1 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
3 * config/aarch64/aarch64-protos.h (aarch64_modes_tieable_p): New.
4 * config/aarch64/aarch64.c
5 (aarch64_cannot_change_mode_class): Weaken conditions.
6 (aarch64_modes_tieable_p): New.
7 * config/aarch64/aarch64.h (MODES_TIEABLE_P): Use it.
9 2014-04-28 Pat Haugen <pthaugen@us.ibm.com>
11 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
12 (loadsync_<mode>): Change mode.
13 (load_quadpti, store_quadpti): New.
14 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
15 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
17 2014-04-28 Martin Jambor <mjambor@suse.cz>
19 * tree-sra.c (sra_modify_expr): Generate new memory accesses with
20 same alias type as the original statement.
21 (subreplacement_assignment_data): New type.
22 (handle_unscalarized_data_in_subtree): New type of parameter,
23 generate new memory accesses with same alias type as the original
25 (load_assign_lhs_subreplacements): Likewise.
26 (sra_modify_constructor_assign): Generate new memory accesses with
27 same alias type as the original statement.
29 2014-04-28 Richard Biener <rguenther@suse.de>
31 * tree-pass.h (TODO_verify_il): Define.
32 (TODO_verify_all): Complete properly.
33 * passes.c (execute_function_todo): Move existing loop-closed
34 SSA verification under TODO_verify_il.
35 (execute_one_pass): Trigger TODO_verify_il at todo-after time.
36 * graphite-sese-to-poly.c (rewrite_cross_bb_scalar_deps):
37 Fix tree sharing issue.
39 2014-04-28 Richard Biener <rguenther@suse.de>
42 * builtins.def (DEF_C11_BUILTIN): Add.
43 (BUILT_IN_ALIGNED_ALLOC): Likewise.
44 * coretypes.h (enum function_class): Add function_c11_misc.
45 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
46 BUILT_IN_ALIGNED_ALLOC like BUILT_IN_MALLOC.
47 (call_may_clobber_ref_p_1): Likewise.
48 * tree-ssa-dce.c (mark_stmt_if_obviously_necessary): Likewise.
49 (mark_all_reaching_defs_necessary_1): Likewise.
50 (propagate_necessity): Likewise.
51 (eliminate_unnecessary_stmts): Likewise.
52 * tree-ssa-ccp.c (evaluate_stmt): Handle BUILT_IN_ALIGNED_ALLOC.
54 2014-04-28 Richard Biener <rguenther@suse.de>
56 * tree-vrp.c (vrp_var_may_overflow): Remove.
57 (vrp_visit_phi_node): Rather than bumping to +-INF possibly
58 with overflow immediately bump to one before that value and
59 let iteration figure out overflow status.
61 2014-04-28 Richard Biener <rguenther@suse.de>
63 * configure.ac: Do valgrind header checks unconditionally.
64 Add --enable-valgrind-annotations.
65 * system.h: Guard valgrind header inclusion with
66 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING.
67 * alloc-pool.c (pool_alloc, pool_free): Use
68 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING
69 to guard possibly dead code.
70 * config.in: Regenerated.
71 * configure: Likewise.
73 2014-04-28 Jeff Law <law@redhat.com>
75 PR tree-optimization/60902
76 * tree-ssa-threadedge.c
77 (record_temporary_equivalences_from_stmts_at_dest): Only iterate
78 over real defs when invalidating outputs from statements that do not
79 produce useful outputs for threading.
81 2014-04-28 Richard Biener <rguenther@suse.de>
83 PR tree-optimization/60979
84 * graphite-scop-detection.c (scopdet_basic_block_info): Reject
85 SCOPs that end in a block with a successor with abnormal
88 2014-04-28 Richard Biener <rguenther@suse.de>
90 * tree-pass.h (execute_pass_list): Adjust prototype.
91 * passes.c (pass_manager::execute_early_local_passes):
93 (do_per_function): Change callback signature, push all actual
95 (do_per_function_toporder): Likewise.
96 (execute_function_dump): Adjust.
97 (execute_function_todo): Likewise.
98 (clear_last_verified): Likewise.
99 (verify_curr_properties): Likewise.
100 (update_properties_after_pass): Likewise.
101 (execute_pass_list_1): Split out from ...
102 (execute_pass_list): ... here. Adjust.
103 (execute_ipa_pass_list): Likewise.
104 * cgraphunit.c (cgraph_add_new_function): Adjust.
105 (analyze_function): Likewise.
106 (expand_function): Likewise.
107 * cgraph.c (release_function_body): Free dominance info
108 here instead of asserting it was magically freed elsewhere.
110 2014-04-28 Eric Botcazou <ebotcazou@adacore.com>
112 * configure.ac: Tweak GAS check for LEON instructions on SPARC.
113 * configure: Regenerate.
114 * config/sparc/sparc.opt (muser-mode): New option.
115 * config/sparc/sync.md (atomic_compare_and_swap<mode>_1): Do not enable
117 (atomic_compare_and_swap_leon3_1): New instruction for LEON3.
118 * doc/invoke.texi (SPARC options): Document -muser-mode.
120 2014-04-27 Richard Sandiford <rdsandiford@googlemail.com>
122 * cselib.c (find_slot_memmode): Delete.
123 (cselib_hasher): Change compare_type to a struct.
124 (cselib_hasher::equal): Update accordingly. Don't expect wrapped
126 (preserve_constants_and_equivs): Adjust for new compare_type.
127 (cselib_find_slot): Likewise. Take the mode of the rtx as argument.
128 (wrap_constant): Delete.
129 (cselib_lookup_mem, cselib_lookup_1): Update calls to cselib_find_slot.
131 2014-04-26 Markus Trippelsdorf <markus@trippelsdorf.de>
133 * doc/install.texi (Building with profile feedback): Remove
136 2014-04-26 Tom de Vries <tom@codesourcery.com>
138 * config/i386/i386.md (define_expand "ldexpxf3"): Fix out-of-bounds
141 2014-04-25 Cary Coutant <ccoutant@google.com>
144 * dwarf2out.c (should_move_die_to_comdat): A type definition
145 can contain a subprogram definition, but don't move it to a
147 (clone_as_declaration): Copy DW_AT_abstract_origin attribute.
148 (generate_skeleton_bottom_up): Remove DW_AT_object_pointer attribute
150 (clone_tree_hash): Rename to...
151 (clone_tree_partial): ...this; change callers. Copy
152 DW_TAG_subprogram DIEs as declarations.
153 (copy_decls_walk): Don't copy children of a declaration into a
156 2014-04-25 H.J. Lu <hongjiu.lu@intel.com>
159 * config/i386/i386.md (*movsf_internal): Set MODE to SI for
162 2014-04-25 Jiong Wang <jiong.wang@arm.com>
164 * config/arm/predicates.md (call_insn_operand): Add long_call check.
165 * config/arm/arm.md (sibcall, sibcall_value): Force the address to
167 * config/arm/arm.c (arm_function_ok_for_sibcall): Remove long_call
170 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
172 * config/arm/arm.c (arm_cortex_a8_tune): Initialise
175 2014-04-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
177 PR tree-optimization/60930
178 * gimple-ssa-strength-reduction.c (create_mul_imm_cand): Reject
179 creating a multiply candidate by folding two constant
180 multiplicands when the result overflows.
182 2014-04-25 Jakub Jelinek <jakub@redhat.com>
184 PR tree-optimization/60960
185 * tree-vect-generic.c (expand_vector_operation): Only call
186 expand_vector_divmod if type's mode satisfies VECTOR_MODE_P.
188 2014-04-25 Tom de Vries <tom@codesourcery.com>
190 * expr.c (clobber_reg_mode): New function.
191 * expr.h (clobber_reg): New function.
193 2014-04-25 Tom de Vries <tom@codesourcery.com>
195 * rtlanal.c (find_all_hard_reg_sets): Note INSN_CALL_FUNCTION_USAGE
198 2014-04-25 Radovan Obradovic <robradovic@mips.com>
199 Tom de Vries <tom@codesourcery.com>
201 * rtlanal.c (find_all_hard_reg_sets): Add bool implicit parameter and
203 * rtl.h (find_all_hard_reg_sets): Add bool parameter.
204 * haifa-sched.c (recompute_todo_spec, check_clobbered_conditions): Add
205 new argument to find_all_hard_reg_sets call.
207 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
209 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p):
210 Use HOST_WIDE_INT_C for mask literal.
211 (aarch_rev16_shleft_mask_imm_p): Likewise.
213 2014-04-25 Eric Botcazou <ebotcazou@adacore.com>
216 * config/sparc/sparc.md (ashlsi3_extend): Delete.
218 2014-04-25 Marc Glisse <marc.glisse@inria.fr>
220 PR preprocessor/56540
221 * config/i386/i386-c.c (ix86_target_macros): Define
222 __SIZEOF_FLOAT80__ and __SIZEOF_FLOAT128__.
224 2014-04-25 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
226 * configure.ac (tga_func): Remove.
227 (LIB_TLS_SPEC): Remove.
228 * configure: Regenerate.
229 * config.in: Regenerate.
230 * config/sol2.h (LIB_SPEC): Don't use LIB_TLS_SPEC.
232 2014-04-25 Richard Biener <rguenther@suse.de>
235 * tree-ssa-structalias.c (ipa_pta_execute): Compute direct
236 call stmt use/clobber sets during stmt walk instead of
237 walking the possibly incomplete set of caller edges.
239 2014-04-25 Richard Biener <rguenther@suse.de>
242 * passes.c (apply_ipa_transforms): Inline into only caller ...
243 (execute_one_pass): ... here. Properly bring in function
244 bodies for nodes we want to apply IPA transforms to.
246 2014-04-24 Cong Hou <congh@google.com>
248 PR tree-optimization/60896
249 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Pick up
250 all statements in PATTERN_DEF_SEQ in recognized widen-mult pattern.
251 (vect_mark_pattern_stmts): Set the def type of all statements in
252 PATTERN_DEF_SEQ as vect_internal_def.
254 2014-04-24 Michael Meissner <meissner@linux.vnet.ibm.com>
256 * doc/extend.texi (PowerPC Built-in Functions): Document new
257 powerpc extended divide, bcd, pack/unpack 128-bit, builtin
259 (PowerPC AltiVec/VSX Built-in Functions): Likewise.
261 * config/rs6000/predicates.md (const_0_to_3_operand): New
262 predicate to match 0..3 integer constants.
264 * config/rs6000/rs6000-builtin.def (BU_DFP_MISC_1): Add new macros
265 to support adding miscellaneous builtin functions.
266 (BU_DFP_MISC_2): Likewise.
267 (BU_P7_MISC_1): Likewise.
268 (BU_P7_MISC_2): Likewise.
269 (BU_P8V_MISC_3): Likewise.
270 (BU_MISC_1): Likewise.
271 (BU_MISC_2): Likewise.
272 (DIVWE): Add extended divide builtin functions.
280 (DXEX): Add decimal floating-point builtin functions.
292 (CDTBCD): Add new BCD builtin functions.
296 (BCDADD_LT): Likewise.
297 (BCDADD_EQ): Likewise.
298 (BCDADD_GT): Likewise.
299 (BCDADD_OV): Likewise.
301 (BCDSUB_LT): Likewise.
302 (BCDSUB_EQ): Likewise.
303 (BCDSUB_GT): Likewise.
304 (BCDSUB_OV): Likewise.
305 (PACK_TD): Add new pack/unpack 128-bit type builtin functions.
306 (UNPACK_TD): Likewise.
308 (UNPACK_TF): Likewise.
309 (UNPACK_TF_0): Likewise.
310 (UNPACK_TF_1): Likewise.
311 (PACK_V1TI): Likewise.
312 (UNPACK_V1TI): Likewise.
314 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Add
315 support for decimal floating point builtin functions.
316 (rs6000_expand_ternop_builtin): Add checks for the new builtin
317 functions that take constant arguments.
318 (rs6000_invalid_builtin): Add decimal floating point builtin
320 (rs6000_init_builtins): Setup long double, _Decimal64, and
321 _Decimal128 types for new builtin functions.
322 (builtin_function_type): Set the unsigned flags appropriately for
323 the new builtin functions.
324 (rs6000_opt_masks): Add support for decimal floating point builtin
327 * config/rs6000/rs6000.h (RS6000_BTM_DFP): Add support for decimal
328 floating point builtin functions.
329 (RS6000_BTM_COMMON): Likewise.
330 (RS6000_BTI_long_double): Likewise.
331 (RS6000_BTI_dfloat64): Likewise.
332 (RS6000_BTI_dfloat128): Likewise.
333 (long_double_type_internal_node): Likewise.
334 (dfloat64_type_internal_node): Likewise.
335 (dfloat128_type_internal_node): Likewise.
337 * config/rs6000/altivec.h (UNSPEC_BCDADD): Add support for ISA
338 2.07 bcd arithmetic instructions.
339 (UNSPEC_BCDSUB): Likewise.
340 (UNSPEC_BCD_OVERFLOW): Likewise.
341 (UNSPEC_BCD_ADD_SUB): Likewise.
342 (bcd_add_sub): Likewise.
343 (BCD_TEST): Likewise.
344 (bcd<bcd_add_sub>): Likewise.
345 (bcd<bcd_add_sub>_test): Likewise.
346 (bcd<bcd_add_sub>_test2): Likewise.
347 (bcd<bcd_add_sub>_<code>): Likewise.
348 (peephole2 for combined bcd ops): Likewise.
350 * config/rs6000/dfp.md (UNSPEC_DDEDPD): Add support for new
351 decimal floating point builtin functions.
352 (UNSPEC_DENBCD): Likewise.
353 (UNSPEC_DXEX): Likewise.
354 (UNSPEC_DIEX): Likewise.
355 (UNSPEC_DSCLI): Likewise.
356 (UNSPEC_DSCRI): Likewise.
357 (D64_D128): Likewise.
358 (dfp_suffix): Likewise.
359 (dfp_ddedpd_<mode>): Likewise.
360 (dfp_denbcd_<mode>): Likewise.
361 (dfp_dxex_<mode>): Likewise.
362 (dfp_diex_<mode>): Likewise.
363 (dfp_dscli_<mode>): Likewise.
364 (dfp_dscri_<mode>): Likewise.
366 * config/rs6000/rs6000.md (UNSPEC_ADDG6S): Add support for new BCD
368 (UNSPEC_CDTBCD): Likewise.
369 (UNSPEC_CBCDTD): Likewise.
370 (UNSPEC_DIVE): Add support for new extended divide builtin
372 (UNSPEC_DIVEO): Likewise.
373 (UNSPEC_DIVEU): Likewise.
374 (UNSPEC_DIVEUO): Likewise.
375 (UNSPEC_UNPACK_128BIT): Add support for new builtin functions to
376 pack/unpack 128-bit types.
377 (UNSPEC_PACK_128BIT): Likewise.
378 (idiv_ldiv): New mode attribute to set the 32/64-bit divide type.
379 (udiv<mode>3): Use idiv_ldiv mode attribute.
380 (div<mode>3): Likewise.
381 (addg6s): Add new BCD builtin functions.
384 (UNSPEC_DIV_EXTEND): Add support for new extended divide
386 (div_extend): Likewise.
387 (div<div_extend>_<mode>"): Likewise.
388 (FP128_64): Add support for new builtin functions to pack/unpack
390 (unpack<mode>): Likewise.
391 (unpacktf_0): Likewise.
392 (unpacktf_1): Likewise.
393 (unpack<mode>_dm): Likewise.
394 (unpack<mode>_nodm): Likewise.
395 (pack<mode>): Likewise.
396 (unpackv1ti): Likewise.
397 (packv1ti): Likewise.
399 2014-04-24 Vishnu K S <Vishnu.k_s@atmel.com>
401 * gcc/config/avr/avr.c: Add comment on why -fdelete-null-pointer-checks
404 2014-04-24 Jakub Jelinek <jakub@redhat.com>
406 * tree.h (OMP_CLAUSE_LINEAR_GIMPLE_SEQ): Define.
407 * gimplify.c (omp_is_private): Change last argument's type to int.
408 Only diagnose lastprivate if the simd argument is 1, only diagnose
409 linear if the simd argument is 2.
410 (gimplify_omp_for): Adjust omp_is_private callers. When adding
411 lastprivate or private, add the clause to OMP_FOR_CLAUSES. Pass
412 GOVD_EXPLICIT to omp_add_variable. For simd with collapse == 1
413 create OMP_CLAUSE_LINEAR rather than OMP_CLAUSE_PRIVATE for var.
414 If var != decl and decl is in OMP_CLAUSE_LINEAR, gimplify decl
415 increment to OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
416 * omp-low.c (scan_sharing_clauses, lower_lastprivate_clauses): Handle
417 OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
418 * tree-nested.c (convert_nonlocal_omp_clauses,
419 convert_local_omp_clauses): Handle OMP_CLAUSE_LINEAR.
421 2014-04-24 Segher Boessenkool <segher@kernel.crashing.org>
424 * config/m68k/m68k.md (extendplussidi): Don't allow memory for
427 2014-04-24 Dimitris Papavasiliou <dpapavas@gmail.com>
429 * flag-types.h (enum ivar_visibility): Add.
431 2014-04-24 Trevor Saunders <tsaunders@mozilla.com>
433 * config/sh/sh_treg_combine.c (sh_treg_combine::execute): Take
436 2014-04-24 Alan Lawrence <alan.lawrence@arm.com>
438 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Enable for bigendian.
440 2014-04-24 Radovan Obradovic <robradovic@mips.com>
441 Tom de Vries <tom@codesourcery.com>
443 * reg-notes.def (REG_NOTE (CALL_DECL)): New reg-note REG_CALL_DECL.
444 * calls.c (expand_call, emit_library_call_value_1): Add REG_CALL_DECL
446 * combine.c (distribute_notes): Handle REG_CALL_DECL reg-note.
447 * emit-rtl.c (try_split): Same.
449 2014-04-24 Radovan Obradovic <robradovic@mips.com>
450 Tom de Vries <tom@codesourcery.com>
452 * common.opt (fuse-caller-save): New option.
454 2014-04-24 Tejas Belagod <tejas.belagod@arm.com>
456 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Reverse order of elements
459 2014-04-24 Richard Biener <rguenther@suse.de>
461 * expr.c (expand_expr_real_1): Avoid gimple_assign_rhs_to_tree
462 during TER and instead use the sepops interface for expanding
463 non-GIMPLE_SINGLE_RHS.
465 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
467 * config/i386/sol2.h (ASM_PREFERRED_EH_DATA_FORMAT): Only redefine
468 if not HAVE_AS_IX86_DIFF_SECT_DELTA.
470 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
472 * configure.ac (gcc_cv_as_cfi_directive): Support Solaris/x86
473 assembler 64-bit option.
474 * configure: Regenerate.
476 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
478 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Check
479 TARGET_SIMD rather than TARGET_GENERAL_REGS_ONLY.
480 (TARGET_SIMD): Take AARCH64_ISA_SIMD into account.
481 (TARGET_FLOAT): Take AARCH64_ISA_FP into account.
482 (TARGET_CRYPTO): Take TARGET_SIMD into account.
484 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
486 * config/aarch64/aarch64-builtins.c
487 (aarch64_builtin_vectorized_function): Handle BUILT_IN_BSWAP16,
488 BUILT_IN_BSWAP32, BUILT_IN_BSWAP64.
489 * config/aarch64/aarch64-simd.md (bswap<mode>): New pattern.
490 * config/aarch64/aarch64-simd-builtins.def: Define vector bswap
492 * config/aarch64/iterator.md (VDQHSD): New mode iterator.
493 (Vrevsuff): New mode attribute.
495 2014-04-24 Terry Guo <terry.guo@arm.com>
497 * config/arm/arm.h (machine_function): Define variable
498 after_arm_reorg here.
499 * config/arm/arm.c (after_arm_reorg): Remove the definition.
500 (arm_split_constant): Update the way to access variable
503 (arm_output_function_epilogue): Remove the reset of after_arm_reorg.
505 2014-04-23 Tom de Vries <tom@codesourcery.com>
507 * target-hooks-macros.h: Fix DEFHOOKPOD argument order in comment.
509 2014-04-23 David Malcolm <dmalcolm@redhat.com>
511 * is-a.h: Update comments to reflect the following changes to the
512 "pointerness" of the API, making the template parameter match the
513 return type, allowing use of is-a.h with typedefs of pointers.
514 (is_a_helper::cast): Return a T rather then a pointer to a T, so
515 that the return type matches the parameter to the is_a_helper.
517 (dyn_cast): Likewise.
519 * cgraph.c (cgraph_node_for_asm): Update for removal of implicit
520 pointer from the is-a.h API.
522 * cgraph.h (is_a_helper <cgraph_node>::test): Convert to...
523 (is_a_helper <cgraph_node *>::test): ...this, matching change to
525 (is_a_helper <varpool_node>::test): Likewise, convert to...
526 (is_a_helper <varpool_node *>::test): ...this.
528 (varpool_first_variable): Update for removal of implicit pointer
530 (varpool_next_variable): Likewise.
531 (varpool_first_static_initializer): Likewise.
532 (varpool_next_static_initializer): Likewise.
533 (varpool_first_defined_variable): Likewise.
534 (varpool_next_defined_variable): Likewise.
535 (cgraph_first_defined_function): Likewise.
536 (cgraph_next_defined_function): Likewise.
537 (cgraph_first_function): Likewise.
538 (cgraph_next_function): Likewise.
539 (cgraph_first_function_with_gimple_body): Likewise.
540 (cgraph_next_function_with_gimple_body): Likewise.
541 (cgraph_alias_target): Likewise.
542 (varpool_alias_target): Likewise.
543 (cgraph_function_or_thunk_node): Likewise.
544 (varpool_variable_node): Likewise.
545 (symtab_real_symbol_p): Likewise.
546 * cgraphunit.c (referred_to_p): Likewise.
547 (analyze_functions): Likewise.
548 (handle_alias_pairs): Likewise.
549 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
550 * gimple-ssa.h (gimple_vuse_op): Likewise.
551 (gimple_vdef_op): Likewise.
552 * gimple-streamer-in.c (input_gimple_stmt): Likewise.
553 * gimple.c (gimple_build_asm_1): Likewise.
554 (gimple_build_try): Likewise.
555 (gimple_build_resx): Likewise.
556 (gimple_build_eh_dispatch): Likewise.
557 (gimple_build_omp_for): Likewise.
558 (gimple_omp_for_set_clauses): Likewise.
560 * gimple.h (is_a_helper <gimple_statement_asm>::test): Convert to...
561 (is_a_helper <gimple_statement_asm *>::test): ...this.
562 (is_a_helper <gimple_statement_bind>::test): Convert to...
563 (is_a_helper <gimple_statement_bind *>::test): ...this.
564 (is_a_helper <gimple_statement_call>::test): Convert to...
565 (is_a_helper <gimple_statement_call *>::test): ...this.
566 (is_a_helper <gimple_statement_catch>::test): Convert to...
567 (is_a_helper <gimple_statement_catch *>::test): ...this.
568 (is_a_helper <gimple_statement_resx>::test): Convert to...
569 (is_a_helper <gimple_statement_resx *>::test): ...this.
570 (is_a_helper <gimple_statement_eh_dispatch>::test): Convert to...
571 (is_a_helper <gimple_statement_eh_dispatch *>::test): ...this.
572 (is_a_helper <gimple_statement_eh_else>::test): Convert to...
573 (is_a_helper <gimple_statement_eh_else *>::test): ...this.
574 (is_a_helper <gimple_statement_eh_filter>::test): Convert to...
575 (is_a_helper <gimple_statement_eh_filter *>::test): ...this.
576 (is_a_helper <gimple_statement_eh_mnt>::test): Convert to...
577 (is_a_helper <gimple_statement_eh_mnt *>::test): ...this.
578 (is_a_helper <gimple_statement_omp_atomic_load>::test): Convert to...
579 (is_a_helper <gimple_statement_omp_atomic_load *>::test): ...this.
580 (is_a_helper <gimple_statement_omp_atomic_store>::test): Convert to...
581 (is_a_helper <gimple_statement_omp_atomic_store *>::test): ...this.
582 (is_a_helper <gimple_statement_omp_return>::test): Convert to...
583 (is_a_helper <gimple_statement_omp_return *>::test): ...this.
584 (is_a_helper <gimple_statement_omp_continue>::test): Convert to...
585 (is_a_helper <gimple_statement_omp_continue *>::test): ...this.
586 (is_a_helper <gimple_statement_omp_critical>::test): Convert to...
587 (is_a_helper <gimple_statement_omp_critical *>::test): ...this.
588 (is_a_helper <gimple_statement_omp_for>::test): Convert to...
589 (is_a_helper <gimple_statement_omp_for *>::test): ...this.
590 (is_a_helper <gimple_statement_omp_taskreg>::test): Convert to...
591 (is_a_helper <gimple_statement_omp_taskreg *>::test): ...this.
592 (is_a_helper <gimple_statement_omp_parallel>::test): Convert to...
593 (is_a_helper <gimple_statement_omp_parallel *>::test): ...this.
594 (is_a_helper <gimple_statement_omp_target>::test): Convert to...
595 (is_a_helper <gimple_statement_omp_target *>::test): ...this.
596 (is_a_helper <gimple_statement_omp_sections>::test): Convert to...
597 (is_a_helper <gimple_statement_omp_sections *>::test): ...this.
598 (is_a_helper <gimple_statement_omp_single>::test): Convert to...
599 (is_a_helper <gimple_statement_omp_single *>::test): ...this.
600 (is_a_helper <gimple_statement_omp_teams>::test): Convert to...
601 (is_a_helper <gimple_statement_omp_teams *>::test): ...this.
602 (is_a_helper <gimple_statement_omp_task>::test): Convert to...
603 (is_a_helper <gimple_statement_omp_task *>::test): ...this.
604 (is_a_helper <gimple_statement_phi>::test): Convert to...
605 (is_a_helper <gimple_statement_phi *>::test): ...this.
606 (is_a_helper <gimple_statement_transaction>::test): Convert to...
607 (is_a_helper <gimple_statement_transaction *>::test): ...this.
608 (is_a_helper <gimple_statement_try>::test): Convert to...
609 (is_a_helper <gimple_statement_try *>::test): ...this.
610 (is_a_helper <gimple_statement_wce>::test): Convert to...
611 (is_a_helper <gimple_statement_wce *>::test): ...this.
612 (is_a_helper <const gimple_statement_asm>::test): Convert to...
613 (is_a_helper <const gimple_statement_asm *>::test): ...this.
614 (is_a_helper <const gimple_statement_bind>::test): Convert to...
615 (is_a_helper <const gimple_statement_bind *>::test): ...this.
616 (is_a_helper <const gimple_statement_call>::test): Convert to...
617 (is_a_helper <const gimple_statement_call *>::test): ...this.
618 (is_a_helper <const gimple_statement_catch>::test): Convert to...
619 (is_a_helper <const gimple_statement_catch *>::test): ...this.
620 (is_a_helper <const gimple_statement_resx>::test): Convert to...
621 (is_a_helper <const gimple_statement_resx *>::test): ...this.
622 (is_a_helper <const gimple_statement_eh_dispatch>::test):
624 (is_a_helper <const gimple_statement_eh_dispatch *>::test):
626 (is_a_helper <const gimple_statement_eh_filter>::test): Convert
628 (is_a_helper <const gimple_statement_eh_filter *>::test): ...this.
629 (is_a_helper <const gimple_statement_omp_atomic_load>::test):
631 (is_a_helper <const gimple_statement_omp_atomic_load *>::test):
633 (is_a_helper <const gimple_statement_omp_atomic_store>::test):
635 (is_a_helper <const gimple_statement_omp_atomic_store *>::test):
637 (is_a_helper <const gimple_statement_omp_return>::test): Convert
639 (is_a_helper <const gimple_statement_omp_return *>::test): ...this.
640 (is_a_helper <const gimple_statement_omp_continue>::test): Convert
642 (is_a_helper <const gimple_statement_omp_continue *>::test): ...this.
643 (is_a_helper <const gimple_statement_omp_critical>::test): Convert
645 (is_a_helper <const gimple_statement_omp_critical *>::test): ...this.
646 (is_a_helper <const gimple_statement_omp_for>::test): Convert to...
647 (is_a_helper <const gimple_statement_omp_for *>::test): ...this.
648 (is_a_helper <const gimple_statement_omp_taskreg>::test): Convert
650 (is_a_helper <const gimple_statement_omp_taskreg *>::test): ...this.
651 (is_a_helper <const gimple_statement_omp_parallel>::test): Convert
653 (is_a_helper <const gimple_statement_omp_parallel *>::test): ...this.
654 (is_a_helper <const gimple_statement_omp_target>::test): Convert
656 (is_a_helper <const gimple_statement_omp_target *>::test): ...this.
657 (is_a_helper <const gimple_statement_omp_sections>::test): Convert
659 (is_a_helper <const gimple_statement_omp_sections *>::test): ...this.
660 (is_a_helper <const gimple_statement_omp_single>::test): Convert
662 (is_a_helper <const gimple_statement_omp_single *>::test): ...this.
663 (is_a_helper <const gimple_statement_omp_teams>::test): Convert
665 (is_a_helper <const gimple_statement_omp_teams *>::test): ...this.
666 (is_a_helper <const gimple_statement_omp_task>::test): Convert
668 (is_a_helper <const gimple_statement_omp_task *>::test): ...this.
669 (is_a_helper <const gimple_statement_phi>::test): Convert to...
670 (is_a_helper <const gimple_statement_phi *>::test): ...this.
671 (is_a_helper <const gimple_statement_transaction>::test): Convert
673 (is_a_helper <const gimple_statement_transaction *>::test): ...this.
674 (is_a_helper <const gimple_statement_with_ops>::test): Convert
676 (is_a_helper <const gimple_statement_with_ops *>::test): ...this.
677 (is_a_helper <gimple_statement_with_ops>::test): Convert to...
678 (is_a_helper <gimple_statement_with_ops *>::test): ...this.
679 (is_a_helper <const gimple_statement_with_memory_ops>::test):
681 (is_a_helper <const gimple_statement_with_memory_ops *>::test):
683 (is_a_helper <gimple_statement_with_memory_ops>::test): Convert
685 (is_a_helper <gimple_statement_with_memory_ops *>::test): ...this.
687 (gimple_use_ops): Update for removal of implicit pointer from the
689 (gimple_set_use_ops): Likewise.
690 (gimple_vuse): Likewise.
691 (gimple_vdef): Likewise.
692 (gimple_vuse_ptr): Likewise.
693 (gimple_vdef_ptr): Likewise.
694 (gimple_set_vuse): Likewise.
695 (gimple_set_vdef): Likewise.
696 (gimple_omp_return_set_lhs): Likewise.
697 (gimple_omp_return_lhs): Likewise.
698 (gimple_omp_return_lhs_ptr): Likewise.
699 (gimple_call_fntype): Likewise.
700 (gimple_call_set_fntype): Likewise.
701 (gimple_call_set_internal_fn): Likewise.
702 (gimple_call_use_set): Likewise.
703 (gimple_call_clobber_set): Likewise.
704 (gimple_bind_vars): Likewise.
705 (gimple_bind_set_vars): Likewise.
706 (gimple_bind_body_ptr): Likewise.
707 (gimple_bind_set_body): Likewise.
708 (gimple_bind_add_stmt): Likewise.
709 (gimple_bind_block): Likewise.
710 (gimple_bind_set_block): Likewise.
711 (gimple_asm_ninputs): Likewise.
712 (gimple_asm_noutputs): Likewise.
713 (gimple_asm_nclobbers): Likewise.
714 (gimple_asm_nlabels): Likewise.
715 (gimple_asm_input_op): Likewise.
716 (gimple_asm_input_op_ptr): Likewise.
717 (gimple_asm_output_op): Likewise.
718 (gimple_asm_output_op_ptr): Likewise.
719 (gimple_asm_set_output_op): Likewise.
720 (gimple_asm_clobber_op): Likewise.
721 (gimple_asm_set_clobber_op): Likewise.
722 (gimple_asm_label_op): Likewise.
723 (gimple_asm_set_label_op): Likewise.
724 (gimple_asm_string): Likewise.
725 (gimple_catch_types): Likewise.
726 (gimple_catch_types_ptr): Likewise.
727 (gimple_catch_handler_ptr): Likewise.
728 (gimple_catch_set_types): Likewise.
729 (gimple_catch_set_handler): Likewise.
730 (gimple_eh_filter_types): Likewise.
731 (gimple_eh_filter_types_ptr): Likewise.
732 (gimple_eh_filter_failure_ptr): Likewise.
733 (gimple_eh_filter_set_types): Likewise.
734 (gimple_eh_filter_set_failure): Likewise.
735 (gimple_eh_must_not_throw_fndecl): Likewise.
736 (gimple_eh_must_not_throw_set_fndecl): Likewise.
737 (gimple_eh_else_n_body_ptr): Likewise.
738 (gimple_eh_else_e_body_ptr): Likewise.
739 (gimple_eh_else_set_n_body): Likewise.
740 (gimple_eh_else_set_e_body): Likewise.
741 (gimple_try_eval_ptr): Likewise.
742 (gimple_try_cleanup_ptr): Likewise.
743 (gimple_try_set_eval): Likewise.
744 (gimple_try_set_cleanup): Likewise.
745 (gimple_wce_cleanup_ptr): Likewise.
746 (gimple_wce_set_cleanup): Likewise.
747 (gimple_phi_capacity): Likewise.
748 (gimple_phi_num_args): Likewise.
749 (gimple_phi_result): Likewise.
750 (gimple_phi_result_ptr): Likewise.
751 (gimple_phi_set_result): Likewise.
752 (gimple_phi_arg): Likewise.
753 (gimple_phi_set_arg): Likewise.
754 (gimple_resx_region): Likewise.
755 (gimple_resx_set_region): Likewise.
756 (gimple_eh_dispatch_region): Likewise.
757 (gimple_eh_dispatch_set_region): Likewise.
758 (gimple_omp_critical_name): Likewise.
759 (gimple_omp_critical_name_ptr): Likewise.
760 (gimple_omp_critical_set_name): Likewise.
761 (gimple_omp_for_clauses): Likewise.
762 (gimple_omp_for_clauses_ptr): Likewise.
763 (gimple_omp_for_set_clauses): Likewise.
764 (gimple_omp_for_collapse): Likewise.
765 (gimple_omp_for_index): Likewise.
766 (gimple_omp_for_index_ptr): Likewise.
767 (gimple_omp_for_set_index): Likewise.
768 (gimple_omp_for_initial): Likewise.
769 (gimple_omp_for_initial_ptr): Likewise.
770 (gimple_omp_for_set_initial): Likewise.
771 (gimple_omp_for_final): Likewise.
772 (gimple_omp_for_final_ptr): Likewise.
773 (gimple_omp_for_set_final): Likewise.
774 (gimple_omp_for_incr): Likewise.
775 (gimple_omp_for_incr_ptr): Likewise.
776 (gimple_omp_for_set_incr): Likewise.
777 (gimple_omp_for_pre_body_ptr): Likewise.
778 (gimple_omp_for_set_pre_body): Likewise.
779 (gimple_omp_parallel_clauses): Likewise.
780 (gimple_omp_parallel_clauses_ptr): Likewise.
781 (gimple_omp_parallel_set_clauses): Likewise.
782 (gimple_omp_parallel_child_fn): Likewise.
783 (gimple_omp_parallel_child_fn_ptr): Likewise.
784 (gimple_omp_parallel_set_child_fn): Likewise.
785 (gimple_omp_parallel_data_arg): Likewise.
786 (gimple_omp_parallel_data_arg_ptr): Likewise.
787 (gimple_omp_parallel_set_data_arg): Likewise.
788 (gimple_omp_task_clauses): Likewise.
789 (gimple_omp_task_clauses_ptr): Likewise.
790 (gimple_omp_task_set_clauses): Likewise.
791 (gimple_omp_task_child_fn): Likewise.
792 (gimple_omp_task_child_fn_ptr): Likewise.
793 (gimple_omp_task_set_child_fn): Likewise.
794 (gimple_omp_task_data_arg): Likewise.
795 (gimple_omp_task_data_arg_ptr): Likewise.
796 (gimple_omp_task_set_data_arg): Likewise.
797 (gimple_omp_taskreg_clauses): Likewise.
798 (gimple_omp_taskreg_clauses_ptr): Likewise.
799 (gimple_omp_taskreg_set_clauses): Likewise.
800 (gimple_omp_taskreg_child_fn): Likewise.
801 (gimple_omp_taskreg_child_fn_ptr): Likewise.
802 (gimple_omp_taskreg_set_child_fn): Likewise.
803 (gimple_omp_taskreg_data_arg): Likewise.
804 (gimple_omp_taskreg_data_arg_ptr): Likewise.
805 (gimple_omp_taskreg_set_data_arg): Likewise.
806 (gimple_omp_task_copy_fn): Likewise.
807 (gimple_omp_task_copy_fn_ptr): Likewise.
808 (gimple_omp_task_set_copy_fn): Likewise.
809 (gimple_omp_task_arg_size): Likewise.
810 (gimple_omp_task_arg_size_ptr): Likewise.
811 (gimple_omp_task_set_arg_size): Likewise.
812 (gimple_omp_task_arg_align): Likewise.
813 (gimple_omp_task_arg_align_ptr): Likewise.
814 (gimple_omp_task_set_arg_align): Likewise.
815 (gimple_omp_single_clauses): Likewise.
816 (gimple_omp_single_clauses_ptr): Likewise.
817 (gimple_omp_single_set_clauses): Likewise.
818 (gimple_omp_target_clauses): Likewise.
819 (gimple_omp_target_clauses_ptr): Likewise.
820 (gimple_omp_target_set_clauses): Likewise.
821 (gimple_omp_target_child_fn): Likewise.
822 (gimple_omp_target_child_fn_ptr): Likewise.
823 (gimple_omp_target_set_child_fn): Likewise.
824 (gimple_omp_target_data_arg): Likewise.
825 (gimple_omp_target_data_arg_ptr): Likewise.
826 (gimple_omp_target_set_data_arg): Likewise.
827 (gimple_omp_teams_clauses): Likewise.
828 (gimple_omp_teams_clauses_ptr): Likewise.
829 (gimple_omp_teams_set_clauses): Likewise.
830 (gimple_omp_sections_clauses): Likewise.
831 (gimple_omp_sections_clauses_ptr): Likewise.
832 (gimple_omp_sections_set_clauses): Likewise.
833 (gimple_omp_sections_control): Likewise.
834 (gimple_omp_sections_control_ptr): Likewise.
835 (gimple_omp_sections_set_control): Likewise.
836 (gimple_omp_for_set_cond): Likewise.
837 (gimple_omp_for_cond): Likewise.
838 (gimple_omp_atomic_store_set_val): Likewise.
839 (gimple_omp_atomic_store_val): Likewise.
840 (gimple_omp_atomic_store_val_ptr): Likewise.
841 (gimple_omp_atomic_load_set_lhs): Likewise.
842 (gimple_omp_atomic_load_lhs): Likewise.
843 (gimple_omp_atomic_load_lhs_ptr): Likewise.
844 (gimple_omp_atomic_load_set_rhs): Likewise.
845 (gimple_omp_atomic_load_rhs): Likewise.
846 (gimple_omp_atomic_load_rhs_ptr): Likewise.
847 (gimple_omp_continue_control_def): Likewise.
848 (gimple_omp_continue_control_def_ptr): Likewise.
849 (gimple_omp_continue_set_control_def): Likewise.
850 (gimple_omp_continue_control_use): Likewise.
851 (gimple_omp_continue_control_use_ptr): Likewise.
852 (gimple_omp_continue_set_control_use): Likewise.
853 (gimple_transaction_body_ptr): Likewise.
854 (gimple_transaction_label): Likewise.
855 (gimple_transaction_label_ptr): Likewise.
856 (gimple_transaction_set_body): Likewise.
857 (gimple_transaction_set_label): Likewise.
859 * ipa-devirt.c (build_type_inheritance_graph): Likewise.
860 * ipa-inline-analysis.c (inline_write_summary): Likewise.
861 * ipa-ref.c (ipa_record_reference): Likewise.
862 * ipa-reference.c (analyze_function): Likewise.
863 (ipa_reference_write_optimization_summary): Likewise.
864 * ipa.c (symtab_remove_unreachable_nodes): Likewise.
865 (address_taken_from_non_vtable_p): Likewise.
866 (comdat_can_be_unshared_p_1): Likewise.
867 * lto-cgraph.c (lto_output_ref): Likewise.
868 (add_references): Likewise.
869 (compute_ltrans_boundary): Likewise.
870 (output_symtab): Likewise.
871 (input_ref): Likewise.
872 (input_cgraph_1): Likewise.
873 (output_cgraph_opt_summary): Likewise.
874 * lto-streamer-out.c (lto_output): Likewise.
875 (output_symbol_p): Likewise.
876 * lto-streamer.h (lsei_next_function_in_partition): Likewise.
877 (lsei_start_function_in_partition): Likewise.
878 (lsei_next_variable_in_partition): Likewise.
879 (lsei_start_variable_in_partition): Likewise.
880 * symtab.c (insert_to_assembler_name_hash): Likewise.
881 (unlink_from_assembler_name_hash): Likewise.
882 (symtab_unregister_node): Likewise.
883 (symtab_remove_node): Likewise.
884 (dump_symtab_node): Likewise.
885 (verify_symtab_base): Likewise.
886 (verify_symtab_node): Likewise.
887 (symtab_make_decl_local): Likewise.
888 (symtab_alias_ultimate_target): Likewise.
889 (symtab_resolve_alias): Likewise.
890 (symtab_get_symbol_partitioning_class): Likewise.
891 * tree-phinodes.c (allocate_phi_node): Likewise.
892 (reserve_phi_args_for_new_edge): Likewise.
893 (remove_phi_args): Likewise.
894 * varpool.c (varpool_node_for_asm): Likewise.
895 (varpool_remove_unreferenced_decls): Likewise.
897 2014-04-23 Jeff Law <law@redhat.com>
899 PR tree-optimization/60902
900 * tree-ssa-threadedge.c
901 (record_temporary_equivalences_from_stmts_at_dest): Make sure to
902 invalidate outputs from statements that do not produce useful
903 outputs for threading.
905 2014-04-23 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
907 * config/aarch64/aarch64.md (stack_protect_set, stack_protect_test)
908 (stack_protect_set_<mode>, stack_protect_test_<mode>): Add
909 machine descriptions for Stack Smashing Protector.
911 2014-04-23 Richard Earnshaw <rearnsha@arm.com>
913 * aarch64.md (<optab>_rol<mode>3): New pattern.
914 (<optab>_rolsi3_uxtw): Likewise.
915 * aarch64.c (aarch64_strip_shift): Handle ROTATE and ROTATERT.
917 2014-04-23 James Greenhalgh <james.greenhalgh@arm.com>
919 * config/arm/arm.c (arm_cortex_a57_tune): Initialize all fields.
920 (arm_cortex_a12_tune): Likewise.
922 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
924 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle BSWAP.
926 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
928 * config/arm/arm.md (arm_rev16si2): New pattern.
929 (arm_rev16si2_alt): Likewise.
930 * config/arm/arm.c (arm_new_rtx_costs): Handle rev16 case.
932 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
934 * config/aarch64/aarch64.md (rev16<mode>2): New pattern.
935 (rev16<mode>2_alt): Likewise.
936 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle rev16 case.
937 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p): New.
938 (aarch_rev16_shleft_mask_imm_p): Likewise.
939 (aarch_rev16_p_1): Likewise.
940 (aarch_rev16_p): Likewise.
941 * config/arm/aarch-common-protos.h (aarch_rev16_p): Declare extern.
942 (aarch_rev16_shright_mask_imm_p): Likewise.
943 (aarch_rev16_shleft_mask_imm_p): Likewise.
945 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
947 * config/arm/aarch-common-protos.h (alu_cost_table): Add rev field.
948 * config/arm/aarch-cost-tables.h (generic_extra_costs): Specify
950 (cortex_a53_extra_costs): Likewise.
951 (cortex_a57_extra_costs): Likewise.
952 * config/arm/arm.c (cortexa9_extra_costs): Likewise.
953 (cortexa7_extra_costs): Likewise.
954 (cortexa8_extra_costs): Likewise.
955 (cortexa12_extra_costs): Likewise.
956 (cortexa15_extra_costs): Likewise.
957 (v7m_extra_costs): Likewise.
958 (arm_new_rtx_costs): Handle BSWAP.
960 2013-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
962 * config/arm/arm.c (cortexa8_extra_costs): New table.
963 (arm_cortex_a8_tune): New tuning struct.
964 * config/arm/arm-cores.def (cortex-a8): Use cortex_a8 tuning struct.
966 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
968 * config/arm/arm.c (arm_new_rtx_costs): Handle FMA.
970 2014-04-23 Richard Biener <rguenther@suse.de>
972 * Makefile.in (OBJS): Remove loop-unswitch.o.
973 * loop-unswitch.c: Delete.
974 * tree-pass.h (make_pass_rtl_unswitch): Remove.
975 * passes.def (pass_rtl_unswitch): Likewise.
976 * loop-init.c (gate_rtl_unswitch): Likewise.
977 (rtl_unswitch): Likewise.
978 (pass_data_rtl_unswitch): Likewise.
979 (pass_rtl_unswitch): Likewise.
980 (make_pass_rtl_unswitch): Likewise.
981 * rtl.h (reversed_condition): Likewise.
982 (compare_and_jump_seq): Likewise.
983 * loop-iv.c (reversed_condition): Move here from loop-unswitch.c
985 * loop-unroll.c (compare_and_jump_seq): Likewise.
987 2014-04-23 Richard Biener <rguenther@suse.de>
989 PR tree-optimization/60903
990 * tree-ssa-loop-im.c (analyze_memory_references): Remove
991 commented code block.
992 (execute_sm_if_changed): Properly apply IRREDUCIBLE_LOOP
993 loop flags to newly created BBs and edges.
995 2014-04-23 Nick Clifton <nickc@redhat.com>
997 * config/msp430/msp430.c (msp430_handle_option): Move function
999 (msp430_option_override): Simplify mcu and mcpu option handling.
1000 (msp430_is_f5_mcu): Rename to msp430_use_f5_series_hwmult. Add
1001 support for -mhwmult command line option.
1002 (has_32bit_hwmult): Rename to use_32bit_hwmult. Add support for
1003 -mhwmult command line option.
1004 (msp430_hwmult_enabled): Delete.
1005 (msp43o_output_labelref): Add support for -mhwmult command line
1007 * config/msp430/msp430.md (mulhisi3, umulhisi3, mulsidi3)
1008 (umulsidi3): Likewise.
