1 2014-04-28 Richard Biener <rguenther@suse.de>
3 * tree-vrp.c (vrp_var_may_overflow): Remove.
4 (vrp_visit_phi_node): Rather than bumping to +-INF possibly
5 with overflow immediately bump to one before that value and
6 let iteration figure out overflow status.
8 2014-04-28 Richard Biener <rguenther@suse.de>
10 * configure.ac: Do valgrind header checks unconditionally.
11 Add --enable-valgrind-annotations.
12 * system.h: Guard valgrind header inclusion with
13 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING.
14 * alloc-pool.c (pool_alloc, pool_free): Use
15 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING
16 to guard possibly dead code.
17 * config.in: Regenerated.
18 * configure: Likewise.
20 2014-04-28 Jeff Law <law@redhat.com>
22 PR tree-optimization/60902
23 * tree-ssa-threadedge.c
24 (record_temporary_equivalences_from_stmts_at_dest): Only iterate
25 over real defs when invalidating outputs from statements that do not
26 produce useful outputs for threading.
28 2014-04-28 Richard Biener <rguenther@suse.de>
30 PR tree-optimization/60979
31 * graphite-scop-detection.c (scopdet_basic_block_info): Reject
32 SCOPs that end in a block with a successor with abnormal
35 2014-04-28 Richard Biener <rguenther@suse.de>
37 * tree-pass.h (execute_pass_list): Adjust prototype.
38 * passes.c (pass_manager::execute_early_local_passes):
40 (do_per_function): Change callback signature, push all actual
42 (do_per_function_toporder): Likewise.
43 (execute_function_dump): Adjust.
44 (execute_function_todo): Likewise.
45 (clear_last_verified): Likewise.
46 (verify_curr_properties): Likewise.
47 (update_properties_after_pass): Likewise.
48 (execute_pass_list_1): Split out from ...
49 (execute_pass_list): ... here. Adjust.
50 (execute_ipa_pass_list): Likewise.
51 * cgraphunit.c (cgraph_add_new_function): Adjust.
52 (analyze_function): Likewise.
53 (expand_function): Likewise.
54 * cgraph.c (release_function_body): Free dominance info
55 here instead of asserting it was magically freed elsewhere.
57 2014-04-28 Eric Botcazou <ebotcazou@adacore.com>
59 * configure.ac: Tweak GAS check for LEON instructions on SPARC.
60 * configure: Regenerate.
61 * config/sparc/sparc.opt (muser-mode): New option.
62 * config/sparc/sync.md (atomic_compare_and_swap<mode>_1): Do not enable
64 (atomic_compare_and_swap_leon3_1): New instruction for LEON3.
65 * doc/invoke.texi (SPARC options): Document -muser-mode.
67 2014-04-27 Richard Sandiford <rdsandiford@googlemail.com>
69 * cselib.c (find_slot_memmode): Delete.
70 (cselib_hasher): Change compare_type to a struct.
71 (cselib_hasher::equal): Update accordingly. Don't expect wrapped
73 (preserve_constants_and_equivs): Adjust for new compare_type.
74 (cselib_find_slot): Likewise. Take the mode of the rtx as argument.
75 (wrap_constant): Delete.
76 (cselib_lookup_mem, cselib_lookup_1): Update calls to cselib_find_slot.
78 2014-04-26 Markus Trippelsdorf <markus@trippelsdorf.de>
80 * doc/install.texi (Building with profile feedback): Remove
83 2014-04-26 Tom de Vries <tom@codesourcery.com>
85 * config/i386/i386.md (define_expand "ldexpxf3"): Fix out-of-bounds
88 2014-04-25 Cary Coutant <ccoutant@google.com>
91 * dwarf2out.c (should_move_die_to_comdat): A type definition
92 can contain a subprogram definition, but don't move it to a
94 (clone_as_declaration): Copy DW_AT_abstract_origin attribute.
95 (generate_skeleton_bottom_up): Remove DW_AT_object_pointer attribute
97 (clone_tree_hash): Rename to...
98 (clone_tree_partial): ...this; change callers. Copy
99 DW_TAG_subprogram DIEs as declarations.
100 (copy_decls_walk): Don't copy children of a declaration into a
103 2014-04-25 H.J. Lu <hongjiu.lu@intel.com>
106 * config/i386/i386.md (*movsf_internal): Set MODE to SI for
109 2014-04-25 Jiong Wang <jiong.wang@arm.com>
111 * config/arm/predicates.md (call_insn_operand): Add long_call check.
112 * config/arm/arm.md (sibcall, sibcall_value): Force the address to
114 * config/arm/arm.c (arm_function_ok_for_sibcall): Remove long_call
117 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
119 * config/arm/arm.c (arm_cortex_a8_tune): Initialise
122 2014-04-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
124 PR tree-optimization/60930
125 * gimple-ssa-strength-reduction.c (create_mul_imm_cand): Reject
126 creating a multiply candidate by folding two constant
127 multiplicands when the result overflows.
129 2014-04-25 Jakub Jelinek <jakub@redhat.com>
131 PR tree-optimization/60960
132 * tree-vect-generic.c (expand_vector_operation): Only call
133 expand_vector_divmod if type's mode satisfies VECTOR_MODE_P.
135 2014-04-25 Tom de Vries <tom@codesourcery.com>
137 * expr.c (clobber_reg_mode): New function.
138 * expr.h (clobber_reg): New function.
140 2014-04-25 Tom de Vries <tom@codesourcery.com>
142 * rtlanal.c (find_all_hard_reg_sets): Note INSN_CALL_FUNCTION_USAGE
145 2014-04-25 Radovan Obradovic <robradovic@mips.com>
146 Tom de Vries <tom@codesourcery.com>
148 * rtlanal.c (find_all_hard_reg_sets): Add bool implicit parameter and
150 * rtl.h (find_all_hard_reg_sets): Add bool parameter.
151 * haifa-sched.c (recompute_todo_spec, check_clobbered_conditions): Add
152 new argument to find_all_hard_reg_sets call.
154 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
156 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p):
157 Use HOST_WIDE_INT_C for mask literal.
158 (aarch_rev16_shleft_mask_imm_p): Likewise.
160 2014-04-25 Eric Botcazou <ebotcazou@adacore.com>
163 * config/sparc/sparc.md (ashlsi3_extend): Delete.
165 2014-04-25 Marc Glisse <marc.glisse@inria.fr>
167 PR preprocessor/56540
168 * config/i386/i386-c.c (ix86_target_macros): Define
169 __SIZEOF_FLOAT80__ and __SIZEOF_FLOAT128__.
171 2014-04-25 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
173 * configure.ac (tga_func): Remove.
174 (LIB_TLS_SPEC): Remove.
175 * configure: Regenerate.
176 * config.in: Regenerate.
177 * config/sol2.h (LIB_SPEC): Don't use LIB_TLS_SPEC.
179 2014-04-25 Richard Biener <rguenther@suse.de>
182 * tree-ssa-structalias.c (ipa_pta_execute): Compute direct
183 call stmt use/clobber sets during stmt walk instead of
184 walking the possibly incomplete set of caller edges.
186 2014-04-25 Richard Biener <rguenther@suse.de>
189 * passes.c (apply_ipa_transforms): Inline into only caller ...
190 (execute_one_pass): ... here. Properly bring in function
191 bodies for nodes we want to apply IPA transforms to.
193 2014-04-24 Cong Hou <congh@google.com>
195 PR tree-optimization/60896
196 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Pick up
197 all statements in PATTERN_DEF_SEQ in recognized widen-mult pattern.
198 (vect_mark_pattern_stmts): Set the def type of all statements in
199 PATTERN_DEF_SEQ as vect_internal_def.
201 2014-04-24 Michael Meissner <meissner@linux.vnet.ibm.com>
203 * doc/extend.texi (PowerPC Built-in Functions): Document new
204 powerpc extended divide, bcd, pack/unpack 128-bit, builtin
206 (PowerPC AltiVec/VSX Built-in Functions): Likewise.
208 * config/rs6000/predicates.md (const_0_to_3_operand): New
209 predicate to match 0..3 integer constants.
211 * config/rs6000/rs6000-builtin.def (BU_DFP_MISC_1): Add new macros
212 to support adding miscellaneous builtin functions.
213 (BU_DFP_MISC_2): Likewise.
214 (BU_P7_MISC_1): Likewise.
215 (BU_P7_MISC_2): Likewise.
216 (BU_P8V_MISC_3): Likewise.
217 (BU_MISC_1): Likewise.
218 (BU_MISC_2): Likewise.
219 (DIVWE): Add extended divide builtin functions.
227 (DXEX): Add decimal floating-point builtin functions.
239 (CDTBCD): Add new BCD builtin functions.
243 (BCDADD_LT): Likewise.
244 (BCDADD_EQ): Likewise.
245 (BCDADD_GT): Likewise.
246 (BCDADD_OV): Likewise.
248 (BCDSUB_LT): Likewise.
249 (BCDSUB_EQ): Likewise.
250 (BCDSUB_GT): Likewise.
251 (BCDSUB_OV): Likewise.
252 (PACK_TD): Add new pack/unpack 128-bit type builtin functions.
253 (UNPACK_TD): Likewise.
255 (UNPACK_TF): Likewise.
256 (UNPACK_TF_0): Likewise.
257 (UNPACK_TF_1): Likewise.
258 (PACK_V1TI): Likewise.
259 (UNPACK_V1TI): Likewise.
261 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Add
262 support for decimal floating point builtin functions.
263 (rs6000_expand_ternop_builtin): Add checks for the new builtin
264 functions that take constant arguments.
265 (rs6000_invalid_builtin): Add decimal floating point builtin
267 (rs6000_init_builtins): Setup long double, _Decimal64, and
268 _Decimal128 types for new builtin functions.
269 (builtin_function_type): Set the unsigned flags appropriately for
270 the new builtin functions.
271 (rs6000_opt_masks): Add support for decimal floating point builtin
274 * config/rs6000/rs6000.h (RS6000_BTM_DFP): Add support for decimal
275 floating point builtin functions.
276 (RS6000_BTM_COMMON): Likewise.
277 (RS6000_BTI_long_double): Likewise.
278 (RS6000_BTI_dfloat64): Likewise.
279 (RS6000_BTI_dfloat128): Likewise.
280 (long_double_type_internal_node): Likewise.
281 (dfloat64_type_internal_node): Likewise.
282 (dfloat128_type_internal_node): Likewise.
284 * config/rs6000/altivec.h (UNSPEC_BCDADD): Add support for ISA
285 2.07 bcd arithmetic instructions.
286 (UNSPEC_BCDSUB): Likewise.
287 (UNSPEC_BCD_OVERFLOW): Likewise.
288 (UNSPEC_BCD_ADD_SUB): Likewise.
289 (bcd_add_sub): Likewise.
290 (BCD_TEST): Likewise.
291 (bcd<bcd_add_sub>): Likewise.
292 (bcd<bcd_add_sub>_test): Likewise.
293 (bcd<bcd_add_sub>_test2): Likewise.
294 (bcd<bcd_add_sub>_<code>): Likewise.
295 (peephole2 for combined bcd ops): Likewise.
297 * config/rs6000/dfp.md (UNSPEC_DDEDPD): Add support for new
298 decimal floating point builtin functions.
299 (UNSPEC_DENBCD): Likewise.
300 (UNSPEC_DXEX): Likewise.
301 (UNSPEC_DIEX): Likewise.
302 (UNSPEC_DSCLI): Likewise.
303 (UNSPEC_DSCRI): Likewise.
304 (D64_D128): Likewise.
305 (dfp_suffix): Likewise.
306 (dfp_ddedpd_<mode>): Likewise.
307 (dfp_denbcd_<mode>): Likewise.
308 (dfp_dxex_<mode>): Likewise.
309 (dfp_diex_<mode>): Likewise.
310 (dfp_dscli_<mode>): Likewise.
311 (dfp_dscri_<mode>): Likewise.
313 * config/rs6000/rs6000.md (UNSPEC_ADDG6S): Add support for new BCD
315 (UNSPEC_CDTBCD): Likewise.
316 (UNSPEC_CBCDTD): Likewise.
317 (UNSPEC_DIVE): Add support for new extended divide builtin
319 (UNSPEC_DIVEO): Likewise.
320 (UNSPEC_DIVEU): Likewise.
321 (UNSPEC_DIVEUO): Likewise.
322 (UNSPEC_UNPACK_128BIT): Add support for new builtin functions to
323 pack/unpack 128-bit types.
324 (UNSPEC_PACK_128BIT): Likewise.
325 (idiv_ldiv): New mode attribute to set the 32/64-bit divide type.
326 (udiv<mode>3): Use idiv_ldiv mode attribute.
327 (div<mode>3): Likewise.
328 (addg6s): Add new BCD builtin functions.
331 (UNSPEC_DIV_EXTEND): Add support for new extended divide
333 (div_extend): Likewise.
334 (div<div_extend>_<mode>"): Likewise.
335 (FP128_64): Add support for new builtin functions to pack/unpack
337 (unpack<mode>): Likewise.
338 (unpacktf_0): Likewise.
339 (unpacktf_1): Likewise.
340 (unpack<mode>_dm): Likewise.
341 (unpack<mode>_nodm): Likewise.
342 (pack<mode>): Likewise.
343 (unpackv1ti): Likewise.
344 (packv1ti): Likewise.
346 2014-04-24 Vishnu K S <Vishnu.k_s@atmel.com>
348 * gcc/config/avr/avr.c: Add comment on why -fdelete-null-pointer-checks
351 2014-04-24 Jakub Jelinek <jakub@redhat.com>
353 * tree.h (OMP_CLAUSE_LINEAR_GIMPLE_SEQ): Define.
354 * gimplify.c (omp_is_private): Change last argument's type to int.
355 Only diagnose lastprivate if the simd argument is 1, only diagnose
356 linear if the simd argument is 2.
357 (gimplify_omp_for): Adjust omp_is_private callers. When adding
358 lastprivate or private, add the clause to OMP_FOR_CLAUSES. Pass
359 GOVD_EXPLICIT to omp_add_variable. For simd with collapse == 1
360 create OMP_CLAUSE_LINEAR rather than OMP_CLAUSE_PRIVATE for var.
361 If var != decl and decl is in OMP_CLAUSE_LINEAR, gimplify decl
362 increment to OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
363 * omp-low.c (scan_sharing_clauses, lower_lastprivate_clauses): Handle
364 OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
365 * tree-nested.c (convert_nonlocal_omp_clauses,
366 convert_local_omp_clauses): Handle OMP_CLAUSE_LINEAR.
368 2014-04-24 Segher Boessenkool <segher@kernel.crashing.org>
371 * config/m68k/m68k.md (extendplussidi): Don't allow memory for
374 2014-04-24 Dimitris Papavasiliou <dpapavas@gmail.com>
376 * flag-types.h (enum ivar_visibility): Add.
378 2014-04-24 Trevor Saunders <tsaunders@mozilla.com>
380 * config/sh/sh_treg_combine.c (sh_treg_combine::execute): Take
383 2014-04-24 Alan Lawrence <alan.lawrence@arm.com>
385 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Enable for bigendian.
387 2014-04-24 Radovan Obradovic <robradovic@mips.com>
388 Tom de Vries <tom@codesourcery.com>
390 * reg-notes.def (REG_NOTE (CALL_DECL)): New reg-note REG_CALL_DECL.
391 * calls.c (expand_call, emit_library_call_value_1): Add REG_CALL_DECL
393 * combine.c (distribute_notes): Handle REG_CALL_DECL reg-note.
394 * emit-rtl.c (try_split): Same.
396 2014-04-24 Radovan Obradovic <robradovic@mips.com>
397 Tom de Vries <tom@codesourcery.com>
399 * common.opt (fuse-caller-save): New option.
401 2014-04-24 Tejas Belagod <tejas.belagod@arm.com>
403 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Reverse order of elements
406 2014-04-24 Richard Biener <rguenther@suse.de>
408 * expr.c (expand_expr_real_1): Avoid gimple_assign_rhs_to_tree
409 during TER and instead use the sepops interface for expanding
410 non-GIMPLE_SINGLE_RHS.
412 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
414 * config/i386/sol2.h (ASM_PREFERRED_EH_DATA_FORMAT): Only redefine
415 if not HAVE_AS_IX86_DIFF_SECT_DELTA.
417 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
419 * configure.ac (gcc_cv_as_cfi_directive): Support Solaris/x86
420 assembler 64-bit option.
421 * configure: Regenerate.
423 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
425 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Check
426 TARGET_SIMD rather than TARGET_GENERAL_REGS_ONLY.
427 (TARGET_SIMD): Take AARCH64_ISA_SIMD into account.
428 (TARGET_FLOAT): Take AARCH64_ISA_FP into account.
429 (TARGET_CRYPTO): Take TARGET_SIMD into account.
431 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
433 * config/aarch64/aarch64-builtins.c
434 (aarch64_builtin_vectorized_function): Handle BUILT_IN_BSWAP16,
435 BUILT_IN_BSWAP32, BUILT_IN_BSWAP64.
436 * config/aarch64/aarch64-simd.md (bswap<mode>): New pattern.
437 * config/aarch64/aarch64-simd-builtins.def: Define vector bswap
439 * config/aarch64/iterator.md (VDQHSD): New mode iterator.
440 (Vrevsuff): New mode attribute.
442 2014-04-24 Terry Guo <terry.guo@arm.com>
444 * config/arm/arm.h (machine_function): Define variable
445 after_arm_reorg here.
446 * config/arm/arm.c (after_arm_reorg): Remove the definition.
447 (arm_split_constant): Update the way to access variable
450 (arm_output_function_epilogue): Remove the reset of after_arm_reorg.
452 2014-04-23 Tom de Vries <tom@codesourcery.com>
454 * target-hooks-macros.h: Fix DEFHOOKPOD argument order in comment.
456 2014-04-23 David Malcolm <dmalcolm@redhat.com>
458 * is-a.h: Update comments to reflect the following changes to the
459 "pointerness" of the API, making the template parameter match the
460 return type, allowing use of is-a.h with typedefs of pointers.
461 (is_a_helper::cast): Return a T rather then a pointer to a T, so
462 that the return type matches the parameter to the is_a_helper.
464 (dyn_cast): Likewise.
466 * cgraph.c (cgraph_node_for_asm): Update for removal of implicit
467 pointer from the is-a.h API.
469 * cgraph.h (is_a_helper <cgraph_node>::test): Convert to...
470 (is_a_helper <cgraph_node *>::test): ...this, matching change to
472 (is_a_helper <varpool_node>::test): Likewise, convert to...
473 (is_a_helper <varpool_node *>::test): ...this.
475 (varpool_first_variable): Update for removal of implicit pointer
477 (varpool_next_variable): Likewise.
478 (varpool_first_static_initializer): Likewise.
479 (varpool_next_static_initializer): Likewise.
480 (varpool_first_defined_variable): Likewise.
481 (varpool_next_defined_variable): Likewise.
482 (cgraph_first_defined_function): Likewise.
483 (cgraph_next_defined_function): Likewise.
484 (cgraph_first_function): Likewise.
485 (cgraph_next_function): Likewise.
486 (cgraph_first_function_with_gimple_body): Likewise.
487 (cgraph_next_function_with_gimple_body): Likewise.
488 (cgraph_alias_target): Likewise.
489 (varpool_alias_target): Likewise.
490 (cgraph_function_or_thunk_node): Likewise.
491 (varpool_variable_node): Likewise.
492 (symtab_real_symbol_p): Likewise.
493 * cgraphunit.c (referred_to_p): Likewise.
494 (analyze_functions): Likewise.
495 (handle_alias_pairs): Likewise.
496 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
497 * gimple-ssa.h (gimple_vuse_op): Likewise.
498 (gimple_vdef_op): Likewise.
499 * gimple-streamer-in.c (input_gimple_stmt): Likewise.
500 * gimple.c (gimple_build_asm_1): Likewise.
501 (gimple_build_try): Likewise.
502 (gimple_build_resx): Likewise.
503 (gimple_build_eh_dispatch): Likewise.
504 (gimple_build_omp_for): Likewise.
505 (gimple_omp_for_set_clauses): Likewise.
507 * gimple.h (is_a_helper <gimple_statement_asm>::test): Convert to...
508 (is_a_helper <gimple_statement_asm *>::test): ...this.
509 (is_a_helper <gimple_statement_bind>::test): Convert to...
510 (is_a_helper <gimple_statement_bind *>::test): ...this.
511 (is_a_helper <gimple_statement_call>::test): Convert to...
512 (is_a_helper <gimple_statement_call *>::test): ...this.
513 (is_a_helper <gimple_statement_catch>::test): Convert to...
514 (is_a_helper <gimple_statement_catch *>::test): ...this.
515 (is_a_helper <gimple_statement_resx>::test): Convert to...
516 (is_a_helper <gimple_statement_resx *>::test): ...this.
517 (is_a_helper <gimple_statement_eh_dispatch>::test): Convert to...
518 (is_a_helper <gimple_statement_eh_dispatch *>::test): ...this.
519 (is_a_helper <gimple_statement_eh_else>::test): Convert to...
520 (is_a_helper <gimple_statement_eh_else *>::test): ...this.
521 (is_a_helper <gimple_statement_eh_filter>::test): Convert to...
522 (is_a_helper <gimple_statement_eh_filter *>::test): ...this.
523 (is_a_helper <gimple_statement_eh_mnt>::test): Convert to...
524 (is_a_helper <gimple_statement_eh_mnt *>::test): ...this.
525 (is_a_helper <gimple_statement_omp_atomic_load>::test): Convert to...
526 (is_a_helper <gimple_statement_omp_atomic_load *>::test): ...this.
527 (is_a_helper <gimple_statement_omp_atomic_store>::test): Convert to...
528 (is_a_helper <gimple_statement_omp_atomic_store *>::test): ...this.
529 (is_a_helper <gimple_statement_omp_return>::test): Convert to...
530 (is_a_helper <gimple_statement_omp_return *>::test): ...this.
531 (is_a_helper <gimple_statement_omp_continue>::test): Convert to...
532 (is_a_helper <gimple_statement_omp_continue *>::test): ...this.
533 (is_a_helper <gimple_statement_omp_critical>::test): Convert to...
534 (is_a_helper <gimple_statement_omp_critical *>::test): ...this.
535 (is_a_helper <gimple_statement_omp_for>::test): Convert to...
536 (is_a_helper <gimple_statement_omp_for *>::test): ...this.
537 (is_a_helper <gimple_statement_omp_taskreg>::test): Convert to...
538 (is_a_helper <gimple_statement_omp_taskreg *>::test): ...this.
539 (is_a_helper <gimple_statement_omp_parallel>::test): Convert to...
540 (is_a_helper <gimple_statement_omp_parallel *>::test): ...this.
541 (is_a_helper <gimple_statement_omp_target>::test): Convert to...
542 (is_a_helper <gimple_statement_omp_target *>::test): ...this.
543 (is_a_helper <gimple_statement_omp_sections>::test): Convert to...
544 (is_a_helper <gimple_statement_omp_sections *>::test): ...this.
545 (is_a_helper <gimple_statement_omp_single>::test): Convert to...
546 (is_a_helper <gimple_statement_omp_single *>::test): ...this.
547 (is_a_helper <gimple_statement_omp_teams>::test): Convert to...
548 (is_a_helper <gimple_statement_omp_teams *>::test): ...this.
549 (is_a_helper <gimple_statement_omp_task>::test): Convert to...
550 (is_a_helper <gimple_statement_omp_task *>::test): ...this.
551 (is_a_helper <gimple_statement_phi>::test): Convert to...
552 (is_a_helper <gimple_statement_phi *>::test): ...this.
553 (is_a_helper <gimple_statement_transaction>::test): Convert to...
554 (is_a_helper <gimple_statement_transaction *>::test): ...this.
555 (is_a_helper <gimple_statement_try>::test): Convert to...
556 (is_a_helper <gimple_statement_try *>::test): ...this.
557 (is_a_helper <gimple_statement_wce>::test): Convert to...
558 (is_a_helper <gimple_statement_wce *>::test): ...this.
559 (is_a_helper <const gimple_statement_asm>::test): Convert to...
560 (is_a_helper <const gimple_statement_asm *>::test): ...this.
561 (is_a_helper <const gimple_statement_bind>::test): Convert to...
562 (is_a_helper <const gimple_statement_bind *>::test): ...this.
563 (is_a_helper <const gimple_statement_call>::test): Convert to...
564 (is_a_helper <const gimple_statement_call *>::test): ...this.
565 (is_a_helper <const gimple_statement_catch>::test): Convert to...
566 (is_a_helper <const gimple_statement_catch *>::test): ...this.
567 (is_a_helper <const gimple_statement_resx>::test): Convert to...
568 (is_a_helper <const gimple_statement_resx *>::test): ...this.
569 (is_a_helper <const gimple_statement_eh_dispatch>::test):
571 (is_a_helper <const gimple_statement_eh_dispatch *>::test):
573 (is_a_helper <const gimple_statement_eh_filter>::test): Convert
575 (is_a_helper <const gimple_statement_eh_filter *>::test): ...this.
576 (is_a_helper <const gimple_statement_omp_atomic_load>::test):
578 (is_a_helper <const gimple_statement_omp_atomic_load *>::test):
580 (is_a_helper <const gimple_statement_omp_atomic_store>::test):
582 (is_a_helper <const gimple_statement_omp_atomic_store *>::test):
584 (is_a_helper <const gimple_statement_omp_return>::test): Convert
586 (is_a_helper <const gimple_statement_omp_return *>::test): ...this.
587 (is_a_helper <const gimple_statement_omp_continue>::test): Convert
589 (is_a_helper <const gimple_statement_omp_continue *>::test): ...this.
590 (is_a_helper <const gimple_statement_omp_critical>::test): Convert
592 (is_a_helper <const gimple_statement_omp_critical *>::test): ...this.
593 (is_a_helper <const gimple_statement_omp_for>::test): Convert to...
594 (is_a_helper <const gimple_statement_omp_for *>::test): ...this.
595 (is_a_helper <const gimple_statement_omp_taskreg>::test): Convert
597 (is_a_helper <const gimple_statement_omp_taskreg *>::test): ...this.
598 (is_a_helper <const gimple_statement_omp_parallel>::test): Convert
600 (is_a_helper <const gimple_statement_omp_parallel *>::test): ...this.
601 (is_a_helper <const gimple_statement_omp_target>::test): Convert
603 (is_a_helper <const gimple_statement_omp_target *>::test): ...this.
604 (is_a_helper <const gimple_statement_omp_sections>::test): Convert
606 (is_a_helper <const gimple_statement_omp_sections *>::test): ...this.
607 (is_a_helper <const gimple_statement_omp_single>::test): Convert
609 (is_a_helper <const gimple_statement_omp_single *>::test): ...this.
610 (is_a_helper <const gimple_statement_omp_teams>::test): Convert
612 (is_a_helper <const gimple_statement_omp_teams *>::test): ...this.
