1 2014-04-28 Richard Biener <rguenther@suse.de>
3 * tree-pass.h (TODO_verify_il): Define.
4 (TODO_verify_all): Complete properly.
5 * passes.c (execute_function_todo): Move existing loop-closed
6 SSA verification under TODO_verify_il.
7 (execute_one_pass): Trigger TODO_verify_il at todo-after time.
8 * graphite-sese-to-poly.c (rewrite_cross_bb_scalar_deps):
9 Fix tree sharing issue.
11 2014-04-28 Richard Biener <rguenther@suse.de>
14 * builtins.def (DEF_C11_BUILTIN): Add.
15 (BUILT_IN_ALIGNED_ALLOC): Likewise.
16 * coretypes.h (enum function_class): Add function_c11_misc.
17 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
18 BUILT_IN_ALIGNED_ALLOC like BUILT_IN_MALLOC.
19 (call_may_clobber_ref_p_1): Likewise.
20 * tree-ssa-dce.c (mark_stmt_if_obviously_necessary): Likewise.
21 (mark_all_reaching_defs_necessary_1): Likewise.
22 (propagate_necessity): Likewise.
23 (eliminate_unnecessary_stmts): Likewise.
24 * tree-ssa-ccp.c (evaluate_stmt): Handle BUILT_IN_ALIGNED_ALLOC.
26 2014-04-28 Richard Biener <rguenther@suse.de>
28 * tree-vrp.c (vrp_var_may_overflow): Remove.
29 (vrp_visit_phi_node): Rather than bumping to +-INF possibly
30 with overflow immediately bump to one before that value and
31 let iteration figure out overflow status.
33 2014-04-28 Richard Biener <rguenther@suse.de>
35 * configure.ac: Do valgrind header checks unconditionally.
36 Add --enable-valgrind-annotations.
37 * system.h: Guard valgrind header inclusion with
38 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING.
39 * alloc-pool.c (pool_alloc, pool_free): Use
40 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING
41 to guard possibly dead code.
42 * config.in: Regenerated.
43 * configure: Likewise.
45 2014-04-28 Jeff Law <law@redhat.com>
47 PR tree-optimization/60902
48 * tree-ssa-threadedge.c
49 (record_temporary_equivalences_from_stmts_at_dest): Only iterate
50 over real defs when invalidating outputs from statements that do not
51 produce useful outputs for threading.
53 2014-04-28 Richard Biener <rguenther@suse.de>
55 PR tree-optimization/60979
56 * graphite-scop-detection.c (scopdet_basic_block_info): Reject
57 SCOPs that end in a block with a successor with abnormal
60 2014-04-28 Richard Biener <rguenther@suse.de>
62 * tree-pass.h (execute_pass_list): Adjust prototype.
63 * passes.c (pass_manager::execute_early_local_passes):
65 (do_per_function): Change callback signature, push all actual
67 (do_per_function_toporder): Likewise.
68 (execute_function_dump): Adjust.
69 (execute_function_todo): Likewise.
70 (clear_last_verified): Likewise.
71 (verify_curr_properties): Likewise.
72 (update_properties_after_pass): Likewise.
73 (execute_pass_list_1): Split out from ...
74 (execute_pass_list): ... here. Adjust.
75 (execute_ipa_pass_list): Likewise.
76 * cgraphunit.c (cgraph_add_new_function): Adjust.
77 (analyze_function): Likewise.
78 (expand_function): Likewise.
79 * cgraph.c (release_function_body): Free dominance info
80 here instead of asserting it was magically freed elsewhere.
82 2014-04-28 Eric Botcazou <ebotcazou@adacore.com>
84 * configure.ac: Tweak GAS check for LEON instructions on SPARC.
85 * configure: Regenerate.
86 * config/sparc/sparc.opt (muser-mode): New option.
87 * config/sparc/sync.md (atomic_compare_and_swap<mode>_1): Do not enable
89 (atomic_compare_and_swap_leon3_1): New instruction for LEON3.
90 * doc/invoke.texi (SPARC options): Document -muser-mode.
92 2014-04-27 Richard Sandiford <rdsandiford@googlemail.com>
94 * cselib.c (find_slot_memmode): Delete.
95 (cselib_hasher): Change compare_type to a struct.
96 (cselib_hasher::equal): Update accordingly. Don't expect wrapped
98 (preserve_constants_and_equivs): Adjust for new compare_type.
99 (cselib_find_slot): Likewise. Take the mode of the rtx as argument.
100 (wrap_constant): Delete.
101 (cselib_lookup_mem, cselib_lookup_1): Update calls to cselib_find_slot.
103 2014-04-26 Markus Trippelsdorf <markus@trippelsdorf.de>
105 * doc/install.texi (Building with profile feedback): Remove
108 2014-04-26 Tom de Vries <tom@codesourcery.com>
110 * config/i386/i386.md (define_expand "ldexpxf3"): Fix out-of-bounds
113 2014-04-25 Cary Coutant <ccoutant@google.com>
116 * dwarf2out.c (should_move_die_to_comdat): A type definition
117 can contain a subprogram definition, but don't move it to a
119 (clone_as_declaration): Copy DW_AT_abstract_origin attribute.
120 (generate_skeleton_bottom_up): Remove DW_AT_object_pointer attribute
122 (clone_tree_hash): Rename to...
123 (clone_tree_partial): ...this; change callers. Copy
124 DW_TAG_subprogram DIEs as declarations.
125 (copy_decls_walk): Don't copy children of a declaration into a
128 2014-04-25 H.J. Lu <hongjiu.lu@intel.com>
131 * config/i386/i386.md (*movsf_internal): Set MODE to SI for
134 2014-04-25 Jiong Wang <jiong.wang@arm.com>
136 * config/arm/predicates.md (call_insn_operand): Add long_call check.
137 * config/arm/arm.md (sibcall, sibcall_value): Force the address to
139 * config/arm/arm.c (arm_function_ok_for_sibcall): Remove long_call
142 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
144 * config/arm/arm.c (arm_cortex_a8_tune): Initialise
147 2014-04-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
149 PR tree-optimization/60930
150 * gimple-ssa-strength-reduction.c (create_mul_imm_cand): Reject
151 creating a multiply candidate by folding two constant
152 multiplicands when the result overflows.
154 2014-04-25 Jakub Jelinek <jakub@redhat.com>
156 PR tree-optimization/60960
157 * tree-vect-generic.c (expand_vector_operation): Only call
158 expand_vector_divmod if type's mode satisfies VECTOR_MODE_P.
160 2014-04-25 Tom de Vries <tom@codesourcery.com>
162 * expr.c (clobber_reg_mode): New function.
163 * expr.h (clobber_reg): New function.
165 2014-04-25 Tom de Vries <tom@codesourcery.com>
167 * rtlanal.c (find_all_hard_reg_sets): Note INSN_CALL_FUNCTION_USAGE
170 2014-04-25 Radovan Obradovic <robradovic@mips.com>
171 Tom de Vries <tom@codesourcery.com>
173 * rtlanal.c (find_all_hard_reg_sets): Add bool implicit parameter and
175 * rtl.h (find_all_hard_reg_sets): Add bool parameter.
176 * haifa-sched.c (recompute_todo_spec, check_clobbered_conditions): Add
177 new argument to find_all_hard_reg_sets call.
179 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
181 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p):
182 Use HOST_WIDE_INT_C for mask literal.
183 (aarch_rev16_shleft_mask_imm_p): Likewise.
185 2014-04-25 Eric Botcazou <ebotcazou@adacore.com>
188 * config/sparc/sparc.md (ashlsi3_extend): Delete.
190 2014-04-25 Marc Glisse <marc.glisse@inria.fr>
192 PR preprocessor/56540
193 * config/i386/i386-c.c (ix86_target_macros): Define
194 __SIZEOF_FLOAT80__ and __SIZEOF_FLOAT128__.
196 2014-04-25 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
198 * configure.ac (tga_func): Remove.
199 (LIB_TLS_SPEC): Remove.
200 * configure: Regenerate.
201 * config.in: Regenerate.
202 * config/sol2.h (LIB_SPEC): Don't use LIB_TLS_SPEC.
204 2014-04-25 Richard Biener <rguenther@suse.de>
207 * tree-ssa-structalias.c (ipa_pta_execute): Compute direct
208 call stmt use/clobber sets during stmt walk instead of
209 walking the possibly incomplete set of caller edges.
211 2014-04-25 Richard Biener <rguenther@suse.de>
214 * passes.c (apply_ipa_transforms): Inline into only caller ...
215 (execute_one_pass): ... here. Properly bring in function
216 bodies for nodes we want to apply IPA transforms to.
218 2014-04-24 Cong Hou <congh@google.com>
220 PR tree-optimization/60896
221 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Pick up
222 all statements in PATTERN_DEF_SEQ in recognized widen-mult pattern.
223 (vect_mark_pattern_stmts): Set the def type of all statements in
224 PATTERN_DEF_SEQ as vect_internal_def.
226 2014-04-24 Michael Meissner <meissner@linux.vnet.ibm.com>
228 * doc/extend.texi (PowerPC Built-in Functions): Document new
229 powerpc extended divide, bcd, pack/unpack 128-bit, builtin
231 (PowerPC AltiVec/VSX Built-in Functions): Likewise.
233 * config/rs6000/predicates.md (const_0_to_3_operand): New
234 predicate to match 0..3 integer constants.
236 * config/rs6000/rs6000-builtin.def (BU_DFP_MISC_1): Add new macros
237 to support adding miscellaneous builtin functions.
238 (BU_DFP_MISC_2): Likewise.
239 (BU_P7_MISC_1): Likewise.
240 (BU_P7_MISC_2): Likewise.
241 (BU_P8V_MISC_3): Likewise.
242 (BU_MISC_1): Likewise.
243 (BU_MISC_2): Likewise.
244 (DIVWE): Add extended divide builtin functions.
252 (DXEX): Add decimal floating-point builtin functions.
264 (CDTBCD): Add new BCD builtin functions.
268 (BCDADD_LT): Likewise.
269 (BCDADD_EQ): Likewise.
270 (BCDADD_GT): Likewise.
271 (BCDADD_OV): Likewise.
273 (BCDSUB_LT): Likewise.
274 (BCDSUB_EQ): Likewise.
275 (BCDSUB_GT): Likewise.
276 (BCDSUB_OV): Likewise.
277 (PACK_TD): Add new pack/unpack 128-bit type builtin functions.
278 (UNPACK_TD): Likewise.
280 (UNPACK_TF): Likewise.
281 (UNPACK_TF_0): Likewise.
282 (UNPACK_TF_1): Likewise.
283 (PACK_V1TI): Likewise.
284 (UNPACK_V1TI): Likewise.
286 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Add
287 support for decimal floating point builtin functions.
288 (rs6000_expand_ternop_builtin): Add checks for the new builtin
289 functions that take constant arguments.
290 (rs6000_invalid_builtin): Add decimal floating point builtin
292 (rs6000_init_builtins): Setup long double, _Decimal64, and
293 _Decimal128 types for new builtin functions.
294 (builtin_function_type): Set the unsigned flags appropriately for
295 the new builtin functions.
296 (rs6000_opt_masks): Add support for decimal floating point builtin
299 * config/rs6000/rs6000.h (RS6000_BTM_DFP): Add support for decimal
300 floating point builtin functions.
301 (RS6000_BTM_COMMON): Likewise.
302 (RS6000_BTI_long_double): Likewise.
303 (RS6000_BTI_dfloat64): Likewise.
304 (RS6000_BTI_dfloat128): Likewise.
305 (long_double_type_internal_node): Likewise.
306 (dfloat64_type_internal_node): Likewise.
307 (dfloat128_type_internal_node): Likewise.
309 * config/rs6000/altivec.h (UNSPEC_BCDADD): Add support for ISA
310 2.07 bcd arithmetic instructions.
311 (UNSPEC_BCDSUB): Likewise.
312 (UNSPEC_BCD_OVERFLOW): Likewise.
313 (UNSPEC_BCD_ADD_SUB): Likewise.
314 (bcd_add_sub): Likewise.
315 (BCD_TEST): Likewise.
316 (bcd<bcd_add_sub>): Likewise.
317 (bcd<bcd_add_sub>_test): Likewise.
318 (bcd<bcd_add_sub>_test2): Likewise.
319 (bcd<bcd_add_sub>_<code>): Likewise.
320 (peephole2 for combined bcd ops): Likewise.
322 * config/rs6000/dfp.md (UNSPEC_DDEDPD): Add support for new
323 decimal floating point builtin functions.
324 (UNSPEC_DENBCD): Likewise.
325 (UNSPEC_DXEX): Likewise.
326 (UNSPEC_DIEX): Likewise.
327 (UNSPEC_DSCLI): Likewise.
328 (UNSPEC_DSCRI): Likewise.
329 (D64_D128): Likewise.
330 (dfp_suffix): Likewise.
331 (dfp_ddedpd_<mode>): Likewise.
332 (dfp_denbcd_<mode>): Likewise.
333 (dfp_dxex_<mode>): Likewise.
334 (dfp_diex_<mode>): Likewise.
335 (dfp_dscli_<mode>): Likewise.
336 (dfp_dscri_<mode>): Likewise.
338 * config/rs6000/rs6000.md (UNSPEC_ADDG6S): Add support for new BCD
340 (UNSPEC_CDTBCD): Likewise.
341 (UNSPEC_CBCDTD): Likewise.
342 (UNSPEC_DIVE): Add support for new extended divide builtin
344 (UNSPEC_DIVEO): Likewise.
345 (UNSPEC_DIVEU): Likewise.
346 (UNSPEC_DIVEUO): Likewise.
347 (UNSPEC_UNPACK_128BIT): Add support for new builtin functions to
348 pack/unpack 128-bit types.
349 (UNSPEC_PACK_128BIT): Likewise.
350 (idiv_ldiv): New mode attribute to set the 32/64-bit divide type.
351 (udiv<mode>3): Use idiv_ldiv mode attribute.
352 (div<mode>3): Likewise.
353 (addg6s): Add new BCD builtin functions.
356 (UNSPEC_DIV_EXTEND): Add support for new extended divide
358 (div_extend): Likewise.
359 (div<div_extend>_<mode>"): Likewise.
360 (FP128_64): Add support for new builtin functions to pack/unpack
362 (unpack<mode>): Likewise.
363 (unpacktf_0): Likewise.
364 (unpacktf_1): Likewise.
365 (unpack<mode>_dm): Likewise.
366 (unpack<mode>_nodm): Likewise.
367 (pack<mode>): Likewise.
368 (unpackv1ti): Likewise.
369 (packv1ti): Likewise.
371 2014-04-24 Vishnu K S <Vishnu.k_s@atmel.com>
373 * gcc/config/avr/avr.c: Add comment on why -fdelete-null-pointer-checks
376 2014-04-24 Jakub Jelinek <jakub@redhat.com>
378 * tree.h (OMP_CLAUSE_LINEAR_GIMPLE_SEQ): Define.
379 * gimplify.c (omp_is_private): Change last argument's type to int.
380 Only diagnose lastprivate if the simd argument is 1, only diagnose
381 linear if the simd argument is 2.
382 (gimplify_omp_for): Adjust omp_is_private callers. When adding
383 lastprivate or private, add the clause to OMP_FOR_CLAUSES. Pass
384 GOVD_EXPLICIT to omp_add_variable. For simd with collapse == 1
385 create OMP_CLAUSE_LINEAR rather than OMP_CLAUSE_PRIVATE for var.
386 If var != decl and decl is in OMP_CLAUSE_LINEAR, gimplify decl
387 increment to OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
388 * omp-low.c (scan_sharing_clauses, lower_lastprivate_clauses): Handle
389 OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
390 * tree-nested.c (convert_nonlocal_omp_clauses,
391 convert_local_omp_clauses): Handle OMP_CLAUSE_LINEAR.
393 2014-04-24 Segher Boessenkool <segher@kernel.crashing.org>
396 * config/m68k/m68k.md (extendplussidi): Don't allow memory for
399 2014-04-24 Dimitris Papavasiliou <dpapavas@gmail.com>
401 * flag-types.h (enum ivar_visibility): Add.
403 2014-04-24 Trevor Saunders <tsaunders@mozilla.com>
405 * config/sh/sh_treg_combine.c (sh_treg_combine::execute): Take
408 2014-04-24 Alan Lawrence <alan.lawrence@arm.com>
410 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Enable for bigendian.
412 2014-04-24 Radovan Obradovic <robradovic@mips.com>
413 Tom de Vries <tom@codesourcery.com>
415 * reg-notes.def (REG_NOTE (CALL_DECL)): New reg-note REG_CALL_DECL.
416 * calls.c (expand_call, emit_library_call_value_1): Add REG_CALL_DECL
418 * combine.c (distribute_notes): Handle REG_CALL_DECL reg-note.
419 * emit-rtl.c (try_split): Same.
421 2014-04-24 Radovan Obradovic <robradovic@mips.com>
422 Tom de Vries <tom@codesourcery.com>
424 * common.opt (fuse-caller-save): New option.
426 2014-04-24 Tejas Belagod <tejas.belagod@arm.com>
428 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Reverse order of elements
431 2014-04-24 Richard Biener <rguenther@suse.de>
433 * expr.c (expand_expr_real_1): Avoid gimple_assign_rhs_to_tree
434 during TER and instead use the sepops interface for expanding
435 non-GIMPLE_SINGLE_RHS.
437 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
439 * config/i386/sol2.h (ASM_PREFERRED_EH_DATA_FORMAT): Only redefine
440 if not HAVE_AS_IX86_DIFF_SECT_DELTA.
442 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
444 * configure.ac (gcc_cv_as_cfi_directive): Support Solaris/x86
445 assembler 64-bit option.
446 * configure: Regenerate.
448 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
450 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Check
451 TARGET_SIMD rather than TARGET_GENERAL_REGS_ONLY.
452 (TARGET_SIMD): Take AARCH64_ISA_SIMD into account.
453 (TARGET_FLOAT): Take AARCH64_ISA_FP into account.
454 (TARGET_CRYPTO): Take TARGET_SIMD into account.
456 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
458 * config/aarch64/aarch64-builtins.c
459 (aarch64_builtin_vectorized_function): Handle BUILT_IN_BSWAP16,
460 BUILT_IN_BSWAP32, BUILT_IN_BSWAP64.
461 * config/aarch64/aarch64-simd.md (bswap<mode>): New pattern.
462 * config/aarch64/aarch64-simd-builtins.def: Define vector bswap
464 * config/aarch64/iterator.md (VDQHSD): New mode iterator.
465 (Vrevsuff): New mode attribute.
467 2014-04-24 Terry Guo <terry.guo@arm.com>
469 * config/arm/arm.h (machine_function): Define variable
470 after_arm_reorg here.
471 * config/arm/arm.c (after_arm_reorg): Remove the definition.
472 (arm_split_constant): Update the way to access variable
475 (arm_output_function_epilogue): Remove the reset of after_arm_reorg.
477 2014-04-23 Tom de Vries <tom@codesourcery.com>
479 * target-hooks-macros.h: Fix DEFHOOKPOD argument order in comment.
481 2014-04-23 David Malcolm <dmalcolm@redhat.com>
483 * is-a.h: Update comments to reflect the following changes to the
484 "pointerness" of the API, making the template parameter match the
485 return type, allowing use of is-a.h with typedefs of pointers.
486 (is_a_helper::cast): Return a T rather then a pointer to a T, so
487 that the return type matches the parameter to the is_a_helper.
489 (dyn_cast): Likewise.
491 * cgraph.c (cgraph_node_for_asm): Update for removal of implicit
492 pointer from the is-a.h API.
494 * cgraph.h (is_a_helper <cgraph_node>::test): Convert to...
495 (is_a_helper <cgraph_node *>::test): ...this, matching change to
497 (is_a_helper <varpool_node>::test): Likewise, convert to...
498 (is_a_helper <varpool_node *>::test): ...this.
500 (varpool_first_variable): Update for removal of implicit pointer
502 (varpool_next_variable): Likewise.
503 (varpool_first_static_initializer): Likewise.
504 (varpool_next_static_initializer): Likewise.
505 (varpool_first_defined_variable): Likewise.
506 (varpool_next_defined_variable): Likewise.
507 (cgraph_first_defined_function): Likewise.
508 (cgraph_next_defined_function): Likewise.
509 (cgraph_first_function): Likewise.
510 (cgraph_next_function): Likewise.
511 (cgraph_first_function_with_gimple_body): Likewise.
512 (cgraph_next_function_with_gimple_body): Likewise.
513 (cgraph_alias_target): Likewise.
514 (varpool_alias_target): Likewise.
515 (cgraph_function_or_thunk_node): Likewise.
516 (varpool_variable_node): Likewise.
517 (symtab_real_symbol_p): Likewise.
518 * cgraphunit.c (referred_to_p): Likewise.
519 (analyze_functions): Likewise.
520 (handle_alias_pairs): Likewise.
521 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
522 * gimple-ssa.h (gimple_vuse_op): Likewise.
523 (gimple_vdef_op): Likewise.
524 * gimple-streamer-in.c (input_gimple_stmt): Likewise.
525 * gimple.c (gimple_build_asm_1): Likewise.
526 (gimple_build_try): Likewise.
527 (gimple_build_resx): Likewise.
528 (gimple_build_eh_dispatch): Likewise.
529 (gimple_build_omp_for): Likewise.
530 (gimple_omp_for_set_clauses): Likewise.
532 * gimple.h (is_a_helper <gimple_statement_asm>::test): Convert to...
533 (is_a_helper <gimple_statement_asm *>::test): ...this.
534 (is_a_helper <gimple_statement_bind>::test): Convert to...
535 (is_a_helper <gimple_statement_bind *>::test): ...this.
536 (is_a_helper <gimple_statement_call>::test): Convert to...
537 (is_a_helper <gimple_statement_call *>::test): ...this.
538 (is_a_helper <gimple_statement_catch>::test): Convert to...
539 (is_a_helper <gimple_statement_catch *>::test): ...this.
540 (is_a_helper <gimple_statement_resx>::test): Convert to...
541 (is_a_helper <gimple_statement_resx *>::test): ...this.
542 (is_a_helper <gimple_statement_eh_dispatch>::test): Convert to...
543 (is_a_helper <gimple_statement_eh_dispatch *>::test): ...this.
544 (is_a_helper <gimple_statement_eh_else>::test): Convert to...
545 (is_a_helper <gimple_statement_eh_else *>::test): ...this.
546 (is_a_helper <gimple_statement_eh_filter>::test): Convert to...
547 (is_a_helper <gimple_statement_eh_filter *>::test): ...this.
548 (is_a_helper <gimple_statement_eh_mnt>::test): Convert to...
549 (is_a_helper <gimple_statement_eh_mnt *>::test): ...this.
550 (is_a_helper <gimple_statement_omp_atomic_load>::test): Convert to...
551 (is_a_helper <gimple_statement_omp_atomic_load *>::test): ...this.
552 (is_a_helper <gimple_statement_omp_atomic_store>::test): Convert to...
553 (is_a_helper <gimple_statement_omp_atomic_store *>::test): ...this.
554 (is_a_helper <gimple_statement_omp_return>::test): Convert to...
555 (is_a_helper <gimple_statement_omp_return *>::test): ...this.
556 (is_a_helper <gimple_statement_omp_continue>::test): Convert to...
557 (is_a_helper <gimple_statement_omp_continue *>::test): ...this.
558 (is_a_helper <gimple_statement_omp_critical>::test): Convert to...
559 (is_a_helper <gimple_statement_omp_critical *>::test): ...this.
560 (is_a_helper <gimple_statement_omp_for>::test): Convert to...
561 (is_a_helper <gimple_statement_omp_for *>::test): ...this.
562 (is_a_helper <gimple_statement_omp_taskreg>::test): Convert to...
563 (is_a_helper <gimple_statement_omp_taskreg *>::test): ...this.
564 (is_a_helper <gimple_statement_omp_parallel>::test): Convert to...
565 (is_a_helper <gimple_statement_omp_parallel *>::test): ...this.
566 (is_a_helper <gimple_statement_omp_target>::test): Convert to...
567 (is_a_helper <gimple_statement_omp_target *>::test): ...this.
568 (is_a_helper <gimple_statement_omp_sections>::test): Convert to...
569 (is_a_helper <gimple_statement_omp_sections *>::test): ...this.
570 (is_a_helper <gimple_statement_omp_single>::test): Convert to...
571 (is_a_helper <gimple_statement_omp_single *>::test): ...this.
572 (is_a_helper <gimple_statement_omp_teams>::test): Convert to...
573 (is_a_helper <gimple_statement_omp_teams *>::test): ...this.
574 (is_a_helper <gimple_statement_omp_task>::test): Convert to...
575 (is_a_helper <gimple_statement_omp_task *>::test): ...this.
576 (is_a_helper <gimple_statement_phi>::test): Convert to...
577 (is_a_helper <gimple_statement_phi *>::test): ...this.
578 (is_a_helper <gimple_statement_transaction>::test): Convert to...
579 (is_a_helper <gimple_statement_transaction *>::test): ...this.
580 (is_a_helper <gimple_statement_try>::test): Convert to...
581 (is_a_helper <gimple_statement_try *>::test): ...this.
582 (is_a_helper <gimple_statement_wce>::test): Convert to...
583 (is_a_helper <gimple_statement_wce *>::test): ...this.
584 (is_a_helper <const gimple_statement_asm>::test): Convert to...
585 (is_a_helper <const gimple_statement_asm *>::test): ...this.
586 (is_a_helper <const gimple_statement_bind>::test): Convert to...
587 (is_a_helper <const gimple_statement_bind *>::test): ...this.
588 (is_a_helper <const gimple_statement_call>::test): Convert to...
589 (is_a_helper <const gimple_statement_call *>::test): ...this.
590 (is_a_helper <const gimple_statement_catch>::test): Convert to...
591 (is_a_helper <const gimple_statement_catch *>::test): ...this.
592 (is_a_helper <const gimple_statement_resx>::test): Convert to...
593 (is_a_helper <const gimple_statement_resx *>::test): ...this.
594 (is_a_helper <const gimple_statement_eh_dispatch>::test):
596 (is_a_helper <const gimple_statement_eh_dispatch *>::test):
598 (is_a_helper <const gimple_statement_eh_filter>::test): Convert
600 (is_a_helper <const gimple_statement_eh_filter *>::test): ...this.
601 (is_a_helper <const gimple_statement_omp_atomic_load>::test):
603 (is_a_helper <const gimple_statement_omp_atomic_load *>::test):
605 (is_a_helper <const gimple_statement_omp_atomic_store>::test):
607 (is_a_helper <const gimple_statement_omp_atomic_store *>::test):
609 (is_a_helper <const gimple_statement_omp_return>::test): Convert
611 (is_a_helper <const gimple_statement_omp_return *>::test): ...this.
612 (is_a_helper <const gimple_statement_omp_continue>::test): Convert
614 (is_a_helper <const gimple_statement_omp_continue *>::test): ...this.
615 (is_a_helper <const gimple_statement_omp_critical>::test): Convert
617 (is_a_helper <const gimple_statement_omp_critical *>::test): ...this.
618 (is_a_helper <const gimple_statement_omp_for>::test): Convert to...
619 (is_a_helper <const gimple_statement_omp_for *>::test): ...this.
620 (is_a_helper <const gimple_statement_omp_taskreg>::test): Convert
622 (is_a_helper <const gimple_statement_omp_taskreg *>::test): ...this.
623 (is_a_helper <const gimple_statement_omp_parallel>::test): Convert
625 (is_a_helper <const gimple_statement_omp_parallel *>::test): ...this.
626 (is_a_helper <const gimple_statement_omp_target>::test): Convert
628 (is_a_helper <const gimple_statement_omp_target *>::test): ...this.
629 (is_a_helper <const gimple_statement_omp_sections>::test): Convert
631 (is_a_helper <const gimple_statement_omp_sections *>::test): ...this.
632 (is_a_helper <const gimple_statement_omp_single>::test): Convert
634 (is_a_helper <const gimple_statement_omp_single *>::test): ...this.
635 (is_a_helper <const gimple_statement_omp_teams>::test): Convert
637 (is_a_helper <const gimple_statement_omp_teams *>::test): ...this.
638 (is_a_helper <const gimple_statement_omp_task>::test): Convert
640 (is_a_helper <const gimple_statement_omp_task *>::test): ...this.
641 (is_a_helper <const gimple_statement_phi>::test): Convert to...
642 (is_a_helper <const gimple_statement_phi *>::test): ...this.
643 (is_a_helper <const gimple_statement_transaction>::test): Convert
645 (is_a_helper <const gimple_statement_transaction *>::test): ...this.
646 (is_a_helper <const gimple_statement_with_ops>::test): Convert
648 (is_a_helper <const gimple_statement_with_ops *>::test): ...this.
649 (is_a_helper <gimple_statement_with_ops>::test): Convert to...
650 (is_a_helper <gimple_statement_with_ops *>::test): ...this.
651 (is_a_helper <const gimple_statement_with_memory_ops>::test):
653 (is_a_helper <const gimple_statement_with_memory_ops *>::test):
655 (is_a_helper <gimple_statement_with_memory_ops>::test): Convert
657 (is_a_helper <gimple_statement_with_memory_ops *>::test): ...this.
659 (gimple_use_ops): Update for removal of implicit pointer from the
661 (gimple_set_use_ops): Likewise.
662 (gimple_vuse): Likewise.
663 (gimple_vdef): Likewise.
664 (gimple_vuse_ptr): Likewise.
665 (gimple_vdef_ptr): Likewise.
666 (gimple_set_vuse): Likewise.
667 (gimple_set_vdef): Likewise.
668 (gimple_omp_return_set_lhs): Likewise.
669 (gimple_omp_return_lhs): Likewise.
670 (gimple_omp_return_lhs_ptr): Likewise.
671 (gimple_call_fntype): Likewise.
672 (gimple_call_set_fntype): Likewise.
673 (gimple_call_set_internal_fn): Likewise.
674 (gimple_call_use_set): Likewise.
675 (gimple_call_clobber_set): Likewise.
676 (gimple_bind_vars): Likewise.
677 (gimple_bind_set_vars): Likewise.
678 (gimple_bind_body_ptr): Likewise.
679 (gimple_bind_set_body): Likewise.
680 (gimple_bind_add_stmt): Likewise.
681 (gimple_bind_block): Likewise.
682 (gimple_bind_set_block): Likewise.
683 (gimple_asm_ninputs): Likewise.
684 (gimple_asm_noutputs): Likewise.
685 (gimple_asm_nclobbers): Likewise.
686 (gimple_asm_nlabels): Likewise.
687 (gimple_asm_input_op): Likewise.
688 (gimple_asm_input_op_ptr): Likewise.
689 (gimple_asm_output_op): Likewise.
690 (gimple_asm_output_op_ptr): Likewise.
691 (gimple_asm_set_output_op): Likewise.
692 (gimple_asm_clobber_op): Likewise.
693 (gimple_asm_set_clobber_op): Likewise.
694 (gimple_asm_label_op): Likewise.
695 (gimple_asm_set_label_op): Likewise.
696 (gimple_asm_string): Likewise.
697 (gimple_catch_types): Likewise.
698 (gimple_catch_types_ptr): Likewise.
699 (gimple_catch_handler_ptr): Likewise.
700 (gimple_catch_set_types): Likewise.
701 (gimple_catch_set_handler): Likewise.
702 (gimple_eh_filter_types): Likewise.
703 (gimple_eh_filter_types_ptr): Likewise.
704 (gimple_eh_filter_failure_ptr): Likewise.
705 (gimple_eh_filter_set_types): Likewise.
706 (gimple_eh_filter_set_failure): Likewise.
707 (gimple_eh_must_not_throw_fndecl): Likewise.
708 (gimple_eh_must_not_throw_set_fndecl): Likewise.
709 (gimple_eh_else_n_body_ptr): Likewise.
710 (gimple_eh_else_e_body_ptr): Likewise.
711 (gimple_eh_else_set_n_body): Likewise.
712 (gimple_eh_else_set_e_body): Likewise.
713 (gimple_try_eval_ptr): Likewise.
714 (gimple_try_cleanup_ptr): Likewise.
715 (gimple_try_set_eval): Likewise.
716 (gimple_try_set_cleanup): Likewise.
717 (gimple_wce_cleanup_ptr): Likewise.
718 (gimple_wce_set_cleanup): Likewise.
719 (gimple_phi_capacity): Likewise.
720 (gimple_phi_num_args): Likewise.
721 (gimple_phi_result): Likewise.
722 (gimple_phi_result_ptr): Likewise.
723 (gimple_phi_set_result): Likewise.
724 (gimple_phi_arg): Likewise.
725 (gimple_phi_set_arg): Likewise.
726 (gimple_resx_region): Likewise.
727 (gimple_resx_set_region): Likewise.
728 (gimple_eh_dispatch_region): Likewise.
729 (gimple_eh_dispatch_set_region): Likewise.
730 (gimple_omp_critical_name): Likewise.
731 (gimple_omp_critical_name_ptr): Likewise.
732 (gimple_omp_critical_set_name): Likewise.
733 (gimple_omp_for_clauses): Likewise.
734 (gimple_omp_for_clauses_ptr): Likewise.
735 (gimple_omp_for_set_clauses): Likewise.
736 (gimple_omp_for_collapse): Likewise.
737 (gimple_omp_for_index): Likewise.
738 (gimple_omp_for_index_ptr): Likewise.
739 (gimple_omp_for_set_index): Likewise.
740 (gimple_omp_for_initial): Likewise.
741 (gimple_omp_for_initial_ptr): Likewise.
742 (gimple_omp_for_set_initial): Likewise.
743 (gimple_omp_for_final): Likewise.
744 (gimple_omp_for_final_ptr): Likewise.
745 (gimple_omp_for_set_final): Likewise.
746 (gimple_omp_for_incr): Likewise.
747 (gimple_omp_for_incr_ptr): Likewise.
748 (gimple_omp_for_set_incr): Likewise.
749 (gimple_omp_for_pre_body_ptr): Likewise.
750 (gimple_omp_for_set_pre_body): Likewise.
751 (gimple_omp_parallel_clauses): Likewise.
752 (gimple_omp_parallel_clauses_ptr): Likewise.
753 (gimple_omp_parallel_set_clauses): Likewise.
754 (gimple_omp_parallel_child_fn): Likewise.
755 (gimple_omp_parallel_child_fn_ptr): Likewise.
756 (gimple_omp_parallel_set_child_fn): Likewise.
757 (gimple_omp_parallel_data_arg): Likewise.
758 (gimple_omp_parallel_data_arg_ptr): Likewise.
759 (gimple_omp_parallel_set_data_arg): Likewise.
760 (gimple_omp_task_clauses): Likewise.
761 (gimple_omp_task_clauses_ptr): Likewise.
762 (gimple_omp_task_set_clauses): Likewise.
763 (gimple_omp_task_child_fn): Likewise.
764 (gimple_omp_task_child_fn_ptr): Likewise.
765 (gimple_omp_task_set_child_fn): Likewise.
766 (gimple_omp_task_data_arg): Likewise.
767 (gimple_omp_task_data_arg_ptr): Likewise.
768 (gimple_omp_task_set_data_arg): Likewise.
769 (gimple_omp_taskreg_clauses): Likewise.
770 (gimple_omp_taskreg_clauses_ptr): Likewise.
771 (gimple_omp_taskreg_set_clauses): Likewise.
772 (gimple_omp_taskreg_child_fn): Likewise.
773 (gimple_omp_taskreg_child_fn_ptr): Likewise.
774 (gimple_omp_taskreg_set_child_fn): Likewise.
775 (gimple_omp_taskreg_data_arg): Likewise.
776 (gimple_omp_taskreg_data_arg_ptr): Likewise.
777 (gimple_omp_taskreg_set_data_arg): Likewise.
778 (gimple_omp_task_copy_fn): Likewise.
779 (gimple_omp_task_copy_fn_ptr): Likewise.
780 (gimple_omp_task_set_copy_fn): Likewise.
781 (gimple_omp_task_arg_size): Likewise.
782 (gimple_omp_task_arg_size_ptr): Likewise.
783 (gimple_omp_task_set_arg_size): Likewise.
784 (gimple_omp_task_arg_align): Likewise.
785 (gimple_omp_task_arg_align_ptr): Likewise.
786 (gimple_omp_task_set_arg_align): Likewise.
787 (gimple_omp_single_clauses): Likewise.
788 (gimple_omp_single_clauses_ptr): Likewise.
789 (gimple_omp_single_set_clauses): Likewise.
790 (gimple_omp_target_clauses): Likewise.
791 (gimple_omp_target_clauses_ptr): Likewise.
792 (gimple_omp_target_set_clauses): Likewise.
793 (gimple_omp_target_child_fn): Likewise.
794 (gimple_omp_target_child_fn_ptr): Likewise.
795 (gimple_omp_target_set_child_fn): Likewise.
796 (gimple_omp_target_data_arg): Likewise.
797 (gimple_omp_target_data_arg_ptr): Likewise.
798 (gimple_omp_target_set_data_arg): Likewise.
799 (gimple_omp_teams_clauses): Likewise.
800 (gimple_omp_teams_clauses_ptr): Likewise.
801 (gimple_omp_teams_set_clauses): Likewise.
802 (gimple_omp_sections_clauses): Likewise.
803 (gimple_omp_sections_clauses_ptr): Likewise.
804 (gimple_omp_sections_set_clauses): Likewise.
805 (gimple_omp_sections_control): Likewise.
806 (gimple_omp_sections_control_ptr): Likewise.
807 (gimple_omp_sections_set_control): Likewise.
808 (gimple_omp_for_set_cond): Likewise.
809 (gimple_omp_for_cond): Likewise.
810 (gimple_omp_atomic_store_set_val): Likewise.
811 (gimple_omp_atomic_store_val): Likewise.
812 (gimple_omp_atomic_store_val_ptr): Likewise.
813 (gimple_omp_atomic_load_set_lhs): Likewise.
814 (gimple_omp_atomic_load_lhs): Likewise.
815 (gimple_omp_atomic_load_lhs_ptr): Likewise.
816 (gimple_omp_atomic_load_set_rhs): Likewise.
817 (gimple_omp_atomic_load_rhs): Likewise.
818 (gimple_omp_atomic_load_rhs_ptr): Likewise.
819 (gimple_omp_continue_control_def): Likewise.
820 (gimple_omp_continue_control_def_ptr): Likewise.
821 (gimple_omp_continue_set_control_def): Likewise.
822 (gimple_omp_continue_control_use): Likewise.
823 (gimple_omp_continue_control_use_ptr): Likewise.
824 (gimple_omp_continue_set_control_use): Likewise.
825 (gimple_transaction_body_ptr): Likewise.
826 (gimple_transaction_label): Likewise.
827 (gimple_transaction_label_ptr): Likewise.
828 (gimple_transaction_set_body): Likewise.
829 (gimple_transaction_set_label): Likewise.
831 * ipa-devirt.c (build_type_inheritance_graph): Likewise.
832 * ipa-inline-analysis.c (inline_write_summary): Likewise.
833 * ipa-ref.c (ipa_record_reference): Likewise.
834 * ipa-reference.c (analyze_function): Likewise.
835 (ipa_reference_write_optimization_summary): Likewise.
836 * ipa.c (symtab_remove_unreachable_nodes): Likewise.
837 (address_taken_from_non_vtable_p): Likewise.
838 (comdat_can_be_unshared_p_1): Likewise.
839 * lto-cgraph.c (lto_output_ref): Likewise.
840 (add_references): Likewise.
841 (compute_ltrans_boundary): Likewise.
842 (output_symtab): Likewise.
843 (input_ref): Likewise.
844 (input_cgraph_1): Likewise.
845 (output_cgraph_opt_summary): Likewise.
846 * lto-streamer-out.c (lto_output): Likewise.
847 (output_symbol_p): Likewise.
848 * lto-streamer.h (lsei_next_function_in_partition): Likewise.
849 (lsei_start_function_in_partition): Likewise.
850 (lsei_next_variable_in_partition): Likewise.
851 (lsei_start_variable_in_partition): Likewise.
852 * symtab.c (insert_to_assembler_name_hash): Likewise.
853 (unlink_from_assembler_name_hash): Likewise.
854 (symtab_unregister_node): Likewise.
855 (symtab_remove_node): Likewise.
856 (dump_symtab_node): Likewise.
857 (verify_symtab_base): Likewise.
858 (verify_symtab_node): Likewise.
859 (symtab_make_decl_local): Likewise.
860 (symtab_alias_ultimate_target): Likewise.
861 (symtab_resolve_alias): Likewise.
862 (symtab_get_symbol_partitioning_class): Likewise.
863 * tree-phinodes.c (allocate_phi_node): Likewise.
864 (reserve_phi_args_for_new_edge): Likewise.
865 (remove_phi_args): Likewise.
866 * varpool.c (varpool_node_for_asm): Likewise.
867 (varpool_remove_unreferenced_decls): Likewise.
869 2014-04-23 Jeff Law <law@redhat.com>
871 PR tree-optimization/60902
872 * tree-ssa-threadedge.c
873 (record_temporary_equivalences_from_stmts_at_dest): Make sure to
874 invalidate outputs from statements that do not produce useful
875 outputs for threading.
877 2014-04-23 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
879 * config/aarch64/aarch64.md (stack_protect_set, stack_protect_test)
880 (stack_protect_set_<mode>, stack_protect_test_<mode>): Add
881 machine descriptions for Stack Smashing Protector.
883 2014-04-23 Richard Earnshaw <rearnsha@arm.com>
885 * aarch64.md (<optab>_rol<mode>3): New pattern.
886 (<optab>_rolsi3_uxtw): Likewise.
887 * aarch64.c (aarch64_strip_shift): Handle ROTATE and ROTATERT.
889 2014-04-23 James Greenhalgh <james.greenhalgh@arm.com>
891 * config/arm/arm.c (arm_cortex_a57_tune): Initialize all fields.
892 (arm_cortex_a12_tune): Likewise.
894 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
896 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle BSWAP.
898 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
900 * config/arm/arm.md (arm_rev16si2): New pattern.
901 (arm_rev16si2_alt): Likewise.
902 * config/arm/arm.c (arm_new_rtx_costs): Handle rev16 case.
904 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
906 * config/aarch64/aarch64.md (rev16<mode>2): New pattern.
907 (rev16<mode>2_alt): Likewise.
908 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle rev16 case.
909 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p): New.
910 (aarch_rev16_shleft_mask_imm_p): Likewise.
911 (aarch_rev16_p_1): Likewise.
912 (aarch_rev16_p): Likewise.
913 * config/arm/aarch-common-protos.h (aarch_rev16_p): Declare extern.
914 (aarch_rev16_shright_mask_imm_p): Likewise.
915 (aarch_rev16_shleft_mask_imm_p): Likewise.
917 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
919 * config/arm/aarch-common-protos.h (alu_cost_table): Add rev field.
920 * config/arm/aarch-cost-tables.h (generic_extra_costs): Specify
922 (cortex_a53_extra_costs): Likewise.
923 (cortex_a57_extra_costs): Likewise.
924 * config/arm/arm.c (cortexa9_extra_costs): Likewise.
925 (cortexa7_extra_costs): Likewise.
926 (cortexa8_extra_costs): Likewise.
927 (cortexa12_extra_costs): Likewise.
928 (cortexa15_extra_costs): Likewise.
929 (v7m_extra_costs): Likewise.
930 (arm_new_rtx_costs): Handle BSWAP.
932 2013-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
934 * config/arm/arm.c (cortexa8_extra_costs): New table.
935 (arm_cortex_a8_tune): New tuning struct.
936 * config/arm/arm-cores.def (cortex-a8): Use cortex_a8 tuning struct.
938 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
940 * config/arm/arm.c (arm_new_rtx_costs): Handle FMA.
942 2014-04-23 Richard Biener <rguenther@suse.de>
944 * Makefile.in (OBJS): Remove loop-unswitch.o.
945 * loop-unswitch.c: Delete.
946 * tree-pass.h (make_pass_rtl_unswitch): Remove.
947 * passes.def (pass_rtl_unswitch): Likewise.
948 * loop-init.c (gate_rtl_unswitch): Likewise.
949 (rtl_unswitch): Likewise.
950 (pass_data_rtl_unswitch): Likewise.
951 (pass_rtl_unswitch): Likewise.
952 (make_pass_rtl_unswitch): Likewise.
953 * rtl.h (reversed_condition): Likewise.
954 (compare_and_jump_seq): Likewise.
955 * loop-iv.c (reversed_condition): Move here from loop-unswitch.c
957 * loop-unroll.c (compare_and_jump_seq): Likewise.
959 2014-04-23 Richard Biener <rguenther@suse.de>
961 PR tree-optimization/60903
962 * tree-ssa-loop-im.c (analyze_memory_references): Remove
963 commented code block.
964 (execute_sm_if_changed): Properly apply IRREDUCIBLE_LOOP
965 loop flags to newly created BBs and edges.
967 2014-04-23 Nick Clifton <nickc@redhat.com>
969 * config/msp430/msp430.c (msp430_handle_option): Move function
971 (msp430_option_override): Simplify mcu and mcpu option handling.
972 (msp430_is_f5_mcu): Rename to msp430_use_f5_series_hwmult. Add
973 support for -mhwmult command line option.
974 (has_32bit_hwmult): Rename to use_32bit_hwmult. Add support for
975 -mhwmult command line option.
976 (msp430_hwmult_enabled): Delete.
977 (msp43o_output_labelref): Add support for -mhwmult command line
979 * config/msp430/msp430.md (mulhisi3, umulhisi3, mulsidi3)
980 (umulsidi3): Likewise.
981 * config/msp430/msp430.opt (mmcu): Add Report attribute.
982 (mcpu, mlarge, msmall): Likewise.
983 (mhwmult): New option.
984 * config/msp430/msp430-protos.h (msp430_hwmult_enabled): Remove
986 (msp430_is_f5_mcu): Remove prototype.
987 (msp430_use_f5_series_hwmult): Add prototype.
988 * config/msp430/msp430-opts.h: New file.
989 * common/config/msp430: New directory.
990 * common/config/msp430/msp430-common.c: New file.
991 * config.gcc (msp430): Remove target_has_targetm_common.
992 * doc/invoke.texi: Document -mhwmult command line option.
994 2014-04-23 Nick Clifton <nickc@redhat.com>
996 * config/i386/cygwin.h (ENDFILE_SPEC): Include
997 default-manifest.o if it can be found in the search path.
998 * config/i386/mingw32.h (ENDFILE_SPEC): Likewise.
