (expand_vec_perm_even_odd_1): Avoid byte shuffles
1973 for TARGET_SLOW_PSHUFB
1975 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
1977 * config/i386/i386.c (slm_cost): Adjust vec_to_scalar_cost.
1978 * config/i386/i386.c (intel_cost): Ditto.
1980 2014-04-17 Joey Ye <joey.ye@arm.com>
1982 * opts.c (OPT_fif_conversion, OPT_fif_conversion2): Disable for Og.
1984 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
1986 * opts.c (common_handle_option): Disable -fipa-reference coorectly
1989 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
1991 * ipa-devirt.c (odr_type_d): Add field all_derivations_known.
1992 (type_all_derivations_known_p): New predicate.
1993 (type_all_ctors_visible_p): New predicate.
1994 (type_possibly_instantiated_p): New predicate.
1995 (get_odr_type): Compute all_derivations_known.
1996 (dump_odr_type): Dump the flag.
1997 (maybe_record_type): Cleanup.
1998 (record_target_from_binfo): Add bases_to_consider array;
1999 record bases for types w/o instances and skip CXX destructor.
2000 (possible_polymorphic_call_targets_1): Add bases_to_consider
2001 and consider_construction parameters; check if type may have instance.
2002 (get_polymorphic_call_info): Set maybe_in_construction to true
2003 when we know nothing.
2004 (record_targets_from_bases): Skip CXX destructors; they are
2005 never called for types in construction.
2006 (possible_polymorphic_call_targets): Do not record target when
2007 type may not have instance.
2009 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
2012 * ipa.c (symtab_remove_unreachable_nodes): Mark targets of
2013 external aliases alive, too.
2015 2014-04-16 Andrew Pinski <apinski@cavium.com>
2017 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Change aarch64 ilp32
2020 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
2022 * final.c (compute_alignments): Do not apply loop alignment to a block
2023 falling through to the exit.
2025 2014-04-16 Catherine Moore <clm@codesourcery.com>
2027 * mips.md (*mov<mode>_internal, *movhi_internal, *movqi_internal):
2028 Adjust constraints for microMIPS store patterns.
2030 2014-04-16 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
2032 * config/avr/avr-mcus.def: Correct typo for atxmega256a3bu macro.
2034 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
2036 * tree-ssa-operands.c (create_vop_var): Set DECL_IGNORED_P.
2037 (append_use): Run at -O0.
2038 (append_vdef): Likewise.
2039 * tree-ssa-ter.c (ter_is_replaceable_p): Do not special-case -O0.
2040 * tree-ssa-uninit.c (warn_uninitialized_vars): Remove obsolete comment.
2042 2014-04-16 Jakub Jelinek <jakub@redhat.com>
2044 PR tree-optimization/60844
2045 * tree-ssa-reassoc.c (reassoc_remove_stmt): New function.
2046 (propagate_op_to_single_use, remove_visited_stmt_chain,
2047 linearize_expr, repropagate_negates, reassociate_bb): Use it
2048 instead of gsi_remove.
2050 2014-04-16 Martin Jambor <mjambor@suse.cz>
2052 * cgraphclones.c (cgraph_create_virtual_clone): Duplicate
2053 ipa_transforms_to_apply.
2054 (cgraph_function_versioning): Assert that old_node has empty
2055 ipa_transforms_to_apply.
2056 * trans-mem.c (ipa_tm_create_version): Likewise.
2057 * tree-inline.c (tree_function_versioning): Do not duplicate
2058 ipa_transforms_to_apply.
2060 2014-04-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2063 * configure.ac (set_have_as_tls): Merge i[34567]86-*-* and
2065 Pass necessary as flags on 64-bit Solaris/x86.
2066 Use lowercase relocs for x86_64-*-*.
2067 * configure: Regenerate.
2069 2014-04-15 Jan Hubicka <jh@suse.cz>
2071 * ipa-devirt.c (referenced_from_vtable_p): New predicate.
2072 (maybe_record_node, likely_target_p): Use it.
2074 2014-04-15 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2077 Revert following patch
2079 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
2082 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
2083 software floating point or no floating point registers, do not
2084 allow any type in the FPRs. Eliminate a test for SPE SIMD types
2085 in GPRs that occurs after we tested for GPRs that would never be
2088 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
2089 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
2090 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
2091 specifically allow DDmode, since that does not use the SPE SIMD
2094 2014-03-21 Mark Wielaard <mjw@redhat.com>
2096 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_const_value
2097 as unsigned or int depending on type and value used.
2099 2014-04-15 Richard Biener <rguenther@suse.de>
2101 PR rtl-optimization/56965
2102 * alias.c (ncr_compar, nonoverlapping_component_refs_p): Move ...
2103 * tree-ssa-alias.c (ncr_compar, nonoverlapping_component_refs_p):
2105 * alias.c (true_dependence_1): Do not call
2106 nonoverlapping_component_refs_p.
2107 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Call
2108 nonoverlapping_component_refs_p.
2109 (indirect_refs_may_alias_p): Likewise.
2111 2014-04-15 Teresa Johnson <tejohnson@google.com>
2113 * cfg.c (dump_bb_info): Fix flags check.
2114 * tree-cfg.c (remove_bb): Only dump TDF_BLOCKS when removing.
2116 2014-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2118 PR rtl-optimization/60663
2119 * config/arm/arm.c (arm_new_rtx_costs): Improve ASM_OPERANDS case,
2122 2014-04-15 Richard Biener <rguenther@suse.de>
2124 * lto-streamer.h (LTO_major_version): Bump to 4.
2126 2014-04-15 Richard Biener <rguenther@suse.de>
2128 * common.opt (lto_partition_model): New enum.
2129 (flto-partition=): Merge separate options with a single with argument,
2130 add -flto-partition=one support.
2131 * flag-types.h (enum lto_partition_model): Declare.
2132 * opts.c (finish_options): Remove duplicate -flto-partition=
2134 * lto-wrapper.c (run_gcc): Adjust.
2136 2014-04-15 Richard Biener <rguenther@suse.de>
2138 * alias.c (ncr_compar): New function.
2139 (nonoverlapping_component_refs_p): Re-implement in O (n log n).
2141 2014-04-15 Richard Biener <rguenther@suse.de>
2143 * alias.c (record_component_aliases): Do not walk BINFOs.
2145 2014-04-15 Richard Biener <rguenther@suse.de>
2147 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
2148 Add struct function argument and adjust.
2149 (find_func_aliases_for_call): Likewise.
2150 (find_func_aliases): Likewise.
2151 (find_func_clobbers): Likewise.
2152 (intra_create_variable_infos): Likewise.
2153 (compute_points_to_sets): Likewise.
2154 (ipa_pta_execute): Adjust. Do not push/pop cfun.
2156 2014-04-15 Richard Biener <rguenther@suse.de>
2158 * tree.c (iterative_hash_expr): Use enum tree_code_class
2159 to store TREE_CODE_CLASS.
2160 (tree_block): Likewise.
2161 (tree_set_block): Likewise.
2162 * tree.h (fold_build_pointer_plus_loc): Use
2163 convert_to_ptrofftype_loc.
2165 2014-04-15 Jakub Jelinek <jakub@redhat.com>
2168 * Makefile.in (PLUGIN_HEADERS): Add various headers that have been
2171 2014-04-15 Eric Botcazou <ebotcazou@adacore.com>
2173 * cfgloop.h (struct loop): Move force_vectorize down.
2174 * gimplify.c (gimple_boolify) <ANNOTATE_EXPR>: Handle new kinds.
2175 (gimplify_expr) <ANNOTATE_EXPR>: Minor tweak.
2176 * lto-streamer-in.c (input_cfg): Read dont_vectorize field.
2177 * lto-streamer-out.c (output_cfg): Write dont_vectorize field.
2178 * tree-cfg.c (replace_loop_annotate): Revamp and handle new kinds.
2179 * tree-core.h (enum annot_expr_kind): Add new kind values.
2180 * tree-inline.c (copy_loops): Copy dont_vectorize field and reorder.
2181 * tree-pretty-print.c (dump_generic_node) <ANNOTATE_EXPR>: Handle new
2183 * tree.def (ANNOTATE_EXPR): Tweak comment.
2185 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
2187 * ipa-devirt.c (maybe_record_node): Ignore all non-methods (including
2190 2014-04-14 Paolo Carlini <paolo.carlini@oracle.com>
2192 * tree.h (TYPE_IDENTIFIER): Declare.
2193 * tree.c (subrange_type_for_debug_p): Use it.
2194 * godump.c (go_format_type): Likewise.
2195 * dwarf2out.c (is_cxx_auto, modified_type_die,
2196 gen_type_die_with_usage, gen_type_die_with_usage): Likewise.
2197 * dbxout.c (dbxout_type, dbxout_symbol): Likewise.
2199 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
2202 * varpool.c (varpool_remove_node): Do not alter decls when streaming.
2204 2014-04-14 Uros Bizjak <ubizjak@gmail.com>
2206 * config/i386/i386.c (examine_argument): Return bool. Return true if
2207 parameter should be passed in memory.
2208 <case X86_64_COMPLEX_X87_CLASS>: Adjust.
2209 (construct_container): Update calls to examine_argument.
2210 (function_arg_advance_64): Ditto.
2211 (return_in_memory_32): Merge with ix86_return_in_memory.
2212 (return_in_memory_64): Ditto.
2213 (return_in_memory_ms_64): Ditto.
2215 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
2217 * ipa-utils.c (ipa_merge_profiles): Merge profile_id.
2218 * coverage.c (coverage_compute_profile_id): Handle externally visible
2221 2014-04-14 Martin Jambor <mjambor@suse.cz>
2223 * tree-sra.c (ipa_sra_preliminary_function_checks): Skip
2224 DECL_DISREGARD_INLINE_LIMITS functions.
2226 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
2229 * config/i386/i386.md (*fixuns_trunc<mode>_1): Revert the last change.
2231 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
2234 * config/i386/i386.md (*fixuns_trunc<mode>_1): Check
2235 optimize_insn_for_speed_p instead of
2236 optimize_function_for_speed_p.
2238 2014-04-14 Yufeng Zhang <yufeng.zhang@arm.com>
2240 * doc/invoke.texi (free): Document AArch64.
2242 2014-04-14 Richard Biener <rguenther@suse.de>
2244 PR tree-optimization/60042
2245 * tree-ssa-pre.c (inhibit_phi_insertion): Remove.
2246 (insert_into_preds_of_block): Do not prevent PHI insertion
2247 for REFERENCE exprs here ...
2248 (eliminate_dom_walker::before_dom_children): ... but prevent
2249 their use here under similar conditions when applied to the
2250 IL after PRE optimizations.
2252 2014-04-14 Richard Biener <rguenther@suse.de>
2254 * passes.def: Move early points-to after early SRA.
2256 2014-04-14 Richard Biener <rguenther@suse.de>
2258 * tree-ssa-forwprop.c (simplify_gimple_switch): Enhance
2259 check for which sign-changes we allow when forwarding
2260 a converted value into a switch.
2262 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
2264 * stor-layout.c (place_field): Finalize non-constant offset for the
2267 2014-04-14 Richard Biener <rguenther@suse.de>
2269 * tree-switch-conversion.c (lshift_cheap_p): Get speed_p
2271 (expand_switch_using_bit_tests_p): Likewise.
2272 (process_switch): Compute and pass on speed_p based on the
2274 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Use
2275 optimize_bb_for_speed_p.
2277 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
2279 * cfgloop.h (struct loop): Rename force_vect into force_vectorize.
2280 * function.h (struct function): Rename has_force_vect_loops into
2281 has_force_vectorize_loops.
2282 * lto-streamer-in.c (input_cfg): Adjust for renaming.
2283 (input_struct_function_base): Likewise.
2284 * lto-streamer-out.c (output_cfg): Likewise.
2285 (output_struct_function_base): Likewise.
2286 * omp-low.c (expand_omp_simd): Likewise.
2287 * tree-cfg.c (move_sese_region_to_fn): Likewise.
2288 * tree-if-conv.c (ifcvt_can_use_mask_load_store): Likewise.
2289 (version_loop_for_if_conversion): Likewise.
2290 (tree_if_conversion): Likewise.
2291 (main_tree_if_conversion): Likewise.
2292 (gate_tree_if_conversion): Likewise.
2293 * tree-inline.c (copy_loops): Likewise.
2294 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Likewise.
2295 * tree-ssa-loop.c (tree_loop_vectorize): Likewise.
2296 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Likewise.
2297 * tree-vect-loop.c (vect_estimate_min_profitable_iters): Likewise.
2298 * tree-vectorizer.c (vectorize_loops): Likewise.
2299 * tree-vectorizer.h (unlimited_cost_model): Likewise.
2301 2014-04-14 Richard Biener <rguenther@suse.de>
2304 * lto-streamer-out.c (wrap_refs): New function.
2305 (lto_output): Wrap symbol references in global initializes in
2306 type-preserving MEM_REFs.
2308 2014-04-14 Christian Bruel <christian.bruel@st.com>
2310 * config/sh/sh-mem.cc (sh_expand_strlen): Unroll last word.
2312 2014-04-14 Christian Bruel <christian.bruel@st.com>
2314 * config/sh/sh.md (setmemqi): New expand pattern.
2315 * config/sh/sh.h (CLEAR_RATIO): Define.
2316 * config/sh/sh-mem.cc (sh_expand_setmem): Define.
2317 * config/sh/sh-protos.h (sh_expand_setmem): Declare.
2319 2014-04-14 Richard Biener <rguenther@suse.de>
2322 * fold-const.c (negate_expr_p): Don't negate directional rounding
2324 (fold_negate_expr): Likewise.
2326 2014-04-14 Richard Biener <rguenther@suse.de>
2328 PR tree-optimization/59817
2329 PR tree-optimization/60453
2330 * graphite-scop-detection.c (graphite_can_represent_scev): Complete
2331 recursion to catch all CHRECs in the scalar evolution and restrict
2332 the predicate for the remains appropriately.
2334 2014-04-12 Catherine Moore <clm@codesourcery.com>
2336 * config/mips/constraints.md: Add new register constraint "kb".
2337 * config/mips/mips.md (*mov<mode>_internal): Use constraint "kb".
2338 (*movhi_internal): Likewise.
2339 (*movqi_internal): Likewise.
2340 * config/mips/mips.h (M16_STORE_REGS): New register class.
2341 (REG_CLASS_NAMES): Add M16_STORE_REGS.
2342 (REG_CLASS_CONTENTS): Likewise.
2343 * config/mips/mips.c (mips_regno_to_class): Add M16_STORE_REGS.
2345 2014-04-11 Tobias Burnus <burnus@net-b.de>
2348 * doc/invoke.texi (-Wformat-signedness): Document it.
2349 (Wformat=2): Mention that this enables -Wformat-signedness.
2351 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
2353 * common/config/epiphany/epiphany-common.c
2354 (epiphany_option_optimization_table): Enable section anchors by
2355 default at -O1 or higher.
2356 * config/epiphany/epiphany.c (TARGET_MAX_ANCHOR_OFFSET): Define.
2357 (TARGET_MIN_ANCHOR_OFFSET): Likewise.
2358 (epiphany_rtx_costs) <SET>: For binary operators, the set as such
2359 carries no extra cost.
2360 (epiphany_legitimate_address_p): For BLKmode, apply SImode check.
2361 * config/epiphany/epiphany.h (ASM_OUTPUT_DEF): Define.
2362 * config/epiphany/predicates.md (memclob_operand): New predicate.
2363 * config/epiphany/epiphany.md (stack_adjust_add, stack_adjust_str):
2364 Use memclob_operand predicate and X constraint for operand 3.
2366 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
2368 * config/epiphany/epiphany.c (epiphany_rtx_cost): Compare
2369 with CC_N_NE / CC_C_LTU / CC_C_GTU carries no extra cost for
2372 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
2374 PR rtl-optimization/60651
2375 * mode-switching.c (optimize_mode_switching): Make sure to emit
2376 sets of a lower numbered entity before sets of a higher numbered
2377 entity to a mode of the same or lower priority.
2378 When creating a seginfo for a basic block that starts with a code
2379 label, move the insertion point past the code label.
2380 (new_seginfo): Document and enforce requirement that
2381 NOTE_INSN_BASIC_BLOCK only appears for empty blocks.
2382 * doc/tm.texi.in: Document ordering constraint for emitted mode sets.
2383 * doc/tm.texi: Regenerate.
2385 2014-01-11 Joern Rennecke <joern.rennecke@embecosm.com>
2388 * config/arc/arc.c (arc_save_restore): Fix assert typo.
2390 2013-04-11 Jakub Jelinek <jakub@redhat.com>
2392 * BASE-VER: Set to 4.10.0.
2394 2014-04-11 Tobias Burnus <burnus@net-b.de>
2397 * doc/bugreport.texi (Bugs): Remove nodes pointing to the nirvana.
2398 * doc/gcc.texi (Service): Update description in the @menu
2399 * doc/invoke.texi (Option Summary): Remove misplaced and
2402 2014-04-11 Steve Ellcey <sellcey@mips.com>
2403 Jakub Jelinek <jakub@redhat.com>
2406 * expr.c (convert_move): Use emit_store_flag_force instead of
2407 emit_store_flag. Pass lowpart_mode instead of VOIDmode as 5th
2410 2014-04-11 Richard Biener <rguenther@suse.de>
2413 * varasm.c (assemble_alias): Avoid endless error reporting
2414 recursion by setting TREE_ASM_WRITTEN.
2416 2014-04-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2418 * config/s390/s390.md: Add a splitter for NOT rtx.
2420 2014-04-11 Jakub Jelinek <jakub@redhat.com>
2422 PR rtl-optimization/60663
2423 * cse.c (cse_insn): Set src_volatile on ASM_OPERANDS in PARALLEL.
2425 2014-04-10 Jan Hubicka <hubicka@ucw.cz>
2426 Jakub Jelinek <jakub@redhat.com>
2429 * ipa.c (function_and_variable_visibility): Copy forced_by_abi
2430 flag from decl_node to node.
2432 2014-04-10 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2435 * config/arm/arm.c (TARGET_CONST_NOT_OK_FOR_DEBUG_P): Define
2436 (arm_const_not_ok_for_debug_p): Reject MINUS with SYM_REF's
2437 ameliorating the cases where it can be.
2439 2014-04-09 David Edelsohn <dje.gcc@gmail.com>
2442 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
2444 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
2445 (loadsync_<mode>): Change mode.
2446 (load_quadpti, store_quadpti): New.
2447 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
2448 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
2449 * config/rs6000/predicates.md (quad_memory_operand): !TARGET_SYNC_TI.
2451 2014-04-09 Cong Hou <congh@google.com>
2454 * doc/sourcebuild.texi (vect_widen_mult_si_to_di_pattern): Add
2457 2014-04-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2459 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Use vnand
2460 instead of vnor to exploit possible fusion opportunity in the
2462 (altivec_expand_vec_perm_const_le): Likewise.
2464 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
2466 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
2467 (loadsync_<mode>): Change mode.
2468 (load_quadpti, store_quadpti): New.
2469 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
2470 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
2472 2014-04-08 Richard Sandiford <rdsandiford@googlemail.com>
2475 * config/rs6000/vsx.md (vsx_xscvdpspn_scalar): Change input to DImode.
2476 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Update accordingly.
2477 Use gen_rtx_REG rather than simplify_gen_subreg for op0_di.
2479 2014-04-08 Richard Biener <rguenther@suse.de>
2482 * tree-pretty-print.c (pp_double_int): For HWI32 hosts with
2483 a 64bit widest int print double-int similar to on HWI64 hosts.
2485 2014-04-08 Richard Biener <rguenther@suse.de>
2487 PR tree-optimization/60785
2488 * graphite-sese-to-poly.c (rewrite_phi_out_of_ssa): Treat
2489 default defs properly.
2491 2014-04-08 Nathan Sidwell <nathan@codesourcery.com>
2493 * doc/invoke (Wnon-virtual-dtor): Update to match implementation.
2494 (Weffc++): Likewise.
2496 2014-04-07 Jan Hubicka <hubcika@ucw.cz>
2498 * ipa-devirt.c (maybe_record_node): When node is not recorded,
2499 set completep to false rather than true.
2501 2014-04-07 Douglas B Rupp <rupp@adacore.com>
2504 * config/arm/arm.h (ASM_PREFERRED_EH_DATA_FORMAT): Expose from
2505 ARM_TARGET2_DWARF_FORMAT.
2507 2014-04-07 Charles Baylis <charles.baylis@linaro.org>
2510 * config/arm/arm.h (ASM_OUTPUT_CASE_END): Remove.
2511 (LABEL_ALIGN_AFTER_BARRIER): Align barriers which occur after
2514 2014-04-07 Richard Biener <rguenther@suse.de>
2516 PR tree-optimization/60766
2517 * tree-ssa-loop-ivopts.c (cand_value_at): Compute in an unsigned type.
2518 (may_eliminate_iv): Convert cand_value_at result to desired type.
2520 2014-04-07 Jason Merrill <jason@redhat.com>
2523 * common.opt (-fno-gnu-unique): Add.
2524 * config/elfos.h (USE_GNU_UNIQUE_OBJECT): Check it.
2526 2014-04-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2528 * haifa-sched.c: Fix outdated function reference and minor
2529 grammar errors in introductory comment.
2531 2014-04-07 Richard Biener <rguenther@suse.de>
2534 * tree-ssa-operands.c (maybe_add_call_vops): Also add VDEFs
2536 * tree-cfgcleanup.c (fixup_noreturn_call): Do not remove VDEFs.
2538 2014-04-06 John David Anglin <danglin@gcc.gnu.org>
2541 * config/pa/pa.c (pa_output_function_epilogue): Skip address and code
2542 size accounting for thunks.
2543 (pa_asm_output_mi_thunk): Use final_start_function() and
2544 final_end_function() to output function start and end directives.
2546 2014-04-05 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
2548 * config/avr/avr-arch.h (avr_mcu_t): Add dev_attribute field to have
2549 device specific ISA/ feature information. Remove short_sp and
2550 errata_skip ds. Add avr_device_specific_features enum to have device
2552 * config/avr/avr-c.c (avr_cpu_cpp_builtins): use dev_attribute to check
2553 errata_skip. Add __AVR_ISA_RMW__ builtin macro if RMW ISA available.
2554 * config/avr/avr-devices.c (avr_mcu_types): Update AVR_MCU macro for
2555 updated device specific info.
2556 * config/avr/avr-mcus.def: Merge device specific details to
2557 dev_attribute field.
2558 * config/avr/avr.c (avr_2word_insn_p): use dev_attribute field to check
2560 * config/avr/avr.h (AVR_HAVE_8BIT_SP): same for short sp info.
2561 * config/avr/driver-avr.c (avr_device_to_as): Pass -mrmw option to
2562 assembler if RMW isa supported by current device.
2563 * config/avr/genmultilib.awk: Update as device info structure changed.
2564 * doc/invoke.texi: Add info for __AVR_ISA_RMW__ builtin macro
2566 2014-04-04 Cong Hou <congh@google.com>
2568 PR tree-optimization/60656
2569 * tree-vect-stmts.c (supportable_widening_operation):
2570 Fix a bug that elements in a vector with vect_used_by_reduction
2571 property are incorrectly reordered when the operation on it is not
2572 consistant with the one in reduction operation.
2574 2014-04-04 John David Anglin <danglin@gcc.gnu.org>
2576 PR rtl-optimization/60155
2577 * gcse.c (record_set_data): New function.
2578 (single_set_gcse): New function.
2579 (gcse_emit_move_after): Use single_set_gcse instead of single_set.
2580 (hoist_code): Likewise.
2581 (get_pressure_class_and_nregs): Likewise.
2583 2014-04-04 Eric Botcazou <ebotcazou@adacore.com>
2585 * explow.c (probe_stack_range): Emit a final optimization blockage.
2587 2014-04-04 Anthony Green <green@moxielogic.com>
2589 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2): Fix
2592 2014-04-04 Jan Hubicka <hubicka@ucw.cz>
2595 * lto-cgraph.c (input_overwrite_node): Check that partitioning
2596 flags are set only during streaming.
2597 * ipa.c (process_references, walk_polymorphic_call_targets,
2598 symtab_remove_unreachable_nodes): Drop bodies of always inline
2599 after early inlining.
2600 (symtab_remove_unreachable_nodes): Remove always_inline attribute.
2602 2014-04-04 Jakub Jelinek <jakub@redhat.com>
2603 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2606 * dwarf2out.c (const_ok_for_output_1): Reject expressions
2609 2014-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2612 * config/arm/cortex-a53.md (cortex_a53_fdivs): Reduce reservation
2614 (cortex_a53_fdivd): Likewise.
2616 2014-04-04 Martin Jambor <mjambor@suse.cz>
2619 * cgraph.h (cgraph_clone_node): New parameter added to declaration.
2621 * cgraph.c (clone_of_p): Also return true if thunks match.
2622 (verify_edge_corresponds_to_fndecl): Removed extraneous call to
2623 cgraph_function_or_thunk_node and an obsolete comment.
2624 * cgraphclones.c (build_function_type_skip_args): Moved upwards in the
2626 (build_function_decl_skip_args): Likewise.
2627 (set_new_clone_decl_and_node_flags): New function.
2628 (duplicate_thunk_for_node): Likewise.
2629 (redirect_edge_duplicating_thunks): Likewise.
2630 (cgraph_clone_node): New parameter args_to_skip, pass it to
2631 redirect_edge_duplicating_thunks which is called instead of
2632 cgraph_redirect_edge_callee.
2633 (cgraph_create_virtual_clone): Pass args_to_skip to cgraph_clone_node,
2634 moved setting of a lot of flags to set_new_clone_decl_and_node_flags.
2636 2014-04-04 Jeff Law <law@redhat.com>
2639 * config/arm/predicates.md (const_int_I_operand): New predicate.
2640 (const_int_M_operand): Similarly.
2641 * config/arm/arm.md (insv_zero): Use const_int_M_operand instead of
2643 (insv_t2, extv_reg, extzv_t2): Likewise.
2644 (load_multiple_with_writeback): Similarly for const_int_I_operand.
2645 (pop_multiple_with_writeback_and_return): Likewise.
2646 (vfp_pop_multiple_with_writeback): Likewise
2648 2014-04-04 Richard Biener <rguenther@suse.de>
2651 * tree-ssanames.c (make_ssa_name_fn): Fix assert.
2652 * gimple.c (gimple_set_bb): Avoid ICEing for NULL cfun for
2654 * gimplify.h (gimple_add_tmp_var_fn): Declare.
2655 * gimplify.c (gimple_add_tmp_var_fn): New function.
2656 * gimple-expr.h (create_tmp_reg_fn): Declare.
2657 * gimple-expr.c (create_tmp_reg_fn): New function.
2658 * gimple-low.c (record_vars_into): Don't change cfun.
2659 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix
2660 code generation without cfun.
2662 2014-04-04 Thomas Schwinge <thomas@codesourcery.com>
2665 * Makefile.in (install-driver): Fix shell scripting.
2667 2014-04-03 Cong Hou <congh@google.com>
2669 PR tree-optimization/60505
2670 * tree-vectorizer.h (struct _stmt_vec_info): Add th field as the
2671 threshold of number of iterations below which no vectorization
2673 * tree-vect-loop.c (new_loop_vec_info):
2674 Initialize LOOP_VINFO_COST_MODEL_THRESHOLD.
2675 * tree-vect-loop.c (vect_analyze_loop_operations):
2676 Set LOOP_VINFO_COST_MODEL_THRESHOLD.
2677 * tree-vect-loop.c (vect_transform_loop):
2678 Use LOOP_VINFO_COST_MODEL_THRESHOLD.
2679 * tree-vect-loop.c (vect_analyze_loop_2): Check the maximum number
2680 of iterations of the loop and see if we should build the epilogue.
2682 2014-04-03 Richard Biener <rguenther@suse.de>
2684 * tree-streamer.h (struct streamer_tree_cache_d): Add next_idx member.
2685 (streamer_tree_cache_create): Adjust.
2686 * tree-streamer.c (streamer_tree_cache_add_to_node_array): Adjust
2687 to allow optional nodes array.
2688 (streamer_tree_cache_insert_1): Use next_idx to assign idx.
2689 (streamer_tree_cache_append): Likewise.
2690 (streamer_tree_cache_create): Create nodes array optionally
2691 as specified by parameter.
2692 * lto-streamer-out.c (create_output_block): Avoid maintaining
2693 the node array in the writer cache.
2694 (DFS_write_tree): Remove assertion.
2695 (produce_asm_for_decls): Free the out decl state hash table early.
2696 * lto-streamer-in.c (lto_data_in_create): Adjust for
2697 streamer_tree_cache_create prototype change.
2699 2014-04-03 Richard Biener <rguenther@suse.de>
2701 * tree-streamer-out.c (streamer_write_chain): Do not temporarily
2702 set TREE_CHAIN to NULL_TREE.
2704 2014-04-03 Richard Biener <rguenther@suse.de>
2706 PR tree-optimization/60740
2707 * graphite-scop-detection.c (stmt_simple_for_scop_p): Iterate
2708 over all GIMPLE_COND operands.
2710 2014-04-03 Nathan Sidwell <nathan@codesourcery.com>
2712 * doc/invoke.texi (Wnon-virtual-dtor): Adjust documentation.
2713 (Weffc++): Remove Scott's numbering, merge lists and reference
2716 2014-04-03 Nick Clifton <nickc@redhat.com>
2718 * config/rl78/rl78-expand.md (movqi): Handle (SUBREG (SYMBOL_REF))
2721 2014-04-03 Martin Jambor <mjambor@suse.cz>
2723 * ipa-cp.c (ipcp_verify_propagated_values): Also dump symtab and
2724 mention gcc_unreachable before failing.
2725 * ipa.c (symtab_remove_unreachable_nodes): Also print order of
2728 2014-04-02 Jan Hubicka <hubicka@ucw.cz>
2731 * ipa-devirt.c (get_polymorphic_call_info): Do not ICE on type
2732 inconsistent code and instead mark the context inconsistent.
2733 (possible_polymorphic_call_targets): For inconsistent contexts
2734 return empty complete list.
2736 2014-04-02 Anthony Green <green@moxielogic.com>
2738 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2)
2739 (extendqisi2, extendhisi2): Define.
2740 * config/moxie/moxie.h (DEFAULT_SIGNED_CHAR): Change to 0.
2741 (WCHAR_TYPE): Change to unsigned int.
2743 2014-04-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2745 PR tree-optimization/60733
2746 * gimple-ssa-strength-reduction.c (ncd_with_phi): Change required
2747 insertion point for PHI candidates to be the end of the feeding
2748 block for the PHI argument.
2750 2014-04-02 Vladimir Makarov <vmakarov@redhat.com>
2752 PR rtl-optimization/60650
2753 * lra-constraints.c (process_alt_operands): Decrease reject for
2754 earlyclobber matching.
2756 2014-04-02 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2758 * config/s390/s390.c (s390_expand_insv): Use GET_MODE_BITSIZE.
2760 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
2762 * config/spu/spu.c (pad_bb): Do not crash when the last
2763 insn is CODE_FOR_blockage.
2765 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
2767 * config/spu/spu.md ("insv"): Fail if bitoffset+bitsize
2768 lies outside the target mode.
2770 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
2773 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
2774 software floating point or no floating point registers, do not
2775 allow any type in the FPRs. Eliminate a test for SPE SIMD types
2776 in GPRs that occurs after we tested for GPRs that would never be
2779 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
2780 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
2781 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
2782 specifically allow DDmode, since that does not use the SPE SIMD
2785 2014-04-02 Richard Biener <rguenther@suse.de>
2788 * optabs.c (expand_abs_nojump): Honor flag_trapv only for
2789 MODE_INTs. Properly use negv_optab.
2790 (expand_abs): Likewise.
2792 2014-04-02 Richard Biener <rguenther@suse.de>
2795 * Makefile.in (install-driver): Guard extra installs with special
2798 2014-04-01 Michael Meissner <meissner@linux.vnet.ibm.com>
2800 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
2803 2014-04-01 Richard Henderson <rth@redhat.com>
2806 * config/i386/i386.md (*float<SWI48><MODEF>2_sse): Leave the second
2807 alternative enabled before register allocation.
2809 2014-04-01 Chung-Lin Tang <cltang@codesourcery.com>
2811 * config/nios2/nios2.md (unspec): Remove UNSPEC_TLS, UNSPEC_TLS_LDM.
2812 * config/nios2/nios2.c (nios2_function_profiler): Fix addi operand
2814 (nios2_large_got_address): Remove unneeded 'sym' parameter.
2815 (nios2_got_address): Update nios2_large_got_address call site.
2816 (nios2_delegitimize_address): New function.
2817 (TARGET_DELEGITIMIZE_ADDRESS): Define to nios2_delegitimize_address.
2818 * config/nios2/linux.h (GLIBC_DYNAMIC_LINKER): Define.
2819 (LINK_SPEC): Specify dynamic linker using GNU_USER_DYNAMIC_LINKER.
2821 2014-04-01 Martin Husemann <martin@duskware.de>
2823 * config/mips/netbsd.h (TARGET_OS_CPP_BUILTINS): Define __mips_o32
2826 2014-04-01 Richard Sandiford <rdsandiford@googlemail.com>
2828 PR rtl-optimization/60604
2829 * recog.c (general_operand): Incorporate REG_CANNOT_CHANGE_MODE_P
2830 check from register_operand.
2831 (register_operand): Redefine in terms of general_operand.
2832 (nonmemory_operand): Use register_operand for the non-constant cases.
2834 2014-04-01 Richard Biener <rguenther@suse.de>
2836 * gimple.h (struct gimple_statement_base): Align subcode to 16 bits.
2838 2014-04-01 Sebastian Huber <sebastian.huber@embedded-brains.de>
2840 * doc/invoke.texi (mapp-regs): Clarify.
2842 2014-03-31 Ulrich Drepper <drepper@gmail.com>
2844 * config/i386/avx512fintrin.h (__v32hi): Define type.
2845 (__v64qi): Likewise.
2846 (_mm512_set1_epi8): Define.
2847 (_mm512_set1_epi16): Define.
2848 (_mm512_set4_epi32): Define.
2849 (_mm512_set4_epi64): Define.
2850 (_mm512_set4_pd): Define.
2851 (_mm512_set4_ps): Define.
2852 (_mm512_setr4_epi64): Define.
2853 (_mm512_setr4_epi32): Define.
2854 (_mm512_setr4_pd): Define.
2855 (_mm512_setr4_ps): Define.
2856 (_mm512_setzero_epi32): Define.
2858 2014-03-31 Martin Jambor <mjambor@suse.cz>
2861 * tree-sra.c (callsite_has_enough_arguments_p): Renamed to
2862 callsite_arguments_match_p. Updated all callers. Also check types of
2863 corresponding formal parameters and actual arguments.
2864 (not_all_callers_have_enough_arguments_p) Renamed to
2865 some_callers_have_mismatched_arguments_p.
2867 2014-03-31 Yuri Rumyantsev <ysrumyan@gmail.com>
2869 * tree-inline.c (copy_loops): Add missed copy of 'safelen'.
2871 2014-03-31 Kugan Vivekanandarajah <kuganv@linaro.org>
2874 * aarch64/aarch64.c (aarch64_classify_address): Fix alignment for
2877 2014-03-30 Uros Bizjak <ubizjak@gmail.com>
2879 * config/i386/sse.md (FMAMODE_NOVF512): New mode iterator.
2880 (<sd_mask_codefor>fma_fmadd_<mode><sd_maskz_name><round_name>):
2882 <sd_mask_codefor>fma_fmadd_<VF_512:mode><sd_maskz_name><round_name>.