1009 * config/msp430/msp430.opt (mmcu): Add Report attribute.
1010 (mcpu, mlarge, msmall): Likewise.
1011 (mhwmult): New option.
1012 * config/msp430/msp430-protos.h (msp430_hwmult_enabled): Remove
1014 (msp430_is_f5_mcu): Remove prototype.
1015 (msp430_use_f5_series_hwmult): Add prototype.
1016 * config/msp430/msp430-opts.h: New file.
1017 * common/config/msp430: New directory.
1018 * common/config/msp430/msp430-common.c: New file.
1019 * config.gcc (msp430): Remove target_has_targetm_common.
1020 * doc/invoke.texi: Document -mhwmult command line option.
1022 2014-04-23 Nick Clifton <nickc@redhat.com>
1024 * config/i386/cygwin.h (ENDFILE_SPEC): Include
1025 default-manifest.o if it can be found in the search path.
1026 * config/i386/mingw32.h (ENDFILE_SPEC): Likewise.
1028 2014-04-23 Terry Guo <terry.guo@arm.com>
1030 * config/arm/arm.h (ASM_APP_OFF): Re-define it in a cleaner way.
1032 2014-04-23 Richard Biener <rguenther@suse.de>
1035 * tree-inline.c (declare_return_variable): Use mark_addressable.
1037 2014-04-23 Richard Biener <rguenther@suse.de>
1040 * loop-init.c (loop_optimizer_init): Make sure to apply
1041 LOOPS_MAY_HAVE_MULTIPLE_LATCHES before fixing up loops.
1043 2014-04-22 Jakub Jelinek <jakub@redhat.com>
1046 * common.opt (fsanitize-recover, fsanitize-undefined-trap-on-error):
1048 * gcc.c (sanitize_spec_function): Don't return "" for "undefined"
1049 if flag_sanitize_undefined_trap_on_error.
1050 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW_ABORT,
1051 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS_ABORT,
1052 BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE_ABORT,
1053 BUILT_IN_UBSAN_HANDLE_TYPE_MISMATCH_ABORT,
1054 BUILT_IN_UBSAN_HANDLE_ADD_OVERFLOW_ABORT,
1055 BUILT_IN_UBSAN_HANDLE_SUB_OVERFLOW_ABORT,
1056 BUILT_IN_UBSAN_HANDLE_MUL_OVERFLOW_ABORT,
1057 BUILT_IN_UBSAN_HANDLE_NEGATE_OVERFLOW_ABORT,
1058 BUILT_IN_UBSAN_HANDLE_LOAD_INVALID_VALUE_ABORT): New builtins.
1059 * ubsan.c (ubsan_instrument_unreachable): Return
1060 __builtin_trap () if flag_sanitize_undefined_trap_on_error.
1061 (ubsan_expand_null_ifn): Emit __builtin_trap ()
1062 if flag_sanitize_undefined_trap_on_error and
1063 __ubsan_handle_type_mismatch_abort if !flag_sanitize_recover.
1064 (ubsan_expand_null_ifn, ubsan_build_overflow_builtin,
1065 instrument_bool_enum_load): Emit __builtin_trap () if
1066 flag_sanitize_undefined_trap_on_error and
1067 __builtin_handle_*_abort () if !flag_sanitize_recover.
1068 * doc/invoke.texi (-fsanitize-recover,
1069 -fsanitize-undefined-trap-on-error): Document.
1071 2014-04-22 Christian Bruel <christian.bruel@st.com>
1073 * config/sh/sh.md (mov<mode>): Replace movQIHI.
1074 Force immediates to SImode.
1076 2014-04-22 Sandra Loosemore <sandra@codesourcery.com>
1078 * config/nios2/nios2.md (UNSPEC_ROUND): New.
1080 * config/nios2/nios2.opt (mno-custom-round, mcustom-round=): New.
1081 * config/nios2/nios2-opts.h (N2FPU_ALL_CODES): Add round.
1082 * config/nios2/nios2.c (N2F_NO_ERRNO): Define.
1083 (nios2_fpu_insn): Add entry for round.
1084 (N2FPU_NO_ERRNO_P): Define.
1085 (nios2_custom_check_insns): Add check for N2F_NO_ERRNO and
1087 * doc/invoke.texi (Nios II Options): Document -mcustom-round.
1089 2014-04-22 Richard Henderson <rth@redhat.com>
1091 * config/aarch64/aarch64 (addti3, subti3): New expanders.
1092 (add<GPI>3_compare0): Remove leading * from name.
1093 (add<GPI>3_carryin): Likewise.
1094 (sub<GPI>3_compare0): Likewise.
1095 (sub<GPI>3_carryin): Likewise.
1096 (<su_optab>mulditi3): New expander.
1097 (multi3): New expander.
1098 (madd<GPI>): Remove leading * from name.
1100 2014-04-22 Martin Jambor <mjambor@suse.cz>
1102 * cgraphclones.c (cgraph_function_versioning): Copy
1103 ipa_transforms_to_apply instead of asserting it is empty.
1105 2014-04-22 H.J. Lu <hongjiu.lu@intel.com>
1108 * config/i386/i386.c (ix86_expand_set_or_movmem): Call counter_mode
1109 on count_exp to get mode.
1111 2014-04-22 Andrew Pinski <apinski@cavium.com>
1113 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
1114 Handle TLS for ILP32.
1115 * config/aarch64/aarch64.md (tlsie_small): Rename to ...
1116 (tlsie_small_<mode>): this and handle PTR.
1117 (tlsie_small_sidi): New pattern.
1118 (tlsle_small): Change to an expand to handle ILP32.
1119 (tlsle_small_<mode>): New pattern.
1120 (tlsdesc_small): Rename to ...
1121 (tlsdesc_small_<mode>): this and handle PTR.
1123 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1125 * config/aarch64/aarch64.c (TARGET_FLAGS_REGNUM): Define.
1127 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1129 * config/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
1130 (aarch64_types_signed_unsigned_qualifiers): Qualifier added.
1131 (aarch64_types_signed_poly_qualifiers): Likewise.
1132 (aarch64_types_unsigned_signed_qualifiers): Likewise.
1133 (aarch64_types_poly_signed_qualifiers): Likewise.
1134 (TYPES_REINTERP_SS): Type macro added.
1135 (TYPES_REINTERP_SU): Likewise.
1136 (TYPES_REINTERP_SP): Likewise.
1137 (TYPES_REINTERP_US): Likewise.
1138 (TYPES_REINTERP_PS): Likewise.
1139 (aarch64_fold_builtin): New expression folding added.
1140 * config/aarch64/aarch64-simd-builtins.def (REINTERP):
1141 Declarations removed.
1142 (REINTERP_SS): Declarations added.
1143 (REINTERP_US): Likewise.
1144 (REINTERP_PS): Likewise.
1145 (REINTERP_SU): Likewise.
1146 (REINTERP_SP): Likewise.
1147 * config/aarch64/arm_neon.h (vreinterpret_p8_f64): Implemented.
1148 (vreinterpretq_p8_f64): Likewise.
1149 (vreinterpret_p16_f64): Likewise.
1150 (vreinterpretq_p16_f64): Likewise.
1151 (vreinterpret_f32_f64): Likewise.
1152 (vreinterpretq_f32_f64): Likewise.
1153 (vreinterpret_f64_f32): Likewise.
1154 (vreinterpret_f64_p8): Likewise.
1155 (vreinterpret_f64_p16): Likewise.
1156 (vreinterpret_f64_s8): Likewise.
1157 (vreinterpret_f64_s16): Likewise.
1158 (vreinterpret_f64_s32): Likewise.
1159 (vreinterpret_f64_s64): Likewise.
1160 (vreinterpret_f64_u8): Likewise.
1161 (vreinterpret_f64_u16): Likewise.
1162 (vreinterpret_f64_u32): Likewise.
1163 (vreinterpret_f64_u64): Likewise.
1164 (vreinterpretq_f64_f32): Likewise.
1165 (vreinterpretq_f64_p8): Likewise.
1166 (vreinterpretq_f64_p16): Likewise.
1167 (vreinterpretq_f64_s8): Likewise.
1168 (vreinterpretq_f64_s16): Likewise.
1169 (vreinterpretq_f64_s32): Likewise.
1170 (vreinterpretq_f64_s64): Likewise.
1171 (vreinterpretq_f64_u8): Likewise.
1172 (vreinterpretq_f64_u16): Likewise.
1173 (vreinterpretq_f64_u32): Likewise.
1174 (vreinterpretq_f64_u64): Likewise.
1175 (vreinterpret_s64_f64): Likewise.
1176 (vreinterpretq_s64_f64): Likewise.
1177 (vreinterpret_u64_f64): Likewise.
1178 (vreinterpretq_u64_f64): Likewise.
1179 (vreinterpret_s8_f64): Likewise.
1180 (vreinterpretq_s8_f64): Likewise.
1181 (vreinterpret_s16_f64): Likewise.
1182 (vreinterpretq_s16_f64): Likewise.
1183 (vreinterpret_s32_f64): Likewise.
1184 (vreinterpretq_s32_f64): Likewise.
1185 (vreinterpret_u8_f64): Likewise.
1186 (vreinterpretq_u8_f64): Likewise.
1187 (vreinterpret_u16_f64): Likewise.
1188 (vreinterpretq_u16_f64): Likewise.
1189 (vreinterpret_u32_f64): Likewise.
1190 (vreinterpretq_u32_f64): Likewise.
1192 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1194 * config/aarch64/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
1195 * config/aarch64/aarch64/aarch64-simd-builtins.def (REINTERP): Removed.
1196 (vreinterpret_p8_s8): Likewise.
1197 * config/aarch64/aarch64/arm_neon.h (vreinterpret_p8_s8): Uses cast.
1198 (vreinterpret_p8_s16): Likewise.
1199 (vreinterpret_p8_s32): Likewise.
1200 (vreinterpret_p8_s64): Likewise.
1201 (vreinterpret_p8_f32): Likewise.
1202 (vreinterpret_p8_u8): Likewise.
1203 (vreinterpret_p8_u16): Likewise.
1204 (vreinterpret_p8_u32): Likewise.
1205 (vreinterpret_p8_u64): Likewise.
1206 (vreinterpret_p8_p16): Likewise.
1207 (vreinterpretq_p8_s8): Likewise.
1208 (vreinterpretq_p8_s16): Likewise.
1209 (vreinterpretq_p8_s32): Likewise.
1210 (vreinterpretq_p8_s64): Likewise.
1211 (vreinterpretq_p8_f32): Likewise.
1212 (vreinterpretq_p8_u8): Likewise.
1213 (vreinterpretq_p8_u16): Likewise.
1214 (vreinterpretq_p8_u32): Likewise.
1215 (vreinterpretq_p8_u64): Likewise.
1216 (vreinterpretq_p8_p16): Likewise.
1217 (vreinterpret_p16_s8): Likewise.
1218 (vreinterpret_p16_s16): Likewise.
1219 (vreinterpret_p16_s32): Likewise.
1220 (vreinterpret_p16_s64): Likewise.
1221 (vreinterpret_p16_f32): Likewise.
1222 (vreinterpret_p16_u8): Likewise.
1223 (vreinterpret_p16_u16): Likewise.
1224 (vreinterpret_p16_u32): Likewise.
1225 (vreinterpret_p16_u64): Likewise.
1226 (vreinterpret_p16_p8): Likewise.
1227 (vreinterpretq_p16_s8): Likewise.
1228 (vreinterpretq_p16_s16): Likewise.
1229 (vreinterpretq_p16_s32): Likewise.
1230 (vreinterpretq_p16_s64): Likewise.
1231 (vreinterpretq_p16_f32): Likewise.
1232 (vreinterpretq_p16_u8): Likewise.
1233 (vreinterpretq_p16_u16): Likewise.
1234 (vreinterpretq_p16_u32): Likewise.
1235 (vreinterpretq_p16_u64): Likewise.
1236 (vreinterpretq_p16_p8): Likewise.
1237 (vreinterpret_f32_s8): Likewise.
1238 (vreinterpret_f32_s16): Likewise.
1239 (vreinterpret_f32_s32): Likewise.
1240 (vreinterpret_f32_s64): Likewise.
1241 (vreinterpret_f32_u8): Likewise.
1242 (vreinterpret_f32_u16): Likewise.
1243 (vreinterpret_f32_u32): Likewise.
1244 (vreinterpret_f32_u64): Likewise.
1245 (vreinterpret_f32_p8): Likewise.
1246 (vreinterpret_f32_p16): Likewise.
1247 (vreinterpretq_f32_s8): Likewise.
1248 (vreinterpretq_f32_s16): Likewise.
1249 (vreinterpretq_f32_s32): Likewise.
1250 (vreinterpretq_f32_s64): Likewise.
1251 (vreinterpretq_f32_u8): Likewise.
1252 (vreinterpretq_f32_u16): Likewise.
1253 (vreinterpretq_f32_u32): Likewise.
1254 (vreinterpretq_f32_u64): Likewise.
1255 (vreinterpretq_f32_p8): Likewise.
1256 (vreinterpretq_f32_p16): Likewise.
1257 (vreinterpret_s64_s8): Likewise.
1258 (vreinterpret_s64_s16): Likewise.
1259 (vreinterpret_s64_s32): Likewise.
1260 (vreinterpret_s64_f32): Likewise.
1261 (vreinterpret_s64_u8): Likewise.
1262 (vreinterpret_s64_u16): Likewise.
1263 (vreinterpret_s64_u32): Likewise.
1264 (vreinterpret_s64_u64): Likewise.
1265 (vreinterpret_s64_p8): Likewise.
1266 (vreinterpret_s64_p16): Likewise.
1267 (vreinterpretq_s64_s8): Likewise.
1268 (vreinterpretq_s64_s16): Likewise.
1269 (vreinterpretq_s64_s32): Likewise.
1270 (vreinterpretq_s64_f32): Likewise.
1271 (vreinterpretq_s64_u8): Likewise.
1272 (vreinterpretq_s64_u16): Likewise.
1273 (vreinterpretq_s64_u32): Likewise.
1274 (vreinterpretq_s64_u64): Likewise.
1275 (vreinterpretq_s64_p8): Likewise.
1276 (vreinterpretq_s64_p16): Likewise.
1277 (vreinterpret_u64_s8): Likewise.
1278 (vreinterpret_u64_s16): Likewise.
1279 (vreinterpret_u64_s32): Likewise.
1280 (vreinterpret_u64_s64): Likewise.
1281 (vreinterpret_u64_f32): Likewise.
1282 (vreinterpret_u64_u8): Likewise.
1283 (vreinterpret_u64_u16): Likewise.
1284 (vreinterpret_u64_u32): Likewise.
1285 (vreinterpret_u64_p8): Likewise.
1286 (vreinterpret_u64_p16): Likewise.
1287 (vreinterpretq_u64_s8): Likewise.
1288 (vreinterpretq_u64_s16): Likewise.
1289 (vreinterpretq_u64_s32): Likewise.
1290 (vreinterpretq_u64_s64): Likewise.
1291 (vreinterpretq_u64_f32): Likewise.
1292 (vreinterpretq_u64_u8): Likewise.
1293 (vreinterpretq_u64_u16): Likewise.
1294 (vreinterpretq_u64_u32): Likewise.
1295 (vreinterpretq_u64_p8): Likewise.
1296 (vreinterpretq_u64_p16): Likewise.
1297 (vreinterpret_s8_s16): Likewise.
1298 (vreinterpret_s8_s32): Likewise.
1299 (vreinterpret_s8_s64): Likewise.
1300 (vreinterpret_s8_f32): Likewise.
1301 (vreinterpret_s8_u8): Likewise.
1302 (vreinterpret_s8_u16): Likewise.
1303 (vreinterpret_s8_u32): Likewise.
1304 (vreinterpret_s8_u64): Likewise.
1305 (vreinterpret_s8_p8): Likewise.
1306 (vreinterpret_s8_p16): Likewise.
1307 (vreinterpretq_s8_s16): Likewise.
1308 (vreinterpretq_s8_s32): Likewise.
1309 (vreinterpretq_s8_s64): Likewise.
1310 (vreinterpretq_s8_f32): Likewise.
1311 (vreinterpretq_s8_u8): Likewise.
1312 (vreinterpretq_s8_u16): Likewise.
1313 (vreinterpretq_s8_u32): Likewise.
1314 (vreinterpretq_s8_u64): Likewise.
1315 (vreinterpretq_s8_p8): Likewise.
1316 (vreinterpretq_s8_p16): Likewise.
1317 (vreinterpret_s16_s8): Likewise.
1318 (vreinterpret_s16_s32): Likewise.
1319 (vreinterpret_s16_s64): Likewise.
1320 (vreinterpret_s16_f32): Likewise.
1321 (vreinterpret_s16_u8): Likewise.
1322 (vreinterpret_s16_u16): Likewise.
1323 (vreinterpret_s16_u32): Likewise.
1324 (vreinterpret_s16_u64): Likewise.
1325 (vreinterpret_s16_p8): Likewise.
1326 (vreinterpret_s16_p16): Likewise.
1327 (vreinterpretq_s16_s8): Likewise.
1328 (vreinterpretq_s16_s32): Likewise.
1329 (vreinterpretq_s16_s64): Likewise.
1330 (vreinterpretq_s16_f32): Likewise.
1331 (vreinterpretq_s16_u8): Likewise.
1332 (vreinterpretq_s16_u16): Likewise.
1333 (vreinterpretq_s16_u32): Likewise.
1334 (vreinterpretq_s16_u64): Likewise.
1335 (vreinterpretq_s16_p8): Likewise.
1336 (vreinterpretq_s16_p16): Likewise.
1337 (vreinterpret_s32_s8): Likewise.
1338 (vreinterpret_s32_s16): Likewise.
1339 (vreinterpret_s32_s64): Likewise.
1340 (vreinterpret_s32_f32): Likewise.
1341 (vreinterpret_s32_u8): Likewise.
1342 (vreinterpret_s32_u16): Likewise.
1343 (vreinterpret_s32_u32): Likewise.
1344 (vreinterpret_s32_u64): Likewise.
1345 (vreinterpret_s32_p8): Likewise.
1346 (vreinterpret_s32_p16): Likewise.
1347 (vreinterpretq_s32_s8): Likewise.
1348 (vreinterpretq_s32_s16): Likewise.
1349 (vreinterpretq_s32_s64): Likewise.
1350 (vreinterpretq_s32_f32): Likewise.
1351 (vreinterpretq_s32_u8): Likewise.
1352 (vreinterpretq_s32_u16): Likewise.
1353 (vreinterpretq_s32_u32): Likewise.
1354 (vreinterpretq_s32_u64): Likewise.
1355 (vreinterpretq_s32_p8): Likewise.
1356 (vreinterpretq_s32_p16): Likewise.
1357 (vreinterpret_u8_s8): Likewise.
1358 (vreinterpret_u8_s16): Likewise.
1359 (vreinterpret_u8_s32): Likewise.
1360 (vreinterpret_u8_s64): Likewise.
1361 (vreinterpret_u8_f32): Likewise.
1362 (vreinterpret_u8_u16): Likewise.
1363 (vreinterpret_u8_u32): Likewise.
1364 (vreinterpret_u8_u64): Likewise.
1365 (vreinterpret_u8_p8): Likewise.
1366 (vreinterpret_u8_p16): Likewise.
1367 (vreinterpretq_u8_s8): Likewise.
1368 (vreinterpretq_u8_s16): Likewise.
1369 (vreinterpretq_u8_s32): Likewise.
1370 (vreinterpretq_u8_s64): Likewise.
1371 (vreinterpretq_u8_f32): Likewise.
1372 (vreinterpretq_u8_u16): Likewise.
1373 (vreinterpretq_u8_u32): Likewise.
1374 (vreinterpretq_u8_u64): Likewise.
1375 (vreinterpretq_u8_p8): Likewise.
1376 (vreinterpretq_u8_p16): Likewise.
1377 (vreinterpret_u16_s8): Likewise.
1378 (vreinterpret_u16_s16): Likewise.
1379 (vreinterpret_u16_s32): Likewise.
1380 (vreinterpret_u16_s64): Likewise.
1381 (vreinterpret_u16_f32): Likewise.
1382 (vreinterpret_u16_u8): Likewise.
1383 (vreinterpret_u16_u32): Likewise.
1384 (vreinterpret_u16_u64): Likewise.
1385 (vreinterpret_u16_p8): Likewise.
1386 (vreinterpret_u16_p16): Likewise.
1387 (vreinterpretq_u16_s8): Likewise.
1388 (vreinterpretq_u16_s16): Likewise.
1389 (vreinterpretq_u16_s32): Likewise.
1390 (vreinterpretq_u16_s64): Likewise.
1391 (vreinterpretq_u16_f32): Likewise.
1392 (vreinterpretq_u16_u8): Likewise.
1393 (vreinterpretq_u16_u32): Likewise.
1394 (vreinterpretq_u16_u64): Likewise.
1395 (vreinterpretq_u16_p8): Likewise.
1396 (vreinterpretq_u16_p16): Likewise.
1397 (vreinterpret_u32_s8): Likewise.
1398 (vreinterpret_u32_s16): Likewise.
1399 (vreinterpret_u32_s32): Likewise.
1400 (vreinterpret_u32_s64): Likewise.
1401 (vreinterpret_u32_f32): Likewise.
1402 (vreinterpret_u32_u8): Likewise.
1403 (vreinterpret_u32_u16): Likewise.
1404 (vreinterpret_u32_u64): Likewise.
1405 (vreinterpret_u32_p8): Likewise.
1406 (vreinterpret_u32_p16): Likewise.
1407 (vreinterpretq_u32_s8): Likewise.
1408 (vreinterpretq_u32_s16): Likewise.
1409 (vreinterpretq_u32_s32): Likewise.
1410 (vreinterpretq_u32_s64): Likewise.
1411 (vreinterpretq_u32_f32): Likewise.
1412 (vreinterpretq_u32_u8): Likewise.
1413 (vreinterpretq_u32_u16): Likewise.
1414 (vreinterpretq_u32_u64): Likewise.
1415 (vreinterpretq_u32_p8): Likewise.
1416 (vreinterpretq_u32_p16): Likewise.
1418 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1420 * gcc/config/aarch64/aarch64-simd.md (aarch64_s<optab><mode>):
1422 * config/aarch64/aarch64-simd-builtins.def (sqneg): Iterator
1425 * config/aarch64/arm_neon.h (vqneg_s64): New intrinsic.
1426 (vqnegd_s64): Likewise.
1427 (vqabs_s64): Likewise.
1428 (vqabsd_s64): Likewise.
1430 2014-04-22 Richard Henderson <rth@redhat.com>
1432 * config/sparc/sparc.c (sparc_init_modes): Hoist GET_MODE_SIZE
1433 computation to the top of the loop.
1435 2014-04-22 Renlin <renlin.li@arm.com>
1436 Jiong Wang <jiong.wang@arm.com>
1438 * config/aarch64/aarch64.h (aarch64_frame): Delete "fp_lr_offset".
1439 * config/aarch64/aarch64.c (aarch64_layout_frame)
1440 (aarch64_initial_elimination_offset): Likewise.
1442 2014-04-22 Marcus Shawcroft <marcus.shawcroft@arm.com>
1444 * config/aarch64/aarch64.c (aarch64_initial_elimination_offset):
1447 2014-04-22 Richard Sandiford <rdsandiford@googlemail.com>
1449 * machmode.h (bitwise_mode_for_mode): Declare.
1450 * stor-layout.h (bitwise_type_for_mode): Likewise.
1451 * stor-layout.c (bitwise_mode_for_mode): New function.
1452 (bitwise_type_for_mode): Likewise.
1453 * builtins.c (fold_builtin_memory_op): Use it instead of
1454 int_mode_for_mode and build_nonstandard_integer_type.
1456 2014-04-22 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1458 * config.gcc (enable_obsolete): Remove *-*-solaris2.9*.
1459 (*-*-solaris2.[0-9] | *-*-solaris2.[0-9].*): Mark unsupported.
1460 (*-*-solaris2*): Simplify.
1461 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Likewise.
1462 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Remove
1463 *-*-solaris2.9* handling.
1465 * configure.ac (gcc_cv_as_hidden): Remove test for Solaris 9/x86
1467 (gcc_cv_ld_hidden): Remove *-*-solaris2.9* handling.
1468 (ld_tls_support): Remove i?86-*-solaris2.9, sparc*-*-solaris2.9
1470 (gcc_cv_as_gstabs_flag): Remove workaround for Solaris 9/x86 as bug.
1471 * configure: Regenerate.
1473 * config/i386/sol2-9.h: Remove.
1475 * doc/install.texi (Specific, i?86-*-solaris2.9): Remove.
1476 (Specific, *-*-solaris2*): Mention Solaris 9 support removal.
1477 Remove Solaris 9 references.
1479 2014-04-22 Vidya Praveen <vidyapraveen@arm.com>
1481 * aarch64.md (float<GPI:mode><GPF:mode>2): Remove.
1482 (floatuns<GPI:mode><GPF:mode>2): Remove.
1483 (<optab><fcvt_target><GPF:mode>2): New pattern for equal width float
1484 and floatuns conversions.
1485 (<optab><fcvt_iesize><GPF:mode>2): New pattern for inequal width float
1486 and floatuns conversions.
1487 * iterators.md (fcvt_target, FCVT_TARGET): Support SF and DF modes.
1488 (w1,w2): New mode attributes for inequal width conversions.
1490 2014-04-22 Renlin Li <Renlin.Li@arm.com>
1492 * config/aarch64/aarch64.c (aarch64_print_operand_address): Adjust
1493 the output asm format.
1495 2014-04-22 James Greenhalgh <james.greenhalgh@arm.com>
1497 * config/aarch64/aarch64-simd.md
1498 (aarch64_cm<optab>di): Always split.
1499 (*aarch64_cm<optab>di): New.
1500 (aarch64_cmtstdi): Always split.
1501 (*aarch64_cmtstdi): New.
1503 2014-04-22 Jakub Jelinek <jakub@redhat.com>
1505 PR tree-optimization/60823
1506 * omp-low.c (ipa_simd_modify_function_body): Go through
1507 all SSA_NAMEs and for those refering to vector arguments
1508 which are going to be replaced adjust SSA_NAME_VAR and,
1509 if it is a default definition, change it into a non-default
1510 definition assigned at the beginning of function from new_decl.
1511 (ipa_simd_modify_stmt_ops): Rewritten.
1512 * tree-dfa.c (set_ssa_default_def): When removing default def,
1513 check for NULL loc instead of NULL *loc.
1515 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1517 * config/arm/arm.c (arm_hard_regno_mode_ok): Loosen
1518 restrictions on core registers for DImode values in Thumb2.
1520 2014-04-22 Ian Bolton <ian.bolton@arm.com>
1522 * config/arm/arm.md (*anddi_notdi_zesidi): New pattern.
1523 * config/arm/thumb2.md (*iordi_notdi_zesidi): New pattern.
1525 2014-04-22 Ian Bolton <ian.bolton@arm.com>
1527 * config/arm/thumb2.md (*iordi_notdi_di): New pattern.
1528 (*iordi_notzesidi_di): Likewise.
1529 (*iordi_notsesidi_di): Likewise.
1531 2014-04-22 Ian Bolton <ian.bolton@arm.com>
1533 * config/arm/arm-protos.h (tune_params): New struct members.
1534 * config/arm/arm.c: Initialise tune_params per processor.
1535 (thumb2_reorg): Suppress conversion from t32 to t16 when optimizing
1536 for speed, based on new tune_params.
1538 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1540 * config/aarch64/aarch64-builtins.c (BUILTIN_VDQF_DF): Macro
1542 * config/aarch64/aarch64-simd-builtins.def (frintn): Use added
1544 * config/aarch64/aarch64-simd.md (<frint_pattern>): Comment
1546 * config/aarch64/aarch64.md (<frint_pattern>): Likewise.
1547 * config/aarch64/arm_neon.h (vrnd_f64): Added.
1548 (vrnda_f64): Likewise.
1549 (vrndi_f64): Likewise.
1550 (vrndm_f64): Likewise.
1551 (vrndn_f64): Likewise.
1552 (vrndp_f64): Likewise.
1553 (vrndx_f64): Likewise.
1555 2014-04-22 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1557 * config/arm/arm.c (arm_print_operand, thumb_exit): Make sure
1558 GET_MODE_SIZE argument is enum machine_mode.
1560 2014-04-22 Jakub Jelinek <jakub@redhat.com>
1563 * config/sparc/sparc.c (sparc_init_modes): Pass enum machine_mode
1564 value instead of int to GET_MODE_CLASS and GET_MODE_SIZE macros.
1566 2014-04-22 Lin Zuojian <manjian2006@gmail.com>
1569 * asan.c (asan_emit_stack_protection): Force the base to align to
1570 appropriate bits if STRICT_ALIGNMENT. Set shadow_mem align to
1571 appropriate bits if STRICT_ALIGNMENT.
1572 * cfgexpand.c (expand_stack_vars): Set base_align appropriately
1574 (expand_used_vars): Leave a space in the stack frame for alignment
1575 if STRICT_ALIGNMENT.
1577 2014-04-21 David Malcolm <dmalcolm@redhat.com>
1579 * gimple.h (gimple_assign_single_p): Accept a const_gimple rather
1581 (gimple_store_p): Likewise.
1582 (gimple_assign_load_p): Likewise.
1583 (gimple_assign_cast_p): Likewise.
1584 (gimple_clobber_p): Likewise.
1586 * doc/gimple.texi (gimple_assign_cast_p): Accept a const_gimple
1587 rather than a gimple.
1588 (gimple_assign_cast_p): Likewise.
1590 2014-04-21 Michael Meissner <meissner@linux.vnet.ibm.com>
1593 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64 case):
1594 If mode is DDmode and TARGET_E500_DOUBLE allow move.
1596 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print some
1597 more debug information for E500 if -mdebug=reg.
1599 2014-04-21 Uros Bizjak <ubizjak@gmail.com>
1602 * config/i386/i386.c (ix86_expand_builtin)
1603 <case IX86_BUILTIN_RDRAND{16,32,64}_STEP>: Use temporary
1604 register for target RTX.
1605 <case IX86_BUILTIN_RDSEED{16,32,64}_STEP>: Ditto.
1607 2014-04-18 Cong Hou <congh@google.com>
1609 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): Enhance
1610 the widen-mult pattern by handling two operands with different sizes,
1611 and operands whose size is smaller than half of the result type.
1613 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
1615 * ipa-inline.h (INLINE_HINT_known_hot): New hint.
1616 * ipa-inline-analysis.c (dump_inline_hints): Dump it.
1617 (do_estimate_edge_time): Compute it.
1618 * ipa-inline.c (want_inline_small_function_p): Bypass
1619 INLINE_INSNS_AUTO/SINGLE limits for calls that are known to be hot.
1621 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
1623 * ipa-inline.c (spec_rem): New static variable.
1624 (dump_overall_stats): New function.
1625 (dump_inline_stats): New function.
1627 2014-04-18 Richard Henderson <rth@redhat.com>
1629 * config/aarch64/aarch64.c (aarch64_register_move_cost): Pass a mode
1630 to GET_MODE_SIZE, not a reg_class_t.
1632 2014-04-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1634 * config/rs6000/vsx.md (vsx_xxmrghw_<mode>): Adjust for little-endian.
1635 (vsx_xxmrglw_<mode>): Likewise.
1637 2014-04-17 Michael Meissner <meissner@linux.vnet.ibm.com>
1640 * config/rs6000/rs6000.c (rs6000_setup_reg_addr_masks): Make sure
1641 GET_MODE_SIZE gets passed an enum machine_mode type and not integer.
1642 (rs6000_init_hard_regno_mode_ok): Likewise.
1644 2014-04-17 Jan Hubicka <hubicka@ucw.cz>
1646 * ipa-inline.c (inline_small_functions): Account only non-cold
1648 * doc/invoke.texi (inline-unit-growth): Update documentation.
1650 2014-04-17 Pat Haugen <pthaugen@us.ibm.com>
1652 * config/rs6000/rs6000.md (addti3, subti3): New.
1654 2014-04-17 H.J. Lu <hongjiu.lu@intel.com>
1657 * config/i386/i386.c (ix86_expand_clear): Remove outdated
1658 comment. Check optimize_insn_for_size_p instead of
1659 optimize_insn_for_speed_p.
1661 2014-04-17 Martin Jambor <mjambor@suse.cz>
1663 * gimple-iterator.c (gsi_start_edge): New function.
1664 * gimple-iterator.h (gsi_start_edge): Declare.
1665 * tree-sra.c (single_non_eh_succ): New function.
1666 (disqualify_ops_if_throwing_stmt): Renamed to
1667 disqualify_if_bad_bb_terminating_stmt. Allow throwing statements
1668 having one non-EH successor BB.
1669 (sra_modify_expr): If stmt ends bb, use single non-EH successor to
1670 generate loads into replacements.
1671 (sra_modify_assign): Likewise and and also use the simple path for
1673 (sra_modify_function_body): Commit statements on edges.
1675 2014-04-17 Richard Biener <rguenther@suse.de>
1678 * tree-ssa-propagate.c (valid_gimple_rhs_p): Allow vector
1679 comparison results and add clarifying comment.
1681 2014-04-17 Jakub Jelinek <jakub@redhat.com>
1683 * genmodes.c (struct mode_data): Add need_bytesize_adj field.
1684 (blank_mode): Initialize it.
1685 (emit_mode_size_inline, emit_mode_nunits_inline,
1686 emit_mode_inner_inline): New functions.
1687 (emit_insn_modes_h): Call them and surround their output with
1688 #if GCC_VERSION >= 4001 ... #endif.
1689 * machmode.h (GET_MODE_SIZE, GET_MODE_NUNITS, GET_MODE_INNER):
1690 For GCC_VERSION >= 4001 use mode_*_inline routines instead of
1691 mode_* arrays if the argument is __builtin_constant_p.
1692 * lower-subreg.c (dump_choices): Make sure GET_MODE_SIZE argument
1693 is enum machine_mode.
1695 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1697 * passes.c (opt_pass::execute): Adjust.
1698 (pass_manager::execute_pass_mode_switching): Likewise.
1699 (early_local_passes::execute): Likewise.
1700 (execute_one_pass): Pass cfun to the pass's execute method.
1701 * tree-pass.h (opt_pass::execute): Add function * argument.
1702 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
1703 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
1704 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
1705 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
1706 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
1707 config/sparc/sparc.c, cprop.c, dce.c, df-core.c, dse.c, dwarf2cfi.c,
1708 except.c, final.c, function.c, fwprop.c, gcse.c, gimple-low.c,
1709 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
1710 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
1711 ipa-inline-analysis.c, ipa-inline.c, ipa-profile.c, ipa-pure-const.c,
1712 ipa-reference.c, ipa-split.c, ipa.c, ira.c, jump.c, loop-init.c,
1713 lower-subreg.c, mode-switching.c, omp-low.c, postreload-gcse.c,
1714 postreload.c, predict.c, recog.c, ree.c, reg-stack.c, regcprop.c,
1715 reginfo.c, regrename.c, reorg.c, sched-rgn.c, stack-ptr-mod.c,
1716 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
1717 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
1718 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
1719 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-ssa-ccp.c,
1720 tree-ssa-copy.c, tree-ssa-copyrename.c, tree-ssa-dce.c,
1721 tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
1722 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
1723 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
1724 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
1725 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
1726 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
1727 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
1728 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
1729 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
1730 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
1733 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1735 * passes.c (opt_pass::gate): Take function * argument.
1736 (gate_all_early_local_passes): Merge into
1737 (early_local_passes::gate): this.
1738 (gate_all_early_optimizations): Merge into
1739 (all_early_optimizations::gate): this.
1740 (gate_all_optimizations): Mege into
1741 (all_optimizations::gate): this.
1742 (gate_all_optimizations_g): Merge into
1743 (all_optimizations_g::gate): this.
1744 (gate_rest_of_compilation): Mege into
1745 (rest_of_compilation::gate): this.
1746 (gate_postreload): Merge into
1747 (postreload::gate): this.
1748 (dump_one_pass): Pass cfun to the pass's gate method.
1749 (execute_ipa_summary_passes): Likewise.
1750 (execute_one_pass): Likewise.
1751 (ipa_write_summaries_2): Likewise.
1752 (ipa_write_optimization_summaries_1): Likewise.
1753 (ipa_read_summaries_1): Likewise.
1754 (ipa_read_optimization_summaries_1): Likewise.
1755 (execute_ipa_stmt_fixups): Likewise.
1756 * tree-pass.h (opt_pass::gate): Add function * argument.
1757 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c,
1758 combine-stack-adj.c, combine.c, compare-elim.c,
1759 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
1760 config/rl78/rl78.c, config/sh/sh_optimize_sett_clrt.cc,
1761 config/sh/sh_treg_combine.cc, config/sparc/sparc.c, cprop.c, cse.c,
1762 dce.c, df-core.c, dse.c, dwarf2cfi.c, except.c, fwprop.c, gcse.c,
1763 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
1764 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
1765 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
1766 loop-init.c, lower-subreg.c, mode-switching.c, modulo-sched.c,
1767 omp-low.c, postreload-gcse.c, postreload.c, predict.c, recog.c, ree.c,
1768 reg-stack.c, regcprop.c, regrename.c, reorg.c, sched-rgn.c,
1769 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
1770 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
1771 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c,
1772 tree-nrv.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
1773 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
1774 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
1775 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
1776 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
1777 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
1778 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
1779 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
1780 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
1781 tree-ssa.c, tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
1782 tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c, tsan.c, ubsan.c,
1783 var-tracking.c, vtable-verify.c, web.c: Adjust.
1785 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1787 * configure.ac: Check for -Woverloaded-virtual and enable it if found.
1788 * configure: Regenerate.
1790 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1792 * passes.c (dump_one_pass): don't check pass->has_gate.
1793 (execute_ipa_summary_passes): Likewise.
1794 (execute_one_pass): Likewise.
1795 (ipa_write_summaries_2): Likewise.
1796 (ipa_write_optimization_summaries_1): Likewise.
1797 (ipa_read_optimization_summaries_1): Likewise.
1798 (execute_ipa_stmt_fixups): Likewise.
1799 * tree-pass.h (pass_data::has_gate): Remove.
1800 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
1801 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
1802 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
1803 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
1804 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
1805 config/sh/sh_optimize_sett_clrt.cc, config/sh/sh_treg_combine.cc,
1806 config/sparc/sparc.c, cprop.c, cse.c, dce.c, df-core.c, dse.c,
1807 dwarf2cfi.c, except.c, final.c, function.c, fwprop.c, gcse.c,
1808 gimple-low.c, gimple-ssa-isolate-paths.c,
1809 gimple-ssa-strength-reduction.c, graphite.c, ifcvt.c, init-regs.c,
1810 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
1811 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
1812 ira.c, jump.c, loop-init.c, lower-subreg.c, mode-switching.c,
1813 modulo-sched.c, omp-low.c, postreload-gcse.c, postreload.c, predict.c,
1814 recog.c, ree.c, reg-stack.c, regcprop.c, reginfo.c, regrename.c,
1815 reorg.c, sched-rgn.c, stack-ptr-mod.c, store-motion.c, tracer.c,
1816 trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
1817 tree-complex.c, tree-eh.c, tree-emutls.c, tree-if-conv.c,
1818 tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
1819 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
1820 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
1821 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
1822 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
1823 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
1824 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
1825 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
1826 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
1827 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
1828 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
1829 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
1830 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
1833 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1835 * pass_manager.h (pass_manager::register_dump_files_1): Remove
1837 * passes.c (pass_manager::register_dump_files_1): Merge into
1838 (pass_manager::register_dump_files): this, and remove its handling of
1839 properties since the pass always has the properties anyway.
1840 (pass_manager::pass_manager): Adjust.
1842 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1844 * pass_manager.h (pass_manager::register_dump_files_1): Adjust.
1845 * passes.c (pass_manager::register_dump_files_1): Remove dead code
1846 dealing with properties.
1847 (pass_manager::register_dump_files): Adjust.
1849 2014-03-20 Mark Wielaard <mjw@redhat.com>
1851 * dwarf2out.c (add_bound_info): If HOST_WIDE_INT is big enough,
1852 then represent the bound as normal constant value.
1854 2014-04-17 Jakub Jelinek <jakub@redhat.com>
1857 Forward port from 4.8 branch
1858 2013-07-19 Kirill Yukhin <kirill.yukhin@intel.com>
1860 * config/i386/bmiintrin.h (_blsi_u32): New.
1864 (_blsmsk_u32): Ditto.
1865 (_blsmsk_u64): Ditto.
1866 (_tzcnt_u32): Ditto.
1867 (_tzcnt_u64): Ditto.
1869 2014-04-17 Kito Cheng <kito@0xlab.org>
1871 * gcc.c (used_arg): Prevent out of bound access for multilib_options.
1873 2014-04-17 Richard Biener <rguenther@suse.de>
1876 * tree-ssa-propagate.c (valid_gimple_rhs_p): Only allow effective
1877 boolean results for comparisons.
1879 2014-04-17 Richard Biener <rguenther@suse.de>
1881 PR tree-optimization/60836
1882 * tree-vect-loop.c (vect_create_epilog_for_reduction): Force
1883 initial PHI args to be gimple values.
1885 2014-04-17 Richard Biener <rguenther@suse.de>
1887 PR tree-optimization/60841
1888 * tree-vect-data-refs.c (vect_analyze_data_refs): Count stmts.
1889 * tree-vect-loop.c (vect_analyze_loop_2): Pass down number
1890 of stmts to SLP build.
1891 * tree-vect-slp.c (vect_slp_analyze_bb_1): Likewise.
1892 (vect_analyze_slp): Likewise.
1893 (vect_analyze_slp_instance): Likewise.
1894 (vect_build_slp_tree): Limit overall SLP tree growth.
1895 * tree-vectorizer.h (vect_analyze_data_refs,
1896 vect_analyze_slp): Adjust prototypes.
1898 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
1900 * config/i386/i386.c (x86_add_stmt_cost): Fix vector cost model for
1903 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
1905 * config/i386/x86-tune.def (TARGET_SLOW_PSHUFB): New tune definition.
1906 * config/i386/i386.h (TARGET_SLOW_PSHUFB): New tune flag.
1907 * config/i386/i386.c (expand_vec_perm_even_odd_1): Avoid byte shuffles
1908 for TARGET_SLOW_PSHUFB
1910 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
1912 * config/i386/i386.c (slm_cost): Adjust vec_to_scalar_cost.
1913 * config/i386/i386.c (intel_cost): Ditto.
1915 2014-04-17 Joey Ye <joey.ye@arm.com>
1917 * opts.c (OPT_fif_conversion, OPT_fif_conversion2): Disable for Og.
1919 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
1921 * opts.c (common_handle_option): Disable -fipa-reference coorectly
1924 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
1926 * ipa-devirt.c (odr_type_d): Add field all_derivations_known.
1927 (type_all_derivations_known_p): New predicate.
1928 (type_all_ctors_visible_p): New predicate.
1929 (type_possibly_instantiated_p): New predicate.
1930 (get_odr_type): Compute all_derivations_known.
1931 (dump_odr_type): Dump the flag.
1932 (maybe_record_type): Cleanup.
1933 (record_target_from_binfo): Add bases_to_consider array;
1934 record bases for types w/o instances and skip CXX destructor.
1935 (possible_polymorphic_call_targets_1): Add bases_to_consider
1936 and consider_construction parameters; check if type may have instance.
1937 (get_polymorphic_call_info): Set maybe_in_construction to true
1938 when we know nothing.
1939 (record_targets_from_bases): Skip CXX destructors; they are
1940 never called for types in construction.
1941 (possible_polymorphic_call_targets): Do not record target when
1942 type may not have instance.
1944 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
1947 * ipa.c (symtab_remove_unreachable_nodes): Mark targets of
1948 external aliases alive, too.
1950 2014-04-16 Andrew Pinski <apinski@cavium.com>
1952 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Change aarch64 ilp32
1955 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
1957 * final.c (compute_alignments): Do not apply loop alignment to a block
1958 falling through to the exit.
1960 2014-04-16 Catherine Moore <clm@codesourcery.com>
1962 * mips.md (*mov<mode>_internal, *movhi_internal, *movqi_internal):
1963 Adjust constraints for microMIPS store patterns.
1965 2014-04-16 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
1967 * config/avr/avr-mcus.def: Correct typo for atxmega256a3bu macro.
1969 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
1971 * tree-ssa-operands.c (create_vop_var): Set DECL_IGNORED_P.
1972 (append_use): Run at -O0.
1973 (append_vdef): Likewise.
1974 * tree-ssa-ter.c (ter_is_replaceable_p): Do not special-case -O0.
1975 * tree-ssa-uninit.c (warn_uninitialized_vars): Remove obsolete comment.
1977 2014-04-16 Jakub Jelinek <jakub@redhat.com>
1979 PR tree-optimization/60844
1980 * tree-ssa-reassoc.c (reassoc_remove_stmt): New function.
1981 (propagate_op_to_single_use, remove_visited_stmt_chain,
1982 linearize_expr, repropagate_negates, reassociate_bb): Use it
1983 instead of gsi_remove.
1985 2014-04-16 Martin Jambor <mjambor@suse.cz>
1987 * cgraphclones.c (cgraph_create_virtual_clone): Duplicate
1988 ipa_transforms_to_apply.
1989 (cgraph_function_versioning): Assert that old_node has empty
1990 ipa_transforms_to_apply.
1991 * trans-mem.c (ipa_tm_create_version): Likewise.
1992 * tree-inline.c (tree_function_versioning): Do not duplicate
1993 ipa_transforms_to_apply.
1995 2014-04-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1998 * configure.ac (set_have_as_tls): Merge i[34567]86-*-* and
2000 Pass necessary as flags on 64-bit Solaris/x86.
2001 Use lowercase relocs for x86_64-*-*.
2002 * configure: Regenerate.
2004 2014-04-15 Jan Hubicka <jh@suse.cz>
2006 * ipa-devirt.c (referenced_from_vtable_p): New predicate.
2007 (maybe_record_node, likely_target_p): Use it.
2009 2014-04-15 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2012 Revert following patch
2014 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
2017 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
2018 software floating point or no floating point registers, do not
2019 allow any type in the FPRs. Eliminate a test for SPE SIMD types
2020 in GPRs that occurs after we tested for GPRs that would never be
2023 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
2024 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
2025 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
2026 specifically allow DDmode, since that does not use the SPE SIMD
2029 2014-03-21 Mark Wielaard <mjw@redhat.com>
2031 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_const_value
2032 as unsigned or int depending on type and value used.