1000 2014-04-23 Terry Guo <terry.guo@arm.com>
1002 * config/arm/arm.h (ASM_APP_OFF): Re-define it in a cleaner way.
1004 2014-04-23 Richard Biener <rguenther@suse.de>
1007 * tree-inline.c (declare_return_variable): Use mark_addressable.
1009 2014-04-23 Richard Biener <rguenther@suse.de>
1012 * loop-init.c (loop_optimizer_init): Make sure to apply
1013 LOOPS_MAY_HAVE_MULTIPLE_LATCHES before fixing up loops.
1015 2014-04-22 Jakub Jelinek <jakub@redhat.com>
1018 * common.opt (fsanitize-recover, fsanitize-undefined-trap-on-error):
1020 * gcc.c (sanitize_spec_function): Don't return "" for "undefined"
1021 if flag_sanitize_undefined_trap_on_error.
1022 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW_ABORT,
1023 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS_ABORT,
1024 BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE_ABORT,
1025 BUILT_IN_UBSAN_HANDLE_TYPE_MISMATCH_ABORT,
1026 BUILT_IN_UBSAN_HANDLE_ADD_OVERFLOW_ABORT,
1027 BUILT_IN_UBSAN_HANDLE_SUB_OVERFLOW_ABORT,
1028 BUILT_IN_UBSAN_HANDLE_MUL_OVERFLOW_ABORT,
1029 BUILT_IN_UBSAN_HANDLE_NEGATE_OVERFLOW_ABORT,
1030 BUILT_IN_UBSAN_HANDLE_LOAD_INVALID_VALUE_ABORT): New builtins.
1031 * ubsan.c (ubsan_instrument_unreachable): Return
1032 __builtin_trap () if flag_sanitize_undefined_trap_on_error.
1033 (ubsan_expand_null_ifn): Emit __builtin_trap ()
1034 if flag_sanitize_undefined_trap_on_error and
1035 __ubsan_handle_type_mismatch_abort if !flag_sanitize_recover.
1036 (ubsan_expand_null_ifn, ubsan_build_overflow_builtin,
1037 instrument_bool_enum_load): Emit __builtin_trap () if
1038 flag_sanitize_undefined_trap_on_error and
1039 __builtin_handle_*_abort () if !flag_sanitize_recover.
1040 * doc/invoke.texi (-fsanitize-recover,
1041 -fsanitize-undefined-trap-on-error): Document.
1043 2014-04-22 Christian Bruel <christian.bruel@st.com>
1045 * config/sh/sh.md (mov<mode>): Replace movQIHI.
1046 Force immediates to SImode.
1048 2014-04-22 Sandra Loosemore <sandra@codesourcery.com>
1050 * config/nios2/nios2.md (UNSPEC_ROUND): New.
1052 * config/nios2/nios2.opt (mno-custom-round, mcustom-round=): New.
1053 * config/nios2/nios2-opts.h (N2FPU_ALL_CODES): Add round.
1054 * config/nios2/nios2.c (N2F_NO_ERRNO): Define.
1055 (nios2_fpu_insn): Add entry for round.
1056 (N2FPU_NO_ERRNO_P): Define.
1057 (nios2_custom_check_insns): Add check for N2F_NO_ERRNO and
1059 * doc/invoke.texi (Nios II Options): Document -mcustom-round.
1061 2014-04-22 Richard Henderson <rth@redhat.com>
1063 * config/aarch64/aarch64 (addti3, subti3): New expanders.
1064 (add<GPI>3_compare0): Remove leading * from name.
1065 (add<GPI>3_carryin): Likewise.
1066 (sub<GPI>3_compare0): Likewise.
1067 (sub<GPI>3_carryin): Likewise.
1068 (<su_optab>mulditi3): New expander.
1069 (multi3): New expander.
1070 (madd<GPI>): Remove leading * from name.
1072 2014-04-22 Martin Jambor <mjambor@suse.cz>
1074 * cgraphclones.c (cgraph_function_versioning): Copy
1075 ipa_transforms_to_apply instead of asserting it is empty.
1077 2014-04-22 H.J. Lu <hongjiu.lu@intel.com>
1080 * config/i386/i386.c (ix86_expand_set_or_movmem): Call counter_mode
1081 on count_exp to get mode.
1083 2014-04-22 Andrew Pinski <apinski@cavium.com>
1085 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
1086 Handle TLS for ILP32.
1087 * config/aarch64/aarch64.md (tlsie_small): Rename to ...
1088 (tlsie_small_<mode>): this and handle PTR.
1089 (tlsie_small_sidi): New pattern.
1090 (tlsle_small): Change to an expand to handle ILP32.
1091 (tlsle_small_<mode>): New pattern.
1092 (tlsdesc_small): Rename to ...
1093 (tlsdesc_small_<mode>): this and handle PTR.
1095 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1097 * config/aarch64/aarch64.c (TARGET_FLAGS_REGNUM): Define.
1099 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1101 * config/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
1102 (aarch64_types_signed_unsigned_qualifiers): Qualifier added.
1103 (aarch64_types_signed_poly_qualifiers): Likewise.
1104 (aarch64_types_unsigned_signed_qualifiers): Likewise.
1105 (aarch64_types_poly_signed_qualifiers): Likewise.
1106 (TYPES_REINTERP_SS): Type macro added.
1107 (TYPES_REINTERP_SU): Likewise.
1108 (TYPES_REINTERP_SP): Likewise.
1109 (TYPES_REINTERP_US): Likewise.
1110 (TYPES_REINTERP_PS): Likewise.
1111 (aarch64_fold_builtin): New expression folding added.
1112 * config/aarch64/aarch64-simd-builtins.def (REINTERP):
1113 Declarations removed.
1114 (REINTERP_SS): Declarations added.
1115 (REINTERP_US): Likewise.
1116 (REINTERP_PS): Likewise.
1117 (REINTERP_SU): Likewise.
1118 (REINTERP_SP): Likewise.
1119 * config/aarch64/arm_neon.h (vreinterpret_p8_f64): Implemented.
1120 (vreinterpretq_p8_f64): Likewise.
1121 (vreinterpret_p16_f64): Likewise.
1122 (vreinterpretq_p16_f64): Likewise.
1123 (vreinterpret_f32_f64): Likewise.
1124 (vreinterpretq_f32_f64): Likewise.
1125 (vreinterpret_f64_f32): Likewise.
1126 (vreinterpret_f64_p8): Likewise.
1127 (vreinterpret_f64_p16): Likewise.
1128 (vreinterpret_f64_s8): Likewise.
1129 (vreinterpret_f64_s16): Likewise.
1130 (vreinterpret_f64_s32): Likewise.
1131 (vreinterpret_f64_s64): Likewise.
1132 (vreinterpret_f64_u8): Likewise.
1133 (vreinterpret_f64_u16): Likewise.
1134 (vreinterpret_f64_u32): Likewise.
1135 (vreinterpret_f64_u64): Likewise.
1136 (vreinterpretq_f64_f32): Likewise.
1137 (vreinterpretq_f64_p8): Likewise.
1138 (vreinterpretq_f64_p16): Likewise.
1139 (vreinterpretq_f64_s8): Likewise.
1140 (vreinterpretq_f64_s16): Likewise.
1141 (vreinterpretq_f64_s32): Likewise.
1142 (vreinterpretq_f64_s64): Likewise.
1143 (vreinterpretq_f64_u8): Likewise.
1144 (vreinterpretq_f64_u16): Likewise.
1145 (vreinterpretq_f64_u32): Likewise.
1146 (vreinterpretq_f64_u64): Likewise.
1147 (vreinterpret_s64_f64): Likewise.
1148 (vreinterpretq_s64_f64): Likewise.
1149 (vreinterpret_u64_f64): Likewise.
1150 (vreinterpretq_u64_f64): Likewise.
1151 (vreinterpret_s8_f64): Likewise.
1152 (vreinterpretq_s8_f64): Likewise.
1153 (vreinterpret_s16_f64): Likewise.
1154 (vreinterpretq_s16_f64): Likewise.
1155 (vreinterpret_s32_f64): Likewise.
1156 (vreinterpretq_s32_f64): Likewise.
1157 (vreinterpret_u8_f64): Likewise.
1158 (vreinterpretq_u8_f64): Likewise.
1159 (vreinterpret_u16_f64): Likewise.
1160 (vreinterpretq_u16_f64): Likewise.
1161 (vreinterpret_u32_f64): Likewise.
1162 (vreinterpretq_u32_f64): Likewise.
1164 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1166 * config/aarch64/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
1167 * config/aarch64/aarch64/aarch64-simd-builtins.def (REINTERP): Removed.
1168 (vreinterpret_p8_s8): Likewise.
1169 * config/aarch64/aarch64/arm_neon.h (vreinterpret_p8_s8): Uses cast.
1170 (vreinterpret_p8_s16): Likewise.
1171 (vreinterpret_p8_s32): Likewise.
1172 (vreinterpret_p8_s64): Likewise.
1173 (vreinterpret_p8_f32): Likewise.
1174 (vreinterpret_p8_u8): Likewise.
1175 (vreinterpret_p8_u16): Likewise.
1176 (vreinterpret_p8_u32): Likewise.
1177 (vreinterpret_p8_u64): Likewise.
1178 (vreinterpret_p8_p16): Likewise.
1179 (vreinterpretq_p8_s8): Likewise.
1180 (vreinterpretq_p8_s16): Likewise.
1181 (vreinterpretq_p8_s32): Likewise.
1182 (vreinterpretq_p8_s64): Likewise.
1183 (vreinterpretq_p8_f32): Likewise.
1184 (vreinterpretq_p8_u8): Likewise.
1185 (vreinterpretq_p8_u16): Likewise.
1186 (vreinterpretq_p8_u32): Likewise.
1187 (vreinterpretq_p8_u64): Likewise.
1188 (vreinterpretq_p8_p16): Likewise.
1189 (vreinterpret_p16_s8): Likewise.
1190 (vreinterpret_p16_s16): Likewise.
1191 (vreinterpret_p16_s32): Likewise.
1192 (vreinterpret_p16_s64): Likewise.
1193 (vreinterpret_p16_f32): Likewise.
1194 (vreinterpret_p16_u8): Likewise.
1195 (vreinterpret_p16_u16): Likewise.
1196 (vreinterpret_p16_u32): Likewise.
1197 (vreinterpret_p16_u64): Likewise.
1198 (vreinterpret_p16_p8): Likewise.
1199 (vreinterpretq_p16_s8): Likewise.
1200 (vreinterpretq_p16_s16): Likewise.
1201 (vreinterpretq_p16_s32): Likewise.
1202 (vreinterpretq_p16_s64): Likewise.
1203 (vreinterpretq_p16_f32): Likewise.
1204 (vreinterpretq_p16_u8): Likewise.
1205 (vreinterpretq_p16_u16): Likewise.
1206 (vreinterpretq_p16_u32): Likewise.
1207 (vreinterpretq_p16_u64): Likewise.
1208 (vreinterpretq_p16_p8): Likewise.
1209 (vreinterpret_f32_s8): Likewise.
1210 (vreinterpret_f32_s16): Likewise.
1211 (vreinterpret_f32_s32): Likewise.
1212 (vreinterpret_f32_s64): Likewise.
1213 (vreinterpret_f32_u8): Likewise.
1214 (vreinterpret_f32_u16): Likewise.
1215 (vreinterpret_f32_u32): Likewise.
1216 (vreinterpret_f32_u64): Likewise.
1217 (vreinterpret_f32_p8): Likewise.
1218 (vreinterpret_f32_p16): Likewise.
1219 (vreinterpretq_f32_s8): Likewise.
1220 (vreinterpretq_f32_s16): Likewise.
1221 (vreinterpretq_f32_s32): Likewise.
1222 (vreinterpretq_f32_s64): Likewise.
1223 (vreinterpretq_f32_u8): Likewise.
1224 (vreinterpretq_f32_u16): Likewise.
1225 (vreinterpretq_f32_u32): Likewise.
1226 (vreinterpretq_f32_u64): Likewise.
1227 (vreinterpretq_f32_p8): Likewise.
1228 (vreinterpretq_f32_p16): Likewise.
1229 (vreinterpret_s64_s8): Likewise.
1230 (vreinterpret_s64_s16): Likewise.
1231 (vreinterpret_s64_s32): Likewise.
1232 (vreinterpret_s64_f32): Likewise.
1233 (vreinterpret_s64_u8): Likewise.
1234 (vreinterpret_s64_u16): Likewise.
1235 (vreinterpret_s64_u32): Likewise.
1236 (vreinterpret_s64_u64): Likewise.
1237 (vreinterpret_s64_p8): Likewise.
1238 (vreinterpret_s64_p16): Likewise.
1239 (vreinterpretq_s64_s8): Likewise.
1240 (vreinterpretq_s64_s16): Likewise.
1241 (vreinterpretq_s64_s32): Likewise.
1242 (vreinterpretq_s64_f32): Likewise.
1243 (vreinterpretq_s64_u8): Likewise.
1244 (vreinterpretq_s64_u16): Likewise.
1245 (vreinterpretq_s64_u32): Likewise.
1246 (vreinterpretq_s64_u64): Likewise.
1247 (vreinterpretq_s64_p8): Likewise.
1248 (vreinterpretq_s64_p16): Likewise.
1249 (vreinterpret_u64_s8): Likewise.
1250 (vreinterpret_u64_s16): Likewise.
1251 (vreinterpret_u64_s32): Likewise.
1252 (vreinterpret_u64_s64): Likewise.
1253 (vreinterpret_u64_f32): Likewise.
1254 (vreinterpret_u64_u8): Likewise.
1255 (vreinterpret_u64_u16): Likewise.
1256 (vreinterpret_u64_u32): Likewise.
1257 (vreinterpret_u64_p8): Likewise.
1258 (vreinterpret_u64_p16): Likewise.
1259 (vreinterpretq_u64_s8): Likewise.
1260 (vreinterpretq_u64_s16): Likewise.
1261 (vreinterpretq_u64_s32): Likewise.
1262 (vreinterpretq_u64_s64): Likewise.
1263 (vreinterpretq_u64_f32): Likewise.
1264 (vreinterpretq_u64_u8): Likewise.
1265 (vreinterpretq_u64_u16): Likewise.
1266 (vreinterpretq_u64_u32): Likewise.
1267 (vreinterpretq_u64_p8): Likewise.
1268 (vreinterpretq_u64_p16): Likewise.
1269 (vreinterpret_s8_s16): Likewise.
1270 (vreinterpret_s8_s32): Likewise.
1271 (vreinterpret_s8_s64): Likewise.
1272 (vreinterpret_s8_f32): Likewise.
1273 (vreinterpret_s8_u8): Likewise.
1274 (vreinterpret_s8_u16): Likewise.
1275 (vreinterpret_s8_u32): Likewise.
1276 (vreinterpret_s8_u64): Likewise.
1277 (vreinterpret_s8_p8): Likewise.
1278 (vreinterpret_s8_p16): Likewise.
1279 (vreinterpretq_s8_s16): Likewise.
1280 (vreinterpretq_s8_s32): Likewise.
1281 (vreinterpretq_s8_s64): Likewise.
1282 (vreinterpretq_s8_f32): Likewise.
1283 (vreinterpretq_s8_u8): Likewise.
1284 (vreinterpretq_s8_u16): Likewise.
1285 (vreinterpretq_s8_u32): Likewise.
1286 (vreinterpretq_s8_u64): Likewise.
1287 (vreinterpretq_s8_p8): Likewise.
1288 (vreinterpretq_s8_p16): Likewise.
1289 (vreinterpret_s16_s8): Likewise.
1290 (vreinterpret_s16_s32): Likewise.
1291 (vreinterpret_s16_s64): Likewise.
1292 (vreinterpret_s16_f32): Likewise.
1293 (vreinterpret_s16_u8): Likewise.
1294 (vreinterpret_s16_u16): Likewise.
1295 (vreinterpret_s16_u32): Likewise.
1296 (vreinterpret_s16_u64): Likewise.
1297 (vreinterpret_s16_p8): Likewise.
1298 (vreinterpret_s16_p16): Likewise.
1299 (vreinterpretq_s16_s8): Likewise.
1300 (vreinterpretq_s16_s32): Likewise.
1301 (vreinterpretq_s16_s64): Likewise.
1302 (vreinterpretq_s16_f32): Likewise.
1303 (vreinterpretq_s16_u8): Likewise.
1304 (vreinterpretq_s16_u16): Likewise.
1305 (vreinterpretq_s16_u32): Likewise.
1306 (vreinterpretq_s16_u64): Likewise.
1307 (vreinterpretq_s16_p8): Likewise.
1308 (vreinterpretq_s16_p16): Likewise.
1309 (vreinterpret_s32_s8): Likewise.
1310 (vreinterpret_s32_s16): Likewise.
1311 (vreinterpret_s32_s64): Likewise.
1312 (vreinterpret_s32_f32): Likewise.
1313 (vreinterpret_s32_u8): Likewise.
1314 (vreinterpret_s32_u16): Likewise.
1315 (vreinterpret_s32_u32): Likewise.
1316 (vreinterpret_s32_u64): Likewise.
1317 (vreinterpret_s32_p8): Likewise.
1318 (vreinterpret_s32_p16): Likewise.
1319 (vreinterpretq_s32_s8): Likewise.
1320 (vreinterpretq_s32_s16): Likewise.
1321 (vreinterpretq_s32_s64): Likewise.
1322 (vreinterpretq_s32_f32): Likewise.
1323 (vreinterpretq_s32_u8): Likewise.
1324 (vreinterpretq_s32_u16): Likewise.
1325 (vreinterpretq_s32_u32): Likewise.
1326 (vreinterpretq_s32_u64): Likewise.
1327 (vreinterpretq_s32_p8): Likewise.
1328 (vreinterpretq_s32_p16): Likewise.
1329 (vreinterpret_u8_s8): Likewise.
1330 (vreinterpret_u8_s16): Likewise.
1331 (vreinterpret_u8_s32): Likewise.
1332 (vreinterpret_u8_s64): Likewise.
1333 (vreinterpret_u8_f32): Likewise.
1334 (vreinterpret_u8_u16): Likewise.
1335 (vreinterpret_u8_u32): Likewise.
1336 (vreinterpret_u8_u64): Likewise.
1337 (vreinterpret_u8_p8): Likewise.
1338 (vreinterpret_u8_p16): Likewise.
1339 (vreinterpretq_u8_s8): Likewise.
1340 (vreinterpretq_u8_s16): Likewise.
1341 (vreinterpretq_u8_s32): Likewise.
1342 (vreinterpretq_u8_s64): Likewise.
1343 (vreinterpretq_u8_f32): Likewise.
1344 (vreinterpretq_u8_u16): Likewise.
1345 (vreinterpretq_u8_u32): Likewise.
1346 (vreinterpretq_u8_u64): Likewise.
1347 (vreinterpretq_u8_p8): Likewise.
1348 (vreinterpretq_u8_p16): Likewise.
1349 (vreinterpret_u16_s8): Likewise.
1350 (vreinterpret_u16_s16): Likewise.
1351 (vreinterpret_u16_s32): Likewise.
1352 (vreinterpret_u16_s64): Likewise.
1353 (vreinterpret_u16_f32): Likewise.
1354 (vreinterpret_u16_u8): Likewise.
1355 (vreinterpret_u16_u32): Likewise.
1356 (vreinterpret_u16_u64): Likewise.
1357 (vreinterpret_u16_p8): Likewise.
1358 (vreinterpret_u16_p16): Likewise.
1359 (vreinterpretq_u16_s8): Likewise.
1360 (vreinterpretq_u16_s16): Likewise.
1361 (vreinterpretq_u16_s32): Likewise.
1362 (vreinterpretq_u16_s64): Likewise.
1363 (vreinterpretq_u16_f32): Likewise.
1364 (vreinterpretq_u16_u8): Likewise.
1365 (vreinterpretq_u16_u32): Likewise.
1366 (vreinterpretq_u16_u64): Likewise.
1367 (vreinterpretq_u16_p8): Likewise.
1368 (vreinterpretq_u16_p16): Likewise.
1369 (vreinterpret_u32_s8): Likewise.
1370 (vreinterpret_u32_s16): Likewise.
1371 (vreinterpret_u32_s32): Likewise.
1372 (vreinterpret_u32_s64): Likewise.
1373 (vreinterpret_u32_f32): Likewise.
1374 (vreinterpret_u32_u8): Likewise.
1375 (vreinterpret_u32_u16): Likewise.
1376 (vreinterpret_u32_u64): Likewise.
1377 (vreinterpret_u32_p8): Likewise.
1378 (vreinterpret_u32_p16): Likewise.
1379 (vreinterpretq_u32_s8): Likewise.
1380 (vreinterpretq_u32_s16): Likewise.
1381 (vreinterpretq_u32_s32): Likewise.
1382 (vreinterpretq_u32_s64): Likewise.
1383 (vreinterpretq_u32_f32): Likewise.
1384 (vreinterpretq_u32_u8): Likewise.
1385 (vreinterpretq_u32_u16): Likewise.
1386 (vreinterpretq_u32_u64): Likewise.
1387 (vreinterpretq_u32_p8): Likewise.
1388 (vreinterpretq_u32_p16): Likewise.
1390 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1392 * gcc/config/aarch64/aarch64-simd.md (aarch64_s<optab><mode>):
1394 * config/aarch64/aarch64-simd-builtins.def (sqneg): Iterator
1397 * config/aarch64/arm_neon.h (vqneg_s64): New intrinsic.
1398 (vqnegd_s64): Likewise.
1399 (vqabs_s64): Likewise.
1400 (vqabsd_s64): Likewise.
1402 2014-04-22 Richard Henderson <rth@redhat.com>
1404 * config/sparc/sparc.c (sparc_init_modes): Hoist GET_MODE_SIZE
1405 computation to the top of the loop.
1407 2014-04-22 Renlin <renlin.li@arm.com>
1408 Jiong Wang <jiong.wang@arm.com>
1410 * config/aarch64/aarch64.h (aarch64_frame): Delete "fp_lr_offset".
1411 * config/aarch64/aarch64.c (aarch64_layout_frame)
1412 (aarch64_initial_elimination_offset): Likewise.
1414 2014-04-22 Marcus Shawcroft <marcus.shawcroft@arm.com>
1416 * config/aarch64/aarch64.c (aarch64_initial_elimination_offset):
1419 2014-04-22 Richard Sandiford <rdsandiford@googlemail.com>
1421 * machmode.h (bitwise_mode_for_mode): Declare.
1422 * stor-layout.h (bitwise_type_for_mode): Likewise.
1423 * stor-layout.c (bitwise_mode_for_mode): New function.
1424 (bitwise_type_for_mode): Likewise.
1425 * builtins.c (fold_builtin_memory_op): Use it instead of
1426 int_mode_for_mode and build_nonstandard_integer_type.
1428 2014-04-22 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1430 * config.gcc (enable_obsolete): Remove *-*-solaris2.9*.
1431 (*-*-solaris2.[0-9] | *-*-solaris2.[0-9].*): Mark unsupported.
1432 (*-*-solaris2*): Simplify.
1433 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Likewise.
1434 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Remove
1435 *-*-solaris2.9* handling.
1437 * configure.ac (gcc_cv_as_hidden): Remove test for Solaris 9/x86
1439 (gcc_cv_ld_hidden): Remove *-*-solaris2.9* handling.
1440 (ld_tls_support): Remove i?86-*-solaris2.9, sparc*-*-solaris2.9
1442 (gcc_cv_as_gstabs_flag): Remove workaround for Solaris 9/x86 as bug.
1443 * configure: Regenerate.
1445 * config/i386/sol2-9.h: Remove.
1447 * doc/install.texi (Specific, i?86-*-solaris2.9): Remove.
1448 (Specific, *-*-solaris2*): Mention Solaris 9 support removal.
1449 Remove Solaris 9 references.
1451 2014-04-22 Vidya Praveen <vidyapraveen@arm.com>
1453 * aarch64.md (float<GPI:mode><GPF:mode>2): Remove.
1454 (floatuns<GPI:mode><GPF:mode>2): Remove.
1455 (<optab><fcvt_target><GPF:mode>2): New pattern for equal width float
1456 and floatuns conversions.
1457 (<optab><fcvt_iesize><GPF:mode>2): New pattern for inequal width float
1458 and floatuns conversions.
1459 * iterators.md (fcvt_target, FCVT_TARGET): Support SF and DF modes.
1460 (w1,w2): New mode attributes for inequal width conversions.
1462 2014-04-22 Renlin Li <Renlin.Li@arm.com>
1464 * config/aarch64/aarch64.c (aarch64_print_operand_address): Adjust
1465 the output asm format.
1467 2014-04-22 James Greenhalgh <james.greenhalgh@arm.com>
1469 * config/aarch64/aarch64-simd.md
1470 (aarch64_cm<optab>di): Always split.
1471 (*aarch64_cm<optab>di): New.
1472 (aarch64_cmtstdi): Always split.
1473 (*aarch64_cmtstdi): New.
1475 2014-04-22 Jakub Jelinek <jakub@redhat.com>
1477 PR tree-optimization/60823
1478 * omp-low.c (ipa_simd_modify_function_body): Go through
1479 all SSA_NAMEs and for those refering to vector arguments
1480 which are going to be replaced adjust SSA_NAME_VAR and,
1481 if it is a default definition, change it into a non-default
1482 definition assigned at the beginning of function from new_decl.
1483 (ipa_simd_modify_stmt_ops): Rewritten.
1484 * tree-dfa.c (set_ssa_default_def): When removing default def,
1485 check for NULL loc instead of NULL *loc.
1487 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1489 * config/arm/arm.c (arm_hard_regno_mode_ok): Loosen
1490 restrictions on core registers for DImode values in Thumb2.
1492 2014-04-22 Ian Bolton <ian.bolton@arm.com>
1494 * config/arm/arm.md (*anddi_notdi_zesidi): New pattern.
1495 * config/arm/thumb2.md (*iordi_notdi_zesidi): New pattern.
1497 2014-04-22 Ian Bolton <ian.bolton@arm.com>
1499 * config/arm/thumb2.md (*iordi_notdi_di): New pattern.
1500 (*iordi_notzesidi_di): Likewise.
1501 (*iordi_notsesidi_di): Likewise.
1503 2014-04-22 Ian Bolton <ian.bolton@arm.com>
1505 * config/arm/arm-protos.h (tune_params): New struct members.
1506 * config/arm/arm.c: Initialise tune_params per processor.
1507 (thumb2_reorg): Suppress conversion from t32 to t16 when optimizing
1508 for speed, based on new tune_params.
1510 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1512 * config/aarch64/aarch64-builtins.c (BUILTIN_VDQF_DF): Macro
1514 * config/aarch64/aarch64-simd-builtins.def (frintn): Use added
1516 * config/aarch64/aarch64-simd.md (<frint_pattern>): Comment
1518 * config/aarch64/aarch64.md (<frint_pattern>): Likewise.
1519 * config/aarch64/arm_neon.h (vrnd_f64): Added.
1520 (vrnda_f64): Likewise.
1521 (vrndi_f64): Likewise.
1522 (vrndm_f64): Likewise.
1523 (vrndn_f64): Likewise.
1524 (vrndp_f64): Likewise.
1525 (vrndx_f64): Likewise.
1527 2014-04-22 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1529 * config/arm/arm.c (arm_print_operand, thumb_exit): Make sure
1530 GET_MODE_SIZE argument is enum machine_mode.
1532 2014-04-22 Jakub Jelinek <jakub@redhat.com>
1535 * config/sparc/sparc.c (sparc_init_modes): Pass enum machine_mode
1536 value instead of int to GET_MODE_CLASS and GET_MODE_SIZE macros.
1538 2014-04-22 Lin Zuojian <manjian2006@gmail.com>
1541 * asan.c (asan_emit_stack_protection): Force the base to align to
1542 appropriate bits if STRICT_ALIGNMENT. Set shadow_mem align to
1543 appropriate bits if STRICT_ALIGNMENT.
1544 * cfgexpand.c (expand_stack_vars): Set base_align appropriately
1546 (expand_used_vars): Leave a space in the stack frame for alignment
1547 if STRICT_ALIGNMENT.
1549 2014-04-21 David Malcolm <dmalcolm@redhat.com>
1551 * gimple.h (gimple_assign_single_p): Accept a const_gimple rather
1553 (gimple_store_p): Likewise.
1554 (gimple_assign_load_p): Likewise.
1555 (gimple_assign_cast_p): Likewise.
1556 (gimple_clobber_p): Likewise.
1558 * doc/gimple.texi (gimple_assign_cast_p): Accept a const_gimple
1559 rather than a gimple.
1560 (gimple_assign_cast_p): Likewise.
1562 2014-04-21 Michael Meissner <meissner@linux.vnet.ibm.com>
1565 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64 case):
1566 If mode is DDmode and TARGET_E500_DOUBLE allow move.
1568 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print some
1569 more debug information for E500 if -mdebug=reg.
1571 2014-04-21 Uros Bizjak <ubizjak@gmail.com>
1574 * config/i386/i386.c (ix86_expand_builtin)
1575 <case IX86_BUILTIN_RDRAND{16,32,64}_STEP>: Use temporary
1576 register for target RTX.
1577 <case IX86_BUILTIN_RDSEED{16,32,64}_STEP>: Ditto.
1579 2014-04-18 Cong Hou <congh@google.com>
1581 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): Enhance
1582 the widen-mult pattern by handling two operands with different sizes,
1583 and operands whose size is smaller than half of the result type.
1585 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
1587 * ipa-inline.h (INLINE_HINT_known_hot): New hint.
1588 * ipa-inline-analysis.c (dump_inline_hints): Dump it.
1589 (do_estimate_edge_time): Compute it.
1590 * ipa-inline.c (want_inline_small_function_p): Bypass
1591 INLINE_INSNS_AUTO/SINGLE limits for calls that are known to be hot.
1593 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
1595 * ipa-inline.c (spec_rem): New static variable.
1596 (dump_overall_stats): New function.
1597 (dump_inline_stats): New function.
1599 2014-04-18 Richard Henderson <rth@redhat.com>
1601 * config/aarch64/aarch64.c (aarch64_register_move_cost): Pass a mode
1602 to GET_MODE_SIZE, not a reg_class_t.
1604 2014-04-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1606 * config/rs6000/vsx.md (vsx_xxmrghw_<mode>): Adjust for little-endian.
1607 (vsx_xxmrglw_<mode>): Likewise.
1609 2014-04-17 Michael Meissner <meissner@linux.vnet.ibm.com>
1612 * config/rs6000/rs6000.c (rs6000_setup_reg_addr_masks): Make sure
1613 GET_MODE_SIZE gets passed an enum machine_mode type and not integer.
1614 (rs6000_init_hard_regno_mode_ok): Likewise.
1616 2014-04-17 Jan Hubicka <hubicka@ucw.cz>
1618 * ipa-inline.c (inline_small_functions): Account only non-cold
1620 * doc/invoke.texi (inline-unit-growth): Update documentation.
1622 2014-04-17 Pat Haugen <pthaugen@us.ibm.com>
1624 * config/rs6000/rs6000.md (addti3, subti3): New.
1626 2014-04-17 H.J. Lu <hongjiu.lu@intel.com>
1629 * config/i386/i386.c (ix86_expand_clear): Remove outdated
1630 comment. Check optimize_insn_for_size_p instead of
1631 optimize_insn_for_speed_p.
1633 2014-04-17 Martin Jambor <mjambor@suse.cz>
1635 * gimple-iterator.c (gsi_start_edge): New function.
1636 * gimple-iterator.h (gsi_start_edge): Declare.
1637 * tree-sra.c (single_non_eh_succ): New function.
1638 (disqualify_ops_if_throwing_stmt): Renamed to
1639 disqualify_if_bad_bb_terminating_stmt. Allow throwing statements
1640 having one non-EH successor BB.
1641 (sra_modify_expr): If stmt ends bb, use single non-EH successor to
1642 generate loads into replacements.
1643 (sra_modify_assign): Likewise and and also use the simple path for
1645 (sra_modify_function_body): Commit statements on edges.
1647 2014-04-17 Richard Biener <rguenther@suse.de>
1650 * tree-ssa-propagate.c (valid_gimple_rhs_p): Allow vector
1651 comparison results and add clarifying comment.
1653 2014-04-17 Jakub Jelinek <jakub@redhat.com>
1655 * genmodes.c (struct mode_data): Add need_bytesize_adj field.
1656 (blank_mode): Initialize it.
1657 (emit_mode_size_inline, emit_mode_nunits_inline,
1658 emit_mode_inner_inline): New functions.
1659 (emit_insn_modes_h): Call them and surround their output with
1660 #if GCC_VERSION >= 4001 ... #endif.
1661 * machmode.h (GET_MODE_SIZE, GET_MODE_NUNITS, GET_MODE_INNER):
1662 For GCC_VERSION >= 4001 use mode_*_inline routines instead of
1663 mode_* arrays if the argument is __builtin_constant_p.
1664 * lower-subreg.c (dump_choices): Make sure GET_MODE_SIZE argument
1665 is enum machine_mode.
1667 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1669 * passes.c (opt_pass::execute): Adjust.
1670 (pass_manager::execute_pass_mode_switching): Likewise.
1671 (early_local_passes::execute): Likewise.
1672 (execute_one_pass): Pass cfun to the pass's execute method.
1673 * tree-pass.h (opt_pass::execute): Add function * argument.
1674 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
1675 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
1676 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
1677 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
1678 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
1679 config/sparc/sparc.c, cprop.c, dce.c, df-core.c, dse.c, dwarf2cfi.c,
1680 except.c, final.c, function.c, fwprop.c, gcse.c, gimple-low.c,
1681 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
1682 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
1683 ipa-inline-analysis.c, ipa-inline.c, ipa-profile.c, ipa-pure-const.c,
1684 ipa-reference.c, ipa-split.c, ipa.c, ira.c, jump.c, loop-init.c,
1685 lower-subreg.c, mode-switching.c, omp-low.c, postreload-gcse.c,
1686 postreload.c, predict.c, recog.c, ree.c, reg-stack.c, regcprop.c,
1687 reginfo.c, regrename.c, reorg.c, sched-rgn.c, stack-ptr-mod.c,
1688 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
1689 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
1690 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
1691 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-ssa-ccp.c,
1692 tree-ssa-copy.c, tree-ssa-copyrename.c, tree-ssa-dce.c,
1693 tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
1694 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
1695 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
1696 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
1697 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
1698 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
1699 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
1700 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
1701 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
1702 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
1705 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1707 * passes.c (opt_pass::gate): Take function * argument.
1708 (gate_all_early_local_passes): Merge into
1709 (early_local_passes::gate): this.
1710 (gate_all_early_optimizations): Merge into
1711 (all_early_optimizations::gate): this.
1712 (gate_all_optimizations): Mege into
1713 (all_optimizations::gate): this.
1714 (gate_all_optimizations_g): Merge into
1715 (all_optimizations_g::gate): this.
1716 (gate_rest_of_compilation): Mege into
1717 (rest_of_compilation::gate): this.
1718 (gate_postreload): Merge into
1719 (postreload::gate): this.
1720 (dump_one_pass): Pass cfun to the pass's gate method.
1721 (execute_ipa_summary_passes): Likewise.
1722 (execute_one_pass): Likewise.
1723 (ipa_write_summaries_2): Likewise.
1724 (ipa_write_optimization_summaries_1): Likewise.
1725 (ipa_read_summaries_1): Likewise.
1726 (ipa_read_optimization_summaries_1): Likewise.
1727 (execute_ipa_stmt_fixups): Likewise.
1728 * tree-pass.h (opt_pass::gate): Add function * argument.
1729 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c,
1730 combine-stack-adj.c, combine.c, compare-elim.c,
1731 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
1732 config/rl78/rl78.c, config/sh/sh_optimize_sett_clrt.cc,
1733 config/sh/sh_treg_combine.cc, config/sparc/sparc.c, cprop.c, cse.c,
1734 dce.c, df-core.c, dse.c, dwarf2cfi.c, except.c, fwprop.c, gcse.c,
1735 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
1736 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
1737 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
1738 loop-init.c, lower-subreg.c, mode-switching.c, modulo-sched.c,
1739 omp-low.c, postreload-gcse.c, postreload.c, predict.c, recog.c, ree.c,
1740 reg-stack.c, regcprop.c, regrename.c, reorg.c, sched-rgn.c,
1741 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
1742 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
1743 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c,
1744 tree-nrv.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
1745 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
1746 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
1747 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
1748 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
1749 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
1750 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
1751 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
1752 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
1753 tree-ssa.c, tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
1754 tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c, tsan.c, ubsan.c,
1755 var-tracking.c, vtable-verify.c, web.c: Adjust.
1757 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1759 * configure.ac: Check for -Woverloaded-virtual and enable it if found.
1760 * configure: Regenerate.
1762 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1764 * passes.c (dump_one_pass): don't check pass->has_gate.
1765 (execute_ipa_summary_passes): Likewise.
1766 (execute_one_pass): Likewise.
1767 (ipa_write_summaries_2): Likewise.
1768 (ipa_write_optimization_summaries_1): Likewise.
1769 (ipa_read_optimization_summaries_1): Likewise.
1770 (execute_ipa_stmt_fixups): Likewise.
1771 * tree-pass.h (pass_data::has_gate): Remove.
1772 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
1773 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
1774 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
1775 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
1776 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
1777 config/sh/sh_optimize_sett_clrt.cc, config/sh/sh_treg_combine.cc,
1778 config/sparc/sparc.c, cprop.c, cse.c, dce.c, df-core.c, dse.c,
1779 dwarf2cfi.c, except.c, final.c, function.c, fwprop.c, gcse.c,
1780 gimple-low.c, gimple-ssa-isolate-paths.c,
1781 gimple-ssa-strength-reduction.c, graphite.c, ifcvt.c, init-regs.c,
1782 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
1783 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
1784 ira.c, jump.c, loop-init.c, lower-subreg.c, mode-switching.c,
1785 modulo-sched.c, omp-low.c, postreload-gcse.c, postreload.c, predict.c,
1786 recog.c, ree.c, reg-stack.c, regcprop.c, reginfo.c, regrename.c,
1787 reorg.c, sched-rgn.c, stack-ptr-mod.c, store-motion.c, tracer.c,
1788 trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
1789 tree-complex.c, tree-eh.c, tree-emutls.c, tree-if-conv.c,
1790 tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
1791 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
1792 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
1793 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
1794 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
1795 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
1796 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
1797 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
1798 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
1799 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
1800 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
1801 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
1802 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
1805 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1807 * pass_manager.h (pass_manager::register_dump_files_1): Remove
1809 * passes.c (pass_manager::register_dump_files_1): Merge into
1810 (pass_manager::register_dump_files): this, and remove its handling of
1811 properties since the pass always has the properties anyway.
1812 (pass_manager::pass_manager): Adjust.
1814 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1816 * pass_manager.h (pass_manager::register_dump_files_1): Adjust.
1817 * passes.c (pass_manager::register_dump_files_1): Remove dead code
1818 dealing with properties.
1819 (pass_manager::register_dump_files): Adjust.
1821 2014-03-20 Mark Wielaard <mjw@redhat.com>
1823 * dwarf2out.c (add_bound_info): If HOST_WIDE_INT is big enough,
1824 then represent the bound as normal constant value.
1826 2014-04-17 Jakub Jelinek <jakub@redhat.com>
1829 Forward port from 4.8 branch
1830 2013-07-19 Kirill Yukhin <kirill.yukhin@intel.com>
1832 * config/i386/bmiintrin.h (_blsi_u32): New.
1836 (_blsmsk_u32): Ditto.
1837 (_blsmsk_u64): Ditto.
1838 (_tzcnt_u32): Ditto.
1839 (_tzcnt_u64): Ditto.
1841 2014-04-17 Kito Cheng <kito@0xlab.org>
1843 * gcc.c (used_arg): Prevent out of bound access for multilib_options.
1845 2014-04-17 Richard Biener <rguenther@suse.de>
1848 * tree-ssa-propagate.c (valid_gimple_rhs_p): Only allow effective
1849 boolean results for comparisons.
1851 2014-04-17 Richard Biener <rguenther@suse.de>
1853 PR tree-optimization/60836
1854 * tree-vect-loop.c (vect_create_epilog_for_reduction): Force
1855 initial PHI args to be gimple values.
1857 2014-04-17 Richard Biener <rguenther@suse.de>
1859 PR tree-optimization/60841
1860 * tree-vect-data-refs.c (vect_analyze_data_refs): Count stmts.
1861 * tree-vect-loop.c (vect_analyze_loop_2): Pass down number
1862 of stmts to SLP build.
1863 * tree-vect-slp.c (vect_slp_analyze_bb_1): Likewise.
1864 (vect_analyze_slp): Likewise.
1865 (vect_analyze_slp_instance): Likewise.
1866 (vect_build_slp_tree): Limit overall SLP tree growth.
1867 * tree-vectorizer.h (vect_analyze_data_refs,
1868 vect_analyze_slp): Adjust prototypes.
1870 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
1872 * config/i386/i386.c (x86_add_stmt_cost): Fix vector cost model for
1875 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
1877 * config/i386/x86-tune.def (TARGET_SLOW_PSHUFB): New tune definition.
1878 * config/i386/i386.h (TARGET_SLOW_PSHUFB): New tune flag.
1879 * config/i386/i386.c (expand_vec_perm_even_odd_1): Avoid byte shuffles
1880 for TARGET_SLOW_PSHUFB
1882 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
1884 * config/i386/i386.c (slm_cost): Adjust vec_to_scalar_cost.
1885 * config/i386/i386.c (intel_cost): Ditto.
1887 2014-04-17 Joey Ye <joey.ye@arm.com>
1889 * opts.c (OPT_fif_conversion, OPT_fif_conversion2): Disable for Og.
1891 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
1893 * opts.c (common_handle_option): Disable -fipa-reference coorectly
1896 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
1898 * ipa-devirt.c (odr_type_d): Add field all_derivations_known.
1899 (type_all_derivations_known_p): New predicate.
1900 (type_all_ctors_visible_p): New predicate.
1901 (type_possibly_instantiated_p): New predicate.
1902 (get_odr_type): Compute all_derivations_known.
1903 (dump_odr_type): Dump the flag.
1904 (maybe_record_type): Cleanup.
1905 (record_target_from_binfo): Add bases_to_consider array;
1906 record bases for types w/o instances and skip CXX destructor.
1907 (possible_polymorphic_call_targets_1): Add bases_to_consider
1908 and consider_construction parameters; check if type may have instance.
1909 (get_polymorphic_call_info): Set maybe_in_construction to true
1910 when we know nothing.
1911 (record_targets_from_bases): Skip CXX destructors; they are
1912 never called for types in construction.
1913 (possible_polymorphic_call_targets): Do not record target when
1914 type may not have instance.
1916 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
1919 * ipa.c (symtab_remove_unreachable_nodes): Mark targets of
1920 external aliases alive, too.
1922 2014-04-16 Andrew Pinski <apinski@cavium.com>
1924 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Change aarch64 ilp32
1927 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
1929 * final.c (compute_alignments): Do not apply loop alignment to a block
1930 falling through to the exit.
1932 2014-04-16 Catherine Moore <clm@codesourcery.com>
1934 * mips.md (*mov<mode>_internal, *movhi_internal, *movqi_internal):
1935 Adjust constraints for microMIPS store patterns.
1937 2014-04-16 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
1939 * config/avr/avr-mcus.def: Correct typo for atxmega256a3bu macro.
1941 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
1943 * tree-ssa-operands.c (create_vop_var): Set DECL_IGNORED_P.
1944 (append_use): Run at -O0.
1945 (append_vdef): Likewise.
1946 * tree-ssa-ter.c (ter_is_replaceable_p): Do not special-case -O0.
1947 * tree-ssa-uninit.c (warn_uninitialized_vars): Remove obsolete comment.
1949 2014-04-16 Jakub Jelinek <jakub@redhat.com>
1951 PR tree-optimization/60844
1952 * tree-ssa-reassoc.c (reassoc_remove_stmt): New function.
1953 (propagate_op_to_single_use, remove_visited_stmt_chain,
1954 linearize_expr, repropagate_negates, reassociate_bb): Use it
1955 instead of gsi_remove.
1957 2014-04-16 Martin Jambor <mjambor@suse.cz>
1959 * cgraphclones.c (cgraph_create_virtual_clone): Duplicate
1960 ipa_transforms_to_apply.
1961 (cgraph_function_versioning): Assert that old_node has empty
1962 ipa_transforms_to_apply.
1963 * trans-mem.c (ipa_tm_create_version): Likewise.
1964 * tree-inline.c (tree_function_versioning): Do not duplicate
1965 ipa_transforms_to_apply.
1967 2014-04-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1970 * configure.ac (set_have_as_tls): Merge i[34567]86-*-* and
1972 Pass necessary as flags on 64-bit Solaris/x86.
1973 Use lowercase relocs for x86_64-*-*.
1974 * configure: Regenerate.
1976 2014-04-15 Jan Hubicka <jh@suse.cz>
1978 * ipa-devirt.c (referenced_from_vtable_p): New predicate.
1979 (maybe_record_node, likely_target_p): Use it.
1981 2014-04-15 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1984 Revert following patch
1986 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
1989 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
1990 software floating point or no floating point registers, do not
1991 allow any type in the FPRs. Eliminate a test for SPE SIMD types
1992 in GPRs that occurs after we tested for GPRs that would never be
1995 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
1996 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
1997 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
1998 specifically allow DDmode, since that does not use the SPE SIMD
2001 2014-03-21 Mark Wielaard <mjw@redhat.com>
2003 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_const_value
2004 as unsigned or int depending on type and value used.
2006 2014-04-15 Richard Biener <rguenther@suse.de>
2008 PR rtl-optimization/56965
2009 * alias.c (ncr_compar, nonoverlapping_component_refs_p): Move ...
2010 * tree-ssa-alias.c (ncr_compar, nonoverlapping_component_refs_p):
2012 * alias.c (true_dependence_1): Do not call
2013 nonoverlapping_component_refs_p.
2014 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Call
2015 nonoverlapping_component_refs_p.
2016 (indirect_refs_may_alias_p): Likewise.
2018 2014-04-15 Teresa Johnson <tejohnson@google.com>
2020 * cfg.c (dump_bb_info): Fix flags check.
2021 * tree-cfg.c (remove_bb): Only dump TDF_BLOCKS when removing.