613 (is_a_helper <const gimple_statement_omp_task>::test): Convert
615 (is_a_helper <const gimple_statement_omp_task *>::test): ...this.
616 (is_a_helper <const gimple_statement_phi>::test): Convert to...
617 (is_a_helper <const gimple_statement_phi *>::test): ...this.
618 (is_a_helper <const gimple_statement_transaction>::test): Convert
620 (is_a_helper <const gimple_statement_transaction *>::test): ...this.
621 (is_a_helper <const gimple_statement_with_ops>::test): Convert
623 (is_a_helper <const gimple_statement_with_ops *>::test): ...this.
624 (is_a_helper <gimple_statement_with_ops>::test): Convert to...
625 (is_a_helper <gimple_statement_with_ops *>::test): ...this.
626 (is_a_helper <const gimple_statement_with_memory_ops>::test):
628 (is_a_helper <const gimple_statement_with_memory_ops *>::test):
630 (is_a_helper <gimple_statement_with_memory_ops>::test): Convert
632 (is_a_helper <gimple_statement_with_memory_ops *>::test): ...this.
634 (gimple_use_ops): Update for removal of implicit pointer from the
636 (gimple_set_use_ops): Likewise.
637 (gimple_vuse): Likewise.
638 (gimple_vdef): Likewise.
639 (gimple_vuse_ptr): Likewise.
640 (gimple_vdef_ptr): Likewise.
641 (gimple_set_vuse): Likewise.
642 (gimple_set_vdef): Likewise.
643 (gimple_omp_return_set_lhs): Likewise.
644 (gimple_omp_return_lhs): Likewise.
645 (gimple_omp_return_lhs_ptr): Likewise.
646 (gimple_call_fntype): Likewise.
647 (gimple_call_set_fntype): Likewise.
648 (gimple_call_set_internal_fn): Likewise.
649 (gimple_call_use_set): Likewise.
650 (gimple_call_clobber_set): Likewise.
651 (gimple_bind_vars): Likewise.
652 (gimple_bind_set_vars): Likewise.
653 (gimple_bind_body_ptr): Likewise.
654 (gimple_bind_set_body): Likewise.
655 (gimple_bind_add_stmt): Likewise.
656 (gimple_bind_block): Likewise.
657 (gimple_bind_set_block): Likewise.
658 (gimple_asm_ninputs): Likewise.
659 (gimple_asm_noutputs): Likewise.
660 (gimple_asm_nclobbers): Likewise.
661 (gimple_asm_nlabels): Likewise.
662 (gimple_asm_input_op): Likewise.
663 (gimple_asm_input_op_ptr): Likewise.
664 (gimple_asm_output_op): Likewise.
665 (gimple_asm_output_op_ptr): Likewise.
666 (gimple_asm_set_output_op): Likewise.
667 (gimple_asm_clobber_op): Likewise.
668 (gimple_asm_set_clobber_op): Likewise.
669 (gimple_asm_label_op): Likewise.
670 (gimple_asm_set_label_op): Likewise.
671 (gimple_asm_string): Likewise.
672 (gimple_catch_types): Likewise.
673 (gimple_catch_types_ptr): Likewise.
674 (gimple_catch_handler_ptr): Likewise.
675 (gimple_catch_set_types): Likewise.
676 (gimple_catch_set_handler): Likewise.
677 (gimple_eh_filter_types): Likewise.
678 (gimple_eh_filter_types_ptr): Likewise.
679 (gimple_eh_filter_failure_ptr): Likewise.
680 (gimple_eh_filter_set_types): Likewise.
681 (gimple_eh_filter_set_failure): Likewise.
682 (gimple_eh_must_not_throw_fndecl): Likewise.
683 (gimple_eh_must_not_throw_set_fndecl): Likewise.
684 (gimple_eh_else_n_body_ptr): Likewise.
685 (gimple_eh_else_e_body_ptr): Likewise.
686 (gimple_eh_else_set_n_body): Likewise.
687 (gimple_eh_else_set_e_body): Likewise.
688 (gimple_try_eval_ptr): Likewise.
689 (gimple_try_cleanup_ptr): Likewise.
690 (gimple_try_set_eval): Likewise.
691 (gimple_try_set_cleanup): Likewise.
692 (gimple_wce_cleanup_ptr): Likewise.
693 (gimple_wce_set_cleanup): Likewise.
694 (gimple_phi_capacity): Likewise.
695 (gimple_phi_num_args): Likewise.
696 (gimple_phi_result): Likewise.
697 (gimple_phi_result_ptr): Likewise.
698 (gimple_phi_set_result): Likewise.
699 (gimple_phi_arg): Likewise.
700 (gimple_phi_set_arg): Likewise.
701 (gimple_resx_region): Likewise.
702 (gimple_resx_set_region): Likewise.
703 (gimple_eh_dispatch_region): Likewise.
704 (gimple_eh_dispatch_set_region): Likewise.
705 (gimple_omp_critical_name): Likewise.
706 (gimple_omp_critical_name_ptr): Likewise.
707 (gimple_omp_critical_set_name): Likewise.
708 (gimple_omp_for_clauses): Likewise.
709 (gimple_omp_for_clauses_ptr): Likewise.
710 (gimple_omp_for_set_clauses): Likewise.
711 (gimple_omp_for_collapse): Likewise.
712 (gimple_omp_for_index): Likewise.
713 (gimple_omp_for_index_ptr): Likewise.
714 (gimple_omp_for_set_index): Likewise.
715 (gimple_omp_for_initial): Likewise.
716 (gimple_omp_for_initial_ptr): Likewise.
717 (gimple_omp_for_set_initial): Likewise.
718 (gimple_omp_for_final): Likewise.
719 (gimple_omp_for_final_ptr): Likewise.
720 (gimple_omp_for_set_final): Likewise.
721 (gimple_omp_for_incr): Likewise.
722 (gimple_omp_for_incr_ptr): Likewise.
723 (gimple_omp_for_set_incr): Likewise.
724 (gimple_omp_for_pre_body_ptr): Likewise.
725 (gimple_omp_for_set_pre_body): Likewise.
726 (gimple_omp_parallel_clauses): Likewise.
727 (gimple_omp_parallel_clauses_ptr): Likewise.
728 (gimple_omp_parallel_set_clauses): Likewise.
729 (gimple_omp_parallel_child_fn): Likewise.
730 (gimple_omp_parallel_child_fn_ptr): Likewise.
731 (gimple_omp_parallel_set_child_fn): Likewise.
732 (gimple_omp_parallel_data_arg): Likewise.
733 (gimple_omp_parallel_data_arg_ptr): Likewise.
734 (gimple_omp_parallel_set_data_arg): Likewise.
735 (gimple_omp_task_clauses): Likewise.
736 (gimple_omp_task_clauses_ptr): Likewise.
737 (gimple_omp_task_set_clauses): Likewise.
738 (gimple_omp_task_child_fn): Likewise.
739 (gimple_omp_task_child_fn_ptr): Likewise.
740 (gimple_omp_task_set_child_fn): Likewise.
741 (gimple_omp_task_data_arg): Likewise.
742 (gimple_omp_task_data_arg_ptr): Likewise.
743 (gimple_omp_task_set_data_arg): Likewise.
744 (gimple_omp_taskreg_clauses): Likewise.
745 (gimple_omp_taskreg_clauses_ptr): Likewise.
746 (gimple_omp_taskreg_set_clauses): Likewise.
747 (gimple_omp_taskreg_child_fn): Likewise.
748 (gimple_omp_taskreg_child_fn_ptr): Likewise.
749 (gimple_omp_taskreg_set_child_fn): Likewise.
750 (gimple_omp_taskreg_data_arg): Likewise.
751 (gimple_omp_taskreg_data_arg_ptr): Likewise.
752 (gimple_omp_taskreg_set_data_arg): Likewise.
753 (gimple_omp_task_copy_fn): Likewise.
754 (gimple_omp_task_copy_fn_ptr): Likewise.
755 (gimple_omp_task_set_copy_fn): Likewise.
756 (gimple_omp_task_arg_size): Likewise.
757 (gimple_omp_task_arg_size_ptr): Likewise.
758 (gimple_omp_task_set_arg_size): Likewise.
759 (gimple_omp_task_arg_align): Likewise.
760 (gimple_omp_task_arg_align_ptr): Likewise.
761 (gimple_omp_task_set_arg_align): Likewise.
762 (gimple_omp_single_clauses): Likewise.
763 (gimple_omp_single_clauses_ptr): Likewise.
764 (gimple_omp_single_set_clauses): Likewise.
765 (gimple_omp_target_clauses): Likewise.
766 (gimple_omp_target_clauses_ptr): Likewise.
767 (gimple_omp_target_set_clauses): Likewise.
768 (gimple_omp_target_child_fn): Likewise.
769 (gimple_omp_target_child_fn_ptr): Likewise.
770 (gimple_omp_target_set_child_fn): Likewise.
771 (gimple_omp_target_data_arg): Likewise.
772 (gimple_omp_target_data_arg_ptr): Likewise.
773 (gimple_omp_target_set_data_arg): Likewise.
774 (gimple_omp_teams_clauses): Likewise.
775 (gimple_omp_teams_clauses_ptr): Likewise.
776 (gimple_omp_teams_set_clauses): Likewise.
777 (gimple_omp_sections_clauses): Likewise.
778 (gimple_omp_sections_clauses_ptr): Likewise.
779 (gimple_omp_sections_set_clauses): Likewise.
780 (gimple_omp_sections_control): Likewise.
781 (gimple_omp_sections_control_ptr): Likewise.
782 (gimple_omp_sections_set_control): Likewise.
783 (gimple_omp_for_set_cond): Likewise.
784 (gimple_omp_for_cond): Likewise.
785 (gimple_omp_atomic_store_set_val): Likewise.
786 (gimple_omp_atomic_store_val): Likewise.
787 (gimple_omp_atomic_store_val_ptr): Likewise.
788 (gimple_omp_atomic_load_set_lhs): Likewise.
789 (gimple_omp_atomic_load_lhs): Likewise.
790 (gimple_omp_atomic_load_lhs_ptr): Likewise.
791 (gimple_omp_atomic_load_set_rhs): Likewise.
792 (gimple_omp_atomic_load_rhs): Likewise.
793 (gimple_omp_atomic_load_rhs_ptr): Likewise.
794 (gimple_omp_continue_control_def): Likewise.
795 (gimple_omp_continue_control_def_ptr): Likewise.
796 (gimple_omp_continue_set_control_def): Likewise.
797 (gimple_omp_continue_control_use): Likewise.
798 (gimple_omp_continue_control_use_ptr): Likewise.
799 (gimple_omp_continue_set_control_use): Likewise.
800 (gimple_transaction_body_ptr): Likewise.
801 (gimple_transaction_label): Likewise.
802 (gimple_transaction_label_ptr): Likewise.
803 (gimple_transaction_set_body): Likewise.
804 (gimple_transaction_set_label): Likewise.
806 * ipa-devirt.c (build_type_inheritance_graph): Likewise.
807 * ipa-inline-analysis.c (inline_write_summary): Likewise.
808 * ipa-ref.c (ipa_record_reference): Likewise.
809 * ipa-reference.c (analyze_function): Likewise.
810 (ipa_reference_write_optimization_summary): Likewise.
811 * ipa.c (symtab_remove_unreachable_nodes): Likewise.
812 (address_taken_from_non_vtable_p): Likewise.
813 (comdat_can_be_unshared_p_1): Likewise.
814 * lto-cgraph.c (lto_output_ref): Likewise.
815 (add_references): Likewise.
816 (compute_ltrans_boundary): Likewise.
817 (output_symtab): Likewise.
818 (input_ref): Likewise.
819 (input_cgraph_1): Likewise.
820 (output_cgraph_opt_summary): Likewise.
821 * lto-streamer-out.c (lto_output): Likewise.
822 (output_symbol_p): Likewise.
823 * lto-streamer.h (lsei_next_function_in_partition): Likewise.
824 (lsei_start_function_in_partition): Likewise.
825 (lsei_next_variable_in_partition): Likewise.
826 (lsei_start_variable_in_partition): Likewise.
827 * symtab.c (insert_to_assembler_name_hash): Likewise.
828 (unlink_from_assembler_name_hash): Likewise.
829 (symtab_unregister_node): Likewise.
830 (symtab_remove_node): Likewise.
831 (dump_symtab_node): Likewise.
832 (verify_symtab_base): Likewise.
833 (verify_symtab_node): Likewise.
834 (symtab_make_decl_local): Likewise.
835 (symtab_alias_ultimate_target): Likewise.
836 (symtab_resolve_alias): Likewise.
837 (symtab_get_symbol_partitioning_class): Likewise.
838 * tree-phinodes.c (allocate_phi_node): Likewise.
839 (reserve_phi_args_for_new_edge): Likewise.
840 (remove_phi_args): Likewise.
841 * varpool.c (varpool_node_for_asm): Likewise.
842 (varpool_remove_unreferenced_decls): Likewise.
844 2014-04-23 Jeff Law <law@redhat.com>
846 PR tree-optimization/60902
847 * tree-ssa-threadedge.c
848 (record_temporary_equivalences_from_stmts_at_dest): Make sure to
849 invalidate outputs from statements that do not produce useful
850 outputs for threading.
852 2014-04-23 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
854 * config/aarch64/aarch64.md (stack_protect_set, stack_protect_test)
855 (stack_protect_set_<mode>, stack_protect_test_<mode>): Add
856 machine descriptions for Stack Smashing Protector.
858 2014-04-23 Richard Earnshaw <rearnsha@arm.com>
860 * aarch64.md (<optab>_rol<mode>3): New pattern.
861 (<optab>_rolsi3_uxtw): Likewise.
862 * aarch64.c (aarch64_strip_shift): Handle ROTATE and ROTATERT.
864 2014-04-23 James Greenhalgh <james.greenhalgh@arm.com>
866 * config/arm/arm.c (arm_cortex_a57_tune): Initialize all fields.
867 (arm_cortex_a12_tune): Likewise.
869 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
871 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle BSWAP.
873 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
875 * config/arm/arm.md (arm_rev16si2): New pattern.
876 (arm_rev16si2_alt): Likewise.
877 * config/arm/arm.c (arm_new_rtx_costs): Handle rev16 case.
879 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
881 * config/aarch64/aarch64.md (rev16<mode>2): New pattern.
882 (rev16<mode>2_alt): Likewise.
883 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle rev16 case.
884 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p): New.
885 (aarch_rev16_shleft_mask_imm_p): Likewise.
886 (aarch_rev16_p_1): Likewise.
887 (aarch_rev16_p): Likewise.
888 * config/arm/aarch-common-protos.h (aarch_rev16_p): Declare extern.
889 (aarch_rev16_shright_mask_imm_p): Likewise.
890 (aarch_rev16_shleft_mask_imm_p): Likewise.
892 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
894 * config/arm/aarch-common-protos.h (alu_cost_table): Add rev field.
895 * config/arm/aarch-cost-tables.h (generic_extra_costs): Specify
897 (cortex_a53_extra_costs): Likewise.
898 (cortex_a57_extra_costs): Likewise.
899 * config/arm/arm.c (cortexa9_extra_costs): Likewise.
900 (cortexa7_extra_costs): Likewise.
901 (cortexa8_extra_costs): Likewise.
902 (cortexa12_extra_costs): Likewise.
903 (cortexa15_extra_costs): Likewise.
904 (v7m_extra_costs): Likewise.
905 (arm_new_rtx_costs): Handle BSWAP.
907 2013-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
909 * config/arm/arm.c (cortexa8_extra_costs): New table.
910 (arm_cortex_a8_tune): New tuning struct.
911 * config/arm/arm-cores.def (cortex-a8): Use cortex_a8 tuning struct.
913 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
915 * config/arm/arm.c (arm_new_rtx_costs): Handle FMA.
917 2014-04-23 Richard Biener <rguenther@suse.de>
919 * Makefile.in (OBJS): Remove loop-unswitch.o.
920 * loop-unswitch.c: Delete.
921 * tree-pass.h (make_pass_rtl_unswitch): Remove.
922 * passes.def (pass_rtl_unswitch): Likewise.
923 * loop-init.c (gate_rtl_unswitch): Likewise.
924 (rtl_unswitch): Likewise.
925 (pass_data_rtl_unswitch): Likewise.
926 (pass_rtl_unswitch): Likewise.
927 (make_pass_rtl_unswitch): Likewise.
928 * rtl.h (reversed_condition): Likewise.
929 (compare_and_jump_seq): Likewise.
930 * loop-iv.c (reversed_condition): Move here from loop-unswitch.c
932 * loop-unroll.c (compare_and_jump_seq): Likewise.
934 2014-04-23 Richard Biener <rguenther@suse.de>
936 PR tree-optimization/60903
937 * tree-ssa-loop-im.c (analyze_memory_references): Remove
938 commented code block.
939 (execute_sm_if_changed): Properly apply IRREDUCIBLE_LOOP
940 loop flags to newly created BBs and edges.
942 2014-04-23 Nick Clifton <nickc@redhat.com>
944 * config/msp430/msp430.c (msp430_handle_option): Move function
946 (msp430_option_override): Simplify mcu and mcpu option handling.
947 (msp430_is_f5_mcu): Rename to msp430_use_f5_series_hwmult. Add
948 support for -mhwmult command line option.
949 (has_32bit_hwmult): Rename to use_32bit_hwmult. Add support for
950 -mhwmult command line option.
951 (msp430_hwmult_enabled): Delete.
952 (msp43o_output_labelref): Add support for -mhwmult command line
954 * config/msp430/msp430.md (mulhisi3, umulhisi3, mulsidi3)
955 (umulsidi3): Likewise.
956 * config/msp430/msp430.opt (mmcu): Add Report attribute.
957 (mcpu, mlarge, msmall): Likewise.
958 (mhwmult): New option.
959 * config/msp430/msp430-protos.h (msp430_hwmult_enabled): Remove
961 (msp430_is_f5_mcu): Remove prototype.
962 (msp430_use_f5_series_hwmult): Add prototype.
963 * config/msp430/msp430-opts.h: New file.
964 * common/config/msp430: New directory.
965 * common/config/msp430/msp430-common.c: New file.
966 * config.gcc (msp430): Remove target_has_targetm_common.
967 * doc/invoke.texi: Document -mhwmult command line option.
969 2014-04-23 Nick Clifton <nickc@redhat.com>
971 * config/i386/cygwin.h (ENDFILE_SPEC): Include
972 default-manifest.o if it can be found in the search path.
973 * config/i386/mingw32.h (ENDFILE_SPEC): Likewise.
975 2014-04-23 Terry Guo <terry.guo@arm.com>
977 * config/arm/arm.h (ASM_APP_OFF): Re-define it in a cleaner way.
979 2014-04-23 Richard Biener <rguenther@suse.de>
982 * tree-inline.c (declare_return_variable): Use mark_addressable.
984 2014-04-23 Richard Biener <rguenther@suse.de>
987 * loop-init.c (loop_optimizer_init): Make sure to apply
988 LOOPS_MAY_HAVE_MULTIPLE_LATCHES before fixing up loops.
990 2014-04-22 Jakub Jelinek <jakub@redhat.com>
993 * common.opt (fsanitize-recover, fsanitize-undefined-trap-on-error):
995 * gcc.c (sanitize_spec_function): Don't return "" for "undefined"
996 if flag_sanitize_undefined_trap_on_error.
997 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW_ABORT,
998 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS_ABORT,
999 BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE_ABORT,
1000 BUILT_IN_UBSAN_HANDLE_TYPE_MISMATCH_ABORT,
1001 BUILT_IN_UBSAN_HANDLE_ADD_OVERFLOW_ABORT,
1002 BUILT_IN_UBSAN_HANDLE_SUB_OVERFLOW_ABORT,
1003 BUILT_IN_UBSAN_HANDLE_MUL_OVERFLOW_ABORT,
1004 BUILT_IN_UBSAN_HANDLE_NEGATE_OVERFLOW_ABORT,
1005 BUILT_IN_UBSAN_HANDLE_LOAD_INVALID_VALUE_ABORT): New builtins.
1006 * ubsan.c (ubsan_instrument_unreachable): Return
1007 __builtin_trap () if flag_sanitize_undefined_trap_on_error.
1008 (ubsan_expand_null_ifn): Emit __builtin_trap ()
1009 if flag_sanitize_undefined_trap_on_error and
1010 __ubsan_handle_type_mismatch_abort if !flag_sanitize_recover.
1011 (ubsan_expand_null_ifn, ubsan_build_overflow_builtin,
1012 instrument_bool_enum_load): Emit __builtin_trap () if
1013 flag_sanitize_undefined_trap_on_error and
1014 __builtin_handle_*_abort () if !flag_sanitize_recover.
1015 * doc/invoke.texi (-fsanitize-recover,
1016 -fsanitize-undefined-trap-on-error): Document.
1018 2014-04-22 Christian Bruel <christian.bruel@st.com>
1020 * config/sh/sh.md (mov<mode>): Replace movQIHI.
1021 Force immediates to SImode.
1023 2014-04-22 Sandra Loosemore <sandra@codesourcery.com>
1025 * config/nios2/nios2.md (UNSPEC_ROUND): New.
1027 * config/nios2/nios2.opt (mno-custom-round, mcustom-round=): New.
1028 * config/nios2/nios2-opts.h (N2FPU_ALL_CODES): Add round.
1029 * config/nios2/nios2.c (N2F_NO_ERRNO): Define.
1030 (nios2_fpu_insn): Add entry for round.
1031 (N2FPU_NO_ERRNO_P): Define.
1032 (nios2_custom_check_insns): Add check for N2F_NO_ERRNO and
1034 * doc/invoke.texi (Nios II Options): Document -mcustom-round.
1036 2014-04-22 Richard Henderson <rth@redhat.com>
1038 * config/aarch64/aarch64 (addti3, subti3): New expanders.
1039 (add<GPI>3_compare0): Remove leading * from name.
1040 (add<GPI>3_carryin): Likewise.
1041 (sub<GPI>3_compare0): Likewise.
1042 (sub<GPI>3_carryin): Likewise.
1043 (<su_optab>mulditi3): New expander.
1044 (multi3): New expander.
1045 (madd<GPI>): Remove leading * from name.
1047 2014-04-22 Martin Jambor <mjambor@suse.cz>
1049 * cgraphclones.c (cgraph_function_versioning): Copy
1050 ipa_transforms_to_apply instead of asserting it is empty.
1052 2014-04-22 H.J. Lu <hongjiu.lu@intel.com>
1055 * config/i386/i386.c (ix86_expand_set_or_movmem): Call counter_mode
1056 on count_exp to get mode.
1058 2014-04-22 Andrew Pinski <apinski@cavium.com>
1060 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
1061 Handle TLS for ILP32.
1062 * config/aarch64/aarch64.md (tlsie_small): Rename to ...
1063 (tlsie_small_<mode>): this and handle PTR.
1064 (tlsie_small_sidi): New pattern.
1065 (tlsle_small): Change to an expand to handle ILP32.
1066 (tlsle_small_<mode>): New pattern.
1067 (tlsdesc_small): Rename to ...
1068 (tlsdesc_small_<mode>): this and handle PTR.
1070 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1072 * config/aarch64/aarch64.c (TARGET_FLAGS_REGNUM): Define.
1074 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1076 * config/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
1077 (aarch64_types_signed_unsigned_qualifiers): Qualifier added.
1078 (aarch64_types_signed_poly_qualifiers): Likewise.
1079 (aarch64_types_unsigned_signed_qualifiers): Likewise.
1080 (aarch64_types_poly_signed_qualifiers): Likewise.
1081 (TYPES_REINTERP_SS): Type macro added.
1082 (TYPES_REINTERP_SU): Likewise.
1083 (TYPES_REINTERP_SP): Likewise.
1084 (TYPES_REINTERP_US): Likewise.
1085 (TYPES_REINTERP_PS): Likewise.
1086 (aarch64_fold_builtin): New expression folding added.
1087 * config/aarch64/aarch64-simd-builtins.def (REINTERP):
1088 Declarations removed.
1089 (REINTERP_SS): Declarations added.
1090 (REINTERP_US): Likewise.
1091 (REINTERP_PS): Likewise.
1092 (REINTERP_SU): Likewise.
1093 (REINTERP_SP): Likewise.
1094 * config/aarch64/arm_neon.h (vreinterpret_p8_f64): Implemented.
1095 (vreinterpretq_p8_f64): Likewise.
1096 (vreinterpret_p16_f64): Likewise.
1097 (vreinterpretq_p16_f64): Likewise.
1098 (vreinterpret_f32_f64): Likewise.
1099 (vreinterpretq_f32_f64): Likewise.
1100 (vreinterpret_f64_f32): Likewise.
1101 (vreinterpret_f64_p8): Likewise.
1102 (vreinterpret_f64_p16): Likewise.
1103 (vreinterpret_f64_s8): Likewise.
1104 (vreinterpret_f64_s16): Likewise.
1105 (vreinterpret_f64_s32): Likewise.
1106 (vreinterpret_f64_s64): Likewise.
1107 (vreinterpret_f64_u8): Likewise.
1108 (vreinterpret_f64_u16): Likewise.
1109 (vreinterpret_f64_u32): Likewise.
1110 (vreinterpret_f64_u64): Likewise.
1111 (vreinterpretq_f64_f32): Likewise.
1112 (vreinterpretq_f64_p8): Likewise.
1113 (vreinterpretq_f64_p16): Likewise.
1114 (vreinterpretq_f64_s8): Likewise.
1115 (vreinterpretq_f64_s16): Likewise.
1116 (vreinterpretq_f64_s32): Likewise.
1117 (vreinterpretq_f64_s64): Likewise.
1118 (vreinterpretq_f64_u8): Likewise.
1119 (vreinterpretq_f64_u16): Likewise.
1120 (vreinterpretq_f64_u32): Likewise.
1121 (vreinterpretq_f64_u64): Likewise.
1122 (vreinterpret_s64_f64): Likewise.
1123 (vreinterpretq_s64_f64): Likewise.
1124 (vreinterpret_u64_f64): Likewise.
1125 (vreinterpretq_u64_f64): Likewise.
1126 (vreinterpret_s8_f64): Likewise.
1127 (vreinterpretq_s8_f64): Likewise.
1128 (vreinterpret_s16_f64): Likewise.
1129 (vreinterpretq_s16_f64): Likewise.
1130 (vreinterpret_s32_f64): Likewise.
1131 (vreinterpretq_s32_f64): Likewise.
1132 (vreinterpret_u8_f64): Likewise.
1133 (vreinterpretq_u8_f64): Likewise.
1134 (vreinterpret_u16_f64): Likewise.
1135 (vreinterpretq_u16_f64): Likewise.
1136 (vreinterpret_u32_f64): Likewise.
1137 (vreinterpretq_u32_f64): Likewise.
1139 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1141 * config/aarch64/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
1142 * config/aarch64/aarch64/aarch64-simd-builtins.def (REINTERP): Removed.
1143 (vreinterpret_p8_s8): Likewise.
