1 2014-04-28 Richard Biener <rguenther@suse.de>
3 * configure.ac: Do valgrind header checks unconditionally.
4 Add --enable-valgrind-annotations.
5 * system.h: Guard valgrind header inclusion with
6 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING.
7 * alloc-pool.c (pool_alloc, pool_free): Use
8 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING
9 to guard possibly dead code.
10 * config.in: Regenerated.
11 * configure: Likewise.
13 2014-04-28 Jeff Law <law@redhat.com>
15 PR tree-optimization/60902
16 * tree-ssa-threadedge.c
17 (record_temporary_equivalences_from_stmts_at_dest): Only iterate
18 over real defs when invalidating outputs from statements that do not
19 produce useful outputs for threading.
21 2014-04-28 Richard Biener <rguenther@suse.de>
23 PR tree-optimization/60979
24 * graphite-scop-detection.c (scopdet_basic_block_info): Reject
25 SCOPs that end in a block with a successor with abnormal
28 2014-04-28 Richard Biener <rguenther@suse.de>
30 * tree-pass.h (execute_pass_list): Adjust prototype.
31 * passes.c (pass_manager::execute_early_local_passes):
33 (do_per_function): Change callback signature, push all actual
35 (do_per_function_toporder): Likewise.
36 (execute_function_dump): Adjust.
37 (execute_function_todo): Likewise.
38 (clear_last_verified): Likewise.
39 (verify_curr_properties): Likewise.
40 (update_properties_after_pass): Likewise.
41 (execute_pass_list_1): Split out from ...
42 (execute_pass_list): ... here. Adjust.
43 (execute_ipa_pass_list): Likewise.
44 * cgraphunit.c (cgraph_add_new_function): Adjust.
45 (analyze_function): Likewise.
46 (expand_function): Likewise.
47 * cgraph.c (release_function_body): Free dominance info
48 here instead of asserting it was magically freed elsewhere.
50 2014-04-28 Eric Botcazou <ebotcazou@adacore.com>
52 * configure.ac: Tweak GAS check for LEON instructions on SPARC.
53 * configure: Regenerate.
54 * config/sparc/sparc.opt (muser-mode): New option.
55 * config/sparc/sync.md (atomic_compare_and_swap<mode>_1): Do not enable
57 (atomic_compare_and_swap_leon3_1): New instruction for LEON3.
58 * doc/invoke.texi (SPARC options): Document -muser-mode.
60 2014-04-27 Richard Sandiford <rdsandiford@googlemail.com>
62 * cselib.c (find_slot_memmode): Delete.
63 (cselib_hasher): Change compare_type to a struct.
64 (cselib_hasher::equal): Update accordingly. Don't expect wrapped
66 (preserve_constants_and_equivs): Adjust for new compare_type.
67 (cselib_find_slot): Likewise. Take the mode of the rtx as argument.
68 (wrap_constant): Delete.
69 (cselib_lookup_mem, cselib_lookup_1): Update calls to cselib_find_slot.
71 2014-04-26 Markus Trippelsdorf <markus@trippelsdorf.de>
73 * doc/install.texi (Building with profile feedback): Remove
76 2014-04-26 Tom de Vries <tom@codesourcery.com>
78 * config/i386/i386.md (define_expand "ldexpxf3"): Fix out-of-bounds
81 2014-04-25 Cary Coutant <ccoutant@google.com>
84 * dwarf2out.c (should_move_die_to_comdat): A type definition
85 can contain a subprogram definition, but don't move it to a
87 (clone_as_declaration): Copy DW_AT_abstract_origin attribute.
88 (generate_skeleton_bottom_up): Remove DW_AT_object_pointer attribute
90 (clone_tree_hash): Rename to...
91 (clone_tree_partial): ...this; change callers. Copy
92 DW_TAG_subprogram DIEs as declarations.
93 (copy_decls_walk): Don't copy children of a declaration into a
96 2014-04-25 H.J. Lu <hongjiu.lu@intel.com>
99 * config/i386/i386.md (*movsf_internal): Set MODE to SI for
102 2014-04-25 Jiong Wang <jiong.wang@arm.com>
104 * config/arm/predicates.md (call_insn_operand): Add long_call check.
105 * config/arm/arm.md (sibcall, sibcall_value): Force the address to
107 * config/arm/arm.c (arm_function_ok_for_sibcall): Remove long_call
110 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
112 * config/arm/arm.c (arm_cortex_a8_tune): Initialise
115 2014-04-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
117 PR tree-optimization/60930
118 * gimple-ssa-strength-reduction.c (create_mul_imm_cand): Reject
119 creating a multiply candidate by folding two constant
120 multiplicands when the result overflows.
122 2014-04-25 Jakub Jelinek <jakub@redhat.com>
124 PR tree-optimization/60960
125 * tree-vect-generic.c (expand_vector_operation): Only call
126 expand_vector_divmod if type's mode satisfies VECTOR_MODE_P.
128 2014-04-25 Tom de Vries <tom@codesourcery.com>
130 * expr.c (clobber_reg_mode): New function.
131 * expr.h (clobber_reg): New function.
133 2014-04-25 Tom de Vries <tom@codesourcery.com>
135 * rtlanal.c (find_all_hard_reg_sets): Note INSN_CALL_FUNCTION_USAGE
138 2014-04-25 Radovan Obradovic <robradovic@mips.com>
139 Tom de Vries <tom@codesourcery.com>
141 * rtlanal.c (find_all_hard_reg_sets): Add bool implicit parameter and
143 * rtl.h (find_all_hard_reg_sets): Add bool parameter.
144 * haifa-sched.c (recompute_todo_spec, check_clobbered_conditions): Add
145 new argument to find_all_hard_reg_sets call.
147 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
149 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p):
150 Use HOST_WIDE_INT_C for mask literal.
151 (aarch_rev16_shleft_mask_imm_p): Likewise.
153 2014-04-25 Eric Botcazou <ebotcazou@adacore.com>
156 * config/sparc/sparc.md (ashlsi3_extend): Delete.
158 2014-04-25 Marc Glisse <marc.glisse@inria.fr>
160 PR preprocessor/56540
161 * config/i386/i386-c.c (ix86_target_macros): Define
162 __SIZEOF_FLOAT80__ and __SIZEOF_FLOAT128__.
164 2014-04-25 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
166 * configure.ac (tga_func): Remove.
167 (LIB_TLS_SPEC): Remove.
168 * configure: Regenerate.
169 * config.in: Regenerate.
170 * config/sol2.h (LIB_SPEC): Don't use LIB_TLS_SPEC.
172 2014-04-25 Richard Biener <rguenther@suse.de>
175 * tree-ssa-structalias.c (ipa_pta_execute): Compute direct
176 call stmt use/clobber sets during stmt walk instead of
177 walking the possibly incomplete set of caller edges.
179 2014-04-25 Richard Biener <rguenther@suse.de>
182 * passes.c (apply_ipa_transforms): Inline into only caller ...
183 (execute_one_pass): ... here. Properly bring in function
184 bodies for nodes we want to apply IPA transforms to.
186 2014-04-24 Cong Hou <congh@google.com>
188 PR tree-optimization/60896
189 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Pick up
190 all statements in PATTERN_DEF_SEQ in recognized widen-mult pattern.
191 (vect_mark_pattern_stmts): Set the def type of all statements in
192 PATTERN_DEF_SEQ as vect_internal_def.
194 2014-04-24 Michael Meissner <meissner@linux.vnet.ibm.com>
196 * doc/extend.texi (PowerPC Built-in Functions): Document new
197 powerpc extended divide, bcd, pack/unpack 128-bit, builtin
199 (PowerPC AltiVec/VSX Built-in Functions): Likewise.
201 * config/rs6000/predicates.md (const_0_to_3_operand): New
202 predicate to match 0..3 integer constants.
204 * config/rs6000/rs6000-builtin.def (BU_DFP_MISC_1): Add new macros
205 to support adding miscellaneous builtin functions.
206 (BU_DFP_MISC_2): Likewise.
207 (BU_P7_MISC_1): Likewise.
208 (BU_P7_MISC_2): Likewise.
209 (BU_P8V_MISC_3): Likewise.
210 (BU_MISC_1): Likewise.
211 (BU_MISC_2): Likewise.
212 (DIVWE): Add extended divide builtin functions.
220 (DXEX): Add decimal floating-point builtin functions.
232 (CDTBCD): Add new BCD builtin functions.
236 (BCDADD_LT): Likewise.
237 (BCDADD_EQ): Likewise.
238 (BCDADD_GT): Likewise.
239 (BCDADD_OV): Likewise.
241 (BCDSUB_LT): Likewise.
242 (BCDSUB_EQ): Likewise.
243 (BCDSUB_GT): Likewise.
244 (BCDSUB_OV): Likewise.
245 (PACK_TD): Add new pack/unpack 128-bit type builtin functions.
246 (UNPACK_TD): Likewise.
248 (UNPACK_TF): Likewise.
249 (UNPACK_TF_0): Likewise.
250 (UNPACK_TF_1): Likewise.
251 (PACK_V1TI): Likewise.
252 (UNPACK_V1TI): Likewise.
254 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Add
255 support for decimal floating point builtin functions.
256 (rs6000_expand_ternop_builtin): Add checks for the new builtin
257 functions that take constant arguments.
258 (rs6000_invalid_builtin): Add decimal floating point builtin
260 (rs6000_init_builtins): Setup long double, _Decimal64, and
261 _Decimal128 types for new builtin functions.
262 (builtin_function_type): Set the unsigned flags appropriately for
263 the new builtin functions.
264 (rs6000_opt_masks): Add support for decimal floating point builtin
267 * config/rs6000/rs6000.h (RS6000_BTM_DFP): Add support for decimal
268 floating point builtin functions.
269 (RS6000_BTM_COMMON): Likewise.
270 (RS6000_BTI_long_double): Likewise.
271 (RS6000_BTI_dfloat64): Likewise.
272 (RS6000_BTI_dfloat128): Likewise.
273 (long_double_type_internal_node): Likewise.
274 (dfloat64_type_internal_node): Likewise.
275 (dfloat128_type_internal_node): Likewise.
277 * config/rs6000/altivec.h (UNSPEC_BCDADD): Add support for ISA
278 2.07 bcd arithmetic instructions.
279 (UNSPEC_BCDSUB): Likewise.
280 (UNSPEC_BCD_OVERFLOW): Likewise.
281 (UNSPEC_BCD_ADD_SUB): Likewise.
282 (bcd_add_sub): Likewise.
283 (BCD_TEST): Likewise.
284 (bcd<bcd_add_sub>): Likewise.
285 (bcd<bcd_add_sub>_test): Likewise.
286 (bcd<bcd_add_sub>_test2): Likewise.
287 (bcd<bcd_add_sub>_<code>): Likewise.
288 (peephole2 for combined bcd ops): Likewise.
290 * config/rs6000/dfp.md (UNSPEC_DDEDPD): Add support for new
291 decimal floating point builtin functions.
292 (UNSPEC_DENBCD): Likewise.
293 (UNSPEC_DXEX): Likewise.
294 (UNSPEC_DIEX): Likewise.
295 (UNSPEC_DSCLI): Likewise.
296 (UNSPEC_DSCRI): Likewise.
297 (D64_D128): Likewise.
298 (dfp_suffix): Likewise.
299 (dfp_ddedpd_<mode>): Likewise.
300 (dfp_denbcd_<mode>): Likewise.
301 (dfp_dxex_<mode>): Likewise.
302 (dfp_diex_<mode>): Likewise.
303 (dfp_dscli_<mode>): Likewise.
304 (dfp_dscri_<mode>): Likewise.
306 * config/rs6000/rs6000.md (UNSPEC_ADDG6S): Add support for new BCD
308 (UNSPEC_CDTBCD): Likewise.
309 (UNSPEC_CBCDTD): Likewise.
310 (UNSPEC_DIVE): Add support for new extended divide builtin
312 (UNSPEC_DIVEO): Likewise.
313 (UNSPEC_DIVEU): Likewise.
314 (UNSPEC_DIVEUO): Likewise.
315 (UNSPEC_UNPACK_128BIT): Add support for new builtin functions to
316 pack/unpack 128-bit types.
317 (UNSPEC_PACK_128BIT): Likewise.
318 (idiv_ldiv): New mode attribute to set the 32/64-bit divide type.
319 (udiv<mode>3): Use idiv_ldiv mode attribute.
320 (div<mode>3): Likewise.
321 (addg6s): Add new BCD builtin functions.
324 (UNSPEC_DIV_EXTEND): Add support for new extended divide
326 (div_extend): Likewise.
327 (div<div_extend>_<mode>"): Likewise.
328 (FP128_64): Add support for new builtin functions to pack/unpack
330 (unpack<mode>): Likewise.
331 (unpacktf_0): Likewise.
332 (unpacktf_1): Likewise.
333 (unpack<mode>_dm): Likewise.
334 (unpack<mode>_nodm): Likewise.
335 (pack<mode>): Likewise.
336 (unpackv1ti): Likewise.
337 (packv1ti): Likewise.
339 2014-04-24 Vishnu K S <Vishnu.k_s@atmel.com>
341 * gcc/config/avr/avr.c: Add comment on why -fdelete-null-pointer-checks
344 2014-04-24 Jakub Jelinek <jakub@redhat.com>
346 * tree.h (OMP_CLAUSE_LINEAR_GIMPLE_SEQ): Define.
347 * gimplify.c (omp_is_private): Change last argument's type to int.
348 Only diagnose lastprivate if the simd argument is 1, only diagnose
349 linear if the simd argument is 2.
350 (gimplify_omp_for): Adjust omp_is_private callers. When adding
351 lastprivate or private, add the clause to OMP_FOR_CLAUSES. Pass
352 GOVD_EXPLICIT to omp_add_variable. For simd with collapse == 1
353 create OMP_CLAUSE_LINEAR rather than OMP_CLAUSE_PRIVATE for var.
354 If var != decl and decl is in OMP_CLAUSE_LINEAR, gimplify decl
355 increment to OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
356 * omp-low.c (scan_sharing_clauses, lower_lastprivate_clauses): Handle
357 OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
358 * tree-nested.c (convert_nonlocal_omp_clauses,
359 convert_local_omp_clauses): Handle OMP_CLAUSE_LINEAR.
361 2014-04-24 Segher Boessenkool <segher@kernel.crashing.org>
364 * config/m68k/m68k.md (extendplussidi): Don't allow memory for
367 2014-04-24 Dimitris Papavasiliou <dpapavas@gmail.com>
369 * flag-types.h (enum ivar_visibility): Add.
371 2014-04-24 Trevor Saunders <tsaunders@mozilla.com>
373 * config/sh/sh_treg_combine.c (sh_treg_combine::execute): Take
376 2014-04-24 Alan Lawrence <alan.lawrence@arm.com>
378 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Enable for bigendian.
380 2014-04-24 Radovan Obradovic <robradovic@mips.com>
381 Tom de Vries <tom@codesourcery.com>
383 * reg-notes.def (REG_NOTE (CALL_DECL)): New reg-note REG_CALL_DECL.
384 * calls.c (expand_call, emit_library_call_value_1): Add REG_CALL_DECL
386 * combine.c (distribute_notes): Handle REG_CALL_DECL reg-note.
387 * emit-rtl.c (try_split): Same.
389 2014-04-24 Radovan Obradovic <robradovic@mips.com>
390 Tom de Vries <tom@codesourcery.com>
392 * common.opt (fuse-caller-save): New option.
394 2014-04-24 Tejas Belagod <tejas.belagod@arm.com>
396 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Reverse order of elements
399 2014-04-24 Richard Biener <rguenther@suse.de>
401 * expr.c (expand_expr_real_1): Avoid gimple_assign_rhs_to_tree
402 during TER and instead use the sepops interface for expanding
403 non-GIMPLE_SINGLE_RHS.
405 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
407 * config/i386/sol2.h (ASM_PREFERRED_EH_DATA_FORMAT): Only redefine
408 if not HAVE_AS_IX86_DIFF_SECT_DELTA.
410 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
412 * configure.ac (gcc_cv_as_cfi_directive): Support Solaris/x86
413 assembler 64-bit option.
414 * configure: Regenerate.
416 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
418 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Check
419 TARGET_SIMD rather than TARGET_GENERAL_REGS_ONLY.
420 (TARGET_SIMD): Take AARCH64_ISA_SIMD into account.
421 (TARGET_FLOAT): Take AARCH64_ISA_FP into account.
422 (TARGET_CRYPTO): Take TARGET_SIMD into account.
424 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
426 * config/aarch64/aarch64-builtins.c
427 (aarch64_builtin_vectorized_function): Handle BUILT_IN_BSWAP16,
428 BUILT_IN_BSWAP32, BUILT_IN_BSWAP64.
429 * config/aarch64/aarch64-simd.md (bswap<mode>): New pattern.
430 * config/aarch64/aarch64-simd-builtins.def: Define vector bswap
432 * config/aarch64/iterator.md (VDQHSD): New mode iterator.
433 (Vrevsuff): New mode attribute.
435 2014-04-24 Terry Guo <terry.guo@arm.com>
437 * config/arm/arm.h (machine_function): Define variable
438 after_arm_reorg here.
439 * config/arm/arm.c (after_arm_reorg): Remove the definition.
440 (arm_split_constant): Update the way to access variable
443 (arm_output_function_epilogue): Remove the reset of after_arm_reorg.
445 2014-04-23 Tom de Vries <tom@codesourcery.com>
447 * target-hooks-macros.h: Fix DEFHOOKPOD argument order in comment.
449 2014-04-23 David Malcolm <dmalcolm@redhat.com>
451 * is-a.h: Update comments to reflect the following changes to the
452 "pointerness" of the API, making the template parameter match the
453 return type, allowing use of is-a.h with typedefs of pointers.
454 (is_a_helper::cast): Return a T rather then a pointer to a T, so
455 that the return type matches the parameter to the is_a_helper.
457 (dyn_cast): Likewise.
459 * cgraph.c (cgraph_node_for_asm): Update for removal of implicit
460 pointer from the is-a.h API.
462 * cgraph.h (is_a_helper <cgraph_node>::test): Convert to...
463 (is_a_helper <cgraph_node *>::test): ...this, matching change to
465 (is_a_helper <varpool_node>::test): Likewise, convert to...
466 (is_a_helper <varpool_node *>::test): ...this.
468 (varpool_first_variable): Update for removal of implicit pointer
470 (varpool_next_variable): Likewise.
471 (varpool_first_static_initializer): Likewise.
472 (varpool_next_static_initializer): Likewise.
473 (varpool_first_defined_variable): Likewise.
474 (varpool_next_defined_variable): Likewise.
475 (cgraph_first_defined_function): Likewise.
476 (cgraph_next_defined_function): Likewise.
477 (cgraph_first_function): Likewise.
478 (cgraph_next_function): Likewise.
479 (cgraph_first_function_with_gimple_body): Likewise.
480 (cgraph_next_function_with_gimple_body): Likewise.
481 (cgraph_alias_target): Likewise.
482 (varpool_alias_target): Likewise.
483 (cgraph_function_or_thunk_node): Likewise.
484 (varpool_variable_node): Likewise.
485 (symtab_real_symbol_p): Likewise.
486 * cgraphunit.c (referred_to_p): Likewise.
487 (analyze_functions): Likewise.
488 (handle_alias_pairs): Likewise.
489 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
490 * gimple-ssa.h (gimple_vuse_op): Likewise.
491 (gimple_vdef_op): Likewise.
492 * gimple-streamer-in.c (input_gimple_stmt): Likewise.
493 * gimple.c (gimple_build_asm_1): Likewise.
494 (gimple_build_try): Likewise.
495 (gimple_build_resx): Likewise.
496 (gimple_build_eh_dispatch): Likewise.
497 (gimple_build_omp_for): Likewise.
498 (gimple_omp_for_set_clauses): Likewise.
500 * gimple.h (is_a_helper <gimple_statement_asm>::test): Convert to...
501 (is_a_helper <gimple_statement_asm *>::test): ...this.
502 (is_a_helper <gimple_statement_bind>::test): Convert to...
503 (is_a_helper <gimple_statement_bind *>::test): ...this.
504 (is_a_helper <gimple_statement_call>::test): Convert to...
505 (is_a_helper <gimple_statement_call *>::test): ...this.
506 (is_a_helper <gimple_statement_catch>::test): Convert to...
507 (is_a_helper <gimple_statement_catch *>::test): ...this.
508 (is_a_helper <gimple_statement_resx>::test): Convert to...
509 (is_a_helper <gimple_statement_resx *>::test): ...this.
510 (is_a_helper <gimple_statement_eh_dispatch>::test): Convert to...
511 (is_a_helper <gimple_statement_eh_dispatch *>::test): ...this.
512 (is_a_helper <gimple_statement_eh_else>::test): Convert to...
513 (is_a_helper <gimple_statement_eh_else *>::test): ...this.
514 (is_a_helper <gimple_statement_eh_filter>::test): Convert to...
515 (is_a_helper <gimple_statement_eh_filter *>::test): ...this.
516 (is_a_helper <gimple_statement_eh_mnt>::test): Convert to...
517 (is_a_helper <gimple_statement_eh_mnt *>::test): ...this.
518 (is_a_helper <gimple_statement_omp_atomic_load>::test): Convert to...
519 (is_a_helper <gimple_statement_omp_atomic_load *>::test): ...this.
520 (is_a_helper <gimple_statement_omp_atomic_store>::test): Convert to...
521 (is_a_helper <gimple_statement_omp_atomic_store *>::test): ...this.
522 (is_a_helper <gimple_statement_omp_return>::test): Convert to...
523 (is_a_helper <gimple_statement_omp_return *>::test): ...this.
524 (is_a_helper <gimple_statement_omp_continue>::test): Convert to...
525 (is_a_helper <gimple_statement_omp_continue *>::test): ...this.
526 (is_a_helper <gimple_statement_omp_critical>::test): Convert to...
527 (is_a_helper <gimple_statement_omp_critical *>::test): ...this.
528 (is_a_helper <gimple_statement_omp_for>::test): Convert to...
529 (is_a_helper <gimple_statement_omp_for *>::test): ...this.
530 (is_a_helper <gimple_statement_omp_taskreg>::test): Convert to...
531 (is_a_helper <gimple_statement_omp_taskreg *>::test): ...this.
532 (is_a_helper <gimple_statement_omp_parallel>::test): Convert to...
533 (is_a_helper <gimple_statement_omp_parallel *>::test): ...this.
534 (is_a_helper <gimple_statement_omp_target>::test): Convert to...
535 (is_a_helper <gimple_statement_omp_target *>::test): ...this.
536 (is_a_helper <gimple_statement_omp_sections>::test): Convert to...
537 (is_a_helper <gimple_statement_omp_sections *>::test): ...this.
538 (is_a_helper <gimple_statement_omp_single>::test): Convert to...
539 (is_a_helper <gimple_statement_omp_single *>::test): ...this.
540 (is_a_helper <gimple_statement_omp_teams>::test): Convert to...
541 (is_a_helper <gimple_statement_omp_teams *>::test): ...this.
542 (is_a_helper <gimple_statement_omp_task>::test): Convert to...
543 (is_a_helper <gimple_statement_omp_task *>::test): ...this.
544 (is_a_helper <gimple_statement_phi>::test): Convert to...
545 (is_a_helper <gimple_statement_phi *>::test): ...this.
546 (is_a_helper <gimple_statement_transaction>::test): Convert to...
547 (is_a_helper <gimple_statement_transaction *>::test): ...this.
548 (is_a_helper <gimple_statement_try>::test): Convert to...
549 (is_a_helper <gimple_statement_try *>::test): ...this.
550 (is_a_helper <gimple_statement_wce>::test): Convert to...
551 (is_a_helper <gimple_statement_wce *>::test): ...this.
552 (is_a_helper <const gimple_statement_asm>::test): Convert to...
553 (is_a_helper <const gimple_statement_asm *>::test): ...this.
554 (is_a_helper <const gimple_statement_bind>::test): Convert to...
555 (is_a_helper <const gimple_statement_bind *>::test): ...this.
556 (is_a_helper <const gimple_statement_call>::test): Convert to...
557 (is_a_helper <const gimple_statement_call *>::test): ...this.
558 (is_a_helper <const gimple_statement_catch>::test): Convert to...
559 (is_a_helper <const gimple_statement_catch *>::test): ...this.
560 (is_a_helper <const gimple_statement_resx>::test): Convert to...
561 (is_a_helper <const gimple_statement_resx *>::test): ...this.
562 (is_a_helper <const gimple_statement_eh_dispatch>::test):
564 (is_a_helper <const gimple_statement_eh_dispatch *>::test):
566 (is_a_helper <const gimple_statement_eh_filter>::test): Convert
568 (is_a_helper <const gimple_statement_eh_filter *>::test): ...this.
569 (is_a_helper <const gimple_statement_omp_atomic_load>::test):
571 (is_a_helper <const gimple_statement_omp_atomic_load *>::test):
573 (is_a_helper <const gimple_statement_omp_atomic_store>::test):
575 (is_a_helper <const gimple_statement_omp_atomic_store *>::test):
577 (is_a_helper <const gimple_statement_omp_return>::test): Convert
579 (is_a_helper <const gimple_statement_omp_return *>::test): ...this.
580 (is_a_helper <const gimple_statement_omp_continue>::test): Convert
582 (is_a_helper <const gimple_statement_omp_continue *>::test): ...this.
583 (is_a_helper <const gimple_statement_omp_critical>::test): Convert
585 (is_a_helper <const gimple_statement_omp_critical *>::test): ...this.
586 (is_a_helper <const gimple_statement_omp_for>::test): Convert to...
587 (is_a_helper <const gimple_statement_omp_for *>::test): ...this.
588 (is_a_helper <const gimple_statement_omp_taskreg>::test): Convert
590 (is_a_helper <const gimple_statement_omp_taskreg *>::test): ...this.
591 (is_a_helper <const gimple_statement_omp_parallel>::test): Convert
593 (is_a_helper <const gimple_statement_omp_parallel *>::test): ...this.
594 (is_a_helper <const gimple_statement_omp_target>::test): Convert
596 (is_a_helper <const gimple_statement_omp_target *>::test): ...this.
597 (is_a_helper <const gimple_statement_omp_sections>::test): Convert
599 (is_a_helper <const gimple_statement_omp_sections *>::test): ...this.
600 (is_a_helper <const gimple_statement_omp_single>::test): Convert
602 (is_a_helper <const gimple_statement_omp_single *>::test): ...this.
603 (is_a_helper <const gimple_statement_omp_teams>::test): Convert
605 (is_a_helper <const gimple_statement_omp_teams *>::test): ...this.
606 (is_a_helper <const gimple_statement_omp_task>::test): Convert
608 (is_a_helper <const gimple_statement_omp_task *>::test): ...this.
609 (is_a_helper <const gimple_statement_phi>::test): Convert to...
610 (is_a_helper <const gimple_statement_phi *>::test): ...this.
611 (is_a_helper <const gimple_statement_transaction>::test): Convert
613 (is_a_helper <const gimple_statement_transaction *>::test): ...this.
614 (is_a_helper <const gimple_statement_with_ops>::test): Convert
616 (is_a_helper <const gimple_statement_with_ops *>::test): ...this.
617 (is_a_helper <gimple_statement_with_ops>::test): Convert to...
618 (is_a_helper <gimple_statement_with_ops *>::test): ...this.
619 (is_a_helper <const gimple_statement_with_memory_ops>::test):
621 (is_a_helper <const gimple_statement_with_memory_ops *>::test):
623 (is_a_helper <gimple_statement_with_memory_ops>::test): Convert
625 (is_a_helper <gimple_statement_with_memory_ops *>::test): ...this.
627 (gimple_use_ops): Update for removal of implicit pointer from the
629 (gimple_set_use_ops): Likewise.
630 (gimple_vuse): Likewise.
631 (gimple_vdef): Likewise.
632 (gimple_vuse_ptr): Likewise.
633 (gimple_vdef_ptr): Likewise.
634 (gimple_set_vuse): Likewise.
635 (gimple_set_vdef): Likewise.
636 (gimple_omp_return_set_lhs): Likewise.
637 (gimple_omp_return_lhs): Likewise.
638 (gimple_omp_return_lhs_ptr): Likewise.
639 (gimple_call_fntype): Likewise.
640 (gimple_call_set_fntype): Likewise.
641 (gimple_call_set_internal_fn): Likewise.
642 (gimple_call_use_set): Likewise.
643 (gimple_call_clobber_set): Likewise.
644 (gimple_bind_vars): Likewise.
645 (gimple_bind_set_vars): Likewise.
646 (gimple_bind_body_ptr): Likewise.
647 (gimple_bind_set_body): Likewise.
648 (gimple_bind_add_stmt): Likewise.
649 (gimple_bind_block): Likewise.
650 (gimple_bind_set_block): Likewise.
651 (gimple_asm_ninputs): Likewise.
652 (gimple_asm_noutputs): Likewise.
653 (gimple_asm_nclobbers): Likewise.
654 (gimple_asm_nlabels): Likewise.
655 (gimple_asm_input_op): Likewise.
656 (gimple_asm_input_op_ptr): Likewise.
657 (gimple_asm_output_op): Likewise.
658 (gimple_asm_output_op_ptr): Likewise.
659 (gimple_asm_set_output_op): Likewise.
660 (gimple_asm_clobber_op): Likewise.
661 (gimple_asm_set_clobber_op): Likewise.
662 (gimple_asm_label_op): Likewise.
663 (gimple_asm_set_label_op): Likewise.
664 (gimple_asm_string): Likewise.
665 (gimple_catch_types): Likewise.
666 (gimple_catch_types_ptr): Likewise.
667 (gimple_catch_handler_ptr): Likewise.
668 (gimple_catch_set_types): Likewise.
669 (gimple_catch_set_handler): Likewise.
670 (gimple_eh_filter_types): Likewise.
671 (gimple_eh_filter_types_ptr): Likewise.
672 (gimple_eh_filter_failure_ptr): Likewise.
673 (gimple_eh_filter_set_types): Likewise.
674 (gimple_eh_filter_set_failure): Likewise.
675 (gimple_eh_must_not_throw_fndecl): Likewise.
676 (gimple_eh_must_not_throw_set_fndecl): Likewise.
677 (gimple_eh_else_n_body_ptr): Likewise.
678 (gimple_eh_else_e_body_ptr): Likewise.
679 (gimple_eh_else_set_n_body): Likewise.
680 (gimple_eh_else_set_e_body): Likewise.
681 (gimple_try_eval_ptr): Likewise.
682 (gimple_try_cleanup_ptr): Likewise.
683 (gimple_try_set_eval): Likewise.
684 (gimple_try_set_cleanup): Likewise.
685 (gimple_wce_cleanup_ptr): Likewise.
686 (gimple_wce_set_cleanup): Likewise.
687 (gimple_phi_capacity): Likewise.
688 (gimple_phi_num_args): Likewise.
689 (gimple_phi_result): Likewise.
690 (gimple_phi_result_ptr): Likewise.
691 (gimple_phi_set_result): Likewise.
692 (gimple_phi_arg): Likewise.
693 (gimple_phi_set_arg): Likewise.
694 (gimple_resx_region): Likewise.
695 (gimple_resx_set_region): Likewise.
696 (gimple_eh_dispatch_region): Likewise.
697 (gimple_eh_dispatch_set_region): Likewise.
698 (gimple_omp_critical_name): Likewise.
699 (gimple_omp_critical_name_ptr): Likewise.
700 (gimple_omp_critical_set_name): Likewise.
701 (gimple_omp_for_clauses): Likewise.
702 (gimple_omp_for_clauses_ptr): Likewise.
703 (gimple_omp_for_set_clauses): Likewise.
704 (gimple_omp_for_collapse): Likewise.
705 (gimple_omp_for_index): Likewise.
706 (gimple_omp_for_index_ptr): Likewise.
707 (gimple_omp_for_set_index): Likewise.
708 (gimple_omp_for_initial): Likewise.
709 (gimple_omp_for_initial_ptr): Likewise.
710 (gimple_omp_for_set_initial): Likewise.
711 (gimple_omp_for_final): Likewise.
712 (gimple_omp_for_final_ptr): Likewise.
713 (gimple_omp_for_set_final): Likewise.
714 (gimple_omp_for_incr): Likewise.
715 (gimple_omp_for_incr_ptr): Likewise.
716 (gimple_omp_for_set_incr): Likewise.
717 (gimple_omp_for_pre_body_ptr): Likewise.
718 (gimple_omp_for_set_pre_body): Likewise.
719 (gimple_omp_parallel_clauses): Likewise.
720 (gimple_omp_parallel_clauses_ptr): Likewise.
721 (gimple_omp_parallel_set_clauses): Likewise.
722 (gimple_omp_parallel_child_fn): Likewise.
723 (gimple_omp_parallel_child_fn_ptr): Likewise.
724 (gimple_omp_parallel_set_child_fn): Likewise.
725 (gimple_omp_parallel_data_arg): Likewise.
726 (gimple_omp_parallel_data_arg_ptr): Likewise.
727 (gimple_omp_parallel_set_data_arg): Likewise.
728 (gimple_omp_task_clauses): Likewise.
729 (gimple_omp_task_clauses_ptr): Likewise.
730 (gimple_omp_task_set_clauses): Likewise.
731 (gimple_omp_task_child_fn): Likewise.
732 (gimple_omp_task_child_fn_ptr): Likewise.
733 (gimple_omp_task_set_child_fn): Likewise.
734 (gimple_omp_task_data_arg): Likewise.
735 (gimple_omp_task_data_arg_ptr): Likewise.
736 (gimple_omp_task_set_data_arg): Likewise.
737 (gimple_omp_taskreg_clauses): Likewise.
738 (gimple_omp_taskreg_clauses_ptr): Likewise.
739 (gimple_omp_taskreg_set_clauses): Likewise.
740 (gimple_omp_taskreg_child_fn): Likewise.
741 (gimple_omp_taskreg_child_fn_ptr): Likewise.
742 (gimple_omp_taskreg_set_child_fn): Likewise.
743 (gimple_omp_taskreg_data_arg): Likewise.
744 (gimple_omp_taskreg_data_arg_ptr): Likewise.
745 (gimple_omp_taskreg_set_data_arg): Likewise.
746 (gimple_omp_task_copy_fn): Likewise.
747 (gimple_omp_task_copy_fn_ptr): Likewise.
748 (gimple_omp_task_set_copy_fn): Likewise.
749 (gimple_omp_task_arg_size): Likewise.
750 (gimple_omp_task_arg_size_ptr): Likewise.
751 (gimple_omp_task_set_arg_size): Likewise.
752 (gimple_omp_task_arg_align): Likewise.
753 (gimple_omp_task_arg_align_ptr): Likewise.
754 (gimple_omp_task_set_arg_align): Likewise.
755 (gimple_omp_single_clauses): Likewise.
756 (gimple_omp_single_clauses_ptr): Likewise.
757 (gimple_omp_single_set_clauses): Likewise.
758 (gimple_omp_target_clauses): Likewise.
759 (gimple_omp_target_clauses_ptr): Likewise.
760 (gimple_omp_target_set_clauses): Likewise.
761 (gimple_omp_target_child_fn): Likewise.
762 (gimple_omp_target_child_fn_ptr): Likewise.
763 (gimple_omp_target_set_child_fn): Likewise.
764 (gimple_omp_target_data_arg): Likewise.
765 (gimple_omp_target_data_arg_ptr): Likewise.
766 (gimple_omp_target_set_data_arg): Likewise.
767 (gimple_omp_teams_clauses): Likewise.
768 (gimple_omp_teams_clauses_ptr): Likewise.
769 (gimple_omp_teams_set_clauses): Likewise.
770 (gimple_omp_sections_clauses): Likewise.
771 (gimple_omp_sections_clauses_ptr): Likewise.
772 (gimple_omp_sections_set_clauses): Likewise.
773 (gimple_omp_sections_control): Likewise.
774 (gimple_omp_sections_control_ptr): Likewise.
775 (gimple_omp_sections_set_control): Likewise.
776 (gimple_omp_for_set_cond): Likewise.
777 (gimple_omp_for_cond): Likewise.
778 (gimple_omp_atomic_store_set_val): Likewise.
779 (gimple_omp_atomic_store_val): Likewise.
780 (gimple_omp_atomic_store_val_ptr): Likewise.
781 (gimple_omp_atomic_load_set_lhs): Likewise.
782 (gimple_omp_atomic_load_lhs): Likewise.
783 (gimple_omp_atomic_load_lhs_ptr): Likewise.
784 (gimple_omp_atomic_load_set_rhs): Likewise.
785 (gimple_omp_atomic_load_rhs): Likewise.
786 (gimple_omp_atomic_load_rhs_ptr): Likewise.
787 (gimple_omp_continue_control_def): Likewise.
788 (gimple_omp_continue_control_def_ptr): Likewise.
789 (gimple_omp_continue_set_control_def): Likewise.
790 (gimple_omp_continue_control_use): Likewise.
791 (gimple_omp_continue_control_use_ptr): Likewise.
792 (gimple_omp_continue_set_control_use): Likewise.
793 (gimple_transaction_body_ptr): Likewise.
794 (gimple_transaction_label): Likewise.
795 (gimple_transaction_label_ptr): Likewise.
796 (gimple_transaction_set_body): Likewise.
797 (gimple_transaction_set_label): Likewise.
799 * ipa-devirt.c (build_type_inheritance_graph): Likewise.
800 * ipa-inline-analysis.c (inline_write_summary): Likewise.
801 * ipa-ref.c (ipa_record_reference): Likewise.
802 * ipa-reference.c (analyze_function): Likewise.
803 (ipa_reference_write_optimization_summary): Likewise.
804 * ipa.c (symtab_remove_unreachable_nodes): Likewise.
805 (address_taken_from_non_vtable_p): Likewise.
806 (comdat_can_be_unshared_p_1): Likewise.
807 * lto-cgraph.c (lto_output_ref): Likewise.
808 (add_references): Likewise.
809 (compute_ltrans_boundary): Likewise.
810 (output_symtab): Likewise.
811 (input_ref): Likewise.
812 (input_cgraph_1): Likewise.
813 (output_cgraph_opt_summary): Likewise.
814 * lto-streamer-out.c (lto_output): Likewise.
815 (output_symbol_p): Likewise.
816 * lto-streamer.h (lsei_next_function_in_partition): Likewise.
817 (lsei_start_function_in_partition): Likewise.
818 (lsei_next_variable_in_partition): Likewise.
819 (lsei_start_variable_in_partition): Likewise.
820 * symtab.c (insert_to_assembler_name_hash): Likewise.
821 (unlink_from_assembler_name_hash): Likewise.
822 (symtab_unregister_node): Likewise.
823 (symtab_remove_node): Likewise.
824 (dump_symtab_node): Likewise.
825 (verify_symtab_base): Likewise.
826 (verify_symtab_node): Likewise.
827 (symtab_make_decl_local): Likewise.
828 (symtab_alias_ultimate_target): Likewise.
829 (symtab_resolve_alias): Likewise.
830 (symtab_get_symbol_partitioning_class): Likewise.
831 * tree-phinodes.c (allocate_phi_node): Likewise.
832 (reserve_phi_args_for_new_edge): Likewise.
833 (remove_phi_args): Likewise.
834 * varpool.c (varpool_node_for_asm): Likewise.
835 (varpool_remove_unreferenced_decls): Likewise.
837 2014-04-23 Jeff Law <law@redhat.com>
839 PR tree-optimization/60902
840 * tree-ssa-threadedge.c
841 (record_temporary_equivalences_from_stmts_at_dest): Make sure to
842 invalidate outputs from statements that do not produce useful
843 outputs for threading.
845 2014-04-23 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
847 * config/aarch64/aarch64.md (stack_protect_set, stack_protect_test)
848 (stack_protect_set_<mode>, stack_protect_test_<mode>): Add
849 machine descriptions for Stack Smashing Protector.
851 2014-04-23 Richard Earnshaw <rearnsha@arm.com>
853 * aarch64.md (<optab>_rol<mode>3): New pattern.
854 (<optab>_rolsi3_uxtw): Likewise.
855 * aarch64.c (aarch64_strip_shift): Handle ROTATE and ROTATERT.
857 2014-04-23 James Greenhalgh <james.greenhalgh@arm.com>
859 * config/arm/arm.c (arm_cortex_a57_tune): Initialize all fields.
860 (arm_cortex_a12_tune): Likewise.
862 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
864 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle BSWAP.
866 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
868 * config/arm/arm.md (arm_rev16si2): New pattern.
869 (arm_rev16si2_alt): Likewise.
870 * config/arm/arm.c (arm_new_rtx_costs): Handle rev16 case.
872 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
874 * config/aarch64/aarch64.md (rev16<mode>2): New pattern.
875 (rev16<mode>2_alt): Likewise.
876 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle rev16 case.
877 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p): New.
878 (aarch_rev16_shleft_mask_imm_p): Likewise.
879 (aarch_rev16_p_1): Likewise.
880 (aarch_rev16_p): Likewise.
881 * config/arm/aarch-common-protos.h (aarch_rev16_p): Declare extern.
882 (aarch_rev16_shright_mask_imm_p): Likewise.
883 (aarch_rev16_shleft_mask_imm_p): Likewise.
885 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
887 * config/arm/aarch-common-protos.h (alu_cost_table): Add rev field.
888 * config/arm/aarch-cost-tables.h (generic_extra_costs): Specify
890 (cortex_a53_extra_costs): Likewise.
891 (cortex_a57_extra_costs): Likewise.
892 * config/arm/arm.c (cortexa9_extra_costs): Likewise.
893 (cortexa7_extra_costs): Likewise.
894 (cortexa8_extra_costs): Likewise.
895 (cortexa12_extra_costs): Likewise.
896 (cortexa15_extra_costs): Likewise.
897 (v7m_extra_costs): Likewise.
898 (arm_new_rtx_costs): Handle BSWAP.
900 2013-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
902 * config/arm/arm.c (cortexa8_extra_costs): New table.
903 (arm_cortex_a8_tune): New tuning struct.
904 * config/arm/arm-cores.def (cortex-a8): Use cortex_a8 tuning struct.
906 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
908 * config/arm/arm.c (arm_new_rtx_costs): Handle FMA.
910 2014-04-23 Richard Biener <rguenther@suse.de>
912 * Makefile.in (OBJS): Remove loop-unswitch.o.
913 * loop-unswitch.c: Delete.
914 * tree-pass.h (make_pass_rtl_unswitch): Remove.
915 * passes.def (pass_rtl_unswitch): Likewise.
916 * loop-init.c (gate_rtl_unswitch): Likewise.
917 (rtl_unswitch): Likewise.
918 (pass_data_rtl_unswitch): Likewise.
919 (pass_rtl_unswitch): Likewise.
920 (make_pass_rtl_unswitch): Likewise.
921 * rtl.h (reversed_condition): Likewise.
922 (compare_and_jump_seq): Likewise.
923 * loop-iv.c (reversed_condition): Move here from loop-unswitch.c
925 * loop-unroll.c (compare_and_jump_seq): Likewise.
927 2014-04-23 Richard Biener <rguenther@suse.de>
929 PR tree-optimization/60903
930 * tree-ssa-loop-im.c (analyze_memory_references): Remove
931 commented code block.
932 (execute_sm_if_changed): Properly apply IRREDUCIBLE_LOOP
933 loop flags to newly created BBs and edges.
935 2014-04-23 Nick Clifton <nickc@redhat.com>
937 * config/msp430/msp430.c (msp430_handle_option): Move function
939 (msp430_option_override): Simplify mcu and mcpu option handling.
940 (msp430_is_f5_mcu): Rename to msp430_use_f5_series_hwmult. Add
941 support for -mhwmult command line option.
942 (has_32bit_hwmult): Rename to use_32bit_hwmult. Add support for
943 -mhwmult command line option.
944 (msp430_hwmult_enabled): Delete.
945 (msp43o_output_labelref): Add support for -mhwmult command line
947 * config/msp430/msp430.md (mulhisi3, umulhisi3, mulsidi3)
948 (umulsidi3): Likewise.
949 * config/msp430/msp430.opt (mmcu): Add Report attribute.
950 (mcpu, mlarge, msmall): Likewise.
951 (mhwmult): New option.
952 * config/msp430/msp430-protos.h (msp430_hwmult_enabled): Remove
954 (msp430_is_f5_mcu): Remove prototype.
955 (msp430_use_f5_series_hwmult): Add prototype.
956 * config/msp430/msp430-opts.h: New file.
957 * common/config/msp430: New directory.
958 * common/config/msp430/msp430-common.c: New file.
959 * config.gcc (msp430): Remove target_has_targetm_common.
960 * doc/invoke.texi: Document -mhwmult command line option.
962 2014-04-23 Nick Clifton <nickc@redhat.com>
964 * config/i386/cygwin.h (ENDFILE_SPEC): Include
965 default-manifest.o if it can be found in the search path.
966 * config/i386/mingw32.h (ENDFILE_SPEC): Likewise.
968 2014-04-23 Terry Guo <terry.guo@arm.com>
970 * config/arm/arm.h (ASM_APP_OFF): Re-define it in a cleaner way.
972 2014-04-23 Richard Biener <rguenther@suse.de>
975 * tree-inline.c (declare_return_variable): Use mark_addressable.
977 2014-04-23 Richard Biener <rguenther@suse.de>
980 * loop-init.c (loop_optimizer_init): Make sure to apply
981 LOOPS_MAY_HAVE_MULTIPLE_LATCHES before fixing up loops.
983 2014-04-22 Jakub Jelinek <jakub@redhat.com>
986 * common.opt (fsanitize-recover, fsanitize-undefined-trap-on-error):
988 * gcc.c (sanitize_spec_function): Don't return "" for "undefined"
989 if flag_sanitize_undefined_trap_on_error.
990 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW_ABORT,
991 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS_ABORT,
992 BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE_ABORT,
993 BUILT_IN_UBSAN_HANDLE_TYPE_MISMATCH_ABORT,
994 BUILT_IN_UBSAN_HANDLE_ADD_OVERFLOW_ABORT,
995 BUILT_IN_UBSAN_HANDLE_SUB_OVERFLOW_ABORT,
996 BUILT_IN_UBSAN_HANDLE_MUL_OVERFLOW_ABORT,
997 BUILT_IN_UBSAN_HANDLE_NEGATE_OVERFLOW_ABORT,
998 BUILT_IN_UBSAN_HANDLE_LOAD_INVALID_VALUE_ABORT): New builtins.
999 * ubsan.c (ubsan_instrument_unreachable): Return
1000 __builtin_trap () if flag_sanitize_undefined_trap_on_error.