2883 Use FMAMODE_NOVF512 mode iterator.
2884 (<sd_mask_codefor>fma_fmsub_<mode><sd_maskz_name><round_name>): Ditto.
2885 (<sd_mask_codefor>fma_fnmadd_<mode><sd_maskz_name><round_name>): Ditto.
2886 (<sd_mask_codefor>fma_fnmsub_<mode><sd_maskz_name><round_name>): Ditto.
2887 (<sd_mask_codefor>fma_fmaddsub_<mode><sd_maskz_name><round_name>):
2889 <sd_mask_codefor>fma_fmaddsub_<VF_512:mode><sd_maskz_name><round_name>.
2890 Use VF_128_256 mode iterator.
2891 (<sd_mask_codefor>fma_fmsubadd_<mode><sd_maskz_name><round_name>):
2894 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
2896 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Clear
2897 static chain if needed.
2899 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
2902 * lra-constraints.c (index_part_to_reg): New.
2903 (process_address): Use it.
2905 2014-03-27 Jeff Law <law@redhat.com>
2906 Jakub Jelinek <jakub@redhat.com>
2909 * expr.c (do_tablejump): Use simplify_gen_binary rather than
2910 gen_rtx_{PLUS,MULT} to build up the address expression.
2912 * i386/i386.c (ix86_legitimize_address): Use copy_addr_to_reg to avoid
2913 creating non-canonical RTL.
2915 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
2918 * ipa-inline.c (want_inline_small_function_p): Short circuit large
2919 functions; reorganize to make cheap checks first.
2920 (inline_small_functions): Do not estimate growth when dumping;
2922 * ipa-inline.h (inline_summary): Add min_size.
2923 (growth_likely_positive): New function.
2924 * ipa-inline-analysis.c (dump_inline_summary): Add min_size.
2925 (set_cond_stmt_execution_predicate): Cleanup.
2926 (estimate_edge_size_and_time): Compute min_size.
2927 (estimate_calls_size_and_time): Likewise.
2928 (estimate_node_size_and_time): Likewise.
2929 (inline_update_overall_summary): Update min_size.
2930 (do_estimate_edge_time): Likewise.
2931 (do_estimate_edge_size): Update.
2932 (do_estimate_edge_hints): Update.
2933 (growth_likely_positive): New function.
2935 2014-03-28 Jakub Jelinek <jakub@redhat.com>
2938 * config/i386/i386.c (ix86_copy_addr_to_reg): Call copy_addr_to_reg
2939 also if addr has VOIDmode.
2941 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2943 * config/arm/aarch-common.c (aarch_crypto_can_dual_issue): New.
2944 * config/arm/aarch-common-protos.h (aarch_crypto_can_dual_issue):
2946 * config/arm/cortex-a53.md: Add reservations and bypass for crypto
2947 instructions as well as AdvancedSIMD loads.
2949 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2951 * config/aarch64/aarch64-simd.md (aarch64_crypto_aes<aes_op>v16qi):
2952 Use crypto_aese type.
2953 (aarch64_crypto_aes<aesmc_op>v16qi): Use crypto_aesmc type.
2954 * config/arm/arm.md (is_neon_type): Replace crypto_aes with
2955 crypto_aese, crypto_aesmc. Move to types.md.
2956 * config/arm/types.md (crypto_aes): Split into crypto_aese,
2958 * config/arm/iterators.md (crypto_type): Likewise.
2960 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
2962 * cgraph.c: Include expr.h and tree-dfa.h.
2963 (cgraph_redirect_edge_call_stmt_to_callee): If call in noreturn;
2966 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
2969 * lra-assigns.c (find_hard_regno_for): Remove unavailable hard
2970 regs from checking multi-reg pseudos.
2972 2014-03-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2974 * config/arm/t-aprofile (MULTILIB_MATCHES): Correct A12 rule.
2976 2014-03-28 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
2978 * config/rs6000/rs6000.c (fusion_gpr_load_p): Refuse optimization
2979 if it would clobber the stack pointer, even temporarily.
2981 2014-03-28 Eric Botcazou <ebotcazou@adacore.com>
2983 * mode-switching.c: Make small adjustments to the top comment.
2985 2014-03-27 Michael Meissner <meissner@linux.vnet.ibm.com>
2987 * config/rs6000/constraints.md (wD constraint): New constraint to
2988 match the constant integer to get the top DImode/DFmode out of a
2989 vector in a VSX register.
2991 * config/rs6000/predicates.md (vsx_scalar_64bit): New predicate to
2992 match the constant integer to get the top DImode/DFmode out of a
2993 vector in a VSX register.
2995 * config/rs6000/rs6000-builtins.def (VBPERMQ): Add vbpermq builtin
2998 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
3001 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
3002 -mdebug=reg, print value of VECTOR_ELEMENT_SCALAR_64BIT.
3004 * config/rs6000/vsx.md (vsx_extract_<mode>, V2DI/V2DF modes):
3005 Optimize vec_extract of 64-bit values, where the value being
3006 extracted is in the top word, where we can use scalar
3007 instructions. Add direct move and store support. Combine the big
3008 endian/little endian vector select load support into a single insn.
3009 (vsx_extract_<mode>_internal1): Likewise.
3010 (vsx_extract_<mode>_internal2): Likewise.
3011 (vsx_extract_<mode>_load): Likewise.
3012 (vsx_extract_<mode>_store): Likewise.
3013 (vsx_extract_<mode>_zero): Delete, big and little endian insns are
3014 combined into vsx_extract_<mode>_load.
3015 (vsx_extract_<mode>_one_le): Likewise.
3017 * config/rs6000/rs6000.h (VECTOR_ELEMENT_SCALAR_64BIT): Macro to
3018 define the top 64-bit vector element.
3020 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wD
3023 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
3024 Document vec_vbpermq builtin.
3027 * config/rs6000/altivec.h (vec_xxsldwi): Add missing define to
3028 enable use of xxsldwi and xxpermdi builtin functions.
3029 (vec_xxpermdi): Likewise.
3031 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
3032 Document use of vec_xxsldwi and vec_xxpermdi builtins.
3034 2014-03-27 Vladimir Makarov <vmakarov@redhat.com>
3036 PR rtl-optimization/60650
3037 * lra-assign.c (find_hard_regno_for, spill_for): Add parameter
3039 (find_spills_for): New.
3040 (assign_by_spills): Pass the new parameter to find_hard_regno_for.
3041 Spill all pseudos on the second iteration.
3043 2014-03-27 Marek Polacek <polacek@redhat.com>
3046 * doc/extend.texi (ffs Builtins): Change unsigned types to signed
3049 2014-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3051 * config/s390/s390.c (s390_can_use_return_insn): Check for
3052 call-saved FPRs on 31 bit.
3054 2014-03-27 Jakub Jelinek <jakub@redhat.com>
3057 * omp-low.c (lower_omp_1): For gimple_clobber_p stmts,
3058 if they need regimplification, just drop them instead of
3059 calling gimple_regimplify_operands on them.
3061 2014-03-27 Marcus Shawcroft <marcus.shawcroft@arm.com>
3064 * config/aarch64/aarch64.c (faked_omit_frame_pointer): Remove.
3065 (aarch64_frame_pointer_required): Adjust logic.
3066 (aarch64_can_eliminate): Adjust logic.
3067 (aarch64_override_options_after_change): Adjust logic.
3069 2014-03-27 Dehao Chen <dehao@google.com>
3071 * ipa-inline.c (early_inliner): Update node's inline info.
3073 2014-03-26 Dehao Chen <dehao@google.com>
3075 * dojump.c (do_compare_rtx_and_jump): Sets correct probability for
3076 compiler inserted conditional jumps for NAN float check.
3078 2014-03-26 Jakub Jelinek <jakub@redhat.com>
3080 * ubsan.h (ubsan_create_data): Change second argument's type
3081 to const location_t *.
3082 * ubsan.c (ubsan_source_location): If xloc.file is NULL, set it to
3084 (ubsan_create_data): Change second argument to const location_t *PLOC.
3085 Create Loc field whenever PLOC is non-NULL.
3086 (ubsan_instrument_unreachable, ubsan_expand_null_ifn,
3087 ubsan_build_overflow_builtin, instrument_bool_enum_load): Adjust
3091 * real.c (real_to_integer2): Change type of low to UHWI.
3093 2014-03-26 Tobias Burnus <burnus@net-b.de>
3095 * gcc.c (LINK_COMMAND_SPEC): Use libcilkrts.spec for -fcilkplus.
3096 (CILK_SELF_SPECS): New define.
3097 (driver_self_specs): Use it.
3099 2014-03-26 Richard Biener <rguenther@suse.de>
3101 * tree-pretty-print.c (percent_K_format): Implement special
3102 case for LTO and its stripped down BLOCK tree.
3104 2014-03-26 Jakub Jelinek <jakub@redhat.com>
3107 * ubsan.c (instrument_si_overflow): Instrument ABS_EXPR.
3109 * tree-vrp.c (simplify_internal_call_using_ranges): If only
3110 one range is range_int_cst_p, but not both, at least optimize
3111 addition/subtraction of 0 and multiplication by 0 or 1.
3112 * gimple-fold.c (gimple_fold_call): Fold
3113 IFN_UBSAN_CHECK_{ADD,SUB,MUL}.
3114 (gimple_fold_stmt_to_constant_1): If both op0 and op1 aren't
3115 INTEGER_CSTs, try to fold at least x * 0 and y - y.
3117 2014-03-26 Eric Botcazou <ebotcazou@adacore.com>
3119 PR rtl-optimization/60452
3120 * rtlanal.c (rtx_addr_can_trap_p_1): Fix head comment.
3121 <case REG>: Return 1 for invalid offsets from the frame pointer.
3123 2014-03-26 Marek Polacek <polacek@redhat.com>
3126 * doc/extend.texi (C Extensions): Mention variable-length arrays in
3129 2014-03-26 Marek Polacek <polacek@redhat.com>
3132 * doc/extend.texi (Designated Inits): Describe what happens to omitted
3135 2014-03-26 Marek Polacek <polacek@redhat.com>
3138 * ira-color.c (update_conflict_hard_regno_costs): Perform the
3139 multiplication in unsigned type.
3141 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
3143 * doc/install.texi: Document nds32le-*-elf and nds32be-*-elf.
3145 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
3147 * doc/contrib.texi: Add myself as Andes nds32 port contributor.
3149 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
3152 * cif-code.def (UNREACHABLE) New code.
3153 * ipa-inline.c (inline_small_functions): Skip edges to
3154 __builtlin_unreachable.
3155 (estimate_edge_growth): Allow edges to __builtlin_unreachable.
3156 * ipa-inline-analysis.c (edge_set_predicate): Redirect edges with false
3157 predicate to __bulitin_unreachable.
3158 (set_cond_stmt_execution_predicate): Fix issue when
3159 invert_tree_comparison returns ERROR_MARK.
3160 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow): Do not
3161 propagate to inline clones.
3162 * cgraph.c (verify_edge_corresponds_to_fndecl): Allow redirection
3164 * ipa-cp.c (create_specialized_node): Be ready for new node to appear.
3165 * cgraphclones.c (cgraph_clone_node): If call destination is already
3166 ureachable, do not redirect it back.
3167 * tree-inline.c (fold_marked_statements): Hanlde calls becoming
3170 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
3172 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow):
3173 Do not modify inline clones.
3175 2014-03-25 Jakub Jelinek <jakub@redhat.com>
3177 * config/i386/i386.md (general_sext_operand): New mode attr.
3178 (addv<mode>4, subv<mode>4, mulv<mode>4): If operands[2] is CONST_INT,
3179 don't generate (sign_extend (const_int)).
3180 (*addv<mode>4, *subv<mode>4, *mulv<mode>4): Disallow CONST_INT_P
3181 operands[2]. Use We constraint instead of <i> and
3182 <general_sext_operand> predicate instead of <general_operand>.
3183 (*addv<mode>4_1, *subv<mode>4_1, *mulv<mode>4_1): New insns.
3184 * config/i386/constraints.md (We): New constraint.
3185 * config/i386/predicates.md (x86_64_sext_operand,
3186 sext_operand): New predicates.
3188 2014-03-25 Martin Jambor <mjambor@suse.cz>
3191 * ipa-cp.c (ipa_get_indirect_edge_target_1): Redirect type
3192 inconsistent devirtualizations to __builtin_unreachable.
3194 2014-03-25 Marek Polacek <polacek@redhat.com>
3197 * doc/extend.texi (Example of asm with clobbered asm reg): Fix typo.
3199 2014-03-25 Alan Lawrence <alan.lawrence@arm.com>
3201 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Reverse
3202 order of elements for big-endian.
3204 2014-03-25 Richard Biener <rguenther@suse.de>
3207 * gimplify-me.c (gimple_regimplify_operands): Update the
3210 2014-03-25 Martin Jambor <mjambor@suse.cz>
3213 * lto-cgraph.c (lto_output_node): Stream body_removed flag.
3214 (lto_output_varpool_node): Likewise.
3215 (input_overwrite_node): Likewise.
3216 (input_varpool_node): Likewise.
3218 2014-03-25 Richard Biener <rguenther@suse.de>
3220 * lto-wrapper.c (merge_and_complain): Handle OPT_fPIE like OPT_fpie.
3221 (run_gcc): Likewise.
3223 2014-03-25 Jakub Jelinek <jakub@redhat.com>
3225 * combine.c (simplify_compare_const): Add MODE argument.
3226 Handle mode_width 0 as very large mode_width.
3227 (try_combine, simplify_comparison): Adjust callers.
3229 * cselib.c (cselib_hash_rtx): Perform addition in unsigned
3230 type to avoid signed integer overflow.
3231 * explow.c (plus_constant): Likewise.
3233 2014-03-25 Dominik Vogt <vogt@linux.vnet.ibm.com>
3235 * doc/generic.texi: Correct typos.
3237 2014-03-24 Tobias Burnus <burnus@net-b.de>
3239 * doc/invoke.texi (-flto): Expand section about
3240 using static libraries with LTO.
3242 2014-03-24 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3244 PR rtl-optimization/60501
3245 * optabs.def (addptr3_optab): New optab.
3246 * optabs.c (gen_addptr3_insn, have_addptr3_insn): New function.
3247 * doc/md.texi ("addptrm3"): Document new RTL standard expander.
3248 * expr.h (gen_addptr3_insn, have_addptr3_insn): Add prototypes.
3250 * lra.c (emit_add3_insn): Use the addptr pattern if available.
3252 * config/s390/s390.md ("addptrdi3", "addptrsi3"): New expanders.
3254 2014-03-24 Ulrich Drepper <drepper@gmail.com>
3256 * config/i386/avx512fintrin.h: Define _mm512_set1_ps and
3259 * config/i386/avxintrin.h (_mm256_undefined_si256): Define.
3260 (_mm256_undefined_ps): Define.
3261 (_mm256_undefined_pd): Define.
3262 * config/i386/emmintrin.h (_mm_undefined_si128): Define.
3263 (_mm_undefined_pd): Define.
3264 * config/i386/xmmintrin.h (_mm_undefined_ps): Define.
3265 * config/i386/avx512fintrin.h (_mm512_undefined_si512): Define.
3266 (_mm512_undefined_ps): Define.
3267 (_mm512_undefined_pd): Define.
3268 Use _mm*_undefined_*.
3269 * config/i386/avx2intrin.h: Use _mm*_undefined_*.
3271 2014-03-24 Alex Velenko <Alex.Velenko@arm.com>
3273 * config/aarch64/aarch64-simd-builtins.def (lshr): DI mode excluded.
3274 (lshr_simd): DI mode added.
3275 * config/aarch64/aarch64-simd.md (aarch64_lshr_simddi): New pattern.
3276 (aarch64_ushr_simddi): Likewise.
3277 * config/aarch64/aarch64.md (UNSPEC_USHR64): New unspec.
3278 * config/aarch64/arm_neon.h (vshr_n_u64): Intrinsic fixed.
3279 (vshrd_n_u64): Likewise.
3281 2014-03-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3283 * Makefile.in (s-macro_list): Depend on cc1.
3285 2014-03-23 Teresa Johnson <tejohnson@google.com>
3287 * ipa-utils.c (ipa_print_order): Use specified dump file.
3289 2014-03-23 Eric Botcazou <ebotcazou@adacore.com>
3291 PR rtl-optimization/60601
3292 * bb-reorder.c (fix_up_fall_thru_edges): Test EDGE_FALLTHRU everywhere.
3294 * gcc.c (eval_spec_function): Initialize save_growing_value.
3296 2014-03-22 Jakub Jelinek <jakub@redhat.com>
3299 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): For
3300 code == MINUS_EXPR, never swap op0 with op1.
3302 * toplev.c (init_local_tick): Avoid signed integer multiplication
3304 * genautomata.c (reserv_sets_hash_value): Fix rotate idiom, avoid
3305 shift by first operand's bitsize.
3307 2014-03-21 Jakub Jelinek <jakub@redhat.com>
3310 * config/i386/i386.h (TARGET_64BIT_P): If not TARGET_BI_ARCH,
3312 * config/i386/darwin.h (TARGET_64BIT_P): Redefine to
3313 TARGET_ISA_64BIT_P(x).
3315 2014-03-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3317 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Generate a
3318 pattern for vector nor instead of subtract from splat(-1).
3319 (altivec_expand_vec_perm_const_le): Likewise.
3321 2014-03-21 Richard Henderson <rth@twiddle.net>
3324 * ifcvt.c (dead_or_predicable): Return FALSE if there are any frame
3325 related insns after epilogue_completed.
3327 2014-03-21 Martin Jambor <mjambor@suse.cz>
3330 * cgraph.h (symtab_node): New flag body_removed.
3331 * ipa.c (symtab_remove_unreachable_nodes): Set body_removed flag
3332 when removing bodies.
3333 * symtab.c (dump_symtab_base): Dump body_removed flag.
3334 * cgraph.c (verify_edge_corresponds_to_fndecl): Skip nodes which
3335 had their bodies removed.
3337 2014-03-21 Martin Jambor <mjambor@suse.cz>
3340 * ipa.c (symtab_remove_unreachable_nodes): Clear thunk flag of nodes
3343 2014-03-21 Richard Biener <rguenther@suse.de>
3345 PR tree-optimization/60577
3346 * tree-core.h (struct tree_base): Document nothrow_flag use
3347 in VAR_DECL_NONALIASED.
3348 * tree.h (VAR_DECL_NONALIASED): New.
3349 (may_be_aliased): Adjust.
3350 * coverage.c (build_var): Set VAR_DECL_NONALIASED.
3352 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
3354 * expr.c (expand_expr_real_1): Remove outdated comment.
3356 2014-03-20 Jakub Jelinek <jakub@redhat.com>
3359 * ira.c (adjust_cleared_regs): Call copy_rtx on
3360 *reg_equiv[REGNO (loc)].src_p before passing it to
3361 simplify_replace_fn_rtx.
3364 * config/i386/i386.c (x86_output_mi_thunk): Surround UNSPEC_GOT
3365 into CONST, put pic register as first operand of PLUS. Use
3366 gen_const_mem for both 32-bit and 64-bit PIC got loads.
3368 2014-03-20 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3370 * config/aarch64/aarch64.c (MEMORY_MOVE_COST): Delete.
3372 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
3374 * config/sparc/sparc.c (sparc_do_work_around_errata): Implement work
3375 around for store forwarding issue in the FPU on the UT699.
3376 * config/sparc/sparc.md (in_branch_delay): Return false for single FP
3377 loads and operations if -mfix-ut699 is specified.
3378 (divtf3_hq): Tweak attribute.
3379 (sqrttf2_hq): Likewise.
3381 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
3383 * calls.c (store_one_arg): Remove incorrect const qualification on the
3384 type of the temporary.
3385 * cfgexpand.c (expand_return): Likewise.
3386 * expr.c (expand_constructor): Likewise.
3387 (expand_expr_real_1): Likewise.
3389 2014-03-20 Zhenqiang Chen <zhenqiang.chen@linaro.org>
3391 * config/arm/arm.c (arm_dwarf_register_span): Update the element number
3394 2014-03-19 Kaz Kojima <kkojima@gcc.gnu.org>
3397 * config/sh/sh.md (udivsi3_i1): Clobber R1 register.
3399 2014-03-19 James Greenhalgh <james.greenhalgh@arm.com>
3401 * config/arm/aarch-common-protos.h
3402 (alu_cost_table): Fix spelling of "extend".
3403 * config/arm/arm.c (arm_new_rtx_costs): Fix spelling of "extend".
3405 2014-03-19 Richard Biener <rguenther@suse.de>
3408 * tree-core.h (tree_type_common): Re-order pointer members
3409 to reduce recursion depth during GC walks.
3411 2014-03-19 Marek Polacek <polacek@redhat.com>
3414 * ubsan.c (ubsan_type_descriptor): Check that DECL_NAME is nonnull
3415 before accessing it.
3417 2014-03-19 Richard Biener <rguenther@suse.de>
3420 * lto-streamer-in.c (input_function): In WPA stage do not drop
3423 2014-03-19 Jakub Jelinek <jakub@redhat.com>
3425 PR tree-optimization/60559
3426 * vectorizable_mask_load_store): Replace scalar MASK_LOAD
3427 with build_zero_cst assignment.
3429 2014-03-18 Kai Tietz <ktietz@redhat.com>
3431 PR rtl-optimization/56356
3432 * sdbout.c (sdbout_parms): Verify that parms'
3433 incoming argument is valid.
3434 (sdbout_reg_parms): Likewise.
3436 2014-03-18 Richard Henderson <rth@redhat.com>
3439 * config/i386/i386.md (*float<SWI48x><MODEF>2_i387): Move down to
3440 be shadowed by *float<SWI48><MODEF>2_sse. Test X87_ENABLE_FLOAT.
3441 (*float<SWI48><MODEF>2_sse): Check X87_ENABLE_FLOAT for alternative 0.
3443 2014-03-18 Basile Starynkevitch <basile@starynkevitch.net>
3445 * plugin.def: Improve comment for PLUGIN_INCLUDE_FILE.
3446 * doc/plugins.texi (Plugin callbacks): Mention PLUGIN_INCLUDE_FILE.
3447 Italicize plugin event names in description. Explain that
3448 PLUGIN_PRAGMAS has no sense for lto1. Explain PLUGIN_INCLUDE_FILE.
3449 Remind that no GCC functions should be called after PLUGIN_FINISH.
3450 Explain what pragmas with expansion are.
3452 2014-03-18 Martin Liska <mliska@suse.cz>
3454 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Added case when
3455 gimple call statement is update.
3456 * gimple-fold.c (gimple_fold_call): Changed order for GIMPLE_ASSIGN and
3457 GIMPLE_CALL, where gsi iterator still points to GIMPLE CALL.
3459 2014-03-18 Jakub Jelinek <jakub@redhat.com>
3462 * ubsan.c (ubsan_instrument_unreachable): Call
3463 initialize_sanitizer_builtins.
3464 (ubsan_pass): Likewise.
3467 * ubsan.c (ubsan_type_descriptor, ubsan_create_data): Call
3468 varpool_finalize_decl instead of rest_of_decl_compilation.
3470 2014-03-18 Richard Biener <rguenther@suse.de>
3472 * df-problems.c (df_rd_confluence_n): Avoid bitmap_copy
3473 by using bitmap_and_compl instead of bitmap_and_compl_into.
3474 (df_rd_transfer_function): Likewise.
3476 2014-03-18 Richard Biener <rguenther@suse.de>
3478 * doc/lto.texi (fresolution): Fix typo.
3480 2014-03-18 Richard Biener <rguenther@suse.de>
3482 * doc/invoke.texi (flto): Update for changes in 4.9.
3484 2014-03-18 Richard Biener <rguenther@suse.de>
3486 * doc/loop.texi: Remove section on the removed lambda framework.
3487 Update loop docs with recent changes in preserving loop structure.
3489 2014-03-18 Richard Biener <rguenther@suse.de>
3491 * doc/lto.texi (-fresolution): Document.
3493 2014-03-18 Richard Biener <rguenther@suse.de>
3495 * doc/contrib.texi: Adjust my name.
3497 2014-03-18 Jakub Jelinek <jakub@redhat.com>
3500 * internal-fn.c: Include diagnostic-core.h.
3501 (expand_BUILTIN_EXPECT): New function.
3502 * gimplify.c (gimplify_call_expr): Use false instead of FALSE.
3503 (gimplify_modify_expr): Gimplify 3 argument __builtin_expect into
3504 IFN_BUILTIN_EXPECT call instead of __builtin_expect builtin call.
3505 * ipa-inline-analysis.c (find_foldable_builtin_expect): Handle
3507 * predict.c (expr_expected_value_1): Handle IFN_BUILTIN_EXPECT.
3508 Revert 3 argument __builtin_expect code.
3509 (strip_predict_hints): Handle IFN_BUILTIN_EXPECT.
3510 * gimple-fold.c (gimple_fold_call): Likewise.
3511 * tree.h (fold_builtin_expect): New prototype.
3512 * builtins.c (build_builtin_expect_predicate): Add predictor
3513 argument, if non-NULL, create 3 argument __builtin_expect.
3514 (fold_builtin_expect): No longer static. Add ARG2 argument,
3515 pass it through to build_builtin_expect_predicate.
3516 (fold_builtin_2): Adjust caller.
3517 (fold_builtin_3): Handle BUILT_IN_EXPECT.
3518 * internal-fn.def (BUILTIN_EXPECT): New.
3520 2014-03-18 Tobias Burnus <burnus@net-b.de>
3523 * predict.def (PRED_FORTRAN_OVERFLOW, PRED_FORTRAN_FAIL_ALLOC,
3524 PRED_FORTRAN_FAIL_IO, PRED_FORTRAN_WARN_ONCE, PRED_FORTRAN_SIZE_ZERO,
3525 PRED_FORTRAN_INVALID_BOUND, PRED_FORTRAN_ABSENT_DUMMY): Add.
3527 2014-03-18 Jan Hubicka <hubicka@ucw.cz>
3530 * predict.c (combine_predictions_for_bb): Fix up formatting.
3531 (expr_expected_value_1, expr_expected_value): Add predictor argument,
3532 fill what it points to if non-NULL.
3533 (tree_predict_by_opcode): Adjust caller, use the predictor.
3534 * predict.def (PRED_COMPARE_AND_SWAP): Add.
3536 2014-03-18 Eric Botcazou <ebotcazou@adacore.com>
3538 * config/sparc/sparc.c (sparc_do_work_around_errata): Speed up and use
3539 proper constant for the store mode.
3541 2014-03-18 Ilya Enkovich <ilya.enkovich@intel.com>
3543 * symtab.c (change_decl_assembler_name): Fix transparent alias
3546 2014-03-16 Renlin Li <Renlin.Li@arm.com>
3548 * config/aarch64/aarch64.c: Correct the comments about the
3549 aarch64 stack layout.
3551 2014-03-18 Thomas Schwinge <thomas@codesourcery.com>
3553 * omp-low.c (lower_rec_input_clauses) <build_omp_barrier>: Restore
3554 check for GF_OMP_FOR_KIND_FOR.
3556 2013-03-18 Kirill Yukhin <kirill.yukhin@intel.com>
3558 * config/i386/i386.h (ADDITIONAL_REGISTER_NAMES): Add
3559 ymm and zmm register names.
3561 2014-03-17 Jakub Jelinek <jakub@redhat.com>
3564 * config/i386/i386.c (ix86_expand_epilogue): Adjust REG_CFA_ADJUST_CFA
3565 note creation for the 2010-08-31 changes.
3567 2014-03-17 Marek Polacek <polacek@redhat.com>
3570 * omp-low.c (omp_max_vf): Treat -fno-tree-loop-optimize the same
3571 as -fno-tree-loop-vectorize.
3572 (expand_omp_simd): Likewise.
3574 2014-03-15 Eric Botcazou <ebotcazou@adacore.com>
3576 * config/sparc/sparc-protos.h (tls_call_delay): Delete.
3577 (eligible_for_call_delay): New prototype.
3578 * config/sparc/sparc.c (tls_call_delay): Rename into...
3579 (eligible_for_call_delay): ...this. Return false if the instruction
3580 cannot be put in the delay slot of a branch.
3581 (eligible_for_restore_insn): Simplify.
3582 (eligible_for_return_delay): Return false if the instruction cannot be
3583 put in the delay slot of a branch and simplify.
3584 (eligible_for_sibcall_delay): Return false if the instruction cannot be
3585 put in the delay slot of a branch.
3586 * config/sparc/sparc.md (fix_ut699): New attribute.
3587 (tls_call_delay): Delete.
3588 (in_call_delay): Reimplement.
3589 (eligible_for_sibcall_delay): Rename into...
3590 (in_sibcall_delay): ...this.
3591 (eligible_for_return_delay): Rename into...
3592 (in_return_delay): ...this.
3593 (in_branch_delay): Reimplement.
3594 (in_uncond_branch_delay): Delete.
3595 (in_annul_branch_delay): Delete.
3597 2014-03-14 Richard Henderson <rth@redhat.com>
3600 * config/i386/i386.md (floathi<X87MODEF>2): Delete expander; rename
3601 define_insn from *floathi<X87MODEF>2_i387; allow nonimmediate_operand.
3602 (*floathi<X87MODEF>2_i387_with_temp): Remove.
3603 (floathi splitters): Remove.
3604 (float<SWI48x>xf2): New pattern.
3605 (float<SWI48><MODEF>2): Rename from float<SWI48x><X87MODEF>2. Drop
3606 code that tried to handle DImode for 32-bit, but which was excluded
3607 by the pattern's condition. Drop allocation of stack temporary.
3608 (*floatsi<MODEF>2_vector_mixed_with_temp): Remove.
3609 (*float<SWI48><MODEF>2_mixed_with_temp): Remove.
3610 (*float<SWI48><MODEF>2_mixed_interunit): Remove.
3611 (*float<SWI48><MODEF>2_mixed_nointerunit): Remove.
3612 (*floatsi<MODEF>2_vector_sse_with_temp): Remove.
3613 (*float<SWI48><MODEF>2_sse_with_temp): Remove.
3614 (*float<SWI48><MODEF>2_sse_interunit): Remove.
3615 (*float<SWI48><MODEF>2_sse_nointerunit): Remove.
3616 (*float<SWI48x><X87MODEF>2_i387_with_temp): Remove.
3617 (*float<SWI48x><X87MODEF>2_i387): Remove.
3618 (all float _with_temp splitters): Remove.
3619 (*float<SWI48x><MODEF>2_i387): New pattern.
3620 (*float<SWI48><MODEF>2_sse): New pattern.
3621 (float TARGET_USE_VECTOR_CONVERTS splitters): Merge them.
3622 (float TARGET_SSE_PARTIAL_REG_DEPENDENCY splitters): Merge them.
3624 2014-03-14 Jakub Jelinek <jakub@redhat.com>
3625 Marek Polacek <polacek@redhat.com>
3628 * common.opt (dump_base_name_prefixed): New Variable.
3629 * opts.c (finish_options): Don't prepend directory to x_dump_base_name
3630 if x_dump_base_name_prefixed is already set, set it at the end.
3632 2014-03-14 Vladimir Makarov <vmakarov@redhat.com>
3634 PR rtl-optimization/60508
3635 * lra-constraints.c (get_reload_reg): Add new parameter
3637 (process_addr_reg, simplify_operand_subreg, curr_insn_transform):
3638 Pass the new parameter values.
3640 2014-03-14 Richard Biener <rguenther@suse.de>
3642 * common.opt: Revert unintented changes from r205065.
3645 2014-03-14 Richard Biener <rguenther@suse.de>
3648 * cfghooks.c (split_block): Properly adjust all loops the
3649 block was a latch of.
3651 2014-03-14 Martin Jambor <mjambor@suse.cz>
3654 * ipa-prop.c (ipa_modify_call_arguments): Fix iteration condition
3657 2014-03-14 Georg-Johann Lay <avr@gjlay.de>
3660 * config/avr/avr.c (avr_set_current_function): Pass function name
3661 through default_strip_name_encoding before sanity checking instead
3662 of skipping the first char of the assembler name.
3664 2014-03-13 Richard Henderson <rth@redhat.com>
3667 * config/i386/i386.c (ix86_split_fp_branch): Remove pushed argument.
3668 (ix86_force_to_memory, ix86_free_from_memory): Remove.
3669 * config/i386/i386-protos.h: Likewise.
3670 * config/i386/i386.md (floathi<X87MODEF>2): Use assign_386_stack_local
3671 in the expander instead of a splitter.
3672 (float<SWI48x><X87MODEF>2): Use assign_386_stack_local if there is
3673 any possibility of requiring a memory.
3674 (*floatsi<MODEF>2_vector_mixed): Remove, and the splitters.
3675 (*floatsi<MODEF>2_vector_sse): Remove, and the splitters.
3676 (fp branch splitters): Update for ix86_split_fp_branch.
3677 (*jcc<X87MODEF>_<SWI24>_i387): Remove r/f alternative.
3678 (*jcc<X87MODEF>_<SWI24>_r_i387): Likewise.
3679 (splitter for jcc<X87MODEF>_<SWI24>_i387 r/f): Remove.
3680 (*fop_<MODEF>_2_i387): Remove f/r alternative.
3681 (*fop_<MODEF>_3_i387): Likewise.
3682 (*fop_xf_2_i387, *fop_xf_3_i387): Likewise.
3683 (splitters for the fop_* register patterns): Remove.
3684 (fscalexf4_i387): Rename from *fscalexf4_i387.
3685 (ldexpxf3): Use gen_floatsixf2 and gen_fscalexf4_i387.
3687 2014-03-13 Jakub Jelinek <jakub@redhat.com>
3689 PR tree-optimization/59779
3690 * tree-dfa.c (get_ref_base_and_extent): Use double_int
3691 type for bitsize and maxsize instead of HOST_WIDE_INT.
3693 2014-03-13 Steven Bosscher <steven@gcc.gnu.org>
3695 PR rtl-optimization/57320
3696 * function.c (rest_of_handle_thread_prologue_and_epilogue): Cleanup
3697 the CFG after thread_prologue_and_epilogue_insns.
3699 2014-03-13 Vladimir Makarov <vmakarov@redhat.com>
3701 PR rtl-optimization/57189
3702 * lra-constraints.c (process_alt_operands): Disfavor spilling
3705 2014-03-13 Cesar Philippidis <cesar@codesourcery.com>
3707 * lto-wrapper.c (maybe_unlink_file): Suppress diagnostic messages.
3709 2014-03-13 Jakub Jelinek <jakub@redhat.com>
3711 PR tree-optimization/59025
3713 * tree-ssa-reassoc.c (sort_by_operand_rank): For SSA_NAMEs with the
3714 same rank, sort by bb_rank and gimple_uid of SSA_NAME_DEF_STMT first.
3716 2014-03-13 Georg-Johann Lay <avr@gjlay.de>
3719 * config/avr/avr.c (avr_out_plus): Swap cc_plus and cc_minus in
3720 calls of avr_out_plus_1.
3722 2014-03-13 Bin Cheng <bin.cheng@arm.com>
3724 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Record
3725 BB's single pred and update the father loop's latch info later.
3727 2014-03-12 Michael Meissner <meissner@linux.vnet.ibm.com>
3729 * config/rs6000/vector.md (VEC_L): Add V1TI mode to vector types.
3733 (VEC_base): Likewise.
3734 (mov<MODE>, VEC_M modes): If we are loading TImode into VSX
3735 registers, we need to swap double words in little endian mode.