2023 2014-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2025 PR rtl-optimization/60663
2026 * config/arm/arm.c (arm_new_rtx_costs): Improve ASM_OPERANDS case,
2029 2014-04-15 Richard Biener <rguenther@suse.de>
2031 * lto-streamer.h (LTO_major_version): Bump to 4.
2033 2014-04-15 Richard Biener <rguenther@suse.de>
2035 * common.opt (lto_partition_model): New enum.
2036 (flto-partition=): Merge separate options with a single with argument,
2037 add -flto-partition=one support.
2038 * flag-types.h (enum lto_partition_model): Declare.
2039 * opts.c (finish_options): Remove duplicate -flto-partition=
2041 * lto-wrapper.c (run_gcc): Adjust.
2043 2014-04-15 Richard Biener <rguenther@suse.de>
2045 * alias.c (ncr_compar): New function.
2046 (nonoverlapping_component_refs_p): Re-implement in O (n log n).
2048 2014-04-15 Richard Biener <rguenther@suse.de>
2050 * alias.c (record_component_aliases): Do not walk BINFOs.
2052 2014-04-15 Richard Biener <rguenther@suse.de>
2054 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
2055 Add struct function argument and adjust.
2056 (find_func_aliases_for_call): Likewise.
2057 (find_func_aliases): Likewise.
2058 (find_func_clobbers): Likewise.
2059 (intra_create_variable_infos): Likewise.
2060 (compute_points_to_sets): Likewise.
2061 (ipa_pta_execute): Adjust. Do not push/pop cfun.
2063 2014-04-15 Richard Biener <rguenther@suse.de>
2065 * tree.c (iterative_hash_expr): Use enum tree_code_class
2066 to store TREE_CODE_CLASS.
2067 (tree_block): Likewise.
2068 (tree_set_block): Likewise.
2069 * tree.h (fold_build_pointer_plus_loc): Use
2070 convert_to_ptrofftype_loc.
2072 2014-04-15 Jakub Jelinek <jakub@redhat.com>
2075 * Makefile.in (PLUGIN_HEADERS): Add various headers that have been
2078 2014-04-15 Eric Botcazou <ebotcazou@adacore.com>
2080 * cfgloop.h (struct loop): Move force_vectorize down.
2081 * gimplify.c (gimple_boolify) <ANNOTATE_EXPR>: Handle new kinds.
2082 (gimplify_expr) <ANNOTATE_EXPR>: Minor tweak.
2083 * lto-streamer-in.c (input_cfg): Read dont_vectorize field.
2084 * lto-streamer-out.c (output_cfg): Write dont_vectorize field.
2085 * tree-cfg.c (replace_loop_annotate): Revamp and handle new kinds.
2086 * tree-core.h (enum annot_expr_kind): Add new kind values.
2087 * tree-inline.c (copy_loops): Copy dont_vectorize field and reorder.
2088 * tree-pretty-print.c (dump_generic_node) <ANNOTATE_EXPR>: Handle new
2090 * tree.def (ANNOTATE_EXPR): Tweak comment.
2092 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
2094 * ipa-devirt.c (maybe_record_node): Ignore all non-methods (including
2097 2014-04-14 Paolo Carlini <paolo.carlini@oracle.com>
2099 * tree.h (TYPE_IDENTIFIER): Declare.
2100 * tree.c (subrange_type_for_debug_p): Use it.
2101 * godump.c (go_format_type): Likewise.
2102 * dwarf2out.c (is_cxx_auto, modified_type_die,
2103 gen_type_die_with_usage, gen_type_die_with_usage): Likewise.
2104 * dbxout.c (dbxout_type, dbxout_symbol): Likewise.
2106 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
2109 * varpool.c (varpool_remove_node): Do not alter decls when streaming.
2111 2014-04-14 Uros Bizjak <ubizjak@gmail.com>
2113 * config/i386/i386.c (examine_argument): Return bool. Return true if
2114 parameter should be passed in memory.
2115 <case X86_64_COMPLEX_X87_CLASS>: Adjust.
2116 (construct_container): Update calls to examine_argument.
2117 (function_arg_advance_64): Ditto.
2118 (return_in_memory_32): Merge with ix86_return_in_memory.
2119 (return_in_memory_64): Ditto.
2120 (return_in_memory_ms_64): Ditto.
2122 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
2124 * ipa-utils.c (ipa_merge_profiles): Merge profile_id.
2125 * coverage.c (coverage_compute_profile_id): Handle externally visible
2128 2014-04-14 Martin Jambor <mjambor@suse.cz>
2130 * tree-sra.c (ipa_sra_preliminary_function_checks): Skip
2131 DECL_DISREGARD_INLINE_LIMITS functions.
2133 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
2136 * config/i386/i386.md (*fixuns_trunc<mode>_1): Revert the last change.
2138 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
2141 * config/i386/i386.md (*fixuns_trunc<mode>_1): Check
2142 optimize_insn_for_speed_p instead of
2143 optimize_function_for_speed_p.
2145 2014-04-14 Yufeng Zhang <yufeng.zhang@arm.com>
2147 * doc/invoke.texi (free): Document AArch64.
2149 2014-04-14 Richard Biener <rguenther@suse.de>
2151 PR tree-optimization/60042
2152 * tree-ssa-pre.c (inhibit_phi_insertion): Remove.
2153 (insert_into_preds_of_block): Do not prevent PHI insertion
2154 for REFERENCE exprs here ...
2155 (eliminate_dom_walker::before_dom_children): ... but prevent
2156 their use here under similar conditions when applied to the
2157 IL after PRE optimizations.
2159 2014-04-14 Richard Biener <rguenther@suse.de>
2161 * passes.def: Move early points-to after early SRA.
2163 2014-04-14 Richard Biener <rguenther@suse.de>
2165 * tree-ssa-forwprop.c (simplify_gimple_switch): Enhance
2166 check for which sign-changes we allow when forwarding
2167 a converted value into a switch.
2169 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
2171 * stor-layout.c (place_field): Finalize non-constant offset for the
2174 2014-04-14 Richard Biener <rguenther@suse.de>
2176 * tree-switch-conversion.c (lshift_cheap_p): Get speed_p
2178 (expand_switch_using_bit_tests_p): Likewise.
2179 (process_switch): Compute and pass on speed_p based on the
2181 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Use
2182 optimize_bb_for_speed_p.
2184 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
2186 * cfgloop.h (struct loop): Rename force_vect into force_vectorize.
2187 * function.h (struct function): Rename has_force_vect_loops into
2188 has_force_vectorize_loops.
2189 * lto-streamer-in.c (input_cfg): Adjust for renaming.
2190 (input_struct_function_base): Likewise.
2191 * lto-streamer-out.c (output_cfg): Likewise.
2192 (output_struct_function_base): Likewise.
2193 * omp-low.c (expand_omp_simd): Likewise.
2194 * tree-cfg.c (move_sese_region_to_fn): Likewise.
2195 * tree-if-conv.c (ifcvt_can_use_mask_load_store): Likewise.
2196 (version_loop_for_if_conversion): Likewise.
2197 (tree_if_conversion): Likewise.
2198 (main_tree_if_conversion): Likewise.
2199 (gate_tree_if_conversion): Likewise.
2200 * tree-inline.c (copy_loops): Likewise.
2201 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Likewise.
2202 * tree-ssa-loop.c (tree_loop_vectorize): Likewise.
2203 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Likewise.
2204 * tree-vect-loop.c (vect_estimate_min_profitable_iters): Likewise.
2205 * tree-vectorizer.c (vectorize_loops): Likewise.
2206 * tree-vectorizer.h (unlimited_cost_model): Likewise.
2208 2014-04-14 Richard Biener <rguenther@suse.de>
2211 * lto-streamer-out.c (wrap_refs): New function.
2212 (lto_output): Wrap symbol references in global initializes in
2213 type-preserving MEM_REFs.
2215 2014-04-14 Christian Bruel <christian.bruel@st.com>
2217 * config/sh/sh-mem.cc (sh_expand_strlen): Unroll last word.
2219 2014-04-14 Christian Bruel <christian.bruel@st.com>
2221 * config/sh/sh.md (setmemqi): New expand pattern.
2222 * config/sh/sh.h (CLEAR_RATIO): Define.
2223 * config/sh/sh-mem.cc (sh_expand_setmem): Define.
2224 * config/sh/sh-protos.h (sh_expand_setmem): Declare.
2226 2014-04-14 Richard Biener <rguenther@suse.de>
2229 * fold-const.c (negate_expr_p): Don't negate directional rounding
2231 (fold_negate_expr): Likewise.
2233 2014-04-14 Richard Biener <rguenther@suse.de>
2235 PR tree-optimization/59817
2236 PR tree-optimization/60453
2237 * graphite-scop-detection.c (graphite_can_represent_scev): Complete
2238 recursion to catch all CHRECs in the scalar evolution and restrict
2239 the predicate for the remains appropriately.
2241 2014-04-12 Catherine Moore <clm@codesourcery.com>
2243 * config/mips/constraints.md: Add new register constraint "kb".
2244 * config/mips/mips.md (*mov<mode>_internal): Use constraint "kb".
2245 (*movhi_internal): Likewise.
2246 (*movqi_internal): Likewise.
2247 * config/mips/mips.h (M16_STORE_REGS): New register class.
2248 (REG_CLASS_NAMES): Add M16_STORE_REGS.
2249 (REG_CLASS_CONTENTS): Likewise.
2250 * config/mips/mips.c (mips_regno_to_class): Add M16_STORE_REGS.
2252 2014-04-11 Tobias Burnus <burnus@net-b.de>
2255 * doc/invoke.texi (-Wformat-signedness): Document it.
2256 (Wformat=2): Mention that this enables -Wformat-signedness.
2258 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
2260 * common/config/epiphany/epiphany-common.c
2261 (epiphany_option_optimization_table): Enable section anchors by
2262 default at -O1 or higher.
2263 * config/epiphany/epiphany.c (TARGET_MAX_ANCHOR_OFFSET): Define.
2264 (TARGET_MIN_ANCHOR_OFFSET): Likewise.
2265 (epiphany_rtx_costs) <SET>: For binary operators, the set as such
2266 carries no extra cost.
2267 (epiphany_legitimate_address_p): For BLKmode, apply SImode check.
2268 * config/epiphany/epiphany.h (ASM_OUTPUT_DEF): Define.
2269 * config/epiphany/predicates.md (memclob_operand): New predicate.
2270 * config/epiphany/epiphany.md (stack_adjust_add, stack_adjust_str):
2271 Use memclob_operand predicate and X constraint for operand 3.
2273 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
2275 * config/epiphany/epiphany.c (epiphany_rtx_cost): Compare
2276 with CC_N_NE / CC_C_LTU / CC_C_GTU carries no extra cost for
2279 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
2281 PR rtl-optimization/60651
2282 * mode-switching.c (optimize_mode_switching): Make sure to emit
2283 sets of a lower numbered entity before sets of a higher numbered
2284 entity to a mode of the same or lower priority.
2285 When creating a seginfo for a basic block that starts with a code
2286 label, move the insertion point past the code label.
2287 (new_seginfo): Document and enforce requirement that
2288 NOTE_INSN_BASIC_BLOCK only appears for empty blocks.
2289 * doc/tm.texi.in: Document ordering constraint for emitted mode sets.
2290 * doc/tm.texi: Regenerate.
2292 2014-01-11 Joern Rennecke <joern.rennecke@embecosm.com>
2295 * config/arc/arc.c (arc_save_restore): Fix assert typo.
2297 2013-04-11 Jakub Jelinek <jakub@redhat.com>
2299 * BASE-VER: Set to 4.10.0.
2301 2014-04-11 Tobias Burnus <burnus@net-b.de>
2304 * doc/bugreport.texi (Bugs): Remove nodes pointing to the nirvana.
2305 * doc/gcc.texi (Service): Update description in the @menu
2306 * doc/invoke.texi (Option Summary): Remove misplaced and
2309 2014-04-11 Steve Ellcey <sellcey@mips.com>
2310 Jakub Jelinek <jakub@redhat.com>
2313 * expr.c (convert_move): Use emit_store_flag_force instead of
2314 emit_store_flag. Pass lowpart_mode instead of VOIDmode as 5th
2317 2014-04-11 Richard Biener <rguenther@suse.de>
2320 * varasm.c (assemble_alias): Avoid endless error reporting
2321 recursion by setting TREE_ASM_WRITTEN.
2323 2014-04-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2325 * config/s390/s390.md: Add a splitter for NOT rtx.
2327 2014-04-11 Jakub Jelinek <jakub@redhat.com>
2329 PR rtl-optimization/60663
2330 * cse.c (cse_insn): Set src_volatile on ASM_OPERANDS in PARALLEL.
2332 2014-04-10 Jan Hubicka <hubicka@ucw.cz>
2333 Jakub Jelinek <jakub@redhat.com>
2336 * ipa.c (function_and_variable_visibility): Copy forced_by_abi
2337 flag from decl_node to node.
2339 2014-04-10 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2342 * config/arm/arm.c (TARGET_CONST_NOT_OK_FOR_DEBUG_P): Define
2343 (arm_const_not_ok_for_debug_p): Reject MINUS with SYM_REF's
2344 ameliorating the cases where it can be.
2346 2014-04-09 David Edelsohn <dje.gcc@gmail.com>
2349 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
2351 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
2352 (loadsync_<mode>): Change mode.
2353 (load_quadpti, store_quadpti): New.
2354 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
2355 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
2356 * config/rs6000/predicates.md (quad_memory_operand): !TARGET_SYNC_TI.
2358 2014-04-09 Cong Hou <congh@google.com>
2361 * doc/sourcebuild.texi (vect_widen_mult_si_to_di_pattern): Add
2364 2014-04-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2366 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Use vnand
2367 instead of vnor to exploit possible fusion opportunity in the
2369 (altivec_expand_vec_perm_const_le): Likewise.
2371 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
2373 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
2374 (loadsync_<mode>): Change mode.
2375 (load_quadpti, store_quadpti): New.
2376 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
2377 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
2379 2014-04-08 Richard Sandiford <rdsandiford@googlemail.com>
2382 * config/rs6000/vsx.md (vsx_xscvdpspn_scalar): Change input to DImode.
2383 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Update accordingly.
2384 Use gen_rtx_REG rather than simplify_gen_subreg for op0_di.
2386 2014-04-08 Richard Biener <rguenther@suse.de>
2389 * tree-pretty-print.c (pp_double_int): For HWI32 hosts with
2390 a 64bit widest int print double-int similar to on HWI64 hosts.
2392 2014-04-08 Richard Biener <rguenther@suse.de>
2394 PR tree-optimization/60785
2395 * graphite-sese-to-poly.c (rewrite_phi_out_of_ssa): Treat
2396 default defs properly.
2398 2014-04-08 Nathan Sidwell <nathan@codesourcery.com>
2400 * doc/invoke (Wnon-virtual-dtor): Update to match implementation.
2401 (Weffc++): Likewise.
2403 2014-04-07 Jan Hubicka <hubcika@ucw.cz>
2405 * ipa-devirt.c (maybe_record_node): When node is not recorded,
2406 set completep to false rather than true.
2408 2014-04-07 Douglas B Rupp <rupp@adacore.com>
2411 * config/arm/arm.h (ASM_PREFERRED_EH_DATA_FORMAT): Expose from
2412 ARM_TARGET2_DWARF_FORMAT.
2414 2014-04-07 Charles Baylis <charles.baylis@linaro.org>
2417 * config/arm/arm.h (ASM_OUTPUT_CASE_END): Remove.
2418 (LABEL_ALIGN_AFTER_BARRIER): Align barriers which occur after
2421 2014-04-07 Richard Biener <rguenther@suse.de>
2423 PR tree-optimization/60766
2424 * tree-ssa-loop-ivopts.c (cand_value_at): Compute in an unsigned type.
2425 (may_eliminate_iv): Convert cand_value_at result to desired type.
2427 2014-04-07 Jason Merrill <jason@redhat.com>
2430 * common.opt (-fno-gnu-unique): Add.
2431 * config/elfos.h (USE_GNU_UNIQUE_OBJECT): Check it.
2433 2014-04-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2435 * haifa-sched.c: Fix outdated function reference and minor
2436 grammar errors in introductory comment.
2438 2014-04-07 Richard Biener <rguenther@suse.de>
2441 * tree-ssa-operands.c (maybe_add_call_vops): Also add VDEFs
2443 * tree-cfgcleanup.c (fixup_noreturn_call): Do not remove VDEFs.
2445 2014-04-06 John David Anglin <danglin@gcc.gnu.org>
2448 * config/pa/pa.c (pa_output_function_epilogue): Skip address and code
2449 size accounting for thunks.
2450 (pa_asm_output_mi_thunk): Use final_start_function() and
2451 final_end_function() to output function start and end directives.
2453 2014-04-05 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
2455 * config/avr/avr-arch.h (avr_mcu_t): Add dev_attribute field to have
2456 device specific ISA/ feature information. Remove short_sp and
2457 errata_skip ds. Add avr_device_specific_features enum to have device
2459 * config/avr/avr-c.c (avr_cpu_cpp_builtins): use dev_attribute to check
2460 errata_skip. Add __AVR_ISA_RMW__ builtin macro if RMW ISA available.
2461 * config/avr/avr-devices.c (avr_mcu_types): Update AVR_MCU macro for
2462 updated device specific info.
2463 * config/avr/avr-mcus.def: Merge device specific details to
2464 dev_attribute field.
2465 * config/avr/avr.c (avr_2word_insn_p): use dev_attribute field to check
2467 * config/avr/avr.h (AVR_HAVE_8BIT_SP): same for short sp info.
2468 * config/avr/driver-avr.c (avr_device_to_as): Pass -mrmw option to
2469 assembler if RMW isa supported by current device.
2470 * config/avr/genmultilib.awk: Update as device info structure changed.
2471 * doc/invoke.texi: Add info for __AVR_ISA_RMW__ builtin macro
2473 2014-04-04 Cong Hou <congh@google.com>
2475 PR tree-optimization/60656
2476 * tree-vect-stmts.c (supportable_widening_operation):
2477 Fix a bug that elements in a vector with vect_used_by_reduction
2478 property are incorrectly reordered when the operation on it is not
2479 consistant with the one in reduction operation.
2481 2014-04-04 John David Anglin <danglin@gcc.gnu.org>
2483 PR rtl-optimization/60155
2484 * gcse.c (record_set_data): New function.
2485 (single_set_gcse): New function.
2486 (gcse_emit_move_after): Use single_set_gcse instead of single_set.
2487 (hoist_code): Likewise.
2488 (get_pressure_class_and_nregs): Likewise.
2490 2014-04-04 Eric Botcazou <ebotcazou@adacore.com>
2492 * explow.c (probe_stack_range): Emit a final optimization blockage.
2494 2014-04-04 Anthony Green <green@moxielogic.com>
2496 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2): Fix
2499 2014-04-04 Jan Hubicka <hubicka@ucw.cz>
2502 * lto-cgraph.c (input_overwrite_node): Check that partitioning
2503 flags are set only during streaming.
2504 * ipa.c (process_references, walk_polymorphic_call_targets,
2505 symtab_remove_unreachable_nodes): Drop bodies of always inline
2506 after early inlining.
2507 (symtab_remove_unreachable_nodes): Remove always_inline attribute.
2509 2014-04-04 Jakub Jelinek <jakub@redhat.com>
2510 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2513 * dwarf2out.c (const_ok_for_output_1): Reject expressions
2516 2014-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2519 * config/arm/cortex-a53.md (cortex_a53_fdivs): Reduce reservation
2521 (cortex_a53_fdivd): Likewise.
2523 2014-04-04 Martin Jambor <mjambor@suse.cz>
2526 * cgraph.h (cgraph_clone_node): New parameter added to declaration.
2528 * cgraph.c (clone_of_p): Also return true if thunks match.
2529 (verify_edge_corresponds_to_fndecl): Removed extraneous call to
2530 cgraph_function_or_thunk_node and an obsolete comment.
2531 * cgraphclones.c (build_function_type_skip_args): Moved upwards in the
2533 (build_function_decl_skip_args): Likewise.
2534 (set_new_clone_decl_and_node_flags): New function.
2535 (duplicate_thunk_for_node): Likewise.
2536 (redirect_edge_duplicating_thunks): Likewise.
2537 (cgraph_clone_node): New parameter args_to_skip, pass it to
2538 redirect_edge_duplicating_thunks which is called instead of
2539 cgraph_redirect_edge_callee.
2540 (cgraph_create_virtual_clone): Pass args_to_skip to cgraph_clone_node,
2541 moved setting of a lot of flags to set_new_clone_decl_and_node_flags.
2543 2014-04-04 Jeff Law <law@redhat.com>
2546 * config/arm/predicates.md (const_int_I_operand): New predicate.
2547 (const_int_M_operand): Similarly.
2548 * config/arm/arm.md (insv_zero): Use const_int_M_operand instead of
2550 (insv_t2, extv_reg, extzv_t2): Likewise.
2551 (load_multiple_with_writeback): Similarly for const_int_I_operand.
2552 (pop_multiple_with_writeback_and_return): Likewise.
2553 (vfp_pop_multiple_with_writeback): Likewise
2555 2014-04-04 Richard Biener <rguenther@suse.de>
2558 * tree-ssanames.c (make_ssa_name_fn): Fix assert.
2559 * gimple.c (gimple_set_bb): Avoid ICEing for NULL cfun for
2561 * gimplify.h (gimple_add_tmp_var_fn): Declare.
2562 * gimplify.c (gimple_add_tmp_var_fn): New function.
2563 * gimple-expr.h (create_tmp_reg_fn): Declare.
2564 * gimple-expr.c (create_tmp_reg_fn): New function.
2565 * gimple-low.c (record_vars_into): Don't change cfun.
2566 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix
2567 code generation without cfun.
2569 2014-04-04 Thomas Schwinge <thomas@codesourcery.com>
2572 * Makefile.in (install-driver): Fix shell scripting.
2574 2014-04-03 Cong Hou <congh@google.com>
2576 PR tree-optimization/60505
2577 * tree-vectorizer.h (struct _stmt_vec_info): Add th field as the
2578 threshold of number of iterations below which no vectorization
2580 * tree-vect-loop.c (new_loop_vec_info):
2581 Initialize LOOP_VINFO_COST_MODEL_THRESHOLD.
2582 * tree-vect-loop.c (vect_analyze_loop_operations):
2583 Set LOOP_VINFO_COST_MODEL_THRESHOLD.
2584 * tree-vect-loop.c (vect_transform_loop):
2585 Use LOOP_VINFO_COST_MODEL_THRESHOLD.
2586 * tree-vect-loop.c (vect_analyze_loop_2): Check the maximum number
2587 of iterations of the loop and see if we should build the epilogue.
2589 2014-04-03 Richard Biener <rguenther@suse.de>
2591 * tree-streamer.h (struct streamer_tree_cache_d): Add next_idx member.
2592 (streamer_tree_cache_create): Adjust.
2593 * tree-streamer.c (streamer_tree_cache_add_to_node_array): Adjust
2594 to allow optional nodes array.
2595 (streamer_tree_cache_insert_1): Use next_idx to assign idx.
2596 (streamer_tree_cache_append): Likewise.
2597 (streamer_tree_cache_create): Create nodes array optionally
2598 as specified by parameter.
2599 * lto-streamer-out.c (create_output_block): Avoid maintaining
2600 the node array in the writer cache.
2601 (DFS_write_tree): Remove assertion.
2602 (produce_asm_for_decls): Free the out decl state hash table early.
2603 * lto-streamer-in.c (lto_data_in_create): Adjust for
2604 streamer_tree_cache_create prototype change.
2606 2014-04-03 Richard Biener <rguenther@suse.de>
2608 * tree-streamer-out.c (streamer_write_chain): Do not temporarily
2609 set TREE_CHAIN to NULL_TREE.
2611 2014-04-03 Richard Biener <rguenther@suse.de>
2613 PR tree-optimization/60740
2614 * graphite-scop-detection.c (stmt_simple_for_scop_p): Iterate
2615 over all GIMPLE_COND operands.
2617 2014-04-03 Nathan Sidwell <nathan@codesourcery.com>
2619 * doc/invoke.texi (Wnon-virtual-dtor): Adjust documentation.
2620 (Weffc++): Remove Scott's numbering, merge lists and reference
2623 2014-04-03 Nick Clifton <nickc@redhat.com>
2625 * config/rl78/rl78-expand.md (movqi): Handle (SUBREG (SYMBOL_REF))
2628 2014-04-03 Martin Jambor <mjambor@suse.cz>
2630 * ipa-cp.c (ipcp_verify_propagated_values): Also dump symtab and
2631 mention gcc_unreachable before failing.
2632 * ipa.c (symtab_remove_unreachable_nodes): Also print order of
2635 2014-04-02 Jan Hubicka <hubicka@ucw.cz>
2638 * ipa-devirt.c (get_polymorphic_call_info): Do not ICE on type
2639 inconsistent code and instead mark the context inconsistent.
2640 (possible_polymorphic_call_targets): For inconsistent contexts
2641 return empty complete list.
2643 2014-04-02 Anthony Green <green@moxielogic.com>
2645 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2)
2646 (extendqisi2, extendhisi2): Define.
2647 * config/moxie/moxie.h (DEFAULT_SIGNED_CHAR): Change to 0.
2648 (WCHAR_TYPE): Change to unsigned int.
2650 2014-04-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2652 PR tree-optimization/60733
2653 * gimple-ssa-strength-reduction.c (ncd_with_phi): Change required
2654 insertion point for PHI candidates to be the end of the feeding
2655 block for the PHI argument.
2657 2014-04-02 Vladimir Makarov <vmakarov@redhat.com>
2659 PR rtl-optimization/60650
2660 * lra-constraints.c (process_alt_operands): Decrease reject for
2661 earlyclobber matching.
2663 2014-04-02 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2665 * config/s390/s390.c (s390_expand_insv): Use GET_MODE_BITSIZE.
2667 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
2669 * config/spu/spu.c (pad_bb): Do not crash when the last
2670 insn is CODE_FOR_blockage.
2672 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
2674 * config/spu/spu.md ("insv"): Fail if bitoffset+bitsize
2675 lies outside the target mode.
2677 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
2680 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
2681 software floating point or no floating point registers, do not
2682 allow any type in the FPRs. Eliminate a test for SPE SIMD types
2683 in GPRs that occurs after we tested for GPRs that would never be
2686 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
2687 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
2688 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
2689 specifically allow DDmode, since that does not use the SPE SIMD
2692 2014-04-02 Richard Biener <rguenther@suse.de>
2695 * optabs.c (expand_abs_nojump): Honor flag_trapv only for
2696 MODE_INTs. Properly use negv_optab.
2697 (expand_abs): Likewise.
2699 2014-04-02 Richard Biener <rguenther@suse.de>
2702 * Makefile.in (install-driver): Guard extra installs with special
2705 2014-04-01 Michael Meissner <meissner@linux.vnet.ibm.com>
2707 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
2710 2014-04-01 Richard Henderson <rth@redhat.com>
2713 * config/i386/i386.md (*float<SWI48><MODEF>2_sse): Leave the second
2714 alternative enabled before register allocation.
2716 2014-04-01 Chung-Lin Tang <cltang@codesourcery.com>
2718 * config/nios2/nios2.md (unspec): Remove UNSPEC_TLS, UNSPEC_TLS_LDM.
2719 * config/nios2/nios2.c (nios2_function_profiler): Fix addi operand
2721 (nios2_large_got_address): Remove unneeded 'sym' parameter.
2722 (nios2_got_address): Update nios2_large_got_address call site.
2723 (nios2_delegitimize_address): New function.
2724 (TARGET_DELEGITIMIZE_ADDRESS): Define to nios2_delegitimize_address.
2725 * config/nios2/linux.h (GLIBC_DYNAMIC_LINKER): Define.
2726 (LINK_SPEC): Specify dynamic linker using GNU_USER_DYNAMIC_LINKER.
2728 2014-04-01 Martin Husemann <martin@duskware.de>
2730 * config/mips/netbsd.h (TARGET_OS_CPP_BUILTINS): Define __mips_o32
2733 2014-04-01 Richard Sandiford <rdsandiford@googlemail.com>
2735 PR rtl-optimization/60604
2736 * recog.c (general_operand): Incorporate REG_CANNOT_CHANGE_MODE_P
2737 check from register_operand.
2738 (register_operand): Redefine in terms of general_operand.
2739 (nonmemory_operand): Use register_operand for the non-constant cases.
2741 2014-04-01 Richard Biener <rguenther@suse.de>
2743 * gimple.h (struct gimple_statement_base): Align subcode to 16 bits.
2745 2014-04-01 Sebastian Huber <sebastian.huber@embedded-brains.de>
2747 * doc/invoke.texi (mapp-regs): Clarify.
2749 2014-03-31 Ulrich Drepper <drepper@gmail.com>
2751 * config/i386/avx512fintrin.h (__v32hi): Define type.
2752 (__v64qi): Likewise.
2753 (_mm512_set1_epi8): Define.
2754 (_mm512_set1_epi16): Define.
2755 (_mm512_set4_epi32): Define.
2756 (_mm512_set4_epi64): Define.
2757 (_mm512_set4_pd): Define.
2758 (_mm512_set4_ps): Define.
2759 (_mm512_setr4_epi64): Define.
2760 (_mm512_setr4_epi32): Define.
2761 (_mm512_setr4_pd): Define.
2762 (_mm512_setr4_ps): Define.
2763 (_mm512_setzero_epi32): Define.
2765 2014-03-31 Martin Jambor <mjambor@suse.cz>
2768 * tree-sra.c (callsite_has_enough_arguments_p): Renamed to
2769 callsite_arguments_match_p. Updated all callers. Also check types of
2770 corresponding formal parameters and actual arguments.
2771 (not_all_callers_have_enough_arguments_p) Renamed to
2772 some_callers_have_mismatched_arguments_p.
2774 2014-03-31 Yuri Rumyantsev <ysrumyan@gmail.com>
2776 * tree-inline.c (copy_loops): Add missed copy of 'safelen'.
2778 2014-03-31 Kugan Vivekanandarajah <kuganv@linaro.org>
2781 * aarch64/aarch64.c (aarch64_classify_address): Fix alignment for
2784 2014-03-30 Uros Bizjak <ubizjak@gmail.com>
2786 * config/i386/sse.md (FMAMODE_NOVF512): New mode iterator.
2787 (<sd_mask_codefor>fma_fmadd_<mode><sd_maskz_name><round_name>):
2789 <sd_mask_codefor>fma_fmadd_<VF_512:mode><sd_maskz_name><round_name>.
2790 Use FMAMODE_NOVF512 mode iterator.
2791 (<sd_mask_codefor>fma_fmsub_<mode><sd_maskz_name><round_name>): Ditto.
2792 (<sd_mask_codefor>fma_fnmadd_<mode><sd_maskz_name><round_name>): Ditto.
2793 (<sd_mask_codefor>fma_fnmsub_<mode><sd_maskz_name><round_name>): Ditto.
2794 (<sd_mask_codefor>fma_fmaddsub_<mode><sd_maskz_name><round_name>):
2796 <sd_mask_codefor>fma_fmaddsub_<VF_512:mode><sd_maskz_name><round_name>.
2797 Use VF_128_256 mode iterator.
2798 (<sd_mask_codefor>fma_fmsubadd_<mode><sd_maskz_name><round_name>):
2801 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
2803 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Clear
2804 static chain if needed.
2806 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
2809 * lra-constraints.c (index_part_to_reg): New.
2810 (process_address): Use it.
2812 2014-03-27 Jeff Law <law@redhat.com>
2813 Jakub Jelinek <jakub@redhat.com>
2816 * expr.c (do_tablejump): Use simplify_gen_binary rather than
2817 gen_rtx_{PLUS,MULT} to build up the address expression.
2819 * i386/i386.c (ix86_legitimize_address): Use copy_addr_to_reg to avoid
2820 creating non-canonical RTL.
2822 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
2825 * ipa-inline.c (want_inline_small_function_p): Short circuit large
2826 functions; reorganize to make cheap checks first.
2827 (inline_small_functions): Do not estimate growth when dumping;
2829 * ipa-inline.h (inline_summary): Add min_size.
2830 (growth_likely_positive): New function.
2831 * ipa-inline-analysis.c (dump_inline_summary): Add min_size.
2832 (set_cond_stmt_execution_predicate): Cleanup.
2833 (estimate_edge_size_and_time): Compute min_size.
2834 (estimate_calls_size_and_time): Likewise.
2835 (estimate_node_size_and_time): Likewise.
2836 (inline_update_overall_summary): Update min_size.
2837 (do_estimate_edge_time): Likewise.
2838 (do_estimate_edge_size): Update.
2839 (do_estimate_edge_hints): Update.
2840 (growth_likely_positive): New function.
2842 2014-03-28 Jakub Jelinek <jakub@redhat.com>
2845 * config/i386/i386.c (ix86_copy_addr_to_reg): Call copy_addr_to_reg
2846 also if addr has VOIDmode.
2848 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2850 * config/arm/aarch-common.c (aarch_crypto_can_dual_issue): New.
2851 * config/arm/aarch-common-protos.h (aarch_crypto_can_dual_issue):
2853 * config/arm/cortex-a53.md: Add reservations and bypass for crypto
2854 instructions as well as AdvancedSIMD loads.
2856 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2858 * config/aarch64/aarch64-simd.md (aarch64_crypto_aes<aes_op>v16qi):
2859 Use crypto_aese type.
2860 (aarch64_crypto_aes<aesmc_op>v16qi): Use crypto_aesmc type.
2861 * config/arm/arm.md (is_neon_type): Replace crypto_aes with
2862 crypto_aese, crypto_aesmc. Move to types.md.
2863 * config/arm/types.md (crypto_aes): Split into crypto_aese,
2865 * config/arm/iterators.md (crypto_type): Likewise.
2867 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
2869 * cgraph.c: Include expr.h and tree-dfa.h.
2870 (cgraph_redirect_edge_call_stmt_to_callee): If call in noreturn;
2873 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
2876 * lra-assigns.c (find_hard_regno_for): Remove unavailable hard
2877 regs from checking multi-reg pseudos.
2879 2014-03-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2881 * config/arm/t-aprofile (MULTILIB_MATCHES): Correct A12 rule.
2883 2014-03-28 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
2885 * config/rs6000/rs6000.c (fusion_gpr_load_p): Refuse optimization
2886 if it would clobber the stack pointer, even temporarily.
2888 2014-03-28 Eric Botcazou <ebotcazou@adacore.com>
2890 * mode-switching.c: Make small adjustments to the top comment.
2892 2014-03-27 Michael Meissner <meissner@linux.vnet.ibm.com>
2894 * config/rs6000/constraints.md (wD constraint): New constraint to
2895 match the constant integer to get the top DImode/DFmode out of a
2896 vector in a VSX register.
2898 * config/rs6000/predicates.md (vsx_scalar_64bit): New predicate to
2899 match the constant integer to get the top DImode/DFmode out of a
2900 vector in a VSX register.
2902 * config/rs6000/rs6000-builtins.def (VBPERMQ): Add vbpermq builtin
2905 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
2908 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
2909 -mdebug=reg, print value of VECTOR_ELEMENT_SCALAR_64BIT.
2911 * config/rs6000/vsx.md (vsx_extract_<mode>, V2DI/V2DF modes):
2912 Optimize vec_extract of 64-bit values, where the value being
2913 extracted is in the top word, where we can use scalar
2914 instructions. Add direct move and store support. Combine the big
2915 endian/little endian vector select load support into a single insn.
2916 (vsx_extract_<mode>_internal1): Likewise.
2917 (vsx_extract_<mode>_internal2): Likewise.
2918 (vsx_extract_<mode>_load): Likewise.
2919 (vsx_extract_<mode>_store): Likewise.
2920 (vsx_extract_<mode>_zero): Delete, big and little endian insns are
2921 combined into vsx_extract_<mode>_load.
2922 (vsx_extract_<mode>_one_le): Likewise.
2924 * config/rs6000/rs6000.h (VECTOR_ELEMENT_SCALAR_64BIT): Macro to
2925 define the top 64-bit vector element.
2927 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wD
2930 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
2931 Document vec_vbpermq builtin.
2934 * config/rs6000/altivec.h (vec_xxsldwi): Add missing define to
2935 enable use of xxsldwi and xxpermdi builtin functions.
2936 (vec_xxpermdi): Likewise.
2938 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
2939 Document use of vec_xxsldwi and vec_xxpermdi builtins.
2941 2014-03-27 Vladimir Makarov <vmakarov@redhat.com>
2943 PR rtl-optimization/60650
2944 * lra-assign.c (find_hard_regno_for, spill_for): Add parameter
2946 (find_spills_for): New.
2947 (assign_by_spills): Pass the new parameter to find_hard_regno_for.
2948 Spill all pseudos on the second iteration.
2950 2014-03-27 Marek Polacek <polacek@redhat.com>
2953 * doc/extend.texi (ffs Builtins): Change unsigned types to signed
2956 2014-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2958 * config/s390/s390.c (s390_can_use_return_insn): Check for
2959 call-saved FPRs on 31 bit.
2961 2014-03-27 Jakub Jelinek <jakub@redhat.com>
2964 * omp-low.c (lower_omp_1): For gimple_clobber_p stmts,
2965 if they need regimplification, just drop them instead of
2966 calling gimple_regimplify_operands on them.
2968 2014-03-27 Marcus Shawcroft <marcus.shawcroft@arm.com>
2971 * config/aarch64/aarch64.c (faked_omit_frame_pointer): Remove.
2972 (aarch64_frame_pointer_required): Adjust logic.
2973 (aarch64_can_eliminate): Adjust logic.
2974 (aarch64_override_options_after_change): Adjust logic.
2976 2014-03-27 Dehao Chen <dehao@google.com>
2978 * ipa-inline.c (early_inliner): Update node's inline info.
2980 2014-03-26 Dehao Chen <dehao@google.com>
2982 * dojump.c (do_compare_rtx_and_jump): Sets correct probability for
2983 compiler inserted conditional jumps for NAN float check.
2985 2014-03-26 Jakub Jelinek <jakub@redhat.com>
2987 * ubsan.h (ubsan_create_data): Change second argument's type
2988 to const location_t *.
2989 * ubsan.c (ubsan_source_location): If xloc.file is NULL, set it to
2991 (ubsan_create_data): Change second argument to const location_t *PLOC.
2992 Create Loc field whenever PLOC is non-NULL.
2993 (ubsan_instrument_unreachable, ubsan_expand_null_ifn,
2994 ubsan_build_overflow_builtin, instrument_bool_enum_load): Adjust
2998 * real.c (real_to_integer2): Change type of low to UHWI.
3000 2014-03-26 Tobias Burnus <burnus@net-b.de>
3002 * gcc.c (LINK_COMMAND_SPEC): Use libcilkrts.spec for -fcilkplus.
3003 (CILK_SELF_SPECS): New define.
3004 (driver_self_specs): Use it.
3006 2014-03-26 Richard Biener <rguenther@suse.de>
3008 * tree-pretty-print.c (percent_K_format): Implement special
3009 case for LTO and its stripped down BLOCK tree.
3011 2014-03-26 Jakub Jelinek <jakub@redhat.com>
3014 * ubsan.c (instrument_si_overflow): Instrument ABS_EXPR.
3016 * tree-vrp.c (simplify_internal_call_using_ranges): If only
3017 one range is range_int_cst_p, but not both, at least optimize
3018 addition/subtraction of 0 and multiplication by 0 or 1.
3019 * gimple-fold.c (gimple_fold_call): Fold
3020 IFN_UBSAN_CHECK_{ADD,SUB,MUL}.
3021 (gimple_fold_stmt_to_constant_1): If both op0 and op1 aren't
3022 INTEGER_CSTs, try to fold at least x * 0 and y - y.
3024 2014-03-26 Eric Botcazou <ebotcazou@adacore.com>
3026 PR rtl-optimization/60452
3027 * rtlanal.c (rtx_addr_can_trap_p_1): Fix head comment.
3028 <case REG>: Return 1 for invalid offsets from the frame pointer.
3030 2014-03-26 Marek Polacek <polacek@redhat.com>
3033 * doc/extend.texi (C Extensions): Mention variable-length arrays in
3036 2014-03-26 Marek Polacek <polacek@redhat.com>
3039 * doc/extend.texi (Designated Inits): Describe what happens to omitted
3042 2014-03-26 Marek Polacek <polacek@redhat.com>
3045 * ira-color.c (update_conflict_hard_regno_costs): Perform the
3046 multiplication in unsigned type.
3048 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
3050 * doc/install.texi: Document nds32le-*-elf and nds32be-*-elf.
3052 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
3054 * doc/contrib.texi: Add myself as Andes nds32 port contributor.
3056 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
3059 * cif-code.def (UNREACHABLE) New code.
3060 * ipa-inline.c (inline_small_functions): Skip edges to
3061 __builtlin_unreachable.
3062 (estimate_edge_growth): Allow edges to __builtlin_unreachable.
3063 * ipa-inline-analysis.c (edge_set_predicate): Redirect edges with false
3064 predicate to __bulitin_unreachable.
3065 (set_cond_stmt_execution_predicate): Fix issue when
3066 invert_tree_comparison returns ERROR_MARK.
3067 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow): Do not
3068 propagate to inline clones.
3069 * cgraph.c (verify_edge_corresponds_to_fndecl): Allow redirection
3071 * ipa-cp.c (create_specialized_node): Be ready for new node to appear.
3072 * cgraphclones.c (cgraph_clone_node): If call destination is already
3073 ureachable, do not redirect it back.
3074 * tree-inline.c (fold_marked_statements): Hanlde calls becoming
3077 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
3079 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow):
3080 Do not modify inline clones.
3082 2014-03-25 Jakub Jelinek <jakub@redhat.com>
3084 * config/i386/i386.md (general_sext_operand): New mode attr.
3085 (addv<mode>4, subv<mode>4, mulv<mode>4): If operands[2] is CONST_INT,
3086 don't generate (sign_extend (const_int)).
3087 (*addv<mode>4, *subv<mode>4, *mulv<mode>4): Disallow CONST_INT_P
3088 operands[2]. Use We constraint instead of <i> and
3089 <general_sext_operand> predicate instead of <general_operand>.
3090 (*addv<mode>4_1, *subv<mode>4_1, *mulv<mode>4_1): New insns.
3091 * config/i386/constraints.md (We): New constraint.
3092 * config/i386/predicates.md (x86_64_sext_operand,
3093 sext_operand): New predicates.
3095 2014-03-25 Martin Jambor <mjambor@suse.cz>
3098 * ipa-cp.c (ipa_get_indirect_edge_target_1): Redirect type
3099 inconsistent devirtualizations to __builtin_unreachable.
3101 2014-03-25 Marek Polacek <polacek@redhat.com>
3104 * doc/extend.texi (Example of asm with clobbered asm reg): Fix typo.
3106 2014-03-25 Alan Lawrence <alan.lawrence@arm.com>
3108 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Reverse
3109 order of elements for big-endian.
3111 2014-03-25 Richard Biener <rguenther@suse.de>
3114 * gimplify-me.c (gimple_regimplify_operands): Update the
3117 2014-03-25 Martin Jambor <mjambor@suse.cz>
3120 * lto-cgraph.c (lto_output_node): Stream body_removed flag.
3121 (lto_output_varpool_node): Likewise.
3122 (input_overwrite_node): Likewise.
3123 (input_varpool_node): Likewise.
3125 2014-03-25 Richard Biener <rguenther@suse.de>
3127 * lto-wrapper.c (merge_and_complain): Handle OPT_fPIE like OPT_fpie.
3128 (run_gcc): Likewise.
3130 2014-03-25 Jakub Jelinek <jakub@redhat.com>
3132 * combine.c (simplify_compare_const): Add MODE argument.
3133 Handle mode_width 0 as very large mode_width.
3134 (try_combine, simplify_comparison): Adjust callers.
3136 * cselib.c (cselib_hash_rtx): Perform addition in unsigned
3137 type to avoid signed integer overflow.
3138 * explow.c (plus_constant): Likewise.
3140 2014-03-25 Dominik Vogt <vogt@linux.vnet.ibm.com>
3142 * doc/generic.texi: Correct typos.
3144 2014-03-24 Tobias Burnus <burnus@net-b.de>
3146 * doc/invoke.texi (-flto): Expand section about
3147 using static libraries with LTO.
3149 2014-03-24 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3151 PR rtl-optimization/60501
3152 * optabs.def (addptr3_optab): New optab.
3153 * optabs.c (gen_addptr3_insn, have_addptr3_insn): New function.
3154 * doc/md.texi ("addptrm3"): Document new RTL standard expander.
3155 * expr.h (gen_addptr3_insn, have_addptr3_insn): Add prototypes.
3157 * lra.c (emit_add3_insn): Use the addptr pattern if available.
3159 * config/s390/s390.md ("addptrdi3", "addptrsi3"): New expanders.
3161 2014-03-24 Ulrich Drepper <drepper@gmail.com>
3163 * config/i386/avx512fintrin.h: Define _mm512_set1_ps and
3166 * config/i386/avxintrin.h (_mm256_undefined_si256): Define.
3167 (_mm256_undefined_ps): Define.
3168 (_mm256_undefined_pd): Define.
3169 * config/i386/emmintrin.h (_mm_undefined_si128): Define.
3170 (_mm_undefined_pd): Define.
3171 * config/i386/xmmintrin.h (_mm_undefined_ps): Define.
3172 * config/i386/avx512fintrin.h (_mm512_undefined_si512): Define.
3173 (_mm512_undefined_ps): Define.
3174 (_mm512_undefined_pd): Define.
3175 Use _mm*_undefined_*.
3176 * config/i386/avx2intrin.h: Use _mm*_undefined_*.
3178 2014-03-24 Alex Velenko <Alex.Velenko@arm.com>
3180 * config/aarch64/aarch64-simd-builtins.def (lshr): DI mode excluded.
3181 (lshr_simd): DI mode added.
3182 * config/aarch64/aarch64-simd.md (aarch64_lshr_simddi): New pattern.
3183 (aarch64_ushr_simddi): Likewise.
3184 * config/aarch64/aarch64.md (UNSPEC_USHR64): New unspec.
3185 * config/aarch64/arm_neon.h (vshr_n_u64): Intrinsic fixed.
3186 (vshrd_n_u64): Likewise.
3188 2014-03-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3190 * Makefile.in (s-macro_list): Depend on cc1.
3192 2014-03-23 Teresa Johnson <tejohnson@google.com>
3194 * ipa-utils.c (ipa_print_order): Use specified dump file.