1001 (ubsan_expand_null_ifn): Emit __builtin_trap ()
1002 if flag_sanitize_undefined_trap_on_error and
1003 __ubsan_handle_type_mismatch_abort if !flag_sanitize_recover.
1004 (ubsan_expand_null_ifn, ubsan_build_overflow_builtin,
1005 instrument_bool_enum_load): Emit __builtin_trap () if
1006 flag_sanitize_undefined_trap_on_error and
1007 __builtin_handle_*_abort () if !flag_sanitize_recover.
1008 * doc/invoke.texi (-fsanitize-recover,
1009 -fsanitize-undefined-trap-on-error): Document.
1011 2014-04-22 Christian Bruel <christian.bruel@st.com>
1013 * config/sh/sh.md (mov<mode>): Replace movQIHI.
1014 Force immediates to SImode.
1016 2014-04-22 Sandra Loosemore <sandra@codesourcery.com>
1018 * config/nios2/nios2.md (UNSPEC_ROUND): New.
1020 * config/nios2/nios2.opt (mno-custom-round, mcustom-round=): New.
1021 * config/nios2/nios2-opts.h (N2FPU_ALL_CODES): Add round.
1022 * config/nios2/nios2.c (N2F_NO_ERRNO): Define.
1023 (nios2_fpu_insn): Add entry for round.
1024 (N2FPU_NO_ERRNO_P): Define.
1025 (nios2_custom_check_insns): Add check for N2F_NO_ERRNO and
1027 * doc/invoke.texi (Nios II Options): Document -mcustom-round.
1029 2014-04-22 Richard Henderson <rth@redhat.com>
1031 * config/aarch64/aarch64 (addti3, subti3): New expanders.
1032 (add<GPI>3_compare0): Remove leading * from name.
1033 (add<GPI>3_carryin): Likewise.
1034 (sub<GPI>3_compare0): Likewise.
1035 (sub<GPI>3_carryin): Likewise.
1036 (<su_optab>mulditi3): New expander.
1037 (multi3): New expander.
1038 (madd<GPI>): Remove leading * from name.
1040 2014-04-22 Martin Jambor <mjambor@suse.cz>
1042 * cgraphclones.c (cgraph_function_versioning): Copy
1043 ipa_transforms_to_apply instead of asserting it is empty.
1045 2014-04-22 H.J. Lu <hongjiu.lu@intel.com>
1048 * config/i386/i386.c (ix86_expand_set_or_movmem): Call counter_mode
1049 on count_exp to get mode.
1051 2014-04-22 Andrew Pinski <apinski@cavium.com>
1053 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
1054 Handle TLS for ILP32.
1055 * config/aarch64/aarch64.md (tlsie_small): Rename to ...
1056 (tlsie_small_<mode>): this and handle PTR.
1057 (tlsie_small_sidi): New pattern.
1058 (tlsle_small): Change to an expand to handle ILP32.
1059 (tlsle_small_<mode>): New pattern.
1060 (tlsdesc_small): Rename to ...
1061 (tlsdesc_small_<mode>): this and handle PTR.
1063 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1065 * config/aarch64/aarch64.c (TARGET_FLAGS_REGNUM): Define.
1067 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1069 * config/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
1070 (aarch64_types_signed_unsigned_qualifiers): Qualifier added.
1071 (aarch64_types_signed_poly_qualifiers): Likewise.
1072 (aarch64_types_unsigned_signed_qualifiers): Likewise.
1073 (aarch64_types_poly_signed_qualifiers): Likewise.
1074 (TYPES_REINTERP_SS): Type macro added.
1075 (TYPES_REINTERP_SU): Likewise.
1076 (TYPES_REINTERP_SP): Likewise.
1077 (TYPES_REINTERP_US): Likewise.
1078 (TYPES_REINTERP_PS): Likewise.
1079 (aarch64_fold_builtin): New expression folding added.
1080 * config/aarch64/aarch64-simd-builtins.def (REINTERP):
1081 Declarations removed.
1082 (REINTERP_SS): Declarations added.
1083 (REINTERP_US): Likewise.
1084 (REINTERP_PS): Likewise.
1085 (REINTERP_SU): Likewise.
1086 (REINTERP_SP): Likewise.
1087 * config/aarch64/arm_neon.h (vreinterpret_p8_f64): Implemented.
1088 (vreinterpretq_p8_f64): Likewise.
1089 (vreinterpret_p16_f64): Likewise.
1090 (vreinterpretq_p16_f64): Likewise.
1091 (vreinterpret_f32_f64): Likewise.
1092 (vreinterpretq_f32_f64): Likewise.
1093 (vreinterpret_f64_f32): Likewise.
1094 (vreinterpret_f64_p8): Likewise.
1095 (vreinterpret_f64_p16): Likewise.
1096 (vreinterpret_f64_s8): Likewise.
1097 (vreinterpret_f64_s16): Likewise.
1098 (vreinterpret_f64_s32): Likewise.
1099 (vreinterpret_f64_s64): Likewise.
1100 (vreinterpret_f64_u8): Likewise.
1101 (vreinterpret_f64_u16): Likewise.
1102 (vreinterpret_f64_u32): Likewise.
1103 (vreinterpret_f64_u64): Likewise.
1104 (vreinterpretq_f64_f32): Likewise.
1105 (vreinterpretq_f64_p8): Likewise.
1106 (vreinterpretq_f64_p16): Likewise.
1107 (vreinterpretq_f64_s8): Likewise.
1108 (vreinterpretq_f64_s16): Likewise.
1109 (vreinterpretq_f64_s32): Likewise.
1110 (vreinterpretq_f64_s64): Likewise.
1111 (vreinterpretq_f64_u8): Likewise.
1112 (vreinterpretq_f64_u16): Likewise.
1113 (vreinterpretq_f64_u32): Likewise.
1114 (vreinterpretq_f64_u64): Likewise.
1115 (vreinterpret_s64_f64): Likewise.
1116 (vreinterpretq_s64_f64): Likewise.
1117 (vreinterpret_u64_f64): Likewise.
1118 (vreinterpretq_u64_f64): Likewise.
1119 (vreinterpret_s8_f64): Likewise.
1120 (vreinterpretq_s8_f64): Likewise.
1121 (vreinterpret_s16_f64): Likewise.
1122 (vreinterpretq_s16_f64): Likewise.
1123 (vreinterpret_s32_f64): Likewise.
1124 (vreinterpretq_s32_f64): Likewise.
1125 (vreinterpret_u8_f64): Likewise.
1126 (vreinterpretq_u8_f64): Likewise.
1127 (vreinterpret_u16_f64): Likewise.
1128 (vreinterpretq_u16_f64): Likewise.
1129 (vreinterpret_u32_f64): Likewise.
1130 (vreinterpretq_u32_f64): Likewise.
1132 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1134 * config/aarch64/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
1135 * config/aarch64/aarch64/aarch64-simd-builtins.def (REINTERP): Removed.
1136 (vreinterpret_p8_s8): Likewise.
1137 * config/aarch64/aarch64/arm_neon.h (vreinterpret_p8_s8): Uses cast.
1138 (vreinterpret_p8_s16): Likewise.
1139 (vreinterpret_p8_s32): Likewise.
1140 (vreinterpret_p8_s64): Likewise.
1141 (vreinterpret_p8_f32): Likewise.
1142 (vreinterpret_p8_u8): Likewise.
1143 (vreinterpret_p8_u16): Likewise.
1144 (vreinterpret_p8_u32): Likewise.
1145 (vreinterpret_p8_u64): Likewise.
1146 (vreinterpret_p8_p16): Likewise.
1147 (vreinterpretq_p8_s8): Likewise.
1148 (vreinterpretq_p8_s16): Likewise.
1149 (vreinterpretq_p8_s32): Likewise.
1150 (vreinterpretq_p8_s64): Likewise.
1151 (vreinterpretq_p8_f32): Likewise.
1152 (vreinterpretq_p8_u8): Likewise.
1153 (vreinterpretq_p8_u16): Likewise.
1154 (vreinterpretq_p8_u32): Likewise.
1155 (vreinterpretq_p8_u64): Likewise.
1156 (vreinterpretq_p8_p16): Likewise.
1157 (vreinterpret_p16_s8): Likewise.
1158 (vreinterpret_p16_s16): Likewise.
1159 (vreinterpret_p16_s32): Likewise.
1160 (vreinterpret_p16_s64): Likewise.
1161 (vreinterpret_p16_f32): Likewise.
1162 (vreinterpret_p16_u8): Likewise.
1163 (vreinterpret_p16_u16): Likewise.
1164 (vreinterpret_p16_u32): Likewise.
1165 (vreinterpret_p16_u64): Likewise.
1166 (vreinterpret_p16_p8): Likewise.
1167 (vreinterpretq_p16_s8): Likewise.
1168 (vreinterpretq_p16_s16): Likewise.
1169 (vreinterpretq_p16_s32): Likewise.
1170 (vreinterpretq_p16_s64): Likewise.
1171 (vreinterpretq_p16_f32): Likewise.
1172 (vreinterpretq_p16_u8): Likewise.
1173 (vreinterpretq_p16_u16): Likewise.
1174 (vreinterpretq_p16_u32): Likewise.
1175 (vreinterpretq_p16_u64): Likewise.
1176 (vreinterpretq_p16_p8): Likewise.
1177 (vreinterpret_f32_s8): Likewise.
1178 (vreinterpret_f32_s16): Likewise.
1179 (vreinterpret_f32_s32): Likewise.
1180 (vreinterpret_f32_s64): Likewise.
1181 (vreinterpret_f32_u8): Likewise.
1182 (vreinterpret_f32_u16): Likewise.
1183 (vreinterpret_f32_u32): Likewise.
1184 (vreinterpret_f32_u64): Likewise.
1185 (vreinterpret_f32_p8): Likewise.
1186 (vreinterpret_f32_p16): Likewise.
1187 (vreinterpretq_f32_s8): Likewise.
1188 (vreinterpretq_f32_s16): Likewise.
1189 (vreinterpretq_f32_s32): Likewise.
1190 (vreinterpretq_f32_s64): Likewise.
1191 (vreinterpretq_f32_u8): Likewise.
1192 (vreinterpretq_f32_u16): Likewise.
1193 (vreinterpretq_f32_u32): Likewise.
1194 (vreinterpretq_f32_u64): Likewise.
1195 (vreinterpretq_f32_p8): Likewise.
1196 (vreinterpretq_f32_p16): Likewise.
1197 (vreinterpret_s64_s8): Likewise.
1198 (vreinterpret_s64_s16): Likewise.
1199 (vreinterpret_s64_s32): Likewise.
1200 (vreinterpret_s64_f32): Likewise.
1201 (vreinterpret_s64_u8): Likewise.
1202 (vreinterpret_s64_u16): Likewise.
1203 (vreinterpret_s64_u32): Likewise.
1204 (vreinterpret_s64_u64): Likewise.
1205 (vreinterpret_s64_p8): Likewise.
1206 (vreinterpret_s64_p16): Likewise.
1207 (vreinterpretq_s64_s8): Likewise.
1208 (vreinterpretq_s64_s16): Likewise.
1209 (vreinterpretq_s64_s32): Likewise.
1210 (vreinterpretq_s64_f32): Likewise.
1211 (vreinterpretq_s64_u8): Likewise.
1212 (vreinterpretq_s64_u16): Likewise.
1213 (vreinterpretq_s64_u32): Likewise.
1214 (vreinterpretq_s64_u64): Likewise.
1215 (vreinterpretq_s64_p8): Likewise.
1216 (vreinterpretq_s64_p16): Likewise.
1217 (vreinterpret_u64_s8): Likewise.
1218 (vreinterpret_u64_s16): Likewise.
1219 (vreinterpret_u64_s32): Likewise.
1220 (vreinterpret_u64_s64): Likewise.
1221 (vreinterpret_u64_f32): Likewise.
1222 (vreinterpret_u64_u8): Likewise.
1223 (vreinterpret_u64_u16): Likewise.
1224 (vreinterpret_u64_u32): Likewise.
1225 (vreinterpret_u64_p8): Likewise.
1226 (vreinterpret_u64_p16): Likewise.
1227 (vreinterpretq_u64_s8): Likewise.
1228 (vreinterpretq_u64_s16): Likewise.
1229 (vreinterpretq_u64_s32): Likewise.
1230 (vreinterpretq_u64_s64): Likewise.
1231 (vreinterpretq_u64_f32): Likewise.
1232 (vreinterpretq_u64_u8): Likewise.
1233 (vreinterpretq_u64_u16): Likewise.
1234 (vreinterpretq_u64_u32): Likewise.
1235 (vreinterpretq_u64_p8): Likewise.
1236 (vreinterpretq_u64_p16): Likewise.
1237 (vreinterpret_s8_s16): Likewise.
1238 (vreinterpret_s8_s32): Likewise.
1239 (vreinterpret_s8_s64): Likewise.
1240 (vreinterpret_s8_f32): Likewise.
1241 (vreinterpret_s8_u8): Likewise.
1242 (vreinterpret_s8_u16): Likewise.
1243 (vreinterpret_s8_u32): Likewise.
1244 (vreinterpret_s8_u64): Likewise.
1245 (vreinterpret_s8_p8): Likewise.
1246 (vreinterpret_s8_p16): Likewise.
1247 (vreinterpretq_s8_s16): Likewise.
1248 (vreinterpretq_s8_s32): Likewise.
1249 (vreinterpretq_s8_s64): Likewise.
1250 (vreinterpretq_s8_f32): Likewise.
1251 (vreinterpretq_s8_u8): Likewise.
1252 (vreinterpretq_s8_u16): Likewise.
1253 (vreinterpretq_s8_u32): Likewise.
1254 (vreinterpretq_s8_u64): Likewise.
1255 (vreinterpretq_s8_p8): Likewise.
1256 (vreinterpretq_s8_p16): Likewise.
1257 (vreinterpret_s16_s8): Likewise.
1258 (vreinterpret_s16_s32): Likewise.
1259 (vreinterpret_s16_s64): Likewise.
1260 (vreinterpret_s16_f32): Likewise.
1261 (vreinterpret_s16_u8): Likewise.
1262 (vreinterpret_s16_u16): Likewise.
1263 (vreinterpret_s16_u32): Likewise.
1264 (vreinterpret_s16_u64): Likewise.
1265 (vreinterpret_s16_p8): Likewise.
1266 (vreinterpret_s16_p16): Likewise.
1267 (vreinterpretq_s16_s8): Likewise.
1268 (vreinterpretq_s16_s32): Likewise.
1269 (vreinterpretq_s16_s64): Likewise.
1270 (vreinterpretq_s16_f32): Likewise.
1271 (vreinterpretq_s16_u8): Likewise.
1272 (vreinterpretq_s16_u16): Likewise.
1273 (vreinterpretq_s16_u32): Likewise.
1274 (vreinterpretq_s16_u64): Likewise.
1275 (vreinterpretq_s16_p8): Likewise.
1276 (vreinterpretq_s16_p16): Likewise.
1277 (vreinterpret_s32_s8): Likewise.
1278 (vreinterpret_s32_s16): Likewise.
1279 (vreinterpret_s32_s64): Likewise.
1280 (vreinterpret_s32_f32): Likewise.
1281 (vreinterpret_s32_u8): Likewise.
1282 (vreinterpret_s32_u16): Likewise.
1283 (vreinterpret_s32_u32): Likewise.
1284 (vreinterpret_s32_u64): Likewise.
1285 (vreinterpret_s32_p8): Likewise.
1286 (vreinterpret_s32_p16): Likewise.
1287 (vreinterpretq_s32_s8): Likewise.
1288 (vreinterpretq_s32_s16): Likewise.
1289 (vreinterpretq_s32_s64): Likewise.
1290 (vreinterpretq_s32_f32): Likewise.
1291 (vreinterpretq_s32_u8): Likewise.
1292 (vreinterpretq_s32_u16): Likewise.
1293 (vreinterpretq_s32_u32): Likewise.
1294 (vreinterpretq_s32_u64): Likewise.
1295 (vreinterpretq_s32_p8): Likewise.
1296 (vreinterpretq_s32_p16): Likewise.
1297 (vreinterpret_u8_s8): Likewise.
1298 (vreinterpret_u8_s16): Likewise.
1299 (vreinterpret_u8_s32): Likewise.
1300 (vreinterpret_u8_s64): Likewise.
1301 (vreinterpret_u8_f32): Likewise.
1302 (vreinterpret_u8_u16): Likewise.
1303 (vreinterpret_u8_u32): Likewise.
1304 (vreinterpret_u8_u64): Likewise.
1305 (vreinterpret_u8_p8): Likewise.
1306 (vreinterpret_u8_p16): Likewise.
1307 (vreinterpretq_u8_s8): Likewise.
1308 (vreinterpretq_u8_s16): Likewise.
1309 (vreinterpretq_u8_s32): Likewise.
1310 (vreinterpretq_u8_s64): Likewise.
1311 (vreinterpretq_u8_f32): Likewise.
1312 (vreinterpretq_u8_u16): Likewise.
1313 (vreinterpretq_u8_u32): Likewise.
1314 (vreinterpretq_u8_u64): Likewise.
1315 (vreinterpretq_u8_p8): Likewise.
1316 (vreinterpretq_u8_p16): Likewise.
1317 (vreinterpret_u16_s8): Likewise.
1318 (vreinterpret_u16_s16): Likewise.
1319 (vreinterpret_u16_s32): Likewise.
1320 (vreinterpret_u16_s64): Likewise.
1321 (vreinterpret_u16_f32): Likewise.
1322 (vreinterpret_u16_u8): Likewise.
1323 (vreinterpret_u16_u32): Likewise.
1324 (vreinterpret_u16_u64): Likewise.
1325 (vreinterpret_u16_p8): Likewise.
1326 (vreinterpret_u16_p16): Likewise.
1327 (vreinterpretq_u16_s8): Likewise.
1328 (vreinterpretq_u16_s16): Likewise.
1329 (vreinterpretq_u16_s32): Likewise.
1330 (vreinterpretq_u16_s64): Likewise.
1331 (vreinterpretq_u16_f32): Likewise.
1332 (vreinterpretq_u16_u8): Likewise.
1333 (vreinterpretq_u16_u32): Likewise.
1334 (vreinterpretq_u16_u64): Likewise.
1335 (vreinterpretq_u16_p8): Likewise.
1336 (vreinterpretq_u16_p16): Likewise.
1337 (vreinterpret_u32_s8): Likewise.
1338 (vreinterpret_u32_s16): Likewise.
1339 (vreinterpret_u32_s32): Likewise.
1340 (vreinterpret_u32_s64): Likewise.
1341 (vreinterpret_u32_f32): Likewise.
1342 (vreinterpret_u32_u8): Likewise.
1343 (vreinterpret_u32_u16): Likewise.
1344 (vreinterpret_u32_u64): Likewise.
1345 (vreinterpret_u32_p8): Likewise.
1346 (vreinterpret_u32_p16): Likewise.
1347 (vreinterpretq_u32_s8): Likewise.
1348 (vreinterpretq_u32_s16): Likewise.
1349 (vreinterpretq_u32_s32): Likewise.
1350 (vreinterpretq_u32_s64): Likewise.
1351 (vreinterpretq_u32_f32): Likewise.
1352 (vreinterpretq_u32_u8): Likewise.
1353 (vreinterpretq_u32_u16): Likewise.
1354 (vreinterpretq_u32_u64): Likewise.
1355 (vreinterpretq_u32_p8): Likewise.
1356 (vreinterpretq_u32_p16): Likewise.
1358 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1360 * gcc/config/aarch64/aarch64-simd.md (aarch64_s<optab><mode>):
1362 * config/aarch64/aarch64-simd-builtins.def (sqneg): Iterator
1365 * config/aarch64/arm_neon.h (vqneg_s64): New intrinsic.
1366 (vqnegd_s64): Likewise.
1367 (vqabs_s64): Likewise.
1368 (vqabsd_s64): Likewise.
1370 2014-04-22 Richard Henderson <rth@redhat.com>
1372 * config/sparc/sparc.c (sparc_init_modes): Hoist GET_MODE_SIZE
1373 computation to the top of the loop.
1375 2014-04-22 Renlin <renlin.li@arm.com>
1376 Jiong Wang <jiong.wang@arm.com>
1378 * config/aarch64/aarch64.h (aarch64_frame): Delete "fp_lr_offset".
1379 * config/aarch64/aarch64.c (aarch64_layout_frame)
1380 (aarch64_initial_elimination_offset): Likewise.
1382 2014-04-22 Marcus Shawcroft <marcus.shawcroft@arm.com>
1384 * config/aarch64/aarch64.c (aarch64_initial_elimination_offset):
1387 2014-04-22 Richard Sandiford <rdsandiford@googlemail.com>
1389 * machmode.h (bitwise_mode_for_mode): Declare.
1390 * stor-layout.h (bitwise_type_for_mode): Likewise.
1391 * stor-layout.c (bitwise_mode_for_mode): New function.
1392 (bitwise_type_for_mode): Likewise.
1393 * builtins.c (fold_builtin_memory_op): Use it instead of
1394 int_mode_for_mode and build_nonstandard_integer_type.
1396 2014-04-22 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1398 * config.gcc (enable_obsolete): Remove *-*-solaris2.9*.
1399 (*-*-solaris2.[0-9] | *-*-solaris2.[0-9].*): Mark unsupported.
1400 (*-*-solaris2*): Simplify.
1401 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Likewise.
1402 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Remove
1403 *-*-solaris2.9* handling.
1405 * configure.ac (gcc_cv_as_hidden): Remove test for Solaris 9/x86
1407 (gcc_cv_ld_hidden): Remove *-*-solaris2.9* handling.
1408 (ld_tls_support): Remove i?86-*-solaris2.9, sparc*-*-solaris2.9
1410 (gcc_cv_as_gstabs_flag): Remove workaround for Solaris 9/x86 as bug.
1411 * configure: Regenerate.
1413 * config/i386/sol2-9.h: Remove.
1415 * doc/install.texi (Specific, i?86-*-solaris2.9): Remove.
1416 (Specific, *-*-solaris2*): Mention Solaris 9 support removal.
1417 Remove Solaris 9 references.
1419 2014-04-22 Vidya Praveen <vidyapraveen@arm.com>
1421 * aarch64.md (float<GPI:mode><GPF:mode>2): Remove.
1422 (floatuns<GPI:mode><GPF:mode>2): Remove.
1423 (<optab><fcvt_target><GPF:mode>2): New pattern for equal width float
1424 and floatuns conversions.
1425 (<optab><fcvt_iesize><GPF:mode>2): New pattern for inequal width float
1426 and floatuns conversions.
1427 * iterators.md (fcvt_target, FCVT_TARGET): Support SF and DF modes.
1428 (w1,w2): New mode attributes for inequal width conversions.
1430 2014-04-22 Renlin Li <Renlin.Li@arm.com>
1432 * config/aarch64/aarch64.c (aarch64_print_operand_address): Adjust
1433 the output asm format.
1435 2014-04-22 James Greenhalgh <james.greenhalgh@arm.com>
1437 * config/aarch64/aarch64-simd.md
1438 (aarch64_cm<optab>di): Always split.
1439 (*aarch64_cm<optab>di): New.
1440 (aarch64_cmtstdi): Always split.
1441 (*aarch64_cmtstdi): New.
1443 2014-04-22 Jakub Jelinek <jakub@redhat.com>
1445 PR tree-optimization/60823
1446 * omp-low.c (ipa_simd_modify_function_body): Go through
1447 all SSA_NAMEs and for those refering to vector arguments
1448 which are going to be replaced adjust SSA_NAME_VAR and,
1449 if it is a default definition, change it into a non-default
1450 definition assigned at the beginning of function from new_decl.
1451 (ipa_simd_modify_stmt_ops): Rewritten.
1452 * tree-dfa.c (set_ssa_default_def): When removing default def,
1453 check for NULL loc instead of NULL *loc.
1455 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1457 * config/arm/arm.c (arm_hard_regno_mode_ok): Loosen
1458 restrictions on core registers for DImode values in Thumb2.
1460 2014-04-22 Ian Bolton <ian.bolton@arm.com>
1462 * config/arm/arm.md (*anddi_notdi_zesidi): New pattern.
1463 * config/arm/thumb2.md (*iordi_notdi_zesidi): New pattern.
1465 2014-04-22 Ian Bolton <ian.bolton@arm.com>
1467 * config/arm/thumb2.md (*iordi_notdi_di): New pattern.
1468 (*iordi_notzesidi_di): Likewise.
1469 (*iordi_notsesidi_di): Likewise.
1471 2014-04-22 Ian Bolton <ian.bolton@arm.com>
1473 * config/arm/arm-protos.h (tune_params): New struct members.
1474 * config/arm/arm.c: Initialise tune_params per processor.
1475 (thumb2_reorg): Suppress conversion from t32 to t16 when optimizing
1476 for speed, based on new tune_params.
1478 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1480 * config/aarch64/aarch64-builtins.c (BUILTIN_VDQF_DF): Macro
1482 * config/aarch64/aarch64-simd-builtins.def (frintn): Use added
1484 * config/aarch64/aarch64-simd.md (<frint_pattern>): Comment
1486 * config/aarch64/aarch64.md (<frint_pattern>): Likewise.
1487 * config/aarch64/arm_neon.h (vrnd_f64): Added.
1488 (vrnda_f64): Likewise.
1489 (vrndi_f64): Likewise.
1490 (vrndm_f64): Likewise.
1491 (vrndn_f64): Likewise.
1492 (vrndp_f64): Likewise.
1493 (vrndx_f64): Likewise.
1495 2014-04-22 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1497 * config/arm/arm.c (arm_print_operand, thumb_exit): Make sure
1498 GET_MODE_SIZE argument is enum machine_mode.
1500 2014-04-22 Jakub Jelinek <jakub@redhat.com>
1503 * config/sparc/sparc.c (sparc_init_modes): Pass enum machine_mode
1504 value instead of int to GET_MODE_CLASS and GET_MODE_SIZE macros.
1506 2014-04-22 Lin Zuojian <manjian2006@gmail.com>
1509 * asan.c (asan_emit_stack_protection): Force the base to align to
1510 appropriate bits if STRICT_ALIGNMENT. Set shadow_mem align to
1511 appropriate bits if STRICT_ALIGNMENT.
1512 * cfgexpand.c (expand_stack_vars): Set base_align appropriately
1514 (expand_used_vars): Leave a space in the stack frame for alignment
1515 if STRICT_ALIGNMENT.
1517 2014-04-21 David Malcolm <dmalcolm@redhat.com>
1519 * gimple.h (gimple_assign_single_p): Accept a const_gimple rather
1521 (gimple_store_p): Likewise.
1522 (gimple_assign_load_p): Likewise.
1523 (gimple_assign_cast_p): Likewise.
1524 (gimple_clobber_p): Likewise.
1526 * doc/gimple.texi (gimple_assign_cast_p): Accept a const_gimple
1527 rather than a gimple.
1528 (gimple_assign_cast_p): Likewise.
1530 2014-04-21 Michael Meissner <meissner@linux.vnet.ibm.com>
1533 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64 case):
1534 If mode is DDmode and TARGET_E500_DOUBLE allow move.
1536 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print some
1537 more debug information for E500 if -mdebug=reg.
1539 2014-04-21 Uros Bizjak <ubizjak@gmail.com>
1542 * config/i386/i386.c (ix86_expand_builtin)
1543 <case IX86_BUILTIN_RDRAND{16,32,64}_STEP>: Use temporary
1544 register for target RTX.
1545 <case IX86_BUILTIN_RDSEED{16,32,64}_STEP>: Ditto.
1547 2014-04-18 Cong Hou <congh@google.com>
1549 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): Enhance
1550 the widen-mult pattern by handling two operands with different sizes,
1551 and operands whose size is smaller than half of the result type.
1553 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
1555 * ipa-inline.h (INLINE_HINT_known_hot): New hint.
1556 * ipa-inline-analysis.c (dump_inline_hints): Dump it.
1557 (do_estimate_edge_time): Compute it.
1558 * ipa-inline.c (want_inline_small_function_p): Bypass
1559 INLINE_INSNS_AUTO/SINGLE limits for calls that are known to be hot.
1561 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
1563 * ipa-inline.c (spec_rem): New static variable.
1564 (dump_overall_stats): New function.
1565 (dump_inline_stats): New function.
1567 2014-04-18 Richard Henderson <rth@redhat.com>
1569 * config/aarch64/aarch64.c (aarch64_register_move_cost): Pass a mode
1570 to GET_MODE_SIZE, not a reg_class_t.
1572 2014-04-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1574 * config/rs6000/vsx.md (vsx_xxmrghw_<mode>): Adjust for little-endian.
1575 (vsx_xxmrglw_<mode>): Likewise.
1577 2014-04-17 Michael Meissner <meissner@linux.vnet.ibm.com>
1580 * config/rs6000/rs6000.c (rs6000_setup_reg_addr_masks): Make sure
1581 GET_MODE_SIZE gets passed an enum machine_mode type and not integer.
1582 (rs6000_init_hard_regno_mode_ok): Likewise.
1584 2014-04-17 Jan Hubicka <hubicka@ucw.cz>
1586 * ipa-inline.c (inline_small_functions): Account only non-cold
1588 * doc/invoke.texi (inline-unit-growth): Update documentation.
1590 2014-04-17 Pat Haugen <pthaugen@us.ibm.com>
1592 * config/rs6000/rs6000.md (addti3, subti3): New.
1594 2014-04-17 H.J. Lu <hongjiu.lu@intel.com>
1597 * config/i386/i386.c (ix86_expand_clear): Remove outdated
1598 comment. Check optimize_insn_for_size_p instead of
1599 optimize_insn_for_speed_p.
1601 2014-04-17 Martin Jambor <mjambor@suse.cz>
1603 * gimple-iterator.c (gsi_start_edge): New function.
1604 * gimple-iterator.h (gsi_start_edge): Declare.
1605 * tree-sra.c (single_non_eh_succ): New function.
1606 (disqualify_ops_if_throwing_stmt): Renamed to
1607 disqualify_if_bad_bb_terminating_stmt. Allow throwing statements
1608 having one non-EH successor BB.
1609 (sra_modify_expr): If stmt ends bb, use single non-EH successor to
1610 generate loads into replacements.
1611 (sra_modify_assign): Likewise and and also use the simple path for
1613 (sra_modify_function_body): Commit statements on edges.
1615 2014-04-17 Richard Biener <rguenther@suse.de>
1618 * tree-ssa-propagate.c (valid_gimple_rhs_p): Allow vector
1619 comparison results and add clarifying comment.
1621 2014-04-17 Jakub Jelinek <jakub@redhat.com>
1623 * genmodes.c (struct mode_data): Add need_bytesize_adj field.
1624 (blank_mode): Initialize it.
1625 (emit_mode_size_inline, emit_mode_nunits_inline,
1626 emit_mode_inner_inline): New functions.
1627 (emit_insn_modes_h): Call them and surround their output with
1628 #if GCC_VERSION >= 4001 ... #endif.
1629 * machmode.h (GET_MODE_SIZE, GET_MODE_NUNITS, GET_MODE_INNER):
1630 For GCC_VERSION >= 4001 use mode_*_inline routines instead of
1631 mode_* arrays if the argument is __builtin_constant_p.
1632 * lower-subreg.c (dump_choices): Make sure GET_MODE_SIZE argument
1633 is enum machine_mode.
1635 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1637 * passes.c (opt_pass::execute): Adjust.
1638 (pass_manager::execute_pass_mode_switching): Likewise.
1639 (early_local_passes::execute): Likewise.
1640 (execute_one_pass): Pass cfun to the pass's execute method.
1641 * tree-pass.h (opt_pass::execute): Add function * argument.
1642 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
1643 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
1644 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
1645 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
1646 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
1647 config/sparc/sparc.c, cprop.c, dce.c, df-core.c, dse.c, dwarf2cfi.c,
1648 except.c, final.c, function.c, fwprop.c, gcse.c, gimple-low.c,
1649 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
1650 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
1651 ipa-inline-analysis.c, ipa-inline.c, ipa-profile.c, ipa-pure-const.c,
1652 ipa-reference.c, ipa-split.c, ipa.c, ira.c, jump.c, loop-init.c,
1653 lower-subreg.c, mode-switching.c, omp-low.c, postreload-gcse.c,
1654 postreload.c, predict.c, recog.c, ree.c, reg-stack.c, regcprop.c,
1655 reginfo.c, regrename.c, reorg.c, sched-rgn.c, stack-ptr-mod.c,
1656 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
1657 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
1658 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
1659 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-ssa-ccp.c,
1660 tree-ssa-copy.c, tree-ssa-copyrename.c, tree-ssa-dce.c,
1661 tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
1662 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
1663 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
1664 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
1665 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
1666 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
1667 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
1668 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
1669 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
1670 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
1673 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1675 * passes.c (opt_pass::gate): Take function * argument.
1676 (gate_all_early_local_passes): Merge into
1677 (early_local_passes::gate): this.
1678 (gate_all_early_optimizations): Merge into
1679 (all_early_optimizations::gate): this.
1680 (gate_all_optimizations): Mege into
1681 (all_optimizations::gate): this.
1682 (gate_all_optimizations_g): Merge into
1683 (all_optimizations_g::gate): this.
1684 (gate_rest_of_compilation): Mege into
1685 (rest_of_compilation::gate): this.
1686 (gate_postreload): Merge into
1687 (postreload::gate): this.
1688 (dump_one_pass): Pass cfun to the pass's gate method.
1689 (execute_ipa_summary_passes): Likewise.
1690 (execute_one_pass): Likewise.
1691 (ipa_write_summaries_2): Likewise.
1692 (ipa_write_optimization_summaries_1): Likewise.
1693 (ipa_read_summaries_1): Likewise.
1694 (ipa_read_optimization_summaries_1): Likewise.
1695 (execute_ipa_stmt_fixups): Likewise.
1696 * tree-pass.h (opt_pass::gate): Add function * argument.
1697 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c,
1698 combine-stack-adj.c, combine.c, compare-elim.c,
1699 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
1700 config/rl78/rl78.c, config/sh/sh_optimize_sett_clrt.cc,
1701 config/sh/sh_treg_combine.cc, config/sparc/sparc.c, cprop.c, cse.c,
1702 dce.c, df-core.c, dse.c, dwarf2cfi.c, except.c, fwprop.c, gcse.c,
1703 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
1704 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
1705 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
1706 loop-init.c, lower-subreg.c, mode-switching.c, modulo-sched.c,
1707 omp-low.c, postreload-gcse.c, postreload.c, predict.c, recog.c, ree.c,
1708 reg-stack.c, regcprop.c, regrename.c, reorg.c, sched-rgn.c,
1709 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
1710 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
1711 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c,
1712 tree-nrv.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
1713 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
1714 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
1715 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
1716 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
1717 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
1718 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
1719 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
1720 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
1721 tree-ssa.c, tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
1722 tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c, tsan.c, ubsan.c,
1723 var-tracking.c, vtable-verify.c, web.c: Adjust.
1725 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1727 * configure.ac: Check for -Woverloaded-virtual and enable it if found.
1728 * configure: Regenerate.
1730 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1732 * passes.c (dump_one_pass): don't check pass->has_gate.
1733 (execute_ipa_summary_passes): Likewise.
1734 (execute_one_pass): Likewise.
1735 (ipa_write_summaries_2): Likewise.
1736 (ipa_write_optimization_summaries_1): Likewise.
1737 (ipa_read_optimization_summaries_1): Likewise.
1738 (execute_ipa_stmt_fixups): Likewise.
1739 * tree-pass.h (pass_data::has_gate): Remove.
1740 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
1741 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
1742 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
1743 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
1744 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
1745 config/sh/sh_optimize_sett_clrt.cc, config/sh/sh_treg_combine.cc,
1746 config/sparc/sparc.c, cprop.c, cse.c, dce.c, df-core.c, dse.c,
1747 dwarf2cfi.c, except.c, final.c, function.c, fwprop.c, gcse.c,
1748 gimple-low.c, gimple-ssa-isolate-paths.c,
1749 gimple-ssa-strength-reduction.c, graphite.c, ifcvt.c, init-regs.c,
1750 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
1751 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
1752 ira.c, jump.c, loop-init.c, lower-subreg.c, mode-switching.c,
1753 modulo-sched.c, omp-low.c, postreload-gcse.c, postreload.c, predict.c,
1754 recog.c, ree.c, reg-stack.c, regcprop.c, reginfo.c, regrename.c,
1755 reorg.c, sched-rgn.c, stack-ptr-mod.c, store-motion.c, tracer.c,
1756 trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
1757 tree-complex.c, tree-eh.c, tree-emutls.c, tree-if-conv.c,
1758 tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
1759 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
1760 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
1761 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
1762 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
1763 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
1764 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
1765 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
1766 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
1767 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
1768 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
1769 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
1770 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
1773 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1775 * pass_manager.h (pass_manager::register_dump_files_1): Remove
1777 * passes.c (pass_manager::register_dump_files_1): Merge into
1778 (pass_manager::register_dump_files): this, and remove its handling of
1779 properties since the pass always has the properties anyway.
1780 (pass_manager::pass_manager): Adjust.
1782 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1784 * pass_manager.h (pass_manager::register_dump_files_1): Adjust.
1785 * passes.c (pass_manager::register_dump_files_1): Remove dead code
1786 dealing with properties.
1787 (pass_manager::register_dump_files): Adjust.
1789 2014-03-20 Mark Wielaard <mjw@redhat.com>
1791 * dwarf2out.c (add_bound_info): If HOST_WIDE_INT is big enough,
1792 then represent the bound as normal constant value.
1794 2014-04-17 Jakub Jelinek <jakub@redhat.com>
1797 Forward port from 4.8 branch
1798 2013-07-19 Kirill Yukhin <kirill.yukhin@intel.com>
1800 * config/i386/bmiintrin.h (_blsi_u32): New.
1804 (_blsmsk_u32): Ditto.
1805 (_blsmsk_u64): Ditto.
1806 (_tzcnt_u32): Ditto.
1807 (_tzcnt_u64): Ditto.
1809 2014-04-17 Kito Cheng <kito@0xlab.org>
1811 * gcc.c (used_arg): Prevent out of bound access for multilib_options.
1813 2014-04-17 Richard Biener <rguenther@suse.de>
1816 * tree-ssa-propagate.c (valid_gimple_rhs_p): Only allow effective
1817 boolean results for comparisons.
1819 2014-04-17 Richard Biener <rguenther@suse.de>
1821 PR tree-optimization/60836
1822 * tree-vect-loop.c (vect_create_epilog_for_reduction): Force
1823 initial PHI args to be gimple values.
1825 2014-04-17 Richard Biener <rguenther@suse.de>
1827 PR tree-optimization/60841
1828 * tree-vect-data-refs.c (vect_analyze_data_refs): Count stmts.
1829 * tree-vect-loop.c (vect_analyze_loop_2): Pass down number
1830 of stmts to SLP build.
1831 * tree-vect-slp.c (vect_slp_analyze_bb_1): Likewise.
1832 (vect_analyze_slp): Likewise.
1833 (vect_analyze_slp_instance): Likewise.
1834 (vect_build_slp_tree): Limit overall SLP tree growth.
1835 * tree-vectorizer.h (vect_analyze_data_refs,
1836 vect_analyze_slp): Adjust prototypes.
1838 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
1840 * config/i386/i386.c (x86_add_stmt_cost): Fix vector cost model for
1843 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
1845 * config/i386/x86-tune.def (TARGET_SLOW_PSHUFB): New tune definition.
1846 * config/i386/i386.h (TARGET_SLOW_PSHUFB): New tune flag.
1847 * config/i386/i386.c (expand_vec_perm_even_odd_1): Avoid byte shuffles
1848 for TARGET_SLOW_PSHUFB
1850 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
1852 * config/i386/i386.c (slm_cost): Adjust vec_to_scalar_cost.
1853 * config/i386/i386.c (intel_cost): Ditto.
1855 2014-04-17 Joey Ye <joey.ye@arm.com>
1857 * opts.c (OPT_fif_conversion, OPT_fif_conversion2): Disable for Og.
1859 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
1861 * opts.c (common_handle_option): Disable -fipa-reference coorectly
1864 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
1866 * ipa-devirt.c (odr_type_d): Add field all_derivations_known.
1867 (type_all_derivations_known_p): New predicate.
1868 (type_all_ctors_visible_p): New predicate.
1869 (type_possibly_instantiated_p): New predicate.
1870 (get_odr_type): Compute all_derivations_known.
1871 (dump_odr_type): Dump the flag.
1872 (maybe_record_type): Cleanup.
1873 (record_target_from_binfo): Add bases_to_consider array;
1874 record bases for types w/o instances and skip CXX destructor.
1875 (possible_polymorphic_call_targets_1): Add bases_to_consider
1876 and consider_construction parameters; check if type may have instance.
1877 (get_polymorphic_call_info): Set maybe_in_construction to true
1878 when we know nothing.
1879 (record_targets_from_bases): Skip CXX destructors; they are
1880 never called for types in construction.
1881 (possible_polymorphic_call_targets): Do not record target when
1882 type may not have instance.
1884 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
1887 * ipa.c (symtab_remove_unreachable_nodes): Mark targets of
1888 external aliases alive, too.
1890 2014-04-16 Andrew Pinski <apinski@cavium.com>
1892 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Change aarch64 ilp32
1895 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
1897 * final.c (compute_alignments): Do not apply loop alignment to a block
1898 falling through to the exit.
1900 2014-04-16 Catherine Moore <clm@codesourcery.com>
1902 * mips.md (*mov<mode>_internal, *movhi_internal, *movqi_internal):
1903 Adjust constraints for microMIPS store patterns.
1905 2014-04-16 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
1907 * config/avr/avr-mcus.def: Correct typo for atxmega256a3bu macro.
1909 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
1911 * tree-ssa-operands.c (create_vop_var): Set DECL_IGNORED_P.
1912 (append_use): Run at -O0.
1913 (append_vdef): Likewise.
1914 * tree-ssa-ter.c (ter_is_replaceable_p): Do not special-case -O0.
1915 * tree-ssa-uninit.c (warn_uninitialized_vars): Remove obsolete comment.
1917 2014-04-16 Jakub Jelinek <jakub@redhat.com>
1919 PR tree-optimization/60844
1920 * tree-ssa-reassoc.c (reassoc_remove_stmt): New function.
1921 (propagate_op_to_single_use, remove_visited_stmt_chain,
1922 linearize_expr, repropagate_negates, reassociate_bb): Use it
1923 instead of gsi_remove.
1925 2014-04-16 Martin Jambor <mjambor@suse.cz>
1927 * cgraphclones.c (cgraph_create_virtual_clone): Duplicate
1928 ipa_transforms_to_apply.
1929 (cgraph_function_versioning): Assert that old_node has empty
1930 ipa_transforms_to_apply.
1931 * trans-mem.c (ipa_tm_create_version): Likewise.
1932 * tree-inline.c (tree_function_versioning): Do not duplicate
1933 ipa_transforms_to_apply.
1935 2014-04-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1938 * configure.ac (set_have_as_tls): Merge i[34567]86-*-* and
1940 Pass necessary as flags on 64-bit Solaris/x86.
1941 Use lowercase relocs for x86_64-*-*.
1942 * configure: Regenerate.
1944 2014-04-15 Jan Hubicka <jh@suse.cz>
1946 * ipa-devirt.c (referenced_from_vtable_p): New predicate.
1947 (maybe_record_node, likely_target_p): Use it.
1949 2014-04-15 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1952 Revert following patch
1954 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
1957 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
1958 software floating point or no floating point registers, do not
1959 allow any type in the FPRs. Eliminate a test for SPE SIMD types
1960 in GPRs that occurs after we tested for GPRs that would never be
1963 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
1964 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
1965 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
1966 specifically allow DDmode, since that does not use the SPE SIMD
1969 2014-03-21 Mark Wielaard <mjw@redhat.com>
1971 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_const_value
1972 as unsigned or int depending on type and value used.
1974 2014-04-15 Richard Biener <rguenther@suse.de>
1976 PR rtl-optimization/56965
1977 * alias.c (ncr_compar, nonoverlapping_component_refs_p): Move ...
1978 * tree-ssa-alias.c (ncr_compar, nonoverlapping_component_refs_p):
1980 * alias.c (true_dependence_1): Do not call
1981 nonoverlapping_component_refs_p.
1982 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Call
1983 nonoverlapping_component_refs_p.
1984 (indirect_refs_may_alias_p): Likewise.
1986 2014-04-15 Teresa Johnson <tejohnson@google.com>
1988 * cfg.c (dump_bb_info): Fix flags check.
1989 * tree-cfg.c (remove_bb): Only dump TDF_BLOCKS when removing.
1991 2014-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1993 PR rtl-optimization/60663
1994 * config/arm/arm.c (arm_new_rtx_costs): Improve ASM_OPERANDS case,
1997 2014-04-15 Richard Biener <rguenther@suse.de>
1999 * lto-streamer.h (LTO_major_version): Bump to 4.
2001 2014-04-15 Richard Biener <rguenther@suse.de>
2003 * common.opt (lto_partition_model): New enum.
2004 (flto-partition=): Merge separate options with a single with argument,
2005 add -flto-partition=one support.
2006 * flag-types.h (enum lto_partition_model): Declare.
2007 * opts.c (finish_options): Remove duplicate -flto-partition=
2009 * lto-wrapper.c (run_gcc): Adjust.
2011 2014-04-15 Richard Biener <rguenther@suse.de>
2013 * alias.c (ncr_compar): New function.
2014 (nonoverlapping_component_refs_p): Re-implement in O (n log n).
2016 2014-04-15 Richard Biener <rguenther@suse.de>
2018 * alias.c (record_component_aliases): Do not walk BINFOs.
2020 2014-04-15 Richard Biener <rguenther@suse.de>
2022 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
2023 Add struct function argument and adjust.
2024 (find_func_aliases_for_call): Likewise.
2025 (find_func_aliases): Likewise.
2026 (find_func_clobbers): Likewise.