3737 * config/rs6000/rs6000-modes.def (V1TImode): Add new vector mode
3738 to be a container mode for 128-bit integer operations added in ISA
3739 2.07. Unlike TImode and PTImode, the preferred register set is
3740 the Altivec/VMX registers for the 128-bit operations.
3742 * config/rs6000/rs6000-protos.h (rs6000_move_128bit_ok_p): Add
3744 (rs6000_split_128bit_ok_p): Likewise.
3746 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_3): Add new support
3747 macros for creating ISA 2.07 normal and overloaded builtin
3748 functions with 3 arguments.
3749 (BU_P8V_OVERLOAD_3): Likewise.
3750 (VPERM_1T): Add support for V1TImode in 128-bit vector operations
3751 for use as overloaded functions.
3752 (VPERM_1TI_UNS): Likewise.
3753 (VSEL_1TI): Likewise.
3754 (VSEL_1TI_UNS): Likewise.
3755 (ST_INTERNAL_1ti): Likewise.
3756 (LD_INTERNAL_1ti): Likewise.
3757 (XXSEL_1TI): Likewise.
3758 (XXSEL_1TI_UNS): Likewise.
3759 (VPERM_1TI): Likewise.
3760 (VPERM_1TI_UNS): Likewise.
3761 (XXPERMDI_1TI): Likewise.
3762 (SET_1TI): Likewise.
3763 (LXVD2X_V1TI): Likewise.
3764 (STXVD2X_V1TI): Likewise.
3765 (VEC_INIT_V1TI): Likewise.
3766 (VEC_SET_V1TI): Likewise.
3767 (VEC_EXT_V1TI): Likewise.
3768 (EQV_V1TI): Likewise.
3769 (NAND_V1TI): Likewise.
3770 (ORC_V1TI): Likewise.
3771 (VADDCUQ): Add support for 128-bit integer arithmetic instructions
3772 added in ISA 2.07. Add both normal 'altivec' builtins, and the
3774 (VADDUQM): Likewise.
3775 (VSUBCUQ): Likewise.
3776 (VADDEUQM): Likewise.
3777 (VADDECUQ): Likewise.
3778 (VSUBEUQM): Likewise.
3779 (VSUBECUQ): Likewise.
3781 * config/rs6000/rs6000-c.c (__int128_type): New static to hold
3782 __int128_t and __uint128_t types.
3783 (__uint128_type): Likewise.
3784 (altivec_categorize_keyword): Add support for vector __int128_t,
3785 vector __uint128_t, vector __int128, and vector unsigned __int128
3786 as a container type for TImode operations that need to be done in
3787 VSX/Altivec registers.
3788 (rs6000_macro_to_expand): Likewise.
3789 (altivec_overloaded_builtins): Add ISA 2.07 overloaded functions
3790 to support 128-bit integer instructions vaddcuq, vadduqm,
3791 vaddecuq, vaddeuqm, vsubcuq, vsubuqm, vsubecuq, vsubeuqm.
3792 (altivec_resolve_overloaded_builtin): Add support for V1TImode.
3794 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Add support
3795 for V1TImode, and set up preferences to use VSX/Altivec registers.
3796 Setup VSX reload handlers.
3797 (rs6000_debug_reg_global): Likewise.
3798 (rs6000_init_hard_regno_mode_ok): Likewise.
3799 (rs6000_preferred_simd_mode): Likewise.
3800 (vspltis_constant): Do not allow V1TImode as easy altivec constants.
3801 (easy_altivec_constant): Likewise.
3802 (output_vec_const_move): Likewise.
3803 (rs6000_expand_vector_set): Convert V1TImode set and extract to
3805 (rs6000_expand_vector_extract): Likewise.
3806 (reg_offset_addressing_ok_p): Setup V1TImode to use VSX reg+reg
3808 (rs6000_const_vec): Add support for V1TImode.
3809 (rs6000_emit_le_vsx_load): Swap double words when loading or
3810 storing TImode/V1TImode.
3811 (rs6000_emit_le_vsx_store): Likewise.
3812 (rs6000_emit_le_vsx_move): Likewise.
3813 (rs6000_emit_move): Add support for V1TImode.
3814 (altivec_expand_ld_builtin): Likewise.
3815 (altivec_expand_st_builtin): Likewise.
3816 (altivec_expand_vec_init_builtin): Likewise.
3817 (altivec_expand_builtin): Likewise.
3818 (rs6000_init_builtins): Add support for V1TImode type. Add
3819 support for ISA 2.07 128-bit integer builtins. Define type names
3820 for the VSX/Altivec vector types.
3821 (altivec_init_builtins): Add support for overloaded vector
3822 functions with V1TImode type.
3823 (rs6000_preferred_reload_class): Prefer Altivec registers for V1TImode.
3824 (rs6000_move_128bit_ok_p): Move 128-bit move/split validation to
3826 (rs6000_split_128bit_ok_p): Likewise.
3827 (rs6000_handle_altivec_attribute): Create V1TImode from vector
3828 __int128_t and vector __uint128_t.
3830 * config/rs6000/vsx.md (VSX_L): Add V1TImode to vector iterators
3831 and mode attributes.
3838 (VS_scalar): Likewise.
3839 (VS_double): Likewise.
3840 (vsx_set_v1ti): New builtin function to create V1TImode from TImode.
3842 * config/rs6000/rs6000.h (TARGET_VADDUQM): New macro to say whether
3843 we support the ISA 2.07 128-bit integer arithmetic instructions.
3844 (ALTIVEC_OR_VSX_VECTOR_MODE): Add V1TImode.
3845 (enum rs6000_builtin_type_index): Add fields to hold V1TImode
3846 and TImode types for use with the builtin functions.
3847 (V1TI_type_node): Likewise.
3848 (unsigned_V1TI_type_node): Likewise.
3849 (intTI_type_internal_node): Likewise.
3850 (uintTI_type_internal_node): Likewise.
3852 * config/rs6000/altivec.md (UNSPEC_VADDCUQ): New unspecs for ISA 2.07
3853 128-bit builtin functions.
3854 (UNSPEC_VADDEUQM): Likewise.
3855 (UNSPEC_VADDECUQ): Likewise.
3856 (UNSPEC_VSUBCUQ): Likewise.
3857 (UNSPEC_VSUBEUQM): Likewise.
3858 (UNSPEC_VSUBECUQ): Likewise.
3859 (VM): Add V1TImode to vector mode iterators.
3861 (VI_unit): Likewise.
3862 (altivec_vadduqm): Add ISA 2.07 128-bit binary builtins.
3863 (altivec_vaddcuq): Likewise.
3864 (altivec_vsubuqm): Likewise.
3865 (altivec_vsubcuq): Likewise.
3866 (altivec_vaddeuqm): Likewise.
3867 (altivec_vaddecuq): Likewise.
3868 (altivec_vsubeuqm): Likewise.
3869 (altivec_vsubecuq): Likewise.
3871 * config/rs6000/rs6000.md (FMOVE128_GPR): Add V1TImode to vector
3873 (BOOL_128): Likewise.
3874 (BOOL_REGS_OUTPUT): Likewise.
3875 (BOOL_REGS_OP1): Likewise.
3876 (BOOL_REGS_OP2): Likewise.
3877 (BOOL_REGS_UNARY): Likewise.
3878 (BOOL_REGS_AND_CR0): Likewise.
3880 * config/rs6000/altivec.h (vec_vaddcuq): Add support for ISA 2.07
3881 128-bit integer builtin support.
3882 (vec_vadduqm): Likewise.
3883 (vec_vaddecuq): Likewise.
3884 (vec_vaddeuqm): Likewise.
3885 (vec_vsubecuq): Likewise.
3886 (vec_vsubeuqm): Likewise.
3887 (vec_vsubcuq): Likewise.
3888 (vec_vsubuqm): Likewise.
3890 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
3891 Document vec_vaddcuq, vec_vadduqm, vec_vaddecuq, vec_vaddeuqm,
3892 vec_subecuq, vec_subeuqm, vec_vsubcuq, vec_vsubeqm builtins adding
3893 128-bit integer add/subtract to ISA 2.07.
3895 2014-03-12 Joern Rennecke <joern.rennecke@embecosm.com>
3897 * config/arc/arc.c (arc_predicate_delay_insns):
3898 Fix third argument passed to conditionalize_nonjump.
3900 2014-03-12 Yufeng Zhang <yufeng.zhang@arm.com>
3902 * config/aarch64/aarch64-builtins.c
3903 (aarch64_builtin_vectorized_function): Add BUILT_IN_LFLOORF,
3904 BUILT_IN_LLFLOOR, BUILT_IN_LCEILF and BUILT_IN_LLCEIL.
3905 * config/aarch64/arm_neon.h (vcvtaq_u64_f64): Call __builtin_llfloor
3906 instead of __builtin_lfloor.
3907 (vcvtnq_u64_f64): Call __builtin_llceil instead of __builtin_lceil.
3909 2014-03-12 Jakub Jelinek <jakub@redhat.com>
3911 * tree-ssa-ifcombine.c (forwarder_block_to): New function.
3912 (tree_ssa_ifcombine_bb_1): New function.
3913 (tree_ssa_ifcombine_bb): Use it. Handle also cases where else_bb
3914 is an empty forwarder block to then_bb or vice versa and then_bb
3915 and else_bb are effectively swapped.
3917 2014-03-12 Christian Bruel <christian.bruel@st.com>
3920 * config/arm/arm.c (arm_emit_vfp_multi_reg_pop): Emit a
3921 REG_CFA_DEF_CFA note.
3922 (arm_expand_epilogue_apcs_frame): call arm_add_cfa_adjust_cfa_note.
3923 (arm_unwind_emit): Allow REG_CFA_DEF_CFA.
3925 2014-03-12 Thomas Preud'homme <thomas.preudhomme@arm.com>
3927 PR tree-optimization/60454
3928 * tree-ssa-math-opts.c (find_bswap_1): Fix bswap detection.
3930 2014-03-12 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3932 * config.gcc (aarch64*-*-*): Use ISA flags from aarch64-arches.def.
3933 Do not define target_cpu_default2 to generic.
3934 * config/aarch64/aarch64.h (TARGET_CPU_DEFAULT): Use generic cpu.
3935 * config/aarch64/aarch64.c (aarch64_override_options): Update comment.
3936 * config/aarch64/aarch64-arches.def (armv8-a): Use generic cpu.
3938 2014-03-12 Jakub Jelinek <jakub@redhat.com>
3939 Marc Glisse <marc.glisse@inria.fr>
3941 PR tree-optimization/60502
3942 * tree-ssa-reassoc.c (eliminate_not_pairs): Use build_all_ones_cst
3943 instead of build_low_bits_mask.
3945 2014-03-12 Jakub Jelinek <jakub@redhat.com>
3948 * tree-vrp.c (register_edge_assert_for_1): Don't add assert
3949 if there are multiple uses, but op doesn't live on E edge.
3950 * tree-cfg.c (assert_unreachable_fallthru_edge_p): Also ignore
3951 clobber stmts before __builtin_unreachable.
3953 2014-03-11 Richard Sandiford <rdsandiford@googlemail.com>
3955 * builtins.c (expand_builtin_setjmp_receiver): Use and clobber
3956 hard_frame_pointer_rtx.
3957 * cse.c (cse_insn): Remove volatile check.
3958 * cselib.c (cselib_process_insn): Likewise.
3959 * dse.c (scan_insn): Likewise.
3961 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
3963 * config/arc/arc.c (conditionalize_nonjump): New function,
3965 (arc_ifcvt): ... this.
3966 (arc_predicate_delay_insns): Use it.
3968 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
3970 * config/arc/predicates.md (extend_operand): During/after reload,
3971 allow const_int_operand.
3972 * config/arc/arc.md (mulsidi3_700): Use extend_operand predicate.
3973 (umulsidi3_700): Likewise. Change operand 2 constraint back to "cL".
3974 (mulsi3_highpart): Change operand 2 constraint alternatives 2 and 3
3976 (umulsi3_highpart_i): Likewise.
3978 2014-03-11 Richard Biener <rguenther@suse.de>
3980 * tree-ssa-structalias.c (get_constraint_for_ptr_offset):
3981 Add asserts to guard possible wrong-code bugs.
3983 2014-03-11 Richard Biener <rguenther@suse.de>
3985 PR tree-optimization/60429
3986 PR tree-optimization/60485
3987 * tree-ssa-structalias.c (set_union_with_increment): Properly
3988 take into account all fields that overlap the shifted vars.
3989 (do_sd_constraint): Likewise.
3990 (do_ds_constraint): Likewise.
3991 (get_constraint_for_ptr_offset): Likewise.
3993 2014-03-11 Chung-Lin Tang <cltang@codesourcery.com>
3995 * config/nios2/nios2.c (machine_function): Add fp_save_offset field.
3996 (nios2_compute_frame_layout):
3997 Add calculation of cfun->machine->fp_save_offset.
3998 (nios2_expand_prologue): Correct setting of frame pointer register
4000 (nios2_expand_epilogue): Update recovery of stack pointer from
4001 frame pointer accordingly.
4002 (nios2_initial_elimination_offset): Update calculation of offset
4003 for eliminating to HARD_FRAME_POINTER_REGNUM.
4005 2014-03-10 Jakub Jelinek <jakub@redhat.com>
4008 * ipa.c (symtab_remove_unreachable_nodes): Don't call
4009 cgraph_get_create_node on VAR_DECLs.
4011 2014-03-10 Richard Biener <rguenther@suse.de>
4014 * tree.c (signed_or_unsigned_type_for): Handle OFFSET_TYPEs.
4016 2014-03-08 Douglas B Rupp <rupp@gnat.com>
4018 * config/vms/vms.opt (vms_float_format): New variable.
4020 2014-03-08 Tobias Burnus <burnus@net-b.de>
4022 * doc/invoke.texi (-fcilkplus): Update implementation status.
4024 2014-03-08 Paulo Matos <paulo@matos-sorge.com>
4025 Richard Biener <rguenther@suse.de>
4027 * lto-wrapper.c (merge_and_complain): Ensure -fshort-double is used
4028 consistently accross all TUs.
4029 (run_gcc): Enable -fshort-double automatically at link at link-time
4030 and disallow override.
4032 2014-03-08 Richard Sandiford <rdsandiford@googlemail.com>
4035 * config/mips/mips.c (mips_option_override): Promote -mpaired-single
4036 warning to an error. Disable TARGET_PAIRED_SINGLE and TARGET_MIPS3D
4037 if they can't be used.
4039 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4041 * configure.ac (HAVE_AS_IX86_TLSLDMPLT): Improve test
4042 for Solaris 11/x86 ld.
4043 * configure: Regenerate.
4045 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4047 * configure.ac (TLS_SECTION_ASM_FLAG): Save as tls_section_flag.
4048 (LIB_TLS_SPEC): Save as ld_tls_libs.
4049 (HAVE_AS_IX86_TLSLDMPLT): Define as 1/0.
4050 (HAVE_AS_IX86_TLSLDM): New test.
4051 * configure, config.in: Regenerate.
4052 * config/i386/i386.c (legitimize_tls_address): Fall back to
4053 TLS_MODEL_GLOBAL_DYNAMIC on 32-bit Solaris/x86 if tool chain
4054 cannot support TLS_MODEL_LOCAL_DYNAMIC.
4055 * config/i386/i386.md (*tls_local_dynamic_base_32_gnu): Use if
4056 instead of #ifdef in HAVE_AS_IX86_TLSLDMPLT test.
4058 2014-03-07 Paulo Matos <paulo@matos-sorge.com>
4060 * common.opt (fira-loop-pressure): Mark as optimization.
4062 2014-03-07 Thomas Schwinge <thomas@codesourcery.com>
4064 * langhooks.c (lhd_omp_mappable_type): The error_mark_node is not
4065 an OpenMP mappable type.
4067 2014-03-06 Matthias Klose <doko@ubuntu.com>
4069 * Makefile.in (s-mlib): Only pass MULTIARCH_DIRNAME if
4070 MULTILIB_OSDIRNAMES is not defined.
4072 2014-03-06 Jakub Jelinek <jakub@redhat.com>
4073 Meador Inge <meadori@codesourcery.com>
4076 * config/arm/arm.c (arm_tls_symbol_p): Remove.
4077 (arm_legitimize_address): Call legitimize_tls_address for any
4078 arm_tls_referenced_p expression, handle constant addend. Call it
4079 before testing for !TARGET_ARM.
4080 (thumb_legitimize_address): Don't handle arm_tls_symbol_p here.
4082 2014-03-06 Richard Biener <rguenther@suse.de>
4088 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
4090 * tree-streamer.c (record_common_node): Assert we don't record
4091 nodes with type double.
4092 (preload_common_node): Skip type double, complex double and double
4093 pointer since it is now frontend dependent due to fshort-double option.
4095 2014-03-06 Richard Biener <rguenther@suse.de>
4097 * gcc.c (PLUGIN_COND): Always enable unless -fno-use-linker-plugin
4098 or -fno-lto is specified and the linker has full plugin support.
4099 * collect2.c (lto_mode): Default to LTO_MODE_WHOPR if LTO is enabled.
4100 (main): Remove -flto processing, adjust lto_mode using use_plugin late.
4101 * lto-wrapper.c (merge_and_complain): Merge compile-time
4102 optimization levels.
4103 (run_gcc): And pass it through to the link options.
4105 2014-03-06 Alexandre Oliva <aoliva@redhat.com>
4109 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4111 * cselib.c (remove_useless_values): Skip to avoid quadratic
4112 behavior if the condition moved from...
4113 (cselib_process_insn): ... here holds.
4115 2014-03-05 Jakub Jelinek <jakub@redhat.com>
4118 * Makefile.in (PLUGIN_HEADERS): Add tree-phinodes.h, stor-layout.h,
4119 ssa-iterators.h, $(RESOURCE_H) and tree-cfgcleanup.h.
4122 * config/i386/t-i386 (OPTIONS_H_EXTRA): Add stringop.def.
4123 (TM_H): Add x86-tune.def.
4125 2014-03-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4127 * config/aarch64/aarch64.c (generic_tunings):
4128 Use cortexa57_extra_costs.
4130 2014-03-05 Jakub Jelinek <jakub@redhat.com>
4133 * cfgexpand.c (expand_used_vars): Do not assume all SSA_NAMEs
4134 of PARM/RESULT_DECLs must be coalesced with optimize && in_lto_p.
4135 * tree-ssa-coalesce.c (coalesce_ssa_name): Use MUST_COALESCE_COST - 1
4138 2014-03-04 Heiher <r@hev.cc>
4140 * config/mips/mips-cpus.def (loongson3a): Mark as a MIPS64r2 processor.
4141 * config/mips/mips.h (MIPS_ISA_LEVEL_SPEC): Adjust accordingly.
4143 2014-03-04 Uros Bizjak <ubizjak@gmail.com>
4145 * config/i386/predicates.md (const2356_operand): Change to ...
4146 (const2367_operand): ... this.
4147 * config/i386/sse.md (avx512pf_scatterpf<mode>sf): Use
4149 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
4150 (*avx512pf_scatterpf<mode>sf): Ditto.
4151 (avx512pf_scatterpf<mode>df): Ditto.
4152 (*avx512pf_scatterpf<mode>df_mask): Ditto.
4153 (*avx512pf_scatterpf<mode>df): Ditto.
4154 * config/i386/i386.c (ix86_expand_builtin): Update
4155 incorrect hint operand error message.
4157 2014-03-04 Richard Biener <rguenther@suse.de>
4159 * lto-section-in.c (lto_get_section_data): Fix const cast.
4161 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
4163 * tree-streamer.c (record_common_node): Assert we don't record
4164 nodes with type double.
4165 (preload_common_node): Skip type double, complex double and double
4166 pointer since it is now frontend dependent due to fshort-double option.
4168 2014-03-04 Richard Biener <rguenther@suse.de>
4171 * lto-streamer-in.c (lto_read_body): Remove LTO bytecode version check.
4172 (lto_input_toplevel_asms): Likewise.
4173 * lto-section-in.c (lto_get_section_data): Instead do it here
4176 2014-03-04 Richard Biener <rguenther@suse.de>
4178 PR tree-optimization/60382
4179 * tree-vect-loop.c (vect_is_simple_reduction_1): Do not consider
4180 dead PHIs a reduction.
4182 2014-03-03 Uros Bizjak <ubizjak@gmail.com>
4184 * config/i386/xmmintrin.h (enum _mm_hint) <_MM_HINT_ET0>: Correct
4186 (_mm_prefetch): Move out of GCC target("sse") pragma.
4187 * config/i386/prfchwintrin.h (_m_prefetchw): Move out of
4188 GCC target("prfchw") pragma.
4189 * config/i386/i386.md (prefetch): Emit prefetchwt1 only
4191 * config/i386/i386.c (ix86_option_override_internal): Enable
4192 -mprfchw with -mprefetchwt1.
4194 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
4196 * config/arc/arc.md (casesi_load) <length attribute alternative 0>:
4199 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
4201 * opts.h (CL_PCH_IGNORE): Define.
4202 * targhooks.c (option_affects_pch_p):
4203 Return false for options that have CL_PCH_IGNORE set.
4204 * opt-functions.awk: Process PchIgnore.
4205 * doc/options.texi: Document PchIgnore.
4207 * config/arc/arc.opt (misize): Add PchIgnore property.
4209 2014-03-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4211 * config/rs6000/rs6000.c (rs6000_preferred_reload_class): Disallow
4212 reload of PLUS rtx's outside of GENERAL_REGS or BASE_REGS; relax
4213 constraint on constants to permit them being loaded into
4214 GENERAL_REGS or BASE_REGS.
4216 2014-03-03 Nick Clifton <nickc@redhat.com>
4218 * config/rl78/rl78-real.md (cbranchsi4_real_signed): Add
4219 anti-cacnonical alternatives.
4220 (negandhi3_real): New pattern.
4221 * config/rl78/rl78-virt.md (negandhi3_virt): New pattern.
4223 2014-03-03 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
4225 * config/avr/avr-mcus.def: Remove atxmega16x1.
4226 * config/avr/avr-tables.opt: Regenerate.
4227 * config/avr/t-multilib: Regenerate.
4228 * doc/avr-mmcu.texi: Regenerate.
4230 2014-03-03 Tobias Grosser <tobias@grosser.es>
4231 Mircea Namolaru <mircea.namolaru@inria.fr>
4233 PR tree-optimization/58028
4234 * graphite-clast-to-gimple.c (set_cloog_options): Don't remove
4237 2014-03-03 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4239 * config/arm/neon.md (*movmisalign<mode>): Legitimize addresses
4240 not handled by recognizers.
4242 2014-03-03 Jakub Jelinek <jakub@redhat.com>
4245 * function.c (expand_function_end): Don't emit
4246 clobber_return_register sequence if clobber_after is a BARRIER.
4247 * cfgexpand.c (construct_exit_block): Append instructions before
4248 return_label to prev_bb.
4250 2014-03-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4252 * config/rs6000/constraints.md: Document reserved use of "wc".
4254 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
4257 * ipa.c (function_and_variable_visibility): When dissolving comdat
4258 group, also set all symbols to local.
4260 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
4265 2013-12-14 Jan Hubicka <jh@suse.cz>
4267 * ipa-prop.c (stmt_may_be_vtbl_ptr_store): Skip clobbers.
4269 2014-03-02 Jon Beniston <jon@beniston.com>
4275 * config/lm32/lm32.c (lm32_legitimate_constant_p): Remove, as incorrect.
4276 (TARGET_LEGITIMATE_CONSTANT_P): Undefine, as not needed.
4277 * config/lm32/lm32.md (movsi_insn): Add 32-bit immediate support.
4278 (simple_return, *simple_return): New patterns
4279 * config/lm32/predicates.md (movsi_rhs_operand): Remove as obsolete.
4280 * configure.ac (force_sjlj_exceptions): Force sjlj exceptions for lm32.
4282 2014-03-01 Paolo Carlini <paolo.carlini@oracle.com>
4284 * dwarf2out.c (gen_subprogram_die): Tidy.
4286 2014-03-01 Oleg Endo <olegendo@gcc.gnu.org>
4289 * config/sh/sh.md (*mov_t_msb_neg): Split into ...
4290 (*mov_t_msb_neg_negc): ... this new insn.
4292 2014-02-28 Jason Merrill <jason@redhat.com>
4295 * ipa-devirt.c (ipa_devirt): Don't choose an implicitly-declared
4298 2014-02-28 Paolo Carlini <paolo.carlini@oracle.com>
4301 * dwarf2out.c (decltype_auto_die): New static.
4302 (gen_subprogram_die): Handle 'decltype(auto)' like 'auto'.
4303 (gen_type_die_with_usage): Handle 'decltype(auto)'.
4304 (is_cxx_auto): Likewise.
4306 2014-02-28 Ian Bolton <ian.bolton@arm.com>
4308 * config/aarch64/aarch64.h: Define __ARM_NEON by default if
4309 we are not using general regs only.
4311 2014-02-28 Richard Biener <rguenther@suse.de>
4314 * tree-cfgcleanup.c (tree_forwarder_block_p): Restrict
4315 previous fix and only allow to remove trivial pre-headers
4316 and latches. Also honor LOOPS_MAY_HAVE_MULTIPLE_LATCHES.
4317 (remove_forwarder_block): Properly update the latch of a loop.
4319 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4322 * cselib.c (cselib_hasher::equal): Special-case VALUE lookup.
4323 (cselib_preserved_hash_table): New.
4324 (preserve_constants_and_equivs): Move preserved vals to it.
4325 (cselib_find_slot): Look it up first.
4326 (cselib_init): Initialize it.
4327 (cselib_finish): Release it.
4328 (dump_cselib_table): Dump it.
4330 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4333 * cselib.c (remove_useless_values): Skip to avoid quadratic
4334 behavior if the condition moved from...
4335 (cselib_process_insn): ... here holds.
4337 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4340 * var-tracking.c (vt_initialize): Apply the same condition to
4341 preserve the CFA base value.
4343 2014-02-28 Joey Ye <joey.ye@arm.com>
4346 * config/arm/arm.c (thumb_far_jump_used_p): Don't change
4347 if reload in progress or completed.
4349 2014-02-28 Tobias Burnus <burnus@net-b.de>
4352 * tree-pretty-print.c (dump_generic_node, print_declaration): Handle
4355 2014-02-27 H.J. Lu <hongjiu.lu@intel.com>
4357 * doc/tm.texi.in (Condition Code Status): Update documention for
4358 relative locations of cc0-setter and cc0-user.
4360 2014-02-27 Jeff Law <law@redhat.com>
4362 PR rtl-optimization/52714
4363 * combine.c (try_combine): When splitting an unrecognized PARALLEL
4364 into two independent simple sets, if I3 is a jump, ensure the
4365 pattern we place into I3 is a (set (pc) ...).
4367 2014-02-27 Mikael Pettersson <mikpe@it.uu.se>
4368 Jeff Law <law@redhat.com>
4370 PR rtl-optimization/49847
4371 * cse.c (fold_rtx) Handle case where cc0 setter and cc0 user
4372 are in different blocks.
4373 * doc/tm.texi (Condition Code Status): Update documention for
4374 relative locations of cc0-setter and cc0-user.
4376 2014-02-27 Vladimir Makarov <vmakarov@redhat.com>
4379 * lra.c (lra_emit_add): Check SUBREG too.
4381 2014-02-27 Andreas Schwab <schwab@suse.de>
4383 * config/m68k/m68k.c (m68k_option_override): Disable
4384 -flive-range-shrinkage for classic m68k.
4385 (m68k_override_options_after_change): Likewise.
4387 2014-02-27 Marek Polacek <polacek@redhat.com>
4390 * tree-ssa-uninit.c (gate_warn_uninitialized): Run the pass even for
4391 -Wmaybe-uninitialized.
4393 2014-02-27 Alan Modra <amodra@gmail.com>
4396 * reload1.c (emit_input_reload_insns): When reload_override_in,
4397 set old to rl->in_reg when rl->in_reg is a subreg.
4399 2014-02-26 Richard Biener <rguenther@suse.de>
4402 * lra-assigns.c (spill_for): Avoid mixed-sign comparison.
4404 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
4406 * common/config/i386/predicates.md (const1256_operand): Remove.
4407 (const2356_operand): New.
4408 (const_1_to_2_operand): Remove.
4409 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
4410 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
4411 (*avx512pf_gatherpf<mode>sf): Ditto.
4412 (avx512pf_gatherpf<mode>df): Ditto.
4413 (*avx512pf_gatherpf<mode>df_mask): Ditto.
4414 (*avx512pf_gatherpf<mode>df): Ditto.
4415 (avx512pf_scatterpf<mode>sf): Ditto.
4416 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
4417 (*avx512pf_scatterpf<mode>sf): Ditto.
4418 (avx512pf_scatterpf<mode>df): Ditto.
4419 (*avx512pf_scatterpf<mode>df_mask): Ditto.
4420 (*avx512pf_scatterpf<mode>df): Ditto.
4421 * common/config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET0.
4423 2014-02-26 Ilya Tocar <ilya.tocar@intel.com>
4425 * config/i386/avx512fintrin.h (_mm512_testn_epi32_mask),
4426 (_mm512_mask_testn_epi32_mask), (_mm512_testn_epi64_mask),
4427 (_mm512_mask_testn_epi64_mask): Move to ...
4428 * config/i386/avx512cdintrin.h: Here.
4429 * config/i386/i386.c (bdesc_args): Change MASK_ISA for testnm.
4430 * config/i386/sse.md (avx512f_vmscalef<mode><round_name>): Remove %.
4431 (avx512f_scalef<mode><mask_name><round_name>): Ditto.
4432 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Change conditon to
4433 TARGET_AVX512F from TARGET_AVX512CD.
4435 2014-02-26 Richard Biener <rguenther@suse.de>
4438 * ipa.c (walk_polymorphic_call_targets): Properly guard
4439 call to inline_update_overall_summary.
4441 2014-02-26 Bin Cheng <bin.cheng@arm.com>
4444 * tree-cfgcleanup.c (tree_forwarder_block_p): Protect loop preheaders
4445 and latches only if requested. Fix latch if it is removed.
4446 * tree-ssa-dom.c (tree_ssa_dominator_optimize): Set
4447 LOOPS_HAVE_PREHEADERS.
4449 2014-02-25 Andrew Pinski <apinski@cavium.com>
4451 * builtins.c (expand_builtin_thread_pointer): Create a new target
4452 when the target is NULL.
4454 2014-02-25 Vladimir Makarov <vmakarov@redhat.com>
4456 PR rtl-optimization/60317
4457 * params.def (PARAM_LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
4458 * params.h (LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
4459 * lra-assigns.c: Include params.h.
4460 (spill_for): Use LRA_MAX_CONSIDERED_RELOAD_PSEUDOS as guard for
4461 other reload pseudos considerations.
4463 2014-02-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4465 * config/rs6000/vector.md (*vector_unordered<mode>): Change split
4466 to use canonical form for nor<mode>3.
4468 2014-02-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4471 * config/arm/arm.h (CANNOT_CHANGE_MODE_CLASS): Allow 128 to 64-bit
4474 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
4476 * common/config/i386/i386-common.c (OPTION_MASK_ISA_PREFETCHWT1_SET),
4477 (OPTION_MASK_ISA_PREFETCHWT1_UNSET): New.
4478 (ix86_handle_option): Handle OPT_mprefetchwt1.
4479 * config/i386/cpuid.h (bit_PREFETCHWT1): New.
4480 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
4482 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
4483 OPTION_MASK_ISA_PREFETCHWT1.
4484 * config/i386/i386.c (ix86_target_string): Handle mprefetchwt1.
4485 (PTA_PREFETCHWT1): New.
4486 (ix86_option_override_internal): Handle PTA_PREFETCHWT1.
4487 (ix86_valid_target_attribute_inner_p): Handle OPT_mprefetchwt1.
4488 * config/i386/i386.h (TARGET_PREFETCHWT1, TARGET_PREFETCHWT1_P): New.
4489 * config/i386/i386.md (prefetch): Check TARGET_PREFETCHWT1
4490 (*prefetch_avx512pf_<mode>_: Change into ...
4491 (*prefetch_prefetchwt1_<mode>: This.
4492 * config/i386/i386.opt (mprefetchwt1): New.
4493 * config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET1.
4494 (_mm_prefetch): Handle intent to write.
4495 * doc/invoke.texi (mprefetchwt1), (mno-prefetchwt1): Doccument.
4497 2014-02-25 Richard Biener <rguenther@suse.de>
4500 * emit-rtl.c (mem_attrs_htab): Remove.
4501 (mem_attrs_htab_hash): Likewise.
4502 (mem_attrs_htab_eq): Likewise.
4503 (set_mem_attrs): Always allocate new mem-attrs when something changed.
4504 (init_emit_once): Do not allocate mem_attrs_htab.
4506 2014-02-25 Richard Biener <rguenther@suse.de>
4509 * lto-opts.c (lto_write_options): Output non-explicit conservative
4510 -fwrapv, -fno-trapv and -fno-strict-overflow.
4511 * lto-wrapper.c (merge_and_complain): Handle merging those options.
4512 (run_gcc): And pass them through.
4514 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
4516 * sel-sched.c (calculate_new_fences): New parameter ptime.
4517 Calculate it as a maximum over all fence cycles.
4518 (sel_sched_region_2): Adjust the call to calculate_new_fences.
4519 Print the final schedule timing when sched_verbose.
4521 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
4523 PR rtl-optimization/60292
4524 * sel-sched.c (fill_vec_av_set): Do not reset target availability
4525 bit fot the fence instruction.
4527 2014-02-24 Alangi Derick <alangiderick@gmail.com>
4529 * calls.h: Fix typo in comment.
4531 2014-02-24 John David Anglin <danglin@gcc.gnu.org>
4533 * config/pa/pa.c (pa_output_move_double): Don't valididate when
4534 adjusting offsetable addresses.
4536 2014-02-24 Guozhi Wei <carrot@google.com>
4538 * sparseset.h (sparseset_pop): Fix the wrong index.
4540 2014-02-24 Walter Lee <walt@tilera.com>
4542 * config.gcc (tilepro-*-*): Change to tilepro*-*-*.
4543 (tilegx-*-linux*): Change to tilegx*-*-linux*; Support tilegxbe
4545 * common/config/tilegx/tilegx-common.c
4546 (TARGET_DEFAULT_TARGET_FLAGS): Define.
4547 * config/tilegx/linux.h (ASM_SPEC): Add endian_spec.
4549 * config/tilegx/sync.md (atomic_test_and_set): Handle big endian.
4550 * config/tilegx/tilegx.c (tilegx_return_in_msb): New.
4551 (tilegx_gimplify_va_arg_expr): Handle big endian.
4552 (tilegx_expand_unaligned_load): Ditto.
4553 (tilegx_expand_unaligned_store): Ditto.
4554 (TARGET_RETURN_IN_MSB): New.
4555 * config/tilegx/tilegx.h (TARGET_DEFAULT): New.
4556 (TARGET_ENDIAN_DEFAULT): New.
4557 (TARGET_BIG_ENDIAN): Handle big endian.
4558 (BYTES_BIG_ENDIAN): Ditto.
4559 (WORDS_BIG_ENDIAN): Ditto.
4560 (FLOAT_WORDS_BIG_ENDIAN): Ditto.
4563 * config/tilegx/tilegx.md (extv): Handle big endian.
4565 (insn_st<n>): Ditto.
4566 (insn_st<n>_add<bitsuffix>): Ditto.
4567 (insn_stnt<n>): Ditto.
4568 (insn_stnt<n>_add<bitsuffix>):Ditto.
4569 (vec_interleave_highv8qi): Handle big endian.