3196 2014-03-23 Eric Botcazou <ebotcazou@adacore.com>
3198 PR rtl-optimization/60601
3199 * bb-reorder.c (fix_up_fall_thru_edges): Test EDGE_FALLTHRU everywhere.
3201 * gcc.c (eval_spec_function): Initialize save_growing_value.
3203 2014-03-22 Jakub Jelinek <jakub@redhat.com>
3206 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): For
3207 code == MINUS_EXPR, never swap op0 with op1.
3209 * toplev.c (init_local_tick): Avoid signed integer multiplication
3211 * genautomata.c (reserv_sets_hash_value): Fix rotate idiom, avoid
3212 shift by first operand's bitsize.
3214 2014-03-21 Jakub Jelinek <jakub@redhat.com>
3217 * config/i386/i386.h (TARGET_64BIT_P): If not TARGET_BI_ARCH,
3219 * config/i386/darwin.h (TARGET_64BIT_P): Redefine to
3220 TARGET_ISA_64BIT_P(x).
3222 2014-03-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3224 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Generate a
3225 pattern for vector nor instead of subtract from splat(-1).
3226 (altivec_expand_vec_perm_const_le): Likewise.
3228 2014-03-21 Richard Henderson <rth@twiddle.net>
3231 * ifcvt.c (dead_or_predicable): Return FALSE if there are any frame
3232 related insns after epilogue_completed.
3234 2014-03-21 Martin Jambor <mjambor@suse.cz>
3237 * cgraph.h (symtab_node): New flag body_removed.
3238 * ipa.c (symtab_remove_unreachable_nodes): Set body_removed flag
3239 when removing bodies.
3240 * symtab.c (dump_symtab_base): Dump body_removed flag.
3241 * cgraph.c (verify_edge_corresponds_to_fndecl): Skip nodes which
3242 had their bodies removed.
3244 2014-03-21 Martin Jambor <mjambor@suse.cz>
3247 * ipa.c (symtab_remove_unreachable_nodes): Clear thunk flag of nodes
3250 2014-03-21 Richard Biener <rguenther@suse.de>
3252 PR tree-optimization/60577
3253 * tree-core.h (struct tree_base): Document nothrow_flag use
3254 in VAR_DECL_NONALIASED.
3255 * tree.h (VAR_DECL_NONALIASED): New.
3256 (may_be_aliased): Adjust.
3257 * coverage.c (build_var): Set VAR_DECL_NONALIASED.
3259 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
3261 * expr.c (expand_expr_real_1): Remove outdated comment.
3263 2014-03-20 Jakub Jelinek <jakub@redhat.com>
3266 * ira.c (adjust_cleared_regs): Call copy_rtx on
3267 *reg_equiv[REGNO (loc)].src_p before passing it to
3268 simplify_replace_fn_rtx.
3271 * config/i386/i386.c (x86_output_mi_thunk): Surround UNSPEC_GOT
3272 into CONST, put pic register as first operand of PLUS. Use
3273 gen_const_mem for both 32-bit and 64-bit PIC got loads.
3275 2014-03-20 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3277 * config/aarch64/aarch64.c (MEMORY_MOVE_COST): Delete.
3279 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
3281 * config/sparc/sparc.c (sparc_do_work_around_errata): Implement work
3282 around for store forwarding issue in the FPU on the UT699.
3283 * config/sparc/sparc.md (in_branch_delay): Return false for single FP
3284 loads and operations if -mfix-ut699 is specified.
3285 (divtf3_hq): Tweak attribute.
3286 (sqrttf2_hq): Likewise.
3288 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
3290 * calls.c (store_one_arg): Remove incorrect const qualification on the
3291 type of the temporary.
3292 * cfgexpand.c (expand_return): Likewise.
3293 * expr.c (expand_constructor): Likewise.
3294 (expand_expr_real_1): Likewise.
3296 2014-03-20 Zhenqiang Chen <zhenqiang.chen@linaro.org>
3298 * config/arm/arm.c (arm_dwarf_register_span): Update the element number
3301 2014-03-19 Kaz Kojima <kkojima@gcc.gnu.org>
3304 * config/sh/sh.md (udivsi3_i1): Clobber R1 register.
3306 2014-03-19 James Greenhalgh <james.greenhalgh@arm.com>
3308 * config/arm/aarch-common-protos.h
3309 (alu_cost_table): Fix spelling of "extend".
3310 * config/arm/arm.c (arm_new_rtx_costs): Fix spelling of "extend".
3312 2014-03-19 Richard Biener <rguenther@suse.de>
3315 * tree-core.h (tree_type_common): Re-order pointer members
3316 to reduce recursion depth during GC walks.
3318 2014-03-19 Marek Polacek <polacek@redhat.com>
3321 * ubsan.c (ubsan_type_descriptor): Check that DECL_NAME is nonnull
3322 before accessing it.
3324 2014-03-19 Richard Biener <rguenther@suse.de>
3327 * lto-streamer-in.c (input_function): In WPA stage do not drop
3330 2014-03-19 Jakub Jelinek <jakub@redhat.com>
3332 PR tree-optimization/60559
3333 * vectorizable_mask_load_store): Replace scalar MASK_LOAD
3334 with build_zero_cst assignment.
3336 2014-03-18 Kai Tietz <ktietz@redhat.com>
3338 PR rtl-optimization/56356
3339 * sdbout.c (sdbout_parms): Verify that parms'
3340 incoming argument is valid.
3341 (sdbout_reg_parms): Likewise.
3343 2014-03-18 Richard Henderson <rth@redhat.com>
3346 * config/i386/i386.md (*float<SWI48x><MODEF>2_i387): Move down to
3347 be shadowed by *float<SWI48><MODEF>2_sse. Test X87_ENABLE_FLOAT.
3348 (*float<SWI48><MODEF>2_sse): Check X87_ENABLE_FLOAT for alternative 0.
3350 2014-03-18 Basile Starynkevitch <basile@starynkevitch.net>
3352 * plugin.def: Improve comment for PLUGIN_INCLUDE_FILE.
3353 * doc/plugins.texi (Plugin callbacks): Mention PLUGIN_INCLUDE_FILE.
3354 Italicize plugin event names in description. Explain that
3355 PLUGIN_PRAGMAS has no sense for lto1. Explain PLUGIN_INCLUDE_FILE.
3356 Remind that no GCC functions should be called after PLUGIN_FINISH.
3357 Explain what pragmas with expansion are.
3359 2014-03-18 Martin Liska <mliska@suse.cz>
3361 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Added case when
3362 gimple call statement is update.
3363 * gimple-fold.c (gimple_fold_call): Changed order for GIMPLE_ASSIGN and
3364 GIMPLE_CALL, where gsi iterator still points to GIMPLE CALL.
3366 2014-03-18 Jakub Jelinek <jakub@redhat.com>
3369 * ubsan.c (ubsan_instrument_unreachable): Call
3370 initialize_sanitizer_builtins.
3371 (ubsan_pass): Likewise.
3374 * ubsan.c (ubsan_type_descriptor, ubsan_create_data): Call
3375 varpool_finalize_decl instead of rest_of_decl_compilation.
3377 2014-03-18 Richard Biener <rguenther@suse.de>
3379 * df-problems.c (df_rd_confluence_n): Avoid bitmap_copy
3380 by using bitmap_and_compl instead of bitmap_and_compl_into.
3381 (df_rd_transfer_function): Likewise.
3383 2014-03-18 Richard Biener <rguenther@suse.de>
3385 * doc/lto.texi (fresolution): Fix typo.
3387 2014-03-18 Richard Biener <rguenther@suse.de>
3389 * doc/invoke.texi (flto): Update for changes in 4.9.
3391 2014-03-18 Richard Biener <rguenther@suse.de>
3393 * doc/loop.texi: Remove section on the removed lambda framework.
3394 Update loop docs with recent changes in preserving loop structure.
3396 2014-03-18 Richard Biener <rguenther@suse.de>
3398 * doc/lto.texi (-fresolution): Document.
3400 2014-03-18 Richard Biener <rguenther@suse.de>
3402 * doc/contrib.texi: Adjust my name.
3404 2014-03-18 Jakub Jelinek <jakub@redhat.com>
3407 * internal-fn.c: Include diagnostic-core.h.
3408 (expand_BUILTIN_EXPECT): New function.
3409 * gimplify.c (gimplify_call_expr): Use false instead of FALSE.
3410 (gimplify_modify_expr): Gimplify 3 argument __builtin_expect into
3411 IFN_BUILTIN_EXPECT call instead of __builtin_expect builtin call.
3412 * ipa-inline-analysis.c (find_foldable_builtin_expect): Handle
3414 * predict.c (expr_expected_value_1): Handle IFN_BUILTIN_EXPECT.
3415 Revert 3 argument __builtin_expect code.
3416 (strip_predict_hints): Handle IFN_BUILTIN_EXPECT.
3417 * gimple-fold.c (gimple_fold_call): Likewise.
3418 * tree.h (fold_builtin_expect): New prototype.
3419 * builtins.c (build_builtin_expect_predicate): Add predictor
3420 argument, if non-NULL, create 3 argument __builtin_expect.
3421 (fold_builtin_expect): No longer static. Add ARG2 argument,
3422 pass it through to build_builtin_expect_predicate.
3423 (fold_builtin_2): Adjust caller.
3424 (fold_builtin_3): Handle BUILT_IN_EXPECT.
3425 * internal-fn.def (BUILTIN_EXPECT): New.
3427 2014-03-18 Tobias Burnus <burnus@net-b.de>
3430 * predict.def (PRED_FORTRAN_OVERFLOW, PRED_FORTRAN_FAIL_ALLOC,
3431 PRED_FORTRAN_FAIL_IO, PRED_FORTRAN_WARN_ONCE, PRED_FORTRAN_SIZE_ZERO,
3432 PRED_FORTRAN_INVALID_BOUND, PRED_FORTRAN_ABSENT_DUMMY): Add.
3434 2014-03-18 Jan Hubicka <hubicka@ucw.cz>
3437 * predict.c (combine_predictions_for_bb): Fix up formatting.
3438 (expr_expected_value_1, expr_expected_value): Add predictor argument,
3439 fill what it points to if non-NULL.
3440 (tree_predict_by_opcode): Adjust caller, use the predictor.
3441 * predict.def (PRED_COMPARE_AND_SWAP): Add.
3443 2014-03-18 Eric Botcazou <ebotcazou@adacore.com>
3445 * config/sparc/sparc.c (sparc_do_work_around_errata): Speed up and use
3446 proper constant for the store mode.
3448 2014-03-18 Ilya Enkovich <ilya.enkovich@intel.com>
3450 * symtab.c (change_decl_assembler_name): Fix transparent alias
3453 2014-03-16 Renlin Li <Renlin.Li@arm.com>
3455 * config/aarch64/aarch64.c: Correct the comments about the
3456 aarch64 stack layout.
3458 2014-03-18 Thomas Schwinge <thomas@codesourcery.com>
3460 * omp-low.c (lower_rec_input_clauses) <build_omp_barrier>: Restore
3461 check for GF_OMP_FOR_KIND_FOR.
3463 2013-03-18 Kirill Yukhin <kirill.yukhin@intel.com>
3465 * config/i386/i386.h (ADDITIONAL_REGISTER_NAMES): Add
3466 ymm and zmm register names.
3468 2014-03-17 Jakub Jelinek <jakub@redhat.com>
3471 * config/i386/i386.c (ix86_expand_epilogue): Adjust REG_CFA_ADJUST_CFA
3472 note creation for the 2010-08-31 changes.
3474 2014-03-17 Marek Polacek <polacek@redhat.com>
3477 * omp-low.c (omp_max_vf): Treat -fno-tree-loop-optimize the same
3478 as -fno-tree-loop-vectorize.
3479 (expand_omp_simd): Likewise.
3481 2014-03-15 Eric Botcazou <ebotcazou@adacore.com>
3483 * config/sparc/sparc-protos.h (tls_call_delay): Delete.
3484 (eligible_for_call_delay): New prototype.
3485 * config/sparc/sparc.c (tls_call_delay): Rename into...
3486 (eligible_for_call_delay): ...this. Return false if the instruction
3487 cannot be put in the delay slot of a branch.
3488 (eligible_for_restore_insn): Simplify.
3489 (eligible_for_return_delay): Return false if the instruction cannot be
3490 put in the delay slot of a branch and simplify.
3491 (eligible_for_sibcall_delay): Return false if the instruction cannot be
3492 put in the delay slot of a branch.
3493 * config/sparc/sparc.md (fix_ut699): New attribute.
3494 (tls_call_delay): Delete.
3495 (in_call_delay): Reimplement.
3496 (eligible_for_sibcall_delay): Rename into...
3497 (in_sibcall_delay): ...this.
3498 (eligible_for_return_delay): Rename into...
3499 (in_return_delay): ...this.
3500 (in_branch_delay): Reimplement.
3501 (in_uncond_branch_delay): Delete.
3502 (in_annul_branch_delay): Delete.
3504 2014-03-14 Richard Henderson <rth@redhat.com>
3507 * config/i386/i386.md (floathi<X87MODEF>2): Delete expander; rename
3508 define_insn from *floathi<X87MODEF>2_i387; allow nonimmediate_operand.
3509 (*floathi<X87MODEF>2_i387_with_temp): Remove.
3510 (floathi splitters): Remove.
3511 (float<SWI48x>xf2): New pattern.
3512 (float<SWI48><MODEF>2): Rename from float<SWI48x><X87MODEF>2. Drop
3513 code that tried to handle DImode for 32-bit, but which was excluded
3514 by the pattern's condition. Drop allocation of stack temporary.
3515 (*floatsi<MODEF>2_vector_mixed_with_temp): Remove.
3516 (*float<SWI48><MODEF>2_mixed_with_temp): Remove.
3517 (*float<SWI48><MODEF>2_mixed_interunit): Remove.
3518 (*float<SWI48><MODEF>2_mixed_nointerunit): Remove.
3519 (*floatsi<MODEF>2_vector_sse_with_temp): Remove.
3520 (*float<SWI48><MODEF>2_sse_with_temp): Remove.
3521 (*float<SWI48><MODEF>2_sse_interunit): Remove.
3522 (*float<SWI48><MODEF>2_sse_nointerunit): Remove.
3523 (*float<SWI48x><X87MODEF>2_i387_with_temp): Remove.
3524 (*float<SWI48x><X87MODEF>2_i387): Remove.
3525 (all float _with_temp splitters): Remove.
3526 (*float<SWI48x><MODEF>2_i387): New pattern.
3527 (*float<SWI48><MODEF>2_sse): New pattern.
3528 (float TARGET_USE_VECTOR_CONVERTS splitters): Merge them.
3529 (float TARGET_SSE_PARTIAL_REG_DEPENDENCY splitters): Merge them.
3531 2014-03-14 Jakub Jelinek <jakub@redhat.com>
3532 Marek Polacek <polacek@redhat.com>
3535 * common.opt (dump_base_name_prefixed): New Variable.
3536 * opts.c (finish_options): Don't prepend directory to x_dump_base_name
3537 if x_dump_base_name_prefixed is already set, set it at the end.
3539 2014-03-14 Vladimir Makarov <vmakarov@redhat.com>
3541 PR rtl-optimization/60508
3542 * lra-constraints.c (get_reload_reg): Add new parameter
3544 (process_addr_reg, simplify_operand_subreg, curr_insn_transform):
3545 Pass the new parameter values.
3547 2014-03-14 Richard Biener <rguenther@suse.de>
3549 * common.opt: Revert unintented changes from r205065.
3552 2014-03-14 Richard Biener <rguenther@suse.de>
3555 * cfghooks.c (split_block): Properly adjust all loops the
3556 block was a latch of.
3558 2014-03-14 Martin Jambor <mjambor@suse.cz>
3561 * ipa-prop.c (ipa_modify_call_arguments): Fix iteration condition
3564 2014-03-14 Georg-Johann Lay <avr@gjlay.de>
3567 * config/avr/avr.c (avr_set_current_function): Pass function name
3568 through default_strip_name_encoding before sanity checking instead
3569 of skipping the first char of the assembler name.
3571 2014-03-13 Richard Henderson <rth@redhat.com>
3574 * config/i386/i386.c (ix86_split_fp_branch): Remove pushed argument.
3575 (ix86_force_to_memory, ix86_free_from_memory): Remove.
3576 * config/i386/i386-protos.h: Likewise.
3577 * config/i386/i386.md (floathi<X87MODEF>2): Use assign_386_stack_local
3578 in the expander instead of a splitter.
3579 (float<SWI48x><X87MODEF>2): Use assign_386_stack_local if there is
3580 any possibility of requiring a memory.
3581 (*floatsi<MODEF>2_vector_mixed): Remove, and the splitters.
3582 (*floatsi<MODEF>2_vector_sse): Remove, and the splitters.
3583 (fp branch splitters): Update for ix86_split_fp_branch.
3584 (*jcc<X87MODEF>_<SWI24>_i387): Remove r/f alternative.
3585 (*jcc<X87MODEF>_<SWI24>_r_i387): Likewise.
3586 (splitter for jcc<X87MODEF>_<SWI24>_i387 r/f): Remove.
3587 (*fop_<MODEF>_2_i387): Remove f/r alternative.
3588 (*fop_<MODEF>_3_i387): Likewise.
3589 (*fop_xf_2_i387, *fop_xf_3_i387): Likewise.
3590 (splitters for the fop_* register patterns): Remove.
3591 (fscalexf4_i387): Rename from *fscalexf4_i387.
3592 (ldexpxf3): Use gen_floatsixf2 and gen_fscalexf4_i387.
3594 2014-03-13 Jakub Jelinek <jakub@redhat.com>
3596 PR tree-optimization/59779
3597 * tree-dfa.c (get_ref_base_and_extent): Use double_int
3598 type for bitsize and maxsize instead of HOST_WIDE_INT.
3600 2014-03-13 Steven Bosscher <steven@gcc.gnu.org>
3602 PR rtl-optimization/57320
3603 * function.c (rest_of_handle_thread_prologue_and_epilogue): Cleanup
3604 the CFG after thread_prologue_and_epilogue_insns.
3606 2014-03-13 Vladimir Makarov <vmakarov@redhat.com>
3608 PR rtl-optimization/57189
3609 * lra-constraints.c (process_alt_operands): Disfavor spilling
3612 2014-03-13 Cesar Philippidis <cesar@codesourcery.com>
3614 * lto-wrapper.c (maybe_unlink_file): Suppress diagnostic messages.
3616 2014-03-13 Jakub Jelinek <jakub@redhat.com>
3618 PR tree-optimization/59025
3620 * tree-ssa-reassoc.c (sort_by_operand_rank): For SSA_NAMEs with the
3621 same rank, sort by bb_rank and gimple_uid of SSA_NAME_DEF_STMT first.
3623 2014-03-13 Georg-Johann Lay <avr@gjlay.de>
3626 * config/avr/avr.c (avr_out_plus): Swap cc_plus and cc_minus in
3627 calls of avr_out_plus_1.
3629 2014-03-13 Bin Cheng <bin.cheng@arm.com>
3631 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Record
3632 BB's single pred and update the father loop's latch info later.
3634 2014-03-12 Michael Meissner <meissner@linux.vnet.ibm.com>
3636 * config/rs6000/vector.md (VEC_L): Add V1TI mode to vector types.
3640 (VEC_base): Likewise.
3641 (mov<MODE>, VEC_M modes): If we are loading TImode into VSX
3642 registers, we need to swap double words in little endian mode.
3644 * config/rs6000/rs6000-modes.def (V1TImode): Add new vector mode
3645 to be a container mode for 128-bit integer operations added in ISA
3646 2.07. Unlike TImode and PTImode, the preferred register set is
3647 the Altivec/VMX registers for the 128-bit operations.
3649 * config/rs6000/rs6000-protos.h (rs6000_move_128bit_ok_p): Add
3651 (rs6000_split_128bit_ok_p): Likewise.
3653 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_3): Add new support
3654 macros for creating ISA 2.07 normal and overloaded builtin
3655 functions with 3 arguments.
3656 (BU_P8V_OVERLOAD_3): Likewise.
3657 (VPERM_1T): Add support for V1TImode in 128-bit vector operations
3658 for use as overloaded functions.
3659 (VPERM_1TI_UNS): Likewise.
3660 (VSEL_1TI): Likewise.
3661 (VSEL_1TI_UNS): Likewise.
3662 (ST_INTERNAL_1ti): Likewise.
3663 (LD_INTERNAL_1ti): Likewise.
3664 (XXSEL_1TI): Likewise.
3665 (XXSEL_1TI_UNS): Likewise.
3666 (VPERM_1TI): Likewise.
3667 (VPERM_1TI_UNS): Likewise.
3668 (XXPERMDI_1TI): Likewise.
3669 (SET_1TI): Likewise.
3670 (LXVD2X_V1TI): Likewise.
3671 (STXVD2X_V1TI): Likewise.
3672 (VEC_INIT_V1TI): Likewise.
3673 (VEC_SET_V1TI): Likewise.
3674 (VEC_EXT_V1TI): Likewise.
3675 (EQV_V1TI): Likewise.
3676 (NAND_V1TI): Likewise.
3677 (ORC_V1TI): Likewise.
3678 (VADDCUQ): Add support for 128-bit integer arithmetic instructions
3679 added in ISA 2.07. Add both normal 'altivec' builtins, and the
3681 (VADDUQM): Likewise.
3682 (VSUBCUQ): Likewise.
3683 (VADDEUQM): Likewise.
3684 (VADDECUQ): Likewise.
3685 (VSUBEUQM): Likewise.
3686 (VSUBECUQ): Likewise.
3688 * config/rs6000/rs6000-c.c (__int128_type): New static to hold
3689 __int128_t and __uint128_t types.
3690 (__uint128_type): Likewise.
3691 (altivec_categorize_keyword): Add support for vector __int128_t,
3692 vector __uint128_t, vector __int128, and vector unsigned __int128
3693 as a container type for TImode operations that need to be done in
3694 VSX/Altivec registers.
3695 (rs6000_macro_to_expand): Likewise.
3696 (altivec_overloaded_builtins): Add ISA 2.07 overloaded functions
3697 to support 128-bit integer instructions vaddcuq, vadduqm,
3698 vaddecuq, vaddeuqm, vsubcuq, vsubuqm, vsubecuq, vsubeuqm.
3699 (altivec_resolve_overloaded_builtin): Add support for V1TImode.
3701 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Add support
3702 for V1TImode, and set up preferences to use VSX/Altivec registers.
3703 Setup VSX reload handlers.
3704 (rs6000_debug_reg_global): Likewise.
3705 (rs6000_init_hard_regno_mode_ok): Likewise.
3706 (rs6000_preferred_simd_mode): Likewise.
3707 (vspltis_constant): Do not allow V1TImode as easy altivec constants.
3708 (easy_altivec_constant): Likewise.
3709 (output_vec_const_move): Likewise.
3710 (rs6000_expand_vector_set): Convert V1TImode set and extract to
3712 (rs6000_expand_vector_extract): Likewise.
3713 (reg_offset_addressing_ok_p): Setup V1TImode to use VSX reg+reg
3715 (rs6000_const_vec): Add support for V1TImode.
3716 (rs6000_emit_le_vsx_load): Swap double words when loading or
3717 storing TImode/V1TImode.
3718 (rs6000_emit_le_vsx_store): Likewise.
3719 (rs6000_emit_le_vsx_move): Likewise.
3720 (rs6000_emit_move): Add support for V1TImode.
3721 (altivec_expand_ld_builtin): Likewise.
3722 (altivec_expand_st_builtin): Likewise.
3723 (altivec_expand_vec_init_builtin): Likewise.
3724 (altivec_expand_builtin): Likewise.
3725 (rs6000_init_builtins): Add support for V1TImode type. Add
3726 support for ISA 2.07 128-bit integer builtins. Define type names
3727 for the VSX/Altivec vector types.
3728 (altivec_init_builtins): Add support for overloaded vector
3729 functions with V1TImode type.
3730 (rs6000_preferred_reload_class): Prefer Altivec registers for V1TImode.
3731 (rs6000_move_128bit_ok_p): Move 128-bit move/split validation to
3733 (rs6000_split_128bit_ok_p): Likewise.
3734 (rs6000_handle_altivec_attribute): Create V1TImode from vector
3735 __int128_t and vector __uint128_t.
3737 * config/rs6000/vsx.md (VSX_L): Add V1TImode to vector iterators
3738 and mode attributes.
3745 (VS_scalar): Likewise.
3746 (VS_double): Likewise.
3747 (vsx_set_v1ti): New builtin function to create V1TImode from TImode.
3749 * config/rs6000/rs6000.h (TARGET_VADDUQM): New macro to say whether
3750 we support the ISA 2.07 128-bit integer arithmetic instructions.
3751 (ALTIVEC_OR_VSX_VECTOR_MODE): Add V1TImode.
3752 (enum rs6000_builtin_type_index): Add fields to hold V1TImode
3753 and TImode types for use with the builtin functions.
3754 (V1TI_type_node): Likewise.
3755 (unsigned_V1TI_type_node): Likewise.
3756 (intTI_type_internal_node): Likewise.
3757 (uintTI_type_internal_node): Likewise.
3759 * config/rs6000/altivec.md (UNSPEC_VADDCUQ): New unspecs for ISA 2.07
3760 128-bit builtin functions.
3761 (UNSPEC_VADDEUQM): Likewise.
3762 (UNSPEC_VADDECUQ): Likewise.
3763 (UNSPEC_VSUBCUQ): Likewise.
3764 (UNSPEC_VSUBEUQM): Likewise.
3765 (UNSPEC_VSUBECUQ): Likewise.
3766 (VM): Add V1TImode to vector mode iterators.
3768 (VI_unit): Likewise.
3769 (altivec_vadduqm): Add ISA 2.07 128-bit binary builtins.
3770 (altivec_vaddcuq): Likewise.
3771 (altivec_vsubuqm): Likewise.
3772 (altivec_vsubcuq): Likewise.
3773 (altivec_vaddeuqm): Likewise.
3774 (altivec_vaddecuq): Likewise.
3775 (altivec_vsubeuqm): Likewise.
3776 (altivec_vsubecuq): Likewise.
3778 * config/rs6000/rs6000.md (FMOVE128_GPR): Add V1TImode to vector
3780 (BOOL_128): Likewise.
3781 (BOOL_REGS_OUTPUT): Likewise.
3782 (BOOL_REGS_OP1): Likewise.
3783 (BOOL_REGS_OP2): Likewise.
3784 (BOOL_REGS_UNARY): Likewise.
3785 (BOOL_REGS_AND_CR0): Likewise.
3787 * config/rs6000/altivec.h (vec_vaddcuq): Add support for ISA 2.07
3788 128-bit integer builtin support.
3789 (vec_vadduqm): Likewise.
3790 (vec_vaddecuq): Likewise.
3791 (vec_vaddeuqm): Likewise.
3792 (vec_vsubecuq): Likewise.
3793 (vec_vsubeuqm): Likewise.
3794 (vec_vsubcuq): Likewise.
3795 (vec_vsubuqm): Likewise.
3797 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
3798 Document vec_vaddcuq, vec_vadduqm, vec_vaddecuq, vec_vaddeuqm,
3799 vec_subecuq, vec_subeuqm, vec_vsubcuq, vec_vsubeqm builtins adding
3800 128-bit integer add/subtract to ISA 2.07.
3802 2014-03-12 Joern Rennecke <joern.rennecke@embecosm.com>
3804 * config/arc/arc.c (arc_predicate_delay_insns):
3805 Fix third argument passed to conditionalize_nonjump.
3807 2014-03-12 Yufeng Zhang <yufeng.zhang@arm.com>
3809 * config/aarch64/aarch64-builtins.c
3810 (aarch64_builtin_vectorized_function): Add BUILT_IN_LFLOORF,
3811 BUILT_IN_LLFLOOR, BUILT_IN_LCEILF and BUILT_IN_LLCEIL.
3812 * config/aarch64/arm_neon.h (vcvtaq_u64_f64): Call __builtin_llfloor
3813 instead of __builtin_lfloor.
3814 (vcvtnq_u64_f64): Call __builtin_llceil instead of __builtin_lceil.
3816 2014-03-12 Jakub Jelinek <jakub@redhat.com>
3818 * tree-ssa-ifcombine.c (forwarder_block_to): New function.
3819 (tree_ssa_ifcombine_bb_1): New function.
3820 (tree_ssa_ifcombine_bb): Use it. Handle also cases where else_bb
3821 is an empty forwarder block to then_bb or vice versa and then_bb
3822 and else_bb are effectively swapped.
3824 2014-03-12 Christian Bruel <christian.bruel@st.com>
3827 * config/arm/arm.c (arm_emit_vfp_multi_reg_pop): Emit a
3828 REG_CFA_DEF_CFA note.
3829 (arm_expand_epilogue_apcs_frame): call arm_add_cfa_adjust_cfa_note.
3830 (arm_unwind_emit): Allow REG_CFA_DEF_CFA.
3832 2014-03-12 Thomas Preud'homme <thomas.preudhomme@arm.com>
3834 PR tree-optimization/60454
3835 * tree-ssa-math-opts.c (find_bswap_1): Fix bswap detection.
3837 2014-03-12 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3839 * config.gcc (aarch64*-*-*): Use ISA flags from aarch64-arches.def.
3840 Do not define target_cpu_default2 to generic.
3841 * config/aarch64/aarch64.h (TARGET_CPU_DEFAULT): Use generic cpu.
3842 * config/aarch64/aarch64.c (aarch64_override_options): Update comment.
3843 * config/aarch64/aarch64-arches.def (armv8-a): Use generic cpu.
3845 2014-03-12 Jakub Jelinek <jakub@redhat.com>
3846 Marc Glisse <marc.glisse@inria.fr>
3848 PR tree-optimization/60502
3849 * tree-ssa-reassoc.c (eliminate_not_pairs): Use build_all_ones_cst
3850 instead of build_low_bits_mask.
3852 2014-03-12 Jakub Jelinek <jakub@redhat.com>
3855 * tree-vrp.c (register_edge_assert_for_1): Don't add assert
3856 if there are multiple uses, but op doesn't live on E edge.
3857 * tree-cfg.c (assert_unreachable_fallthru_edge_p): Also ignore
3858 clobber stmts before __builtin_unreachable.
3860 2014-03-11 Richard Sandiford <rdsandiford@googlemail.com>
3862 * builtins.c (expand_builtin_setjmp_receiver): Use and clobber
3863 hard_frame_pointer_rtx.
3864 * cse.c (cse_insn): Remove volatile check.
3865 * cselib.c (cselib_process_insn): Likewise.
3866 * dse.c (scan_insn): Likewise.
3868 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
3870 * config/arc/arc.c (conditionalize_nonjump): New function,
3872 (arc_ifcvt): ... this.
3873 (arc_predicate_delay_insns): Use it.
3875 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
3877 * config/arc/predicates.md (extend_operand): During/after reload,
3878 allow const_int_operand.
3879 * config/arc/arc.md (mulsidi3_700): Use extend_operand predicate.
3880 (umulsidi3_700): Likewise. Change operand 2 constraint back to "cL".
3881 (mulsi3_highpart): Change operand 2 constraint alternatives 2 and 3
3883 (umulsi3_highpart_i): Likewise.
3885 2014-03-11 Richard Biener <rguenther@suse.de>
3887 * tree-ssa-structalias.c (get_constraint_for_ptr_offset):
3888 Add asserts to guard possible wrong-code bugs.
3890 2014-03-11 Richard Biener <rguenther@suse.de>
3892 PR tree-optimization/60429
3893 PR tree-optimization/60485
3894 * tree-ssa-structalias.c (set_union_with_increment): Properly
3895 take into account all fields that overlap the shifted vars.
3896 (do_sd_constraint): Likewise.
3897 (do_ds_constraint): Likewise.
3898 (get_constraint_for_ptr_offset): Likewise.
3900 2014-03-11 Chung-Lin Tang <cltang@codesourcery.com>
3902 * config/nios2/nios2.c (machine_function): Add fp_save_offset field.
3903 (nios2_compute_frame_layout):
3904 Add calculation of cfun->machine->fp_save_offset.
3905 (nios2_expand_prologue): Correct setting of frame pointer register
3907 (nios2_expand_epilogue): Update recovery of stack pointer from
3908 frame pointer accordingly.
3909 (nios2_initial_elimination_offset): Update calculation of offset
3910 for eliminating to HARD_FRAME_POINTER_REGNUM.
3912 2014-03-10 Jakub Jelinek <jakub@redhat.com>
3915 * ipa.c (symtab_remove_unreachable_nodes): Don't call
3916 cgraph_get_create_node on VAR_DECLs.
3918 2014-03-10 Richard Biener <rguenther@suse.de>
3921 * tree.c (signed_or_unsigned_type_for): Handle OFFSET_TYPEs.
3923 2014-03-08 Douglas B Rupp <rupp@gnat.com>
3925 * config/vms/vms.opt (vms_float_format): New variable.
3927 2014-03-08 Tobias Burnus <burnus@net-b.de>
3929 * doc/invoke.texi (-fcilkplus): Update implementation status.
3931 2014-03-08 Paulo Matos <paulo@matos-sorge.com>
3932 Richard Biener <rguenther@suse.de>
3934 * lto-wrapper.c (merge_and_complain): Ensure -fshort-double is used
3935 consistently accross all TUs.
3936 (run_gcc): Enable -fshort-double automatically at link at link-time
3937 and disallow override.
3939 2014-03-08 Richard Sandiford <rdsandiford@googlemail.com>
3942 * config/mips/mips.c (mips_option_override): Promote -mpaired-single
3943 warning to an error. Disable TARGET_PAIRED_SINGLE and TARGET_MIPS3D
3944 if they can't be used.
3946 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3948 * configure.ac (HAVE_AS_IX86_TLSLDMPLT): Improve test
3949 for Solaris 11/x86 ld.
3950 * configure: Regenerate.
3952 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3954 * configure.ac (TLS_SECTION_ASM_FLAG): Save as tls_section_flag.
3955 (LIB_TLS_SPEC): Save as ld_tls_libs.
3956 (HAVE_AS_IX86_TLSLDMPLT): Define as 1/0.
3957 (HAVE_AS_IX86_TLSLDM): New test.
3958 * configure, config.in: Regenerate.
3959 * config/i386/i386.c (legitimize_tls_address): Fall back to
3960 TLS_MODEL_GLOBAL_DYNAMIC on 32-bit Solaris/x86 if tool chain
3961 cannot support TLS_MODEL_LOCAL_DYNAMIC.
3962 * config/i386/i386.md (*tls_local_dynamic_base_32_gnu): Use if
3963 instead of #ifdef in HAVE_AS_IX86_TLSLDMPLT test.
3965 2014-03-07 Paulo Matos <paulo@matos-sorge.com>
3967 * common.opt (fira-loop-pressure): Mark as optimization.
3969 2014-03-07 Thomas Schwinge <thomas@codesourcery.com>
3971 * langhooks.c (lhd_omp_mappable_type): The error_mark_node is not
3972 an OpenMP mappable type.
3974 2014-03-06 Matthias Klose <doko@ubuntu.com>
3976 * Makefile.in (s-mlib): Only pass MULTIARCH_DIRNAME if
3977 MULTILIB_OSDIRNAMES is not defined.
3979 2014-03-06 Jakub Jelinek <jakub@redhat.com>
3980 Meador Inge <meadori@codesourcery.com>
3983 * config/arm/arm.c (arm_tls_symbol_p): Remove.
3984 (arm_legitimize_address): Call legitimize_tls_address for any
3985 arm_tls_referenced_p expression, handle constant addend. Call it
3986 before testing for !TARGET_ARM.
3987 (thumb_legitimize_address): Don't handle arm_tls_symbol_p here.
3989 2014-03-06 Richard Biener <rguenther@suse.de>
3995 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
3997 * tree-streamer.c (record_common_node): Assert we don't record
3998 nodes with type double.
3999 (preload_common_node): Skip type double, complex double and double
4000 pointer since it is now frontend dependent due to fshort-double option.
4002 2014-03-06 Richard Biener <rguenther@suse.de>
4004 * gcc.c (PLUGIN_COND): Always enable unless -fno-use-linker-plugin
4005 or -fno-lto is specified and the linker has full plugin support.
4006 * collect2.c (lto_mode): Default to LTO_MODE_WHOPR if LTO is enabled.
4007 (main): Remove -flto processing, adjust lto_mode using use_plugin late.
4008 * lto-wrapper.c (merge_and_complain): Merge compile-time
4009 optimization levels.
4010 (run_gcc): And pass it through to the link options.
4012 2014-03-06 Alexandre Oliva <aoliva@redhat.com>
4016 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4018 * cselib.c (remove_useless_values): Skip to avoid quadratic
4019 behavior if the condition moved from...
4020 (cselib_process_insn): ... here holds.
4022 2014-03-05 Jakub Jelinek <jakub@redhat.com>
4025 * Makefile.in (PLUGIN_HEADERS): Add tree-phinodes.h, stor-layout.h,
4026 ssa-iterators.h, $(RESOURCE_H) and tree-cfgcleanup.h.
4029 * config/i386/t-i386 (OPTIONS_H_EXTRA): Add stringop.def.
4030 (TM_H): Add x86-tune.def.
4032 2014-03-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4034 * config/aarch64/aarch64.c (generic_tunings):
4035 Use cortexa57_extra_costs.
4037 2014-03-05 Jakub Jelinek <jakub@redhat.com>
4040 * cfgexpand.c (expand_used_vars): Do not assume all SSA_NAMEs
4041 of PARM/RESULT_DECLs must be coalesced with optimize && in_lto_p.
4042 * tree-ssa-coalesce.c (coalesce_ssa_name): Use MUST_COALESCE_COST - 1
4045 2014-03-04 Heiher <r@hev.cc>
4047 * config/mips/mips-cpus.def (loongson3a): Mark as a MIPS64r2 processor.
4048 * config/mips/mips.h (MIPS_ISA_LEVEL_SPEC): Adjust accordingly.
4050 2014-03-04 Uros Bizjak <ubizjak@gmail.com>
4052 * config/i386/predicates.md (const2356_operand): Change to ...
4053 (const2367_operand): ... this.
4054 * config/i386/sse.md (avx512pf_scatterpf<mode>sf): Use
4056 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
4057 (*avx512pf_scatterpf<mode>sf): Ditto.
4058 (avx512pf_scatterpf<mode>df): Ditto.
4059 (*avx512pf_scatterpf<mode>df_mask): Ditto.
4060 (*avx512pf_scatterpf<mode>df): Ditto.
4061 * config/i386/i386.c (ix86_expand_builtin): Update
4062 incorrect hint operand error message.
4064 2014-03-04 Richard Biener <rguenther@suse.de>
4066 * lto-section-in.c (lto_get_section_data): Fix const cast.
4068 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
4070 * tree-streamer.c (record_common_node): Assert we don't record
4071 nodes with type double.
4072 (preload_common_node): Skip type double, complex double and double
4073 pointer since it is now frontend dependent due to fshort-double option.
4075 2014-03-04 Richard Biener <rguenther@suse.de>
4078 * lto-streamer-in.c (lto_read_body): Remove LTO bytecode version check.
4079 (lto_input_toplevel_asms): Likewise.
4080 * lto-section-in.c (lto_get_section_data): Instead do it here
4083 2014-03-04 Richard Biener <rguenther@suse.de>
4085 PR tree-optimization/60382
4086 * tree-vect-loop.c (vect_is_simple_reduction_1): Do not consider
4087 dead PHIs a reduction.
4089 2014-03-03 Uros Bizjak <ubizjak@gmail.com>
4091 * config/i386/xmmintrin.h (enum _mm_hint) <_MM_HINT_ET0>: Correct
4093 (_mm_prefetch): Move out of GCC target("sse") pragma.
4094 * config/i386/prfchwintrin.h (_m_prefetchw): Move out of
4095 GCC target("prfchw") pragma.
4096 * config/i386/i386.md (prefetch): Emit prefetchwt1 only
4098 * config/i386/i386.c (ix86_option_override_internal): Enable
4099 -mprfchw with -mprefetchwt1.
4101 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
4103 * config/arc/arc.md (casesi_load) <length attribute alternative 0>:
4106 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
4108 * opts.h (CL_PCH_IGNORE): Define.
4109 * targhooks.c (option_affects_pch_p):
4110 Return false for options that have CL_PCH_IGNORE set.
4111 * opt-functions.awk: Process PchIgnore.
4112 * doc/options.texi: Document PchIgnore.
4114 * config/arc/arc.opt (misize): Add PchIgnore property.
4116 2014-03-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4118 * config/rs6000/rs6000.c (rs6000_preferred_reload_class): Disallow
4119 reload of PLUS rtx's outside of GENERAL_REGS or BASE_REGS; relax
4120 constraint on constants to permit them being loaded into
4121 GENERAL_REGS or BASE_REGS.
4123 2014-03-03 Nick Clifton <nickc@redhat.com>
4125 * config/rl78/rl78-real.md (cbranchsi4_real_signed): Add
4126 anti-cacnonical alternatives.
4127 (negandhi3_real): New pattern.
4128 * config/rl78/rl78-virt.md (negandhi3_virt): New pattern.
4130 2014-03-03 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
4132 * config/avr/avr-mcus.def: Remove atxmega16x1.
4133 * config/avr/avr-tables.opt: Regenerate.
4134 * config/avr/t-multilib: Regenerate.
4135 * doc/avr-mmcu.texi: Regenerate.
4137 2014-03-03 Tobias Grosser <tobias@grosser.es>
4138 Mircea Namolaru <mircea.namolaru@inria.fr>
4140 PR tree-optimization/58028
4141 * graphite-clast-to-gimple.c (set_cloog_options): Don't remove
4144 2014-03-03 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4146 * config/arm/neon.md (*movmisalign<mode>): Legitimize addresses
4147 not handled by recognizers.
4149 2014-03-03 Jakub Jelinek <jakub@redhat.com>
4152 * function.c (expand_function_end): Don't emit
4153 clobber_return_register sequence if clobber_after is a BARRIER.
4154 * cfgexpand.c (construct_exit_block): Append instructions before
4155 return_label to prev_bb.
4157 2014-03-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4159 * config/rs6000/constraints.md: Document reserved use of "wc".
4161 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
4164 * ipa.c (function_and_variable_visibility): When dissolving comdat
4165 group, also set all symbols to local.
4167 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
4172 2013-12-14 Jan Hubicka <jh@suse.cz>
4174 * ipa-prop.c (stmt_may_be_vtbl_ptr_store): Skip clobbers.
4176 2014-03-02 Jon Beniston <jon@beniston.com>
4182 * config/lm32/lm32.c (lm32_legitimate_constant_p): Remove, as incorrect.
4183 (TARGET_LEGITIMATE_CONSTANT_P): Undefine, as not needed.
4184 * config/lm32/lm32.md (movsi_insn): Add 32-bit immediate support.
4185 (simple_return, *simple_return): New patterns
4186 * config/lm32/predicates.md (movsi_rhs_operand): Remove as obsolete.
4187 * configure.ac (force_sjlj_exceptions): Force sjlj exceptions for lm32.
4189 2014-03-01 Paolo Carlini <paolo.carlini@oracle.com>
4191 * dwarf2out.c (gen_subprogram_die): Tidy.
4193 2014-03-01 Oleg Endo <olegendo@gcc.gnu.org>
4196 * config/sh/sh.md (*mov_t_msb_neg): Split into ...
4197 (*mov_t_msb_neg_negc): ... this new insn.
4199 2014-02-28 Jason Merrill <jason@redhat.com>
4202 * ipa-devirt.c (ipa_devirt): Don't choose an implicitly-declared
4205 2014-02-28 Paolo Carlini <paolo.carlini@oracle.com>
4208 * dwarf2out.c (decltype_auto_die): New static.
4209 (gen_subprogram_die): Handle 'decltype(auto)' like 'auto'.
4210 (gen_type_die_with_usage): Handle 'decltype(auto)'.
4211 (is_cxx_auto): Likewise.
4213 2014-02-28 Ian Bolton <ian.bolton@arm.com>
4215 * config/aarch64/aarch64.h: Define __ARM_NEON by default if
4216 we are not using general regs only.
4218 2014-02-28 Richard Biener <rguenther@suse.de>
4221 * tree-cfgcleanup.c (tree_forwarder_block_p): Restrict
4222 previous fix and only allow to remove trivial pre-headers
4223 and latches. Also honor LOOPS_MAY_HAVE_MULTIPLE_LATCHES.
4224 (remove_forwarder_block): Properly update the latch of a loop.
4226 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4229 * cselib.c (cselib_hasher::equal): Special-case VALUE lookup.
4230 (cselib_preserved_hash_table): New.
4231 (preserve_constants_and_equivs): Move preserved vals to it.
4232 (cselib_find_slot): Look it up first.
4233 (cselib_init): Initialize it.
4234 (cselib_finish): Release it.
4235 (dump_cselib_table): Dump it.
4237 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4240 * cselib.c (remove_useless_values): Skip to avoid quadratic
4241 behavior if the condition moved from...
4242 (cselib_process_insn): ... here holds.
4244 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4247 * var-tracking.c (vt_initialize): Apply the same condition to
4248 preserve the CFA base value.
4250 2014-02-28 Joey Ye <joey.ye@arm.com>
4253 * config/arm/arm.c (thumb_far_jump_used_p): Don't change
4254 if reload in progress or completed.
4256 2014-02-28 Tobias Burnus <burnus@net-b.de>
4259 * tree-pretty-print.c (dump_generic_node, print_declaration): Handle
4262 2014-02-27 H.J. Lu <hongjiu.lu@intel.com>
4264 * doc/tm.texi.in (Condition Code Status): Update documention for
4265 relative locations of cc0-setter and cc0-user.
4267 2014-02-27 Jeff Law <law@redhat.com>
4269 PR rtl-optimization/52714
4270 * combine.c (try_combine): When splitting an unrecognized PARALLEL
4271 into two independent simple sets, if I3 is a jump, ensure the
4272 pattern we place into I3 is a (set (pc) ...).
4274 2014-02-27 Mikael Pettersson <mikpe@it.uu.se>
4275 Jeff Law <law@redhat.com>
4277 PR rtl-optimization/49847
4278 * cse.c (fold_rtx) Handle case where cc0 setter and cc0 user
4279 are in different blocks.
4280 * doc/tm.texi (Condition Code Status): Update documention for
4281 relative locations of cc0-setter and cc0-user.
4283 2014-02-27 Vladimir Makarov <vmakarov@redhat.com>
4286 * lra.c (lra_emit_add): Check SUBREG too.