2027 (intra_create_variable_infos): Likewise.
2028 (compute_points_to_sets): Likewise.
2029 (ipa_pta_execute): Adjust. Do not push/pop cfun.
2031 2014-04-15 Richard Biener <rguenther@suse.de>
2033 * tree.c (iterative_hash_expr): Use enum tree_code_class
2034 to store TREE_CODE_CLASS.
2035 (tree_block): Likewise.
2036 (tree_set_block): Likewise.
2037 * tree.h (fold_build_pointer_plus_loc): Use
2038 convert_to_ptrofftype_loc.
2040 2014-04-15 Jakub Jelinek <jakub@redhat.com>
2043 * Makefile.in (PLUGIN_HEADERS): Add various headers that have been
2046 2014-04-15 Eric Botcazou <ebotcazou@adacore.com>
2048 * cfgloop.h (struct loop): Move force_vectorize down.
2049 * gimplify.c (gimple_boolify) <ANNOTATE_EXPR>: Handle new kinds.
2050 (gimplify_expr) <ANNOTATE_EXPR>: Minor tweak.
2051 * lto-streamer-in.c (input_cfg): Read dont_vectorize field.
2052 * lto-streamer-out.c (output_cfg): Write dont_vectorize field.
2053 * tree-cfg.c (replace_loop_annotate): Revamp and handle new kinds.
2054 * tree-core.h (enum annot_expr_kind): Add new kind values.
2055 * tree-inline.c (copy_loops): Copy dont_vectorize field and reorder.
2056 * tree-pretty-print.c (dump_generic_node) <ANNOTATE_EXPR>: Handle new
2058 * tree.def (ANNOTATE_EXPR): Tweak comment.
2060 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
2062 * ipa-devirt.c (maybe_record_node): Ignore all non-methods (including
2065 2014-04-14 Paolo Carlini <paolo.carlini@oracle.com>
2067 * tree.h (TYPE_IDENTIFIER): Declare.
2068 * tree.c (subrange_type_for_debug_p): Use it.
2069 * godump.c (go_format_type): Likewise.
2070 * dwarf2out.c (is_cxx_auto, modified_type_die,
2071 gen_type_die_with_usage, gen_type_die_with_usage): Likewise.
2072 * dbxout.c (dbxout_type, dbxout_symbol): Likewise.
2074 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
2077 * varpool.c (varpool_remove_node): Do not alter decls when streaming.
2079 2014-04-14 Uros Bizjak <ubizjak@gmail.com>
2081 * config/i386/i386.c (examine_argument): Return bool. Return true if
2082 parameter should be passed in memory.
2083 <case X86_64_COMPLEX_X87_CLASS>: Adjust.
2084 (construct_container): Update calls to examine_argument.
2085 (function_arg_advance_64): Ditto.
2086 (return_in_memory_32): Merge with ix86_return_in_memory.
2087 (return_in_memory_64): Ditto.
2088 (return_in_memory_ms_64): Ditto.
2090 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
2092 * ipa-utils.c (ipa_merge_profiles): Merge profile_id.
2093 * coverage.c (coverage_compute_profile_id): Handle externally visible
2096 2014-04-14 Martin Jambor <mjambor@suse.cz>
2098 * tree-sra.c (ipa_sra_preliminary_function_checks): Skip
2099 DECL_DISREGARD_INLINE_LIMITS functions.
2101 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
2104 * config/i386/i386.md (*fixuns_trunc<mode>_1): Revert the last change.
2106 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
2109 * config/i386/i386.md (*fixuns_trunc<mode>_1): Check
2110 optimize_insn_for_speed_p instead of
2111 optimize_function_for_speed_p.
2113 2014-04-14 Yufeng Zhang <yufeng.zhang@arm.com>
2115 * doc/invoke.texi (free): Document AArch64.
2117 2014-04-14 Richard Biener <rguenther@suse.de>
2119 PR tree-optimization/60042
2120 * tree-ssa-pre.c (inhibit_phi_insertion): Remove.
2121 (insert_into_preds_of_block): Do not prevent PHI insertion
2122 for REFERENCE exprs here ...
2123 (eliminate_dom_walker::before_dom_children): ... but prevent
2124 their use here under similar conditions when applied to the
2125 IL after PRE optimizations.
2127 2014-04-14 Richard Biener <rguenther@suse.de>
2129 * passes.def: Move early points-to after early SRA.
2131 2014-04-14 Richard Biener <rguenther@suse.de>
2133 * tree-ssa-forwprop.c (simplify_gimple_switch): Enhance
2134 check for which sign-changes we allow when forwarding
2135 a converted value into a switch.
2137 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
2139 * stor-layout.c (place_field): Finalize non-constant offset for the
2142 2014-04-14 Richard Biener <rguenther@suse.de>
2144 * tree-switch-conversion.c (lshift_cheap_p): Get speed_p
2146 (expand_switch_using_bit_tests_p): Likewise.
2147 (process_switch): Compute and pass on speed_p based on the
2149 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Use
2150 optimize_bb_for_speed_p.
2152 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
2154 * cfgloop.h (struct loop): Rename force_vect into force_vectorize.
2155 * function.h (struct function): Rename has_force_vect_loops into
2156 has_force_vectorize_loops.
2157 * lto-streamer-in.c (input_cfg): Adjust for renaming.
2158 (input_struct_function_base): Likewise.
2159 * lto-streamer-out.c (output_cfg): Likewise.
2160 (output_struct_function_base): Likewise.
2161 * omp-low.c (expand_omp_simd): Likewise.
2162 * tree-cfg.c (move_sese_region_to_fn): Likewise.
2163 * tree-if-conv.c (ifcvt_can_use_mask_load_store): Likewise.
2164 (version_loop_for_if_conversion): Likewise.
2165 (tree_if_conversion): Likewise.
2166 (main_tree_if_conversion): Likewise.
2167 (gate_tree_if_conversion): Likewise.
2168 * tree-inline.c (copy_loops): Likewise.
2169 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Likewise.
2170 * tree-ssa-loop.c (tree_loop_vectorize): Likewise.
2171 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Likewise.
2172 * tree-vect-loop.c (vect_estimate_min_profitable_iters): Likewise.
2173 * tree-vectorizer.c (vectorize_loops): Likewise.
2174 * tree-vectorizer.h (unlimited_cost_model): Likewise.
2176 2014-04-14 Richard Biener <rguenther@suse.de>
2179 * lto-streamer-out.c (wrap_refs): New function.
2180 (lto_output): Wrap symbol references in global initializes in
2181 type-preserving MEM_REFs.
2183 2014-04-14 Christian Bruel <christian.bruel@st.com>
2185 * config/sh/sh-mem.cc (sh_expand_strlen): Unroll last word.
2187 2014-04-14 Christian Bruel <christian.bruel@st.com>
2189 * config/sh/sh.md (setmemqi): New expand pattern.
2190 * config/sh/sh.h (CLEAR_RATIO): Define.
2191 * config/sh/sh-mem.cc (sh_expand_setmem): Define.
2192 * config/sh/sh-protos.h (sh_expand_setmem): Declare.
2194 2014-04-14 Richard Biener <rguenther@suse.de>
2197 * fold-const.c (negate_expr_p): Don't negate directional rounding
2199 (fold_negate_expr): Likewise.
2201 2014-04-14 Richard Biener <rguenther@suse.de>
2203 PR tree-optimization/59817
2204 PR tree-optimization/60453
2205 * graphite-scop-detection.c (graphite_can_represent_scev): Complete
2206 recursion to catch all CHRECs in the scalar evolution and restrict
2207 the predicate for the remains appropriately.
2209 2014-04-12 Catherine Moore <clm@codesourcery.com>
2211 * config/mips/constraints.md: Add new register constraint "kb".
2212 * config/mips/mips.md (*mov<mode>_internal): Use constraint "kb".
2213 (*movhi_internal): Likewise.
2214 (*movqi_internal): Likewise.
2215 * config/mips/mips.h (M16_STORE_REGS): New register class.
2216 (REG_CLASS_NAMES): Add M16_STORE_REGS.
2217 (REG_CLASS_CONTENTS): Likewise.
2218 * config/mips/mips.c (mips_regno_to_class): Add M16_STORE_REGS.
2220 2014-04-11 Tobias Burnus <burnus@net-b.de>
2223 * doc/invoke.texi (-Wformat-signedness): Document it.
2224 (Wformat=2): Mention that this enables -Wformat-signedness.
2226 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
2228 * common/config/epiphany/epiphany-common.c
2229 (epiphany_option_optimization_table): Enable section anchors by
2230 default at -O1 or higher.
2231 * config/epiphany/epiphany.c (TARGET_MAX_ANCHOR_OFFSET): Define.
2232 (TARGET_MIN_ANCHOR_OFFSET): Likewise.
2233 (epiphany_rtx_costs) <SET>: For binary operators, the set as such
2234 carries no extra cost.
2235 (epiphany_legitimate_address_p): For BLKmode, apply SImode check.
2236 * config/epiphany/epiphany.h (ASM_OUTPUT_DEF): Define.
2237 * config/epiphany/predicates.md (memclob_operand): New predicate.
2238 * config/epiphany/epiphany.md (stack_adjust_add, stack_adjust_str):
2239 Use memclob_operand predicate and X constraint for operand 3.
2241 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
2243 * config/epiphany/epiphany.c (epiphany_rtx_cost): Compare
2244 with CC_N_NE / CC_C_LTU / CC_C_GTU carries no extra cost for
2247 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
2249 PR rtl-optimization/60651
2250 * mode-switching.c (optimize_mode_switching): Make sure to emit
2251 sets of a lower numbered entity before sets of a higher numbered
2252 entity to a mode of the same or lower priority.
2253 When creating a seginfo for a basic block that starts with a code
2254 label, move the insertion point past the code label.
2255 (new_seginfo): Document and enforce requirement that
2256 NOTE_INSN_BASIC_BLOCK only appears for empty blocks.
2257 * doc/tm.texi.in: Document ordering constraint for emitted mode sets.
2258 * doc/tm.texi: Regenerate.
2260 2014-01-11 Joern Rennecke <joern.rennecke@embecosm.com>
2263 * config/arc/arc.c (arc_save_restore): Fix assert typo.
2265 2013-04-11 Jakub Jelinek <jakub@redhat.com>
2267 * BASE-VER: Set to 4.10.0.
2269 2014-04-11 Tobias Burnus <burnus@net-b.de>
2272 * doc/bugreport.texi (Bugs): Remove nodes pointing to the nirvana.
2273 * doc/gcc.texi (Service): Update description in the @menu
2274 * doc/invoke.texi (Option Summary): Remove misplaced and
2277 2014-04-11 Steve Ellcey <sellcey@mips.com>
2278 Jakub Jelinek <jakub@redhat.com>
2281 * expr.c (convert_move): Use emit_store_flag_force instead of
2282 emit_store_flag. Pass lowpart_mode instead of VOIDmode as 5th
2285 2014-04-11 Richard Biener <rguenther@suse.de>
2288 * varasm.c (assemble_alias): Avoid endless error reporting
2289 recursion by setting TREE_ASM_WRITTEN.
2291 2014-04-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2293 * config/s390/s390.md: Add a splitter for NOT rtx.
2295 2014-04-11 Jakub Jelinek <jakub@redhat.com>
2297 PR rtl-optimization/60663
2298 * cse.c (cse_insn): Set src_volatile on ASM_OPERANDS in PARALLEL.
2300 2014-04-10 Jan Hubicka <hubicka@ucw.cz>
2301 Jakub Jelinek <jakub@redhat.com>
2304 * ipa.c (function_and_variable_visibility): Copy forced_by_abi
2305 flag from decl_node to node.
2307 2014-04-10 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2310 * config/arm/arm.c (TARGET_CONST_NOT_OK_FOR_DEBUG_P): Define
2311 (arm_const_not_ok_for_debug_p): Reject MINUS with SYM_REF's
2312 ameliorating the cases where it can be.
2314 2014-04-09 David Edelsohn <dje.gcc@gmail.com>
2317 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
2319 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
2320 (loadsync_<mode>): Change mode.
2321 (load_quadpti, store_quadpti): New.
2322 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
2323 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
2324 * config/rs6000/predicates.md (quad_memory_operand): !TARGET_SYNC_TI.
2326 2014-04-09 Cong Hou <congh@google.com>
2329 * doc/sourcebuild.texi (vect_widen_mult_si_to_di_pattern): Add
2332 2014-04-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2334 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Use vnand
2335 instead of vnor to exploit possible fusion opportunity in the
2337 (altivec_expand_vec_perm_const_le): Likewise.
2339 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
2341 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
2342 (loadsync_<mode>): Change mode.
2343 (load_quadpti, store_quadpti): New.
2344 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
2345 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
2347 2014-04-08 Richard Sandiford <rdsandiford@googlemail.com>
2350 * config/rs6000/vsx.md (vsx_xscvdpspn_scalar): Change input to DImode.
2351 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Update accordingly.
2352 Use gen_rtx_REG rather than simplify_gen_subreg for op0_di.
2354 2014-04-08 Richard Biener <rguenther@suse.de>
2357 * tree-pretty-print.c (pp_double_int): For HWI32 hosts with
2358 a 64bit widest int print double-int similar to on HWI64 hosts.
2360 2014-04-08 Richard Biener <rguenther@suse.de>
2362 PR tree-optimization/60785
2363 * graphite-sese-to-poly.c (rewrite_phi_out_of_ssa): Treat
2364 default defs properly.
2366 2014-04-08 Nathan Sidwell <nathan@codesourcery.com>
2368 * doc/invoke (Wnon-virtual-dtor): Update to match implementation.
2369 (Weffc++): Likewise.
2371 2014-04-07 Jan Hubicka <hubcika@ucw.cz>
2373 * ipa-devirt.c (maybe_record_node): When node is not recorded,
2374 set completep to false rather than true.
2376 2014-04-07 Douglas B Rupp <rupp@adacore.com>
2379 * config/arm/arm.h (ASM_PREFERRED_EH_DATA_FORMAT): Expose from
2380 ARM_TARGET2_DWARF_FORMAT.
2382 2014-04-07 Charles Baylis <charles.baylis@linaro.org>
2385 * config/arm/arm.h (ASM_OUTPUT_CASE_END): Remove.
2386 (LABEL_ALIGN_AFTER_BARRIER): Align barriers which occur after
2389 2014-04-07 Richard Biener <rguenther@suse.de>
2391 PR tree-optimization/60766
2392 * tree-ssa-loop-ivopts.c (cand_value_at): Compute in an unsigned type.
2393 (may_eliminate_iv): Convert cand_value_at result to desired type.
2395 2014-04-07 Jason Merrill <jason@redhat.com>
2398 * common.opt (-fno-gnu-unique): Add.
2399 * config/elfos.h (USE_GNU_UNIQUE_OBJECT): Check it.
2401 2014-04-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2403 * haifa-sched.c: Fix outdated function reference and minor
2404 grammar errors in introductory comment.
2406 2014-04-07 Richard Biener <rguenther@suse.de>
2409 * tree-ssa-operands.c (maybe_add_call_vops): Also add VDEFs
2411 * tree-cfgcleanup.c (fixup_noreturn_call): Do not remove VDEFs.
2413 2014-04-06 John David Anglin <danglin@gcc.gnu.org>
2416 * config/pa/pa.c (pa_output_function_epilogue): Skip address and code
2417 size accounting for thunks.
2418 (pa_asm_output_mi_thunk): Use final_start_function() and
2419 final_end_function() to output function start and end directives.
2421 2014-04-05 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
2423 * config/avr/avr-arch.h (avr_mcu_t): Add dev_attribute field to have
2424 device specific ISA/ feature information. Remove short_sp and
2425 errata_skip ds. Add avr_device_specific_features enum to have device
2427 * config/avr/avr-c.c (avr_cpu_cpp_builtins): use dev_attribute to check
2428 errata_skip. Add __AVR_ISA_RMW__ builtin macro if RMW ISA available.
2429 * config/avr/avr-devices.c (avr_mcu_types): Update AVR_MCU macro for
2430 updated device specific info.
2431 * config/avr/avr-mcus.def: Merge device specific details to
2432 dev_attribute field.
2433 * config/avr/avr.c (avr_2word_insn_p): use dev_attribute field to check
2435 * config/avr/avr.h (AVR_HAVE_8BIT_SP): same for short sp info.
2436 * config/avr/driver-avr.c (avr_device_to_as): Pass -mrmw option to
2437 assembler if RMW isa supported by current device.
2438 * config/avr/genmultilib.awk: Update as device info structure changed.
2439 * doc/invoke.texi: Add info for __AVR_ISA_RMW__ builtin macro
2441 2014-04-04 Cong Hou <congh@google.com>
2443 PR tree-optimization/60656
2444 * tree-vect-stmts.c (supportable_widening_operation):
2445 Fix a bug that elements in a vector with vect_used_by_reduction
2446 property are incorrectly reordered when the operation on it is not
2447 consistant with the one in reduction operation.
2449 2014-04-04 John David Anglin <danglin@gcc.gnu.org>
2451 PR rtl-optimization/60155
2452 * gcse.c (record_set_data): New function.
2453 (single_set_gcse): New function.
2454 (gcse_emit_move_after): Use single_set_gcse instead of single_set.
2455 (hoist_code): Likewise.
2456 (get_pressure_class_and_nregs): Likewise.
2458 2014-04-04 Eric Botcazou <ebotcazou@adacore.com>
2460 * explow.c (probe_stack_range): Emit a final optimization blockage.
2462 2014-04-04 Anthony Green <green@moxielogic.com>
2464 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2): Fix
2467 2014-04-04 Jan Hubicka <hubicka@ucw.cz>
2470 * lto-cgraph.c (input_overwrite_node): Check that partitioning
2471 flags are set only during streaming.
2472 * ipa.c (process_references, walk_polymorphic_call_targets,
2473 symtab_remove_unreachable_nodes): Drop bodies of always inline
2474 after early inlining.
2475 (symtab_remove_unreachable_nodes): Remove always_inline attribute.
2477 2014-04-04 Jakub Jelinek <jakub@redhat.com>
2478 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2481 * dwarf2out.c (const_ok_for_output_1): Reject expressions
2484 2014-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2487 * config/arm/cortex-a53.md (cortex_a53_fdivs): Reduce reservation
2489 (cortex_a53_fdivd): Likewise.
2491 2014-04-04 Martin Jambor <mjambor@suse.cz>
2494 * cgraph.h (cgraph_clone_node): New parameter added to declaration.
2496 * cgraph.c (clone_of_p): Also return true if thunks match.
2497 (verify_edge_corresponds_to_fndecl): Removed extraneous call to
2498 cgraph_function_or_thunk_node and an obsolete comment.
2499 * cgraphclones.c (build_function_type_skip_args): Moved upwards in the
2501 (build_function_decl_skip_args): Likewise.
2502 (set_new_clone_decl_and_node_flags): New function.
2503 (duplicate_thunk_for_node): Likewise.
2504 (redirect_edge_duplicating_thunks): Likewise.
2505 (cgraph_clone_node): New parameter args_to_skip, pass it to
2506 redirect_edge_duplicating_thunks which is called instead of
2507 cgraph_redirect_edge_callee.
2508 (cgraph_create_virtual_clone): Pass args_to_skip to cgraph_clone_node,
2509 moved setting of a lot of flags to set_new_clone_decl_and_node_flags.
2511 2014-04-04 Jeff Law <law@redhat.com>
2514 * config/arm/predicates.md (const_int_I_operand): New predicate.
2515 (const_int_M_operand): Similarly.
2516 * config/arm/arm.md (insv_zero): Use const_int_M_operand instead of
2518 (insv_t2, extv_reg, extzv_t2): Likewise.
2519 (load_multiple_with_writeback): Similarly for const_int_I_operand.
2520 (pop_multiple_with_writeback_and_return): Likewise.
2521 (vfp_pop_multiple_with_writeback): Likewise
2523 2014-04-04 Richard Biener <rguenther@suse.de>
2526 * tree-ssanames.c (make_ssa_name_fn): Fix assert.
2527 * gimple.c (gimple_set_bb): Avoid ICEing for NULL cfun for
2529 * gimplify.h (gimple_add_tmp_var_fn): Declare.
2530 * gimplify.c (gimple_add_tmp_var_fn): New function.
2531 * gimple-expr.h (create_tmp_reg_fn): Declare.
2532 * gimple-expr.c (create_tmp_reg_fn): New function.
2533 * gimple-low.c (record_vars_into): Don't change cfun.
2534 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix
2535 code generation without cfun.
2537 2014-04-04 Thomas Schwinge <thomas@codesourcery.com>
2540 * Makefile.in (install-driver): Fix shell scripting.
2542 2014-04-03 Cong Hou <congh@google.com>
2544 PR tree-optimization/60505
2545 * tree-vectorizer.h (struct _stmt_vec_info): Add th field as the
2546 threshold of number of iterations below which no vectorization
2548 * tree-vect-loop.c (new_loop_vec_info):
2549 Initialize LOOP_VINFO_COST_MODEL_THRESHOLD.
2550 * tree-vect-loop.c (vect_analyze_loop_operations):
2551 Set LOOP_VINFO_COST_MODEL_THRESHOLD.
2552 * tree-vect-loop.c (vect_transform_loop):
2553 Use LOOP_VINFO_COST_MODEL_THRESHOLD.
2554 * tree-vect-loop.c (vect_analyze_loop_2): Check the maximum number
2555 of iterations of the loop and see if we should build the epilogue.
2557 2014-04-03 Richard Biener <rguenther@suse.de>
2559 * tree-streamer.h (struct streamer_tree_cache_d): Add next_idx member.
2560 (streamer_tree_cache_create): Adjust.
2561 * tree-streamer.c (streamer_tree_cache_add_to_node_array): Adjust
2562 to allow optional nodes array.
2563 (streamer_tree_cache_insert_1): Use next_idx to assign idx.
2564 (streamer_tree_cache_append): Likewise.
2565 (streamer_tree_cache_create): Create nodes array optionally
2566 as specified by parameter.
2567 * lto-streamer-out.c (create_output_block): Avoid maintaining
2568 the node array in the writer cache.
2569 (DFS_write_tree): Remove assertion.
2570 (produce_asm_for_decls): Free the out decl state hash table early.
2571 * lto-streamer-in.c (lto_data_in_create): Adjust for
2572 streamer_tree_cache_create prototype change.
2574 2014-04-03 Richard Biener <rguenther@suse.de>
2576 * tree-streamer-out.c (streamer_write_chain): Do not temporarily
2577 set TREE_CHAIN to NULL_TREE.
2579 2014-04-03 Richard Biener <rguenther@suse.de>
2581 PR tree-optimization/60740
2582 * graphite-scop-detection.c (stmt_simple_for_scop_p): Iterate
2583 over all GIMPLE_COND operands.
2585 2014-04-03 Nathan Sidwell <nathan@codesourcery.com>
2587 * doc/invoke.texi (Wnon-virtual-dtor): Adjust documentation.
2588 (Weffc++): Remove Scott's numbering, merge lists and reference
2591 2014-04-03 Nick Clifton <nickc@redhat.com>
2593 * config/rl78/rl78-expand.md (movqi): Handle (SUBREG (SYMBOL_REF))
2596 2014-04-03 Martin Jambor <mjambor@suse.cz>
2598 * ipa-cp.c (ipcp_verify_propagated_values): Also dump symtab and
2599 mention gcc_unreachable before failing.
2600 * ipa.c (symtab_remove_unreachable_nodes): Also print order of
2603 2014-04-02 Jan Hubicka <hubicka@ucw.cz>
2606 * ipa-devirt.c (get_polymorphic_call_info): Do not ICE on type
2607 inconsistent code and instead mark the context inconsistent.
2608 (possible_polymorphic_call_targets): For inconsistent contexts
2609 return empty complete list.
2611 2014-04-02 Anthony Green <green@moxielogic.com>
2613 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2)
2614 (extendqisi2, extendhisi2): Define.
2615 * config/moxie/moxie.h (DEFAULT_SIGNED_CHAR): Change to 0.
2616 (WCHAR_TYPE): Change to unsigned int.
2618 2014-04-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2620 PR tree-optimization/60733
2621 * gimple-ssa-strength-reduction.c (ncd_with_phi): Change required
2622 insertion point for PHI candidates to be the end of the feeding
2623 block for the PHI argument.
2625 2014-04-02 Vladimir Makarov <vmakarov@redhat.com>
2627 PR rtl-optimization/60650
2628 * lra-constraints.c (process_alt_operands): Decrease reject for
2629 earlyclobber matching.
2631 2014-04-02 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2633 * config/s390/s390.c (s390_expand_insv): Use GET_MODE_BITSIZE.
2635 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
2637 * config/spu/spu.c (pad_bb): Do not crash when the last
2638 insn is CODE_FOR_blockage.
2640 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
2642 * config/spu/spu.md ("insv"): Fail if bitoffset+bitsize
2643 lies outside the target mode.
2645 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
2648 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
2649 software floating point or no floating point registers, do not
2650 allow any type in the FPRs. Eliminate a test for SPE SIMD types
2651 in GPRs that occurs after we tested for GPRs that would never be
2654 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
2655 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
2656 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
2657 specifically allow DDmode, since that does not use the SPE SIMD
2660 2014-04-02 Richard Biener <rguenther@suse.de>
2663 * optabs.c (expand_abs_nojump): Honor flag_trapv only for
2664 MODE_INTs. Properly use negv_optab.
2665 (expand_abs): Likewise.
2667 2014-04-02 Richard Biener <rguenther@suse.de>
2670 * Makefile.in (install-driver): Guard extra installs with special
2673 2014-04-01 Michael Meissner <meissner@linux.vnet.ibm.com>
2675 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
2678 2014-04-01 Richard Henderson <rth@redhat.com>
2681 * config/i386/i386.md (*float<SWI48><MODEF>2_sse): Leave the second
2682 alternative enabled before register allocation.
2684 2014-04-01 Chung-Lin Tang <cltang@codesourcery.com>
2686 * config/nios2/nios2.md (unspec): Remove UNSPEC_TLS, UNSPEC_TLS_LDM.
2687 * config/nios2/nios2.c (nios2_function_profiler): Fix addi operand
2689 (nios2_large_got_address): Remove unneeded 'sym' parameter.
2690 (nios2_got_address): Update nios2_large_got_address call site.
2691 (nios2_delegitimize_address): New function.
2692 (TARGET_DELEGITIMIZE_ADDRESS): Define to nios2_delegitimize_address.
2693 * config/nios2/linux.h (GLIBC_DYNAMIC_LINKER): Define.
2694 (LINK_SPEC): Specify dynamic linker using GNU_USER_DYNAMIC_LINKER.
2696 2014-04-01 Martin Husemann <martin@duskware.de>
2698 * config/mips/netbsd.h (TARGET_OS_CPP_BUILTINS): Define __mips_o32
2701 2014-04-01 Richard Sandiford <rdsandiford@googlemail.com>
2703 PR rtl-optimization/60604
2704 * recog.c (general_operand): Incorporate REG_CANNOT_CHANGE_MODE_P
2705 check from register_operand.
2706 (register_operand): Redefine in terms of general_operand.
2707 (nonmemory_operand): Use register_operand for the non-constant cases.
2709 2014-04-01 Richard Biener <rguenther@suse.de>
2711 * gimple.h (struct gimple_statement_base): Align subcode to 16 bits.
2713 2014-04-01 Sebastian Huber <sebastian.huber@embedded-brains.de>
2715 * doc/invoke.texi (mapp-regs): Clarify.
2717 2014-03-31 Ulrich Drepper <drepper@gmail.com>
2719 * config/i386/avx512fintrin.h (__v32hi): Define type.
2720 (__v64qi): Likewise.
2721 (_mm512_set1_epi8): Define.
2722 (_mm512_set1_epi16): Define.
2723 (_mm512_set4_epi32): Define.
2724 (_mm512_set4_epi64): Define.
2725 (_mm512_set4_pd): Define.
2726 (_mm512_set4_ps): Define.
2727 (_mm512_setr4_epi64): Define.
2728 (_mm512_setr4_epi32): Define.
2729 (_mm512_setr4_pd): Define.
2730 (_mm512_setr4_ps): Define.
2731 (_mm512_setzero_epi32): Define.
2733 2014-03-31 Martin Jambor <mjambor@suse.cz>
2736 * tree-sra.c (callsite_has_enough_arguments_p): Renamed to
2737 callsite_arguments_match_p. Updated all callers. Also check types of
2738 corresponding formal parameters and actual arguments.
2739 (not_all_callers_have_enough_arguments_p) Renamed to
2740 some_callers_have_mismatched_arguments_p.
2742 2014-03-31 Yuri Rumyantsev <ysrumyan@gmail.com>
2744 * tree-inline.c (copy_loops): Add missed copy of 'safelen'.
2746 2014-03-31 Kugan Vivekanandarajah <kuganv@linaro.org>
2749 * aarch64/aarch64.c (aarch64_classify_address): Fix alignment for
2752 2014-03-30 Uros Bizjak <ubizjak@gmail.com>
2754 * config/i386/sse.md (FMAMODE_NOVF512): New mode iterator.
2755 (<sd_mask_codefor>fma_fmadd_<mode><sd_maskz_name><round_name>):
2757 <sd_mask_codefor>fma_fmadd_<VF_512:mode><sd_maskz_name><round_name>.
2758 Use FMAMODE_NOVF512 mode iterator.
2759 (<sd_mask_codefor>fma_fmsub_<mode><sd_maskz_name><round_name>): Ditto.
2760 (<sd_mask_codefor>fma_fnmadd_<mode><sd_maskz_name><round_name>): Ditto.
2761 (<sd_mask_codefor>fma_fnmsub_<mode><sd_maskz_name><round_name>): Ditto.
2762 (<sd_mask_codefor>fma_fmaddsub_<mode><sd_maskz_name><round_name>):
2764 <sd_mask_codefor>fma_fmaddsub_<VF_512:mode><sd_maskz_name><round_name>.
2765 Use VF_128_256 mode iterator.
2766 (<sd_mask_codefor>fma_fmsubadd_<mode><sd_maskz_name><round_name>):
2769 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
2771 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Clear
2772 static chain if needed.
2774 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
2777 * lra-constraints.c (index_part_to_reg): New.
2778 (process_address): Use it.
2780 2014-03-27 Jeff Law <law@redhat.com>
2781 Jakub Jelinek <jakub@redhat.com>
2784 * expr.c (do_tablejump): Use simplify_gen_binary rather than
2785 gen_rtx_{PLUS,MULT} to build up the address expression.
2787 * i386/i386.c (ix86_legitimize_address): Use copy_addr_to_reg to avoid
2788 creating non-canonical RTL.
2790 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
2793 * ipa-inline.c (want_inline_small_function_p): Short circuit large
2794 functions; reorganize to make cheap checks first.
2795 (inline_small_functions): Do not estimate growth when dumping;
2797 * ipa-inline.h (inline_summary): Add min_size.
2798 (growth_likely_positive): New function.
2799 * ipa-inline-analysis.c (dump_inline_summary): Add min_size.
2800 (set_cond_stmt_execution_predicate): Cleanup.
2801 (estimate_edge_size_and_time): Compute min_size.
2802 (estimate_calls_size_and_time): Likewise.
2803 (estimate_node_size_and_time): Likewise.
2804 (inline_update_overall_summary): Update min_size.
2805 (do_estimate_edge_time): Likewise.
2806 (do_estimate_edge_size): Update.
2807 (do_estimate_edge_hints): Update.
2808 (growth_likely_positive): New function.
2810 2014-03-28 Jakub Jelinek <jakub@redhat.com>
2813 * config/i386/i386.c (ix86_copy_addr_to_reg): Call copy_addr_to_reg
2814 also if addr has VOIDmode.
2816 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2818 * config/arm/aarch-common.c (aarch_crypto_can_dual_issue): New.
2819 * config/arm/aarch-common-protos.h (aarch_crypto_can_dual_issue):
2821 * config/arm/cortex-a53.md: Add reservations and bypass for crypto
2822 instructions as well as AdvancedSIMD loads.
2824 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2826 * config/aarch64/aarch64-simd.md (aarch64_crypto_aes<aes_op>v16qi):
2827 Use crypto_aese type.
2828 (aarch64_crypto_aes<aesmc_op>v16qi): Use crypto_aesmc type.
2829 * config/arm/arm.md (is_neon_type): Replace crypto_aes with
2830 crypto_aese, crypto_aesmc. Move to types.md.
2831 * config/arm/types.md (crypto_aes): Split into crypto_aese,
2833 * config/arm/iterators.md (crypto_type): Likewise.
2835 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
2837 * cgraph.c: Include expr.h and tree-dfa.h.
2838 (cgraph_redirect_edge_call_stmt_to_callee): If call in noreturn;
2841 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
2844 * lra-assigns.c (find_hard_regno_for): Remove unavailable hard
2845 regs from checking multi-reg pseudos.
2847 2014-03-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2849 * config/arm/t-aprofile (MULTILIB_MATCHES): Correct A12 rule.
2851 2014-03-28 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
2853 * config/rs6000/rs6000.c (fusion_gpr_load_p): Refuse optimization
2854 if it would clobber the stack pointer, even temporarily.
2856 2014-03-28 Eric Botcazou <ebotcazou@adacore.com>
2858 * mode-switching.c: Make small adjustments to the top comment.
2860 2014-03-27 Michael Meissner <meissner@linux.vnet.ibm.com>
2862 * config/rs6000/constraints.md (wD constraint): New constraint to
2863 match the constant integer to get the top DImode/DFmode out of a
2864 vector in a VSX register.
2866 * config/rs6000/predicates.md (vsx_scalar_64bit): New predicate to
2867 match the constant integer to get the top DImode/DFmode out of a
2868 vector in a VSX register.
2870 * config/rs6000/rs6000-builtins.def (VBPERMQ): Add vbpermq builtin
2873 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
2876 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
2877 -mdebug=reg, print value of VECTOR_ELEMENT_SCALAR_64BIT.
2879 * config/rs6000/vsx.md (vsx_extract_<mode>, V2DI/V2DF modes):
2880 Optimize vec_extract of 64-bit values, where the value being
2881 extracted is in the top word, where we can use scalar
2882 instructions. Add direct move and store support. Combine the big
2883 endian/little endian vector select load support into a single insn.
2884 (vsx_extract_<mode>_internal1): Likewise.
2885 (vsx_extract_<mode>_internal2): Likewise.
2886 (vsx_extract_<mode>_load): Likewise.
2887 (vsx_extract_<mode>_store): Likewise.
2888 (vsx_extract_<mode>_zero): Delete, big and little endian insns are
2889 combined into vsx_extract_<mode>_load.
2890 (vsx_extract_<mode>_one_le): Likewise.
2892 * config/rs6000/rs6000.h (VECTOR_ELEMENT_SCALAR_64BIT): Macro to
2893 define the top 64-bit vector element.
2895 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wD
2898 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
2899 Document vec_vbpermq builtin.
2902 * config/rs6000/altivec.h (vec_xxsldwi): Add missing define to
2903 enable use of xxsldwi and xxpermdi builtin functions.
2904 (vec_xxpermdi): Likewise.
2906 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
2907 Document use of vec_xxsldwi and vec_xxpermdi builtins.
2909 2014-03-27 Vladimir Makarov <vmakarov@redhat.com>
2911 PR rtl-optimization/60650
2912 * lra-assign.c (find_hard_regno_for, spill_for): Add parameter
2914 (find_spills_for): New.
2915 (assign_by_spills): Pass the new parameter to find_hard_regno_for.
2916 Spill all pseudos on the second iteration.
2918 2014-03-27 Marek Polacek <polacek@redhat.com>
2921 * doc/extend.texi (ffs Builtins): Change unsigned types to signed
2924 2014-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2926 * config/s390/s390.c (s390_can_use_return_insn): Check for
2927 call-saved FPRs on 31 bit.
2929 2014-03-27 Jakub Jelinek <jakub@redhat.com>
2932 * omp-low.c (lower_omp_1): For gimple_clobber_p stmts,
2933 if they need regimplification, just drop them instead of
2934 calling gimple_regimplify_operands on them.
2936 2014-03-27 Marcus Shawcroft <marcus.shawcroft@arm.com>
2939 * config/aarch64/aarch64.c (faked_omit_frame_pointer): Remove.
2940 (aarch64_frame_pointer_required): Adjust logic.
2941 (aarch64_can_eliminate): Adjust logic.
2942 (aarch64_override_options_after_change): Adjust logic.
2944 2014-03-27 Dehao Chen <dehao@google.com>
2946 * ipa-inline.c (early_inliner): Update node's inline info.
2948 2014-03-26 Dehao Chen <dehao@google.com>
2950 * dojump.c (do_compare_rtx_and_jump): Sets correct probability for
2951 compiler inserted conditional jumps for NAN float check.
2953 2014-03-26 Jakub Jelinek <jakub@redhat.com>
2955 * ubsan.h (ubsan_create_data): Change second argument's type
2956 to const location_t *.
2957 * ubsan.c (ubsan_source_location): If xloc.file is NULL, set it to
2959 (ubsan_create_data): Change second argument to const location_t *PLOC.
2960 Create Loc field whenever PLOC is non-NULL.
2961 (ubsan_instrument_unreachable, ubsan_expand_null_ifn,
2962 ubsan_build_overflow_builtin, instrument_bool_enum_load): Adjust
2966 * real.c (real_to_integer2): Change type of low to UHWI.
2968 2014-03-26 Tobias Burnus <burnus@net-b.de>
2970 * gcc.c (LINK_COMMAND_SPEC): Use libcilkrts.spec for -fcilkplus.
2971 (CILK_SELF_SPECS): New define.
2972 (driver_self_specs): Use it.
2974 2014-03-26 Richard Biener <rguenther@suse.de>
2976 * tree-pretty-print.c (percent_K_format): Implement special
2977 case for LTO and its stripped down BLOCK tree.
2979 2014-03-26 Jakub Jelinek <jakub@redhat.com>
2982 * ubsan.c (instrument_si_overflow): Instrument ABS_EXPR.
2984 * tree-vrp.c (simplify_internal_call_using_ranges): If only
2985 one range is range_int_cst_p, but not both, at least optimize
2986 addition/subtraction of 0 and multiplication by 0 or 1.
2987 * gimple-fold.c (gimple_fold_call): Fold
2988 IFN_UBSAN_CHECK_{ADD,SUB,MUL}.
2989 (gimple_fold_stmt_to_constant_1): If both op0 and op1 aren't
2990 INTEGER_CSTs, try to fold at least x * 0 and y - y.
2992 2014-03-26 Eric Botcazou <ebotcazou@adacore.com>
2994 PR rtl-optimization/60452
2995 * rtlanal.c (rtx_addr_can_trap_p_1): Fix head comment.
2996 <case REG>: Return 1 for invalid offsets from the frame pointer.
2998 2014-03-26 Marek Polacek <polacek@redhat.com>
3001 * doc/extend.texi (C Extensions): Mention variable-length arrays in
3004 2014-03-26 Marek Polacek <polacek@redhat.com>
3007 * doc/extend.texi (Designated Inits): Describe what happens to omitted
3010 2014-03-26 Marek Polacek <polacek@redhat.com>
3013 * ira-color.c (update_conflict_hard_regno_costs): Perform the
3014 multiplication in unsigned type.
3016 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
3018 * doc/install.texi: Document nds32le-*-elf and nds32be-*-elf.
3020 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
3022 * doc/contrib.texi: Add myself as Andes nds32 port contributor.
3024 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
3027 * cif-code.def (UNREACHABLE) New code.
3028 * ipa-inline.c (inline_small_functions): Skip edges to
3029 __builtlin_unreachable.
3030 (estimate_edge_growth): Allow edges to __builtlin_unreachable.
3031 * ipa-inline-analysis.c (edge_set_predicate): Redirect edges with false
3032 predicate to __bulitin_unreachable.
3033 (set_cond_stmt_execution_predicate): Fix issue when
3034 invert_tree_comparison returns ERROR_MARK.
3035 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow): Do not
3036 propagate to inline clones.
3037 * cgraph.c (verify_edge_corresponds_to_fndecl): Allow redirection
3039 * ipa-cp.c (create_specialized_node): Be ready for new node to appear.
3040 * cgraphclones.c (cgraph_clone_node): If call destination is already
3041 ureachable, do not redirect it back.
3042 * tree-inline.c (fold_marked_statements): Hanlde calls becoming
3045 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
3047 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow):
3048 Do not modify inline clones.
3050 2014-03-25 Jakub Jelinek <jakub@redhat.com>
3052 * config/i386/i386.md (general_sext_operand): New mode attr.
3053 (addv<mode>4, subv<mode>4, mulv<mode>4): If operands[2] is CONST_INT,
3054 don't generate (sign_extend (const_int)).
3055 (*addv<mode>4, *subv<mode>4, *mulv<mode>4): Disallow CONST_INT_P
3056 operands[2]. Use We constraint instead of <i> and
3057 <general_sext_operand> predicate instead of <general_operand>.
3058 (*addv<mode>4_1, *subv<mode>4_1, *mulv<mode>4_1): New insns.
3059 * config/i386/constraints.md (We): New constraint.
3060 * config/i386/predicates.md (x86_64_sext_operand,
3061 sext_operand): New predicates.
3063 2014-03-25 Martin Jambor <mjambor@suse.cz>
3066 * ipa-cp.c (ipa_get_indirect_edge_target_1): Redirect type
3067 inconsistent devirtualizations to __builtin_unreachable.
3069 2014-03-25 Marek Polacek <polacek@redhat.com>
3072 * doc/extend.texi (Example of asm with clobbered asm reg): Fix typo.
3074 2014-03-25 Alan Lawrence <alan.lawrence@arm.com>
3076 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Reverse
3077 order of elements for big-endian.
3079 2014-03-25 Richard Biener <rguenther@suse.de>
3082 * gimplify-me.c (gimple_regimplify_operands): Update the
3085 2014-03-25 Martin Jambor <mjambor@suse.cz>
3088 * lto-cgraph.c (lto_output_node): Stream body_removed flag.
3089 (lto_output_varpool_node): Likewise.
3090 (input_overwrite_node): Likewise.
3091 (input_varpool_node): Likewise.
3093 2014-03-25 Richard Biener <rguenther@suse.de>
3095 * lto-wrapper.c (merge_and_complain): Handle OPT_fPIE like OPT_fpie.
3096 (run_gcc): Likewise.
3098 2014-03-25 Jakub Jelinek <jakub@redhat.com>
3100 * combine.c (simplify_compare_const): Add MODE argument.
3101 Handle mode_width 0 as very large mode_width.
3102 (try_combine, simplify_comparison): Adjust callers.
3104 * cselib.c (cselib_hash_rtx): Perform addition in unsigned
3105 type to avoid signed integer overflow.
3106 * explow.c (plus_constant): Likewise.
3108 2014-03-25 Dominik Vogt <vogt@linux.vnet.ibm.com>
3110 * doc/generic.texi: Correct typos.
3112 2014-03-24 Tobias Burnus <burnus@net-b.de>
3114 * doc/invoke.texi (-flto): Expand section about
3115 using static libraries with LTO.
3117 2014-03-24 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3119 PR rtl-optimization/60501
3120 * optabs.def (addptr3_optab): New optab.
3121 * optabs.c (gen_addptr3_insn, have_addptr3_insn): New function.
3122 * doc/md.texi ("addptrm3"): Document new RTL standard expander.
3123 * expr.h (gen_addptr3_insn, have_addptr3_insn): Add prototypes.
3125 * lra.c (emit_add3_insn): Use the addptr pattern if available.
3127 * config/s390/s390.md ("addptrdi3", "addptrsi3"): New expanders.
3129 2014-03-24 Ulrich Drepper <drepper@gmail.com>
3131 * config/i386/avx512fintrin.h: Define _mm512_set1_ps and
3134 * config/i386/avxintrin.h (_mm256_undefined_si256): Define.
3135 (_mm256_undefined_ps): Define.
3136 (_mm256_undefined_pd): Define.
3137 * config/i386/emmintrin.h (_mm_undefined_si128): Define.
3138 (_mm_undefined_pd): Define.
3139 * config/i386/xmmintrin.h (_mm_undefined_ps): Define.
3140 * config/i386/avx512fintrin.h (_mm512_undefined_si512): Define.
3141 (_mm512_undefined_ps): Define.
3142 (_mm512_undefined_pd): Define.
3143 Use _mm*_undefined_*.
3144 * config/i386/avx2intrin.h: Use _mm*_undefined_*.
3146 2014-03-24 Alex Velenko <Alex.Velenko@arm.com>
3148 * config/aarch64/aarch64-simd-builtins.def (lshr): DI mode excluded.
3149 (lshr_simd): DI mode added.
3150 * config/aarch64/aarch64-simd.md (aarch64_lshr_simddi): New pattern.
3151 (aarch64_ushr_simddi): Likewise.
3152 * config/aarch64/aarch64.md (UNSPEC_USHR64): New unspec.
3153 * config/aarch64/arm_neon.h (vshr_n_u64): Intrinsic fixed.
3154 (vshrd_n_u64): Likewise.
3156 2014-03-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3158 * Makefile.in (s-macro_list): Depend on cc1.
3160 2014-03-23 Teresa Johnson <tejohnson@google.com>
3162 * ipa-utils.c (ipa_print_order): Use specified dump file.
3164 2014-03-23 Eric Botcazou <ebotcazou@adacore.com>
3166 PR rtl-optimization/60601
3167 * bb-reorder.c (fix_up_fall_thru_edges): Test EDGE_FALLTHRU everywhere.
3169 * gcc.c (eval_spec_function): Initialize save_growing_value.
3171 2014-03-22 Jakub Jelinek <jakub@redhat.com>
3174 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): For
3175 code == MINUS_EXPR, never swap op0 with op1.
3177 * toplev.c (init_local_tick): Avoid signed integer multiplication
3179 * genautomata.c (reserv_sets_hash_value): Fix rotate idiom, avoid
3180 shift by first operand's bitsize.
3182 2014-03-21 Jakub Jelinek <jakub@redhat.com>
3185 * config/i386/i386.h (TARGET_64BIT_P): If not TARGET_BI_ARCH,
3187 * config/i386/darwin.h (TARGET_64BIT_P): Redefine to
3188 TARGET_ISA_64BIT_P(x).
3190 2014-03-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3192 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Generate a
3193 pattern for vector nor instead of subtract from splat(-1).