2034 2014-04-15 Richard Biener <rguenther@suse.de>
2036 PR rtl-optimization/56965
2037 * alias.c (ncr_compar, nonoverlapping_component_refs_p): Move ...
2038 * tree-ssa-alias.c (ncr_compar, nonoverlapping_component_refs_p):
2040 * alias.c (true_dependence_1): Do not call
2041 nonoverlapping_component_refs_p.
2042 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Call
2043 nonoverlapping_component_refs_p.
2044 (indirect_refs_may_alias_p): Likewise.
2046 2014-04-15 Teresa Johnson <tejohnson@google.com>
2048 * cfg.c (dump_bb_info): Fix flags check.
2049 * tree-cfg.c (remove_bb): Only dump TDF_BLOCKS when removing.
2051 2014-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2053 PR rtl-optimization/60663
2054 * config/arm/arm.c (arm_new_rtx_costs): Improve ASM_OPERANDS case,
2057 2014-04-15 Richard Biener <rguenther@suse.de>
2059 * lto-streamer.h (LTO_major_version): Bump to 4.
2061 2014-04-15 Richard Biener <rguenther@suse.de>
2063 * common.opt (lto_partition_model): New enum.
2064 (flto-partition=): Merge separate options with a single with argument,
2065 add -flto-partition=one support.
2066 * flag-types.h (enum lto_partition_model): Declare.
2067 * opts.c (finish_options): Remove duplicate -flto-partition=
2069 * lto-wrapper.c (run_gcc): Adjust.
2071 2014-04-15 Richard Biener <rguenther@suse.de>
2073 * alias.c (ncr_compar): New function.
2074 (nonoverlapping_component_refs_p): Re-implement in O (n log n).
2076 2014-04-15 Richard Biener <rguenther@suse.de>
2078 * alias.c (record_component_aliases): Do not walk BINFOs.
2080 2014-04-15 Richard Biener <rguenther@suse.de>
2082 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
2083 Add struct function argument and adjust.
2084 (find_func_aliases_for_call): Likewise.
2085 (find_func_aliases): Likewise.
2086 (find_func_clobbers): Likewise.
2087 (intra_create_variable_infos): Likewise.
2088 (compute_points_to_sets): Likewise.
2089 (ipa_pta_execute): Adjust. Do not push/pop cfun.
2091 2014-04-15 Richard Biener <rguenther@suse.de>
2093 * tree.c (iterative_hash_expr): Use enum tree_code_class
2094 to store TREE_CODE_CLASS.
2095 (tree_block): Likewise.
2096 (tree_set_block): Likewise.
2097 * tree.h (fold_build_pointer_plus_loc): Use
2098 convert_to_ptrofftype_loc.
2100 2014-04-15 Jakub Jelinek <jakub@redhat.com>
2103 * Makefile.in (PLUGIN_HEADERS): Add various headers that have been
2106 2014-04-15 Eric Botcazou <ebotcazou@adacore.com>
2108 * cfgloop.h (struct loop): Move force_vectorize down.
2109 * gimplify.c (gimple_boolify) <ANNOTATE_EXPR>: Handle new kinds.
2110 (gimplify_expr) <ANNOTATE_EXPR>: Minor tweak.
2111 * lto-streamer-in.c (input_cfg): Read dont_vectorize field.
2112 * lto-streamer-out.c (output_cfg): Write dont_vectorize field.
2113 * tree-cfg.c (replace_loop_annotate): Revamp and handle new kinds.
2114 * tree-core.h (enum annot_expr_kind): Add new kind values.
2115 * tree-inline.c (copy_loops): Copy dont_vectorize field and reorder.
2116 * tree-pretty-print.c (dump_generic_node) <ANNOTATE_EXPR>: Handle new
2118 * tree.def (ANNOTATE_EXPR): Tweak comment.
2120 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
2122 * ipa-devirt.c (maybe_record_node): Ignore all non-methods (including
2125 2014-04-14 Paolo Carlini <paolo.carlini@oracle.com>
2127 * tree.h (TYPE_IDENTIFIER): Declare.
2128 * tree.c (subrange_type_for_debug_p): Use it.
2129 * godump.c (go_format_type): Likewise.
2130 * dwarf2out.c (is_cxx_auto, modified_type_die,
2131 gen_type_die_with_usage, gen_type_die_with_usage): Likewise.
2132 * dbxout.c (dbxout_type, dbxout_symbol): Likewise.
2134 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
2137 * varpool.c (varpool_remove_node): Do not alter decls when streaming.
2139 2014-04-14 Uros Bizjak <ubizjak@gmail.com>
2141 * config/i386/i386.c (examine_argument): Return bool. Return true if
2142 parameter should be passed in memory.
2143 <case X86_64_COMPLEX_X87_CLASS>: Adjust.
2144 (construct_container): Update calls to examine_argument.
2145 (function_arg_advance_64): Ditto.
2146 (return_in_memory_32): Merge with ix86_return_in_memory.
2147 (return_in_memory_64): Ditto.
2148 (return_in_memory_ms_64): Ditto.
2150 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
2152 * ipa-utils.c (ipa_merge_profiles): Merge profile_id.
2153 * coverage.c (coverage_compute_profile_id): Handle externally visible
2156 2014-04-14 Martin Jambor <mjambor@suse.cz>
2158 * tree-sra.c (ipa_sra_preliminary_function_checks): Skip
2159 DECL_DISREGARD_INLINE_LIMITS functions.
2161 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
2164 * config/i386/i386.md (*fixuns_trunc<mode>_1): Revert the last change.
2166 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
2169 * config/i386/i386.md (*fixuns_trunc<mode>_1): Check
2170 optimize_insn_for_speed_p instead of
2171 optimize_function_for_speed_p.
2173 2014-04-14 Yufeng Zhang <yufeng.zhang@arm.com>
2175 * doc/invoke.texi (free): Document AArch64.
2177 2014-04-14 Richard Biener <rguenther@suse.de>
2179 PR tree-optimization/60042
2180 * tree-ssa-pre.c (inhibit_phi_insertion): Remove.
2181 (insert_into_preds_of_block): Do not prevent PHI insertion
2182 for REFERENCE exprs here ...
2183 (eliminate_dom_walker::before_dom_children): ... but prevent
2184 their use here under similar conditions when applied to the
2185 IL after PRE optimizations.
2187 2014-04-14 Richard Biener <rguenther@suse.de>
2189 * passes.def: Move early points-to after early SRA.
2191 2014-04-14 Richard Biener <rguenther@suse.de>
2193 * tree-ssa-forwprop.c (simplify_gimple_switch): Enhance
2194 check for which sign-changes we allow when forwarding
2195 a converted value into a switch.
2197 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
2199 * stor-layout.c (place_field): Finalize non-constant offset for the
2202 2014-04-14 Richard Biener <rguenther@suse.de>
2204 * tree-switch-conversion.c (lshift_cheap_p): Get speed_p
2206 (expand_switch_using_bit_tests_p): Likewise.
2207 (process_switch): Compute and pass on speed_p based on the
2209 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Use
2210 optimize_bb_for_speed_p.
2212 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
2214 * cfgloop.h (struct loop): Rename force_vect into force_vectorize.
2215 * function.h (struct function): Rename has_force_vect_loops into
2216 has_force_vectorize_loops.
2217 * lto-streamer-in.c (input_cfg): Adjust for renaming.
2218 (input_struct_function_base): Likewise.
2219 * lto-streamer-out.c (output_cfg): Likewise.
2220 (output_struct_function_base): Likewise.
2221 * omp-low.c (expand_omp_simd): Likewise.
2222 * tree-cfg.c (move_sese_region_to_fn): Likewise.
2223 * tree-if-conv.c (ifcvt_can_use_mask_load_store): Likewise.
2224 (version_loop_for_if_conversion): Likewise.
2225 (tree_if_conversion): Likewise.
2226 (main_tree_if_conversion): Likewise.
2227 (gate_tree_if_conversion): Likewise.
2228 * tree-inline.c (copy_loops): Likewise.
2229 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Likewise.
2230 * tree-ssa-loop.c (tree_loop_vectorize): Likewise.
2231 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Likewise.
2232 * tree-vect-loop.c (vect_estimate_min_profitable_iters): Likewise.
2233 * tree-vectorizer.c (vectorize_loops): Likewise.
2234 * tree-vectorizer.h (unlimited_cost_model): Likewise.
2236 2014-04-14 Richard Biener <rguenther@suse.de>
2239 * lto-streamer-out.c (wrap_refs): New function.
2240 (lto_output): Wrap symbol references in global initializes in
2241 type-preserving MEM_REFs.
2243 2014-04-14 Christian Bruel <christian.bruel@st.com>
2245 * config/sh/sh-mem.cc (sh_expand_strlen): Unroll last word.
2247 2014-04-14 Christian Bruel <christian.bruel@st.com>
2249 * config/sh/sh.md (setmemqi): New expand pattern.
2250 * config/sh/sh.h (CLEAR_RATIO): Define.
2251 * config/sh/sh-mem.cc (sh_expand_setmem): Define.
2252 * config/sh/sh-protos.h (sh_expand_setmem): Declare.
2254 2014-04-14 Richard Biener <rguenther@suse.de>
2257 * fold-const.c (negate_expr_p): Don't negate directional rounding
2259 (fold_negate_expr): Likewise.
2261 2014-04-14 Richard Biener <rguenther@suse.de>
2263 PR tree-optimization/59817
2264 PR tree-optimization/60453
2265 * graphite-scop-detection.c (graphite_can_represent_scev): Complete
2266 recursion to catch all CHRECs in the scalar evolution and restrict
2267 the predicate for the remains appropriately.
2269 2014-04-12 Catherine Moore <clm@codesourcery.com>
2271 * config/mips/constraints.md: Add new register constraint "kb".
2272 * config/mips/mips.md (*mov<mode>_internal): Use constraint "kb".
2273 (*movhi_internal): Likewise.
2274 (*movqi_internal): Likewise.
2275 * config/mips/mips.h (M16_STORE_REGS): New register class.
2276 (REG_CLASS_NAMES): Add M16_STORE_REGS.
2277 (REG_CLASS_CONTENTS): Likewise.
2278 * config/mips/mips.c (mips_regno_to_class): Add M16_STORE_REGS.
2280 2014-04-11 Tobias Burnus <burnus@net-b.de>
2283 * doc/invoke.texi (-Wformat-signedness): Document it.
2284 (Wformat=2): Mention that this enables -Wformat-signedness.
2286 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
2288 * common/config/epiphany/epiphany-common.c
2289 (epiphany_option_optimization_table): Enable section anchors by
2290 default at -O1 or higher.
2291 * config/epiphany/epiphany.c (TARGET_MAX_ANCHOR_OFFSET): Define.
2292 (TARGET_MIN_ANCHOR_OFFSET): Likewise.
2293 (epiphany_rtx_costs) <SET>: For binary operators, the set as such
2294 carries no extra cost.
2295 (epiphany_legitimate_address_p): For BLKmode, apply SImode check.
2296 * config/epiphany/epiphany.h (ASM_OUTPUT_DEF): Define.
2297 * config/epiphany/predicates.md (memclob_operand): New predicate.
2298 * config/epiphany/epiphany.md (stack_adjust_add, stack_adjust_str):
2299 Use memclob_operand predicate and X constraint for operand 3.
2301 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
2303 * config/epiphany/epiphany.c (epiphany_rtx_cost): Compare
2304 with CC_N_NE / CC_C_LTU / CC_C_GTU carries no extra cost for
2307 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
2309 PR rtl-optimization/60651
2310 * mode-switching.c (optimize_mode_switching): Make sure to emit
2311 sets of a lower numbered entity before sets of a higher numbered
2312 entity to a mode of the same or lower priority.
2313 When creating a seginfo for a basic block that starts with a code
2314 label, move the insertion point past the code label.
2315 (new_seginfo): Document and enforce requirement that
2316 NOTE_INSN_BASIC_BLOCK only appears for empty blocks.
2317 * doc/tm.texi.in: Document ordering constraint for emitted mode sets.
2318 * doc/tm.texi: Regenerate.
2320 2014-01-11 Joern Rennecke <joern.rennecke@embecosm.com>
2323 * config/arc/arc.c (arc_save_restore): Fix assert typo.
2325 2013-04-11 Jakub Jelinek <jakub@redhat.com>
2327 * BASE-VER: Set to 4.10.0.
2329 2014-04-11 Tobias Burnus <burnus@net-b.de>
2332 * doc/bugreport.texi (Bugs): Remove nodes pointing to the nirvana.
2333 * doc/gcc.texi (Service): Update description in the @menu
2334 * doc/invoke.texi (Option Summary): Remove misplaced and
2337 2014-04-11 Steve Ellcey <sellcey@mips.com>
2338 Jakub Jelinek <jakub@redhat.com>
2341 * expr.c (convert_move): Use emit_store_flag_force instead of
2342 emit_store_flag. Pass lowpart_mode instead of VOIDmode as 5th
2345 2014-04-11 Richard Biener <rguenther@suse.de>
2348 * varasm.c (assemble_alias): Avoid endless error reporting
2349 recursion by setting TREE_ASM_WRITTEN.
2351 2014-04-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2353 * config/s390/s390.md: Add a splitter for NOT rtx.
2355 2014-04-11 Jakub Jelinek <jakub@redhat.com>
2357 PR rtl-optimization/60663
2358 * cse.c (cse_insn): Set src_volatile on ASM_OPERANDS in PARALLEL.
2360 2014-04-10 Jan Hubicka <hubicka@ucw.cz>
2361 Jakub Jelinek <jakub@redhat.com>
2364 * ipa.c (function_and_variable_visibility): Copy forced_by_abi
2365 flag from decl_node to node.
2367 2014-04-10 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2370 * config/arm/arm.c (TARGET_CONST_NOT_OK_FOR_DEBUG_P): Define
2371 (arm_const_not_ok_for_debug_p): Reject MINUS with SYM_REF's
2372 ameliorating the cases where it can be.
2374 2014-04-09 David Edelsohn <dje.gcc@gmail.com>
2377 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
2379 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
2380 (loadsync_<mode>): Change mode.
2381 (load_quadpti, store_quadpti): New.
2382 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
2383 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
2384 * config/rs6000/predicates.md (quad_memory_operand): !TARGET_SYNC_TI.
2386 2014-04-09 Cong Hou <congh@google.com>
2389 * doc/sourcebuild.texi (vect_widen_mult_si_to_di_pattern): Add
2392 2014-04-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2394 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Use vnand
2395 instead of vnor to exploit possible fusion opportunity in the
2397 (altivec_expand_vec_perm_const_le): Likewise.
2399 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
2401 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
2402 (loadsync_<mode>): Change mode.
2403 (load_quadpti, store_quadpti): New.
2404 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
2405 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
2407 2014-04-08 Richard Sandiford <rdsandiford@googlemail.com>
2410 * config/rs6000/vsx.md (vsx_xscvdpspn_scalar): Change input to DImode.
2411 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Update accordingly.
2412 Use gen_rtx_REG rather than simplify_gen_subreg for op0_di.
2414 2014-04-08 Richard Biener <rguenther@suse.de>
2417 * tree-pretty-print.c (pp_double_int): For HWI32 hosts with
2418 a 64bit widest int print double-int similar to on HWI64 hosts.
2420 2014-04-08 Richard Biener <rguenther@suse.de>
2422 PR tree-optimization/60785
2423 * graphite-sese-to-poly.c (rewrite_phi_out_of_ssa): Treat
2424 default defs properly.
2426 2014-04-08 Nathan Sidwell <nathan@codesourcery.com>
2428 * doc/invoke (Wnon-virtual-dtor): Update to match implementation.
2429 (Weffc++): Likewise.
2431 2014-04-07 Jan Hubicka <hubcika@ucw.cz>
2433 * ipa-devirt.c (maybe_record_node): When node is not recorded,
2434 set completep to false rather than true.
2436 2014-04-07 Douglas B Rupp <rupp@adacore.com>
2439 * config/arm/arm.h (ASM_PREFERRED_EH_DATA_FORMAT): Expose from
2440 ARM_TARGET2_DWARF_FORMAT.
2442 2014-04-07 Charles Baylis <charles.baylis@linaro.org>
2445 * config/arm/arm.h (ASM_OUTPUT_CASE_END): Remove.
2446 (LABEL_ALIGN_AFTER_BARRIER): Align barriers which occur after
2449 2014-04-07 Richard Biener <rguenther@suse.de>
2451 PR tree-optimization/60766
2452 * tree-ssa-loop-ivopts.c (cand_value_at): Compute in an unsigned type.
2453 (may_eliminate_iv): Convert cand_value_at result to desired type.
2455 2014-04-07 Jason Merrill <jason@redhat.com>
2458 * common.opt (-fno-gnu-unique): Add.
2459 * config/elfos.h (USE_GNU_UNIQUE_OBJECT): Check it.
2461 2014-04-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2463 * haifa-sched.c: Fix outdated function reference and minor
2464 grammar errors in introductory comment.
2466 2014-04-07 Richard Biener <rguenther@suse.de>
2469 * tree-ssa-operands.c (maybe_add_call_vops): Also add VDEFs
2471 * tree-cfgcleanup.c (fixup_noreturn_call): Do not remove VDEFs.
2473 2014-04-06 John David Anglin <danglin@gcc.gnu.org>
2476 * config/pa/pa.c (pa_output_function_epilogue): Skip address and code
2477 size accounting for thunks.
2478 (pa_asm_output_mi_thunk): Use final_start_function() and
2479 final_end_function() to output function start and end directives.
2481 2014-04-05 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
2483 * config/avr/avr-arch.h (avr_mcu_t): Add dev_attribute field to have
2484 device specific ISA/ feature information. Remove short_sp and
2485 errata_skip ds. Add avr_device_specific_features enum to have device
2487 * config/avr/avr-c.c (avr_cpu_cpp_builtins): use dev_attribute to check
2488 errata_skip. Add __AVR_ISA_RMW__ builtin macro if RMW ISA available.
2489 * config/avr/avr-devices.c (avr_mcu_types): Update AVR_MCU macro for
2490 updated device specific info.
2491 * config/avr/avr-mcus.def: Merge device specific details to
2492 dev_attribute field.
2493 * config/avr/avr.c (avr_2word_insn_p): use dev_attribute field to check
2495 * config/avr/avr.h (AVR_HAVE_8BIT_SP): same for short sp info.
2496 * config/avr/driver-avr.c (avr_device_to_as): Pass -mrmw option to
2497 assembler if RMW isa supported by current device.
2498 * config/avr/genmultilib.awk: Update as device info structure changed.
2499 * doc/invoke.texi: Add info for __AVR_ISA_RMW__ builtin macro
2501 2014-04-04 Cong Hou <congh@google.com>
2503 PR tree-optimization/60656
2504 * tree-vect-stmts.c (supportable_widening_operation):
2505 Fix a bug that elements in a vector with vect_used_by_reduction
2506 property are incorrectly reordered when the operation on it is not
2507 consistant with the one in reduction operation.
2509 2014-04-04 John David Anglin <danglin@gcc.gnu.org>
2511 PR rtl-optimization/60155
2512 * gcse.c (record_set_data): New function.
2513 (single_set_gcse): New function.
2514 (gcse_emit_move_after): Use single_set_gcse instead of single_set.
2515 (hoist_code): Likewise.
2516 (get_pressure_class_and_nregs): Likewise.
2518 2014-04-04 Eric Botcazou <ebotcazou@adacore.com>
2520 * explow.c (probe_stack_range): Emit a final optimization blockage.
2522 2014-04-04 Anthony Green <green@moxielogic.com>
2524 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2): Fix
2527 2014-04-04 Jan Hubicka <hubicka@ucw.cz>
2530 * lto-cgraph.c (input_overwrite_node): Check that partitioning
2531 flags are set only during streaming.
2532 * ipa.c (process_references, walk_polymorphic_call_targets,
2533 symtab_remove_unreachable_nodes): Drop bodies of always inline
2534 after early inlining.
2535 (symtab_remove_unreachable_nodes): Remove always_inline attribute.
2537 2014-04-04 Jakub Jelinek <jakub@redhat.com>
2538 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2541 * dwarf2out.c (const_ok_for_output_1): Reject expressions
2544 2014-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2547 * config/arm/cortex-a53.md (cortex_a53_fdivs): Reduce reservation
2549 (cortex_a53_fdivd): Likewise.
2551 2014-04-04 Martin Jambor <mjambor@suse.cz>
2554 * cgraph.h (cgraph_clone_node): New parameter added to declaration.
2556 * cgraph.c (clone_of_p): Also return true if thunks match.
2557 (verify_edge_corresponds_to_fndecl): Removed extraneous call to
2558 cgraph_function_or_thunk_node and an obsolete comment.
2559 * cgraphclones.c (build_function_type_skip_args): Moved upwards in the
2561 (build_function_decl_skip_args): Likewise.
2562 (set_new_clone_decl_and_node_flags): New function.
2563 (duplicate_thunk_for_node): Likewise.
2564 (redirect_edge_duplicating_thunks): Likewise.
2565 (cgraph_clone_node): New parameter args_to_skip, pass it to
2566 redirect_edge_duplicating_thunks which is called instead of
2567 cgraph_redirect_edge_callee.
2568 (cgraph_create_virtual_clone): Pass args_to_skip to cgraph_clone_node,
2569 moved setting of a lot of flags to set_new_clone_decl_and_node_flags.
2571 2014-04-04 Jeff Law <law@redhat.com>
2574 * config/arm/predicates.md (const_int_I_operand): New predicate.
2575 (const_int_M_operand): Similarly.
2576 * config/arm/arm.md (insv_zero): Use const_int_M_operand instead of
2578 (insv_t2, extv_reg, extzv_t2): Likewise.
2579 (load_multiple_with_writeback): Similarly for const_int_I_operand.
2580 (pop_multiple_with_writeback_and_return): Likewise.
2581 (vfp_pop_multiple_with_writeback): Likewise
2583 2014-04-04 Richard Biener <rguenther@suse.de>
2586 * tree-ssanames.c (make_ssa_name_fn): Fix assert.
2587 * gimple.c (gimple_set_bb): Avoid ICEing for NULL cfun for
2589 * gimplify.h (gimple_add_tmp_var_fn): Declare.
2590 * gimplify.c (gimple_add_tmp_var_fn): New function.
2591 * gimple-expr.h (create_tmp_reg_fn): Declare.
2592 * gimple-expr.c (create_tmp_reg_fn): New function.
2593 * gimple-low.c (record_vars_into): Don't change cfun.
2594 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix
2595 code generation without cfun.
2597 2014-04-04 Thomas Schwinge <thomas@codesourcery.com>
2600 * Makefile.in (install-driver): Fix shell scripting.
2602 2014-04-03 Cong Hou <congh@google.com>
2604 PR tree-optimization/60505
2605 * tree-vectorizer.h (struct _stmt_vec_info): Add th field as the
2606 threshold of number of iterations below which no vectorization
2608 * tree-vect-loop.c (new_loop_vec_info):
2609 Initialize LOOP_VINFO_COST_MODEL_THRESHOLD.
2610 * tree-vect-loop.c (vect_analyze_loop_operations):
2611 Set LOOP_VINFO_COST_MODEL_THRESHOLD.
2612 * tree-vect-loop.c (vect_transform_loop):
2613 Use LOOP_VINFO_COST_MODEL_THRESHOLD.
2614 * tree-vect-loop.c (vect_analyze_loop_2): Check the maximum number
2615 of iterations of the loop and see if we should build the epilogue.
2617 2014-04-03 Richard Biener <rguenther@suse.de>
2619 * tree-streamer.h (struct streamer_tree_cache_d): Add next_idx member.
2620 (streamer_tree_cache_create): Adjust.
2621 * tree-streamer.c (streamer_tree_cache_add_to_node_array): Adjust
2622 to allow optional nodes array.
2623 (streamer_tree_cache_insert_1): Use next_idx to assign idx.
2624 (streamer_tree_cache_append): Likewise.
2625 (streamer_tree_cache_create): Create nodes array optionally
2626 as specified by parameter.
2627 * lto-streamer-out.c (create_output_block): Avoid maintaining
2628 the node array in the writer cache.
2629 (DFS_write_tree): Remove assertion.
2630 (produce_asm_for_decls): Free the out decl state hash table early.
2631 * lto-streamer-in.c (lto_data_in_create): Adjust for
2632 streamer_tree_cache_create prototype change.
2634 2014-04-03 Richard Biener <rguenther@suse.de>
2636 * tree-streamer-out.c (streamer_write_chain): Do not temporarily
2637 set TREE_CHAIN to NULL_TREE.
2639 2014-04-03 Richard Biener <rguenther@suse.de>
2641 PR tree-optimization/60740
2642 * graphite-scop-detection.c (stmt_simple_for_scop_p): Iterate
2643 over all GIMPLE_COND operands.
2645 2014-04-03 Nathan Sidwell <nathan@codesourcery.com>
2647 * doc/invoke.texi (Wnon-virtual-dtor): Adjust documentation.
2648 (Weffc++): Remove Scott's numbering, merge lists and reference
2651 2014-04-03 Nick Clifton <nickc@redhat.com>
2653 * config/rl78/rl78-expand.md (movqi): Handle (SUBREG (SYMBOL_REF))
2656 2014-04-03 Martin Jambor <mjambor@suse.cz>
2658 * ipa-cp.c (ipcp_verify_propagated_values): Also dump symtab and
2659 mention gcc_unreachable before failing.
2660 * ipa.c (symtab_remove_unreachable_nodes): Also print order of
2663 2014-04-02 Jan Hubicka <hubicka@ucw.cz>
2666 * ipa-devirt.c (get_polymorphic_call_info): Do not ICE on type
2667 inconsistent code and instead mark the context inconsistent.
2668 (possible_polymorphic_call_targets): For inconsistent contexts
2669 return empty complete list.
2671 2014-04-02 Anthony Green <green@moxielogic.com>
2673 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2)
2674 (extendqisi2, extendhisi2): Define.
2675 * config/moxie/moxie.h (DEFAULT_SIGNED_CHAR): Change to 0.
2676 (WCHAR_TYPE): Change to unsigned int.
2678 2014-04-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2680 PR tree-optimization/60733
2681 * gimple-ssa-strength-reduction.c (ncd_with_phi): Change required
2682 insertion point for PHI candidates to be the end of the feeding
2683 block for the PHI argument.
2685 2014-04-02 Vladimir Makarov <vmakarov@redhat.com>
2687 PR rtl-optimization/60650
2688 * lra-constraints.c (process_alt_operands): Decrease reject for
2689 earlyclobber matching.
2691 2014-04-02 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2693 * config/s390/s390.c (s390_expand_insv): Use GET_MODE_BITSIZE.
2695 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
2697 * config/spu/spu.c (pad_bb): Do not crash when the last
2698 insn is CODE_FOR_blockage.
2700 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
2702 * config/spu/spu.md ("insv"): Fail if bitoffset+bitsize
2703 lies outside the target mode.
2705 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
2708 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
2709 software floating point or no floating point registers, do not
2710 allow any type in the FPRs. Eliminate a test for SPE SIMD types
2711 in GPRs that occurs after we tested for GPRs that would never be
2714 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
2715 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
2716 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
2717 specifically allow DDmode, since that does not use the SPE SIMD
2720 2014-04-02 Richard Biener <rguenther@suse.de>
2723 * optabs.c (expand_abs_nojump): Honor flag_trapv only for
2724 MODE_INTs. Properly use negv_optab.
2725 (expand_abs): Likewise.
2727 2014-04-02 Richard Biener <rguenther@suse.de>
2730 * Makefile.in (install-driver): Guard extra installs with special
2733 2014-04-01 Michael Meissner <meissner@linux.vnet.ibm.com>
2735 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
2738 2014-04-01 Richard Henderson <rth@redhat.com>
2741 * config/i386/i386.md (*float<SWI48><MODEF>2_sse): Leave the second
2742 alternative enabled before register allocation.
2744 2014-04-01 Chung-Lin Tang <cltang@codesourcery.com>
2746 * config/nios2/nios2.md (unspec): Remove UNSPEC_TLS, UNSPEC_TLS_LDM.
2747 * config/nios2/nios2.c (nios2_function_profiler): Fix addi operand
2749 (nios2_large_got_address): Remove unneeded 'sym' parameter.
2750 (nios2_got_address): Update nios2_large_got_address call site.
2751 (nios2_delegitimize_address): New function.
2752 (TARGET_DELEGITIMIZE_ADDRESS): Define to nios2_delegitimize_address.
2753 * config/nios2/linux.h (GLIBC_DYNAMIC_LINKER): Define.
2754 (LINK_SPEC): Specify dynamic linker using GNU_USER_DYNAMIC_LINKER.
2756 2014-04-01 Martin Husemann <martin@duskware.de>
2758 * config/mips/netbsd.h (TARGET_OS_CPP_BUILTINS): Define __mips_o32
2761 2014-04-01 Richard Sandiford <rdsandiford@googlemail.com>
2763 PR rtl-optimization/60604
2764 * recog.c (general_operand): Incorporate REG_CANNOT_CHANGE_MODE_P
2765 check from register_operand.
2766 (register_operand): Redefine in terms of general_operand.
2767 (nonmemory_operand): Use register_operand for the non-constant cases.
2769 2014-04-01 Richard Biener <rguenther@suse.de>
2771 * gimple.h (struct gimple_statement_base): Align subcode to 16 bits.
2773 2014-04-01 Sebastian Huber <sebastian.huber@embedded-brains.de>
2775 * doc/invoke.texi (mapp-regs): Clarify.
2777 2014-03-31 Ulrich Drepper <drepper@gmail.com>
2779 * config/i386/avx512fintrin.h (__v32hi): Define type.
2780 (__v64qi): Likewise.
2781 (_mm512_set1_epi8): Define.
2782 (_mm512_set1_epi16): Define.
2783 (_mm512_set4_epi32): Define.
2784 (_mm512_set4_epi64): Define.
2785 (_mm512_set4_pd): Define.
2786 (_mm512_set4_ps): Define.
2787 (_mm512_setr4_epi64): Define.
2788 (_mm512_setr4_epi32): Define.
2789 (_mm512_setr4_pd): Define.
2790 (_mm512_setr4_ps): Define.
2791 (_mm512_setzero_epi32): Define.
2793 2014-03-31 Martin Jambor <mjambor@suse.cz>
2796 * tree-sra.c (callsite_has_enough_arguments_p): Renamed to
2797 callsite_arguments_match_p. Updated all callers. Also check types of
2798 corresponding formal parameters and actual arguments.
2799 (not_all_callers_have_enough_arguments_p) Renamed to
2800 some_callers_have_mismatched_arguments_p.
2802 2014-03-31 Yuri Rumyantsev <ysrumyan@gmail.com>
2804 * tree-inline.c (copy_loops): Add missed copy of 'safelen'.
2806 2014-03-31 Kugan Vivekanandarajah <kuganv@linaro.org>
2809 * aarch64/aarch64.c (aarch64_classify_address): Fix alignment for
2812 2014-03-30 Uros Bizjak <ubizjak@gmail.com>
2814 * config/i386/sse.md (FMAMODE_NOVF512): New mode iterator.
2815 (<sd_mask_codefor>fma_fmadd_<mode><sd_maskz_name><round_name>):
2817 <sd_mask_codefor>fma_fmadd_<VF_512:mode><sd_maskz_name><round_name>.
2818 Use FMAMODE_NOVF512 mode iterator.
2819 (<sd_mask_codefor>fma_fmsub_<mode><sd_maskz_name><round_name>): Ditto.
2820 (<sd_mask_codefor>fma_fnmadd_<mode><sd_maskz_name><round_name>): Ditto.
2821 (<sd_mask_codefor>fma_fnmsub_<mode><sd_maskz_name><round_name>): Ditto.
2822 (<sd_mask_codefor>fma_fmaddsub_<mode><sd_maskz_name><round_name>):
2824 <sd_mask_codefor>fma_fmaddsub_<VF_512:mode><sd_maskz_name><round_name>.
2825 Use VF_128_256 mode iterator.
2826 (<sd_mask_codefor>fma_fmsubadd_<mode><sd_maskz_name><round_name>):
2829 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
2831 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Clear
2832 static chain if needed.
2834 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
2837 * lra-constraints.c (index_part_to_reg): New.
2838 (process_address): Use it.
2840 2014-03-27 Jeff Law <law@redhat.com>
2841 Jakub Jelinek <jakub@redhat.com>
2844 * expr.c (do_tablejump): Use simplify_gen_binary rather than
2845 gen_rtx_{PLUS,MULT} to build up the address expression.
2847 * i386/i386.c (ix86_legitimize_address): Use copy_addr_to_reg to avoid
2848 creating non-canonical RTL.
2850 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
2853 * ipa-inline.c (want_inline_small_function_p): Short circuit large
2854 functions; reorganize to make cheap checks first.
2855 (inline_small_functions): Do not estimate growth when dumping;
2857 * ipa-inline.h (inline_summary): Add min_size.
2858 (growth_likely_positive): New function.
2859 * ipa-inline-analysis.c (dump_inline_summary): Add min_size.
2860 (set_cond_stmt_execution_predicate): Cleanup.
2861 (estimate_edge_size_and_time): Compute min_size.
2862 (estimate_calls_size_and_time): Likewise.
2863 (estimate_node_size_and_time): Likewise.
2864 (inline_update_overall_summary): Update min_size.
2865 (do_estimate_edge_time): Likewise.
2866 (do_estimate_edge_size): Update.
2867 (do_estimate_edge_hints): Update.
2868 (growth_likely_positive): New function.
2870 2014-03-28 Jakub Jelinek <jakub@redhat.com>
2873 * config/i386/i386.c (ix86_copy_addr_to_reg): Call copy_addr_to_reg
2874 also if addr has VOIDmode.
2876 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2878 * config/arm/aarch-common.c (aarch_crypto_can_dual_issue): New.
2879 * config/arm/aarch-common-protos.h (aarch_crypto_can_dual_issue):
2881 * config/arm/cortex-a53.md: Add reservations and bypass for crypto
2882 instructions as well as AdvancedSIMD loads.
2884 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2886 * config/aarch64/aarch64-simd.md (aarch64_crypto_aes<aes_op>v16qi):
2887 Use crypto_aese type.
2888 (aarch64_crypto_aes<aesmc_op>v16qi): Use crypto_aesmc type.
2889 * config/arm/arm.md (is_neon_type): Replace crypto_aes with
2890 crypto_aese, crypto_aesmc. Move to types.md.
2891 * config/arm/types.md (crypto_aes): Split into crypto_aese,
2893 * config/arm/iterators.md (crypto_type): Likewise.
2895 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
2897 * cgraph.c: Include expr.h and tree-dfa.h.
2898 (cgraph_redirect_edge_call_stmt_to_callee): If call in noreturn;
2901 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
2904 * lra-assigns.c (find_hard_regno_for): Remove unavailable hard
2905 regs from checking multi-reg pseudos.
2907 2014-03-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2909 * config/arm/t-aprofile (MULTILIB_MATCHES): Correct A12 rule.
2911 2014-03-28 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
2913 * config/rs6000/rs6000.c (fusion_gpr_load_p): Refuse optimization
2914 if it would clobber the stack pointer, even temporarily.
2916 2014-03-28 Eric Botcazou <ebotcazou@adacore.com>
2918 * mode-switching.c: Make small adjustments to the top comment.
2920 2014-03-27 Michael Meissner <meissner@linux.vnet.ibm.com>
2922 * config/rs6000/constraints.md (wD constraint): New constraint to
2923 match the constant integer to get the top DImode/DFmode out of a
2924 vector in a VSX register.
2926 * config/rs6000/predicates.md (vsx_scalar_64bit): New predicate to
2927 match the constant integer to get the top DImode/DFmode out of a
2928 vector in a VSX register.
2930 * config/rs6000/rs6000-builtins.def (VBPERMQ): Add vbpermq builtin
2933 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
2936 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
2937 -mdebug=reg, print value of VECTOR_ELEMENT_SCALAR_64BIT.
2939 * config/rs6000/vsx.md (vsx_extract_<mode>, V2DI/V2DF modes):
2940 Optimize vec_extract of 64-bit values, where the value being
2941 extracted is in the top word, where we can use scalar
2942 instructions. Add direct move and store support. Combine the big
2943 endian/little endian vector select load support into a single insn.
2944 (vsx_extract_<mode>_internal1): Likewise.
2945 (vsx_extract_<mode>_internal2): Likewise.
2946 (vsx_extract_<mode>_load): Likewise.
2947 (vsx_extract_<mode>_store): Likewise.
2948 (vsx_extract_<mode>_zero): Delete, big and little endian insns are
2949 combined into vsx_extract_<mode>_load.
2950 (vsx_extract_<mode>_one_le): Likewise.
2952 * config/rs6000/rs6000.h (VECTOR_ELEMENT_SCALAR_64BIT): Macro to
2953 define the top 64-bit vector element.
2955 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wD
2958 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
2959 Document vec_vbpermq builtin.
2962 * config/rs6000/altivec.h (vec_xxsldwi): Add missing define to
2963 enable use of xxsldwi and xxpermdi builtin functions.
2964 (vec_xxpermdi): Likewise.
2966 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
2967 Document use of vec_xxsldwi and vec_xxpermdi builtins.
2969 2014-03-27 Vladimir Makarov <vmakarov@redhat.com>
2971 PR rtl-optimization/60650
2972 * lra-assign.c (find_hard_regno_for, spill_for): Add parameter
2974 (find_spills_for): New.
2975 (assign_by_spills): Pass the new parameter to find_hard_regno_for.
2976 Spill all pseudos on the second iteration.
2978 2014-03-27 Marek Polacek <polacek@redhat.com>
2981 * doc/extend.texi (ffs Builtins): Change unsigned types to signed
2984 2014-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2986 * config/s390/s390.c (s390_can_use_return_insn): Check for
2987 call-saved FPRs on 31 bit.
2989 2014-03-27 Jakub Jelinek <jakub@redhat.com>
2992 * omp-low.c (lower_omp_1): For gimple_clobber_p stmts,
2993 if they need regimplification, just drop them instead of
2994 calling gimple_regimplify_operands on them.
2996 2014-03-27 Marcus Shawcroft <marcus.shawcroft@arm.com>
2999 * config/aarch64/aarch64.c (faked_omit_frame_pointer): Remove.
3000 (aarch64_frame_pointer_required): Adjust logic.
3001 (aarch64_can_eliminate): Adjust logic.
3002 (aarch64_override_options_after_change): Adjust logic.
3004 2014-03-27 Dehao Chen <dehao@google.com>
3006 * ipa-inline.c (early_inliner): Update node's inline info.
3008 2014-03-26 Dehao Chen <dehao@google.com>
3010 * dojump.c (do_compare_rtx_and_jump): Sets correct probability for
3011 compiler inserted conditional jumps for NAN float check.
3013 2014-03-26 Jakub Jelinek <jakub@redhat.com>
3015 * ubsan.h (ubsan_create_data): Change second argument's type
3016 to const location_t *.
3017 * ubsan.c (ubsan_source_location): If xloc.file is NULL, set it to
3019 (ubsan_create_data): Change second argument to const location_t *PLOC.
3020 Create Loc field whenever PLOC is non-NULL.
3021 (ubsan_instrument_unreachable, ubsan_expand_null_ifn,
3022 ubsan_build_overflow_builtin, instrument_bool_enum_load): Adjust
3026 * real.c (real_to_integer2): Change type of low to UHWI.
3028 2014-03-26 Tobias Burnus <burnus@net-b.de>
3030 * gcc.c (LINK_COMMAND_SPEC): Use libcilkrts.spec for -fcilkplus.
3031 (CILK_SELF_SPECS): New define.
3032 (driver_self_specs): Use it.
3034 2014-03-26 Richard Biener <rguenther@suse.de>
3036 * tree-pretty-print.c (percent_K_format): Implement special
3037 case for LTO and its stripped down BLOCK tree.
3039 2014-03-26 Jakub Jelinek <jakub@redhat.com>
3042 * ubsan.c (instrument_si_overflow): Instrument ABS_EXPR.
3044 * tree-vrp.c (simplify_internal_call_using_ranges): If only
3045 one range is range_int_cst_p, but not both, at least optimize
3046 addition/subtraction of 0 and multiplication by 0 or 1.
3047 * gimple-fold.c (gimple_fold_call): Fold
3048 IFN_UBSAN_CHECK_{ADD,SUB,MUL}.
3049 (gimple_fold_stmt_to_constant_1): If both op0 and op1 aren't
3050 INTEGER_CSTs, try to fold at least x * 0 and y - y.
3052 2014-03-26 Eric Botcazou <ebotcazou@adacore.com>
3054 PR rtl-optimization/60452
3055 * rtlanal.c (rtx_addr_can_trap_p_1): Fix head comment.
3056 <case REG>: Return 1 for invalid offsets from the frame pointer.
3058 2014-03-26 Marek Polacek <polacek@redhat.com>
3061 * doc/extend.texi (C Extensions): Mention variable-length arrays in
3064 2014-03-26 Marek Polacek <polacek@redhat.com>
3067 * doc/extend.texi (Designated Inits): Describe what happens to omitted
3070 2014-03-26 Marek Polacek <polacek@redhat.com>
3073 * ira-color.c (update_conflict_hard_regno_costs): Perform the
3074 multiplication in unsigned type.
3076 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
3078 * doc/install.texi: Document nds32le-*-elf and nds32be-*-elf.
3080 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
3082 * doc/contrib.texi: Add myself as Andes nds32 port contributor.
3084 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
3087 * cif-code.def (UNREACHABLE) New code.
3088 * ipa-inline.c (inline_small_functions): Skip edges to
3089 __builtlin_unreachable.
3090 (estimate_edge_growth): Allow edges to __builtlin_unreachable.
3091 * ipa-inline-analysis.c (edge_set_predicate): Redirect edges with false
3092 predicate to __bulitin_unreachable.
3093 (set_cond_stmt_execution_predicate): Fix issue when
3094 invert_tree_comparison returns ERROR_MARK.
3095 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow): Do not
3096 propagate to inline clones.
3097 * cgraph.c (verify_edge_corresponds_to_fndecl): Allow redirection
3099 * ipa-cp.c (create_specialized_node): Be ready for new node to appear.
3100 * cgraphclones.c (cgraph_clone_node): If call destination is already
3101 ureachable, do not redirect it back.
3102 * tree-inline.c (fold_marked_statements): Hanlde calls becoming
3105 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
3107 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow):
3108 Do not modify inline clones.
3110 2014-03-25 Jakub Jelinek <jakub@redhat.com>
3112 * config/i386/i386.md (general_sext_operand): New mode attr.
3113 (addv<mode>4, subv<mode>4, mulv<mode>4): If operands[2] is CONST_INT,
3114 don't generate (sign_extend (const_int)).
3115 (*addv<mode>4, *subv<mode>4, *mulv<mode>4): Disallow CONST_INT_P
3116 operands[2]. Use We constraint instead of <i> and
3117 <general_sext_operand> predicate instead of <general_operand>.
3118 (*addv<mode>4_1, *subv<mode>4_1, *mulv<mode>4_1): New insns.
3119 * config/i386/constraints.md (We): New constraint.
3120 * config/i386/predicates.md (x86_64_sext_operand,
3121 sext_operand): New predicates.
3123 2014-03-25 Martin Jambor <mjambor@suse.cz>
3126 * ipa-cp.c (ipa_get_indirect_edge_target_1): Redirect type
3127 inconsistent devirtualizations to __builtin_unreachable.
3129 2014-03-25 Marek Polacek <polacek@redhat.com>
3132 * doc/extend.texi (Example of asm with clobbered asm reg): Fix typo.
3134 2014-03-25 Alan Lawrence <alan.lawrence@arm.com>
3136 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Reverse
3137 order of elements for big-endian.
3139 2014-03-25 Richard Biener <rguenther@suse.de>
3142 * gimplify-me.c (gimple_regimplify_operands): Update the
3145 2014-03-25 Martin Jambor <mjambor@suse.cz>
3148 * lto-cgraph.c (lto_output_node): Stream body_removed flag.
3149 (lto_output_varpool_node): Likewise.
3150 (input_overwrite_node): Likewise.
3151 (input_varpool_node): Likewise.
3153 2014-03-25 Richard Biener <rguenther@suse.de>
3155 * lto-wrapper.c (merge_and_complain): Handle OPT_fPIE like OPT_fpie.
3156 (run_gcc): Likewise.
3158 2014-03-25 Jakub Jelinek <jakub@redhat.com>
3160 * combine.c (simplify_compare_const): Add MODE argument.
3161 Handle mode_width 0 as very large mode_width.
3162 (try_combine, simplify_comparison): Adjust callers.
3164 * cselib.c (cselib_hash_rtx): Perform addition in unsigned
3165 type to avoid signed integer overflow.
3166 * explow.c (plus_constant): Likewise.
3168 2014-03-25 Dominik Vogt <vogt@linux.vnet.ibm.com>
3170 * doc/generic.texi: Correct typos.
3172 2014-03-24 Tobias Burnus <burnus@net-b.de>
3174 * doc/invoke.texi (-flto): Expand section about
3175 using static libraries with LTO.
3177 2014-03-24 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3179 PR rtl-optimization/60501
3180 * optabs.def (addptr3_optab): New optab.
3181 * optabs.c (gen_addptr3_insn, have_addptr3_insn): New function.
3182 * doc/md.texi ("addptrm3"): Document new RTL standard expander.
3183 * expr.h (gen_addptr3_insn, have_addptr3_insn): Add prototypes.
3185 * lra.c (emit_add3_insn): Use the addptr pattern if available.
3187 * config/s390/s390.md ("addptrdi3", "addptrsi3"): New expanders.
3189 2014-03-24 Ulrich Drepper <drepper@gmail.com>
3191 * config/i386/avx512fintrin.h: Define _mm512_set1_ps and
3194 * config/i386/avxintrin.h (_mm256_undefined_si256): Define.
3195 (_mm256_undefined_ps): Define.
3196 (_mm256_undefined_pd): Define.
3197 * config/i386/emmintrin.h (_mm_undefined_si128): Define.
3198 (_mm_undefined_pd): Define.
3199 * config/i386/xmmintrin.h (_mm_undefined_ps): Define.
3200 * config/i386/avx512fintrin.h (_mm512_undefined_si512): Define.
3201 (_mm512_undefined_ps): Define.
3202 (_mm512_undefined_pd): Define.
3203 Use _mm*_undefined_*.
3204 * config/i386/avx2intrin.h: Use _mm*_undefined_*.
3206 2014-03-24 Alex Velenko <Alex.Velenko@arm.com>
3208 * config/aarch64/aarch64-simd-builtins.def (lshr): DI mode excluded.