1144 * config/aarch64/aarch64/arm_neon.h (vreinterpret_p8_s8): Uses cast.
1145 (vreinterpret_p8_s16): Likewise.
1146 (vreinterpret_p8_s32): Likewise.
1147 (vreinterpret_p8_s64): Likewise.
1148 (vreinterpret_p8_f32): Likewise.
1149 (vreinterpret_p8_u8): Likewise.
1150 (vreinterpret_p8_u16): Likewise.
1151 (vreinterpret_p8_u32): Likewise.
1152 (vreinterpret_p8_u64): Likewise.
1153 (vreinterpret_p8_p16): Likewise.
1154 (vreinterpretq_p8_s8): Likewise.
1155 (vreinterpretq_p8_s16): Likewise.
1156 (vreinterpretq_p8_s32): Likewise.
1157 (vreinterpretq_p8_s64): Likewise.
1158 (vreinterpretq_p8_f32): Likewise.
1159 (vreinterpretq_p8_u8): Likewise.
1160 (vreinterpretq_p8_u16): Likewise.
1161 (vreinterpretq_p8_u32): Likewise.
1162 (vreinterpretq_p8_u64): Likewise.
1163 (vreinterpretq_p8_p16): Likewise.
1164 (vreinterpret_p16_s8): Likewise.
1165 (vreinterpret_p16_s16): Likewise.
1166 (vreinterpret_p16_s32): Likewise.
1167 (vreinterpret_p16_s64): Likewise.
1168 (vreinterpret_p16_f32): Likewise.
1169 (vreinterpret_p16_u8): Likewise.
1170 (vreinterpret_p16_u16): Likewise.
1171 (vreinterpret_p16_u32): Likewise.
1172 (vreinterpret_p16_u64): Likewise.
1173 (vreinterpret_p16_p8): Likewise.
1174 (vreinterpretq_p16_s8): Likewise.
1175 (vreinterpretq_p16_s16): Likewise.
1176 (vreinterpretq_p16_s32): Likewise.
1177 (vreinterpretq_p16_s64): Likewise.
1178 (vreinterpretq_p16_f32): Likewise.
1179 (vreinterpretq_p16_u8): Likewise.
1180 (vreinterpretq_p16_u16): Likewise.
1181 (vreinterpretq_p16_u32): Likewise.
1182 (vreinterpretq_p16_u64): Likewise.
1183 (vreinterpretq_p16_p8): Likewise.
1184 (vreinterpret_f32_s8): Likewise.
1185 (vreinterpret_f32_s16): Likewise.
1186 (vreinterpret_f32_s32): Likewise.
1187 (vreinterpret_f32_s64): Likewise.
1188 (vreinterpret_f32_u8): Likewise.
1189 (vreinterpret_f32_u16): Likewise.
1190 (vreinterpret_f32_u32): Likewise.
1191 (vreinterpret_f32_u64): Likewise.
1192 (vreinterpret_f32_p8): Likewise.
1193 (vreinterpret_f32_p16): Likewise.
1194 (vreinterpretq_f32_s8): Likewise.
1195 (vreinterpretq_f32_s16): Likewise.
1196 (vreinterpretq_f32_s32): Likewise.
1197 (vreinterpretq_f32_s64): Likewise.
1198 (vreinterpretq_f32_u8): Likewise.
1199 (vreinterpretq_f32_u16): Likewise.
1200 (vreinterpretq_f32_u32): Likewise.
1201 (vreinterpretq_f32_u64): Likewise.
1202 (vreinterpretq_f32_p8): Likewise.
1203 (vreinterpretq_f32_p16): Likewise.
1204 (vreinterpret_s64_s8): Likewise.
1205 (vreinterpret_s64_s16): Likewise.
1206 (vreinterpret_s64_s32): Likewise.
1207 (vreinterpret_s64_f32): Likewise.
1208 (vreinterpret_s64_u8): Likewise.
1209 (vreinterpret_s64_u16): Likewise.
1210 (vreinterpret_s64_u32): Likewise.
1211 (vreinterpret_s64_u64): Likewise.
1212 (vreinterpret_s64_p8): Likewise.
1213 (vreinterpret_s64_p16): Likewise.
1214 (vreinterpretq_s64_s8): Likewise.
1215 (vreinterpretq_s64_s16): Likewise.
1216 (vreinterpretq_s64_s32): Likewise.
1217 (vreinterpretq_s64_f32): Likewise.
1218 (vreinterpretq_s64_u8): Likewise.
1219 (vreinterpretq_s64_u16): Likewise.
1220 (vreinterpretq_s64_u32): Likewise.
1221 (vreinterpretq_s64_u64): Likewise.
1222 (vreinterpretq_s64_p8): Likewise.
1223 (vreinterpretq_s64_p16): Likewise.
1224 (vreinterpret_u64_s8): Likewise.
1225 (vreinterpret_u64_s16): Likewise.
1226 (vreinterpret_u64_s32): Likewise.
1227 (vreinterpret_u64_s64): Likewise.
1228 (vreinterpret_u64_f32): Likewise.
1229 (vreinterpret_u64_u8): Likewise.
1230 (vreinterpret_u64_u16): Likewise.
1231 (vreinterpret_u64_u32): Likewise.
1232 (vreinterpret_u64_p8): Likewise.
1233 (vreinterpret_u64_p16): Likewise.
1234 (vreinterpretq_u64_s8): Likewise.
1235 (vreinterpretq_u64_s16): Likewise.
1236 (vreinterpretq_u64_s32): Likewise.
1237 (vreinterpretq_u64_s64): Likewise.
1238 (vreinterpretq_u64_f32): Likewise.
1239 (vreinterpretq_u64_u8): Likewise.
1240 (vreinterpretq_u64_u16): Likewise.
1241 (vreinterpretq_u64_u32): Likewise.
1242 (vreinterpretq_u64_p8): Likewise.
1243 (vreinterpretq_u64_p16): Likewise.
1244 (vreinterpret_s8_s16): Likewise.
1245 (vreinterpret_s8_s32): Likewise.
1246 (vreinterpret_s8_s64): Likewise.
1247 (vreinterpret_s8_f32): Likewise.
1248 (vreinterpret_s8_u8): Likewise.
1249 (vreinterpret_s8_u16): Likewise.
1250 (vreinterpret_s8_u32): Likewise.
1251 (vreinterpret_s8_u64): Likewise.
1252 (vreinterpret_s8_p8): Likewise.
1253 (vreinterpret_s8_p16): Likewise.
1254 (vreinterpretq_s8_s16): Likewise.
1255 (vreinterpretq_s8_s32): Likewise.
1256 (vreinterpretq_s8_s64): Likewise.
1257 (vreinterpretq_s8_f32): Likewise.
1258 (vreinterpretq_s8_u8): Likewise.
1259 (vreinterpretq_s8_u16): Likewise.
1260 (vreinterpretq_s8_u32): Likewise.
1261 (vreinterpretq_s8_u64): Likewise.
1262 (vreinterpretq_s8_p8): Likewise.
1263 (vreinterpretq_s8_p16): Likewise.
1264 (vreinterpret_s16_s8): Likewise.
1265 (vreinterpret_s16_s32): Likewise.
1266 (vreinterpret_s16_s64): Likewise.
1267 (vreinterpret_s16_f32): Likewise.
1268 (vreinterpret_s16_u8): Likewise.
1269 (vreinterpret_s16_u16): Likewise.
1270 (vreinterpret_s16_u32): Likewise.
1271 (vreinterpret_s16_u64): Likewise.
1272 (vreinterpret_s16_p8): Likewise.
1273 (vreinterpret_s16_p16): Likewise.
1274 (vreinterpretq_s16_s8): Likewise.
1275 (vreinterpretq_s16_s32): Likewise.
1276 (vreinterpretq_s16_s64): Likewise.
1277 (vreinterpretq_s16_f32): Likewise.
1278 (vreinterpretq_s16_u8): Likewise.
1279 (vreinterpretq_s16_u16): Likewise.
1280 (vreinterpretq_s16_u32): Likewise.
1281 (vreinterpretq_s16_u64): Likewise.
1282 (vreinterpretq_s16_p8): Likewise.
1283 (vreinterpretq_s16_p16): Likewise.
1284 (vreinterpret_s32_s8): Likewise.
1285 (vreinterpret_s32_s16): Likewise.
1286 (vreinterpret_s32_s64): Likewise.
1287 (vreinterpret_s32_f32): Likewise.
1288 (vreinterpret_s32_u8): Likewise.
1289 (vreinterpret_s32_u16): Likewise.
1290 (vreinterpret_s32_u32): Likewise.
1291 (vreinterpret_s32_u64): Likewise.
1292 (vreinterpret_s32_p8): Likewise.
1293 (vreinterpret_s32_p16): Likewise.
1294 (vreinterpretq_s32_s8): Likewise.
1295 (vreinterpretq_s32_s16): Likewise.
1296 (vreinterpretq_s32_s64): Likewise.
1297 (vreinterpretq_s32_f32): Likewise.
1298 (vreinterpretq_s32_u8): Likewise.
1299 (vreinterpretq_s32_u16): Likewise.
1300 (vreinterpretq_s32_u32): Likewise.
1301 (vreinterpretq_s32_u64): Likewise.
1302 (vreinterpretq_s32_p8): Likewise.
1303 (vreinterpretq_s32_p16): Likewise.
1304 (vreinterpret_u8_s8): Likewise.
1305 (vreinterpret_u8_s16): Likewise.
1306 (vreinterpret_u8_s32): Likewise.
1307 (vreinterpret_u8_s64): Likewise.
1308 (vreinterpret_u8_f32): Likewise.
1309 (vreinterpret_u8_u16): Likewise.
1310 (vreinterpret_u8_u32): Likewise.
1311 (vreinterpret_u8_u64): Likewise.
1312 (vreinterpret_u8_p8): Likewise.
1313 (vreinterpret_u8_p16): Likewise.
1314 (vreinterpretq_u8_s8): Likewise.
1315 (vreinterpretq_u8_s16): Likewise.
1316 (vreinterpretq_u8_s32): Likewise.
1317 (vreinterpretq_u8_s64): Likewise.
1318 (vreinterpretq_u8_f32): Likewise.
1319 (vreinterpretq_u8_u16): Likewise.
1320 (vreinterpretq_u8_u32): Likewise.
1321 (vreinterpretq_u8_u64): Likewise.
1322 (vreinterpretq_u8_p8): Likewise.
1323 (vreinterpretq_u8_p16): Likewise.
1324 (vreinterpret_u16_s8): Likewise.
1325 (vreinterpret_u16_s16): Likewise.
1326 (vreinterpret_u16_s32): Likewise.
1327 (vreinterpret_u16_s64): Likewise.
1328 (vreinterpret_u16_f32): Likewise.
1329 (vreinterpret_u16_u8): Likewise.
1330 (vreinterpret_u16_u32): Likewise.
1331 (vreinterpret_u16_u64): Likewise.
1332 (vreinterpret_u16_p8): Likewise.
1333 (vreinterpret_u16_p16): Likewise.
1334 (vreinterpretq_u16_s8): Likewise.
1335 (vreinterpretq_u16_s16): Likewise.
1336 (vreinterpretq_u16_s32): Likewise.
1337 (vreinterpretq_u16_s64): Likewise.
1338 (vreinterpretq_u16_f32): Likewise.
1339 (vreinterpretq_u16_u8): Likewise.
1340 (vreinterpretq_u16_u32): Likewise.
1341 (vreinterpretq_u16_u64): Likewise.
1342 (vreinterpretq_u16_p8): Likewise.
1343 (vreinterpretq_u16_p16): Likewise.
1344 (vreinterpret_u32_s8): Likewise.
1345 (vreinterpret_u32_s16): Likewise.
1346 (vreinterpret_u32_s32): Likewise.
1347 (vreinterpret_u32_s64): Likewise.
1348 (vreinterpret_u32_f32): Likewise.
1349 (vreinterpret_u32_u8): Likewise.
1350 (vreinterpret_u32_u16): Likewise.
1351 (vreinterpret_u32_u64): Likewise.
1352 (vreinterpret_u32_p8): Likewise.
1353 (vreinterpret_u32_p16): Likewise.
1354 (vreinterpretq_u32_s8): Likewise.
1355 (vreinterpretq_u32_s16): Likewise.
1356 (vreinterpretq_u32_s32): Likewise.
1357 (vreinterpretq_u32_s64): Likewise.
1358 (vreinterpretq_u32_f32): Likewise.
1359 (vreinterpretq_u32_u8): Likewise.
1360 (vreinterpretq_u32_u16): Likewise.
1361 (vreinterpretq_u32_u64): Likewise.
1362 (vreinterpretq_u32_p8): Likewise.
1363 (vreinterpretq_u32_p16): Likewise.
1365 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1367 * gcc/config/aarch64/aarch64-simd.md (aarch64_s<optab><mode>):
1369 * config/aarch64/aarch64-simd-builtins.def (sqneg): Iterator
1372 * config/aarch64/arm_neon.h (vqneg_s64): New intrinsic.
1373 (vqnegd_s64): Likewise.
1374 (vqabs_s64): Likewise.
1375 (vqabsd_s64): Likewise.
1377 2014-04-22 Richard Henderson <rth@redhat.com>
1379 * config/sparc/sparc.c (sparc_init_modes): Hoist GET_MODE_SIZE
1380 computation to the top of the loop.
1382 2014-04-22 Renlin <renlin.li@arm.com>
1383 Jiong Wang <jiong.wang@arm.com>
1385 * config/aarch64/aarch64.h (aarch64_frame): Delete "fp_lr_offset".
1386 * config/aarch64/aarch64.c (aarch64_layout_frame)
1387 (aarch64_initial_elimination_offset): Likewise.
1389 2014-04-22 Marcus Shawcroft <marcus.shawcroft@arm.com>
1391 * config/aarch64/aarch64.c (aarch64_initial_elimination_offset):
1394 2014-04-22 Richard Sandiford <rdsandiford@googlemail.com>
1396 * machmode.h (bitwise_mode_for_mode): Declare.
1397 * stor-layout.h (bitwise_type_for_mode): Likewise.
1398 * stor-layout.c (bitwise_mode_for_mode): New function.
1399 (bitwise_type_for_mode): Likewise.
1400 * builtins.c (fold_builtin_memory_op): Use it instead of
1401 int_mode_for_mode and build_nonstandard_integer_type.
1403 2014-04-22 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1405 * config.gcc (enable_obsolete): Remove *-*-solaris2.9*.
1406 (*-*-solaris2.[0-9] | *-*-solaris2.[0-9].*): Mark unsupported.
1407 (*-*-solaris2*): Simplify.
1408 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Likewise.
1409 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Remove
1410 *-*-solaris2.9* handling.
1412 * configure.ac (gcc_cv_as_hidden): Remove test for Solaris 9/x86
1414 (gcc_cv_ld_hidden): Remove *-*-solaris2.9* handling.
1415 (ld_tls_support): Remove i?86-*-solaris2.9, sparc*-*-solaris2.9
1417 (gcc_cv_as_gstabs_flag): Remove workaround for Solaris 9/x86 as bug.
1418 * configure: Regenerate.
1420 * config/i386/sol2-9.h: Remove.
1422 * doc/install.texi (Specific, i?86-*-solaris2.9): Remove.
1423 (Specific, *-*-solaris2*): Mention Solaris 9 support removal.
1424 Remove Solaris 9 references.
1426 2014-04-22 Vidya Praveen <vidyapraveen@arm.com>
1428 * aarch64.md (float<GPI:mode><GPF:mode>2): Remove.
1429 (floatuns<GPI:mode><GPF:mode>2): Remove.
1430 (<optab><fcvt_target><GPF:mode>2): New pattern for equal width float
1431 and floatuns conversions.
1432 (<optab><fcvt_iesize><GPF:mode>2): New pattern for inequal width float
1433 and floatuns conversions.
1434 * iterators.md (fcvt_target, FCVT_TARGET): Support SF and DF modes.
1435 (w1,w2): New mode attributes for inequal width conversions.
1437 2014-04-22 Renlin Li <Renlin.Li@arm.com>
1439 * config/aarch64/aarch64.c (aarch64_print_operand_address): Adjust
1440 the output asm format.
1442 2014-04-22 James Greenhalgh <james.greenhalgh@arm.com>
1444 * config/aarch64/aarch64-simd.md
1445 (aarch64_cm<optab>di): Always split.
1446 (*aarch64_cm<optab>di): New.
1447 (aarch64_cmtstdi): Always split.
1448 (*aarch64_cmtstdi): New.
1450 2014-04-22 Jakub Jelinek <jakub@redhat.com>
1452 PR tree-optimization/60823
1453 * omp-low.c (ipa_simd_modify_function_body): Go through
1454 all SSA_NAMEs and for those refering to vector arguments
1455 which are going to be replaced adjust SSA_NAME_VAR and,
1456 if it is a default definition, change it into a non-default
1457 definition assigned at the beginning of function from new_decl.
1458 (ipa_simd_modify_stmt_ops): Rewritten.
1459 * tree-dfa.c (set_ssa_default_def): When removing default def,
1460 check for NULL loc instead of NULL *loc.
1462 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1464 * config/arm/arm.c (arm_hard_regno_mode_ok): Loosen
1465 restrictions on core registers for DImode values in Thumb2.
1467 2014-04-22 Ian Bolton <ian.bolton@arm.com>
1469 * config/arm/arm.md (*anddi_notdi_zesidi): New pattern.
1470 * config/arm/thumb2.md (*iordi_notdi_zesidi): New pattern.
1472 2014-04-22 Ian Bolton <ian.bolton@arm.com>
1474 * config/arm/thumb2.md (*iordi_notdi_di): New pattern.
1475 (*iordi_notzesidi_di): Likewise.
1476 (*iordi_notsesidi_di): Likewise.
1478 2014-04-22 Ian Bolton <ian.bolton@arm.com>
1480 * config/arm/arm-protos.h (tune_params): New struct members.
1481 * config/arm/arm.c: Initialise tune_params per processor.
1482 (thumb2_reorg): Suppress conversion from t32 to t16 when optimizing
1483 for speed, based on new tune_params.
1485 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1487 * config/aarch64/aarch64-builtins.c (BUILTIN_VDQF_DF): Macro
1489 * config/aarch64/aarch64-simd-builtins.def (frintn): Use added
1491 * config/aarch64/aarch64-simd.md (<frint_pattern>): Comment
1493 * config/aarch64/aarch64.md (<frint_pattern>): Likewise.
1494 * config/aarch64/arm_neon.h (vrnd_f64): Added.
1495 (vrnda_f64): Likewise.
1496 (vrndi_f64): Likewise.
1497 (vrndm_f64): Likewise.
1498 (vrndn_f64): Likewise.
1499 (vrndp_f64): Likewise.
1500 (vrndx_f64): Likewise.
1502 2014-04-22 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1504 * config/arm/arm.c (arm_print_operand, thumb_exit): Make sure
1505 GET_MODE_SIZE argument is enum machine_mode.
1507 2014-04-22 Jakub Jelinek <jakub@redhat.com>
1510 * config/sparc/sparc.c (sparc_init_modes): Pass enum machine_mode
1511 value instead of int to GET_MODE_CLASS and GET_MODE_SIZE macros.
1513 2014-04-22 Lin Zuojian <manjian2006@gmail.com>
1516 * asan.c (asan_emit_stack_protection): Force the base to align to
1517 appropriate bits if STRICT_ALIGNMENT. Set shadow_mem align to
1518 appropriate bits if STRICT_ALIGNMENT.
1519 * cfgexpand.c (expand_stack_vars): Set base_align appropriately
1521 (expand_used_vars): Leave a space in the stack frame for alignment
1522 if STRICT_ALIGNMENT.
1524 2014-04-21 David Malcolm <dmalcolm@redhat.com>
1526 * gimple.h (gimple_assign_single_p): Accept a const_gimple rather
1528 (gimple_store_p): Likewise.
1529 (gimple_assign_load_p): Likewise.
1530 (gimple_assign_cast_p): Likewise.
1531 (gimple_clobber_p): Likewise.
1533 * doc/gimple.texi (gimple_assign_cast_p): Accept a const_gimple
1534 rather than a gimple.
1535 (gimple_assign_cast_p): Likewise.
1537 2014-04-21 Michael Meissner <meissner@linux.vnet.ibm.com>
1540 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64 case):
1541 If mode is DDmode and TARGET_E500_DOUBLE allow move.
1543 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print some
1544 more debug information for E500 if -mdebug=reg.
1546 2014-04-21 Uros Bizjak <ubizjak@gmail.com>
1549 * config/i386/i386.c (ix86_expand_builtin)
1550 <case IX86_BUILTIN_RDRAND{16,32,64}_STEP>: Use temporary
1551 register for target RTX.
1552 <case IX86_BUILTIN_RDSEED{16,32,64}_STEP>: Ditto.
1554 2014-04-18 Cong Hou <congh@google.com>
1556 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): Enhance
1557 the widen-mult pattern by handling two operands with different sizes,
1558 and operands whose size is smaller than half of the result type.
1560 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
1562 * ipa-inline.h (INLINE_HINT_known_hot): New hint.
1563 * ipa-inline-analysis.c (dump_inline_hints): Dump it.
1564 (do_estimate_edge_time): Compute it.
1565 * ipa-inline.c (want_inline_small_function_p): Bypass
1566 INLINE_INSNS_AUTO/SINGLE limits for calls that are known to be hot.
1568 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
1570 * ipa-inline.c (spec_rem): New static variable.
1571 (dump_overall_stats): New function.
1572 (dump_inline_stats): New function.
1574 2014-04-18 Richard Henderson <rth@redhat.com>
1576 * config/aarch64/aarch64.c (aarch64_register_move_cost): Pass a mode
1577 to GET_MODE_SIZE, not a reg_class_t.
1579 2014-04-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1581 * config/rs6000/vsx.md (vsx_xxmrghw_<mode>): Adjust for little-endian.
1582 (vsx_xxmrglw_<mode>): Likewise.
1584 2014-04-17 Michael Meissner <meissner@linux.vnet.ibm.com>
1587 * config/rs6000/rs6000.c (rs6000_setup_reg_addr_masks): Make sure
1588 GET_MODE_SIZE gets passed an enum machine_mode type and not integer.
1589 (rs6000_init_hard_regno_mode_ok): Likewise.
1591 2014-04-17 Jan Hubicka <hubicka@ucw.cz>
1593 * ipa-inline.c (inline_small_functions): Account only non-cold
1595 * doc/invoke.texi (inline-unit-growth): Update documentation.
1597 2014-04-17 Pat Haugen <pthaugen@us.ibm.com>
1599 * config/rs6000/rs6000.md (addti3, subti3): New.
1601 2014-04-17 H.J. Lu <hongjiu.lu@intel.com>
1604 * config/i386/i386.c (ix86_expand_clear): Remove outdated
1605 comment. Check optimize_insn_for_size_p instead of
1606 optimize_insn_for_speed_p.
1608 2014-04-17 Martin Jambor <mjambor@suse.cz>
1610 * gimple-iterator.c (gsi_start_edge): New function.
1611 * gimple-iterator.h (gsi_start_edge): Declare.
1612 * tree-sra.c (single_non_eh_succ): New function.
1613 (disqualify_ops_if_throwing_stmt): Renamed to
1614 disqualify_if_bad_bb_terminating_stmt. Allow throwing statements
1615 having one non-EH successor BB.
1616 (sra_modify_expr): If stmt ends bb, use single non-EH successor to
1617 generate loads into replacements.
1618 (sra_modify_assign): Likewise and and also use the simple path for
1620 (sra_modify_function_body): Commit statements on edges.
1622 2014-04-17 Richard Biener <rguenther@suse.de>
1625 * tree-ssa-propagate.c (valid_gimple_rhs_p): Allow vector
1626 comparison results and add clarifying comment.
1628 2014-04-17 Jakub Jelinek <jakub@redhat.com>
1630 * genmodes.c (struct mode_data): Add need_bytesize_adj field.
1631 (blank_mode): Initialize it.
1632 (emit_mode_size_inline, emit_mode_nunits_inline,
1633 emit_mode_inner_inline): New functions.
1634 (emit_insn_modes_h): Call them and surround their output with
1635 #if GCC_VERSION >= 4001 ... #endif.
1636 * machmode.h (GET_MODE_SIZE, GET_MODE_NUNITS, GET_MODE_INNER):
1637 For GCC_VERSION >= 4001 use mode_*_inline routines instead of
1638 mode_* arrays if the argument is __builtin_constant_p.
1639 * lower-subreg.c (dump_choices): Make sure GET_MODE_SIZE argument
1640 is enum machine_mode.
1642 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1644 * passes.c (opt_pass::execute): Adjust.
1645 (pass_manager::execute_pass_mode_switching): Likewise.
1646 (early_local_passes::execute): Likewise.
1647 (execute_one_pass): Pass cfun to the pass's execute method.
1648 * tree-pass.h (opt_pass::execute): Add function * argument.
1649 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
1650 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
1651 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
1652 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
1653 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
1654 config/sparc/sparc.c, cprop.c, dce.c, df-core.c, dse.c, dwarf2cfi.c,
1655 except.c, final.c, function.c, fwprop.c, gcse.c, gimple-low.c,
1656 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
1657 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
1658 ipa-inline-analysis.c, ipa-inline.c, ipa-profile.c, ipa-pure-const.c,
1659 ipa-reference.c, ipa-split.c, ipa.c, ira.c, jump.c, loop-init.c,
1660 lower-subreg.c, mode-switching.c, omp-low.c, postreload-gcse.c,
1661 postreload.c, predict.c, recog.c, ree.c, reg-stack.c, regcprop.c,
1662 reginfo.c, regrename.c, reorg.c, sched-rgn.c, stack-ptr-mod.c,
1663 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
1664 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
1665 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
1666 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-ssa-ccp.c,
1667 tree-ssa-copy.c, tree-ssa-copyrename.c, tree-ssa-dce.c,
1668 tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
1669 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
1670 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
1671 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
1672 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
1673 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
1674 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
1675 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
1676 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
1677 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
1680 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1682 * passes.c (opt_pass::gate): Take function * argument.
1683 (gate_all_early_local_passes): Merge into
1684 (early_local_passes::gate): this.
1685 (gate_all_early_optimizations): Merge into
1686 (all_early_optimizations::gate): this.
1687 (gate_all_optimizations): Mege into
1688 (all_optimizations::gate): this.
1689 (gate_all_optimizations_g): Merge into
1690 (all_optimizations_g::gate): this.
1691 (gate_rest_of_compilation): Mege into
1692 (rest_of_compilation::gate): this.
1693 (gate_postreload): Merge into
1694 (postreload::gate): this.
1695 (dump_one_pass): Pass cfun to the pass's gate method.
1696 (execute_ipa_summary_passes): Likewise.
1697 (execute_one_pass): Likewise.
1698 (ipa_write_summaries_2): Likewise.
1699 (ipa_write_optimization_summaries_1): Likewise.
1700 (ipa_read_summaries_1): Likewise.