3194 (altivec_expand_vec_perm_const_le): Likewise.
3196 2014-03-21 Richard Henderson <rth@twiddle.net>
3199 * ifcvt.c (dead_or_predicable): Return FALSE if there are any frame
3200 related insns after epilogue_completed.
3202 2014-03-21 Martin Jambor <mjambor@suse.cz>
3205 * cgraph.h (symtab_node): New flag body_removed.
3206 * ipa.c (symtab_remove_unreachable_nodes): Set body_removed flag
3207 when removing bodies.
3208 * symtab.c (dump_symtab_base): Dump body_removed flag.
3209 * cgraph.c (verify_edge_corresponds_to_fndecl): Skip nodes which
3210 had their bodies removed.
3212 2014-03-21 Martin Jambor <mjambor@suse.cz>
3215 * ipa.c (symtab_remove_unreachable_nodes): Clear thunk flag of nodes
3218 2014-03-21 Richard Biener <rguenther@suse.de>
3220 PR tree-optimization/60577
3221 * tree-core.h (struct tree_base): Document nothrow_flag use
3222 in VAR_DECL_NONALIASED.
3223 * tree.h (VAR_DECL_NONALIASED): New.
3224 (may_be_aliased): Adjust.
3225 * coverage.c (build_var): Set VAR_DECL_NONALIASED.
3227 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
3229 * expr.c (expand_expr_real_1): Remove outdated comment.
3231 2014-03-20 Jakub Jelinek <jakub@redhat.com>
3234 * ira.c (adjust_cleared_regs): Call copy_rtx on
3235 *reg_equiv[REGNO (loc)].src_p before passing it to
3236 simplify_replace_fn_rtx.
3239 * config/i386/i386.c (x86_output_mi_thunk): Surround UNSPEC_GOT
3240 into CONST, put pic register as first operand of PLUS. Use
3241 gen_const_mem for both 32-bit and 64-bit PIC got loads.
3243 2014-03-20 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3245 * config/aarch64/aarch64.c (MEMORY_MOVE_COST): Delete.
3247 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
3249 * config/sparc/sparc.c (sparc_do_work_around_errata): Implement work
3250 around for store forwarding issue in the FPU on the UT699.
3251 * config/sparc/sparc.md (in_branch_delay): Return false for single FP
3252 loads and operations if -mfix-ut699 is specified.
3253 (divtf3_hq): Tweak attribute.
3254 (sqrttf2_hq): Likewise.
3256 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
3258 * calls.c (store_one_arg): Remove incorrect const qualification on the
3259 type of the temporary.
3260 * cfgexpand.c (expand_return): Likewise.
3261 * expr.c (expand_constructor): Likewise.
3262 (expand_expr_real_1): Likewise.
3264 2014-03-20 Zhenqiang Chen <zhenqiang.chen@linaro.org>
3266 * config/arm/arm.c (arm_dwarf_register_span): Update the element number
3269 2014-03-19 Kaz Kojima <kkojima@gcc.gnu.org>
3272 * config/sh/sh.md (udivsi3_i1): Clobber R1 register.
3274 2014-03-19 James Greenhalgh <james.greenhalgh@arm.com>
3276 * config/arm/aarch-common-protos.h
3277 (alu_cost_table): Fix spelling of "extend".
3278 * config/arm/arm.c (arm_new_rtx_costs): Fix spelling of "extend".
3280 2014-03-19 Richard Biener <rguenther@suse.de>
3283 * tree-core.h (tree_type_common): Re-order pointer members
3284 to reduce recursion depth during GC walks.
3286 2014-03-19 Marek Polacek <polacek@redhat.com>
3289 * ubsan.c (ubsan_type_descriptor): Check that DECL_NAME is nonnull
3290 before accessing it.
3292 2014-03-19 Richard Biener <rguenther@suse.de>
3295 * lto-streamer-in.c (input_function): In WPA stage do not drop
3298 2014-03-19 Jakub Jelinek <jakub@redhat.com>
3300 PR tree-optimization/60559
3301 * vectorizable_mask_load_store): Replace scalar MASK_LOAD
3302 with build_zero_cst assignment.
3304 2014-03-18 Kai Tietz <ktietz@redhat.com>
3306 PR rtl-optimization/56356
3307 * sdbout.c (sdbout_parms): Verify that parms'
3308 incoming argument is valid.
3309 (sdbout_reg_parms): Likewise.
3311 2014-03-18 Richard Henderson <rth@redhat.com>
3314 * config/i386/i386.md (*float<SWI48x><MODEF>2_i387): Move down to
3315 be shadowed by *float<SWI48><MODEF>2_sse. Test X87_ENABLE_FLOAT.
3316 (*float<SWI48><MODEF>2_sse): Check X87_ENABLE_FLOAT for alternative 0.
3318 2014-03-18 Basile Starynkevitch <basile@starynkevitch.net>
3320 * plugin.def: Improve comment for PLUGIN_INCLUDE_FILE.
3321 * doc/plugins.texi (Plugin callbacks): Mention PLUGIN_INCLUDE_FILE.
3322 Italicize plugin event names in description. Explain that
3323 PLUGIN_PRAGMAS has no sense for lto1. Explain PLUGIN_INCLUDE_FILE.
3324 Remind that no GCC functions should be called after PLUGIN_FINISH.
3325 Explain what pragmas with expansion are.
3327 2014-03-18 Martin Liska <mliska@suse.cz>
3329 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Added case when
3330 gimple call statement is update.
3331 * gimple-fold.c (gimple_fold_call): Changed order for GIMPLE_ASSIGN and
3332 GIMPLE_CALL, where gsi iterator still points to GIMPLE CALL.
3334 2014-03-18 Jakub Jelinek <jakub@redhat.com>
3337 * ubsan.c (ubsan_instrument_unreachable): Call
3338 initialize_sanitizer_builtins.
3339 (ubsan_pass): Likewise.
3342 * ubsan.c (ubsan_type_descriptor, ubsan_create_data): Call
3343 varpool_finalize_decl instead of rest_of_decl_compilation.
3345 2014-03-18 Richard Biener <rguenther@suse.de>
3347 * df-problems.c (df_rd_confluence_n): Avoid bitmap_copy
3348 by using bitmap_and_compl instead of bitmap_and_compl_into.
3349 (df_rd_transfer_function): Likewise.
3351 2014-03-18 Richard Biener <rguenther@suse.de>
3353 * doc/lto.texi (fresolution): Fix typo.
3355 2014-03-18 Richard Biener <rguenther@suse.de>
3357 * doc/invoke.texi (flto): Update for changes in 4.9.
3359 2014-03-18 Richard Biener <rguenther@suse.de>
3361 * doc/loop.texi: Remove section on the removed lambda framework.
3362 Update loop docs with recent changes in preserving loop structure.
3364 2014-03-18 Richard Biener <rguenther@suse.de>
3366 * doc/lto.texi (-fresolution): Document.
3368 2014-03-18 Richard Biener <rguenther@suse.de>
3370 * doc/contrib.texi: Adjust my name.
3372 2014-03-18 Jakub Jelinek <jakub@redhat.com>
3375 * internal-fn.c: Include diagnostic-core.h.
3376 (expand_BUILTIN_EXPECT): New function.
3377 * gimplify.c (gimplify_call_expr): Use false instead of FALSE.
3378 (gimplify_modify_expr): Gimplify 3 argument __builtin_expect into
3379 IFN_BUILTIN_EXPECT call instead of __builtin_expect builtin call.
3380 * ipa-inline-analysis.c (find_foldable_builtin_expect): Handle
3382 * predict.c (expr_expected_value_1): Handle IFN_BUILTIN_EXPECT.
3383 Revert 3 argument __builtin_expect code.
3384 (strip_predict_hints): Handle IFN_BUILTIN_EXPECT.
3385 * gimple-fold.c (gimple_fold_call): Likewise.
3386 * tree.h (fold_builtin_expect): New prototype.
3387 * builtins.c (build_builtin_expect_predicate): Add predictor
3388 argument, if non-NULL, create 3 argument __builtin_expect.
3389 (fold_builtin_expect): No longer static. Add ARG2 argument,
3390 pass it through to build_builtin_expect_predicate.
3391 (fold_builtin_2): Adjust caller.
3392 (fold_builtin_3): Handle BUILT_IN_EXPECT.
3393 * internal-fn.def (BUILTIN_EXPECT): New.
3395 2014-03-18 Tobias Burnus <burnus@net-b.de>
3398 * predict.def (PRED_FORTRAN_OVERFLOW, PRED_FORTRAN_FAIL_ALLOC,
3399 PRED_FORTRAN_FAIL_IO, PRED_FORTRAN_WARN_ONCE, PRED_FORTRAN_SIZE_ZERO,
3400 PRED_FORTRAN_INVALID_BOUND, PRED_FORTRAN_ABSENT_DUMMY): Add.
3402 2014-03-18 Jan Hubicka <hubicka@ucw.cz>
3405 * predict.c (combine_predictions_for_bb): Fix up formatting.
3406 (expr_expected_value_1, expr_expected_value): Add predictor argument,
3407 fill what it points to if non-NULL.
3408 (tree_predict_by_opcode): Adjust caller, use the predictor.
3409 * predict.def (PRED_COMPARE_AND_SWAP): Add.
3411 2014-03-18 Eric Botcazou <ebotcazou@adacore.com>
3413 * config/sparc/sparc.c (sparc_do_work_around_errata): Speed up and use
3414 proper constant for the store mode.
3416 2014-03-18 Ilya Enkovich <ilya.enkovich@intel.com>
3418 * symtab.c (change_decl_assembler_name): Fix transparent alias
3421 2014-03-16 Renlin Li <Renlin.Li@arm.com>
3423 * config/aarch64/aarch64.c: Correct the comments about the
3424 aarch64 stack layout.
3426 2014-03-18 Thomas Schwinge <thomas@codesourcery.com>
3428 * omp-low.c (lower_rec_input_clauses) <build_omp_barrier>: Restore
3429 check for GF_OMP_FOR_KIND_FOR.
3431 2013-03-18 Kirill Yukhin <kirill.yukhin@intel.com>
3433 * config/i386/i386.h (ADDITIONAL_REGISTER_NAMES): Add
3434 ymm and zmm register names.
3436 2014-03-17 Jakub Jelinek <jakub@redhat.com>
3439 * config/i386/i386.c (ix86_expand_epilogue): Adjust REG_CFA_ADJUST_CFA
3440 note creation for the 2010-08-31 changes.
3442 2014-03-17 Marek Polacek <polacek@redhat.com>
3445 * omp-low.c (omp_max_vf): Treat -fno-tree-loop-optimize the same
3446 as -fno-tree-loop-vectorize.
3447 (expand_omp_simd): Likewise.
3449 2014-03-15 Eric Botcazou <ebotcazou@adacore.com>
3451 * config/sparc/sparc-protos.h (tls_call_delay): Delete.
3452 (eligible_for_call_delay): New prototype.
3453 * config/sparc/sparc.c (tls_call_delay): Rename into...
3454 (eligible_for_call_delay): ...this. Return false if the instruction
3455 cannot be put in the delay slot of a branch.
3456 (eligible_for_restore_insn): Simplify.
3457 (eligible_for_return_delay): Return false if the instruction cannot be
3458 put in the delay slot of a branch and simplify.
3459 (eligible_for_sibcall_delay): Return false if the instruction cannot be
3460 put in the delay slot of a branch.
3461 * config/sparc/sparc.md (fix_ut699): New attribute.
3462 (tls_call_delay): Delete.
3463 (in_call_delay): Reimplement.
3464 (eligible_for_sibcall_delay): Rename into...
3465 (in_sibcall_delay): ...this.
3466 (eligible_for_return_delay): Rename into...
3467 (in_return_delay): ...this.
3468 (in_branch_delay): Reimplement.
3469 (in_uncond_branch_delay): Delete.
3470 (in_annul_branch_delay): Delete.
3472 2014-03-14 Richard Henderson <rth@redhat.com>
3475 * config/i386/i386.md (floathi<X87MODEF>2): Delete expander; rename
3476 define_insn from *floathi<X87MODEF>2_i387; allow nonimmediate_operand.
3477 (*floathi<X87MODEF>2_i387_with_temp): Remove.
3478 (floathi splitters): Remove.
3479 (float<SWI48x>xf2): New pattern.
3480 (float<SWI48><MODEF>2): Rename from float<SWI48x><X87MODEF>2. Drop
3481 code that tried to handle DImode for 32-bit, but which was excluded
3482 by the pattern's condition. Drop allocation of stack temporary.
3483 (*floatsi<MODEF>2_vector_mixed_with_temp): Remove.
3484 (*float<SWI48><MODEF>2_mixed_with_temp): Remove.
3485 (*float<SWI48><MODEF>2_mixed_interunit): Remove.
3486 (*float<SWI48><MODEF>2_mixed_nointerunit): Remove.
3487 (*floatsi<MODEF>2_vector_sse_with_temp): Remove.
3488 (*float<SWI48><MODEF>2_sse_with_temp): Remove.
3489 (*float<SWI48><MODEF>2_sse_interunit): Remove.
3490 (*float<SWI48><MODEF>2_sse_nointerunit): Remove.
3491 (*float<SWI48x><X87MODEF>2_i387_with_temp): Remove.
3492 (*float<SWI48x><X87MODEF>2_i387): Remove.
3493 (all float _with_temp splitters): Remove.
3494 (*float<SWI48x><MODEF>2_i387): New pattern.
3495 (*float<SWI48><MODEF>2_sse): New pattern.
3496 (float TARGET_USE_VECTOR_CONVERTS splitters): Merge them.
3497 (float TARGET_SSE_PARTIAL_REG_DEPENDENCY splitters): Merge them.
3499 2014-03-14 Jakub Jelinek <jakub@redhat.com>
3500 Marek Polacek <polacek@redhat.com>
3503 * common.opt (dump_base_name_prefixed): New Variable.
3504 * opts.c (finish_options): Don't prepend directory to x_dump_base_name
3505 if x_dump_base_name_prefixed is already set, set it at the end.
3507 2014-03-14 Vladimir Makarov <vmakarov@redhat.com>
3509 PR rtl-optimization/60508
3510 * lra-constraints.c (get_reload_reg): Add new parameter
3512 (process_addr_reg, simplify_operand_subreg, curr_insn_transform):
3513 Pass the new parameter values.
3515 2014-03-14 Richard Biener <rguenther@suse.de>
3517 * common.opt: Revert unintented changes from r205065.
3520 2014-03-14 Richard Biener <rguenther@suse.de>
3523 * cfghooks.c (split_block): Properly adjust all loops the
3524 block was a latch of.
3526 2014-03-14 Martin Jambor <mjambor@suse.cz>
3529 * ipa-prop.c (ipa_modify_call_arguments): Fix iteration condition
3532 2014-03-14 Georg-Johann Lay <avr@gjlay.de>
3535 * config/avr/avr.c (avr_set_current_function): Pass function name
3536 through default_strip_name_encoding before sanity checking instead
3537 of skipping the first char of the assembler name.
3539 2014-03-13 Richard Henderson <rth@redhat.com>
3542 * config/i386/i386.c (ix86_split_fp_branch): Remove pushed argument.
3543 (ix86_force_to_memory, ix86_free_from_memory): Remove.
3544 * config/i386/i386-protos.h: Likewise.
3545 * config/i386/i386.md (floathi<X87MODEF>2): Use assign_386_stack_local
3546 in the expander instead of a splitter.
3547 (float<SWI48x><X87MODEF>2): Use assign_386_stack_local if there is
3548 any possibility of requiring a memory.
3549 (*floatsi<MODEF>2_vector_mixed): Remove, and the splitters.
3550 (*floatsi<MODEF>2_vector_sse): Remove, and the splitters.
3551 (fp branch splitters): Update for ix86_split_fp_branch.
3552 (*jcc<X87MODEF>_<SWI24>_i387): Remove r/f alternative.
3553 (*jcc<X87MODEF>_<SWI24>_r_i387): Likewise.
3554 (splitter for jcc<X87MODEF>_<SWI24>_i387 r/f): Remove.
3555 (*fop_<MODEF>_2_i387): Remove f/r alternative.
3556 (*fop_<MODEF>_3_i387): Likewise.
3557 (*fop_xf_2_i387, *fop_xf_3_i387): Likewise.
3558 (splitters for the fop_* register patterns): Remove.
3559 (fscalexf4_i387): Rename from *fscalexf4_i387.
3560 (ldexpxf3): Use gen_floatsixf2 and gen_fscalexf4_i387.
3562 2014-03-13 Jakub Jelinek <jakub@redhat.com>
3564 PR tree-optimization/59779
3565 * tree-dfa.c (get_ref_base_and_extent): Use double_int
3566 type for bitsize and maxsize instead of HOST_WIDE_INT.
3568 2014-03-13 Steven Bosscher <steven@gcc.gnu.org>
3570 PR rtl-optimization/57320
3571 * function.c (rest_of_handle_thread_prologue_and_epilogue): Cleanup
3572 the CFG after thread_prologue_and_epilogue_insns.
3574 2014-03-13 Vladimir Makarov <vmakarov@redhat.com>
3576 PR rtl-optimization/57189
3577 * lra-constraints.c (process_alt_operands): Disfavor spilling
3580 2014-03-13 Cesar Philippidis <cesar@codesourcery.com>
3582 * lto-wrapper.c (maybe_unlink_file): Suppress diagnostic messages.
3584 2014-03-13 Jakub Jelinek <jakub@redhat.com>
3586 PR tree-optimization/59025
3588 * tree-ssa-reassoc.c (sort_by_operand_rank): For SSA_NAMEs with the
3589 same rank, sort by bb_rank and gimple_uid of SSA_NAME_DEF_STMT first.
3591 2014-03-13 Georg-Johann Lay <avr@gjlay.de>
3594 * config/avr/avr.c (avr_out_plus): Swap cc_plus and cc_minus in
3595 calls of avr_out_plus_1.
3597 2014-03-13 Bin Cheng <bin.cheng@arm.com>
3599 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Record
3600 BB's single pred and update the father loop's latch info later.
3602 2014-03-12 Michael Meissner <meissner@linux.vnet.ibm.com>
3604 * config/rs6000/vector.md (VEC_L): Add V1TI mode to vector types.
3608 (VEC_base): Likewise.
3609 (mov<MODE>, VEC_M modes): If we are loading TImode into VSX
3610 registers, we need to swap double words in little endian mode.
3612 * config/rs6000/rs6000-modes.def (V1TImode): Add new vector mode
3613 to be a container mode for 128-bit integer operations added in ISA
3614 2.07. Unlike TImode and PTImode, the preferred register set is
3615 the Altivec/VMX registers for the 128-bit operations.
3617 * config/rs6000/rs6000-protos.h (rs6000_move_128bit_ok_p): Add
3619 (rs6000_split_128bit_ok_p): Likewise.
3621 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_3): Add new support
3622 macros for creating ISA 2.07 normal and overloaded builtin
3623 functions with 3 arguments.
3624 (BU_P8V_OVERLOAD_3): Likewise.
3625 (VPERM_1T): Add support for V1TImode in 128-bit vector operations
3626 for use as overloaded functions.
3627 (VPERM_1TI_UNS): Likewise.
3628 (VSEL_1TI): Likewise.
3629 (VSEL_1TI_UNS): Likewise.
3630 (ST_INTERNAL_1ti): Likewise.
3631 (LD_INTERNAL_1ti): Likewise.
3632 (XXSEL_1TI): Likewise.
3633 (XXSEL_1TI_UNS): Likewise.
3634 (VPERM_1TI): Likewise.
3635 (VPERM_1TI_UNS): Likewise.
3636 (XXPERMDI_1TI): Likewise.
3637 (SET_1TI): Likewise.
3638 (LXVD2X_V1TI): Likewise.
3639 (STXVD2X_V1TI): Likewise.
3640 (VEC_INIT_V1TI): Likewise.
3641 (VEC_SET_V1TI): Likewise.
3642 (VEC_EXT_V1TI): Likewise.
3643 (EQV_V1TI): Likewise.
3644 (NAND_V1TI): Likewise.
3645 (ORC_V1TI): Likewise.
3646 (VADDCUQ): Add support for 128-bit integer arithmetic instructions
3647 added in ISA 2.07. Add both normal 'altivec' builtins, and the
3649 (VADDUQM): Likewise.
3650 (VSUBCUQ): Likewise.
3651 (VADDEUQM): Likewise.
3652 (VADDECUQ): Likewise.
3653 (VSUBEUQM): Likewise.
3654 (VSUBECUQ): Likewise.
3656 * config/rs6000/rs6000-c.c (__int128_type): New static to hold
3657 __int128_t and __uint128_t types.
3658 (__uint128_type): Likewise.
3659 (altivec_categorize_keyword): Add support for vector __int128_t,
3660 vector __uint128_t, vector __int128, and vector unsigned __int128
3661 as a container type for TImode operations that need to be done in
3662 VSX/Altivec registers.
3663 (rs6000_macro_to_expand): Likewise.
3664 (altivec_overloaded_builtins): Add ISA 2.07 overloaded functions
3665 to support 128-bit integer instructions vaddcuq, vadduqm,
3666 vaddecuq, vaddeuqm, vsubcuq, vsubuqm, vsubecuq, vsubeuqm.
3667 (altivec_resolve_overloaded_builtin): Add support for V1TImode.
3669 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Add support
3670 for V1TImode, and set up preferences to use VSX/Altivec registers.
3671 Setup VSX reload handlers.
3672 (rs6000_debug_reg_global): Likewise.
3673 (rs6000_init_hard_regno_mode_ok): Likewise.
3674 (rs6000_preferred_simd_mode): Likewise.
3675 (vspltis_constant): Do not allow V1TImode as easy altivec constants.
3676 (easy_altivec_constant): Likewise.
3677 (output_vec_const_move): Likewise.
3678 (rs6000_expand_vector_set): Convert V1TImode set and extract to
3680 (rs6000_expand_vector_extract): Likewise.
3681 (reg_offset_addressing_ok_p): Setup V1TImode to use VSX reg+reg
3683 (rs6000_const_vec): Add support for V1TImode.
3684 (rs6000_emit_le_vsx_load): Swap double words when loading or
3685 storing TImode/V1TImode.
3686 (rs6000_emit_le_vsx_store): Likewise.
3687 (rs6000_emit_le_vsx_move): Likewise.
3688 (rs6000_emit_move): Add support for V1TImode.
3689 (altivec_expand_ld_builtin): Likewise.
3690 (altivec_expand_st_builtin): Likewise.
3691 (altivec_expand_vec_init_builtin): Likewise.
3692 (altivec_expand_builtin): Likewise.
3693 (rs6000_init_builtins): Add support for V1TImode type. Add
3694 support for ISA 2.07 128-bit integer builtins. Define type names
3695 for the VSX/Altivec vector types.
3696 (altivec_init_builtins): Add support for overloaded vector
3697 functions with V1TImode type.
3698 (rs6000_preferred_reload_class): Prefer Altivec registers for V1TImode.
3699 (rs6000_move_128bit_ok_p): Move 128-bit move/split validation to
3701 (rs6000_split_128bit_ok_p): Likewise.
3702 (rs6000_handle_altivec_attribute): Create V1TImode from vector
3703 __int128_t and vector __uint128_t.
3705 * config/rs6000/vsx.md (VSX_L): Add V1TImode to vector iterators
3706 and mode attributes.
3713 (VS_scalar): Likewise.
3714 (VS_double): Likewise.
3715 (vsx_set_v1ti): New builtin function to create V1TImode from TImode.
3717 * config/rs6000/rs6000.h (TARGET_VADDUQM): New macro to say whether
3718 we support the ISA 2.07 128-bit integer arithmetic instructions.
3719 (ALTIVEC_OR_VSX_VECTOR_MODE): Add V1TImode.
3720 (enum rs6000_builtin_type_index): Add fields to hold V1TImode
3721 and TImode types for use with the builtin functions.
3722 (V1TI_type_node): Likewise.
3723 (unsigned_V1TI_type_node): Likewise.
3724 (intTI_type_internal_node): Likewise.
3725 (uintTI_type_internal_node): Likewise.
3727 * config/rs6000/altivec.md (UNSPEC_VADDCUQ): New unspecs for ISA 2.07
3728 128-bit builtin functions.
3729 (UNSPEC_VADDEUQM): Likewise.
3730 (UNSPEC_VADDECUQ): Likewise.
3731 (UNSPEC_VSUBCUQ): Likewise.
3732 (UNSPEC_VSUBEUQM): Likewise.
3733 (UNSPEC_VSUBECUQ): Likewise.
3734 (VM): Add V1TImode to vector mode iterators.
3736 (VI_unit): Likewise.
3737 (altivec_vadduqm): Add ISA 2.07 128-bit binary builtins.
3738 (altivec_vaddcuq): Likewise.
3739 (altivec_vsubuqm): Likewise.
3740 (altivec_vsubcuq): Likewise.
3741 (altivec_vaddeuqm): Likewise.
3742 (altivec_vaddecuq): Likewise.
3743 (altivec_vsubeuqm): Likewise.
3744 (altivec_vsubecuq): Likewise.
3746 * config/rs6000/rs6000.md (FMOVE128_GPR): Add V1TImode to vector
3748 (BOOL_128): Likewise.
3749 (BOOL_REGS_OUTPUT): Likewise.
3750 (BOOL_REGS_OP1): Likewise.
3751 (BOOL_REGS_OP2): Likewise.
3752 (BOOL_REGS_UNARY): Likewise.
3753 (BOOL_REGS_AND_CR0): Likewise.
3755 * config/rs6000/altivec.h (vec_vaddcuq): Add support for ISA 2.07
3756 128-bit integer builtin support.
3757 (vec_vadduqm): Likewise.
3758 (vec_vaddecuq): Likewise.
3759 (vec_vaddeuqm): Likewise.
3760 (vec_vsubecuq): Likewise.
3761 (vec_vsubeuqm): Likewise.
3762 (vec_vsubcuq): Likewise.
3763 (vec_vsubuqm): Likewise.
3765 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
3766 Document vec_vaddcuq, vec_vadduqm, vec_vaddecuq, vec_vaddeuqm,
3767 vec_subecuq, vec_subeuqm, vec_vsubcuq, vec_vsubeqm builtins adding
3768 128-bit integer add/subtract to ISA 2.07.
3770 2014-03-12 Joern Rennecke <joern.rennecke@embecosm.com>
3772 * config/arc/arc.c (arc_predicate_delay_insns):
3773 Fix third argument passed to conditionalize_nonjump.
3775 2014-03-12 Yufeng Zhang <yufeng.zhang@arm.com>
3777 * config/aarch64/aarch64-builtins.c
3778 (aarch64_builtin_vectorized_function): Add BUILT_IN_LFLOORF,
3779 BUILT_IN_LLFLOOR, BUILT_IN_LCEILF and BUILT_IN_LLCEIL.
3780 * config/aarch64/arm_neon.h (vcvtaq_u64_f64): Call __builtin_llfloor
3781 instead of __builtin_lfloor.
3782 (vcvtnq_u64_f64): Call __builtin_llceil instead of __builtin_lceil.
3784 2014-03-12 Jakub Jelinek <jakub@redhat.com>
3786 * tree-ssa-ifcombine.c (forwarder_block_to): New function.
3787 (tree_ssa_ifcombine_bb_1): New function.
3788 (tree_ssa_ifcombine_bb): Use it. Handle also cases where else_bb
3789 is an empty forwarder block to then_bb or vice versa and then_bb
3790 and else_bb are effectively swapped.
3792 2014-03-12 Christian Bruel <christian.bruel@st.com>
3795 * config/arm/arm.c (arm_emit_vfp_multi_reg_pop): Emit a
3796 REG_CFA_DEF_CFA note.
3797 (arm_expand_epilogue_apcs_frame): call arm_add_cfa_adjust_cfa_note.
3798 (arm_unwind_emit): Allow REG_CFA_DEF_CFA.
3800 2014-03-12 Thomas Preud'homme <thomas.preudhomme@arm.com>
3802 PR tree-optimization/60454
3803 * tree-ssa-math-opts.c (find_bswap_1): Fix bswap detection.
3805 2014-03-12 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3807 * config.gcc (aarch64*-*-*): Use ISA flags from aarch64-arches.def.
3808 Do not define target_cpu_default2 to generic.
3809 * config/aarch64/aarch64.h (TARGET_CPU_DEFAULT): Use generic cpu.
3810 * config/aarch64/aarch64.c (aarch64_override_options): Update comment.
3811 * config/aarch64/aarch64-arches.def (armv8-a): Use generic cpu.
3813 2014-03-12 Jakub Jelinek <jakub@redhat.com>
3814 Marc Glisse <marc.glisse@inria.fr>
3816 PR tree-optimization/60502
3817 * tree-ssa-reassoc.c (eliminate_not_pairs): Use build_all_ones_cst
3818 instead of build_low_bits_mask.
3820 2014-03-12 Jakub Jelinek <jakub@redhat.com>
3823 * tree-vrp.c (register_edge_assert_for_1): Don't add assert
3824 if there are multiple uses, but op doesn't live on E edge.
3825 * tree-cfg.c (assert_unreachable_fallthru_edge_p): Also ignore
3826 clobber stmts before __builtin_unreachable.
3828 2014-03-11 Richard Sandiford <rdsandiford@googlemail.com>
3830 * builtins.c (expand_builtin_setjmp_receiver): Use and clobber
3831 hard_frame_pointer_rtx.
3832 * cse.c (cse_insn): Remove volatile check.
3833 * cselib.c (cselib_process_insn): Likewise.
3834 * dse.c (scan_insn): Likewise.
3836 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
3838 * config/arc/arc.c (conditionalize_nonjump): New function,
3840 (arc_ifcvt): ... this.
3841 (arc_predicate_delay_insns): Use it.
3843 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
3845 * config/arc/predicates.md (extend_operand): During/after reload,
3846 allow const_int_operand.
3847 * config/arc/arc.md (mulsidi3_700): Use extend_operand predicate.
3848 (umulsidi3_700): Likewise. Change operand 2 constraint back to "cL".
3849 (mulsi3_highpart): Change operand 2 constraint alternatives 2 and 3
3851 (umulsi3_highpart_i): Likewise.
3853 2014-03-11 Richard Biener <rguenther@suse.de>
3855 * tree-ssa-structalias.c (get_constraint_for_ptr_offset):
3856 Add asserts to guard possible wrong-code bugs.
3858 2014-03-11 Richard Biener <rguenther@suse.de>
3860 PR tree-optimization/60429
3861 PR tree-optimization/60485
3862 * tree-ssa-structalias.c (set_union_with_increment): Properly
3863 take into account all fields that overlap the shifted vars.
3864 (do_sd_constraint): Likewise.
3865 (do_ds_constraint): Likewise.
3866 (get_constraint_for_ptr_offset): Likewise.
3868 2014-03-11 Chung-Lin Tang <cltang@codesourcery.com>
3870 * config/nios2/nios2.c (machine_function): Add fp_save_offset field.
3871 (nios2_compute_frame_layout):
3872 Add calculation of cfun->machine->fp_save_offset.
3873 (nios2_expand_prologue): Correct setting of frame pointer register
3875 (nios2_expand_epilogue): Update recovery of stack pointer from
3876 frame pointer accordingly.
3877 (nios2_initial_elimination_offset): Update calculation of offset
3878 for eliminating to HARD_FRAME_POINTER_REGNUM.
3880 2014-03-10 Jakub Jelinek <jakub@redhat.com>
3883 * ipa.c (symtab_remove_unreachable_nodes): Don't call
3884 cgraph_get_create_node on VAR_DECLs.
3886 2014-03-10 Richard Biener <rguenther@suse.de>
3889 * tree.c (signed_or_unsigned_type_for): Handle OFFSET_TYPEs.
3891 2014-03-08 Douglas B Rupp <rupp@gnat.com>
3893 * config/vms/vms.opt (vms_float_format): New variable.
3895 2014-03-08 Tobias Burnus <burnus@net-b.de>
3897 * doc/invoke.texi (-fcilkplus): Update implementation status.
3899 2014-03-08 Paulo Matos <paulo@matos-sorge.com>
3900 Richard Biener <rguenther@suse.de>
3902 * lto-wrapper.c (merge_and_complain): Ensure -fshort-double is used
3903 consistently accross all TUs.
3904 (run_gcc): Enable -fshort-double automatically at link at link-time
3905 and disallow override.
3907 2014-03-08 Richard Sandiford <rdsandiford@googlemail.com>
3910 * config/mips/mips.c (mips_option_override): Promote -mpaired-single
3911 warning to an error. Disable TARGET_PAIRED_SINGLE and TARGET_MIPS3D
3912 if they can't be used.
3914 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3916 * configure.ac (HAVE_AS_IX86_TLSLDMPLT): Improve test
3917 for Solaris 11/x86 ld.
3918 * configure: Regenerate.
3920 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3922 * configure.ac (TLS_SECTION_ASM_FLAG): Save as tls_section_flag.
3923 (LIB_TLS_SPEC): Save as ld_tls_libs.
3924 (HAVE_AS_IX86_TLSLDMPLT): Define as 1/0.
3925 (HAVE_AS_IX86_TLSLDM): New test.
3926 * configure, config.in: Regenerate.
3927 * config/i386/i386.c (legitimize_tls_address): Fall back to
3928 TLS_MODEL_GLOBAL_DYNAMIC on 32-bit Solaris/x86 if tool chain
3929 cannot support TLS_MODEL_LOCAL_DYNAMIC.
3930 * config/i386/i386.md (*tls_local_dynamic_base_32_gnu): Use if
3931 instead of #ifdef in HAVE_AS_IX86_TLSLDMPLT test.
3933 2014-03-07 Paulo Matos <paulo@matos-sorge.com>
3935 * common.opt (fira-loop-pressure): Mark as optimization.
3937 2014-03-07 Thomas Schwinge <thomas@codesourcery.com>
3939 * langhooks.c (lhd_omp_mappable_type): The error_mark_node is not
3940 an OpenMP mappable type.
3942 2014-03-06 Matthias Klose <doko@ubuntu.com>
3944 * Makefile.in (s-mlib): Only pass MULTIARCH_DIRNAME if
3945 MULTILIB_OSDIRNAMES is not defined.
3947 2014-03-06 Jakub Jelinek <jakub@redhat.com>
3948 Meador Inge <meadori@codesourcery.com>
3951 * config/arm/arm.c (arm_tls_symbol_p): Remove.
3952 (arm_legitimize_address): Call legitimize_tls_address for any
3953 arm_tls_referenced_p expression, handle constant addend. Call it
3954 before testing for !TARGET_ARM.
3955 (thumb_legitimize_address): Don't handle arm_tls_symbol_p here.
3957 2014-03-06 Richard Biener <rguenther@suse.de>
3963 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
3965 * tree-streamer.c (record_common_node): Assert we don't record
3966 nodes with type double.
3967 (preload_common_node): Skip type double, complex double and double
3968 pointer since it is now frontend dependent due to fshort-double option.
3970 2014-03-06 Richard Biener <rguenther@suse.de>
3972 * gcc.c (PLUGIN_COND): Always enable unless -fno-use-linker-plugin
3973 or -fno-lto is specified and the linker has full plugin support.
3974 * collect2.c (lto_mode): Default to LTO_MODE_WHOPR if LTO is enabled.
3975 (main): Remove -flto processing, adjust lto_mode using use_plugin late.
3976 * lto-wrapper.c (merge_and_complain): Merge compile-time
3977 optimization levels.
3978 (run_gcc): And pass it through to the link options.
3980 2014-03-06 Alexandre Oliva <aoliva@redhat.com>
3984 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
3986 * cselib.c (remove_useless_values): Skip to avoid quadratic
3987 behavior if the condition moved from...
3988 (cselib_process_insn): ... here holds.
3990 2014-03-05 Jakub Jelinek <jakub@redhat.com>
3993 * Makefile.in (PLUGIN_HEADERS): Add tree-phinodes.h, stor-layout.h,
3994 ssa-iterators.h, $(RESOURCE_H) and tree-cfgcleanup.h.
3997 * config/i386/t-i386 (OPTIONS_H_EXTRA): Add stringop.def.
3998 (TM_H): Add x86-tune.def.
4000 2014-03-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4002 * config/aarch64/aarch64.c (generic_tunings):
4003 Use cortexa57_extra_costs.
4005 2014-03-05 Jakub Jelinek <jakub@redhat.com>
4008 * cfgexpand.c (expand_used_vars): Do not assume all SSA_NAMEs
4009 of PARM/RESULT_DECLs must be coalesced with optimize && in_lto_p.
4010 * tree-ssa-coalesce.c (coalesce_ssa_name): Use MUST_COALESCE_COST - 1
4013 2014-03-04 Heiher <r@hev.cc>
4015 * config/mips/mips-cpus.def (loongson3a): Mark as a MIPS64r2 processor.
4016 * config/mips/mips.h (MIPS_ISA_LEVEL_SPEC): Adjust accordingly.
4018 2014-03-04 Uros Bizjak <ubizjak@gmail.com>
4020 * config/i386/predicates.md (const2356_operand): Change to ...
4021 (const2367_operand): ... this.
4022 * config/i386/sse.md (avx512pf_scatterpf<mode>sf): Use
4024 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
4025 (*avx512pf_scatterpf<mode>sf): Ditto.
4026 (avx512pf_scatterpf<mode>df): Ditto.
4027 (*avx512pf_scatterpf<mode>df_mask): Ditto.
4028 (*avx512pf_scatterpf<mode>df): Ditto.
4029 * config/i386/i386.c (ix86_expand_builtin): Update
4030 incorrect hint operand error message.
4032 2014-03-04 Richard Biener <rguenther@suse.de>
4034 * lto-section-in.c (lto_get_section_data): Fix const cast.
4036 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
4038 * tree-streamer.c (record_common_node): Assert we don't record
4039 nodes with type double.
4040 (preload_common_node): Skip type double, complex double and double
4041 pointer since it is now frontend dependent due to fshort-double option.
4043 2014-03-04 Richard Biener <rguenther@suse.de>
4046 * lto-streamer-in.c (lto_read_body): Remove LTO bytecode version check.
4047 (lto_input_toplevel_asms): Likewise.
4048 * lto-section-in.c (lto_get_section_data): Instead do it here
4051 2014-03-04 Richard Biener <rguenther@suse.de>
4053 PR tree-optimization/60382
4054 * tree-vect-loop.c (vect_is_simple_reduction_1): Do not consider
4055 dead PHIs a reduction.
4057 2014-03-03 Uros Bizjak <ubizjak@gmail.com>
4059 * config/i386/xmmintrin.h (enum _mm_hint) <_MM_HINT_ET0>: Correct
4061 (_mm_prefetch): Move out of GCC target("sse") pragma.
4062 * config/i386/prfchwintrin.h (_m_prefetchw): Move out of
4063 GCC target("prfchw") pragma.
4064 * config/i386/i386.md (prefetch): Emit prefetchwt1 only
4066 * config/i386/i386.c (ix86_option_override_internal): Enable
4067 -mprfchw with -mprefetchwt1.
4069 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
4071 * config/arc/arc.md (casesi_load) <length attribute alternative 0>:
4074 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
4076 * opts.h (CL_PCH_IGNORE): Define.
4077 * targhooks.c (option_affects_pch_p):
4078 Return false for options that have CL_PCH_IGNORE set.
4079 * opt-functions.awk: Process PchIgnore.
4080 * doc/options.texi: Document PchIgnore.
4082 * config/arc/arc.opt (misize): Add PchIgnore property.
4084 2014-03-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4086 * config/rs6000/rs6000.c (rs6000_preferred_reload_class): Disallow
4087 reload of PLUS rtx's outside of GENERAL_REGS or BASE_REGS; relax
4088 constraint on constants to permit them being loaded into
4089 GENERAL_REGS or BASE_REGS.
4091 2014-03-03 Nick Clifton <nickc@redhat.com>
4093 * config/rl78/rl78-real.md (cbranchsi4_real_signed): Add
4094 anti-cacnonical alternatives.
4095 (negandhi3_real): New pattern.
4096 * config/rl78/rl78-virt.md (negandhi3_virt): New pattern.
4098 2014-03-03 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
4100 * config/avr/avr-mcus.def: Remove atxmega16x1.
4101 * config/avr/avr-tables.opt: Regenerate.
4102 * config/avr/t-multilib: Regenerate.
4103 * doc/avr-mmcu.texi: Regenerate.
4105 2014-03-03 Tobias Grosser <tobias@grosser.es>
4106 Mircea Namolaru <mircea.namolaru@inria.fr>
4108 PR tree-optimization/58028
4109 * graphite-clast-to-gimple.c (set_cloog_options): Don't remove
4112 2014-03-03 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4114 * config/arm/neon.md (*movmisalign<mode>): Legitimize addresses
4115 not handled by recognizers.
4117 2014-03-03 Jakub Jelinek <jakub@redhat.com>
4120 * function.c (expand_function_end): Don't emit
4121 clobber_return_register sequence if clobber_after is a BARRIER.
4122 * cfgexpand.c (construct_exit_block): Append instructions before
4123 return_label to prev_bb.
4125 2014-03-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4127 * config/rs6000/constraints.md: Document reserved use of "wc".
4129 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
4132 * ipa.c (function_and_variable_visibility): When dissolving comdat
4133 group, also set all symbols to local.
4135 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
4140 2013-12-14 Jan Hubicka <jh@suse.cz>
4142 * ipa-prop.c (stmt_may_be_vtbl_ptr_store): Skip clobbers.
4144 2014-03-02 Jon Beniston <jon@beniston.com>
4150 * config/lm32/lm32.c (lm32_legitimate_constant_p): Remove, as incorrect.
4151 (TARGET_LEGITIMATE_CONSTANT_P): Undefine, as not needed.
4152 * config/lm32/lm32.md (movsi_insn): Add 32-bit immediate support.
4153 (simple_return, *simple_return): New patterns
4154 * config/lm32/predicates.md (movsi_rhs_operand): Remove as obsolete.
4155 * configure.ac (force_sjlj_exceptions): Force sjlj exceptions for lm32.
4157 2014-03-01 Paolo Carlini <paolo.carlini@oracle.com>
4159 * dwarf2out.c (gen_subprogram_die): Tidy.
4161 2014-03-01 Oleg Endo <olegendo@gcc.gnu.org>
4164 * config/sh/sh.md (*mov_t_msb_neg): Split into ...
4165 (*mov_t_msb_neg_negc): ... this new insn.
4167 2014-02-28 Jason Merrill <jason@redhat.com>
4170 * ipa-devirt.c (ipa_devirt): Don't choose an implicitly-declared
4173 2014-02-28 Paolo Carlini <paolo.carlini@oracle.com>
4176 * dwarf2out.c (decltype_auto_die): New static.
4177 (gen_subprogram_die): Handle 'decltype(auto)' like 'auto'.
4178 (gen_type_die_with_usage): Handle 'decltype(auto)'.
4179 (is_cxx_auto): Likewise.
4181 2014-02-28 Ian Bolton <ian.bolton@arm.com>
4183 * config/aarch64/aarch64.h: Define __ARM_NEON by default if
4184 we are not using general regs only.
4186 2014-02-28 Richard Biener <rguenther@suse.de>
4189 * tree-cfgcleanup.c (tree_forwarder_block_p): Restrict
4190 previous fix and only allow to remove trivial pre-headers
4191 and latches. Also honor LOOPS_MAY_HAVE_MULTIPLE_LATCHES.
4192 (remove_forwarder_block): Properly update the latch of a loop.
4194 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4197 * cselib.c (cselib_hasher::equal): Special-case VALUE lookup.
4198 (cselib_preserved_hash_table): New.
4199 (preserve_constants_and_equivs): Move preserved vals to it.
4200 (cselib_find_slot): Look it up first.
4201 (cselib_init): Initialize it.
4202 (cselib_finish): Release it.
4203 (dump_cselib_table): Dump it.
4205 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4208 * cselib.c (remove_useless_values): Skip to avoid quadratic
4209 behavior if the condition moved from...
4210 (cselib_process_insn): ... here holds.
4212 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4215 * var-tracking.c (vt_initialize): Apply the same condition to
4216 preserve the CFA base value.
4218 2014-02-28 Joey Ye <joey.ye@arm.com>
4221 * config/arm/arm.c (thumb_far_jump_used_p): Don't change
4222 if reload in progress or completed.
4224 2014-02-28 Tobias Burnus <burnus@net-b.de>
4227 * tree-pretty-print.c (dump_generic_node, print_declaration): Handle
4230 2014-02-27 H.J. Lu <hongjiu.lu@intel.com>
4232 * doc/tm.texi.in (Condition Code Status): Update documention for
4233 relative locations of cc0-setter and cc0-user.
4235 2014-02-27 Jeff Law <law@redhat.com>
4237 PR rtl-optimization/52714
4238 * combine.c (try_combine): When splitting an unrecognized PARALLEL
4239 into two independent simple sets, if I3 is a jump, ensure the
4240 pattern we place into I3 is a (set (pc) ...).
4242 2014-02-27 Mikael Pettersson <mikpe@it.uu.se>
4243 Jeff Law <law@redhat.com>
4245 PR rtl-optimization/49847
4246 * cse.c (fold_rtx) Handle case where cc0 setter and cc0 user
4247 are in different blocks.
4248 * doc/tm.texi (Condition Code Status): Update documention for
4249 relative locations of cc0-setter and cc0-user.
4251 2014-02-27 Vladimir Makarov <vmakarov@redhat.com>
4254 * lra.c (lra_emit_add): Check SUBREG too.
4256 2014-02-27 Andreas Schwab <schwab@suse.de>
4258 * config/m68k/m68k.c (m68k_option_override): Disable
4259 -flive-range-shrinkage for classic m68k.
4260 (m68k_override_options_after_change): Likewise.
4262 2014-02-27 Marek Polacek <polacek@redhat.com>
4265 * tree-ssa-uninit.c (gate_warn_uninitialized): Run the pass even for
4266 -Wmaybe-uninitialized.
4268 2014-02-27 Alan Modra <amodra@gmail.com>
4271 * reload1.c (emit_input_reload_insns): When reload_override_in,
4272 set old to rl->in_reg when rl->in_reg is a subreg.
4274 2014-02-26 Richard Biener <rguenther@suse.de>
4277 * lra-assigns.c (spill_for): Avoid mixed-sign comparison.