4288 2014-02-27 Andreas Schwab <schwab@suse.de>
4290 * config/m68k/m68k.c (m68k_option_override): Disable
4291 -flive-range-shrinkage for classic m68k.
4292 (m68k_override_options_after_change): Likewise.
4294 2014-02-27 Marek Polacek <polacek@redhat.com>
4297 * tree-ssa-uninit.c (gate_warn_uninitialized): Run the pass even for
4298 -Wmaybe-uninitialized.
4300 2014-02-27 Alan Modra <amodra@gmail.com>
4303 * reload1.c (emit_input_reload_insns): When reload_override_in,
4304 set old to rl->in_reg when rl->in_reg is a subreg.
4306 2014-02-26 Richard Biener <rguenther@suse.de>
4309 * lra-assigns.c (spill_for): Avoid mixed-sign comparison.
4311 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
4313 * common/config/i386/predicates.md (const1256_operand): Remove.
4314 (const2356_operand): New.
4315 (const_1_to_2_operand): Remove.
4316 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
4317 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
4318 (*avx512pf_gatherpf<mode>sf): Ditto.
4319 (avx512pf_gatherpf<mode>df): Ditto.
4320 (*avx512pf_gatherpf<mode>df_mask): Ditto.
4321 (*avx512pf_gatherpf<mode>df): Ditto.
4322 (avx512pf_scatterpf<mode>sf): Ditto.
4323 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
4324 (*avx512pf_scatterpf<mode>sf): Ditto.
4325 (avx512pf_scatterpf<mode>df): Ditto.
4326 (*avx512pf_scatterpf<mode>df_mask): Ditto.
4327 (*avx512pf_scatterpf<mode>df): Ditto.
4328 * common/config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET0.
4330 2014-02-26 Ilya Tocar <ilya.tocar@intel.com>
4332 * config/i386/avx512fintrin.h (_mm512_testn_epi32_mask),
4333 (_mm512_mask_testn_epi32_mask), (_mm512_testn_epi64_mask),
4334 (_mm512_mask_testn_epi64_mask): Move to ...
4335 * config/i386/avx512cdintrin.h: Here.
4336 * config/i386/i386.c (bdesc_args): Change MASK_ISA for testnm.
4337 * config/i386/sse.md (avx512f_vmscalef<mode><round_name>): Remove %.
4338 (avx512f_scalef<mode><mask_name><round_name>): Ditto.
4339 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Change conditon to
4340 TARGET_AVX512F from TARGET_AVX512CD.
4342 2014-02-26 Richard Biener <rguenther@suse.de>
4345 * ipa.c (walk_polymorphic_call_targets): Properly guard
4346 call to inline_update_overall_summary.
4348 2014-02-26 Bin Cheng <bin.cheng@arm.com>
4351 * tree-cfgcleanup.c (tree_forwarder_block_p): Protect loop preheaders
4352 and latches only if requested. Fix latch if it is removed.
4353 * tree-ssa-dom.c (tree_ssa_dominator_optimize): Set
4354 LOOPS_HAVE_PREHEADERS.
4356 2014-02-25 Andrew Pinski <apinski@cavium.com>
4358 * builtins.c (expand_builtin_thread_pointer): Create a new target
4359 when the target is NULL.
4361 2014-02-25 Vladimir Makarov <vmakarov@redhat.com>
4363 PR rtl-optimization/60317
4364 * params.def (PARAM_LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
4365 * params.h (LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
4366 * lra-assigns.c: Include params.h.
4367 (spill_for): Use LRA_MAX_CONSIDERED_RELOAD_PSEUDOS as guard for
4368 other reload pseudos considerations.
4370 2014-02-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4372 * config/rs6000/vector.md (*vector_unordered<mode>): Change split
4373 to use canonical form for nor<mode>3.
4375 2014-02-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4378 * config/arm/arm.h (CANNOT_CHANGE_MODE_CLASS): Allow 128 to 64-bit
4381 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
4383 * common/config/i386/i386-common.c (OPTION_MASK_ISA_PREFETCHWT1_SET),
4384 (OPTION_MASK_ISA_PREFETCHWT1_UNSET): New.
4385 (ix86_handle_option): Handle OPT_mprefetchwt1.
4386 * config/i386/cpuid.h (bit_PREFETCHWT1): New.
4387 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
4389 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
4390 OPTION_MASK_ISA_PREFETCHWT1.
4391 * config/i386/i386.c (ix86_target_string): Handle mprefetchwt1.
4392 (PTA_PREFETCHWT1): New.
4393 (ix86_option_override_internal): Handle PTA_PREFETCHWT1.
4394 (ix86_valid_target_attribute_inner_p): Handle OPT_mprefetchwt1.
4395 * config/i386/i386.h (TARGET_PREFETCHWT1, TARGET_PREFETCHWT1_P): New.
4396 * config/i386/i386.md (prefetch): Check TARGET_PREFETCHWT1
4397 (*prefetch_avx512pf_<mode>_: Change into ...
4398 (*prefetch_prefetchwt1_<mode>: This.
4399 * config/i386/i386.opt (mprefetchwt1): New.
4400 * config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET1.
4401 (_mm_prefetch): Handle intent to write.
4402 * doc/invoke.texi (mprefetchwt1), (mno-prefetchwt1): Doccument.
4404 2014-02-25 Richard Biener <rguenther@suse.de>
4407 * emit-rtl.c (mem_attrs_htab): Remove.
4408 (mem_attrs_htab_hash): Likewise.
4409 (mem_attrs_htab_eq): Likewise.
4410 (set_mem_attrs): Always allocate new mem-attrs when something changed.
4411 (init_emit_once): Do not allocate mem_attrs_htab.
4413 2014-02-25 Richard Biener <rguenther@suse.de>
4416 * lto-opts.c (lto_write_options): Output non-explicit conservative
4417 -fwrapv, -fno-trapv and -fno-strict-overflow.
4418 * lto-wrapper.c (merge_and_complain): Handle merging those options.
4419 (run_gcc): And pass them through.
4421 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
4423 * sel-sched.c (calculate_new_fences): New parameter ptime.
4424 Calculate it as a maximum over all fence cycles.
4425 (sel_sched_region_2): Adjust the call to calculate_new_fences.
4426 Print the final schedule timing when sched_verbose.
4428 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
4430 PR rtl-optimization/60292
4431 * sel-sched.c (fill_vec_av_set): Do not reset target availability
4432 bit fot the fence instruction.
4434 2014-02-24 Alangi Derick <alangiderick@gmail.com>
4436 * calls.h: Fix typo in comment.
4438 2014-02-24 John David Anglin <danglin@gcc.gnu.org>
4440 * config/pa/pa.c (pa_output_move_double): Don't valididate when
4441 adjusting offsetable addresses.
4443 2014-02-24 Guozhi Wei <carrot@google.com>
4445 * sparseset.h (sparseset_pop): Fix the wrong index.
4447 2014-02-24 Walter Lee <walt@tilera.com>
4449 * config.gcc (tilepro-*-*): Change to tilepro*-*-*.
4450 (tilegx-*-linux*): Change to tilegx*-*-linux*; Support tilegxbe
4452 * common/config/tilegx/tilegx-common.c
4453 (TARGET_DEFAULT_TARGET_FLAGS): Define.
4454 * config/tilegx/linux.h (ASM_SPEC): Add endian_spec.
4456 * config/tilegx/sync.md (atomic_test_and_set): Handle big endian.
4457 * config/tilegx/tilegx.c (tilegx_return_in_msb): New.
4458 (tilegx_gimplify_va_arg_expr): Handle big endian.
4459 (tilegx_expand_unaligned_load): Ditto.
4460 (tilegx_expand_unaligned_store): Ditto.
4461 (TARGET_RETURN_IN_MSB): New.
4462 * config/tilegx/tilegx.h (TARGET_DEFAULT): New.
4463 (TARGET_ENDIAN_DEFAULT): New.
4464 (TARGET_BIG_ENDIAN): Handle big endian.
4465 (BYTES_BIG_ENDIAN): Ditto.
4466 (WORDS_BIG_ENDIAN): Ditto.
4467 (FLOAT_WORDS_BIG_ENDIAN): Ditto.
4470 * config/tilegx/tilegx.md (extv): Handle big endian.
4472 (insn_st<n>): Ditto.
4473 (insn_st<n>_add<bitsuffix>): Ditto.
4474 (insn_stnt<n>): Ditto.
4475 (insn_stnt<n>_add<bitsuffix>):Ditto.
4476 (vec_interleave_highv8qi): Handle big endian.
4477 (vec_interleave_highv8qi_be): New.
4478 (vec_interleave_highv8qi_le): New.
4479 (insn_v1int_h): Handle big endian.
4480 (vec_interleave_lowv8qi): Handle big endian.
4481 (vec_interleave_lowv8qi_be): New.
4482 (vec_interleave_lowv8qi_le): New.
4483 (insn_v1int_l): Handle big endian.
4484 (vec_interleave_highv4hi): Handle big endian.
4485 (vec_interleave_highv4hi_be): New.
4486 (vec_interleave_highv4hi_le): New.
4487 (insn_v2int_h): Handle big endian.
4488 (vec_interleave_lowv4hi): Handle big endian.
4489 (vec_interleave_lowv4hi_be): New.
4490 (vec_interleave_lowv4hi_le): New.
4491 (insn_v2int_l): Handle big endian.
4492 (vec_interleave_highv2si): Handle big endian.
4493 (vec_interleave_highv2si_be): New.
4494 (vec_interleave_highv2si_le): New.
4495 (insn_v4int_h): Handle big endian.
4496 (vec_interleave_lowv2si): Handle big endian.
4497 (vec_interleave_lowv2si_be): New.
4498 (vec_interleave_lowv2si_le): New.
4499 (insn_v4int_l): Handle big endian.
4500 * config/tilegx/tilegx.opt (mbig-endian): New option.
4501 (mlittle-endian): New option.
4502 * doc/install.texi: Document tilegxbe-linux.
4503 * doc/invoke.texi: Document -mbig-endian and -mlittle-endian.
4505 2014-02-24 Martin Jambor <mjambor@suse.cz>
4508 * ipa-cp.c (propagate_constants_accross_call): Bail out early if
4509 there are no parameter descriptors.
4511 2014-02-24 Andrey Belevantsev <abel@ispras.ru>
4513 PR rtl-optimization/60268
4514 * sched-rgn.c (haifa_find_rgns): Move the nr_regions_initial variable
4515 initialization to ...
4516 (sched_rgn_init): ... here.
4517 (schedule_region): Check for SCHED_PRESSURE_NONE earlier.
4519 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
4521 * config/microblaze/microblaze.md: Correct ashrsi_reg / lshrsi_reg
4524 2014-02-23 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
4526 * config/microblaze/microblaze.h: Remove SECONDARY_MEMORY_NEEDED
4529 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
4531 * /config/microblaze/microblaze.c: Add microblaze_asm_output_mi_thunk,
4532 define TARGET_ASM_OUTPUT_MI_THUNK and TARGET_ASM_CAN_OUTPUT_MI_THUNK.
4534 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
4536 * config/microblaze/predicates.md: Add cmp_op predicate.
4537 * config/microblaze/microblaze.md: Add branch_compare instruction
4538 which uses cmp_op predicate and emits cmp insn before branch.
4539 * config/microblaze/microblaze.c (microblaze_emit_compare): Rename
4540 to microblaze_expand_conditional_branch and consolidate logic.
4541 (microblaze_expand_conditional_branch): emit branch_compare
4542 insn instead of handling cmp op separate from branch insn.
4544 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4546 * config/rs6000/rs6000.c (rs6000_emit_le_vsx_move): Relax assert
4549 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4551 * config/rs6000/altivec.md (altivec_lve<VI_char>x): Replace
4552 define_insn with define_expand and new define_insn
4553 *altivec_lve<VI_char>x_internal.
4554 (altivec_stve<VI_char>x): Replace define_insn with define_expand
4555 and new define_insn *altivec_stve<VI_char>x_internal.
4556 * config/rs6000/rs6000-protos.h (altivec_expand_stvex_be): New
4558 * config/rs6000/rs6000.c (altivec_expand_lvx_be): Document use by
4560 (altivec_expand_stvex_be): New function.
4562 2014-02-22 Joern Rennecke <joern.rennecke@embecosm.com>
4564 * config/avr/avr.c (avr_can_eliminate): Allow elimination from
4565 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM if !frame_pointer_needed.
4566 * config/avr/avr.c (ELIMINABLE_REGS): Add elimination from
4567 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM.
4569 2014-02-21 Vladimir Makarov <vmakarov@redhat.com>
4572 * lra-constraints.c (inherit_reload_reg): Use lra_emit_move
4573 instead of emit_move_insn.
4575 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4577 * config/rs6000/altivec.md (altivec_vsumsws): Replace second
4579 (reduc_uplus_v16qi): Use gen_altivec_vsumsws_direct instead of
4580 gen_altivec_vsumsws.
4582 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4584 * config/rs6000/altivec.md (altivec_lvxl): Rename as
4585 *altivec_lvxl_<mode>_internal and use VM2 iterator instead of V4SI.
4586 (altivec_lvxl_<mode>): New define_expand incorporating
4587 -maltivec=be semantics where needed.
4588 (altivec_lvx): Rename as *altivec_lvx_<mode>_internal.
4589 (altivec_lvx_<mode>): New define_expand incorporating -maltivec=be
4590 semantics where needed.
4591 (altivec_stvx): Rename as *altivec_stvx_<mode>_internal.
4592 (altivec_stvx_<mode>): New define_expand incorporating
4593 -maltivec=be semantics where needed.
4594 (altivec_stvxl): Rename as *altivec_stvxl_<mode>_internal and use
4595 VM2 iterator instead of V4SI.
4596 (altivec_stvxl_<mode>): New define_expand incorporating
4597 -maltivec=be semantics where needed.
4598 * config/rs6000/rs6000-builtin.def: Add new built-in definitions
4599 LVXL_V2DF, LVXL_V2DI, LVXL_V4SF, LVXL_V4SI, LVXL_V8HI, LVXL_V16QI,
4600 LVX_V2DF, LVX_V2DI, LVX_V4SF, LVX_V4SI, LVX_V8HI, LVX_V16QI, STVX_V2DF,
4601 STVX_V2DI, STVX_V4SF, STVX_V4SI, STVX_V8HI, STVX_V16QI, STVXL_V2DF,
4602 STVXL_V2DI, STVXL_V4SF, STVXL_V4SI, STVXL_V8HI, STVXL_V16QI.
4603 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Replace
4604 ALTIVEC_BUILTIN_LVX with ALTIVEC_BUILTIN_LVX_<MODE> throughout;
4605 similarly for ALTIVEC_BUILTIN_LVXL, ALTIVEC_BUILTIN_STVX, and
4606 ALTIVEC_BUILTIN_STVXL.
4607 * config/rs6000/rs6000-protos.h (altivec_expand_lvx_be): New prototype.
4608 (altivec_expand_stvx_be): Likewise.
4609 * config/rs6000/rs6000.c (swap_selector_for_mode): New function.
4610 (altivec_expand_lvx_be): Likewise.
4611 (altivec_expand_stvx_be): Likewise.
4612 (altivec_expand_builtin): Add cases for
4613 ALTIVEC_BUILTIN_STVX_<MODE>, ALTIVEC_BUILTIN_STVXL_<MODE>,
4614 ALTIVEC_BUILTIN_LVXL_<MODE>, and ALTIVEC_BUILTIN_LVX_<MODE>.
4615 (altivec_init_builtins): Add definitions for
4616 __builtin_altivec_lvxl_<mode>, __builtin_altivec_lvx_<mode>,
4617 __builtin_altivec_stvx_<mode>, and __builtin_altivec_stvxl_<mode>.
4619 2014-02-21 Catherine Moore <clm@codesourcery.com>
4621 * doc/invoke.texi (mvirt, mno-virt): Document.
4622 * config/mips/mips.opt (mvirt): New option.
4623 * config/mips/mips.h (ASM_SPEC): Pass mvirt to the assembler.
4625 2014-02-21 Richard Biener <rguenther@suse.de>
4627 PR tree-optimization/60276
4628 * tree-vectorizer.h (struct _stmt_vec_info): Add min_neg_dist field.
4629 (STMT_VINFO_MIN_NEG_DIST): New macro.
4630 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Record
4631 STMT_VINFO_MIN_NEG_DIST.
4632 * tree-vect-stmts.c (vectorizable_load): Verify if assumptions
4633 made for negative dependence distances still hold.
4635 2014-02-21 Richard Biener <rguenther@suse.de>
4638 * tree-ssa-live.c (mark_all_vars_used_1): Do not walk
4639 DECL_INITIAL for globals not in the current function context.
4641 2014-02-21 Jakub Jelinek <jakub@redhat.com>
4643 PR tree-optimization/56490
4644 * params.def (PARAM_UNINIT_CONTROL_DEP_ATTEMPTS): New param.
4645 * tree-ssa-uninit.c: Include params.h.
4646 (compute_control_dep_chain): Add num_calls argument, return false
4647 if it exceed PARAM_UNINIT_CONTROL_DEP_ATTEMPTS param, pass
4648 num_calls to recursive call.
4649 (find_predicates): Change dep_chain into normal array,
4650 cur_chain into auto_vec<edge, MAX_CHAIN_LEN + 1>, add num_calls
4651 variable and adjust compute_control_dep_chain caller.
4652 (find_def_preds): Likewise.
4654 2014-02-21 Thomas Schwinge <thomas@codesourcery.com>
4656 * gimple-pretty-print.c (dump_gimple_omp_for) [flags & TDF_RAW]
4657 <case GF_OMP_FOR_KIND_CILKSIMD>: Add missing break statement.
4659 2014-02-21 Nick Clifton <nickc@redhat.com>
4661 * config/stormy16/stormy16.md (pushdqi1): Add mode to post_inc.
4662 (pushhi1): Likewise.
4663 (popqi1): Add mode to pre_dec.
4666 2014-02-21 Jakub Jelinek <jakub@redhat.com>
4668 * config/i386/i386.c (ix86_expand_vec_perm): Use V8SImode
4669 mode for mask of V8SFmode permutation.
4671 2014-02-20 Richard Henderson <rth@redhat.com>
4674 * builtins.c (expand_builtin_atomic_compare_exchange): Always make
4675 a new pseudo for OLDVAL.
4677 2014-02-20 Jakub Jelinek <jakub@redhat.com>
4680 * config/i386/i386.c (expand_vec_perm_interleave2): Don't call
4681 gen_reg_rtx if d->testing_p.
4682 (expand_vec_perm_pshufb2, expand_vec_perm_broadcast_1): Return early
4683 if d->testing_p and we will certainly return true.
4684 (expand_vec_perm_even_odd_1): Likewise. Don't call gen_reg_rtx
4687 2014-02-20 Uros Bizjak <ubizjak@gmail.com>
4689 * emit-rtl.c (gen_reg_rtx): Assert that
4690 crtl->emit.regno_pointer_align_length is non-zero.
4692 2014-02-20 Richard Henderson <rth@redhat.com>
4695 * builtins.c (expand_builtin_atomic_compare_exchange): Conditionalize
4696 on failure the store back into EXPECT.
4698 2014-02-20 Chung-Lin Tang <cltang@codesourcery.com>
4699 Sandra Loosemore <sandra@codesourcery.com>
4701 * config/nios2/nios2.md (unspec): Add UNSPEC_PIC_GOTOFF_SYM enum.
4702 * config/nios2/nios2.c (nios2_function_profiler): Add
4703 -fPIC (flag_pic == 2) support.
4704 (nios2_handle_custom_fpu_cfg): Fix warning parameter.
4705 (nios2_large_offset_p): New function.
4706 (nios2_unspec_reloc_p): Move up position, update to use
4707 nios2_large_offset_p.
4708 (nios2_unspec_address): Remove function.
4709 (nios2_unspec_offset): New function.
4710 (nios2_large_got_address): New function.
4711 (nios2_got_address): Add large offset support.
4712 (nios2_legitimize_tls_address): Update usage of removed and new
4714 (nios2_symbol_binds_local_p): New function.
4715 (nios2_load_pic_address): Add -fPIC (flag_pic == 2) support.
4716 (nios2_legitimize_address): Update to use nios2_large_offset_p.
4717 (nios2_emit_move_sequence): Avoid legitimizing (const (unspec ...)).
4718 (nios2_print_operand): Merge H/L processing, add hiadj/lo
4719 processing for (const (unspec ...)).
4720 (nios2_unspec_reloc_name): Add UNSPEC_PIC_GOTOFF_SYM case.
4722 2014-02-20 Richard Biener <rguenther@suse.de>
4724 * tree-cfg.c (replace_uses_by): Mark altered BBs before
4725 doing the substitution.
4726 (verify_gimple_assign_single): Also verify bare MEM_REFs on the lhs.
4728 2014-02-20 Martin Jambor <mjambor@suse.cz>
4731 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Uce correct
4732 info when checking whether lattices are bottom.
4734 2014-02-20 Richard Biener <rguenther@suse.de>
4737 * tree-eh.c (execute_cleanup_eh_1): Also cleanup empty EH
4740 2014-02-20 Jan Hubicka <hubicka@ucw.cz>
4743 * ipa-inline-transform.c (clone_inlined_nodes): Add freq_scale
4744 parameter specifying the scaling.
4745 (inline_call): Update.
4746 (want_inline_recursively): Guard division by zero.
4747 (recursive_inlining): Update.
4748 * ipa-inline.h (clone_inlined_nodes): Update.
4750 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
4753 * config/i386/i386.c (classify_argument): Pass structures of size
4754 64 bytes or less in register.
4756 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
4757 Kirill Yukhin <kirill.yukhin@intel.com>
4759 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): Swap operands.
4760 (_mm_rcp28_round_ss): Ditto.
4761 (_mm_rsqrt28_round_sd): Ditto.
4762 (_mm_rsqrt28_round_ss): Ditto.
4763 * config/i386/avx512erintrin.h (_mm_rcp14_round_sd): Ditto.
4764 (_mm_rcp14_round_ss): Ditto.
4765 (_mm_rsqrt14_round_sd): Ditto.
4766 (_mm_rsqrt14_round_ss): Ditto.
4767 * config/i386/sse.md (rsqrt14<mode>): Put nonimmediate operand as
4768 the first input operand, get rid of match_dup.
4769 (avx512er_exp2<mode><mask_name><round_saeonly_name>): Set type
4771 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
4773 (avx512er_vmrcp28<mode><round_saeonly_name>): Put nonimmediate
4774 operand as the first input operand, set type attribute.
4775 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
4777 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Put nonimmediate
4778 operand as the first input operand, set type attribute.
4780 2014-02-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4782 * config/rs6000/rs6000.c (vspltis_constant): Fix most significant
4785 2014-02-19 H.J. Lu <hongjiu.lu@intel.com>
4788 * config/i386/i386.c (construct_container): Remove TFmode check
4789 for X86_64_INTEGER_CLASS.
4791 2014-02-19 Uros Bizjak <ubizjak@gmail.com>
4794 * config/i386/i386.c (type_natural_mode): Warn for ABI changes
4795 only when -Wpsabi is enabled.
4797 2014-02-19 Michael Hudson-Doyle <michael.hudson@linaro.org>
4800 * config/aarch64/aarch64.c (aarch64_pass_by_reference): The rules for
4801 passing arrays in registers are the same as for structs, so remove the
4802 special case for them.
4804 2014-02-19 Eric Botcazou <ebotcazou@adacore.com>
4806 * expr.c (expand_expr_real_1) <case VIEW_CONVERT_EXPR>: For a bit-field
4807 destination type, extract only the valid bits if the source type is not
4808 integral and has a different mode.
4810 2014-02-19 Richard Biener <rguenther@suse.de>
4813 * tree-inline.c (estimate_num_insns): Avoid calling cgraph_get_node
4816 2014-02-19 Richard Biener <rguenther@suse.de>
4819 * ipa-prop.c: Include stringpool.h and tree-ssanames.h.
4820 (ipa_modify_call_arguments): Emit an argument load explicitely and
4821 preserve virtual SSA form there and for the replacement call.
4822 Do not update SSA form nor free dominance info.
4824 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
4826 * ipa.c (function_and_variable_visibility): Also clear WEAK
4827 flag when disolving COMDAT_GROUP.
4829 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
4831 * ipa-prop.h (ipa_ancestor_jf_data): Update ocmment.
4832 * ipa-prop.c (ipa_set_jf_known_type): Return early when
4834 (ipa_set_ancestor_jf): Set type to NULL hwen it is not preserved;
4835 do more sanity checks.
4836 (detect_type_change): Return true when giving up early.
4837 (compute_complex_assign_jump_func): Fix type parameter of
4838 ipa_set_ancestor_jf.
4839 (compute_complex_ancestor_jump_func): Likewise.
4840 (update_jump_functions_after_inlining): Fix updating of
4842 * ipa-cp.c (ipa_get_jf_ancestor_result): Be ready for type to be NULL.
4844 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
4846 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Also remove
4847 inline clones when edge disappears.
4849 2014-02-18 Michael Meissner <meissner@linux.vnet.ibm.com>
4852 * config/rs6000/rs6000.md (mov<mode>_64bit, TF/TDmode moves):
4853 Split 64-bit moves into 2 patterns. Do not allow the use of
4854 direct move for TDmode in little endian, since the decimal value
4855 has little endian bytes within a word, but the 64-bit pieces are
4856 ordered in a big endian fashion, and normal subreg's of TDmode are
4858 (mov<mode>_64bit_dm): Likewise.
4859 (movtd_64bit_nodm): Likewise.
4861 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
4863 PR tree-optimization/60174
4864 * tree-ssa-reassoc.c (init_range_entry): Do not look into the defining
4865 statement of an SSA_NAME that occurs in an abnormal PHI node.
4867 2014-02-18 Jakub Jelinek <jakub@redhat.com>
4870 * final.c (SEEN_BB): Remove.
4871 (SEEN_NOTE, SEEN_EMITTED): Renumber.
4872 (final_scan_insn): Don't force_source_line on second
4873 NOTE_INSN_BASIC_BLOCK.
4875 2014-02-18 Uros Bizjak <ubizjak@gmail.com>
4878 * config/i386/i386.h (struct ix86_args): Add warn_avx512f.
4879 * config/i386/i386.c (init_cumulative_args): Initialize warn_avx512f.
4880 (type_natural_mode): Warn ABI change when %zmm register is not
4881 available for AVX512F vector value passing.
4883 2014-02-18 Kai Tietz <ktietz@redhat.com>
4886 * config/i386/i386.c (ix86_expand_prologue): Use value in
4887 rax register as displacement when restoring %r10 or %rax.
4888 Fix wrong offset when restoring both registers.
4890 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
4892 * ipa-prop.c (compute_complex_ancestor_jump_func): Replace overzealous
4893 assertion with conditional return.
4895 2014-02-18 Jakub Jelinek <jakub@redhat.com>
4896 Uros Bizjak <ubizjak@gmail.com>
4899 * config/i386/driver-i386.c (host_detect_local_cpu): If
4900 YMM state is not saved by the OS, also clear has_f16c. Move
4901 CPUID 0x80000001 handling before YMM state saving checking.
4903 2014-02-18 Andrey Belevantsev <abel@ispras.ru>
4905 PR rtl-optimization/58960
4906 * haifa-sched.c (alloc_global_sched_pressure_data): New,
4907 factored out from ...
4908 (sched_init): ... here.
4909 (free_global_sched_pressure_data): New, factored out from ...
4910 (sched_finish): ... here.
4911 * sched-int.h (free_global_sched_pressure_data): Declare.
4912 * sched-rgn.c (nr_regions_initial): New static global.
4913 (haifa_find_rgns): Initialize it.
4914 (schedule_region): Disable sched-pressure for the newly
4917 2014-02-17 Richard Biener <rguenther@suse.de>
4919 * tree-vect-stmts.c (free_stmt_vec_info): Clear BB and
4920 release SSA defs of pattern stmts.
4922 2014-02-17 Richard Biener <rguenther@suse.de>
4924 * tree-inline.c (expand_call_inline): Release the virtual
4925 operand defined by the call we are about to inline.
4927 2014-02-17 Richard Biener <rguenther@suse.de>
4929 * tree-ssa.c (verify_ssa): If verify_def found an error, ICE.
4931 2014-02-17 Kirill Yukhin <kirill.yukhin@intel.com>
4932 Ilya Tocar <ilya.tocar@intel.com>
4934 * config/i386/avx512fintrin.h (_mm512_maskz_permutexvar_epi64): Swap
4935 arguments order in builtin.
4936 (_mm512_permutexvar_epi64): Ditto.
4937 (_mm512_mask_permutexvar_epi64): Ditto
4938 (_mm512_maskz_permutexvar_epi32): Ditto
4939 (_mm512_permutexvar_epi32): Ditto
4940 (_mm512_mask_permutexvar_epi32): Ditto
4942 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4944 * config/rs6000/altivec.md (p8_vmrgew): Handle little endian targets.
4945 (p8_vmrgow): Likewise.
4947 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4949 * config/rs6000/vsx.md (vsx_xxpermdi_<mode>): Handle little
4952 2014-02-15 Michael Meissner <meissner@linux.vnet.ibm.com>
4955 * config/rs6000/rs6000.md (rreg): Add TFmode, TDmode constraints.
4956 (mov<mode>_internal, TFmode/TDmode): Split TFmode/TDmode moves
4957 into 64-bit and 32-bit moves. On 64-bit moves, add support for
4958 using direct move instructions on ISA 2.07. Also adjust
4959 instruction length for 64-bit.
4960 (mov<mode>_64bit, TFmode/TDmode): Likewise.
4961 (mov<mode>_32bit, TFmode/TDmode): Likewise.
4963 2014-02-15 Alan Modra <amodra@gmail.com>
4967 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
4968 find_replacement on parts of insn rtl that might be reloaded.
4970 2014-02-15 Richard Biener <rguenther@suse.de>
4972 PR tree-optimization/60183
4973 * tree-ssa-phiprop.c (propagate_with_phi): Avoid speculating loads.
4974 (tree_ssa_phiprop): Calculate and free post-dominators.
4976 2014-02-14 Jeff Law <law@redhat.com>
4978 PR rtl-optimization/60131
4979 * ree.c (get_extended_src_reg): New function.
4980 (combine_reaching_defs): Use it rather than assuming location of REG.
4981 (find_and_remove_re): Verify first operand of extension is
4982 a REG before adding the insns to the copy list.
4984 2014-02-14 Roland McGrath <mcgrathr@google.com>
4986 * configure.ac (HAVE_AS_IX86_UD2): New test for 'ud2' mnemonic.
4987 * configure: Regenerated.
4988 * config.in: Regenerated.
4989 * config/i386/i386.md (trap) [HAVE_AS_IX86_UD2]: Use the mnemonic
4990 instead of ASM_SHORT.
4992 2014-02-14 Vladimir Makarov <vmakarov@redhat.com>
4993 Richard Earnshaw <rearnsha@arm.com>
4995 PR rtl-optimization/59535
4996 * lra-constraints.c (process_alt_operands): Encourage alternative
4997 when unassigned pseudo class is superset of the alternative class.
4998 (inherit_reload_reg): Don't inherit when optimizing for code size.
4999 * config/arm/arm.h (MODE_BASE_REG_CLASS): Add version for LRA
5000 returning CORE_REGS for anything but Thumb1 and BASE_REGS for
5001 modes not less than 4 for Thumb1.
5003 2014-02-14 Kyle McMartin <kyle@redhat.com>
5006 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Define for AArch64.
5008 2014-02-14 Richard Biener <rguenther@suse.de>
5010 * cilk-common.c (cilk_arrow): Build a MEM_REF, not an INDIRECT_REF.
5011 (get_frame_arg): Drop the assert with langhook types_compatible_p.
5012 Do not strip INDIRECT_REFs.
5014 2014-02-14 Richard Biener <rguenther@suse.de>
5017 * lto-streamer-out.c (DFS_write_tree_body): Do not follow
5018 DECL_FUNCTION_SPECIFIC_TARGET.
5019 (hash_tree): Do not hash DECL_FUNCTION_SPECIFIC_TARGET.
5020 * tree-streamer-out.c (pack_ts_target_option): Remove.
5021 (streamer_pack_tree_bitfields): Do not stream TS_TARGET_OPTION.
5022 (write_ts_function_decl_tree_pointers): Do not stream
5023 DECL_FUNCTION_SPECIFIC_TARGET.
5024 * tree-streamer-in.c (unpack_ts_target_option): Remove.
5025 (unpack_value_fields): Do not stream TS_TARGET_OPTION.
5026 (lto_input_ts_function_decl_tree_pointers): Do not stream
5027 DECL_FUNCTION_SPECIFIC_TARGET.
5029 2014-02-14 Jakub Jelinek <jakub@redhat.com>
5031 * tree-vect-loop.c (vect_is_slp_reduction): Don't set use_stmt twice.
5032 (get_initial_def_for_induction, vectorizable_induction): Ignore
5033 debug stmts when looking for exit_phi.
5034 (vectorizable_live_operation): Fix up condition.
5036 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
5038 * config/nds32/nds32.c (nds32_asm_function_prologue): Do not use
5039 nreverse() because it changes the content of original tree list.
5041 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
5043 * config/nds32/t-mlibs (MULTILIB_OPTIONS): Fix typo in comment.
5044 * config/nds32/nds32.c (nds32_merge_decl_attributes): Likewise.
5046 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
5048 * config/nds32/nds32.c (nds32_naked_function_p): Follow the
5049 GNU coding standards.
5051 2014-02-13 Jakub Jelinek <jakub@redhat.com>
5054 * dwarf2out.c (gen_subprogram_die): Don't call
5055 add_calling_convention_attribute if subr_die is old_die.
5057 2014-02-13 Sharad Singhai <singhai@google.com>
5059 * doc/optinfo.texi: Fix order of nodes.
5061 2014-02-13 Uros Bizjak <ubizjak@gmail.com>
5063 * config/i386/sse.md (xop_vmfrcz<mode>2): Generate const0 in
5064 operands[2], not operands[3].
5066 2014-02-13 Richard Biener <rguenther@suse.de>
5069 * doc/install.texi (ISL): Update recommended version to 0.12.2,
5070 mention the possibility of an in-tree build.
5071 (CLooG): Update recommended version to 0.18.1, mention the
5072 possibility of an in-tree build and clarify that the ISL
5073 bundled with CLooG does not work.
5075 2014-02-13 Jakub Jelinek <jakub@redhat.com>
5078 * expr.c (compress_float_constant): If x is a hard register,
5079 extend into a pseudo and then move to x.
5081 2014-02-13 Dominik Vogt <vogt@linux.vnet.ibm.com>
5083 * config/s390/s390.c (s390_asm_output_function_label): Fix crash
5084 caused by bad second argument to warning_at() with -mhotpatch and
5085 nested functions (e.g. with gfortran).
5087 2014-02-13 Richard Sandiford <rdsandiford@googlemail.com>
5089 * opts.c (option_name): Remove "enabled by default" rider.
5091 2014-02-12 John David Anglin <danglin@gcc.gnu.org>
5093 * config/pa/pa.c (pa_option_override): Remove auto increment FIXME.
5095 2014-02-12 H.J. Lu <hongjiu.lu@intel.com>
5096 Uros Bizjak <ubizjak@gmail.com>
5099 * configure.ac (HAVE_AS_GOTOFF_IN_DATA): Pass --32 to GNU assembler.
5100 * configure: Regenerated.
5102 2014-02-12 Richard Biener <rguenther@suse.de>
5104 * vec.c (vec_prefix::calculate_allocation): Move as
5105 inline variant to vec.h.
5106 (vec_prefix::calculate_allocation_1): New out-of-line version.
5107 * vec.h (vec_prefix::calculate_allocation_1): Declare.
5108 (vec_prefix::m_has_auto_buf): Rename to ...
5109 (vec_prefix::m_using_auto_storage): ... this.
5110 (vec_prefix::calculate_allocation): Inline the easy cases
5111 and dispatch to calculate_allocation_1 which doesn't need the
5113 (va_heap::reserve): Use gcc_checking_assert.
5114 (vec<T, A, vl_embed>::embedded_init): Add argument to initialize
5115 m_using_auto_storage.
5116 (auto_vec): Change m_vecpfx member to a vec<T, va_heap, vl_embed>
5118 (vec<T, va_heap, vl_ptr>::reserve): Remove redundant check.
5119 (vec<T, va_heap, vl_ptr>::release): Avoid casting.
5120 (vec<T, va_heap, vl_ptr>::using_auto_storage): Simplify.
5122 2014-02-12 Richard Biener <rguenther@suse.de>
5124 * gcse.c (compute_transp): break from loop over canon_modify_mem_list
5125 when we found a dependence.
5127 2014-02-12 Thomas Schwinge <thomas@codesourcery.com>
5129 * gimplify.c (gimplify_call_expr, gimplify_modify_expr): Move
5131 (maybe_fold_stmt): ... into this new function.
5132 * omp-low.c (lower_omp): Update comment.
5134 * omp-low.c (lower_omp_target): Add clobber for sizes array, after
5137 * omp-low.c (diagnose_sb_0): Make sure label_ctx is valid to
5140 2014-02-12 James Greenhalgh <james.greenhalgh@arm.com>
5142 * config/arm/aarch-cost-tables.h (generic_extra_costs): Fix
5143 identifiers in comments.
5144 (cortexa53_extra_costs): Likewise.
5145 * config/arm/arm.c (cortexa9_extra_costs): Fix identifiers in comments.
5146 (cortexa7_extra_costs): Likewise.
5147 (cortexa12_extra_costs): Likewise.
5148 (cortexa15_extra_costs): Likewise.
5149 (v7m_extra_costs): Likewise.
5151 2014-02-12 Richard Biener <rguenther@suse.de>
5154 * gimple-low.c (lower_builtin_posix_memalign): Lower conditional
5155 of posix_memalign being successful.
5156 (lower_stmt): Restrict lowering of posix_memalign to when
5157 -ftree-bit-ccp is enabled.
5159 2014-02-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
5161 * config/avr/avr-c.c (avr_resolve_overloaded_builtin): Pass vNULL for
5163 * config/spu/spu-c.c (spu_resolve_overloaded_builtin): Likewise.
5165 2014-02-12 Eric Botcazou <ebotcazou@adacore.com>
5167 PR rtl-optimization/60116
5168 * combine.c (try_combine): Also remove dangling REG_DEAD notes on the
5169 other_insn once the combination has been validated.
5171 2014-02-11 Jan Hubicka <hubicka@ucw.cz>
5174 * ipa-utils.h (possible_polymorphic_call_targets): Update prototype
5176 * ipa-devirt.c: Include demangle.h
5177 (odr_violation_reported): New static variable.
5178 (add_type_duplicate): Update odr_violations.
5179 (maybe_record_node): Add completep parameter; update it.
5180 (record_target_from_binfo): Add COMPLETEP parameter;
5181 update it as needed.
5182 (possible_polymorphic_call_targets_1): Likewise.
5183 (struct polymorphic_call_target_d): Add nonconstruction_targets;
5184 rename FINAL to COMPLETE.
5185 (record_targets_from_bases): Sanity check we found the binfo;
5186 fix COMPLETEP updating.
5187 (possible_polymorphic_call_targets): Add NONCONSTRUTION_TARGETSP
5188 parameter, fix computing of COMPLETEP.
5189 (dump_possible_polymorphic_call_targets): Imrove readability of dump;
5190 at LTO time do demangling.
5191 (ipa_devirt): Use nonconstruction_targets; Improve dumps.
5192 * gimple-fold.c (gimple_get_virt_method_for_vtable): Add can_refer
5194 (gimple_get_virt_method_for_binfo): Likewise.
5195 * gimple-fold.h (gimple_get_virt_method_for_binfo,
5196 gimple_get_virt_method_for_vtable): Update prototypes.
5198 2014-02-11 Vladimir Makarov <vmakarov@redhat.com>
5201 * genautomata.c (add_presence_absence): Fix typo with
5202 {final_}presence_list.
5204 2014-02-11 Michael Meissner <meissner@linux.vnet.ibm.com>
5207 * config/rs6000/rs6000.md (128-bit GPR splitter): Add a splitter
5208 for VSX/Altivec vectors that land in GPR registers.
5210 2014-02-11 Richard Henderson <rth@redhat.com>
5211 Jakub Jelinek <jakub@redhat.com>
5214 * tree-sra.c (load_assign_lhs_subreplacements): Add VIEW_CONVERT_EXPR
5215 around drhs if type conversion to lacc->type is not useless.
5217 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5219 * config/aarch64/aarch64-cores.def (cortex-a57): Use cortexa57
5221 (cortex-a57.cortex-a53): Likewise.
5222 * config/aarch64/aarch64.c (cortexa57_tunings): New tuning struct.
5224 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5226 * config/arm/thumb2.md (*thumb2_movhi_insn): Add alternatives for
5229 2014-02-11 Renlin Li <Renlin.Li@arm.com>
5231 * doc/sourcebuild.texi: Document check_effective_target_arm_vfp3_ok and
5232 add_options_for_arm_vfp3.
5234 2014-02-11 Jeff Law <law@redhat.com>
5237 * expr.c (expand_expr_addr_expr_1): Handle expand_expr returning an
5238 object with an undesirable mode.
5240 2014-02-11 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5243 * config/i386/sol2-9.h: New file.
5244 * config.gcc (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*,
5245 *-*-solaris2.9*): Use it.
5247 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
5249 * config/microblaze/microblaze.md: Add movsi4_rev insn pattern.
5250 * config/microblaze/predicates.md: Add reg_or_mem_operand predicate.
5252 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
5254 * config/microblaze/microblaze.c: Extend mcpu version format
5256 2014-02-10 David Holsgrove <david.holsgrove@xilinx.com>
5258 * config/microblaze/microblaze.h: Define SIZE_TYPE and PTRDIFF_TYPE.
5260 2014-02-10 Richard Henderson <rth@redhat.com>
5263 * calls.c (expand_call): Don't double-push for reg_parm_stack_space.
5264 * config/i386/i386.c (init_cumulative_args): Remove sorry for 64-bit
5265 ms-abi vs -mno-accumulate-outgoing-args.
5266 (ix86_expand_prologue): Unconditionally call ix86_eax_live_at_start_p.
5267 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Fix comment with
5270 2014-02-10 Bernd Edlinger <bernd.edlinger@hotmail.de>
5273 * cfgexpand.c (expand_asm_operands): Attach source location to
5274 ASM_INPUT rtx objects.
5275 * print-rtl.c (print_rtx): Check for UNKNOWN_LOCATION.
5277 2014-02-10 Nick Clifton <nickc@redhat.com>
5279 * config/mn10300/mn10300.c (popcount): New function.
5280 (mn10300_expand_prologue): Include saved registers in stack usage
5283 2014-02-10 Jeff Law <law@redhat.com>
5286 * reload1.c (emit_input_reload_insns): Do not create invalid RTL
5287 when changing the SET_DEST of a prior insn to avoid an input reload.
5289 2014-02-10 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
5291 * config/rs6000/sysv4.h (ENDIAN_SELECT): Do not attempt to enforce
5292 big-endian mode for -mcall-aixdesc, -mcall-freebsd, -mcall-netbsd,
5293 -mcall-openbsd, or -mcall-linux.
5294 (CC1_ENDIAN_BIG_SPEC): Remove.
5295 (CC1_ENDIAN_LITTLE_SPEC): Remove.
5296 (CC1_ENDIAN_DEFAULT_SPEC): Remove.
5297 (CC1_SPEC): Remove (always empty) %cc1_endian_... spec.
5298 (SUBTARGET_EXTRA_SPECS): Remove %cc1_endian_big, %cc1_endian_little,
5299 and %cc1_endian_default.
5300 * config/rs6000/sysv4le.h (CC1_ENDIAN_DEFAULT_SPEC): Remove.
5302 2014-02-10 Richard Biener <rguenther@suse.de>
5304 PR tree-optimization/60115
5305 * tree-eh.c (tree_could_trap_p): Unify TARGET_MEM_REF and
5306 MEM_REF handling. Properly verify that the accesses are not
5307 out of the objects bound.
5309 2014-02-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5311 * config/aarch64/aarch64.c (aarch64_override_options): Fix typo from
5314 2014-02-10 Eric Botcazou <ebotcazou@adacore.com>
5316 * ipa-devirt.c (get_polymorphic_call_info_from_invariant): Return
5317 proper constants and fix formatting.
5318 (possible_polymorphic_call_targets): Fix formatting.
5320 2014-02-10 Kirill Yukhin <kirill.yukhin@intel.com>
5321 Ilya Tocar <ilya.tocar@intel.com>
5323 * config/i386/avx512fintrin.h (_mm512_storeu_epi64): Removed.
5324 (_mm512_loadu_epi32): Renamed into...
5325 (_mm512_loadu_si512): This.
5326 (_mm512_storeu_epi32): Renamed into...
5327 (_mm512_storeu_si512): This.
5328 (_mm512_maskz_ceil_ps): Removed.
5329 (_mm512_maskz_ceil_pd): Ditto.
5330 (_mm512_maskz_floor_ps): Ditto.
5331 (_mm512_maskz_floor_pd): Ditto.
5332 (_mm512_floor_round_ps): Ditto.
5333 (_mm512_floor_round_pd): Ditto.
5334 (_mm512_ceil_round_ps): Ditto.
5335 (_mm512_ceil_round_pd): Ditto.
5336 (_mm512_mask_floor_round_ps): Ditto.
5337 (_mm512_mask_floor_round_pd): Ditto.
5338 (_mm512_mask_ceil_round_ps): Ditto.
5339 (_mm512_mask_ceil_round_pd): Ditto.
5340 (_mm512_maskz_floor_round_ps): Ditto.
5341 (_mm512_maskz_floor_round_pd): Ditto.
5342 (_mm512_maskz_ceil_round_ps): Ditto.
5343 (_mm512_maskz_ceil_round_pd): Ditto.
5344 (_mm512_expand_pd): Ditto.
5345 (_mm512_expand_ps): Ditto.
5346 * config/i386/i386.c (ix86_builtins): Remove
5347 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK.
5348 (bdesc_args): Ditto.
5349 * config/i386/predicates.md (const1256_operand): New.
5350 (const_1_to_2_operand): Ditto.
5351 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
5352 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
5353 (*avx512pf_gatherpf<mode>sf): Ditto.
5354 (avx512pf_gatherpf<mode>df): Ditto.
5355 (*avx512pf_gatherpf<mode>df_mask): Ditto.