1701 (ipa_read_optimization_summaries_1): Likewise.
1702 (execute_ipa_stmt_fixups): Likewise.
1703 * tree-pass.h (opt_pass::gate): Add function * argument.
1704 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c,
1705 combine-stack-adj.c, combine.c, compare-elim.c,
1706 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
1707 config/rl78/rl78.c, config/sh/sh_optimize_sett_clrt.cc,
1708 config/sh/sh_treg_combine.cc, config/sparc/sparc.c, cprop.c, cse.c,
1709 dce.c, df-core.c, dse.c, dwarf2cfi.c, except.c, fwprop.c, gcse.c,
1710 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
1711 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
1712 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
1713 loop-init.c, lower-subreg.c, mode-switching.c, modulo-sched.c,
1714 omp-low.c, postreload-gcse.c, postreload.c, predict.c, recog.c, ree.c,
1715 reg-stack.c, regcprop.c, regrename.c, reorg.c, sched-rgn.c,
1716 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
1717 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
1718 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c,
1719 tree-nrv.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
1720 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
1721 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
1722 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
1723 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
1724 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
1725 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
1726 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
1727 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
1728 tree-ssa.c, tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
1729 tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c, tsan.c, ubsan.c,
1730 var-tracking.c, vtable-verify.c, web.c: Adjust.
1732 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1734 * configure.ac: Check for -Woverloaded-virtual and enable it if found.
1735 * configure: Regenerate.
1737 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1739 * passes.c (dump_one_pass): don't check pass->has_gate.
1740 (execute_ipa_summary_passes): Likewise.
1741 (execute_one_pass): Likewise.
1742 (ipa_write_summaries_2): Likewise.
1743 (ipa_write_optimization_summaries_1): Likewise.
1744 (ipa_read_optimization_summaries_1): Likewise.
1745 (execute_ipa_stmt_fixups): Likewise.
1746 * tree-pass.h (pass_data::has_gate): Remove.
1747 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
1748 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
1749 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
1750 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
1751 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
1752 config/sh/sh_optimize_sett_clrt.cc, config/sh/sh_treg_combine.cc,
1753 config/sparc/sparc.c, cprop.c, cse.c, dce.c, df-core.c, dse.c,
1754 dwarf2cfi.c, except.c, final.c, function.c, fwprop.c, gcse.c,
1755 gimple-low.c, gimple-ssa-isolate-paths.c,
1756 gimple-ssa-strength-reduction.c, graphite.c, ifcvt.c, init-regs.c,
1757 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
1758 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
1759 ira.c, jump.c, loop-init.c, lower-subreg.c, mode-switching.c,
1760 modulo-sched.c, omp-low.c, postreload-gcse.c, postreload.c, predict.c,
1761 recog.c, ree.c, reg-stack.c, regcprop.c, reginfo.c, regrename.c,
1762 reorg.c, sched-rgn.c, stack-ptr-mod.c, store-motion.c, tracer.c,
1763 trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
1764 tree-complex.c, tree-eh.c, tree-emutls.c, tree-if-conv.c,
1765 tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
1766 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
1767 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
1768 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
1769 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
1770 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
1771 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
1772 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
1773 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
1774 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
1775 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
1776 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
1777 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
1780 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1782 * pass_manager.h (pass_manager::register_dump_files_1): Remove
1784 * passes.c (pass_manager::register_dump_files_1): Merge into
1785 (pass_manager::register_dump_files): this, and remove its handling of
1786 properties since the pass always has the properties anyway.
1787 (pass_manager::pass_manager): Adjust.
1789 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1791 * pass_manager.h (pass_manager::register_dump_files_1): Adjust.
1792 * passes.c (pass_manager::register_dump_files_1): Remove dead code
1793 dealing with properties.
1794 (pass_manager::register_dump_files): Adjust.
1796 2014-03-20 Mark Wielaard <mjw@redhat.com>
1798 * dwarf2out.c (add_bound_info): If HOST_WIDE_INT is big enough,
1799 then represent the bound as normal constant value.
1801 2014-04-17 Jakub Jelinek <jakub@redhat.com>
1804 Forward port from 4.8 branch
1805 2013-07-19 Kirill Yukhin <kirill.yukhin@intel.com>
1807 * config/i386/bmiintrin.h (_blsi_u32): New.
1811 (_blsmsk_u32): Ditto.
1812 (_blsmsk_u64): Ditto.
1813 (_tzcnt_u32): Ditto.
1814 (_tzcnt_u64): Ditto.
1816 2014-04-17 Kito Cheng <kito@0xlab.org>
1818 * gcc.c (used_arg): Prevent out of bound access for multilib_options.
1820 2014-04-17 Richard Biener <rguenther@suse.de>
1823 * tree-ssa-propagate.c (valid_gimple_rhs_p): Only allow effective
1824 boolean results for comparisons.
1826 2014-04-17 Richard Biener <rguenther@suse.de>
1828 PR tree-optimization/60836
1829 * tree-vect-loop.c (vect_create_epilog_for_reduction): Force
1830 initial PHI args to be gimple values.
1832 2014-04-17 Richard Biener <rguenther@suse.de>
1834 PR tree-optimization/60841
1835 * tree-vect-data-refs.c (vect_analyze_data_refs): Count stmts.
1836 * tree-vect-loop.c (vect_analyze_loop_2): Pass down number
1837 of stmts to SLP build.
1838 * tree-vect-slp.c (vect_slp_analyze_bb_1): Likewise.
1839 (vect_analyze_slp): Likewise.
1840 (vect_analyze_slp_instance): Likewise.
1841 (vect_build_slp_tree): Limit overall SLP tree growth.
1842 * tree-vectorizer.h (vect_analyze_data_refs,
1843 vect_analyze_slp): Adjust prototypes.
1845 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
1847 * config/i386/i386.c (x86_add_stmt_cost): Fix vector cost model for
1850 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
1852 * config/i386/x86-tune.def (TARGET_SLOW_PSHUFB): New tune definition.
1853 * config/i386/i386.h (TARGET_SLOW_PSHUFB): New tune flag.
1854 * config/i386/i386.c (expand_vec_perm_even_odd_1): Avoid byte shuffles
1855 for TARGET_SLOW_PSHUFB
1857 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
1859 * config/i386/i386.c (slm_cost): Adjust vec_to_scalar_cost.
1860 * config/i386/i386.c (intel_cost): Ditto.
1862 2014-04-17 Joey Ye <joey.ye@arm.com>
1864 * opts.c (OPT_fif_conversion, OPT_fif_conversion2): Disable for Og.
1866 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
1868 * opts.c (common_handle_option): Disable -fipa-reference coorectly
1871 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
1873 * ipa-devirt.c (odr_type_d): Add field all_derivations_known.
1874 (type_all_derivations_known_p): New predicate.
1875 (type_all_ctors_visible_p): New predicate.
1876 (type_possibly_instantiated_p): New predicate.
1877 (get_odr_type): Compute all_derivations_known.
1878 (dump_odr_type): Dump the flag.
1879 (maybe_record_type): Cleanup.
1880 (record_target_from_binfo): Add bases_to_consider array;
1881 record bases for types w/o instances and skip CXX destructor.
1882 (possible_polymorphic_call_targets_1): Add bases_to_consider
1883 and consider_construction parameters; check if type may have instance.
1884 (get_polymorphic_call_info): Set maybe_in_construction to true
1885 when we know nothing.
1886 (record_targets_from_bases): Skip CXX destructors; they are
1887 never called for types in construction.
1888 (possible_polymorphic_call_targets): Do not record target when
1889 type may not have instance.
1891 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
1894 * ipa.c (symtab_remove_unreachable_nodes): Mark targets of
1895 external aliases alive, too.
1897 2014-04-16 Andrew Pinski <apinski@cavium.com>
1899 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Change aarch64 ilp32
1902 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
1904 * final.c (compute_alignments): Do not apply loop alignment to a block
1905 falling through to the exit.
1907 2014-04-16 Catherine Moore <clm@codesourcery.com>
1909 * mips.md (*mov<mode>_internal, *movhi_internal, *movqi_internal):
1910 Adjust constraints for microMIPS store patterns.
1912 2014-04-16 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
1914 * config/avr/avr-mcus.def: Correct typo for atxmega256a3bu macro.
1916 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
1918 * tree-ssa-operands.c (create_vop_var): Set DECL_IGNORED_P.
1919 (append_use): Run at -O0.
1920 (append_vdef): Likewise.
1921 * tree-ssa-ter.c (ter_is_replaceable_p): Do not special-case -O0.
1922 * tree-ssa-uninit.c (warn_uninitialized_vars): Remove obsolete comment.
1924 2014-04-16 Jakub Jelinek <jakub@redhat.com>
1926 PR tree-optimization/60844
1927 * tree-ssa-reassoc.c (reassoc_remove_stmt): New function.
1928 (propagate_op_to_single_use, remove_visited_stmt_chain,
1929 linearize_expr, repropagate_negates, reassociate_bb): Use it
1930 instead of gsi_remove.
1932 2014-04-16 Martin Jambor <mjambor@suse.cz>
1934 * cgraphclones.c (cgraph_create_virtual_clone): Duplicate
1935 ipa_transforms_to_apply.
1936 (cgraph_function_versioning): Assert that old_node has empty
1937 ipa_transforms_to_apply.
1938 * trans-mem.c (ipa_tm_create_version): Likewise.
1939 * tree-inline.c (tree_function_versioning): Do not duplicate
1940 ipa_transforms_to_apply.
1942 2014-04-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1945 * configure.ac (set_have_as_tls): Merge i[34567]86-*-* and
1947 Pass necessary as flags on 64-bit Solaris/x86.
1948 Use lowercase relocs for x86_64-*-*.
1949 * configure: Regenerate.
1951 2014-04-15 Jan Hubicka <jh@suse.cz>
1953 * ipa-devirt.c (referenced_from_vtable_p): New predicate.
1954 (maybe_record_node, likely_target_p): Use it.
1956 2014-04-15 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1959 Revert following patch
1961 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
1964 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
1965 software floating point or no floating point registers, do not
1966 allow any type in the FPRs. Eliminate a test for SPE SIMD types
1967 in GPRs that occurs after we tested for GPRs that would never be
1970 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
1971 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
1972 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
1973 specifically allow DDmode, since that does not use the SPE SIMD
1976 2014-03-21 Mark Wielaard <mjw@redhat.com>
1978 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_const_value
1979 as unsigned or int depending on type and value used.
1981 2014-04-15 Richard Biener <rguenther@suse.de>
1983 PR rtl-optimization/56965
1984 * alias.c (ncr_compar, nonoverlapping_component_refs_p): Move ...
1985 * tree-ssa-alias.c (ncr_compar, nonoverlapping_component_refs_p):
1987 * alias.c (true_dependence_1): Do not call
1988 nonoverlapping_component_refs_p.
1989 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Call
1990 nonoverlapping_component_refs_p.
1991 (indirect_refs_may_alias_p): Likewise.
1993 2014-04-15 Teresa Johnson <tejohnson@google.com>
1995 * cfg.c (dump_bb_info): Fix flags check.
1996 * tree-cfg.c (remove_bb): Only dump TDF_BLOCKS when removing.
1998 2014-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2000 PR rtl-optimization/60663
2001 * config/arm/arm.c (arm_new_rtx_costs): Improve ASM_OPERANDS case,
2004 2014-04-15 Richard Biener <rguenther@suse.de>
2006 * lto-streamer.h (LTO_major_version): Bump to 4.
2008 2014-04-15 Richard Biener <rguenther@suse.de>
2010 * common.opt (lto_partition_model): New enum.
2011 (flto-partition=): Merge separate options with a single with argument,
2012 add -flto-partition=one support.
2013 * flag-types.h (enum lto_partition_model): Declare.
2014 * opts.c (finish_options): Remove duplicate -flto-partition=
2016 * lto-wrapper.c (run_gcc): Adjust.
2018 2014-04-15 Richard Biener <rguenther@suse.de>
2020 * alias.c (ncr_compar): New function.
2021 (nonoverlapping_component_refs_p): Re-implement in O (n log n).
2023 2014-04-15 Richard Biener <rguenther@suse.de>
2025 * alias.c (record_component_aliases): Do not walk BINFOs.
2027 2014-04-15 Richard Biener <rguenther@suse.de>
2029 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
2030 Add struct function argument and adjust.
2031 (find_func_aliases_for_call): Likewise.
2032 (find_func_aliases): Likewise.
2033 (find_func_clobbers): Likewise.
2034 (intra_create_variable_infos): Likewise.
2035 (compute_points_to_sets): Likewise.
2036 (ipa_pta_execute): Adjust. Do not push/pop cfun.
2038 2014-04-15 Richard Biener <rguenther@suse.de>
2040 * tree.c (iterative_hash_expr): Use enum tree_code_class
2041 to store TREE_CODE_CLASS.
2042 (tree_block): Likewise.
2043 (tree_set_block): Likewise.
2044 * tree.h (fold_build_pointer_plus_loc): Use
2045 convert_to_ptrofftype_loc.
2047 2014-04-15 Jakub Jelinek <jakub@redhat.com>
2050 * Makefile.in (PLUGIN_HEADERS): Add various headers that have been
2053 2014-04-15 Eric Botcazou <ebotcazou@adacore.com>
2055 * cfgloop.h (struct loop): Move force_vectorize down.
2056 * gimplify.c (gimple_boolify) <ANNOTATE_EXPR>: Handle new kinds.
2057 (gimplify_expr) <ANNOTATE_EXPR>: Minor tweak.
2058 * lto-streamer-in.c (input_cfg): Read dont_vectorize field.
2059 * lto-streamer-out.c (output_cfg): Write dont_vectorize field.
2060 * tree-cfg.c (replace_loop_annotate): Revamp and handle new kinds.
2061 * tree-core.h (enum annot_expr_kind): Add new kind values.
2062 * tree-inline.c (copy_loops): Copy dont_vectorize field and reorder.
2063 * tree-pretty-print.c (dump_generic_node) <ANNOTATE_EXPR>: Handle new
2065 * tree.def (ANNOTATE_EXPR): Tweak comment.
2067 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
2069 * ipa-devirt.c (maybe_record_node): Ignore all non-methods (including
2072 2014-04-14 Paolo Carlini <paolo.carlini@oracle.com>
2074 * tree.h (TYPE_IDENTIFIER): Declare.
2075 * tree.c (subrange_type_for_debug_p): Use it.
2076 * godump.c (go_format_type): Likewise.
2077 * dwarf2out.c (is_cxx_auto, modified_type_die,
2078 gen_type_die_with_usage, gen_type_die_with_usage): Likewise.
2079 * dbxout.c (dbxout_type, dbxout_symbol): Likewise.
2081 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
2084 * varpool.c (varpool_remove_node): Do not alter decls when streaming.
2086 2014-04-14 Uros Bizjak <ubizjak@gmail.com>
2088 * config/i386/i386.c (examine_argument): Return bool. Return true if
2089 parameter should be passed in memory.
2090 <case X86_64_COMPLEX_X87_CLASS>: Adjust.
2091 (construct_container): Update calls to examine_argument.
2092 (function_arg_advance_64): Ditto.
2093 (return_in_memory_32): Merge with ix86_return_in_memory.
2094 (return_in_memory_64): Ditto.
2095 (return_in_memory_ms_64): Ditto.
2097 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
2099 * ipa-utils.c (ipa_merge_profiles): Merge profile_id.
2100 * coverage.c (coverage_compute_profile_id): Handle externally visible
2103 2014-04-14 Martin Jambor <mjambor@suse.cz>
2105 * tree-sra.c (ipa_sra_preliminary_function_checks): Skip
2106 DECL_DISREGARD_INLINE_LIMITS functions.
2108 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
2111 * config/i386/i386.md (*fixuns_trunc<mode>_1): Revert the last change.
2113 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
2116 * config/i386/i386.md (*fixuns_trunc<mode>_1): Check
2117 optimize_insn_for_speed_p instead of
2118 optimize_function_for_speed_p.
2120 2014-04-14 Yufeng Zhang <yufeng.zhang@arm.com>
2122 * doc/invoke.texi (free): Document AArch64.
2124 2014-04-14 Richard Biener <rguenther@suse.de>
2126 PR tree-optimization/60042
2127 * tree-ssa-pre.c (inhibit_phi_insertion): Remove.
2128 (insert_into_preds_of_block): Do not prevent PHI insertion
2129 for REFERENCE exprs here ...
2130 (eliminate_dom_walker::before_dom_children): ... but prevent
2131 their use here under similar conditions when applied to the
2132 IL after PRE optimizations.
2134 2014-04-14 Richard Biener <rguenther@suse.de>
2136 * passes.def: Move early points-to after early SRA.
2138 2014-04-14 Richard Biener <rguenther@suse.de>
2140 * tree-ssa-forwprop.c (simplify_gimple_switch): Enhance
2141 check for which sign-changes we allow when forwarding
2142 a converted value into a switch.
2144 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
2146 * stor-layout.c (place_field): Finalize non-constant offset for the
2149 2014-04-14 Richard Biener <rguenther@suse.de>
2151 * tree-switch-conversion.c (lshift_cheap_p): Get speed_p
2153 (expand_switch_using_bit_tests_p): Likewise.
2154 (process_switch): Compute and pass on speed_p based on the
2156 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Use
2157 optimize_bb_for_speed_p.
2159 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
2161 * cfgloop.h (struct loop): Rename force_vect into force_vectorize.
2162 * function.h (struct function): Rename has_force_vect_loops into
2163 has_force_vectorize_loops.
2164 * lto-streamer-in.c (input_cfg): Adjust for renaming.
2165 (input_struct_function_base): Likewise.
2166 * lto-streamer-out.c (output_cfg): Likewise.
2167 (output_struct_function_base): Likewise.
2168 * omp-low.c (expand_omp_simd): Likewise.
2169 * tree-cfg.c (move_sese_region_to_fn): Likewise.
2170 * tree-if-conv.c (ifcvt_can_use_mask_load_store): Likewise.
2171 (version_loop_for_if_conversion): Likewise.
2172 (tree_if_conversion): Likewise.
2173 (main_tree_if_conversion): Likewise.
2174 (gate_tree_if_conversion): Likewise.
2175 * tree-inline.c (copy_loops): Likewise.
2176 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Likewise.
2177 * tree-ssa-loop.c (tree_loop_vectorize): Likewise.
2178 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Likewise.
2179 * tree-vect-loop.c (vect_estimate_min_profitable_iters): Likewise.
2180 * tree-vectorizer.c (vectorize_loops): Likewise.
2181 * tree-vectorizer.h (unlimited_cost_model): Likewise.
2183 2014-04-14 Richard Biener <rguenther@suse.de>
2186 * lto-streamer-out.c (wrap_refs): New function.
2187 (lto_output): Wrap symbol references in global initializes in
2188 type-preserving MEM_REFs.
2190 2014-04-14 Christian Bruel <christian.bruel@st.com>
2192 * config/sh/sh-mem.cc (sh_expand_strlen): Unroll last word.
2194 2014-04-14 Christian Bruel <christian.bruel@st.com>
2196 * config/sh/sh.md (setmemqi): New expand pattern.
2197 * config/sh/sh.h (CLEAR_RATIO): Define.
2198 * config/sh/sh-mem.cc (sh_expand_setmem): Define.
2199 * config/sh/sh-protos.h (sh_expand_setmem): Declare.
2201 2014-04-14 Richard Biener <rguenther@suse.de>
2204 * fold-const.c (negate_expr_p): Don't negate directional rounding
2206 (fold_negate_expr): Likewise.
2208 2014-04-14 Richard Biener <rguenther@suse.de>
2210 PR tree-optimization/59817
2211 PR tree-optimization/60453
2212 * graphite-scop-detection.c (graphite_can_represent_scev): Complete
2213 recursion to catch all CHRECs in the scalar evolution and restrict
2214 the predicate for the remains appropriately.
2216 2014-04-12 Catherine Moore <clm@codesourcery.com>
2218 * config/mips/constraints.md: Add new register constraint "kb".
2219 * config/mips/mips.md (*mov<mode>_internal): Use constraint "kb".
2220 (*movhi_internal): Likewise.
2221 (*movqi_internal): Likewise.
2222 * config/mips/mips.h (M16_STORE_REGS): New register class.
2223 (REG_CLASS_NAMES): Add M16_STORE_REGS.
2224 (REG_CLASS_CONTENTS): Likewise.
2225 * config/mips/mips.c (mips_regno_to_class): Add M16_STORE_REGS.
2227 2014-04-11 Tobias Burnus <burnus@net-b.de>
2230 * doc/invoke.texi (-Wformat-signedness): Document it.
2231 (Wformat=2): Mention that this enables -Wformat-signedness.
2233 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
2235 * common/config/epiphany/epiphany-common.c
2236 (epiphany_option_optimization_table): Enable section anchors by
2237 default at -O1 or higher.
2238 * config/epiphany/epiphany.c (TARGET_MAX_ANCHOR_OFFSET): Define.
2239 (TARGET_MIN_ANCHOR_OFFSET): Likewise.
2240 (epiphany_rtx_costs) <SET>: For binary operators, the set as such
2241 carries no extra cost.
2242 (epiphany_legitimate_address_p): For BLKmode, apply SImode check.
2243 * config/epiphany/epiphany.h (ASM_OUTPUT_DEF): Define.
2244 * config/epiphany/predicates.md (memclob_operand): New predicate.
2245 * config/epiphany/epiphany.md (stack_adjust_add, stack_adjust_str):
2246 Use memclob_operand predicate and X constraint for operand 3.
2248 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
2250 * config/epiphany/epiphany.c (epiphany_rtx_cost): Compare
2251 with CC_N_NE / CC_C_LTU / CC_C_GTU carries no extra cost for
2254 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
2256 PR rtl-optimization/60651
2257 * mode-switching.c (optimize_mode_switching): Make sure to emit
2258 sets of a lower numbered entity before sets of a higher numbered
2259 entity to a mode of the same or lower priority.
2260 When creating a seginfo for a basic block that starts with a code
2261 label, move the insertion point past the code label.
2262 (new_seginfo): Document and enforce requirement that
2263 NOTE_INSN_BASIC_BLOCK only appears for empty blocks.
2264 * doc/tm.texi.in: Document ordering constraint for emitted mode sets.
2265 * doc/tm.texi: Regenerate.
2267 2014-01-11 Joern Rennecke <joern.rennecke@embecosm.com>
2270 * config/arc/arc.c (arc_save_restore): Fix assert typo.
2272 2013-04-11 Jakub Jelinek <jakub@redhat.com>
2274 * BASE-VER: Set to 4.10.0.
2276 2014-04-11 Tobias Burnus <burnus@net-b.de>
2279 * doc/bugreport.texi (Bugs): Remove nodes pointing to the nirvana.
2280 * doc/gcc.texi (Service): Update description in the @menu
2281 * doc/invoke.texi (Option Summary): Remove misplaced and
2284 2014-04-11 Steve Ellcey <sellcey@mips.com>
2285 Jakub Jelinek <jakub@redhat.com>
2288 * expr.c (convert_move): Use emit_store_flag_force instead of
2289 emit_store_flag. Pass lowpart_mode instead of VOIDmode as 5th
2292 2014-04-11 Richard Biener <rguenther@suse.de>
2295 * varasm.c (assemble_alias): Avoid endless error reporting
2296 recursion by setting TREE_ASM_WRITTEN.
2298 2014-04-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2300 * config/s390/s390.md: Add a splitter for NOT rtx.
2302 2014-04-11 Jakub Jelinek <jakub@redhat.com>
2304 PR rtl-optimization/60663
2305 * cse.c (cse_insn): Set src_volatile on ASM_OPERANDS in PARALLEL.
2307 2014-04-10 Jan Hubicka <hubicka@ucw.cz>
2308 Jakub Jelinek <jakub@redhat.com>
2311 * ipa.c (function_and_variable_visibility): Copy forced_by_abi
2312 flag from decl_node to node.
2314 2014-04-10 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2317 * config/arm/arm.c (TARGET_CONST_NOT_OK_FOR_DEBUG_P): Define
2318 (arm_const_not_ok_for_debug_p): Reject MINUS with SYM_REF's
2319 ameliorating the cases where it can be.
2321 2014-04-09 David Edelsohn <dje.gcc@gmail.com>
2324 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
2326 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
2327 (loadsync_<mode>): Change mode.
2328 (load_quadpti, store_quadpti): New.
2329 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
2330 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
2331 * config/rs6000/predicates.md (quad_memory_operand): !TARGET_SYNC_TI.
2333 2014-04-09 Cong Hou <congh@google.com>
2336 * doc/sourcebuild.texi (vect_widen_mult_si_to_di_pattern): Add
2339 2014-04-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2341 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Use vnand
2342 instead of vnor to exploit possible fusion opportunity in the
2344 (altivec_expand_vec_perm_const_le): Likewise.
2346 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
2348 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
2349 (loadsync_<mode>): Change mode.
2350 (load_quadpti, store_quadpti): New.
2351 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
2352 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
2354 2014-04-08 Richard Sandiford <rdsandiford@googlemail.com>
2357 * config/rs6000/vsx.md (vsx_xscvdpspn_scalar): Change input to DImode.
2358 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Update accordingly.
2359 Use gen_rtx_REG rather than simplify_gen_subreg for op0_di.
2361 2014-04-08 Richard Biener <rguenther@suse.de>
2364 * tree-pretty-print.c (pp_double_int): For HWI32 hosts with
2365 a 64bit widest int print double-int similar to on HWI64 hosts.
2367 2014-04-08 Richard Biener <rguenther@suse.de>
2369 PR tree-optimization/60785
2370 * graphite-sese-to-poly.c (rewrite_phi_out_of_ssa): Treat
2371 default defs properly.
2373 2014-04-08 Nathan Sidwell <nathan@codesourcery.com>
2375 * doc/invoke (Wnon-virtual-dtor): Update to match implementation.
2376 (Weffc++): Likewise.
2378 2014-04-07 Jan Hubicka <hubcika@ucw.cz>
2380 * ipa-devirt.c (maybe_record_node): When node is not recorded,
2381 set completep to false rather than true.
2383 2014-04-07 Douglas B Rupp <rupp@adacore.com>
2386 * config/arm/arm.h (ASM_PREFERRED_EH_DATA_FORMAT): Expose from
2387 ARM_TARGET2_DWARF_FORMAT.
2389 2014-04-07 Charles Baylis <charles.baylis@linaro.org>
2392 * config/arm/arm.h (ASM_OUTPUT_CASE_END): Remove.
2393 (LABEL_ALIGN_AFTER_BARRIER): Align barriers which occur after
2396 2014-04-07 Richard Biener <rguenther@suse.de>
2398 PR tree-optimization/60766
2399 * tree-ssa-loop-ivopts.c (cand_value_at): Compute in an unsigned type.
2400 (may_eliminate_iv): Convert cand_value_at result to desired type.
2402 2014-04-07 Jason Merrill <jason@redhat.com>
2405 * common.opt (-fno-gnu-unique): Add.