3209 (lshr_simd): DI mode added.
3210 * config/aarch64/aarch64-simd.md (aarch64_lshr_simddi): New pattern.
3211 (aarch64_ushr_simddi): Likewise.
3212 * config/aarch64/aarch64.md (UNSPEC_USHR64): New unspec.
3213 * config/aarch64/arm_neon.h (vshr_n_u64): Intrinsic fixed.
3214 (vshrd_n_u64): Likewise.
3216 2014-03-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3218 * Makefile.in (s-macro_list): Depend on cc1.
3220 2014-03-23 Teresa Johnson <tejohnson@google.com>
3222 * ipa-utils.c (ipa_print_order): Use specified dump file.
3224 2014-03-23 Eric Botcazou <ebotcazou@adacore.com>
3226 PR rtl-optimization/60601
3227 * bb-reorder.c (fix_up_fall_thru_edges): Test EDGE_FALLTHRU everywhere.
3229 * gcc.c (eval_spec_function): Initialize save_growing_value.
3231 2014-03-22 Jakub Jelinek <jakub@redhat.com>
3234 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): For
3235 code == MINUS_EXPR, never swap op0 with op1.
3237 * toplev.c (init_local_tick): Avoid signed integer multiplication
3239 * genautomata.c (reserv_sets_hash_value): Fix rotate idiom, avoid
3240 shift by first operand's bitsize.
3242 2014-03-21 Jakub Jelinek <jakub@redhat.com>
3245 * config/i386/i386.h (TARGET_64BIT_P): If not TARGET_BI_ARCH,
3247 * config/i386/darwin.h (TARGET_64BIT_P): Redefine to
3248 TARGET_ISA_64BIT_P(x).
3250 2014-03-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3252 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Generate a
3253 pattern for vector nor instead of subtract from splat(-1).
3254 (altivec_expand_vec_perm_const_le): Likewise.
3256 2014-03-21 Richard Henderson <rth@twiddle.net>
3259 * ifcvt.c (dead_or_predicable): Return FALSE if there are any frame
3260 related insns after epilogue_completed.
3262 2014-03-21 Martin Jambor <mjambor@suse.cz>
3265 * cgraph.h (symtab_node): New flag body_removed.
3266 * ipa.c (symtab_remove_unreachable_nodes): Set body_removed flag
3267 when removing bodies.
3268 * symtab.c (dump_symtab_base): Dump body_removed flag.
3269 * cgraph.c (verify_edge_corresponds_to_fndecl): Skip nodes which
3270 had their bodies removed.
3272 2014-03-21 Martin Jambor <mjambor@suse.cz>
3275 * ipa.c (symtab_remove_unreachable_nodes): Clear thunk flag of nodes
3278 2014-03-21 Richard Biener <rguenther@suse.de>
3280 PR tree-optimization/60577
3281 * tree-core.h (struct tree_base): Document nothrow_flag use
3282 in VAR_DECL_NONALIASED.
3283 * tree.h (VAR_DECL_NONALIASED): New.
3284 (may_be_aliased): Adjust.
3285 * coverage.c (build_var): Set VAR_DECL_NONALIASED.
3287 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
3289 * expr.c (expand_expr_real_1): Remove outdated comment.
3291 2014-03-20 Jakub Jelinek <jakub@redhat.com>
3294 * ira.c (adjust_cleared_regs): Call copy_rtx on
3295 *reg_equiv[REGNO (loc)].src_p before passing it to
3296 simplify_replace_fn_rtx.
3299 * config/i386/i386.c (x86_output_mi_thunk): Surround UNSPEC_GOT
3300 into CONST, put pic register as first operand of PLUS. Use
3301 gen_const_mem for both 32-bit and 64-bit PIC got loads.
3303 2014-03-20 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3305 * config/aarch64/aarch64.c (MEMORY_MOVE_COST): Delete.
3307 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
3309 * config/sparc/sparc.c (sparc_do_work_around_errata): Implement work
3310 around for store forwarding issue in the FPU on the UT699.
3311 * config/sparc/sparc.md (in_branch_delay): Return false for single FP
3312 loads and operations if -mfix-ut699 is specified.
3313 (divtf3_hq): Tweak attribute.
3314 (sqrttf2_hq): Likewise.
3316 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
3318 * calls.c (store_one_arg): Remove incorrect const qualification on the
3319 type of the temporary.
3320 * cfgexpand.c (expand_return): Likewise.
3321 * expr.c (expand_constructor): Likewise.
3322 (expand_expr_real_1): Likewise.
3324 2014-03-20 Zhenqiang Chen <zhenqiang.chen@linaro.org>
3326 * config/arm/arm.c (arm_dwarf_register_span): Update the element number
3329 2014-03-19 Kaz Kojima <kkojima@gcc.gnu.org>
3332 * config/sh/sh.md (udivsi3_i1): Clobber R1 register.
3334 2014-03-19 James Greenhalgh <james.greenhalgh@arm.com>
3336 * config/arm/aarch-common-protos.h
3337 (alu_cost_table): Fix spelling of "extend".
3338 * config/arm/arm.c (arm_new_rtx_costs): Fix spelling of "extend".
3340 2014-03-19 Richard Biener <rguenther@suse.de>
3343 * tree-core.h (tree_type_common): Re-order pointer members
3344 to reduce recursion depth during GC walks.
3346 2014-03-19 Marek Polacek <polacek@redhat.com>
3349 * ubsan.c (ubsan_type_descriptor): Check that DECL_NAME is nonnull
3350 before accessing it.
3352 2014-03-19 Richard Biener <rguenther@suse.de>
3355 * lto-streamer-in.c (input_function): In WPA stage do not drop
3358 2014-03-19 Jakub Jelinek <jakub@redhat.com>
3360 PR tree-optimization/60559
3361 * vectorizable_mask_load_store): Replace scalar MASK_LOAD
3362 with build_zero_cst assignment.
3364 2014-03-18 Kai Tietz <ktietz@redhat.com>
3366 PR rtl-optimization/56356
3367 * sdbout.c (sdbout_parms): Verify that parms'
3368 incoming argument is valid.
3369 (sdbout_reg_parms): Likewise.
3371 2014-03-18 Richard Henderson <rth@redhat.com>
3374 * config/i386/i386.md (*float<SWI48x><MODEF>2_i387): Move down to
3375 be shadowed by *float<SWI48><MODEF>2_sse. Test X87_ENABLE_FLOAT.
3376 (*float<SWI48><MODEF>2_sse): Check X87_ENABLE_FLOAT for alternative 0.
3378 2014-03-18 Basile Starynkevitch <basile@starynkevitch.net>
3380 * plugin.def: Improve comment for PLUGIN_INCLUDE_FILE.
3381 * doc/plugins.texi (Plugin callbacks): Mention PLUGIN_INCLUDE_FILE.
3382 Italicize plugin event names in description. Explain that
3383 PLUGIN_PRAGMAS has no sense for lto1. Explain PLUGIN_INCLUDE_FILE.
3384 Remind that no GCC functions should be called after PLUGIN_FINISH.
3385 Explain what pragmas with expansion are.
3387 2014-03-18 Martin Liska <mliska@suse.cz>
3389 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Added case when
3390 gimple call statement is update.
3391 * gimple-fold.c (gimple_fold_call): Changed order for GIMPLE_ASSIGN and
3392 GIMPLE_CALL, where gsi iterator still points to GIMPLE CALL.
3394 2014-03-18 Jakub Jelinek <jakub@redhat.com>
3397 * ubsan.c (ubsan_instrument_unreachable): Call
3398 initialize_sanitizer_builtins.
3399 (ubsan_pass): Likewise.
3402 * ubsan.c (ubsan_type_descriptor, ubsan_create_data): Call
3403 varpool_finalize_decl instead of rest_of_decl_compilation.
3405 2014-03-18 Richard Biener <rguenther@suse.de>
3407 * df-problems.c (df_rd_confluence_n): Avoid bitmap_copy
3408 by using bitmap_and_compl instead of bitmap_and_compl_into.
3409 (df_rd_transfer_function): Likewise.
3411 2014-03-18 Richard Biener <rguenther@suse.de>
3413 * doc/lto.texi (fresolution): Fix typo.
3415 2014-03-18 Richard Biener <rguenther@suse.de>
3417 * doc/invoke.texi (flto): Update for changes in 4.9.
3419 2014-03-18 Richard Biener <rguenther@suse.de>
3421 * doc/loop.texi: Remove section on the removed lambda framework.
3422 Update loop docs with recent changes in preserving loop structure.
3424 2014-03-18 Richard Biener <rguenther@suse.de>
3426 * doc/lto.texi (-fresolution): Document.
3428 2014-03-18 Richard Biener <rguenther@suse.de>
3430 * doc/contrib.texi: Adjust my name.
3432 2014-03-18 Jakub Jelinek <jakub@redhat.com>
3435 * internal-fn.c: Include diagnostic-core.h.
3436 (expand_BUILTIN_EXPECT): New function.
3437 * gimplify.c (gimplify_call_expr): Use false instead of FALSE.
3438 (gimplify_modify_expr): Gimplify 3 argument __builtin_expect into
3439 IFN_BUILTIN_EXPECT call instead of __builtin_expect builtin call.
3440 * ipa-inline-analysis.c (find_foldable_builtin_expect): Handle
3442 * predict.c (expr_expected_value_1): Handle IFN_BUILTIN_EXPECT.
3443 Revert 3 argument __builtin_expect code.
3444 (strip_predict_hints): Handle IFN_BUILTIN_EXPECT.
3445 * gimple-fold.c (gimple_fold_call): Likewise.
3446 * tree.h (fold_builtin_expect): New prototype.
3447 * builtins.c (build_builtin_expect_predicate): Add predictor
3448 argument, if non-NULL, create 3 argument __builtin_expect.
3449 (fold_builtin_expect): No longer static. Add ARG2 argument,
3450 pass it through to build_builtin_expect_predicate.
3451 (fold_builtin_2): Adjust caller.
3452 (fold_builtin_3): Handle BUILT_IN_EXPECT.
3453 * internal-fn.def (BUILTIN_EXPECT): New.
3455 2014-03-18 Tobias Burnus <burnus@net-b.de>
3458 * predict.def (PRED_FORTRAN_OVERFLOW, PRED_FORTRAN_FAIL_ALLOC,
3459 PRED_FORTRAN_FAIL_IO, PRED_FORTRAN_WARN_ONCE, PRED_FORTRAN_SIZE_ZERO,
3460 PRED_FORTRAN_INVALID_BOUND, PRED_FORTRAN_ABSENT_DUMMY): Add.
3462 2014-03-18 Jan Hubicka <hubicka@ucw.cz>
3465 * predict.c (combine_predictions_for_bb): Fix up formatting.
3466 (expr_expected_value_1, expr_expected_value): Add predictor argument,
3467 fill what it points to if non-NULL.
3468 (tree_predict_by_opcode): Adjust caller, use the predictor.
3469 * predict.def (PRED_COMPARE_AND_SWAP): Add.
3471 2014-03-18 Eric Botcazou <ebotcazou@adacore.com>
3473 * config/sparc/sparc.c (sparc_do_work_around_errata): Speed up and use
3474 proper constant for the store mode.
3476 2014-03-18 Ilya Enkovich <ilya.enkovich@intel.com>
3478 * symtab.c (change_decl_assembler_name): Fix transparent alias
3481 2014-03-16 Renlin Li <Renlin.Li@arm.com>
3483 * config/aarch64/aarch64.c: Correct the comments about the
3484 aarch64 stack layout.
3486 2014-03-18 Thomas Schwinge <thomas@codesourcery.com>
3488 * omp-low.c (lower_rec_input_clauses) <build_omp_barrier>: Restore
3489 check for GF_OMP_FOR_KIND_FOR.
3491 2013-03-18 Kirill Yukhin <kirill.yukhin@intel.com>
3493 * config/i386/i386.h (ADDITIONAL_REGISTER_NAMES): Add
3494 ymm and zmm register names.
3496 2014-03-17 Jakub Jelinek <jakub@redhat.com>
3499 * config/i386/i386.c (ix86_expand_epilogue): Adjust REG_CFA_ADJUST_CFA
3500 note creation for the 2010-08-31 changes.
3502 2014-03-17 Marek Polacek <polacek@redhat.com>
3505 * omp-low.c (omp_max_vf): Treat -fno-tree-loop-optimize the same
3506 as -fno-tree-loop-vectorize.
3507 (expand_omp_simd): Likewise.
3509 2014-03-15 Eric Botcazou <ebotcazou@adacore.com>
3511 * config/sparc/sparc-protos.h (tls_call_delay): Delete.
3512 (eligible_for_call_delay): New prototype.
3513 * config/sparc/sparc.c (tls_call_delay): Rename into...
3514 (eligible_for_call_delay): ...this. Return false if the instruction
3515 cannot be put in the delay slot of a branch.
3516 (eligible_for_restore_insn): Simplify.
3517 (eligible_for_return_delay): Return false if the instruction cannot be
3518 put in the delay slot of a branch and simplify.
3519 (eligible_for_sibcall_delay): Return false if the instruction cannot be
3520 put in the delay slot of a branch.
3521 * config/sparc/sparc.md (fix_ut699): New attribute.
3522 (tls_call_delay): Delete.
3523 (in_call_delay): Reimplement.
3524 (eligible_for_sibcall_delay): Rename into...
3525 (in_sibcall_delay): ...this.
3526 (eligible_for_return_delay): Rename into...
3527 (in_return_delay): ...this.
3528 (in_branch_delay): Reimplement.
3529 (in_uncond_branch_delay): Delete.
3530 (in_annul_branch_delay): Delete.
3532 2014-03-14 Richard Henderson <rth@redhat.com>
3535 * config/i386/i386.md (floathi<X87MODEF>2): Delete expander; rename
3536 define_insn from *floathi<X87MODEF>2_i387; allow nonimmediate_operand.
3537 (*floathi<X87MODEF>2_i387_with_temp): Remove.
3538 (floathi splitters): Remove.
3539 (float<SWI48x>xf2): New pattern.
3540 (float<SWI48><MODEF>2): Rename from float<SWI48x><X87MODEF>2. Drop
3541 code that tried to handle DImode for 32-bit, but which was excluded
3542 by the pattern's condition. Drop allocation of stack temporary.
3543 (*floatsi<MODEF>2_vector_mixed_with_temp): Remove.
3544 (*float<SWI48><MODEF>2_mixed_with_temp): Remove.
3545 (*float<SWI48><MODEF>2_mixed_interunit): Remove.
3546 (*float<SWI48><MODEF>2_mixed_nointerunit): Remove.
3547 (*floatsi<MODEF>2_vector_sse_with_temp): Remove.
3548 (*float<SWI48><MODEF>2_sse_with_temp): Remove.
3549 (*float<SWI48><MODEF>2_sse_interunit): Remove.
3550 (*float<SWI48><MODEF>2_sse_nointerunit): Remove.
3551 (*float<SWI48x><X87MODEF>2_i387_with_temp): Remove.
3552 (*float<SWI48x><X87MODEF>2_i387): Remove.
3553 (all float _with_temp splitters): Remove.
3554 (*float<SWI48x><MODEF>2_i387): New pattern.
3555 (*float<SWI48><MODEF>2_sse): New pattern.
3556 (float TARGET_USE_VECTOR_CONVERTS splitters): Merge them.
3557 (float TARGET_SSE_PARTIAL_REG_DEPENDENCY splitters): Merge them.
3559 2014-03-14 Jakub Jelinek <jakub@redhat.com>
3560 Marek Polacek <polacek@redhat.com>
3563 * common.opt (dump_base_name_prefixed): New Variable.
3564 * opts.c (finish_options): Don't prepend directory to x_dump_base_name
3565 if x_dump_base_name_prefixed is already set, set it at the end.
3567 2014-03-14 Vladimir Makarov <vmakarov@redhat.com>
3569 PR rtl-optimization/60508
3570 * lra-constraints.c (get_reload_reg): Add new parameter
3572 (process_addr_reg, simplify_operand_subreg, curr_insn_transform):
3573 Pass the new parameter values.
3575 2014-03-14 Richard Biener <rguenther@suse.de>
3577 * common.opt: Revert unintented changes from r205065.
3580 2014-03-14 Richard Biener <rguenther@suse.de>
3583 * cfghooks.c (split_block): Properly adjust all loops the
3584 block was a latch of.
3586 2014-03-14 Martin Jambor <mjambor@suse.cz>
3589 * ipa-prop.c (ipa_modify_call_arguments): Fix iteration condition
3592 2014-03-14 Georg-Johann Lay <avr@gjlay.de>
3595 * config/avr/avr.c (avr_set_current_function): Pass function name
3596 through default_strip_name_encoding before sanity checking instead
3597 of skipping the first char of the assembler name.
3599 2014-03-13 Richard Henderson <rth@redhat.com>
3602 * config/i386/i386.c (ix86_split_fp_branch): Remove pushed argument.
3603 (ix86_force_to_memory, ix86_free_from_memory): Remove.
3604 * config/i386/i386-protos.h: Likewise.
3605 * config/i386/i386.md (floathi<X87MODEF>2): Use assign_386_stack_local
3606 in the expander instead of a splitter.
3607 (float<SWI48x><X87MODEF>2): Use assign_386_stack_local if there is
3608 any possibility of requiring a memory.
3609 (*floatsi<MODEF>2_vector_mixed): Remove, and the splitters.
3610 (*floatsi<MODEF>2_vector_sse): Remove, and the splitters.
3611 (fp branch splitters): Update for ix86_split_fp_branch.
3612 (*jcc<X87MODEF>_<SWI24>_i387): Remove r/f alternative.
3613 (*jcc<X87MODEF>_<SWI24>_r_i387): Likewise.
3614 (splitter for jcc<X87MODEF>_<SWI24>_i387 r/f): Remove.
3615 (*fop_<MODEF>_2_i387): Remove f/r alternative.
3616 (*fop_<MODEF>_3_i387): Likewise.
3617 (*fop_xf_2_i387, *fop_xf_3_i387): Likewise.
3618 (splitters for the fop_* register patterns): Remove.
3619 (fscalexf4_i387): Rename from *fscalexf4_i387.
3620 (ldexpxf3): Use gen_floatsixf2 and gen_fscalexf4_i387.
3622 2014-03-13 Jakub Jelinek <jakub@redhat.com>
3624 PR tree-optimization/59779
3625 * tree-dfa.c (get_ref_base_and_extent): Use double_int
3626 type for bitsize and maxsize instead of HOST_WIDE_INT.
3628 2014-03-13 Steven Bosscher <steven@gcc.gnu.org>
3630 PR rtl-optimization/57320
3631 * function.c (rest_of_handle_thread_prologue_and_epilogue): Cleanup
3632 the CFG after thread_prologue_and_epilogue_insns.
3634 2014-03-13 Vladimir Makarov <vmakarov@redhat.com>
3636 PR rtl-optimization/57189
3637 * lra-constraints.c (process_alt_operands): Disfavor spilling
3640 2014-03-13 Cesar Philippidis <cesar@codesourcery.com>
3642 * lto-wrapper.c (maybe_unlink_file): Suppress diagnostic messages.
3644 2014-03-13 Jakub Jelinek <jakub@redhat.com>
3646 PR tree-optimization/59025
3648 * tree-ssa-reassoc.c (sort_by_operand_rank): For SSA_NAMEs with the
3649 same rank, sort by bb_rank and gimple_uid of SSA_NAME_DEF_STMT first.
3651 2014-03-13 Georg-Johann Lay <avr@gjlay.de>
3654 * config/avr/avr.c (avr_out_plus): Swap cc_plus and cc_minus in
3655 calls of avr_out_plus_1.
3657 2014-03-13 Bin Cheng <bin.cheng@arm.com>
3659 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Record
3660 BB's single pred and update the father loop's latch info later.
3662 2014-03-12 Michael Meissner <meissner@linux.vnet.ibm.com>
3664 * config/rs6000/vector.md (VEC_L): Add V1TI mode to vector types.
3668 (VEC_base): Likewise.
3669 (mov<MODE>, VEC_M modes): If we are loading TImode into VSX
3670 registers, we need to swap double words in little endian mode.
3672 * config/rs6000/rs6000-modes.def (V1TImode): Add new vector mode
3673 to be a container mode for 128-bit integer operations added in ISA
3674 2.07. Unlike TImode and PTImode, the preferred register set is
3675 the Altivec/VMX registers for the 128-bit operations.
3677 * config/rs6000/rs6000-protos.h (rs6000_move_128bit_ok_p): Add
3679 (rs6000_split_128bit_ok_p): Likewise.
3681 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_3): Add new support
3682 macros for creating ISA 2.07 normal and overloaded builtin
3683 functions with 3 arguments.
3684 (BU_P8V_OVERLOAD_3): Likewise.
3685 (VPERM_1T): Add support for V1TImode in 128-bit vector operations
3686 for use as overloaded functions.
3687 (VPERM_1TI_UNS): Likewise.
3688 (VSEL_1TI): Likewise.
3689 (VSEL_1TI_UNS): Likewise.
3690 (ST_INTERNAL_1ti): Likewise.
3691 (LD_INTERNAL_1ti): Likewise.
3692 (XXSEL_1TI): Likewise.
3693 (XXSEL_1TI_UNS): Likewise.
3694 (VPERM_1TI): Likewise.
3695 (VPERM_1TI_UNS): Likewise.
3696 (XXPERMDI_1TI): Likewise.
3697 (SET_1TI): Likewise.
3698 (LXVD2X_V1TI): Likewise.
3699 (STXVD2X_V1TI): Likewise.
3700 (VEC_INIT_V1TI): Likewise.
3701 (VEC_SET_V1TI): Likewise.
3702 (VEC_EXT_V1TI): Likewise.
3703 (EQV_V1TI): Likewise.
3704 (NAND_V1TI): Likewise.
3705 (ORC_V1TI): Likewise.
3706 (VADDCUQ): Add support for 128-bit integer arithmetic instructions
3707 added in ISA 2.07. Add both normal 'altivec' builtins, and the
3709 (VADDUQM): Likewise.
3710 (VSUBCUQ): Likewise.
3711 (VADDEUQM): Likewise.
3712 (VADDECUQ): Likewise.
3713 (VSUBEUQM): Likewise.
3714 (VSUBECUQ): Likewise.
3716 * config/rs6000/rs6000-c.c (__int128_type): New static to hold
3717 __int128_t and __uint128_t types.
3718 (__uint128_type): Likewise.
3719 (altivec_categorize_keyword): Add support for vector __int128_t,
3720 vector __uint128_t, vector __int128, and vector unsigned __int128
3721 as a container type for TImode operations that need to be done in
3722 VSX/Altivec registers.
3723 (rs6000_macro_to_expand): Likewise.
3724 (altivec_overloaded_builtins): Add ISA 2.07 overloaded functions
3725 to support 128-bit integer instructions vaddcuq, vadduqm,
3726 vaddecuq, vaddeuqm, vsubcuq, vsubuqm, vsubecuq, vsubeuqm.
3727 (altivec_resolve_overloaded_builtin): Add support for V1TImode.
3729 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Add support
3730 for V1TImode, and set up preferences to use VSX/Altivec registers.
3731 Setup VSX reload handlers.
3732 (rs6000_debug_reg_global): Likewise.
3733 (rs6000_init_hard_regno_mode_ok): Likewise.
3734 (rs6000_preferred_simd_mode): Likewise.
3735 (vspltis_constant): Do not allow V1TImode as easy altivec constants.
3736 (easy_altivec_constant): Likewise.
3737 (output_vec_const_move): Likewise.
3738 (rs6000_expand_vector_set): Convert V1TImode set and extract to
3740 (rs6000_expand_vector_extract): Likewise.
3741 (reg_offset_addressing_ok_p): Setup V1TImode to use VSX reg+reg
3743 (rs6000_const_vec): Add support for V1TImode.
3744 (rs6000_emit_le_vsx_load): Swap double words when loading or
3745 storing TImode/V1TImode.
3746 (rs6000_emit_le_vsx_store): Likewise.
3747 (rs6000_emit_le_vsx_move): Likewise.
3748 (rs6000_emit_move): Add support for V1TImode.
3749 (altivec_expand_ld_builtin): Likewise.
3750 (altivec_expand_st_builtin): Likewise.
3751 (altivec_expand_vec_init_builtin): Likewise.
3752 (altivec_expand_builtin): Likewise.
3753 (rs6000_init_builtins): Add support for V1TImode type. Add
3754 support for ISA 2.07 128-bit integer builtins. Define type names
3755 for the VSX/Altivec vector types.
3756 (altivec_init_builtins): Add support for overloaded vector
3757 functions with V1TImode type.
3758 (rs6000_preferred_reload_class): Prefer Altivec registers for V1TImode.
3759 (rs6000_move_128bit_ok_p): Move 128-bit move/split validation to
3761 (rs6000_split_128bit_ok_p): Likewise.
3762 (rs6000_handle_altivec_attribute): Create V1TImode from vector
3763 __int128_t and vector __uint128_t.
3765 * config/rs6000/vsx.md (VSX_L): Add V1TImode to vector iterators
3766 and mode attributes.
3773 (VS_scalar): Likewise.
3774 (VS_double): Likewise.
3775 (vsx_set_v1ti): New builtin function to create V1TImode from TImode.
3777 * config/rs6000/rs6000.h (TARGET_VADDUQM): New macro to say whether
3778 we support the ISA 2.07 128-bit integer arithmetic instructions.
3779 (ALTIVEC_OR_VSX_VECTOR_MODE): Add V1TImode.
3780 (enum rs6000_builtin_type_index): Add fields to hold V1TImode
3781 and TImode types for use with the builtin functions.
3782 (V1TI_type_node): Likewise.
3783 (unsigned_V1TI_type_node): Likewise.
3784 (intTI_type_internal_node): Likewise.
3785 (uintTI_type_internal_node): Likewise.
3787 * config/rs6000/altivec.md (UNSPEC_VADDCUQ): New unspecs for ISA 2.07
3788 128-bit builtin functions.
3789 (UNSPEC_VADDEUQM): Likewise.
3790 (UNSPEC_VADDECUQ): Likewise.
3791 (UNSPEC_VSUBCUQ): Likewise.
3792 (UNSPEC_VSUBEUQM): Likewise.
3793 (UNSPEC_VSUBECUQ): Likewise.
3794 (VM): Add V1TImode to vector mode iterators.
3796 (VI_unit): Likewise.
3797 (altivec_vadduqm): Add ISA 2.07 128-bit binary builtins.
3798 (altivec_vaddcuq): Likewise.
3799 (altivec_vsubuqm): Likewise.
3800 (altivec_vsubcuq): Likewise.
3801 (altivec_vaddeuqm): Likewise.
3802 (altivec_vaddecuq): Likewise.
3803 (altivec_vsubeuqm): Likewise.
3804 (altivec_vsubecuq): Likewise.
3806 * config/rs6000/rs6000.md (FMOVE128_GPR): Add V1TImode to vector
3808 (BOOL_128): Likewise.
3809 (BOOL_REGS_OUTPUT): Likewise.
3810 (BOOL_REGS_OP1): Likewise.
3811 (BOOL_REGS_OP2): Likewise.
3812 (BOOL_REGS_UNARY): Likewise.
3813 (BOOL_REGS_AND_CR0): Likewise.
3815 * config/rs6000/altivec.h (vec_vaddcuq): Add support for ISA 2.07
3816 128-bit integer builtin support.
3817 (vec_vadduqm): Likewise.
3818 (vec_vaddecuq): Likewise.
3819 (vec_vaddeuqm): Likewise.
3820 (vec_vsubecuq): Likewise.
3821 (vec_vsubeuqm): Likewise.
3822 (vec_vsubcuq): Likewise.
3823 (vec_vsubuqm): Likewise.
3825 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
3826 Document vec_vaddcuq, vec_vadduqm, vec_vaddecuq, vec_vaddeuqm,
3827 vec_subecuq, vec_subeuqm, vec_vsubcuq, vec_vsubeqm builtins adding
3828 128-bit integer add/subtract to ISA 2.07.
3830 2014-03-12 Joern Rennecke <joern.rennecke@embecosm.com>
3832 * config/arc/arc.c (arc_predicate_delay_insns):
3833 Fix third argument passed to conditionalize_nonjump.
3835 2014-03-12 Yufeng Zhang <yufeng.zhang@arm.com>
3837 * config/aarch64/aarch64-builtins.c
3838 (aarch64_builtin_vectorized_function): Add BUILT_IN_LFLOORF,
3839 BUILT_IN_LLFLOOR, BUILT_IN_LCEILF and BUILT_IN_LLCEIL.
3840 * config/aarch64/arm_neon.h (vcvtaq_u64_f64): Call __builtin_llfloor
3841 instead of __builtin_lfloor.
3842 (vcvtnq_u64_f64): Call __builtin_llceil instead of __builtin_lceil.
3844 2014-03-12 Jakub Jelinek <jakub@redhat.com>
3846 * tree-ssa-ifcombine.c (forwarder_block_to): New function.
3847 (tree_ssa_ifcombine_bb_1): New function.
3848 (tree_ssa_ifcombine_bb): Use it. Handle also cases where else_bb
3849 is an empty forwarder block to then_bb or vice versa and then_bb
3850 and else_bb are effectively swapped.
3852 2014-03-12 Christian Bruel <christian.bruel@st.com>
3855 * config/arm/arm.c (arm_emit_vfp_multi_reg_pop): Emit a
3856 REG_CFA_DEF_CFA note.
3857 (arm_expand_epilogue_apcs_frame): call arm_add_cfa_adjust_cfa_note.
3858 (arm_unwind_emit): Allow REG_CFA_DEF_CFA.
3860 2014-03-12 Thomas Preud'homme <thomas.preudhomme@arm.com>
3862 PR tree-optimization/60454
3863 * tree-ssa-math-opts.c (find_bswap_1): Fix bswap detection.
3865 2014-03-12 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3867 * config.gcc (aarch64*-*-*): Use ISA flags from aarch64-arches.def.
3868 Do not define target_cpu_default2 to generic.
3869 * config/aarch64/aarch64.h (TARGET_CPU_DEFAULT): Use generic cpu.
3870 * config/aarch64/aarch64.c (aarch64_override_options): Update comment.
3871 * config/aarch64/aarch64-arches.def (armv8-a): Use generic cpu.
3873 2014-03-12 Jakub Jelinek <jakub@redhat.com>
3874 Marc Glisse <marc.glisse@inria.fr>
3876 PR tree-optimization/60502
3877 * tree-ssa-reassoc.c (eliminate_not_pairs): Use build_all_ones_cst
3878 instead of build_low_bits_mask.
3880 2014-03-12 Jakub Jelinek <jakub@redhat.com>
3883 * tree-vrp.c (register_edge_assert_for_1): Don't add assert
3884 if there are multiple uses, but op doesn't live on E edge.
3885 * tree-cfg.c (assert_unreachable_fallthru_edge_p): Also ignore
3886 clobber stmts before __builtin_unreachable.
3888 2014-03-11 Richard Sandiford <rdsandiford@googlemail.com>
3890 * builtins.c (expand_builtin_setjmp_receiver): Use and clobber
3891 hard_frame_pointer_rtx.
3892 * cse.c (cse_insn): Remove volatile check.
3893 * cselib.c (cselib_process_insn): Likewise.
3894 * dse.c (scan_insn): Likewise.
3896 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
3898 * config/arc/arc.c (conditionalize_nonjump): New function,
3900 (arc_ifcvt): ... this.
3901 (arc_predicate_delay_insns): Use it.
3903 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
3905 * config/arc/predicates.md (extend_operand): During/after reload,
3906 allow const_int_operand.
3907 * config/arc/arc.md (mulsidi3_700): Use extend_operand predicate.
3908 (umulsidi3_700): Likewise. Change operand 2 constraint back to "cL".
3909 (mulsi3_highpart): Change operand 2 constraint alternatives 2 and 3
3911 (umulsi3_highpart_i): Likewise.
3913 2014-03-11 Richard Biener <rguenther@suse.de>
3915 * tree-ssa-structalias.c (get_constraint_for_ptr_offset):
3916 Add asserts to guard possible wrong-code bugs.
3918 2014-03-11 Richard Biener <rguenther@suse.de>
3920 PR tree-optimization/60429
3921 PR tree-optimization/60485
3922 * tree-ssa-structalias.c (set_union_with_increment): Properly
3923 take into account all fields that overlap the shifted vars.
3924 (do_sd_constraint): Likewise.
3925 (do_ds_constraint): Likewise.
3926 (get_constraint_for_ptr_offset): Likewise.
3928 2014-03-11 Chung-Lin Tang <cltang@codesourcery.com>
3930 * config/nios2/nios2.c (machine_function): Add fp_save_offset field.
3931 (nios2_compute_frame_layout):
3932 Add calculation of cfun->machine->fp_save_offset.
3933 (nios2_expand_prologue): Correct setting of frame pointer register
3935 (nios2_expand_epilogue): Update recovery of stack pointer from
3936 frame pointer accordingly.
3937 (nios2_initial_elimination_offset): Update calculation of offset
3938 for eliminating to HARD_FRAME_POINTER_REGNUM.
3940 2014-03-10 Jakub Jelinek <jakub@redhat.com>
3943 * ipa.c (symtab_remove_unreachable_nodes): Don't call
3944 cgraph_get_create_node on VAR_DECLs.
3946 2014-03-10 Richard Biener <rguenther@suse.de>
3949 * tree.c (signed_or_unsigned_type_for): Handle OFFSET_TYPEs.
3951 2014-03-08 Douglas B Rupp <rupp@gnat.com>
3953 * config/vms/vms.opt (vms_float_format): New variable.
3955 2014-03-08 Tobias Burnus <burnus@net-b.de>
3957 * doc/invoke.texi (-fcilkplus): Update implementation status.
3959 2014-03-08 Paulo Matos <paulo@matos-sorge.com>
3960 Richard Biener <rguenther@suse.de>
3962 * lto-wrapper.c (merge_and_complain): Ensure -fshort-double is used
3963 consistently accross all TUs.
3964 (run_gcc): Enable -fshort-double automatically at link at link-time
3965 and disallow override.
3967 2014-03-08 Richard Sandiford <rdsandiford@googlemail.com>
3970 * config/mips/mips.c (mips_option_override): Promote -mpaired-single
3971 warning to an error. Disable TARGET_PAIRED_SINGLE and TARGET_MIPS3D
3972 if they can't be used.
3974 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3976 * configure.ac (HAVE_AS_IX86_TLSLDMPLT): Improve test
3977 for Solaris 11/x86 ld.
3978 * configure: Regenerate.
3980 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3982 * configure.ac (TLS_SECTION_ASM_FLAG): Save as tls_section_flag.
3983 (LIB_TLS_SPEC): Save as ld_tls_libs.
3984 (HAVE_AS_IX86_TLSLDMPLT): Define as 1/0.
3985 (HAVE_AS_IX86_TLSLDM): New test.
3986 * configure, config.in: Regenerate.
3987 * config/i386/i386.c (legitimize_tls_address): Fall back to
3988 TLS_MODEL_GLOBAL_DYNAMIC on 32-bit Solaris/x86 if tool chain
3989 cannot support TLS_MODEL_LOCAL_DYNAMIC.
3990 * config/i386/i386.md (*tls_local_dynamic_base_32_gnu): Use if
3991 instead of #ifdef in HAVE_AS_IX86_TLSLDMPLT test.
3993 2014-03-07 Paulo Matos <paulo@matos-sorge.com>
3995 * common.opt (fira-loop-pressure): Mark as optimization.
3997 2014-03-07 Thomas Schwinge <thomas@codesourcery.com>
3999 * langhooks.c (lhd_omp_mappable_type): The error_mark_node is not
4000 an OpenMP mappable type.
4002 2014-03-06 Matthias Klose <doko@ubuntu.com>
4004 * Makefile.in (s-mlib): Only pass MULTIARCH_DIRNAME if
4005 MULTILIB_OSDIRNAMES is not defined.
4007 2014-03-06 Jakub Jelinek <jakub@redhat.com>
4008 Meador Inge <meadori@codesourcery.com>
4011 * config/arm/arm.c (arm_tls_symbol_p): Remove.
4012 (arm_legitimize_address): Call legitimize_tls_address for any
4013 arm_tls_referenced_p expression, handle constant addend. Call it
4014 before testing for !TARGET_ARM.
4015 (thumb_legitimize_address): Don't handle arm_tls_symbol_p here.
4017 2014-03-06 Richard Biener <rguenther@suse.de>
4023 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
4025 * tree-streamer.c (record_common_node): Assert we don't record
4026 nodes with type double.
4027 (preload_common_node): Skip type double, complex double and double
4028 pointer since it is now frontend dependent due to fshort-double option.
4030 2014-03-06 Richard Biener <rguenther@suse.de>
4032 * gcc.c (PLUGIN_COND): Always enable unless -fno-use-linker-plugin
4033 or -fno-lto is specified and the linker has full plugin support.
4034 * collect2.c (lto_mode): Default to LTO_MODE_WHOPR if LTO is enabled.
4035 (main): Remove -flto processing, adjust lto_mode using use_plugin late.
4036 * lto-wrapper.c (merge_and_complain): Merge compile-time
4037 optimization levels.
4038 (run_gcc): And pass it through to the link options.
4040 2014-03-06 Alexandre Oliva <aoliva@redhat.com>
4044 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4046 * cselib.c (remove_useless_values): Skip to avoid quadratic
4047 behavior if the condition moved from...
4048 (cselib_process_insn): ... here holds.
4050 2014-03-05 Jakub Jelinek <jakub@redhat.com>
4053 * Makefile.in (PLUGIN_HEADERS): Add tree-phinodes.h, stor-layout.h,
4054 ssa-iterators.h, $(RESOURCE_H) and tree-cfgcleanup.h.
4057 * config/i386/t-i386 (OPTIONS_H_EXTRA): Add stringop.def.
4058 (TM_H): Add x86-tune.def.
4060 2014-03-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4062 * config/aarch64/aarch64.c (generic_tunings):
4063 Use cortexa57_extra_costs.
4065 2014-03-05 Jakub Jelinek <jakub@redhat.com>
4068 * cfgexpand.c (expand_used_vars): Do not assume all SSA_NAMEs
4069 of PARM/RESULT_DECLs must be coalesced with optimize && in_lto_p.
4070 * tree-ssa-coalesce.c (coalesce_ssa_name): Use MUST_COALESCE_COST - 1
4073 2014-03-04 Heiher <r@hev.cc>
4075 * config/mips/mips-cpus.def (loongson3a): Mark as a MIPS64r2 processor.
4076 * config/mips/mips.h (MIPS_ISA_LEVEL_SPEC): Adjust accordingly.
4078 2014-03-04 Uros Bizjak <ubizjak@gmail.com>
4080 * config/i386/predicates.md (const2356_operand): Change to ...
4081 (const2367_operand): ... this.
4082 * config/i386/sse.md (avx512pf_scatterpf<mode>sf): Use
4084 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
4085 (*avx512pf_scatterpf<mode>sf): Ditto.
4086 (avx512pf_scatterpf<mode>df): Ditto.
4087 (*avx512pf_scatterpf<mode>df_mask): Ditto.
4088 (*avx512pf_scatterpf<mode>df): Ditto.
4089 * config/i386/i386.c (ix86_expand_builtin): Update
4090 incorrect hint operand error message.
4092 2014-03-04 Richard Biener <rguenther@suse.de>
4094 * lto-section-in.c (lto_get_section_data): Fix const cast.
4096 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
4098 * tree-streamer.c (record_common_node): Assert we don't record
4099 nodes with type double.
4100 (preload_common_node): Skip type double, complex double and double
4101 pointer since it is now frontend dependent due to fshort-double option.
4103 2014-03-04 Richard Biener <rguenther@suse.de>
4106 * lto-streamer-in.c (lto_read_body): Remove LTO bytecode version check.
4107 (lto_input_toplevel_asms): Likewise.
4108 * lto-section-in.c (lto_get_section_data): Instead do it here
4111 2014-03-04 Richard Biener <rguenther@suse.de>
4113 PR tree-optimization/60382
4114 * tree-vect-loop.c (vect_is_simple_reduction_1): Do not consider
4115 dead PHIs a reduction.
4117 2014-03-03 Uros Bizjak <ubizjak@gmail.com>
4119 * config/i386/xmmintrin.h (enum _mm_hint) <_MM_HINT_ET0>: Correct
4121 (_mm_prefetch): Move out of GCC target("sse") pragma.
4122 * config/i386/prfchwintrin.h (_m_prefetchw): Move out of
4123 GCC target("prfchw") pragma.
4124 * config/i386/i386.md (prefetch): Emit prefetchwt1 only
4126 * config/i386/i386.c (ix86_option_override_internal): Enable
4127 -mprfchw with -mprefetchwt1.
4129 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
4131 * config/arc/arc.md (casesi_load) <length attribute alternative 0>:
4134 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
4136 * opts.h (CL_PCH_IGNORE): Define.
4137 * targhooks.c (option_affects_pch_p):
4138 Return false for options that have CL_PCH_IGNORE set.
4139 * opt-functions.awk: Process PchIgnore.
4140 * doc/options.texi: Document PchIgnore.
4142 * config/arc/arc.opt (misize): Add PchIgnore property.
4144 2014-03-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4146 * config/rs6000/rs6000.c (rs6000_preferred_reload_class): Disallow
4147 reload of PLUS rtx's outside of GENERAL_REGS or BASE_REGS; relax
4148 constraint on constants to permit them being loaded into
4149 GENERAL_REGS or BASE_REGS.
4151 2014-03-03 Nick Clifton <nickc@redhat.com>
4153 * config/rl78/rl78-real.md (cbranchsi4_real_signed): Add
4154 anti-cacnonical alternatives.
4155 (negandhi3_real): New pattern.
4156 * config/rl78/rl78-virt.md (negandhi3_virt): New pattern.
4158 2014-03-03 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
4160 * config/avr/avr-mcus.def: Remove atxmega16x1.
4161 * config/avr/avr-tables.opt: Regenerate.
4162 * config/avr/t-multilib: Regenerate.
4163 * doc/avr-mmcu.texi: Regenerate.
4165 2014-03-03 Tobias Grosser <tobias@grosser.es>
4166 Mircea Namolaru <mircea.namolaru@inria.fr>
4168 PR tree-optimization/58028
4169 * graphite-clast-to-gimple.c (set_cloog_options): Don't remove
4172 2014-03-03 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4174 * config/arm/neon.md (*movmisalign<mode>): Legitimize addresses
4175 not handled by recognizers.
4177 2014-03-03 Jakub Jelinek <jakub@redhat.com>
4180 * function.c (expand_function_end): Don't emit
4181 clobber_return_register sequence if clobber_after is a BARRIER.
4182 * cfgexpand.c (construct_exit_block): Append instructions before
4183 return_label to prev_bb.
4185 2014-03-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4187 * config/rs6000/constraints.md: Document reserved use of "wc".
4189 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
4192 * ipa.c (function_and_variable_visibility): When dissolving comdat
4193 group, also set all symbols to local.
4195 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
4200 2013-12-14 Jan Hubicka <jh@suse.cz>
4202 * ipa-prop.c (stmt_may_be_vtbl_ptr_store): Skip clobbers.
4204 2014-03-02 Jon Beniston <jon@beniston.com>
4210 * config/lm32/lm32.c (lm32_legitimate_constant_p): Remove, as incorrect.
4211 (TARGET_LEGITIMATE_CONSTANT_P): Undefine, as not needed.
4212 * config/lm32/lm32.md (movsi_insn): Add 32-bit immediate support.
4213 (simple_return, *simple_return): New patterns
4214 * config/lm32/predicates.md (movsi_rhs_operand): Remove as obsolete.
4215 * configure.ac (force_sjlj_exceptions): Force sjlj exceptions for lm32.
4217 2014-03-01 Paolo Carlini <paolo.carlini@oracle.com>
4219 * dwarf2out.c (gen_subprogram_die): Tidy.
4221 2014-03-01 Oleg Endo <olegendo@gcc.gnu.org>
4224 * config/sh/sh.md (*mov_t_msb_neg): Split into ...
4225 (*mov_t_msb_neg_negc): ... this new insn.
4227 2014-02-28 Jason Merrill <jason@redhat.com>
4230 * ipa-devirt.c (ipa_devirt): Don't choose an implicitly-declared
4233 2014-02-28 Paolo Carlini <paolo.carlini@oracle.com>
4236 * dwarf2out.c (decltype_auto_die): New static.
4237 (gen_subprogram_die): Handle 'decltype(auto)' like 'auto'.
4238 (gen_type_die_with_usage): Handle 'decltype(auto)'.
4239 (is_cxx_auto): Likewise.
4241 2014-02-28 Ian Bolton <ian.bolton@arm.com>
4243 * config/aarch64/aarch64.h: Define __ARM_NEON by default if
4244 we are not using general regs only.
4246 2014-02-28 Richard Biener <rguenther@suse.de>
4249 * tree-cfgcleanup.c (tree_forwarder_block_p): Restrict
4250 previous fix and only allow to remove trivial pre-headers
4251 and latches. Also honor LOOPS_MAY_HAVE_MULTIPLE_LATCHES.
4252 (remove_forwarder_block): Properly update the latch of a loop.
4254 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4257 * cselib.c (cselib_hasher::equal): Special-case VALUE lookup.
4258 (cselib_preserved_hash_table): New.
4259 (preserve_constants_and_equivs): Move preserved vals to it.
4260 (cselib_find_slot): Look it up first.
4261 (cselib_init): Initialize it.
4262 (cselib_finish): Release it.
4263 (dump_cselib_table): Dump it.
4265 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4268 * cselib.c (remove_useless_values): Skip to avoid quadratic
4269 behavior if the condition moved from...
4270 (cselib_process_insn): ... here holds.
4272 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4275 * var-tracking.c (vt_initialize): Apply the same condition to
4276 preserve the CFA base value.
4278 2014-02-28 Joey Ye <joey.ye@arm.com>
4281 * config/arm/arm.c (thumb_far_jump_used_p): Don't change
4282 if reload in progress or completed.
4284 2014-02-28 Tobias Burnus <burnus@net-b.de>
4287 * tree-pretty-print.c (dump_generic_node, print_declaration): Handle
4290 2014-02-27 H.J. Lu <hongjiu.lu@intel.com>
4292 * doc/tm.texi.in (Condition Code Status): Update documention for
4293 relative locations of cc0-setter and cc0-user.
4295 2014-02-27 Jeff Law <law@redhat.com>
4297 PR rtl-optimization/52714
4298 * combine.c (try_combine): When splitting an unrecognized PARALLEL
4299 into two independent simple sets, if I3 is a jump, ensure the
4300 pattern we place into I3 is a (set (pc) ...).