4570 (vec_interleave_highv8qi_be): New.
4571 (vec_interleave_highv8qi_le): New.
4572 (insn_v1int_h): Handle big endian.
4573 (vec_interleave_lowv8qi): Handle big endian.
4574 (vec_interleave_lowv8qi_be): New.
4575 (vec_interleave_lowv8qi_le): New.
4576 (insn_v1int_l): Handle big endian.
4577 (vec_interleave_highv4hi): Handle big endian.
4578 (vec_interleave_highv4hi_be): New.
4579 (vec_interleave_highv4hi_le): New.
4580 (insn_v2int_h): Handle big endian.
4581 (vec_interleave_lowv4hi): Handle big endian.
4582 (vec_interleave_lowv4hi_be): New.
4583 (vec_interleave_lowv4hi_le): New.
4584 (insn_v2int_l): Handle big endian.
4585 (vec_interleave_highv2si): Handle big endian.
4586 (vec_interleave_highv2si_be): New.
4587 (vec_interleave_highv2si_le): New.
4588 (insn_v4int_h): Handle big endian.
4589 (vec_interleave_lowv2si): Handle big endian.
4590 (vec_interleave_lowv2si_be): New.
4591 (vec_interleave_lowv2si_le): New.
4592 (insn_v4int_l): Handle big endian.
4593 * config/tilegx/tilegx.opt (mbig-endian): New option.
4594 (mlittle-endian): New option.
4595 * doc/install.texi: Document tilegxbe-linux.
4596 * doc/invoke.texi: Document -mbig-endian and -mlittle-endian.
4598 2014-02-24 Martin Jambor <mjambor@suse.cz>
4601 * ipa-cp.c (propagate_constants_accross_call): Bail out early if
4602 there are no parameter descriptors.
4604 2014-02-24 Andrey Belevantsev <abel@ispras.ru>
4606 PR rtl-optimization/60268
4607 * sched-rgn.c (haifa_find_rgns): Move the nr_regions_initial variable
4608 initialization to ...
4609 (sched_rgn_init): ... here.
4610 (schedule_region): Check for SCHED_PRESSURE_NONE earlier.
4612 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
4614 * config/microblaze/microblaze.md: Correct ashrsi_reg / lshrsi_reg
4617 2014-02-23 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
4619 * config/microblaze/microblaze.h: Remove SECONDARY_MEMORY_NEEDED
4622 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
4624 * /config/microblaze/microblaze.c: Add microblaze_asm_output_mi_thunk,
4625 define TARGET_ASM_OUTPUT_MI_THUNK and TARGET_ASM_CAN_OUTPUT_MI_THUNK.
4627 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
4629 * config/microblaze/predicates.md: Add cmp_op predicate.
4630 * config/microblaze/microblaze.md: Add branch_compare instruction
4631 which uses cmp_op predicate and emits cmp insn before branch.
4632 * config/microblaze/microblaze.c (microblaze_emit_compare): Rename
4633 to microblaze_expand_conditional_branch and consolidate logic.
4634 (microblaze_expand_conditional_branch): emit branch_compare
4635 insn instead of handling cmp op separate from branch insn.
4637 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4639 * config/rs6000/rs6000.c (rs6000_emit_le_vsx_move): Relax assert
4642 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4644 * config/rs6000/altivec.md (altivec_lve<VI_char>x): Replace
4645 define_insn with define_expand and new define_insn
4646 *altivec_lve<VI_char>x_internal.
4647 (altivec_stve<VI_char>x): Replace define_insn with define_expand
4648 and new define_insn *altivec_stve<VI_char>x_internal.
4649 * config/rs6000/rs6000-protos.h (altivec_expand_stvex_be): New
4651 * config/rs6000/rs6000.c (altivec_expand_lvx_be): Document use by
4653 (altivec_expand_stvex_be): New function.
4655 2014-02-22 Joern Rennecke <joern.rennecke@embecosm.com>
4657 * config/avr/avr.c (avr_can_eliminate): Allow elimination from
4658 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM if !frame_pointer_needed.
4659 * config/avr/avr.c (ELIMINABLE_REGS): Add elimination from
4660 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM.
4662 2014-02-21 Vladimir Makarov <vmakarov@redhat.com>
4665 * lra-constraints.c (inherit_reload_reg): Use lra_emit_move
4666 instead of emit_move_insn.
4668 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4670 * config/rs6000/altivec.md (altivec_vsumsws): Replace second
4672 (reduc_uplus_v16qi): Use gen_altivec_vsumsws_direct instead of
4673 gen_altivec_vsumsws.
4675 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4677 * config/rs6000/altivec.md (altivec_lvxl): Rename as
4678 *altivec_lvxl_<mode>_internal and use VM2 iterator instead of V4SI.
4679 (altivec_lvxl_<mode>): New define_expand incorporating
4680 -maltivec=be semantics where needed.
4681 (altivec_lvx): Rename as *altivec_lvx_<mode>_internal.
4682 (altivec_lvx_<mode>): New define_expand incorporating -maltivec=be
4683 semantics where needed.
4684 (altivec_stvx): Rename as *altivec_stvx_<mode>_internal.
4685 (altivec_stvx_<mode>): New define_expand incorporating
4686 -maltivec=be semantics where needed.
4687 (altivec_stvxl): Rename as *altivec_stvxl_<mode>_internal and use
4688 VM2 iterator instead of V4SI.
4689 (altivec_stvxl_<mode>): New define_expand incorporating
4690 -maltivec=be semantics where needed.
4691 * config/rs6000/rs6000-builtin.def: Add new built-in definitions
4692 LVXL_V2DF, LVXL_V2DI, LVXL_V4SF, LVXL_V4SI, LVXL_V8HI, LVXL_V16QI,
4693 LVX_V2DF, LVX_V2DI, LVX_V4SF, LVX_V4SI, LVX_V8HI, LVX_V16QI, STVX_V2DF,
4694 STVX_V2DI, STVX_V4SF, STVX_V4SI, STVX_V8HI, STVX_V16QI, STVXL_V2DF,
4695 STVXL_V2DI, STVXL_V4SF, STVXL_V4SI, STVXL_V8HI, STVXL_V16QI.
4696 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Replace
4697 ALTIVEC_BUILTIN_LVX with ALTIVEC_BUILTIN_LVX_<MODE> throughout;
4698 similarly for ALTIVEC_BUILTIN_LVXL, ALTIVEC_BUILTIN_STVX, and
4699 ALTIVEC_BUILTIN_STVXL.
4700 * config/rs6000/rs6000-protos.h (altivec_expand_lvx_be): New prototype.
4701 (altivec_expand_stvx_be): Likewise.
4702 * config/rs6000/rs6000.c (swap_selector_for_mode): New function.
4703 (altivec_expand_lvx_be): Likewise.
4704 (altivec_expand_stvx_be): Likewise.
4705 (altivec_expand_builtin): Add cases for
4706 ALTIVEC_BUILTIN_STVX_<MODE>, ALTIVEC_BUILTIN_STVXL_<MODE>,
4707 ALTIVEC_BUILTIN_LVXL_<MODE>, and ALTIVEC_BUILTIN_LVX_<MODE>.
4708 (altivec_init_builtins): Add definitions for
4709 __builtin_altivec_lvxl_<mode>, __builtin_altivec_lvx_<mode>,
4710 __builtin_altivec_stvx_<mode>, and __builtin_altivec_stvxl_<mode>.
4712 2014-02-21 Catherine Moore <clm@codesourcery.com>
4714 * doc/invoke.texi (mvirt, mno-virt): Document.
4715 * config/mips/mips.opt (mvirt): New option.
4716 * config/mips/mips.h (ASM_SPEC): Pass mvirt to the assembler.
4718 2014-02-21 Richard Biener <rguenther@suse.de>
4720 PR tree-optimization/60276
4721 * tree-vectorizer.h (struct _stmt_vec_info): Add min_neg_dist field.
4722 (STMT_VINFO_MIN_NEG_DIST): New macro.
4723 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Record
4724 STMT_VINFO_MIN_NEG_DIST.
4725 * tree-vect-stmts.c (vectorizable_load): Verify if assumptions
4726 made for negative dependence distances still hold.
4728 2014-02-21 Richard Biener <rguenther@suse.de>
4731 * tree-ssa-live.c (mark_all_vars_used_1): Do not walk
4732 DECL_INITIAL for globals not in the current function context.
4734 2014-02-21 Jakub Jelinek <jakub@redhat.com>
4736 PR tree-optimization/56490
4737 * params.def (PARAM_UNINIT_CONTROL_DEP_ATTEMPTS): New param.
4738 * tree-ssa-uninit.c: Include params.h.
4739 (compute_control_dep_chain): Add num_calls argument, return false
4740 if it exceed PARAM_UNINIT_CONTROL_DEP_ATTEMPTS param, pass
4741 num_calls to recursive call.
4742 (find_predicates): Change dep_chain into normal array,
4743 cur_chain into auto_vec<edge, MAX_CHAIN_LEN + 1>, add num_calls
4744 variable and adjust compute_control_dep_chain caller.
4745 (find_def_preds): Likewise.
4747 2014-02-21 Thomas Schwinge <thomas@codesourcery.com>
4749 * gimple-pretty-print.c (dump_gimple_omp_for) [flags & TDF_RAW]
4750 <case GF_OMP_FOR_KIND_CILKSIMD>: Add missing break statement.
4752 2014-02-21 Nick Clifton <nickc@redhat.com>
4754 * config/stormy16/stormy16.md (pushdqi1): Add mode to post_inc.
4755 (pushhi1): Likewise.
4756 (popqi1): Add mode to pre_dec.
4759 2014-02-21 Jakub Jelinek <jakub@redhat.com>
4761 * config/i386/i386.c (ix86_expand_vec_perm): Use V8SImode
4762 mode for mask of V8SFmode permutation.
4764 2014-02-20 Richard Henderson <rth@redhat.com>
4767 * builtins.c (expand_builtin_atomic_compare_exchange): Always make
4768 a new pseudo for OLDVAL.
4770 2014-02-20 Jakub Jelinek <jakub@redhat.com>
4773 * config/i386/i386.c (expand_vec_perm_interleave2): Don't call
4774 gen_reg_rtx if d->testing_p.
4775 (expand_vec_perm_pshufb2, expand_vec_perm_broadcast_1): Return early
4776 if d->testing_p and we will certainly return true.
4777 (expand_vec_perm_even_odd_1): Likewise. Don't call gen_reg_rtx
4780 2014-02-20 Uros Bizjak <ubizjak@gmail.com>
4782 * emit-rtl.c (gen_reg_rtx): Assert that
4783 crtl->emit.regno_pointer_align_length is non-zero.
4785 2014-02-20 Richard Henderson <rth@redhat.com>
4788 * builtins.c (expand_builtin_atomic_compare_exchange): Conditionalize
4789 on failure the store back into EXPECT.
4791 2014-02-20 Chung-Lin Tang <cltang@codesourcery.com>
4792 Sandra Loosemore <sandra@codesourcery.com>
4794 * config/nios2/nios2.md (unspec): Add UNSPEC_PIC_GOTOFF_SYM enum.
4795 * config/nios2/nios2.c (nios2_function_profiler): Add
4796 -fPIC (flag_pic == 2) support.
4797 (nios2_handle_custom_fpu_cfg): Fix warning parameter.
4798 (nios2_large_offset_p): New function.
4799 (nios2_unspec_reloc_p): Move up position, update to use
4800 nios2_large_offset_p.
4801 (nios2_unspec_address): Remove function.
4802 (nios2_unspec_offset): New function.
4803 (nios2_large_got_address): New function.
4804 (nios2_got_address): Add large offset support.
4805 (nios2_legitimize_tls_address): Update usage of removed and new
4807 (nios2_symbol_binds_local_p): New function.
4808 (nios2_load_pic_address): Add -fPIC (flag_pic == 2) support.
4809 (nios2_legitimize_address): Update to use nios2_large_offset_p.
4810 (nios2_emit_move_sequence): Avoid legitimizing (const (unspec ...)).
4811 (nios2_print_operand): Merge H/L processing, add hiadj/lo
4812 processing for (const (unspec ...)).
4813 (nios2_unspec_reloc_name): Add UNSPEC_PIC_GOTOFF_SYM case.
4815 2014-02-20 Richard Biener <rguenther@suse.de>
4817 * tree-cfg.c (replace_uses_by): Mark altered BBs before
4818 doing the substitution.
4819 (verify_gimple_assign_single): Also verify bare MEM_REFs on the lhs.
4821 2014-02-20 Martin Jambor <mjambor@suse.cz>
4824 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Uce correct
4825 info when checking whether lattices are bottom.
4827 2014-02-20 Richard Biener <rguenther@suse.de>
4830 * tree-eh.c (execute_cleanup_eh_1): Also cleanup empty EH
4833 2014-02-20 Jan Hubicka <hubicka@ucw.cz>
4836 * ipa-inline-transform.c (clone_inlined_nodes): Add freq_scale
4837 parameter specifying the scaling.
4838 (inline_call): Update.
4839 (want_inline_recursively): Guard division by zero.
4840 (recursive_inlining): Update.
4841 * ipa-inline.h (clone_inlined_nodes): Update.
4843 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
4846 * config/i386/i386.c (classify_argument): Pass structures of size
4847 64 bytes or less in register.
4849 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
4850 Kirill Yukhin <kirill.yukhin@intel.com>
4852 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): Swap operands.
4853 (_mm_rcp28_round_ss): Ditto.
4854 (_mm_rsqrt28_round_sd): Ditto.
4855 (_mm_rsqrt28_round_ss): Ditto.
4856 * config/i386/avx512erintrin.h (_mm_rcp14_round_sd): Ditto.
4857 (_mm_rcp14_round_ss): Ditto.
4858 (_mm_rsqrt14_round_sd): Ditto.
4859 (_mm_rsqrt14_round_ss): Ditto.
4860 * config/i386/sse.md (rsqrt14<mode>): Put nonimmediate operand as
4861 the first input operand, get rid of match_dup.
4862 (avx512er_exp2<mode><mask_name><round_saeonly_name>): Set type
4864 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
4866 (avx512er_vmrcp28<mode><round_saeonly_name>): Put nonimmediate
4867 operand as the first input operand, set type attribute.
4868 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
4870 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Put nonimmediate
4871 operand as the first input operand, set type attribute.
4873 2014-02-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4875 * config/rs6000/rs6000.c (vspltis_constant): Fix most significant
4878 2014-02-19 H.J. Lu <hongjiu.lu@intel.com>
4881 * config/i386/i386.c (construct_container): Remove TFmode check
4882 for X86_64_INTEGER_CLASS.
4884 2014-02-19 Uros Bizjak <ubizjak@gmail.com>
4887 * config/i386/i386.c (type_natural_mode): Warn for ABI changes
4888 only when -Wpsabi is enabled.
4890 2014-02-19 Michael Hudson-Doyle <michael.hudson@linaro.org>
4893 * config/aarch64/aarch64.c (aarch64_pass_by_reference): The rules for
4894 passing arrays in registers are the same as for structs, so remove the
4895 special case for them.
4897 2014-02-19 Eric Botcazou <ebotcazou@adacore.com>
4899 * expr.c (expand_expr_real_1) <case VIEW_CONVERT_EXPR>: For a bit-field
4900 destination type, extract only the valid bits if the source type is not
4901 integral and has a different mode.
4903 2014-02-19 Richard Biener <rguenther@suse.de>
4906 * tree-inline.c (estimate_num_insns): Avoid calling cgraph_get_node
4909 2014-02-19 Richard Biener <rguenther@suse.de>
4912 * ipa-prop.c: Include stringpool.h and tree-ssanames.h.
4913 (ipa_modify_call_arguments): Emit an argument load explicitely and
4914 preserve virtual SSA form there and for the replacement call.
4915 Do not update SSA form nor free dominance info.
4917 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
4919 * ipa.c (function_and_variable_visibility): Also clear WEAK
4920 flag when disolving COMDAT_GROUP.
4922 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
4924 * ipa-prop.h (ipa_ancestor_jf_data): Update ocmment.
4925 * ipa-prop.c (ipa_set_jf_known_type): Return early when
4927 (ipa_set_ancestor_jf): Set type to NULL hwen it is not preserved;
4928 do more sanity checks.
4929 (detect_type_change): Return true when giving up early.
4930 (compute_complex_assign_jump_func): Fix type parameter of
4931 ipa_set_ancestor_jf.
4932 (compute_complex_ancestor_jump_func): Likewise.
4933 (update_jump_functions_after_inlining): Fix updating of
4935 * ipa-cp.c (ipa_get_jf_ancestor_result): Be ready for type to be NULL.
4937 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
4939 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Also remove
4940 inline clones when edge disappears.
4942 2014-02-18 Michael Meissner <meissner@linux.vnet.ibm.com>
4945 * config/rs6000/rs6000.md (mov<mode>_64bit, TF/TDmode moves):
4946 Split 64-bit moves into 2 patterns. Do not allow the use of
4947 direct move for TDmode in little endian, since the decimal value
4948 has little endian bytes within a word, but the 64-bit pieces are
4949 ordered in a big endian fashion, and normal subreg's of TDmode are
4951 (mov<mode>_64bit_dm): Likewise.
4952 (movtd_64bit_nodm): Likewise.
4954 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
4956 PR tree-optimization/60174
4957 * tree-ssa-reassoc.c (init_range_entry): Do not look into the defining
4958 statement of an SSA_NAME that occurs in an abnormal PHI node.
4960 2014-02-18 Jakub Jelinek <jakub@redhat.com>
4963 * final.c (SEEN_BB): Remove.
4964 (SEEN_NOTE, SEEN_EMITTED): Renumber.
4965 (final_scan_insn): Don't force_source_line on second
4966 NOTE_INSN_BASIC_BLOCK.
4968 2014-02-18 Uros Bizjak <ubizjak@gmail.com>
4971 * config/i386/i386.h (struct ix86_args): Add warn_avx512f.
4972 * config/i386/i386.c (init_cumulative_args): Initialize warn_avx512f.
4973 (type_natural_mode): Warn ABI change when %zmm register is not
4974 available for AVX512F vector value passing.
4976 2014-02-18 Kai Tietz <ktietz@redhat.com>
4979 * config/i386/i386.c (ix86_expand_prologue): Use value in
4980 rax register as displacement when restoring %r10 or %rax.
4981 Fix wrong offset when restoring both registers.
4983 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
4985 * ipa-prop.c (compute_complex_ancestor_jump_func): Replace overzealous
4986 assertion with conditional return.
4988 2014-02-18 Jakub Jelinek <jakub@redhat.com>
4989 Uros Bizjak <ubizjak@gmail.com>
4992 * config/i386/driver-i386.c (host_detect_local_cpu): If
4993 YMM state is not saved by the OS, also clear has_f16c. Move
4994 CPUID 0x80000001 handling before YMM state saving checking.
4996 2014-02-18 Andrey Belevantsev <abel@ispras.ru>
4998 PR rtl-optimization/58960
4999 * haifa-sched.c (alloc_global_sched_pressure_data): New,
5000 factored out from ...
5001 (sched_init): ... here.
5002 (free_global_sched_pressure_data): New, factored out from ...
5003 (sched_finish): ... here.
5004 * sched-int.h (free_global_sched_pressure_data): Declare.
5005 * sched-rgn.c (nr_regions_initial): New static global.
5006 (haifa_find_rgns): Initialize it.
5007 (schedule_region): Disable sched-pressure for the newly
5010 2014-02-17 Richard Biener <rguenther@suse.de>
5012 * tree-vect-stmts.c (free_stmt_vec_info): Clear BB and
5013 release SSA defs of pattern stmts.
5015 2014-02-17 Richard Biener <rguenther@suse.de>
5017 * tree-inline.c (expand_call_inline): Release the virtual
5018 operand defined by the call we are about to inline.
5020 2014-02-17 Richard Biener <rguenther@suse.de>
5022 * tree-ssa.c (verify_ssa): If verify_def found an error, ICE.
5024 2014-02-17 Kirill Yukhin <kirill.yukhin@intel.com>
5025 Ilya Tocar <ilya.tocar@intel.com>
5027 * config/i386/avx512fintrin.h (_mm512_maskz_permutexvar_epi64): Swap
5028 arguments order in builtin.
5029 (_mm512_permutexvar_epi64): Ditto.
5030 (_mm512_mask_permutexvar_epi64): Ditto
5031 (_mm512_maskz_permutexvar_epi32): Ditto
5032 (_mm512_permutexvar_epi32): Ditto
5033 (_mm512_mask_permutexvar_epi32): Ditto
5035 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5037 * config/rs6000/altivec.md (p8_vmrgew): Handle little endian targets.
5038 (p8_vmrgow): Likewise.
5040 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5042 * config/rs6000/vsx.md (vsx_xxpermdi_<mode>): Handle little
5045 2014-02-15 Michael Meissner <meissner@linux.vnet.ibm.com>
5048 * config/rs6000/rs6000.md (rreg): Add TFmode, TDmode constraints.
5049 (mov<mode>_internal, TFmode/TDmode): Split TFmode/TDmode moves
5050 into 64-bit and 32-bit moves. On 64-bit moves, add support for
5051 using direct move instructions on ISA 2.07. Also adjust
5052 instruction length for 64-bit.
5053 (mov<mode>_64bit, TFmode/TDmode): Likewise.
5054 (mov<mode>_32bit, TFmode/TDmode): Likewise.
5056 2014-02-15 Alan Modra <amodra@gmail.com>
5060 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
5061 find_replacement on parts of insn rtl that might be reloaded.
5063 2014-02-15 Richard Biener <rguenther@suse.de>
5065 PR tree-optimization/60183
5066 * tree-ssa-phiprop.c (propagate_with_phi): Avoid speculating loads.
5067 (tree_ssa_phiprop): Calculate and free post-dominators.
5069 2014-02-14 Jeff Law <law@redhat.com>
5071 PR rtl-optimization/60131
5072 * ree.c (get_extended_src_reg): New function.
5073 (combine_reaching_defs): Use it rather than assuming location of REG.
5074 (find_and_remove_re): Verify first operand of extension is
5075 a REG before adding the insns to the copy list.
5077 2014-02-14 Roland McGrath <mcgrathr@google.com>
5079 * configure.ac (HAVE_AS_IX86_UD2): New test for 'ud2' mnemonic.
5080 * configure: Regenerated.
5081 * config.in: Regenerated.
5082 * config/i386/i386.md (trap) [HAVE_AS_IX86_UD2]: Use the mnemonic
5083 instead of ASM_SHORT.
5085 2014-02-14 Vladimir Makarov <vmakarov@redhat.com>
5086 Richard Earnshaw <rearnsha@arm.com>
5088 PR rtl-optimization/59535
5089 * lra-constraints.c (process_alt_operands): Encourage alternative
5090 when unassigned pseudo class is superset of the alternative class.
5091 (inherit_reload_reg): Don't inherit when optimizing for code size.
5092 * config/arm/arm.h (MODE_BASE_REG_CLASS): Add version for LRA
5093 returning CORE_REGS for anything but Thumb1 and BASE_REGS for
5094 modes not less than 4 for Thumb1.
5096 2014-02-14 Kyle McMartin <kyle@redhat.com>
5099 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Define for AArch64.
5101 2014-02-14 Richard Biener <rguenther@suse.de>
5103 * cilk-common.c (cilk_arrow): Build a MEM_REF, not an INDIRECT_REF.
5104 (get_frame_arg): Drop the assert with langhook types_compatible_p.
5105 Do not strip INDIRECT_REFs.
5107 2014-02-14 Richard Biener <rguenther@suse.de>
5110 * lto-streamer-out.c (DFS_write_tree_body): Do not follow
5111 DECL_FUNCTION_SPECIFIC_TARGET.
5112 (hash_tree): Do not hash DECL_FUNCTION_SPECIFIC_TARGET.
5113 * tree-streamer-out.c (pack_ts_target_option): Remove.
5114 (streamer_pack_tree_bitfields): Do not stream TS_TARGET_OPTION.
5115 (write_ts_function_decl_tree_pointers): Do not stream
5116 DECL_FUNCTION_SPECIFIC_TARGET.
5117 * tree-streamer-in.c (unpack_ts_target_option): Remove.
5118 (unpack_value_fields): Do not stream TS_TARGET_OPTION.
5119 (lto_input_ts_function_decl_tree_pointers): Do not stream
5120 DECL_FUNCTION_SPECIFIC_TARGET.
5122 2014-02-14 Jakub Jelinek <jakub@redhat.com>
5124 * tree-vect-loop.c (vect_is_slp_reduction): Don't set use_stmt twice.
5125 (get_initial_def_for_induction, vectorizable_induction): Ignore
5126 debug stmts when looking for exit_phi.
5127 (vectorizable_live_operation): Fix up condition.
5129 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
5131 * config/nds32/nds32.c (nds32_asm_function_prologue): Do not use
5132 nreverse() because it changes the content of original tree list.
5134 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
5136 * config/nds32/t-mlibs (MULTILIB_OPTIONS): Fix typo in comment.
5137 * config/nds32/nds32.c (nds32_merge_decl_attributes): Likewise.
5139 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
5141 * config/nds32/nds32.c (nds32_naked_function_p): Follow the
5142 GNU coding standards.
5144 2014-02-13 Jakub Jelinek <jakub@redhat.com>
5147 * dwarf2out.c (gen_subprogram_die): Don't call
5148 add_calling_convention_attribute if subr_die is old_die.
5150 2014-02-13 Sharad Singhai <singhai@google.com>
5152 * doc/optinfo.texi: Fix order of nodes.
5154 2014-02-13 Uros Bizjak <ubizjak@gmail.com>
5156 * config/i386/sse.md (xop_vmfrcz<mode>2): Generate const0 in
5157 operands[2], not operands[3].
5159 2014-02-13 Richard Biener <rguenther@suse.de>
5162 * doc/install.texi (ISL): Update recommended version to 0.12.2,
5163 mention the possibility of an in-tree build.
5164 (CLooG): Update recommended version to 0.18.1, mention the
5165 possibility of an in-tree build and clarify that the ISL
5166 bundled with CLooG does not work.
5168 2014-02-13 Jakub Jelinek <jakub@redhat.com>
5171 * expr.c (compress_float_constant): If x is a hard register,
5172 extend into a pseudo and then move to x.
5174 2014-02-13 Dominik Vogt <vogt@linux.vnet.ibm.com>
5176 * config/s390/s390.c (s390_asm_output_function_label): Fix crash
5177 caused by bad second argument to warning_at() with -mhotpatch and
5178 nested functions (e.g. with gfortran).
5180 2014-02-13 Richard Sandiford <rdsandiford@googlemail.com>
5182 * opts.c (option_name): Remove "enabled by default" rider.
5184 2014-02-12 John David Anglin <danglin@gcc.gnu.org>
5186 * config/pa/pa.c (pa_option_override): Remove auto increment FIXME.
5188 2014-02-12 H.J. Lu <hongjiu.lu@intel.com>
5189 Uros Bizjak <ubizjak@gmail.com>
5192 * configure.ac (HAVE_AS_GOTOFF_IN_DATA): Pass --32 to GNU assembler.
5193 * configure: Regenerated.
5195 2014-02-12 Richard Biener <rguenther@suse.de>
5197 * vec.c (vec_prefix::calculate_allocation): Move as
5198 inline variant to vec.h.
5199 (vec_prefix::calculate_allocation_1): New out-of-line version.
5200 * vec.h (vec_prefix::calculate_allocation_1): Declare.
5201 (vec_prefix::m_has_auto_buf): Rename to ...
5202 (vec_prefix::m_using_auto_storage): ... this.
5203 (vec_prefix::calculate_allocation): Inline the easy cases
5204 and dispatch to calculate_allocation_1 which doesn't need the
5206 (va_heap::reserve): Use gcc_checking_assert.
5207 (vec<T, A, vl_embed>::embedded_init): Add argument to initialize
5208 m_using_auto_storage.
5209 (auto_vec): Change m_vecpfx member to a vec<T, va_heap, vl_embed>
5211 (vec<T, va_heap, vl_ptr>::reserve): Remove redundant check.
5212 (vec<T, va_heap, vl_ptr>::release): Avoid casting.
5213 (vec<T, va_heap, vl_ptr>::using_auto_storage): Simplify.
5215 2014-02-12 Richard Biener <rguenther@suse.de>
5217 * gcse.c (compute_transp): break from loop over canon_modify_mem_list
5218 when we found a dependence.
5220 2014-02-12 Thomas Schwinge <thomas@codesourcery.com>
5222 * gimplify.c (gimplify_call_expr, gimplify_modify_expr): Move
5224 (maybe_fold_stmt): ... into this new function.
5225 * omp-low.c (lower_omp): Update comment.
5227 * omp-low.c (lower_omp_target): Add clobber for sizes array, after
5230 * omp-low.c (diagnose_sb_0): Make sure label_ctx is valid to
5233 2014-02-12 James Greenhalgh <james.greenhalgh@arm.com>
5235 * config/arm/aarch-cost-tables.h (generic_extra_costs): Fix
5236 identifiers in comments.
5237 (cortexa53_extra_costs): Likewise.
5238 * config/arm/arm.c (cortexa9_extra_costs): Fix identifiers in comments.
5239 (cortexa7_extra_costs): Likewise.
5240 (cortexa12_extra_costs): Likewise.
5241 (cortexa15_extra_costs): Likewise.
5242 (v7m_extra_costs): Likewise.
5244 2014-02-12 Richard Biener <rguenther@suse.de>
5247 * gimple-low.c (lower_builtin_posix_memalign): Lower conditional
5248 of posix_memalign being successful.
5249 (lower_stmt): Restrict lowering of posix_memalign to when
5250 -ftree-bit-ccp is enabled.
5252 2014-02-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
5254 * config/avr/avr-c.c (avr_resolve_overloaded_builtin): Pass vNULL for
5256 * config/spu/spu-c.c (spu_resolve_overloaded_builtin): Likewise.
5258 2014-02-12 Eric Botcazou <ebotcazou@adacore.com>
5260 PR rtl-optimization/60116
5261 * combine.c (try_combine): Also remove dangling REG_DEAD notes on the
5262 other_insn once the combination has been validated.
5264 2014-02-11 Jan Hubicka <hubicka@ucw.cz>
5267 * ipa-utils.h (possible_polymorphic_call_targets): Update prototype
5269 * ipa-devirt.c: Include demangle.h
5270 (odr_violation_reported): New static variable.
5271 (add_type_duplicate): Update odr_violations.
5272 (maybe_record_node): Add completep parameter; update it.
5273 (record_target_from_binfo): Add COMPLETEP parameter;
5274 update it as needed.
5275 (possible_polymorphic_call_targets_1): Likewise.
5276 (struct polymorphic_call_target_d): Add nonconstruction_targets;
5277 rename FINAL to COMPLETE.
5278 (record_targets_from_bases): Sanity check we found the binfo;
5279 fix COMPLETEP updating.
5280 (possible_polymorphic_call_targets): Add NONCONSTRUTION_TARGETSP
5281 parameter, fix computing of COMPLETEP.
5282 (dump_possible_polymorphic_call_targets): Imrove readability of dump;
5283 at LTO time do demangling.
5284 (ipa_devirt): Use nonconstruction_targets; Improve dumps.
5285 * gimple-fold.c (gimple_get_virt_method_for_vtable): Add can_refer
5287 (gimple_get_virt_method_for_binfo): Likewise.
5288 * gimple-fold.h (gimple_get_virt_method_for_binfo,
5289 gimple_get_virt_method_for_vtable): Update prototypes.
5291 2014-02-11 Vladimir Makarov <vmakarov@redhat.com>
5294 * genautomata.c (add_presence_absence): Fix typo with
5295 {final_}presence_list.
5297 2014-02-11 Michael Meissner <meissner@linux.vnet.ibm.com>
5300 * config/rs6000/rs6000.md (128-bit GPR splitter): Add a splitter
5301 for VSX/Altivec vectors that land in GPR registers.
5303 2014-02-11 Richard Henderson <rth@redhat.com>
5304 Jakub Jelinek <jakub@redhat.com>
5307 * tree-sra.c (load_assign_lhs_subreplacements): Add VIEW_CONVERT_EXPR
5308 around drhs if type conversion to lacc->type is not useless.
5310 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5312 * config/aarch64/aarch64-cores.def (cortex-a57): Use cortexa57
5314 (cortex-a57.cortex-a53): Likewise.
5315 * config/aarch64/aarch64.c (cortexa57_tunings): New tuning struct.
5317 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5319 * config/arm/thumb2.md (*thumb2_movhi_insn): Add alternatives for
5322 2014-02-11 Renlin Li <Renlin.Li@arm.com>
5324 * doc/sourcebuild.texi: Document check_effective_target_arm_vfp3_ok and
5325 add_options_for_arm_vfp3.
5327 2014-02-11 Jeff Law <law@redhat.com>
5330 * expr.c (expand_expr_addr_expr_1): Handle expand_expr returning an
5331 object with an undesirable mode.
5333 2014-02-11 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5336 * config/i386/sol2-9.h: New file.
5337 * config.gcc (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*,
5338 *-*-solaris2.9*): Use it.
5340 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
5342 * config/microblaze/microblaze.md: Add movsi4_rev insn pattern.
5343 * config/microblaze/predicates.md: Add reg_or_mem_operand predicate.
5345 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
5347 * config/microblaze/microblaze.c: Extend mcpu version format
5349 2014-02-10 David Holsgrove <david.holsgrove@xilinx.com>
5351 * config/microblaze/microblaze.h: Define SIZE_TYPE and PTRDIFF_TYPE.
5353 2014-02-10 Richard Henderson <rth@redhat.com>
5356 * calls.c (expand_call): Don't double-push for reg_parm_stack_space.
5357 * config/i386/i386.c (init_cumulative_args): Remove sorry for 64-bit
5358 ms-abi vs -mno-accumulate-outgoing-args.
5359 (ix86_expand_prologue): Unconditionally call ix86_eax_live_at_start_p.
5360 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Fix comment with
5363 2014-02-10 Bernd Edlinger <bernd.edlinger@hotmail.de>
5366 * cfgexpand.c (expand_asm_operands): Attach source location to
5367 ASM_INPUT rtx objects.
5368 * print-rtl.c (print_rtx): Check for UNKNOWN_LOCATION.
5370 2014-02-10 Nick Clifton <nickc@redhat.com>
5372 * config/mn10300/mn10300.c (popcount): New function.
5373 (mn10300_expand_prologue): Include saved registers in stack usage
5376 2014-02-10 Jeff Law <law@redhat.com>
5379 * reload1.c (emit_input_reload_insns): Do not create invalid RTL
5380 when changing the SET_DEST of a prior insn to avoid an input reload.
5382 2014-02-10 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
5384 * config/rs6000/sysv4.h (ENDIAN_SELECT): Do not attempt to enforce
5385 big-endian mode for -mcall-aixdesc, -mcall-freebsd, -mcall-netbsd,
5386 -mcall-openbsd, or -mcall-linux.
5387 (CC1_ENDIAN_BIG_SPEC): Remove.
5388 (CC1_ENDIAN_LITTLE_SPEC): Remove.
5389 (CC1_ENDIAN_DEFAULT_SPEC): Remove.
5390 (CC1_SPEC): Remove (always empty) %cc1_endian_... spec.
5391 (SUBTARGET_EXTRA_SPECS): Remove %cc1_endian_big, %cc1_endian_little,
5392 and %cc1_endian_default.
5393 * config/rs6000/sysv4le.h (CC1_ENDIAN_DEFAULT_SPEC): Remove.