2406 * config/elfos.h (USE_GNU_UNIQUE_OBJECT): Check it.
2408 2014-04-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2410 * haifa-sched.c: Fix outdated function reference and minor
2411 grammar errors in introductory comment.
2413 2014-04-07 Richard Biener <rguenther@suse.de>
2416 * tree-ssa-operands.c (maybe_add_call_vops): Also add VDEFs
2418 * tree-cfgcleanup.c (fixup_noreturn_call): Do not remove VDEFs.
2420 2014-04-06 John David Anglin <danglin@gcc.gnu.org>
2423 * config/pa/pa.c (pa_output_function_epilogue): Skip address and code
2424 size accounting for thunks.
2425 (pa_asm_output_mi_thunk): Use final_start_function() and
2426 final_end_function() to output function start and end directives.
2428 2014-04-05 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
2430 * config/avr/avr-arch.h (avr_mcu_t): Add dev_attribute field to have
2431 device specific ISA/ feature information. Remove short_sp and
2432 errata_skip ds. Add avr_device_specific_features enum to have device
2434 * config/avr/avr-c.c (avr_cpu_cpp_builtins): use dev_attribute to check
2435 errata_skip. Add __AVR_ISA_RMW__ builtin macro if RMW ISA available.
2436 * config/avr/avr-devices.c (avr_mcu_types): Update AVR_MCU macro for
2437 updated device specific info.
2438 * config/avr/avr-mcus.def: Merge device specific details to
2439 dev_attribute field.
2440 * config/avr/avr.c (avr_2word_insn_p): use dev_attribute field to check
2442 * config/avr/avr.h (AVR_HAVE_8BIT_SP): same for short sp info.
2443 * config/avr/driver-avr.c (avr_device_to_as): Pass -mrmw option to
2444 assembler if RMW isa supported by current device.
2445 * config/avr/genmultilib.awk: Update as device info structure changed.
2446 * doc/invoke.texi: Add info for __AVR_ISA_RMW__ builtin macro
2448 2014-04-04 Cong Hou <congh@google.com>
2450 PR tree-optimization/60656
2451 * tree-vect-stmts.c (supportable_widening_operation):
2452 Fix a bug that elements in a vector with vect_used_by_reduction
2453 property are incorrectly reordered when the operation on it is not
2454 consistant with the one in reduction operation.
2456 2014-04-04 John David Anglin <danglin@gcc.gnu.org>
2458 PR rtl-optimization/60155
2459 * gcse.c (record_set_data): New function.
2460 (single_set_gcse): New function.
2461 (gcse_emit_move_after): Use single_set_gcse instead of single_set.
2462 (hoist_code): Likewise.
2463 (get_pressure_class_and_nregs): Likewise.
2465 2014-04-04 Eric Botcazou <ebotcazou@adacore.com>
2467 * explow.c (probe_stack_range): Emit a final optimization blockage.
2469 2014-04-04 Anthony Green <green@moxielogic.com>
2471 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2): Fix
2474 2014-04-04 Jan Hubicka <hubicka@ucw.cz>
2477 * lto-cgraph.c (input_overwrite_node): Check that partitioning
2478 flags are set only during streaming.
2479 * ipa.c (process_references, walk_polymorphic_call_targets,
2480 symtab_remove_unreachable_nodes): Drop bodies of always inline
2481 after early inlining.
2482 (symtab_remove_unreachable_nodes): Remove always_inline attribute.
2484 2014-04-04 Jakub Jelinek <jakub@redhat.com>
2485 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2488 * dwarf2out.c (const_ok_for_output_1): Reject expressions
2491 2014-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2494 * config/arm/cortex-a53.md (cortex_a53_fdivs): Reduce reservation
2496 (cortex_a53_fdivd): Likewise.
2498 2014-04-04 Martin Jambor <mjambor@suse.cz>
2501 * cgraph.h (cgraph_clone_node): New parameter added to declaration.
2503 * cgraph.c (clone_of_p): Also return true if thunks match.
2504 (verify_edge_corresponds_to_fndecl): Removed extraneous call to
2505 cgraph_function_or_thunk_node and an obsolete comment.
2506 * cgraphclones.c (build_function_type_skip_args): Moved upwards in the
2508 (build_function_decl_skip_args): Likewise.
2509 (set_new_clone_decl_and_node_flags): New function.
2510 (duplicate_thunk_for_node): Likewise.
2511 (redirect_edge_duplicating_thunks): Likewise.
2512 (cgraph_clone_node): New parameter args_to_skip, pass it to
2513 redirect_edge_duplicating_thunks which is called instead of
2514 cgraph_redirect_edge_callee.
2515 (cgraph_create_virtual_clone): Pass args_to_skip to cgraph_clone_node,
2516 moved setting of a lot of flags to set_new_clone_decl_and_node_flags.
2518 2014-04-04 Jeff Law <law@redhat.com>
2521 * config/arm/predicates.md (const_int_I_operand): New predicate.
2522 (const_int_M_operand): Similarly.
2523 * config/arm/arm.md (insv_zero): Use const_int_M_operand instead of
2525 (insv_t2, extv_reg, extzv_t2): Likewise.
2526 (load_multiple_with_writeback): Similarly for const_int_I_operand.
2527 (pop_multiple_with_writeback_and_return): Likewise.
2528 (vfp_pop_multiple_with_writeback): Likewise
2530 2014-04-04 Richard Biener <rguenther@suse.de>
2533 * tree-ssanames.c (make_ssa_name_fn): Fix assert.
2534 * gimple.c (gimple_set_bb): Avoid ICEing for NULL cfun for
2536 * gimplify.h (gimple_add_tmp_var_fn): Declare.
2537 * gimplify.c (gimple_add_tmp_var_fn): New function.
2538 * gimple-expr.h (create_tmp_reg_fn): Declare.
2539 * gimple-expr.c (create_tmp_reg_fn): New function.
2540 * gimple-low.c (record_vars_into): Don't change cfun.
2541 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix
2542 code generation without cfun.
2544 2014-04-04 Thomas Schwinge <thomas@codesourcery.com>
2547 * Makefile.in (install-driver): Fix shell scripting.
2549 2014-04-03 Cong Hou <congh@google.com>
2551 PR tree-optimization/60505
2552 * tree-vectorizer.h (struct _stmt_vec_info): Add th field as the
2553 threshold of number of iterations below which no vectorization
2555 * tree-vect-loop.c (new_loop_vec_info):
2556 Initialize LOOP_VINFO_COST_MODEL_THRESHOLD.
2557 * tree-vect-loop.c (vect_analyze_loop_operations):
2558 Set LOOP_VINFO_COST_MODEL_THRESHOLD.
2559 * tree-vect-loop.c (vect_transform_loop):
2560 Use LOOP_VINFO_COST_MODEL_THRESHOLD.
2561 * tree-vect-loop.c (vect_analyze_loop_2): Check the maximum number
2562 of iterations of the loop and see if we should build the epilogue.
2564 2014-04-03 Richard Biener <rguenther@suse.de>
2566 * tree-streamer.h (struct streamer_tree_cache_d): Add next_idx member.
2567 (streamer_tree_cache_create): Adjust.
2568 * tree-streamer.c (streamer_tree_cache_add_to_node_array): Adjust
2569 to allow optional nodes array.
2570 (streamer_tree_cache_insert_1): Use next_idx to assign idx.
2571 (streamer_tree_cache_append): Likewise.
2572 (streamer_tree_cache_create): Create nodes array optionally
2573 as specified by parameter.
2574 * lto-streamer-out.c (create_output_block): Avoid maintaining
2575 the node array in the writer cache.
2576 (DFS_write_tree): Remove assertion.
2577 (produce_asm_for_decls): Free the out decl state hash table early.
2578 * lto-streamer-in.c (lto_data_in_create): Adjust for
2579 streamer_tree_cache_create prototype change.
2581 2014-04-03 Richard Biener <rguenther@suse.de>
2583 * tree-streamer-out.c (streamer_write_chain): Do not temporarily
2584 set TREE_CHAIN to NULL_TREE.
2586 2014-04-03 Richard Biener <rguenther@suse.de>
2588 PR tree-optimization/60740
2589 * graphite-scop-detection.c (stmt_simple_for_scop_p): Iterate
2590 over all GIMPLE_COND operands.
2592 2014-04-03 Nathan Sidwell <nathan@codesourcery.com>
2594 * doc/invoke.texi (Wnon-virtual-dtor): Adjust documentation.
2595 (Weffc++): Remove Scott's numbering, merge lists and reference
2598 2014-04-03 Nick Clifton <nickc@redhat.com>
2600 * config/rl78/rl78-expand.md (movqi): Handle (SUBREG (SYMBOL_REF))
2603 2014-04-03 Martin Jambor <mjambor@suse.cz>
2605 * ipa-cp.c (ipcp_verify_propagated_values): Also dump symtab and
2606 mention gcc_unreachable before failing.
2607 * ipa.c (symtab_remove_unreachable_nodes): Also print order of
2610 2014-04-02 Jan Hubicka <hubicka@ucw.cz>
2613 * ipa-devirt.c (get_polymorphic_call_info): Do not ICE on type
2614 inconsistent code and instead mark the context inconsistent.
2615 (possible_polymorphic_call_targets): For inconsistent contexts
2616 return empty complete list.
2618 2014-04-02 Anthony Green <green@moxielogic.com>
2620 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2)
2621 (extendqisi2, extendhisi2): Define.
2622 * config/moxie/moxie.h (DEFAULT_SIGNED_CHAR): Change to 0.
2623 (WCHAR_TYPE): Change to unsigned int.
2625 2014-04-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2627 PR tree-optimization/60733
2628 * gimple-ssa-strength-reduction.c (ncd_with_phi): Change required
2629 insertion point for PHI candidates to be the end of the feeding
2630 block for the PHI argument.
2632 2014-04-02 Vladimir Makarov <vmakarov@redhat.com>
2634 PR rtl-optimization/60650
2635 * lra-constraints.c (process_alt_operands): Decrease reject for
2636 earlyclobber matching.
2638 2014-04-02 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2640 * config/s390/s390.c (s390_expand_insv): Use GET_MODE_BITSIZE.
2642 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
2644 * config/spu/spu.c (pad_bb): Do not crash when the last
2645 insn is CODE_FOR_blockage.
2647 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
2649 * config/spu/spu.md ("insv"): Fail if bitoffset+bitsize
2650 lies outside the target mode.
2652 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
2655 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
2656 software floating point or no floating point registers, do not
2657 allow any type in the FPRs. Eliminate a test for SPE SIMD types
2658 in GPRs that occurs after we tested for GPRs that would never be
2661 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
2662 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
2663 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
2664 specifically allow DDmode, since that does not use the SPE SIMD
2667 2014-04-02 Richard Biener <rguenther@suse.de>
2670 * optabs.c (expand_abs_nojump): Honor flag_trapv only for
2671 MODE_INTs. Properly use negv_optab.
2672 (expand_abs): Likewise.
2674 2014-04-02 Richard Biener <rguenther@suse.de>
2677 * Makefile.in (install-driver): Guard extra installs with special
2680 2014-04-01 Michael Meissner <meissner@linux.vnet.ibm.com>
2682 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
2685 2014-04-01 Richard Henderson <rth@redhat.com>
2688 * config/i386/i386.md (*float<SWI48><MODEF>2_sse): Leave the second
2689 alternative enabled before register allocation.
2691 2014-04-01 Chung-Lin Tang <cltang@codesourcery.com>
2693 * config/nios2/nios2.md (unspec): Remove UNSPEC_TLS, UNSPEC_TLS_LDM.
2694 * config/nios2/nios2.c (nios2_function_profiler): Fix addi operand
2696 (nios2_large_got_address): Remove unneeded 'sym' parameter.
2697 (nios2_got_address): Update nios2_large_got_address call site.
2698 (nios2_delegitimize_address): New function.
2699 (TARGET_DELEGITIMIZE_ADDRESS): Define to nios2_delegitimize_address.
2700 * config/nios2/linux.h (GLIBC_DYNAMIC_LINKER): Define.
2701 (LINK_SPEC): Specify dynamic linker using GNU_USER_DYNAMIC_LINKER.
2703 2014-04-01 Martin Husemann <martin@duskware.de>
2705 * config/mips/netbsd.h (TARGET_OS_CPP_BUILTINS): Define __mips_o32
2708 2014-04-01 Richard Sandiford <rdsandiford@googlemail.com>
2710 PR rtl-optimization/60604
2711 * recog.c (general_operand): Incorporate REG_CANNOT_CHANGE_MODE_P
2712 check from register_operand.
2713 (register_operand): Redefine in terms of general_operand.
2714 (nonmemory_operand): Use register_operand for the non-constant cases.
2716 2014-04-01 Richard Biener <rguenther@suse.de>
2718 * gimple.h (struct gimple_statement_base): Align subcode to 16 bits.
2720 2014-04-01 Sebastian Huber <sebastian.huber@embedded-brains.de>
2722 * doc/invoke.texi (mapp-regs): Clarify.
2724 2014-03-31 Ulrich Drepper <drepper@gmail.com>
2726 * config/i386/avx512fintrin.h (__v32hi): Define type.
2727 (__v64qi): Likewise.
2728 (_mm512_set1_epi8): Define.
2729 (_mm512_set1_epi16): Define.
2730 (_mm512_set4_epi32): Define.
2731 (_mm512_set4_epi64): Define.
2732 (_mm512_set4_pd): Define.
2733 (_mm512_set4_ps): Define.
2734 (_mm512_setr4_epi64): Define.
2735 (_mm512_setr4_epi32): Define.
2736 (_mm512_setr4_pd): Define.
2737 (_mm512_setr4_ps): Define.
2738 (_mm512_setzero_epi32): Define.
2740 2014-03-31 Martin Jambor <mjambor@suse.cz>
2743 * tree-sra.c (callsite_has_enough_arguments_p): Renamed to
2744 callsite_arguments_match_p. Updated all callers. Also check types of
2745 corresponding formal parameters and actual arguments.
2746 (not_all_callers_have_enough_arguments_p) Renamed to
2747 some_callers_have_mismatched_arguments_p.
2749 2014-03-31 Yuri Rumyantsev <ysrumyan@gmail.com>
2751 * tree-inline.c (copy_loops): Add missed copy of 'safelen'.
2753 2014-03-31 Kugan Vivekanandarajah <kuganv@linaro.org>
2756 * aarch64/aarch64.c (aarch64_classify_address): Fix alignment for
2759 2014-03-30 Uros Bizjak <ubizjak@gmail.com>
2761 * config/i386/sse.md (FMAMODE_NOVF512): New mode iterator.
2762 (<sd_mask_codefor>fma_fmadd_<mode><sd_maskz_name><round_name>):
2764 <sd_mask_codefor>fma_fmadd_<VF_512:mode><sd_maskz_name><round_name>.
2765 Use FMAMODE_NOVF512 mode iterator.
2766 (<sd_mask_codefor>fma_fmsub_<mode><sd_maskz_name><round_name>): Ditto.
2767 (<sd_mask_codefor>fma_fnmadd_<mode><sd_maskz_name><round_name>): Ditto.
2768 (<sd_mask_codefor>fma_fnmsub_<mode><sd_maskz_name><round_name>): Ditto.
2769 (<sd_mask_codefor>fma_fmaddsub_<mode><sd_maskz_name><round_name>):
2771 <sd_mask_codefor>fma_fmaddsub_<VF_512:mode><sd_maskz_name><round_name>.
2772 Use VF_128_256 mode iterator.
2773 (<sd_mask_codefor>fma_fmsubadd_<mode><sd_maskz_name><round_name>):
2776 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
2778 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Clear
2779 static chain if needed.
2781 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
2784 * lra-constraints.c (index_part_to_reg): New.
2785 (process_address): Use it.
2787 2014-03-27 Jeff Law <law@redhat.com>
2788 Jakub Jelinek <jakub@redhat.com>
2791 * expr.c (do_tablejump): Use simplify_gen_binary rather than
2792 gen_rtx_{PLUS,MULT} to build up the address expression.
2794 * i386/i386.c (ix86_legitimize_address): Use copy_addr_to_reg to avoid
2795 creating non-canonical RTL.
2797 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
2800 * ipa-inline.c (want_inline_small_function_p): Short circuit large
2801 functions; reorganize to make cheap checks first.
2802 (inline_small_functions): Do not estimate growth when dumping;
2804 * ipa-inline.h (inline_summary): Add min_size.
2805 (growth_likely_positive): New function.
2806 * ipa-inline-analysis.c (dump_inline_summary): Add min_size.
2807 (set_cond_stmt_execution_predicate): Cleanup.
2808 (estimate_edge_size_and_time): Compute min_size.
2809 (estimate_calls_size_and_time): Likewise.
2810 (estimate_node_size_and_time): Likewise.
2811 (inline_update_overall_summary): Update min_size.
2812 (do_estimate_edge_time): Likewise.
2813 (do_estimate_edge_size): Update.
2814 (do_estimate_edge_hints): Update.
2815 (growth_likely_positive): New function.
2817 2014-03-28 Jakub Jelinek <jakub@redhat.com>
2820 * config/i386/i386.c (ix86_copy_addr_to_reg): Call copy_addr_to_reg
2821 also if addr has VOIDmode.
2823 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2825 * config/arm/aarch-common.c (aarch_crypto_can_dual_issue): New.
2826 * config/arm/aarch-common-protos.h (aarch_crypto_can_dual_issue):
2828 * config/arm/cortex-a53.md: Add reservations and bypass for crypto
2829 instructions as well as AdvancedSIMD loads.
2831 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2833 * config/aarch64/aarch64-simd.md (aarch64_crypto_aes<aes_op>v16qi):
2834 Use crypto_aese type.
2835 (aarch64_crypto_aes<aesmc_op>v16qi): Use crypto_aesmc type.
2836 * config/arm/arm.md (is_neon_type): Replace crypto_aes with
2837 crypto_aese, crypto_aesmc. Move to types.md.
2838 * config/arm/types.md (crypto_aes): Split into crypto_aese,
2840 * config/arm/iterators.md (crypto_type): Likewise.
2842 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
2844 * cgraph.c: Include expr.h and tree-dfa.h.
2845 (cgraph_redirect_edge_call_stmt_to_callee): If call in noreturn;
2848 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
2851 * lra-assigns.c (find_hard_regno_for): Remove unavailable hard
2852 regs from checking multi-reg pseudos.
2854 2014-03-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2856 * config/arm/t-aprofile (MULTILIB_MATCHES): Correct A12 rule.
2858 2014-03-28 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
2860 * config/rs6000/rs6000.c (fusion_gpr_load_p): Refuse optimization
2861 if it would clobber the stack pointer, even temporarily.
2863 2014-03-28 Eric Botcazou <ebotcazou@adacore.com>
2865 * mode-switching.c: Make small adjustments to the top comment.
2867 2014-03-27 Michael Meissner <meissner@linux.vnet.ibm.com>
2869 * config/rs6000/constraints.md (wD constraint): New constraint to
2870 match the constant integer to get the top DImode/DFmode out of a
2871 vector in a VSX register.
2873 * config/rs6000/predicates.md (vsx_scalar_64bit): New predicate to
2874 match the constant integer to get the top DImode/DFmode out of a
2875 vector in a VSX register.
2877 * config/rs6000/rs6000-builtins.def (VBPERMQ): Add vbpermq builtin
2880 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
2883 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
2884 -mdebug=reg, print value of VECTOR_ELEMENT_SCALAR_64BIT.
2886 * config/rs6000/vsx.md (vsx_extract_<mode>, V2DI/V2DF modes):
2887 Optimize vec_extract of 64-bit values, where the value being
2888 extracted is in the top word, where we can use scalar
2889 instructions. Add direct move and store support. Combine the big
2890 endian/little endian vector select load support into a single insn.
2891 (vsx_extract_<mode>_internal1): Likewise.
2892 (vsx_extract_<mode>_internal2): Likewise.
2893 (vsx_extract_<mode>_load): Likewise.
2894 (vsx_extract_<mode>_store): Likewise.
2895 (vsx_extract_<mode>_zero): Delete, big and little endian insns are
2896 combined into vsx_extract_<mode>_load.
2897 (vsx_extract_<mode>_one_le): Likewise.
2899 * config/rs6000/rs6000.h (VECTOR_ELEMENT_SCALAR_64BIT): Macro to
2900 define the top 64-bit vector element.
2902 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wD
2905 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
2906 Document vec_vbpermq builtin.
2909 * config/rs6000/altivec.h (vec_xxsldwi): Add missing define to
2910 enable use of xxsldwi and xxpermdi builtin functions.
2911 (vec_xxpermdi): Likewise.
2913 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
2914 Document use of vec_xxsldwi and vec_xxpermdi builtins.
2916 2014-03-27 Vladimir Makarov <vmakarov@redhat.com>
2918 PR rtl-optimization/60650
2919 * lra-assign.c (find_hard_regno_for, spill_for): Add parameter
2921 (find_spills_for): New.
2922 (assign_by_spills): Pass the new parameter to find_hard_regno_for.
2923 Spill all pseudos on the second iteration.
2925 2014-03-27 Marek Polacek <polacek@redhat.com>
2928 * doc/extend.texi (ffs Builtins): Change unsigned types to signed
2931 2014-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2933 * config/s390/s390.c (s390_can_use_return_insn): Check for
2934 call-saved FPRs on 31 bit.
2936 2014-03-27 Jakub Jelinek <jakub@redhat.com>
2939 * omp-low.c (lower_omp_1): For gimple_clobber_p stmts,
2940 if they need regimplification, just drop them instead of
2941 calling gimple_regimplify_operands on them.
2943 2014-03-27 Marcus Shawcroft <marcus.shawcroft@arm.com>
2946 * config/aarch64/aarch64.c (faked_omit_frame_pointer): Remove.
2947 (aarch64_frame_pointer_required): Adjust logic.
2948 (aarch64_can_eliminate): Adjust logic.
2949 (aarch64_override_options_after_change): Adjust logic.
2951 2014-03-27 Dehao Chen <dehao@google.com>
2953 * ipa-inline.c (early_inliner): Update node's inline info.
2955 2014-03-26 Dehao Chen <dehao@google.com>
2957 * dojump.c (do_compare_rtx_and_jump): Sets correct probability for
2958 compiler inserted conditional jumps for NAN float check.
2960 2014-03-26 Jakub Jelinek <jakub@redhat.com>
2962 * ubsan.h (ubsan_create_data): Change second argument's type
2963 to const location_t *.
2964 * ubsan.c (ubsan_source_location): If xloc.file is NULL, set it to
2966 (ubsan_create_data): Change second argument to const location_t *PLOC.
2967 Create Loc field whenever PLOC is non-NULL.
2968 (ubsan_instrument_unreachable, ubsan_expand_null_ifn,
2969 ubsan_build_overflow_builtin, instrument_bool_enum_load): Adjust
2973 * real.c (real_to_integer2): Change type of low to UHWI.
2975 2014-03-26 Tobias Burnus <burnus@net-b.de>
2977 * gcc.c (LINK_COMMAND_SPEC): Use libcilkrts.spec for -fcilkplus.
2978 (CILK_SELF_SPECS): New define.
2979 (driver_self_specs): Use it.
2981 2014-03-26 Richard Biener <rguenther@suse.de>
2983 * tree-pretty-print.c (percent_K_format): Implement special
2984 case for LTO and its stripped down BLOCK tree.
2986 2014-03-26 Jakub Jelinek <jakub@redhat.com>
2989 * ubsan.c (instrument_si_overflow): Instrument ABS_EXPR.
2991 * tree-vrp.c (simplify_internal_call_using_ranges): If only
2992 one range is range_int_cst_p, but not both, at least optimize
2993 addition/subtraction of 0 and multiplication by 0 or 1.
2994 * gimple-fold.c (gimple_fold_call): Fold
2995 IFN_UBSAN_CHECK_{ADD,SUB,MUL}.
2996 (gimple_fold_stmt_to_constant_1): If both op0 and op1 aren't
2997 INTEGER_CSTs, try to fold at least x * 0 and y - y.
2999 2014-03-26 Eric Botcazou <ebotcazou@adacore.com>
3001 PR rtl-optimization/60452
3002 * rtlanal.c (rtx_addr_can_trap_p_1): Fix head comment.
3003 <case REG>: Return 1 for invalid offsets from the frame pointer.
3005 2014-03-26 Marek Polacek <polacek@redhat.com>
3008 * doc/extend.texi (C Extensions): Mention variable-length arrays in
3011 2014-03-26 Marek Polacek <polacek@redhat.com>
3014 * doc/extend.texi (Designated Inits): Describe what happens to omitted
3017 2014-03-26 Marek Polacek <polacek@redhat.com>
3020 * ira-color.c (update_conflict_hard_regno_costs): Perform the
3021 multiplication in unsigned type.
3023 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
3025 * doc/install.texi: Document nds32le-*-elf and nds32be-*-elf.
3027 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
3029 * doc/contrib.texi: Add myself as Andes nds32 port contributor.
3031 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
3034 * cif-code.def (UNREACHABLE) New code.
3035 * ipa-inline.c (inline_small_functions): Skip edges to
3036 __builtlin_unreachable.
3037 (estimate_edge_growth): Allow edges to __builtlin_unreachable.
3038 * ipa-inline-analysis.c (edge_set_predicate): Redirect edges with false
3039 predicate to __bulitin_unreachable.
3040 (set_cond_stmt_execution_predicate): Fix issue when
3041 invert_tree_comparison returns ERROR_MARK.
3042 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow): Do not
3043 propagate to inline clones.
3044 * cgraph.c (verify_edge_corresponds_to_fndecl): Allow redirection
3046 * ipa-cp.c (create_specialized_node): Be ready for new node to appear.
3047 * cgraphclones.c (cgraph_clone_node): If call destination is already
3048 ureachable, do not redirect it back.
3049 * tree-inline.c (fold_marked_statements): Hanlde calls becoming
3052 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
3054 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow):
3055 Do not modify inline clones.
3057 2014-03-25 Jakub Jelinek <jakub@redhat.com>
3059 * config/i386/i386.md (general_sext_operand): New mode attr.
3060 (addv<mode>4, subv<mode>4, mulv<mode>4): If operands[2] is CONST_INT,
3061 don't generate (sign_extend (const_int)).
3062 (*addv<mode>4, *subv<mode>4, *mulv<mode>4): Disallow CONST_INT_P
3063 operands[2]. Use We constraint instead of <i> and
3064 <general_sext_operand> predicate instead of <general_operand>.
3065 (*addv<mode>4_1, *subv<mode>4_1, *mulv<mode>4_1): New insns.
3066 * config/i386/constraints.md (We): New constraint.
3067 * config/i386/predicates.md (x86_64_sext_operand,
3068 sext_operand): New predicates.
3070 2014-03-25 Martin Jambor <mjambor@suse.cz>
3073 * ipa-cp.c (ipa_get_indirect_edge_target_1): Redirect type
3074 inconsistent devirtualizations to __builtin_unreachable.