5356 (*avx512pf_gatherpf<mode>df): Ditto.
5357 (avx512pf_scatterpf<mode>sf): Ditto.
5358 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
5359 (*avx512pf_scatterpf<mode>sf): Ditto.
5360 (avx512pf_scatterpf<mode>df): Ditto.
5361 (*avx512pf_scatterpf<mode>df_mask): Ditto.
5362 (*avx512pf_scatterpf<mode>df): Ditto.
5363 (avx512f_expand<mode>): Removed.
5364 (<shift_insn><mode>3<mask_name>): Change predicate type.
5366 2014-02-08 Jakub Jelinek <jakub@redhat.com>
5368 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers
5369 not at the end of datarefs vector use ordered_remove to avoid
5370 reordering datarefs vector.
5373 * gimplify.c (gimplify_bind_expr): In ORT_SIMD region
5374 mark local addressable non-static vars as GOVD_PRIVATE
5375 instead of GOVD_LOCAL.
5376 * omp-low.c (lower_omp_for): Move gimple_bind_vars
5377 and BLOCK_VARS of gimple_bind_block to new_stmt rather
5381 * tree-ssa-ccp.c (surely_varying_stmt_p): Don't return true
5382 if TYPE_ATTRIBUTES (gimple_call_fntype ()) contain
5383 assume_aligned or alloc_align attributes.
5384 (bit_value_assume_aligned): Add ATTR, PTRVAL and ALLOC_ALIGN
5385 arguments. Handle also assume_aligned and alloc_align attributes.
5386 (evaluate_stmt): Adjust bit_value_assume_aligned caller. Handle
5387 calls to functions with assume_aligned or alloc_align attributes.
5388 * doc/extend.texi: Document assume_aligned and alloc_align attributes.
5390 2014-02-08 Terry Guo <terry.guo@arm.com>
5392 * doc/invoke.texi: Document ARM -march=armv7e-m.
5394 2014-02-08 Jakub Jelinek <jakub@redhat.com>
5396 * cilk-common.c (cilk_init_builtins): Clear TREE_NOTHROW
5397 flag on __cilkrts_rethrow builtin.
5400 * ipa-cp.c (determine_versionability): Fail at -O0
5401 or __attribute__((optimize (0))) or -fno-ipa-cp functions.
5402 * tree-sra.c (ipa_sra_preliminary_function_checks): Similarly.
5405 2014-02-04 Jakub Jelinek <jakub@redhat.com>
5408 * tree-inline.c (copy_forbidden): Fail for
5409 __attribute__((optimize (0))) functions.
5411 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
5413 * varpool.c: Include pointer-set.h.
5414 (varpool_remove_unreferenced_decls): Variables in other partitions
5415 will not be output; be however careful to not lose information
5418 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
5420 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do O(1)
5421 lookup in the vtable constructor.
5423 2014-02-07 Jeff Law <law@redhat.com>
5426 * config/m68k/m68k.md (ashldi_extsi): Turn into a
5427 define_insn_and_split.
5429 * ipa-inline.c (inline_small_functions): Fix typos.
5431 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5433 * config/s390/s390-protos.h (s390_can_use_simple_return_insn)
5434 (s390_can_use_return_insn): Declare.
5435 * config/s390/s390.h (EPILOGUE_USES): Define.
5436 * config/s390/s390.c (s390_mainpool_start): Allow two main_pool
5438 (s390_chunkify_start): Handle return JUMP_LABELs.
5439 (s390_early_mach): Emit a main_pool instruction on the entry edge.
5440 (s300_set_up_by_prologue, s390_can_use_simple_return_insn)
5441 (s390_can_use_return_insn): New functions.
5442 (s390_fix_long_loop_prediction): Handle conditional returns.
5443 (TARGET_SET_UP_BY_PROLOGUE): Define.
5444 * config/s390/s390.md (ANY_RETURN): New code iterator.
5445 (*creturn, *csimple_return, return, simple_return): New patterns.
5447 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5449 * config/s390/s390.c (s390_restore_gprs_from_fprs): Add REG_CFA_RESTORE
5450 notes to each restore. Also add REG_CFA_DEF_CFA when restoring %r15.
5451 (s390_optimize_prologue): Don't clear RTX_FRAME_RELATED_P. Update the
5452 REG_CFA_RESTORE list when deciding not to restore a register.
5454 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5456 * config/s390/s390.c: Include tree-pass.h and context.h.
5457 (s390_early_mach): New function, split out from...
5458 (s390_emit_prologue): ...here.
5459 (pass_data_s390_early_mach): New pass structure.
5460 (pass_s390_early_mach): New class.
5461 (s390_option_override): Create and register early_mach pass.
5462 Move to end of file.
5464 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5466 * var-tracking.c (vt_stack_adjustments): Don't require stack_adjusts
5467 to match for the exit block.
5469 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5471 * config/s390/s390.md ("atomic_load<mode>", "atomic_store<mode>")
5472 ("atomic_compare_and_swap<mode>", "atomic_fetch_<atomic><mode>"):
5473 Reject misaligned operands.
5475 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5477 * optabs.c (expand_atomic_compare_and_swap): Allow expander to fail.
5479 2014-02-07 Richard Biener <rguenther@suse.de>
5482 * gimple-low.c (lower_builtin_posix_memalign): New function.
5483 (lower_stmt): Call it to lower posix_memalign in a way
5484 to make alignment info accessible.
5486 2014-02-07 Jakub Jelinek <jakub@redhat.com>
5489 * tree.c (build_common_builtin_nodes): Set ECF_LEAF for
5490 __builtin_setjmp_receiver.
5492 2014-02-07 Richard Biener <rguenther@suse.de>
5495 * builtin-types.def (BT_FN_INT_PTRPTR_SIZE_SIZE): Add.
5496 * builtins.def (BUILT_IN_POSIX_MEMALIGN): Likewise.
5497 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
5498 Handle BUILT_IN_POSIX_MEMALIGN.
5499 (find_func_clobbers): Likewise.
5500 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
5501 (call_may_clobber_ref_p_1): Likewise.
5503 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5506 * ipa-devirt.c (record_target_from_binfo): Remove overactive
5509 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5512 * lto-cgraph.c (lto_output_node): Use
5513 symtab_get_symbol_partitioning_class.
5514 (lto_output_varpool_node): likewise.
5515 (symtab_get_symbol_partitioning_class): Move here from
5517 * cgraph.h (symbol_partitioning_class): Likewise.
5518 (symtab_get_symbol_partitioning_class): Declare.
5520 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5522 * ggc.h (ggc_internal_cleared_alloc): New macro.
5523 * vec.h (vec_safe_copy): Handle memory stats.
5524 * omp-low.c (simd_clone_struct_alloc): Use ggc_internal_cleared_alloc.
5525 * target-globals.c (save_target_globals): Likewise.
5527 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5530 * expr.c (emit_move_resolve_push): Export; be bit more selective
5531 on when to clear alias set.
5532 * expr.h (emit_move_resolve_push): Declare.
5533 * function.h (struct function): Add tail_call_marked.
5534 * tree-tailcall.c (optimize_tail_call): Set tail_call_marked.
5535 * config/i386/i386-protos.h (ix86_expand_push): Remove.
5536 * config/i386/i386.md (TImode move expander): De not call
5538 (FP push expanders): Preserve memory attributes.
5539 * config/i386/sse.md (push<mode>1): Remove.
5540 * config/i386/i386.c (ix86_expand_vector_move): Handle push operation.
5541 (ix86_expand_push): Remove.
5542 * config/i386/mmx.md (push<mode>1): Remove.
5544 2014-02-06 Jakub Jelinek <jakub@redhat.com>
5546 PR rtl-optimization/60030
5547 * internal-fn.c (ubsan_expand_si_overflow_mul_check): Surround
5548 lopart with paradoxical subreg before shifting it up by hprec.
5550 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5552 * config/arm/aarch-cost-tables.h (cortexa57_extra_costs): New table.
5553 Remove extra newline at end of file.
5554 * config/arm/arm.c (arm_cortex_a57_tune): New tuning struct.
5555 (arm_issue_rate): Handle cortexa57.
5556 * config/arm/arm-cores.def (cortex-a57): Use cortex_a57 tuning.
5557 (cortex-a57.cortex-a53): Likewise.
5559 2014-02-06 Jakub Jelinek <jakub@redhat.com>
5562 * config/arm/arm.c (emit_multi_reg_push): Add dwarf_regs_mask argument,
5563 don't record in REG_FRAME_RELATED_EXPR registers not set in that
5565 (arm_expand_prologue): Adjust all callers.
5566 (arm_unwind_emit_sequence): Allow saved, but not important for unwind
5567 info, registers also at the lowest numbered registers side. Use
5568 gcc_assert instead of abort, and SET_SRC/SET_DEST macros instead of
5572 * var-tracking.c (adjust_mems): Before adding a SET to
5573 amd->side_effects, adjust it's SET_SRC using simplify_replace_fn_rtx.
5575 2014-02-06 Alan Modra <amodra@gmail.com>
5578 * config/rs6000/rs6000.c (rs6000_secondary_memory_needed_mode): Only
5579 change SDmode to DDmode when lra_in_progress.
5581 2014-02-06 Jakub Jelinek <jakub@redhat.com>
5584 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers, call
5585 free_data_ref on the dr first, and before goto again also set dr
5586 to the next dr. For simd_lane_access, free old datarefs[i] before
5587 overwriting it. For get_vectype_for_scalar_type failure, don't
5588 free_data_ref if simd_lane_access.
5590 * Makefile.in (prefix.o, cppbuiltin.o): Depend on $(BASEVER).
5593 * tree.h (opts_for_fn): New inline function.
5594 (opt_for_fn): Define.
5595 * config/i386/i386.c (ix86_function_regparm): Use
5596 opt_for_fn (decl, optimize) instead of optimize.
5598 2014-02-06 Marcus Shawcroft <marcus.shawcroft@arm.com>
5600 * config/aarch64/aarch64.c (aarch64_classify_symbol): Fix logic
5601 for SYMBOL_REF in large memory model.
5603 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5605 * config/aarch64/aarch64-cores.def (cortex-a53): Specify CRC32
5607 (cortex-a57): Likewise.
5608 (cortex-a57.cortex-a53): Likewise.
5610 2014-02-06 Yury Gribov <y.gribov@samsung.com>
5611 Kugan Vivekanandarajah <kuganv@linaro.org>
5613 * config/arm/arm.c (arm_vector_alignment_reachable): Check
5615 * config/arm/arm.c (arm_builtin_support_vector_misalignment): Likewise.
5617 2014-02-06 Richard Biener <rguenther@suse.de>
5619 * tree-cfg.c (gimple_duplicate_sese_region): Fix ordering of
5620 set_loop_copy and initialize_original_copy_tables.
5622 2014-02-06 Alex Velenko <Alex.Velenko@arm.com>
5624 * config/aarch64/aarch64-simd.md
5625 (aarch64_ashr_simddi): Change QI to SI.
5627 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
5628 Jakub Jelinek <jakub@redhat.com>
5631 * ipa-inline-analysis.c (compute_bb_predicates): Ensure monotonicity
5634 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5636 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Change
5637 CODE_FOR_altivec_vpku[hw]um to
5638 CODE_FOR_altivec_vpku[hw]um_direct.
5639 * config/rs6000/altivec.md (vec_unpacks_hi_<VP_small_lc>): Change
5640 UNSPEC_VUNPACK_HI_SIGN to UNSPEC_VUNPACK_HI_SIGN_DIRECT.
5641 (vec_unpacks_lo_<VP_small_lc>): Change UNSPEC_VUNPACK_LO_SIGN to
5642 UNSPEC_VUNPACK_LO_SIGN_DIRECT.
5644 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5646 * config/rs6000/altivec.md (altivec_vsum2sws): Adjust code
5647 generation for -maltivec=be.
5648 (altivec_vsumsws): Simplify redundant test.
5650 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5652 * altivec.md (UNSPEC_VPACK_UNS_UNS_MOD_DIRECT): New unspec.
5653 (UNSPEC_VUNPACK_HI_SIGN_DIRECT): Likewise.
5654 (UNSPEC_VUNPACK_LO_SIGN_DIRECT): Likewise.
5655 (mulv8hi3): Use gen_altivec_vpkuwum_direct instead of
5656 gen_altivec_vpkuwum.
5657 (altivec_vpkpx): Test for VECTOR_ELT_ORDER_BIG instead of for
5659 (altivec_vpks<VI_char>ss): Likewise.
5660 (altivec_vpks<VI_char>us): Likewise.
5661 (altivec_vpku<VI_char>us): Likewise.
5662 (altivec_vpku<VI_char>um): Likewise.
5663 (altivec_vpku<VI_char>um_direct): New (copy of
5664 altivec_vpku<VI_char>um that still relies on BYTES_BIG_ENDIAN, for
5666 (altivec_vupkhs<VU_char>): Emit vupkls* instead of vupkhs* when
5667 target is little endian and -maltivec=be is not specified.
5668 (*altivec_vupkhs<VU_char>_direct): New (copy of
5669 altivec_vupkhs<VU_char> that always emits vupkhs*, for internal use).
5670 (altivec_vupkls<VU_char>): Emit vupkhs* instead of vupkls* when
5671 target is little endian and -maltivec=be is not specified.
5672 (*altivec_vupkls<VU_char>_direct): New (copy of
5673 altivec_vupkls<VU_char> that always emits vupkls*, for internal use).
5674 (altivec_vupkhpx): Emit vupklpx instead of vupkhpx when target is
5675 little endian and -maltivec=be is not specified.
5676 (altivec_vupklpx): Emit vupkhpx instead of vupklpx when target is
5677 little endian and -maltivec=be is not specified.
5679 2014-02-05 Richard Henderson <rth@redhat.com>
5682 * combine-stack-adj.c: Revert r206943.
5683 * sched-int.h (struct deps_desc): Add last_args_size.
5684 * sched-deps.c (init_deps): Initialize it.
5685 (sched_analyze_insn): Add OUTPUT dependencies between insns that
5686 contain REG_ARGS_SIZE notes.
5688 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
5690 * lto-cgraph.c (asm_nodes_output): Make global.
5691 * lto-wrapper.c (run_gcc): Pass down paralelizm to WPA.
5692 * gcc.c (AS_NEEDS_DASH_FOR_PIPED_INPUT): Allow WPA parameter
5693 (driver_handle_option): Handle OPT_fwpa.
5695 2014-02-05 Jakub Jelinek <jakub@redhat.com>
5698 * ipa-devirt.c (possible_polymorphic_call_targets): Fix
5699 a comment typo and formatting issue. If odr_hash hasn't been
5700 created, return vNULL and set *completep to false.
5703 * tree-eh.c (cleanup_empty_eh): Bail out on totally empty
5704 bb with no successors.
5706 2014-02-05 James Greenhalgh <james.greenhalgh@arm.com>
5709 * doc/invoke.texi (-march): Clarify documentation for ARM.
5713 2014-02-05 Richard Biener <rguenther@suse.de>
5715 * tree-vect-loop.c (vect_analyze_loop_2): Be more informative
5716 when not vectorizing because of too many alias checks.
5717 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
5718 Add more verboseness, avoid duplicate MSG_MISSED_OPTIMIZATION.
5720 2014-02-05 Nick Clifton <nickc@redhat.com>
5722 * config/mn10300/mn10300.c (mn10300_hard_regno_mode_ok): Do not
5723 accept extended registers in any mode when compiling for the MN10300.
5725 2014-02-05 Yury Gribov <y.gribov@samsung.com>
5727 * cif-code.def (ATTRIBUTE_MISMATCH): New CIF code.
5728 * ipa-inline.c (report_inline_failed_reason): Handle mismatched
5729 sanitization attributes.
5730 (can_inline_edge_p): Likewise.
5731 (sanitize_attrs_match_for_inline_p): New function.
5733 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
5735 * ipa-prop.c (detect_type_change): Shor circuit testing of
5736 type changes on THIS pointer.
5738 2014-02-04 John David Anglin <danglin@gcc.gnu.org>
5741 * config/pa/pa.c (legitimize_tls_address): Return original address
5742 if not passed a SYMBOL_REF rtx.
5743 (hppa_legitimize_address): Call legitimize_tls_address for all TLS
5745 (pa_emit_move_sequence): Simplify TLS source operands.
5746 (pa_legitimate_constant_p): Reject all TLS constants.
5747 * config/pa/pa.h (PA_SYMBOL_REF_TLS_P): Correct comment.
5748 (CONSTANT_ADDRESS_P): Reject TLS CONST addresses.
5750 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
5752 * ipa.c (function_and_variable_visibility): Decompose DECL_ONE_ONLY
5753 groups when we know they are controlled by LTO.
5754 * varasm.c (default_binds_local_p_1): If object is in other partition,
5755 it will be resolved locally.
5757 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
5759 * config/host-linux.c (linux_gt_pch_use_address): Don't
5760 use SSIZE_MAX because it is not always defined.
5762 2014-02-04 Vladimir Makarov <vmakarov@redhat.com>
5765 * lra-constraints.c (need_for_split_p): Use more 3 reloads as
5766 threshold for pseudo splitting.
5767 (update_ebb_live_info): Process call argument hard registers and
5768 hard registers from insn definition too.
5769 (max_small_class_regs_num): New constant.
5770 (inherit_in_ebb): Update live hard regs through EBBs. Update
5771 reloads_num only for small register classes. Don't split for
5774 2014-02-04 Markus Trippelsdorf <markus@trippelsdorf.de>
5777 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that target
5780 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
5782 * gimple-fold.c (can_refer_decl_in_current_unit_p): Default
5785 2014-02-04 Marek Polacek <polacek@redhat.com>
5787 * gdbinit.in (pel): Define.
5789 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
5791 * doc/invoke.texi (fstrict-volatile-bitfields): Clarify current
5794 2014-02-04 Richard Biener <rguenther@suse.de>
5797 * lto-streamer-out.c (tree_is_indexable): Force NAMELIST_DECLs
5798 in function context local.
5799 (lto_output_tree_ref): Do not write trees from lto_output_tree_ref.
5800 * lto-streamer-in.c (lto_input_tree_ref): Handle LTO_namelist_decl_ref
5801 similar to LTO_imported_decl_ref.
5803 2014-02-04 Jakub Jelinek <jakub@redhat.com>
5805 PR tree-optimization/60002
5806 * cgraphclones.c (build_function_decl_skip_args): Clear
5809 PR tree-optimization/60023
5810 * tree-if-conv.c (predicate_mem_writes): Pass true instead of
5811 false to gsi_replace.
5812 * tree-vect-stmts.c (vect_finish_stmt_generation): If stmt
5813 has been in some EH region and vec_stmt could throw, add
5814 vec_stmt into the same EH region.
5815 * tree-data-ref.c (get_references_in_stmt): If IFN_MASK_LOAD
5816 has no lhs, ignore it.
5817 * internal-fn.c (expand_MASK_LOAD): Likewise.
5820 * tree-inline.c (copy_forbidden): Fail for
5821 __attribute__((optimize (0))) functions.
5824 * omp-low.c (simd_clone_struct_copy): If from->inbranch
5825 is set, copy one less argument.
5826 (expand_simd_clones): Don't subtract clone_info->inbranch
5827 from simd_clone_struct_alloc argument.
5829 PR rtl-optimization/57915
5830 * recog.c (simplify_while_replacing): If all unary/binary/relational
5831 operation arguments are constant, attempt to simplify those.
5834 * expmed.c (expand_mult): For MODE_VECTOR_INT multiplication
5835 if there is no vashl<mode>3 or ashl<mode>3 insn, skip_synth.
5837 2014-02-04 Richard Biener <rguenther@suse.de>
5839 PR tree-optimization/60012
5840 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Apply
5841 TBAA disambiguation to all DDRs.
5843 2014-02-04 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5846 * config/sol2.h (LINK_LIBGCC_MAPFILE_SPEC): Define.
5847 (LINK_SPEC): Use it for -shared, -shared-libgcc.
5849 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
5852 * tree.c (get_binfo_at_offset): Do not get confused by empty classes;
5854 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
5856 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Remove.
5857 * gimple-fold.h (gimple_extract_devirt_binfo_from_cst): Remove.
5859 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
5862 * ipa-cp.c (ipa_get_indirect_edge_target_1): Use ipa-devirt
5863 to figure out targets of polymorphic calls with known decl.
5864 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
5865 * ipa-utils.h (get_polymorphic_call_info_from_invariant): Declare.
5866 * ipa-devirt.c (get_polymorphic_call_info_for_decl): Break out from ...
5867 (get_polymorphic_call_info): ... here.
5868 (get_polymorphic_call_info_from_invariant): New function.
5870 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
5872 * ipa-cp.c (ipa_get_indirect_edge_target_1): Do direct
5873 lookup via vtable pointer; check for type consistency
5874 and turn inconsitent facts into UNREACHABLE.
5875 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
5876 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do not ICE on
5877 type inconsistent querries; return UNREACHABLE instead.
5879 2014-02-03 Richard Henderson <rth@twiddle.net>
5882 * tree-ssa-uninit.c (push_to_worklist): Don't re-push if we've
5883 already processed this node.
5884 (normalize_one_pred_1): Pass along mark_set.
5885 (normalize_one_pred): Create and destroy a pointer_set_t.
5886 (normalize_one_pred_chain): Likewise.
5888 2014-02-03 Laurent Aflonsi <laurent.alfonsi@st.com>
5890 PR gcov-profile/58602
5891 * gcc/gcov-io.c (gcov_open): Open with truncation when mode < 0.
5893 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
5896 * ipa-cp.c (ipa_get_indirect_edge_target_1): Give up on
5897 -fno-devirtualize; try to devirtualize by the knowledge of
5898 virtual table pointer given by aggregate propagation.
5899 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
5900 (ipa_print_node_jump_functions): Dump also offset that
5901 is relevant for polymorphic calls.
5902 (determine_known_aggregate_parts): Add arg_type parameter; use it
5903 instead of determining the type from pointer type.
5904 (ipa_compute_jump_functions_for_edge): Update call of
5905 determine_known_aggregate_parts.
5906 * gimple-fold.c (gimple_get_virt_method_for_vtable): Break out from ...
5907 (gimple_get_virt_method_for_binfo): ... here; simplify using
5908 vtable_pointer_value_to_vtable.
5909 * gimple-fold.h (gimple_get_virt_method_for_vtable): Declare.
5910 * ipa-devirt.c (subbinfo_with_vtable_at_offset): Turn OFFSET parameter
5911 to unsigned HOST_WIDE_INT; use vtable_pointer_value_to_vtable.
5912 (vtable_pointer_value_to_vtable): Break out from ...; handle also
5914 (vtable_pointer_value_to_binfo): ... here.
5915 * ipa-utils.h (vtable_pointer_value_to_vtable): Declare.
5917 2014-02-03 Teresa Johnson <tejohnson@google.com>
5919 * tree-vect-slp.c (vect_supported_load_permutation_p): Avoid
5920 redef of outer loop index variable.
5922 2014-02-03 Marc Glisse <marc.glisse@inria.fr>
5926 * doc/extend.texi (Function Attributes): Typo.
5928 2014-02-03 Cong Hou <congh@google.com>
5930 PR tree-optimization/60000
5931 * tree-vect-loop.c (vect_transform_loop): Set pattern_def_seq to NULL
5932 if the vectorized statement is a store. A store statement can only
5933 appear at the end of pattern statements.
5935 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
5937 * config/i386/i386.c (flag_opts): Add -mlong-double-128.
5938 (ix86_option_override_internal): Default long double to 64-bit for
5939 32-bit Bionic and to 128-bit for 64-bit Bionic.
5941 * config/i386/i386.h (LONG_DOUBLE_TYPE_SIZE): Use 128 if
5942 TARGET_LONG_DOUBLE_128 is true.
5943 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Likewise.
5945 * config/i386/i386.opt (mlong-double-80): Negate -mlong-double-64.
5946 (mlong-double-64): Negate -mlong-double-128.
5947 (mlong-double-128): New option.
5949 * config/i386/i386-c.c (ix86_target_macros): Define
5950 __LONG_DOUBLE_128__ for TARGET_LONG_DOUBLE_128.
5952 * doc/invoke.texi: Document -mlong-double-128.
5954 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
5956 PR rtl-optimization/60024
5957 * sel-sched.c (init_regs_for_mode): Check if mode is OK first.
5959 2014-02-03 Markus Trippelsdorf <markus@trippelsdorf.de>
5961 * doc/invoke.texi (fprofile-reorder-functions): Fix typo.
5963 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
5965 PR rtl-optimization/57662
5966 * sel-sched.c (code_motion_path_driver): Do not mark already not
5967 existing blocks in the visiting bitmap.
5969 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
5971 * sel-sched-ir.c (sel_gen_insn_from_expr_after): Reset INSN_DELETED_P
5972 on the insn being emitted.
5974 2014-02-03 James Greenhalgh <james.greenhalgh@arm.com>
5975 Will Deacon <will.deacon@arm.com>
5977 * doc/gimple.texi (gimple_asm_clear_volatile): Remove.
5979 2014-02-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5981 * config/arm/arm-tables.opt: Regenerate.
5983 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5985 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): Generalize
5986 for vector types other than V16QImode.
5987 * config/rs6000/altivec.md (altivec_vperm_<mode>): Change to a
5988 define_expand, and call altivec_expand_vec_perm_le when producing
5989 code with little endian element order.
5990 (*altivec_vperm_<mode>_internal): New insn having previous
5991 behavior of altivec_vperm_<mode>.
5992 (altivec_vperm_<mode>_uns): Change to a define_expand, and call
5993 altivec_expand_vec_perm_le when producing code with little endian
5995 (*altivec_vperm_<mode>_uns_internal): New insn having previous
5996 behavior of altivec_vperm_<mode>_uns.
5998 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6000 * config/rs6000/altivec.md (UNSPEC_VSUMSWS_DIRECT): New unspec.
6001 (altivec_vsumsws): Add handling for -maltivec=be with a little
6003 (altivec_vsumsws_direct): New.
6004 (reduc_splus_<mode>): Call gen_altivec_vsumsws_direct instead of
6005 gen_altivec_vsumsws.
6007 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
6009 * ipa-devirt.c (subbinfo_with_vtable_at_offset,
6010 vtable_pointer_value_to_binfo): New functions.
6011 * ipa-utils.h (vtable_pointer_value_to_binfo): Declare.
6012 * ipa-prop.c (extr_type_from_vtbl_ptr_store): Use it.
6014 2014-02-02 Sandra Loosemore <sandra@codesourcery.com>
6016 * config/nios2/nios2.md (load_got_register): Initialize GOT
6017 pointer from _gp_got instead of _GLOBAL_OFFSET_TABLE_.
6018 * config/nios2/nios2.c (nios2_function_profiler): Likewise.
6020 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
6022 * ipa-prop.c (update_jump_functions_after_inlining): When type is not
6023 preserverd by passthrough, do not propagate the type.
6025 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
6027 * config/mips/mips.c (MIPS_GET_FCSR, MIPS_SET_FCSR): New macros.
6028 (mips_atomic_assign_expand_fenv): New function.
6029 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): Define.
6031 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
6033 * doc/extend.texi (__builtin_mips_get_fcsr): Document.
6034 (__builtin_mips_set_fcsr): Likewise.
6035 * config/mips/mips-ftypes.def: Add MIPS_VOID_FTYPE_USI and
6036 MIPS_USI_FTYPE_VOID.
6037 * config/mips/mips-protos.h (mips16_expand_get_fcsr): Declare
6038 (mips16_expand_set_fcsr): Likewise.
6039 * config/mips/mips.c (mips16_get_fcsr_stub): New variable.
6040 (mips16_set_fcsr_stub): Likewise.
6041 (mips16_get_fcsr_one_only_stub): New class.
6042 (mips16_set_fcsr_one_only_stub): Likewise.
6043 (mips16_expand_get_fcsr, mips16_expand_set_fcsr): New functions.
6044 (mips_code_end): Output the get_fcsr and set_fcsr stubs, if needed.
6045 (BUILTIN_AVAIL_MIPS16, AVAIL_ALL): New macros.
6046 (hard_float): New availability predicate.
6047 (mips_builtins): Add get_fcsr and set_fcsr.
6048 (mips_expand_builtin): Check BUILTIN_AVAIL_MIPS16.
6049 * config/mips/mips.md (UNSPEC_GET_FCSR, UNSPEC_SET_FCSR): New unspecs.
6050 (GET_FCSR_REGNUM, SET_FCSR_REGNUM): New constants.
6051 (mips_get_fcsr, *mips_get_fcsr, mips_get_fcsr_mips16_<mode>)
6052 (mips_set_fcsr, *mips_set_fcsr, mips_set_fcsr_mips16_<mode>): New
6055 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
6057 * config/mips/mips.c (mips_one_only_stub): New class.
6058 (mips_need_mips16_rdhwr_p): Replace with...
6059 (mips16_rdhwr_stub): ...this new variable.
6060 (mips16_stub_call_address): New function.
6061 (mips16_rdhwr_one_only_stub): New class.
6062 (mips_expand_thread_pointer): Use mips16_stub_call_address.
6063 (mips_output_mips16_rdhwr): Delete.
6064 (mips_finish_stub): New function.
6065 (mips_code_end): Use it to handle rdhwr stubs.
6067 2014-02-02 Uros Bizjak <ubizjak@gmail.com>
6070 * config/i386/i386.c (classify_argument): Fix handling of bit_offset
6071 when calculating size of integer atomic types.
6073 2014-02-02 H.J. Lu <hongjiu.lu@intel.com>
6075 * ipa-inline-analysis.c (true_predicate_p): Fix a typo in comments.
6077 2014-02-01 Jakub Jelinek <jakub@redhat.com>
6079 PR tree-optimization/60003
6080 * gimple-low.c (lower_builtin_setjmp): Set cfun->has_nonlocal_label.
6081 * profile.c (branch_prob): Use gimple_call_builtin_p
6082 to check for BUILT_IN_SETJMP_RECEIVER.
6083 * tree-inline.c (copy_bb): Call notice_special_calls.
6085 2014-01-31 Vladimir Makarov <vmakarov@redhat.com>
6088 * lra-constraints.c (process_alt_operands): Update reload_sum only
6091 2014-01-31 Richard Henderson <rth@redhat.com>
6094 * tree-eh.c (lower_try_finally_switch): Delay lowering finally block
6095 until after else_eh is processed.
6097 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
6099 * config/i386/avx512fintrin.h (_MM_FROUND_TO_NEAREST_INT),
6100 (_MM_FROUND_TO_NEG_INF), (_MM_FROUND_TO_POS_INF),
6101 (_MM_FROUND_TO_ZERO), (_MM_FROUND_CUR_DIRECTION): Are already defined
6102 in smmintrin.h, remove them.
6103 (_MM_FROUND_NO_EXC): Same as above, bit also wrong value.
6104 * config/i386/i386.c (ix86_print_operand): Split sae and rounding.
6105 * config/i386/i386.md (ROUND_SAE): Fix value.
6106 * config/i386/predicates.md (const_4_or_8_to_11_operand): New.
6107 (const48_operand): New.
6108 * config/i386/subst.md (round), (round_expand): Use
6109 const_4_or_8_to_11_operand.
6110 (round_saeonly), (round_saeonly_expand): Use const48_operand.
6112 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
6114 * config/i386/constraints.md (Yk): Swap meaning with k.
6115 * config/i386/i386.md (movhi_internal): Change Yk to k.
6116 (movqi_internal): Ditto.
6117 (*k<logic><mode>): Ditto.
6120 (kandn<mode>): Ditto.
6121 (*<code>hi_1): Ditto.
6122 (*<code>qi_1): Ditto.
6123 (kxnor<mode>): Ditto.
6124 (kortestzhi): Ditto.
6125 (kortestchi): Ditto.
6127 (*one_cmplhi2_1): Ditto.
6128 (*one_cmplqi2_1): Ditto.
6129 * config/i386/sse.md (): Change k to Yk.
6130 (avx512f_load<mode>_mask): Ditto.
6131 (avx512f_blendm<mode>): Ditto.
6132 (avx512f_store<mode>_mask): Ditto.
6133 (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
6134 (avx512f_storedqu<mode>_mask): Ditto.
6135 (avx512f_cmp<mode>3<mask_scalar_merge_name><round_saeonly_name>):
6137 (avx512f_ucmp<mode>3<mask_scalar_merge_name>): Ditto.
6138 (avx512f_vmcmp<mode>3<round_saeonly_name>): Ditto.
6139 (avx512f_vmcmp<mode>3_mask<round_saeonly_name>): Ditto.
6140 (avx512f_maskcmp<mode>3): Ditto.
6141 (avx512f_fmadd_<mode>_mask<round_name>): Ditto.
6142 (avx512f_fmadd_<mode>_mask3<round_name>): Ditto.
6143 (avx512f_fmsub_<mode>_mask<round_name>): Ditto.
6144 (avx512f_fmsub_<mode>_mask3<round_name>): Ditto.
6145 (avx512f_fnmadd_<mode>_mask<round_name>): Ditto.
6146 (avx512f_fnmadd_<mode>_mask3<round_name>): Ditto.
6147 (avx512f_fnmsub_<mode>_mask<round_name>): Ditto.
6148 (avx512f_fnmsub_<mode>_mask3<round_name>): Ditto.
6149 (avx512f_fmaddsub_<mode>_mask<round_name>): Ditto.
6150 (avx512f_fmaddsub_<mode>_mask3<round_name>): Ditto.
6151 (avx512f_fmsubadd_<mode>_mask<round_name>): Ditto.
6152 (avx512f_fmsubadd_<mode>_mask3<round_name>): Ditto.
6153 (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
6154 (vec_extract_lo_<mode>_maskm): Ditto.
6155 (vec_extract_hi_<mode>_maskm): Ditto.
6156 (avx512f_vternlog<mode>_mask): Ditto.
6157 (avx512f_fixupimm<mode>_mask<round_saeonly_name>): Ditto.
6158 (avx512f_sfixupimm<mode>_mask<round_saeonly_name>): Ditto.
6159 (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
6160 (avx512f_<code>v8div16qi2_mask): Ditto.
6161 (avx512f_<code>v8div16qi2_mask_store): Ditto.
6162 (avx512f_eq<mode>3<mask_scalar_merge_name>_1): Ditto.
6163 (avx512f_gt<mode>3<mask_scalar_merge_name>): Ditto.
6164 (avx512f_testm<mode>3<mask_scalar_merge_name>): Ditto.
6165 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Ditto.
6166 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
6167 (*avx512pf_gatherpf<mode>df_mask): Ditto.
6168 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
6169 (*avx512pf_scatterpf<mode>df_mask): Ditto.
6170 (avx512cd_maskb_vec_dupv8di): Ditto.
6171 (avx512cd_maskw_vec_dupv16si): Ditto.
6172 (avx512f_vpermi2var<mode>3_maskz): Ditto.
6173 (avx512f_vpermi2var<mode>3_mask): Ditto.
6174 (avx512f_vpermi2var<mode>3_mask): Ditto.
6175 (avx512f_vpermt2var<mode>3_maskz): Ditto.
6176 (*avx512f_gathersi<mode>): Ditto.
6177 (*avx512f_gathersi<mode>_2): Ditto.
6178 (*avx512f_gatherdi<mode>): Ditto.
6179 (*avx512f_gatherdi<mode>_2): Ditto.
6180 (*avx512f_scattersi<mode>): Ditto.
6181 (*avx512f_scatterdi<mode>): Ditto.
6182 (avx512f_compress<mode>_mask): Ditto.
6183 (avx512f_compressstore<mode>_mask): Ditto.
6184 (avx512f_expand<mode>_mask): Ditto.
6185 * config/i386/subst.md (mask): Change k to Yk.
6186 (mask_scalar_merge): Ditto.
6189 2014-01-31 Marc Glisse <marc.glisse@inria.fr>
6191 * doc/extend.texi (Vector Extensions): Document ?: in C++.
6193 2014-01-31 Richard Biener <rguenther@suse.de>
6196 * builtins.c (fold_builtin_memory_op): Make sure to not
6197 use a floating-point mode or a boolean or enumeral type for
6200 2014-01-30 DJ Delorie <dj@redhat.com>
6202 * config/msp430/msp430.h (LIB_SPEC): Add -lcrt
6203 * config/msp430/msp430.md (msp430_refsym_need_exit): New.
6204 * config/msp430/msp430.c (msp430_expand_epilogue): Call it
6205 whenever main() has an epilogue.
6207 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6209 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Remove
6210 unused variable "field".
6211 * config/rs6000/vsx.md (vsx_mergel_<mode>): Add missing DONE.
6212 (vsx_mergeh_<mode>): Likewise.
6213 * config/rs6000/altivec.md (altivec_vmrghb): Likewise.
6214 (altivec_vmrghh): Likewise.
6215 (altivec_vmrghw): Likewise.
6216 (altivec_vmrglb): Likewise.
6217 (altivec_vmrglh): Likewise.
6218 (altivec_vmrglw): Likewise.
6219 (altivec_vspltb): Add missing uses.
6220 (altivec_vsplth): Likewise.
6221 (altivec_vspltw): Likewise.
6222 (altivec_vspltsf): Likewise.
6224 2014-01-30 Jakub Jelinek <jakub@redhat.com>
6227 * ifcvt.c (cond_exec_process_insns): Don't conditionalize
6228 frame related instructions.
6230 2014-01-30 Vladimir Makarov <vmakarov@redhat.com>
6232 PR rtl-optimization/59959
6233 * lra-constrains.c (simplify_operand_subreg): Assign NO_REGS to
6234 any reload of register whose subreg is invalid.
6236 2014-01-30 Jakub Jelinek <jakub@redhat.com>
6238 * config/i386/f16cintrin.h (_cvtsh_ss): Avoid -Wnarrowing warning.
6239 * config/i386/avx512fintrin.h (_mm512_mask_cvtusepi64_storeu_epi32):
6240 Add missing return type - void.
6242 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6244 * gcc/config/rs6000/rs6000.c (rs6000_expand_vector_init): Use
6245 gen_vsx_xxspltw_v4sf_direct instead of gen_vsx_xxspltw_v4sf;
6246 remove element index adjustment for endian (now handled in vsx.md
6248 (altivec_expand_vec_perm_const): Use
6249 gen_altivec_vsplt[bhw]_direct instead of gen_altivec_vsplt[bhw].
6250 * gcc/config/rs6000/vsx.md (UNSPEC_VSX_XXSPLTW): New unspec.
6251 (vsx_xxspltw_<mode>): Adjust element index for little endian.
6252 * gcc/config/rs6000/altivec.md (altivec_vspltb): Divide into a
6253 define_expand and a new define_insn *altivec_vspltb_internal;
6254 adjust for -maltivec=be on a little endian target.
6255 (altivec_vspltb_direct): New.
6256 (altivec_vsplth): Divide into a define_expand and a new
6257 define_insn *altivec_vsplth_internal; adjust for -maltivec=be on a
6258 little endian target.
6259 (altivec_vsplth_direct): New.
6260 (altivec_vspltw): Divide into a define_expand and a new
6261 define_insn *altivec_vspltw_internal; adjust for -maltivec=be on a
6262 little endian target.
6263 (altivec_vspltw_direct): New.
6264 (altivec_vspltsf): Divide into a define_expand and a new
6265 define_insn *altivec_vspltsf_internal; adjust for -maltivec=be on
6266 a little endian target.
6268 2014-01-30 Richard Biener <rguenther@suse.de>
6270 PR tree-optimization/59993
6271 * tree-ssa-forwprop.c (associate_pointerplus): Check we
6272 can propagate form the earlier stmt and avoid the transform
6273 when the intermediate result is needed.
6275 2014-01-30 Alangi Derick <alangiderick@gmail.com>
6277 * README.Portability: Fix typo.
6279 2014-01-30 David Holsgrove <david.holsgrove@xilinx.com>
6281 * config/microblaze/microblaze.md(cstoresf4, cbranchsf4): Replace
6282 comparison_operator with ordered_comparison_operator.
6284 2014-01-30 Nick Clifton <nickc@redhat.com>
6286 * config/mn10300/mn10300-protos.h (mn10300_store_multiple_operation_p):
6287 Rename to mn10300_store_multiple_regs.
6288 * config/mn10300/mn10300.c: Likewise.
6289 * config/mn10300/mn10300.md (store_movm): Fix typo: call
6290 store_multiple_regs.
6291 * config/mn10300/predicates.md (mn10300_store_multiple_operation):
6292 Call mn10300_store_multiple_regs.
6294 2014-01-30 Nick Clifton <nickc@redhat.com>
6295 DJ Delorie <dj@redhat.com>
6297 * config/rl78/rl78.c (register_sizes): Make the "upper half" of
6298 %fp 2 to keep registers after it properly word-aligned.
6299 (rl78_alloc_physical_registers_umul): Handle the case where both
6300 input operands are the same.
6302 2014-01-30 Richard Biener <rguenther@suse.de>
6304 PR tree-optimization/59903
6305 * tree-vect-loop.c (vect_transform_loop): Guard multiple-types
6308 2014-01-30 Jason Merrill <jason@redhat.com>
6311 * tree.c (walk_type_fields): Handle VECTOR_TYPE.
6314 * cgraphunit.c (expand_thunk): Copy volatile arg to a temporary.
6316 2014-01-30 Richard Biener <rguenther@suse.de>
6318 PR tree-optimization/59951
6319 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Skip uses in debug insns.
6321 2014-01-30 Savin Zlobec <savin.zlobec@gmail.com>
6324 * config/nios2/nios2.c (nios2_fpu_insn_asm): Fix asm output of
6325 SFmode to DFmode case.
6327 2014-01-29 DJ Delorie <dj@redhat.com>
6329 * config/msp430/msp430.opt (-minrt): New.
6330 * config/msp430/msp430.h (STARTFILE_SPEC): Link alternate runtime
6332 (ENDFILE_SPEC): Likewise.
6334 2014-01-29 Jan Hubicka <hubicka@ucw.cz>
6336 * ipa-inline-analysis.c (clobber_only_eh_bb_p): New function.
6337 (estimate_function_body_sizes): Use it.
6339 2014-01-29 Paolo Carlini <paolo.carlini@oracle.com>
6342 * dwarf2out.c (is_cxx_auto): New.
6343 (is_base_type): Use it.
6344 (gen_type_die_with_usage): Likewise.
6346 2014-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6348 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Use
6349 CODE_FOR_altivec_vmrg*_direct rather than CODE_FOR_altivec_vmrg*.
6350 * config/rs6000/vsx.md (vsx_mergel_<mode>): Adjust for
6351 -maltivec=be with LE targets.
6352 (vsx_mergeh_<mode>): Likewise.
6353 * config/rs6000/altivec.md (UNSPEC_VMRG[HL]_DIRECT): New unspecs.
6354 (mulv8hi3): Use gen_altivec_vmrg[hl]w_direct.
6355 (altivec_vmrghb): Replace with define_expand and new
6356 *altivec_vmrghb_internal insn; adjust for -maltivec=be with LE targets.
6357 (altivec_vmrghb_direct): New define_insn.
6358 (altivec_vmrghh): Replace with define_expand and new
6359 *altivec_vmrghh_internal insn; adjust for -maltivec=be with LE targets.
6360 (altivec_vmrghh_direct): New define_insn.
6361 (altivec_vmrghw): Replace with define_expand and new
6362 *altivec_vmrghw_internal insn; adjust for -maltivec=be with LE targets.
6363 (altivec_vmrghw_direct): New define_insn.
6364 (*altivec_vmrghsf): Adjust for endianness.
6365 (altivec_vmrglb): Replace with define_expand and new
6366 *altivec_vmrglb_internal insn; adjust for -maltivec=be with LE targets.
6367 (altivec_vmrglb_direct): New define_insn.
6368 (altivec_vmrglh): Replace with define_expand and new
6369 *altivec_vmrglh_internal insn; adjust for -maltivec=be with LE targets.
6370 (altivec_vmrglh_direct): New define_insn.
6371 (altivec_vmrglw): Replace with define_expand and new
6372 *altivec_vmrglw_internal insn; adjust for -maltivec=be with LE targets.
6373 (altivec_vmrglw_direct): New define_insn.
6374 (*altivec_vmrglsf): Adjust for endianness.
6375 (vec_widen_umult_hi_v16qi): Use gen_altivec_vmrghh_direct.
6376 (vec_widen_umult_lo_v16qi): Use gen_altivec_vmrglh_direct.
6377 (vec_widen_smult_hi_v16qi): Use gen_altivec_vmrghh_direct.
6378 (vec_widen_smult_lo_v16qi): Use gen_altivec_vmrglh_direct.
6379 (vec_widen_umult_hi_v8hi): Use gen_altivec_vmrghw_direct.
6380 (vec_widen_umult_lo_v8hi): Use gen_altivec_vmrglw_direct.
6381 (vec_widen_smult_hi_v8hi): Use gen_altivec_vmrghw_direct.
6382 (vec_widen_smult_lo_v8hi): Use gen_altivec_vmrglw_direct.
6384 2014-01-29 Marcus Shawcroft <marcus.shawcroft@arm.com>
6386 * config/aarch64/aarch64.c (aarch64_expand_mov_immediate)
6387 (aarch64_legitimate_address_p, aarch64_class_max_nregs): Adjust
6390 2014-01-29 Richard Biener <rguenther@suse.de>
6392 PR tree-optimization/58742
6393 * tree-ssa-forwprop.c (associate_pointerplus): Rename to
6394 associate_pointerplus_align.
6395 (associate_pointerplus_diff): New function.
6396 (associate_pointerplus): Likewise. Call associate_pointerplus_align
6397 and associate_pointerplus_diff.
6399 2014-01-29 Richard Biener <rguenther@suse.de>
6401 * lto-streamer.h (LTO_major_version): Bump to 3.
6402 (LTO_minor_version): Reset to 0.
6404 2014-01-29 Renlin Li <Renlin.Li@arm.com>
6406 * config/arm/arm-arches.def (ARM_ARCH): Add armv7ve arch.
6407 * config/arm/arm.c (FL_FOR_ARCH7VE): New.
6408 (arm_file_start): Generate correct asm header for armv7ve.
6409 * config/arm/bpabi.h: Add multilib support for armv7ve.
6410 * config/arm/driver-arm.c: Change the architectures of cortex-a7
6411 and cortex-a15 to armv7ve.
6412 * config/arm/t-aprofile: Add multilib support for armv7ve.
6413 * doc/invoke.texi: Document -march=armv7ve.