5395 2014-02-10 Richard Biener <rguenther@suse.de>
5397 PR tree-optimization/60115
5398 * tree-eh.c (tree_could_trap_p): Unify TARGET_MEM_REF and
5399 MEM_REF handling. Properly verify that the accesses are not
5400 out of the objects bound.
5402 2014-02-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5404 * config/aarch64/aarch64.c (aarch64_override_options): Fix typo from
5407 2014-02-10 Eric Botcazou <ebotcazou@adacore.com>
5409 * ipa-devirt.c (get_polymorphic_call_info_from_invariant): Return
5410 proper constants and fix formatting.
5411 (possible_polymorphic_call_targets): Fix formatting.
5413 2014-02-10 Kirill Yukhin <kirill.yukhin@intel.com>
5414 Ilya Tocar <ilya.tocar@intel.com>
5416 * config/i386/avx512fintrin.h (_mm512_storeu_epi64): Removed.
5417 (_mm512_loadu_epi32): Renamed into...
5418 (_mm512_loadu_si512): This.
5419 (_mm512_storeu_epi32): Renamed into...
5420 (_mm512_storeu_si512): This.
5421 (_mm512_maskz_ceil_ps): Removed.
5422 (_mm512_maskz_ceil_pd): Ditto.
5423 (_mm512_maskz_floor_ps): Ditto.
5424 (_mm512_maskz_floor_pd): Ditto.
5425 (_mm512_floor_round_ps): Ditto.
5426 (_mm512_floor_round_pd): Ditto.
5427 (_mm512_ceil_round_ps): Ditto.
5428 (_mm512_ceil_round_pd): Ditto.
5429 (_mm512_mask_floor_round_ps): Ditto.
5430 (_mm512_mask_floor_round_pd): Ditto.
5431 (_mm512_mask_ceil_round_ps): Ditto.
5432 (_mm512_mask_ceil_round_pd): Ditto.
5433 (_mm512_maskz_floor_round_ps): Ditto.
5434 (_mm512_maskz_floor_round_pd): Ditto.
5435 (_mm512_maskz_ceil_round_ps): Ditto.
5436 (_mm512_maskz_ceil_round_pd): Ditto.
5437 (_mm512_expand_pd): Ditto.
5438 (_mm512_expand_ps): Ditto.
5439 * config/i386/i386.c (ix86_builtins): Remove
5440 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK.
5441 (bdesc_args): Ditto.
5442 * config/i386/predicates.md (const1256_operand): New.
5443 (const_1_to_2_operand): Ditto.
5444 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
5445 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
5446 (*avx512pf_gatherpf<mode>sf): Ditto.
5447 (avx512pf_gatherpf<mode>df): Ditto.
5448 (*avx512pf_gatherpf<mode>df_mask): Ditto.
5449 (*avx512pf_gatherpf<mode>df): Ditto.
5450 (avx512pf_scatterpf<mode>sf): Ditto.
5451 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
5452 (*avx512pf_scatterpf<mode>sf): Ditto.
5453 (avx512pf_scatterpf<mode>df): Ditto.
5454 (*avx512pf_scatterpf<mode>df_mask): Ditto.
5455 (*avx512pf_scatterpf<mode>df): Ditto.
5456 (avx512f_expand<mode>): Removed.
5457 (<shift_insn><mode>3<mask_name>): Change predicate type.
5459 2014-02-08 Jakub Jelinek <jakub@redhat.com>
5461 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers
5462 not at the end of datarefs vector use ordered_remove to avoid
5463 reordering datarefs vector.
5466 * gimplify.c (gimplify_bind_expr): In ORT_SIMD region
5467 mark local addressable non-static vars as GOVD_PRIVATE
5468 instead of GOVD_LOCAL.
5469 * omp-low.c (lower_omp_for): Move gimple_bind_vars
5470 and BLOCK_VARS of gimple_bind_block to new_stmt rather
5474 * tree-ssa-ccp.c (surely_varying_stmt_p): Don't return true
5475 if TYPE_ATTRIBUTES (gimple_call_fntype ()) contain
5476 assume_aligned or alloc_align attributes.
5477 (bit_value_assume_aligned): Add ATTR, PTRVAL and ALLOC_ALIGN
5478 arguments. Handle also assume_aligned and alloc_align attributes.
5479 (evaluate_stmt): Adjust bit_value_assume_aligned caller. Handle
5480 calls to functions with assume_aligned or alloc_align attributes.
5481 * doc/extend.texi: Document assume_aligned and alloc_align attributes.
5483 2014-02-08 Terry Guo <terry.guo@arm.com>
5485 * doc/invoke.texi: Document ARM -march=armv7e-m.
5487 2014-02-08 Jakub Jelinek <jakub@redhat.com>
5489 * cilk-common.c (cilk_init_builtins): Clear TREE_NOTHROW
5490 flag on __cilkrts_rethrow builtin.
5493 * ipa-cp.c (determine_versionability): Fail at -O0
5494 or __attribute__((optimize (0))) or -fno-ipa-cp functions.
5495 * tree-sra.c (ipa_sra_preliminary_function_checks): Similarly.
5498 2014-02-04 Jakub Jelinek <jakub@redhat.com>
5501 * tree-inline.c (copy_forbidden): Fail for
5502 __attribute__((optimize (0))) functions.
5504 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
5506 * varpool.c: Include pointer-set.h.
5507 (varpool_remove_unreferenced_decls): Variables in other partitions
5508 will not be output; be however careful to not lose information
5511 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
5513 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do O(1)
5514 lookup in the vtable constructor.
5516 2014-02-07 Jeff Law <law@redhat.com>
5519 * config/m68k/m68k.md (ashldi_extsi): Turn into a
5520 define_insn_and_split.
5522 * ipa-inline.c (inline_small_functions): Fix typos.
5524 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5526 * config/s390/s390-protos.h (s390_can_use_simple_return_insn)
5527 (s390_can_use_return_insn): Declare.
5528 * config/s390/s390.h (EPILOGUE_USES): Define.
5529 * config/s390/s390.c (s390_mainpool_start): Allow two main_pool
5531 (s390_chunkify_start): Handle return JUMP_LABELs.
5532 (s390_early_mach): Emit a main_pool instruction on the entry edge.
5533 (s300_set_up_by_prologue, s390_can_use_simple_return_insn)
5534 (s390_can_use_return_insn): New functions.
5535 (s390_fix_long_loop_prediction): Handle conditional returns.
5536 (TARGET_SET_UP_BY_PROLOGUE): Define.
5537 * config/s390/s390.md (ANY_RETURN): New code iterator.
5538 (*creturn, *csimple_return, return, simple_return): New patterns.
5540 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5542 * config/s390/s390.c (s390_restore_gprs_from_fprs): Add REG_CFA_RESTORE
5543 notes to each restore. Also add REG_CFA_DEF_CFA when restoring %r15.
5544 (s390_optimize_prologue): Don't clear RTX_FRAME_RELATED_P. Update the
5545 REG_CFA_RESTORE list when deciding not to restore a register.
5547 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5549 * config/s390/s390.c: Include tree-pass.h and context.h.
5550 (s390_early_mach): New function, split out from...
5551 (s390_emit_prologue): ...here.
5552 (pass_data_s390_early_mach): New pass structure.
5553 (pass_s390_early_mach): New class.
5554 (s390_option_override): Create and register early_mach pass.
5555 Move to end of file.
5557 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5559 * var-tracking.c (vt_stack_adjustments): Don't require stack_adjusts
5560 to match for the exit block.
5562 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5564 * config/s390/s390.md ("atomic_load<mode>", "atomic_store<mode>")
5565 ("atomic_compare_and_swap<mode>", "atomic_fetch_<atomic><mode>"):
5566 Reject misaligned operands.
5568 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5570 * optabs.c (expand_atomic_compare_and_swap): Allow expander to fail.
5572 2014-02-07 Richard Biener <rguenther@suse.de>
5575 * gimple-low.c (lower_builtin_posix_memalign): New function.
5576 (lower_stmt): Call it to lower posix_memalign in a way
5577 to make alignment info accessible.
5579 2014-02-07 Jakub Jelinek <jakub@redhat.com>
5582 * tree.c (build_common_builtin_nodes): Set ECF_LEAF for
5583 __builtin_setjmp_receiver.
5585 2014-02-07 Richard Biener <rguenther@suse.de>
5588 * builtin-types.def (BT_FN_INT_PTRPTR_SIZE_SIZE): Add.
5589 * builtins.def (BUILT_IN_POSIX_MEMALIGN): Likewise.
5590 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
5591 Handle BUILT_IN_POSIX_MEMALIGN.
5592 (find_func_clobbers): Likewise.
5593 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
5594 (call_may_clobber_ref_p_1): Likewise.
5596 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5599 * ipa-devirt.c (record_target_from_binfo): Remove overactive
5602 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5605 * lto-cgraph.c (lto_output_node): Use
5606 symtab_get_symbol_partitioning_class.
5607 (lto_output_varpool_node): likewise.
5608 (symtab_get_symbol_partitioning_class): Move here from
5610 * cgraph.h (symbol_partitioning_class): Likewise.
5611 (symtab_get_symbol_partitioning_class): Declare.
5613 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5615 * ggc.h (ggc_internal_cleared_alloc): New macro.
5616 * vec.h (vec_safe_copy): Handle memory stats.
5617 * omp-low.c (simd_clone_struct_alloc): Use ggc_internal_cleared_alloc.
5618 * target-globals.c (save_target_globals): Likewise.
5620 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5623 * expr.c (emit_move_resolve_push): Export; be bit more selective
5624 on when to clear alias set.
5625 * expr.h (emit_move_resolve_push): Declare.
5626 * function.h (struct function): Add tail_call_marked.
5627 * tree-tailcall.c (optimize_tail_call): Set tail_call_marked.
5628 * config/i386/i386-protos.h (ix86_expand_push): Remove.
5629 * config/i386/i386.md (TImode move expander): De not call
5631 (FP push expanders): Preserve memory attributes.
5632 * config/i386/sse.md (push<mode>1): Remove.
5633 * config/i386/i386.c (ix86_expand_vector_move): Handle push operation.
5634 (ix86_expand_push): Remove.
5635 * config/i386/mmx.md (push<mode>1): Remove.
5637 2014-02-06 Jakub Jelinek <jakub@redhat.com>
5639 PR rtl-optimization/60030
5640 * internal-fn.c (ubsan_expand_si_overflow_mul_check): Surround
5641 lopart with paradoxical subreg before shifting it up by hprec.
5643 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5645 * config/arm/aarch-cost-tables.h (cortexa57_extra_costs): New table.
5646 Remove extra newline at end of file.
5647 * config/arm/arm.c (arm_cortex_a57_tune): New tuning struct.
5648 (arm_issue_rate): Handle cortexa57.
5649 * config/arm/arm-cores.def (cortex-a57): Use cortex_a57 tuning.
5650 (cortex-a57.cortex-a53): Likewise.
5652 2014-02-06 Jakub Jelinek <jakub@redhat.com>
5655 * config/arm/arm.c (emit_multi_reg_push): Add dwarf_regs_mask argument,
5656 don't record in REG_FRAME_RELATED_EXPR registers not set in that
5658 (arm_expand_prologue): Adjust all callers.
5659 (arm_unwind_emit_sequence): Allow saved, but not important for unwind
5660 info, registers also at the lowest numbered registers side. Use
5661 gcc_assert instead of abort, and SET_SRC/SET_DEST macros instead of
5665 * var-tracking.c (adjust_mems): Before adding a SET to
5666 amd->side_effects, adjust it's SET_SRC using simplify_replace_fn_rtx.
5668 2014-02-06 Alan Modra <amodra@gmail.com>
5671 * config/rs6000/rs6000.c (rs6000_secondary_memory_needed_mode): Only
5672 change SDmode to DDmode when lra_in_progress.
5674 2014-02-06 Jakub Jelinek <jakub@redhat.com>
5677 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers, call
5678 free_data_ref on the dr first, and before goto again also set dr
5679 to the next dr. For simd_lane_access, free old datarefs[i] before
5680 overwriting it. For get_vectype_for_scalar_type failure, don't
5681 free_data_ref if simd_lane_access.
5683 * Makefile.in (prefix.o, cppbuiltin.o): Depend on $(BASEVER).
5686 * tree.h (opts_for_fn): New inline function.
5687 (opt_for_fn): Define.
5688 * config/i386/i386.c (ix86_function_regparm): Use
5689 opt_for_fn (decl, optimize) instead of optimize.
5691 2014-02-06 Marcus Shawcroft <marcus.shawcroft@arm.com>
5693 * config/aarch64/aarch64.c (aarch64_classify_symbol): Fix logic
5694 for SYMBOL_REF in large memory model.
5696 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5698 * config/aarch64/aarch64-cores.def (cortex-a53): Specify CRC32
5700 (cortex-a57): Likewise.
5701 (cortex-a57.cortex-a53): Likewise.
5703 2014-02-06 Yury Gribov <y.gribov@samsung.com>
5704 Kugan Vivekanandarajah <kuganv@linaro.org>
5706 * config/arm/arm.c (arm_vector_alignment_reachable): Check
5708 * config/arm/arm.c (arm_builtin_support_vector_misalignment): Likewise.
5710 2014-02-06 Richard Biener <rguenther@suse.de>
5712 * tree-cfg.c (gimple_duplicate_sese_region): Fix ordering of
5713 set_loop_copy and initialize_original_copy_tables.
5715 2014-02-06 Alex Velenko <Alex.Velenko@arm.com>
5717 * config/aarch64/aarch64-simd.md
5718 (aarch64_ashr_simddi): Change QI to SI.
5720 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
5721 Jakub Jelinek <jakub@redhat.com>
5724 * ipa-inline-analysis.c (compute_bb_predicates): Ensure monotonicity
5727 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5729 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Change
5730 CODE_FOR_altivec_vpku[hw]um to
5731 CODE_FOR_altivec_vpku[hw]um_direct.
5732 * config/rs6000/altivec.md (vec_unpacks_hi_<VP_small_lc>): Change
5733 UNSPEC_VUNPACK_HI_SIGN to UNSPEC_VUNPACK_HI_SIGN_DIRECT.
5734 (vec_unpacks_lo_<VP_small_lc>): Change UNSPEC_VUNPACK_LO_SIGN to
5735 UNSPEC_VUNPACK_LO_SIGN_DIRECT.
5737 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5739 * config/rs6000/altivec.md (altivec_vsum2sws): Adjust code
5740 generation for -maltivec=be.
5741 (altivec_vsumsws): Simplify redundant test.
5743 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5745 * altivec.md (UNSPEC_VPACK_UNS_UNS_MOD_DIRECT): New unspec.
5746 (UNSPEC_VUNPACK_HI_SIGN_DIRECT): Likewise.
5747 (UNSPEC_VUNPACK_LO_SIGN_DIRECT): Likewise.
5748 (mulv8hi3): Use gen_altivec_vpkuwum_direct instead of
5749 gen_altivec_vpkuwum.
5750 (altivec_vpkpx): Test for VECTOR_ELT_ORDER_BIG instead of for
5752 (altivec_vpks<VI_char>ss): Likewise.
5753 (altivec_vpks<VI_char>us): Likewise.
5754 (altivec_vpku<VI_char>us): Likewise.
5755 (altivec_vpku<VI_char>um): Likewise.
5756 (altivec_vpku<VI_char>um_direct): New (copy of
5757 altivec_vpku<VI_char>um that still relies on BYTES_BIG_ENDIAN, for
5759 (altivec_vupkhs<VU_char>): Emit vupkls* instead of vupkhs* when
5760 target is little endian and -maltivec=be is not specified.
5761 (*altivec_vupkhs<VU_char>_direct): New (copy of
5762 altivec_vupkhs<VU_char> that always emits vupkhs*, for internal use).
5763 (altivec_vupkls<VU_char>): Emit vupkhs* instead of vupkls* when
5764 target is little endian and -maltivec=be is not specified.
5765 (*altivec_vupkls<VU_char>_direct): New (copy of
5766 altivec_vupkls<VU_char> that always emits vupkls*, for internal use).
5767 (altivec_vupkhpx): Emit vupklpx instead of vupkhpx when target is
5768 little endian and -maltivec=be is not specified.
5769 (altivec_vupklpx): Emit vupkhpx instead of vupklpx when target is
5770 little endian and -maltivec=be is not specified.
5772 2014-02-05 Richard Henderson <rth@redhat.com>
5775 * combine-stack-adj.c: Revert r206943.
5776 * sched-int.h (struct deps_desc): Add last_args_size.
5777 * sched-deps.c (init_deps): Initialize it.
5778 (sched_analyze_insn): Add OUTPUT dependencies between insns that
5779 contain REG_ARGS_SIZE notes.
5781 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
5783 * lto-cgraph.c (asm_nodes_output): Make global.
5784 * lto-wrapper.c (run_gcc): Pass down paralelizm to WPA.
5785 * gcc.c (AS_NEEDS_DASH_FOR_PIPED_INPUT): Allow WPA parameter
5786 (driver_handle_option): Handle OPT_fwpa.
5788 2014-02-05 Jakub Jelinek <jakub@redhat.com>
5791 * ipa-devirt.c (possible_polymorphic_call_targets): Fix
5792 a comment typo and formatting issue. If odr_hash hasn't been
5793 created, return vNULL and set *completep to false.
5796 * tree-eh.c (cleanup_empty_eh): Bail out on totally empty
5797 bb with no successors.
5799 2014-02-05 James Greenhalgh <james.greenhalgh@arm.com>
5802 * doc/invoke.texi (-march): Clarify documentation for ARM.
5806 2014-02-05 Richard Biener <rguenther@suse.de>
5808 * tree-vect-loop.c (vect_analyze_loop_2): Be more informative
5809 when not vectorizing because of too many alias checks.
5810 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
5811 Add more verboseness, avoid duplicate MSG_MISSED_OPTIMIZATION.
5813 2014-02-05 Nick Clifton <nickc@redhat.com>
5815 * config/mn10300/mn10300.c (mn10300_hard_regno_mode_ok): Do not
5816 accept extended registers in any mode when compiling for the MN10300.
5818 2014-02-05 Yury Gribov <y.gribov@samsung.com>
5820 * cif-code.def (ATTRIBUTE_MISMATCH): New CIF code.
5821 * ipa-inline.c (report_inline_failed_reason): Handle mismatched
5822 sanitization attributes.
5823 (can_inline_edge_p): Likewise.
5824 (sanitize_attrs_match_for_inline_p): New function.
5826 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
5828 * ipa-prop.c (detect_type_change): Shor circuit testing of
5829 type changes on THIS pointer.
5831 2014-02-04 John David Anglin <danglin@gcc.gnu.org>
5834 * config/pa/pa.c (legitimize_tls_address): Return original address
5835 if not passed a SYMBOL_REF rtx.
5836 (hppa_legitimize_address): Call legitimize_tls_address for all TLS
5838 (pa_emit_move_sequence): Simplify TLS source operands.
5839 (pa_legitimate_constant_p): Reject all TLS constants.
5840 * config/pa/pa.h (PA_SYMBOL_REF_TLS_P): Correct comment.
5841 (CONSTANT_ADDRESS_P): Reject TLS CONST addresses.
5843 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
5845 * ipa.c (function_and_variable_visibility): Decompose DECL_ONE_ONLY
5846 groups when we know they are controlled by LTO.
5847 * varasm.c (default_binds_local_p_1): If object is in other partition,
5848 it will be resolved locally.
5850 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
5852 * config/host-linux.c (linux_gt_pch_use_address): Don't
5853 use SSIZE_MAX because it is not always defined.
5855 2014-02-04 Vladimir Makarov <vmakarov@redhat.com>
5858 * lra-constraints.c (need_for_split_p): Use more 3 reloads as
5859 threshold for pseudo splitting.
5860 (update_ebb_live_info): Process call argument hard registers and
5861 hard registers from insn definition too.
5862 (max_small_class_regs_num): New constant.
5863 (inherit_in_ebb): Update live hard regs through EBBs. Update
5864 reloads_num only for small register classes. Don't split for
5867 2014-02-04 Markus Trippelsdorf <markus@trippelsdorf.de>
5870 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that target
5873 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
5875 * gimple-fold.c (can_refer_decl_in_current_unit_p): Default
5878 2014-02-04 Marek Polacek <polacek@redhat.com>
5880 * gdbinit.in (pel): Define.
5882 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
5884 * doc/invoke.texi (fstrict-volatile-bitfields): Clarify current
5887 2014-02-04 Richard Biener <rguenther@suse.de>
5890 * lto-streamer-out.c (tree_is_indexable): Force NAMELIST_DECLs
5891 in function context local.
5892 (lto_output_tree_ref): Do not write trees from lto_output_tree_ref.
5893 * lto-streamer-in.c (lto_input_tree_ref): Handle LTO_namelist_decl_ref
5894 similar to LTO_imported_decl_ref.
5896 2014-02-04 Jakub Jelinek <jakub@redhat.com>
5898 PR tree-optimization/60002
5899 * cgraphclones.c (build_function_decl_skip_args): Clear
5902 PR tree-optimization/60023
5903 * tree-if-conv.c (predicate_mem_writes): Pass true instead of
5904 false to gsi_replace.
5905 * tree-vect-stmts.c (vect_finish_stmt_generation): If stmt
5906 has been in some EH region and vec_stmt could throw, add
5907 vec_stmt into the same EH region.
5908 * tree-data-ref.c (get_references_in_stmt): If IFN_MASK_LOAD
5909 has no lhs, ignore it.
5910 * internal-fn.c (expand_MASK_LOAD): Likewise.
5913 * tree-inline.c (copy_forbidden): Fail for
5914 __attribute__((optimize (0))) functions.
5917 * omp-low.c (simd_clone_struct_copy): If from->inbranch
5918 is set, copy one less argument.
5919 (expand_simd_clones): Don't subtract clone_info->inbranch
5920 from simd_clone_struct_alloc argument.
5922 PR rtl-optimization/57915
5923 * recog.c (simplify_while_replacing): If all unary/binary/relational
5924 operation arguments are constant, attempt to simplify those.
5927 * expmed.c (expand_mult): For MODE_VECTOR_INT multiplication
5928 if there is no vashl<mode>3 or ashl<mode>3 insn, skip_synth.
5930 2014-02-04 Richard Biener <rguenther@suse.de>
5932 PR tree-optimization/60012
5933 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Apply
5934 TBAA disambiguation to all DDRs.
5936 2014-02-04 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5939 * config/sol2.h (LINK_LIBGCC_MAPFILE_SPEC): Define.
5940 (LINK_SPEC): Use it for -shared, -shared-libgcc.
5942 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
5945 * tree.c (get_binfo_at_offset): Do not get confused by empty classes;
5947 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
5949 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Remove.
5950 * gimple-fold.h (gimple_extract_devirt_binfo_from_cst): Remove.
5952 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
5955 * ipa-cp.c (ipa_get_indirect_edge_target_1): Use ipa-devirt
5956 to figure out targets of polymorphic calls with known decl.
5957 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
5958 * ipa-utils.h (get_polymorphic_call_info_from_invariant): Declare.
5959 * ipa-devirt.c (get_polymorphic_call_info_for_decl): Break out from ...
5960 (get_polymorphic_call_info): ... here.
5961 (get_polymorphic_call_info_from_invariant): New function.
5963 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
5965 * ipa-cp.c (ipa_get_indirect_edge_target_1): Do direct
5966 lookup via vtable pointer; check for type consistency
5967 and turn inconsitent facts into UNREACHABLE.
5968 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
5969 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do not ICE on
5970 type inconsistent querries; return UNREACHABLE instead.
5972 2014-02-03 Richard Henderson <rth@twiddle.net>
5975 * tree-ssa-uninit.c (push_to_worklist): Don't re-push if we've
5976 already processed this node.
5977 (normalize_one_pred_1): Pass along mark_set.
5978 (normalize_one_pred): Create and destroy a pointer_set_t.
5979 (normalize_one_pred_chain): Likewise.
5981 2014-02-03 Laurent Aflonsi <laurent.alfonsi@st.com>
5983 PR gcov-profile/58602
5984 * gcc/gcov-io.c (gcov_open): Open with truncation when mode < 0.
5986 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
5989 * ipa-cp.c (ipa_get_indirect_edge_target_1): Give up on
5990 -fno-devirtualize; try to devirtualize by the knowledge of
5991 virtual table pointer given by aggregate propagation.
5992 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
5993 (ipa_print_node_jump_functions): Dump also offset that
5994 is relevant for polymorphic calls.
5995 (determine_known_aggregate_parts): Add arg_type parameter; use it
5996 instead of determining the type from pointer type.
5997 (ipa_compute_jump_functions_for_edge): Update call of
5998 determine_known_aggregate_parts.
5999 * gimple-fold.c (gimple_get_virt_method_for_vtable): Break out from ...
6000 (gimple_get_virt_method_for_binfo): ... here; simplify using
6001 vtable_pointer_value_to_vtable.
6002 * gimple-fold.h (gimple_get_virt_method_for_vtable): Declare.
6003 * ipa-devirt.c (subbinfo_with_vtable_at_offset): Turn OFFSET parameter
6004 to unsigned HOST_WIDE_INT; use vtable_pointer_value_to_vtable.
6005 (vtable_pointer_value_to_vtable): Break out from ...; handle also
6007 (vtable_pointer_value_to_binfo): ... here.
6008 * ipa-utils.h (vtable_pointer_value_to_vtable): Declare.
6010 2014-02-03 Teresa Johnson <tejohnson@google.com>
6012 * tree-vect-slp.c (vect_supported_load_permutation_p): Avoid
6013 redef of outer loop index variable.
6015 2014-02-03 Marc Glisse <marc.glisse@inria.fr>
6019 * doc/extend.texi (Function Attributes): Typo.
6021 2014-02-03 Cong Hou <congh@google.com>
6023 PR tree-optimization/60000
6024 * tree-vect-loop.c (vect_transform_loop): Set pattern_def_seq to NULL
6025 if the vectorized statement is a store. A store statement can only
6026 appear at the end of pattern statements.
6028 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
6030 * config/i386/i386.c (flag_opts): Add -mlong-double-128.
6031 (ix86_option_override_internal): Default long double to 64-bit for
6032 32-bit Bionic and to 128-bit for 64-bit Bionic.
6034 * config/i386/i386.h (LONG_DOUBLE_TYPE_SIZE): Use 128 if
6035 TARGET_LONG_DOUBLE_128 is true.
6036 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Likewise.
6038 * config/i386/i386.opt (mlong-double-80): Negate -mlong-double-64.
6039 (mlong-double-64): Negate -mlong-double-128.
6040 (mlong-double-128): New option.
6042 * config/i386/i386-c.c (ix86_target_macros): Define
6043 __LONG_DOUBLE_128__ for TARGET_LONG_DOUBLE_128.
6045 * doc/invoke.texi: Document -mlong-double-128.
6047 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
6049 PR rtl-optimization/60024
6050 * sel-sched.c (init_regs_for_mode): Check if mode is OK first.
6052 2014-02-03 Markus Trippelsdorf <markus@trippelsdorf.de>
6054 * doc/invoke.texi (fprofile-reorder-functions): Fix typo.
6056 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
6058 PR rtl-optimization/57662
6059 * sel-sched.c (code_motion_path_driver): Do not mark already not
6060 existing blocks in the visiting bitmap.
6062 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
6064 * sel-sched-ir.c (sel_gen_insn_from_expr_after): Reset INSN_DELETED_P
6065 on the insn being emitted.
6067 2014-02-03 James Greenhalgh <james.greenhalgh@arm.com>
6068 Will Deacon <will.deacon@arm.com>
6070 * doc/gimple.texi (gimple_asm_clear_volatile): Remove.
6072 2014-02-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6074 * config/arm/arm-tables.opt: Regenerate.
6076 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6078 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): Generalize
6079 for vector types other than V16QImode.
6080 * config/rs6000/altivec.md (altivec_vperm_<mode>): Change to a
6081 define_expand, and call altivec_expand_vec_perm_le when producing
6082 code with little endian element order.
6083 (*altivec_vperm_<mode>_internal): New insn having previous
6084 behavior of altivec_vperm_<mode>.
6085 (altivec_vperm_<mode>_uns): Change to a define_expand, and call
6086 altivec_expand_vec_perm_le when producing code with little endian
6088 (*altivec_vperm_<mode>_uns_internal): New insn having previous
6089 behavior of altivec_vperm_<mode>_uns.
6091 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6093 * config/rs6000/altivec.md (UNSPEC_VSUMSWS_DIRECT): New unspec.
6094 (altivec_vsumsws): Add handling for -maltivec=be with a little
6096 (altivec_vsumsws_direct): New.
6097 (reduc_splus_<mode>): Call gen_altivec_vsumsws_direct instead of
6098 gen_altivec_vsumsws.
6100 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
6102 * ipa-devirt.c (subbinfo_with_vtable_at_offset,
6103 vtable_pointer_value_to_binfo): New functions.
6104 * ipa-utils.h (vtable_pointer_value_to_binfo): Declare.
6105 * ipa-prop.c (extr_type_from_vtbl_ptr_store): Use it.
6107 2014-02-02 Sandra Loosemore <sandra@codesourcery.com>
6109 * config/nios2/nios2.md (load_got_register): Initialize GOT
6110 pointer from _gp_got instead of _GLOBAL_OFFSET_TABLE_.
6111 * config/nios2/nios2.c (nios2_function_profiler): Likewise.
6113 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
6115 * ipa-prop.c (update_jump_functions_after_inlining): When type is not
6116 preserverd by passthrough, do not propagate the type.
6118 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
6120 * config/mips/mips.c (MIPS_GET_FCSR, MIPS_SET_FCSR): New macros.
6121 (mips_atomic_assign_expand_fenv): New function.
6122 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): Define.
6124 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
6126 * doc/extend.texi (__builtin_mips_get_fcsr): Document.
6127 (__builtin_mips_set_fcsr): Likewise.
6128 * config/mips/mips-ftypes.def: Add MIPS_VOID_FTYPE_USI and
6129 MIPS_USI_FTYPE_VOID.
6130 * config/mips/mips-protos.h (mips16_expand_get_fcsr): Declare
6131 (mips16_expand_set_fcsr): Likewise.
6132 * config/mips/mips.c (mips16_get_fcsr_stub): New variable.
6133 (mips16_set_fcsr_stub): Likewise.
6134 (mips16_get_fcsr_one_only_stub): New class.
6135 (mips16_set_fcsr_one_only_stub): Likewise.
6136 (mips16_expand_get_fcsr, mips16_expand_set_fcsr): New functions.
6137 (mips_code_end): Output the get_fcsr and set_fcsr stubs, if needed.
6138 (BUILTIN_AVAIL_MIPS16, AVAIL_ALL): New macros.
6139 (hard_float): New availability predicate.
6140 (mips_builtins): Add get_fcsr and set_fcsr.
6141 (mips_expand_builtin): Check BUILTIN_AVAIL_MIPS16.
6142 * config/mips/mips.md (UNSPEC_GET_FCSR, UNSPEC_SET_FCSR): New unspecs.
6143 (GET_FCSR_REGNUM, SET_FCSR_REGNUM): New constants.
6144 (mips_get_fcsr, *mips_get_fcsr, mips_get_fcsr_mips16_<mode>)
6145 (mips_set_fcsr, *mips_set_fcsr, mips_set_fcsr_mips16_<mode>): New
6148 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
6150 * config/mips/mips.c (mips_one_only_stub): New class.
6151 (mips_need_mips16_rdhwr_p): Replace with...
6152 (mips16_rdhwr_stub): ...this new variable.
6153 (mips16_stub_call_address): New function.
6154 (mips16_rdhwr_one_only_stub): New class.
6155 (mips_expand_thread_pointer): Use mips16_stub_call_address.
6156 (mips_output_mips16_rdhwr): Delete.
6157 (mips_finish_stub): New function.
6158 (mips_code_end): Use it to handle rdhwr stubs.
6160 2014-02-02 Uros Bizjak <ubizjak@gmail.com>
6163 * config/i386/i386.c (classify_argument): Fix handling of bit_offset
6164 when calculating size of integer atomic types.
6166 2014-02-02 H.J. Lu <hongjiu.lu@intel.com>
6168 * ipa-inline-analysis.c (true_predicate_p): Fix a typo in comments.
6170 2014-02-01 Jakub Jelinek <jakub@redhat.com>
6172 PR tree-optimization/60003
6173 * gimple-low.c (lower_builtin_setjmp): Set cfun->has_nonlocal_label.
6174 * profile.c (branch_prob): Use gimple_call_builtin_p
6175 to check for BUILT_IN_SETJMP_RECEIVER.
6176 * tree-inline.c (copy_bb): Call notice_special_calls.
6178 2014-01-31 Vladimir Makarov <vmakarov@redhat.com>
6181 * lra-constraints.c (process_alt_operands): Update reload_sum only
6184 2014-01-31 Richard Henderson <rth@redhat.com>
6187 * tree-eh.c (lower_try_finally_switch): Delay lowering finally block
6188 until after else_eh is processed.
6190 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
6192 * config/i386/avx512fintrin.h (_MM_FROUND_TO_NEAREST_INT),
6193 (_MM_FROUND_TO_NEG_INF), (_MM_FROUND_TO_POS_INF),
6194 (_MM_FROUND_TO_ZERO), (_MM_FROUND_CUR_DIRECTION): Are already defined
6195 in smmintrin.h, remove them.
6196 (_MM_FROUND_NO_EXC): Same as above, bit also wrong value.
6197 * config/i386/i386.c (ix86_print_operand): Split sae and rounding.
6198 * config/i386/i386.md (ROUND_SAE): Fix value.
6199 * config/i386/predicates.md (const_4_or_8_to_11_operand): New.
6200 (const48_operand): New.
6201 * config/i386/subst.md (round), (round_expand): Use
6202 const_4_or_8_to_11_operand.
6203 (round_saeonly), (round_saeonly_expand): Use const48_operand.
6205 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
6207 * config/i386/constraints.md (Yk): Swap meaning with k.
6208 * config/i386/i386.md (movhi_internal): Change Yk to k.
6209 (movqi_internal): Ditto.
6210 (*k<logic><mode>): Ditto.
6213 (kandn<mode>): Ditto.
6214 (*<code>hi_1): Ditto.
6215 (*<code>qi_1): Ditto.
6216 (kxnor<mode>): Ditto.
6217 (kortestzhi): Ditto.
6218 (kortestchi): Ditto.
6220 (*one_cmplhi2_1): Ditto.
6221 (*one_cmplqi2_1): Ditto.
6222 * config/i386/sse.md (): Change k to Yk.
6223 (avx512f_load<mode>_mask): Ditto.
6224 (avx512f_blendm<mode>): Ditto.
6225 (avx512f_store<mode>_mask): Ditto.
6226 (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
6227 (avx512f_storedqu<mode>_mask): Ditto.
6228 (avx512f_cmp<mode>3<mask_scalar_merge_name><round_saeonly_name>):
6230 (avx512f_ucmp<mode>3<mask_scalar_merge_name>): Ditto.
6231 (avx512f_vmcmp<mode>3<round_saeonly_name>): Ditto.
6232 (avx512f_vmcmp<mode>3_mask<round_saeonly_name>): Ditto.
6233 (avx512f_maskcmp<mode>3): Ditto.
6234 (avx512f_fmadd_<mode>_mask<round_name>): Ditto.
6235 (avx512f_fmadd_<mode>_mask3<round_name>): Ditto.
6236 (avx512f_fmsub_<mode>_mask<round_name>): Ditto.
6237 (avx512f_fmsub_<mode>_mask3<round_name>): Ditto.
6238 (avx512f_fnmadd_<mode>_mask<round_name>): Ditto.