3076 2014-03-25 Marek Polacek <polacek@redhat.com>
3079 * doc/extend.texi (Example of asm with clobbered asm reg): Fix typo.
3081 2014-03-25 Alan Lawrence <alan.lawrence@arm.com>
3083 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Reverse
3084 order of elements for big-endian.
3086 2014-03-25 Richard Biener <rguenther@suse.de>
3089 * gimplify-me.c (gimple_regimplify_operands): Update the
3092 2014-03-25 Martin Jambor <mjambor@suse.cz>
3095 * lto-cgraph.c (lto_output_node): Stream body_removed flag.
3096 (lto_output_varpool_node): Likewise.
3097 (input_overwrite_node): Likewise.
3098 (input_varpool_node): Likewise.
3100 2014-03-25 Richard Biener <rguenther@suse.de>
3102 * lto-wrapper.c (merge_and_complain): Handle OPT_fPIE like OPT_fpie.
3103 (run_gcc): Likewise.
3105 2014-03-25 Jakub Jelinek <jakub@redhat.com>
3107 * combine.c (simplify_compare_const): Add MODE argument.
3108 Handle mode_width 0 as very large mode_width.
3109 (try_combine, simplify_comparison): Adjust callers.
3111 * cselib.c (cselib_hash_rtx): Perform addition in unsigned
3112 type to avoid signed integer overflow.
3113 * explow.c (plus_constant): Likewise.
3115 2014-03-25 Dominik Vogt <vogt@linux.vnet.ibm.com>
3117 * doc/generic.texi: Correct typos.
3119 2014-03-24 Tobias Burnus <burnus@net-b.de>
3121 * doc/invoke.texi (-flto): Expand section about
3122 using static libraries with LTO.
3124 2014-03-24 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3126 PR rtl-optimization/60501
3127 * optabs.def (addptr3_optab): New optab.
3128 * optabs.c (gen_addptr3_insn, have_addptr3_insn): New function.
3129 * doc/md.texi ("addptrm3"): Document new RTL standard expander.
3130 * expr.h (gen_addptr3_insn, have_addptr3_insn): Add prototypes.
3132 * lra.c (emit_add3_insn): Use the addptr pattern if available.
3134 * config/s390/s390.md ("addptrdi3", "addptrsi3"): New expanders.
3136 2014-03-24 Ulrich Drepper <drepper@gmail.com>
3138 * config/i386/avx512fintrin.h: Define _mm512_set1_ps and
3141 * config/i386/avxintrin.h (_mm256_undefined_si256): Define.
3142 (_mm256_undefined_ps): Define.
3143 (_mm256_undefined_pd): Define.
3144 * config/i386/emmintrin.h (_mm_undefined_si128): Define.
3145 (_mm_undefined_pd): Define.
3146 * config/i386/xmmintrin.h (_mm_undefined_ps): Define.
3147 * config/i386/avx512fintrin.h (_mm512_undefined_si512): Define.
3148 (_mm512_undefined_ps): Define.
3149 (_mm512_undefined_pd): Define.
3150 Use _mm*_undefined_*.
3151 * config/i386/avx2intrin.h: Use _mm*_undefined_*.
3153 2014-03-24 Alex Velenko <Alex.Velenko@arm.com>
3155 * config/aarch64/aarch64-simd-builtins.def (lshr): DI mode excluded.
3156 (lshr_simd): DI mode added.
3157 * config/aarch64/aarch64-simd.md (aarch64_lshr_simddi): New pattern.
3158 (aarch64_ushr_simddi): Likewise.
3159 * config/aarch64/aarch64.md (UNSPEC_USHR64): New unspec.
3160 * config/aarch64/arm_neon.h (vshr_n_u64): Intrinsic fixed.
3161 (vshrd_n_u64): Likewise.
3163 2014-03-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3165 * Makefile.in (s-macro_list): Depend on cc1.
3167 2014-03-23 Teresa Johnson <tejohnson@google.com>
3169 * ipa-utils.c (ipa_print_order): Use specified dump file.
3171 2014-03-23 Eric Botcazou <ebotcazou@adacore.com>
3173 PR rtl-optimization/60601
3174 * bb-reorder.c (fix_up_fall_thru_edges): Test EDGE_FALLTHRU everywhere.
3176 * gcc.c (eval_spec_function): Initialize save_growing_value.
3178 2014-03-22 Jakub Jelinek <jakub@redhat.com>
3181 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): For
3182 code == MINUS_EXPR, never swap op0 with op1.
3184 * toplev.c (init_local_tick): Avoid signed integer multiplication
3186 * genautomata.c (reserv_sets_hash_value): Fix rotate idiom, avoid
3187 shift by first operand's bitsize.
3189 2014-03-21 Jakub Jelinek <jakub@redhat.com>
3192 * config/i386/i386.h (TARGET_64BIT_P): If not TARGET_BI_ARCH,
3194 * config/i386/darwin.h (TARGET_64BIT_P): Redefine to
3195 TARGET_ISA_64BIT_P(x).
3197 2014-03-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3199 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Generate a
3200 pattern for vector nor instead of subtract from splat(-1).
3201 (altivec_expand_vec_perm_const_le): Likewise.
3203 2014-03-21 Richard Henderson <rth@twiddle.net>
3206 * ifcvt.c (dead_or_predicable): Return FALSE if there are any frame
3207 related insns after epilogue_completed.
3209 2014-03-21 Martin Jambor <mjambor@suse.cz>
3212 * cgraph.h (symtab_node): New flag body_removed.
3213 * ipa.c (symtab_remove_unreachable_nodes): Set body_removed flag
3214 when removing bodies.
3215 * symtab.c (dump_symtab_base): Dump body_removed flag.
3216 * cgraph.c (verify_edge_corresponds_to_fndecl): Skip nodes which
3217 had their bodies removed.
3219 2014-03-21 Martin Jambor <mjambor@suse.cz>
3222 * ipa.c (symtab_remove_unreachable_nodes): Clear thunk flag of nodes
3225 2014-03-21 Richard Biener <rguenther@suse.de>
3227 PR tree-optimization/60577
3228 * tree-core.h (struct tree_base): Document nothrow_flag use
3229 in VAR_DECL_NONALIASED.
3230 * tree.h (VAR_DECL_NONALIASED): New.
3231 (may_be_aliased): Adjust.
3232 * coverage.c (build_var): Set VAR_DECL_NONALIASED.
3234 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
3236 * expr.c (expand_expr_real_1): Remove outdated comment.
3238 2014-03-20 Jakub Jelinek <jakub@redhat.com>
3241 * ira.c (adjust_cleared_regs): Call copy_rtx on
3242 *reg_equiv[REGNO (loc)].src_p before passing it to
3243 simplify_replace_fn_rtx.
3246 * config/i386/i386.c (x86_output_mi_thunk): Surround UNSPEC_GOT
3247 into CONST, put pic register as first operand of PLUS. Use
3248 gen_const_mem for both 32-bit and 64-bit PIC got loads.
3250 2014-03-20 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3252 * config/aarch64/aarch64.c (MEMORY_MOVE_COST): Delete.
3254 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
3256 * config/sparc/sparc.c (sparc_do_work_around_errata): Implement work
3257 around for store forwarding issue in the FPU on the UT699.
3258 * config/sparc/sparc.md (in_branch_delay): Return false for single FP
3259 loads and operations if -mfix-ut699 is specified.
3260 (divtf3_hq): Tweak attribute.
3261 (sqrttf2_hq): Likewise.
3263 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
3265 * calls.c (store_one_arg): Remove incorrect const qualification on the
3266 type of the temporary.
3267 * cfgexpand.c (expand_return): Likewise.
3268 * expr.c (expand_constructor): Likewise.
3269 (expand_expr_real_1): Likewise.
3271 2014-03-20 Zhenqiang Chen <zhenqiang.chen@linaro.org>
3273 * config/arm/arm.c (arm_dwarf_register_span): Update the element number
3276 2014-03-19 Kaz Kojima <kkojima@gcc.gnu.org>
3279 * config/sh/sh.md (udivsi3_i1): Clobber R1 register.
3281 2014-03-19 James Greenhalgh <james.greenhalgh@arm.com>
3283 * config/arm/aarch-common-protos.h
3284 (alu_cost_table): Fix spelling of "extend".
3285 * config/arm/arm.c (arm_new_rtx_costs): Fix spelling of "extend".
3287 2014-03-19 Richard Biener <rguenther@suse.de>
3290 * tree-core.h (tree_type_common): Re-order pointer members
3291 to reduce recursion depth during GC walks.
3293 2014-03-19 Marek Polacek <polacek@redhat.com>
3296 * ubsan.c (ubsan_type_descriptor): Check that DECL_NAME is nonnull
3297 before accessing it.
3299 2014-03-19 Richard Biener <rguenther@suse.de>
3302 * lto-streamer-in.c (input_function): In WPA stage do not drop
3305 2014-03-19 Jakub Jelinek <jakub@redhat.com>
3307 PR tree-optimization/60559
3308 * vectorizable_mask_load_store): Replace scalar MASK_LOAD
3309 with build_zero_cst assignment.
3311 2014-03-18 Kai Tietz <ktietz@redhat.com>
3313 PR rtl-optimization/56356
3314 * sdbout.c (sdbout_parms): Verify that parms'
3315 incoming argument is valid.
3316 (sdbout_reg_parms): Likewise.
3318 2014-03-18 Richard Henderson <rth@redhat.com>
3321 * config/i386/i386.md (*float<SWI48x><MODEF>2_i387): Move down to
3322 be shadowed by *float<SWI48><MODEF>2_sse. Test X87_ENABLE_FLOAT.
3323 (*float<SWI48><MODEF>2_sse): Check X87_ENABLE_FLOAT for alternative 0.
3325 2014-03-18 Basile Starynkevitch <basile@starynkevitch.net>
3327 * plugin.def: Improve comment for PLUGIN_INCLUDE_FILE.
3328 * doc/plugins.texi (Plugin callbacks): Mention PLUGIN_INCLUDE_FILE.
3329 Italicize plugin event names in description. Explain that
3330 PLUGIN_PRAGMAS has no sense for lto1. Explain PLUGIN_INCLUDE_FILE.
3331 Remind that no GCC functions should be called after PLUGIN_FINISH.
3332 Explain what pragmas with expansion are.
3334 2014-03-18 Martin Liska <mliska@suse.cz>
3336 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Added case when
3337 gimple call statement is update.
3338 * gimple-fold.c (gimple_fold_call): Changed order for GIMPLE_ASSIGN and
3339 GIMPLE_CALL, where gsi iterator still points to GIMPLE CALL.
3341 2014-03-18 Jakub Jelinek <jakub@redhat.com>
3344 * ubsan.c (ubsan_instrument_unreachable): Call
3345 initialize_sanitizer_builtins.
3346 (ubsan_pass): Likewise.
3349 * ubsan.c (ubsan_type_descriptor, ubsan_create_data): Call
3350 varpool_finalize_decl instead of rest_of_decl_compilation.
3352 2014-03-18 Richard Biener <rguenther@suse.de>
3354 * df-problems.c (df_rd_confluence_n): Avoid bitmap_copy
3355 by using bitmap_and_compl instead of bitmap_and_compl_into.
3356 (df_rd_transfer_function): Likewise.
3358 2014-03-18 Richard Biener <rguenther@suse.de>
3360 * doc/lto.texi (fresolution): Fix typo.
3362 2014-03-18 Richard Biener <rguenther@suse.de>
3364 * doc/invoke.texi (flto): Update for changes in 4.9.
3366 2014-03-18 Richard Biener <rguenther@suse.de>
3368 * doc/loop.texi: Remove section on the removed lambda framework.
3369 Update loop docs with recent changes in preserving loop structure.
3371 2014-03-18 Richard Biener <rguenther@suse.de>
3373 * doc/lto.texi (-fresolution): Document.
3375 2014-03-18 Richard Biener <rguenther@suse.de>
3377 * doc/contrib.texi: Adjust my name.
3379 2014-03-18 Jakub Jelinek <jakub@redhat.com>
3382 * internal-fn.c: Include diagnostic-core.h.
3383 (expand_BUILTIN_EXPECT): New function.
3384 * gimplify.c (gimplify_call_expr): Use false instead of FALSE.
3385 (gimplify_modify_expr): Gimplify 3 argument __builtin_expect into
3386 IFN_BUILTIN_EXPECT call instead of __builtin_expect builtin call.
3387 * ipa-inline-analysis.c (find_foldable_builtin_expect): Handle
3389 * predict.c (expr_expected_value_1): Handle IFN_BUILTIN_EXPECT.
3390 Revert 3 argument __builtin_expect code.
3391 (strip_predict_hints): Handle IFN_BUILTIN_EXPECT.
3392 * gimple-fold.c (gimple_fold_call): Likewise.
3393 * tree.h (fold_builtin_expect): New prototype.
3394 * builtins.c (build_builtin_expect_predicate): Add predictor
3395 argument, if non-NULL, create 3 argument __builtin_expect.
3396 (fold_builtin_expect): No longer static. Add ARG2 argument,
3397 pass it through to build_builtin_expect_predicate.
3398 (fold_builtin_2): Adjust caller.
3399 (fold_builtin_3): Handle BUILT_IN_EXPECT.
3400 * internal-fn.def (BUILTIN_EXPECT): New.
3402 2014-03-18 Tobias Burnus <burnus@net-b.de>
3405 * predict.def (PRED_FORTRAN_OVERFLOW, PRED_FORTRAN_FAIL_ALLOC,
3406 PRED_FORTRAN_FAIL_IO, PRED_FORTRAN_WARN_ONCE, PRED_FORTRAN_SIZE_ZERO,
3407 PRED_FORTRAN_INVALID_BOUND, PRED_FORTRAN_ABSENT_DUMMY): Add.
3409 2014-03-18 Jan Hubicka <hubicka@ucw.cz>
3412 * predict.c (combine_predictions_for_bb): Fix up formatting.
3413 (expr_expected_value_1, expr_expected_value): Add predictor argument,
3414 fill what it points to if non-NULL.
3415 (tree_predict_by_opcode): Adjust caller, use the predictor.
3416 * predict.def (PRED_COMPARE_AND_SWAP): Add.
3418 2014-03-18 Eric Botcazou <ebotcazou@adacore.com>
3420 * config/sparc/sparc.c (sparc_do_work_around_errata): Speed up and use
3421 proper constant for the store mode.
3423 2014-03-18 Ilya Enkovich <ilya.enkovich@intel.com>
3425 * symtab.c (change_decl_assembler_name): Fix transparent alias
3428 2014-03-16 Renlin Li <Renlin.Li@arm.com>
3430 * config/aarch64/aarch64.c: Correct the comments about the
3431 aarch64 stack layout.
3433 2014-03-18 Thomas Schwinge <thomas@codesourcery.com>
3435 * omp-low.c (lower_rec_input_clauses) <build_omp_barrier>: Restore
3436 check for GF_OMP_FOR_KIND_FOR.
3438 2013-03-18 Kirill Yukhin <kirill.yukhin@intel.com>
3440 * config/i386/i386.h (ADDITIONAL_REGISTER_NAMES): Add
3441 ymm and zmm register names.
3443 2014-03-17 Jakub Jelinek <jakub@redhat.com>
3446 * config/i386/i386.c (ix86_expand_epilogue): Adjust REG_CFA_ADJUST_CFA
3447 note creation for the 2010-08-31 changes.
3449 2014-03-17 Marek Polacek <polacek@redhat.com>
3452 * omp-low.c (omp_max_vf): Treat -fno-tree-loop-optimize the same
3453 as -fno-tree-loop-vectorize.
3454 (expand_omp_simd): Likewise.
3456 2014-03-15 Eric Botcazou <ebotcazou@adacore.com>
3458 * config/sparc/sparc-protos.h (tls_call_delay): Delete.
3459 (eligible_for_call_delay): New prototype.
3460 * config/sparc/sparc.c (tls_call_delay): Rename into...
3461 (eligible_for_call_delay): ...this. Return false if the instruction
3462 cannot be put in the delay slot of a branch.
3463 (eligible_for_restore_insn): Simplify.
3464 (eligible_for_return_delay): Return false if the instruction cannot be
3465 put in the delay slot of a branch and simplify.
3466 (eligible_for_sibcall_delay): Return false if the instruction cannot be
3467 put in the delay slot of a branch.
3468 * config/sparc/sparc.md (fix_ut699): New attribute.
3469 (tls_call_delay): Delete.
3470 (in_call_delay): Reimplement.
3471 (eligible_for_sibcall_delay): Rename into...
3472 (in_sibcall_delay): ...this.
3473 (eligible_for_return_delay): Rename into...
3474 (in_return_delay): ...this.
3475 (in_branch_delay): Reimplement.
3476 (in_uncond_branch_delay): Delete.
3477 (in_annul_branch_delay): Delete.
3479 2014-03-14 Richard Henderson <rth@redhat.com>
3482 * config/i386/i386.md (floathi<X87MODEF>2): Delete expander; rename
3483 define_insn from *floathi<X87MODEF>2_i387; allow nonimmediate_operand.
3484 (*floathi<X87MODEF>2_i387_with_temp): Remove.
3485 (floathi splitters): Remove.
3486 (float<SWI48x>xf2): New pattern.
3487 (float<SWI48><MODEF>2): Rename from float<SWI48x><X87MODEF>2. Drop
3488 code that tried to handle DImode for 32-bit, but which was excluded
3489 by the pattern's condition. Drop allocation of stack temporary.
3490 (*floatsi<MODEF>2_vector_mixed_with_temp): Remove.
3491 (*float<SWI48><MODEF>2_mixed_with_temp): Remove.
3492 (*float<SWI48><MODEF>2_mixed_interunit): Remove.
3493 (*float<SWI48><MODEF>2_mixed_nointerunit): Remove.
3494 (*floatsi<MODEF>2_vector_sse_with_temp): Remove.
3495 (*float<SWI48><MODEF>2_sse_with_temp): Remove.
3496 (*float<SWI48><MODEF>2_sse_interunit): Remove.
3497 (*float<SWI48><MODEF>2_sse_nointerunit): Remove.
3498 (*float<SWI48x><X87MODEF>2_i387_with_temp): Remove.
3499 (*float<SWI48x><X87MODEF>2_i387): Remove.
3500 (all float _with_temp splitters): Remove.
3501 (*float<SWI48x><MODEF>2_i387): New pattern.
3502 (*float<SWI48><MODEF>2_sse): New pattern.
3503 (float TARGET_USE_VECTOR_CONVERTS splitters): Merge them.
3504 (float TARGET_SSE_PARTIAL_REG_DEPENDENCY splitters): Merge them.
3506 2014-03-14 Jakub Jelinek <jakub@redhat.com>
3507 Marek Polacek <polacek@redhat.com>
3510 * common.opt (dump_base_name_prefixed): New Variable.
3511 * opts.c (finish_options): Don't prepend directory to x_dump_base_name
3512 if x_dump_base_name_prefixed is already set, set it at the end.
3514 2014-03-14 Vladimir Makarov <vmakarov@redhat.com>
3516 PR rtl-optimization/60508
3517 * lra-constraints.c (get_reload_reg): Add new parameter
3519 (process_addr_reg, simplify_operand_subreg, curr_insn_transform):
3520 Pass the new parameter values.
3522 2014-03-14 Richard Biener <rguenther@suse.de>
3524 * common.opt: Revert unintented changes from r205065.
3527 2014-03-14 Richard Biener <rguenther@suse.de>
3530 * cfghooks.c (split_block): Properly adjust all loops the
3531 block was a latch of.
3533 2014-03-14 Martin Jambor <mjambor@suse.cz>
3536 * ipa-prop.c (ipa_modify_call_arguments): Fix iteration condition
3539 2014-03-14 Georg-Johann Lay <avr@gjlay.de>
3542 * config/avr/avr.c (avr_set_current_function): Pass function name
3543 through default_strip_name_encoding before sanity checking instead
3544 of skipping the first char of the assembler name.
3546 2014-03-13 Richard Henderson <rth@redhat.com>
3549 * config/i386/i386.c (ix86_split_fp_branch): Remove pushed argument.
3550 (ix86_force_to_memory, ix86_free_from_memory): Remove.
3551 * config/i386/i386-protos.h: Likewise.
3552 * config/i386/i386.md (floathi<X87MODEF>2): Use assign_386_stack_local
3553 in the expander instead of a splitter.
3554 (float<SWI48x><X87MODEF>2): Use assign_386_stack_local if there is
3555 any possibility of requiring a memory.
3556 (*floatsi<MODEF>2_vector_mixed): Remove, and the splitters.
3557 (*floatsi<MODEF>2_vector_sse): Remove, and the splitters.
3558 (fp branch splitters): Update for ix86_split_fp_branch.
3559 (*jcc<X87MODEF>_<SWI24>_i387): Remove r/f alternative.
3560 (*jcc<X87MODEF>_<SWI24>_r_i387): Likewise.
3561 (splitter for jcc<X87MODEF>_<SWI24>_i387 r/f): Remove.
3562 (*fop_<MODEF>_2_i387): Remove f/r alternative.
3563 (*fop_<MODEF>_3_i387): Likewise.
3564 (*fop_xf_2_i387, *fop_xf_3_i387): Likewise.
3565 (splitters for the fop_* register patterns): Remove.
3566 (fscalexf4_i387): Rename from *fscalexf4_i387.
3567 (ldexpxf3): Use gen_floatsixf2 and gen_fscalexf4_i387.
3569 2014-03-13 Jakub Jelinek <jakub@redhat.com>
3571 PR tree-optimization/59779
3572 * tree-dfa.c (get_ref_base_and_extent): Use double_int
3573 type for bitsize and maxsize instead of HOST_WIDE_INT.
3575 2014-03-13 Steven Bosscher <steven@gcc.gnu.org>
3577 PR rtl-optimization/57320
3578 * function.c (rest_of_handle_thread_prologue_and_epilogue): Cleanup
3579 the CFG after thread_prologue_and_epilogue_insns.
3581 2014-03-13 Vladimir Makarov <vmakarov@redhat.com>
3583 PR rtl-optimization/57189
3584 * lra-constraints.c (process_alt_operands): Disfavor spilling
3587 2014-03-13 Cesar Philippidis <cesar@codesourcery.com>
3589 * lto-wrapper.c (maybe_unlink_file): Suppress diagnostic messages.
3591 2014-03-13 Jakub Jelinek <jakub@redhat.com>
3593 PR tree-optimization/59025
3595 * tree-ssa-reassoc.c (sort_by_operand_rank): For SSA_NAMEs with the
3596 same rank, sort by bb_rank and gimple_uid of SSA_NAME_DEF_STMT first.
3598 2014-03-13 Georg-Johann Lay <avr@gjlay.de>
3601 * config/avr/avr.c (avr_out_plus): Swap cc_plus and cc_minus in
3602 calls of avr_out_plus_1.
3604 2014-03-13 Bin Cheng <bin.cheng@arm.com>
3606 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Record
3607 BB's single pred and update the father loop's latch info later.
3609 2014-03-12 Michael Meissner <meissner@linux.vnet.ibm.com>
3611 * config/rs6000/vector.md (VEC_L): Add V1TI mode to vector types.
3615 (VEC_base): Likewise.
3616 (mov<MODE>, VEC_M modes): If we are loading TImode into VSX
3617 registers, we need to swap double words in little endian mode.
3619 * config/rs6000/rs6000-modes.def (V1TImode): Add new vector mode
3620 to be a container mode for 128-bit integer operations added in ISA
3621 2.07. Unlike TImode and PTImode, the preferred register set is
3622 the Altivec/VMX registers for the 128-bit operations.
3624 * config/rs6000/rs6000-protos.h (rs6000_move_128bit_ok_p): Add
3626 (rs6000_split_128bit_ok_p): Likewise.
3628 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_3): Add new support
3629 macros for creating ISA 2.07 normal and overloaded builtin
3630 functions with 3 arguments.
3631 (BU_P8V_OVERLOAD_3): Likewise.
3632 (VPERM_1T): Add support for V1TImode in 128-bit vector operations
3633 for use as overloaded functions.
3634 (VPERM_1TI_UNS): Likewise.
3635 (VSEL_1TI): Likewise.
3636 (VSEL_1TI_UNS): Likewise.
3637 (ST_INTERNAL_1ti): Likewise.
3638 (LD_INTERNAL_1ti): Likewise.
3639 (XXSEL_1TI): Likewise.
3640 (XXSEL_1TI_UNS): Likewise.
3641 (VPERM_1TI): Likewise.
3642 (VPERM_1TI_UNS): Likewise.
3643 (XXPERMDI_1TI): Likewise.
3644 (SET_1TI): Likewise.
3645 (LXVD2X_V1TI): Likewise.
3646 (STXVD2X_V1TI): Likewise.
3647 (VEC_INIT_V1TI): Likewise.
3648 (VEC_SET_V1TI): Likewise.
3649 (VEC_EXT_V1TI): Likewise.
3650 (EQV_V1TI): Likewise.
3651 (NAND_V1TI): Likewise.
3652 (ORC_V1TI): Likewise.
3653 (VADDCUQ): Add support for 128-bit integer arithmetic instructions
3654 added in ISA 2.07. Add both normal 'altivec' builtins, and the
3656 (VADDUQM): Likewise.
3657 (VSUBCUQ): Likewise.
3658 (VADDEUQM): Likewise.
3659 (VADDECUQ): Likewise.
3660 (VSUBEUQM): Likewise.
3661 (VSUBECUQ): Likewise.
3663 * config/rs6000/rs6000-c.c (__int128_type): New static to hold
3664 __int128_t and __uint128_t types.
3665 (__uint128_type): Likewise.
3666 (altivec_categorize_keyword): Add support for vector __int128_t,
3667 vector __uint128_t, vector __int128, and vector unsigned __int128
3668 as a container type for TImode operations that need to be done in
3669 VSX/Altivec registers.
3670 (rs6000_macro_to_expand): Likewise.
3671 (altivec_overloaded_builtins): Add ISA 2.07 overloaded functions
3672 to support 128-bit integer instructions vaddcuq, vadduqm,
3673 vaddecuq, vaddeuqm, vsubcuq, vsubuqm, vsubecuq, vsubeuqm.
3674 (altivec_resolve_overloaded_builtin): Add support for V1TImode.
3676 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Add support
3677 for V1TImode, and set up preferences to use VSX/Altivec registers.
3678 Setup VSX reload handlers.
3679 (rs6000_debug_reg_global): Likewise.
3680 (rs6000_init_hard_regno_mode_ok): Likewise.
3681 (rs6000_preferred_simd_mode): Likewise.
3682 (vspltis_constant): Do not allow V1TImode as easy altivec constants.
3683 (easy_altivec_constant): Likewise.
3684 (output_vec_const_move): Likewise.
3685 (rs6000_expand_vector_set): Convert V1TImode set and extract to
3687 (rs6000_expand_vector_extract): Likewise.
3688 (reg_offset_addressing_ok_p): Setup V1TImode to use VSX reg+reg
3690 (rs6000_const_vec): Add support for V1TImode.