4279 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
4281 * common/config/i386/predicates.md (const1256_operand): Remove.
4282 (const2356_operand): New.
4283 (const_1_to_2_operand): Remove.
4284 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
4285 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
4286 (*avx512pf_gatherpf<mode>sf): Ditto.
4287 (avx512pf_gatherpf<mode>df): Ditto.
4288 (*avx512pf_gatherpf<mode>df_mask): Ditto.
4289 (*avx512pf_gatherpf<mode>df): Ditto.
4290 (avx512pf_scatterpf<mode>sf): Ditto.
4291 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
4292 (*avx512pf_scatterpf<mode>sf): Ditto.
4293 (avx512pf_scatterpf<mode>df): Ditto.
4294 (*avx512pf_scatterpf<mode>df_mask): Ditto.
4295 (*avx512pf_scatterpf<mode>df): Ditto.
4296 * common/config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET0.
4298 2014-02-26 Ilya Tocar <ilya.tocar@intel.com>
4300 * config/i386/avx512fintrin.h (_mm512_testn_epi32_mask),
4301 (_mm512_mask_testn_epi32_mask), (_mm512_testn_epi64_mask),
4302 (_mm512_mask_testn_epi64_mask): Move to ...
4303 * config/i386/avx512cdintrin.h: Here.
4304 * config/i386/i386.c (bdesc_args): Change MASK_ISA for testnm.
4305 * config/i386/sse.md (avx512f_vmscalef<mode><round_name>): Remove %.
4306 (avx512f_scalef<mode><mask_name><round_name>): Ditto.
4307 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Change conditon to
4308 TARGET_AVX512F from TARGET_AVX512CD.
4310 2014-02-26 Richard Biener <rguenther@suse.de>
4313 * ipa.c (walk_polymorphic_call_targets): Properly guard
4314 call to inline_update_overall_summary.
4316 2014-02-26 Bin Cheng <bin.cheng@arm.com>
4319 * tree-cfgcleanup.c (tree_forwarder_block_p): Protect loop preheaders
4320 and latches only if requested. Fix latch if it is removed.
4321 * tree-ssa-dom.c (tree_ssa_dominator_optimize): Set
4322 LOOPS_HAVE_PREHEADERS.
4324 2014-02-25 Andrew Pinski <apinski@cavium.com>
4326 * builtins.c (expand_builtin_thread_pointer): Create a new target
4327 when the target is NULL.
4329 2014-02-25 Vladimir Makarov <vmakarov@redhat.com>
4331 PR rtl-optimization/60317
4332 * params.def (PARAM_LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
4333 * params.h (LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
4334 * lra-assigns.c: Include params.h.
4335 (spill_for): Use LRA_MAX_CONSIDERED_RELOAD_PSEUDOS as guard for
4336 other reload pseudos considerations.
4338 2014-02-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4340 * config/rs6000/vector.md (*vector_unordered<mode>): Change split
4341 to use canonical form for nor<mode>3.
4343 2014-02-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4346 * config/arm/arm.h (CANNOT_CHANGE_MODE_CLASS): Allow 128 to 64-bit
4349 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
4351 * common/config/i386/i386-common.c (OPTION_MASK_ISA_PREFETCHWT1_SET),
4352 (OPTION_MASK_ISA_PREFETCHWT1_UNSET): New.
4353 (ix86_handle_option): Handle OPT_mprefetchwt1.
4354 * config/i386/cpuid.h (bit_PREFETCHWT1): New.
4355 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
4357 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
4358 OPTION_MASK_ISA_PREFETCHWT1.
4359 * config/i386/i386.c (ix86_target_string): Handle mprefetchwt1.
4360 (PTA_PREFETCHWT1): New.
4361 (ix86_option_override_internal): Handle PTA_PREFETCHWT1.
4362 (ix86_valid_target_attribute_inner_p): Handle OPT_mprefetchwt1.
4363 * config/i386/i386.h (TARGET_PREFETCHWT1, TARGET_PREFETCHWT1_P): New.
4364 * config/i386/i386.md (prefetch): Check TARGET_PREFETCHWT1
4365 (*prefetch_avx512pf_<mode>_: Change into ...
4366 (*prefetch_prefetchwt1_<mode>: This.
4367 * config/i386/i386.opt (mprefetchwt1): New.
4368 * config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET1.
4369 (_mm_prefetch): Handle intent to write.
4370 * doc/invoke.texi (mprefetchwt1), (mno-prefetchwt1): Doccument.
4372 2014-02-25 Richard Biener <rguenther@suse.de>
4375 * emit-rtl.c (mem_attrs_htab): Remove.
4376 (mem_attrs_htab_hash): Likewise.
4377 (mem_attrs_htab_eq): Likewise.
4378 (set_mem_attrs): Always allocate new mem-attrs when something changed.
4379 (init_emit_once): Do not allocate mem_attrs_htab.
4381 2014-02-25 Richard Biener <rguenther@suse.de>
4384 * lto-opts.c (lto_write_options): Output non-explicit conservative
4385 -fwrapv, -fno-trapv and -fno-strict-overflow.
4386 * lto-wrapper.c (merge_and_complain): Handle merging those options.
4387 (run_gcc): And pass them through.
4389 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
4391 * sel-sched.c (calculate_new_fences): New parameter ptime.
4392 Calculate it as a maximum over all fence cycles.
4393 (sel_sched_region_2): Adjust the call to calculate_new_fences.
4394 Print the final schedule timing when sched_verbose.
4396 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
4398 PR rtl-optimization/60292
4399 * sel-sched.c (fill_vec_av_set): Do not reset target availability
4400 bit fot the fence instruction.
4402 2014-02-24 Alangi Derick <alangiderick@gmail.com>
4404 * calls.h: Fix typo in comment.
4406 2014-02-24 John David Anglin <danglin@gcc.gnu.org>
4408 * config/pa/pa.c (pa_output_move_double): Don't valididate when
4409 adjusting offsetable addresses.
4411 2014-02-24 Guozhi Wei <carrot@google.com>
4413 * sparseset.h (sparseset_pop): Fix the wrong index.
4415 2014-02-24 Walter Lee <walt@tilera.com>
4417 * config.gcc (tilepro-*-*): Change to tilepro*-*-*.
4418 (tilegx-*-linux*): Change to tilegx*-*-linux*; Support tilegxbe
4420 * common/config/tilegx/tilegx-common.c
4421 (TARGET_DEFAULT_TARGET_FLAGS): Define.
4422 * config/tilegx/linux.h (ASM_SPEC): Add endian_spec.
4424 * config/tilegx/sync.md (atomic_test_and_set): Handle big endian.
4425 * config/tilegx/tilegx.c (tilegx_return_in_msb): New.
4426 (tilegx_gimplify_va_arg_expr): Handle big endian.
4427 (tilegx_expand_unaligned_load): Ditto.
4428 (tilegx_expand_unaligned_store): Ditto.
4429 (TARGET_RETURN_IN_MSB): New.
4430 * config/tilegx/tilegx.h (TARGET_DEFAULT): New.
4431 (TARGET_ENDIAN_DEFAULT): New.
4432 (TARGET_BIG_ENDIAN): Handle big endian.
4433 (BYTES_BIG_ENDIAN): Ditto.
4434 (WORDS_BIG_ENDIAN): Ditto.
4435 (FLOAT_WORDS_BIG_ENDIAN): Ditto.
4438 * config/tilegx/tilegx.md (extv): Handle big endian.
4440 (insn_st<n>): Ditto.
4441 (insn_st<n>_add<bitsuffix>): Ditto.
4442 (insn_stnt<n>): Ditto.
4443 (insn_stnt<n>_add<bitsuffix>):Ditto.
4444 (vec_interleave_highv8qi): Handle big endian.
4445 (vec_interleave_highv8qi_be): New.
4446 (vec_interleave_highv8qi_le): New.
4447 (insn_v1int_h): Handle big endian.
4448 (vec_interleave_lowv8qi): Handle big endian.
4449 (vec_interleave_lowv8qi_be): New.
4450 (vec_interleave_lowv8qi_le): New.
4451 (insn_v1int_l): Handle big endian.
4452 (vec_interleave_highv4hi): Handle big endian.
4453 (vec_interleave_highv4hi_be): New.
4454 (vec_interleave_highv4hi_le): New.
4455 (insn_v2int_h): Handle big endian.
4456 (vec_interleave_lowv4hi): Handle big endian.
4457 (vec_interleave_lowv4hi_be): New.
4458 (vec_interleave_lowv4hi_le): New.
4459 (insn_v2int_l): Handle big endian.
4460 (vec_interleave_highv2si): Handle big endian.
4461 (vec_interleave_highv2si_be): New.
4462 (vec_interleave_highv2si_le): New.
4463 (insn_v4int_h): Handle big endian.
4464 (vec_interleave_lowv2si): Handle big endian.
4465 (vec_interleave_lowv2si_be): New.
4466 (vec_interleave_lowv2si_le): New.
4467 (insn_v4int_l): Handle big endian.
4468 * config/tilegx/tilegx.opt (mbig-endian): New option.
4469 (mlittle-endian): New option.
4470 * doc/install.texi: Document tilegxbe-linux.
4471 * doc/invoke.texi: Document -mbig-endian and -mlittle-endian.
4473 2014-02-24 Martin Jambor <mjambor@suse.cz>
4476 * ipa-cp.c (propagate_constants_accross_call): Bail out early if
4477 there are no parameter descriptors.
4479 2014-02-24 Andrey Belevantsev <abel@ispras.ru>
4481 PR rtl-optimization/60268
4482 * sched-rgn.c (haifa_find_rgns): Move the nr_regions_initial variable
4483 initialization to ...
4484 (sched_rgn_init): ... here.
4485 (schedule_region): Check for SCHED_PRESSURE_NONE earlier.
4487 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
4489 * config/microblaze/microblaze.md: Correct ashrsi_reg / lshrsi_reg
4492 2014-02-23 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
4494 * config/microblaze/microblaze.h: Remove SECONDARY_MEMORY_NEEDED
4497 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
4499 * /config/microblaze/microblaze.c: Add microblaze_asm_output_mi_thunk,
4500 define TARGET_ASM_OUTPUT_MI_THUNK and TARGET_ASM_CAN_OUTPUT_MI_THUNK.
4502 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
4504 * config/microblaze/predicates.md: Add cmp_op predicate.
4505 * config/microblaze/microblaze.md: Add branch_compare instruction
4506 which uses cmp_op predicate and emits cmp insn before branch.
4507 * config/microblaze/microblaze.c (microblaze_emit_compare): Rename
4508 to microblaze_expand_conditional_branch and consolidate logic.
4509 (microblaze_expand_conditional_branch): emit branch_compare
4510 insn instead of handling cmp op separate from branch insn.
4512 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4514 * config/rs6000/rs6000.c (rs6000_emit_le_vsx_move): Relax assert
4517 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4519 * config/rs6000/altivec.md (altivec_lve<VI_char>x): Replace
4520 define_insn with define_expand and new define_insn
4521 *altivec_lve<VI_char>x_internal.
4522 (altivec_stve<VI_char>x): Replace define_insn with define_expand
4523 and new define_insn *altivec_stve<VI_char>x_internal.
4524 * config/rs6000/rs6000-protos.h (altivec_expand_stvex_be): New
4526 * config/rs6000/rs6000.c (altivec_expand_lvx_be): Document use by
4528 (altivec_expand_stvex_be): New function.
4530 2014-02-22 Joern Rennecke <joern.rennecke@embecosm.com>
4532 * config/avr/avr.c (avr_can_eliminate): Allow elimination from
4533 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM if !frame_pointer_needed.
4534 * config/avr/avr.c (ELIMINABLE_REGS): Add elimination from
4535 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM.
4537 2014-02-21 Vladimir Makarov <vmakarov@redhat.com>
4540 * lra-constraints.c (inherit_reload_reg): Use lra_emit_move
4541 instead of emit_move_insn.
4543 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4545 * config/rs6000/altivec.md (altivec_vsumsws): Replace second
4547 (reduc_uplus_v16qi): Use gen_altivec_vsumsws_direct instead of
4548 gen_altivec_vsumsws.
4550 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4552 * config/rs6000/altivec.md (altivec_lvxl): Rename as
4553 *altivec_lvxl_<mode>_internal and use VM2 iterator instead of V4SI.
4554 (altivec_lvxl_<mode>): New define_expand incorporating
4555 -maltivec=be semantics where needed.
4556 (altivec_lvx): Rename as *altivec_lvx_<mode>_internal.
4557 (altivec_lvx_<mode>): New define_expand incorporating -maltivec=be
4558 semantics where needed.
4559 (altivec_stvx): Rename as *altivec_stvx_<mode>_internal.
4560 (altivec_stvx_<mode>): New define_expand incorporating
4561 -maltivec=be semantics where needed.
4562 (altivec_stvxl): Rename as *altivec_stvxl_<mode>_internal and use
4563 VM2 iterator instead of V4SI.
4564 (altivec_stvxl_<mode>): New define_expand incorporating
4565 -maltivec=be semantics where needed.
4566 * config/rs6000/rs6000-builtin.def: Add new built-in definitions
4567 LVXL_V2DF, LVXL_V2DI, LVXL_V4SF, LVXL_V4SI, LVXL_V8HI, LVXL_V16QI,
4568 LVX_V2DF, LVX_V2DI, LVX_V4SF, LVX_V4SI, LVX_V8HI, LVX_V16QI, STVX_V2DF,
4569 STVX_V2DI, STVX_V4SF, STVX_V4SI, STVX_V8HI, STVX_V16QI, STVXL_V2DF,
4570 STVXL_V2DI, STVXL_V4SF, STVXL_V4SI, STVXL_V8HI, STVXL_V16QI.
4571 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Replace
4572 ALTIVEC_BUILTIN_LVX with ALTIVEC_BUILTIN_LVX_<MODE> throughout;
4573 similarly for ALTIVEC_BUILTIN_LVXL, ALTIVEC_BUILTIN_STVX, and
4574 ALTIVEC_BUILTIN_STVXL.
4575 * config/rs6000/rs6000-protos.h (altivec_expand_lvx_be): New prototype.
4576 (altivec_expand_stvx_be): Likewise.
4577 * config/rs6000/rs6000.c (swap_selector_for_mode): New function.
4578 (altivec_expand_lvx_be): Likewise.
4579 (altivec_expand_stvx_be): Likewise.
4580 (altivec_expand_builtin): Add cases for
4581 ALTIVEC_BUILTIN_STVX_<MODE>, ALTIVEC_BUILTIN_STVXL_<MODE>,
4582 ALTIVEC_BUILTIN_LVXL_<MODE>, and ALTIVEC_BUILTIN_LVX_<MODE>.
4583 (altivec_init_builtins): Add definitions for
4584 __builtin_altivec_lvxl_<mode>, __builtin_altivec_lvx_<mode>,
4585 __builtin_altivec_stvx_<mode>, and __builtin_altivec_stvxl_<mode>.
4587 2014-02-21 Catherine Moore <clm@codesourcery.com>
4589 * doc/invoke.texi (mvirt, mno-virt): Document.
4590 * config/mips/mips.opt (mvirt): New option.
4591 * config/mips/mips.h (ASM_SPEC): Pass mvirt to the assembler.
4593 2014-02-21 Richard Biener <rguenther@suse.de>
4595 PR tree-optimization/60276
4596 * tree-vectorizer.h (struct _stmt_vec_info): Add min_neg_dist field.
4597 (STMT_VINFO_MIN_NEG_DIST): New macro.
4598 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Record
4599 STMT_VINFO_MIN_NEG_DIST.
4600 * tree-vect-stmts.c (vectorizable_load): Verify if assumptions
4601 made for negative dependence distances still hold.
4603 2014-02-21 Richard Biener <rguenther@suse.de>
4606 * tree-ssa-live.c (mark_all_vars_used_1): Do not walk
4607 DECL_INITIAL for globals not in the current function context.
4609 2014-02-21 Jakub Jelinek <jakub@redhat.com>
4611 PR tree-optimization/56490
4612 * params.def (PARAM_UNINIT_CONTROL_DEP_ATTEMPTS): New param.
4613 * tree-ssa-uninit.c: Include params.h.
4614 (compute_control_dep_chain): Add num_calls argument, return false
4615 if it exceed PARAM_UNINIT_CONTROL_DEP_ATTEMPTS param, pass
4616 num_calls to recursive call.
4617 (find_predicates): Change dep_chain into normal array,
4618 cur_chain into auto_vec<edge, MAX_CHAIN_LEN + 1>, add num_calls
4619 variable and adjust compute_control_dep_chain caller.
4620 (find_def_preds): Likewise.
4622 2014-02-21 Thomas Schwinge <thomas@codesourcery.com>
4624 * gimple-pretty-print.c (dump_gimple_omp_for) [flags & TDF_RAW]
4625 <case GF_OMP_FOR_KIND_CILKSIMD>: Add missing break statement.
4627 2014-02-21 Nick Clifton <nickc@redhat.com>
4629 * config/stormy16/stormy16.md (pushdqi1): Add mode to post_inc.
4630 (pushhi1): Likewise.
4631 (popqi1): Add mode to pre_dec.
4634 2014-02-21 Jakub Jelinek <jakub@redhat.com>
4636 * config/i386/i386.c (ix86_expand_vec_perm): Use V8SImode
4637 mode for mask of V8SFmode permutation.
4639 2014-02-20 Richard Henderson <rth@redhat.com>
4642 * builtins.c (expand_builtin_atomic_compare_exchange): Always make
4643 a new pseudo for OLDVAL.
4645 2014-02-20 Jakub Jelinek <jakub@redhat.com>
4648 * config/i386/i386.c (expand_vec_perm_interleave2): Don't call
4649 gen_reg_rtx if d->testing_p.
4650 (expand_vec_perm_pshufb2, expand_vec_perm_broadcast_1): Return early
4651 if d->testing_p and we will certainly return true.
4652 (expand_vec_perm_even_odd_1): Likewise. Don't call gen_reg_rtx
4655 2014-02-20 Uros Bizjak <ubizjak@gmail.com>
4657 * emit-rtl.c (gen_reg_rtx): Assert that
4658 crtl->emit.regno_pointer_align_length is non-zero.
4660 2014-02-20 Richard Henderson <rth@redhat.com>
4663 * builtins.c (expand_builtin_atomic_compare_exchange): Conditionalize
4664 on failure the store back into EXPECT.
4666 2014-02-20 Chung-Lin Tang <cltang@codesourcery.com>
4667 Sandra Loosemore <sandra@codesourcery.com>
4669 * config/nios2/nios2.md (unspec): Add UNSPEC_PIC_GOTOFF_SYM enum.
4670 * config/nios2/nios2.c (nios2_function_profiler): Add
4671 -fPIC (flag_pic == 2) support.
4672 (nios2_handle_custom_fpu_cfg): Fix warning parameter.
4673 (nios2_large_offset_p): New function.
4674 (nios2_unspec_reloc_p): Move up position, update to use
4675 nios2_large_offset_p.
4676 (nios2_unspec_address): Remove function.
4677 (nios2_unspec_offset): New function.
4678 (nios2_large_got_address): New function.
4679 (nios2_got_address): Add large offset support.
4680 (nios2_legitimize_tls_address): Update usage of removed and new
4682 (nios2_symbol_binds_local_p): New function.
4683 (nios2_load_pic_address): Add -fPIC (flag_pic == 2) support.
4684 (nios2_legitimize_address): Update to use nios2_large_offset_p.
4685 (nios2_emit_move_sequence): Avoid legitimizing (const (unspec ...)).
4686 (nios2_print_operand): Merge H/L processing, add hiadj/lo
4687 processing for (const (unspec ...)).
4688 (nios2_unspec_reloc_name): Add UNSPEC_PIC_GOTOFF_SYM case.
4690 2014-02-20 Richard Biener <rguenther@suse.de>
4692 * tree-cfg.c (replace_uses_by): Mark altered BBs before
4693 doing the substitution.
4694 (verify_gimple_assign_single): Also verify bare MEM_REFs on the lhs.
4696 2014-02-20 Martin Jambor <mjambor@suse.cz>
4699 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Uce correct
4700 info when checking whether lattices are bottom.
4702 2014-02-20 Richard Biener <rguenther@suse.de>
4705 * tree-eh.c (execute_cleanup_eh_1): Also cleanup empty EH
4708 2014-02-20 Jan Hubicka <hubicka@ucw.cz>
4711 * ipa-inline-transform.c (clone_inlined_nodes): Add freq_scale
4712 parameter specifying the scaling.
4713 (inline_call): Update.
4714 (want_inline_recursively): Guard division by zero.
4715 (recursive_inlining): Update.
4716 * ipa-inline.h (clone_inlined_nodes): Update.
4718 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
4721 * config/i386/i386.c (classify_argument): Pass structures of size
4722 64 bytes or less in register.
4724 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
4725 Kirill Yukhin <kirill.yukhin@intel.com>
4727 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): Swap operands.
4728 (_mm_rcp28_round_ss): Ditto.
4729 (_mm_rsqrt28_round_sd): Ditto.
4730 (_mm_rsqrt28_round_ss): Ditto.
4731 * config/i386/avx512erintrin.h (_mm_rcp14_round_sd): Ditto.
4732 (_mm_rcp14_round_ss): Ditto.
4733 (_mm_rsqrt14_round_sd): Ditto.
4734 (_mm_rsqrt14_round_ss): Ditto.
4735 * config/i386/sse.md (rsqrt14<mode>): Put nonimmediate operand as
4736 the first input operand, get rid of match_dup.
4737 (avx512er_exp2<mode><mask_name><round_saeonly_name>): Set type
4739 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
4741 (avx512er_vmrcp28<mode><round_saeonly_name>): Put nonimmediate
4742 operand as the first input operand, set type attribute.
4743 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
4745 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Put nonimmediate
4746 operand as the first input operand, set type attribute.
4748 2014-02-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4750 * config/rs6000/rs6000.c (vspltis_constant): Fix most significant
4753 2014-02-19 H.J. Lu <hongjiu.lu@intel.com>
4756 * config/i386/i386.c (construct_container): Remove TFmode check
4757 for X86_64_INTEGER_CLASS.
4759 2014-02-19 Uros Bizjak <ubizjak@gmail.com>
4762 * config/i386/i386.c (type_natural_mode): Warn for ABI changes
4763 only when -Wpsabi is enabled.
4765 2014-02-19 Michael Hudson-Doyle <michael.hudson@linaro.org>
4768 * config/aarch64/aarch64.c (aarch64_pass_by_reference): The rules for
4769 passing arrays in registers are the same as for structs, so remove the
4770 special case for them.
4772 2014-02-19 Eric Botcazou <ebotcazou@adacore.com>
4774 * expr.c (expand_expr_real_1) <case VIEW_CONVERT_EXPR>: For a bit-field
4775 destination type, extract only the valid bits if the source type is not
4776 integral and has a different mode.
4778 2014-02-19 Richard Biener <rguenther@suse.de>
4781 * tree-inline.c (estimate_num_insns): Avoid calling cgraph_get_node
4784 2014-02-19 Richard Biener <rguenther@suse.de>
4787 * ipa-prop.c: Include stringpool.h and tree-ssanames.h.
4788 (ipa_modify_call_arguments): Emit an argument load explicitely and
4789 preserve virtual SSA form there and for the replacement call.
4790 Do not update SSA form nor free dominance info.
4792 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
4794 * ipa.c (function_and_variable_visibility): Also clear WEAK
4795 flag when disolving COMDAT_GROUP.
4797 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
4799 * ipa-prop.h (ipa_ancestor_jf_data): Update ocmment.
4800 * ipa-prop.c (ipa_set_jf_known_type): Return early when
4802 (ipa_set_ancestor_jf): Set type to NULL hwen it is not preserved;
4803 do more sanity checks.
4804 (detect_type_change): Return true when giving up early.
4805 (compute_complex_assign_jump_func): Fix type parameter of
4806 ipa_set_ancestor_jf.
4807 (compute_complex_ancestor_jump_func): Likewise.
4808 (update_jump_functions_after_inlining): Fix updating of
4810 * ipa-cp.c (ipa_get_jf_ancestor_result): Be ready for type to be NULL.
4812 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
4814 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Also remove
4815 inline clones when edge disappears.
4817 2014-02-18 Michael Meissner <meissner@linux.vnet.ibm.com>
4820 * config/rs6000/rs6000.md (mov<mode>_64bit, TF/TDmode moves):
4821 Split 64-bit moves into 2 patterns. Do not allow the use of
4822 direct move for TDmode in little endian, since the decimal value
4823 has little endian bytes within a word, but the 64-bit pieces are
4824 ordered in a big endian fashion, and normal subreg's of TDmode are
4826 (mov<mode>_64bit_dm): Likewise.
4827 (movtd_64bit_nodm): Likewise.
4829 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
4831 PR tree-optimization/60174
4832 * tree-ssa-reassoc.c (init_range_entry): Do not look into the defining
4833 statement of an SSA_NAME that occurs in an abnormal PHI node.
4835 2014-02-18 Jakub Jelinek <jakub@redhat.com>
4838 * final.c (SEEN_BB): Remove.
4839 (SEEN_NOTE, SEEN_EMITTED): Renumber.
4840 (final_scan_insn): Don't force_source_line on second
4841 NOTE_INSN_BASIC_BLOCK.
4843 2014-02-18 Uros Bizjak <ubizjak@gmail.com>
4846 * config/i386/i386.h (struct ix86_args): Add warn_avx512f.
4847 * config/i386/i386.c (init_cumulative_args): Initialize warn_avx512f.
4848 (type_natural_mode): Warn ABI change when %zmm register is not
4849 available for AVX512F vector value passing.
4851 2014-02-18 Kai Tietz <ktietz@redhat.com>
4854 * config/i386/i386.c (ix86_expand_prologue): Use value in
4855 rax register as displacement when restoring %r10 or %rax.
4856 Fix wrong offset when restoring both registers.
4858 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
4860 * ipa-prop.c (compute_complex_ancestor_jump_func): Replace overzealous
4861 assertion with conditional return.
4863 2014-02-18 Jakub Jelinek <jakub@redhat.com>
4864 Uros Bizjak <ubizjak@gmail.com>
4867 * config/i386/driver-i386.c (host_detect_local_cpu): If
4868 YMM state is not saved by the OS, also clear has_f16c. Move
4869 CPUID 0x80000001 handling before YMM state saving checking.
4871 2014-02-18 Andrey Belevantsev <abel@ispras.ru>
4873 PR rtl-optimization/58960
4874 * haifa-sched.c (alloc_global_sched_pressure_data): New,
4875 factored out from ...
4876 (sched_init): ... here.
4877 (free_global_sched_pressure_data): New, factored out from ...
4878 (sched_finish): ... here.
4879 * sched-int.h (free_global_sched_pressure_data): Declare.
4880 * sched-rgn.c (nr_regions_initial): New static global.
4881 (haifa_find_rgns): Initialize it.
4882 (schedule_region): Disable sched-pressure for the newly
4885 2014-02-17 Richard Biener <rguenther@suse.de>
4887 * tree-vect-stmts.c (free_stmt_vec_info): Clear BB and
4888 release SSA defs of pattern stmts.
4890 2014-02-17 Richard Biener <rguenther@suse.de>
4892 * tree-inline.c (expand_call_inline): Release the virtual
4893 operand defined by the call we are about to inline.
4895 2014-02-17 Richard Biener <rguenther@suse.de>
4897 * tree-ssa.c (verify_ssa): If verify_def found an error, ICE.
4899 2014-02-17 Kirill Yukhin <kirill.yukhin@intel.com>
4900 Ilya Tocar <ilya.tocar@intel.com>
4902 * config/i386/avx512fintrin.h (_mm512_maskz_permutexvar_epi64): Swap
4903 arguments order in builtin.
4904 (_mm512_permutexvar_epi64): Ditto.
4905 (_mm512_mask_permutexvar_epi64): Ditto
4906 (_mm512_maskz_permutexvar_epi32): Ditto
4907 (_mm512_permutexvar_epi32): Ditto
4908 (_mm512_mask_permutexvar_epi32): Ditto
4910 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4912 * config/rs6000/altivec.md (p8_vmrgew): Handle little endian targets.
4913 (p8_vmrgow): Likewise.
4915 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4917 * config/rs6000/vsx.md (vsx_xxpermdi_<mode>): Handle little
4920 2014-02-15 Michael Meissner <meissner@linux.vnet.ibm.com>
4923 * config/rs6000/rs6000.md (rreg): Add TFmode, TDmode constraints.
4924 (mov<mode>_internal, TFmode/TDmode): Split TFmode/TDmode moves
4925 into 64-bit and 32-bit moves. On 64-bit moves, add support for
4926 using direct move instructions on ISA 2.07. Also adjust
4927 instruction length for 64-bit.
4928 (mov<mode>_64bit, TFmode/TDmode): Likewise.
4929 (mov<mode>_32bit, TFmode/TDmode): Likewise.
4931 2014-02-15 Alan Modra <amodra@gmail.com>
4935 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
4936 find_replacement on parts of insn rtl that might be reloaded.
4938 2014-02-15 Richard Biener <rguenther@suse.de>
4940 PR tree-optimization/60183
4941 * tree-ssa-phiprop.c (propagate_with_phi): Avoid speculating loads.
4942 (tree_ssa_phiprop): Calculate and free post-dominators.
4944 2014-02-14 Jeff Law <law@redhat.com>
4946 PR rtl-optimization/60131
4947 * ree.c (get_extended_src_reg): New function.
4948 (combine_reaching_defs): Use it rather than assuming location of REG.
4949 (find_and_remove_re): Verify first operand of extension is
4950 a REG before adding the insns to the copy list.
4952 2014-02-14 Roland McGrath <mcgrathr@google.com>
4954 * configure.ac (HAVE_AS_IX86_UD2): New test for 'ud2' mnemonic.
4955 * configure: Regenerated.
4956 * config.in: Regenerated.
4957 * config/i386/i386.md (trap) [HAVE_AS_IX86_UD2]: Use the mnemonic
4958 instead of ASM_SHORT.
4960 2014-02-14 Vladimir Makarov <vmakarov@redhat.com>
4961 Richard Earnshaw <rearnsha@arm.com>
4963 PR rtl-optimization/59535
4964 * lra-constraints.c (process_alt_operands): Encourage alternative
4965 when unassigned pseudo class is superset of the alternative class.
4966 (inherit_reload_reg): Don't inherit when optimizing for code size.
4967 * config/arm/arm.h (MODE_BASE_REG_CLASS): Add version for LRA
4968 returning CORE_REGS for anything but Thumb1 and BASE_REGS for
4969 modes not less than 4 for Thumb1.
4971 2014-02-14 Kyle McMartin <kyle@redhat.com>
4974 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Define for AArch64.
4976 2014-02-14 Richard Biener <rguenther@suse.de>
4978 * cilk-common.c (cilk_arrow): Build a MEM_REF, not an INDIRECT_REF.
4979 (get_frame_arg): Drop the assert with langhook types_compatible_p.
4980 Do not strip INDIRECT_REFs.
4982 2014-02-14 Richard Biener <rguenther@suse.de>
4985 * lto-streamer-out.c (DFS_write_tree_body): Do not follow
4986 DECL_FUNCTION_SPECIFIC_TARGET.
4987 (hash_tree): Do not hash DECL_FUNCTION_SPECIFIC_TARGET.
4988 * tree-streamer-out.c (pack_ts_target_option): Remove.
4989 (streamer_pack_tree_bitfields): Do not stream TS_TARGET_OPTION.
4990 (write_ts_function_decl_tree_pointers): Do not stream
4991 DECL_FUNCTION_SPECIFIC_TARGET.
4992 * tree-streamer-in.c (unpack_ts_target_option): Remove.
4993 (unpack_value_fields): Do not stream TS_TARGET_OPTION.
4994 (lto_input_ts_function_decl_tree_pointers): Do not stream
4995 DECL_FUNCTION_SPECIFIC_TARGET.
4997 2014-02-14 Jakub Jelinek <jakub@redhat.com>
4999 * tree-vect-loop.c (vect_is_slp_reduction): Don't set use_stmt twice.
5000 (get_initial_def_for_induction, vectorizable_induction): Ignore
5001 debug stmts when looking for exit_phi.
5002 (vectorizable_live_operation): Fix up condition.
5004 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
5006 * config/nds32/nds32.c (nds32_asm_function_prologue): Do not use
5007 nreverse() because it changes the content of original tree list.
5009 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
5011 * config/nds32/t-mlibs (MULTILIB_OPTIONS): Fix typo in comment.
5012 * config/nds32/nds32.c (nds32_merge_decl_attributes): Likewise.
5014 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
5016 * config/nds32/nds32.c (nds32_naked_function_p): Follow the
5017 GNU coding standards.
5019 2014-02-13 Jakub Jelinek <jakub@redhat.com>
5022 * dwarf2out.c (gen_subprogram_die): Don't call
5023 add_calling_convention_attribute if subr_die is old_die.
5025 2014-02-13 Sharad Singhai <singhai@google.com>
5027 * doc/optinfo.texi: Fix order of nodes.
5029 2014-02-13 Uros Bizjak <ubizjak@gmail.com>
5031 * config/i386/sse.md (xop_vmfrcz<mode>2): Generate const0 in
5032 operands[2], not operands[3].
5034 2014-02-13 Richard Biener <rguenther@suse.de>
5037 * doc/install.texi (ISL): Update recommended version to 0.12.2,
5038 mention the possibility of an in-tree build.
5039 (CLooG): Update recommended version to 0.18.1, mention the
5040 possibility of an in-tree build and clarify that the ISL
5041 bundled with CLooG does not work.
5043 2014-02-13 Jakub Jelinek <jakub@redhat.com>
5046 * expr.c (compress_float_constant): If x is a hard register,
5047 extend into a pseudo and then move to x.
5049 2014-02-13 Dominik Vogt <vogt@linux.vnet.ibm.com>
5051 * config/s390/s390.c (s390_asm_output_function_label): Fix crash
5052 caused by bad second argument to warning_at() with -mhotpatch and
5053 nested functions (e.g. with gfortran).
5055 2014-02-13 Richard Sandiford <rdsandiford@googlemail.com>
5057 * opts.c (option_name): Remove "enabled by default" rider.
5059 2014-02-12 John David Anglin <danglin@gcc.gnu.org>
5061 * config/pa/pa.c (pa_option_override): Remove auto increment FIXME.
5063 2014-02-12 H.J. Lu <hongjiu.lu@intel.com>
5064 Uros Bizjak <ubizjak@gmail.com>
5067 * configure.ac (HAVE_AS_GOTOFF_IN_DATA): Pass --32 to GNU assembler.
5068 * configure: Regenerated.
5070 2014-02-12 Richard Biener <rguenther@suse.de>
5072 * vec.c (vec_prefix::calculate_allocation): Move as
5073 inline variant to vec.h.
5074 (vec_prefix::calculate_allocation_1): New out-of-line version.
5075 * vec.h (vec_prefix::calculate_allocation_1): Declare.
5076 (vec_prefix::m_has_auto_buf): Rename to ...
5077 (vec_prefix::m_using_auto_storage): ... this.
5078 (vec_prefix::calculate_allocation): Inline the easy cases
5079 and dispatch to calculate_allocation_1 which doesn't need the
5081 (va_heap::reserve): Use gcc_checking_assert.
5082 (vec<T, A, vl_embed>::embedded_init): Add argument to initialize
5083 m_using_auto_storage.
5084 (auto_vec): Change m_vecpfx member to a vec<T, va_heap, vl_embed>
5086 (vec<T, va_heap, vl_ptr>::reserve): Remove redundant check.
5087 (vec<T, va_heap, vl_ptr>::release): Avoid casting.
5088 (vec<T, va_heap, vl_ptr>::using_auto_storage): Simplify.
5090 2014-02-12 Richard Biener <rguenther@suse.de>
5092 * gcse.c (compute_transp): break from loop over canon_modify_mem_list
5093 when we found a dependence.
5095 2014-02-12 Thomas Schwinge <thomas@codesourcery.com>
5097 * gimplify.c (gimplify_call_expr, gimplify_modify_expr): Move
5099 (maybe_fold_stmt): ... into this new function.
5100 * omp-low.c (lower_omp): Update comment.
5102 * omp-low.c (lower_omp_target): Add clobber for sizes array, after
5105 * omp-low.c (diagnose_sb_0): Make sure label_ctx is valid to
5108 2014-02-12 James Greenhalgh <james.greenhalgh@arm.com>
5110 * config/arm/aarch-cost-tables.h (generic_extra_costs): Fix
5111 identifiers in comments.
5112 (cortexa53_extra_costs): Likewise.
5113 * config/arm/arm.c (cortexa9_extra_costs): Fix identifiers in comments.
5114 (cortexa7_extra_costs): Likewise.
5115 (cortexa12_extra_costs): Likewise.
5116 (cortexa15_extra_costs): Likewise.
5117 (v7m_extra_costs): Likewise.
5119 2014-02-12 Richard Biener <rguenther@suse.de>
5122 * gimple-low.c (lower_builtin_posix_memalign): Lower conditional
5123 of posix_memalign being successful.
5124 (lower_stmt): Restrict lowering of posix_memalign to when
5125 -ftree-bit-ccp is enabled.
5127 2014-02-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
5129 * config/avr/avr-c.c (avr_resolve_overloaded_builtin): Pass vNULL for
5131 * config/spu/spu-c.c (spu_resolve_overloaded_builtin): Likewise.
5133 2014-02-12 Eric Botcazou <ebotcazou@adacore.com>
5135 PR rtl-optimization/60116
5136 * combine.c (try_combine): Also remove dangling REG_DEAD notes on the
5137 other_insn once the combination has been validated.
5139 2014-02-11 Jan Hubicka <hubicka@ucw.cz>
5142 * ipa-utils.h (possible_polymorphic_call_targets): Update prototype
5144 * ipa-devirt.c: Include demangle.h
5145 (odr_violation_reported): New static variable.
5146 (add_type_duplicate): Update odr_violations.
5147 (maybe_record_node): Add completep parameter; update it.
5148 (record_target_from_binfo): Add COMPLETEP parameter;
5149 update it as needed.
5150 (possible_polymorphic_call_targets_1): Likewise.
5151 (struct polymorphic_call_target_d): Add nonconstruction_targets;
5152 rename FINAL to COMPLETE.
5153 (record_targets_from_bases): Sanity check we found the binfo;
5154 fix COMPLETEP updating.
5155 (possible_polymorphic_call_targets): Add NONCONSTRUTION_TARGETSP
5156 parameter, fix computing of COMPLETEP.
5157 (dump_possible_polymorphic_call_targets): Imrove readability of dump;
5158 at LTO time do demangling.
5159 (ipa_devirt): Use nonconstruction_targets; Improve dumps.
5160 * gimple-fold.c (gimple_get_virt_method_for_vtable): Add can_refer
5162 (gimple_get_virt_method_for_binfo): Likewise.
5163 * gimple-fold.h (gimple_get_virt_method_for_binfo,
5164 gimple_get_virt_method_for_vtable): Update prototypes.
5166 2014-02-11 Vladimir Makarov <vmakarov@redhat.com>
5169 * genautomata.c (add_presence_absence): Fix typo with
5170 {final_}presence_list.
5172 2014-02-11 Michael Meissner <meissner@linux.vnet.ibm.com>
5175 * config/rs6000/rs6000.md (128-bit GPR splitter): Add a splitter
5176 for VSX/Altivec vectors that land in GPR registers.
5178 2014-02-11 Richard Henderson <rth@redhat.com>
5179 Jakub Jelinek <jakub@redhat.com>
5182 * tree-sra.c (load_assign_lhs_subreplacements): Add VIEW_CONVERT_EXPR
5183 around drhs if type conversion to lacc->type is not useless.
5185 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5187 * config/aarch64/aarch64-cores.def (cortex-a57): Use cortexa57
5189 (cortex-a57.cortex-a53): Likewise.
5190 * config/aarch64/aarch64.c (cortexa57_tunings): New tuning struct.
5192 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5194 * config/arm/thumb2.md (*thumb2_movhi_insn): Add alternatives for
5197 2014-02-11 Renlin Li <Renlin.Li@arm.com>
5199 * doc/sourcebuild.texi: Document check_effective_target_arm_vfp3_ok and
5200 add_options_for_arm_vfp3.
5202 2014-02-11 Jeff Law <law@redhat.com>
5205 * expr.c (expand_expr_addr_expr_1): Handle expand_expr returning an
5206 object with an undesirable mode.
5208 2014-02-11 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5211 * config/i386/sol2-9.h: New file.
5212 * config.gcc (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*,
5213 *-*-solaris2.9*): Use it.
5215 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
5217 * config/microblaze/microblaze.md: Add movsi4_rev insn pattern.
5218 * config/microblaze/predicates.md: Add reg_or_mem_operand predicate.
5220 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
5222 * config/microblaze/microblaze.c: Extend mcpu version format
5224 2014-02-10 David Holsgrove <david.holsgrove@xilinx.com>
5226 * config/microblaze/microblaze.h: Define SIZE_TYPE and PTRDIFF_TYPE.
5228 2014-02-10 Richard Henderson <rth@redhat.com>
5231 * calls.c (expand_call): Don't double-push for reg_parm_stack_space.
5232 * config/i386/i386.c (init_cumulative_args): Remove sorry for 64-bit
5233 ms-abi vs -mno-accumulate-outgoing-args.
5234 (ix86_expand_prologue): Unconditionally call ix86_eax_live_at_start_p.
5235 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Fix comment with
5238 2014-02-10 Bernd Edlinger <bernd.edlinger@hotmail.de>
5241 * cfgexpand.c (expand_asm_operands): Attach source location to
5242 ASM_INPUT rtx objects.
5243 * print-rtl.c (print_rtx): Check for UNKNOWN_LOCATION.
5245 2014-02-10 Nick Clifton <nickc@redhat.com>
5247 * config/mn10300/mn10300.c (popcount): New function.
5248 (mn10300_expand_prologue): Include saved registers in stack usage
5251 2014-02-10 Jeff Law <law@redhat.com>
5254 * reload1.c (emit_input_reload_insns): Do not create invalid RTL
5255 when changing the SET_DEST of a prior insn to avoid an input reload.
5257 2014-02-10 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
5259 * config/rs6000/sysv4.h (ENDIAN_SELECT): Do not attempt to enforce
5260 big-endian mode for -mcall-aixdesc, -mcall-freebsd, -mcall-netbsd,
5261 -mcall-openbsd, or -mcall-linux.
5262 (CC1_ENDIAN_BIG_SPEC): Remove.
5263 (CC1_ENDIAN_LITTLE_SPEC): Remove.
5264 (CC1_ENDIAN_DEFAULT_SPEC): Remove.
5265 (CC1_SPEC): Remove (always empty) %cc1_endian_... spec.
5266 (SUBTARGET_EXTRA_SPECS): Remove %cc1_endian_big, %cc1_endian_little,
5267 and %cc1_endian_default.
5268 * config/rs6000/sysv4le.h (CC1_ENDIAN_DEFAULT_SPEC): Remove.
5270 2014-02-10 Richard Biener <rguenther@suse.de>
5272 PR tree-optimization/60115
5273 * tree-eh.c (tree_could_trap_p): Unify TARGET_MEM_REF and
5274 MEM_REF handling. Properly verify that the accesses are not
5275 out of the objects bound.
5277 2014-02-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5279 * config/aarch64/aarch64.c (aarch64_override_options): Fix typo from
5282 2014-02-10 Eric Botcazou <ebotcazou@adacore.com>
5284 * ipa-devirt.c (get_polymorphic_call_info_from_invariant): Return
5285 proper constants and fix formatting.
5286 (possible_polymorphic_call_targets): Fix formatting.
5288 2014-02-10 Kirill Yukhin <kirill.yukhin@intel.com>
5289 Ilya Tocar <ilya.tocar@intel.com>
5291 * config/i386/avx512fintrin.h (_mm512_storeu_epi64): Removed.
5292 (_mm512_loadu_epi32): Renamed into...
5293 (_mm512_loadu_si512): This.
5294 (_mm512_storeu_epi32): Renamed into...
5295 (_mm512_storeu_si512): This.
5296 (_mm512_maskz_ceil_ps): Removed.
5297 (_mm512_maskz_ceil_pd): Ditto.
5298 (_mm512_maskz_floor_ps): Ditto.
5299 (_mm512_maskz_floor_pd): Ditto.
5300 (_mm512_floor_round_ps): Ditto.
5301 (_mm512_floor_round_pd): Ditto.
5302 (_mm512_ceil_round_ps): Ditto.
5303 (_mm512_ceil_round_pd): Ditto.
5304 (_mm512_mask_floor_round_ps): Ditto.
5305 (_mm512_mask_floor_round_pd): Ditto.
5306 (_mm512_mask_ceil_round_ps): Ditto.
5307 (_mm512_mask_ceil_round_pd): Ditto.
5308 (_mm512_maskz_floor_round_ps): Ditto.
5309 (_mm512_maskz_floor_round_pd): Ditto.
5310 (_mm512_maskz_ceil_round_ps): Ditto.
5311 (_mm512_maskz_ceil_round_pd): Ditto.
5312 (_mm512_expand_pd): Ditto.
5313 (_mm512_expand_ps): Ditto.
5314 * config/i386/i386.c (ix86_builtins): Remove
5315 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK.
5316 (bdesc_args): Ditto.
5317 * config/i386/predicates.md (const1256_operand): New.
5318 (const_1_to_2_operand): Ditto.
5319 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
5320 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
5321 (*avx512pf_gatherpf<mode>sf): Ditto.
5322 (avx512pf_gatherpf<mode>df): Ditto.
5323 (*avx512pf_gatherpf<mode>df_mask): Ditto.
5324 (*avx512pf_gatherpf<mode>df): Ditto.
5325 (avx512pf_scatterpf<mode>sf): Ditto.
5326 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
5327 (*avx512pf_scatterpf<mode>sf): Ditto.
5328 (avx512pf_scatterpf<mode>df): Ditto.
5329 (*avx512pf_scatterpf<mode>df_mask): Ditto.
5330 (*avx512pf_scatterpf<mode>df): Ditto.
5331 (avx512f_expand<mode>): Removed.
5332 (<shift_insn><mode>3<mask_name>): Change predicate type.