6239 (avx512f_fnmadd_<mode>_mask3<round_name>): Ditto.
6240 (avx512f_fnmsub_<mode>_mask<round_name>): Ditto.
6241 (avx512f_fnmsub_<mode>_mask3<round_name>): Ditto.
6242 (avx512f_fmaddsub_<mode>_mask<round_name>): Ditto.
6243 (avx512f_fmaddsub_<mode>_mask3<round_name>): Ditto.
6244 (avx512f_fmsubadd_<mode>_mask<round_name>): Ditto.
6245 (avx512f_fmsubadd_<mode>_mask3<round_name>): Ditto.
6246 (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
6247 (vec_extract_lo_<mode>_maskm): Ditto.
6248 (vec_extract_hi_<mode>_maskm): Ditto.
6249 (avx512f_vternlog<mode>_mask): Ditto.
6250 (avx512f_fixupimm<mode>_mask<round_saeonly_name>): Ditto.
6251 (avx512f_sfixupimm<mode>_mask<round_saeonly_name>): Ditto.
6252 (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
6253 (avx512f_<code>v8div16qi2_mask): Ditto.
6254 (avx512f_<code>v8div16qi2_mask_store): Ditto.
6255 (avx512f_eq<mode>3<mask_scalar_merge_name>_1): Ditto.
6256 (avx512f_gt<mode>3<mask_scalar_merge_name>): Ditto.
6257 (avx512f_testm<mode>3<mask_scalar_merge_name>): Ditto.
6258 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Ditto.
6259 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
6260 (*avx512pf_gatherpf<mode>df_mask): Ditto.
6261 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
6262 (*avx512pf_scatterpf<mode>df_mask): Ditto.
6263 (avx512cd_maskb_vec_dupv8di): Ditto.
6264 (avx512cd_maskw_vec_dupv16si): Ditto.
6265 (avx512f_vpermi2var<mode>3_maskz): Ditto.
6266 (avx512f_vpermi2var<mode>3_mask): Ditto.
6267 (avx512f_vpermi2var<mode>3_mask): Ditto.
6268 (avx512f_vpermt2var<mode>3_maskz): Ditto.
6269 (*avx512f_gathersi<mode>): Ditto.
6270 (*avx512f_gathersi<mode>_2): Ditto.
6271 (*avx512f_gatherdi<mode>): Ditto.
6272 (*avx512f_gatherdi<mode>_2): Ditto.
6273 (*avx512f_scattersi<mode>): Ditto.
6274 (*avx512f_scatterdi<mode>): Ditto.
6275 (avx512f_compress<mode>_mask): Ditto.
6276 (avx512f_compressstore<mode>_mask): Ditto.
6277 (avx512f_expand<mode>_mask): Ditto.
6278 * config/i386/subst.md (mask): Change k to Yk.
6279 (mask_scalar_merge): Ditto.
6282 2014-01-31 Marc Glisse <marc.glisse@inria.fr>
6284 * doc/extend.texi (Vector Extensions): Document ?: in C++.
6286 2014-01-31 Richard Biener <rguenther@suse.de>
6289 * builtins.c (fold_builtin_memory_op): Make sure to not
6290 use a floating-point mode or a boolean or enumeral type for
6293 2014-01-30 DJ Delorie <dj@redhat.com>
6295 * config/msp430/msp430.h (LIB_SPEC): Add -lcrt
6296 * config/msp430/msp430.md (msp430_refsym_need_exit): New.
6297 * config/msp430/msp430.c (msp430_expand_epilogue): Call it
6298 whenever main() has an epilogue.
6300 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6302 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Remove
6303 unused variable "field".
6304 * config/rs6000/vsx.md (vsx_mergel_<mode>): Add missing DONE.
6305 (vsx_mergeh_<mode>): Likewise.
6306 * config/rs6000/altivec.md (altivec_vmrghb): Likewise.
6307 (altivec_vmrghh): Likewise.
6308 (altivec_vmrghw): Likewise.
6309 (altivec_vmrglb): Likewise.
6310 (altivec_vmrglh): Likewise.
6311 (altivec_vmrglw): Likewise.
6312 (altivec_vspltb): Add missing uses.
6313 (altivec_vsplth): Likewise.
6314 (altivec_vspltw): Likewise.
6315 (altivec_vspltsf): Likewise.
6317 2014-01-30 Jakub Jelinek <jakub@redhat.com>
6320 * ifcvt.c (cond_exec_process_insns): Don't conditionalize
6321 frame related instructions.
6323 2014-01-30 Vladimir Makarov <vmakarov@redhat.com>
6325 PR rtl-optimization/59959
6326 * lra-constrains.c (simplify_operand_subreg): Assign NO_REGS to
6327 any reload of register whose subreg is invalid.
6329 2014-01-30 Jakub Jelinek <jakub@redhat.com>
6331 * config/i386/f16cintrin.h (_cvtsh_ss): Avoid -Wnarrowing warning.
6332 * config/i386/avx512fintrin.h (_mm512_mask_cvtusepi64_storeu_epi32):
6333 Add missing return type - void.
6335 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6337 * gcc/config/rs6000/rs6000.c (rs6000_expand_vector_init): Use
6338 gen_vsx_xxspltw_v4sf_direct instead of gen_vsx_xxspltw_v4sf;
6339 remove element index adjustment for endian (now handled in vsx.md
6341 (altivec_expand_vec_perm_const): Use
6342 gen_altivec_vsplt[bhw]_direct instead of gen_altivec_vsplt[bhw].
6343 * gcc/config/rs6000/vsx.md (UNSPEC_VSX_XXSPLTW): New unspec.
6344 (vsx_xxspltw_<mode>): Adjust element index for little endian.
6345 * gcc/config/rs6000/altivec.md (altivec_vspltb): Divide into a
6346 define_expand and a new define_insn *altivec_vspltb_internal;
6347 adjust for -maltivec=be on a little endian target.
6348 (altivec_vspltb_direct): New.
6349 (altivec_vsplth): Divide into a define_expand and a new
6350 define_insn *altivec_vsplth_internal; adjust for -maltivec=be on a
6351 little endian target.
6352 (altivec_vsplth_direct): New.
6353 (altivec_vspltw): Divide into a define_expand and a new
6354 define_insn *altivec_vspltw_internal; adjust for -maltivec=be on a
6355 little endian target.
6356 (altivec_vspltw_direct): New.
6357 (altivec_vspltsf): Divide into a define_expand and a new
6358 define_insn *altivec_vspltsf_internal; adjust for -maltivec=be on
6359 a little endian target.
6361 2014-01-30 Richard Biener <rguenther@suse.de>
6363 PR tree-optimization/59993
6364 * tree-ssa-forwprop.c (associate_pointerplus): Check we
6365 can propagate form the earlier stmt and avoid the transform
6366 when the intermediate result is needed.
6368 2014-01-30 Alangi Derick <alangiderick@gmail.com>
6370 * README.Portability: Fix typo.
6372 2014-01-30 David Holsgrove <david.holsgrove@xilinx.com>
6374 * config/microblaze/microblaze.md(cstoresf4, cbranchsf4): Replace
6375 comparison_operator with ordered_comparison_operator.
6377 2014-01-30 Nick Clifton <nickc@redhat.com>
6379 * config/mn10300/mn10300-protos.h (mn10300_store_multiple_operation_p):
6380 Rename to mn10300_store_multiple_regs.
6381 * config/mn10300/mn10300.c: Likewise.
6382 * config/mn10300/mn10300.md (store_movm): Fix typo: call
6383 store_multiple_regs.
6384 * config/mn10300/predicates.md (mn10300_store_multiple_operation):
6385 Call mn10300_store_multiple_regs.
6387 2014-01-30 Nick Clifton <nickc@redhat.com>
6388 DJ Delorie <dj@redhat.com>
6390 * config/rl78/rl78.c (register_sizes): Make the "upper half" of
6391 %fp 2 to keep registers after it properly word-aligned.
6392 (rl78_alloc_physical_registers_umul): Handle the case where both
6393 input operands are the same.
6395 2014-01-30 Richard Biener <rguenther@suse.de>
6397 PR tree-optimization/59903
6398 * tree-vect-loop.c (vect_transform_loop): Guard multiple-types
6401 2014-01-30 Jason Merrill <jason@redhat.com>
6404 * tree.c (walk_type_fields): Handle VECTOR_TYPE.
6407 * cgraphunit.c (expand_thunk): Copy volatile arg to a temporary.
6409 2014-01-30 Richard Biener <rguenther@suse.de>
6411 PR tree-optimization/59951
6412 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Skip uses in debug insns.
6414 2014-01-30 Savin Zlobec <savin.zlobec@gmail.com>
6417 * config/nios2/nios2.c (nios2_fpu_insn_asm): Fix asm output of
6418 SFmode to DFmode case.
6420 2014-01-29 DJ Delorie <dj@redhat.com>
6422 * config/msp430/msp430.opt (-minrt): New.
6423 * config/msp430/msp430.h (STARTFILE_SPEC): Link alternate runtime
6425 (ENDFILE_SPEC): Likewise.
6427 2014-01-29 Jan Hubicka <hubicka@ucw.cz>
6429 * ipa-inline-analysis.c (clobber_only_eh_bb_p): New function.
6430 (estimate_function_body_sizes): Use it.
6432 2014-01-29 Paolo Carlini <paolo.carlini@oracle.com>
6435 * dwarf2out.c (is_cxx_auto): New.
6436 (is_base_type): Use it.
6437 (gen_type_die_with_usage): Likewise.
6439 2014-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6441 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Use
6442 CODE_FOR_altivec_vmrg*_direct rather than CODE_FOR_altivec_vmrg*.
6443 * config/rs6000/vsx.md (vsx_mergel_<mode>): Adjust for
6444 -maltivec=be with LE targets.
6445 (vsx_mergeh_<mode>): Likewise.
6446 * config/rs6000/altivec.md (UNSPEC_VMRG[HL]_DIRECT): New unspecs.
6447 (mulv8hi3): Use gen_altivec_vmrg[hl]w_direct.
6448 (altivec_vmrghb): Replace with define_expand and new
6449 *altivec_vmrghb_internal insn; adjust for -maltivec=be with LE targets.
6450 (altivec_vmrghb_direct): New define_insn.
6451 (altivec_vmrghh): Replace with define_expand and new
6452 *altivec_vmrghh_internal insn; adjust for -maltivec=be with LE targets.
6453 (altivec_vmrghh_direct): New define_insn.
6454 (altivec_vmrghw): Replace with define_expand and new
6455 *altivec_vmrghw_internal insn; adjust for -maltivec=be with LE targets.
6456 (altivec_vmrghw_direct): New define_insn.
6457 (*altivec_vmrghsf): Adjust for endianness.
6458 (altivec_vmrglb): Replace with define_expand and new
6459 *altivec_vmrglb_internal insn; adjust for -maltivec=be with LE targets.
6460 (altivec_vmrglb_direct): New define_insn.
6461 (altivec_vmrglh): Replace with define_expand and new
6462 *altivec_vmrglh_internal insn; adjust for -maltivec=be with LE targets.
6463 (altivec_vmrglh_direct): New define_insn.
6464 (altivec_vmrglw): Replace with define_expand and new
6465 *altivec_vmrglw_internal insn; adjust for -maltivec=be with LE targets.
6466 (altivec_vmrglw_direct): New define_insn.
6467 (*altivec_vmrglsf): Adjust for endianness.
6468 (vec_widen_umult_hi_v16qi): Use gen_altivec_vmrghh_direct.
6469 (vec_widen_umult_lo_v16qi): Use gen_altivec_vmrglh_direct.
6470 (vec_widen_smult_hi_v16qi): Use gen_altivec_vmrghh_direct.
6471 (vec_widen_smult_lo_v16qi): Use gen_altivec_vmrglh_direct.
6472 (vec_widen_umult_hi_v8hi): Use gen_altivec_vmrghw_direct.
6473 (vec_widen_umult_lo_v8hi): Use gen_altivec_vmrglw_direct.
6474 (vec_widen_smult_hi_v8hi): Use gen_altivec_vmrghw_direct.
6475 (vec_widen_smult_lo_v8hi): Use gen_altivec_vmrglw_direct.
6477 2014-01-29 Marcus Shawcroft <marcus.shawcroft@arm.com>
6479 * config/aarch64/aarch64.c (aarch64_expand_mov_immediate)
6480 (aarch64_legitimate_address_p, aarch64_class_max_nregs): Adjust
6483 2014-01-29 Richard Biener <rguenther@suse.de>
6485 PR tree-optimization/58742
6486 * tree-ssa-forwprop.c (associate_pointerplus): Rename to
6487 associate_pointerplus_align.
6488 (associate_pointerplus_diff): New function.
6489 (associate_pointerplus): Likewise. Call associate_pointerplus_align
6490 and associate_pointerplus_diff.
6492 2014-01-29 Richard Biener <rguenther@suse.de>
6494 * lto-streamer.h (LTO_major_version): Bump to 3.
6495 (LTO_minor_version): Reset to 0.
6497 2014-01-29 Renlin Li <Renlin.Li@arm.com>
6499 * config/arm/arm-arches.def (ARM_ARCH): Add armv7ve arch.
6500 * config/arm/arm.c (FL_FOR_ARCH7VE): New.
6501 (arm_file_start): Generate correct asm header for armv7ve.
6502 * config/arm/bpabi.h: Add multilib support for armv7ve.
6503 * config/arm/driver-arm.c: Change the architectures of cortex-a7
6504 and cortex-a15 to armv7ve.
6505 * config/arm/t-aprofile: Add multilib support for armv7ve.
6506 * doc/invoke.texi: Document -march=armv7ve.
6508 2014-01-29 Richard Biener <rguenther@suse.de>
6510 PR tree-optimization/58742
6511 * tree-ssa-forwprop.c (associate_plusminus): Return true
6512 if we changed sth, defer EH cleanup to ...
6513 (ssa_forward_propagate_and_combine): ... here. Call simplify_mult.
6514 (simplify_mult): New function.
6516 2014-01-29 Jakub Jelinek <jakub@redhat.com>
6519 PR tree-optimization/59920
6520 * tree.c (build_common_builtin_nodes): Remove
6521 __builtin_setjmp_dispatcher initialization.
6522 * omp-low.h (make_gimple_omp_edges): Add a new int * argument.
6523 * profile.c (branch_prob): Use gsi_start_nondebug_after_labels_bb
6524 instead of gsi_after_labels + manually skipping debug stmts.
6525 Don't ignore bbs with BUILT_IN_SETJMP_DISPATCHER, instead
6526 ignore bbs with IFN_ABNORMAL_DISPATCHER.
6527 * tree-inline.c (copy_edges_for_bb): Remove
6528 can_make_abnormal_goto argument, instead add abnormal_goto_dest
6529 argument. Ignore computed_goto_p stmts. Don't call
6530 make_abnormal_goto_edges. If a call might need abnormal edges
6531 for non-local gotos, see if it already has an edge to
6532 IFN_ABNORMAL_DISPATCHER or if it is IFN_ABNORMAL_DISPATCHER
6533 with true argument, don't do anything then, otherwise add
6534 EDGE_ABNORMAL from the call's bb to abnormal_goto_dest.
6535 (copy_cfg_body): Compute abnormal_goto_dest, adjust copy_edges_for_bb
6537 * gimple-low.c (struct lower_data): Remove calls_builtin_setjmp.
6538 (lower_function_body): Don't emit __builtin_setjmp_dispatcher.
6539 (lower_stmt): Don't set data->calls_builtin_setjmp.
6540 (lower_builtin_setjmp): Adjust comment.
6541 * builtins.def (BUILT_IN_SETJMP_DISPATCHER): Remove.
6542 * tree-cfg.c (found_computed_goto): Remove.
6543 (factor_computed_gotos): Remove.
6544 (make_goto_expr_edges): Return bool, true for computed gotos.
6545 Don't call make_abnormal_goto_edges.
6546 (build_gimple_cfg): Don't set found_computed_goto, don't call
6547 factor_computed_gotos.
6548 (computed_goto_p): No longer static.
6549 (make_blocks): Don't set found_computed_goto.
6550 (get_abnormal_succ_dispatcher, handle_abnormal_edges): New functions.
6551 (make_edges): If make_goto_expr_edges returns true, push bb
6552 into ab_edge_goto vector, for stmt_can_make_abnormal_goto calls
6553 instead of calling make_abnormal_goto_edges push bb into ab_edge_call
6554 vector. Record mapping between bbs and OpenMP regions if there
6555 are any, adjust make_gimple_omp_edges caller. Call
6556 handle_abnormal_edges.
6557 (make_abnormal_goto_edges): Remove.
6558 * tree-cfg.h (make_abnormal_goto_edges): Remove.
6559 (computed_goto_p, get_abnormal_succ_dispatcher): New prototypes.
6560 * internal-fn.c (expand_ABNORMAL_DISPATCHER): New function.
6561 * builtins.c (expand_builtin): Don't handle BUILT_IN_SETJMP_DISPATCHER.
6562 * internal-fn.def (ABNORMAL_DISPATCHER): New.
6563 * omp-low.c (make_gimple_omp_edges): Add region_idx argument, when
6564 filling *region also set *region_idx to (*region)->entry->index.
6567 * read-rtl.c (read_rtx_code): Clear all of RTX_CODE_SIZE (code).
6568 For REGs set ORIGINAL_REGNO.
6570 2014-01-29 Bingfeng Mei <bmei@broadcom.com>
6572 * doc/md.texi: Mention that a target shouldn't implement
6573 vec_widen_(s|u)mul_even/odd pair if it is less efficient
6576 2014-01-29 Jakub Jelinek <jakub@redhat.com>
6578 PR tree-optimization/59594
6579 * tree-vect-data-refs.c (vect_analyze_data_ref_accesses): Sort
6580 a copy of the datarefs vector rather than the vector itself.
6582 2014-01-28 Jason Merrill <jason@redhat.com>
6585 * dwarf2out.c (auto_die): New static.
6586 (gen_type_die_with_usage): Handle C++1y 'auto'.
6587 (gen_subprogram_die): If in-class DIE had 'auto', emit type again
6590 2014-01-28 H.J. Lu <hongjiu.lu@intel.com>
6593 * config/i386/gnu-user64.h (SPEC_32): Add "m16|" to "m32".
6594 (SPEC_X32): Likewise.
6595 (SPEC_64): Likewise.
6596 * config/i386/i386.c (ix86_option_override_internal): Turn off
6597 OPTION_MASK_ISA_64BIT, OPTION_MASK_ABI_X32 and OPTION_MASK_ABI_64
6599 (x86_file_start): Output .code16gcc for TARGET_16BIT.
6600 * config/i386/i386.h (TARGET_16BIT): New macro.
6601 (TARGET_16BIT_P): Likewise.
6602 * config/i386/i386.opt: Add m16.
6603 * doc/invoke.texi: Document -m16.
6605 2014-01-28 Jakub Jelinek <jakub@redhat.com>
6607 PR preprocessor/59935
6608 * input.c (location_get_source_line): Bail out on when line number
6609 is zero, and test the return value of lookup_or_add_file_to_cache_tab.
6611 2014-01-28 Richard Biener <rguenther@suse.de>
6613 PR tree-optimization/58742
6614 * tree-ssa-forwprop.c (associate_plusminus): Handle
6615 pointer subtraction of the form (T)(P + A) - (T)P.
6617 2014-01-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6619 * config/arm/arm.c (arm_new_rtx_costs): Remove useless statement
6622 2014-01-28 Richard Biener <rguenther@suse.de>
6625 2014-01-28 Richard Biener <rguenther@suse.de>
6627 PR rtl-optimization/45364
6628 PR rtl-optimization/59890
6629 * var-tracking.c (local_get_addr_clear_given_value): Handle
6630 already cleared slot.
6631 (val_reset): Handle not allocated local_get_addr_cache.
6632 (vt_find_locations): Use post-order on the inverted CFG.
6634 2014-01-28 Richard Biener <rguenther@suse.de>
6636 * tree-data-ref.h (ddr_is_anti_dependent, ddrs_have_anti_deps): Remove.
6638 2014-01-28 Richard Biener <rguenther@suse.de>
6640 PR rtl-optimization/45364
6641 PR rtl-optimization/59890
6642 * var-tracking.c (local_get_addr_clear_given_value): Handle
6643 already cleared slot.
6644 (val_reset): Handle not allocated local_get_addr_cache.
6645 (vt_find_locations): Use post-order on the inverted CFG.
6647 2014-01-28 Alan Modra <amodra@gmail.com>
6649 * Makefile.in (BUILD_CPPFLAGS): Do not use ALL_CPPFLAGS.
6650 * configure.ac <recursive call for build != host>: Define
6651 GENERATOR_FILE. Comment. Use CXX_FOR_BUILD, CXXFLAGS_FOR_BUILD
6652 and LD_FOR_BUILD too.
6653 * configure: Regenerate.
6655 2014-01-27 Allan Sandfeld Jensen <sandfeld@kde.org>
6657 * config/i386/i386.c (get_builtin_code_for_version): Separate
6658 Westmere from Nehalem, Ivy Bridge from Sandy Bridge and
6659 Broadwell from Haswell.
6661 2014-01-27 Steve Ellcey <sellcey@mips.com>
6663 * common/config/mips/mips-common.c (TARGET_DEFAULT_TARGET_FLAGS):
6664 Remove TARGET_FP_EXCEPTIONS_DEFAULT and MASK_FUSED_MADD.
6665 * config/mips/mips.c (mips_option_override): Change setting
6667 * config/mips/mips.h (TARGET_FP_EXCEPTIONS_DEFAULT): Remove.
6668 * config/mips/mips.opt (DSP, DSPR2, FP_EXCEPTIONS, FUSED_MADD, MIPS3D):
6669 Change from Mask to Var.
6671 2014-01-27 Jeff Law <law@redhat.com>
6673 * ipa-inline.c (inline_small_functions): Fix typo.
6675 2014-01-27 Ilya Tocar <ilya.tocar@intel.com>
6677 * config/i386/avx512fintrin.h (_mm512_mask_cvtepi32_storeu_epi8): New.
6678 (_mm512_mask_cvtsepi32_storeu_epi8): Ditto.
6679 (_mm512_mask_cvtusepi32_storeu_epi8): Ditto.
6680 (_mm512_mask_cvtepi32_storeu_epi16): Ditto.
6681 (_mm512_mask_cvtsepi32_storeu_epi16): Ditto.
6682 (_mm512_mask_cvtusepi32_storeu_epi16): Ditto.
6683 (_mm512_mask_cvtepi64_storeu_epi32): Ditto.
6684 (_mm512_mask_cvtsepi64_storeu_epi32): Ditto.
6685 (_mm512_mask_cvtusepi64_storeu_epi32): Ditto.
6686 (_mm512_mask_cvtepi64_storeu_epi16): Ditto.
6687 (_mm512_mask_cvtsepi64_storeu_epi16): Ditto.
6688 (_mm512_mask_cvtusepi64_storeu_epi16): Ditto.
6689 (_mm512_mask_cvtepi64_storeu_epi8): Ditto.
6690 (_mm512_mask_cvtsepi64_storeu_epi8): Ditto.
6691 (_mm512_mask_cvtusepi64_storeu_epi8): Ditto.
6692 (_mm512_storeu_epi64): Ditto.
6693 (_mm512_cmpge_epi32_mask): Ditto.
6694 (_mm512_cmpge_epu32_mask): Ditto.
6695 (_mm512_cmpge_epi64_mask): Ditto.
6696 (_mm512_cmpge_epu64_mask): Ditto.
6697 (_mm512_cmple_epi32_mask): Ditto.
6698 (_mm512_cmple_epu32_mask): Ditto.
6699 (_mm512_cmple_epi64_mask): Ditto.
6700 (_mm512_cmple_epu64_mask): Ditto.
6701 (_mm512_cmplt_epi32_mask): Ditto.
6702 (_mm512_cmplt_epu32_mask): Ditto.
6703 (_mm512_cmplt_epi64_mask): Ditto.
6704 (_mm512_cmplt_epu64_mask): Ditto.
6705 (_mm512_cmpneq_epi32_mask): Ditto.
6706 (_mm512_cmpneq_epu32_mask): Ditto.
6707 (_mm512_cmpneq_epi64_mask): Ditto.
6708 (_mm512_cmpneq_epu64_mask): Ditto.
6709 (_mm512_expand_pd): Ditto.
6710 (_mm512_expand_ps): Ditto.
6711 * config/i386/i386-builtin-types.def: Add PV16QI, PV16QI, PV16HI,
6712 VOID_PV8SI_V8DI_QI, VOID_PV8HI_V8DI_QI, VOID_PV16QI_V8DI_QI,
6713 VOID_PV16QI_V16SI_HI, VOID_PV16HI_V16SI_HI.
6714 * config/i386/i386.c (ix86_builtins): Add
6715 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK,
6716 IX86_BUILTIN_PMOVDB512_MEM, IX86_BUILTIN_PMOVDW512_MEM,
6717 IX86_BUILTIN_PMOVQB512_MEM, IX86_BUILTIN_PMOVQD512_MEM,
6718 IX86_BUILTIN_PMOVQW512_MEM, IX86_BUILTIN_PMOVSDB512_MEM,
6719 IX86_BUILTIN_PMOVSDW512_MEM, IX86_BUILTIN_PMOVSQB512_MEM,
6720 IX86_BUILTIN_PMOVSQD512_MEM, IX86_BUILTIN_PMOVSQW512_MEM,
6721 IX86_BUILTIN_PMOVUSDB512_MEM, IX86_BUILTIN_PMOVUSDW512_MEM,
6722 IX86_BUILTIN_PMOVUSQB512_MEM, IX86_BUILTIN_PMOVUSQD512_MEM,
6723 IX86_BUILTIN_PMOVUSQW512_MEM.
6724 (bdesc_special_args): Add __builtin_ia32_pmovusqd512mem_mask,
6725 __builtin_ia32_pmovsqd512mem_mask,
6726 __builtin_ia32_pmovqd512mem_mask,
6727 __builtin_ia32_pmovusqw512mem_mask,
6728 __builtin_ia32_pmovsqw512mem_mask,
6729 __builtin_ia32_pmovqw512mem_mask,
6730 __builtin_ia32_pmovusdw512mem_mask,
6731 __builtin_ia32_pmovsdw512mem_mask,
6732 __builtin_ia32_pmovdw512mem_mask,
6733 __builtin_ia32_pmovqb512mem_mask,
6734 __builtin_ia32_pmovusqb512mem_mask,
6735 __builtin_ia32_pmovsqb512mem_mask,
6736 __builtin_ia32_pmovusdb512mem_mask,
6737 __builtin_ia32_pmovsdb512mem_mask,
6738 __builtin_ia32_pmovdb512mem_mask.
6739 (bdesc_args): Add __builtin_ia32_expanddf512,
6740 __builtin_ia32_expandsf512.
6741 (ix86_expand_special_args_builtin): Handle VOID_FTYPE_PV8SI_V8DI_QI,
6742 VOID_FTYPE_PV8HI_V8DI_QI, VOID_FTYPE_PV16HI_V16SI_HI,
6743 VOID_FTYPE_PV16QI_V8DI_QI, VOID_FTYPE_PV16QI_V16SI_HI.
6744 * config/i386/sse.md (unspec): Add UNSPEC_EXPAND_NOMASK.
6745 (avx512f_<code><pmov_src_lower><mode>2_mask_store): New.
6746 (*avx512f_<code>v8div16qi2_store_mask): Renamed to ...
6747 (avx512f_<code>v8div16qi2_mask_store): This.
6748 (avx512f_expand<mode>): New.
6750 2014-01-27 Kirill Yukhin <kirill.yukhin@intel.com>
6752 * config/i386/avx512pfintrin.h (_mm512_mask_prefetch_i32gather_pd):
6754 (_mm512_mask_prefetch_i64gather_pd): Ditto.
6755 (_mm512_prefetch_i32scatter_pd): Ditto.
6756 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
6757 (_mm512_prefetch_i64scatter_pd): Ditto.
6758 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
6759 (_mm512_mask_prefetch_i32gather_ps): Fix operand type.
6760 (_mm512_mask_prefetch_i64gather_ps): Ditto.
6761 (_mm512_prefetch_i32scatter_ps): Ditto.
6762 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
6763 (_mm512_prefetch_i64scatter_ps): Ditto.
6764 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
6765 * config/i386/i386-builtin-types.def: Define
6766 VOID_FTYPE_QI_V8SI_PCINT64_INT_INT
6767 and VOID_FTYPE_QI_V8DI_PCINT64_INT_INT.
6768 * config/i386/i386.c (ix86_builtins): Define IX86_BUILTIN_GATHERPFQPD,
6769 IX86_BUILTIN_GATHERPFDPD, IX86_BUILTIN_SCATTERPFDPD,
6770 IX86_BUILTIN_SCATTERPFQPD.
6771 (ix86_init_mmx_sse_builtins): Define __builtin_ia32_gatherpfdpd,
6772 __builtin_ia32_gatherpfdps, __builtin_ia32_gatherpfqpd,
6773 __builtin_ia32_gatherpfqps, __builtin_ia32_scatterpfdpd,
6774 __builtin_ia32_scatterpfdps, __builtin_ia32_scatterpfqpd,
6775 __builtin_ia32_scatterpfqps.
6776 (ix86_expand_builtin): Expand new built-ins.
6777 * config/i386/sse.md (avx512pf_gatherpf<mode>): Add SF suffix,
6778 fix memory access data type.
6779 (*avx512pf_gatherpf<mode>_mask): Ditto.
6780 (*avx512pf_gatherpf<mode>): Ditto.
6781 (avx512pf_scatterpf<mode>): Ditto.
6782 (*avx512pf_scatterpf<mode>_mask): Ditto.
6783 (*avx512pf_scatterpf<mode>): Ditto.
6784 (GATHER_SCATTER_SF_MEM_MODE): New.
6785 (avx512pf_gatherpf<mode>df): Ditto.
6786 (*avx512pf_gatherpf<mode>df_mask): Ditto.
6787 (*avx512pf_scatterpf<mode>df): Ditto.
6789 2014-01-27 Jakub Jelinek <jakub@redhat.com>
6792 * expmed.h (expmed_mode_index): Rework so that analysis and optimziers
6793 know when the MODE_PARTIAL_INT and MODE_VECTOR_INT cases can never be
6796 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
6798 * common/config/arm/arm-common.c
6799 (arm_rewrite_mcpu): Handle multiple names.
6801 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
6803 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
6805 * gimple-builder.h (create_gimple_tmp): Delete.
6807 2014-01-27 Christian Bruel <christian.bruel@st.com>
6809 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Fix remaining bytes after
6812 2014-01-26 John David Anglin <danglin@gcc.gnu.org>
6814 * config/pa/pa.md (call): Generate indirect long calls to non-local
6815 functions when outputing 32-bit code.
6816 (call_value): Likewise except for special call to buggy powf function.
6818 * config/pa/pa.c (pa_attr_length_indirect_call): Adjust length of
6819 portable runtime and PIC indirect calls.
6820 (pa_output_indirect_call): Remove unnecessary nop from portable runtime
6821 and PIC call sequences. Use ldo instead of blr to set return register
6822 in PIC call sequence.
6824 2014-01-25 Walter Lee <walt@tilera.com>
6826 * config/tilegx/sync.md (atomic_fetch_sub): Fix negation and
6827 avoid clobbering a live register.
6829 2014-01-25 Walter Lee <walt@tilera.com>
6831 * config/tilegx/tilegx-c.c (tilegx_cpu_cpp_builtins):
6832 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2}.
6833 * config/tilegx/tilepro-c.c (tilepro_cpu_cpp_builtins):
6834 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2,4,8}.
6836 2014-01-25 Walter Lee <walt@tilera.com>
6838 * config/tilegx/tilegx.c (tilegx_function_arg): Start 16-byte
6839 arguments on even registers.
6840 (tilegx_gimplify_va_arg_expr): Align 16-byte var args to
6842 * config/tilegx/tilegx.h (STACK_BOUNDARY): Change to 16 bytes.
6843 (BIGGEST_ALIGNMENT): Ditto.
6844 (BIGGEST_FIELD_ALIGNMENT): Ditto.
6846 2014-01-25 Walter Lee <walt@tilera.com>
6848 * config/tilegx/tilegx.c (tilegx_gen_bundles): Delete barrier
6849 insns before bundling.
6850 * config/tilegx/tilegx.md (tile_network_barrier): Update comment.
6852 2014-01-25 Walter Lee <walt@tilera.com>
6854 * config/tilegx/tilegx.c (tilegx_expand_builtin): Set
6855 PREFETCH_SCHEDULE_BARRIER_P to true for prefetches.
6856 * config/tilepro/tilepro.c (tilepro_expand_builtin): Ditto.
6858 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
6860 * config/mips/constraints.md (kl): Delete.
6861 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Turn into
6862 define expands, using...
6863 (divmod<mode>4_mips16, udivmod<mode>4_mips16): ...these new
6864 instructions for MIPS16.
6865 (*divmod<mode>4, *udivmod<mode>4): New patterns, taken from the
6866 non-MIPS16 version of the old divmod<mode>4 and udivmod<mode>4.
6868 2014-01-25 Walter Lee <walt@tilera.com>
6870 * config/tilepro/tilepro.md (ctzdi2): Use register_operand predicate.
6874 2014-01-25 Walter Lee <walt@tilera.com>
6876 * config/tilegx/tilegx.c (tilegx_expand_to_rtl_hook): New.
6877 (TARGET_EXPAND_TO_RTL_HOOK): Define.
6879 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
6881 * rtlanal.c (canonicalize_condition): Split out duplicated mode check.
6884 2014-01-25 Jakub Jelinek <jakub@redhat.com>
6886 * print-rtl.c (in_call_function_usage): New var.
6887 (print_rtx): When in CALL_INSN_FUNCTION_USAGE, always print
6888 EXPR_LIST mode as mode and not as reg note name.
6891 * cfgloopmanip.c (copy_loop_info): If
6892 loop->warned_aggressive_loop_optimizations, make sure
6893 the flag is set in target loop too.
6895 2014-01-24 Balaji V. Iyer <balaji.v.iyer@intel.com>
6897 * builtins.c (is_builtin_name): Renamed flag_enable_cilkplus to
6899 * builtins.def: Likewise.
6900 * cilk.h (fn_contains_cilk_spawn_p): Likewise.
6901 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Likewise.
6902 * ira.c (ira_setup_eliminable_regset): Likewise.
6903 * omp-low.c (gate_expand_omp): Likewise.
6904 (execute_lower_omp): Likewise.
6905 (diagnose_sb_0): Likewise.
6906 (gate_diagnose_omp_blocks): Likewise.
6907 (simd_clone_clauses_extract): Likewise.
6910 2014-01-24 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6912 * config/rs6000/rs6000.c (rs6000_expand_vec_perm_const_1): Remove
6913 correction for little endian...
6914 * config/rs6000/vsx.md (vsx_xxpermdi2_<mode>_1): ...and move it to
6917 2014-01-24 Jeff Law <law@redhat.com>
6919 PR tree-optimization/59919
6920 * tree-vrp.c (find_assert_locations_1): Do not register asserts
6921 for non-returning calls.
6923 2014-01-24 James Greenhalgh <james.greenhalgh@arm.com>
6925 * common/config/aarch64/aarch64-common.c
6926 (aarch64_rewrite_mcpu): Handle multiple names.