3691 (rs6000_emit_le_vsx_load): Swap double words when loading or
3692 storing TImode/V1TImode.
3693 (rs6000_emit_le_vsx_store): Likewise.
3694 (rs6000_emit_le_vsx_move): Likewise.
3695 (rs6000_emit_move): Add support for V1TImode.
3696 (altivec_expand_ld_builtin): Likewise.
3697 (altivec_expand_st_builtin): Likewise.
3698 (altivec_expand_vec_init_builtin): Likewise.
3699 (altivec_expand_builtin): Likewise.
3700 (rs6000_init_builtins): Add support for V1TImode type. Add
3701 support for ISA 2.07 128-bit integer builtins. Define type names
3702 for the VSX/Altivec vector types.
3703 (altivec_init_builtins): Add support for overloaded vector
3704 functions with V1TImode type.
3705 (rs6000_preferred_reload_class): Prefer Altivec registers for V1TImode.
3706 (rs6000_move_128bit_ok_p): Move 128-bit move/split validation to
3708 (rs6000_split_128bit_ok_p): Likewise.
3709 (rs6000_handle_altivec_attribute): Create V1TImode from vector
3710 __int128_t and vector __uint128_t.
3712 * config/rs6000/vsx.md (VSX_L): Add V1TImode to vector iterators
3713 and mode attributes.
3720 (VS_scalar): Likewise.
3721 (VS_double): Likewise.
3722 (vsx_set_v1ti): New builtin function to create V1TImode from TImode.
3724 * config/rs6000/rs6000.h (TARGET_VADDUQM): New macro to say whether
3725 we support the ISA 2.07 128-bit integer arithmetic instructions.
3726 (ALTIVEC_OR_VSX_VECTOR_MODE): Add V1TImode.
3727 (enum rs6000_builtin_type_index): Add fields to hold V1TImode
3728 and TImode types for use with the builtin functions.
3729 (V1TI_type_node): Likewise.
3730 (unsigned_V1TI_type_node): Likewise.
3731 (intTI_type_internal_node): Likewise.
3732 (uintTI_type_internal_node): Likewise.
3734 * config/rs6000/altivec.md (UNSPEC_VADDCUQ): New unspecs for ISA 2.07
3735 128-bit builtin functions.
3736 (UNSPEC_VADDEUQM): Likewise.
3737 (UNSPEC_VADDECUQ): Likewise.
3738 (UNSPEC_VSUBCUQ): Likewise.
3739 (UNSPEC_VSUBEUQM): Likewise.
3740 (UNSPEC_VSUBECUQ): Likewise.
3741 (VM): Add V1TImode to vector mode iterators.
3743 (VI_unit): Likewise.
3744 (altivec_vadduqm): Add ISA 2.07 128-bit binary builtins.
3745 (altivec_vaddcuq): Likewise.
3746 (altivec_vsubuqm): Likewise.
3747 (altivec_vsubcuq): Likewise.
3748 (altivec_vaddeuqm): Likewise.
3749 (altivec_vaddecuq): Likewise.
3750 (altivec_vsubeuqm): Likewise.
3751 (altivec_vsubecuq): Likewise.
3753 * config/rs6000/rs6000.md (FMOVE128_GPR): Add V1TImode to vector
3755 (BOOL_128): Likewise.
3756 (BOOL_REGS_OUTPUT): Likewise.
3757 (BOOL_REGS_OP1): Likewise.
3758 (BOOL_REGS_OP2): Likewise.
3759 (BOOL_REGS_UNARY): Likewise.
3760 (BOOL_REGS_AND_CR0): Likewise.
3762 * config/rs6000/altivec.h (vec_vaddcuq): Add support for ISA 2.07
3763 128-bit integer builtin support.
3764 (vec_vadduqm): Likewise.
3765 (vec_vaddecuq): Likewise.
3766 (vec_vaddeuqm): Likewise.
3767 (vec_vsubecuq): Likewise.
3768 (vec_vsubeuqm): Likewise.
3769 (vec_vsubcuq): Likewise.
3770 (vec_vsubuqm): Likewise.
3772 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
3773 Document vec_vaddcuq, vec_vadduqm, vec_vaddecuq, vec_vaddeuqm,
3774 vec_subecuq, vec_subeuqm, vec_vsubcuq, vec_vsubeqm builtins adding
3775 128-bit integer add/subtract to ISA 2.07.
3777 2014-03-12 Joern Rennecke <joern.rennecke@embecosm.com>
3779 * config/arc/arc.c (arc_predicate_delay_insns):
3780 Fix third argument passed to conditionalize_nonjump.
3782 2014-03-12 Yufeng Zhang <yufeng.zhang@arm.com>
3784 * config/aarch64/aarch64-builtins.c
3785 (aarch64_builtin_vectorized_function): Add BUILT_IN_LFLOORF,
3786 BUILT_IN_LLFLOOR, BUILT_IN_LCEILF and BUILT_IN_LLCEIL.
3787 * config/aarch64/arm_neon.h (vcvtaq_u64_f64): Call __builtin_llfloor
3788 instead of __builtin_lfloor.
3789 (vcvtnq_u64_f64): Call __builtin_llceil instead of __builtin_lceil.
3791 2014-03-12 Jakub Jelinek <jakub@redhat.com>
3793 * tree-ssa-ifcombine.c (forwarder_block_to): New function.
3794 (tree_ssa_ifcombine_bb_1): New function.
3795 (tree_ssa_ifcombine_bb): Use it. Handle also cases where else_bb
3796 is an empty forwarder block to then_bb or vice versa and then_bb
3797 and else_bb are effectively swapped.
3799 2014-03-12 Christian Bruel <christian.bruel@st.com>
3802 * config/arm/arm.c (arm_emit_vfp_multi_reg_pop): Emit a
3803 REG_CFA_DEF_CFA note.
3804 (arm_expand_epilogue_apcs_frame): call arm_add_cfa_adjust_cfa_note.
3805 (arm_unwind_emit): Allow REG_CFA_DEF_CFA.
3807 2014-03-12 Thomas Preud'homme <thomas.preudhomme@arm.com>
3809 PR tree-optimization/60454
3810 * tree-ssa-math-opts.c (find_bswap_1): Fix bswap detection.
3812 2014-03-12 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3814 * config.gcc (aarch64*-*-*): Use ISA flags from aarch64-arches.def.
3815 Do not define target_cpu_default2 to generic.
3816 * config/aarch64/aarch64.h (TARGET_CPU_DEFAULT): Use generic cpu.
3817 * config/aarch64/aarch64.c (aarch64_override_options): Update comment.
3818 * config/aarch64/aarch64-arches.def (armv8-a): Use generic cpu.
3820 2014-03-12 Jakub Jelinek <jakub@redhat.com>
3821 Marc Glisse <marc.glisse@inria.fr>
3823 PR tree-optimization/60502
3824 * tree-ssa-reassoc.c (eliminate_not_pairs): Use build_all_ones_cst
3825 instead of build_low_bits_mask.
3827 2014-03-12 Jakub Jelinek <jakub@redhat.com>
3830 * tree-vrp.c (register_edge_assert_for_1): Don't add assert
3831 if there are multiple uses, but op doesn't live on E edge.
3832 * tree-cfg.c (assert_unreachable_fallthru_edge_p): Also ignore
3833 clobber stmts before __builtin_unreachable.
3835 2014-03-11 Richard Sandiford <rdsandiford@googlemail.com>
3837 * builtins.c (expand_builtin_setjmp_receiver): Use and clobber
3838 hard_frame_pointer_rtx.
3839 * cse.c (cse_insn): Remove volatile check.
3840 * cselib.c (cselib_process_insn): Likewise.
3841 * dse.c (scan_insn): Likewise.
3843 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
3845 * config/arc/arc.c (conditionalize_nonjump): New function,
3847 (arc_ifcvt): ... this.
3848 (arc_predicate_delay_insns): Use it.
3850 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
3852 * config/arc/predicates.md (extend_operand): During/after reload,
3853 allow const_int_operand.
3854 * config/arc/arc.md (mulsidi3_700): Use extend_operand predicate.
3855 (umulsidi3_700): Likewise. Change operand 2 constraint back to "cL".
3856 (mulsi3_highpart): Change operand 2 constraint alternatives 2 and 3
3858 (umulsi3_highpart_i): Likewise.
3860 2014-03-11 Richard Biener <rguenther@suse.de>
3862 * tree-ssa-structalias.c (get_constraint_for_ptr_offset):
3863 Add asserts to guard possible wrong-code bugs.
3865 2014-03-11 Richard Biener <rguenther@suse.de>
3867 PR tree-optimization/60429
3868 PR tree-optimization/60485
3869 * tree-ssa-structalias.c (set_union_with_increment): Properly
3870 take into account all fields that overlap the shifted vars.
3871 (do_sd_constraint): Likewise.
3872 (do_ds_constraint): Likewise.
3873 (get_constraint_for_ptr_offset): Likewise.
3875 2014-03-11 Chung-Lin Tang <cltang@codesourcery.com>
3877 * config/nios2/nios2.c (machine_function): Add fp_save_offset field.
3878 (nios2_compute_frame_layout):
3879 Add calculation of cfun->machine->fp_save_offset.
3880 (nios2_expand_prologue): Correct setting of frame pointer register
3882 (nios2_expand_epilogue): Update recovery of stack pointer from
3883 frame pointer accordingly.
3884 (nios2_initial_elimination_offset): Update calculation of offset
3885 for eliminating to HARD_FRAME_POINTER_REGNUM.
3887 2014-03-10 Jakub Jelinek <jakub@redhat.com>
3890 * ipa.c (symtab_remove_unreachable_nodes): Don't call
3891 cgraph_get_create_node on VAR_DECLs.
3893 2014-03-10 Richard Biener <rguenther@suse.de>
3896 * tree.c (signed_or_unsigned_type_for): Handle OFFSET_TYPEs.
3898 2014-03-08 Douglas B Rupp <rupp@gnat.com>
3900 * config/vms/vms.opt (vms_float_format): New variable.
3902 2014-03-08 Tobias Burnus <burnus@net-b.de>
3904 * doc/invoke.texi (-fcilkplus): Update implementation status.
3906 2014-03-08 Paulo Matos <paulo@matos-sorge.com>
3907 Richard Biener <rguenther@suse.de>
3909 * lto-wrapper.c (merge_and_complain): Ensure -fshort-double is used
3910 consistently accross all TUs.
3911 (run_gcc): Enable -fshort-double automatically at link at link-time
3912 and disallow override.
3914 2014-03-08 Richard Sandiford <rdsandiford@googlemail.com>
3917 * config/mips/mips.c (mips_option_override): Promote -mpaired-single
3918 warning to an error. Disable TARGET_PAIRED_SINGLE and TARGET_MIPS3D
3919 if they can't be used.
3921 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3923 * configure.ac (HAVE_AS_IX86_TLSLDMPLT): Improve test
3924 for Solaris 11/x86 ld.
3925 * configure: Regenerate.
3927 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3929 * configure.ac (TLS_SECTION_ASM_FLAG): Save as tls_section_flag.
3930 (LIB_TLS_SPEC): Save as ld_tls_libs.
3931 (HAVE_AS_IX86_TLSLDMPLT): Define as 1/0.
3932 (HAVE_AS_IX86_TLSLDM): New test.
3933 * configure, config.in: Regenerate.
3934 * config/i386/i386.c (legitimize_tls_address): Fall back to
3935 TLS_MODEL_GLOBAL_DYNAMIC on 32-bit Solaris/x86 if tool chain
3936 cannot support TLS_MODEL_LOCAL_DYNAMIC.
3937 * config/i386/i386.md (*tls_local_dynamic_base_32_gnu): Use if
3938 instead of #ifdef in HAVE_AS_IX86_TLSLDMPLT test.
3940 2014-03-07 Paulo Matos <paulo@matos-sorge.com>
3942 * common.opt (fira-loop-pressure): Mark as optimization.
3944 2014-03-07 Thomas Schwinge <thomas@codesourcery.com>
3946 * langhooks.c (lhd_omp_mappable_type): The error_mark_node is not
3947 an OpenMP mappable type.
3949 2014-03-06 Matthias Klose <doko@ubuntu.com>
3951 * Makefile.in (s-mlib): Only pass MULTIARCH_DIRNAME if
3952 MULTILIB_OSDIRNAMES is not defined.
3954 2014-03-06 Jakub Jelinek <jakub@redhat.com>
3955 Meador Inge <meadori@codesourcery.com>
3958 * config/arm/arm.c (arm_tls_symbol_p): Remove.
3959 (arm_legitimize_address): Call legitimize_tls_address for any
3960 arm_tls_referenced_p expression, handle constant addend. Call it
3961 before testing for !TARGET_ARM.
3962 (thumb_legitimize_address): Don't handle arm_tls_symbol_p here.
3964 2014-03-06 Richard Biener <rguenther@suse.de>
3970 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
3972 * tree-streamer.c (record_common_node): Assert we don't record
3973 nodes with type double.
3974 (preload_common_node): Skip type double, complex double and double
3975 pointer since it is now frontend dependent due to fshort-double option.
3977 2014-03-06 Richard Biener <rguenther@suse.de>
3979 * gcc.c (PLUGIN_COND): Always enable unless -fno-use-linker-plugin
3980 or -fno-lto is specified and the linker has full plugin support.
3981 * collect2.c (lto_mode): Default to LTO_MODE_WHOPR if LTO is enabled.
3982 (main): Remove -flto processing, adjust lto_mode using use_plugin late.
3983 * lto-wrapper.c (merge_and_complain): Merge compile-time
3984 optimization levels.
3985 (run_gcc): And pass it through to the link options.
3987 2014-03-06 Alexandre Oliva <aoliva@redhat.com>
3991 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
3993 * cselib.c (remove_useless_values): Skip to avoid quadratic
3994 behavior if the condition moved from...
3995 (cselib_process_insn): ... here holds.
3997 2014-03-05 Jakub Jelinek <jakub@redhat.com>
4000 * Makefile.in (PLUGIN_HEADERS): Add tree-phinodes.h, stor-layout.h,
4001 ssa-iterators.h, $(RESOURCE_H) and tree-cfgcleanup.h.
4004 * config/i386/t-i386 (OPTIONS_H_EXTRA): Add stringop.def.
4005 (TM_H): Add x86-tune.def.
4007 2014-03-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4009 * config/aarch64/aarch64.c (generic_tunings):
4010 Use cortexa57_extra_costs.
4012 2014-03-05 Jakub Jelinek <jakub@redhat.com>
4015 * cfgexpand.c (expand_used_vars): Do not assume all SSA_NAMEs
4016 of PARM/RESULT_DECLs must be coalesced with optimize && in_lto_p.
4017 * tree-ssa-coalesce.c (coalesce_ssa_name): Use MUST_COALESCE_COST - 1
4020 2014-03-04 Heiher <r@hev.cc>
4022 * config/mips/mips-cpus.def (loongson3a): Mark as a MIPS64r2 processor.
4023 * config/mips/mips.h (MIPS_ISA_LEVEL_SPEC): Adjust accordingly.
4025 2014-03-04 Uros Bizjak <ubizjak@gmail.com>
4027 * config/i386/predicates.md (const2356_operand): Change to ...
4028 (const2367_operand): ... this.
4029 * config/i386/sse.md (avx512pf_scatterpf<mode>sf): Use
4031 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
4032 (*avx512pf_scatterpf<mode>sf): Ditto.
4033 (avx512pf_scatterpf<mode>df): Ditto.
4034 (*avx512pf_scatterpf<mode>df_mask): Ditto.
4035 (*avx512pf_scatterpf<mode>df): Ditto.
4036 * config/i386/i386.c (ix86_expand_builtin): Update
4037 incorrect hint operand error message.
4039 2014-03-04 Richard Biener <rguenther@suse.de>
4041 * lto-section-in.c (lto_get_section_data): Fix const cast.
4043 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
4045 * tree-streamer.c (record_common_node): Assert we don't record
4046 nodes with type double.
4047 (preload_common_node): Skip type double, complex double and double
4048 pointer since it is now frontend dependent due to fshort-double option.
4050 2014-03-04 Richard Biener <rguenther@suse.de>
4053 * lto-streamer-in.c (lto_read_body): Remove LTO bytecode version check.
4054 (lto_input_toplevel_asms): Likewise.
4055 * lto-section-in.c (lto_get_section_data): Instead do it here
4058 2014-03-04 Richard Biener <rguenther@suse.de>
4060 PR tree-optimization/60382
4061 * tree-vect-loop.c (vect_is_simple_reduction_1): Do not consider
4062 dead PHIs a reduction.
4064 2014-03-03 Uros Bizjak <ubizjak@gmail.com>
4066 * config/i386/xmmintrin.h (enum _mm_hint) <_MM_HINT_ET0>: Correct
4068 (_mm_prefetch): Move out of GCC target("sse") pragma.
4069 * config/i386/prfchwintrin.h (_m_prefetchw): Move out of
4070 GCC target("prfchw") pragma.
4071 * config/i386/i386.md (prefetch): Emit prefetchwt1 only
4073 * config/i386/i386.c (ix86_option_override_internal): Enable
4074 -mprfchw with -mprefetchwt1.
4076 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
4078 * config/arc/arc.md (casesi_load) <length attribute alternative 0>:
4081 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
4083 * opts.h (CL_PCH_IGNORE): Define.
4084 * targhooks.c (option_affects_pch_p):
4085 Return false for options that have CL_PCH_IGNORE set.
4086 * opt-functions.awk: Process PchIgnore.
4087 * doc/options.texi: Document PchIgnore.
4089 * config/arc/arc.opt (misize): Add PchIgnore property.
4091 2014-03-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4093 * config/rs6000/rs6000.c (rs6000_preferred_reload_class): Disallow
4094 reload of PLUS rtx's outside of GENERAL_REGS or BASE_REGS; relax
4095 constraint on constants to permit them being loaded into
4096 GENERAL_REGS or BASE_REGS.
4098 2014-03-03 Nick Clifton <nickc@redhat.com>
4100 * config/rl78/rl78-real.md (cbranchsi4_real_signed): Add
4101 anti-cacnonical alternatives.
4102 (negandhi3_real): New pattern.
4103 * config/rl78/rl78-virt.md (negandhi3_virt): New pattern.
4105 2014-03-03 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
4107 * config/avr/avr-mcus.def: Remove atxmega16x1.
4108 * config/avr/avr-tables.opt: Regenerate.
4109 * config/avr/t-multilib: Regenerate.
4110 * doc/avr-mmcu.texi: Regenerate.
4112 2014-03-03 Tobias Grosser <tobias@grosser.es>
4113 Mircea Namolaru <mircea.namolaru@inria.fr>
4115 PR tree-optimization/58028
4116 * graphite-clast-to-gimple.c (set_cloog_options): Don't remove
4119 2014-03-03 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4121 * config/arm/neon.md (*movmisalign<mode>): Legitimize addresses
4122 not handled by recognizers.
4124 2014-03-03 Jakub Jelinek <jakub@redhat.com>
4127 * function.c (expand_function_end): Don't emit
4128 clobber_return_register sequence if clobber_after is a BARRIER.
4129 * cfgexpand.c (construct_exit_block): Append instructions before
4130 return_label to prev_bb.
4132 2014-03-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4134 * config/rs6000/constraints.md: Document reserved use of "wc".
4136 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
4139 * ipa.c (function_and_variable_visibility): When dissolving comdat
4140 group, also set all symbols to local.
4142 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
4147 2013-12-14 Jan Hubicka <jh@suse.cz>
4149 * ipa-prop.c (stmt_may_be_vtbl_ptr_store): Skip clobbers.
4151 2014-03-02 Jon Beniston <jon@beniston.com>
4157 * config/lm32/lm32.c (lm32_legitimate_constant_p): Remove, as incorrect.
4158 (TARGET_LEGITIMATE_CONSTANT_P): Undefine, as not needed.
4159 * config/lm32/lm32.md (movsi_insn): Add 32-bit immediate support.
4160 (simple_return, *simple_return): New patterns
4161 * config/lm32/predicates.md (movsi_rhs_operand): Remove as obsolete.
4162 * configure.ac (force_sjlj_exceptions): Force sjlj exceptions for lm32.
4164 2014-03-01 Paolo Carlini <paolo.carlini@oracle.com>
4166 * dwarf2out.c (gen_subprogram_die): Tidy.
4168 2014-03-01 Oleg Endo <olegendo@gcc.gnu.org>
4171 * config/sh/sh.md (*mov_t_msb_neg): Split into ...
4172 (*mov_t_msb_neg_negc): ... this new insn.
4174 2014-02-28 Jason Merrill <jason@redhat.com>
4177 * ipa-devirt.c (ipa_devirt): Don't choose an implicitly-declared
4180 2014-02-28 Paolo Carlini <paolo.carlini@oracle.com>
4183 * dwarf2out.c (decltype_auto_die): New static.
4184 (gen_subprogram_die): Handle 'decltype(auto)' like 'auto'.
4185 (gen_type_die_with_usage): Handle 'decltype(auto)'.
4186 (is_cxx_auto): Likewise.
4188 2014-02-28 Ian Bolton <ian.bolton@arm.com>
4190 * config/aarch64/aarch64.h: Define __ARM_NEON by default if
4191 we are not using general regs only.
4193 2014-02-28 Richard Biener <rguenther@suse.de>
4196 * tree-cfgcleanup.c (tree_forwarder_block_p): Restrict
4197 previous fix and only allow to remove trivial pre-headers
4198 and latches. Also honor LOOPS_MAY_HAVE_MULTIPLE_LATCHES.
4199 (remove_forwarder_block): Properly update the latch of a loop.
4201 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4204 * cselib.c (cselib_hasher::equal): Special-case VALUE lookup.
4205 (cselib_preserved_hash_table): New.
4206 (preserve_constants_and_equivs): Move preserved vals to it.
4207 (cselib_find_slot): Look it up first.
4208 (cselib_init): Initialize it.
4209 (cselib_finish): Release it.
4210 (dump_cselib_table): Dump it.
4212 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4215 * cselib.c (remove_useless_values): Skip to avoid quadratic
4216 behavior if the condition moved from...
4217 (cselib_process_insn): ... here holds.
4219 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4222 * var-tracking.c (vt_initialize): Apply the same condition to
4223 preserve the CFA base value.
4225 2014-02-28 Joey Ye <joey.ye@arm.com>
4228 * config/arm/arm.c (thumb_far_jump_used_p): Don't change
4229 if reload in progress or completed.
4231 2014-02-28 Tobias Burnus <burnus@net-b.de>
4234 * tree-pretty-print.c (dump_generic_node, print_declaration): Handle
4237 2014-02-27 H.J. Lu <hongjiu.lu@intel.com>
4239 * doc/tm.texi.in (Condition Code Status): Update documention for
4240 relative locations of cc0-setter and cc0-user.
4242 2014-02-27 Jeff Law <law@redhat.com>
4244 PR rtl-optimization/52714
4245 * combine.c (try_combine): When splitting an unrecognized PARALLEL
4246 into two independent simple sets, if I3 is a jump, ensure the
4247 pattern we place into I3 is a (set (pc) ...).
4249 2014-02-27 Mikael Pettersson <mikpe@it.uu.se>
4250 Jeff Law <law@redhat.com>
4252 PR rtl-optimization/49847
4253 * cse.c (fold_rtx) Handle case where cc0 setter and cc0 user
4254 are in different blocks.
4255 * doc/tm.texi (Condition Code Status): Update documention for
4256 relative locations of cc0-setter and cc0-user.
4258 2014-02-27 Vladimir Makarov <vmakarov@redhat.com>
4261 * lra.c (lra_emit_add): Check SUBREG too.
4263 2014-02-27 Andreas Schwab <schwab@suse.de>
4265 * config/m68k/m68k.c (m68k_option_override): Disable
4266 -flive-range-shrinkage for classic m68k.
4267 (m68k_override_options_after_change): Likewise.
4269 2014-02-27 Marek Polacek <polacek@redhat.com>
4272 * tree-ssa-uninit.c (gate_warn_uninitialized): Run the pass even for
4273 -Wmaybe-uninitialized.
4275 2014-02-27 Alan Modra <amodra@gmail.com>
4278 * reload1.c (emit_input_reload_insns): When reload_override_in,
4279 set old to rl->in_reg when rl->in_reg is a subreg.
4281 2014-02-26 Richard Biener <rguenther@suse.de>
4284 * lra-assigns.c (spill_for): Avoid mixed-sign comparison.
4286 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
4288 * common/config/i386/predicates.md (const1256_operand): Remove.
4289 (const2356_operand): New.
4290 (const_1_to_2_operand): Remove.
4291 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
4292 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
4293 (*avx512pf_gatherpf<mode>sf): Ditto.
4294 (avx512pf_gatherpf<mode>df): Ditto.
4295 (*avx512pf_gatherpf<mode>df_mask): Ditto.
4296 (*avx512pf_gatherpf<mode>df): Ditto.
4297 (avx512pf_scatterpf<mode>sf): Ditto.
4298 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
4299 (*avx512pf_scatterpf<mode>sf): Ditto.
4300 (avx512pf_scatterpf<mode>df): Ditto.
4301 (*avx512pf_scatterpf<mode>df_mask): Ditto.
4302 (*avx512pf_scatterpf<mode>df): Ditto.
4303 * common/config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET0.
4305 2014-02-26 Ilya Tocar <ilya.tocar@intel.com>
4307 * config/i386/avx512fintrin.h (_mm512_testn_epi32_mask),
4308 (_mm512_mask_testn_epi32_mask), (_mm512_testn_epi64_mask),
4309 (_mm512_mask_testn_epi64_mask): Move to ...
4310 * config/i386/avx512cdintrin.h: Here.
4311 * config/i386/i386.c (bdesc_args): Change MASK_ISA for testnm.
4312 * config/i386/sse.md (avx512f_vmscalef<mode><round_name>): Remove %.
4313 (avx512f_scalef<mode><mask_name><round_name>): Ditto.
4314 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Change conditon to
4315 TARGET_AVX512F from TARGET_AVX512CD.
4317 2014-02-26 Richard Biener <rguenther@suse.de>
4320 * ipa.c (walk_polymorphic_call_targets): Properly guard
4321 call to inline_update_overall_summary.
4323 2014-02-26 Bin Cheng <bin.cheng@arm.com>
4326 * tree-cfgcleanup.c (tree_forwarder_block_p): Protect loop preheaders
4327 and latches only if requested. Fix latch if it is removed.
4328 * tree-ssa-dom.c (tree_ssa_dominator_optimize): Set
4329 LOOPS_HAVE_PREHEADERS.
4331 2014-02-25 Andrew Pinski <apinski@cavium.com>
4333 * builtins.c (expand_builtin_thread_pointer): Create a new target
4334 when the target is NULL.
4336 2014-02-25 Vladimir Makarov <vmakarov@redhat.com>
4338 PR rtl-optimization/60317
4339 * params.def (PARAM_LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
4340 * params.h (LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
4341 * lra-assigns.c: Include params.h.