5334 2014-02-08 Jakub Jelinek <jakub@redhat.com>
5336 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers
5337 not at the end of datarefs vector use ordered_remove to avoid
5338 reordering datarefs vector.
5341 * gimplify.c (gimplify_bind_expr): In ORT_SIMD region
5342 mark local addressable non-static vars as GOVD_PRIVATE
5343 instead of GOVD_LOCAL.
5344 * omp-low.c (lower_omp_for): Move gimple_bind_vars
5345 and BLOCK_VARS of gimple_bind_block to new_stmt rather
5349 * tree-ssa-ccp.c (surely_varying_stmt_p): Don't return true
5350 if TYPE_ATTRIBUTES (gimple_call_fntype ()) contain
5351 assume_aligned or alloc_align attributes.
5352 (bit_value_assume_aligned): Add ATTR, PTRVAL and ALLOC_ALIGN
5353 arguments. Handle also assume_aligned and alloc_align attributes.
5354 (evaluate_stmt): Adjust bit_value_assume_aligned caller. Handle
5355 calls to functions with assume_aligned or alloc_align attributes.
5356 * doc/extend.texi: Document assume_aligned and alloc_align attributes.
5358 2014-02-08 Terry Guo <terry.guo@arm.com>
5360 * doc/invoke.texi: Document ARM -march=armv7e-m.
5362 2014-02-08 Jakub Jelinek <jakub@redhat.com>
5364 * cilk-common.c (cilk_init_builtins): Clear TREE_NOTHROW
5365 flag on __cilkrts_rethrow builtin.
5368 * ipa-cp.c (determine_versionability): Fail at -O0
5369 or __attribute__((optimize (0))) or -fno-ipa-cp functions.
5370 * tree-sra.c (ipa_sra_preliminary_function_checks): Similarly.
5373 2014-02-04 Jakub Jelinek <jakub@redhat.com>
5376 * tree-inline.c (copy_forbidden): Fail for
5377 __attribute__((optimize (0))) functions.
5379 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
5381 * varpool.c: Include pointer-set.h.
5382 (varpool_remove_unreferenced_decls): Variables in other partitions
5383 will not be output; be however careful to not lose information
5386 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
5388 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do O(1)
5389 lookup in the vtable constructor.
5391 2014-02-07 Jeff Law <law@redhat.com>
5394 * config/m68k/m68k.md (ashldi_extsi): Turn into a
5395 define_insn_and_split.
5397 * ipa-inline.c (inline_small_functions): Fix typos.
5399 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5401 * config/s390/s390-protos.h (s390_can_use_simple_return_insn)
5402 (s390_can_use_return_insn): Declare.
5403 * config/s390/s390.h (EPILOGUE_USES): Define.
5404 * config/s390/s390.c (s390_mainpool_start): Allow two main_pool
5406 (s390_chunkify_start): Handle return JUMP_LABELs.
5407 (s390_early_mach): Emit a main_pool instruction on the entry edge.
5408 (s300_set_up_by_prologue, s390_can_use_simple_return_insn)
5409 (s390_can_use_return_insn): New functions.
5410 (s390_fix_long_loop_prediction): Handle conditional returns.
5411 (TARGET_SET_UP_BY_PROLOGUE): Define.
5412 * config/s390/s390.md (ANY_RETURN): New code iterator.
5413 (*creturn, *csimple_return, return, simple_return): New patterns.
5415 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5417 * config/s390/s390.c (s390_restore_gprs_from_fprs): Add REG_CFA_RESTORE
5418 notes to each restore. Also add REG_CFA_DEF_CFA when restoring %r15.
5419 (s390_optimize_prologue): Don't clear RTX_FRAME_RELATED_P. Update the
5420 REG_CFA_RESTORE list when deciding not to restore a register.
5422 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5424 * config/s390/s390.c: Include tree-pass.h and context.h.
5425 (s390_early_mach): New function, split out from...
5426 (s390_emit_prologue): ...here.
5427 (pass_data_s390_early_mach): New pass structure.
5428 (pass_s390_early_mach): New class.
5429 (s390_option_override): Create and register early_mach pass.
5430 Move to end of file.
5432 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5434 * var-tracking.c (vt_stack_adjustments): Don't require stack_adjusts
5435 to match for the exit block.
5437 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5439 * config/s390/s390.md ("atomic_load<mode>", "atomic_store<mode>")
5440 ("atomic_compare_and_swap<mode>", "atomic_fetch_<atomic><mode>"):
5441 Reject misaligned operands.
5443 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5445 * optabs.c (expand_atomic_compare_and_swap): Allow expander to fail.
5447 2014-02-07 Richard Biener <rguenther@suse.de>
5450 * gimple-low.c (lower_builtin_posix_memalign): New function.
5451 (lower_stmt): Call it to lower posix_memalign in a way
5452 to make alignment info accessible.
5454 2014-02-07 Jakub Jelinek <jakub@redhat.com>
5457 * tree.c (build_common_builtin_nodes): Set ECF_LEAF for
5458 __builtin_setjmp_receiver.
5460 2014-02-07 Richard Biener <rguenther@suse.de>
5463 * builtin-types.def (BT_FN_INT_PTRPTR_SIZE_SIZE): Add.
5464 * builtins.def (BUILT_IN_POSIX_MEMALIGN): Likewise.
5465 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
5466 Handle BUILT_IN_POSIX_MEMALIGN.
5467 (find_func_clobbers): Likewise.
5468 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
5469 (call_may_clobber_ref_p_1): Likewise.
5471 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5474 * ipa-devirt.c (record_target_from_binfo): Remove overactive
5477 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5480 * lto-cgraph.c (lto_output_node): Use
5481 symtab_get_symbol_partitioning_class.
5482 (lto_output_varpool_node): likewise.
5483 (symtab_get_symbol_partitioning_class): Move here from
5485 * cgraph.h (symbol_partitioning_class): Likewise.
5486 (symtab_get_symbol_partitioning_class): Declare.
5488 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5490 * ggc.h (ggc_internal_cleared_alloc): New macro.
5491 * vec.h (vec_safe_copy): Handle memory stats.
5492 * omp-low.c (simd_clone_struct_alloc): Use ggc_internal_cleared_alloc.
5493 * target-globals.c (save_target_globals): Likewise.
5495 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5498 * expr.c (emit_move_resolve_push): Export; be bit more selective
5499 on when to clear alias set.
5500 * expr.h (emit_move_resolve_push): Declare.
5501 * function.h (struct function): Add tail_call_marked.
5502 * tree-tailcall.c (optimize_tail_call): Set tail_call_marked.
5503 * config/i386/i386-protos.h (ix86_expand_push): Remove.
5504 * config/i386/i386.md (TImode move expander): De not call
5506 (FP push expanders): Preserve memory attributes.
5507 * config/i386/sse.md (push<mode>1): Remove.
5508 * config/i386/i386.c (ix86_expand_vector_move): Handle push operation.
5509 (ix86_expand_push): Remove.
5510 * config/i386/mmx.md (push<mode>1): Remove.
5512 2014-02-06 Jakub Jelinek <jakub@redhat.com>
5514 PR rtl-optimization/60030
5515 * internal-fn.c (ubsan_expand_si_overflow_mul_check): Surround
5516 lopart with paradoxical subreg before shifting it up by hprec.
5518 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5520 * config/arm/aarch-cost-tables.h (cortexa57_extra_costs): New table.
5521 Remove extra newline at end of file.
5522 * config/arm/arm.c (arm_cortex_a57_tune): New tuning struct.
5523 (arm_issue_rate): Handle cortexa57.
5524 * config/arm/arm-cores.def (cortex-a57): Use cortex_a57 tuning.
5525 (cortex-a57.cortex-a53): Likewise.
5527 2014-02-06 Jakub Jelinek <jakub@redhat.com>
5530 * config/arm/arm.c (emit_multi_reg_push): Add dwarf_regs_mask argument,
5531 don't record in REG_FRAME_RELATED_EXPR registers not set in that
5533 (arm_expand_prologue): Adjust all callers.
5534 (arm_unwind_emit_sequence): Allow saved, but not important for unwind
5535 info, registers also at the lowest numbered registers side. Use
5536 gcc_assert instead of abort, and SET_SRC/SET_DEST macros instead of
5540 * var-tracking.c (adjust_mems): Before adding a SET to
5541 amd->side_effects, adjust it's SET_SRC using simplify_replace_fn_rtx.
5543 2014-02-06 Alan Modra <amodra@gmail.com>
5546 * config/rs6000/rs6000.c (rs6000_secondary_memory_needed_mode): Only
5547 change SDmode to DDmode when lra_in_progress.
5549 2014-02-06 Jakub Jelinek <jakub@redhat.com>
5552 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers, call
5553 free_data_ref on the dr first, and before goto again also set dr
5554 to the next dr. For simd_lane_access, free old datarefs[i] before
5555 overwriting it. For get_vectype_for_scalar_type failure, don't
5556 free_data_ref if simd_lane_access.
5558 * Makefile.in (prefix.o, cppbuiltin.o): Depend on $(BASEVER).
5561 * tree.h (opts_for_fn): New inline function.
5562 (opt_for_fn): Define.
5563 * config/i386/i386.c (ix86_function_regparm): Use
5564 opt_for_fn (decl, optimize) instead of optimize.
5566 2014-02-06 Marcus Shawcroft <marcus.shawcroft@arm.com>
5568 * config/aarch64/aarch64.c (aarch64_classify_symbol): Fix logic
5569 for SYMBOL_REF in large memory model.
5571 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5573 * config/aarch64/aarch64-cores.def (cortex-a53): Specify CRC32
5575 (cortex-a57): Likewise.
5576 (cortex-a57.cortex-a53): Likewise.
5578 2014-02-06 Yury Gribov <y.gribov@samsung.com>
5579 Kugan Vivekanandarajah <kuganv@linaro.org>
5581 * config/arm/arm.c (arm_vector_alignment_reachable): Check
5583 * config/arm/arm.c (arm_builtin_support_vector_misalignment): Likewise.
5585 2014-02-06 Richard Biener <rguenther@suse.de>
5587 * tree-cfg.c (gimple_duplicate_sese_region): Fix ordering of
5588 set_loop_copy and initialize_original_copy_tables.
5590 2014-02-06 Alex Velenko <Alex.Velenko@arm.com>
5592 * config/aarch64/aarch64-simd.md
5593 (aarch64_ashr_simddi): Change QI to SI.
5595 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
5596 Jakub Jelinek <jakub@redhat.com>
5599 * ipa-inline-analysis.c (compute_bb_predicates): Ensure monotonicity
5602 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5604 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Change
5605 CODE_FOR_altivec_vpku[hw]um to
5606 CODE_FOR_altivec_vpku[hw]um_direct.
5607 * config/rs6000/altivec.md (vec_unpacks_hi_<VP_small_lc>): Change
5608 UNSPEC_VUNPACK_HI_SIGN to UNSPEC_VUNPACK_HI_SIGN_DIRECT.
5609 (vec_unpacks_lo_<VP_small_lc>): Change UNSPEC_VUNPACK_LO_SIGN to
5610 UNSPEC_VUNPACK_LO_SIGN_DIRECT.
5612 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5614 * config/rs6000/altivec.md (altivec_vsum2sws): Adjust code
5615 generation for -maltivec=be.
5616 (altivec_vsumsws): Simplify redundant test.
5618 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5620 * altivec.md (UNSPEC_VPACK_UNS_UNS_MOD_DIRECT): New unspec.
5621 (UNSPEC_VUNPACK_HI_SIGN_DIRECT): Likewise.
5622 (UNSPEC_VUNPACK_LO_SIGN_DIRECT): Likewise.
5623 (mulv8hi3): Use gen_altivec_vpkuwum_direct instead of
5624 gen_altivec_vpkuwum.
5625 (altivec_vpkpx): Test for VECTOR_ELT_ORDER_BIG instead of for
5627 (altivec_vpks<VI_char>ss): Likewise.
5628 (altivec_vpks<VI_char>us): Likewise.
5629 (altivec_vpku<VI_char>us): Likewise.
5630 (altivec_vpku<VI_char>um): Likewise.
5631 (altivec_vpku<VI_char>um_direct): New (copy of
5632 altivec_vpku<VI_char>um that still relies on BYTES_BIG_ENDIAN, for
5634 (altivec_vupkhs<VU_char>): Emit vupkls* instead of vupkhs* when
5635 target is little endian and -maltivec=be is not specified.
5636 (*altivec_vupkhs<VU_char>_direct): New (copy of
5637 altivec_vupkhs<VU_char> that always emits vupkhs*, for internal use).
5638 (altivec_vupkls<VU_char>): Emit vupkhs* instead of vupkls* when
5639 target is little endian and -maltivec=be is not specified.
5640 (*altivec_vupkls<VU_char>_direct): New (copy of
5641 altivec_vupkls<VU_char> that always emits vupkls*, for internal use).
5642 (altivec_vupkhpx): Emit vupklpx instead of vupkhpx when target is
5643 little endian and -maltivec=be is not specified.
5644 (altivec_vupklpx): Emit vupkhpx instead of vupklpx when target is
5645 little endian and -maltivec=be is not specified.
5647 2014-02-05 Richard Henderson <rth@redhat.com>
5650 * combine-stack-adj.c: Revert r206943.
5651 * sched-int.h (struct deps_desc): Add last_args_size.
5652 * sched-deps.c (init_deps): Initialize it.
5653 (sched_analyze_insn): Add OUTPUT dependencies between insns that
5654 contain REG_ARGS_SIZE notes.
5656 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
5658 * lto-cgraph.c (asm_nodes_output): Make global.
5659 * lto-wrapper.c (run_gcc): Pass down paralelizm to WPA.
5660 * gcc.c (AS_NEEDS_DASH_FOR_PIPED_INPUT): Allow WPA parameter
5661 (driver_handle_option): Handle OPT_fwpa.
5663 2014-02-05 Jakub Jelinek <jakub@redhat.com>
5666 * ipa-devirt.c (possible_polymorphic_call_targets): Fix
5667 a comment typo and formatting issue. If odr_hash hasn't been
5668 created, return vNULL and set *completep to false.
5671 * tree-eh.c (cleanup_empty_eh): Bail out on totally empty
5672 bb with no successors.
5674 2014-02-05 James Greenhalgh <james.greenhalgh@arm.com>
5677 * doc/invoke.texi (-march): Clarify documentation for ARM.
5681 2014-02-05 Richard Biener <rguenther@suse.de>
5683 * tree-vect-loop.c (vect_analyze_loop_2): Be more informative
5684 when not vectorizing because of too many alias checks.
5685 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
5686 Add more verboseness, avoid duplicate MSG_MISSED_OPTIMIZATION.
5688 2014-02-05 Nick Clifton <nickc@redhat.com>
5690 * config/mn10300/mn10300.c (mn10300_hard_regno_mode_ok): Do not
5691 accept extended registers in any mode when compiling for the MN10300.
5693 2014-02-05 Yury Gribov <y.gribov@samsung.com>
5695 * cif-code.def (ATTRIBUTE_MISMATCH): New CIF code.
5696 * ipa-inline.c (report_inline_failed_reason): Handle mismatched
5697 sanitization attributes.
5698 (can_inline_edge_p): Likewise.
5699 (sanitize_attrs_match_for_inline_p): New function.
5701 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
5703 * ipa-prop.c (detect_type_change): Shor circuit testing of
5704 type changes on THIS pointer.
5706 2014-02-04 John David Anglin <danglin@gcc.gnu.org>
5709 * config/pa/pa.c (legitimize_tls_address): Return original address
5710 if not passed a SYMBOL_REF rtx.
5711 (hppa_legitimize_address): Call legitimize_tls_address for all TLS
5713 (pa_emit_move_sequence): Simplify TLS source operands.
5714 (pa_legitimate_constant_p): Reject all TLS constants.
5715 * config/pa/pa.h (PA_SYMBOL_REF_TLS_P): Correct comment.
5716 (CONSTANT_ADDRESS_P): Reject TLS CONST addresses.
5718 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
5720 * ipa.c (function_and_variable_visibility): Decompose DECL_ONE_ONLY
5721 groups when we know they are controlled by LTO.
5722 * varasm.c (default_binds_local_p_1): If object is in other partition,
5723 it will be resolved locally.
5725 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
5727 * config/host-linux.c (linux_gt_pch_use_address): Don't
5728 use SSIZE_MAX because it is not always defined.
5730 2014-02-04 Vladimir Makarov <vmakarov@redhat.com>
5733 * lra-constraints.c (need_for_split_p): Use more 3 reloads as
5734 threshold for pseudo splitting.
5735 (update_ebb_live_info): Process call argument hard registers and
5736 hard registers from insn definition too.
5737 (max_small_class_regs_num): New constant.
5738 (inherit_in_ebb): Update live hard regs through EBBs. Update
5739 reloads_num only for small register classes. Don't split for
5742 2014-02-04 Markus Trippelsdorf <markus@trippelsdorf.de>
5745 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that target
5748 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
5750 * gimple-fold.c (can_refer_decl_in_current_unit_p): Default
5753 2014-02-04 Marek Polacek <polacek@redhat.com>
5755 * gdbinit.in (pel): Define.
5757 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
5759 * doc/invoke.texi (fstrict-volatile-bitfields): Clarify current
5762 2014-02-04 Richard Biener <rguenther@suse.de>
5765 * lto-streamer-out.c (tree_is_indexable): Force NAMELIST_DECLs
5766 in function context local.
5767 (lto_output_tree_ref): Do not write trees from lto_output_tree_ref.
5768 * lto-streamer-in.c (lto_input_tree_ref): Handle LTO_namelist_decl_ref
5769 similar to LTO_imported_decl_ref.
5771 2014-02-04 Jakub Jelinek <jakub@redhat.com>
5773 PR tree-optimization/60002
5774 * cgraphclones.c (build_function_decl_skip_args): Clear
5777 PR tree-optimization/60023
5778 * tree-if-conv.c (predicate_mem_writes): Pass true instead of
5779 false to gsi_replace.
5780 * tree-vect-stmts.c (vect_finish_stmt_generation): If stmt
5781 has been in some EH region and vec_stmt could throw, add
5782 vec_stmt into the same EH region.
5783 * tree-data-ref.c (get_references_in_stmt): If IFN_MASK_LOAD
5784 has no lhs, ignore it.
5785 * internal-fn.c (expand_MASK_LOAD): Likewise.
5788 * tree-inline.c (copy_forbidden): Fail for
5789 __attribute__((optimize (0))) functions.
5792 * omp-low.c (simd_clone_struct_copy): If from->inbranch
5793 is set, copy one less argument.
5794 (expand_simd_clones): Don't subtract clone_info->inbranch
5795 from simd_clone_struct_alloc argument.
5797 PR rtl-optimization/57915
5798 * recog.c (simplify_while_replacing): If all unary/binary/relational
5799 operation arguments are constant, attempt to simplify those.
5802 * expmed.c (expand_mult): For MODE_VECTOR_INT multiplication
5803 if there is no vashl<mode>3 or ashl<mode>3 insn, skip_synth.
5805 2014-02-04 Richard Biener <rguenther@suse.de>
5807 PR tree-optimization/60012
5808 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Apply
5809 TBAA disambiguation to all DDRs.
5811 2014-02-04 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5814 * config/sol2.h (LINK_LIBGCC_MAPFILE_SPEC): Define.
5815 (LINK_SPEC): Use it for -shared, -shared-libgcc.
5817 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
5820 * tree.c (get_binfo_at_offset): Do not get confused by empty classes;
5822 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
5824 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Remove.
5825 * gimple-fold.h (gimple_extract_devirt_binfo_from_cst): Remove.
5827 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
5830 * ipa-cp.c (ipa_get_indirect_edge_target_1): Use ipa-devirt
5831 to figure out targets of polymorphic calls with known decl.
5832 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
5833 * ipa-utils.h (get_polymorphic_call_info_from_invariant): Declare.
5834 * ipa-devirt.c (get_polymorphic_call_info_for_decl): Break out from ...
5835 (get_polymorphic_call_info): ... here.
5836 (get_polymorphic_call_info_from_invariant): New function.
5838 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
5840 * ipa-cp.c (ipa_get_indirect_edge_target_1): Do direct
5841 lookup via vtable pointer; check for type consistency
5842 and turn inconsitent facts into UNREACHABLE.
5843 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
5844 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do not ICE on
5845 type inconsistent querries; return UNREACHABLE instead.
5847 2014-02-03 Richard Henderson <rth@twiddle.net>
5850 * tree-ssa-uninit.c (push_to_worklist): Don't re-push if we've
5851 already processed this node.
5852 (normalize_one_pred_1): Pass along mark_set.
5853 (normalize_one_pred): Create and destroy a pointer_set_t.
5854 (normalize_one_pred_chain): Likewise.
5856 2014-02-03 Laurent Aflonsi <laurent.alfonsi@st.com>
5858 PR gcov-profile/58602
5859 * gcc/gcov-io.c (gcov_open): Open with truncation when mode < 0.
5861 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
5864 * ipa-cp.c (ipa_get_indirect_edge_target_1): Give up on
5865 -fno-devirtualize; try to devirtualize by the knowledge of
5866 virtual table pointer given by aggregate propagation.
5867 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
5868 (ipa_print_node_jump_functions): Dump also offset that
5869 is relevant for polymorphic calls.
5870 (determine_known_aggregate_parts): Add arg_type parameter; use it
5871 instead of determining the type from pointer type.
5872 (ipa_compute_jump_functions_for_edge): Update call of
5873 determine_known_aggregate_parts.
5874 * gimple-fold.c (gimple_get_virt_method_for_vtable): Break out from ...
5875 (gimple_get_virt_method_for_binfo): ... here; simplify using
5876 vtable_pointer_value_to_vtable.
5877 * gimple-fold.h (gimple_get_virt_method_for_vtable): Declare.
5878 * ipa-devirt.c (subbinfo_with_vtable_at_offset): Turn OFFSET parameter
5879 to unsigned HOST_WIDE_INT; use vtable_pointer_value_to_vtable.
5880 (vtable_pointer_value_to_vtable): Break out from ...; handle also
5882 (vtable_pointer_value_to_binfo): ... here.
5883 * ipa-utils.h (vtable_pointer_value_to_vtable): Declare.
5885 2014-02-03 Teresa Johnson <tejohnson@google.com>
5887 * tree-vect-slp.c (vect_supported_load_permutation_p): Avoid
5888 redef of outer loop index variable.
5890 2014-02-03 Marc Glisse <marc.glisse@inria.fr>
5894 * doc/extend.texi (Function Attributes): Typo.
5896 2014-02-03 Cong Hou <congh@google.com>
5898 PR tree-optimization/60000
5899 * tree-vect-loop.c (vect_transform_loop): Set pattern_def_seq to NULL
5900 if the vectorized statement is a store. A store statement can only
5901 appear at the end of pattern statements.
5903 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
5905 * config/i386/i386.c (flag_opts): Add -mlong-double-128.
5906 (ix86_option_override_internal): Default long double to 64-bit for
5907 32-bit Bionic and to 128-bit for 64-bit Bionic.
5909 * config/i386/i386.h (LONG_DOUBLE_TYPE_SIZE): Use 128 if
5910 TARGET_LONG_DOUBLE_128 is true.
5911 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Likewise.
5913 * config/i386/i386.opt (mlong-double-80): Negate -mlong-double-64.
5914 (mlong-double-64): Negate -mlong-double-128.
5915 (mlong-double-128): New option.
5917 * config/i386/i386-c.c (ix86_target_macros): Define
5918 __LONG_DOUBLE_128__ for TARGET_LONG_DOUBLE_128.
5920 * doc/invoke.texi: Document -mlong-double-128.
5922 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
5924 PR rtl-optimization/60024
5925 * sel-sched.c (init_regs_for_mode): Check if mode is OK first.
5927 2014-02-03 Markus Trippelsdorf <markus@trippelsdorf.de>
5929 * doc/invoke.texi (fprofile-reorder-functions): Fix typo.
5931 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
5933 PR rtl-optimization/57662
5934 * sel-sched.c (code_motion_path_driver): Do not mark already not
5935 existing blocks in the visiting bitmap.
5937 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
5939 * sel-sched-ir.c (sel_gen_insn_from_expr_after): Reset INSN_DELETED_P
5940 on the insn being emitted.
5942 2014-02-03 James Greenhalgh <james.greenhalgh@arm.com>
5943 Will Deacon <will.deacon@arm.com>
5945 * doc/gimple.texi (gimple_asm_clear_volatile): Remove.
5947 2014-02-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5949 * config/arm/arm-tables.opt: Regenerate.
5951 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5953 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): Generalize
5954 for vector types other than V16QImode.
5955 * config/rs6000/altivec.md (altivec_vperm_<mode>): Change to a
5956 define_expand, and call altivec_expand_vec_perm_le when producing
5957 code with little endian element order.
5958 (*altivec_vperm_<mode>_internal): New insn having previous
5959 behavior of altivec_vperm_<mode>.
5960 (altivec_vperm_<mode>_uns): Change to a define_expand, and call
5961 altivec_expand_vec_perm_le when producing code with little endian
5963 (*altivec_vperm_<mode>_uns_internal): New insn having previous
5964 behavior of altivec_vperm_<mode>_uns.
5966 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5968 * config/rs6000/altivec.md (UNSPEC_VSUMSWS_DIRECT): New unspec.
5969 (altivec_vsumsws): Add handling for -maltivec=be with a little
5971 (altivec_vsumsws_direct): New.
5972 (reduc_splus_<mode>): Call gen_altivec_vsumsws_direct instead of
5973 gen_altivec_vsumsws.
5975 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
5977 * ipa-devirt.c (subbinfo_with_vtable_at_offset,
5978 vtable_pointer_value_to_binfo): New functions.
5979 * ipa-utils.h (vtable_pointer_value_to_binfo): Declare.
5980 * ipa-prop.c (extr_type_from_vtbl_ptr_store): Use it.
5982 2014-02-02 Sandra Loosemore <sandra@codesourcery.com>
5984 * config/nios2/nios2.md (load_got_register): Initialize GOT
5985 pointer from _gp_got instead of _GLOBAL_OFFSET_TABLE_.
5986 * config/nios2/nios2.c (nios2_function_profiler): Likewise.
5988 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
5990 * ipa-prop.c (update_jump_functions_after_inlining): When type is not
5991 preserverd by passthrough, do not propagate the type.
5993 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
5995 * config/mips/mips.c (MIPS_GET_FCSR, MIPS_SET_FCSR): New macros.
5996 (mips_atomic_assign_expand_fenv): New function.
5997 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): Define.
5999 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
6001 * doc/extend.texi (__builtin_mips_get_fcsr): Document.
6002 (__builtin_mips_set_fcsr): Likewise.
6003 * config/mips/mips-ftypes.def: Add MIPS_VOID_FTYPE_USI and
6004 MIPS_USI_FTYPE_VOID.
6005 * config/mips/mips-protos.h (mips16_expand_get_fcsr): Declare
6006 (mips16_expand_set_fcsr): Likewise.
6007 * config/mips/mips.c (mips16_get_fcsr_stub): New variable.
6008 (mips16_set_fcsr_stub): Likewise.
6009 (mips16_get_fcsr_one_only_stub): New class.
6010 (mips16_set_fcsr_one_only_stub): Likewise.
6011 (mips16_expand_get_fcsr, mips16_expand_set_fcsr): New functions.
6012 (mips_code_end): Output the get_fcsr and set_fcsr stubs, if needed.
6013 (BUILTIN_AVAIL_MIPS16, AVAIL_ALL): New macros.
6014 (hard_float): New availability predicate.
6015 (mips_builtins): Add get_fcsr and set_fcsr.
6016 (mips_expand_builtin): Check BUILTIN_AVAIL_MIPS16.
6017 * config/mips/mips.md (UNSPEC_GET_FCSR, UNSPEC_SET_FCSR): New unspecs.
6018 (GET_FCSR_REGNUM, SET_FCSR_REGNUM): New constants.
6019 (mips_get_fcsr, *mips_get_fcsr, mips_get_fcsr_mips16_<mode>)
6020 (mips_set_fcsr, *mips_set_fcsr, mips_set_fcsr_mips16_<mode>): New
6023 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
6025 * config/mips/mips.c (mips_one_only_stub): New class.
6026 (mips_need_mips16_rdhwr_p): Replace with...
6027 (mips16_rdhwr_stub): ...this new variable.
6028 (mips16_stub_call_address): New function.
6029 (mips16_rdhwr_one_only_stub): New class.
6030 (mips_expand_thread_pointer): Use mips16_stub_call_address.
6031 (mips_output_mips16_rdhwr): Delete.
6032 (mips_finish_stub): New function.
6033 (mips_code_end): Use it to handle rdhwr stubs.
6035 2014-02-02 Uros Bizjak <ubizjak@gmail.com>
6038 * config/i386/i386.c (classify_argument): Fix handling of bit_offset
6039 when calculating size of integer atomic types.
6041 2014-02-02 H.J. Lu <hongjiu.lu@intel.com>
6043 * ipa-inline-analysis.c (true_predicate_p): Fix a typo in comments.
6045 2014-02-01 Jakub Jelinek <jakub@redhat.com>
6047 PR tree-optimization/60003
6048 * gimple-low.c (lower_builtin_setjmp): Set cfun->has_nonlocal_label.
6049 * profile.c (branch_prob): Use gimple_call_builtin_p
6050 to check for BUILT_IN_SETJMP_RECEIVER.
6051 * tree-inline.c (copy_bb): Call notice_special_calls.
6053 2014-01-31 Vladimir Makarov <vmakarov@redhat.com>
6056 * lra-constraints.c (process_alt_operands): Update reload_sum only
6059 2014-01-31 Richard Henderson <rth@redhat.com>
6062 * tree-eh.c (lower_try_finally_switch): Delay lowering finally block
6063 until after else_eh is processed.
6065 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
6067 * config/i386/avx512fintrin.h (_MM_FROUND_TO_NEAREST_INT),
6068 (_MM_FROUND_TO_NEG_INF), (_MM_FROUND_TO_POS_INF),
6069 (_MM_FROUND_TO_ZERO), (_MM_FROUND_CUR_DIRECTION): Are already defined
6070 in smmintrin.h, remove them.
6071 (_MM_FROUND_NO_EXC): Same as above, bit also wrong value.
6072 * config/i386/i386.c (ix86_print_operand): Split sae and rounding.
6073 * config/i386/i386.md (ROUND_SAE): Fix value.
6074 * config/i386/predicates.md (const_4_or_8_to_11_operand): New.
6075 (const48_operand): New.
6076 * config/i386/subst.md (round), (round_expand): Use
6077 const_4_or_8_to_11_operand.
6078 (round_saeonly), (round_saeonly_expand): Use const48_operand.
6080 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
6082 * config/i386/constraints.md (Yk): Swap meaning with k.
6083 * config/i386/i386.md (movhi_internal): Change Yk to k.
6084 (movqi_internal): Ditto.
6085 (*k<logic><mode>): Ditto.
6088 (kandn<mode>): Ditto.
6089 (*<code>hi_1): Ditto.
6090 (*<code>qi_1): Ditto.
6091 (kxnor<mode>): Ditto.
6092 (kortestzhi): Ditto.
6093 (kortestchi): Ditto.
6095 (*one_cmplhi2_1): Ditto.
6096 (*one_cmplqi2_1): Ditto.
6097 * config/i386/sse.md (): Change k to Yk.
6098 (avx512f_load<mode>_mask): Ditto.
6099 (avx512f_blendm<mode>): Ditto.
6100 (avx512f_store<mode>_mask): Ditto.
6101 (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
6102 (avx512f_storedqu<mode>_mask): Ditto.
6103 (avx512f_cmp<mode>3<mask_scalar_merge_name><round_saeonly_name>):
6105 (avx512f_ucmp<mode>3<mask_scalar_merge_name>): Ditto.
6106 (avx512f_vmcmp<mode>3<round_saeonly_name>): Ditto.
6107 (avx512f_vmcmp<mode>3_mask<round_saeonly_name>): Ditto.
6108 (avx512f_maskcmp<mode>3): Ditto.
6109 (avx512f_fmadd_<mode>_mask<round_name>): Ditto.
6110 (avx512f_fmadd_<mode>_mask3<round_name>): Ditto.
6111 (avx512f_fmsub_<mode>_mask<round_name>): Ditto.
6112 (avx512f_fmsub_<mode>_mask3<round_name>): Ditto.
6113 (avx512f_fnmadd_<mode>_mask<round_name>): Ditto.
6114 (avx512f_fnmadd_<mode>_mask3<round_name>): Ditto.
6115 (avx512f_fnmsub_<mode>_mask<round_name>): Ditto.
6116 (avx512f_fnmsub_<mode>_mask3<round_name>): Ditto.
6117 (avx512f_fmaddsub_<mode>_mask<round_name>): Ditto.
6118 (avx512f_fmaddsub_<mode>_mask3<round_name>): Ditto.
6119 (avx512f_fmsubadd_<mode>_mask<round_name>): Ditto.
6120 (avx512f_fmsubadd_<mode>_mask3<round_name>): Ditto.
6121 (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
6122 (vec_extract_lo_<mode>_maskm): Ditto.
6123 (vec_extract_hi_<mode>_maskm): Ditto.
6124 (avx512f_vternlog<mode>_mask): Ditto.
6125 (avx512f_fixupimm<mode>_mask<round_saeonly_name>): Ditto.
6126 (avx512f_sfixupimm<mode>_mask<round_saeonly_name>): Ditto.
6127 (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
6128 (avx512f_<code>v8div16qi2_mask): Ditto.
6129 (avx512f_<code>v8div16qi2_mask_store): Ditto.
6130 (avx512f_eq<mode>3<mask_scalar_merge_name>_1): Ditto.
6131 (avx512f_gt<mode>3<mask_scalar_merge_name>): Ditto.
6132 (avx512f_testm<mode>3<mask_scalar_merge_name>): Ditto.
6133 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Ditto.
6134 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
6135 (*avx512pf_gatherpf<mode>df_mask): Ditto.
6136 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
6137 (*avx512pf_scatterpf<mode>df_mask): Ditto.
6138 (avx512cd_maskb_vec_dupv8di): Ditto.
6139 (avx512cd_maskw_vec_dupv16si): Ditto.
6140 (avx512f_vpermi2var<mode>3_maskz): Ditto.
6141 (avx512f_vpermi2var<mode>3_mask): Ditto.
6142 (avx512f_vpermi2var<mode>3_mask): Ditto.
6143 (avx512f_vpermt2var<mode>3_maskz): Ditto.
6144 (*avx512f_gathersi<mode>): Ditto.
6145 (*avx512f_gathersi<mode>_2): Ditto.
6146 (*avx512f_gatherdi<mode>): Ditto.
6147 (*avx512f_gatherdi<mode>_2): Ditto.
6148 (*avx512f_scattersi<mode>): Ditto.
6149 (*avx512f_scatterdi<mode>): Ditto.
6150 (avx512f_compress<mode>_mask): Ditto.
6151 (avx512f_compressstore<mode>_mask): Ditto.
6152 (avx512f_expand<mode>_mask): Ditto.
6153 * config/i386/subst.md (mask): Change k to Yk.
6154 (mask_scalar_merge): Ditto.
6157 2014-01-31 Marc Glisse <marc.glisse@inria.fr>
6159 * doc/extend.texi (Vector Extensions): Document ?: in C++.
6161 2014-01-31 Richard Biener <rguenther@suse.de>
6164 * builtins.c (fold_builtin_memory_op): Make sure to not
6165 use a floating-point mode or a boolean or enumeral type for
6168 2014-01-30 DJ Delorie <dj@redhat.com>
6170 * config/msp430/msp430.h (LIB_SPEC): Add -lcrt
6171 * config/msp430/msp430.md (msp430_refsym_need_exit): New.
6172 * config/msp430/msp430.c (msp430_expand_epilogue): Call it
6173 whenever main() has an epilogue.
6175 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6177 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Remove
6178 unused variable "field".
6179 * config/rs6000/vsx.md (vsx_mergel_<mode>): Add missing DONE.
6180 (vsx_mergeh_<mode>): Likewise.
6181 * config/rs6000/altivec.md (altivec_vmrghb): Likewise.
6182 (altivec_vmrghh): Likewise.
6183 (altivec_vmrghw): Likewise.
6184 (altivec_vmrglb): Likewise.
6185 (altivec_vmrglh): Likewise.
6186 (altivec_vmrglw): Likewise.
6187 (altivec_vspltb): Add missing uses.
6188 (altivec_vsplth): Likewise.
6189 (altivec_vspltw): Likewise.
6190 (altivec_vspltsf): Likewise.
6192 2014-01-30 Jakub Jelinek <jakub@redhat.com>
6195 * ifcvt.c (cond_exec_process_insns): Don't conditionalize
6196 frame related instructions.
6198 2014-01-30 Vladimir Makarov <vmakarov@redhat.com>
6200 PR rtl-optimization/59959
6201 * lra-constrains.c (simplify_operand_subreg): Assign NO_REGS to
6202 any reload of register whose subreg is invalid.
6204 2014-01-30 Jakub Jelinek <jakub@redhat.com>
6206 * config/i386/f16cintrin.h (_cvtsh_ss): Avoid -Wnarrowing warning.
6207 * config/i386/avx512fintrin.h (_mm512_mask_cvtusepi64_storeu_epi32):
6208 Add missing return type - void.
6210 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6212 * gcc/config/rs6000/rs6000.c (rs6000_expand_vector_init): Use
6213 gen_vsx_xxspltw_v4sf_direct instead of gen_vsx_xxspltw_v4sf;
6214 remove element index adjustment for endian (now handled in vsx.md
6216 (altivec_expand_vec_perm_const): Use
6217 gen_altivec_vsplt[bhw]_direct instead of gen_altivec_vsplt[bhw].
6218 * gcc/config/rs6000/vsx.md (UNSPEC_VSX_XXSPLTW): New unspec.
6219 (vsx_xxspltw_<mode>): Adjust element index for little endian.
6220 * gcc/config/rs6000/altivec.md (altivec_vspltb): Divide into a
6221 define_expand and a new define_insn *altivec_vspltb_internal;
6222 adjust for -maltivec=be on a little endian target.
6223 (altivec_vspltb_direct): New.
6224 (altivec_vsplth): Divide into a define_expand and a new
6225 define_insn *altivec_vsplth_internal; adjust for -maltivec=be on a
6226 little endian target.
6227 (altivec_vsplth_direct): New.
6228 (altivec_vspltw): Divide into a define_expand and a new
6229 define_insn *altivec_vspltw_internal; adjust for -maltivec=be on a
6230 little endian target.
6231 (altivec_vspltw_direct): New.
6232 (altivec_vspltsf): Divide into a define_expand and a new
6233 define_insn *altivec_vspltsf_internal; adjust for -maltivec=be on
6234 a little endian target.
6236 2014-01-30 Richard Biener <rguenther@suse.de>
6238 PR tree-optimization/59993
6239 * tree-ssa-forwprop.c (associate_pointerplus): Check we
6240 can propagate form the earlier stmt and avoid the transform
6241 when the intermediate result is needed.
6243 2014-01-30 Alangi Derick <alangiderick@gmail.com>
6245 * README.Portability: Fix typo.
6247 2014-01-30 David Holsgrove <david.holsgrove@xilinx.com>
6249 * config/microblaze/microblaze.md(cstoresf4, cbranchsf4): Replace
6250 comparison_operator with ordered_comparison_operator.
6252 2014-01-30 Nick Clifton <nickc@redhat.com>
6254 * config/mn10300/mn10300-protos.h (mn10300_store_multiple_operation_p):
6255 Rename to mn10300_store_multiple_regs.
6256 * config/mn10300/mn10300.c: Likewise.
6257 * config/mn10300/mn10300.md (store_movm): Fix typo: call
6258 store_multiple_regs.
6259 * config/mn10300/predicates.md (mn10300_store_multiple_operation):
6260 Call mn10300_store_multiple_regs.
6262 2014-01-30 Nick Clifton <nickc@redhat.com>
6263 DJ Delorie <dj@redhat.com>
6265 * config/rl78/rl78.c (register_sizes): Make the "upper half" of
6266 %fp 2 to keep registers after it properly word-aligned.
6267 (rl78_alloc_physical_registers_umul): Handle the case where both
6268 input operands are the same.
6270 2014-01-30 Richard Biener <rguenther@suse.de>
6272 PR tree-optimization/59903
6273 * tree-vect-loop.c (vect_transform_loop): Guard multiple-types
6276 2014-01-30 Jason Merrill <jason@redhat.com>
6279 * tree.c (walk_type_fields): Handle VECTOR_TYPE.
6282 * cgraphunit.c (expand_thunk): Copy volatile arg to a temporary.
6284 2014-01-30 Richard Biener <rguenther@suse.de>
6286 PR tree-optimization/59951
6287 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Skip uses in debug insns.
6289 2014-01-30 Savin Zlobec <savin.zlobec@gmail.com>
6292 * config/nios2/nios2.c (nios2_fpu_insn_asm): Fix asm output of
6293 SFmode to DFmode case.
6295 2014-01-29 DJ Delorie <dj@redhat.com>
6297 * config/msp430/msp430.opt (-minrt): New.
6298 * config/msp430/msp430.h (STARTFILE_SPEC): Link alternate runtime
6300 (ENDFILE_SPEC): Likewise.
6302 2014-01-29 Jan Hubicka <hubicka@ucw.cz>
6304 * ipa-inline-analysis.c (clobber_only_eh_bb_p): New function.
6305 (estimate_function_body_sizes): Use it.
6307 2014-01-29 Paolo Carlini <paolo.carlini@oracle.com>
6310 * dwarf2out.c (is_cxx_auto): New.
6311 (is_base_type): Use it.
6312 (gen_type_die_with_usage): Likewise.
6314 2014-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6316 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Use
6317 CODE_FOR_altivec_vmrg*_direct rather than CODE_FOR_altivec_vmrg*.
6318 * config/rs6000/vsx.md (vsx_mergel_<mode>): Adjust for
6319 -maltivec=be with LE targets.
6320 (vsx_mergeh_<mode>): Likewise.
6321 * config/rs6000/altivec.md (UNSPEC_VMRG[HL]_DIRECT): New unspecs.
6322 (mulv8hi3): Use gen_altivec_vmrg[hl]w_direct.
6323 (altivec_vmrghb): Replace with define_expand and new
6324 *altivec_vmrghb_internal insn; adjust for -maltivec=be with LE targets.
6325 (altivec_vmrghb_direct): New define_insn.
6326 (altivec_vmrghh): Replace with define_expand and new
6327 *altivec_vmrghh_internal insn; adjust for -maltivec=be with LE targets.
6328 (altivec_vmrghh_direct): New define_insn.
6329 (altivec_vmrghw): Replace with define_expand and new
6330 *altivec_vmrghw_internal insn; adjust for -maltivec=be with LE targets.
6331 (altivec_vmrghw_direct): New define_insn.
6332 (*altivec_vmrghsf): Adjust for endianness.
6333 (altivec_vmrglb): Replace with define_expand and new
6334 *altivec_vmrglb_internal insn; adjust for -maltivec=be with LE targets.
6335 (altivec_vmrglb_direct): New define_insn.
6336 (altivec_vmrglh): Replace with define_expand and new
6337 *altivec_vmrglh_internal insn; adjust for -maltivec=be with LE targets.
6338 (altivec_vmrglh_direct): New define_insn.
6339 (altivec_vmrglw): Replace with define_expand and new
6340 *altivec_vmrglw_internal insn; adjust for -maltivec=be with LE targets.
6341 (altivec_vmrglw_direct): New define_insn.
6342 (*altivec_vmrglsf): Adjust for endianness.
6343 (vec_widen_umult_hi_v16qi): Use gen_altivec_vmrghh_direct.
6344 (vec_widen_umult_lo_v16qi): Use gen_altivec_vmrglh_direct.
6345 (vec_widen_smult_hi_v16qi): Use gen_altivec_vmrghh_direct.
6346 (vec_widen_smult_lo_v16qi): Use gen_altivec_vmrglh_direct.
6347 (vec_widen_umult_hi_v8hi): Use gen_altivec_vmrghw_direct.
6348 (vec_widen_umult_lo_v8hi): Use gen_altivec_vmrglw_direct.
6349 (vec_widen_smult_hi_v8hi): Use gen_altivec_vmrghw_direct.
6350 (vec_widen_smult_lo_v8hi): Use gen_altivec_vmrglw_direct.
6352 2014-01-29 Marcus Shawcroft <marcus.shawcroft@arm.com>
6354 * config/aarch64/aarch64.c (aarch64_expand_mov_immediate)
6355 (aarch64_legitimate_address_p, aarch64_class_max_nregs): Adjust
6358 2014-01-29 Richard Biener <rguenther@suse.de>
6360 PR tree-optimization/58742
6361 * tree-ssa-forwprop.c (associate_pointerplus): Rename to
6362 associate_pointerplus_align.
6363 (associate_pointerplus_diff): New function.
6364 (associate_pointerplus): Likewise. Call associate_pointerplus_align
6365 and associate_pointerplus_diff.
6367 2014-01-29 Richard Biener <rguenther@suse.de>
6369 * lto-streamer.h (LTO_major_version): Bump to 3.
6370 (LTO_minor_version): Reset to 0.
6372 2014-01-29 Renlin Li <Renlin.Li@arm.com>
6374 * config/arm/arm-arches.def (ARM_ARCH): Add armv7ve arch.
6375 * config/arm/arm.c (FL_FOR_ARCH7VE): New.
6376 (arm_file_start): Generate correct asm header for armv7ve.
6377 * config/arm/bpabi.h: Add multilib support for armv7ve.
6378 * config/arm/driver-arm.c: Change the architectures of cortex-a7
6379 and cortex-a15 to armv7ve.
6380 * config/arm/t-aprofile: Add multilib support for armv7ve.
6381 * doc/invoke.texi: Document -march=armv7ve.
6383 2014-01-29 Richard Biener <rguenther@suse.de>
6385 PR tree-optimization/58742
6386 * tree-ssa-forwprop.c (associate_plusminus): Return true
6387 if we changed sth, defer EH cleanup to ...
6388 (ssa_forward_propagate_and_combine): ... here. Call simplify_mult.
6389 (simplify_mult): New function.