4302 2014-02-27 Mikael Pettersson <mikpe@it.uu.se>
4303 Jeff Law <law@redhat.com>
4305 PR rtl-optimization/49847
4306 * cse.c (fold_rtx) Handle case where cc0 setter and cc0 user
4307 are in different blocks.
4308 * doc/tm.texi (Condition Code Status): Update documention for
4309 relative locations of cc0-setter and cc0-user.
4311 2014-02-27 Vladimir Makarov <vmakarov@redhat.com>
4314 * lra.c (lra_emit_add): Check SUBREG too.
4316 2014-02-27 Andreas Schwab <schwab@suse.de>
4318 * config/m68k/m68k.c (m68k_option_override): Disable
4319 -flive-range-shrinkage for classic m68k.
4320 (m68k_override_options_after_change): Likewise.
4322 2014-02-27 Marek Polacek <polacek@redhat.com>
4325 * tree-ssa-uninit.c (gate_warn_uninitialized): Run the pass even for
4326 -Wmaybe-uninitialized.
4328 2014-02-27 Alan Modra <amodra@gmail.com>
4331 * reload1.c (emit_input_reload_insns): When reload_override_in,
4332 set old to rl->in_reg when rl->in_reg is a subreg.
4334 2014-02-26 Richard Biener <rguenther@suse.de>
4337 * lra-assigns.c (spill_for): Avoid mixed-sign comparison.
4339 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
4341 * common/config/i386/predicates.md (const1256_operand): Remove.
4342 (const2356_operand): New.
4343 (const_1_to_2_operand): Remove.
4344 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
4345 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
4346 (*avx512pf_gatherpf<mode>sf): Ditto.
4347 (avx512pf_gatherpf<mode>df): Ditto.
4348 (*avx512pf_gatherpf<mode>df_mask): Ditto.
4349 (*avx512pf_gatherpf<mode>df): Ditto.
4350 (avx512pf_scatterpf<mode>sf): Ditto.
4351 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
4352 (*avx512pf_scatterpf<mode>sf): Ditto.
4353 (avx512pf_scatterpf<mode>df): Ditto.
4354 (*avx512pf_scatterpf<mode>df_mask): Ditto.
4355 (*avx512pf_scatterpf<mode>df): Ditto.
4356 * common/config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET0.
4358 2014-02-26 Ilya Tocar <ilya.tocar@intel.com>
4360 * config/i386/avx512fintrin.h (_mm512_testn_epi32_mask),
4361 (_mm512_mask_testn_epi32_mask), (_mm512_testn_epi64_mask),
4362 (_mm512_mask_testn_epi64_mask): Move to ...
4363 * config/i386/avx512cdintrin.h: Here.
4364 * config/i386/i386.c (bdesc_args): Change MASK_ISA for testnm.
4365 * config/i386/sse.md (avx512f_vmscalef<mode><round_name>): Remove %.
4366 (avx512f_scalef<mode><mask_name><round_name>): Ditto.
4367 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Change conditon to
4368 TARGET_AVX512F from TARGET_AVX512CD.
4370 2014-02-26 Richard Biener <rguenther@suse.de>
4373 * ipa.c (walk_polymorphic_call_targets): Properly guard
4374 call to inline_update_overall_summary.
4376 2014-02-26 Bin Cheng <bin.cheng@arm.com>
4379 * tree-cfgcleanup.c (tree_forwarder_block_p): Protect loop preheaders
4380 and latches only if requested. Fix latch if it is removed.
4381 * tree-ssa-dom.c (tree_ssa_dominator_optimize): Set
4382 LOOPS_HAVE_PREHEADERS.
4384 2014-02-25 Andrew Pinski <apinski@cavium.com>
4386 * builtins.c (expand_builtin_thread_pointer): Create a new target
4387 when the target is NULL.
4389 2014-02-25 Vladimir Makarov <vmakarov@redhat.com>
4391 PR rtl-optimization/60317
4392 * params.def (PARAM_LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
4393 * params.h (LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
4394 * lra-assigns.c: Include params.h.
4395 (spill_for): Use LRA_MAX_CONSIDERED_RELOAD_PSEUDOS as guard for
4396 other reload pseudos considerations.
4398 2014-02-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4400 * config/rs6000/vector.md (*vector_unordered<mode>): Change split
4401 to use canonical form for nor<mode>3.
4403 2014-02-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4406 * config/arm/arm.h (CANNOT_CHANGE_MODE_CLASS): Allow 128 to 64-bit
4409 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
4411 * common/config/i386/i386-common.c (OPTION_MASK_ISA_PREFETCHWT1_SET),
4412 (OPTION_MASK_ISA_PREFETCHWT1_UNSET): New.
4413 (ix86_handle_option): Handle OPT_mprefetchwt1.
4414 * config/i386/cpuid.h (bit_PREFETCHWT1): New.
4415 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
4417 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
4418 OPTION_MASK_ISA_PREFETCHWT1.
4419 * config/i386/i386.c (ix86_target_string): Handle mprefetchwt1.
4420 (PTA_PREFETCHWT1): New.
4421 (ix86_option_override_internal): Handle PTA_PREFETCHWT1.
4422 (ix86_valid_target_attribute_inner_p): Handle OPT_mprefetchwt1.
4423 * config/i386/i386.h (TARGET_PREFETCHWT1, TARGET_PREFETCHWT1_P): New.
4424 * config/i386/i386.md (prefetch): Check TARGET_PREFETCHWT1
4425 (*prefetch_avx512pf_<mode>_: Change into ...
4426 (*prefetch_prefetchwt1_<mode>: This.
4427 * config/i386/i386.opt (mprefetchwt1): New.
4428 * config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET1.
4429 (_mm_prefetch): Handle intent to write.
4430 * doc/invoke.texi (mprefetchwt1), (mno-prefetchwt1): Doccument.
4432 2014-02-25 Richard Biener <rguenther@suse.de>
4435 * emit-rtl.c (mem_attrs_htab): Remove.
4436 (mem_attrs_htab_hash): Likewise.
4437 (mem_attrs_htab_eq): Likewise.
4438 (set_mem_attrs): Always allocate new mem-attrs when something changed.
4439 (init_emit_once): Do not allocate mem_attrs_htab.
4441 2014-02-25 Richard Biener <rguenther@suse.de>
4444 * lto-opts.c (lto_write_options): Output non-explicit conservative
4445 -fwrapv, -fno-trapv and -fno-strict-overflow.
4446 * lto-wrapper.c (merge_and_complain): Handle merging those options.
4447 (run_gcc): And pass them through.
4449 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
4451 * sel-sched.c (calculate_new_fences): New parameter ptime.
4452 Calculate it as a maximum over all fence cycles.
4453 (sel_sched_region_2): Adjust the call to calculate_new_fences.
4454 Print the final schedule timing when sched_verbose.
4456 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
4458 PR rtl-optimization/60292
4459 * sel-sched.c (fill_vec_av_set): Do not reset target availability
4460 bit fot the fence instruction.
4462 2014-02-24 Alangi Derick <alangiderick@gmail.com>
4464 * calls.h: Fix typo in comment.
4466 2014-02-24 John David Anglin <danglin@gcc.gnu.org>
4468 * config/pa/pa.c (pa_output_move_double): Don't valididate when
4469 adjusting offsetable addresses.
4471 2014-02-24 Guozhi Wei <carrot@google.com>
4473 * sparseset.h (sparseset_pop): Fix the wrong index.
4475 2014-02-24 Walter Lee <walt@tilera.com>
4477 * config.gcc (tilepro-*-*): Change to tilepro*-*-*.
4478 (tilegx-*-linux*): Change to tilegx*-*-linux*; Support tilegxbe
4480 * common/config/tilegx/tilegx-common.c
4481 (TARGET_DEFAULT_TARGET_FLAGS): Define.
4482 * config/tilegx/linux.h (ASM_SPEC): Add endian_spec.
4484 * config/tilegx/sync.md (atomic_test_and_set): Handle big endian.
4485 * config/tilegx/tilegx.c (tilegx_return_in_msb): New.
4486 (tilegx_gimplify_va_arg_expr): Handle big endian.
4487 (tilegx_expand_unaligned_load): Ditto.
4488 (tilegx_expand_unaligned_store): Ditto.
4489 (TARGET_RETURN_IN_MSB): New.
4490 * config/tilegx/tilegx.h (TARGET_DEFAULT): New.
4491 (TARGET_ENDIAN_DEFAULT): New.
4492 (TARGET_BIG_ENDIAN): Handle big endian.
4493 (BYTES_BIG_ENDIAN): Ditto.
4494 (WORDS_BIG_ENDIAN): Ditto.
4495 (FLOAT_WORDS_BIG_ENDIAN): Ditto.
4498 * config/tilegx/tilegx.md (extv): Handle big endian.
4500 (insn_st<n>): Ditto.
4501 (insn_st<n>_add<bitsuffix>): Ditto.
4502 (insn_stnt<n>): Ditto.
4503 (insn_stnt<n>_add<bitsuffix>):Ditto.
4504 (vec_interleave_highv8qi): Handle big endian.
4505 (vec_interleave_highv8qi_be): New.
4506 (vec_interleave_highv8qi_le): New.
4507 (insn_v1int_h): Handle big endian.
4508 (vec_interleave_lowv8qi): Handle big endian.
4509 (vec_interleave_lowv8qi_be): New.
4510 (vec_interleave_lowv8qi_le): New.
4511 (insn_v1int_l): Handle big endian.
4512 (vec_interleave_highv4hi): Handle big endian.
4513 (vec_interleave_highv4hi_be): New.
4514 (vec_interleave_highv4hi_le): New.
4515 (insn_v2int_h): Handle big endian.
4516 (vec_interleave_lowv4hi): Handle big endian.
4517 (vec_interleave_lowv4hi_be): New.
4518 (vec_interleave_lowv4hi_le): New.
4519 (insn_v2int_l): Handle big endian.
4520 (vec_interleave_highv2si): Handle big endian.
4521 (vec_interleave_highv2si_be): New.
4522 (vec_interleave_highv2si_le): New.
4523 (insn_v4int_h): Handle big endian.
4524 (vec_interleave_lowv2si): Handle big endian.
4525 (vec_interleave_lowv2si_be): New.
4526 (vec_interleave_lowv2si_le): New.
4527 (insn_v4int_l): Handle big endian.
4528 * config/tilegx/tilegx.opt (mbig-endian): New option.
4529 (mlittle-endian): New option.
4530 * doc/install.texi: Document tilegxbe-linux.
4531 * doc/invoke.texi: Document -mbig-endian and -mlittle-endian.
4533 2014-02-24 Martin Jambor <mjambor@suse.cz>
4536 * ipa-cp.c (propagate_constants_accross_call): Bail out early if
4537 there are no parameter descriptors.
4539 2014-02-24 Andrey Belevantsev <abel@ispras.ru>
4541 PR rtl-optimization/60268
4542 * sched-rgn.c (haifa_find_rgns): Move the nr_regions_initial variable
4543 initialization to ...
4544 (sched_rgn_init): ... here.
4545 (schedule_region): Check for SCHED_PRESSURE_NONE earlier.
4547 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
4549 * config/microblaze/microblaze.md: Correct ashrsi_reg / lshrsi_reg
4552 2014-02-23 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
4554 * config/microblaze/microblaze.h: Remove SECONDARY_MEMORY_NEEDED
4557 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
4559 * /config/microblaze/microblaze.c: Add microblaze_asm_output_mi_thunk,
4560 define TARGET_ASM_OUTPUT_MI_THUNK and TARGET_ASM_CAN_OUTPUT_MI_THUNK.
4562 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
4564 * config/microblaze/predicates.md: Add cmp_op predicate.
4565 * config/microblaze/microblaze.md: Add branch_compare instruction
4566 which uses cmp_op predicate and emits cmp insn before branch.
4567 * config/microblaze/microblaze.c (microblaze_emit_compare): Rename
4568 to microblaze_expand_conditional_branch and consolidate logic.
4569 (microblaze_expand_conditional_branch): emit branch_compare
4570 insn instead of handling cmp op separate from branch insn.
4572 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4574 * config/rs6000/rs6000.c (rs6000_emit_le_vsx_move): Relax assert
4577 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4579 * config/rs6000/altivec.md (altivec_lve<VI_char>x): Replace
4580 define_insn with define_expand and new define_insn
4581 *altivec_lve<VI_char>x_internal.
4582 (altivec_stve<VI_char>x): Replace define_insn with define_expand
4583 and new define_insn *altivec_stve<VI_char>x_internal.
4584 * config/rs6000/rs6000-protos.h (altivec_expand_stvex_be): New
4586 * config/rs6000/rs6000.c (altivec_expand_lvx_be): Document use by
4588 (altivec_expand_stvex_be): New function.
4590 2014-02-22 Joern Rennecke <joern.rennecke@embecosm.com>
4592 * config/avr/avr.c (avr_can_eliminate): Allow elimination from
4593 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM if !frame_pointer_needed.
4594 * config/avr/avr.c (ELIMINABLE_REGS): Add elimination from
4595 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM.
4597 2014-02-21 Vladimir Makarov <vmakarov@redhat.com>
4600 * lra-constraints.c (inherit_reload_reg): Use lra_emit_move
4601 instead of emit_move_insn.
4603 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4605 * config/rs6000/altivec.md (altivec_vsumsws): Replace second
4607 (reduc_uplus_v16qi): Use gen_altivec_vsumsws_direct instead of
4608 gen_altivec_vsumsws.
4610 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4612 * config/rs6000/altivec.md (altivec_lvxl): Rename as
4613 *altivec_lvxl_<mode>_internal and use VM2 iterator instead of V4SI.
4614 (altivec_lvxl_<mode>): New define_expand incorporating
4615 -maltivec=be semantics where needed.
4616 (altivec_lvx): Rename as *altivec_lvx_<mode>_internal.
4617 (altivec_lvx_<mode>): New define_expand incorporating -maltivec=be
4618 semantics where needed.
4619 (altivec_stvx): Rename as *altivec_stvx_<mode>_internal.
4620 (altivec_stvx_<mode>): New define_expand incorporating
4621 -maltivec=be semantics where needed.
4622 (altivec_stvxl): Rename as *altivec_stvxl_<mode>_internal and use
4623 VM2 iterator instead of V4SI.
4624 (altivec_stvxl_<mode>): New define_expand incorporating
4625 -maltivec=be semantics where needed.
4626 * config/rs6000/rs6000-builtin.def: Add new built-in definitions
4627 LVXL_V2DF, LVXL_V2DI, LVXL_V4SF, LVXL_V4SI, LVXL_V8HI, LVXL_V16QI,
4628 LVX_V2DF, LVX_V2DI, LVX_V4SF, LVX_V4SI, LVX_V8HI, LVX_V16QI, STVX_V2DF,
4629 STVX_V2DI, STVX_V4SF, STVX_V4SI, STVX_V8HI, STVX_V16QI, STVXL_V2DF,
4630 STVXL_V2DI, STVXL_V4SF, STVXL_V4SI, STVXL_V8HI, STVXL_V16QI.
4631 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Replace
4632 ALTIVEC_BUILTIN_LVX with ALTIVEC_BUILTIN_LVX_<MODE> throughout;
4633 similarly for ALTIVEC_BUILTIN_LVXL, ALTIVEC_BUILTIN_STVX, and
4634 ALTIVEC_BUILTIN_STVXL.
4635 * config/rs6000/rs6000-protos.h (altivec_expand_lvx_be): New prototype.
4636 (altivec_expand_stvx_be): Likewise.
4637 * config/rs6000/rs6000.c (swap_selector_for_mode): New function.
4638 (altivec_expand_lvx_be): Likewise.
4639 (altivec_expand_stvx_be): Likewise.
4640 (altivec_expand_builtin): Add cases for
4641 ALTIVEC_BUILTIN_STVX_<MODE>, ALTIVEC_BUILTIN_STVXL_<MODE>,
4642 ALTIVEC_BUILTIN_LVXL_<MODE>, and ALTIVEC_BUILTIN_LVX_<MODE>.
4643 (altivec_init_builtins): Add definitions for
4644 __builtin_altivec_lvxl_<mode>, __builtin_altivec_lvx_<mode>,
4645 __builtin_altivec_stvx_<mode>, and __builtin_altivec_stvxl_<mode>.
4647 2014-02-21 Catherine Moore <clm@codesourcery.com>
4649 * doc/invoke.texi (mvirt, mno-virt): Document.
4650 * config/mips/mips.opt (mvirt): New option.
4651 * config/mips/mips.h (ASM_SPEC): Pass mvirt to the assembler.
4653 2014-02-21 Richard Biener <rguenther@suse.de>
4655 PR tree-optimization/60276
4656 * tree-vectorizer.h (struct _stmt_vec_info): Add min_neg_dist field.
4657 (STMT_VINFO_MIN_NEG_DIST): New macro.
4658 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Record
4659 STMT_VINFO_MIN_NEG_DIST.
4660 * tree-vect-stmts.c (vectorizable_load): Verify if assumptions
4661 made for negative dependence distances still hold.
4663 2014-02-21 Richard Biener <rguenther@suse.de>
4666 * tree-ssa-live.c (mark_all_vars_used_1): Do not walk
4667 DECL_INITIAL for globals not in the current function context.
4669 2014-02-21 Jakub Jelinek <jakub@redhat.com>
4671 PR tree-optimization/56490
4672 * params.def (PARAM_UNINIT_CONTROL_DEP_ATTEMPTS): New param.
4673 * tree-ssa-uninit.c: Include params.h.
4674 (compute_control_dep_chain): Add num_calls argument, return false
4675 if it exceed PARAM_UNINIT_CONTROL_DEP_ATTEMPTS param, pass
4676 num_calls to recursive call.
4677 (find_predicates): Change dep_chain into normal array,
4678 cur_chain into auto_vec<edge, MAX_CHAIN_LEN + 1>, add num_calls
4679 variable and adjust compute_control_dep_chain caller.
4680 (find_def_preds): Likewise.
4682 2014-02-21 Thomas Schwinge <thomas@codesourcery.com>
4684 * gimple-pretty-print.c (dump_gimple_omp_for) [flags & TDF_RAW]
4685 <case GF_OMP_FOR_KIND_CILKSIMD>: Add missing break statement.
4687 2014-02-21 Nick Clifton <nickc@redhat.com>
4689 * config/stormy16/stormy16.md (pushdqi1): Add mode to post_inc.
4690 (pushhi1): Likewise.
4691 (popqi1): Add mode to pre_dec.
4694 2014-02-21 Jakub Jelinek <jakub@redhat.com>
4696 * config/i386/i386.c (ix86_expand_vec_perm): Use V8SImode
4697 mode for mask of V8SFmode permutation.
4699 2014-02-20 Richard Henderson <rth@redhat.com>
4702 * builtins.c (expand_builtin_atomic_compare_exchange): Always make
4703 a new pseudo for OLDVAL.
4705 2014-02-20 Jakub Jelinek <jakub@redhat.com>
4708 * config/i386/i386.c (expand_vec_perm_interleave2): Don't call
4709 gen_reg_rtx if d->testing_p.
4710 (expand_vec_perm_pshufb2, expand_vec_perm_broadcast_1): Return early
4711 if d->testing_p and we will certainly return true.
4712 (expand_vec_perm_even_odd_1): Likewise. Don't call gen_reg_rtx
4715 2014-02-20 Uros Bizjak <ubizjak@gmail.com>
4717 * emit-rtl.c (gen_reg_rtx): Assert that
4718 crtl->emit.regno_pointer_align_length is non-zero.
4720 2014-02-20 Richard Henderson <rth@redhat.com>
4723 * builtins.c (expand_builtin_atomic_compare_exchange): Conditionalize
4724 on failure the store back into EXPECT.
4726 2014-02-20 Chung-Lin Tang <cltang@codesourcery.com>
4727 Sandra Loosemore <sandra@codesourcery.com>
4729 * config/nios2/nios2.md (unspec): Add UNSPEC_PIC_GOTOFF_SYM enum.
4730 * config/nios2/nios2.c (nios2_function_profiler): Add
4731 -fPIC (flag_pic == 2) support.
4732 (nios2_handle_custom_fpu_cfg): Fix warning parameter.
4733 (nios2_large_offset_p): New function.
4734 (nios2_unspec_reloc_p): Move up position, update to use
4735 nios2_large_offset_p.
4736 (nios2_unspec_address): Remove function.
4737 (nios2_unspec_offset): New function.
4738 (nios2_large_got_address): New function.
4739 (nios2_got_address): Add large offset support.
4740 (nios2_legitimize_tls_address): Update usage of removed and new
4742 (nios2_symbol_binds_local_p): New function.
4743 (nios2_load_pic_address): Add -fPIC (flag_pic == 2) support.
4744 (nios2_legitimize_address): Update to use nios2_large_offset_p.
4745 (nios2_emit_move_sequence): Avoid legitimizing (const (unspec ...)).
4746 (nios2_print_operand): Merge H/L processing, add hiadj/lo
4747 processing for (const (unspec ...)).
4748 (nios2_unspec_reloc_name): Add UNSPEC_PIC_GOTOFF_SYM case.
4750 2014-02-20 Richard Biener <rguenther@suse.de>
4752 * tree-cfg.c (replace_uses_by): Mark altered BBs before
4753 doing the substitution.
4754 (verify_gimple_assign_single): Also verify bare MEM_REFs on the lhs.
4756 2014-02-20 Martin Jambor <mjambor@suse.cz>
4759 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Uce correct
4760 info when checking whether lattices are bottom.
4762 2014-02-20 Richard Biener <rguenther@suse.de>
4765 * tree-eh.c (execute_cleanup_eh_1): Also cleanup empty EH
4768 2014-02-20 Jan Hubicka <hubicka@ucw.cz>
4771 * ipa-inline-transform.c (clone_inlined_nodes): Add freq_scale
4772 parameter specifying the scaling.
4773 (inline_call): Update.
4774 (want_inline_recursively): Guard division by zero.
4775 (recursive_inlining): Update.
4776 * ipa-inline.h (clone_inlined_nodes): Update.
4778 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
4781 * config/i386/i386.c (classify_argument): Pass structures of size
4782 64 bytes or less in register.
4784 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
4785 Kirill Yukhin <kirill.yukhin@intel.com>
4787 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): Swap operands.
4788 (_mm_rcp28_round_ss): Ditto.
4789 (_mm_rsqrt28_round_sd): Ditto.
4790 (_mm_rsqrt28_round_ss): Ditto.
4791 * config/i386/avx512erintrin.h (_mm_rcp14_round_sd): Ditto.
4792 (_mm_rcp14_round_ss): Ditto.
4793 (_mm_rsqrt14_round_sd): Ditto.
4794 (_mm_rsqrt14_round_ss): Ditto.
4795 * config/i386/sse.md (rsqrt14<mode>): Put nonimmediate operand as
4796 the first input operand, get rid of match_dup.
4797 (avx512er_exp2<mode><mask_name><round_saeonly_name>): Set type
4799 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
4801 (avx512er_vmrcp28<mode><round_saeonly_name>): Put nonimmediate
4802 operand as the first input operand, set type attribute.
4803 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
4805 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Put nonimmediate
4806 operand as the first input operand, set type attribute.
4808 2014-02-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4810 * config/rs6000/rs6000.c (vspltis_constant): Fix most significant
4813 2014-02-19 H.J. Lu <hongjiu.lu@intel.com>
4816 * config/i386/i386.c (construct_container): Remove TFmode check
4817 for X86_64_INTEGER_CLASS.
4819 2014-02-19 Uros Bizjak <ubizjak@gmail.com>
4822 * config/i386/i386.c (type_natural_mode): Warn for ABI changes
4823 only when -Wpsabi is enabled.
4825 2014-02-19 Michael Hudson-Doyle <michael.hudson@linaro.org>
4828 * config/aarch64/aarch64.c (aarch64_pass_by_reference): The rules for
4829 passing arrays in registers are the same as for structs, so remove the
4830 special case for them.
4832 2014-02-19 Eric Botcazou <ebotcazou@adacore.com>
4834 * expr.c (expand_expr_real_1) <case VIEW_CONVERT_EXPR>: For a bit-field
4835 destination type, extract only the valid bits if the source type is not
4836 integral and has a different mode.
4838 2014-02-19 Richard Biener <rguenther@suse.de>
4841 * tree-inline.c (estimate_num_insns): Avoid calling cgraph_get_node
4844 2014-02-19 Richard Biener <rguenther@suse.de>
4847 * ipa-prop.c: Include stringpool.h and tree-ssanames.h.
4848 (ipa_modify_call_arguments): Emit an argument load explicitely and
4849 preserve virtual SSA form there and for the replacement call.
4850 Do not update SSA form nor free dominance info.
4852 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
4854 * ipa.c (function_and_variable_visibility): Also clear WEAK
4855 flag when disolving COMDAT_GROUP.
4857 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
4859 * ipa-prop.h (ipa_ancestor_jf_data): Update ocmment.
4860 * ipa-prop.c (ipa_set_jf_known_type): Return early when
4862 (ipa_set_ancestor_jf): Set type to NULL hwen it is not preserved;
4863 do more sanity checks.
4864 (detect_type_change): Return true when giving up early.
4865 (compute_complex_assign_jump_func): Fix type parameter of
4866 ipa_set_ancestor_jf.
4867 (compute_complex_ancestor_jump_func): Likewise.
4868 (update_jump_functions_after_inlining): Fix updating of
4870 * ipa-cp.c (ipa_get_jf_ancestor_result): Be ready for type to be NULL.
4872 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
4874 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Also remove
4875 inline clones when edge disappears.
4877 2014-02-18 Michael Meissner <meissner@linux.vnet.ibm.com>
4880 * config/rs6000/rs6000.md (mov<mode>_64bit, TF/TDmode moves):
4881 Split 64-bit moves into 2 patterns. Do not allow the use of
4882 direct move for TDmode in little endian, since the decimal value
4883 has little endian bytes within a word, but the 64-bit pieces are
4884 ordered in a big endian fashion, and normal subreg's of TDmode are
4886 (mov<mode>_64bit_dm): Likewise.
4887 (movtd_64bit_nodm): Likewise.
4889 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
4891 PR tree-optimization/60174
4892 * tree-ssa-reassoc.c (init_range_entry): Do not look into the defining
4893 statement of an SSA_NAME that occurs in an abnormal PHI node.
4895 2014-02-18 Jakub Jelinek <jakub@redhat.com>
4898 * final.c (SEEN_BB): Remove.
4899 (SEEN_NOTE, SEEN_EMITTED): Renumber.
4900 (final_scan_insn): Don't force_source_line on second
4901 NOTE_INSN_BASIC_BLOCK.
4903 2014-02-18 Uros Bizjak <ubizjak@gmail.com>
4906 * config/i386/i386.h (struct ix86_args): Add warn_avx512f.
4907 * config/i386/i386.c (init_cumulative_args): Initialize warn_avx512f.
4908 (type_natural_mode): Warn ABI change when %zmm register is not
4909 available for AVX512F vector value passing.
4911 2014-02-18 Kai Tietz <ktietz@redhat.com>
4914 * config/i386/i386.c (ix86_expand_prologue): Use value in
4915 rax register as displacement when restoring %r10 or %rax.
4916 Fix wrong offset when restoring both registers.
4918 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
4920 * ipa-prop.c (compute_complex_ancestor_jump_func): Replace overzealous
4921 assertion with conditional return.
4923 2014-02-18 Jakub Jelinek <jakub@redhat.com>
4924 Uros Bizjak <ubizjak@gmail.com>
4927 * config/i386/driver-i386.c (host_detect_local_cpu): If
4928 YMM state is not saved by the OS, also clear has_f16c. Move
4929 CPUID 0x80000001 handling before YMM state saving checking.
4931 2014-02-18 Andrey Belevantsev <abel@ispras.ru>
4933 PR rtl-optimization/58960
4934 * haifa-sched.c (alloc_global_sched_pressure_data): New,
4935 factored out from ...
4936 (sched_init): ... here.
4937 (free_global_sched_pressure_data): New, factored out from ...
4938 (sched_finish): ... here.
4939 * sched-int.h (free_global_sched_pressure_data): Declare.
4940 * sched-rgn.c (nr_regions_initial): New static global.
4941 (haifa_find_rgns): Initialize it.
4942 (schedule_region): Disable sched-pressure for the newly
4945 2014-02-17 Richard Biener <rguenther@suse.de>
4947 * tree-vect-stmts.c (free_stmt_vec_info): Clear BB and
4948 release SSA defs of pattern stmts.
4950 2014-02-17 Richard Biener <rguenther@suse.de>
4952 * tree-inline.c (expand_call_inline): Release the virtual
4953 operand defined by the call we are about to inline.
4955 2014-02-17 Richard Biener <rguenther@suse.de>
4957 * tree-ssa.c (verify_ssa): If verify_def found an error, ICE.
4959 2014-02-17 Kirill Yukhin <kirill.yukhin@intel.com>
4960 Ilya Tocar <ilya.tocar@intel.com>
4962 * config/i386/avx512fintrin.h (_mm512_maskz_permutexvar_epi64): Swap
4963 arguments order in builtin.
4964 (_mm512_permutexvar_epi64): Ditto.
4965 (_mm512_mask_permutexvar_epi64): Ditto
4966 (_mm512_maskz_permutexvar_epi32): Ditto
4967 (_mm512_permutexvar_epi32): Ditto
4968 (_mm512_mask_permutexvar_epi32): Ditto
4970 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4972 * config/rs6000/altivec.md (p8_vmrgew): Handle little endian targets.
4973 (p8_vmrgow): Likewise.
4975 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4977 * config/rs6000/vsx.md (vsx_xxpermdi_<mode>): Handle little
4980 2014-02-15 Michael Meissner <meissner@linux.vnet.ibm.com>
4983 * config/rs6000/rs6000.md (rreg): Add TFmode, TDmode constraints.
4984 (mov<mode>_internal, TFmode/TDmode): Split TFmode/TDmode moves
4985 into 64-bit and 32-bit moves. On 64-bit moves, add support for
4986 using direct move instructions on ISA 2.07. Also adjust
4987 instruction length for 64-bit.
4988 (mov<mode>_64bit, TFmode/TDmode): Likewise.
4989 (mov<mode>_32bit, TFmode/TDmode): Likewise.
4991 2014-02-15 Alan Modra <amodra@gmail.com>
4995 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
4996 find_replacement on parts of insn rtl that might be reloaded.
4998 2014-02-15 Richard Biener <rguenther@suse.de>
5000 PR tree-optimization/60183
5001 * tree-ssa-phiprop.c (propagate_with_phi): Avoid speculating loads.
5002 (tree_ssa_phiprop): Calculate and free post-dominators.
5004 2014-02-14 Jeff Law <law@redhat.com>
5006 PR rtl-optimization/60131
5007 * ree.c (get_extended_src_reg): New function.
5008 (combine_reaching_defs): Use it rather than assuming location of REG.
5009 (find_and_remove_re): Verify first operand of extension is
5010 a REG before adding the insns to the copy list.
5012 2014-02-14 Roland McGrath <mcgrathr@google.com>
5014 * configure.ac (HAVE_AS_IX86_UD2): New test for 'ud2' mnemonic.
5015 * configure: Regenerated.
5016 * config.in: Regenerated.
5017 * config/i386/i386.md (trap) [HAVE_AS_IX86_UD2]: Use the mnemonic
5018 instead of ASM_SHORT.
5020 2014-02-14 Vladimir Makarov <vmakarov@redhat.com>
5021 Richard Earnshaw <rearnsha@arm.com>
5023 PR rtl-optimization/59535
5024 * lra-constraints.c (process_alt_operands): Encourage alternative
5025 when unassigned pseudo class is superset of the alternative class.
5026 (inherit_reload_reg): Don't inherit when optimizing for code size.
5027 * config/arm/arm.h (MODE_BASE_REG_CLASS): Add version for LRA
5028 returning CORE_REGS for anything but Thumb1 and BASE_REGS for
5029 modes not less than 4 for Thumb1.
5031 2014-02-14 Kyle McMartin <kyle@redhat.com>
5034 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Define for AArch64.
5036 2014-02-14 Richard Biener <rguenther@suse.de>
5038 * cilk-common.c (cilk_arrow): Build a MEM_REF, not an INDIRECT_REF.
5039 (get_frame_arg): Drop the assert with langhook types_compatible_p.
5040 Do not strip INDIRECT_REFs.
5042 2014-02-14 Richard Biener <rguenther@suse.de>
5045 * lto-streamer-out.c (DFS_write_tree_body): Do not follow
5046 DECL_FUNCTION_SPECIFIC_TARGET.
5047 (hash_tree): Do not hash DECL_FUNCTION_SPECIFIC_TARGET.
5048 * tree-streamer-out.c (pack_ts_target_option): Remove.
5049 (streamer_pack_tree_bitfields): Do not stream TS_TARGET_OPTION.
5050 (write_ts_function_decl_tree_pointers): Do not stream
5051 DECL_FUNCTION_SPECIFIC_TARGET.
5052 * tree-streamer-in.c (unpack_ts_target_option): Remove.
5053 (unpack_value_fields): Do not stream TS_TARGET_OPTION.
5054 (lto_input_ts_function_decl_tree_pointers): Do not stream
5055 DECL_FUNCTION_SPECIFIC_TARGET.
5057 2014-02-14 Jakub Jelinek <jakub@redhat.com>
5059 * tree-vect-loop.c (vect_is_slp_reduction): Don't set use_stmt twice.
5060 (get_initial_def_for_induction, vectorizable_induction): Ignore
5061 debug stmts when looking for exit_phi.
5062 (vectorizable_live_operation): Fix up condition.
5064 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
5066 * config/nds32/nds32.c (nds32_asm_function_prologue): Do not use
5067 nreverse() because it changes the content of original tree list.
5069 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
5071 * config/nds32/t-mlibs (MULTILIB_OPTIONS): Fix typo in comment.
5072 * config/nds32/nds32.c (nds32_merge_decl_attributes): Likewise.
5074 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
5076 * config/nds32/nds32.c (nds32_naked_function_p): Follow the
5077 GNU coding standards.
5079 2014-02-13 Jakub Jelinek <jakub@redhat.com>
5082 * dwarf2out.c (gen_subprogram_die): Don't call
5083 add_calling_convention_attribute if subr_die is old_die.
5085 2014-02-13 Sharad Singhai <singhai@google.com>
5087 * doc/optinfo.texi: Fix order of nodes.
5089 2014-02-13 Uros Bizjak <ubizjak@gmail.com>
5091 * config/i386/sse.md (xop_vmfrcz<mode>2): Generate const0 in
5092 operands[2], not operands[3].
5094 2014-02-13 Richard Biener <rguenther@suse.de>
5097 * doc/install.texi (ISL): Update recommended version to 0.12.2,
5098 mention the possibility of an in-tree build.
5099 (CLooG): Update recommended version to 0.18.1, mention the
5100 possibility of an in-tree build and clarify that the ISL
5101 bundled with CLooG does not work.
5103 2014-02-13 Jakub Jelinek <jakub@redhat.com>
5106 * expr.c (compress_float_constant): If x is a hard register,
5107 extend into a pseudo and then move to x.
5109 2014-02-13 Dominik Vogt <vogt@linux.vnet.ibm.com>
5111 * config/s390/s390.c (s390_asm_output_function_label): Fix crash
5112 caused by bad second argument to warning_at() with -mhotpatch and
5113 nested functions (e.g. with gfortran).
5115 2014-02-13 Richard Sandiford <rdsandiford@googlemail.com>
5117 * opts.c (option_name): Remove "enabled by default" rider.
5119 2014-02-12 John David Anglin <danglin@gcc.gnu.org>
5121 * config/pa/pa.c (pa_option_override): Remove auto increment FIXME.
5123 2014-02-12 H.J. Lu <hongjiu.lu@intel.com>
5124 Uros Bizjak <ubizjak@gmail.com>
5127 * configure.ac (HAVE_AS_GOTOFF_IN_DATA): Pass --32 to GNU assembler.
5128 * configure: Regenerated.
5130 2014-02-12 Richard Biener <rguenther@suse.de>
5132 * vec.c (vec_prefix::calculate_allocation): Move as
5133 inline variant to vec.h.
5134 (vec_prefix::calculate_allocation_1): New out-of-line version.
5135 * vec.h (vec_prefix::calculate_allocation_1): Declare.
5136 (vec_prefix::m_has_auto_buf): Rename to ...
5137 (vec_prefix::m_using_auto_storage): ... this.
5138 (vec_prefix::calculate_allocation): Inline the easy cases
5139 and dispatch to calculate_allocation_1 which doesn't need the
5141 (va_heap::reserve): Use gcc_checking_assert.
5142 (vec<T, A, vl_embed>::embedded_init): Add argument to initialize
5143 m_using_auto_storage.
5144 (auto_vec): Change m_vecpfx member to a vec<T, va_heap, vl_embed>
5146 (vec<T, va_heap, vl_ptr>::reserve): Remove redundant check.
5147 (vec<T, va_heap, vl_ptr>::release): Avoid casting.
5148 (vec<T, va_heap, vl_ptr>::using_auto_storage): Simplify.
5150 2014-02-12 Richard Biener <rguenther@suse.de>
5152 * gcse.c (compute_transp): break from loop over canon_modify_mem_list
5153 when we found a dependence.
5155 2014-02-12 Thomas Schwinge <thomas@codesourcery.com>
5157 * gimplify.c (gimplify_call_expr, gimplify_modify_expr): Move
5159 (maybe_fold_stmt): ... into this new function.
5160 * omp-low.c (lower_omp): Update comment.
5162 * omp-low.c (lower_omp_target): Add clobber for sizes array, after
5165 * omp-low.c (diagnose_sb_0): Make sure label_ctx is valid to
5168 2014-02-12 James Greenhalgh <james.greenhalgh@arm.com>
5170 * config/arm/aarch-cost-tables.h (generic_extra_costs): Fix
5171 identifiers in comments.
5172 (cortexa53_extra_costs): Likewise.
5173 * config/arm/arm.c (cortexa9_extra_costs): Fix identifiers in comments.
5174 (cortexa7_extra_costs): Likewise.
5175 (cortexa12_extra_costs): Likewise.
5176 (cortexa15_extra_costs): Likewise.
5177 (v7m_extra_costs): Likewise.
5179 2014-02-12 Richard Biener <rguenther@suse.de>
5182 * gimple-low.c (lower_builtin_posix_memalign): Lower conditional
5183 of posix_memalign being successful.
5184 (lower_stmt): Restrict lowering of posix_memalign to when
5185 -ftree-bit-ccp is enabled.
5187 2014-02-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
5189 * config/avr/avr-c.c (avr_resolve_overloaded_builtin): Pass vNULL for
5191 * config/spu/spu-c.c (spu_resolve_overloaded_builtin): Likewise.
5193 2014-02-12 Eric Botcazou <ebotcazou@adacore.com>
5195 PR rtl-optimization/60116
5196 * combine.c (try_combine): Also remove dangling REG_DEAD notes on the
5197 other_insn once the combination has been validated.
5199 2014-02-11 Jan Hubicka <hubicka@ucw.cz>
5202 * ipa-utils.h (possible_polymorphic_call_targets): Update prototype
5204 * ipa-devirt.c: Include demangle.h
5205 (odr_violation_reported): New static variable.
5206 (add_type_duplicate): Update odr_violations.
5207 (maybe_record_node): Add completep parameter; update it.
5208 (record_target_from_binfo): Add COMPLETEP parameter;
5209 update it as needed.
5210 (possible_polymorphic_call_targets_1): Likewise.
5211 (struct polymorphic_call_target_d): Add nonconstruction_targets;
5212 rename FINAL to COMPLETE.
5213 (record_targets_from_bases): Sanity check we found the binfo;
5214 fix COMPLETEP updating.
5215 (possible_polymorphic_call_targets): Add NONCONSTRUTION_TARGETSP
5216 parameter, fix computing of COMPLETEP.
5217 (dump_possible_polymorphic_call_targets): Imrove readability of dump;
5218 at LTO time do demangling.
5219 (ipa_devirt): Use nonconstruction_targets; Improve dumps.
5220 * gimple-fold.c (gimple_get_virt_method_for_vtable): Add can_refer
5222 (gimple_get_virt_method_for_binfo): Likewise.
5223 * gimple-fold.h (gimple_get_virt_method_for_binfo,
5224 gimple_get_virt_method_for_vtable): Update prototypes.
5226 2014-02-11 Vladimir Makarov <vmakarov@redhat.com>
5229 * genautomata.c (add_presence_absence): Fix typo with
5230 {final_}presence_list.
5232 2014-02-11 Michael Meissner <meissner@linux.vnet.ibm.com>
5235 * config/rs6000/rs6000.md (128-bit GPR splitter): Add a splitter
5236 for VSX/Altivec vectors that land in GPR registers.
5238 2014-02-11 Richard Henderson <rth@redhat.com>
5239 Jakub Jelinek <jakub@redhat.com>
5242 * tree-sra.c (load_assign_lhs_subreplacements): Add VIEW_CONVERT_EXPR
5243 around drhs if type conversion to lacc->type is not useless.
5245 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5247 * config/aarch64/aarch64-cores.def (cortex-a57): Use cortexa57
5249 (cortex-a57.cortex-a53): Likewise.
5250 * config/aarch64/aarch64.c (cortexa57_tunings): New tuning struct.
5252 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5254 * config/arm/thumb2.md (*thumb2_movhi_insn): Add alternatives for
5257 2014-02-11 Renlin Li <Renlin.Li@arm.com>
5259 * doc/sourcebuild.texi: Document check_effective_target_arm_vfp3_ok and
5260 add_options_for_arm_vfp3.
5262 2014-02-11 Jeff Law <law@redhat.com>
5265 * expr.c (expand_expr_addr_expr_1): Handle expand_expr returning an
5266 object with an undesirable mode.
5268 2014-02-11 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5271 * config/i386/sol2-9.h: New file.
5272 * config.gcc (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*,
5273 *-*-solaris2.9*): Use it.
5275 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
5277 * config/microblaze/microblaze.md: Add movsi4_rev insn pattern.
5278 * config/microblaze/predicates.md: Add reg_or_mem_operand predicate.
5280 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
5282 * config/microblaze/microblaze.c: Extend mcpu version format
5284 2014-02-10 David Holsgrove <david.holsgrove@xilinx.com>
5286 * config/microblaze/microblaze.h: Define SIZE_TYPE and PTRDIFF_TYPE.
5288 2014-02-10 Richard Henderson <rth@redhat.com>
5291 * calls.c (expand_call): Don't double-push for reg_parm_stack_space.
5292 * config/i386/i386.c (init_cumulative_args): Remove sorry for 64-bit
5293 ms-abi vs -mno-accumulate-outgoing-args.
5294 (ix86_expand_prologue): Unconditionally call ix86_eax_live_at_start_p.
5295 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Fix comment with
5298 2014-02-10 Bernd Edlinger <bernd.edlinger@hotmail.de>
5301 * cfgexpand.c (expand_asm_operands): Attach source location to
5302 ASM_INPUT rtx objects.
5303 * print-rtl.c (print_rtx): Check for UNKNOWN_LOCATION.
5305 2014-02-10 Nick Clifton <nickc@redhat.com>
5307 * config/mn10300/mn10300.c (popcount): New function.
5308 (mn10300_expand_prologue): Include saved registers in stack usage
5311 2014-02-10 Jeff Law <law@redhat.com>
5314 * reload1.c (emit_input_reload_insns): Do not create invalid RTL
5315 when changing the SET_DEST of a prior insn to avoid an input reload.
5317 2014-02-10 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
5319 * config/rs6000/sysv4.h (ENDIAN_SELECT): Do not attempt to enforce
5320 big-endian mode for -mcall-aixdesc, -mcall-freebsd, -mcall-netbsd,
5321 -mcall-openbsd, or -mcall-linux.
5322 (CC1_ENDIAN_BIG_SPEC): Remove.
5323 (CC1_ENDIAN_LITTLE_SPEC): Remove.
5324 (CC1_ENDIAN_DEFAULT_SPEC): Remove.
5325 (CC1_SPEC): Remove (always empty) %cc1_endian_... spec.
5326 (SUBTARGET_EXTRA_SPECS): Remove %cc1_endian_big, %cc1_endian_little,
5327 and %cc1_endian_default.
5328 * config/rs6000/sysv4le.h (CC1_ENDIAN_DEFAULT_SPEC): Remove.
5330 2014-02-10 Richard Biener <rguenther@suse.de>
5332 PR tree-optimization/60115
5333 * tree-eh.c (tree_could_trap_p): Unify TARGET_MEM_REF and
5334 MEM_REF handling. Properly verify that the accesses are not
5335 out of the objects bound.
5337 2014-02-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5339 * config/aarch64/aarch64.c (aarch64_override_options): Fix typo from
5342 2014-02-10 Eric Botcazou <ebotcazou@adacore.com>
5344 * ipa-devirt.c (get_polymorphic_call_info_from_invariant): Return
5345 proper constants and fix formatting.
5346 (possible_polymorphic_call_targets): Fix formatting.
5348 2014-02-10 Kirill Yukhin <kirill.yukhin@intel.com>
5349 Ilya Tocar <ilya.tocar@intel.com>
5351 * config/i386/avx512fintrin.h (_mm512_storeu_epi64): Removed.
5352 (_mm512_loadu_epi32): Renamed into...
5353 (_mm512_loadu_si512): This.
5354 (_mm512_storeu_epi32): Renamed into...
5355 (_mm512_storeu_si512): This.
5356 (_mm512_maskz_ceil_ps): Removed.
5357 (_mm512_maskz_ceil_pd): Ditto.
5358 (_mm512_maskz_floor_ps): Ditto.
5359 (_mm512_maskz_floor_pd): Ditto.
5360 (_mm512_floor_round_ps): Ditto.
5361 (_mm512_floor_round_pd): Ditto.
5362 (_mm512_ceil_round_ps): Ditto.
5363 (_mm512_ceil_round_pd): Ditto.
5364 (_mm512_mask_floor_round_ps): Ditto.
5365 (_mm512_mask_floor_round_pd): Ditto.
5366 (_mm512_mask_ceil_round_ps): Ditto.
5367 (_mm512_mask_ceil_round_pd): Ditto.
5368 (_mm512_maskz_floor_round_ps): Ditto.
5369 (_mm512_maskz_floor_round_pd): Ditto.
5370 (_mm512_maskz_ceil_round_ps): Ditto.
5371 (_mm512_maskz_ceil_round_pd): Ditto.
5372 (_mm512_expand_pd): Ditto.
5373 (_mm512_expand_ps): Ditto.