6927 * config/aarch64/aarch64.h
6928 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
6930 2014-01-24 Dodji Seketeli <dodji@redhat.com>
6932 * input.c (add_file_to_cache_tab): Handle the case where fopen
6935 2014-01-23 H.J. Lu <hongjiu.lu@intel.com>
6938 * config/i386/i386.md (pushsf splitter): Get stack adjustment
6939 from push operand if code of push isn't PRE_DEC.
6941 2014-01-23 Michael Meissner <meissner@linux.vnet.ibm.com>
6944 * doc/invoke.texi (RS/6000 and PowerPC Options): Document
6945 -mquad-memory-atomic. Update -mquad-memory documentation to say
6946 it is only used for non-atomic loads/stores.
6948 * config/rs6000/predicates.md (quad_int_reg_operand): Allow either
6949 -mquad-memory or -mquad-memory-atomic switches.
6951 * config/rs6000/rs6000-cpus.def (ISA_2_7_MASKS_SERVER): Add
6952 -mquad-memory-atomic to ISA 2.07 support.
6954 * config/rs6000/rs6000.opt (-mquad-memory-atomic): Add new switch
6955 to separate support of normal quad word memory operations (ldq, stq)
6956 from the atomic quad word memory operations.
6958 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
6959 support to separate non-atomic quad word operations from atomic
6960 quad word operations. Disable non-atomic quad word operations in
6961 little endian mode so that we don't have to swap words after the
6962 load and before the store.
6963 (quad_load_store_p): Add comment about atomic quad word support.
6964 (rs6000_opt_masks): Add -mquad-memory-atomic to the list of
6965 options printed with -mdebug=reg.
6967 * config/rs6000/rs6000.h (TARGET_SYNC_TI): Use
6968 -mquad-memory-atomic as the test for whether we have quad word
6969 atomic instructions.
6970 (TARGET_SYNC_HI_QI): If either -mquad-memory-atomic, -mquad-memory,
6971 or -mp8-vector are used, allow byte/half-word atomic operations.
6973 * config/rs6000/sync.md (load_lockedti): Insure that the address
6974 is a proper indexed or indirect address for the lqarx instruction.
6975 On little endian systems, swap the hi/lo registers after the lqarx
6977 (load_lockedpti): Use indexed_or_indirect_operand predicate to
6978 insure the address is valid for the lqarx instruction.
6979 (store_conditionalti): Insure that the address is a proper indexed
6980 or indirect address for the stqcrx. instruction. On little endian
6981 systems, swap the hi/lo registers before doing the stqcrx.
6983 (store_conditionalpti): Use indexed_or_indirect_operand predicate to
6984 insure the address is valid for the stqcrx. instruction.
6986 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
6987 Define __QUAD_MEMORY__ and __QUAD_MEMORY_ATOMIC__ based on what
6988 type of quad memory support is available.
6990 2014-01-23 Vladimir Makarov <vmakarov@redhat.com>
6993 * lra-constraints.c (simplify_operand_subreg): Spill pseudo if
6994 there is a danger of looping.
6996 2014-01-23 Pat Haugen <pthaugen@us.ibm.com>
6998 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
6999 force flag_ira_loop_pressure if set via command line.
7001 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
7003 * config/aarch64/aarch64-simd-builtins.def (ashr): DI mode removed.
7004 (ashr_simd): New builtin handling DI mode.
7005 * config/aarch64/aarch64-simd.md (aarch64_ashr_simddi): New pattern.
7006 (aarch64_sshr_simddi): New match pattern.
7007 * config/aarch64/arm_neon.h (vshr_n_s32): Builtin call modified.
7008 (vshrd_n_s64): Likewise.
7009 * config/aarch64/predicates.md (aarch64_shift_imm64_di): New predicate.
7011 2014-01-23 Nick Clifton <nickc@redhat.com>
7013 * config/msp430/msp430.h (ASM_SPEC): Pass the -mcpu as -mcpu.
7014 (LIB_SPEC): Drop use of memory.ld and peripherals.ld scripts in
7015 favour of mcu specific scripts.
7016 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more matches for
7019 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
7020 Alex Velenko <Alex.Velenko@arm.com>
7022 * config/aarch64/arm_neon.h (vaddv_s8): __LANE0 cleanup.
7023 (vaddv_s16): Likewise.
7024 (vaddv_s32): Likewise.
7025 (vaddv_u8): Likewise.
7026 (vaddv_u16): Likewise.
7027 (vaddv_u32): Likewise.
7028 (vaddvq_s8): Likewise.
7029 (vaddvq_s16): Likewise.
7030 (vaddvq_s32): Likewise.
7031 (vaddvq_s64): Likewise.
7032 (vaddvq_u8): Likewise.
7033 (vaddvq_u16): Likewise.
7034 (vaddvq_u32): Likewise.
7035 (vaddvq_u64): Likewise.
7036 (vaddv_f32): Likewise.
7037 (vaddvq_f32): Likewise.
7038 (vaddvq_f64): Likewise.
7039 (vmaxv_f32): Likewise.
7040 (vmaxv_s8): Likewise.
7041 (vmaxv_s16): Likewise.
7042 (vmaxv_s32): Likewise.
7043 (vmaxv_u8): Likewise.
7044 (vmaxv_u16): Likewise.
7045 (vmaxv_u32): Likewise.
7046 (vmaxvq_f32): Likewise.
7047 (vmaxvq_f64): Likewise.
7048 (vmaxvq_s8): Likewise.
7049 (vmaxvq_s16): Likewise.
7050 (vmaxvq_s32): Likewise.
7051 (vmaxvq_u8): Likewise.
7052 (vmaxvq_u16): Likewise.
7053 (vmaxvq_u32): Likewise.
7054 (vmaxnmv_f32): Likewise.
7055 (vmaxnmvq_f32): Likewise.
7056 (vmaxnmvq_f64): Likewise.
7057 (vminv_f32): Likewise.
7058 (vminv_s8): Likewise.
7059 (vminv_s16): Likewise.
7060 (vminv_s32): Likewise.
7061 (vminv_u8): Likewise.
7062 (vminv_u16): Likewise.
7063 (vminv_u32): Likewise.
7064 (vminvq_f32): Likewise.
7065 (vminvq_f64): Likewise.
7066 (vminvq_s8): Likewise.
7067 (vminvq_s16): Likewise.
7068 (vminvq_s32): Likewise.
7069 (vminvq_u8): Likewise.
7070 (vminvq_u16): Likewise.
7071 (vminvq_u32): Likewise.
7072 (vminnmv_f32): Likewise.
7073 (vminnmvq_f32): Likewise.
7074 (vminnmvq_f64): Likewise.
7076 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
7078 * config/aarch64/aarch64-simd.md
7079 (aarch64_dup_lane<mode>): Correct lane number on big-endian.
7080 (aarch64_dup_lane_<vswap_widthi_name><mode>): Likewise.
7081 (*aarch64_mul3_elt<mode>): Likewise.
7082 (*aarch64_mul3_elt<vswap_width_name><mode>): Likewise.
7083 (*aarch64_mul3_elt_to_64v2df): Likewise.
7084 (*aarch64_mla_elt<mode>): Likewise.
7085 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
7086 (*aarch64_mls_elt<mode>): Likewise.
7087 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
7088 (*aarch64_fma4_elt<mode>): Likewise.
7089 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
7090 (*aarch64_fma4_elt_to_64v2df): Likewise.
7091 (*aarch64_fnma4_elt<mode>): Likewise.
7092 (*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
7093 (*aarch64_fnma4_elt_to_64v2df): Likewise.
7094 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
7095 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
7096 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
7097 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
7098 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
7099 (aarch64_sqdmull_lane<mode>_internal): Likewise.
7100 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
7102 2013-01-23 Alex Velenko <Alex.Velenko@arm.com>
7104 * config/aarch64/aarch64-simd.md
7105 (aarch64_be_checked_get_lane<mode>): New define_expand.
7106 * config/aarch64/aarch64-simd-builtins.def
7107 (BUILTIN_VALL (GETLANE, be_checked_get_lane, 0)):
7108 New builtin definition.
7109 * config/aarch64/arm_neon.h: (__aarch64_vget_lane_any):
7110 Use new safe be builtin.
7112 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
7114 * config/aarch64/aarch64-simd.md (aarch64_be_ld1<mode>):
7116 (aarch64_be_st1<mode>): Likewise.
7117 (aarch_ld1<VALL:mode>): Define_expand modified.
7118 (aarch_st1<VALL:mode>): Likewise.
7119 * config/aarch64/aarch64.md (UNSPEC_LD1): New unspec definition.
7120 (UNSPEC_ST1): Likewise.
7122 2014-01-23 David Holsgrove <david.holsgrove@xilinx.com>
7124 * config/microblaze/microblaze.md: Add trap insn and attribute
7126 2014-01-23 Dodji Seketeli <dodji@redhat.com>
7128 PR preprocessor/58580
7129 * input.h (location_get_source_line): Take an additional line_size
7131 (void diagnostics_file_cache_fini): Declare new function.
7132 * input.c (struct fcache): New type.
7133 (fcache_tab_size, fcache_buffer_size, fcache_line_record_size):
7134 New static constants.
7135 (diagnostic_file_cache_init, total_lines_num)
7136 (lookup_file_in_cache_tab, evicted_cache_tab_entry)
7137 (add_file_to_cache_tab, lookup_or_add_file_to_cache_tab)
7138 (needs_read, needs_grow, maybe_grow, read_data, maybe_read_data)
7139 (get_next_line, read_next_line, goto_next_line, read_line_num):
7140 New static function definitions.
7141 (diagnostic_file_cache_fini): New function.
7142 (location_get_source_line): Take an additional output line_len
7143 parameter. Re-write using lookup_or_add_file_to_cache_tab and
7145 * diagnostic.c (diagnostic_finish): Call
7146 diagnostic_file_cache_fini.
7147 (adjust_line): Take an additional input parameter for the length
7148 of the line, rather than calculating it with strlen.
7149 (diagnostic_show_locus): Adjust the use of
7150 location_get_source_line and adjust_line with respect to their new
7151 signature. While displaying a line now, do not stop at the first
7152 null byte. Rather, display the zero byte as a space and keep
7153 going until we reach the size of the line.
7154 * Makefile.in: Add vec.o to OBJS-libcommon
7156 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
7157 Ilya Tocar <ilya.tocar@intel.com>
7159 * config/i386/avx512fintrin.h (_mm512_kmov): New.
7160 * config/i386/i386.c (IX86_BUILTIN_KMOV16): Ditto.
7161 (__builtin_ia32_kmov16): Ditto.
7162 * config/i386/i386.md (UNSPEC_KMOV): New.
7165 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
7167 * config/i386/avx512fintrin.h (_mm512_loadu_si512): Rename.
7168 (_mm512_storeu_si512): Ditto.
7170 2014-01-23 Richard Sandiford <rdsandiford@googlemail.com>
7173 * rtl.h (get_referenced_operands): Declare.
7174 * recog.c (get_referenced_operands): New function.
7175 * config/mips/mips.c (mips_reorg_process_insns): Check which asm
7176 operands have been referenced when recording LO_SUM references.
7178 2014-01-22 David Holsgrove <david.holsgrove@xilinx.com>
7180 * config/microblaze/microblaze.md: Correct bswaphi2 insn.
7182 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
7184 * config/i386/x86-tune.def (X86_TUNE_ACCUMULATE_OUTGOING_ARGS):
7185 Enable for generic and recent AMD targets.
7187 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
7189 * combine-stack-adj.c (combine_stack_adjustments_for_block): Remove
7190 ARG_SIZE note when adjustment was eliminated.
7192 2014-01-22 Jeff Law <law@redhat.com>
7194 PR tree-optimization/59597
7195 * tree-ssa-threadupdate.c (dump_jump_thread_path): Move to earlier
7196 in file. Accept new argument REGISTERING and use it to modify
7197 dump output appropriately.
7198 (register_jump_thread): Corresponding changes.
7199 (mark_threaded_blocks): Reinstate code to cancel unprofitable
7200 thread paths involving joiner blocks. Add code to dump cancelled
7201 jump threading paths.
7203 2014-01-22 Vladimir Makarov <vmakarov@redhat.com>
7205 PR rtl-optimization/59477
7206 * lra-constraints.c (inherit_in_ebb): Process call for living hard
7207 regs. Update reloads_num and potential_reload_hard_regs for all insns.
7209 2014-01-22 Tom Tromey <tromey@redhat.com>
7211 * config/i386/i386-interix.h (i386_pe_unique_section): Don't use
7213 * config/cr16/cr16-protos.h (notice_update_cc): Don't use PARAMS.
7215 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
7217 PR rtl-optimization/59896
7218 * lra-constraints.c (process_alt_operands): Check unused note for
7219 matched operands of insn with no output reloads.
7221 2014-01-21 Richard Sandiford <rdsandiford@googlemail.com>
7223 * config/mips/mips.c (mips_move_to_gpr_cost): Add M16_REGS case.
7224 (mips_move_from_gpr_cost): Likewise.
7226 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
7228 PR rtl-optimization/59858
7229 * lra-constraints.c (SMALL_REGISTER_CLASS_P): Use
7230 ira_class_hard_regs_num.
7231 (process_alt_operands): Increase reject for dying matched operand.
7233 2014-01-21 Jakub Jelinek <jakub@redhat.com>
7236 * config/i386/i386.c (expand_small_movmem_or_setmem): If mode is
7237 smaller than size, perform several stores or loads and stores
7238 at dst + count - size to store or copy all of size bytes, rather
7239 than just last modesize bytes.
7241 2014-01-20 DJ Delorie <dj@redhat.com>
7243 * config/rl78/rl78.c (rl78_propogate_register_origins): Verify
7244 that CLOBBERs are REGs before propogating their values.
7246 2014-01-20 H.J. Lu <hongjiu.lu@intel.com>
7249 * cgraph.c (cgraph_inline_failed_string): Add type to DEFCIFCODE.
7250 (cgraph_inline_failed_type): New function.
7251 * cgraph.h (DEFCIFCODE): Add type.
7252 (cgraph_inline_failed_type_t): New enum.
7253 (cgraph_inline_failed_type): New prototype.
7254 * cif-code.def: Add CIF_FINAL_NORMAL to OK, FUNCTION_NOT_CONSIDERED,
7255 FUNCTION_NOT_OPTIMIZED, REDEFINED_EXTERN_INLINE,
7256 FUNCTION_NOT_INLINE_CANDIDATE, LARGE_FUNCTION_GROWTH_LIMIT,
7257 LARGE_STACK_FRAME_GROWTH_LIMIT, MAX_INLINE_INSNS_SINGLE_LIMIT,
7258 MAX_INLINE_INSNS_AUTO_LIMIT, INLINE_UNIT_GROWTH_LIMIT,
7259 RECURSIVE_INLINING, UNLIKELY_CALL, NOT_DECLARED_INLINED,
7260 OPTIMIZING_FOR_SIZE, ORIGINALLY_INDIRECT_CALL,
7261 INDIRECT_UNKNOWN_CALL, USES_COMDAT_LOCAL.
7262 Add CIF_FINAL_ERROR to UNSPECIFIED, BODY_NOT_AVAILABLE,
7263 FUNCTION_NOT_INLINABLE, OVERWRITABLE, MISMATCHED_ARGUMENTS,
7264 EH_PERSONALITY, NON_CALL_EXCEPTIONS, TARGET_OPTION_MISMATCH,
7265 OPTIMIZATION_MISMATCH.
7266 * tree-inline.c (expand_call_inline): Emit errors during
7267 early_inlining if cgraph_inline_failed_type returns CIF_FINAL_ERROR.
7269 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
7272 * config/i386/sse.md (*andnot<mode>3<mask_name>): Handle MODE_V16SF
7273 mode attribute in insn output.
7275 2014-01-20 Eric Botcazou <ebotcazou@adacore.com>
7277 * output.h (output_constant): Delete.
7278 * varasm.c (output_constant): Make private.
7280 2014-01-20 Alex Velenko <Alex.Velenko@arm.com>
7282 * config/aarch64/aarch64-simd.md (vec_perm<mode>): Add BE check.
7284 2014-01-20 Jakub Jelinek <jakub@redhat.com>
7287 * tree.h (fold_builtin_strcat): New prototype.
7288 * builtins.c (fold_builtin_strcat): No longer static. Add len
7289 argument, if non-NULL, don't call c_strlen. Optimize
7290 directly into __builtin_memcpy instead of __builtin_strcpy.
7291 (fold_builtin_2): Adjust fold_builtin_strcat caller.
7292 * gimple-fold.c (gimple_fold_builtin): Handle BUILT_IN_STRCAT.
7294 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
7296 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
7297 for SImode_address_operand operands, having only a REG argument.
7299 2014-01-20 Marcus Shawcroft <marcus.shawcroft@arm.com>
7301 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER): Expand
7302 loader name using mbig-endian.
7303 (LINUX_TARGET_LINK_SPEC): Pass linker -m flag.
7305 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
7307 * doc/invoke.texi (-march): Clarify documentation for AArch64.
7311 2014-01-20 Tejas Belagod <tejas.belagod@arm.com>
7313 * config/aarch64/aarch64-protos.h
7314 (aarch64_cannot_change_mode_class_ptr): Declare.
7315 * config/aarch64/aarch64.c (aarch64_cannot_change_mode_class,
7316 aarch64_cannot_change_mode_class_ptr): New.
7317 * config/aarch64/aarch64.h (CANNOT_CHANGE_MODE_CLASS): Change to call
7318 backend hook aarch64_cannot_change_mode_class.
7320 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
7322 * common/config/aarch64/aarch64-common.c
7323 (aarch64_handle_option): Don't handle any option order logic here.
7324 * config/aarch64/aarch64.c (aarch64_parse_arch): Do not override
7325 selected_cpu, warn on architecture version mismatch.
7326 (aarch64_override_options): Fix parsing order for option strings.
7328 2014-01-20 Jan-Benedict Glaw <jbglaw@lug-owl.de>
7329 Iain Sandoe <iain@codesourcery.com>
7332 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Fix unused variable
7333 warning. Amend comment to reflect current functionality.
7335 2014-01-20 Richard Biener <rguenther@suse.de>
7338 * builtins.c (fold_builtin_strcat): Remove case better handled
7339 by tree-ssa-strlen.c.
7341 2014-01-20 Alan Lawrence <alan.lawrence@arm.com>
7343 * config/aarch64/aarch64.opt
7344 (mcpu, march, mtune): Make case-insensitive.
7346 2014-01-20 Jakub Jelinek <jakub@redhat.com>
7349 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
7350 if operands[1] is a REG or ZERO_EXTEND of a REG.
7352 2014-01-19 Jan Hubicka <hubicka@ucw.cz>
7354 * varasm.c (compute_reloc_for_constant): Use targetm.binds_local_p.
7356 2014-01-19 John David Anglin <danglin@gcc.gnu.org>
7358 * config/pa/pa.c (pa_attr_length_millicode_call): Correct length of
7359 long non-pic millicode calls.
7361 2014-01-19 Jan-Benedict Glaw <jbglaw@lug-owl.de>
7363 * config/vax/vax.h (FUNCTION_ARG_REGNO_P): Fix unused variable warning.
7365 2014-01-19 Kito Cheng <kito@0xlab.org>
7367 * builtins.c (expand_movstr): Check movstr expand done or fail.
7369 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
7370 H.J. Lu <hongjiu.lu@intel.com>
7373 * config/i386/i386.md (*lea<mode>): Zero-extend return register
7374 to DImode for zero-extended addresses.
7376 2014-01-19 Jakub Jelinek <jakub@redhat.com>
7378 PR rtl-optimization/57763
7379 * bb-reorder.c (fix_crossing_unconditional_branches): Set JUMP_LABEL
7380 on the new indirect jump_insn and increment LABEL_NUSES (label).
7382 2014-01-18 H.J. Lu <hongjiu.lu@intel.com>
7386 * config.gcc (x86_archs): New variable.
7387 (x86_64_archs): Likewise.
7388 (x86_cpus): Likewise.
7389 Use $x86_archs, $x86_64_archs and $x86_cpus to check valid
7390 --with-arch/--with-cpu= options.
7391 Support --with-arch=/--with-cpu={nehalem,westmere,
7392 sandybridge,ivybridge,haswell,broadwell,bonnell,silvermont}.
7394 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
7396 * config/i386/i386.c (ix86_adjust_cost): Reorder PROCESSOR_K8
7397 and PROCESSOR_ATHLON to simplify code. Move "memory" calculation.
7399 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
7401 * config/i386/i386.md (*swap<mode>): Rename from swap<mode>.
7403 2014-01-18 Jakub Jelinek <jakub@redhat.com>
7406 * config/i386/i386-c.c (ix86_pragma_target_parse): Temporarily
7407 clear cpp_get_options (parse_in)->warn_unused_macros for
7408 ix86_target_macros_internal with cpp_define.
7410 2014-01-18 Richard Sandiford <rdsandiford@googlemail.com>
7412 * jump.c (delete_related_insns): Keep (use (insn))s.
7413 * reorg.c (redundant_insn): Check for barriers too.
7415 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
7417 * config/i386/i386.c (ix86_split_lea_for_addr): Fix a comment typo.
7419 2014-01-17 John David Anglin <danglin@gcc.gnu.org>
7421 * config/pa/pa.c (pa_attr_length_indirect_call): Don't output a short
7422 call to $$dyncall when TARGET_LONG_CALLS is true.
7424 2014-01-17 Jeff Law <law@redhat.com>
7426 * ree.c (combine_set_extension): Temporarily disable test for
7427 changing number of hard registers.
7429 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
7432 * ipa-inline-analysis.c (inline_free_summary):
7433 Do not free summary of aliases.
7435 2014-01-17 Jakub Jelinek <jakub@redhat.com>
7438 * gimplify.c (gimplify_expr): Use create_tmp_var
7439 instead of create_tmp_var_raw. If cond doesn't have
7440 integral type, don't add the IFN_ANNOTATE builtin at all.
7442 2014-01-17 Martin Jambor <mjambor@suse.cz>
7445 * ipa-cp.c (prev_edge_clone): New variable.
7446 (grow_next_edge_clone_vector): Renamed to grow_edge_clone_vectors.
7447 Also resize prev_edge_clone vector.
7448 (ipcp_edge_duplication_hook): Also update prev_edge_clone.
7449 (ipcp_edge_removal_hook): New function.
7450 (ipcp_driver): Register ipcp_edge_removal_hook.
7452 2014-01-17 Andrew Pinski <apinski@cavium.com>
7453 Steve Ellcey <sellcey@mips.com>
7456 * config/mips/mips.c (mips_print_operand): Check operand mode instead
7459 2014-01-17 Jeff Law <law@redhat.com>
7462 * passes.def: Reorder pass_copy_prop, pass_unrolli, pass_ccp sequence
7463 so that pass_ccp runs first.
7465 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
7467 * config/i386/i386.c (ix86_lea_outperforms): Use TARGET_XXX.
7468 (ix86_adjust_cost): Use !TARGET_XXX.
7469 (do_reorder_for_imul): Likewise.
7470 (swap_top_of_ready_list): Likewise.
7471 (ix86_sched_reorder): Likewise.
7473 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
7475 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
7476 PROCESSOR_INTEL. Treat like PROCESSOR_GENERIC.
7477 * config/i386/i386.c (intel_memcpy): New. Duplicate slm_memcpy.
7478 (intel_memset): New. Duplicate slm_memset.
7479 (intel_cost): New. Duplicate slm_cost.
7480 (m_INTEL): New macro.
7481 (processor_target_table): Add "intel".
7482 (ix86_option_override_internal): Replace PROCESSOR_SILVERMONT
7483 with PROCESSOR_INTEL for "intel".
7484 (ix86_lea_outperforms): Support PROCESSOR_INTEL. Duplicate
7485 PROCESSOR_SILVERMONT.
7486 (ix86_issue_rate): Likewise.
7487 (ix86_adjust_cost): Likewise.
7488 (ia32_multipass_dfa_lookahead): Likewise.
7489 (swap_top_of_ready_list): Likewise.
7490 (ix86_sched_reorder): Likewise.
7491 (ix86_avoid_lea_for_addr): Check TARGET_AVOID_LEA_FOR_ADDR
7492 instead of TARGET_OPT_AGU.
7493 * config/i386/i386.h (TARGET_INTEL): New.
7494 (TARGET_AVOID_LEA_FOR_ADDR): Likewise.
7495 (processor_type): Add PROCESSOR_INTEL.
7496 * config/i386/x86-tune.def: Support m_INTEL. Duplicate m_SILVERMONT.
7497 Add X86_TUNE_AVOID_LEA_FOR_ADDR.
7499 2014-01-17 Marek Polacek <polacek@redhat.com>
7502 * gimple-fold.c (fold_array_ctor_reference): Don't fold if element
7505 2014-01-17 Richard Biener <rguenther@suse.de>
7507 PR tree-optimization/46590
7508 * opts.c (default_options_table): Add entries for
7509 OPT_fbranch_count_reg, OPT_fmove_loop_invariants and OPT_ftree_pta,
7510 all enabled at -O1 but not for -Og.
7511 * common.opt (fbranch-count-reg): Remove Init(1).
7512 (fmove-loop-invariants): Likewise.
7513 (ftree-pta): Likewise.
7515 2014-01-17 Jakub Jelinek <jakub@redhat.com>
7517 * config/i386/i386.c (ix86_data_alignment): For compatibility with
7518 (incorrect) GCC 4.8 and earlier alignment assumptions ensure we align
7519 decls to at least the GCC 4.8 used alignments.
7522 * tree-nested.c (convert_nonlocal_reference_stmt,
7523 convert_local_reference_stmt): For NAMELIST_DECLs in gimple_bind_vars
7524 of GIMPLE_BIND stmts, adjust associated decls.
7526 2014-01-17 Richard Biener <rguenther@suse.de>
7528 PR tree-optimization/46590
7529 * vec.h (vec<>::bseach): New member function implementing
7530 binary search according to C89 bsearch.
7531 (vec<>::qsort): Avoid calling ::qsort for vectors with sizes 0 or 1.
7532 * tree-ssa-loop-im.c (struct mem_ref): Make stored member a
7533 bitmap pointer again. Make accesses_in_loop a flat array.
7534 (mem_ref_obstack): New global.
7535 (outermost_indep_loop): Adjust for mem_ref->stored changes.
7536 (mark_ref_stored): Likewise.
7537 (ref_indep_loop_p_2): Likewise.
7538 (set_ref_stored_in_loop): New helper function.
7539 (mem_ref_alloc): Allocate mem_refs on the mem_ref_obstack obstack.
7540 (memref_free): Adjust.
7541 (record_mem_ref_loc): Simplify.
7542 (gather_mem_refs_stmt): Adjust.
7543 (sort_locs_in_loop_postorder_cmp): New function.
7544 (analyze_memory_references): Sort accesses_in_loop after
7545 loop postorder number.
7546 (find_ref_loc_in_loop_cmp): New function.
7547 (for_all_locs_in_loop): Find relevant cluster of locs in
7548 accesses_in_loop and iterate without recursion.
7549 (execute_sm): Avoid uninit warning.
7550 (struct ref_always_accessed): Simplify.
7551 (ref_always_accessed::operator ()): Likewise.
7552 (ref_always_accessed_p): Likewise.
7553 (tree_ssa_lim_initialize): Initialize mem_ref_obstack, compute
7554 loop postorder numbers here.
7555 (tree_ssa_lim_finalize): Free mem_ref_obstack and loop postorder
7558 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
7561 * passes.c (rest_of_decl_compilation): Don't call varpool_finalize_decl
7562 on decls for which assemble_alias has been called.
7564 2014-01-17 Nick Clifton <nickc@redhat.com>
7566 * config/msp430/msp430.opt: (mcpu): New option.
7567 * config/msp430/msp430.c (msp430_mcu_name): Use target_mcu.
7568 (msp430_option_override): Parse target_cpu. If the MCU name
7569 matches a generic string, clear target_mcu.
7570 (msp430_attr): Allow numeric interrupt values up to 63.
7571 (msp430_expand_epilogue): No longer invert operand 1 of gen_popm.
7572 * config/msp430/msp430.h (ASM_SPEC): Convert -mcpu into a -mmcu
7574 * config/msp430/t-msp430: (MULTILIB_MATCHES): Remove mcu matches.
7576 * config/msp430/msp430.md (popm): Use %J rather than %I.
7577 (addsi3): Use msp430_nonimmediate_operand for operand 2.
7578 (addhi_cy_i): Use immediate_operand for operand 2.
7579 * doc/invoke.texi: Document -mcpu option.
7581 2014-01-17 Richard Biener <rguenther@suse.de>
7583 PR rtl-optimization/38518
7584 * df.h (df_analyze_loop): Declare.
7585 * df-core.c: Include cfgloop.h.
7586 (df_analyze_1): Split out main part of df_analyze.
7587 (df_analyze): Adjust.
7588 (loop_inverted_post_order_compute): New function.
7589 (loop_post_order_compute): Likewise.
7590 (df_analyze_loop): New function avoiding whole-function
7592 * loop-invariant.c (find_defs): Use df_analyze_loop.
7593 (find_invariants): Adjust.
7594 * loop-iv.c (iv_analysis_loop_init): Use df_analyze_loop.
7596 2014-01-17 Zhenqiang Chen <zhenqiang.chen@arm.com>
7598 * config/arm/arm.c (arm_v7m_tune): Set max_insns_skipped to 2.
7599 (thumb2_final_prescan_insn): Set max to MAX_INSN_PER_IT_BLOCK.
7601 2014-01-16 Ilya Enkovich <ilya.enkovich@intel.com>
7603 * ipa-ref.c (ipa_remove_stmt_references): Fix references
7604 traversal when removing references.
7606 2014-01-16 Jan Hubicka <hubicka@ucw.cz>
7609 * tree.c (get_binfo_at_offset): Look harder for virtual bases.
7611 2014-01-16 Bernd Schmidt <bernds@codesourcery.com>
7614 * reload.c (find_reloads_address_1): Do not use RELOAD_OTHER when
7615 pushing a reload for an autoinc when we had previously reloaded an
7616 inner part of the address.
7618 2014-01-16 Jakub Jelinek <jakub@redhat.com>
7620 * tree-vectorizer.h (struct _loop_vec_info): Add no_data_dependencies
7622 (LOOP_VINFO_NO_DATA_DEPENDENCIES): Define.
7623 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Clear it
7624 when not giving up or versioning for alias only because of
7626 (vect_analyze_data_ref_dependences): Set to true.
7627 * tree-vect-stmts.c (hoist_defs_of_uses): Return false if def_stmt
7629 (vectorizable_load): Use LOOP_VINFO_NO_DATA_DEPENDENCIES instead of
7630 LOOP_REQUIRES_VERSIONING_FOR_ALIAS, add && !nested_in_vect_loop
7634 * expr.c (expand_expr_real_1): Handle init == NULL_TREE.
7637 * config/i386/i386.c (ix86_expand_builtin): If target doesn't satisfy
7638 operand 0 predicate for gathers, use a new pseudo as subtarget.
7640 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
7643 * lra-constraints.c (process_alt_operands): Add printing debug info.
7644 Check absence of input/output reloads for matched operands too.
7646 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
7648 PR rtl-optimization/59835
7649 * ira.c (ira_init_register_move_cost): Increase cost for
7652 2014-01-16 Alan Lawrence <alan.lawrence@arm.com>
7654 * config/arm/arm.opt (mcpu, march, mtune): Make case-insensitive.
7656 2014-01-16 Richard Earnshaw <rearnsha@arm.com>
7659 * aarch64.c (aarch64_split_128bit_move): Don't lookup REGNO on
7660 non-register objects. Use gen_(high/low)part more consistently.
7663 2014-01-16 Michael Meissner <meissner@linux.vnet.ibm.com>
7666 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Add little
7667 endian support, remove tests for WORDS_BIG_ENDIAN.
7668 (p8_mfvsrd_3_<mode>): Likewise.
7669 (reload_gpr_from_vsx<mode>): Likewise.
7670 (reload_gpr_from_vsxsf): Likewise.
7671 (p8_mfvsrd_4_disf): Likewise.
7673 2014-01-16 Richard Biener <rguenther@suse.de>
7675 PR rtl-optimization/46590
7676 * lcm.c (compute_antinout_edge): Use postorder iteration.
7677 (compute_laterin): Use inverted postorder iteration.
7679 2014-01-16 Nick Clifton <nickc@redhat.com>
7682 * varasm.c (output_constant): Return the number of bytes actually
7684 (output_constructor_array_range): Update the field size with the
7685 number of bytes emitted by output_constant.
7686 (output_constructor_regular_field): Likewise. Also do not
7687 complain if the total number of bytes emitted is now greater
7688 than the expected fieldpos.
7689 * output.h (output_constant): Update prototype and descriptive comment.
7691 2014-01-16 Marek Polacek <polacek@redhat.com>
7694 * cgraph.c (gimple_check_call_args): Don't use DECL_ARG_TYPE if
7695 it is error_mark_node.
7697 2014-01-15 Uros Bizjak <ubizjak@gmail.com>
7699 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use
7700 VALID_AVX256_REG_OR_OI_MODE.
7702 2014-01-15 Pat Haugen <pthaugen@us.ibm.com>
7704 * config/rs6000/rs6000.c (rs6000_output_function_prologue): Check if
7705 current procedure should be profiled.
7707 2014-01-15 Andrew Pinski <apinski@cavium.com>
7709 * config/aarch64/aarch64.c (aarch64_register_move_cost): Correct cost
7710 of moving from/to the STACK_REG register class.
7712 2014-01-15 Richard Henderson <rth@redhat.com>
7715 * reginfo.c (global_regs_decl): Globalize.
7716 * rtl.h (global_regs_decl): Declare.
7717 * ira.c (do_reload): Diagnose frame_pointer_needed and it
7718 reserved via global_regs.
7720 2014-01-15 Teresa Johnson <tejohnson@google.com>
7722 * tree-ssa-sccvn.c (visit_reference_op_call): Handle NULL vdef.
7724 2014-01-15 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
7726 * config/rs6000/altivec.md (mulv8hi3): Explicitly generate vmulesh
7727 and vmulosh rather than call gen_vec_widen_smult_*.
7728 (vec_widen_umult_even_v16qi): Test VECTOR_ELT_ORDER_BIG rather
7729 than BYTES_BIG_ENDIAN to determine use of even or odd instruction.
7730 (vec_widen_smult_even_v16qi): Likewise.
7731 (vec_widen_umult_even_v8hi): Likewise.
7732 (vec_widen_smult_even_v8hi): Likewise.
7733 (vec_widen_umult_odd_v16qi): Likewise.
7734 (vec_widen_smult_odd_v16qi): Likewise.
7735 (vec_widen_umult_odd_v8hi): Likewise.
7736 (vec_widen_smult_odd_v8hi): Likewise.
7737 (vec_widen_umult_hi_v16qi): Explicitly generate vmuleub and
7738 vmuloub rather than call gen_vec_widen_umult_*.
7739 (vec_widen_umult_lo_v16qi): Likewise.
7740 (vec_widen_smult_hi_v16qi): Explicitly generate vmulesb and
7741 vmulosb rather than call gen_vec_widen_smult_*.
7742 (vec_widen_smult_lo_v16qi): Likewise.
7743 (vec_widen_umult_hi_v8hi): Explicitly generate vmuleuh and vmulouh
7744 rather than call gen_vec_widen_umult_*.
7745 (vec_widen_umult_lo_v8hi): Likewise.
7746 (vec_widen_smult_hi_v8hi): Explicitly gnerate vmulesh and vmulosh
7747 rather than call gen_vec_widen_smult_*.
7748 (vec_widen_smult_lo_v8hi): Likewise.