6391 2014-01-29 Jakub Jelinek <jakub@redhat.com>
6394 PR tree-optimization/59920
6395 * tree.c (build_common_builtin_nodes): Remove
6396 __builtin_setjmp_dispatcher initialization.
6397 * omp-low.h (make_gimple_omp_edges): Add a new int * argument.
6398 * profile.c (branch_prob): Use gsi_start_nondebug_after_labels_bb
6399 instead of gsi_after_labels + manually skipping debug stmts.
6400 Don't ignore bbs with BUILT_IN_SETJMP_DISPATCHER, instead
6401 ignore bbs with IFN_ABNORMAL_DISPATCHER.
6402 * tree-inline.c (copy_edges_for_bb): Remove
6403 can_make_abnormal_goto argument, instead add abnormal_goto_dest
6404 argument. Ignore computed_goto_p stmts. Don't call
6405 make_abnormal_goto_edges. If a call might need abnormal edges
6406 for non-local gotos, see if it already has an edge to
6407 IFN_ABNORMAL_DISPATCHER or if it is IFN_ABNORMAL_DISPATCHER
6408 with true argument, don't do anything then, otherwise add
6409 EDGE_ABNORMAL from the call's bb to abnormal_goto_dest.
6410 (copy_cfg_body): Compute abnormal_goto_dest, adjust copy_edges_for_bb
6412 * gimple-low.c (struct lower_data): Remove calls_builtin_setjmp.
6413 (lower_function_body): Don't emit __builtin_setjmp_dispatcher.
6414 (lower_stmt): Don't set data->calls_builtin_setjmp.
6415 (lower_builtin_setjmp): Adjust comment.
6416 * builtins.def (BUILT_IN_SETJMP_DISPATCHER): Remove.
6417 * tree-cfg.c (found_computed_goto): Remove.
6418 (factor_computed_gotos): Remove.
6419 (make_goto_expr_edges): Return bool, true for computed gotos.
6420 Don't call make_abnormal_goto_edges.
6421 (build_gimple_cfg): Don't set found_computed_goto, don't call
6422 factor_computed_gotos.
6423 (computed_goto_p): No longer static.
6424 (make_blocks): Don't set found_computed_goto.
6425 (get_abnormal_succ_dispatcher, handle_abnormal_edges): New functions.
6426 (make_edges): If make_goto_expr_edges returns true, push bb
6427 into ab_edge_goto vector, for stmt_can_make_abnormal_goto calls
6428 instead of calling make_abnormal_goto_edges push bb into ab_edge_call
6429 vector. Record mapping between bbs and OpenMP regions if there
6430 are any, adjust make_gimple_omp_edges caller. Call
6431 handle_abnormal_edges.
6432 (make_abnormal_goto_edges): Remove.
6433 * tree-cfg.h (make_abnormal_goto_edges): Remove.
6434 (computed_goto_p, get_abnormal_succ_dispatcher): New prototypes.
6435 * internal-fn.c (expand_ABNORMAL_DISPATCHER): New function.
6436 * builtins.c (expand_builtin): Don't handle BUILT_IN_SETJMP_DISPATCHER.
6437 * internal-fn.def (ABNORMAL_DISPATCHER): New.
6438 * omp-low.c (make_gimple_omp_edges): Add region_idx argument, when
6439 filling *region also set *region_idx to (*region)->entry->index.
6442 * read-rtl.c (read_rtx_code): Clear all of RTX_CODE_SIZE (code).
6443 For REGs set ORIGINAL_REGNO.
6445 2014-01-29 Bingfeng Mei <bmei@broadcom.com>
6447 * doc/md.texi: Mention that a target shouldn't implement
6448 vec_widen_(s|u)mul_even/odd pair if it is less efficient
6451 2014-01-29 Jakub Jelinek <jakub@redhat.com>
6453 PR tree-optimization/59594
6454 * tree-vect-data-refs.c (vect_analyze_data_ref_accesses): Sort
6455 a copy of the datarefs vector rather than the vector itself.
6457 2014-01-28 Jason Merrill <jason@redhat.com>
6460 * dwarf2out.c (auto_die): New static.
6461 (gen_type_die_with_usage): Handle C++1y 'auto'.
6462 (gen_subprogram_die): If in-class DIE had 'auto', emit type again
6465 2014-01-28 H.J. Lu <hongjiu.lu@intel.com>
6468 * config/i386/gnu-user64.h (SPEC_32): Add "m16|" to "m32".
6469 (SPEC_X32): Likewise.
6470 (SPEC_64): Likewise.
6471 * config/i386/i386.c (ix86_option_override_internal): Turn off
6472 OPTION_MASK_ISA_64BIT, OPTION_MASK_ABI_X32 and OPTION_MASK_ABI_64
6474 (x86_file_start): Output .code16gcc for TARGET_16BIT.
6475 * config/i386/i386.h (TARGET_16BIT): New macro.
6476 (TARGET_16BIT_P): Likewise.
6477 * config/i386/i386.opt: Add m16.
6478 * doc/invoke.texi: Document -m16.
6480 2014-01-28 Jakub Jelinek <jakub@redhat.com>
6482 PR preprocessor/59935
6483 * input.c (location_get_source_line): Bail out on when line number
6484 is zero, and test the return value of lookup_or_add_file_to_cache_tab.
6486 2014-01-28 Richard Biener <rguenther@suse.de>
6488 PR tree-optimization/58742
6489 * tree-ssa-forwprop.c (associate_plusminus): Handle
6490 pointer subtraction of the form (T)(P + A) - (T)P.
6492 2014-01-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6494 * config/arm/arm.c (arm_new_rtx_costs): Remove useless statement
6497 2014-01-28 Richard Biener <rguenther@suse.de>
6500 2014-01-28 Richard Biener <rguenther@suse.de>
6502 PR rtl-optimization/45364
6503 PR rtl-optimization/59890
6504 * var-tracking.c (local_get_addr_clear_given_value): Handle
6505 already cleared slot.
6506 (val_reset): Handle not allocated local_get_addr_cache.
6507 (vt_find_locations): Use post-order on the inverted CFG.
6509 2014-01-28 Richard Biener <rguenther@suse.de>
6511 * tree-data-ref.h (ddr_is_anti_dependent, ddrs_have_anti_deps): Remove.
6513 2014-01-28 Richard Biener <rguenther@suse.de>
6515 PR rtl-optimization/45364
6516 PR rtl-optimization/59890
6517 * var-tracking.c (local_get_addr_clear_given_value): Handle
6518 already cleared slot.
6519 (val_reset): Handle not allocated local_get_addr_cache.
6520 (vt_find_locations): Use post-order on the inverted CFG.
6522 2014-01-28 Alan Modra <amodra@gmail.com>
6524 * Makefile.in (BUILD_CPPFLAGS): Do not use ALL_CPPFLAGS.
6525 * configure.ac <recursive call for build != host>: Define
6526 GENERATOR_FILE. Comment. Use CXX_FOR_BUILD, CXXFLAGS_FOR_BUILD
6527 and LD_FOR_BUILD too.
6528 * configure: Regenerate.
6530 2014-01-27 Allan Sandfeld Jensen <sandfeld@kde.org>
6532 * config/i386/i386.c (get_builtin_code_for_version): Separate
6533 Westmere from Nehalem, Ivy Bridge from Sandy Bridge and
6534 Broadwell from Haswell.
6536 2014-01-27 Steve Ellcey <sellcey@mips.com>
6538 * common/config/mips/mips-common.c (TARGET_DEFAULT_TARGET_FLAGS):
6539 Remove TARGET_FP_EXCEPTIONS_DEFAULT and MASK_FUSED_MADD.
6540 * config/mips/mips.c (mips_option_override): Change setting
6542 * config/mips/mips.h (TARGET_FP_EXCEPTIONS_DEFAULT): Remove.
6543 * config/mips/mips.opt (DSP, DSPR2, FP_EXCEPTIONS, FUSED_MADD, MIPS3D):
6544 Change from Mask to Var.
6546 2014-01-27 Jeff Law <law@redhat.com>
6548 * ipa-inline.c (inline_small_functions): Fix typo.
6550 2014-01-27 Ilya Tocar <ilya.tocar@intel.com>
6552 * config/i386/avx512fintrin.h (_mm512_mask_cvtepi32_storeu_epi8): New.
6553 (_mm512_mask_cvtsepi32_storeu_epi8): Ditto.
6554 (_mm512_mask_cvtusepi32_storeu_epi8): Ditto.
6555 (_mm512_mask_cvtepi32_storeu_epi16): Ditto.
6556 (_mm512_mask_cvtsepi32_storeu_epi16): Ditto.
6557 (_mm512_mask_cvtusepi32_storeu_epi16): Ditto.
6558 (_mm512_mask_cvtepi64_storeu_epi32): Ditto.
6559 (_mm512_mask_cvtsepi64_storeu_epi32): Ditto.
6560 (_mm512_mask_cvtusepi64_storeu_epi32): Ditto.
6561 (_mm512_mask_cvtepi64_storeu_epi16): Ditto.
6562 (_mm512_mask_cvtsepi64_storeu_epi16): Ditto.
6563 (_mm512_mask_cvtusepi64_storeu_epi16): Ditto.
6564 (_mm512_mask_cvtepi64_storeu_epi8): Ditto.
6565 (_mm512_mask_cvtsepi64_storeu_epi8): Ditto.
6566 (_mm512_mask_cvtusepi64_storeu_epi8): Ditto.
6567 (_mm512_storeu_epi64): Ditto.
6568 (_mm512_cmpge_epi32_mask): Ditto.
6569 (_mm512_cmpge_epu32_mask): Ditto.
6570 (_mm512_cmpge_epi64_mask): Ditto.
6571 (_mm512_cmpge_epu64_mask): Ditto.
6572 (_mm512_cmple_epi32_mask): Ditto.
6573 (_mm512_cmple_epu32_mask): Ditto.
6574 (_mm512_cmple_epi64_mask): Ditto.
6575 (_mm512_cmple_epu64_mask): Ditto.
6576 (_mm512_cmplt_epi32_mask): Ditto.
6577 (_mm512_cmplt_epu32_mask): Ditto.
6578 (_mm512_cmplt_epi64_mask): Ditto.
6579 (_mm512_cmplt_epu64_mask): Ditto.
6580 (_mm512_cmpneq_epi32_mask): Ditto.
6581 (_mm512_cmpneq_epu32_mask): Ditto.
6582 (_mm512_cmpneq_epi64_mask): Ditto.
6583 (_mm512_cmpneq_epu64_mask): Ditto.
6584 (_mm512_expand_pd): Ditto.
6585 (_mm512_expand_ps): Ditto.
6586 * config/i386/i386-builtin-types.def: Add PV16QI, PV16QI, PV16HI,
6587 VOID_PV8SI_V8DI_QI, VOID_PV8HI_V8DI_QI, VOID_PV16QI_V8DI_QI,
6588 VOID_PV16QI_V16SI_HI, VOID_PV16HI_V16SI_HI.
6589 * config/i386/i386.c (ix86_builtins): Add
6590 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK,
6591 IX86_BUILTIN_PMOVDB512_MEM, IX86_BUILTIN_PMOVDW512_MEM,
6592 IX86_BUILTIN_PMOVQB512_MEM, IX86_BUILTIN_PMOVQD512_MEM,
6593 IX86_BUILTIN_PMOVQW512_MEM, IX86_BUILTIN_PMOVSDB512_MEM,
6594 IX86_BUILTIN_PMOVSDW512_MEM, IX86_BUILTIN_PMOVSQB512_MEM,
6595 IX86_BUILTIN_PMOVSQD512_MEM, IX86_BUILTIN_PMOVSQW512_MEM,
6596 IX86_BUILTIN_PMOVUSDB512_MEM, IX86_BUILTIN_PMOVUSDW512_MEM,
6597 IX86_BUILTIN_PMOVUSQB512_MEM, IX86_BUILTIN_PMOVUSQD512_MEM,
6598 IX86_BUILTIN_PMOVUSQW512_MEM.
6599 (bdesc_special_args): Add __builtin_ia32_pmovusqd512mem_mask,
6600 __builtin_ia32_pmovsqd512mem_mask,
6601 __builtin_ia32_pmovqd512mem_mask,
6602 __builtin_ia32_pmovusqw512mem_mask,
6603 __builtin_ia32_pmovsqw512mem_mask,
6604 __builtin_ia32_pmovqw512mem_mask,
6605 __builtin_ia32_pmovusdw512mem_mask,
6606 __builtin_ia32_pmovsdw512mem_mask,
6607 __builtin_ia32_pmovdw512mem_mask,
6608 __builtin_ia32_pmovqb512mem_mask,
6609 __builtin_ia32_pmovusqb512mem_mask,
6610 __builtin_ia32_pmovsqb512mem_mask,
6611 __builtin_ia32_pmovusdb512mem_mask,
6612 __builtin_ia32_pmovsdb512mem_mask,
6613 __builtin_ia32_pmovdb512mem_mask.
6614 (bdesc_args): Add __builtin_ia32_expanddf512,
6615 __builtin_ia32_expandsf512.
6616 (ix86_expand_special_args_builtin): Handle VOID_FTYPE_PV8SI_V8DI_QI,
6617 VOID_FTYPE_PV8HI_V8DI_QI, VOID_FTYPE_PV16HI_V16SI_HI,
6618 VOID_FTYPE_PV16QI_V8DI_QI, VOID_FTYPE_PV16QI_V16SI_HI.
6619 * config/i386/sse.md (unspec): Add UNSPEC_EXPAND_NOMASK.
6620 (avx512f_<code><pmov_src_lower><mode>2_mask_store): New.
6621 (*avx512f_<code>v8div16qi2_store_mask): Renamed to ...
6622 (avx512f_<code>v8div16qi2_mask_store): This.
6623 (avx512f_expand<mode>): New.
6625 2014-01-27 Kirill Yukhin <kirill.yukhin@intel.com>
6627 * config/i386/avx512pfintrin.h (_mm512_mask_prefetch_i32gather_pd):
6629 (_mm512_mask_prefetch_i64gather_pd): Ditto.
6630 (_mm512_prefetch_i32scatter_pd): Ditto.
6631 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
6632 (_mm512_prefetch_i64scatter_pd): Ditto.
6633 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
6634 (_mm512_mask_prefetch_i32gather_ps): Fix operand type.
6635 (_mm512_mask_prefetch_i64gather_ps): Ditto.
6636 (_mm512_prefetch_i32scatter_ps): Ditto.
6637 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
6638 (_mm512_prefetch_i64scatter_ps): Ditto.
6639 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
6640 * config/i386/i386-builtin-types.def: Define
6641 VOID_FTYPE_QI_V8SI_PCINT64_INT_INT
6642 and VOID_FTYPE_QI_V8DI_PCINT64_INT_INT.
6643 * config/i386/i386.c (ix86_builtins): Define IX86_BUILTIN_GATHERPFQPD,
6644 IX86_BUILTIN_GATHERPFDPD, IX86_BUILTIN_SCATTERPFDPD,
6645 IX86_BUILTIN_SCATTERPFQPD.
6646 (ix86_init_mmx_sse_builtins): Define __builtin_ia32_gatherpfdpd,
6647 __builtin_ia32_gatherpfdps, __builtin_ia32_gatherpfqpd,
6648 __builtin_ia32_gatherpfqps, __builtin_ia32_scatterpfdpd,
6649 __builtin_ia32_scatterpfdps, __builtin_ia32_scatterpfqpd,
6650 __builtin_ia32_scatterpfqps.
6651 (ix86_expand_builtin): Expand new built-ins.
6652 * config/i386/sse.md (avx512pf_gatherpf<mode>): Add SF suffix,
6653 fix memory access data type.
6654 (*avx512pf_gatherpf<mode>_mask): Ditto.
6655 (*avx512pf_gatherpf<mode>): Ditto.
6656 (avx512pf_scatterpf<mode>): Ditto.
6657 (*avx512pf_scatterpf<mode>_mask): Ditto.
6658 (*avx512pf_scatterpf<mode>): Ditto.
6659 (GATHER_SCATTER_SF_MEM_MODE): New.
6660 (avx512pf_gatherpf<mode>df): Ditto.
6661 (*avx512pf_gatherpf<mode>df_mask): Ditto.
6662 (*avx512pf_scatterpf<mode>df): Ditto.
6664 2014-01-27 Jakub Jelinek <jakub@redhat.com>
6667 * expmed.h (expmed_mode_index): Rework so that analysis and optimziers
6668 know when the MODE_PARTIAL_INT and MODE_VECTOR_INT cases can never be
6671 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
6673 * common/config/arm/arm-common.c
6674 (arm_rewrite_mcpu): Handle multiple names.
6676 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
6678 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
6680 * gimple-builder.h (create_gimple_tmp): Delete.
6682 2014-01-27 Christian Bruel <christian.bruel@st.com>
6684 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Fix remaining bytes after
6687 2014-01-26 John David Anglin <danglin@gcc.gnu.org>
6689 * config/pa/pa.md (call): Generate indirect long calls to non-local
6690 functions when outputing 32-bit code.
6691 (call_value): Likewise except for special call to buggy powf function.
6693 * config/pa/pa.c (pa_attr_length_indirect_call): Adjust length of
6694 portable runtime and PIC indirect calls.
6695 (pa_output_indirect_call): Remove unnecessary nop from portable runtime
6696 and PIC call sequences. Use ldo instead of blr to set return register
6697 in PIC call sequence.
6699 2014-01-25 Walter Lee <walt@tilera.com>
6701 * config/tilegx/sync.md (atomic_fetch_sub): Fix negation and
6702 avoid clobbering a live register.
6704 2014-01-25 Walter Lee <walt@tilera.com>
6706 * config/tilegx/tilegx-c.c (tilegx_cpu_cpp_builtins):
6707 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2}.
6708 * config/tilegx/tilepro-c.c (tilepro_cpu_cpp_builtins):
6709 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2,4,8}.
6711 2014-01-25 Walter Lee <walt@tilera.com>
6713 * config/tilegx/tilegx.c (tilegx_function_arg): Start 16-byte
6714 arguments on even registers.
6715 (tilegx_gimplify_va_arg_expr): Align 16-byte var args to
6717 * config/tilegx/tilegx.h (STACK_BOUNDARY): Change to 16 bytes.
6718 (BIGGEST_ALIGNMENT): Ditto.
6719 (BIGGEST_FIELD_ALIGNMENT): Ditto.
6721 2014-01-25 Walter Lee <walt@tilera.com>
6723 * config/tilegx/tilegx.c (tilegx_gen_bundles): Delete barrier
6724 insns before bundling.
6725 * config/tilegx/tilegx.md (tile_network_barrier): Update comment.
6727 2014-01-25 Walter Lee <walt@tilera.com>
6729 * config/tilegx/tilegx.c (tilegx_expand_builtin): Set
6730 PREFETCH_SCHEDULE_BARRIER_P to true for prefetches.
6731 * config/tilepro/tilepro.c (tilepro_expand_builtin): Ditto.
6733 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
6735 * config/mips/constraints.md (kl): Delete.
6736 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Turn into
6737 define expands, using...
6738 (divmod<mode>4_mips16, udivmod<mode>4_mips16): ...these new
6739 instructions for MIPS16.
6740 (*divmod<mode>4, *udivmod<mode>4): New patterns, taken from the
6741 non-MIPS16 version of the old divmod<mode>4 and udivmod<mode>4.
6743 2014-01-25 Walter Lee <walt@tilera.com>
6745 * config/tilepro/tilepro.md (ctzdi2): Use register_operand predicate.
6749 2014-01-25 Walter Lee <walt@tilera.com>
6751 * config/tilegx/tilegx.c (tilegx_expand_to_rtl_hook): New.
6752 (TARGET_EXPAND_TO_RTL_HOOK): Define.
6754 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
6756 * rtlanal.c (canonicalize_condition): Split out duplicated mode check.
6759 2014-01-25 Jakub Jelinek <jakub@redhat.com>
6761 * print-rtl.c (in_call_function_usage): New var.
6762 (print_rtx): When in CALL_INSN_FUNCTION_USAGE, always print
6763 EXPR_LIST mode as mode and not as reg note name.
6766 * cfgloopmanip.c (copy_loop_info): If
6767 loop->warned_aggressive_loop_optimizations, make sure
6768 the flag is set in target loop too.
6770 2014-01-24 Balaji V. Iyer <balaji.v.iyer@intel.com>
6772 * builtins.c (is_builtin_name): Renamed flag_enable_cilkplus to
6774 * builtins.def: Likewise.
6775 * cilk.h (fn_contains_cilk_spawn_p): Likewise.
6776 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Likewise.
6777 * ira.c (ira_setup_eliminable_regset): Likewise.
6778 * omp-low.c (gate_expand_omp): Likewise.
6779 (execute_lower_omp): Likewise.
6780 (diagnose_sb_0): Likewise.
6781 (gate_diagnose_omp_blocks): Likewise.
6782 (simd_clone_clauses_extract): Likewise.
6785 2014-01-24 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6787 * config/rs6000/rs6000.c (rs6000_expand_vec_perm_const_1): Remove
6788 correction for little endian...
6789 * config/rs6000/vsx.md (vsx_xxpermdi2_<mode>_1): ...and move it to
6792 2014-01-24 Jeff Law <law@redhat.com>
6794 PR tree-optimization/59919
6795 * tree-vrp.c (find_assert_locations_1): Do not register asserts
6796 for non-returning calls.
6798 2014-01-24 James Greenhalgh <james.greenhalgh@arm.com>
6800 * common/config/aarch64/aarch64-common.c
6801 (aarch64_rewrite_mcpu): Handle multiple names.
6802 * config/aarch64/aarch64.h
6803 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
6805 2014-01-24 Dodji Seketeli <dodji@redhat.com>
6807 * input.c (add_file_to_cache_tab): Handle the case where fopen
6810 2014-01-23 H.J. Lu <hongjiu.lu@intel.com>
6813 * config/i386/i386.md (pushsf splitter): Get stack adjustment
6814 from push operand if code of push isn't PRE_DEC.
6816 2014-01-23 Michael Meissner <meissner@linux.vnet.ibm.com>
6819 * doc/invoke.texi (RS/6000 and PowerPC Options): Document
6820 -mquad-memory-atomic. Update -mquad-memory documentation to say
6821 it is only used for non-atomic loads/stores.
6823 * config/rs6000/predicates.md (quad_int_reg_operand): Allow either
6824 -mquad-memory or -mquad-memory-atomic switches.
6826 * config/rs6000/rs6000-cpus.def (ISA_2_7_MASKS_SERVER): Add
6827 -mquad-memory-atomic to ISA 2.07 support.
6829 * config/rs6000/rs6000.opt (-mquad-memory-atomic): Add new switch
6830 to separate support of normal quad word memory operations (ldq, stq)
6831 from the atomic quad word memory operations.
6833 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
6834 support to separate non-atomic quad word operations from atomic
6835 quad word operations. Disable non-atomic quad word operations in
6836 little endian mode so that we don't have to swap words after the
6837 load and before the store.
6838 (quad_load_store_p): Add comment about atomic quad word support.
6839 (rs6000_opt_masks): Add -mquad-memory-atomic to the list of
6840 options printed with -mdebug=reg.
6842 * config/rs6000/rs6000.h (TARGET_SYNC_TI): Use
6843 -mquad-memory-atomic as the test for whether we have quad word
6844 atomic instructions.
6845 (TARGET_SYNC_HI_QI): If either -mquad-memory-atomic, -mquad-memory,
6846 or -mp8-vector are used, allow byte/half-word atomic operations.
6848 * config/rs6000/sync.md (load_lockedti): Insure that the address
6849 is a proper indexed or indirect address for the lqarx instruction.
6850 On little endian systems, swap the hi/lo registers after the lqarx
6852 (load_lockedpti): Use indexed_or_indirect_operand predicate to
6853 insure the address is valid for the lqarx instruction.
6854 (store_conditionalti): Insure that the address is a proper indexed
6855 or indirect address for the stqcrx. instruction. On little endian
6856 systems, swap the hi/lo registers before doing the stqcrx.
6858 (store_conditionalpti): Use indexed_or_indirect_operand predicate to
6859 insure the address is valid for the stqcrx. instruction.
6861 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
6862 Define __QUAD_MEMORY__ and __QUAD_MEMORY_ATOMIC__ based on what
6863 type of quad memory support is available.
6865 2014-01-23 Vladimir Makarov <vmakarov@redhat.com>
6868 * lra-constraints.c (simplify_operand_subreg): Spill pseudo if
6869 there is a danger of looping.
6871 2014-01-23 Pat Haugen <pthaugen@us.ibm.com>
6873 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
6874 force flag_ira_loop_pressure if set via command line.
6876 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
6878 * config/aarch64/aarch64-simd-builtins.def (ashr): DI mode removed.
6879 (ashr_simd): New builtin handling DI mode.
6880 * config/aarch64/aarch64-simd.md (aarch64_ashr_simddi): New pattern.
6881 (aarch64_sshr_simddi): New match pattern.
6882 * config/aarch64/arm_neon.h (vshr_n_s32): Builtin call modified.
6883 (vshrd_n_s64): Likewise.
6884 * config/aarch64/predicates.md (aarch64_shift_imm64_di): New predicate.
6886 2014-01-23 Nick Clifton <nickc@redhat.com>
6888 * config/msp430/msp430.h (ASM_SPEC): Pass the -mcpu as -mcpu.
6889 (LIB_SPEC): Drop use of memory.ld and peripherals.ld scripts in
6890 favour of mcu specific scripts.
6891 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more matches for
6894 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
6895 Alex Velenko <Alex.Velenko@arm.com>
6897 * config/aarch64/arm_neon.h (vaddv_s8): __LANE0 cleanup.
6898 (vaddv_s16): Likewise.
6899 (vaddv_s32): Likewise.
6900 (vaddv_u8): Likewise.
6901 (vaddv_u16): Likewise.
6902 (vaddv_u32): Likewise.
6903 (vaddvq_s8): Likewise.
6904 (vaddvq_s16): Likewise.
6905 (vaddvq_s32): Likewise.
6906 (vaddvq_s64): Likewise.
6907 (vaddvq_u8): Likewise.
6908 (vaddvq_u16): Likewise.
6909 (vaddvq_u32): Likewise.
6910 (vaddvq_u64): Likewise.
6911 (vaddv_f32): Likewise.
6912 (vaddvq_f32): Likewise.
6913 (vaddvq_f64): Likewise.
6914 (vmaxv_f32): Likewise.
6915 (vmaxv_s8): Likewise.
6916 (vmaxv_s16): Likewise.
6917 (vmaxv_s32): Likewise.
6918 (vmaxv_u8): Likewise.
6919 (vmaxv_u16): Likewise.
6920 (vmaxv_u32): Likewise.
6921 (vmaxvq_f32): Likewise.
6922 (vmaxvq_f64): Likewise.
6923 (vmaxvq_s8): Likewise.
6924 (vmaxvq_s16): Likewise.
6925 (vmaxvq_s32): Likewise.
6926 (vmaxvq_u8): Likewise.
6927 (vmaxvq_u16): Likewise.
6928 (vmaxvq_u32): Likewise.
6929 (vmaxnmv_f32): Likewise.
6930 (vmaxnmvq_f32): Likewise.
6931 (vmaxnmvq_f64): Likewise.
6932 (vminv_f32): Likewise.
6933 (vminv_s8): Likewise.
6934 (vminv_s16): Likewise.
6935 (vminv_s32): Likewise.
6936 (vminv_u8): Likewise.
6937 (vminv_u16): Likewise.
6938 (vminv_u32): Likewise.
6939 (vminvq_f32): Likewise.
6940 (vminvq_f64): Likewise.
6941 (vminvq_s8): Likewise.
6942 (vminvq_s16): Likewise.
6943 (vminvq_s32): Likewise.
6944 (vminvq_u8): Likewise.
6945 (vminvq_u16): Likewise.
6946 (vminvq_u32): Likewise.
6947 (vminnmv_f32): Likewise.
6948 (vminnmvq_f32): Likewise.
6949 (vminnmvq_f64): Likewise.
6951 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
6953 * config/aarch64/aarch64-simd.md
6954 (aarch64_dup_lane<mode>): Correct lane number on big-endian.
6955 (aarch64_dup_lane_<vswap_widthi_name><mode>): Likewise.
6956 (*aarch64_mul3_elt<mode>): Likewise.
6957 (*aarch64_mul3_elt<vswap_width_name><mode>): Likewise.
6958 (*aarch64_mul3_elt_to_64v2df): Likewise.
6959 (*aarch64_mla_elt<mode>): Likewise.
6960 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
6961 (*aarch64_mls_elt<mode>): Likewise.
6962 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
6963 (*aarch64_fma4_elt<mode>): Likewise.
6964 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
6965 (*aarch64_fma4_elt_to_64v2df): Likewise.
6966 (*aarch64_fnma4_elt<mode>): Likewise.
6967 (*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
6968 (*aarch64_fnma4_elt_to_64v2df): Likewise.
6969 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
6970 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
6971 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
6972 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
6973 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
6974 (aarch64_sqdmull_lane<mode>_internal): Likewise.
6975 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
6977 2013-01-23 Alex Velenko <Alex.Velenko@arm.com>
6979 * config/aarch64/aarch64-simd.md
6980 (aarch64_be_checked_get_lane<mode>): New define_expand.
6981 * config/aarch64/aarch64-simd-builtins.def
6982 (BUILTIN_VALL (GETLANE, be_checked_get_lane, 0)):
6983 New builtin definition.
6984 * config/aarch64/arm_neon.h: (__aarch64_vget_lane_any):
6985 Use new safe be builtin.
6987 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
6989 * config/aarch64/aarch64-simd.md (aarch64_be_ld1<mode>):
6991 (aarch64_be_st1<mode>): Likewise.
6992 (aarch_ld1<VALL:mode>): Define_expand modified.
6993 (aarch_st1<VALL:mode>): Likewise.
6994 * config/aarch64/aarch64.md (UNSPEC_LD1): New unspec definition.
6995 (UNSPEC_ST1): Likewise.
6997 2014-01-23 David Holsgrove <david.holsgrove@xilinx.com>
6999 * config/microblaze/microblaze.md: Add trap insn and attribute
7001 2014-01-23 Dodji Seketeli <dodji@redhat.com>
7003 PR preprocessor/58580
7004 * input.h (location_get_source_line): Take an additional line_size
7006 (void diagnostics_file_cache_fini): Declare new function.
7007 * input.c (struct fcache): New type.
7008 (fcache_tab_size, fcache_buffer_size, fcache_line_record_size):
7009 New static constants.
7010 (diagnostic_file_cache_init, total_lines_num)
7011 (lookup_file_in_cache_tab, evicted_cache_tab_entry)
7012 (add_file_to_cache_tab, lookup_or_add_file_to_cache_tab)
7013 (needs_read, needs_grow, maybe_grow, read_data, maybe_read_data)
7014 (get_next_line, read_next_line, goto_next_line, read_line_num):
7015 New static function definitions.
7016 (diagnostic_file_cache_fini): New function.
7017 (location_get_source_line): Take an additional output line_len
7018 parameter. Re-write using lookup_or_add_file_to_cache_tab and
7020 * diagnostic.c (diagnostic_finish): Call
7021 diagnostic_file_cache_fini.
7022 (adjust_line): Take an additional input parameter for the length
7023 of the line, rather than calculating it with strlen.
7024 (diagnostic_show_locus): Adjust the use of
7025 location_get_source_line and adjust_line with respect to their new
7026 signature. While displaying a line now, do not stop at the first
7027 null byte. Rather, display the zero byte as a space and keep
7028 going until we reach the size of the line.
7029 * Makefile.in: Add vec.o to OBJS-libcommon
7031 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
7032 Ilya Tocar <ilya.tocar@intel.com>
7034 * config/i386/avx512fintrin.h (_mm512_kmov): New.
7035 * config/i386/i386.c (IX86_BUILTIN_KMOV16): Ditto.
7036 (__builtin_ia32_kmov16): Ditto.
7037 * config/i386/i386.md (UNSPEC_KMOV): New.
7040 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
7042 * config/i386/avx512fintrin.h (_mm512_loadu_si512): Rename.
7043 (_mm512_storeu_si512): Ditto.
7045 2014-01-23 Richard Sandiford <rdsandiford@googlemail.com>
7048 * rtl.h (get_referenced_operands): Declare.
7049 * recog.c (get_referenced_operands): New function.
7050 * config/mips/mips.c (mips_reorg_process_insns): Check which asm
7051 operands have been referenced when recording LO_SUM references.
7053 2014-01-22 David Holsgrove <david.holsgrove@xilinx.com>
7055 * config/microblaze/microblaze.md: Correct bswaphi2 insn.
7057 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
7059 * config/i386/x86-tune.def (X86_TUNE_ACCUMULATE_OUTGOING_ARGS):
7060 Enable for generic and recent AMD targets.
7062 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
7064 * combine-stack-adj.c (combine_stack_adjustments_for_block): Remove
7065 ARG_SIZE note when adjustment was eliminated.
7067 2014-01-22 Jeff Law <law@redhat.com>
7069 PR tree-optimization/59597
7070 * tree-ssa-threadupdate.c (dump_jump_thread_path): Move to earlier
7071 in file. Accept new argument REGISTERING and use it to modify
7072 dump output appropriately.
7073 (register_jump_thread): Corresponding changes.
7074 (mark_threaded_blocks): Reinstate code to cancel unprofitable
7075 thread paths involving joiner blocks. Add code to dump cancelled
7076 jump threading paths.
7078 2014-01-22 Vladimir Makarov <vmakarov@redhat.com>
7080 PR rtl-optimization/59477
7081 * lra-constraints.c (inherit_in_ebb): Process call for living hard
7082 regs. Update reloads_num and potential_reload_hard_regs for all insns.
7084 2014-01-22 Tom Tromey <tromey@redhat.com>
7086 * config/i386/i386-interix.h (i386_pe_unique_section): Don't use
7088 * config/cr16/cr16-protos.h (notice_update_cc): Don't use PARAMS.
7090 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
7092 PR rtl-optimization/59896
7093 * lra-constraints.c (process_alt_operands): Check unused note for
7094 matched operands of insn with no output reloads.
7096 2014-01-21 Richard Sandiford <rdsandiford@googlemail.com>
7098 * config/mips/mips.c (mips_move_to_gpr_cost): Add M16_REGS case.
7099 (mips_move_from_gpr_cost): Likewise.
7101 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
7103 PR rtl-optimization/59858
7104 * lra-constraints.c (SMALL_REGISTER_CLASS_P): Use
7105 ira_class_hard_regs_num.
7106 (process_alt_operands): Increase reject for dying matched operand.
7108 2014-01-21 Jakub Jelinek <jakub@redhat.com>
7111 * config/i386/i386.c (expand_small_movmem_or_setmem): If mode is
7112 smaller than size, perform several stores or loads and stores
7113 at dst + count - size to store or copy all of size bytes, rather
7114 than just last modesize bytes.
7116 2014-01-20 DJ Delorie <dj@redhat.com>
7118 * config/rl78/rl78.c (rl78_propogate_register_origins): Verify
7119 that CLOBBERs are REGs before propogating their values.
7121 2014-01-20 H.J. Lu <hongjiu.lu@intel.com>
7124 * cgraph.c (cgraph_inline_failed_string): Add type to DEFCIFCODE.
7125 (cgraph_inline_failed_type): New function.
7126 * cgraph.h (DEFCIFCODE): Add type.
7127 (cgraph_inline_failed_type_t): New enum.
7128 (cgraph_inline_failed_type): New prototype.
7129 * cif-code.def: Add CIF_FINAL_NORMAL to OK, FUNCTION_NOT_CONSIDERED,
7130 FUNCTION_NOT_OPTIMIZED, REDEFINED_EXTERN_INLINE,
7131 FUNCTION_NOT_INLINE_CANDIDATE, LARGE_FUNCTION_GROWTH_LIMIT,
7132 LARGE_STACK_FRAME_GROWTH_LIMIT, MAX_INLINE_INSNS_SINGLE_LIMIT,
7133 MAX_INLINE_INSNS_AUTO_LIMIT, INLINE_UNIT_GROWTH_LIMIT,
7134 RECURSIVE_INLINING, UNLIKELY_CALL, NOT_DECLARED_INLINED,
7135 OPTIMIZING_FOR_SIZE, ORIGINALLY_INDIRECT_CALL,
7136 INDIRECT_UNKNOWN_CALL, USES_COMDAT_LOCAL.
7137 Add CIF_FINAL_ERROR to UNSPECIFIED, BODY_NOT_AVAILABLE,
7138 FUNCTION_NOT_INLINABLE, OVERWRITABLE, MISMATCHED_ARGUMENTS,
7139 EH_PERSONALITY, NON_CALL_EXCEPTIONS, TARGET_OPTION_MISMATCH,
7140 OPTIMIZATION_MISMATCH.
7141 * tree-inline.c (expand_call_inline): Emit errors during
7142 early_inlining if cgraph_inline_failed_type returns CIF_FINAL_ERROR.
7144 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
7147 * config/i386/sse.md (*andnot<mode>3<mask_name>): Handle MODE_V16SF
7148 mode attribute in insn output.
7150 2014-01-20 Eric Botcazou <ebotcazou@adacore.com>
7152 * output.h (output_constant): Delete.
7153 * varasm.c (output_constant): Make private.
7155 2014-01-20 Alex Velenko <Alex.Velenko@arm.com>
7157 * config/aarch64/aarch64-simd.md (vec_perm<mode>): Add BE check.
7159 2014-01-20 Jakub Jelinek <jakub@redhat.com>
7162 * tree.h (fold_builtin_strcat): New prototype.
7163 * builtins.c (fold_builtin_strcat): No longer static. Add len
7164 argument, if non-NULL, don't call c_strlen. Optimize
7165 directly into __builtin_memcpy instead of __builtin_strcpy.
7166 (fold_builtin_2): Adjust fold_builtin_strcat caller.
7167 * gimple-fold.c (gimple_fold_builtin): Handle BUILT_IN_STRCAT.
7169 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
7171 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
7172 for SImode_address_operand operands, having only a REG argument.
7174 2014-01-20 Marcus Shawcroft <marcus.shawcroft@arm.com>
7176 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER): Expand
7177 loader name using mbig-endian.
7178 (LINUX_TARGET_LINK_SPEC): Pass linker -m flag.
7180 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
7182 * doc/invoke.texi (-march): Clarify documentation for AArch64.
7186 2014-01-20 Tejas Belagod <tejas.belagod@arm.com>
7188 * config/aarch64/aarch64-protos.h
7189 (aarch64_cannot_change_mode_class_ptr): Declare.
7190 * config/aarch64/aarch64.c (aarch64_cannot_change_mode_class,
7191 aarch64_cannot_change_mode_class_ptr): New.
7192 * config/aarch64/aarch64.h (CANNOT_CHANGE_MODE_CLASS): Change to call
7193 backend hook aarch64_cannot_change_mode_class.
7195 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
7197 * common/config/aarch64/aarch64-common.c
7198 (aarch64_handle_option): Don't handle any option order logic here.
7199 * config/aarch64/aarch64.c (aarch64_parse_arch): Do not override
7200 selected_cpu, warn on architecture version mismatch.
7201 (aarch64_override_options): Fix parsing order for option strings.
7203 2014-01-20 Jan-Benedict Glaw <jbglaw@lug-owl.de>
7204 Iain Sandoe <iain@codesourcery.com>
7207 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Fix unused variable
7208 warning. Amend comment to reflect current functionality.
7210 2014-01-20 Richard Biener <rguenther@suse.de>
7213 * builtins.c (fold_builtin_strcat): Remove case better handled
7214 by tree-ssa-strlen.c.
7216 2014-01-20 Alan Lawrence <alan.lawrence@arm.com>
7218 * config/aarch64/aarch64.opt
7219 (mcpu, march, mtune): Make case-insensitive.
7221 2014-01-20 Jakub Jelinek <jakub@redhat.com>
7224 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
7225 if operands[1] is a REG or ZERO_EXTEND of a REG.
7227 2014-01-19 Jan Hubicka <hubicka@ucw.cz>
7229 * varasm.c (compute_reloc_for_constant): Use targetm.binds_local_p.
7231 2014-01-19 John David Anglin <danglin@gcc.gnu.org>
7233 * config/pa/pa.c (pa_attr_length_millicode_call): Correct length of
7234 long non-pic millicode calls.
7236 2014-01-19 Jan-Benedict Glaw <jbglaw@lug-owl.de>
7238 * config/vax/vax.h (FUNCTION_ARG_REGNO_P): Fix unused variable warning.
7240 2014-01-19 Kito Cheng <kito@0xlab.org>
7242 * builtins.c (expand_movstr): Check movstr expand done or fail.
7244 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
7245 H.J. Lu <hongjiu.lu@intel.com>
7248 * config/i386/i386.md (*lea<mode>): Zero-extend return register
7249 to DImode for zero-extended addresses.
7251 2014-01-19 Jakub Jelinek <jakub@redhat.com>
7253 PR rtl-optimization/57763
7254 * bb-reorder.c (fix_crossing_unconditional_branches): Set JUMP_LABEL
7255 on the new indirect jump_insn and increment LABEL_NUSES (label).
7257 2014-01-18 H.J. Lu <hongjiu.lu@intel.com>
7261 * config.gcc (x86_archs): New variable.
7262 (x86_64_archs): Likewise.
7263 (x86_cpus): Likewise.
7264 Use $x86_archs, $x86_64_archs and $x86_cpus to check valid
7265 --with-arch/--with-cpu= options.
7266 Support --with-arch=/--with-cpu={nehalem,westmere,
7267 sandybridge,ivybridge,haswell,broadwell,bonnell,silvermont}.
7269 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
7271 * config/i386/i386.c (ix86_adjust_cost): Reorder PROCESSOR_K8
7272 and PROCESSOR_ATHLON to simplify code. Move "memory" calculation.
7274 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
7276 * config/i386/i386.md (*swap<mode>): Rename from swap<mode>.
7278 2014-01-18 Jakub Jelinek <jakub@redhat.com>
7281 * config/i386/i386-c.c (ix86_pragma_target_parse): Temporarily
7282 clear cpp_get_options (parse_in)->warn_unused_macros for
7283 ix86_target_macros_internal with cpp_define.
7285 2014-01-18 Richard Sandiford <rdsandiford@googlemail.com>
7287 * jump.c (delete_related_insns): Keep (use (insn))s.
7288 * reorg.c (redundant_insn): Check for barriers too.
7290 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
7292 * config/i386/i386.c (ix86_split_lea_for_addr): Fix a comment typo.
7294 2014-01-17 John David Anglin <danglin@gcc.gnu.org>
7296 * config/pa/pa.c (pa_attr_length_indirect_call): Don't output a short
7297 call to $$dyncall when TARGET_LONG_CALLS is true.
7299 2014-01-17 Jeff Law <law@redhat.com>
7301 * ree.c (combine_set_extension): Temporarily disable test for
7302 changing number of hard registers.
7304 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
7307 * ipa-inline-analysis.c (inline_free_summary):
7308 Do not free summary of aliases.
7310 2014-01-17 Jakub Jelinek <jakub@redhat.com>
7313 * gimplify.c (gimplify_expr): Use create_tmp_var
7314 instead of create_tmp_var_raw. If cond doesn't have
7315 integral type, don't add the IFN_ANNOTATE builtin at all.
7317 2014-01-17 Martin Jambor <mjambor@suse.cz>
7320 * ipa-cp.c (prev_edge_clone): New variable.
7321 (grow_next_edge_clone_vector): Renamed to grow_edge_clone_vectors.
7322 Also resize prev_edge_clone vector.
7323 (ipcp_edge_duplication_hook): Also update prev_edge_clone.
7324 (ipcp_edge_removal_hook): New function.
7325 (ipcp_driver): Register ipcp_edge_removal_hook.
7327 2014-01-17 Andrew Pinski <apinski@cavium.com>
7328 Steve Ellcey <sellcey@mips.com>
7331 * config/mips/mips.c (mips_print_operand): Check operand mode instead
7334 2014-01-17 Jeff Law <law@redhat.com>
7337 * passes.def: Reorder pass_copy_prop, pass_unrolli, pass_ccp sequence
7338 so that pass_ccp runs first.
7340 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
7342 * config/i386/i386.c (ix86_lea_outperforms): Use TARGET_XXX.
7343 (ix86_adjust_cost): Use !TARGET_XXX.
7344 (do_reorder_for_imul): Likewise.
7345 (swap_top_of_ready_list): Likewise.
7346 (ix86_sched_reorder): Likewise.
7348 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
7350 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
7351 PROCESSOR_INTEL. Treat like PROCESSOR_GENERIC.
7352 * config/i386/i386.c (intel_memcpy): New. Duplicate slm_memcpy.
7353 (intel_memset): New. Duplicate slm_memset.
7354 (intel_cost): New. Duplicate slm_cost.
7355 (m_INTEL): New macro.
7356 (processor_target_table): Add "intel".
7357 (ix86_option_override_internal): Replace PROCESSOR_SILVERMONT
7358 with PROCESSOR_INTEL for "intel".
7359 (ix86_lea_outperforms): Support PROCESSOR_INTEL. Duplicate
7360 PROCESSOR_SILVERMONT.
7361 (ix86_issue_rate): Likewise.
7362 (ix86_adjust_cost): Likewise.
7363 (ia32_multipass_dfa_lookahead): Likewise.
7364 (swap_top_of_ready_list): Likewise.
7365 (ix86_sched_reorder): Likewise.
7366 (ix86_avoid_lea_for_addr): Check TARGET_AVOID_LEA_FOR_ADDR
7367 instead of TARGET_OPT_AGU.
7368 * config/i386/i386.h (TARGET_INTEL): New.
7369 (TARGET_AVOID_LEA_FOR_ADDR): Likewise.
7370 (processor_type): Add PROCESSOR_INTEL.
7371 * config/i386/x86-tune.def: Support m_INTEL. Duplicate m_SILVERMONT.
7372 Add X86_TUNE_AVOID_LEA_FOR_ADDR.
7374 2014-01-17 Marek Polacek <polacek@redhat.com>
7377 * gimple-fold.c (fold_array_ctor_reference): Don't fold if element
7380 2014-01-17 Richard Biener <rguenther@suse.de>
7382 PR tree-optimization/46590
7383 * opts.c (default_options_table): Add entries for
7384 OPT_fbranch_count_reg, OPT_fmove_loop_invariants and OPT_ftree_pta,
7385 all enabled at -O1 but not for -Og.