5374 * config/i386/i386.c (ix86_builtins): Remove
5375 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK.
5376 (bdesc_args): Ditto.
5377 * config/i386/predicates.md (const1256_operand): New.
5378 (const_1_to_2_operand): Ditto.
5379 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
5380 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
5381 (*avx512pf_gatherpf<mode>sf): Ditto.
5382 (avx512pf_gatherpf<mode>df): Ditto.
5383 (*avx512pf_gatherpf<mode>df_mask): Ditto.
5384 (*avx512pf_gatherpf<mode>df): Ditto.
5385 (avx512pf_scatterpf<mode>sf): Ditto.
5386 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
5387 (*avx512pf_scatterpf<mode>sf): Ditto.
5388 (avx512pf_scatterpf<mode>df): Ditto.
5389 (*avx512pf_scatterpf<mode>df_mask): Ditto.
5390 (*avx512pf_scatterpf<mode>df): Ditto.
5391 (avx512f_expand<mode>): Removed.
5392 (<shift_insn><mode>3<mask_name>): Change predicate type.
5394 2014-02-08 Jakub Jelinek <jakub@redhat.com>
5396 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers
5397 not at the end of datarefs vector use ordered_remove to avoid
5398 reordering datarefs vector.
5401 * gimplify.c (gimplify_bind_expr): In ORT_SIMD region
5402 mark local addressable non-static vars as GOVD_PRIVATE
5403 instead of GOVD_LOCAL.
5404 * omp-low.c (lower_omp_for): Move gimple_bind_vars
5405 and BLOCK_VARS of gimple_bind_block to new_stmt rather
5409 * tree-ssa-ccp.c (surely_varying_stmt_p): Don't return true
5410 if TYPE_ATTRIBUTES (gimple_call_fntype ()) contain
5411 assume_aligned or alloc_align attributes.
5412 (bit_value_assume_aligned): Add ATTR, PTRVAL and ALLOC_ALIGN
5413 arguments. Handle also assume_aligned and alloc_align attributes.
5414 (evaluate_stmt): Adjust bit_value_assume_aligned caller. Handle
5415 calls to functions with assume_aligned or alloc_align attributes.
5416 * doc/extend.texi: Document assume_aligned and alloc_align attributes.
5418 2014-02-08 Terry Guo <terry.guo@arm.com>
5420 * doc/invoke.texi: Document ARM -march=armv7e-m.
5422 2014-02-08 Jakub Jelinek <jakub@redhat.com>
5424 * cilk-common.c (cilk_init_builtins): Clear TREE_NOTHROW
5425 flag on __cilkrts_rethrow builtin.
5428 * ipa-cp.c (determine_versionability): Fail at -O0
5429 or __attribute__((optimize (0))) or -fno-ipa-cp functions.
5430 * tree-sra.c (ipa_sra_preliminary_function_checks): Similarly.
5433 2014-02-04 Jakub Jelinek <jakub@redhat.com>
5436 * tree-inline.c (copy_forbidden): Fail for
5437 __attribute__((optimize (0))) functions.
5439 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
5441 * varpool.c: Include pointer-set.h.
5442 (varpool_remove_unreferenced_decls): Variables in other partitions
5443 will not be output; be however careful to not lose information
5446 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
5448 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do O(1)
5449 lookup in the vtable constructor.
5451 2014-02-07 Jeff Law <law@redhat.com>
5454 * config/m68k/m68k.md (ashldi_extsi): Turn into a
5455 define_insn_and_split.
5457 * ipa-inline.c (inline_small_functions): Fix typos.
5459 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5461 * config/s390/s390-protos.h (s390_can_use_simple_return_insn)
5462 (s390_can_use_return_insn): Declare.
5463 * config/s390/s390.h (EPILOGUE_USES): Define.
5464 * config/s390/s390.c (s390_mainpool_start): Allow two main_pool
5466 (s390_chunkify_start): Handle return JUMP_LABELs.
5467 (s390_early_mach): Emit a main_pool instruction on the entry edge.
5468 (s300_set_up_by_prologue, s390_can_use_simple_return_insn)
5469 (s390_can_use_return_insn): New functions.
5470 (s390_fix_long_loop_prediction): Handle conditional returns.
5471 (TARGET_SET_UP_BY_PROLOGUE): Define.
5472 * config/s390/s390.md (ANY_RETURN): New code iterator.
5473 (*creturn, *csimple_return, return, simple_return): New patterns.
5475 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5477 * config/s390/s390.c (s390_restore_gprs_from_fprs): Add REG_CFA_RESTORE
5478 notes to each restore. Also add REG_CFA_DEF_CFA when restoring %r15.
5479 (s390_optimize_prologue): Don't clear RTX_FRAME_RELATED_P. Update the
5480 REG_CFA_RESTORE list when deciding not to restore a register.
5482 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5484 * config/s390/s390.c: Include tree-pass.h and context.h.
5485 (s390_early_mach): New function, split out from...
5486 (s390_emit_prologue): ...here.
5487 (pass_data_s390_early_mach): New pass structure.
5488 (pass_s390_early_mach): New class.
5489 (s390_option_override): Create and register early_mach pass.
5490 Move to end of file.
5492 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5494 * var-tracking.c (vt_stack_adjustments): Don't require stack_adjusts
5495 to match for the exit block.
5497 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5499 * config/s390/s390.md ("atomic_load<mode>", "atomic_store<mode>")
5500 ("atomic_compare_and_swap<mode>", "atomic_fetch_<atomic><mode>"):
5501 Reject misaligned operands.
5503 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5505 * optabs.c (expand_atomic_compare_and_swap): Allow expander to fail.
5507 2014-02-07 Richard Biener <rguenther@suse.de>
5510 * gimple-low.c (lower_builtin_posix_memalign): New function.
5511 (lower_stmt): Call it to lower posix_memalign in a way
5512 to make alignment info accessible.
5514 2014-02-07 Jakub Jelinek <jakub@redhat.com>
5517 * tree.c (build_common_builtin_nodes): Set ECF_LEAF for
5518 __builtin_setjmp_receiver.
5520 2014-02-07 Richard Biener <rguenther@suse.de>
5523 * builtin-types.def (BT_FN_INT_PTRPTR_SIZE_SIZE): Add.
5524 * builtins.def (BUILT_IN_POSIX_MEMALIGN): Likewise.
5525 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
5526 Handle BUILT_IN_POSIX_MEMALIGN.
5527 (find_func_clobbers): Likewise.
5528 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
5529 (call_may_clobber_ref_p_1): Likewise.
5531 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5534 * ipa-devirt.c (record_target_from_binfo): Remove overactive
5537 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5540 * lto-cgraph.c (lto_output_node): Use
5541 symtab_get_symbol_partitioning_class.
5542 (lto_output_varpool_node): likewise.
5543 (symtab_get_symbol_partitioning_class): Move here from
5545 * cgraph.h (symbol_partitioning_class): Likewise.
5546 (symtab_get_symbol_partitioning_class): Declare.
5548 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5550 * ggc.h (ggc_internal_cleared_alloc): New macro.
5551 * vec.h (vec_safe_copy): Handle memory stats.
5552 * omp-low.c (simd_clone_struct_alloc): Use ggc_internal_cleared_alloc.
5553 * target-globals.c (save_target_globals): Likewise.
5555 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5558 * expr.c (emit_move_resolve_push): Export; be bit more selective
5559 on when to clear alias set.
5560 * expr.h (emit_move_resolve_push): Declare.
5561 * function.h (struct function): Add tail_call_marked.
5562 * tree-tailcall.c (optimize_tail_call): Set tail_call_marked.
5563 * config/i386/i386-protos.h (ix86_expand_push): Remove.
5564 * config/i386/i386.md (TImode move expander): De not call
5566 (FP push expanders): Preserve memory attributes.
5567 * config/i386/sse.md (push<mode>1): Remove.
5568 * config/i386/i386.c (ix86_expand_vector_move): Handle push operation.
5569 (ix86_expand_push): Remove.
5570 * config/i386/mmx.md (push<mode>1): Remove.
5572 2014-02-06 Jakub Jelinek <jakub@redhat.com>
5574 PR rtl-optimization/60030
5575 * internal-fn.c (ubsan_expand_si_overflow_mul_check): Surround
5576 lopart with paradoxical subreg before shifting it up by hprec.
5578 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5580 * config/arm/aarch-cost-tables.h (cortexa57_extra_costs): New table.
5581 Remove extra newline at end of file.
5582 * config/arm/arm.c (arm_cortex_a57_tune): New tuning struct.
5583 (arm_issue_rate): Handle cortexa57.
5584 * config/arm/arm-cores.def (cortex-a57): Use cortex_a57 tuning.
5585 (cortex-a57.cortex-a53): Likewise.
5587 2014-02-06 Jakub Jelinek <jakub@redhat.com>
5590 * config/arm/arm.c (emit_multi_reg_push): Add dwarf_regs_mask argument,
5591 don't record in REG_FRAME_RELATED_EXPR registers not set in that
5593 (arm_expand_prologue): Adjust all callers.
5594 (arm_unwind_emit_sequence): Allow saved, but not important for unwind
5595 info, registers also at the lowest numbered registers side. Use
5596 gcc_assert instead of abort, and SET_SRC/SET_DEST macros instead of
5600 * var-tracking.c (adjust_mems): Before adding a SET to
5601 amd->side_effects, adjust it's SET_SRC using simplify_replace_fn_rtx.
5603 2014-02-06 Alan Modra <amodra@gmail.com>
5606 * config/rs6000/rs6000.c (rs6000_secondary_memory_needed_mode): Only
5607 change SDmode to DDmode when lra_in_progress.
5609 2014-02-06 Jakub Jelinek <jakub@redhat.com>
5612 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers, call
5613 free_data_ref on the dr first, and before goto again also set dr
5614 to the next dr. For simd_lane_access, free old datarefs[i] before
5615 overwriting it. For get_vectype_for_scalar_type failure, don't
5616 free_data_ref if simd_lane_access.
5618 * Makefile.in (prefix.o, cppbuiltin.o): Depend on $(BASEVER).
5621 * tree.h (opts_for_fn): New inline function.
5622 (opt_for_fn): Define.
5623 * config/i386/i386.c (ix86_function_regparm): Use
5624 opt_for_fn (decl, optimize) instead of optimize.
5626 2014-02-06 Marcus Shawcroft <marcus.shawcroft@arm.com>
5628 * config/aarch64/aarch64.c (aarch64_classify_symbol): Fix logic
5629 for SYMBOL_REF in large memory model.
5631 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5633 * config/aarch64/aarch64-cores.def (cortex-a53): Specify CRC32
5635 (cortex-a57): Likewise.
5636 (cortex-a57.cortex-a53): Likewise.
5638 2014-02-06 Yury Gribov <y.gribov@samsung.com>
5639 Kugan Vivekanandarajah <kuganv@linaro.org>
5641 * config/arm/arm.c (arm_vector_alignment_reachable): Check
5643 * config/arm/arm.c (arm_builtin_support_vector_misalignment): Likewise.
5645 2014-02-06 Richard Biener <rguenther@suse.de>
5647 * tree-cfg.c (gimple_duplicate_sese_region): Fix ordering of
5648 set_loop_copy and initialize_original_copy_tables.
5650 2014-02-06 Alex Velenko <Alex.Velenko@arm.com>
5652 * config/aarch64/aarch64-simd.md
5653 (aarch64_ashr_simddi): Change QI to SI.
5655 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
5656 Jakub Jelinek <jakub@redhat.com>
5659 * ipa-inline-analysis.c (compute_bb_predicates): Ensure monotonicity
5662 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5664 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Change
5665 CODE_FOR_altivec_vpku[hw]um to
5666 CODE_FOR_altivec_vpku[hw]um_direct.
5667 * config/rs6000/altivec.md (vec_unpacks_hi_<VP_small_lc>): Change
5668 UNSPEC_VUNPACK_HI_SIGN to UNSPEC_VUNPACK_HI_SIGN_DIRECT.
5669 (vec_unpacks_lo_<VP_small_lc>): Change UNSPEC_VUNPACK_LO_SIGN to
5670 UNSPEC_VUNPACK_LO_SIGN_DIRECT.
5672 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5674 * config/rs6000/altivec.md (altivec_vsum2sws): Adjust code
5675 generation for -maltivec=be.
5676 (altivec_vsumsws): Simplify redundant test.
5678 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5680 * altivec.md (UNSPEC_VPACK_UNS_UNS_MOD_DIRECT): New unspec.
5681 (UNSPEC_VUNPACK_HI_SIGN_DIRECT): Likewise.
5682 (UNSPEC_VUNPACK_LO_SIGN_DIRECT): Likewise.
5683 (mulv8hi3): Use gen_altivec_vpkuwum_direct instead of
5684 gen_altivec_vpkuwum.
5685 (altivec_vpkpx): Test for VECTOR_ELT_ORDER_BIG instead of for
5687 (altivec_vpks<VI_char>ss): Likewise.
5688 (altivec_vpks<VI_char>us): Likewise.
5689 (altivec_vpku<VI_char>us): Likewise.
5690 (altivec_vpku<VI_char>um): Likewise.
5691 (altivec_vpku<VI_char>um_direct): New (copy of
5692 altivec_vpku<VI_char>um that still relies on BYTES_BIG_ENDIAN, for
5694 (altivec_vupkhs<VU_char>): Emit vupkls* instead of vupkhs* when
5695 target is little endian and -maltivec=be is not specified.
5696 (*altivec_vupkhs<VU_char>_direct): New (copy of
5697 altivec_vupkhs<VU_char> that always emits vupkhs*, for internal use).
5698 (altivec_vupkls<VU_char>): Emit vupkhs* instead of vupkls* when
5699 target is little endian and -maltivec=be is not specified.
5700 (*altivec_vupkls<VU_char>_direct): New (copy of
5701 altivec_vupkls<VU_char> that always emits vupkls*, for internal use).
5702 (altivec_vupkhpx): Emit vupklpx instead of vupkhpx when target is
5703 little endian and -maltivec=be is not specified.
5704 (altivec_vupklpx): Emit vupkhpx instead of vupklpx when target is
5705 little endian and -maltivec=be is not specified.
5707 2014-02-05 Richard Henderson <rth@redhat.com>
5710 * combine-stack-adj.c: Revert r206943.
5711 * sched-int.h (struct deps_desc): Add last_args_size.
5712 * sched-deps.c (init_deps): Initialize it.
5713 (sched_analyze_insn): Add OUTPUT dependencies between insns that
5714 contain REG_ARGS_SIZE notes.
5716 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
5718 * lto-cgraph.c (asm_nodes_output): Make global.
5719 * lto-wrapper.c (run_gcc): Pass down paralelizm to WPA.
5720 * gcc.c (AS_NEEDS_DASH_FOR_PIPED_INPUT): Allow WPA parameter
5721 (driver_handle_option): Handle OPT_fwpa.
5723 2014-02-05 Jakub Jelinek <jakub@redhat.com>
5726 * ipa-devirt.c (possible_polymorphic_call_targets): Fix
5727 a comment typo and formatting issue. If odr_hash hasn't been
5728 created, return vNULL and set *completep to false.
5731 * tree-eh.c (cleanup_empty_eh): Bail out on totally empty
5732 bb with no successors.
5734 2014-02-05 James Greenhalgh <james.greenhalgh@arm.com>
5737 * doc/invoke.texi (-march): Clarify documentation for ARM.
5741 2014-02-05 Richard Biener <rguenther@suse.de>
5743 * tree-vect-loop.c (vect_analyze_loop_2): Be more informative
5744 when not vectorizing because of too many alias checks.
5745 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
5746 Add more verboseness, avoid duplicate MSG_MISSED_OPTIMIZATION.
5748 2014-02-05 Nick Clifton <nickc@redhat.com>
5750 * config/mn10300/mn10300.c (mn10300_hard_regno_mode_ok): Do not
5751 accept extended registers in any mode when compiling for the MN10300.
5753 2014-02-05 Yury Gribov <y.gribov@samsung.com>
5755 * cif-code.def (ATTRIBUTE_MISMATCH): New CIF code.
5756 * ipa-inline.c (report_inline_failed_reason): Handle mismatched
5757 sanitization attributes.
5758 (can_inline_edge_p): Likewise.
5759 (sanitize_attrs_match_for_inline_p): New function.
5761 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
5763 * ipa-prop.c (detect_type_change): Shor circuit testing of
5764 type changes on THIS pointer.
5766 2014-02-04 John David Anglin <danglin@gcc.gnu.org>
5769 * config/pa/pa.c (legitimize_tls_address): Return original address
5770 if not passed a SYMBOL_REF rtx.
5771 (hppa_legitimize_address): Call legitimize_tls_address for all TLS
5773 (pa_emit_move_sequence): Simplify TLS source operands.
5774 (pa_legitimate_constant_p): Reject all TLS constants.
5775 * config/pa/pa.h (PA_SYMBOL_REF_TLS_P): Correct comment.
5776 (CONSTANT_ADDRESS_P): Reject TLS CONST addresses.
5778 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
5780 * ipa.c (function_and_variable_visibility): Decompose DECL_ONE_ONLY
5781 groups when we know they are controlled by LTO.
5782 * varasm.c (default_binds_local_p_1): If object is in other partition,
5783 it will be resolved locally.
5785 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
5787 * config/host-linux.c (linux_gt_pch_use_address): Don't
5788 use SSIZE_MAX because it is not always defined.
5790 2014-02-04 Vladimir Makarov <vmakarov@redhat.com>
5793 * lra-constraints.c (need_for_split_p): Use more 3 reloads as
5794 threshold for pseudo splitting.
5795 (update_ebb_live_info): Process call argument hard registers and
5796 hard registers from insn definition too.
5797 (max_small_class_regs_num): New constant.
5798 (inherit_in_ebb): Update live hard regs through EBBs. Update
5799 reloads_num only for small register classes. Don't split for
5802 2014-02-04 Markus Trippelsdorf <markus@trippelsdorf.de>
5805 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that target
5808 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
5810 * gimple-fold.c (can_refer_decl_in_current_unit_p): Default
5813 2014-02-04 Marek Polacek <polacek@redhat.com>
5815 * gdbinit.in (pel): Define.
5817 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
5819 * doc/invoke.texi (fstrict-volatile-bitfields): Clarify current
5822 2014-02-04 Richard Biener <rguenther@suse.de>
5825 * lto-streamer-out.c (tree_is_indexable): Force NAMELIST_DECLs
5826 in function context local.
5827 (lto_output_tree_ref): Do not write trees from lto_output_tree_ref.
5828 * lto-streamer-in.c (lto_input_tree_ref): Handle LTO_namelist_decl_ref
5829 similar to LTO_imported_decl_ref.
5831 2014-02-04 Jakub Jelinek <jakub@redhat.com>
5833 PR tree-optimization/60002
5834 * cgraphclones.c (build_function_decl_skip_args): Clear
5837 PR tree-optimization/60023
5838 * tree-if-conv.c (predicate_mem_writes): Pass true instead of
5839 false to gsi_replace.
5840 * tree-vect-stmts.c (vect_finish_stmt_generation): If stmt
5841 has been in some EH region and vec_stmt could throw, add
5842 vec_stmt into the same EH region.
5843 * tree-data-ref.c (get_references_in_stmt): If IFN_MASK_LOAD
5844 has no lhs, ignore it.
5845 * internal-fn.c (expand_MASK_LOAD): Likewise.
5848 * tree-inline.c (copy_forbidden): Fail for
5849 __attribute__((optimize (0))) functions.
5852 * omp-low.c (simd_clone_struct_copy): If from->inbranch
5853 is set, copy one less argument.
5854 (expand_simd_clones): Don't subtract clone_info->inbranch
5855 from simd_clone_struct_alloc argument.
5857 PR rtl-optimization/57915
5858 * recog.c (simplify_while_replacing): If all unary/binary/relational
5859 operation arguments are constant, attempt to simplify those.
5862 * expmed.c (expand_mult): For MODE_VECTOR_INT multiplication
5863 if there is no vashl<mode>3 or ashl<mode>3 insn, skip_synth.
5865 2014-02-04 Richard Biener <rguenther@suse.de>
5867 PR tree-optimization/60012
5868 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Apply
5869 TBAA disambiguation to all DDRs.
5871 2014-02-04 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5874 * config/sol2.h (LINK_LIBGCC_MAPFILE_SPEC): Define.
5875 (LINK_SPEC): Use it for -shared, -shared-libgcc.
5877 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
5880 * tree.c (get_binfo_at_offset): Do not get confused by empty classes;
5882 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
5884 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Remove.
5885 * gimple-fold.h (gimple_extract_devirt_binfo_from_cst): Remove.
5887 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
5890 * ipa-cp.c (ipa_get_indirect_edge_target_1): Use ipa-devirt
5891 to figure out targets of polymorphic calls with known decl.
5892 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
5893 * ipa-utils.h (get_polymorphic_call_info_from_invariant): Declare.
5894 * ipa-devirt.c (get_polymorphic_call_info_for_decl): Break out from ...
5895 (get_polymorphic_call_info): ... here.
5896 (get_polymorphic_call_info_from_invariant): New function.
5898 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
5900 * ipa-cp.c (ipa_get_indirect_edge_target_1): Do direct
5901 lookup via vtable pointer; check for type consistency
5902 and turn inconsitent facts into UNREACHABLE.
5903 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
5904 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do not ICE on
5905 type inconsistent querries; return UNREACHABLE instead.
5907 2014-02-03 Richard Henderson <rth@twiddle.net>
5910 * tree-ssa-uninit.c (push_to_worklist): Don't re-push if we've
5911 already processed this node.
5912 (normalize_one_pred_1): Pass along mark_set.
5913 (normalize_one_pred): Create and destroy a pointer_set_t.
5914 (normalize_one_pred_chain): Likewise.
5916 2014-02-03 Laurent Aflonsi <laurent.alfonsi@st.com>
5918 PR gcov-profile/58602
5919 * gcc/gcov-io.c (gcov_open): Open with truncation when mode < 0.
5921 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
5924 * ipa-cp.c (ipa_get_indirect_edge_target_1): Give up on
5925 -fno-devirtualize; try to devirtualize by the knowledge of
5926 virtual table pointer given by aggregate propagation.
5927 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
5928 (ipa_print_node_jump_functions): Dump also offset that
5929 is relevant for polymorphic calls.
5930 (determine_known_aggregate_parts): Add arg_type parameter; use it
5931 instead of determining the type from pointer type.
5932 (ipa_compute_jump_functions_for_edge): Update call of
5933 determine_known_aggregate_parts.
5934 * gimple-fold.c (gimple_get_virt_method_for_vtable): Break out from ...
5935 (gimple_get_virt_method_for_binfo): ... here; simplify using
5936 vtable_pointer_value_to_vtable.
5937 * gimple-fold.h (gimple_get_virt_method_for_vtable): Declare.
5938 * ipa-devirt.c (subbinfo_with_vtable_at_offset): Turn OFFSET parameter
5939 to unsigned HOST_WIDE_INT; use vtable_pointer_value_to_vtable.
5940 (vtable_pointer_value_to_vtable): Break out from ...; handle also
5942 (vtable_pointer_value_to_binfo): ... here.
5943 * ipa-utils.h (vtable_pointer_value_to_vtable): Declare.
5945 2014-02-03 Teresa Johnson <tejohnson@google.com>
5947 * tree-vect-slp.c (vect_supported_load_permutation_p): Avoid
5948 redef of outer loop index variable.
5950 2014-02-03 Marc Glisse <marc.glisse@inria.fr>
5954 * doc/extend.texi (Function Attributes): Typo.
5956 2014-02-03 Cong Hou <congh@google.com>
5958 PR tree-optimization/60000
5959 * tree-vect-loop.c (vect_transform_loop): Set pattern_def_seq to NULL
5960 if the vectorized statement is a store. A store statement can only
5961 appear at the end of pattern statements.
5963 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
5965 * config/i386/i386.c (flag_opts): Add -mlong-double-128.
5966 (ix86_option_override_internal): Default long double to 64-bit for
5967 32-bit Bionic and to 128-bit for 64-bit Bionic.
5969 * config/i386/i386.h (LONG_DOUBLE_TYPE_SIZE): Use 128 if
5970 TARGET_LONG_DOUBLE_128 is true.
5971 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Likewise.
5973 * config/i386/i386.opt (mlong-double-80): Negate -mlong-double-64.
5974 (mlong-double-64): Negate -mlong-double-128.
5975 (mlong-double-128): New option.
5977 * config/i386/i386-c.c (ix86_target_macros): Define
5978 __LONG_DOUBLE_128__ for TARGET_LONG_DOUBLE_128.
5980 * doc/invoke.texi: Document -mlong-double-128.
5982 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
5984 PR rtl-optimization/60024
5985 * sel-sched.c (init_regs_for_mode): Check if mode is OK first.
5987 2014-02-03 Markus Trippelsdorf <markus@trippelsdorf.de>
5989 * doc/invoke.texi (fprofile-reorder-functions): Fix typo.
5991 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
5993 PR rtl-optimization/57662
5994 * sel-sched.c (code_motion_path_driver): Do not mark already not
5995 existing blocks in the visiting bitmap.
5997 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
5999 * sel-sched-ir.c (sel_gen_insn_from_expr_after): Reset INSN_DELETED_P
6000 on the insn being emitted.
6002 2014-02-03 James Greenhalgh <james.greenhalgh@arm.com>
6003 Will Deacon <will.deacon@arm.com>
6005 * doc/gimple.texi (gimple_asm_clear_volatile): Remove.
6007 2014-02-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6009 * config/arm/arm-tables.opt: Regenerate.
6011 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6013 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): Generalize
6014 for vector types other than V16QImode.
6015 * config/rs6000/altivec.md (altivec_vperm_<mode>): Change to a
6016 define_expand, and call altivec_expand_vec_perm_le when producing
6017 code with little endian element order.
6018 (*altivec_vperm_<mode>_internal): New insn having previous
6019 behavior of altivec_vperm_<mode>.
6020 (altivec_vperm_<mode>_uns): Change to a define_expand, and call
6021 altivec_expand_vec_perm_le when producing code with little endian
6023 (*altivec_vperm_<mode>_uns_internal): New insn having previous
6024 behavior of altivec_vperm_<mode>_uns.
6026 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6028 * config/rs6000/altivec.md (UNSPEC_VSUMSWS_DIRECT): New unspec.
6029 (altivec_vsumsws): Add handling for -maltivec=be with a little
6031 (altivec_vsumsws_direct): New.
6032 (reduc_splus_<mode>): Call gen_altivec_vsumsws_direct instead of
6033 gen_altivec_vsumsws.
6035 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
6037 * ipa-devirt.c (subbinfo_with_vtable_at_offset,
6038 vtable_pointer_value_to_binfo): New functions.
6039 * ipa-utils.h (vtable_pointer_value_to_binfo): Declare.
6040 * ipa-prop.c (extr_type_from_vtbl_ptr_store): Use it.
6042 2014-02-02 Sandra Loosemore <sandra@codesourcery.com>
6044 * config/nios2/nios2.md (load_got_register): Initialize GOT
6045 pointer from _gp_got instead of _GLOBAL_OFFSET_TABLE_.
6046 * config/nios2/nios2.c (nios2_function_profiler): Likewise.
6048 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
6050 * ipa-prop.c (update_jump_functions_after_inlining): When type is not
6051 preserverd by passthrough, do not propagate the type.
6053 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
6055 * config/mips/mips.c (MIPS_GET_FCSR, MIPS_SET_FCSR): New macros.
6056 (mips_atomic_assign_expand_fenv): New function.
6057 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): Define.
6059 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
6061 * doc/extend.texi (__builtin_mips_get_fcsr): Document.
6062 (__builtin_mips_set_fcsr): Likewise.
6063 * config/mips/mips-ftypes.def: Add MIPS_VOID_FTYPE_USI and
6064 MIPS_USI_FTYPE_VOID.
6065 * config/mips/mips-protos.h (mips16_expand_get_fcsr): Declare
6066 (mips16_expand_set_fcsr): Likewise.
6067 * config/mips/mips.c (mips16_get_fcsr_stub): New variable.
6068 (mips16_set_fcsr_stub): Likewise.
6069 (mips16_get_fcsr_one_only_stub): New class.
6070 (mips16_set_fcsr_one_only_stub): Likewise.
6071 (mips16_expand_get_fcsr, mips16_expand_set_fcsr): New functions.
6072 (mips_code_end): Output the get_fcsr and set_fcsr stubs, if needed.
6073 (BUILTIN_AVAIL_MIPS16, AVAIL_ALL): New macros.
6074 (hard_float): New availability predicate.
6075 (mips_builtins): Add get_fcsr and set_fcsr.
6076 (mips_expand_builtin): Check BUILTIN_AVAIL_MIPS16.
6077 * config/mips/mips.md (UNSPEC_GET_FCSR, UNSPEC_SET_FCSR): New unspecs.
6078 (GET_FCSR_REGNUM, SET_FCSR_REGNUM): New constants.
6079 (mips_get_fcsr, *mips_get_fcsr, mips_get_fcsr_mips16_<mode>)
6080 (mips_set_fcsr, *mips_set_fcsr, mips_set_fcsr_mips16_<mode>): New
6083 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
6085 * config/mips/mips.c (mips_one_only_stub): New class.
6086 (mips_need_mips16_rdhwr_p): Replace with...
6087 (mips16_rdhwr_stub): ...this new variable.
6088 (mips16_stub_call_address): New function.
6089 (mips16_rdhwr_one_only_stub): New class.
6090 (mips_expand_thread_pointer): Use mips16_stub_call_address.
6091 (mips_output_mips16_rdhwr): Delete.
6092 (mips_finish_stub): New function.
6093 (mips_code_end): Use it to handle rdhwr stubs.
6095 2014-02-02 Uros Bizjak <ubizjak@gmail.com>
6098 * config/i386/i386.c (classify_argument): Fix handling of bit_offset
6099 when calculating size of integer atomic types.
6101 2014-02-02 H.J. Lu <hongjiu.lu@intel.com>
6103 * ipa-inline-analysis.c (true_predicate_p): Fix a typo in comments.
6105 2014-02-01 Jakub Jelinek <jakub@redhat.com>
6107 PR tree-optimization/60003
6108 * gimple-low.c (lower_builtin_setjmp): Set cfun->has_nonlocal_label.
6109 * profile.c (branch_prob): Use gimple_call_builtin_p
6110 to check for BUILT_IN_SETJMP_RECEIVER.
6111 * tree-inline.c (copy_bb): Call notice_special_calls.
6113 2014-01-31 Vladimir Makarov <vmakarov@redhat.com>
6116 * lra-constraints.c (process_alt_operands): Update reload_sum only
6119 2014-01-31 Richard Henderson <rth@redhat.com>
6122 * tree-eh.c (lower_try_finally_switch): Delay lowering finally block
6123 until after else_eh is processed.
6125 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
6127 * config/i386/avx512fintrin.h (_MM_FROUND_TO_NEAREST_INT),
6128 (_MM_FROUND_TO_NEG_INF), (_MM_FROUND_TO_POS_INF),
6129 (_MM_FROUND_TO_ZERO), (_MM_FROUND_CUR_DIRECTION): Are already defined
6130 in smmintrin.h, remove them.
6131 (_MM_FROUND_NO_EXC): Same as above, bit also wrong value.
6132 * config/i386/i386.c (ix86_print_operand): Split sae and rounding.
6133 * config/i386/i386.md (ROUND_SAE): Fix value.
6134 * config/i386/predicates.md (const_4_or_8_to_11_operand): New.
6135 (const48_operand): New.
6136 * config/i386/subst.md (round), (round_expand): Use
6137 const_4_or_8_to_11_operand.
6138 (round_saeonly), (round_saeonly_expand): Use const48_operand.
6140 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
6142 * config/i386/constraints.md (Yk): Swap meaning with k.
6143 * config/i386/i386.md (movhi_internal): Change Yk to k.
6144 (movqi_internal): Ditto.
6145 (*k<logic><mode>): Ditto.
6148 (kandn<mode>): Ditto.
6149 (*<code>hi_1): Ditto.
6150 (*<code>qi_1): Ditto.
6151 (kxnor<mode>): Ditto.
6152 (kortestzhi): Ditto.
6153 (kortestchi): Ditto.
6155 (*one_cmplhi2_1): Ditto.
6156 (*one_cmplqi2_1): Ditto.
6157 * config/i386/sse.md (): Change k to Yk.
6158 (avx512f_load<mode>_mask): Ditto.
6159 (avx512f_blendm<mode>): Ditto.
6160 (avx512f_store<mode>_mask): Ditto.
6161 (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
6162 (avx512f_storedqu<mode>_mask): Ditto.
6163 (avx512f_cmp<mode>3<mask_scalar_merge_name><round_saeonly_name>):
6165 (avx512f_ucmp<mode>3<mask_scalar_merge_name>): Ditto.
6166 (avx512f_vmcmp<mode>3<round_saeonly_name>): Ditto.
6167 (avx512f_vmcmp<mode>3_mask<round_saeonly_name>): Ditto.
6168 (avx512f_maskcmp<mode>3): Ditto.
6169 (avx512f_fmadd_<mode>_mask<round_name>): Ditto.
6170 (avx512f_fmadd_<mode>_mask3<round_name>): Ditto.
6171 (avx512f_fmsub_<mode>_mask<round_name>): Ditto.
6172 (avx512f_fmsub_<mode>_mask3<round_name>): Ditto.
6173 (avx512f_fnmadd_<mode>_mask<round_name>): Ditto.
6174 (avx512f_fnmadd_<mode>_mask3<round_name>): Ditto.
6175 (avx512f_fnmsub_<mode>_mask<round_name>): Ditto.
6176 (avx512f_fnmsub_<mode>_mask3<round_name>): Ditto.
6177 (avx512f_fmaddsub_<mode>_mask<round_name>): Ditto.
6178 (avx512f_fmaddsub_<mode>_mask3<round_name>): Ditto.
6179 (avx512f_fmsubadd_<mode>_mask<round_name>): Ditto.
6180 (avx512f_fmsubadd_<mode>_mask3<round_name>): Ditto.
6181 (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
6182 (vec_extract_lo_<mode>_maskm): Ditto.
6183 (vec_extract_hi_<mode>_maskm): Ditto.
6184 (avx512f_vternlog<mode>_mask): Ditto.
6185 (avx512f_fixupimm<mode>_mask<round_saeonly_name>): Ditto.
6186 (avx512f_sfixupimm<mode>_mask<round_saeonly_name>): Ditto.
6187 (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
6188 (avx512f_<code>v8div16qi2_mask): Ditto.
6189 (avx512f_<code>v8div16qi2_mask_store): Ditto.
6190 (avx512f_eq<mode>3<mask_scalar_merge_name>_1): Ditto.
6191 (avx512f_gt<mode>3<mask_scalar_merge_name>): Ditto.
6192 (avx512f_testm<mode>3<mask_scalar_merge_name>): Ditto.
6193 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Ditto.
6194 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
6195 (*avx512pf_gatherpf<mode>df_mask): Ditto.
6196 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
6197 (*avx512pf_scatterpf<mode>df_mask): Ditto.
6198 (avx512cd_maskb_vec_dupv8di): Ditto.
6199 (avx512cd_maskw_vec_dupv16si): Ditto.
6200 (avx512f_vpermi2var<mode>3_maskz): Ditto.
6201 (avx512f_vpermi2var<mode>3_mask): Ditto.
6202 (avx512f_vpermi2var<mode>3_mask): Ditto.
6203 (avx512f_vpermt2var<mode>3_maskz): Ditto.
6204 (*avx512f_gathersi<mode>): Ditto.
6205 (*avx512f_gathersi<mode>_2): Ditto.
6206 (*avx512f_gatherdi<mode>): Ditto.
6207 (*avx512f_gatherdi<mode>_2): Ditto.
6208 (*avx512f_scattersi<mode>): Ditto.
6209 (*avx512f_scatterdi<mode>): Ditto.
6210 (avx512f_compress<mode>_mask): Ditto.
6211 (avx512f_compressstore<mode>_mask): Ditto.
6212 (avx512f_expand<mode>_mask): Ditto.
6213 * config/i386/subst.md (mask): Change k to Yk.
6214 (mask_scalar_merge): Ditto.
6217 2014-01-31 Marc Glisse <marc.glisse@inria.fr>
6219 * doc/extend.texi (Vector Extensions): Document ?: in C++.
6221 2014-01-31 Richard Biener <rguenther@suse.de>
6224 * builtins.c (fold_builtin_memory_op): Make sure to not
6225 use a floating-point mode or a boolean or enumeral type for
6228 2014-01-30 DJ Delorie <dj@redhat.com>
6230 * config/msp430/msp430.h (LIB_SPEC): Add -lcrt
6231 * config/msp430/msp430.md (msp430_refsym_need_exit): New.
6232 * config/msp430/msp430.c (msp430_expand_epilogue): Call it
6233 whenever main() has an epilogue.
6235 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6237 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Remove
6238 unused variable "field".
6239 * config/rs6000/vsx.md (vsx_mergel_<mode>): Add missing DONE.
6240 (vsx_mergeh_<mode>): Likewise.
6241 * config/rs6000/altivec.md (altivec_vmrghb): Likewise.
6242 (altivec_vmrghh): Likewise.
6243 (altivec_vmrghw): Likewise.
6244 (altivec_vmrglb): Likewise.
6245 (altivec_vmrglh): Likewise.
6246 (altivec_vmrglw): Likewise.
6247 (altivec_vspltb): Add missing uses.
6248 (altivec_vsplth): Likewise.
6249 (altivec_vspltw): Likewise.
6250 (altivec_vspltsf): Likewise.
6252 2014-01-30 Jakub Jelinek <jakub@redhat.com>
6255 * ifcvt.c (cond_exec_process_insns): Don't conditionalize
6256 frame related instructions.
6258 2014-01-30 Vladimir Makarov <vmakarov@redhat.com>
6260 PR rtl-optimization/59959
6261 * lra-constrains.c (simplify_operand_subreg): Assign NO_REGS to
6262 any reload of register whose subreg is invalid.
6264 2014-01-30 Jakub Jelinek <jakub@redhat.com>
6266 * config/i386/f16cintrin.h (_cvtsh_ss): Avoid -Wnarrowing warning.
6267 * config/i386/avx512fintrin.h (_mm512_mask_cvtusepi64_storeu_epi32):
6268 Add missing return type - void.
6270 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6272 * gcc/config/rs6000/rs6000.c (rs6000_expand_vector_init): Use
6273 gen_vsx_xxspltw_v4sf_direct instead of gen_vsx_xxspltw_v4sf;
6274 remove element index adjustment for endian (now handled in vsx.md
6276 (altivec_expand_vec_perm_const): Use
6277 gen_altivec_vsplt[bhw]_direct instead of gen_altivec_vsplt[bhw].
6278 * gcc/config/rs6000/vsx.md (UNSPEC_VSX_XXSPLTW): New unspec.
6279 (vsx_xxspltw_<mode>): Adjust element index for little endian.
6280 * gcc/config/rs6000/altivec.md (altivec_vspltb): Divide into a
6281 define_expand and a new define_insn *altivec_vspltb_internal;
6282 adjust for -maltivec=be on a little endian target.
6283 (altivec_vspltb_direct): New.
6284 (altivec_vsplth): Divide into a define_expand and a new
6285 define_insn *altivec_vsplth_internal; adjust for -maltivec=be on a
6286 little endian target.
6287 (altivec_vsplth_direct): New.
6288 (altivec_vspltw): Divide into a define_expand and a new
6289 define_insn *altivec_vspltw_internal; adjust for -maltivec=be on a
6290 little endian target.
6291 (altivec_vspltw_direct): New.
6292 (altivec_vspltsf): Divide into a define_expand and a new
6293 define_insn *altivec_vspltsf_internal; adjust for -maltivec=be on
6294 a little endian target.
6296 2014-01-30 Richard Biener <rguenther@suse.de>
6298 PR tree-optimization/59993
6299 * tree-ssa-forwprop.c (associate_pointerplus): Check we
6300 can propagate form the earlier stmt and avoid the transform
6301 when the intermediate result is needed.
6303 2014-01-30 Alangi Derick <alangiderick@gmail.com>
6305 * README.Portability: Fix typo.
6307 2014-01-30 David Holsgrove <david.holsgrove@xilinx.com>
6309 * config/microblaze/microblaze.md(cstoresf4, cbranchsf4): Replace
6310 comparison_operator with ordered_comparison_operator.
6312 2014-01-30 Nick Clifton <nickc@redhat.com>
6314 * config/mn10300/mn10300-protos.h (mn10300_store_multiple_operation_p):
6315 Rename to mn10300_store_multiple_regs.
6316 * config/mn10300/mn10300.c: Likewise.
6317 * config/mn10300/mn10300.md (store_movm): Fix typo: call
6318 store_multiple_regs.
6319 * config/mn10300/predicates.md (mn10300_store_multiple_operation):
6320 Call mn10300_store_multiple_regs.
6322 2014-01-30 Nick Clifton <nickc@redhat.com>
6323 DJ Delorie <dj@redhat.com>
6325 * config/rl78/rl78.c (register_sizes): Make the "upper half" of
6326 %fp 2 to keep registers after it properly word-aligned.
6327 (rl78_alloc_physical_registers_umul): Handle the case where both
6328 input operands are the same.
6330 2014-01-30 Richard Biener <rguenther@suse.de>
6332 PR tree-optimization/59903
6333 * tree-vect-loop.c (vect_transform_loop): Guard multiple-types
6336 2014-01-30 Jason Merrill <jason@redhat.com>
6339 * tree.c (walk_type_fields): Handle VECTOR_TYPE.
6342 * cgraphunit.c (expand_thunk): Copy volatile arg to a temporary.
6344 2014-01-30 Richard Biener <rguenther@suse.de>
6346 PR tree-optimization/59951
6347 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Skip uses in debug insns.
6349 2014-01-30 Savin Zlobec <savin.zlobec@gmail.com>
6352 * config/nios2/nios2.c (nios2_fpu_insn_asm): Fix asm output of
6353 SFmode to DFmode case.
6355 2014-01-29 DJ Delorie <dj@redhat.com>
6357 * config/msp430/msp430.opt (-minrt): New.
6358 * config/msp430/msp430.h (STARTFILE_SPEC): Link alternate runtime
6360 (ENDFILE_SPEC): Likewise.
6362 2014-01-29 Jan Hubicka <hubicka@ucw.cz>
6364 * ipa-inline-analysis.c (clobber_only_eh_bb_p): New function.
6365 (estimate_function_body_sizes): Use it.
6367 2014-01-29 Paolo Carlini <paolo.carlini@oracle.com>
6370 * dwarf2out.c (is_cxx_auto): New.
6371 (is_base_type): Use it.
6372 (gen_type_die_with_usage): Likewise.
6374 2014-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6376 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Use
6377 CODE_FOR_altivec_vmrg*_direct rather than CODE_FOR_altivec_vmrg*.
6378 * config/rs6000/vsx.md (vsx_mergel_<mode>): Adjust for
6379 -maltivec=be with LE targets.
6380 (vsx_mergeh_<mode>): Likewise.
6381 * config/rs6000/altivec.md (UNSPEC_VMRG[HL]_DIRECT): New unspecs.
6382 (mulv8hi3): Use gen_altivec_vmrg[hl]w_direct.
6383 (altivec_vmrghb): Replace with define_expand and new
6384 *altivec_vmrghb_internal insn; adjust for -maltivec=be with LE targets.
6385 (altivec_vmrghb_direct): New define_insn.
6386 (altivec_vmrghh): Replace with define_expand and new
6387 *altivec_vmrghh_internal insn; adjust for -maltivec=be with LE targets.
6388 (altivec_vmrghh_direct): New define_insn.
6389 (altivec_vmrghw): Replace with define_expand and new
6390 *altivec_vmrghw_internal insn; adjust for -maltivec=be with LE targets.
6391 (altivec_vmrghw_direct): New define_insn.
6392 (*altivec_vmrghsf): Adjust for endianness.
6393 (altivec_vmrglb): Replace with define_expand and new
6394 *altivec_vmrglb_internal insn; adjust for -maltivec=be with LE targets.
6395 (altivec_vmrglb_direct): New define_insn.
6396 (altivec_vmrglh): Replace with define_expand and new
6397 *altivec_vmrglh_internal insn; adjust for -maltivec=be with LE targets.
6398 (altivec_vmrglh_direct): New define_insn.
6399 (altivec_vmrglw): Replace with define_expand and new
6400 *altivec_vmrglw_internal insn; adjust for -maltivec=be with LE targets.
6401 (altivec_vmrglw_direct): New define_insn.
6402 (*altivec_vmrglsf): Adjust for endianness.
6403 (vec_widen_umult_hi_v16qi): Use gen_altivec_vmrghh_direct.
6404 (vec_widen_umult_lo_v16qi): Use gen_altivec_vmrglh_direct.
6405 (vec_widen_smult_hi_v16qi): Use gen_altivec_vmrghh_direct.
6406 (vec_widen_smult_lo_v16qi): Use gen_altivec_vmrglh_direct.
6407 (vec_widen_umult_hi_v8hi): Use gen_altivec_vmrghw_direct.
6408 (vec_widen_umult_lo_v8hi): Use gen_altivec_vmrglw_direct.
6409 (vec_widen_smult_hi_v8hi): Use gen_altivec_vmrghw_direct.
6410 (vec_widen_smult_lo_v8hi): Use gen_altivec_vmrglw_direct.
6412 2014-01-29 Marcus Shawcroft <marcus.shawcroft@arm.com>
6414 * config/aarch64/aarch64.c (aarch64_expand_mov_immediate)
6415 (aarch64_legitimate_address_p, aarch64_class_max_nregs): Adjust
6418 2014-01-29 Richard Biener <rguenther@suse.de>
6420 PR tree-optimization/58742
6421 * tree-ssa-forwprop.c (associate_pointerplus): Rename to
6422 associate_pointerplus_align.
6423 (associate_pointerplus_diff): New function.
6424 (associate_pointerplus): Likewise. Call associate_pointerplus_align
6425 and associate_pointerplus_diff.
6427 2014-01-29 Richard Biener <rguenther@suse.de>
6429 * lto-streamer.h (LTO_major_version): Bump to 3.
6430 (LTO_minor_version): Reset to 0.
6432 2014-01-29 Renlin Li <Renlin.Li@arm.com>
6434 * config/arm/arm-arches.def (ARM_ARCH): Add armv7ve arch.
6435 * config/arm/arm.c (FL_FOR_ARCH7VE): New.
6436 (arm_file_start): Generate correct asm header for armv7ve.
6437 * config/arm/bpabi.h: Add multilib support for armv7ve.