6415 2014-01-29 Richard Biener <rguenther@suse.de>
6417 PR tree-optimization/58742
6418 * tree-ssa-forwprop.c (associate_plusminus): Return true
6419 if we changed sth, defer EH cleanup to ...
6420 (ssa_forward_propagate_and_combine): ... here. Call simplify_mult.
6421 (simplify_mult): New function.
6423 2014-01-29 Jakub Jelinek <jakub@redhat.com>
6426 PR tree-optimization/59920
6427 * tree.c (build_common_builtin_nodes): Remove
6428 __builtin_setjmp_dispatcher initialization.
6429 * omp-low.h (make_gimple_omp_edges): Add a new int * argument.
6430 * profile.c (branch_prob): Use gsi_start_nondebug_after_labels_bb
6431 instead of gsi_after_labels + manually skipping debug stmts.
6432 Don't ignore bbs with BUILT_IN_SETJMP_DISPATCHER, instead
6433 ignore bbs with IFN_ABNORMAL_DISPATCHER.
6434 * tree-inline.c (copy_edges_for_bb): Remove
6435 can_make_abnormal_goto argument, instead add abnormal_goto_dest
6436 argument. Ignore computed_goto_p stmts. Don't call
6437 make_abnormal_goto_edges. If a call might need abnormal edges
6438 for non-local gotos, see if it already has an edge to
6439 IFN_ABNORMAL_DISPATCHER or if it is IFN_ABNORMAL_DISPATCHER
6440 with true argument, don't do anything then, otherwise add
6441 EDGE_ABNORMAL from the call's bb to abnormal_goto_dest.
6442 (copy_cfg_body): Compute abnormal_goto_dest, adjust copy_edges_for_bb
6444 * gimple-low.c (struct lower_data): Remove calls_builtin_setjmp.
6445 (lower_function_body): Don't emit __builtin_setjmp_dispatcher.
6446 (lower_stmt): Don't set data->calls_builtin_setjmp.
6447 (lower_builtin_setjmp): Adjust comment.
6448 * builtins.def (BUILT_IN_SETJMP_DISPATCHER): Remove.
6449 * tree-cfg.c (found_computed_goto): Remove.
6450 (factor_computed_gotos): Remove.
6451 (make_goto_expr_edges): Return bool, true for computed gotos.
6452 Don't call make_abnormal_goto_edges.
6453 (build_gimple_cfg): Don't set found_computed_goto, don't call
6454 factor_computed_gotos.
6455 (computed_goto_p): No longer static.
6456 (make_blocks): Don't set found_computed_goto.
6457 (get_abnormal_succ_dispatcher, handle_abnormal_edges): New functions.
6458 (make_edges): If make_goto_expr_edges returns true, push bb
6459 into ab_edge_goto vector, for stmt_can_make_abnormal_goto calls
6460 instead of calling make_abnormal_goto_edges push bb into ab_edge_call
6461 vector. Record mapping between bbs and OpenMP regions if there
6462 are any, adjust make_gimple_omp_edges caller. Call
6463 handle_abnormal_edges.
6464 (make_abnormal_goto_edges): Remove.
6465 * tree-cfg.h (make_abnormal_goto_edges): Remove.
6466 (computed_goto_p, get_abnormal_succ_dispatcher): New prototypes.
6467 * internal-fn.c (expand_ABNORMAL_DISPATCHER): New function.
6468 * builtins.c (expand_builtin): Don't handle BUILT_IN_SETJMP_DISPATCHER.
6469 * internal-fn.def (ABNORMAL_DISPATCHER): New.
6470 * omp-low.c (make_gimple_omp_edges): Add region_idx argument, when
6471 filling *region also set *region_idx to (*region)->entry->index.
6474 * read-rtl.c (read_rtx_code): Clear all of RTX_CODE_SIZE (code).
6475 For REGs set ORIGINAL_REGNO.
6477 2014-01-29 Bingfeng Mei <bmei@broadcom.com>
6479 * doc/md.texi: Mention that a target shouldn't implement
6480 vec_widen_(s|u)mul_even/odd pair if it is less efficient
6483 2014-01-29 Jakub Jelinek <jakub@redhat.com>
6485 PR tree-optimization/59594
6486 * tree-vect-data-refs.c (vect_analyze_data_ref_accesses): Sort
6487 a copy of the datarefs vector rather than the vector itself.
6489 2014-01-28 Jason Merrill <jason@redhat.com>
6492 * dwarf2out.c (auto_die): New static.
6493 (gen_type_die_with_usage): Handle C++1y 'auto'.
6494 (gen_subprogram_die): If in-class DIE had 'auto', emit type again
6497 2014-01-28 H.J. Lu <hongjiu.lu@intel.com>
6500 * config/i386/gnu-user64.h (SPEC_32): Add "m16|" to "m32".
6501 (SPEC_X32): Likewise.
6502 (SPEC_64): Likewise.
6503 * config/i386/i386.c (ix86_option_override_internal): Turn off
6504 OPTION_MASK_ISA_64BIT, OPTION_MASK_ABI_X32 and OPTION_MASK_ABI_64
6506 (x86_file_start): Output .code16gcc for TARGET_16BIT.
6507 * config/i386/i386.h (TARGET_16BIT): New macro.
6508 (TARGET_16BIT_P): Likewise.
6509 * config/i386/i386.opt: Add m16.
6510 * doc/invoke.texi: Document -m16.
6512 2014-01-28 Jakub Jelinek <jakub@redhat.com>
6514 PR preprocessor/59935
6515 * input.c (location_get_source_line): Bail out on when line number
6516 is zero, and test the return value of lookup_or_add_file_to_cache_tab.
6518 2014-01-28 Richard Biener <rguenther@suse.de>
6520 PR tree-optimization/58742
6521 * tree-ssa-forwprop.c (associate_plusminus): Handle
6522 pointer subtraction of the form (T)(P + A) - (T)P.
6524 2014-01-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6526 * config/arm/arm.c (arm_new_rtx_costs): Remove useless statement
6529 2014-01-28 Richard Biener <rguenther@suse.de>
6532 2014-01-28 Richard Biener <rguenther@suse.de>
6534 PR rtl-optimization/45364
6535 PR rtl-optimization/59890
6536 * var-tracking.c (local_get_addr_clear_given_value): Handle
6537 already cleared slot.
6538 (val_reset): Handle not allocated local_get_addr_cache.
6539 (vt_find_locations): Use post-order on the inverted CFG.
6541 2014-01-28 Richard Biener <rguenther@suse.de>
6543 * tree-data-ref.h (ddr_is_anti_dependent, ddrs_have_anti_deps): Remove.
6545 2014-01-28 Richard Biener <rguenther@suse.de>
6547 PR rtl-optimization/45364
6548 PR rtl-optimization/59890
6549 * var-tracking.c (local_get_addr_clear_given_value): Handle
6550 already cleared slot.
6551 (val_reset): Handle not allocated local_get_addr_cache.
6552 (vt_find_locations): Use post-order on the inverted CFG.
6554 2014-01-28 Alan Modra <amodra@gmail.com>
6556 * Makefile.in (BUILD_CPPFLAGS): Do not use ALL_CPPFLAGS.
6557 * configure.ac <recursive call for build != host>: Define
6558 GENERATOR_FILE. Comment. Use CXX_FOR_BUILD, CXXFLAGS_FOR_BUILD
6559 and LD_FOR_BUILD too.
6560 * configure: Regenerate.
6562 2014-01-27 Allan Sandfeld Jensen <sandfeld@kde.org>
6564 * config/i386/i386.c (get_builtin_code_for_version): Separate
6565 Westmere from Nehalem, Ivy Bridge from Sandy Bridge and
6566 Broadwell from Haswell.
6568 2014-01-27 Steve Ellcey <sellcey@mips.com>
6570 * common/config/mips/mips-common.c (TARGET_DEFAULT_TARGET_FLAGS):
6571 Remove TARGET_FP_EXCEPTIONS_DEFAULT and MASK_FUSED_MADD.
6572 * config/mips/mips.c (mips_option_override): Change setting
6574 * config/mips/mips.h (TARGET_FP_EXCEPTIONS_DEFAULT): Remove.
6575 * config/mips/mips.opt (DSP, DSPR2, FP_EXCEPTIONS, FUSED_MADD, MIPS3D):
6576 Change from Mask to Var.
6578 2014-01-27 Jeff Law <law@redhat.com>
6580 * ipa-inline.c (inline_small_functions): Fix typo.
6582 2014-01-27 Ilya Tocar <ilya.tocar@intel.com>
6584 * config/i386/avx512fintrin.h (_mm512_mask_cvtepi32_storeu_epi8): New.
6585 (_mm512_mask_cvtsepi32_storeu_epi8): Ditto.
6586 (_mm512_mask_cvtusepi32_storeu_epi8): Ditto.
6587 (_mm512_mask_cvtepi32_storeu_epi16): Ditto.
6588 (_mm512_mask_cvtsepi32_storeu_epi16): Ditto.
6589 (_mm512_mask_cvtusepi32_storeu_epi16): Ditto.
6590 (_mm512_mask_cvtepi64_storeu_epi32): Ditto.
6591 (_mm512_mask_cvtsepi64_storeu_epi32): Ditto.
6592 (_mm512_mask_cvtusepi64_storeu_epi32): Ditto.
6593 (_mm512_mask_cvtepi64_storeu_epi16): Ditto.
6594 (_mm512_mask_cvtsepi64_storeu_epi16): Ditto.
6595 (_mm512_mask_cvtusepi64_storeu_epi16): Ditto.
6596 (_mm512_mask_cvtepi64_storeu_epi8): Ditto.
6597 (_mm512_mask_cvtsepi64_storeu_epi8): Ditto.
6598 (_mm512_mask_cvtusepi64_storeu_epi8): Ditto.
6599 (_mm512_storeu_epi64): Ditto.
6600 (_mm512_cmpge_epi32_mask): Ditto.
6601 (_mm512_cmpge_epu32_mask): Ditto.
6602 (_mm512_cmpge_epi64_mask): Ditto.
6603 (_mm512_cmpge_epu64_mask): Ditto.
6604 (_mm512_cmple_epi32_mask): Ditto.
6605 (_mm512_cmple_epu32_mask): Ditto.
6606 (_mm512_cmple_epi64_mask): Ditto.
6607 (_mm512_cmple_epu64_mask): Ditto.
6608 (_mm512_cmplt_epi32_mask): Ditto.
6609 (_mm512_cmplt_epu32_mask): Ditto.
6610 (_mm512_cmplt_epi64_mask): Ditto.
6611 (_mm512_cmplt_epu64_mask): Ditto.
6612 (_mm512_cmpneq_epi32_mask): Ditto.
6613 (_mm512_cmpneq_epu32_mask): Ditto.
6614 (_mm512_cmpneq_epi64_mask): Ditto.
6615 (_mm512_cmpneq_epu64_mask): Ditto.
6616 (_mm512_expand_pd): Ditto.
6617 (_mm512_expand_ps): Ditto.
6618 * config/i386/i386-builtin-types.def: Add PV16QI, PV16QI, PV16HI,
6619 VOID_PV8SI_V8DI_QI, VOID_PV8HI_V8DI_QI, VOID_PV16QI_V8DI_QI,
6620 VOID_PV16QI_V16SI_HI, VOID_PV16HI_V16SI_HI.
6621 * config/i386/i386.c (ix86_builtins): Add
6622 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK,
6623 IX86_BUILTIN_PMOVDB512_MEM, IX86_BUILTIN_PMOVDW512_MEM,
6624 IX86_BUILTIN_PMOVQB512_MEM, IX86_BUILTIN_PMOVQD512_MEM,
6625 IX86_BUILTIN_PMOVQW512_MEM, IX86_BUILTIN_PMOVSDB512_MEM,
6626 IX86_BUILTIN_PMOVSDW512_MEM, IX86_BUILTIN_PMOVSQB512_MEM,
6627 IX86_BUILTIN_PMOVSQD512_MEM, IX86_BUILTIN_PMOVSQW512_MEM,
6628 IX86_BUILTIN_PMOVUSDB512_MEM, IX86_BUILTIN_PMOVUSDW512_MEM,
6629 IX86_BUILTIN_PMOVUSQB512_MEM, IX86_BUILTIN_PMOVUSQD512_MEM,
6630 IX86_BUILTIN_PMOVUSQW512_MEM.
6631 (bdesc_special_args): Add __builtin_ia32_pmovusqd512mem_mask,
6632 __builtin_ia32_pmovsqd512mem_mask,
6633 __builtin_ia32_pmovqd512mem_mask,
6634 __builtin_ia32_pmovusqw512mem_mask,
6635 __builtin_ia32_pmovsqw512mem_mask,
6636 __builtin_ia32_pmovqw512mem_mask,
6637 __builtin_ia32_pmovusdw512mem_mask,
6638 __builtin_ia32_pmovsdw512mem_mask,
6639 __builtin_ia32_pmovdw512mem_mask,
6640 __builtin_ia32_pmovqb512mem_mask,
6641 __builtin_ia32_pmovusqb512mem_mask,
6642 __builtin_ia32_pmovsqb512mem_mask,
6643 __builtin_ia32_pmovusdb512mem_mask,
6644 __builtin_ia32_pmovsdb512mem_mask,
6645 __builtin_ia32_pmovdb512mem_mask.
6646 (bdesc_args): Add __builtin_ia32_expanddf512,
6647 __builtin_ia32_expandsf512.
6648 (ix86_expand_special_args_builtin): Handle VOID_FTYPE_PV8SI_V8DI_QI,
6649 VOID_FTYPE_PV8HI_V8DI_QI, VOID_FTYPE_PV16HI_V16SI_HI,
6650 VOID_FTYPE_PV16QI_V8DI_QI, VOID_FTYPE_PV16QI_V16SI_HI.
6651 * config/i386/sse.md (unspec): Add UNSPEC_EXPAND_NOMASK.
6652 (avx512f_<code><pmov_src_lower><mode>2_mask_store): New.
6653 (*avx512f_<code>v8div16qi2_store_mask): Renamed to ...
6654 (avx512f_<code>v8div16qi2_mask_store): This.
6655 (avx512f_expand<mode>): New.
6657 2014-01-27 Kirill Yukhin <kirill.yukhin@intel.com>
6659 * config/i386/avx512pfintrin.h (_mm512_mask_prefetch_i32gather_pd):
6661 (_mm512_mask_prefetch_i64gather_pd): Ditto.
6662 (_mm512_prefetch_i32scatter_pd): Ditto.
6663 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
6664 (_mm512_prefetch_i64scatter_pd): Ditto.
6665 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
6666 (_mm512_mask_prefetch_i32gather_ps): Fix operand type.
6667 (_mm512_mask_prefetch_i64gather_ps): Ditto.
6668 (_mm512_prefetch_i32scatter_ps): Ditto.
6669 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
6670 (_mm512_prefetch_i64scatter_ps): Ditto.
6671 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
6672 * config/i386/i386-builtin-types.def: Define
6673 VOID_FTYPE_QI_V8SI_PCINT64_INT_INT
6674 and VOID_FTYPE_QI_V8DI_PCINT64_INT_INT.
6675 * config/i386/i386.c (ix86_builtins): Define IX86_BUILTIN_GATHERPFQPD,
6676 IX86_BUILTIN_GATHERPFDPD, IX86_BUILTIN_SCATTERPFDPD,
6677 IX86_BUILTIN_SCATTERPFQPD.
6678 (ix86_init_mmx_sse_builtins): Define __builtin_ia32_gatherpfdpd,
6679 __builtin_ia32_gatherpfdps, __builtin_ia32_gatherpfqpd,
6680 __builtin_ia32_gatherpfqps, __builtin_ia32_scatterpfdpd,
6681 __builtin_ia32_scatterpfdps, __builtin_ia32_scatterpfqpd,
6682 __builtin_ia32_scatterpfqps.
6683 (ix86_expand_builtin): Expand new built-ins.
6684 * config/i386/sse.md (avx512pf_gatherpf<mode>): Add SF suffix,
6685 fix memory access data type.
6686 (*avx512pf_gatherpf<mode>_mask): Ditto.
6687 (*avx512pf_gatherpf<mode>): Ditto.
6688 (avx512pf_scatterpf<mode>): Ditto.
6689 (*avx512pf_scatterpf<mode>_mask): Ditto.
6690 (*avx512pf_scatterpf<mode>): Ditto.
6691 (GATHER_SCATTER_SF_MEM_MODE): New.
6692 (avx512pf_gatherpf<mode>df): Ditto.
6693 (*avx512pf_gatherpf<mode>df_mask): Ditto.
6694 (*avx512pf_scatterpf<mode>df): Ditto.
6696 2014-01-27 Jakub Jelinek <jakub@redhat.com>
6699 * expmed.h (expmed_mode_index): Rework so that analysis and optimziers
6700 know when the MODE_PARTIAL_INT and MODE_VECTOR_INT cases can never be
6703 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
6705 * common/config/arm/arm-common.c
6706 (arm_rewrite_mcpu): Handle multiple names.
6708 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
6710 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
6712 * gimple-builder.h (create_gimple_tmp): Delete.
6714 2014-01-27 Christian Bruel <christian.bruel@st.com>
6716 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Fix remaining bytes after
6719 2014-01-26 John David Anglin <danglin@gcc.gnu.org>
6721 * config/pa/pa.md (call): Generate indirect long calls to non-local
6722 functions when outputing 32-bit code.
6723 (call_value): Likewise except for special call to buggy powf function.
6725 * config/pa/pa.c (pa_attr_length_indirect_call): Adjust length of
6726 portable runtime and PIC indirect calls.
6727 (pa_output_indirect_call): Remove unnecessary nop from portable runtime
6728 and PIC call sequences. Use ldo instead of blr to set return register
6729 in PIC call sequence.
6731 2014-01-25 Walter Lee <walt@tilera.com>
6733 * config/tilegx/sync.md (atomic_fetch_sub): Fix negation and
6734 avoid clobbering a live register.
6736 2014-01-25 Walter Lee <walt@tilera.com>
6738 * config/tilegx/tilegx-c.c (tilegx_cpu_cpp_builtins):
6739 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2}.
6740 * config/tilegx/tilepro-c.c (tilepro_cpu_cpp_builtins):
6741 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2,4,8}.
6743 2014-01-25 Walter Lee <walt@tilera.com>
6745 * config/tilegx/tilegx.c (tilegx_function_arg): Start 16-byte
6746 arguments on even registers.
6747 (tilegx_gimplify_va_arg_expr): Align 16-byte var args to
6749 * config/tilegx/tilegx.h (STACK_BOUNDARY): Change to 16 bytes.
6750 (BIGGEST_ALIGNMENT): Ditto.
6751 (BIGGEST_FIELD_ALIGNMENT): Ditto.
6753 2014-01-25 Walter Lee <walt@tilera.com>
6755 * config/tilegx/tilegx.c (tilegx_gen_bundles): Delete barrier
6756 insns before bundling.
6757 * config/tilegx/tilegx.md (tile_network_barrier): Update comment.
6759 2014-01-25 Walter Lee <walt@tilera.com>
6761 * config/tilegx/tilegx.c (tilegx_expand_builtin): Set
6762 PREFETCH_SCHEDULE_BARRIER_P to true for prefetches.
6763 * config/tilepro/tilepro.c (tilepro_expand_builtin): Ditto.
6765 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
6767 * config/mips/constraints.md (kl): Delete.
6768 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Turn into
6769 define expands, using...
6770 (divmod<mode>4_mips16, udivmod<mode>4_mips16): ...these new
6771 instructions for MIPS16.
6772 (*divmod<mode>4, *udivmod<mode>4): New patterns, taken from the
6773 non-MIPS16 version of the old divmod<mode>4 and udivmod<mode>4.
6775 2014-01-25 Walter Lee <walt@tilera.com>
6777 * config/tilepro/tilepro.md (ctzdi2): Use register_operand predicate.
6781 2014-01-25 Walter Lee <walt@tilera.com>
6783 * config/tilegx/tilegx.c (tilegx_expand_to_rtl_hook): New.
6784 (TARGET_EXPAND_TO_RTL_HOOK): Define.
6786 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
6788 * rtlanal.c (canonicalize_condition): Split out duplicated mode check.
6791 2014-01-25 Jakub Jelinek <jakub@redhat.com>
6793 * print-rtl.c (in_call_function_usage): New var.
6794 (print_rtx): When in CALL_INSN_FUNCTION_USAGE, always print
6795 EXPR_LIST mode as mode and not as reg note name.
6798 * cfgloopmanip.c (copy_loop_info): If
6799 loop->warned_aggressive_loop_optimizations, make sure
6800 the flag is set in target loop too.
6802 2014-01-24 Balaji V. Iyer <balaji.v.iyer@intel.com>
6804 * builtins.c (is_builtin_name): Renamed flag_enable_cilkplus to
6806 * builtins.def: Likewise.
6807 * cilk.h (fn_contains_cilk_spawn_p): Likewise.
6808 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Likewise.
6809 * ira.c (ira_setup_eliminable_regset): Likewise.
6810 * omp-low.c (gate_expand_omp): Likewise.
6811 (execute_lower_omp): Likewise.
6812 (diagnose_sb_0): Likewise.
6813 (gate_diagnose_omp_blocks): Likewise.
6814 (simd_clone_clauses_extract): Likewise.
6817 2014-01-24 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6819 * config/rs6000/rs6000.c (rs6000_expand_vec_perm_const_1): Remove
6820 correction for little endian...
6821 * config/rs6000/vsx.md (vsx_xxpermdi2_<mode>_1): ...and move it to
6824 2014-01-24 Jeff Law <law@redhat.com>
6826 PR tree-optimization/59919
6827 * tree-vrp.c (find_assert_locations_1): Do not register asserts
6828 for non-returning calls.
6830 2014-01-24 James Greenhalgh <james.greenhalgh@arm.com>
6832 * common/config/aarch64/aarch64-common.c
6833 (aarch64_rewrite_mcpu): Handle multiple names.
6834 * config/aarch64/aarch64.h
6835 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
6837 2014-01-24 Dodji Seketeli <dodji@redhat.com>
6839 * input.c (add_file_to_cache_tab): Handle the case where fopen
6842 2014-01-23 H.J. Lu <hongjiu.lu@intel.com>
6845 * config/i386/i386.md (pushsf splitter): Get stack adjustment
6846 from push operand if code of push isn't PRE_DEC.
6848 2014-01-23 Michael Meissner <meissner@linux.vnet.ibm.com>
6851 * doc/invoke.texi (RS/6000 and PowerPC Options): Document
6852 -mquad-memory-atomic. Update -mquad-memory documentation to say
6853 it is only used for non-atomic loads/stores.
6855 * config/rs6000/predicates.md (quad_int_reg_operand): Allow either
6856 -mquad-memory or -mquad-memory-atomic switches.
6858 * config/rs6000/rs6000-cpus.def (ISA_2_7_MASKS_SERVER): Add
6859 -mquad-memory-atomic to ISA 2.07 support.
6861 * config/rs6000/rs6000.opt (-mquad-memory-atomic): Add new switch
6862 to separate support of normal quad word memory operations (ldq, stq)
6863 from the atomic quad word memory operations.
6865 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
6866 support to separate non-atomic quad word operations from atomic
6867 quad word operations. Disable non-atomic quad word operations in
6868 little endian mode so that we don't have to swap words after the
6869 load and before the store.
6870 (quad_load_store_p): Add comment about atomic quad word support.
6871 (rs6000_opt_masks): Add -mquad-memory-atomic to the list of
6872 options printed with -mdebug=reg.
6874 * config/rs6000/rs6000.h (TARGET_SYNC_TI): Use
6875 -mquad-memory-atomic as the test for whether we have quad word
6876 atomic instructions.
6877 (TARGET_SYNC_HI_QI): If either -mquad-memory-atomic, -mquad-memory,
6878 or -mp8-vector are used, allow byte/half-word atomic operations.
6880 * config/rs6000/sync.md (load_lockedti): Insure that the address
6881 is a proper indexed or indirect address for the lqarx instruction.
6882 On little endian systems, swap the hi/lo registers after the lqarx
6884 (load_lockedpti): Use indexed_or_indirect_operand predicate to
6885 insure the address is valid for the lqarx instruction.
6886 (store_conditionalti): Insure that the address is a proper indexed
6887 or indirect address for the stqcrx. instruction. On little endian
6888 systems, swap the hi/lo registers before doing the stqcrx.
6890 (store_conditionalpti): Use indexed_or_indirect_operand predicate to
6891 insure the address is valid for the stqcrx. instruction.
6893 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
6894 Define __QUAD_MEMORY__ and __QUAD_MEMORY_ATOMIC__ based on what
6895 type of quad memory support is available.
6897 2014-01-23 Vladimir Makarov <vmakarov@redhat.com>
6900 * lra-constraints.c (simplify_operand_subreg): Spill pseudo if
6901 there is a danger of looping.
6903 2014-01-23 Pat Haugen <pthaugen@us.ibm.com>
6905 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
6906 force flag_ira_loop_pressure if set via command line.
6908 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
6910 * config/aarch64/aarch64-simd-builtins.def (ashr): DI mode removed.
6911 (ashr_simd): New builtin handling DI mode.
6912 * config/aarch64/aarch64-simd.md (aarch64_ashr_simddi): New pattern.
6913 (aarch64_sshr_simddi): New match pattern.
6914 * config/aarch64/arm_neon.h (vshr_n_s32): Builtin call modified.
6915 (vshrd_n_s64): Likewise.
6916 * config/aarch64/predicates.md (aarch64_shift_imm64_di): New predicate.
6918 2014-01-23 Nick Clifton <nickc@redhat.com>
6920 * config/msp430/msp430.h (ASM_SPEC): Pass the -mcpu as -mcpu.
6921 (LIB_SPEC): Drop use of memory.ld and peripherals.ld scripts in
6922 favour of mcu specific scripts.
6923 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more matches for
6926 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
6927 Alex Velenko <Alex.Velenko@arm.com>
6929 * config/aarch64/arm_neon.h (vaddv_s8): __LANE0 cleanup.
6930 (vaddv_s16): Likewise.
6931 (vaddv_s32): Likewise.
6932 (vaddv_u8): Likewise.
6933 (vaddv_u16): Likewise.
6934 (vaddv_u32): Likewise.
6935 (vaddvq_s8): Likewise.
6936 (vaddvq_s16): Likewise.
6937 (vaddvq_s32): Likewise.
6938 (vaddvq_s64): Likewise.
6939 (vaddvq_u8): Likewise.
6940 (vaddvq_u16): Likewise.
6941 (vaddvq_u32): Likewise.
6942 (vaddvq_u64): Likewise.
6943 (vaddv_f32): Likewise.
6944 (vaddvq_f32): Likewise.
6945 (vaddvq_f64): Likewise.
6946 (vmaxv_f32): Likewise.
6947 (vmaxv_s8): Likewise.
6948 (vmaxv_s16): Likewise.
6949 (vmaxv_s32): Likewise.
6950 (vmaxv_u8): Likewise.
6951 (vmaxv_u16): Likewise.
6952 (vmaxv_u32): Likewise.
6953 (vmaxvq_f32): Likewise.
6954 (vmaxvq_f64): Likewise.
6955 (vmaxvq_s8): Likewise.
6956 (vmaxvq_s16): Likewise.
6957 (vmaxvq_s32): Likewise.
6958 (vmaxvq_u8): Likewise.
6959 (vmaxvq_u16): Likewise.
6960 (vmaxvq_u32): Likewise.
6961 (vmaxnmv_f32): Likewise.
6962 (vmaxnmvq_f32): Likewise.
6963 (vmaxnmvq_f64): Likewise.
6964 (vminv_f32): Likewise.
6965 (vminv_s8): Likewise.
6966 (vminv_s16): Likewise.
6967 (vminv_s32): Likewise.
6968 (vminv_u8): Likewise.
6969 (vminv_u16): Likewise.
6970 (vminv_u32): Likewise.
6971 (vminvq_f32): Likewise.
6972 (vminvq_f64): Likewise.
6973 (vminvq_s8): Likewise.
6974 (vminvq_s16): Likewise.
6975 (vminvq_s32): Likewise.
6976 (vminvq_u8): Likewise.
6977 (vminvq_u16): Likewise.
6978 (vminvq_u32): Likewise.
6979 (vminnmv_f32): Likewise.
6980 (vminnmvq_f32): Likewise.
6981 (vminnmvq_f64): Likewise.
6983 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
6985 * config/aarch64/aarch64-simd.md
6986 (aarch64_dup_lane<mode>): Correct lane number on big-endian.
6987 (aarch64_dup_lane_<vswap_widthi_name><mode>): Likewise.
6988 (*aarch64_mul3_elt<mode>): Likewise.
6989 (*aarch64_mul3_elt<vswap_width_name><mode>): Likewise.
6990 (*aarch64_mul3_elt_to_64v2df): Likewise.
6991 (*aarch64_mla_elt<mode>): Likewise.
6992 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
6993 (*aarch64_mls_elt<mode>): Likewise.
6994 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
6995 (*aarch64_fma4_elt<mode>): Likewise.
6996 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
6997 (*aarch64_fma4_elt_to_64v2df): Likewise.
6998 (*aarch64_fnma4_elt<mode>): Likewise.
6999 (*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
7000 (*aarch64_fnma4_elt_to_64v2df): Likewise.
7001 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
7002 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
7003 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
7004 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
7005 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
7006 (aarch64_sqdmull_lane<mode>_internal): Likewise.
7007 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
7009 2013-01-23 Alex Velenko <Alex.Velenko@arm.com>
7011 * config/aarch64/aarch64-simd.md
7012 (aarch64_be_checked_get_lane<mode>): New define_expand.
7013 * config/aarch64/aarch64-simd-builtins.def
7014 (BUILTIN_VALL (GETLANE, be_checked_get_lane, 0)):
7015 New builtin definition.
7016 * config/aarch64/arm_neon.h: (__aarch64_vget_lane_any):
7017 Use new safe be builtin.
7019 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
7021 * config/aarch64/aarch64-simd.md (aarch64_be_ld1<mode>):
7023 (aarch64_be_st1<mode>): Likewise.
7024 (aarch_ld1<VALL:mode>): Define_expand modified.
7025 (aarch_st1<VALL:mode>): Likewise.
7026 * config/aarch64/aarch64.md (UNSPEC_LD1): New unspec definition.
7027 (UNSPEC_ST1): Likewise.
7029 2014-01-23 David Holsgrove <david.holsgrove@xilinx.com>
7031 * config/microblaze/microblaze.md: Add trap insn and attribute
7033 2014-01-23 Dodji Seketeli <dodji@redhat.com>
7035 PR preprocessor/58580
7036 * input.h (location_get_source_line): Take an additional line_size
7038 (void diagnostics_file_cache_fini): Declare new function.
7039 * input.c (struct fcache): New type.
7040 (fcache_tab_size, fcache_buffer_size, fcache_line_record_size):
7041 New static constants.
7042 (diagnostic_file_cache_init, total_lines_num)
7043 (lookup_file_in_cache_tab, evicted_cache_tab_entry)
7044 (add_file_to_cache_tab, lookup_or_add_file_to_cache_tab)
7045 (needs_read, needs_grow, maybe_grow, read_data, maybe_read_data)
7046 (get_next_line, read_next_line, goto_next_line, read_line_num):
7047 New static function definitions.
7048 (diagnostic_file_cache_fini): New function.
7049 (location_get_source_line): Take an additional output line_len
7050 parameter. Re-write using lookup_or_add_file_to_cache_tab and
7052 * diagnostic.c (diagnostic_finish): Call
7053 diagnostic_file_cache_fini.
7054 (adjust_line): Take an additional input parameter for the length
7055 of the line, rather than calculating it with strlen.
7056 (diagnostic_show_locus): Adjust the use of
7057 location_get_source_line and adjust_line with respect to their new
7058 signature. While displaying a line now, do not stop at the first
7059 null byte. Rather, display the zero byte as a space and keep
7060 going until we reach the size of the line.
7061 * Makefile.in: Add vec.o to OBJS-libcommon
7063 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
7064 Ilya Tocar <ilya.tocar@intel.com>
7066 * config/i386/avx512fintrin.h (_mm512_kmov): New.
7067 * config/i386/i386.c (IX86_BUILTIN_KMOV16): Ditto.
7068 (__builtin_ia32_kmov16): Ditto.
7069 * config/i386/i386.md (UNSPEC_KMOV): New.
7072 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
7074 * config/i386/avx512fintrin.h (_mm512_loadu_si512): Rename.
7075 (_mm512_storeu_si512): Ditto.
7077 2014-01-23 Richard Sandiford <rdsandiford@googlemail.com>
7080 * rtl.h (get_referenced_operands): Declare.
7081 * recog.c (get_referenced_operands): New function.
7082 * config/mips/mips.c (mips_reorg_process_insns): Check which asm
7083 operands have been referenced when recording LO_SUM references.
7085 2014-01-22 David Holsgrove <david.holsgrove@xilinx.com>
7087 * config/microblaze/microblaze.md: Correct bswaphi2 insn.
7089 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
7091 * config/i386/x86-tune.def (X86_TUNE_ACCUMULATE_OUTGOING_ARGS):
7092 Enable for generic and recent AMD targets.
7094 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
7096 * combine-stack-adj.c (combine_stack_adjustments_for_block): Remove
7097 ARG_SIZE note when adjustment was eliminated.
7099 2014-01-22 Jeff Law <law@redhat.com>
7101 PR tree-optimization/59597
7102 * tree-ssa-threadupdate.c (dump_jump_thread_path): Move to earlier
7103 in file. Accept new argument REGISTERING and use it to modify
7104 dump output appropriately.
7105 (register_jump_thread): Corresponding changes.
7106 (mark_threaded_blocks): Reinstate code to cancel unprofitable
7107 thread paths involving joiner blocks. Add code to dump cancelled
7108 jump threading paths.
7110 2014-01-22 Vladimir Makarov <vmakarov@redhat.com>
7112 PR rtl-optimization/59477
7113 * lra-constraints.c (inherit_in_ebb): Process call for living hard
7114 regs. Update reloads_num and potential_reload_hard_regs for all insns.
7116 2014-01-22 Tom Tromey <tromey@redhat.com>
7118 * config/i386/i386-interix.h (i386_pe_unique_section): Don't use
7120 * config/cr16/cr16-protos.h (notice_update_cc): Don't use PARAMS.
7122 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
7124 PR rtl-optimization/59896
7125 * lra-constraints.c (process_alt_operands): Check unused note for
7126 matched operands of insn with no output reloads.
7128 2014-01-21 Richard Sandiford <rdsandiford@googlemail.com>
7130 * config/mips/mips.c (mips_move_to_gpr_cost): Add M16_REGS case.
7131 (mips_move_from_gpr_cost): Likewise.
7133 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
7135 PR rtl-optimization/59858
7136 * lra-constraints.c (SMALL_REGISTER_CLASS_P): Use
7137 ira_class_hard_regs_num.
7138 (process_alt_operands): Increase reject for dying matched operand.
7140 2014-01-21 Jakub Jelinek <jakub@redhat.com>
7143 * config/i386/i386.c (expand_small_movmem_or_setmem): If mode is
7144 smaller than size, perform several stores or loads and stores
7145 at dst + count - size to store or copy all of size bytes, rather
7146 than just last modesize bytes.
7148 2014-01-20 DJ Delorie <dj@redhat.com>
7150 * config/rl78/rl78.c (rl78_propogate_register_origins): Verify
7151 that CLOBBERs are REGs before propogating their values.
7153 2014-01-20 H.J. Lu <hongjiu.lu@intel.com>
7156 * cgraph.c (cgraph_inline_failed_string): Add type to DEFCIFCODE.
7157 (cgraph_inline_failed_type): New function.
7158 * cgraph.h (DEFCIFCODE): Add type.
7159 (cgraph_inline_failed_type_t): New enum.
7160 (cgraph_inline_failed_type): New prototype.
7161 * cif-code.def: Add CIF_FINAL_NORMAL to OK, FUNCTION_NOT_CONSIDERED,
7162 FUNCTION_NOT_OPTIMIZED, REDEFINED_EXTERN_INLINE,
7163 FUNCTION_NOT_INLINE_CANDIDATE, LARGE_FUNCTION_GROWTH_LIMIT,
7164 LARGE_STACK_FRAME_GROWTH_LIMIT, MAX_INLINE_INSNS_SINGLE_LIMIT,
7165 MAX_INLINE_INSNS_AUTO_LIMIT, INLINE_UNIT_GROWTH_LIMIT,
7166 RECURSIVE_INLINING, UNLIKELY_CALL, NOT_DECLARED_INLINED,
7167 OPTIMIZING_FOR_SIZE, ORIGINALLY_INDIRECT_CALL,
7168 INDIRECT_UNKNOWN_CALL, USES_COMDAT_LOCAL.
7169 Add CIF_FINAL_ERROR to UNSPECIFIED, BODY_NOT_AVAILABLE,
7170 FUNCTION_NOT_INLINABLE, OVERWRITABLE, MISMATCHED_ARGUMENTS,
7171 EH_PERSONALITY, NON_CALL_EXCEPTIONS, TARGET_OPTION_MISMATCH,
7172 OPTIMIZATION_MISMATCH.
7173 * tree-inline.c (expand_call_inline): Emit errors during
7174 early_inlining if cgraph_inline_failed_type returns CIF_FINAL_ERROR.
7176 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
7179 * config/i386/sse.md (*andnot<mode>3<mask_name>): Handle MODE_V16SF
7180 mode attribute in insn output.
7182 2014-01-20 Eric Botcazou <ebotcazou@adacore.com>
7184 * output.h (output_constant): Delete.
7185 * varasm.c (output_constant): Make private.
7187 2014-01-20 Alex Velenko <Alex.Velenko@arm.com>
7189 * config/aarch64/aarch64-simd.md (vec_perm<mode>): Add BE check.
7191 2014-01-20 Jakub Jelinek <jakub@redhat.com>
7194 * tree.h (fold_builtin_strcat): New prototype.
7195 * builtins.c (fold_builtin_strcat): No longer static. Add len
7196 argument, if non-NULL, don't call c_strlen. Optimize
7197 directly into __builtin_memcpy instead of __builtin_strcpy.
7198 (fold_builtin_2): Adjust fold_builtin_strcat caller.
7199 * gimple-fold.c (gimple_fold_builtin): Handle BUILT_IN_STRCAT.
7201 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
7203 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
7204 for SImode_address_operand operands, having only a REG argument.
7206 2014-01-20 Marcus Shawcroft <marcus.shawcroft@arm.com>
7208 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER): Expand
7209 loader name using mbig-endian.
7210 (LINUX_TARGET_LINK_SPEC): Pass linker -m flag.
7212 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
7214 * doc/invoke.texi (-march): Clarify documentation for AArch64.
7218 2014-01-20 Tejas Belagod <tejas.belagod@arm.com>
7220 * config/aarch64/aarch64-protos.h
7221 (aarch64_cannot_change_mode_class_ptr): Declare.
7222 * config/aarch64/aarch64.c (aarch64_cannot_change_mode_class,
7223 aarch64_cannot_change_mode_class_ptr): New.
7224 * config/aarch64/aarch64.h (CANNOT_CHANGE_MODE_CLASS): Change to call
7225 backend hook aarch64_cannot_change_mode_class.
7227 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
7229 * common/config/aarch64/aarch64-common.c
7230 (aarch64_handle_option): Don't handle any option order logic here.
7231 * config/aarch64/aarch64.c (aarch64_parse_arch): Do not override
7232 selected_cpu, warn on architecture version mismatch.
7233 (aarch64_override_options): Fix parsing order for option strings.
7235 2014-01-20 Jan-Benedict Glaw <jbglaw@lug-owl.de>
7236 Iain Sandoe <iain@codesourcery.com>
7239 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Fix unused variable
7240 warning. Amend comment to reflect current functionality.
7242 2014-01-20 Richard Biener <rguenther@suse.de>
7245 * builtins.c (fold_builtin_strcat): Remove case better handled
7246 by tree-ssa-strlen.c.
7248 2014-01-20 Alan Lawrence <alan.lawrence@arm.com>
7250 * config/aarch64/aarch64.opt
7251 (mcpu, march, mtune): Make case-insensitive.
7253 2014-01-20 Jakub Jelinek <jakub@redhat.com>
7256 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
7257 if operands[1] is a REG or ZERO_EXTEND of a REG.
7259 2014-01-19 Jan Hubicka <hubicka@ucw.cz>
7261 * varasm.c (compute_reloc_for_constant): Use targetm.binds_local_p.
7263 2014-01-19 John David Anglin <danglin@gcc.gnu.org>
7265 * config/pa/pa.c (pa_attr_length_millicode_call): Correct length of
7266 long non-pic millicode calls.
7268 2014-01-19 Jan-Benedict Glaw <jbglaw@lug-owl.de>
7270 * config/vax/vax.h (FUNCTION_ARG_REGNO_P): Fix unused variable warning.
7272 2014-01-19 Kito Cheng <kito@0xlab.org>
7274 * builtins.c (expand_movstr): Check movstr expand done or fail.
7276 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
7277 H.J. Lu <hongjiu.lu@intel.com>
7280 * config/i386/i386.md (*lea<mode>): Zero-extend return register
7281 to DImode for zero-extended addresses.
7283 2014-01-19 Jakub Jelinek <jakub@redhat.com>
7285 PR rtl-optimization/57763
7286 * bb-reorder.c (fix_crossing_unconditional_branches): Set JUMP_LABEL
7287 on the new indirect jump_insn and increment LABEL_NUSES (label).
7289 2014-01-18 H.J. Lu <hongjiu.lu@intel.com>
7293 * config.gcc (x86_archs): New variable.
7294 (x86_64_archs): Likewise.
7295 (x86_cpus): Likewise.
7296 Use $x86_archs, $x86_64_archs and $x86_cpus to check valid
7297 --with-arch/--with-cpu= options.
7298 Support --with-arch=/--with-cpu={nehalem,westmere,
7299 sandybridge,ivybridge,haswell,broadwell,bonnell,silvermont}.
7301 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
7303 * config/i386/i386.c (ix86_adjust_cost): Reorder PROCESSOR_K8
7304 and PROCESSOR_ATHLON to simplify code. Move "memory" calculation.
7306 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
7308 * config/i386/i386.md (*swap<mode>): Rename from swap<mode>.
7310 2014-01-18 Jakub Jelinek <jakub@redhat.com>
7313 * config/i386/i386-c.c (ix86_pragma_target_parse): Temporarily
7314 clear cpp_get_options (parse_in)->warn_unused_macros for
7315 ix86_target_macros_internal with cpp_define.
7317 2014-01-18 Richard Sandiford <rdsandiford@googlemail.com>
7319 * jump.c (delete_related_insns): Keep (use (insn))s.
7320 * reorg.c (redundant_insn): Check for barriers too.
7322 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
7324 * config/i386/i386.c (ix86_split_lea_for_addr): Fix a comment typo.
7326 2014-01-17 John David Anglin <danglin@gcc.gnu.org>
7328 * config/pa/pa.c (pa_attr_length_indirect_call): Don't output a short
7329 call to $$dyncall when TARGET_LONG_CALLS is true.
7331 2014-01-17 Jeff Law <law@redhat.com>
7333 * ree.c (combine_set_extension): Temporarily disable test for
7334 changing number of hard registers.
7336 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
7339 * ipa-inline-analysis.c (inline_free_summary):
7340 Do not free summary of aliases.
7342 2014-01-17 Jakub Jelinek <jakub@redhat.com>
7345 * gimplify.c (gimplify_expr): Use create_tmp_var
7346 instead of create_tmp_var_raw. If cond doesn't have
7347 integral type, don't add the IFN_ANNOTATE builtin at all.
7349 2014-01-17 Martin Jambor <mjambor@suse.cz>
7352 * ipa-cp.c (prev_edge_clone): New variable.
7353 (grow_next_edge_clone_vector): Renamed to grow_edge_clone_vectors.
7354 Also resize prev_edge_clone vector.
7355 (ipcp_edge_duplication_hook): Also update prev_edge_clone.
7356 (ipcp_edge_removal_hook): New function.
7357 (ipcp_driver): Register ipcp_edge_removal_hook.
7359 2014-01-17 Andrew Pinski <apinski@cavium.com>
7360 Steve Ellcey <sellcey@mips.com>
7363 * config/mips/mips.c (mips_print_operand): Check operand mode instead
7366 2014-01-17 Jeff Law <law@redhat.com>
7369 * passes.def: Reorder pass_copy_prop, pass_unrolli, pass_ccp sequence
7370 so that pass_ccp runs first.
7372 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
7374 * config/i386/i386.c (ix86_lea_outperforms): Use TARGET_XXX.
7375 (ix86_adjust_cost): Use !TARGET_XXX.
7376 (do_reorder_for_imul): Likewise.
7377 (swap_top_of_ready_list): Likewise.
7378 (ix86_sched_reorder): Likewise.
7380 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
7382 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
7383 PROCESSOR_INTEL. Treat like PROCESSOR_GENERIC.
7384 * config/i386/i386.c (intel_memcpy): New. Duplicate slm_memcpy.
7385 (intel_memset): New. Duplicate slm_memset.
7386 (intel_cost): New. Duplicate slm_cost.
7387 (m_INTEL): New macro.
7388 (processor_target_table): Add "intel".
7389 (ix86_option_override_internal): Replace PROCESSOR_SILVERMONT
7390 with PROCESSOR_INTEL for "intel".
7391 (ix86_lea_outperforms): Support PROCESSOR_INTEL. Duplicate
7392 PROCESSOR_SILVERMONT.
7393 (ix86_issue_rate): Likewise.
7394 (ix86_adjust_cost): Likewise.
7395 (ia32_multipass_dfa_lookahead): Likewise.
7396 (swap_top_of_ready_list): Likewise.
7397 (ix86_sched_reorder): Likewise.
7398 (ix86_avoid_lea_for_addr): Check TARGET_AVOID_LEA_FOR_ADDR
7399 instead of TARGET_OPT_AGU.
7400 * config/i386/i386.h (TARGET_INTEL): New.
7401 (TARGET_AVOID_LEA_FOR_ADDR): Likewise.
7402 (processor_type): Add PROCESSOR_INTEL.
7403 * config/i386/x86-tune.def: Support m_INTEL. Duplicate m_SILVERMONT.
7404 Add X86_TUNE_AVOID_LEA_FOR_ADDR.
7406 2014-01-17 Marek Polacek <polacek@redhat.com>
7409 * gimple-fold.c (fold_array_ctor_reference): Don't fold if element
7412 2014-01-17 Richard Biener <rguenther@suse.de>
7414 PR tree-optimization/46590
7415 * opts.c (default_options_table): Add entries for
7416 OPT_fbranch_count_reg, OPT_fmove_loop_invariants and OPT_ftree_pta,
7417 all enabled at -O1 but not for -Og.