7386 * common.opt (fbranch-count-reg): Remove Init(1).
7387 (fmove-loop-invariants): Likewise.
7388 (ftree-pta): Likewise.
7390 2014-01-17 Jakub Jelinek <jakub@redhat.com>
7392 * config/i386/i386.c (ix86_data_alignment): For compatibility with
7393 (incorrect) GCC 4.8 and earlier alignment assumptions ensure we align
7394 decls to at least the GCC 4.8 used alignments.
7397 * tree-nested.c (convert_nonlocal_reference_stmt,
7398 convert_local_reference_stmt): For NAMELIST_DECLs in gimple_bind_vars
7399 of GIMPLE_BIND stmts, adjust associated decls.
7401 2014-01-17 Richard Biener <rguenther@suse.de>
7403 PR tree-optimization/46590
7404 * vec.h (vec<>::bseach): New member function implementing
7405 binary search according to C89 bsearch.
7406 (vec<>::qsort): Avoid calling ::qsort for vectors with sizes 0 or 1.
7407 * tree-ssa-loop-im.c (struct mem_ref): Make stored member a
7408 bitmap pointer again. Make accesses_in_loop a flat array.
7409 (mem_ref_obstack): New global.
7410 (outermost_indep_loop): Adjust for mem_ref->stored changes.
7411 (mark_ref_stored): Likewise.
7412 (ref_indep_loop_p_2): Likewise.
7413 (set_ref_stored_in_loop): New helper function.
7414 (mem_ref_alloc): Allocate mem_refs on the mem_ref_obstack obstack.
7415 (memref_free): Adjust.
7416 (record_mem_ref_loc): Simplify.
7417 (gather_mem_refs_stmt): Adjust.
7418 (sort_locs_in_loop_postorder_cmp): New function.
7419 (analyze_memory_references): Sort accesses_in_loop after
7420 loop postorder number.
7421 (find_ref_loc_in_loop_cmp): New function.
7422 (for_all_locs_in_loop): Find relevant cluster of locs in
7423 accesses_in_loop and iterate without recursion.
7424 (execute_sm): Avoid uninit warning.
7425 (struct ref_always_accessed): Simplify.
7426 (ref_always_accessed::operator ()): Likewise.
7427 (ref_always_accessed_p): Likewise.
7428 (tree_ssa_lim_initialize): Initialize mem_ref_obstack, compute
7429 loop postorder numbers here.
7430 (tree_ssa_lim_finalize): Free mem_ref_obstack and loop postorder
7433 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
7436 * passes.c (rest_of_decl_compilation): Don't call varpool_finalize_decl
7437 on decls for which assemble_alias has been called.
7439 2014-01-17 Nick Clifton <nickc@redhat.com>
7441 * config/msp430/msp430.opt: (mcpu): New option.
7442 * config/msp430/msp430.c (msp430_mcu_name): Use target_mcu.
7443 (msp430_option_override): Parse target_cpu. If the MCU name
7444 matches a generic string, clear target_mcu.
7445 (msp430_attr): Allow numeric interrupt values up to 63.
7446 (msp430_expand_epilogue): No longer invert operand 1 of gen_popm.
7447 * config/msp430/msp430.h (ASM_SPEC): Convert -mcpu into a -mmcu
7449 * config/msp430/t-msp430: (MULTILIB_MATCHES): Remove mcu matches.
7451 * config/msp430/msp430.md (popm): Use %J rather than %I.
7452 (addsi3): Use msp430_nonimmediate_operand for operand 2.
7453 (addhi_cy_i): Use immediate_operand for operand 2.
7454 * doc/invoke.texi: Document -mcpu option.
7456 2014-01-17 Richard Biener <rguenther@suse.de>
7458 PR rtl-optimization/38518
7459 * df.h (df_analyze_loop): Declare.
7460 * df-core.c: Include cfgloop.h.
7461 (df_analyze_1): Split out main part of df_analyze.
7462 (df_analyze): Adjust.
7463 (loop_inverted_post_order_compute): New function.
7464 (loop_post_order_compute): Likewise.
7465 (df_analyze_loop): New function avoiding whole-function
7467 * loop-invariant.c (find_defs): Use df_analyze_loop.
7468 (find_invariants): Adjust.
7469 * loop-iv.c (iv_analysis_loop_init): Use df_analyze_loop.
7471 2014-01-17 Zhenqiang Chen <zhenqiang.chen@arm.com>
7473 * config/arm/arm.c (arm_v7m_tune): Set max_insns_skipped to 2.
7474 (thumb2_final_prescan_insn): Set max to MAX_INSN_PER_IT_BLOCK.
7476 2014-01-16 Ilya Enkovich <ilya.enkovich@intel.com>
7478 * ipa-ref.c (ipa_remove_stmt_references): Fix references
7479 traversal when removing references.
7481 2014-01-16 Jan Hubicka <hubicka@ucw.cz>
7484 * tree.c (get_binfo_at_offset): Look harder for virtual bases.
7486 2014-01-16 Bernd Schmidt <bernds@codesourcery.com>
7489 * reload.c (find_reloads_address_1): Do not use RELOAD_OTHER when
7490 pushing a reload for an autoinc when we had previously reloaded an
7491 inner part of the address.
7493 2014-01-16 Jakub Jelinek <jakub@redhat.com>
7495 * tree-vectorizer.h (struct _loop_vec_info): Add no_data_dependencies
7497 (LOOP_VINFO_NO_DATA_DEPENDENCIES): Define.
7498 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Clear it
7499 when not giving up or versioning for alias only because of
7501 (vect_analyze_data_ref_dependences): Set to true.
7502 * tree-vect-stmts.c (hoist_defs_of_uses): Return false if def_stmt
7504 (vectorizable_load): Use LOOP_VINFO_NO_DATA_DEPENDENCIES instead of
7505 LOOP_REQUIRES_VERSIONING_FOR_ALIAS, add && !nested_in_vect_loop
7509 * expr.c (expand_expr_real_1): Handle init == NULL_TREE.
7512 * config/i386/i386.c (ix86_expand_builtin): If target doesn't satisfy
7513 operand 0 predicate for gathers, use a new pseudo as subtarget.
7515 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
7518 * lra-constraints.c (process_alt_operands): Add printing debug info.
7519 Check absence of input/output reloads for matched operands too.
7521 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
7523 PR rtl-optimization/59835
7524 * ira.c (ira_init_register_move_cost): Increase cost for
7527 2014-01-16 Alan Lawrence <alan.lawrence@arm.com>
7529 * config/arm/arm.opt (mcpu, march, mtune): Make case-insensitive.
7531 2014-01-16 Richard Earnshaw <rearnsha@arm.com>
7534 * aarch64.c (aarch64_split_128bit_move): Don't lookup REGNO on
7535 non-register objects. Use gen_(high/low)part more consistently.
7538 2014-01-16 Michael Meissner <meissner@linux.vnet.ibm.com>
7541 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Add little
7542 endian support, remove tests for WORDS_BIG_ENDIAN.
7543 (p8_mfvsrd_3_<mode>): Likewise.
7544 (reload_gpr_from_vsx<mode>): Likewise.
7545 (reload_gpr_from_vsxsf): Likewise.
7546 (p8_mfvsrd_4_disf): Likewise.
7548 2014-01-16 Richard Biener <rguenther@suse.de>
7550 PR rtl-optimization/46590
7551 * lcm.c (compute_antinout_edge): Use postorder iteration.
7552 (compute_laterin): Use inverted postorder iteration.
7554 2014-01-16 Nick Clifton <nickc@redhat.com>
7557 * varasm.c (output_constant): Return the number of bytes actually
7559 (output_constructor_array_range): Update the field size with the
7560 number of bytes emitted by output_constant.
7561 (output_constructor_regular_field): Likewise. Also do not
7562 complain if the total number of bytes emitted is now greater
7563 than the expected fieldpos.
7564 * output.h (output_constant): Update prototype and descriptive comment.
7566 2014-01-16 Marek Polacek <polacek@redhat.com>
7569 * cgraph.c (gimple_check_call_args): Don't use DECL_ARG_TYPE if
7570 it is error_mark_node.
7572 2014-01-15 Uros Bizjak <ubizjak@gmail.com>
7574 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use
7575 VALID_AVX256_REG_OR_OI_MODE.
7577 2014-01-15 Pat Haugen <pthaugen@us.ibm.com>
7579 * config/rs6000/rs6000.c (rs6000_output_function_prologue): Check if
7580 current procedure should be profiled.
7582 2014-01-15 Andrew Pinski <apinski@cavium.com>
7584 * config/aarch64/aarch64.c (aarch64_register_move_cost): Correct cost
7585 of moving from/to the STACK_REG register class.
7587 2014-01-15 Richard Henderson <rth@redhat.com>
7590 * reginfo.c (global_regs_decl): Globalize.
7591 * rtl.h (global_regs_decl): Declare.
7592 * ira.c (do_reload): Diagnose frame_pointer_needed and it
7593 reserved via global_regs.
7595 2014-01-15 Teresa Johnson <tejohnson@google.com>
7597 * tree-ssa-sccvn.c (visit_reference_op_call): Handle NULL vdef.
7599 2014-01-15 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
7601 * config/rs6000/altivec.md (mulv8hi3): Explicitly generate vmulesh
7602 and vmulosh rather than call gen_vec_widen_smult_*.
7603 (vec_widen_umult_even_v16qi): Test VECTOR_ELT_ORDER_BIG rather
7604 than BYTES_BIG_ENDIAN to determine use of even or odd instruction.
7605 (vec_widen_smult_even_v16qi): Likewise.
7606 (vec_widen_umult_even_v8hi): Likewise.
7607 (vec_widen_smult_even_v8hi): Likewise.
7608 (vec_widen_umult_odd_v16qi): Likewise.
7609 (vec_widen_smult_odd_v16qi): Likewise.
7610 (vec_widen_umult_odd_v8hi): Likewise.
7611 (vec_widen_smult_odd_v8hi): Likewise.
7612 (vec_widen_umult_hi_v16qi): Explicitly generate vmuleub and
7613 vmuloub rather than call gen_vec_widen_umult_*.
7614 (vec_widen_umult_lo_v16qi): Likewise.
7615 (vec_widen_smult_hi_v16qi): Explicitly generate vmulesb and
7616 vmulosb rather than call gen_vec_widen_smult_*.
7617 (vec_widen_smult_lo_v16qi): Likewise.
7618 (vec_widen_umult_hi_v8hi): Explicitly generate vmuleuh and vmulouh
7619 rather than call gen_vec_widen_umult_*.
7620 (vec_widen_umult_lo_v8hi): Likewise.
7621 (vec_widen_smult_hi_v8hi): Explicitly gnerate vmulesh and vmulosh
7622 rather than call gen_vec_widen_smult_*.
7623 (vec_widen_smult_lo_v8hi): Likewise.
7625 2014-01-15 Jeff Law <law@redhat.com>
7627 PR tree-optimization/59747
7628 * ree.c (find_and_remove_re): Properly handle case where a second
7629 eliminated extension requires widening a copy created for elimination
7630 of a prior extension.
7631 (combine_set_extension): Ensure that the number of hard regs needed
7632 for a destination register does not change when we widen it.
7634 2014-01-15 Sebastian Huber <sebastian.huber@embedded-brains.de>
7636 * config.gcc (*-*-rtems*): Add t-rtems to tmake_file.
7637 (arm*-*-uclinux*eabi*): Do not override an existing tmake_file.
7638 (arm*-*-eabi* | arm*-*-symbianelf* | arm*-*-rtems*): Likwise.
7639 (arm*-*-rtems*): Use t-rtems from existing tmake_file.
7640 (avr-*-rtems*): Likewise.
7641 (bfin*-rtems*): Likewise.
7642 (moxie-*-rtems*): Likewise.
7643 (h8300-*-rtems*): Likewise.
7644 (i[34567]86-*-rtems*): Likewise.
7645 (lm32-*-rtems*): Likewise.
7646 (m32r-*-rtems*): Likewise.
7647 (m68k-*-rtems*): Likewise.
7648 (microblaze*-*-rtems*): Likewise.
7649 (mips*-*-rtems*): Likewise.
7650 (powerpc-*-rtems*): Likewise.
7651 (sh-*-rtems*): Likewise.
7652 (sparc-*-rtems*): Likewise.
7653 (sparc64-*-rtems*): Likewise.
7654 (v850-*-rtems*): Likewise.
7655 (m32c-*-rtems*): Likewise.
7657 2014-01-15 Vladimir Makarov <vmakarov@redhat.com>
7659 PR rtl-optimization/59511
7660 * ira.c (ira_init_register_move_cost): Use memory costs for some
7661 cases of register move cost calculations.
7662 * lra-constraints.c (lra_constraints): Use REG_FREQ_FROM_BB
7663 instead of BB frequency.
7664 * lra-coalesce.c (move_freq_compare_func, lra_coalesce): Ditto.
7665 * lra-assigns.c (find_hard_regno_for): Ditto.
7667 2014-01-15 Richard Biener <rguenther@suse.de>
7669 PR tree-optimization/59822
7670 * tree-vect-stmts.c (hoist_defs_of_uses): New function.
7671 (vectorizable_load): Use it to hoist defs of uses of invariant
7672 loads out of the loop.
7674 2014-01-15 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
7675 Kugan Vivekanandarajah <kuganv@linaro.org>
7678 * config/aarch64/aarch64.c (aarch64_build_constant): Fix incorrect
7681 2014-01-15 Richard Biener <rguenther@suse.de>
7683 PR rtl-optimization/59802
7684 * lcm.c (compute_available): Use inverted postorder to seed
7685 the initial worklist.
7687 2014-01-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7690 * config/s390/s390.c (s390_preferred_reload_class): Don't return
7691 ADDR_REGS for invalid symrefs in non-PIC code.
7693 2014-01-15 Jakub Jelinek <jakub@redhat.com>
7696 * builtins.c (determine_block_size): Initialize *probable_max_size
7697 even if len_rtx is CONST_INT.
7699 2014-01-14 Andrew Pinski <apinski@cavium.com>
7701 * config/aarch64/aarch64-protos.h (tune_params): Add issue_rate.
7702 * config/aarch64/aarch64.c (generic_tunings): Add issue rate of 2.
7703 (cortexa53_tunings): Likewise.
7704 (aarch64_sched_issue_rate): New function.
7705 (TARGET_SCHED_ISSUE_RATE): Define.
7707 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
7709 * ira-costs.c (find_costs_and_classes): Add missed
7710 ira_init_register_move_cost_if_necessary.
7712 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
7715 * config/arm/arm.c (arm_coproc_mem_operand): Add lra_in_progress.
7717 2014-01-14 H.J. Lu <hongjiu.lu@intel.com>
7720 * config/i386/i386.c (type_natural_mode): Add a bool parameter
7721 to indicate if type is used for function return value. Warn ABI
7722 change if the vector mode isn't available for function return value.
7723 (ix86_function_arg_advance): Pass false to type_natural_mode.
7724 (ix86_function_arg): Likewise.
7725 (ix86_gimplify_va_arg): Likewise.
7726 (function_arg_32): Don't warn ABI change.
7727 (ix86_function_value): Pass true to type_natural_mode.
7728 (ix86_return_in_memory): Likewise.
7729 (ix86_struct_value_rtx): Removed.
7730 (TARGET_STRUCT_VALUE_RTX): Likewise.
7732 2014-01-14 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7734 * jump.c (redirect_jump_2): Remove REG_CROSSING_JUMP notes when
7735 converting a conditional jump into a conditional return.
7737 2014-01-14 Richard Biener <rguenther@suse.de>
7739 PR tree-optimization/58921
7740 PR tree-optimization/59006
7741 * tree-vect-loop-manip.c (vect_loop_versioning): Remove code
7742 hoisting invariant stmts.
7743 * tree-vect-stmts.c (vectorizable_load): Insert the splat of
7744 invariant loads on the preheader edge if possible.
7746 2014-01-14 Joey Ye <joey.ye@arm.com>
7748 * doc/plugin.texi (Building GCC plugins): Update to C++.
7750 2014-01-14 Kirill Yukhin <kirill.yukhin@intel.com>
7752 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): New.
7753 (_mm_rcp28_round_ss): Ditto.
7754 (_mm_rsqrt28_round_sd): Ditto.
7755 (_mm_rsqrt28_round_ss): Ditto.
7756 (_mm_rcp28_sd): Ditto.
7757 (_mm_rcp28_ss): Ditto.
7758 (_mm_rsqrt28_sd): Ditto.
7759 (_mm_rsqrt28_ss): Ditto.
7760 * config/i386/avx512fintrin.h (_mm512_stream_load_si512): Ditto.
7761 * config/i386/i386-builtin-types.def (V8DI_FTYPE_PV8DI): Ditto.
7762 * config/i386/i386.c (IX86_BUILTIN_MOVNTDQA512): Ditto.
7763 (IX86_BUILTIN_RCP28SD): Ditto.
7764 (IX86_BUILTIN_RCP28SS): Ditto.
7765 (IX86_BUILTIN_RSQRT28SD): Ditto.
7766 (IX86_BUILTIN_RSQRT28SS): Ditto.
7767 (bdesc_special_args): Define __builtin_ia32_movntdqa512,
7768 __builtin_ia32_rcp28sd_round, __builtin_ia32_rcp28ss_round,
7769 __builtin_ia32_rsqrt28sd_round, __builtin_ia32_rsqrt28ss_round.
7770 (ix86_expand_special_args_builtin): Expand new FTYPE.
7771 * config/i386/sse.md (define_mode_attr "sse4_1_avx2"): Expand to V8DI.
7772 (srcp14<mode>): Make insn unary.
7773 (avx512f_vmscalef<mode><round_name>): Use substed predicate.
7774 (avx512f_sgetexp<mode><round_saeonly_name>): Ditto.
7775 (avx512f_rndscale<mode><round_saeonly_name>): Ditto.
7776 (<sse4_1_avx2>_movntdqa): Extend to 512 bits.
7777 (avx512er_exp2<mode><mask_name><round_saeonly_name>):
7778 Fix rounding: make it SAE only.
7779 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
7781 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
7783 (avx512er_vmrcp28<mode><round_saeonly_name>): Ditto.
7784 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Ditto.
7785 (avx512f_getmant<mode><mask_name><round_saeonly_name>): Ditto.
7786 * config/i386/subst.md (round_saeonly_mask_scalar_operand3): Remove.
7787 (round_saeonly_mask_scalar_operand4): Ditto.
7788 (round_saeonly_mask_scalar_op3): Ditto.
7789 (round_saeonly_mask_scalar_op4): Ditto.
7791 2014-01-13 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7793 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
7794 Implement -maltivec=be for vec_insert and vec_extract.
7796 2014-01-10 DJ Delorie <dj@redhat.com>
7798 * config/msp430/msp430.md (call_internal): Don't allow memory
7799 references with SP as the base register.
7800 (call_value_internal): Likewise.
7801 * config/msp430/constraints.md (Yc): New. For memory references
7802 that don't use SP as a base register.
7804 * config/msp430/msp430.c (msp430_print_operand): Add 'J' to mean
7805 "an integer without a # prefix"
7806 * config/msp430/msp430.md (epilogue_helper): Use it.
7808 2014-01-13 Jakub Jelinek <jakub@redhat.com>
7811 * config/i386/i386.c (ix86_vectorize_builtin_gather): Uncomment
7812 AVX512F gather builtins.
7813 * tree-vect-stmts.c (vectorizable_mask_load_store): For now punt
7814 on gather decls with INTEGER_TYPE masktype.
7815 (vectorizable_load): For INTEGER_TYPE masktype, put the INTEGER_CST
7816 directly into the builtin rather than hoisting it before loop.
7818 PR tree-optimization/59387
7819 * tree-scalar-evolution.c: Include gimple-fold.h and gimplify-me.h.
7820 (scev_const_prop): If folded_casts and type has undefined overflow,
7821 use force_gimple_operand instead of force_gimple_operand_gsi and
7822 for each added stmt if it is assign with
7823 arith_code_with_undefined_signed_overflow, call
7824 rewrite_to_defined_overflow.
7825 * tree-ssa-loop-im.c: Don't include gimplify-me.h, include
7826 gimple-fold.h instead.
7827 (arith_code_with_undefined_signed_overflow,
7828 rewrite_to_defined_overflow): Moved to ...
7829 * gimple-fold.c (arith_code_with_undefined_signed_overflow,
7830 rewrite_to_defined_overflow): ... here. No longer static.
7831 Include gimplify-me.h.
7832 * gimple-fold.h (arith_code_with_undefined_signed_overflow,
7833 rewrite_to_defined_overflow): New prototypes.
7835 2014-01-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7837 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Fix typo in description.
7839 2014-01-13 Eric Botcazou <ebotcazou@adacore.com>
7841 * builtins.c (get_object_alignment_2): Minor tweak.
7842 * tree-ssa-loop-ivopts.c (may_be_unaligned_p): Rewrite.
7844 2014-01-13 Christian Bruel <christian.bruel@st.com>
7846 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Unroll small sizes and
7847 optimized non constant lengths.
7849 2014-01-13 Jakub Jelinek <jakub@redhat.com>
7852 * omp-low.c (expand_omp_atomic_pipeline): Expand the initial
7853 load as __atomic_load_N if possible.
7855 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
7857 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): Remove
7859 (rs6000_expand_builtin): Adjust call.
7861 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
7864 * config/rs6000/rs6000.h (SWITCHABLE_TARGET): Define.
7865 * config/rs6000/rs6000.c: Include target-globals.h.
7866 (rs6000_set_current_function): Instead of doing target_reinit
7867 unconditionally, use save_target_globals_default_opts and
7868 restore_target_globals.
7870 * config/rs6000/rs6000-builtin.def (mffs, mtfsf): Add builtins for
7872 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): New.
7873 (rs6000_expand_builtin): Handle mffs and mtfsf.
7874 (rs6000_init_builtins): Define mffs and mtfsf.
7875 * config/rs6000/rs6000.md (UNSPECV_MFFS, UNSPECV_MTFSF): New constants.
7876 (rs6000_mffs): New pattern.
7877 (rs6000_mtfsf): New pattern.
7879 2014-01-11 Bin Cheng <bin.cheng@arm.com>
7881 * tree-ssa-loop-ivopts.c (iv_ca_narrow): New parameter.
7882 Start narrowing with START. Apply candidate-use pair
7883 and check overall cost in narrowing.
7884 (iv_ca_prune): Pass new argument.
7886 2014-01-10 Jeff Law <law@redhat.com>
7889 * ree.c (combine_reaching_defs): Ensure the defining statement
7890 occurs before the extension when optimizing extensions with
7891 different source and destination hard registers.
7893 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
7896 * ipa-devirt.c (build_type_inheritance_graph): Also add types of
7897 vtables into the type inheritance graph.
7899 2014-01-10 Jakub Jelinek <jakub@redhat.com>
7901 PR rtl-optimization/59754
7902 * ree.c (combine_reaching_defs): Disallow !SCALAR_INT_MODE_P
7903 modes in the REGNO != REGNO case.
7905 2014-01-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7907 * config/rs6000/rs6000-builtin.def: Fix pasto for VPKSDUS.
7909 2014-01-10 Jakub Jelinek <jakub@redhat.com>
7911 PR tree-optimization/59745
7912 * tree-predcom.c (tree_predictive_commoning_loop): Call
7913 free_affine_expand_cache if giving up because components is NULL.
7915 * target-globals.c (save_target_globals): Allocate < 4KB structs using
7916 GC in payload of target_globals struct instead of allocating them on
7917 the heap and the larger structs separately using GC.
7918 * target-globals.h (struct target_globals): Make regs, hard_regs,
7919 reload, expmed, ira, ira_int and lra_fields GTY((atomic)) instead
7920 of GTY((skip)) and change type to void *.
7921 (reset_target_globals): Cast loads from those fields to corresponding
7924 2014-01-10 Steve Ellcey <sellcey@mips.com>
7927 * Makefile.in (PLUGIN_HEADERS): Add gimplify.h, gimple-iterator.h,
7928 gimple-ssa.h, fold-const.h, tree-cfg.h, tree-into-ssa.h,
7929 tree-ssanames.h, print-tree.h, varasm.h, and context.h.
7931 2014-01-10 Richard Earnshaw <rearnsha@arm.com>
7934 * aarch64-modes.def (CC_Zmode): New flags mode.
7935 * aarch64.c (aarch64_select_cc_mode): Only allow NEG when the condition
7936 represents an equality.
7937 (aarch64_get_condition_code): Handle CC_Zmode.
7938 * aarch64.md (compare_neg<mode>): Restrict to equality operations.
7940 2014-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7942 * config/s390/s390.c (s390_expand_tbegin): Remove jump over CC
7943 extraction in good case.
7945 2014-01-10 Richard Biener <rguenther@suse.de>
7947 PR tree-optimization/59374
7948 * tree-vect-slp.c (vect_slp_analyze_bb_1): Move dependence
7949 checking after SLP discovery. Mark stmts not participating
7950 in any SLP instance properly.
7952 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7954 * config/arm/arm.c (arm_new_rtx_costs): Use destination mode
7955 when handling a SET rtx.
7957 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7959 * config/arm/arm-cores.def (cortex-a53): Specify FL_CRC32.
7960 (cortex-a57): Likewise.
7961 (cortex-a57.cortex-a53): Likewise. Remove redundant flags.
7963 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7965 * config/arm/arm.c (arm_init_iwmmxt_builtins): Skip
7966 non-iwmmxt builtins.
7968 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
7972 * ipa-devirt.c record_target_from_binfo): Take as argument
7973 stack of binfos and lookup matching one for virtual inheritance.
7974 (possible_polymorphic_call_targets_1): Update.
7976 2014-01-10 Huacai Chen <chenhc@lemote.com>
7978 * config/mips/driver-native.c (host_detect_local_cpu): Handle new
7979 kernel strings for Loongson-2E/2F/3A.
7981 2014-01-10 Jakub Jelinek <jakub@redhat.com>
7984 * tree-vect-data-refs.c (vect_analyze_data_refs): Check
7985 is_gimple_call before calling gimple_call_internal_p.
7987 2014-01-09 Steve Ellcey <sellcey@mips.com>
7989 * Makefile.in (TREE_FLOW_H): Remove.
7990 (TREE_SSA_H): Add file names from tree-flow.h.
7991 * doc/tree-ssa.texi (Annotations): Remove reference to tree-flow.h
7992 * tree.h: Remove tree-flow.h reference.
7993 * hash-table.h: Remove tree-flow.h reference.
7994 * tree-ssa-loop-niter.c (dump_affine_iv): Replace tree-flow.h
7995 reference with tree-ssa-loop.h.
7997 2014-01-09 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7999 * doc/invoke.texi: Add -maltivec={be,le} options, and document
8000 default element-order behavior for -maltivec.
8001 * config/rs6000/rs6000.opt: Add -maltivec={be,le} options.
8002 * config/rs6000/rs6000.c (rs6000_option_override_internal): Ensure
8003 that -maltivec={le,be} implies -maltivec; disallow -maltivec=le
8004 when targeting big endian, at least for now.
8005 * config/rs6000/rs6000.h: Add #define of VECTOR_ELT_ORDER_BIG.
8007 2014-01-09 Jakub Jelinek <jakub@redhat.com>
8010 * cfgexpand.c (expand_one_var): For SSA_NAMEs, if the underlying
8011 var satisfies use_register_for_decl, just take into account type
8012 alignment, rather than decl alignment.
8014 PR tree-optimization/59622
8015 * gimple-fold.c (gimple_fold_call): Fix a typo in message. For
8016 __builtin_unreachable replace the OBJ_TYPE_REF call with a call to
8017 __builtin_unreachable and add if needed a setter of the lhs SSA_NAME.
8018 Don't devirtualize for inplace at all. For targets.length () == 1,
8019 if the call is noreturn and cfun isn't in SSA form yet, clear lhs.
8021 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
8023 * config/i386/i386.md (cpu): Remove the unused btver1.
8025 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
8027 * gdbasan.in: Put a breakpoint on __sanitizer::Report.
8029 2014-01-09 Jakub Jelinek <jakub@redhat.com>
8032 * tree-core.h (struct target_globals): New forward declaration.
8033 (struct tree_target_option): Add globals field.
8034 * tree.h (TREE_TARGET_GLOBALS): Define.
8035 (prepare_target_option_nodes_for_pch): New prototype.
8036 * target-globals.h (struct target_globals): Define even if
8038 * tree.c (prepare_target_option_node_for_pch,
8039 prepare_target_option_nodes_for_pch): New functions.
8040 * config/i386/i386.h (SWITCHABLE_TARGET): Define.
8041 * config/i386/i386.c: Include target-globals.h.
8042 (ix86_set_current_function): Instead of doing target_reinit
8043 unconditionally, use save_target_globals_default_opts and
8044 restore_target_globals.
8046 2014-01-09 Richard Biener <rguenther@suse.de>
8048 PR tree-optimization/59715
8049 * tree-cfg.h (split_critical_edges): Declare.
8050 * tree-cfg.c (split_critical_edges): Export.
8051 * tree-ssa-sink.c (execute_sink_code): Split critical edges.
8053 2014-01-09 Max Ostapenko <m.ostapenko@partner.samsung.com>
8055 * cfgexpand.c (expand_stack_vars): Optionally disable
8056 asan stack protection.
8057 (expand_used_vars): Likewise.
8058 (partition_stack_vars): Likewise.
8059 * asan.c (asan_emit_stack_protection): Optionally disable
8060 after return stack usage.
8061 (instrument_derefs): Optionally disable memory access instrumentation.
8062 (instrument_builtin_call): Likewise.
8063 (instrument_strlen_call): Likewise.
8064 (asan_protect_global): Optionally disable global variables protection.
8065 * doc/invoke.texi: Added doc for new options.
8066 * params.def: Added new options.
8067 * params.h: Likewise.
8069 2014-01-09 Jakub Jelinek <jakub@redhat.com>
8071 PR rtl-optimization/59724
8072 * ifcvt.c (cond_exec_process_if_block): Don't call
8073 flow_find_head_matching_sequence with 0 longest_match.
8074 * cfgcleanup.c (flow_find_head_matching_sequence): Count even
8075 non-active insns if !stop_after.
8076 (try_head_merge_bb): Revert 2014-01-07 changes.
8078 2014-01-08 Jeff Law <law@redhat.com>
8080 * ree.c (get_sub_rtx): New function, extracted from...
8081 (merge_def_and_ext): Here.
8082 (combine_reaching_defs): Use get_sub_rtx.
8084 2014-01-08 Eric Botcazou <ebotcazou@adacore.com>
8086 * cgraph.h (varpool_variable_node): Do not choke on null node.
8088 2014-01-08 Catherine Moore <clm@codesourcery.com>
8090 * config/mips/mips.md (simple_return): Attempt to use JRC
8092 * config/mips/mips.h (MIPS_CALL): Attempt to use JALS for microMIPS.
8094 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
8096 PR rtl-optimization/59137
8097 * reorg.c (steal_delay_list_from_target): Call update_block for
8099 (steal_delay_list_from_fallthrough, relax_delay_slots): Likewise.
8101 2014-01-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8103 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Remove
8104 two duplicate entries.
8106 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
8109 2012-10-07 Richard Sandiford <rdsandiford@googlemail.com>
8111 * config/mips/mips.c (mips_truncated_op_cost): New function.
8112 (mips_rtx_costs): Adjust test for BADDU.
8113 * config/mips/mips.md (*baddu_di<mode>): Push truncates to operands.
8115 2012-10-02 Richard Sandiford <rdsandiford@googlemail.com>
8117 * config/mips/mips.md (*baddu_si_eb, *baddu_si_el): Merge into...
8118 (*baddu_si): ...this new pattern.
8120 2014-01-08 Jakub Jelinek <jakub@redhat.com>
8123 * ipa-prop.c (ipa_analyze_params_uses): Ignore uses in debug stmts.
8125 2014-01-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
8128 * expr.h (expand_expr_real, expand_expr_real_1): Add new parameter
8130 (expand_expr, expand_normal): Adjust.
8131 * expr.c (expand_expr_real, expand_expr_real_1): Add new parameter
8132 inner_reference_p. Use inner_reference_p to expand inner references.
8133 (store_expr): Adjust.
8134 * cfgexpand.c (expand_call_stmt): Adjust.
8136 2014-01-08 Rong Xu <xur@google.com>
8138 * gcov-io.c (gcov_var): Move from gcov-io.h.
8139 (gcov_position): Ditto.
8140 (gcov_is_error): Ditto.
8141 (gcov_rewrite): Ditto.
8142 * gcov-io.h: Refactor. Move gcov_var to gcov-io.h, and libgcov
8143 only part to libgcc/libgcov.h.
8145 2014-01-08 Marek Polacek <polacek@redhat.com>
8148 * omp-low.c (simd_clone_adjust): Don't crash if def is NULL.
8150 2014-01-08 Marek Polacek <polacek@redhat.com>
8153 * ubsan.c (ubsan_type_descriptor): Call strip_array_types on type2.
8155 2014-01-08 Jakub Jelinek <jakub@redhat.com>
8157 PR rtl-optimization/59649
8158 * stor-layout.c (get_mode_bounds): For BImode return
8159 0 and STORE_FLAG_VALUE.
8161 2014-01-08 Richard Biener <rguenther@suse.de>
8164 * gimple.h (is_gimple_builtin_call): Remove.
8165 (gimple_builtin_call_types_compatible_p): New.
8166 (gimple_call_builtin_p): New overload.
8167 * gimple.c (is_gimple_builtin_call): Remove.
8168 (validate_call): Rename to ...
8169 (gimple_builtin_call_types_compatible_p): ... this and export. Also
8171 (validate_type): New static function.
8172 (gimple_call_builtin_p): New overload and adjust.
8173 * gimple-fold.c (gimple_fold_builtin): Fold the return value.
8174 (gimple_fold_call): Likewise. Use gimple_call_builtin_p.
8175 (gimple_fold_stmt_to_constant_1): Likewise.
8176 * tsan.c (instrument_gimple): Use gimple_call_builtin_p.
8178 2014-01-08 Richard Biener <rguenther@suse.de>
8181 * gimplify.c (gimplify_expr): Gimplify register-register type
8182 VIEW_CONVERT_EXPRs to separate stmts.
8184 2014-01-07 Jeff Law <law@redhat.com>
8187 * ree.c (combine_set_extension): Handle case where source
8188 and destination registers in an extension insn are different.
8189 (combine_reaching_defs): Allow source and destination registers
8190 in extension to be different under limited circumstances.
8191 (add_removable_extension): Remove restriction that the
8192 source and destination registers in the extension are the same.
8193 (find_and_remove_re): Emit a copy from the extension's
8194 destination to its source after the defining insn if
8195 the source and destination registers are different.
8198 * ifcvt.c (merge_if_block): If we are merging a block with more than
8199 one successor with a block with no successors, remove any BARRIER
8200 after the second block.
8202 2014-01-07 Dan Xio Qiang <ziyan01@163.com>
8204 * hw-doloop.c (reorg_loops): Release the bitmap obstack.
8206 2014-01-07 John David Anglin <danglin@gcc.gnu.org>
8209 * config/pa/pa.c (pa_legitimate_address_p): Return false before reload
8210 for 14-bit register offsets when INT14_OK_STRICT is false.
8212 2014-01-07 Roland Stigge <stigge@antcom.de>
8213 Michael Meissner <meissner@linux.vnet.ibm.com>
8216 * config/rs6000/rs6000.c (rs6000_legitimate_offset_address_p):
8217 Only check TFmode for SPE constants. Don't check TImode or TDmode.
8219 2014-01-07 James Greenhalgh <james.greenhalgh@arm.com>
8221 * config/aarch64/aarch64-elf.h (ASM_SPEC): Remove identity spec for
8224 2014-01-07 Yufeng Zhang <yufeng.zhang@arm.com>
8226 * config/arm/arm.c (arm_expand_neon_args): Call expand_expr
8227 with EXPAND_MEMORY for NEON_ARG_MEMORY; check if the returned
8228 rtx is const0_rtx or not.
8230 2014-01-07 Richard Sandiford <rdsandiford@googlemail.com>
8233 * target-globals.c (save_target_globals): Remove this_fn_optab
8235 * toplev.c: Include optabs.h.
8236 (target_reinit): Temporarily restore the global options if another
8237 set of options are in force.
8239 2014-01-07 Jakub Jelinek <jakub@redhat.com>
8241 PR rtl-optimization/58668
8242 * cfgcleanup.c (flow_find_cross_jump): Don't count
8243 any jumps if dir_p is NULL. Remove p1 variable, use active_insn_p
8244 to determine what is counted.
8245 (flow_find_head_matching_sequence): Use active_insn_p to determine
8247 (try_head_merge_bb): Adjust for the flow_find_head_matching_sequence
8249 * ifcvt.c (count_bb_insns): Use active_insn_p && !JUMP_P to
8250 determine what is counted.
8252 PR tree-optimization/59643
8253 * tree-predcom.c (split_data_refs_to_components): If one dr is
8254 read and one write, determine_offset fails and the write isn't
8255 in the bad component, just put the read into the bad component.
8257 2014-01-07 Mike Stump <mikestump@comcast.net>
8258 Jakub Jelinek <jakub@redhat.com>
8261 * tree-core.h (struct tree_optimization_option): Change optabs
8262 type from unsigned char * to void *.
8263 * optabs.c (init_tree_optimization_optabs): Adjust
8264 TREE_OPTIMIZATION_OPTABS initialization.
8266 2014-01-06 Jakub Jelinek <jakub@redhat.com>
8269 * config/i386/i386.h (struct machine_function): Add
8270 no_drap_save_restore field.
8271 * config/i386/i386.c (ix86_save_reg): Use
8272 !cfun->machine->no_drap_save_restore instead of
8273 crtl->stack_realign_needed.
8274 (ix86_finalize_stack_realign_flags): Don't clear drap_reg unless
8275 this function clears frame_pointer_needed. Set
8276 cfun->machine->no_drap_save_restore if clearing frame_pointer_needed
8277 and DRAP reg is needed.
8279 2014-01-06 Marek Polacek <polacek@redhat.com>
8282 * doc/implement-c.texi: Mention that other integer types are
8283 permitted as bit-field types in strictly conforming mode.
8285 2014-01-06 Felix Yang <fei.yang0953@gmail.com>
8287 * modulo-sched.c (schedule_reg_moves): Clear distance1_uses if it
8290 2014-01-06 Richard Earnshaw <rearnsha@arm.com>
8292 * aarch64.c (aarch64_rtx_costs): Fix cost calculation for MADD.
8294 2014-01-06 Martin Jambor <mjambor@suse.cz>
8297 * ipa-cp.c (ipcp_discover_new_direct_edges): Changed param_index type
8299 * ipa-prop.c (ipa_print_node_params): Fix indentation.
8301 2014-01-06 Eric Botcazou <ebotcazou@adacore.com>
8305 * var-tracking.c (add_stores): Preserve the value of the source even if
8306 we don't record the store.
8308 2014-01-06 Terry Guo <terry.guo@arm.com>
8310 * config.gcc (arm*-*-*): Check --with-arch against arm-arches.def.
8312 2014-01-05 Iain Sandoe <iain@codesourcery.com>
8315 * config/darwin.c (darwin_function_section): Adjust return values to
8316 correspond to optimisation changes made in r206070.
8318 2014-01-05 Uros Bizjak <ubizjak@gmail.com>
8320 * config/i386/i386.c (ix86_data_alignment): Calculate max_align
8321 from prefetch_block tune setting.
8322 (nocona_cost): Correct size of prefetch block to 64.
8324 2014-01-04 Eric Botcazou <ebotcazou@adacore.com>
8326 * config/arm/arm.c (arm_get_frame_offsets): Revamp long lines.
8327 (arm_expand_epilogue_apcs_frame): Take into account the number of bytes
8328 used to save the static chain register in the computation of the offset
8329 from which the FP registers need to be restored.
8331 2014-01-04 Jakub Jelinek <jakub@redhat.com>
8333 PR tree-optimization/59519
8334 * tree-vect-loop-manip.c (slpeel_update_phi_nodes_for_guard1): Don't
8335 ICE if get_current_def (current_new_name) is already non-NULL, as long
8336 as it is a phi result of some other phi in *new_exit_bb that has
8339 * config/i386/sse.md (avx512f_load<mode>_mask): Emit vmovup{s,d}
8340 or vmovdqu* for misaligned_operand.
8341 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>,
8342 <sse2_avx_avx512f>_loaddqu<mode><mask_name>): Handle <mask_applied>.
8343 * config/i386/i386.c (ix86_expand_special_args_builtin): Set
8344 aligned_mem for AVX512F masked aligned load and store builtins and for
8347 2014-01-03 Bingfeng Mei <bmei@broadcom.com>
8349 PR tree-optimization/59651
8350 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
8351 Address range for negative step should be added by TYPE_SIZE_UNIT.
8353 2014-01-03 Andreas Schwab <schwab@linux-m68k.org>
8355 * config/m68k/m68k.c (handle_move_double): Handle pushes with
8356 overlapping registers also for registers other than the stack pointer.
8358 2014-01-03 Marek Polacek <polacek@redhat.com>
8361 * doc/extend.texi: Fix the return value of __builtin_FUNCTION and
8364 2014-01-03 Jakub Jelinek <jakub@redhat.com>
8367 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Don't consider
8370 * config/i386/i386.md (MODE_SIZE): New mode attribute.
8371 (push splitter): Use <P:MODE_SIZE> instead of
8372 GET_MODE_SIZE (<P:MODE>mode).
8373 (lea splitter): Use <MODE_SIZE> instead of GET_MODE_SIZE (<MODE>mode).
8374 (mov -1, reg peephole2): Likewise.
8375 * config/i386/sse.md (*mov<mode>_internal,
8376 <sse>_storeu<ssemodesuffix><avxsizesuffix>,
8377 <sse2_avx_avx512f>_storedqu<mode>, <sse>_andnot<mode>3,
8378 *<code><mode>3, *andnot<mode>3<mask_name>,
8379 <mask_codefor><code><mode>3<mask_name>): Likewise.
8380 * config/i386/subst.md (mask_mode512bit_condition,
8381 sd_mask_mode512bit_condition): Likewise.
8383 2014-01-02 Xinliang David Li <davidxl@google.com>
8385 PR tree-optimization/59303
8386 * tree-ssa-uninit.c (is_use_properly_guarded): Main cleanup.
8387 (dump_predicates): Better output format.
8388 (pred_equal_p): New function.
8389 (is_neq_relop_p): Ditto.
8390 (is_neq_zero_form_p): Ditto.
8391 (pred_expr_equal_p): Ditto.
8392 (pred_neg_p): Ditto.
8393 (simplify_pred): Ditto.
8394 (simplify_preds_2): Ditto.
8395 (simplify_preds_3): Ditto.
8396 (simplify_preds_4): Ditto.
8397 (simplify_preds): Ditto.
8399 (push_to_worklist): Ditto.
8400 (get_pred_info_from_cmp): Ditto.
8401 (is_degenerated_phi): Ditto.
8402 (normalize_one_pred_1): Ditto.
8403 (normalize_one_pred): Ditto.
8404 (normalize_one_pred_chain): Ditto.
8405 (normalize_preds): Ditto.
8406 (normalize_cond_1): Remove function.
8407 (normalize_cond): Ditto.
8408 (is_gcond_subset_of): Ditto.
8409 (is_subset_of_any): Ditto.
8410 (is_or_set_subset_of): Ditto.
8411 (is_and_set_subset_of): Ditto.
8412 (is_norm_cond_subset_of): Ditto.
8413 (pred_chain_length_cmp): Ditto.
8414 (convert_control_dep_chain_into_preds): Type change.
8415 (find_predicates): Ditto.
8416 (find_def_preds): Ditto.
8417 (destroy_predicates_vecs): Ditto.
8418 (find_matching_predicates_in_rest_chains): Ditto.
8419 (use_pred_not_overlap_with_undef_path_pred): Ditto.
8420 (is_pred_expr_subset): Ditto.
8421 (is_pred_chain_subset_of): Ditto.
8422 (is_included_in): Ditto.
8423 (is_superset_of): Ditto.
8425 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
8427 Update copyright years.
8429 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
8431 * common/config/arc/arc-common.c, config/arc/arc-modes.def,
8432 config/arc/arc-protos.h, config/arc/arc.c, config/arc/arc.h,
8433 config/arc/arc.md, config/arc/arc.opt,
8434 config/arm/arm_neon_builtins.def, config/arm/crypto.def,
8435 config/i386/avx512cdintrin.h, config/i386/avx512erintrin.h,
8436 config/i386/avx512fintrin.h, config/i386/avx512pfintrin.h,
8437 config/i386/btver2.md, config/i386/shaintrin.h, config/i386/slm.md,
8438 config/linux-protos.h, config/linux.c, config/winnt-c.c,
8439 diagnostic-color.c, diagnostic-color.h, gimple-ssa-isolate-paths.c,
8440 vtable-verify.c, vtable-verify.h: Use the standard form for the
8443 2014-01-02 Tobias Burnus <burnus@net-b.de>
8445 * gcc.c (process_command): Update copyright notice dates.
8446 * gcov-dump.c: Ditto.
8448 * doc/cpp.texi: Bump @copying's copyright year.
8449 * doc/cppinternals.texi: Ditto.
8450 * doc/gcc.texi: Ditto.
8451 * doc/gccint.texi: Ditto.
8452 * doc/gcov.texi: Ditto.
8453 * doc/install.texi: Ditto.
8454 * doc/invoke.texi: Ditto.
8456 2014-01-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8458 * config/nios2/nios2.h (BITS_PER_UNIT): Don't define it.
8460 2014-01-01 Jakub Jelinek <jakub@redhat.com>
8462 * config/i386/sse.md (*mov<mode>_internal): Guard
8463 EXT_REX_SSE_REGNO_P (REGNO ()) uses with REG_P.
8465 PR rtl-optimization/59647
8466 * cse.c (cse_process_notes_1): Don't substitute negative VOIDmode
8467 new_rtx into UNSIGNED_FLOAT rtxes.
8469 Copyright (C) 2014 Free Software Foundation, Inc.
8471 Copying and distribution of this file, with or without modification,
8472 are permitted in any medium without royalty provided the copyright
8473 notice and this notice are preserved.