6438 * config/arm/driver-arm.c: Change the architectures of cortex-a7
6439 and cortex-a15 to armv7ve.
6440 * config/arm/t-aprofile: Add multilib support for armv7ve.
6441 * doc/invoke.texi: Document -march=armv7ve.
6443 2014-01-29 Richard Biener <rguenther@suse.de>
6445 PR tree-optimization/58742
6446 * tree-ssa-forwprop.c (associate_plusminus): Return true
6447 if we changed sth, defer EH cleanup to ...
6448 (ssa_forward_propagate_and_combine): ... here. Call simplify_mult.
6449 (simplify_mult): New function.
6451 2014-01-29 Jakub Jelinek <jakub@redhat.com>
6454 PR tree-optimization/59920
6455 * tree.c (build_common_builtin_nodes): Remove
6456 __builtin_setjmp_dispatcher initialization.
6457 * omp-low.h (make_gimple_omp_edges): Add a new int * argument.
6458 * profile.c (branch_prob): Use gsi_start_nondebug_after_labels_bb
6459 instead of gsi_after_labels + manually skipping debug stmts.
6460 Don't ignore bbs with BUILT_IN_SETJMP_DISPATCHER, instead
6461 ignore bbs with IFN_ABNORMAL_DISPATCHER.
6462 * tree-inline.c (copy_edges_for_bb): Remove
6463 can_make_abnormal_goto argument, instead add abnormal_goto_dest
6464 argument. Ignore computed_goto_p stmts. Don't call
6465 make_abnormal_goto_edges. If a call might need abnormal edges
6466 for non-local gotos, see if it already has an edge to
6467 IFN_ABNORMAL_DISPATCHER or if it is IFN_ABNORMAL_DISPATCHER
6468 with true argument, don't do anything then, otherwise add
6469 EDGE_ABNORMAL from the call's bb to abnormal_goto_dest.
6470 (copy_cfg_body): Compute abnormal_goto_dest, adjust copy_edges_for_bb
6472 * gimple-low.c (struct lower_data): Remove calls_builtin_setjmp.
6473 (lower_function_body): Don't emit __builtin_setjmp_dispatcher.
6474 (lower_stmt): Don't set data->calls_builtin_setjmp.
6475 (lower_builtin_setjmp): Adjust comment.
6476 * builtins.def (BUILT_IN_SETJMP_DISPATCHER): Remove.
6477 * tree-cfg.c (found_computed_goto): Remove.
6478 (factor_computed_gotos): Remove.
6479 (make_goto_expr_edges): Return bool, true for computed gotos.
6480 Don't call make_abnormal_goto_edges.
6481 (build_gimple_cfg): Don't set found_computed_goto, don't call
6482 factor_computed_gotos.
6483 (computed_goto_p): No longer static.
6484 (make_blocks): Don't set found_computed_goto.
6485 (get_abnormal_succ_dispatcher, handle_abnormal_edges): New functions.
6486 (make_edges): If make_goto_expr_edges returns true, push bb
6487 into ab_edge_goto vector, for stmt_can_make_abnormal_goto calls
6488 instead of calling make_abnormal_goto_edges push bb into ab_edge_call
6489 vector. Record mapping between bbs and OpenMP regions if there
6490 are any, adjust make_gimple_omp_edges caller. Call
6491 handle_abnormal_edges.
6492 (make_abnormal_goto_edges): Remove.
6493 * tree-cfg.h (make_abnormal_goto_edges): Remove.
6494 (computed_goto_p, get_abnormal_succ_dispatcher): New prototypes.
6495 * internal-fn.c (expand_ABNORMAL_DISPATCHER): New function.
6496 * builtins.c (expand_builtin): Don't handle BUILT_IN_SETJMP_DISPATCHER.
6497 * internal-fn.def (ABNORMAL_DISPATCHER): New.
6498 * omp-low.c (make_gimple_omp_edges): Add region_idx argument, when
6499 filling *region also set *region_idx to (*region)->entry->index.
6502 * read-rtl.c (read_rtx_code): Clear all of RTX_CODE_SIZE (code).
6503 For REGs set ORIGINAL_REGNO.
6505 2014-01-29 Bingfeng Mei <bmei@broadcom.com>
6507 * doc/md.texi: Mention that a target shouldn't implement
6508 vec_widen_(s|u)mul_even/odd pair if it is less efficient
6511 2014-01-29 Jakub Jelinek <jakub@redhat.com>
6513 PR tree-optimization/59594
6514 * tree-vect-data-refs.c (vect_analyze_data_ref_accesses): Sort
6515 a copy of the datarefs vector rather than the vector itself.
6517 2014-01-28 Jason Merrill <jason@redhat.com>
6520 * dwarf2out.c (auto_die): New static.
6521 (gen_type_die_with_usage): Handle C++1y 'auto'.
6522 (gen_subprogram_die): If in-class DIE had 'auto', emit type again
6525 2014-01-28 H.J. Lu <hongjiu.lu@intel.com>
6528 * config/i386/gnu-user64.h (SPEC_32): Add "m16|" to "m32".
6529 (SPEC_X32): Likewise.
6530 (SPEC_64): Likewise.
6531 * config/i386/i386.c (ix86_option_override_internal): Turn off
6532 OPTION_MASK_ISA_64BIT, OPTION_MASK_ABI_X32 and OPTION_MASK_ABI_64
6534 (x86_file_start): Output .code16gcc for TARGET_16BIT.
6535 * config/i386/i386.h (TARGET_16BIT): New macro.
6536 (TARGET_16BIT_P): Likewise.
6537 * config/i386/i386.opt: Add m16.
6538 * doc/invoke.texi: Document -m16.
6540 2014-01-28 Jakub Jelinek <jakub@redhat.com>
6542 PR preprocessor/59935
6543 * input.c (location_get_source_line): Bail out on when line number
6544 is zero, and test the return value of lookup_or_add_file_to_cache_tab.
6546 2014-01-28 Richard Biener <rguenther@suse.de>
6548 PR tree-optimization/58742
6549 * tree-ssa-forwprop.c (associate_plusminus): Handle
6550 pointer subtraction of the form (T)(P + A) - (T)P.
6552 2014-01-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6554 * config/arm/arm.c (arm_new_rtx_costs): Remove useless statement
6557 2014-01-28 Richard Biener <rguenther@suse.de>
6560 2014-01-28 Richard Biener <rguenther@suse.de>
6562 PR rtl-optimization/45364
6563 PR rtl-optimization/59890
6564 * var-tracking.c (local_get_addr_clear_given_value): Handle
6565 already cleared slot.
6566 (val_reset): Handle not allocated local_get_addr_cache.
6567 (vt_find_locations): Use post-order on the inverted CFG.
6569 2014-01-28 Richard Biener <rguenther@suse.de>
6571 * tree-data-ref.h (ddr_is_anti_dependent, ddrs_have_anti_deps): Remove.
6573 2014-01-28 Richard Biener <rguenther@suse.de>
6575 PR rtl-optimization/45364
6576 PR rtl-optimization/59890
6577 * var-tracking.c (local_get_addr_clear_given_value): Handle
6578 already cleared slot.
6579 (val_reset): Handle not allocated local_get_addr_cache.
6580 (vt_find_locations): Use post-order on the inverted CFG.
6582 2014-01-28 Alan Modra <amodra@gmail.com>
6584 * Makefile.in (BUILD_CPPFLAGS): Do not use ALL_CPPFLAGS.
6585 * configure.ac <recursive call for build != host>: Define
6586 GENERATOR_FILE. Comment. Use CXX_FOR_BUILD, CXXFLAGS_FOR_BUILD
6587 and LD_FOR_BUILD too.
6588 * configure: Regenerate.
6590 2014-01-27 Allan Sandfeld Jensen <sandfeld@kde.org>
6592 * config/i386/i386.c (get_builtin_code_for_version): Separate
6593 Westmere from Nehalem, Ivy Bridge from Sandy Bridge and
6594 Broadwell from Haswell.
6596 2014-01-27 Steve Ellcey <sellcey@mips.com>
6598 * common/config/mips/mips-common.c (TARGET_DEFAULT_TARGET_FLAGS):
6599 Remove TARGET_FP_EXCEPTIONS_DEFAULT and MASK_FUSED_MADD.
6600 * config/mips/mips.c (mips_option_override): Change setting
6602 * config/mips/mips.h (TARGET_FP_EXCEPTIONS_DEFAULT): Remove.
6603 * config/mips/mips.opt (DSP, DSPR2, FP_EXCEPTIONS, FUSED_MADD, MIPS3D):
6604 Change from Mask to Var.
6606 2014-01-27 Jeff Law <law@redhat.com>
6608 * ipa-inline.c (inline_small_functions): Fix typo.
6610 2014-01-27 Ilya Tocar <ilya.tocar@intel.com>
6612 * config/i386/avx512fintrin.h (_mm512_mask_cvtepi32_storeu_epi8): New.
6613 (_mm512_mask_cvtsepi32_storeu_epi8): Ditto.
6614 (_mm512_mask_cvtusepi32_storeu_epi8): Ditto.
6615 (_mm512_mask_cvtepi32_storeu_epi16): Ditto.
6616 (_mm512_mask_cvtsepi32_storeu_epi16): Ditto.
6617 (_mm512_mask_cvtusepi32_storeu_epi16): Ditto.
6618 (_mm512_mask_cvtepi64_storeu_epi32): Ditto.
6619 (_mm512_mask_cvtsepi64_storeu_epi32): Ditto.
6620 (_mm512_mask_cvtusepi64_storeu_epi32): Ditto.
6621 (_mm512_mask_cvtepi64_storeu_epi16): Ditto.
6622 (_mm512_mask_cvtsepi64_storeu_epi16): Ditto.
6623 (_mm512_mask_cvtusepi64_storeu_epi16): Ditto.
6624 (_mm512_mask_cvtepi64_storeu_epi8): Ditto.
6625 (_mm512_mask_cvtsepi64_storeu_epi8): Ditto.
6626 (_mm512_mask_cvtusepi64_storeu_epi8): Ditto.
6627 (_mm512_storeu_epi64): Ditto.
6628 (_mm512_cmpge_epi32_mask): Ditto.
6629 (_mm512_cmpge_epu32_mask): Ditto.
6630 (_mm512_cmpge_epi64_mask): Ditto.
6631 (_mm512_cmpge_epu64_mask): Ditto.
6632 (_mm512_cmple_epi32_mask): Ditto.
6633 (_mm512_cmple_epu32_mask): Ditto.
6634 (_mm512_cmple_epi64_mask): Ditto.
6635 (_mm512_cmple_epu64_mask): Ditto.
6636 (_mm512_cmplt_epi32_mask): Ditto.
6637 (_mm512_cmplt_epu32_mask): Ditto.
6638 (_mm512_cmplt_epi64_mask): Ditto.
6639 (_mm512_cmplt_epu64_mask): Ditto.
6640 (_mm512_cmpneq_epi32_mask): Ditto.
6641 (_mm512_cmpneq_epu32_mask): Ditto.
6642 (_mm512_cmpneq_epi64_mask): Ditto.
6643 (_mm512_cmpneq_epu64_mask): Ditto.
6644 (_mm512_expand_pd): Ditto.
6645 (_mm512_expand_ps): Ditto.
6646 * config/i386/i386-builtin-types.def: Add PV16QI, PV16QI, PV16HI,
6647 VOID_PV8SI_V8DI_QI, VOID_PV8HI_V8DI_QI, VOID_PV16QI_V8DI_QI,
6648 VOID_PV16QI_V16SI_HI, VOID_PV16HI_V16SI_HI.
6649 * config/i386/i386.c (ix86_builtins): Add
6650 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK,
6651 IX86_BUILTIN_PMOVDB512_MEM, IX86_BUILTIN_PMOVDW512_MEM,
6652 IX86_BUILTIN_PMOVQB512_MEM, IX86_BUILTIN_PMOVQD512_MEM,
6653 IX86_BUILTIN_PMOVQW512_MEM, IX86_BUILTIN_PMOVSDB512_MEM,
6654 IX86_BUILTIN_PMOVSDW512_MEM, IX86_BUILTIN_PMOVSQB512_MEM,
6655 IX86_BUILTIN_PMOVSQD512_MEM, IX86_BUILTIN_PMOVSQW512_MEM,
6656 IX86_BUILTIN_PMOVUSDB512_MEM, IX86_BUILTIN_PMOVUSDW512_MEM,
6657 IX86_BUILTIN_PMOVUSQB512_MEM, IX86_BUILTIN_PMOVUSQD512_MEM,
6658 IX86_BUILTIN_PMOVUSQW512_MEM.
6659 (bdesc_special_args): Add __builtin_ia32_pmovusqd512mem_mask,
6660 __builtin_ia32_pmovsqd512mem_mask,
6661 __builtin_ia32_pmovqd512mem_mask,
6662 __builtin_ia32_pmovusqw512mem_mask,
6663 __builtin_ia32_pmovsqw512mem_mask,
6664 __builtin_ia32_pmovqw512mem_mask,
6665 __builtin_ia32_pmovusdw512mem_mask,
6666 __builtin_ia32_pmovsdw512mem_mask,
6667 __builtin_ia32_pmovdw512mem_mask,
6668 __builtin_ia32_pmovqb512mem_mask,
6669 __builtin_ia32_pmovusqb512mem_mask,
6670 __builtin_ia32_pmovsqb512mem_mask,
6671 __builtin_ia32_pmovusdb512mem_mask,
6672 __builtin_ia32_pmovsdb512mem_mask,
6673 __builtin_ia32_pmovdb512mem_mask.
6674 (bdesc_args): Add __builtin_ia32_expanddf512,
6675 __builtin_ia32_expandsf512.
6676 (ix86_expand_special_args_builtin): Handle VOID_FTYPE_PV8SI_V8DI_QI,
6677 VOID_FTYPE_PV8HI_V8DI_QI, VOID_FTYPE_PV16HI_V16SI_HI,
6678 VOID_FTYPE_PV16QI_V8DI_QI, VOID_FTYPE_PV16QI_V16SI_HI.
6679 * config/i386/sse.md (unspec): Add UNSPEC_EXPAND_NOMASK.
6680 (avx512f_<code><pmov_src_lower><mode>2_mask_store): New.
6681 (*avx512f_<code>v8div16qi2_store_mask): Renamed to ...
6682 (avx512f_<code>v8div16qi2_mask_store): This.
6683 (avx512f_expand<mode>): New.
6685 2014-01-27 Kirill Yukhin <kirill.yukhin@intel.com>
6687 * config/i386/avx512pfintrin.h (_mm512_mask_prefetch_i32gather_pd):
6689 (_mm512_mask_prefetch_i64gather_pd): Ditto.
6690 (_mm512_prefetch_i32scatter_pd): Ditto.
6691 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
6692 (_mm512_prefetch_i64scatter_pd): Ditto.
6693 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
6694 (_mm512_mask_prefetch_i32gather_ps): Fix operand type.
6695 (_mm512_mask_prefetch_i64gather_ps): Ditto.
6696 (_mm512_prefetch_i32scatter_ps): Ditto.
6697 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
6698 (_mm512_prefetch_i64scatter_ps): Ditto.
6699 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
6700 * config/i386/i386-builtin-types.def: Define
6701 VOID_FTYPE_QI_V8SI_PCINT64_INT_INT
6702 and VOID_FTYPE_QI_V8DI_PCINT64_INT_INT.
6703 * config/i386/i386.c (ix86_builtins): Define IX86_BUILTIN_GATHERPFQPD,
6704 IX86_BUILTIN_GATHERPFDPD, IX86_BUILTIN_SCATTERPFDPD,
6705 IX86_BUILTIN_SCATTERPFQPD.
6706 (ix86_init_mmx_sse_builtins): Define __builtin_ia32_gatherpfdpd,
6707 __builtin_ia32_gatherpfdps, __builtin_ia32_gatherpfqpd,
6708 __builtin_ia32_gatherpfqps, __builtin_ia32_scatterpfdpd,
6709 __builtin_ia32_scatterpfdps, __builtin_ia32_scatterpfqpd,
6710 __builtin_ia32_scatterpfqps.
6711 (ix86_expand_builtin): Expand new built-ins.
6712 * config/i386/sse.md (avx512pf_gatherpf<mode>): Add SF suffix,
6713 fix memory access data type.
6714 (*avx512pf_gatherpf<mode>_mask): Ditto.
6715 (*avx512pf_gatherpf<mode>): Ditto.
6716 (avx512pf_scatterpf<mode>): Ditto.
6717 (*avx512pf_scatterpf<mode>_mask): Ditto.
6718 (*avx512pf_scatterpf<mode>): Ditto.
6719 (GATHER_SCATTER_SF_MEM_MODE): New.
6720 (avx512pf_gatherpf<mode>df): Ditto.
6721 (*avx512pf_gatherpf<mode>df_mask): Ditto.
6722 (*avx512pf_scatterpf<mode>df): Ditto.
6724 2014-01-27 Jakub Jelinek <jakub@redhat.com>
6727 * expmed.h (expmed_mode_index): Rework so that analysis and optimziers
6728 know when the MODE_PARTIAL_INT and MODE_VECTOR_INT cases can never be
6731 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
6733 * common/config/arm/arm-common.c
6734 (arm_rewrite_mcpu): Handle multiple names.
6736 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
6738 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
6740 * gimple-builder.h (create_gimple_tmp): Delete.
6742 2014-01-27 Christian Bruel <christian.bruel@st.com>
6744 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Fix remaining bytes after
6747 2014-01-26 John David Anglin <danglin@gcc.gnu.org>
6749 * config/pa/pa.md (call): Generate indirect long calls to non-local
6750 functions when outputing 32-bit code.
6751 (call_value): Likewise except for special call to buggy powf function.
6753 * config/pa/pa.c (pa_attr_length_indirect_call): Adjust length of
6754 portable runtime and PIC indirect calls.
6755 (pa_output_indirect_call): Remove unnecessary nop from portable runtime
6756 and PIC call sequences. Use ldo instead of blr to set return register
6757 in PIC call sequence.
6759 2014-01-25 Walter Lee <walt@tilera.com>
6761 * config/tilegx/sync.md (atomic_fetch_sub): Fix negation and
6762 avoid clobbering a live register.
6764 2014-01-25 Walter Lee <walt@tilera.com>
6766 * config/tilegx/tilegx-c.c (tilegx_cpu_cpp_builtins):
6767 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2}.
6768 * config/tilegx/tilepro-c.c (tilepro_cpu_cpp_builtins):
6769 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2,4,8}.
6771 2014-01-25 Walter Lee <walt@tilera.com>
6773 * config/tilegx/tilegx.c (tilegx_function_arg): Start 16-byte
6774 arguments on even registers.
6775 (tilegx_gimplify_va_arg_expr): Align 16-byte var args to
6777 * config/tilegx/tilegx.h (STACK_BOUNDARY): Change to 16 bytes.
6778 (BIGGEST_ALIGNMENT): Ditto.
6779 (BIGGEST_FIELD_ALIGNMENT): Ditto.
6781 2014-01-25 Walter Lee <walt@tilera.com>
6783 * config/tilegx/tilegx.c (tilegx_gen_bundles): Delete barrier
6784 insns before bundling.
6785 * config/tilegx/tilegx.md (tile_network_barrier): Update comment.
6787 2014-01-25 Walter Lee <walt@tilera.com>
6789 * config/tilegx/tilegx.c (tilegx_expand_builtin): Set
6790 PREFETCH_SCHEDULE_BARRIER_P to true for prefetches.
6791 * config/tilepro/tilepro.c (tilepro_expand_builtin): Ditto.
6793 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
6795 * config/mips/constraints.md (kl): Delete.
6796 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Turn into
6797 define expands, using...
6798 (divmod<mode>4_mips16, udivmod<mode>4_mips16): ...these new
6799 instructions for MIPS16.
6800 (*divmod<mode>4, *udivmod<mode>4): New patterns, taken from the
6801 non-MIPS16 version of the old divmod<mode>4 and udivmod<mode>4.
6803 2014-01-25 Walter Lee <walt@tilera.com>
6805 * config/tilepro/tilepro.md (ctzdi2): Use register_operand predicate.
6809 2014-01-25 Walter Lee <walt@tilera.com>
6811 * config/tilegx/tilegx.c (tilegx_expand_to_rtl_hook): New.
6812 (TARGET_EXPAND_TO_RTL_HOOK): Define.
6814 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
6816 * rtlanal.c (canonicalize_condition): Split out duplicated mode check.
6819 2014-01-25 Jakub Jelinek <jakub@redhat.com>
6821 * print-rtl.c (in_call_function_usage): New var.
6822 (print_rtx): When in CALL_INSN_FUNCTION_USAGE, always print
6823 EXPR_LIST mode as mode and not as reg note name.
6826 * cfgloopmanip.c (copy_loop_info): If
6827 loop->warned_aggressive_loop_optimizations, make sure
6828 the flag is set in target loop too.
6830 2014-01-24 Balaji V. Iyer <balaji.v.iyer@intel.com>
6832 * builtins.c (is_builtin_name): Renamed flag_enable_cilkplus to
6834 * builtins.def: Likewise.
6835 * cilk.h (fn_contains_cilk_spawn_p): Likewise.
6836 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Likewise.
6837 * ira.c (ira_setup_eliminable_regset): Likewise.
6838 * omp-low.c (gate_expand_omp): Likewise.
6839 (execute_lower_omp): Likewise.
6840 (diagnose_sb_0): Likewise.
6841 (gate_diagnose_omp_blocks): Likewise.
6842 (simd_clone_clauses_extract): Likewise.
6845 2014-01-24 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6847 * config/rs6000/rs6000.c (rs6000_expand_vec_perm_const_1): Remove
6848 correction for little endian...
6849 * config/rs6000/vsx.md (vsx_xxpermdi2_<mode>_1): ...and move it to
6852 2014-01-24 Jeff Law <law@redhat.com>
6854 PR tree-optimization/59919
6855 * tree-vrp.c (find_assert_locations_1): Do not register asserts
6856 for non-returning calls.
6858 2014-01-24 James Greenhalgh <james.greenhalgh@arm.com>
6860 * common/config/aarch64/aarch64-common.c
6861 (aarch64_rewrite_mcpu): Handle multiple names.
6862 * config/aarch64/aarch64.h
6863 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
6865 2014-01-24 Dodji Seketeli <dodji@redhat.com>
6867 * input.c (add_file_to_cache_tab): Handle the case where fopen
6870 2014-01-23 H.J. Lu <hongjiu.lu@intel.com>
6873 * config/i386/i386.md (pushsf splitter): Get stack adjustment
6874 from push operand if code of push isn't PRE_DEC.
6876 2014-01-23 Michael Meissner <meissner@linux.vnet.ibm.com>
6879 * doc/invoke.texi (RS/6000 and PowerPC Options): Document
6880 -mquad-memory-atomic. Update -mquad-memory documentation to say
6881 it is only used for non-atomic loads/stores.
6883 * config/rs6000/predicates.md (quad_int_reg_operand): Allow either
6884 -mquad-memory or -mquad-memory-atomic switches.
6886 * config/rs6000/rs6000-cpus.def (ISA_2_7_MASKS_SERVER): Add
6887 -mquad-memory-atomic to ISA 2.07 support.
6889 * config/rs6000/rs6000.opt (-mquad-memory-atomic): Add new switch
6890 to separate support of normal quad word memory operations (ldq, stq)
6891 from the atomic quad word memory operations.
6893 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
6894 support to separate non-atomic quad word operations from atomic
6895 quad word operations. Disable non-atomic quad word operations in
6896 little endian mode so that we don't have to swap words after the
6897 load and before the store.
6898 (quad_load_store_p): Add comment about atomic quad word support.
6899 (rs6000_opt_masks): Add -mquad-memory-atomic to the list of
6900 options printed with -mdebug=reg.
6902 * config/rs6000/rs6000.h (TARGET_SYNC_TI): Use
6903 -mquad-memory-atomic as the test for whether we have quad word
6904 atomic instructions.
6905 (TARGET_SYNC_HI_QI): If either -mquad-memory-atomic, -mquad-memory,
6906 or -mp8-vector are used, allow byte/half-word atomic operations.
6908 * config/rs6000/sync.md (load_lockedti): Insure that the address
6909 is a proper indexed or indirect address for the lqarx instruction.
6910 On little endian systems, swap the hi/lo registers after the lqarx
6912 (load_lockedpti): Use indexed_or_indirect_operand predicate to
6913 insure the address is valid for the lqarx instruction.
6914 (store_conditionalti): Insure that the address is a proper indexed
6915 or indirect address for the stqcrx. instruction. On little endian
6916 systems, swap the hi/lo registers before doing the stqcrx.
6918 (store_conditionalpti): Use indexed_or_indirect_operand predicate to
6919 insure the address is valid for the stqcrx. instruction.
6921 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
6922 Define __QUAD_MEMORY__ and __QUAD_MEMORY_ATOMIC__ based on what
6923 type of quad memory support is available.
6925 2014-01-23 Vladimir Makarov <vmakarov@redhat.com>
6928 * lra-constraints.c (simplify_operand_subreg): Spill pseudo if
6929 there is a danger of looping.
6931 2014-01-23 Pat Haugen <pthaugen@us.ibm.com>
6933 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
6934 force flag_ira_loop_pressure if set via command line.
6936 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
6938 * config/aarch64/aarch64-simd-builtins.def (ashr): DI mode removed.
6939 (ashr_simd): New builtin handling DI mode.
6940 * config/aarch64/aarch64-simd.md (aarch64_ashr_simddi): New pattern.
6941 (aarch64_sshr_simddi): New match pattern.
6942 * config/aarch64/arm_neon.h (vshr_n_s32): Builtin call modified.
6943 (vshrd_n_s64): Likewise.
6944 * config/aarch64/predicates.md (aarch64_shift_imm64_di): New predicate.
6946 2014-01-23 Nick Clifton <nickc@redhat.com>
6948 * config/msp430/msp430.h (ASM_SPEC): Pass the -mcpu as -mcpu.
6949 (LIB_SPEC): Drop use of memory.ld and peripherals.ld scripts in
6950 favour of mcu specific scripts.
6951 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more matches for
6954 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
6955 Alex Velenko <Alex.Velenko@arm.com>
6957 * config/aarch64/arm_neon.h (vaddv_s8): __LANE0 cleanup.
6958 (vaddv_s16): Likewise.
6959 (vaddv_s32): Likewise.
6960 (vaddv_u8): Likewise.
6961 (vaddv_u16): Likewise.
6962 (vaddv_u32): Likewise.
6963 (vaddvq_s8): Likewise.
6964 (vaddvq_s16): Likewise.
6965 (vaddvq_s32): Likewise.
6966 (vaddvq_s64): Likewise.
6967 (vaddvq_u8): Likewise.
6968 (vaddvq_u16): Likewise.
6969 (vaddvq_u32): Likewise.
6970 (vaddvq_u64): Likewise.
6971 (vaddv_f32): Likewise.
6972 (vaddvq_f32): Likewise.
6973 (vaddvq_f64): Likewise.
6974 (vmaxv_f32): Likewise.
6975 (vmaxv_s8): Likewise.
6976 (vmaxv_s16): Likewise.
6977 (vmaxv_s32): Likewise.
6978 (vmaxv_u8): Likewise.
6979 (vmaxv_u16): Likewise.
6980 (vmaxv_u32): Likewise.
6981 (vmaxvq_f32): Likewise.
6982 (vmaxvq_f64): Likewise.
6983 (vmaxvq_s8): Likewise.
6984 (vmaxvq_s16): Likewise.
6985 (vmaxvq_s32): Likewise.
6986 (vmaxvq_u8): Likewise.
6987 (vmaxvq_u16): Likewise.
6988 (vmaxvq_u32): Likewise.
6989 (vmaxnmv_f32): Likewise.
6990 (vmaxnmvq_f32): Likewise.
6991 (vmaxnmvq_f64): Likewise.
6992 (vminv_f32): Likewise.
6993 (vminv_s8): Likewise.
6994 (vminv_s16): Likewise.
6995 (vminv_s32): Likewise.
6996 (vminv_u8): Likewise.
6997 (vminv_u16): Likewise.
6998 (vminv_u32): Likewise.
6999 (vminvq_f32): Likewise.
7000 (vminvq_f64): Likewise.
7001 (vminvq_s8): Likewise.
7002 (vminvq_s16): Likewise.
7003 (vminvq_s32): Likewise.
7004 (vminvq_u8): Likewise.
7005 (vminvq_u16): Likewise.
7006 (vminvq_u32): Likewise.
7007 (vminnmv_f32): Likewise.
7008 (vminnmvq_f32): Likewise.
7009 (vminnmvq_f64): Likewise.
7011 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
7013 * config/aarch64/aarch64-simd.md
7014 (aarch64_dup_lane<mode>): Correct lane number on big-endian.
7015 (aarch64_dup_lane_<vswap_widthi_name><mode>): Likewise.
7016 (*aarch64_mul3_elt<mode>): Likewise.
7017 (*aarch64_mul3_elt<vswap_width_name><mode>): Likewise.
7018 (*aarch64_mul3_elt_to_64v2df): Likewise.
7019 (*aarch64_mla_elt<mode>): Likewise.
7020 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
7021 (*aarch64_mls_elt<mode>): Likewise.
7022 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
7023 (*aarch64_fma4_elt<mode>): Likewise.
7024 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
7025 (*aarch64_fma4_elt_to_64v2df): Likewise.
7026 (*aarch64_fnma4_elt<mode>): Likewise.
7027 (*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
7028 (*aarch64_fnma4_elt_to_64v2df): Likewise.
7029 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
7030 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
7031 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
7032 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
7033 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
7034 (aarch64_sqdmull_lane<mode>_internal): Likewise.
7035 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
7037 2013-01-23 Alex Velenko <Alex.Velenko@arm.com>
7039 * config/aarch64/aarch64-simd.md
7040 (aarch64_be_checked_get_lane<mode>): New define_expand.
7041 * config/aarch64/aarch64-simd-builtins.def
7042 (BUILTIN_VALL (GETLANE, be_checked_get_lane, 0)):
7043 New builtin definition.
7044 * config/aarch64/arm_neon.h: (__aarch64_vget_lane_any):
7045 Use new safe be builtin.
7047 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
7049 * config/aarch64/aarch64-simd.md (aarch64_be_ld1<mode>):
7051 (aarch64_be_st1<mode>): Likewise.
7052 (aarch_ld1<VALL:mode>): Define_expand modified.
7053 (aarch_st1<VALL:mode>): Likewise.
7054 * config/aarch64/aarch64.md (UNSPEC_LD1): New unspec definition.
7055 (UNSPEC_ST1): Likewise.
7057 2014-01-23 David Holsgrove <david.holsgrove@xilinx.com>
7059 * config/microblaze/microblaze.md: Add trap insn and attribute
7061 2014-01-23 Dodji Seketeli <dodji@redhat.com>
7063 PR preprocessor/58580
7064 * input.h (location_get_source_line): Take an additional line_size
7066 (void diagnostics_file_cache_fini): Declare new function.
7067 * input.c (struct fcache): New type.
7068 (fcache_tab_size, fcache_buffer_size, fcache_line_record_size):
7069 New static constants.
7070 (diagnostic_file_cache_init, total_lines_num)
7071 (lookup_file_in_cache_tab, evicted_cache_tab_entry)
7072 (add_file_to_cache_tab, lookup_or_add_file_to_cache_tab)
7073 (needs_read, needs_grow, maybe_grow, read_data, maybe_read_data)
7074 (get_next_line, read_next_line, goto_next_line, read_line_num):
7075 New static function definitions.
7076 (diagnostic_file_cache_fini): New function.
7077 (location_get_source_line): Take an additional output line_len
7078 parameter. Re-write using lookup_or_add_file_to_cache_tab and
7080 * diagnostic.c (diagnostic_finish): Call
7081 diagnostic_file_cache_fini.
7082 (adjust_line): Take an additional input parameter for the length
7083 of the line, rather than calculating it with strlen.
7084 (diagnostic_show_locus): Adjust the use of
7085 location_get_source_line and adjust_line with respect to their new
7086 signature. While displaying a line now, do not stop at the first
7087 null byte. Rather, display the zero byte as a space and keep
7088 going until we reach the size of the line.
7089 * Makefile.in: Add vec.o to OBJS-libcommon
7091 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
7092 Ilya Tocar <ilya.tocar@intel.com>
7094 * config/i386/avx512fintrin.h (_mm512_kmov): New.
7095 * config/i386/i386.c (IX86_BUILTIN_KMOV16): Ditto.
7096 (__builtin_ia32_kmov16): Ditto.
7097 * config/i386/i386.md (UNSPEC_KMOV): New.
7100 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
7102 * config/i386/avx512fintrin.h (_mm512_loadu_si512): Rename.
7103 (_mm512_storeu_si512): Ditto.
7105 2014-01-23 Richard Sandiford <rdsandiford@googlemail.com>
7108 * rtl.h (get_referenced_operands): Declare.
7109 * recog.c (get_referenced_operands): New function.
7110 * config/mips/mips.c (mips_reorg_process_insns): Check which asm
7111 operands have been referenced when recording LO_SUM references.
7113 2014-01-22 David Holsgrove <david.holsgrove@xilinx.com>
7115 * config/microblaze/microblaze.md: Correct bswaphi2 insn.
7117 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
7119 * config/i386/x86-tune.def (X86_TUNE_ACCUMULATE_OUTGOING_ARGS):
7120 Enable for generic and recent AMD targets.
7122 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
7124 * combine-stack-adj.c (combine_stack_adjustments_for_block): Remove
7125 ARG_SIZE note when adjustment was eliminated.
7127 2014-01-22 Jeff Law <law@redhat.com>
7129 PR tree-optimization/59597
7130 * tree-ssa-threadupdate.c (dump_jump_thread_path): Move to earlier
7131 in file. Accept new argument REGISTERING and use it to modify
7132 dump output appropriately.
7133 (register_jump_thread): Corresponding changes.
7134 (mark_threaded_blocks): Reinstate code to cancel unprofitable
7135 thread paths involving joiner blocks. Add code to dump cancelled
7136 jump threading paths.
7138 2014-01-22 Vladimir Makarov <vmakarov@redhat.com>
7140 PR rtl-optimization/59477
7141 * lra-constraints.c (inherit_in_ebb): Process call for living hard
7142 regs. Update reloads_num and potential_reload_hard_regs for all insns.
7144 2014-01-22 Tom Tromey <tromey@redhat.com>
7146 * config/i386/i386-interix.h (i386_pe_unique_section): Don't use
7148 * config/cr16/cr16-protos.h (notice_update_cc): Don't use PARAMS.
7150 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
7152 PR rtl-optimization/59896
7153 * lra-constraints.c (process_alt_operands): Check unused note for
7154 matched operands of insn with no output reloads.
7156 2014-01-21 Richard Sandiford <rdsandiford@googlemail.com>
7158 * config/mips/mips.c (mips_move_to_gpr_cost): Add M16_REGS case.
7159 (mips_move_from_gpr_cost): Likewise.
7161 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
7163 PR rtl-optimization/59858
7164 * lra-constraints.c (SMALL_REGISTER_CLASS_P): Use
7165 ira_class_hard_regs_num.
7166 (process_alt_operands): Increase reject for dying matched operand.
7168 2014-01-21 Jakub Jelinek <jakub@redhat.com>
7171 * config/i386/i386.c (expand_small_movmem_or_setmem): If mode is
7172 smaller than size, perform several stores or loads and stores
7173 at dst + count - size to store or copy all of size bytes, rather
7174 than just last modesize bytes.
7176 2014-01-20 DJ Delorie <dj@redhat.com>
7178 * config/rl78/rl78.c (rl78_propogate_register_origins): Verify
7179 that CLOBBERs are REGs before propogating their values.
7181 2014-01-20 H.J. Lu <hongjiu.lu@intel.com>
7184 * cgraph.c (cgraph_inline_failed_string): Add type to DEFCIFCODE.
7185 (cgraph_inline_failed_type): New function.
7186 * cgraph.h (DEFCIFCODE): Add type.
7187 (cgraph_inline_failed_type_t): New enum.
7188 (cgraph_inline_failed_type): New prototype.
7189 * cif-code.def: Add CIF_FINAL_NORMAL to OK, FUNCTION_NOT_CONSIDERED,
7190 FUNCTION_NOT_OPTIMIZED, REDEFINED_EXTERN_INLINE,
7191 FUNCTION_NOT_INLINE_CANDIDATE, LARGE_FUNCTION_GROWTH_LIMIT,
7192 LARGE_STACK_FRAME_GROWTH_LIMIT, MAX_INLINE_INSNS_SINGLE_LIMIT,
7193 MAX_INLINE_INSNS_AUTO_LIMIT, INLINE_UNIT_GROWTH_LIMIT,
7194 RECURSIVE_INLINING, UNLIKELY_CALL, NOT_DECLARED_INLINED,
7195 OPTIMIZING_FOR_SIZE, ORIGINALLY_INDIRECT_CALL,
7196 INDIRECT_UNKNOWN_CALL, USES_COMDAT_LOCAL.
7197 Add CIF_FINAL_ERROR to UNSPECIFIED, BODY_NOT_AVAILABLE,
7198 FUNCTION_NOT_INLINABLE, OVERWRITABLE, MISMATCHED_ARGUMENTS,
7199 EH_PERSONALITY, NON_CALL_EXCEPTIONS, TARGET_OPTION_MISMATCH,
7200 OPTIMIZATION_MISMATCH.
7201 * tree-inline.c (expand_call_inline): Emit errors during
7202 early_inlining if cgraph_inline_failed_type returns CIF_FINAL_ERROR.
7204 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
7207 * config/i386/sse.md (*andnot<mode>3<mask_name>): Handle MODE_V16SF
7208 mode attribute in insn output.
7210 2014-01-20 Eric Botcazou <ebotcazou@adacore.com>
7212 * output.h (output_constant): Delete.
7213 * varasm.c (output_constant): Make private.
7215 2014-01-20 Alex Velenko <Alex.Velenko@arm.com>
7217 * config/aarch64/aarch64-simd.md (vec_perm<mode>): Add BE check.
7219 2014-01-20 Jakub Jelinek <jakub@redhat.com>
7222 * tree.h (fold_builtin_strcat): New prototype.
7223 * builtins.c (fold_builtin_strcat): No longer static. Add len
7224 argument, if non-NULL, don't call c_strlen. Optimize
7225 directly into __builtin_memcpy instead of __builtin_strcpy.
7226 (fold_builtin_2): Adjust fold_builtin_strcat caller.
7227 * gimple-fold.c (gimple_fold_builtin): Handle BUILT_IN_STRCAT.
7229 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
7231 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
7232 for SImode_address_operand operands, having only a REG argument.
7234 2014-01-20 Marcus Shawcroft <marcus.shawcroft@arm.com>
7236 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER): Expand
7237 loader name using mbig-endian.
7238 (LINUX_TARGET_LINK_SPEC): Pass linker -m flag.
7240 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
7242 * doc/invoke.texi (-march): Clarify documentation for AArch64.
7246 2014-01-20 Tejas Belagod <tejas.belagod@arm.com>
7248 * config/aarch64/aarch64-protos.h
7249 (aarch64_cannot_change_mode_class_ptr): Declare.
7250 * config/aarch64/aarch64.c (aarch64_cannot_change_mode_class,
7251 aarch64_cannot_change_mode_class_ptr): New.
7252 * config/aarch64/aarch64.h (CANNOT_CHANGE_MODE_CLASS): Change to call
7253 backend hook aarch64_cannot_change_mode_class.
7255 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
7257 * common/config/aarch64/aarch64-common.c
7258 (aarch64_handle_option): Don't handle any option order logic here.
7259 * config/aarch64/aarch64.c (aarch64_parse_arch): Do not override
7260 selected_cpu, warn on architecture version mismatch.
7261 (aarch64_override_options): Fix parsing order for option strings.
7263 2014-01-20 Jan-Benedict Glaw <jbglaw@lug-owl.de>
7264 Iain Sandoe <iain@codesourcery.com>
7267 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Fix unused variable
7268 warning. Amend comment to reflect current functionality.
7270 2014-01-20 Richard Biener <rguenther@suse.de>
7273 * builtins.c (fold_builtin_strcat): Remove case better handled
7274 by tree-ssa-strlen.c.
7276 2014-01-20 Alan Lawrence <alan.lawrence@arm.com>
7278 * config/aarch64/aarch64.opt
7279 (mcpu, march, mtune): Make case-insensitive.
7281 2014-01-20 Jakub Jelinek <jakub@redhat.com>
7284 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
7285 if operands[1] is a REG or ZERO_EXTEND of a REG.
7287 2014-01-19 Jan Hubicka <hubicka@ucw.cz>
7289 * varasm.c (compute_reloc_for_constant): Use targetm.binds_local_p.
7291 2014-01-19 John David Anglin <danglin@gcc.gnu.org>
7293 * config/pa/pa.c (pa_attr_length_millicode_call): Correct length of
7294 long non-pic millicode calls.
7296 2014-01-19 Jan-Benedict Glaw <jbglaw@lug-owl.de>
7298 * config/vax/vax.h (FUNCTION_ARG_REGNO_P): Fix unused variable warning.
7300 2014-01-19 Kito Cheng <kito@0xlab.org>
7302 * builtins.c (expand_movstr): Check movstr expand done or fail.
7304 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
7305 H.J. Lu <hongjiu.lu@intel.com>
7308 * config/i386/i386.md (*lea<mode>): Zero-extend return register
7309 to DImode for zero-extended addresses.
7311 2014-01-19 Jakub Jelinek <jakub@redhat.com>
7313 PR rtl-optimization/57763
7314 * bb-reorder.c (fix_crossing_unconditional_branches): Set JUMP_LABEL
7315 on the new indirect jump_insn and increment LABEL_NUSES (label).
7317 2014-01-18 H.J. Lu <hongjiu.lu@intel.com>
7321 * config.gcc (x86_archs): New variable.
7322 (x86_64_archs): Likewise.
7323 (x86_cpus): Likewise.
7324 Use $x86_archs, $x86_64_archs and $x86_cpus to check valid
7325 --with-arch/--with-cpu= options.
7326 Support --with-arch=/--with-cpu={nehalem,westmere,
7327 sandybridge,ivybridge,haswell,broadwell,bonnell,silvermont}.
7329 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
7331 * config/i386/i386.c (ix86_adjust_cost): Reorder PROCESSOR_K8
7332 and PROCESSOR_ATHLON to simplify code. Move "memory" calculation.
7334 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
7336 * config/i386/i386.md (*swap<mode>): Rename from swap<mode>.
7338 2014-01-18 Jakub Jelinek <jakub@redhat.com>
7341 * config/i386/i386-c.c (ix86_pragma_target_parse): Temporarily
7342 clear cpp_get_options (parse_in)->warn_unused_macros for
7343 ix86_target_macros_internal with cpp_define.
7345 2014-01-18 Richard Sandiford <rdsandiford@googlemail.com>
7347 * jump.c (delete_related_insns): Keep (use (insn))s.
7348 * reorg.c (redundant_insn): Check for barriers too.
7350 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
7352 * config/i386/i386.c (ix86_split_lea_for_addr): Fix a comment typo.
7354 2014-01-17 John David Anglin <danglin@gcc.gnu.org>
7356 * config/pa/pa.c (pa_attr_length_indirect_call): Don't output a short
7357 call to $$dyncall when TARGET_LONG_CALLS is true.
7359 2014-01-17 Jeff Law <law@redhat.com>
7361 * ree.c (combine_set_extension): Temporarily disable test for
7362 changing number of hard registers.
7364 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
7367 * ipa-inline-analysis.c (inline_free_summary):
7368 Do not free summary of aliases.
7370 2014-01-17 Jakub Jelinek <jakub@redhat.com>
7373 * gimplify.c (gimplify_expr): Use create_tmp_var
7374 instead of create_tmp_var_raw. If cond doesn't have
7375 integral type, don't add the IFN_ANNOTATE builtin at all.
7377 2014-01-17 Martin Jambor <mjambor@suse.cz>
7380 * ipa-cp.c (prev_edge_clone): New variable.
7381 (grow_next_edge_clone_vector): Renamed to grow_edge_clone_vectors.
7382 Also resize prev_edge_clone vector.
7383 (ipcp_edge_duplication_hook): Also update prev_edge_clone.
7384 (ipcp_edge_removal_hook): New function.
7385 (ipcp_driver): Register ipcp_edge_removal_hook.
7387 2014-01-17 Andrew Pinski <apinski@cavium.com>
7388 Steve Ellcey <sellcey@mips.com>
7391 * config/mips/mips.c (mips_print_operand): Check operand mode instead
7394 2014-01-17 Jeff Law <law@redhat.com>
7397 * passes.def: Reorder pass_copy_prop, pass_unrolli, pass_ccp sequence
7398 so that pass_ccp runs first.
7400 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
7402 * config/i386/i386.c (ix86_lea_outperforms): Use TARGET_XXX.
7403 (ix86_adjust_cost): Use !TARGET_XXX.
7404 (do_reorder_for_imul): Likewise.
7405 (swap_top_of_ready_list): Likewise.
7406 (ix86_sched_reorder): Likewise.
7408 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
7410 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
7411 PROCESSOR_INTEL. Treat like PROCESSOR_GENERIC.
7412 * config/i386/i386.c (intel_memcpy): New. Duplicate slm_memcpy.
7413 (intel_memset): New. Duplicate slm_memset.
7414 (intel_cost): New. Duplicate slm_cost.
7415 (m_INTEL): New macro.
7416 (processor_target_table): Add "intel".
7417 (ix86_option_override_internal): Replace PROCESSOR_SILVERMONT
7418 with PROCESSOR_INTEL for "intel".
7419 (ix86_lea_outperforms): Support PROCESSOR_INTEL. Duplicate
7420 PROCESSOR_SILVERMONT.
7421 (ix86_issue_rate): Likewise.
7422 (ix86_adjust_cost): Likewise.
7423 (ia32_multipass_dfa_lookahead): Likewise.
7424 (swap_top_of_ready_list): Likewise.
7425 (ix86_sched_reorder): Likewise.
7426 (ix86_avoid_lea_for_addr): Check TARGET_AVOID_LEA_FOR_ADDR
7427 instead of TARGET_OPT_AGU.
7428 * config/i386/i386.h (TARGET_INTEL): New.
7429 (TARGET_AVOID_LEA_FOR_ADDR): Likewise.
7430 (processor_type): Add PROCESSOR_INTEL.
7431 * config/i386/x86-tune.def: Support m_INTEL. Duplicate m_SILVERMONT.