7750 2014-01-15 Jeff Law <law@redhat.com>
7752 PR tree-optimization/59747
7753 * ree.c (find_and_remove_re): Properly handle case where a second
7754 eliminated extension requires widening a copy created for elimination
7755 of a prior extension.
7756 (combine_set_extension): Ensure that the number of hard regs needed
7757 for a destination register does not change when we widen it.
7759 2014-01-15 Sebastian Huber <sebastian.huber@embedded-brains.de>
7761 * config.gcc (*-*-rtems*): Add t-rtems to tmake_file.
7762 (arm*-*-uclinux*eabi*): Do not override an existing tmake_file.
7763 (arm*-*-eabi* | arm*-*-symbianelf* | arm*-*-rtems*): Likwise.
7764 (arm*-*-rtems*): Use t-rtems from existing tmake_file.
7765 (avr-*-rtems*): Likewise.
7766 (bfin*-rtems*): Likewise.
7767 (moxie-*-rtems*): Likewise.
7768 (h8300-*-rtems*): Likewise.
7769 (i[34567]86-*-rtems*): Likewise.
7770 (lm32-*-rtems*): Likewise.
7771 (m32r-*-rtems*): Likewise.
7772 (m68k-*-rtems*): Likewise.
7773 (microblaze*-*-rtems*): Likewise.
7774 (mips*-*-rtems*): Likewise.
7775 (powerpc-*-rtems*): Likewise.
7776 (sh-*-rtems*): Likewise.
7777 (sparc-*-rtems*): Likewise.
7778 (sparc64-*-rtems*): Likewise.
7779 (v850-*-rtems*): Likewise.
7780 (m32c-*-rtems*): Likewise.
7782 2014-01-15 Vladimir Makarov <vmakarov@redhat.com>
7784 PR rtl-optimization/59511
7785 * ira.c (ira_init_register_move_cost): Use memory costs for some
7786 cases of register move cost calculations.
7787 * lra-constraints.c (lra_constraints): Use REG_FREQ_FROM_BB
7788 instead of BB frequency.
7789 * lra-coalesce.c (move_freq_compare_func, lra_coalesce): Ditto.
7790 * lra-assigns.c (find_hard_regno_for): Ditto.
7792 2014-01-15 Richard Biener <rguenther@suse.de>
7794 PR tree-optimization/59822
7795 * tree-vect-stmts.c (hoist_defs_of_uses): New function.
7796 (vectorizable_load): Use it to hoist defs of uses of invariant
7797 loads out of the loop.
7799 2014-01-15 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
7800 Kugan Vivekanandarajah <kuganv@linaro.org>
7803 * config/aarch64/aarch64.c (aarch64_build_constant): Fix incorrect
7806 2014-01-15 Richard Biener <rguenther@suse.de>
7808 PR rtl-optimization/59802
7809 * lcm.c (compute_available): Use inverted postorder to seed
7810 the initial worklist.
7812 2014-01-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7815 * config/s390/s390.c (s390_preferred_reload_class): Don't return
7816 ADDR_REGS for invalid symrefs in non-PIC code.
7818 2014-01-15 Jakub Jelinek <jakub@redhat.com>
7821 * builtins.c (determine_block_size): Initialize *probable_max_size
7822 even if len_rtx is CONST_INT.
7824 2014-01-14 Andrew Pinski <apinski@cavium.com>
7826 * config/aarch64/aarch64-protos.h (tune_params): Add issue_rate.
7827 * config/aarch64/aarch64.c (generic_tunings): Add issue rate of 2.
7828 (cortexa53_tunings): Likewise.
7829 (aarch64_sched_issue_rate): New function.
7830 (TARGET_SCHED_ISSUE_RATE): Define.
7832 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
7834 * ira-costs.c (find_costs_and_classes): Add missed
7835 ira_init_register_move_cost_if_necessary.
7837 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
7840 * config/arm/arm.c (arm_coproc_mem_operand): Add lra_in_progress.
7842 2014-01-14 H.J. Lu <hongjiu.lu@intel.com>
7845 * config/i386/i386.c (type_natural_mode): Add a bool parameter
7846 to indicate if type is used for function return value. Warn ABI
7847 change if the vector mode isn't available for function return value.
7848 (ix86_function_arg_advance): Pass false to type_natural_mode.
7849 (ix86_function_arg): Likewise.
7850 (ix86_gimplify_va_arg): Likewise.
7851 (function_arg_32): Don't warn ABI change.
7852 (ix86_function_value): Pass true to type_natural_mode.
7853 (ix86_return_in_memory): Likewise.
7854 (ix86_struct_value_rtx): Removed.
7855 (TARGET_STRUCT_VALUE_RTX): Likewise.
7857 2014-01-14 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7859 * jump.c (redirect_jump_2): Remove REG_CROSSING_JUMP notes when
7860 converting a conditional jump into a conditional return.
7862 2014-01-14 Richard Biener <rguenther@suse.de>
7864 PR tree-optimization/58921
7865 PR tree-optimization/59006
7866 * tree-vect-loop-manip.c (vect_loop_versioning): Remove code
7867 hoisting invariant stmts.
7868 * tree-vect-stmts.c (vectorizable_load): Insert the splat of
7869 invariant loads on the preheader edge if possible.
7871 2014-01-14 Joey Ye <joey.ye@arm.com>
7873 * doc/plugin.texi (Building GCC plugins): Update to C++.
7875 2014-01-14 Kirill Yukhin <kirill.yukhin@intel.com>
7877 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): New.
7878 (_mm_rcp28_round_ss): Ditto.
7879 (_mm_rsqrt28_round_sd): Ditto.
7880 (_mm_rsqrt28_round_ss): Ditto.
7881 (_mm_rcp28_sd): Ditto.
7882 (_mm_rcp28_ss): Ditto.
7883 (_mm_rsqrt28_sd): Ditto.
7884 (_mm_rsqrt28_ss): Ditto.
7885 * config/i386/avx512fintrin.h (_mm512_stream_load_si512): Ditto.
7886 * config/i386/i386-builtin-types.def (V8DI_FTYPE_PV8DI): Ditto.
7887 * config/i386/i386.c (IX86_BUILTIN_MOVNTDQA512): Ditto.
7888 (IX86_BUILTIN_RCP28SD): Ditto.
7889 (IX86_BUILTIN_RCP28SS): Ditto.
7890 (IX86_BUILTIN_RSQRT28SD): Ditto.
7891 (IX86_BUILTIN_RSQRT28SS): Ditto.
7892 (bdesc_special_args): Define __builtin_ia32_movntdqa512,
7893 __builtin_ia32_rcp28sd_round, __builtin_ia32_rcp28ss_round,
7894 __builtin_ia32_rsqrt28sd_round, __builtin_ia32_rsqrt28ss_round.
7895 (ix86_expand_special_args_builtin): Expand new FTYPE.
7896 * config/i386/sse.md (define_mode_attr "sse4_1_avx2"): Expand to V8DI.
7897 (srcp14<mode>): Make insn unary.
7898 (avx512f_vmscalef<mode><round_name>): Use substed predicate.
7899 (avx512f_sgetexp<mode><round_saeonly_name>): Ditto.
7900 (avx512f_rndscale<mode><round_saeonly_name>): Ditto.
7901 (<sse4_1_avx2>_movntdqa): Extend to 512 bits.
7902 (avx512er_exp2<mode><mask_name><round_saeonly_name>):
7903 Fix rounding: make it SAE only.
7904 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
7906 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
7908 (avx512er_vmrcp28<mode><round_saeonly_name>): Ditto.
7909 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Ditto.
7910 (avx512f_getmant<mode><mask_name><round_saeonly_name>): Ditto.
7911 * config/i386/subst.md (round_saeonly_mask_scalar_operand3): Remove.
7912 (round_saeonly_mask_scalar_operand4): Ditto.
7913 (round_saeonly_mask_scalar_op3): Ditto.
7914 (round_saeonly_mask_scalar_op4): Ditto.
7916 2014-01-13 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7918 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
7919 Implement -maltivec=be for vec_insert and vec_extract.
7921 2014-01-10 DJ Delorie <dj@redhat.com>
7923 * config/msp430/msp430.md (call_internal): Don't allow memory
7924 references with SP as the base register.
7925 (call_value_internal): Likewise.
7926 * config/msp430/constraints.md (Yc): New. For memory references
7927 that don't use SP as a base register.
7929 * config/msp430/msp430.c (msp430_print_operand): Add 'J' to mean
7930 "an integer without a # prefix"
7931 * config/msp430/msp430.md (epilogue_helper): Use it.
7933 2014-01-13 Jakub Jelinek <jakub@redhat.com>
7936 * config/i386/i386.c (ix86_vectorize_builtin_gather): Uncomment
7937 AVX512F gather builtins.
7938 * tree-vect-stmts.c (vectorizable_mask_load_store): For now punt
7939 on gather decls with INTEGER_TYPE masktype.
7940 (vectorizable_load): For INTEGER_TYPE masktype, put the INTEGER_CST
7941 directly into the builtin rather than hoisting it before loop.
7943 PR tree-optimization/59387
7944 * tree-scalar-evolution.c: Include gimple-fold.h and gimplify-me.h.
7945 (scev_const_prop): If folded_casts and type has undefined overflow,
7946 use force_gimple_operand instead of force_gimple_operand_gsi and
7947 for each added stmt if it is assign with
7948 arith_code_with_undefined_signed_overflow, call
7949 rewrite_to_defined_overflow.
7950 * tree-ssa-loop-im.c: Don't include gimplify-me.h, include
7951 gimple-fold.h instead.
7952 (arith_code_with_undefined_signed_overflow,
7953 rewrite_to_defined_overflow): Moved to ...
7954 * gimple-fold.c (arith_code_with_undefined_signed_overflow,
7955 rewrite_to_defined_overflow): ... here. No longer static.
7956 Include gimplify-me.h.
7957 * gimple-fold.h (arith_code_with_undefined_signed_overflow,
7958 rewrite_to_defined_overflow): New prototypes.
7960 2014-01-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7962 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Fix typo in description.
7964 2014-01-13 Eric Botcazou <ebotcazou@adacore.com>
7966 * builtins.c (get_object_alignment_2): Minor tweak.
7967 * tree-ssa-loop-ivopts.c (may_be_unaligned_p): Rewrite.
7969 2014-01-13 Christian Bruel <christian.bruel@st.com>
7971 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Unroll small sizes and
7972 optimized non constant lengths.
7974 2014-01-13 Jakub Jelinek <jakub@redhat.com>
7977 * omp-low.c (expand_omp_atomic_pipeline): Expand the initial
7978 load as __atomic_load_N if possible.
7980 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
7982 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): Remove
7984 (rs6000_expand_builtin): Adjust call.
7986 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
7989 * config/rs6000/rs6000.h (SWITCHABLE_TARGET): Define.
7990 * config/rs6000/rs6000.c: Include target-globals.h.
7991 (rs6000_set_current_function): Instead of doing target_reinit
7992 unconditionally, use save_target_globals_default_opts and
7993 restore_target_globals.
7995 * config/rs6000/rs6000-builtin.def (mffs, mtfsf): Add builtins for
7997 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): New.
7998 (rs6000_expand_builtin): Handle mffs and mtfsf.
7999 (rs6000_init_builtins): Define mffs and mtfsf.
8000 * config/rs6000/rs6000.md (UNSPECV_MFFS, UNSPECV_MTFSF): New constants.
8001 (rs6000_mffs): New pattern.
8002 (rs6000_mtfsf): New pattern.
8004 2014-01-11 Bin Cheng <bin.cheng@arm.com>
8006 * tree-ssa-loop-ivopts.c (iv_ca_narrow): New parameter.
8007 Start narrowing with START. Apply candidate-use pair
8008 and check overall cost in narrowing.
8009 (iv_ca_prune): Pass new argument.
8011 2014-01-10 Jeff Law <law@redhat.com>
8014 * ree.c (combine_reaching_defs): Ensure the defining statement
8015 occurs before the extension when optimizing extensions with
8016 different source and destination hard registers.
8018 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
8021 * ipa-devirt.c (build_type_inheritance_graph): Also add types of
8022 vtables into the type inheritance graph.
8024 2014-01-10 Jakub Jelinek <jakub@redhat.com>
8026 PR rtl-optimization/59754
8027 * ree.c (combine_reaching_defs): Disallow !SCALAR_INT_MODE_P
8028 modes in the REGNO != REGNO case.
8030 2014-01-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8032 * config/rs6000/rs6000-builtin.def: Fix pasto for VPKSDUS.
8034 2014-01-10 Jakub Jelinek <jakub@redhat.com>
8036 PR tree-optimization/59745
8037 * tree-predcom.c (tree_predictive_commoning_loop): Call
8038 free_affine_expand_cache if giving up because components is NULL.
8040 * target-globals.c (save_target_globals): Allocate < 4KB structs using
8041 GC in payload of target_globals struct instead of allocating them on
8042 the heap and the larger structs separately using GC.
8043 * target-globals.h (struct target_globals): Make regs, hard_regs,
8044 reload, expmed, ira, ira_int and lra_fields GTY((atomic)) instead
8045 of GTY((skip)) and change type to void *.
8046 (reset_target_globals): Cast loads from those fields to corresponding
8049 2014-01-10 Steve Ellcey <sellcey@mips.com>
8052 * Makefile.in (PLUGIN_HEADERS): Add gimplify.h, gimple-iterator.h,
8053 gimple-ssa.h, fold-const.h, tree-cfg.h, tree-into-ssa.h,
8054 tree-ssanames.h, print-tree.h, varasm.h, and context.h.
8056 2014-01-10 Richard Earnshaw <rearnsha@arm.com>
8059 * aarch64-modes.def (CC_Zmode): New flags mode.
8060 * aarch64.c (aarch64_select_cc_mode): Only allow NEG when the condition
8061 represents an equality.
8062 (aarch64_get_condition_code): Handle CC_Zmode.
8063 * aarch64.md (compare_neg<mode>): Restrict to equality operations.
8065 2014-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8067 * config/s390/s390.c (s390_expand_tbegin): Remove jump over CC
8068 extraction in good case.
8070 2014-01-10 Richard Biener <rguenther@suse.de>
8072 PR tree-optimization/59374
8073 * tree-vect-slp.c (vect_slp_analyze_bb_1): Move dependence
8074 checking after SLP discovery. Mark stmts not participating
8075 in any SLP instance properly.
8077 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8079 * config/arm/arm.c (arm_new_rtx_costs): Use destination mode
8080 when handling a SET rtx.
8082 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8084 * config/arm/arm-cores.def (cortex-a53): Specify FL_CRC32.
8085 (cortex-a57): Likewise.
8086 (cortex-a57.cortex-a53): Likewise. Remove redundant flags.
8088 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8090 * config/arm/arm.c (arm_init_iwmmxt_builtins): Skip
8091 non-iwmmxt builtins.
8093 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
8097 * ipa-devirt.c record_target_from_binfo): Take as argument
8098 stack of binfos and lookup matching one for virtual inheritance.
8099 (possible_polymorphic_call_targets_1): Update.
8101 2014-01-10 Huacai Chen <chenhc@lemote.com>
8103 * config/mips/driver-native.c (host_detect_local_cpu): Handle new
8104 kernel strings for Loongson-2E/2F/3A.
8106 2014-01-10 Jakub Jelinek <jakub@redhat.com>
8109 * tree-vect-data-refs.c (vect_analyze_data_refs): Check
8110 is_gimple_call before calling gimple_call_internal_p.
8112 2014-01-09 Steve Ellcey <sellcey@mips.com>
8114 * Makefile.in (TREE_FLOW_H): Remove.
8115 (TREE_SSA_H): Add file names from tree-flow.h.
8116 * doc/tree-ssa.texi (Annotations): Remove reference to tree-flow.h
8117 * tree.h: Remove tree-flow.h reference.
8118 * hash-table.h: Remove tree-flow.h reference.
8119 * tree-ssa-loop-niter.c (dump_affine_iv): Replace tree-flow.h
8120 reference with tree-ssa-loop.h.
8122 2014-01-09 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8124 * doc/invoke.texi: Add -maltivec={be,le} options, and document
8125 default element-order behavior for -maltivec.
8126 * config/rs6000/rs6000.opt: Add -maltivec={be,le} options.
8127 * config/rs6000/rs6000.c (rs6000_option_override_internal): Ensure
8128 that -maltivec={le,be} implies -maltivec; disallow -maltivec=le
8129 when targeting big endian, at least for now.
8130 * config/rs6000/rs6000.h: Add #define of VECTOR_ELT_ORDER_BIG.
8132 2014-01-09 Jakub Jelinek <jakub@redhat.com>
8135 * cfgexpand.c (expand_one_var): For SSA_NAMEs, if the underlying
8136 var satisfies use_register_for_decl, just take into account type
8137 alignment, rather than decl alignment.
8139 PR tree-optimization/59622
8140 * gimple-fold.c (gimple_fold_call): Fix a typo in message. For
8141 __builtin_unreachable replace the OBJ_TYPE_REF call with a call to
8142 __builtin_unreachable and add if needed a setter of the lhs SSA_NAME.
8143 Don't devirtualize for inplace at all. For targets.length () == 1,
8144 if the call is noreturn and cfun isn't in SSA form yet, clear lhs.
8146 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
8148 * config/i386/i386.md (cpu): Remove the unused btver1.
8150 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
8152 * gdbasan.in: Put a breakpoint on __sanitizer::Report.
8154 2014-01-09 Jakub Jelinek <jakub@redhat.com>
8157 * tree-core.h (struct target_globals): New forward declaration.
8158 (struct tree_target_option): Add globals field.
8159 * tree.h (TREE_TARGET_GLOBALS): Define.
8160 (prepare_target_option_nodes_for_pch): New prototype.
8161 * target-globals.h (struct target_globals): Define even if
8163 * tree.c (prepare_target_option_node_for_pch,
8164 prepare_target_option_nodes_for_pch): New functions.
8165 * config/i386/i386.h (SWITCHABLE_TARGET): Define.
8166 * config/i386/i386.c: Include target-globals.h.
8167 (ix86_set_current_function): Instead of doing target_reinit
8168 unconditionally, use save_target_globals_default_opts and
8169 restore_target_globals.
8171 2014-01-09 Richard Biener <rguenther@suse.de>
8173 PR tree-optimization/59715
8174 * tree-cfg.h (split_critical_edges): Declare.
8175 * tree-cfg.c (split_critical_edges): Export.
8176 * tree-ssa-sink.c (execute_sink_code): Split critical edges.
8178 2014-01-09 Max Ostapenko <m.ostapenko@partner.samsung.com>
8180 * cfgexpand.c (expand_stack_vars): Optionally disable
8181 asan stack protection.
8182 (expand_used_vars): Likewise.
8183 (partition_stack_vars): Likewise.
8184 * asan.c (asan_emit_stack_protection): Optionally disable
8185 after return stack usage.
8186 (instrument_derefs): Optionally disable memory access instrumentation.
8187 (instrument_builtin_call): Likewise.
8188 (instrument_strlen_call): Likewise.
8189 (asan_protect_global): Optionally disable global variables protection.
8190 * doc/invoke.texi: Added doc for new options.
8191 * params.def: Added new options.
8192 * params.h: Likewise.
8194 2014-01-09 Jakub Jelinek <jakub@redhat.com>
8196 PR rtl-optimization/59724
8197 * ifcvt.c (cond_exec_process_if_block): Don't call
8198 flow_find_head_matching_sequence with 0 longest_match.
8199 * cfgcleanup.c (flow_find_head_matching_sequence): Count even
8200 non-active insns if !stop_after.
8201 (try_head_merge_bb): Revert 2014-01-07 changes.
8203 2014-01-08 Jeff Law <law@redhat.com>
8205 * ree.c (get_sub_rtx): New function, extracted from...
8206 (merge_def_and_ext): Here.
8207 (combine_reaching_defs): Use get_sub_rtx.
8209 2014-01-08 Eric Botcazou <ebotcazou@adacore.com>
8211 * cgraph.h (varpool_variable_node): Do not choke on null node.
8213 2014-01-08 Catherine Moore <clm@codesourcery.com>
8215 * config/mips/mips.md (simple_return): Attempt to use JRC
8217 * config/mips/mips.h (MIPS_CALL): Attempt to use JALS for microMIPS.
8219 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
8221 PR rtl-optimization/59137
8222 * reorg.c (steal_delay_list_from_target): Call update_block for
8224 (steal_delay_list_from_fallthrough, relax_delay_slots): Likewise.
8226 2014-01-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8228 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Remove
8229 two duplicate entries.
8231 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
8234 2012-10-07 Richard Sandiford <rdsandiford@googlemail.com>
8236 * config/mips/mips.c (mips_truncated_op_cost): New function.
8237 (mips_rtx_costs): Adjust test for BADDU.
8238 * config/mips/mips.md (*baddu_di<mode>): Push truncates to operands.
8240 2012-10-02 Richard Sandiford <rdsandiford@googlemail.com>
8242 * config/mips/mips.md (*baddu_si_eb, *baddu_si_el): Merge into...
8243 (*baddu_si): ...this new pattern.
8245 2014-01-08 Jakub Jelinek <jakub@redhat.com>
8248 * ipa-prop.c (ipa_analyze_params_uses): Ignore uses in debug stmts.
8250 2014-01-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
8253 * expr.h (expand_expr_real, expand_expr_real_1): Add new parameter
8255 (expand_expr, expand_normal): Adjust.
8256 * expr.c (expand_expr_real, expand_expr_real_1): Add new parameter
8257 inner_reference_p. Use inner_reference_p to expand inner references.
8258 (store_expr): Adjust.
8259 * cfgexpand.c (expand_call_stmt): Adjust.
8261 2014-01-08 Rong Xu <xur@google.com>
8263 * gcov-io.c (gcov_var): Move from gcov-io.h.
8264 (gcov_position): Ditto.
8265 (gcov_is_error): Ditto.
8266 (gcov_rewrite): Ditto.
8267 * gcov-io.h: Refactor. Move gcov_var to gcov-io.h, and libgcov
8268 only part to libgcc/libgcov.h.
8270 2014-01-08 Marek Polacek <polacek@redhat.com>
8273 * omp-low.c (simd_clone_adjust): Don't crash if def is NULL.
8275 2014-01-08 Marek Polacek <polacek@redhat.com>
8278 * ubsan.c (ubsan_type_descriptor): Call strip_array_types on type2.
8280 2014-01-08 Jakub Jelinek <jakub@redhat.com>
8282 PR rtl-optimization/59649
8283 * stor-layout.c (get_mode_bounds): For BImode return
8284 0 and STORE_FLAG_VALUE.
8286 2014-01-08 Richard Biener <rguenther@suse.de>
8289 * gimple.h (is_gimple_builtin_call): Remove.
8290 (gimple_builtin_call_types_compatible_p): New.
8291 (gimple_call_builtin_p): New overload.
8292 * gimple.c (is_gimple_builtin_call): Remove.
8293 (validate_call): Rename to ...
8294 (gimple_builtin_call_types_compatible_p): ... this and export. Also
8296 (validate_type): New static function.
8297 (gimple_call_builtin_p): New overload and adjust.
8298 * gimple-fold.c (gimple_fold_builtin): Fold the return value.
8299 (gimple_fold_call): Likewise. Use gimple_call_builtin_p.
8300 (gimple_fold_stmt_to_constant_1): Likewise.
8301 * tsan.c (instrument_gimple): Use gimple_call_builtin_p.
8303 2014-01-08 Richard Biener <rguenther@suse.de>
8306 * gimplify.c (gimplify_expr): Gimplify register-register type
8307 VIEW_CONVERT_EXPRs to separate stmts.
8309 2014-01-07 Jeff Law <law@redhat.com>
8312 * ree.c (combine_set_extension): Handle case where source
8313 and destination registers in an extension insn are different.
8314 (combine_reaching_defs): Allow source and destination registers
8315 in extension to be different under limited circumstances.
8316 (add_removable_extension): Remove restriction that the
8317 source and destination registers in the extension are the same.
8318 (find_and_remove_re): Emit a copy from the extension's
8319 destination to its source after the defining insn if
8320 the source and destination registers are different.
8323 * ifcvt.c (merge_if_block): If we are merging a block with more than
8324 one successor with a block with no successors, remove any BARRIER
8325 after the second block.
8327 2014-01-07 Dan Xio Qiang <ziyan01@163.com>
8329 * hw-doloop.c (reorg_loops): Release the bitmap obstack.
8331 2014-01-07 John David Anglin <danglin@gcc.gnu.org>
8334 * config/pa/pa.c (pa_legitimate_address_p): Return false before reload
8335 for 14-bit register offsets when INT14_OK_STRICT is false.
8337 2014-01-07 Roland Stigge <stigge@antcom.de>
8338 Michael Meissner <meissner@linux.vnet.ibm.com>
8341 * config/rs6000/rs6000.c (rs6000_legitimate_offset_address_p):
8342 Only check TFmode for SPE constants. Don't check TImode or TDmode.
8344 2014-01-07 James Greenhalgh <james.greenhalgh@arm.com>
8346 * config/aarch64/aarch64-elf.h (ASM_SPEC): Remove identity spec for
8349 2014-01-07 Yufeng Zhang <yufeng.zhang@arm.com>
8351 * config/arm/arm.c (arm_expand_neon_args): Call expand_expr
8352 with EXPAND_MEMORY for NEON_ARG_MEMORY; check if the returned
8353 rtx is const0_rtx or not.
8355 2014-01-07 Richard Sandiford <rdsandiford@googlemail.com>
8358 * target-globals.c (save_target_globals): Remove this_fn_optab
8360 * toplev.c: Include optabs.h.
8361 (target_reinit): Temporarily restore the global options if another
8362 set of options are in force.
8364 2014-01-07 Jakub Jelinek <jakub@redhat.com>
8366 PR rtl-optimization/58668
8367 * cfgcleanup.c (flow_find_cross_jump): Don't count
8368 any jumps if dir_p is NULL. Remove p1 variable, use active_insn_p
8369 to determine what is counted.
8370 (flow_find_head_matching_sequence): Use active_insn_p to determine
8372 (try_head_merge_bb): Adjust for the flow_find_head_matching_sequence
8374 * ifcvt.c (count_bb_insns): Use active_insn_p && !JUMP_P to
8375 determine what is counted.
8377 PR tree-optimization/59643
8378 * tree-predcom.c (split_data_refs_to_components): If one dr is
8379 read and one write, determine_offset fails and the write isn't
8380 in the bad component, just put the read into the bad component.
8382 2014-01-07 Mike Stump <mikestump@comcast.net>
8383 Jakub Jelinek <jakub@redhat.com>
8386 * tree-core.h (struct tree_optimization_option): Change optabs
8387 type from unsigned char * to void *.
8388 * optabs.c (init_tree_optimization_optabs): Adjust
8389 TREE_OPTIMIZATION_OPTABS initialization.
8391 2014-01-06 Jakub Jelinek <jakub@redhat.com>
8394 * config/i386/i386.h (struct machine_function): Add
8395 no_drap_save_restore field.
8396 * config/i386/i386.c (ix86_save_reg): Use
8397 !cfun->machine->no_drap_save_restore instead of
8398 crtl->stack_realign_needed.
8399 (ix86_finalize_stack_realign_flags): Don't clear drap_reg unless
8400 this function clears frame_pointer_needed. Set
8401 cfun->machine->no_drap_save_restore if clearing frame_pointer_needed
8402 and DRAP reg is needed.
8404 2014-01-06 Marek Polacek <polacek@redhat.com>
8407 * doc/implement-c.texi: Mention that other integer types are
8408 permitted as bit-field types in strictly conforming mode.
8410 2014-01-06 Felix Yang <fei.yang0953@gmail.com>
8412 * modulo-sched.c (schedule_reg_moves): Clear distance1_uses if it
8415 2014-01-06 Richard Earnshaw <rearnsha@arm.com>
8417 * aarch64.c (aarch64_rtx_costs): Fix cost calculation for MADD.
8419 2014-01-06 Martin Jambor <mjambor@suse.cz>
8422 * ipa-cp.c (ipcp_discover_new_direct_edges): Changed param_index type
8424 * ipa-prop.c (ipa_print_node_params): Fix indentation.
8426 2014-01-06 Eric Botcazou <ebotcazou@adacore.com>
8430 * var-tracking.c (add_stores): Preserve the value of the source even if
8431 we don't record the store.
8433 2014-01-06 Terry Guo <terry.guo@arm.com>
8435 * config.gcc (arm*-*-*): Check --with-arch against arm-arches.def.
8437 2014-01-05 Iain Sandoe <iain@codesourcery.com>
8440 * config/darwin.c (darwin_function_section): Adjust return values to
8441 correspond to optimisation changes made in r206070.
8443 2014-01-05 Uros Bizjak <ubizjak@gmail.com>
8445 * config/i386/i386.c (ix86_data_alignment): Calculate max_align
8446 from prefetch_block tune setting.
8447 (nocona_cost): Correct size of prefetch block to 64.
8449 2014-01-04 Eric Botcazou <ebotcazou@adacore.com>
8451 * config/arm/arm.c (arm_get_frame_offsets): Revamp long lines.
8452 (arm_expand_epilogue_apcs_frame): Take into account the number of bytes
8453 used to save the static chain register in the computation of the offset
8454 from which the FP registers need to be restored.
8456 2014-01-04 Jakub Jelinek <jakub@redhat.com>
8458 PR tree-optimization/59519
8459 * tree-vect-loop-manip.c (slpeel_update_phi_nodes_for_guard1): Don't
8460 ICE if get_current_def (current_new_name) is already non-NULL, as long
8461 as it is a phi result of some other phi in *new_exit_bb that has
8464 * config/i386/sse.md (avx512f_load<mode>_mask): Emit vmovup{s,d}
8465 or vmovdqu* for misaligned_operand.
8466 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>,
8467 <sse2_avx_avx512f>_loaddqu<mode><mask_name>): Handle <mask_applied>.
8468 * config/i386/i386.c (ix86_expand_special_args_builtin): Set
8469 aligned_mem for AVX512F masked aligned load and store builtins and for
8472 2014-01-03 Bingfeng Mei <bmei@broadcom.com>
8474 PR tree-optimization/59651
8475 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
8476 Address range for negative step should be added by TYPE_SIZE_UNIT.
8478 2014-01-03 Andreas Schwab <schwab@linux-m68k.org>
8480 * config/m68k/m68k.c (handle_move_double): Handle pushes with
8481 overlapping registers also for registers other than the stack pointer.
8483 2014-01-03 Marek Polacek <polacek@redhat.com>
8486 * doc/extend.texi: Fix the return value of __builtin_FUNCTION and
8489 2014-01-03 Jakub Jelinek <jakub@redhat.com>
8492 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Don't consider
8495 * config/i386/i386.md (MODE_SIZE): New mode attribute.
8496 (push splitter): Use <P:MODE_SIZE> instead of
8497 GET_MODE_SIZE (<P:MODE>mode).
8498 (lea splitter): Use <MODE_SIZE> instead of GET_MODE_SIZE (<MODE>mode).
8499 (mov -1, reg peephole2): Likewise.
8500 * config/i386/sse.md (*mov<mode>_internal,
8501 <sse>_storeu<ssemodesuffix><avxsizesuffix>,
8502 <sse2_avx_avx512f>_storedqu<mode>, <sse>_andnot<mode>3,
8503 *<code><mode>3, *andnot<mode>3<mask_name>,
8504 <mask_codefor><code><mode>3<mask_name>): Likewise.
8505 * config/i386/subst.md (mask_mode512bit_condition,
8506 sd_mask_mode512bit_condition): Likewise.
8508 2014-01-02 Xinliang David Li <davidxl@google.com>
8510 PR tree-optimization/59303
8511 * tree-ssa-uninit.c (is_use_properly_guarded): Main cleanup.
8512 (dump_predicates): Better output format.
8513 (pred_equal_p): New function.
8514 (is_neq_relop_p): Ditto.
8515 (is_neq_zero_form_p): Ditto.
8516 (pred_expr_equal_p): Ditto.
8517 (pred_neg_p): Ditto.
8518 (simplify_pred): Ditto.
8519 (simplify_preds_2): Ditto.
8520 (simplify_preds_3): Ditto.
8521 (simplify_preds_4): Ditto.
8522 (simplify_preds): Ditto.
8524 (push_to_worklist): Ditto.
8525 (get_pred_info_from_cmp): Ditto.
8526 (is_degenerated_phi): Ditto.
8527 (normalize_one_pred_1): Ditto.
8528 (normalize_one_pred): Ditto.
8529 (normalize_one_pred_chain): Ditto.
8530 (normalize_preds): Ditto.
8531 (normalize_cond_1): Remove function.
8532 (normalize_cond): Ditto.
8533 (is_gcond_subset_of): Ditto.
8534 (is_subset_of_any): Ditto.
8535 (is_or_set_subset_of): Ditto.
8536 (is_and_set_subset_of): Ditto.
8537 (is_norm_cond_subset_of): Ditto.
8538 (pred_chain_length_cmp): Ditto.
8539 (convert_control_dep_chain_into_preds): Type change.
8540 (find_predicates): Ditto.
8541 (find_def_preds): Ditto.
8542 (destroy_predicates_vecs): Ditto.
8543 (find_matching_predicates_in_rest_chains): Ditto.
8544 (use_pred_not_overlap_with_undef_path_pred): Ditto.
8545 (is_pred_expr_subset): Ditto.
8546 (is_pred_chain_subset_of): Ditto.
8547 (is_included_in): Ditto.
8548 (is_superset_of): Ditto.
8550 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
8552 Update copyright years.
8554 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
8556 * common/config/arc/arc-common.c, config/arc/arc-modes.def,
8557 config/arc/arc-protos.h, config/arc/arc.c, config/arc/arc.h,
8558 config/arc/arc.md, config/arc/arc.opt,
8559 config/arm/arm_neon_builtins.def, config/arm/crypto.def,
8560 config/i386/avx512cdintrin.h, config/i386/avx512erintrin.h,
8561 config/i386/avx512fintrin.h, config/i386/avx512pfintrin.h,
8562 config/i386/btver2.md, config/i386/shaintrin.h, config/i386/slm.md,
8563 config/linux-protos.h, config/linux.c, config/winnt-c.c,
8564 diagnostic-color.c, diagnostic-color.h, gimple-ssa-isolate-paths.c,
8565 vtable-verify.c, vtable-verify.h: Use the standard form for the
8568 2014-01-02 Tobias Burnus <burnus@net-b.de>
8570 * gcc.c (process_command): Update copyright notice dates.
8571 * gcov-dump.c: Ditto.
8573 * doc/cpp.texi: Bump @copying's copyright year.
8574 * doc/cppinternals.texi: Ditto.
8575 * doc/gcc.texi: Ditto.
8576 * doc/gccint.texi: Ditto.
8577 * doc/gcov.texi: Ditto.
8578 * doc/install.texi: Ditto.
8579 * doc/invoke.texi: Ditto.
8581 2014-01-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8583 * config/nios2/nios2.h (BITS_PER_UNIT): Don't define it.
8585 2014-01-01 Jakub Jelinek <jakub@redhat.com>
8587 * config/i386/sse.md (*mov<mode>_internal): Guard
8588 EXT_REX_SSE_REGNO_P (REGNO ()) uses with REG_P.
8590 PR rtl-optimization/59647
8591 * cse.c (cse_process_notes_1): Don't substitute negative VOIDmode
8592 new_rtx into UNSIGNED_FLOAT rtxes.
8594 Copyright (C) 2014 Free Software Foundation, Inc.
8596 Copying and distribution of this file, with or without modification,
8597 are permitted in any medium without royalty provided the copyright
8598 notice and this notice are preserved.