7418 * common.opt (fbranch-count-reg): Remove Init(1).
7419 (fmove-loop-invariants): Likewise.
7420 (ftree-pta): Likewise.
7422 2014-01-17 Jakub Jelinek <jakub@redhat.com>
7424 * config/i386/i386.c (ix86_data_alignment): For compatibility with
7425 (incorrect) GCC 4.8 and earlier alignment assumptions ensure we align
7426 decls to at least the GCC 4.8 used alignments.
7429 * tree-nested.c (convert_nonlocal_reference_stmt,
7430 convert_local_reference_stmt): For NAMELIST_DECLs in gimple_bind_vars
7431 of GIMPLE_BIND stmts, adjust associated decls.
7433 2014-01-17 Richard Biener <rguenther@suse.de>
7435 PR tree-optimization/46590
7436 * vec.h (vec<>::bseach): New member function implementing
7437 binary search according to C89 bsearch.
7438 (vec<>::qsort): Avoid calling ::qsort for vectors with sizes 0 or 1.
7439 * tree-ssa-loop-im.c (struct mem_ref): Make stored member a
7440 bitmap pointer again. Make accesses_in_loop a flat array.
7441 (mem_ref_obstack): New global.
7442 (outermost_indep_loop): Adjust for mem_ref->stored changes.
7443 (mark_ref_stored): Likewise.
7444 (ref_indep_loop_p_2): Likewise.
7445 (set_ref_stored_in_loop): New helper function.
7446 (mem_ref_alloc): Allocate mem_refs on the mem_ref_obstack obstack.
7447 (memref_free): Adjust.
7448 (record_mem_ref_loc): Simplify.
7449 (gather_mem_refs_stmt): Adjust.
7450 (sort_locs_in_loop_postorder_cmp): New function.
7451 (analyze_memory_references): Sort accesses_in_loop after
7452 loop postorder number.
7453 (find_ref_loc_in_loop_cmp): New function.
7454 (for_all_locs_in_loop): Find relevant cluster of locs in
7455 accesses_in_loop and iterate without recursion.
7456 (execute_sm): Avoid uninit warning.
7457 (struct ref_always_accessed): Simplify.
7458 (ref_always_accessed::operator ()): Likewise.
7459 (ref_always_accessed_p): Likewise.
7460 (tree_ssa_lim_initialize): Initialize mem_ref_obstack, compute
7461 loop postorder numbers here.
7462 (tree_ssa_lim_finalize): Free mem_ref_obstack and loop postorder
7465 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
7468 * passes.c (rest_of_decl_compilation): Don't call varpool_finalize_decl
7469 on decls for which assemble_alias has been called.
7471 2014-01-17 Nick Clifton <nickc@redhat.com>
7473 * config/msp430/msp430.opt: (mcpu): New option.
7474 * config/msp430/msp430.c (msp430_mcu_name): Use target_mcu.
7475 (msp430_option_override): Parse target_cpu. If the MCU name
7476 matches a generic string, clear target_mcu.
7477 (msp430_attr): Allow numeric interrupt values up to 63.
7478 (msp430_expand_epilogue): No longer invert operand 1 of gen_popm.
7479 * config/msp430/msp430.h (ASM_SPEC): Convert -mcpu into a -mmcu
7481 * config/msp430/t-msp430: (MULTILIB_MATCHES): Remove mcu matches.
7483 * config/msp430/msp430.md (popm): Use %J rather than %I.
7484 (addsi3): Use msp430_nonimmediate_operand for operand 2.
7485 (addhi_cy_i): Use immediate_operand for operand 2.
7486 * doc/invoke.texi: Document -mcpu option.
7488 2014-01-17 Richard Biener <rguenther@suse.de>
7490 PR rtl-optimization/38518
7491 * df.h (df_analyze_loop): Declare.
7492 * df-core.c: Include cfgloop.h.
7493 (df_analyze_1): Split out main part of df_analyze.
7494 (df_analyze): Adjust.
7495 (loop_inverted_post_order_compute): New function.
7496 (loop_post_order_compute): Likewise.
7497 (df_analyze_loop): New function avoiding whole-function
7499 * loop-invariant.c (find_defs): Use df_analyze_loop.
7500 (find_invariants): Adjust.
7501 * loop-iv.c (iv_analysis_loop_init): Use df_analyze_loop.
7503 2014-01-17 Zhenqiang Chen <zhenqiang.chen@arm.com>
7505 * config/arm/arm.c (arm_v7m_tune): Set max_insns_skipped to 2.
7506 (thumb2_final_prescan_insn): Set max to MAX_INSN_PER_IT_BLOCK.
7508 2014-01-16 Ilya Enkovich <ilya.enkovich@intel.com>
7510 * ipa-ref.c (ipa_remove_stmt_references): Fix references
7511 traversal when removing references.
7513 2014-01-16 Jan Hubicka <hubicka@ucw.cz>
7516 * tree.c (get_binfo_at_offset): Look harder for virtual bases.
7518 2014-01-16 Bernd Schmidt <bernds@codesourcery.com>
7521 * reload.c (find_reloads_address_1): Do not use RELOAD_OTHER when
7522 pushing a reload for an autoinc when we had previously reloaded an
7523 inner part of the address.
7525 2014-01-16 Jakub Jelinek <jakub@redhat.com>
7527 * tree-vectorizer.h (struct _loop_vec_info): Add no_data_dependencies
7529 (LOOP_VINFO_NO_DATA_DEPENDENCIES): Define.
7530 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Clear it
7531 when not giving up or versioning for alias only because of
7533 (vect_analyze_data_ref_dependences): Set to true.
7534 * tree-vect-stmts.c (hoist_defs_of_uses): Return false if def_stmt
7536 (vectorizable_load): Use LOOP_VINFO_NO_DATA_DEPENDENCIES instead of
7537 LOOP_REQUIRES_VERSIONING_FOR_ALIAS, add && !nested_in_vect_loop
7541 * expr.c (expand_expr_real_1): Handle init == NULL_TREE.
7544 * config/i386/i386.c (ix86_expand_builtin): If target doesn't satisfy
7545 operand 0 predicate for gathers, use a new pseudo as subtarget.
7547 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
7550 * lra-constraints.c (process_alt_operands): Add printing debug info.
7551 Check absence of input/output reloads for matched operands too.
7553 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
7555 PR rtl-optimization/59835
7556 * ira.c (ira_init_register_move_cost): Increase cost for
7559 2014-01-16 Alan Lawrence <alan.lawrence@arm.com>
7561 * config/arm/arm.opt (mcpu, march, mtune): Make case-insensitive.
7563 2014-01-16 Richard Earnshaw <rearnsha@arm.com>
7566 * aarch64.c (aarch64_split_128bit_move): Don't lookup REGNO on
7567 non-register objects. Use gen_(high/low)part more consistently.
7570 2014-01-16 Michael Meissner <meissner@linux.vnet.ibm.com>
7573 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Add little
7574 endian support, remove tests for WORDS_BIG_ENDIAN.
7575 (p8_mfvsrd_3_<mode>): Likewise.
7576 (reload_gpr_from_vsx<mode>): Likewise.
7577 (reload_gpr_from_vsxsf): Likewise.
7578 (p8_mfvsrd_4_disf): Likewise.
7580 2014-01-16 Richard Biener <rguenther@suse.de>
7582 PR rtl-optimization/46590
7583 * lcm.c (compute_antinout_edge): Use postorder iteration.
7584 (compute_laterin): Use inverted postorder iteration.
7586 2014-01-16 Nick Clifton <nickc@redhat.com>
7589 * varasm.c (output_constant): Return the number of bytes actually
7591 (output_constructor_array_range): Update the field size with the
7592 number of bytes emitted by output_constant.
7593 (output_constructor_regular_field): Likewise. Also do not
7594 complain if the total number of bytes emitted is now greater
7595 than the expected fieldpos.
7596 * output.h (output_constant): Update prototype and descriptive comment.
7598 2014-01-16 Marek Polacek <polacek@redhat.com>
7601 * cgraph.c (gimple_check_call_args): Don't use DECL_ARG_TYPE if
7602 it is error_mark_node.
7604 2014-01-15 Uros Bizjak <ubizjak@gmail.com>
7606 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use
7607 VALID_AVX256_REG_OR_OI_MODE.
7609 2014-01-15 Pat Haugen <pthaugen@us.ibm.com>
7611 * config/rs6000/rs6000.c (rs6000_output_function_prologue): Check if
7612 current procedure should be profiled.
7614 2014-01-15 Andrew Pinski <apinski@cavium.com>
7616 * config/aarch64/aarch64.c (aarch64_register_move_cost): Correct cost
7617 of moving from/to the STACK_REG register class.
7619 2014-01-15 Richard Henderson <rth@redhat.com>
7622 * reginfo.c (global_regs_decl): Globalize.
7623 * rtl.h (global_regs_decl): Declare.
7624 * ira.c (do_reload): Diagnose frame_pointer_needed and it
7625 reserved via global_regs.
7627 2014-01-15 Teresa Johnson <tejohnson@google.com>
7629 * tree-ssa-sccvn.c (visit_reference_op_call): Handle NULL vdef.
7631 2014-01-15 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
7633 * config/rs6000/altivec.md (mulv8hi3): Explicitly generate vmulesh
7634 and vmulosh rather than call gen_vec_widen_smult_*.
7635 (vec_widen_umult_even_v16qi): Test VECTOR_ELT_ORDER_BIG rather
7636 than BYTES_BIG_ENDIAN to determine use of even or odd instruction.
7637 (vec_widen_smult_even_v16qi): Likewise.
7638 (vec_widen_umult_even_v8hi): Likewise.
7639 (vec_widen_smult_even_v8hi): Likewise.
7640 (vec_widen_umult_odd_v16qi): Likewise.
7641 (vec_widen_smult_odd_v16qi): Likewise.
7642 (vec_widen_umult_odd_v8hi): Likewise.
7643 (vec_widen_smult_odd_v8hi): Likewise.
7644 (vec_widen_umult_hi_v16qi): Explicitly generate vmuleub and
7645 vmuloub rather than call gen_vec_widen_umult_*.
7646 (vec_widen_umult_lo_v16qi): Likewise.
7647 (vec_widen_smult_hi_v16qi): Explicitly generate vmulesb and
7648 vmulosb rather than call gen_vec_widen_smult_*.
7649 (vec_widen_smult_lo_v16qi): Likewise.
7650 (vec_widen_umult_hi_v8hi): Explicitly generate vmuleuh and vmulouh
7651 rather than call gen_vec_widen_umult_*.
7652 (vec_widen_umult_lo_v8hi): Likewise.
7653 (vec_widen_smult_hi_v8hi): Explicitly gnerate vmulesh and vmulosh
7654 rather than call gen_vec_widen_smult_*.
7655 (vec_widen_smult_lo_v8hi): Likewise.
7657 2014-01-15 Jeff Law <law@redhat.com>
7659 PR tree-optimization/59747
7660 * ree.c (find_and_remove_re): Properly handle case where a second
7661 eliminated extension requires widening a copy created for elimination
7662 of a prior extension.
7663 (combine_set_extension): Ensure that the number of hard regs needed
7664 for a destination register does not change when we widen it.
7666 2014-01-15 Sebastian Huber <sebastian.huber@embedded-brains.de>
7668 * config.gcc (*-*-rtems*): Add t-rtems to tmake_file.
7669 (arm*-*-uclinux*eabi*): Do not override an existing tmake_file.
7670 (arm*-*-eabi* | arm*-*-symbianelf* | arm*-*-rtems*): Likwise.
7671 (arm*-*-rtems*): Use t-rtems from existing tmake_file.
7672 (avr-*-rtems*): Likewise.
7673 (bfin*-rtems*): Likewise.
7674 (moxie-*-rtems*): Likewise.
7675 (h8300-*-rtems*): Likewise.
7676 (i[34567]86-*-rtems*): Likewise.
7677 (lm32-*-rtems*): Likewise.
7678 (m32r-*-rtems*): Likewise.
7679 (m68k-*-rtems*): Likewise.
7680 (microblaze*-*-rtems*): Likewise.
7681 (mips*-*-rtems*): Likewise.
7682 (powerpc-*-rtems*): Likewise.
7683 (sh-*-rtems*): Likewise.
7684 (sparc-*-rtems*): Likewise.
7685 (sparc64-*-rtems*): Likewise.
7686 (v850-*-rtems*): Likewise.
7687 (m32c-*-rtems*): Likewise.
7689 2014-01-15 Vladimir Makarov <vmakarov@redhat.com>
7691 PR rtl-optimization/59511
7692 * ira.c (ira_init_register_move_cost): Use memory costs for some
7693 cases of register move cost calculations.
7694 * lra-constraints.c (lra_constraints): Use REG_FREQ_FROM_BB
7695 instead of BB frequency.
7696 * lra-coalesce.c (move_freq_compare_func, lra_coalesce): Ditto.
7697 * lra-assigns.c (find_hard_regno_for): Ditto.
7699 2014-01-15 Richard Biener <rguenther@suse.de>
7701 PR tree-optimization/59822
7702 * tree-vect-stmts.c (hoist_defs_of_uses): New function.
7703 (vectorizable_load): Use it to hoist defs of uses of invariant
7704 loads out of the loop.
7706 2014-01-15 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
7707 Kugan Vivekanandarajah <kuganv@linaro.org>
7710 * config/aarch64/aarch64.c (aarch64_build_constant): Fix incorrect
7713 2014-01-15 Richard Biener <rguenther@suse.de>
7715 PR rtl-optimization/59802
7716 * lcm.c (compute_available): Use inverted postorder to seed
7717 the initial worklist.
7719 2014-01-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7722 * config/s390/s390.c (s390_preferred_reload_class): Don't return
7723 ADDR_REGS for invalid symrefs in non-PIC code.
7725 2014-01-15 Jakub Jelinek <jakub@redhat.com>
7728 * builtins.c (determine_block_size): Initialize *probable_max_size
7729 even if len_rtx is CONST_INT.
7731 2014-01-14 Andrew Pinski <apinski@cavium.com>
7733 * config/aarch64/aarch64-protos.h (tune_params): Add issue_rate.
7734 * config/aarch64/aarch64.c (generic_tunings): Add issue rate of 2.
7735 (cortexa53_tunings): Likewise.
7736 (aarch64_sched_issue_rate): New function.
7737 (TARGET_SCHED_ISSUE_RATE): Define.
7739 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
7741 * ira-costs.c (find_costs_and_classes): Add missed
7742 ira_init_register_move_cost_if_necessary.
7744 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
7747 * config/arm/arm.c (arm_coproc_mem_operand): Add lra_in_progress.
7749 2014-01-14 H.J. Lu <hongjiu.lu@intel.com>
7752 * config/i386/i386.c (type_natural_mode): Add a bool parameter
7753 to indicate if type is used for function return value. Warn ABI
7754 change if the vector mode isn't available for function return value.
7755 (ix86_function_arg_advance): Pass false to type_natural_mode.
7756 (ix86_function_arg): Likewise.
7757 (ix86_gimplify_va_arg): Likewise.
7758 (function_arg_32): Don't warn ABI change.
7759 (ix86_function_value): Pass true to type_natural_mode.
7760 (ix86_return_in_memory): Likewise.
7761 (ix86_struct_value_rtx): Removed.
7762 (TARGET_STRUCT_VALUE_RTX): Likewise.
7764 2014-01-14 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7766 * jump.c (redirect_jump_2): Remove REG_CROSSING_JUMP notes when
7767 converting a conditional jump into a conditional return.
7769 2014-01-14 Richard Biener <rguenther@suse.de>
7771 PR tree-optimization/58921
7772 PR tree-optimization/59006
7773 * tree-vect-loop-manip.c (vect_loop_versioning): Remove code
7774 hoisting invariant stmts.
7775 * tree-vect-stmts.c (vectorizable_load): Insert the splat of
7776 invariant loads on the preheader edge if possible.
7778 2014-01-14 Joey Ye <joey.ye@arm.com>
7780 * doc/plugin.texi (Building GCC plugins): Update to C++.
7782 2014-01-14 Kirill Yukhin <kirill.yukhin@intel.com>
7784 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): New.
7785 (_mm_rcp28_round_ss): Ditto.
7786 (_mm_rsqrt28_round_sd): Ditto.
7787 (_mm_rsqrt28_round_ss): Ditto.
7788 (_mm_rcp28_sd): Ditto.
7789 (_mm_rcp28_ss): Ditto.
7790 (_mm_rsqrt28_sd): Ditto.
7791 (_mm_rsqrt28_ss): Ditto.
7792 * config/i386/avx512fintrin.h (_mm512_stream_load_si512): Ditto.
7793 * config/i386/i386-builtin-types.def (V8DI_FTYPE_PV8DI): Ditto.
7794 * config/i386/i386.c (IX86_BUILTIN_MOVNTDQA512): Ditto.
7795 (IX86_BUILTIN_RCP28SD): Ditto.
7796 (IX86_BUILTIN_RCP28SS): Ditto.
7797 (IX86_BUILTIN_RSQRT28SD): Ditto.
7798 (IX86_BUILTIN_RSQRT28SS): Ditto.
7799 (bdesc_special_args): Define __builtin_ia32_movntdqa512,
7800 __builtin_ia32_rcp28sd_round, __builtin_ia32_rcp28ss_round,
7801 __builtin_ia32_rsqrt28sd_round, __builtin_ia32_rsqrt28ss_round.
7802 (ix86_expand_special_args_builtin): Expand new FTYPE.
7803 * config/i386/sse.md (define_mode_attr "sse4_1_avx2"): Expand to V8DI.
7804 (srcp14<mode>): Make insn unary.
7805 (avx512f_vmscalef<mode><round_name>): Use substed predicate.
7806 (avx512f_sgetexp<mode><round_saeonly_name>): Ditto.
7807 (avx512f_rndscale<mode><round_saeonly_name>): Ditto.
7808 (<sse4_1_avx2>_movntdqa): Extend to 512 bits.
7809 (avx512er_exp2<mode><mask_name><round_saeonly_name>):
7810 Fix rounding: make it SAE only.
7811 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
7813 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
7815 (avx512er_vmrcp28<mode><round_saeonly_name>): Ditto.
7816 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Ditto.
7817 (avx512f_getmant<mode><mask_name><round_saeonly_name>): Ditto.
7818 * config/i386/subst.md (round_saeonly_mask_scalar_operand3): Remove.
7819 (round_saeonly_mask_scalar_operand4): Ditto.
7820 (round_saeonly_mask_scalar_op3): Ditto.
7821 (round_saeonly_mask_scalar_op4): Ditto.
7823 2014-01-13 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7825 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
7826 Implement -maltivec=be for vec_insert and vec_extract.
7828 2014-01-10 DJ Delorie <dj@redhat.com>
7830 * config/msp430/msp430.md (call_internal): Don't allow memory
7831 references with SP as the base register.
7832 (call_value_internal): Likewise.
7833 * config/msp430/constraints.md (Yc): New. For memory references
7834 that don't use SP as a base register.
7836 * config/msp430/msp430.c (msp430_print_operand): Add 'J' to mean
7837 "an integer without a # prefix"
7838 * config/msp430/msp430.md (epilogue_helper): Use it.
7840 2014-01-13 Jakub Jelinek <jakub@redhat.com>
7843 * config/i386/i386.c (ix86_vectorize_builtin_gather): Uncomment
7844 AVX512F gather builtins.
7845 * tree-vect-stmts.c (vectorizable_mask_load_store): For now punt
7846 on gather decls with INTEGER_TYPE masktype.
7847 (vectorizable_load): For INTEGER_TYPE masktype, put the INTEGER_CST
7848 directly into the builtin rather than hoisting it before loop.
7850 PR tree-optimization/59387
7851 * tree-scalar-evolution.c: Include gimple-fold.h and gimplify-me.h.
7852 (scev_const_prop): If folded_casts and type has undefined overflow,
7853 use force_gimple_operand instead of force_gimple_operand_gsi and
7854 for each added stmt if it is assign with
7855 arith_code_with_undefined_signed_overflow, call
7856 rewrite_to_defined_overflow.
7857 * tree-ssa-loop-im.c: Don't include gimplify-me.h, include
7858 gimple-fold.h instead.
7859 (arith_code_with_undefined_signed_overflow,
7860 rewrite_to_defined_overflow): Moved to ...
7861 * gimple-fold.c (arith_code_with_undefined_signed_overflow,
7862 rewrite_to_defined_overflow): ... here. No longer static.
7863 Include gimplify-me.h.
7864 * gimple-fold.h (arith_code_with_undefined_signed_overflow,
7865 rewrite_to_defined_overflow): New prototypes.
7867 2014-01-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7869 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Fix typo in description.
7871 2014-01-13 Eric Botcazou <ebotcazou@adacore.com>
7873 * builtins.c (get_object_alignment_2): Minor tweak.
7874 * tree-ssa-loop-ivopts.c (may_be_unaligned_p): Rewrite.
7876 2014-01-13 Christian Bruel <christian.bruel@st.com>
7878 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Unroll small sizes and
7879 optimized non constant lengths.
7881 2014-01-13 Jakub Jelinek <jakub@redhat.com>
7884 * omp-low.c (expand_omp_atomic_pipeline): Expand the initial
7885 load as __atomic_load_N if possible.
7887 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
7889 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): Remove
7891 (rs6000_expand_builtin): Adjust call.
7893 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
7896 * config/rs6000/rs6000.h (SWITCHABLE_TARGET): Define.
7897 * config/rs6000/rs6000.c: Include target-globals.h.
7898 (rs6000_set_current_function): Instead of doing target_reinit
7899 unconditionally, use save_target_globals_default_opts and
7900 restore_target_globals.
7902 * config/rs6000/rs6000-builtin.def (mffs, mtfsf): Add builtins for
7904 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): New.
7905 (rs6000_expand_builtin): Handle mffs and mtfsf.
7906 (rs6000_init_builtins): Define mffs and mtfsf.
7907 * config/rs6000/rs6000.md (UNSPECV_MFFS, UNSPECV_MTFSF): New constants.
7908 (rs6000_mffs): New pattern.
7909 (rs6000_mtfsf): New pattern.
7911 2014-01-11 Bin Cheng <bin.cheng@arm.com>
7913 * tree-ssa-loop-ivopts.c (iv_ca_narrow): New parameter.
7914 Start narrowing with START. Apply candidate-use pair
7915 and check overall cost in narrowing.
7916 (iv_ca_prune): Pass new argument.
7918 2014-01-10 Jeff Law <law@redhat.com>
7921 * ree.c (combine_reaching_defs): Ensure the defining statement
7922 occurs before the extension when optimizing extensions with
7923 different source and destination hard registers.
7925 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
7928 * ipa-devirt.c (build_type_inheritance_graph): Also add types of
7929 vtables into the type inheritance graph.
7931 2014-01-10 Jakub Jelinek <jakub@redhat.com>
7933 PR rtl-optimization/59754
7934 * ree.c (combine_reaching_defs): Disallow !SCALAR_INT_MODE_P
7935 modes in the REGNO != REGNO case.
7937 2014-01-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7939 * config/rs6000/rs6000-builtin.def: Fix pasto for VPKSDUS.
7941 2014-01-10 Jakub Jelinek <jakub@redhat.com>
7943 PR tree-optimization/59745
7944 * tree-predcom.c (tree_predictive_commoning_loop): Call
7945 free_affine_expand_cache if giving up because components is NULL.
7947 * target-globals.c (save_target_globals): Allocate < 4KB structs using
7948 GC in payload of target_globals struct instead of allocating them on
7949 the heap and the larger structs separately using GC.
7950 * target-globals.h (struct target_globals): Make regs, hard_regs,
7951 reload, expmed, ira, ira_int and lra_fields GTY((atomic)) instead
7952 of GTY((skip)) and change type to void *.
7953 (reset_target_globals): Cast loads from those fields to corresponding
7956 2014-01-10 Steve Ellcey <sellcey@mips.com>
7959 * Makefile.in (PLUGIN_HEADERS): Add gimplify.h, gimple-iterator.h,
7960 gimple-ssa.h, fold-const.h, tree-cfg.h, tree-into-ssa.h,
7961 tree-ssanames.h, print-tree.h, varasm.h, and context.h.
7963 2014-01-10 Richard Earnshaw <rearnsha@arm.com>
7966 * aarch64-modes.def (CC_Zmode): New flags mode.
7967 * aarch64.c (aarch64_select_cc_mode): Only allow NEG when the condition
7968 represents an equality.
7969 (aarch64_get_condition_code): Handle CC_Zmode.
7970 * aarch64.md (compare_neg<mode>): Restrict to equality operations.
7972 2014-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7974 * config/s390/s390.c (s390_expand_tbegin): Remove jump over CC
7975 extraction in good case.
7977 2014-01-10 Richard Biener <rguenther@suse.de>
7979 PR tree-optimization/59374
7980 * tree-vect-slp.c (vect_slp_analyze_bb_1): Move dependence
7981 checking after SLP discovery. Mark stmts not participating
7982 in any SLP instance properly.
7984 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7986 * config/arm/arm.c (arm_new_rtx_costs): Use destination mode
7987 when handling a SET rtx.
7989 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7991 * config/arm/arm-cores.def (cortex-a53): Specify FL_CRC32.
7992 (cortex-a57): Likewise.
7993 (cortex-a57.cortex-a53): Likewise. Remove redundant flags.
7995 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7997 * config/arm/arm.c (arm_init_iwmmxt_builtins): Skip
7998 non-iwmmxt builtins.
8000 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
8004 * ipa-devirt.c record_target_from_binfo): Take as argument
8005 stack of binfos and lookup matching one for virtual inheritance.
8006 (possible_polymorphic_call_targets_1): Update.
8008 2014-01-10 Huacai Chen <chenhc@lemote.com>
8010 * config/mips/driver-native.c (host_detect_local_cpu): Handle new
8011 kernel strings for Loongson-2E/2F/3A.
8013 2014-01-10 Jakub Jelinek <jakub@redhat.com>
8016 * tree-vect-data-refs.c (vect_analyze_data_refs): Check
8017 is_gimple_call before calling gimple_call_internal_p.
8019 2014-01-09 Steve Ellcey <sellcey@mips.com>
8021 * Makefile.in (TREE_FLOW_H): Remove.
8022 (TREE_SSA_H): Add file names from tree-flow.h.
8023 * doc/tree-ssa.texi (Annotations): Remove reference to tree-flow.h
8024 * tree.h: Remove tree-flow.h reference.
8025 * hash-table.h: Remove tree-flow.h reference.
8026 * tree-ssa-loop-niter.c (dump_affine_iv): Replace tree-flow.h
8027 reference with tree-ssa-loop.h.
8029 2014-01-09 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8031 * doc/invoke.texi: Add -maltivec={be,le} options, and document
8032 default element-order behavior for -maltivec.
8033 * config/rs6000/rs6000.opt: Add -maltivec={be,le} options.
8034 * config/rs6000/rs6000.c (rs6000_option_override_internal): Ensure
8035 that -maltivec={le,be} implies -maltivec; disallow -maltivec=le
8036 when targeting big endian, at least for now.
8037 * config/rs6000/rs6000.h: Add #define of VECTOR_ELT_ORDER_BIG.
8039 2014-01-09 Jakub Jelinek <jakub@redhat.com>
8042 * cfgexpand.c (expand_one_var): For SSA_NAMEs, if the underlying
8043 var satisfies use_register_for_decl, just take into account type
8044 alignment, rather than decl alignment.
8046 PR tree-optimization/59622
8047 * gimple-fold.c (gimple_fold_call): Fix a typo in message. For
8048 __builtin_unreachable replace the OBJ_TYPE_REF call with a call to
8049 __builtin_unreachable and add if needed a setter of the lhs SSA_NAME.
8050 Don't devirtualize for inplace at all. For targets.length () == 1,
8051 if the call is noreturn and cfun isn't in SSA form yet, clear lhs.
8053 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
8055 * config/i386/i386.md (cpu): Remove the unused btver1.
8057 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
8059 * gdbasan.in: Put a breakpoint on __sanitizer::Report.
8061 2014-01-09 Jakub Jelinek <jakub@redhat.com>
8064 * tree-core.h (struct target_globals): New forward declaration.
8065 (struct tree_target_option): Add globals field.
8066 * tree.h (TREE_TARGET_GLOBALS): Define.
8067 (prepare_target_option_nodes_for_pch): New prototype.
8068 * target-globals.h (struct target_globals): Define even if
8070 * tree.c (prepare_target_option_node_for_pch,
8071 prepare_target_option_nodes_for_pch): New functions.
8072 * config/i386/i386.h (SWITCHABLE_TARGET): Define.
8073 * config/i386/i386.c: Include target-globals.h.
8074 (ix86_set_current_function): Instead of doing target_reinit
8075 unconditionally, use save_target_globals_default_opts and
8076 restore_target_globals.
8078 2014-01-09 Richard Biener <rguenther@suse.de>
8080 PR tree-optimization/59715
8081 * tree-cfg.h (split_critical_edges): Declare.
8082 * tree-cfg.c (split_critical_edges): Export.
8083 * tree-ssa-sink.c (execute_sink_code): Split critical edges.
8085 2014-01-09 Max Ostapenko <m.ostapenko@partner.samsung.com>
8087 * cfgexpand.c (expand_stack_vars): Optionally disable
8088 asan stack protection.
8089 (expand_used_vars): Likewise.
8090 (partition_stack_vars): Likewise.
8091 * asan.c (asan_emit_stack_protection): Optionally disable
8092 after return stack usage.
8093 (instrument_derefs): Optionally disable memory access instrumentation.
8094 (instrument_builtin_call): Likewise.
8095 (instrument_strlen_call): Likewise.
8096 (asan_protect_global): Optionally disable global variables protection.
8097 * doc/invoke.texi: Added doc for new options.
8098 * params.def: Added new options.
8099 * params.h: Likewise.
8101 2014-01-09 Jakub Jelinek <jakub@redhat.com>
8103 PR rtl-optimization/59724
8104 * ifcvt.c (cond_exec_process_if_block): Don't call
8105 flow_find_head_matching_sequence with 0 longest_match.
8106 * cfgcleanup.c (flow_find_head_matching_sequence): Count even
8107 non-active insns if !stop_after.
8108 (try_head_merge_bb): Revert 2014-01-07 changes.
8110 2014-01-08 Jeff Law <law@redhat.com>
8112 * ree.c (get_sub_rtx): New function, extracted from...
8113 (merge_def_and_ext): Here.
8114 (combine_reaching_defs): Use get_sub_rtx.
8116 2014-01-08 Eric Botcazou <ebotcazou@adacore.com>
8118 * cgraph.h (varpool_variable_node): Do not choke on null node.
8120 2014-01-08 Catherine Moore <clm@codesourcery.com>
8122 * config/mips/mips.md (simple_return): Attempt to use JRC
8124 * config/mips/mips.h (MIPS_CALL): Attempt to use JALS for microMIPS.
8126 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
8128 PR rtl-optimization/59137
8129 * reorg.c (steal_delay_list_from_target): Call update_block for
8131 (steal_delay_list_from_fallthrough, relax_delay_slots): Likewise.
8133 2014-01-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8135 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Remove
8136 two duplicate entries.
8138 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
8141 2012-10-07 Richard Sandiford <rdsandiford@googlemail.com>
8143 * config/mips/mips.c (mips_truncated_op_cost): New function.
8144 (mips_rtx_costs): Adjust test for BADDU.
8145 * config/mips/mips.md (*baddu_di<mode>): Push truncates to operands.
8147 2012-10-02 Richard Sandiford <rdsandiford@googlemail.com>
8149 * config/mips/mips.md (*baddu_si_eb, *baddu_si_el): Merge into...
8150 (*baddu_si): ...this new pattern.
8152 2014-01-08 Jakub Jelinek <jakub@redhat.com>
8155 * ipa-prop.c (ipa_analyze_params_uses): Ignore uses in debug stmts.
8157 2014-01-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
8160 * expr.h (expand_expr_real, expand_expr_real_1): Add new parameter
8162 (expand_expr, expand_normal): Adjust.
8163 * expr.c (expand_expr_real, expand_expr_real_1): Add new parameter
8164 inner_reference_p. Use inner_reference_p to expand inner references.
8165 (store_expr): Adjust.
8166 * cfgexpand.c (expand_call_stmt): Adjust.
8168 2014-01-08 Rong Xu <xur@google.com>
8170 * gcov-io.c (gcov_var): Move from gcov-io.h.
8171 (gcov_position): Ditto.
8172 (gcov_is_error): Ditto.
8173 (gcov_rewrite): Ditto.
8174 * gcov-io.h: Refactor. Move gcov_var to gcov-io.h, and libgcov
8175 only part to libgcc/libgcov.h.
8177 2014-01-08 Marek Polacek <polacek@redhat.com>
8180 * omp-low.c (simd_clone_adjust): Don't crash if def is NULL.
8182 2014-01-08 Marek Polacek <polacek@redhat.com>
8185 * ubsan.c (ubsan_type_descriptor): Call strip_array_types on type2.
8187 2014-01-08 Jakub Jelinek <jakub@redhat.com>
8189 PR rtl-optimization/59649
8190 * stor-layout.c (get_mode_bounds): For BImode return
8191 0 and STORE_FLAG_VALUE.
8193 2014-01-08 Richard Biener <rguenther@suse.de>
8196 * gimple.h (is_gimple_builtin_call): Remove.
8197 (gimple_builtin_call_types_compatible_p): New.
8198 (gimple_call_builtin_p): New overload.
8199 * gimple.c (is_gimple_builtin_call): Remove.
8200 (validate_call): Rename to ...
8201 (gimple_builtin_call_types_compatible_p): ... this and export. Also
8203 (validate_type): New static function.
8204 (gimple_call_builtin_p): New overload and adjust.
8205 * gimple-fold.c (gimple_fold_builtin): Fold the return value.
8206 (gimple_fold_call): Likewise. Use gimple_call_builtin_p.
8207 (gimple_fold_stmt_to_constant_1): Likewise.
8208 * tsan.c (instrument_gimple): Use gimple_call_builtin_p.
8210 2014-01-08 Richard Biener <rguenther@suse.de>
8213 * gimplify.c (gimplify_expr): Gimplify register-register type
8214 VIEW_CONVERT_EXPRs to separate stmts.
8216 2014-01-07 Jeff Law <law@redhat.com>
8219 * ree.c (combine_set_extension): Handle case where source
8220 and destination registers in an extension insn are different.
8221 (combine_reaching_defs): Allow source and destination registers
8222 in extension to be different under limited circumstances.
8223 (add_removable_extension): Remove restriction that the
8224 source and destination registers in the extension are the same.
8225 (find_and_remove_re): Emit a copy from the extension's
8226 destination to its source after the defining insn if
8227 the source and destination registers are different.
8230 * ifcvt.c (merge_if_block): If we are merging a block with more than
8231 one successor with a block with no successors, remove any BARRIER
8232 after the second block.
8234 2014-01-07 Dan Xio Qiang <ziyan01@163.com>
8236 * hw-doloop.c (reorg_loops): Release the bitmap obstack.
8238 2014-01-07 John David Anglin <danglin@gcc.gnu.org>
8241 * config/pa/pa.c (pa_legitimate_address_p): Return false before reload
8242 for 14-bit register offsets when INT14_OK_STRICT is false.
8244 2014-01-07 Roland Stigge <stigge@antcom.de>
8245 Michael Meissner <meissner@linux.vnet.ibm.com>
8248 * config/rs6000/rs6000.c (rs6000_legitimate_offset_address_p):
8249 Only check TFmode for SPE constants. Don't check TImode or TDmode.
8251 2014-01-07 James Greenhalgh <james.greenhalgh@arm.com>
8253 * config/aarch64/aarch64-elf.h (ASM_SPEC): Remove identity spec for
8256 2014-01-07 Yufeng Zhang <yufeng.zhang@arm.com>
8258 * config/arm/arm.c (arm_expand_neon_args): Call expand_expr
8259 with EXPAND_MEMORY for NEON_ARG_MEMORY; check if the returned
8260 rtx is const0_rtx or not.
8262 2014-01-07 Richard Sandiford <rdsandiford@googlemail.com>
8265 * target-globals.c (save_target_globals): Remove this_fn_optab
8267 * toplev.c: Include optabs.h.
8268 (target_reinit): Temporarily restore the global options if another
8269 set of options are in force.
8271 2014-01-07 Jakub Jelinek <jakub@redhat.com>
8273 PR rtl-optimization/58668
8274 * cfgcleanup.c (flow_find_cross_jump): Don't count
8275 any jumps if dir_p is NULL. Remove p1 variable, use active_insn_p
8276 to determine what is counted.
8277 (flow_find_head_matching_sequence): Use active_insn_p to determine
8279 (try_head_merge_bb): Adjust for the flow_find_head_matching_sequence
8281 * ifcvt.c (count_bb_insns): Use active_insn_p && !JUMP_P to
8282 determine what is counted.
8284 PR tree-optimization/59643
8285 * tree-predcom.c (split_data_refs_to_components): If one dr is
8286 read and one write, determine_offset fails and the write isn't
8287 in the bad component, just put the read into the bad component.
8289 2014-01-07 Mike Stump <mikestump@comcast.net>
8290 Jakub Jelinek <jakub@redhat.com>
8293 * tree-core.h (struct tree_optimization_option): Change optabs
8294 type from unsigned char * to void *.
8295 * optabs.c (init_tree_optimization_optabs): Adjust
8296 TREE_OPTIMIZATION_OPTABS initialization.
8298 2014-01-06 Jakub Jelinek <jakub@redhat.com>
8301 * config/i386/i386.h (struct machine_function): Add
8302 no_drap_save_restore field.
8303 * config/i386/i386.c (ix86_save_reg): Use
8304 !cfun->machine->no_drap_save_restore instead of
8305 crtl->stack_realign_needed.
8306 (ix86_finalize_stack_realign_flags): Don't clear drap_reg unless
8307 this function clears frame_pointer_needed. Set
8308 cfun->machine->no_drap_save_restore if clearing frame_pointer_needed
8309 and DRAP reg is needed.
8311 2014-01-06 Marek Polacek <polacek@redhat.com>
8314 * doc/implement-c.texi: Mention that other integer types are
8315 permitted as bit-field types in strictly conforming mode.
8317 2014-01-06 Felix Yang <fei.yang0953@gmail.com>
8319 * modulo-sched.c (schedule_reg_moves): Clear distance1_uses if it
8322 2014-01-06 Richard Earnshaw <rearnsha@arm.com>
8324 * aarch64.c (aarch64_rtx_costs): Fix cost calculation for MADD.
8326 2014-01-06 Martin Jambor <mjambor@suse.cz>
8329 * ipa-cp.c (ipcp_discover_new_direct_edges): Changed param_index type
8331 * ipa-prop.c (ipa_print_node_params): Fix indentation.
8333 2014-01-06 Eric Botcazou <ebotcazou@adacore.com>
8337 * var-tracking.c (add_stores): Preserve the value of the source even if
8338 we don't record the store.
8340 2014-01-06 Terry Guo <terry.guo@arm.com>
8342 * config.gcc (arm*-*-*): Check --with-arch against arm-arches.def.
8344 2014-01-05 Iain Sandoe <iain@codesourcery.com>
8347 * config/darwin.c (darwin_function_section): Adjust return values to
8348 correspond to optimisation changes made in r206070.
8350 2014-01-05 Uros Bizjak <ubizjak@gmail.com>
8352 * config/i386/i386.c (ix86_data_alignment): Calculate max_align
8353 from prefetch_block tune setting.
8354 (nocona_cost): Correct size of prefetch block to 64.
8356 2014-01-04 Eric Botcazou <ebotcazou@adacore.com>
8358 * config/arm/arm.c (arm_get_frame_offsets): Revamp long lines.
8359 (arm_expand_epilogue_apcs_frame): Take into account the number of bytes
8360 used to save the static chain register in the computation of the offset
8361 from which the FP registers need to be restored.
8363 2014-01-04 Jakub Jelinek <jakub@redhat.com>
8365 PR tree-optimization/59519
8366 * tree-vect-loop-manip.c (slpeel_update_phi_nodes_for_guard1): Don't
8367 ICE if get_current_def (current_new_name) is already non-NULL, as long
8368 as it is a phi result of some other phi in *new_exit_bb that has
8371 * config/i386/sse.md (avx512f_load<mode>_mask): Emit vmovup{s,d}
8372 or vmovdqu* for misaligned_operand.
8373 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>,
8374 <sse2_avx_avx512f>_loaddqu<mode><mask_name>): Handle <mask_applied>.
8375 * config/i386/i386.c (ix86_expand_special_args_builtin): Set
8376 aligned_mem for AVX512F masked aligned load and store builtins and for
8379 2014-01-03 Bingfeng Mei <bmei@broadcom.com>
8381 PR tree-optimization/59651
8382 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
8383 Address range for negative step should be added by TYPE_SIZE_UNIT.
8385 2014-01-03 Andreas Schwab <schwab@linux-m68k.org>
8387 * config/m68k/m68k.c (handle_move_double): Handle pushes with
8388 overlapping registers also for registers other than the stack pointer.
8390 2014-01-03 Marek Polacek <polacek@redhat.com>
8393 * doc/extend.texi: Fix the return value of __builtin_FUNCTION and
8396 2014-01-03 Jakub Jelinek <jakub@redhat.com>
8399 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Don't consider
8402 * config/i386/i386.md (MODE_SIZE): New mode attribute.
8403 (push splitter): Use <P:MODE_SIZE> instead of
8404 GET_MODE_SIZE (<P:MODE>mode).
8405 (lea splitter): Use <MODE_SIZE> instead of GET_MODE_SIZE (<MODE>mode).
8406 (mov -1, reg peephole2): Likewise.
8407 * config/i386/sse.md (*mov<mode>_internal,
8408 <sse>_storeu<ssemodesuffix><avxsizesuffix>,
8409 <sse2_avx_avx512f>_storedqu<mode>, <sse>_andnot<mode>3,
8410 *<code><mode>3, *andnot<mode>3<mask_name>,
8411 <mask_codefor><code><mode>3<mask_name>): Likewise.
8412 * config/i386/subst.md (mask_mode512bit_condition,
8413 sd_mask_mode512bit_condition): Likewise.
8415 2014-01-02 Xinliang David Li <davidxl@google.com>
8417 PR tree-optimization/59303
8418 * tree-ssa-uninit.c (is_use_properly_guarded): Main cleanup.
8419 (dump_predicates): Better output format.
8420 (pred_equal_p): New function.
8421 (is_neq_relop_p): Ditto.
8422 (is_neq_zero_form_p): Ditto.
8423 (pred_expr_equal_p): Ditto.
8424 (pred_neg_p): Ditto.
8425 (simplify_pred): Ditto.
8426 (simplify_preds_2): Ditto.
8427 (simplify_preds_3): Ditto.
8428 (simplify_preds_4): Ditto.
8429 (simplify_preds): Ditto.
8431 (push_to_worklist): Ditto.
8432 (get_pred_info_from_cmp): Ditto.
8433 (is_degenerated_phi): Ditto.
8434 (normalize_one_pred_1): Ditto.
8435 (normalize_one_pred): Ditto.
8436 (normalize_one_pred_chain): Ditto.
8437 (normalize_preds): Ditto.
8438 (normalize_cond_1): Remove function.
8439 (normalize_cond): Ditto.
8440 (is_gcond_subset_of): Ditto.
8441 (is_subset_of_any): Ditto.
8442 (is_or_set_subset_of): Ditto.
8443 (is_and_set_subset_of): Ditto.
8444 (is_norm_cond_subset_of): Ditto.
8445 (pred_chain_length_cmp): Ditto.
8446 (convert_control_dep_chain_into_preds): Type change.
8447 (find_predicates): Ditto.
8448 (find_def_preds): Ditto.
8449 (destroy_predicates_vecs): Ditto.
8450 (find_matching_predicates_in_rest_chains): Ditto.
8451 (use_pred_not_overlap_with_undef_path_pred): Ditto.
8452 (is_pred_expr_subset): Ditto.
8453 (is_pred_chain_subset_of): Ditto.
8454 (is_included_in): Ditto.
8455 (is_superset_of): Ditto.
8457 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
8459 Update copyright years.
8461 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
8463 * common/config/arc/arc-common.c, config/arc/arc-modes.def,
8464 config/arc/arc-protos.h, config/arc/arc.c, config/arc/arc.h,
8465 config/arc/arc.md, config/arc/arc.opt,
8466 config/arm/arm_neon_builtins.def, config/arm/crypto.def,
8467 config/i386/avx512cdintrin.h, config/i386/avx512erintrin.h,
8468 config/i386/avx512fintrin.h, config/i386/avx512pfintrin.h,
8469 config/i386/btver2.md, config/i386/shaintrin.h, config/i386/slm.md,
8470 config/linux-protos.h, config/linux.c, config/winnt-c.c,
8471 diagnostic-color.c, diagnostic-color.h, gimple-ssa-isolate-paths.c,
8472 vtable-verify.c, vtable-verify.h: Use the standard form for the
8475 2014-01-02 Tobias Burnus <burnus@net-b.de>
8477 * gcc.c (process_command): Update copyright notice dates.
8478 * gcov-dump.c: Ditto.
8480 * doc/cpp.texi: Bump @copying's copyright year.
8481 * doc/cppinternals.texi: Ditto.
8482 * doc/gcc.texi: Ditto.
8483 * doc/gccint.texi: Ditto.
8484 * doc/gcov.texi: Ditto.
8485 * doc/install.texi: Ditto.
8486 * doc/invoke.texi: Ditto.
8488 2014-01-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8490 * config/nios2/nios2.h (BITS_PER_UNIT): Don't define it.
8492 2014-01-01 Jakub Jelinek <jakub@redhat.com>
8494 * config/i386/sse.md (*mov<mode>_internal): Guard
8495 EXT_REX_SSE_REGNO_P (REGNO ()) uses with REG_P.
8497 PR rtl-optimization/59647
8498 * cse.c (cse_process_notes_1): Don't substitute negative VOIDmode
8499 new_rtx into UNSIGNED_FLOAT rtxes.
8501 Copyright (C) 2014 Free Software Foundation, Inc.
8503 Copying and distribution of this file, with or without modification,
8504 are permitted in any medium without royalty provided the copyright
8505 notice and this notice are preserved.