7432 Add X86_TUNE_AVOID_LEA_FOR_ADDR.
7434 2014-01-17 Marek Polacek <polacek@redhat.com>
7437 * gimple-fold.c (fold_array_ctor_reference): Don't fold if element
7440 2014-01-17 Richard Biener <rguenther@suse.de>
7442 PR tree-optimization/46590
7443 * opts.c (default_options_table): Add entries for
7444 OPT_fbranch_count_reg, OPT_fmove_loop_invariants and OPT_ftree_pta,
7445 all enabled at -O1 but not for -Og.
7446 * common.opt (fbranch-count-reg): Remove Init(1).
7447 (fmove-loop-invariants): Likewise.
7448 (ftree-pta): Likewise.
7450 2014-01-17 Jakub Jelinek <jakub@redhat.com>
7452 * config/i386/i386.c (ix86_data_alignment): For compatibility with
7453 (incorrect) GCC 4.8 and earlier alignment assumptions ensure we align
7454 decls to at least the GCC 4.8 used alignments.
7457 * tree-nested.c (convert_nonlocal_reference_stmt,
7458 convert_local_reference_stmt): For NAMELIST_DECLs in gimple_bind_vars
7459 of GIMPLE_BIND stmts, adjust associated decls.
7461 2014-01-17 Richard Biener <rguenther@suse.de>
7463 PR tree-optimization/46590
7464 * vec.h (vec<>::bseach): New member function implementing
7465 binary search according to C89 bsearch.
7466 (vec<>::qsort): Avoid calling ::qsort for vectors with sizes 0 or 1.
7467 * tree-ssa-loop-im.c (struct mem_ref): Make stored member a
7468 bitmap pointer again. Make accesses_in_loop a flat array.
7469 (mem_ref_obstack): New global.
7470 (outermost_indep_loop): Adjust for mem_ref->stored changes.
7471 (mark_ref_stored): Likewise.
7472 (ref_indep_loop_p_2): Likewise.
7473 (set_ref_stored_in_loop): New helper function.
7474 (mem_ref_alloc): Allocate mem_refs on the mem_ref_obstack obstack.
7475 (memref_free): Adjust.
7476 (record_mem_ref_loc): Simplify.
7477 (gather_mem_refs_stmt): Adjust.
7478 (sort_locs_in_loop_postorder_cmp): New function.
7479 (analyze_memory_references): Sort accesses_in_loop after
7480 loop postorder number.
7481 (find_ref_loc_in_loop_cmp): New function.
7482 (for_all_locs_in_loop): Find relevant cluster of locs in
7483 accesses_in_loop and iterate without recursion.
7484 (execute_sm): Avoid uninit warning.
7485 (struct ref_always_accessed): Simplify.
7486 (ref_always_accessed::operator ()): Likewise.
7487 (ref_always_accessed_p): Likewise.
7488 (tree_ssa_lim_initialize): Initialize mem_ref_obstack, compute
7489 loop postorder numbers here.
7490 (tree_ssa_lim_finalize): Free mem_ref_obstack and loop postorder
7493 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
7496 * passes.c (rest_of_decl_compilation): Don't call varpool_finalize_decl
7497 on decls for which assemble_alias has been called.
7499 2014-01-17 Nick Clifton <nickc@redhat.com>
7501 * config/msp430/msp430.opt: (mcpu): New option.
7502 * config/msp430/msp430.c (msp430_mcu_name): Use target_mcu.
7503 (msp430_option_override): Parse target_cpu. If the MCU name
7504 matches a generic string, clear target_mcu.
7505 (msp430_attr): Allow numeric interrupt values up to 63.
7506 (msp430_expand_epilogue): No longer invert operand 1 of gen_popm.
7507 * config/msp430/msp430.h (ASM_SPEC): Convert -mcpu into a -mmcu
7509 * config/msp430/t-msp430: (MULTILIB_MATCHES): Remove mcu matches.
7511 * config/msp430/msp430.md (popm): Use %J rather than %I.
7512 (addsi3): Use msp430_nonimmediate_operand for operand 2.
7513 (addhi_cy_i): Use immediate_operand for operand 2.
7514 * doc/invoke.texi: Document -mcpu option.
7516 2014-01-17 Richard Biener <rguenther@suse.de>
7518 PR rtl-optimization/38518
7519 * df.h (df_analyze_loop): Declare.
7520 * df-core.c: Include cfgloop.h.
7521 (df_analyze_1): Split out main part of df_analyze.
7522 (df_analyze): Adjust.
7523 (loop_inverted_post_order_compute): New function.
7524 (loop_post_order_compute): Likewise.
7525 (df_analyze_loop): New function avoiding whole-function
7527 * loop-invariant.c (find_defs): Use df_analyze_loop.
7528 (find_invariants): Adjust.
7529 * loop-iv.c (iv_analysis_loop_init): Use df_analyze_loop.
7531 2014-01-17 Zhenqiang Chen <zhenqiang.chen@arm.com>
7533 * config/arm/arm.c (arm_v7m_tune): Set max_insns_skipped to 2.
7534 (thumb2_final_prescan_insn): Set max to MAX_INSN_PER_IT_BLOCK.
7536 2014-01-16 Ilya Enkovich <ilya.enkovich@intel.com>
7538 * ipa-ref.c (ipa_remove_stmt_references): Fix references
7539 traversal when removing references.
7541 2014-01-16 Jan Hubicka <hubicka@ucw.cz>
7544 * tree.c (get_binfo_at_offset): Look harder for virtual bases.
7546 2014-01-16 Bernd Schmidt <bernds@codesourcery.com>
7549 * reload.c (find_reloads_address_1): Do not use RELOAD_OTHER when
7550 pushing a reload for an autoinc when we had previously reloaded an
7551 inner part of the address.
7553 2014-01-16 Jakub Jelinek <jakub@redhat.com>
7555 * tree-vectorizer.h (struct _loop_vec_info): Add no_data_dependencies
7557 (LOOP_VINFO_NO_DATA_DEPENDENCIES): Define.
7558 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Clear it
7559 when not giving up or versioning for alias only because of
7561 (vect_analyze_data_ref_dependences): Set to true.
7562 * tree-vect-stmts.c (hoist_defs_of_uses): Return false if def_stmt
7564 (vectorizable_load): Use LOOP_VINFO_NO_DATA_DEPENDENCIES instead of
7565 LOOP_REQUIRES_VERSIONING_FOR_ALIAS, add && !nested_in_vect_loop
7569 * expr.c (expand_expr_real_1): Handle init == NULL_TREE.
7572 * config/i386/i386.c (ix86_expand_builtin): If target doesn't satisfy
7573 operand 0 predicate for gathers, use a new pseudo as subtarget.
7575 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
7578 * lra-constraints.c (process_alt_operands): Add printing debug info.
7579 Check absence of input/output reloads for matched operands too.
7581 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
7583 PR rtl-optimization/59835
7584 * ira.c (ira_init_register_move_cost): Increase cost for
7587 2014-01-16 Alan Lawrence <alan.lawrence@arm.com>
7589 * config/arm/arm.opt (mcpu, march, mtune): Make case-insensitive.
7591 2014-01-16 Richard Earnshaw <rearnsha@arm.com>
7594 * aarch64.c (aarch64_split_128bit_move): Don't lookup REGNO on
7595 non-register objects. Use gen_(high/low)part more consistently.
7598 2014-01-16 Michael Meissner <meissner@linux.vnet.ibm.com>
7601 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Add little
7602 endian support, remove tests for WORDS_BIG_ENDIAN.
7603 (p8_mfvsrd_3_<mode>): Likewise.
7604 (reload_gpr_from_vsx<mode>): Likewise.
7605 (reload_gpr_from_vsxsf): Likewise.
7606 (p8_mfvsrd_4_disf): Likewise.
7608 2014-01-16 Richard Biener <rguenther@suse.de>
7610 PR rtl-optimization/46590
7611 * lcm.c (compute_antinout_edge): Use postorder iteration.
7612 (compute_laterin): Use inverted postorder iteration.
7614 2014-01-16 Nick Clifton <nickc@redhat.com>
7617 * varasm.c (output_constant): Return the number of bytes actually
7619 (output_constructor_array_range): Update the field size with the
7620 number of bytes emitted by output_constant.
7621 (output_constructor_regular_field): Likewise. Also do not
7622 complain if the total number of bytes emitted is now greater
7623 than the expected fieldpos.
7624 * output.h (output_constant): Update prototype and descriptive comment.
7626 2014-01-16 Marek Polacek <polacek@redhat.com>
7629 * cgraph.c (gimple_check_call_args): Don't use DECL_ARG_TYPE if
7630 it is error_mark_node.
7632 2014-01-15 Uros Bizjak <ubizjak@gmail.com>
7634 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use
7635 VALID_AVX256_REG_OR_OI_MODE.
7637 2014-01-15 Pat Haugen <pthaugen@us.ibm.com>
7639 * config/rs6000/rs6000.c (rs6000_output_function_prologue): Check if
7640 current procedure should be profiled.
7642 2014-01-15 Andrew Pinski <apinski@cavium.com>
7644 * config/aarch64/aarch64.c (aarch64_register_move_cost): Correct cost
7645 of moving from/to the STACK_REG register class.
7647 2014-01-15 Richard Henderson <rth@redhat.com>
7650 * reginfo.c (global_regs_decl): Globalize.
7651 * rtl.h (global_regs_decl): Declare.
7652 * ira.c (do_reload): Diagnose frame_pointer_needed and it
7653 reserved via global_regs.
7655 2014-01-15 Teresa Johnson <tejohnson@google.com>
7657 * tree-ssa-sccvn.c (visit_reference_op_call): Handle NULL vdef.
7659 2014-01-15 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
7661 * config/rs6000/altivec.md (mulv8hi3): Explicitly generate vmulesh
7662 and vmulosh rather than call gen_vec_widen_smult_*.
7663 (vec_widen_umult_even_v16qi): Test VECTOR_ELT_ORDER_BIG rather
7664 than BYTES_BIG_ENDIAN to determine use of even or odd instruction.
7665 (vec_widen_smult_even_v16qi): Likewise.
7666 (vec_widen_umult_even_v8hi): Likewise.
7667 (vec_widen_smult_even_v8hi): Likewise.
7668 (vec_widen_umult_odd_v16qi): Likewise.
7669 (vec_widen_smult_odd_v16qi): Likewise.
7670 (vec_widen_umult_odd_v8hi): Likewise.
7671 (vec_widen_smult_odd_v8hi): Likewise.
7672 (vec_widen_umult_hi_v16qi): Explicitly generate vmuleub and
7673 vmuloub rather than call gen_vec_widen_umult_*.
7674 (vec_widen_umult_lo_v16qi): Likewise.
7675 (vec_widen_smult_hi_v16qi): Explicitly generate vmulesb and
7676 vmulosb rather than call gen_vec_widen_smult_*.
7677 (vec_widen_smult_lo_v16qi): Likewise.
7678 (vec_widen_umult_hi_v8hi): Explicitly generate vmuleuh and vmulouh
7679 rather than call gen_vec_widen_umult_*.
7680 (vec_widen_umult_lo_v8hi): Likewise.
7681 (vec_widen_smult_hi_v8hi): Explicitly gnerate vmulesh and vmulosh
7682 rather than call gen_vec_widen_smult_*.
7683 (vec_widen_smult_lo_v8hi): Likewise.
7685 2014-01-15 Jeff Law <law@redhat.com>
7687 PR tree-optimization/59747
7688 * ree.c (find_and_remove_re): Properly handle case where a second
7689 eliminated extension requires widening a copy created for elimination
7690 of a prior extension.
7691 (combine_set_extension): Ensure that the number of hard regs needed
7692 for a destination register does not change when we widen it.
7694 2014-01-15 Sebastian Huber <sebastian.huber@embedded-brains.de>
7696 * config.gcc (*-*-rtems*): Add t-rtems to tmake_file.
7697 (arm*-*-uclinux*eabi*): Do not override an existing tmake_file.
7698 (arm*-*-eabi* | arm*-*-symbianelf* | arm*-*-rtems*): Likwise.
7699 (arm*-*-rtems*): Use t-rtems from existing tmake_file.
7700 (avr-*-rtems*): Likewise.
7701 (bfin*-rtems*): Likewise.
7702 (moxie-*-rtems*): Likewise.
7703 (h8300-*-rtems*): Likewise.
7704 (i[34567]86-*-rtems*): Likewise.
7705 (lm32-*-rtems*): Likewise.
7706 (m32r-*-rtems*): Likewise.
7707 (m68k-*-rtems*): Likewise.
7708 (microblaze*-*-rtems*): Likewise.
7709 (mips*-*-rtems*): Likewise.
7710 (powerpc-*-rtems*): Likewise.
7711 (sh-*-rtems*): Likewise.
7712 (sparc-*-rtems*): Likewise.
7713 (sparc64-*-rtems*): Likewise.
7714 (v850-*-rtems*): Likewise.
7715 (m32c-*-rtems*): Likewise.
7717 2014-01-15 Vladimir Makarov <vmakarov@redhat.com>
7719 PR rtl-optimization/59511
7720 * ira.c (ira_init_register_move_cost): Use memory costs for some
7721 cases of register move cost calculations.
7722 * lra-constraints.c (lra_constraints): Use REG_FREQ_FROM_BB
7723 instead of BB frequency.
7724 * lra-coalesce.c (move_freq_compare_func, lra_coalesce): Ditto.
7725 * lra-assigns.c (find_hard_regno_for): Ditto.
7727 2014-01-15 Richard Biener <rguenther@suse.de>
7729 PR tree-optimization/59822
7730 * tree-vect-stmts.c (hoist_defs_of_uses): New function.
7731 (vectorizable_load): Use it to hoist defs of uses of invariant
7732 loads out of the loop.
7734 2014-01-15 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
7735 Kugan Vivekanandarajah <kuganv@linaro.org>
7738 * config/aarch64/aarch64.c (aarch64_build_constant): Fix incorrect
7741 2014-01-15 Richard Biener <rguenther@suse.de>
7743 PR rtl-optimization/59802
7744 * lcm.c (compute_available): Use inverted postorder to seed
7745 the initial worklist.
7747 2014-01-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7750 * config/s390/s390.c (s390_preferred_reload_class): Don't return
7751 ADDR_REGS for invalid symrefs in non-PIC code.
7753 2014-01-15 Jakub Jelinek <jakub@redhat.com>
7756 * builtins.c (determine_block_size): Initialize *probable_max_size
7757 even if len_rtx is CONST_INT.
7759 2014-01-14 Andrew Pinski <apinski@cavium.com>
7761 * config/aarch64/aarch64-protos.h (tune_params): Add issue_rate.
7762 * config/aarch64/aarch64.c (generic_tunings): Add issue rate of 2.
7763 (cortexa53_tunings): Likewise.
7764 (aarch64_sched_issue_rate): New function.
7765 (TARGET_SCHED_ISSUE_RATE): Define.
7767 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
7769 * ira-costs.c (find_costs_and_classes): Add missed
7770 ira_init_register_move_cost_if_necessary.
7772 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
7775 * config/arm/arm.c (arm_coproc_mem_operand): Add lra_in_progress.
7777 2014-01-14 H.J. Lu <hongjiu.lu@intel.com>
7780 * config/i386/i386.c (type_natural_mode): Add a bool parameter
7781 to indicate if type is used for function return value. Warn ABI
7782 change if the vector mode isn't available for function return value.
7783 (ix86_function_arg_advance): Pass false to type_natural_mode.
7784 (ix86_function_arg): Likewise.
7785 (ix86_gimplify_va_arg): Likewise.
7786 (function_arg_32): Don't warn ABI change.
7787 (ix86_function_value): Pass true to type_natural_mode.
7788 (ix86_return_in_memory): Likewise.
7789 (ix86_struct_value_rtx): Removed.
7790 (TARGET_STRUCT_VALUE_RTX): Likewise.
7792 2014-01-14 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7794 * jump.c (redirect_jump_2): Remove REG_CROSSING_JUMP notes when
7795 converting a conditional jump into a conditional return.
7797 2014-01-14 Richard Biener <rguenther@suse.de>
7799 PR tree-optimization/58921
7800 PR tree-optimization/59006
7801 * tree-vect-loop-manip.c (vect_loop_versioning): Remove code
7802 hoisting invariant stmts.
7803 * tree-vect-stmts.c (vectorizable_load): Insert the splat of
7804 invariant loads on the preheader edge if possible.
7806 2014-01-14 Joey Ye <joey.ye@arm.com>
7808 * doc/plugin.texi (Building GCC plugins): Update to C++.
7810 2014-01-14 Kirill Yukhin <kirill.yukhin@intel.com>
7812 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): New.
7813 (_mm_rcp28_round_ss): Ditto.
7814 (_mm_rsqrt28_round_sd): Ditto.
7815 (_mm_rsqrt28_round_ss): Ditto.
7816 (_mm_rcp28_sd): Ditto.
7817 (_mm_rcp28_ss): Ditto.
7818 (_mm_rsqrt28_sd): Ditto.
7819 (_mm_rsqrt28_ss): Ditto.
7820 * config/i386/avx512fintrin.h (_mm512_stream_load_si512): Ditto.
7821 * config/i386/i386-builtin-types.def (V8DI_FTYPE_PV8DI): Ditto.
7822 * config/i386/i386.c (IX86_BUILTIN_MOVNTDQA512): Ditto.
7823 (IX86_BUILTIN_RCP28SD): Ditto.
7824 (IX86_BUILTIN_RCP28SS): Ditto.
7825 (IX86_BUILTIN_RSQRT28SD): Ditto.
7826 (IX86_BUILTIN_RSQRT28SS): Ditto.
7827 (bdesc_special_args): Define __builtin_ia32_movntdqa512,
7828 __builtin_ia32_rcp28sd_round, __builtin_ia32_rcp28ss_round,
7829 __builtin_ia32_rsqrt28sd_round, __builtin_ia32_rsqrt28ss_round.
7830 (ix86_expand_special_args_builtin): Expand new FTYPE.
7831 * config/i386/sse.md (define_mode_attr "sse4_1_avx2"): Expand to V8DI.
7832 (srcp14<mode>): Make insn unary.
7833 (avx512f_vmscalef<mode><round_name>): Use substed predicate.
7834 (avx512f_sgetexp<mode><round_saeonly_name>): Ditto.
7835 (avx512f_rndscale<mode><round_saeonly_name>): Ditto.
7836 (<sse4_1_avx2>_movntdqa): Extend to 512 bits.
7837 (avx512er_exp2<mode><mask_name><round_saeonly_name>):
7838 Fix rounding: make it SAE only.
7839 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
7841 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
7843 (avx512er_vmrcp28<mode><round_saeonly_name>): Ditto.
7844 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Ditto.
7845 (avx512f_getmant<mode><mask_name><round_saeonly_name>): Ditto.
7846 * config/i386/subst.md (round_saeonly_mask_scalar_operand3): Remove.
7847 (round_saeonly_mask_scalar_operand4): Ditto.
7848 (round_saeonly_mask_scalar_op3): Ditto.
7849 (round_saeonly_mask_scalar_op4): Ditto.
7851 2014-01-13 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7853 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
7854 Implement -maltivec=be for vec_insert and vec_extract.
7856 2014-01-10 DJ Delorie <dj@redhat.com>
7858 * config/msp430/msp430.md (call_internal): Don't allow memory
7859 references with SP as the base register.
7860 (call_value_internal): Likewise.
7861 * config/msp430/constraints.md (Yc): New. For memory references
7862 that don't use SP as a base register.
7864 * config/msp430/msp430.c (msp430_print_operand): Add 'J' to mean
7865 "an integer without a # prefix"
7866 * config/msp430/msp430.md (epilogue_helper): Use it.
7868 2014-01-13 Jakub Jelinek <jakub@redhat.com>
7871 * config/i386/i386.c (ix86_vectorize_builtin_gather): Uncomment
7872 AVX512F gather builtins.
7873 * tree-vect-stmts.c (vectorizable_mask_load_store): For now punt
7874 on gather decls with INTEGER_TYPE masktype.
7875 (vectorizable_load): For INTEGER_TYPE masktype, put the INTEGER_CST
7876 directly into the builtin rather than hoisting it before loop.
7878 PR tree-optimization/59387
7879 * tree-scalar-evolution.c: Include gimple-fold.h and gimplify-me.h.
7880 (scev_const_prop): If folded_casts and type has undefined overflow,
7881 use force_gimple_operand instead of force_gimple_operand_gsi and
7882 for each added stmt if it is assign with
7883 arith_code_with_undefined_signed_overflow, call
7884 rewrite_to_defined_overflow.
7885 * tree-ssa-loop-im.c: Don't include gimplify-me.h, include
7886 gimple-fold.h instead.
7887 (arith_code_with_undefined_signed_overflow,
7888 rewrite_to_defined_overflow): Moved to ...
7889 * gimple-fold.c (arith_code_with_undefined_signed_overflow,
7890 rewrite_to_defined_overflow): ... here. No longer static.
7891 Include gimplify-me.h.
7892 * gimple-fold.h (arith_code_with_undefined_signed_overflow,
7893 rewrite_to_defined_overflow): New prototypes.
7895 2014-01-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7897 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Fix typo in description.
7899 2014-01-13 Eric Botcazou <ebotcazou@adacore.com>
7901 * builtins.c (get_object_alignment_2): Minor tweak.
7902 * tree-ssa-loop-ivopts.c (may_be_unaligned_p): Rewrite.
7904 2014-01-13 Christian Bruel <christian.bruel@st.com>
7906 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Unroll small sizes and
7907 optimized non constant lengths.
7909 2014-01-13 Jakub Jelinek <jakub@redhat.com>
7912 * omp-low.c (expand_omp_atomic_pipeline): Expand the initial
7913 load as __atomic_load_N if possible.
7915 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
7917 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): Remove
7919 (rs6000_expand_builtin): Adjust call.
7921 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
7924 * config/rs6000/rs6000.h (SWITCHABLE_TARGET): Define.
7925 * config/rs6000/rs6000.c: Include target-globals.h.
7926 (rs6000_set_current_function): Instead of doing target_reinit
7927 unconditionally, use save_target_globals_default_opts and
7928 restore_target_globals.
7930 * config/rs6000/rs6000-builtin.def (mffs, mtfsf): Add builtins for
7932 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): New.
7933 (rs6000_expand_builtin): Handle mffs and mtfsf.
7934 (rs6000_init_builtins): Define mffs and mtfsf.
7935 * config/rs6000/rs6000.md (UNSPECV_MFFS, UNSPECV_MTFSF): New constants.
7936 (rs6000_mffs): New pattern.
7937 (rs6000_mtfsf): New pattern.
7939 2014-01-11 Bin Cheng <bin.cheng@arm.com>
7941 * tree-ssa-loop-ivopts.c (iv_ca_narrow): New parameter.
7942 Start narrowing with START. Apply candidate-use pair
7943 and check overall cost in narrowing.
7944 (iv_ca_prune): Pass new argument.
7946 2014-01-10 Jeff Law <law@redhat.com>
7949 * ree.c (combine_reaching_defs): Ensure the defining statement
7950 occurs before the extension when optimizing extensions with
7951 different source and destination hard registers.
7953 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
7956 * ipa-devirt.c (build_type_inheritance_graph): Also add types of
7957 vtables into the type inheritance graph.
7959 2014-01-10 Jakub Jelinek <jakub@redhat.com>
7961 PR rtl-optimization/59754
7962 * ree.c (combine_reaching_defs): Disallow !SCALAR_INT_MODE_P
7963 modes in the REGNO != REGNO case.
7965 2014-01-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7967 * config/rs6000/rs6000-builtin.def: Fix pasto for VPKSDUS.
7969 2014-01-10 Jakub Jelinek <jakub@redhat.com>
7971 PR tree-optimization/59745
7972 * tree-predcom.c (tree_predictive_commoning_loop): Call
7973 free_affine_expand_cache if giving up because components is NULL.
7975 * target-globals.c (save_target_globals): Allocate < 4KB structs using
7976 GC in payload of target_globals struct instead of allocating them on
7977 the heap and the larger structs separately using GC.
7978 * target-globals.h (struct target_globals): Make regs, hard_regs,
7979 reload, expmed, ira, ira_int and lra_fields GTY((atomic)) instead
7980 of GTY((skip)) and change type to void *.
7981 (reset_target_globals): Cast loads from those fields to corresponding
7984 2014-01-10 Steve Ellcey <sellcey@mips.com>
7987 * Makefile.in (PLUGIN_HEADERS): Add gimplify.h, gimple-iterator.h,
7988 gimple-ssa.h, fold-const.h, tree-cfg.h, tree-into-ssa.h,
7989 tree-ssanames.h, print-tree.h, varasm.h, and context.h.
7991 2014-01-10 Richard Earnshaw <rearnsha@arm.com>
7994 * aarch64-modes.def (CC_Zmode): New flags mode.
7995 * aarch64.c (aarch64_select_cc_mode): Only allow NEG when the condition
7996 represents an equality.
7997 (aarch64_get_condition_code): Handle CC_Zmode.
7998 * aarch64.md (compare_neg<mode>): Restrict to equality operations.
8000 2014-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8002 * config/s390/s390.c (s390_expand_tbegin): Remove jump over CC
8003 extraction in good case.
8005 2014-01-10 Richard Biener <rguenther@suse.de>
8007 PR tree-optimization/59374
8008 * tree-vect-slp.c (vect_slp_analyze_bb_1): Move dependence
8009 checking after SLP discovery. Mark stmts not participating
8010 in any SLP instance properly.
8012 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8014 * config/arm/arm.c (arm_new_rtx_costs): Use destination mode
8015 when handling a SET rtx.
8017 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8019 * config/arm/arm-cores.def (cortex-a53): Specify FL_CRC32.
8020 (cortex-a57): Likewise.
8021 (cortex-a57.cortex-a53): Likewise. Remove redundant flags.
8023 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8025 * config/arm/arm.c (arm_init_iwmmxt_builtins): Skip
8026 non-iwmmxt builtins.
8028 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
8032 * ipa-devirt.c record_target_from_binfo): Take as argument
8033 stack of binfos and lookup matching one for virtual inheritance.
8034 (possible_polymorphic_call_targets_1): Update.
8036 2014-01-10 Huacai Chen <chenhc@lemote.com>
8038 * config/mips/driver-native.c (host_detect_local_cpu): Handle new
8039 kernel strings for Loongson-2E/2F/3A.
8041 2014-01-10 Jakub Jelinek <jakub@redhat.com>
8044 * tree-vect-data-refs.c (vect_analyze_data_refs): Check
8045 is_gimple_call before calling gimple_call_internal_p.
8047 2014-01-09 Steve Ellcey <sellcey@mips.com>
8049 * Makefile.in (TREE_FLOW_H): Remove.
8050 (TREE_SSA_H): Add file names from tree-flow.h.
8051 * doc/tree-ssa.texi (Annotations): Remove reference to tree-flow.h
8052 * tree.h: Remove tree-flow.h reference.
8053 * hash-table.h: Remove tree-flow.h reference.
8054 * tree-ssa-loop-niter.c (dump_affine_iv): Replace tree-flow.h
8055 reference with tree-ssa-loop.h.
8057 2014-01-09 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8059 * doc/invoke.texi: Add -maltivec={be,le} options, and document
8060 default element-order behavior for -maltivec.
8061 * config/rs6000/rs6000.opt: Add -maltivec={be,le} options.
8062 * config/rs6000/rs6000.c (rs6000_option_override_internal): Ensure
8063 that -maltivec={le,be} implies -maltivec; disallow -maltivec=le
8064 when targeting big endian, at least for now.
8065 * config/rs6000/rs6000.h: Add #define of VECTOR_ELT_ORDER_BIG.
8067 2014-01-09 Jakub Jelinek <jakub@redhat.com>
8070 * cfgexpand.c (expand_one_var): For SSA_NAMEs, if the underlying
8071 var satisfies use_register_for_decl, just take into account type
8072 alignment, rather than decl alignment.
8074 PR tree-optimization/59622
8075 * gimple-fold.c (gimple_fold_call): Fix a typo in message. For
8076 __builtin_unreachable replace the OBJ_TYPE_REF call with a call to
8077 __builtin_unreachable and add if needed a setter of the lhs SSA_NAME.
8078 Don't devirtualize for inplace at all. For targets.length () == 1,
8079 if the call is noreturn and cfun isn't in SSA form yet, clear lhs.
8081 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
8083 * config/i386/i386.md (cpu): Remove the unused btver1.
8085 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
8087 * gdbasan.in: Put a breakpoint on __sanitizer::Report.
8089 2014-01-09 Jakub Jelinek <jakub@redhat.com>
8092 * tree-core.h (struct target_globals): New forward declaration.
8093 (struct tree_target_option): Add globals field.
8094 * tree.h (TREE_TARGET_GLOBALS): Define.
8095 (prepare_target_option_nodes_for_pch): New prototype.
8096 * target-globals.h (struct target_globals): Define even if
8098 * tree.c (prepare_target_option_node_for_pch,
8099 prepare_target_option_nodes_for_pch): New functions.
8100 * config/i386/i386.h (SWITCHABLE_TARGET): Define.
8101 * config/i386/i386.c: Include target-globals.h.
8102 (ix86_set_current_function): Instead of doing target_reinit
8103 unconditionally, use save_target_globals_default_opts and
8104 restore_target_globals.
8106 2014-01-09 Richard Biener <rguenther@suse.de>
8108 PR tree-optimization/59715
8109 * tree-cfg.h (split_critical_edges): Declare.
8110 * tree-cfg.c (split_critical_edges): Export.
8111 * tree-ssa-sink.c (execute_sink_code): Split critical edges.
8113 2014-01-09 Max Ostapenko <m.ostapenko@partner.samsung.com>
8115 * cfgexpand.c (expand_stack_vars): Optionally disable
8116 asan stack protection.
8117 (expand_used_vars): Likewise.
8118 (partition_stack_vars): Likewise.
8119 * asan.c (asan_emit_stack_protection): Optionally disable
8120 after return stack usage.
8121 (instrument_derefs): Optionally disable memory access instrumentation.
8122 (instrument_builtin_call): Likewise.
8123 (instrument_strlen_call): Likewise.
8124 (asan_protect_global): Optionally disable global variables protection.
8125 * doc/invoke.texi: Added doc for new options.
8126 * params.def: Added new options.
8127 * params.h: Likewise.
8129 2014-01-09 Jakub Jelinek <jakub@redhat.com>
8131 PR rtl-optimization/59724
8132 * ifcvt.c (cond_exec_process_if_block): Don't call
8133 flow_find_head_matching_sequence with 0 longest_match.
8134 * cfgcleanup.c (flow_find_head_matching_sequence): Count even
8135 non-active insns if !stop_after.
8136 (try_head_merge_bb): Revert 2014-01-07 changes.
8138 2014-01-08 Jeff Law <law@redhat.com>
8140 * ree.c (get_sub_rtx): New function, extracted from...
8141 (merge_def_and_ext): Here.
8142 (combine_reaching_defs): Use get_sub_rtx.
8144 2014-01-08 Eric Botcazou <ebotcazou@adacore.com>
8146 * cgraph.h (varpool_variable_node): Do not choke on null node.
8148 2014-01-08 Catherine Moore <clm@codesourcery.com>
8150 * config/mips/mips.md (simple_return): Attempt to use JRC
8152 * config/mips/mips.h (MIPS_CALL): Attempt to use JALS for microMIPS.
8154 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
8156 PR rtl-optimization/59137
8157 * reorg.c (steal_delay_list_from_target): Call update_block for
8159 (steal_delay_list_from_fallthrough, relax_delay_slots): Likewise.
8161 2014-01-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8163 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Remove
8164 two duplicate entries.
8166 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
8169 2012-10-07 Richard Sandiford <rdsandiford@googlemail.com>
8171 * config/mips/mips.c (mips_truncated_op_cost): New function.
8172 (mips_rtx_costs): Adjust test for BADDU.
8173 * config/mips/mips.md (*baddu_di<mode>): Push truncates to operands.
8175 2012-10-02 Richard Sandiford <rdsandiford@googlemail.com>
8177 * config/mips/mips.md (*baddu_si_eb, *baddu_si_el): Merge into...
8178 (*baddu_si): ...this new pattern.
8180 2014-01-08 Jakub Jelinek <jakub@redhat.com>
8183 * ipa-prop.c (ipa_analyze_params_uses): Ignore uses in debug stmts.
8185 2014-01-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
8188 * expr.h (expand_expr_real, expand_expr_real_1): Add new parameter
8190 (expand_expr, expand_normal): Adjust.
8191 * expr.c (expand_expr_real, expand_expr_real_1): Add new parameter
8192 inner_reference_p. Use inner_reference_p to expand inner references.
8193 (store_expr): Adjust.
8194 * cfgexpand.c (expand_call_stmt): Adjust.
8196 2014-01-08 Rong Xu <xur@google.com>
8198 * gcov-io.c (gcov_var): Move from gcov-io.h.
8199 (gcov_position): Ditto.
8200 (gcov_is_error): Ditto.
8201 (gcov_rewrite): Ditto.
8202 * gcov-io.h: Refactor. Move gcov_var to gcov-io.h, and libgcov
8203 only part to libgcc/libgcov.h.
8205 2014-01-08 Marek Polacek <polacek@redhat.com>
8208 * omp-low.c (simd_clone_adjust): Don't crash if def is NULL.
8210 2014-01-08 Marek Polacek <polacek@redhat.com>
8213 * ubsan.c (ubsan_type_descriptor): Call strip_array_types on type2.
8215 2014-01-08 Jakub Jelinek <jakub@redhat.com>
8217 PR rtl-optimization/59649
8218 * stor-layout.c (get_mode_bounds): For BImode return
8219 0 and STORE_FLAG_VALUE.
8221 2014-01-08 Richard Biener <rguenther@suse.de>
8224 * gimple.h (is_gimple_builtin_call): Remove.
8225 (gimple_builtin_call_types_compatible_p): New.
8226 (gimple_call_builtin_p): New overload.
8227 * gimple.c (is_gimple_builtin_call): Remove.
8228 (validate_call): Rename to ...
8229 (gimple_builtin_call_types_compatible_p): ... this and export. Also
8231 (validate_type): New static function.
8232 (gimple_call_builtin_p): New overload and adjust.
8233 * gimple-fold.c (gimple_fold_builtin): Fold the return value.
8234 (gimple_fold_call): Likewise. Use gimple_call_builtin_p.
8235 (gimple_fold_stmt_to_constant_1): Likewise.
8236 * tsan.c (instrument_gimple): Use gimple_call_builtin_p.
8238 2014-01-08 Richard Biener <rguenther@suse.de>
8241 * gimplify.c (gimplify_expr): Gimplify register-register type
8242 VIEW_CONVERT_EXPRs to separate stmts.
8244 2014-01-07 Jeff Law <law@redhat.com>
8247 * ree.c (combine_set_extension): Handle case where source
8248 and destination registers in an extension insn are different.
8249 (combine_reaching_defs): Allow source and destination registers
8250 in extension to be different under limited circumstances.
8251 (add_removable_extension): Remove restriction that the
8252 source and destination registers in the extension are the same.
8253 (find_and_remove_re): Emit a copy from the extension's
8254 destination to its source after the defining insn if
8255 the source and destination registers are different.
8258 * ifcvt.c (merge_if_block): If we are merging a block with more than
8259 one successor with a block with no successors, remove any BARRIER
8260 after the second block.
8262 2014-01-07 Dan Xio Qiang <ziyan01@163.com>
8264 * hw-doloop.c (reorg_loops): Release the bitmap obstack.
8266 2014-01-07 John David Anglin <danglin@gcc.gnu.org>
8269 * config/pa/pa.c (pa_legitimate_address_p): Return false before reload
8270 for 14-bit register offsets when INT14_OK_STRICT is false.
8272 2014-01-07 Roland Stigge <stigge@antcom.de>
8273 Michael Meissner <meissner@linux.vnet.ibm.com>
8276 * config/rs6000/rs6000.c (rs6000_legitimate_offset_address_p):
8277 Only check TFmode for SPE constants. Don't check TImode or TDmode.
8279 2014-01-07 James Greenhalgh <james.greenhalgh@arm.com>
8281 * config/aarch64/aarch64-elf.h (ASM_SPEC): Remove identity spec for
8284 2014-01-07 Yufeng Zhang <yufeng.zhang@arm.com>
8286 * config/arm/arm.c (arm_expand_neon_args): Call expand_expr
8287 with EXPAND_MEMORY for NEON_ARG_MEMORY; check if the returned
8288 rtx is const0_rtx or not.
8290 2014-01-07 Richard Sandiford <rdsandiford@googlemail.com>
8293 * target-globals.c (save_target_globals): Remove this_fn_optab
8295 * toplev.c: Include optabs.h.
8296 (target_reinit): Temporarily restore the global options if another
8297 set of options are in force.
8299 2014-01-07 Jakub Jelinek <jakub@redhat.com>
8301 PR rtl-optimization/58668
8302 * cfgcleanup.c (flow_find_cross_jump): Don't count
8303 any jumps if dir_p is NULL. Remove p1 variable, use active_insn_p
8304 to determine what is counted.
8305 (flow_find_head_matching_sequence): Use active_insn_p to determine
8307 (try_head_merge_bb): Adjust for the flow_find_head_matching_sequence
8309 * ifcvt.c (count_bb_insns): Use active_insn_p && !JUMP_P to
8310 determine what is counted.
8312 PR tree-optimization/59643
8313 * tree-predcom.c (split_data_refs_to_components): If one dr is
8314 read and one write, determine_offset fails and the write isn't
8315 in the bad component, just put the read into the bad component.
8317 2014-01-07 Mike Stump <mikestump@comcast.net>
8318 Jakub Jelinek <jakub@redhat.com>
8321 * tree-core.h (struct tree_optimization_option): Change optabs
8322 type from unsigned char * to void *.
8323 * optabs.c (init_tree_optimization_optabs): Adjust
8324 TREE_OPTIMIZATION_OPTABS initialization.
8326 2014-01-06 Jakub Jelinek <jakub@redhat.com>
8329 * config/i386/i386.h (struct machine_function): Add
8330 no_drap_save_restore field.
8331 * config/i386/i386.c (ix86_save_reg): Use
8332 !cfun->machine->no_drap_save_restore instead of
8333 crtl->stack_realign_needed.
8334 (ix86_finalize_stack_realign_flags): Don't clear drap_reg unless
8335 this function clears frame_pointer_needed. Set
8336 cfun->machine->no_drap_save_restore if clearing frame_pointer_needed
8337 and DRAP reg is needed.
8339 2014-01-06 Marek Polacek <polacek@redhat.com>
8342 * doc/implement-c.texi: Mention that other integer types are
8343 permitted as bit-field types in strictly conforming mode.
8345 2014-01-06 Felix Yang <fei.yang0953@gmail.com>
8347 * modulo-sched.c (schedule_reg_moves): Clear distance1_uses if it
8350 2014-01-06 Richard Earnshaw <rearnsha@arm.com>
8352 * aarch64.c (aarch64_rtx_costs): Fix cost calculation for MADD.
8354 2014-01-06 Martin Jambor <mjambor@suse.cz>
8357 * ipa-cp.c (ipcp_discover_new_direct_edges): Changed param_index type
8359 * ipa-prop.c (ipa_print_node_params): Fix indentation.
8361 2014-01-06 Eric Botcazou <ebotcazou@adacore.com>
8365 * var-tracking.c (add_stores): Preserve the value of the source even if
8366 we don't record the store.
8368 2014-01-06 Terry Guo <terry.guo@arm.com>
8370 * config.gcc (arm*-*-*): Check --with-arch against arm-arches.def.
8372 2014-01-05 Iain Sandoe <iain@codesourcery.com>
8375 * config/darwin.c (darwin_function_section): Adjust return values to
8376 correspond to optimisation changes made in r206070.
8378 2014-01-05 Uros Bizjak <ubizjak@gmail.com>
8380 * config/i386/i386.c (ix86_data_alignment): Calculate max_align
8381 from prefetch_block tune setting.
8382 (nocona_cost): Correct size of prefetch block to 64.
8384 2014-01-04 Eric Botcazou <ebotcazou@adacore.com>
8386 * config/arm/arm.c (arm_get_frame_offsets): Revamp long lines.
8387 (arm_expand_epilogue_apcs_frame): Take into account the number of bytes
8388 used to save the static chain register in the computation of the offset
8389 from which the FP registers need to be restored.
8391 2014-01-04 Jakub Jelinek <jakub@redhat.com>
8393 PR tree-optimization/59519
8394 * tree-vect-loop-manip.c (slpeel_update_phi_nodes_for_guard1): Don't
8395 ICE if get_current_def (current_new_name) is already non-NULL, as long
8396 as it is a phi result of some other phi in *new_exit_bb that has
8399 * config/i386/sse.md (avx512f_load<mode>_mask): Emit vmovup{s,d}
8400 or vmovdqu* for misaligned_operand.
8401 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>,
8402 <sse2_avx_avx512f>_loaddqu<mode><mask_name>): Handle <mask_applied>.
8403 * config/i386/i386.c (ix86_expand_special_args_builtin): Set
8404 aligned_mem for AVX512F masked aligned load and store builtins and for
8407 2014-01-03 Bingfeng Mei <bmei@broadcom.com>
8409 PR tree-optimization/59651
8410 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
8411 Address range for negative step should be added by TYPE_SIZE_UNIT.
8413 2014-01-03 Andreas Schwab <schwab@linux-m68k.org>
8415 * config/m68k/m68k.c (handle_move_double): Handle pushes with
8416 overlapping registers also for registers other than the stack pointer.
8418 2014-01-03 Marek Polacek <polacek@redhat.com>
8421 * doc/extend.texi: Fix the return value of __builtin_FUNCTION and
8424 2014-01-03 Jakub Jelinek <jakub@redhat.com>
8427 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Don't consider
8430 * config/i386/i386.md (MODE_SIZE): New mode attribute.
8431 (push splitter): Use <P:MODE_SIZE> instead of
8432 GET_MODE_SIZE (<P:MODE>mode).
8433 (lea splitter): Use <MODE_SIZE> instead of GET_MODE_SIZE (<MODE>mode).
8434 (mov -1, reg peephole2): Likewise.
8435 * config/i386/sse.md (*mov<mode>_internal,
8436 <sse>_storeu<ssemodesuffix><avxsizesuffix>,
8437 <sse2_avx_avx512f>_storedqu<mode>, <sse>_andnot<mode>3,
8438 *<code><mode>3, *andnot<mode>3<mask_name>,
8439 <mask_codefor><code><mode>3<mask_name>): Likewise.
8440 * config/i386/subst.md (mask_mode512bit_condition,
8441 sd_mask_mode512bit_condition): Likewise.
8443 2014-01-02 Xinliang David Li <davidxl@google.com>
8445 PR tree-optimization/59303
8446 * tree-ssa-uninit.c (is_use_properly_guarded): Main cleanup.
8447 (dump_predicates): Better output format.
8448 (pred_equal_p): New function.
8449 (is_neq_relop_p): Ditto.
8450 (is_neq_zero_form_p): Ditto.
8451 (pred_expr_equal_p): Ditto.
8452 (pred_neg_p): Ditto.
8453 (simplify_pred): Ditto.
8454 (simplify_preds_2): Ditto.
8455 (simplify_preds_3): Ditto.
8456 (simplify_preds_4): Ditto.
8457 (simplify_preds): Ditto.
8459 (push_to_worklist): Ditto.
8460 (get_pred_info_from_cmp): Ditto.
8461 (is_degenerated_phi): Ditto.
8462 (normalize_one_pred_1): Ditto.
8463 (normalize_one_pred): Ditto.
8464 (normalize_one_pred_chain): Ditto.
8465 (normalize_preds): Ditto.
8466 (normalize_cond_1): Remove function.
8467 (normalize_cond): Ditto.
8468 (is_gcond_subset_of): Ditto.
8469 (is_subset_of_any): Ditto.
8470 (is_or_set_subset_of): Ditto.
8471 (is_and_set_subset_of): Ditto.
8472 (is_norm_cond_subset_of): Ditto.
8473 (pred_chain_length_cmp): Ditto.
8474 (convert_control_dep_chain_into_preds): Type change.
8475 (find_predicates): Ditto.
8476 (find_def_preds): Ditto.
8477 (destroy_predicates_vecs): Ditto.
8478 (find_matching_predicates_in_rest_chains): Ditto.
8479 (use_pred_not_overlap_with_undef_path_pred): Ditto.
8480 (is_pred_expr_subset): Ditto.
8481 (is_pred_chain_subset_of): Ditto.
8482 (is_included_in): Ditto.
8483 (is_superset_of): Ditto.
8485 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
8487 Update copyright years.
8489 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
8491 * common/config/arc/arc-common.c, config/arc/arc-modes.def,
8492 config/arc/arc-protos.h, config/arc/arc.c, config/arc/arc.h,
8493 config/arc/arc.md, config/arc/arc.opt,
8494 config/arm/arm_neon_builtins.def, config/arm/crypto.def,
8495 config/i386/avx512cdintrin.h, config/i386/avx512erintrin.h,
8496 config/i386/avx512fintrin.h, config/i386/avx512pfintrin.h,
8497 config/i386/btver2.md, config/i386/shaintrin.h, config/i386/slm.md,
8498 config/linux-protos.h, config/linux.c, config/winnt-c.c,
8499 diagnostic-color.c, diagnostic-color.h, gimple-ssa-isolate-paths.c,
8500 vtable-verify.c, vtable-verify.h: Use the standard form for the
8503 2014-01-02 Tobias Burnus <burnus@net-b.de>
8505 * gcc.c (process_command): Update copyright notice dates.
8506 * gcov-dump.c: Ditto.
8508 * doc/cpp.texi: Bump @copying's copyright year.
8509 * doc/cppinternals.texi: Ditto.
8510 * doc/gcc.texi: Ditto.
8511 * doc/gccint.texi: Ditto.
8512 * doc/gcov.texi: Ditto.
8513 * doc/install.texi: Ditto.
8514 * doc/invoke.texi: Ditto.
8516 2014-01-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8518 * config/nios2/nios2.h (BITS_PER_UNIT): Don't define it.
8520 2014-01-01 Jakub Jelinek <jakub@redhat.com>
8522 * config/i386/sse.md (*mov<mode>_internal): Guard
8523 EXT_REX_SSE_REGNO_P (REGNO ()) uses with REG_P.
8525 PR rtl-optimization/59647
8526 * cse.c (cse_process_notes_1): Don't substitute negative VOIDmode
8527 new_rtx into UNSIGNED_FLOAT rtxes.
8529 Copyright (C) 2014 Free Software Foundation, Inc.
8531 Copying and distribution of this file, with or without modification,
8532 are permitted in any medium without royalty provided the copyright
8533 notice and this notice are preserved.