4342 (spill_for): Use LRA_MAX_CONSIDERED_RELOAD_PSEUDOS as guard for
4343 other reload pseudos considerations.
4345 2014-02-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4347 * config/rs6000/vector.md (*vector_unordered<mode>): Change split
4348 to use canonical form for nor<mode>3.
4350 2014-02-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4353 * config/arm/arm.h (CANNOT_CHANGE_MODE_CLASS): Allow 128 to 64-bit
4356 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
4358 * common/config/i386/i386-common.c (OPTION_MASK_ISA_PREFETCHWT1_SET),
4359 (OPTION_MASK_ISA_PREFETCHWT1_UNSET): New.
4360 (ix86_handle_option): Handle OPT_mprefetchwt1.
4361 * config/i386/cpuid.h (bit_PREFETCHWT1): New.
4362 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
4364 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
4365 OPTION_MASK_ISA_PREFETCHWT1.
4366 * config/i386/i386.c (ix86_target_string): Handle mprefetchwt1.
4367 (PTA_PREFETCHWT1): New.
4368 (ix86_option_override_internal): Handle PTA_PREFETCHWT1.
4369 (ix86_valid_target_attribute_inner_p): Handle OPT_mprefetchwt1.
4370 * config/i386/i386.h (TARGET_PREFETCHWT1, TARGET_PREFETCHWT1_P): New.
4371 * config/i386/i386.md (prefetch): Check TARGET_PREFETCHWT1
4372 (*prefetch_avx512pf_<mode>_: Change into ...
4373 (*prefetch_prefetchwt1_<mode>: This.
4374 * config/i386/i386.opt (mprefetchwt1): New.
4375 * config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET1.
4376 (_mm_prefetch): Handle intent to write.
4377 * doc/invoke.texi (mprefetchwt1), (mno-prefetchwt1): Doccument.
4379 2014-02-25 Richard Biener <rguenther@suse.de>
4382 * emit-rtl.c (mem_attrs_htab): Remove.
4383 (mem_attrs_htab_hash): Likewise.
4384 (mem_attrs_htab_eq): Likewise.
4385 (set_mem_attrs): Always allocate new mem-attrs when something changed.
4386 (init_emit_once): Do not allocate mem_attrs_htab.
4388 2014-02-25 Richard Biener <rguenther@suse.de>
4391 * lto-opts.c (lto_write_options): Output non-explicit conservative
4392 -fwrapv, -fno-trapv and -fno-strict-overflow.
4393 * lto-wrapper.c (merge_and_complain): Handle merging those options.
4394 (run_gcc): And pass them through.
4396 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
4398 * sel-sched.c (calculate_new_fences): New parameter ptime.
4399 Calculate it as a maximum over all fence cycles.
4400 (sel_sched_region_2): Adjust the call to calculate_new_fences.
4401 Print the final schedule timing when sched_verbose.
4403 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
4405 PR rtl-optimization/60292
4406 * sel-sched.c (fill_vec_av_set): Do not reset target availability
4407 bit fot the fence instruction.
4409 2014-02-24 Alangi Derick <alangiderick@gmail.com>
4411 * calls.h: Fix typo in comment.
4413 2014-02-24 John David Anglin <danglin@gcc.gnu.org>
4415 * config/pa/pa.c (pa_output_move_double): Don't valididate when
4416 adjusting offsetable addresses.
4418 2014-02-24 Guozhi Wei <carrot@google.com>
4420 * sparseset.h (sparseset_pop): Fix the wrong index.
4422 2014-02-24 Walter Lee <walt@tilera.com>
4424 * config.gcc (tilepro-*-*): Change to tilepro*-*-*.
4425 (tilegx-*-linux*): Change to tilegx*-*-linux*; Support tilegxbe
4427 * common/config/tilegx/tilegx-common.c
4428 (TARGET_DEFAULT_TARGET_FLAGS): Define.
4429 * config/tilegx/linux.h (ASM_SPEC): Add endian_spec.
4431 * config/tilegx/sync.md (atomic_test_and_set): Handle big endian.
4432 * config/tilegx/tilegx.c (tilegx_return_in_msb): New.
4433 (tilegx_gimplify_va_arg_expr): Handle big endian.
4434 (tilegx_expand_unaligned_load): Ditto.
4435 (tilegx_expand_unaligned_store): Ditto.
4436 (TARGET_RETURN_IN_MSB): New.
4437 * config/tilegx/tilegx.h (TARGET_DEFAULT): New.
4438 (TARGET_ENDIAN_DEFAULT): New.
4439 (TARGET_BIG_ENDIAN): Handle big endian.
4440 (BYTES_BIG_ENDIAN): Ditto.
4441 (WORDS_BIG_ENDIAN): Ditto.
4442 (FLOAT_WORDS_BIG_ENDIAN): Ditto.
4445 * config/tilegx/tilegx.md (extv): Handle big endian.
4447 (insn_st<n>): Ditto.
4448 (insn_st<n>_add<bitsuffix>): Ditto.
4449 (insn_stnt<n>): Ditto.
4450 (insn_stnt<n>_add<bitsuffix>):Ditto.
4451 (vec_interleave_highv8qi): Handle big endian.
4452 (vec_interleave_highv8qi_be): New.
4453 (vec_interleave_highv8qi_le): New.
4454 (insn_v1int_h): Handle big endian.
4455 (vec_interleave_lowv8qi): Handle big endian.
4456 (vec_interleave_lowv8qi_be): New.
4457 (vec_interleave_lowv8qi_le): New.
4458 (insn_v1int_l): Handle big endian.
4459 (vec_interleave_highv4hi): Handle big endian.
4460 (vec_interleave_highv4hi_be): New.
4461 (vec_interleave_highv4hi_le): New.
4462 (insn_v2int_h): Handle big endian.
4463 (vec_interleave_lowv4hi): Handle big endian.
4464 (vec_interleave_lowv4hi_be): New.
4465 (vec_interleave_lowv4hi_le): New.
4466 (insn_v2int_l): Handle big endian.
4467 (vec_interleave_highv2si): Handle big endian.
4468 (vec_interleave_highv2si_be): New.
4469 (vec_interleave_highv2si_le): New.
4470 (insn_v4int_h): Handle big endian.
4471 (vec_interleave_lowv2si): Handle big endian.
4472 (vec_interleave_lowv2si_be): New.
4473 (vec_interleave_lowv2si_le): New.
4474 (insn_v4int_l): Handle big endian.
4475 * config/tilegx/tilegx.opt (mbig-endian): New option.
4476 (mlittle-endian): New option.
4477 * doc/install.texi: Document tilegxbe-linux.
4478 * doc/invoke.texi: Document -mbig-endian and -mlittle-endian.
4480 2014-02-24 Martin Jambor <mjambor@suse.cz>
4483 * ipa-cp.c (propagate_constants_accross_call): Bail out early if
4484 there are no parameter descriptors.
4486 2014-02-24 Andrey Belevantsev <abel@ispras.ru>
4488 PR rtl-optimization/60268
4489 * sched-rgn.c (haifa_find_rgns): Move the nr_regions_initial variable
4490 initialization to ...
4491 (sched_rgn_init): ... here.
4492 (schedule_region): Check for SCHED_PRESSURE_NONE earlier.
4494 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
4496 * config/microblaze/microblaze.md: Correct ashrsi_reg / lshrsi_reg
4499 2014-02-23 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
4501 * config/microblaze/microblaze.h: Remove SECONDARY_MEMORY_NEEDED
4504 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
4506 * /config/microblaze/microblaze.c: Add microblaze_asm_output_mi_thunk,
4507 define TARGET_ASM_OUTPUT_MI_THUNK and TARGET_ASM_CAN_OUTPUT_MI_THUNK.
4509 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
4511 * config/microblaze/predicates.md: Add cmp_op predicate.
4512 * config/microblaze/microblaze.md: Add branch_compare instruction
4513 which uses cmp_op predicate and emits cmp insn before branch.
4514 * config/microblaze/microblaze.c (microblaze_emit_compare): Rename
4515 to microblaze_expand_conditional_branch and consolidate logic.
4516 (microblaze_expand_conditional_branch): emit branch_compare
4517 insn instead of handling cmp op separate from branch insn.
4519 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4521 * config/rs6000/rs6000.c (rs6000_emit_le_vsx_move): Relax assert
4524 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4526 * config/rs6000/altivec.md (altivec_lve<VI_char>x): Replace
4527 define_insn with define_expand and new define_insn
4528 *altivec_lve<VI_char>x_internal.
4529 (altivec_stve<VI_char>x): Replace define_insn with define_expand
4530 and new define_insn *altivec_stve<VI_char>x_internal.
4531 * config/rs6000/rs6000-protos.h (altivec_expand_stvex_be): New
4533 * config/rs6000/rs6000.c (altivec_expand_lvx_be): Document use by
4535 (altivec_expand_stvex_be): New function.
4537 2014-02-22 Joern Rennecke <joern.rennecke@embecosm.com>
4539 * config/avr/avr.c (avr_can_eliminate): Allow elimination from
4540 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM if !frame_pointer_needed.
4541 * config/avr/avr.c (ELIMINABLE_REGS): Add elimination from
4542 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM.
4544 2014-02-21 Vladimir Makarov <vmakarov@redhat.com>
4547 * lra-constraints.c (inherit_reload_reg): Use lra_emit_move
4548 instead of emit_move_insn.
4550 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4552 * config/rs6000/altivec.md (altivec_vsumsws): Replace second
4554 (reduc_uplus_v16qi): Use gen_altivec_vsumsws_direct instead of
4555 gen_altivec_vsumsws.
4557 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4559 * config/rs6000/altivec.md (altivec_lvxl): Rename as
4560 *altivec_lvxl_<mode>_internal and use VM2 iterator instead of V4SI.
4561 (altivec_lvxl_<mode>): New define_expand incorporating
4562 -maltivec=be semantics where needed.
4563 (altivec_lvx): Rename as *altivec_lvx_<mode>_internal.
4564 (altivec_lvx_<mode>): New define_expand incorporating -maltivec=be
4565 semantics where needed.
4566 (altivec_stvx): Rename as *altivec_stvx_<mode>_internal.
4567 (altivec_stvx_<mode>): New define_expand incorporating
4568 -maltivec=be semantics where needed.
4569 (altivec_stvxl): Rename as *altivec_stvxl_<mode>_internal and use
4570 VM2 iterator instead of V4SI.
4571 (altivec_stvxl_<mode>): New define_expand incorporating
4572 -maltivec=be semantics where needed.
4573 * config/rs6000/rs6000-builtin.def: Add new built-in definitions
4574 LVXL_V2DF, LVXL_V2DI, LVXL_V4SF, LVXL_V4SI, LVXL_V8HI, LVXL_V16QI,
4575 LVX_V2DF, LVX_V2DI, LVX_V4SF, LVX_V4SI, LVX_V8HI, LVX_V16QI, STVX_V2DF,
4576 STVX_V2DI, STVX_V4SF, STVX_V4SI, STVX_V8HI, STVX_V16QI, STVXL_V2DF,
4577 STVXL_V2DI, STVXL_V4SF, STVXL_V4SI, STVXL_V8HI, STVXL_V16QI.
4578 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Replace
4579 ALTIVEC_BUILTIN_LVX with ALTIVEC_BUILTIN_LVX_<MODE> throughout;
4580 similarly for ALTIVEC_BUILTIN_LVXL, ALTIVEC_BUILTIN_STVX, and
4581 ALTIVEC_BUILTIN_STVXL.
4582 * config/rs6000/rs6000-protos.h (altivec_expand_lvx_be): New prototype.
4583 (altivec_expand_stvx_be): Likewise.
4584 * config/rs6000/rs6000.c (swap_selector_for_mode): New function.
4585 (altivec_expand_lvx_be): Likewise.
4586 (altivec_expand_stvx_be): Likewise.
4587 (altivec_expand_builtin): Add cases for
4588 ALTIVEC_BUILTIN_STVX_<MODE>, ALTIVEC_BUILTIN_STVXL_<MODE>,
4589 ALTIVEC_BUILTIN_LVXL_<MODE>, and ALTIVEC_BUILTIN_LVX_<MODE>.
4590 (altivec_init_builtins): Add definitions for
4591 __builtin_altivec_lvxl_<mode>, __builtin_altivec_lvx_<mode>,
4592 __builtin_altivec_stvx_<mode>, and __builtin_altivec_stvxl_<mode>.
4594 2014-02-21 Catherine Moore <clm@codesourcery.com>
4596 * doc/invoke.texi (mvirt, mno-virt): Document.
4597 * config/mips/mips.opt (mvirt): New option.
4598 * config/mips/mips.h (ASM_SPEC): Pass mvirt to the assembler.
4600 2014-02-21 Richard Biener <rguenther@suse.de>
4602 PR tree-optimization/60276
4603 * tree-vectorizer.h (struct _stmt_vec_info): Add min_neg_dist field.
4604 (STMT_VINFO_MIN_NEG_DIST): New macro.
4605 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Record
4606 STMT_VINFO_MIN_NEG_DIST.
4607 * tree-vect-stmts.c (vectorizable_load): Verify if assumptions
4608 made for negative dependence distances still hold.
4610 2014-02-21 Richard Biener <rguenther@suse.de>
4613 * tree-ssa-live.c (mark_all_vars_used_1): Do not walk
4614 DECL_INITIAL for globals not in the current function context.
4616 2014-02-21 Jakub Jelinek <jakub@redhat.com>
4618 PR tree-optimization/56490
4619 * params.def (PARAM_UNINIT_CONTROL_DEP_ATTEMPTS): New param.
4620 * tree-ssa-uninit.c: Include params.h.
4621 (compute_control_dep_chain): Add num_calls argument, return false
4622 if it exceed PARAM_UNINIT_CONTROL_DEP_ATTEMPTS param, pass
4623 num_calls to recursive call.
4624 (find_predicates): Change dep_chain into normal array,
4625 cur_chain into auto_vec<edge, MAX_CHAIN_LEN + 1>, add num_calls
4626 variable and adjust compute_control_dep_chain caller.
4627 (find_def_preds): Likewise.
4629 2014-02-21 Thomas Schwinge <thomas@codesourcery.com>
4631 * gimple-pretty-print.c (dump_gimple_omp_for) [flags & TDF_RAW]
4632 <case GF_OMP_FOR_KIND_CILKSIMD>: Add missing break statement.
4634 2014-02-21 Nick Clifton <nickc@redhat.com>
4636 * config/stormy16/stormy16.md (pushdqi1): Add mode to post_inc.
4637 (pushhi1): Likewise.
4638 (popqi1): Add mode to pre_dec.
4641 2014-02-21 Jakub Jelinek <jakub@redhat.com>
4643 * config/i386/i386.c (ix86_expand_vec_perm): Use V8SImode
4644 mode for mask of V8SFmode permutation.
4646 2014-02-20 Richard Henderson <rth@redhat.com>
4649 * builtins.c (expand_builtin_atomic_compare_exchange): Always make
4650 a new pseudo for OLDVAL.
4652 2014-02-20 Jakub Jelinek <jakub@redhat.com>
4655 * config/i386/i386.c (expand_vec_perm_interleave2): Don't call
4656 gen_reg_rtx if d->testing_p.
4657 (expand_vec_perm_pshufb2, expand_vec_perm_broadcast_1): Return early
4658 if d->testing_p and we will certainly return true.
4659 (expand_vec_perm_even_odd_1): Likewise. Don't call gen_reg_rtx
4662 2014-02-20 Uros Bizjak <ubizjak@gmail.com>
4664 * emit-rtl.c (gen_reg_rtx): Assert that
4665 crtl->emit.regno_pointer_align_length is non-zero.
4667 2014-02-20 Richard Henderson <rth@redhat.com>
4670 * builtins.c (expand_builtin_atomic_compare_exchange): Conditionalize
4671 on failure the store back into EXPECT.
4673 2014-02-20 Chung-Lin Tang <cltang@codesourcery.com>
4674 Sandra Loosemore <sandra@codesourcery.com>
4676 * config/nios2/nios2.md (unspec): Add UNSPEC_PIC_GOTOFF_SYM enum.
4677 * config/nios2/nios2.c (nios2_function_profiler): Add
4678 -fPIC (flag_pic == 2) support.
4679 (nios2_handle_custom_fpu_cfg): Fix warning parameter.
4680 (nios2_large_offset_p): New function.
4681 (nios2_unspec_reloc_p): Move up position, update to use
4682 nios2_large_offset_p.
4683 (nios2_unspec_address): Remove function.
4684 (nios2_unspec_offset): New function.
4685 (nios2_large_got_address): New function.
4686 (nios2_got_address): Add large offset support.
4687 (nios2_legitimize_tls_address): Update usage of removed and new
4689 (nios2_symbol_binds_local_p): New function.
4690 (nios2_load_pic_address): Add -fPIC (flag_pic == 2) support.
4691 (nios2_legitimize_address): Update to use nios2_large_offset_p.
4692 (nios2_emit_move_sequence): Avoid legitimizing (const (unspec ...)).
4693 (nios2_print_operand): Merge H/L processing, add hiadj/lo
4694 processing for (const (unspec ...)).
4695 (nios2_unspec_reloc_name): Add UNSPEC_PIC_GOTOFF_SYM case.
4697 2014-02-20 Richard Biener <rguenther@suse.de>
4699 * tree-cfg.c (replace_uses_by): Mark altered BBs before
4700 doing the substitution.
4701 (verify_gimple_assign_single): Also verify bare MEM_REFs on the lhs.
4703 2014-02-20 Martin Jambor <mjambor@suse.cz>
4706 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Uce correct
4707 info when checking whether lattices are bottom.
4709 2014-02-20 Richard Biener <rguenther@suse.de>
4712 * tree-eh.c (execute_cleanup_eh_1): Also cleanup empty EH
4715 2014-02-20 Jan Hubicka <hubicka@ucw.cz>
4718 * ipa-inline-transform.c (clone_inlined_nodes): Add freq_scale
4719 parameter specifying the scaling.
4720 (inline_call): Update.
4721 (want_inline_recursively): Guard division by zero.
4722 (recursive_inlining): Update.
4723 * ipa-inline.h (clone_inlined_nodes): Update.
4725 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
4728 * config/i386/i386.c (classify_argument): Pass structures of size
4729 64 bytes or less in register.
4731 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
4732 Kirill Yukhin <kirill.yukhin@intel.com>
4734 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): Swap operands.
4735 (_mm_rcp28_round_ss): Ditto.
4736 (_mm_rsqrt28_round_sd): Ditto.
4737 (_mm_rsqrt28_round_ss): Ditto.
4738 * config/i386/avx512erintrin.h (_mm_rcp14_round_sd): Ditto.
4739 (_mm_rcp14_round_ss): Ditto.
4740 (_mm_rsqrt14_round_sd): Ditto.
4741 (_mm_rsqrt14_round_ss): Ditto.
4742 * config/i386/sse.md (rsqrt14<mode>): Put nonimmediate operand as
4743 the first input operand, get rid of match_dup.
4744 (avx512er_exp2<mode><mask_name><round_saeonly_name>): Set type
4746 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
4748 (avx512er_vmrcp28<mode><round_saeonly_name>): Put nonimmediate
4749 operand as the first input operand, set type attribute.
4750 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
4752 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Put nonimmediate
4753 operand as the first input operand, set type attribute.
4755 2014-02-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4757 * config/rs6000/rs6000.c (vspltis_constant): Fix most significant
4760 2014-02-19 H.J. Lu <hongjiu.lu@intel.com>
4763 * config/i386/i386.c (construct_container): Remove TFmode check
4764 for X86_64_INTEGER_CLASS.
4766 2014-02-19 Uros Bizjak <ubizjak@gmail.com>
4769 * config/i386/i386.c (type_natural_mode): Warn for ABI changes
4770 only when -Wpsabi is enabled.
4772 2014-02-19 Michael Hudson-Doyle <michael.hudson@linaro.org>
4775 * config/aarch64/aarch64.c (aarch64_pass_by_reference): The rules for
4776 passing arrays in registers are the same as for structs, so remove the
4777 special case for them.
4779 2014-02-19 Eric Botcazou <ebotcazou@adacore.com>
4781 * expr.c (expand_expr_real_1) <case VIEW_CONVERT_EXPR>: For a bit-field
4782 destination type, extract only the valid bits if the source type is not
4783 integral and has a different mode.
4785 2014-02-19 Richard Biener <rguenther@suse.de>
4788 * tree-inline.c (estimate_num_insns): Avoid calling cgraph_get_node
4791 2014-02-19 Richard Biener <rguenther@suse.de>
4794 * ipa-prop.c: Include stringpool.h and tree-ssanames.h.
4795 (ipa_modify_call_arguments): Emit an argument load explicitely and
4796 preserve virtual SSA form there and for the replacement call.
4797 Do not update SSA form nor free dominance info.
4799 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
4801 * ipa.c (function_and_variable_visibility): Also clear WEAK
4802 flag when disolving COMDAT_GROUP.
4804 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
4806 * ipa-prop.h (ipa_ancestor_jf_data): Update ocmment.
4807 * ipa-prop.c (ipa_set_jf_known_type): Return early when
4809 (ipa_set_ancestor_jf): Set type to NULL hwen it is not preserved;
4810 do more sanity checks.
4811 (detect_type_change): Return true when giving up early.
4812 (compute_complex_assign_jump_func): Fix type parameter of
4813 ipa_set_ancestor_jf.
4814 (compute_complex_ancestor_jump_func): Likewise.
4815 (update_jump_functions_after_inlining): Fix updating of
4817 * ipa-cp.c (ipa_get_jf_ancestor_result): Be ready for type to be NULL.
4819 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
4821 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Also remove
4822 inline clones when edge disappears.
4824 2014-02-18 Michael Meissner <meissner@linux.vnet.ibm.com>
4827 * config/rs6000/rs6000.md (mov<mode>_64bit, TF/TDmode moves):
4828 Split 64-bit moves into 2 patterns. Do not allow the use of
4829 direct move for TDmode in little endian, since the decimal value
4830 has little endian bytes within a word, but the 64-bit pieces are
4831 ordered in a big endian fashion, and normal subreg's of TDmode are
4833 (mov<mode>_64bit_dm): Likewise.
4834 (movtd_64bit_nodm): Likewise.
4836 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
4838 PR tree-optimization/60174
4839 * tree-ssa-reassoc.c (init_range_entry): Do not look into the defining
4840 statement of an SSA_NAME that occurs in an abnormal PHI node.
4842 2014-02-18 Jakub Jelinek <jakub@redhat.com>
4845 * final.c (SEEN_BB): Remove.
4846 (SEEN_NOTE, SEEN_EMITTED): Renumber.
4847 (final_scan_insn): Don't force_source_line on second
4848 NOTE_INSN_BASIC_BLOCK.
4850 2014-02-18 Uros Bizjak <ubizjak@gmail.com>
4853 * config/i386/i386.h (struct ix86_args): Add warn_avx512f.
4854 * config/i386/i386.c (init_cumulative_args): Initialize warn_avx512f.
4855 (type_natural_mode): Warn ABI change when %zmm register is not
4856 available for AVX512F vector value passing.
4858 2014-02-18 Kai Tietz <ktietz@redhat.com>
4861 * config/i386/i386.c (ix86_expand_prologue): Use value in
4862 rax register as displacement when restoring %r10 or %rax.
4863 Fix wrong offset when restoring both registers.
4865 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
4867 * ipa-prop.c (compute_complex_ancestor_jump_func): Replace overzealous
4868 assertion with conditional return.
4870 2014-02-18 Jakub Jelinek <jakub@redhat.com>
4871 Uros Bizjak <ubizjak@gmail.com>
4874 * config/i386/driver-i386.c (host_detect_local_cpu): If
4875 YMM state is not saved by the OS, also clear has_f16c. Move
4876 CPUID 0x80000001 handling before YMM state saving checking.
4878 2014-02-18 Andrey Belevantsev <abel@ispras.ru>
4880 PR rtl-optimization/58960
4881 * haifa-sched.c (alloc_global_sched_pressure_data): New,
4882 factored out from ...
4883 (sched_init): ... here.
4884 (free_global_sched_pressure_data): New, factored out from ...
4885 (sched_finish): ... here.
4886 * sched-int.h (free_global_sched_pressure_data): Declare.
4887 * sched-rgn.c (nr_regions_initial): New static global.
4888 (haifa_find_rgns): Initialize it.
4889 (schedule_region): Disable sched-pressure for the newly
4892 2014-02-17 Richard Biener <rguenther@suse.de>
4894 * tree-vect-stmts.c (free_stmt_vec_info): Clear BB and
4895 release SSA defs of pattern stmts.
4897 2014-02-17 Richard Biener <rguenther@suse.de>
4899 * tree-inline.c (expand_call_inline): Release the virtual
4900 operand defined by the call we are about to inline.
4902 2014-02-17 Richard Biener <rguenther@suse.de>
4904 * tree-ssa.c (verify_ssa): If verify_def found an error, ICE.
4906 2014-02-17 Kirill Yukhin <kirill.yukhin@intel.com>
4907 Ilya Tocar <ilya.tocar@intel.com>
4909 * config/i386/avx512fintrin.h (_mm512_maskz_permutexvar_epi64): Swap
4910 arguments order in builtin.
4911 (_mm512_permutexvar_epi64): Ditto.
4912 (_mm512_mask_permutexvar_epi64): Ditto
4913 (_mm512_maskz_permutexvar_epi32): Ditto
4914 (_mm512_permutexvar_epi32): Ditto
4915 (_mm512_mask_permutexvar_epi32): Ditto
4917 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4919 * config/rs6000/altivec.md (p8_vmrgew): Handle little endian targets.
4920 (p8_vmrgow): Likewise.
4922 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4924 * config/rs6000/vsx.md (vsx_xxpermdi_<mode>): Handle little
4927 2014-02-15 Michael Meissner <meissner@linux.vnet.ibm.com>
4930 * config/rs6000/rs6000.md (rreg): Add TFmode, TDmode constraints.
4931 (mov<mode>_internal, TFmode/TDmode): Split TFmode/TDmode moves
4932 into 64-bit and 32-bit moves. On 64-bit moves, add support for
4933 using direct move instructions on ISA 2.07. Also adjust
4934 instruction length for 64-bit.
4935 (mov<mode>_64bit, TFmode/TDmode): Likewise.
4936 (mov<mode>_32bit, TFmode/TDmode): Likewise.
4938 2014-02-15 Alan Modra <amodra@gmail.com>
4942 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
4943 find_replacement on parts of insn rtl that might be reloaded.
4945 2014-02-15 Richard Biener <rguenther@suse.de>
4947 PR tree-optimization/60183
4948 * tree-ssa-phiprop.c (propagate_with_phi): Avoid speculating loads.
4949 (tree_ssa_phiprop): Calculate and free post-dominators.
4951 2014-02-14 Jeff Law <law@redhat.com>
4953 PR rtl-optimization/60131
4954 * ree.c (get_extended_src_reg): New function.
4955 (combine_reaching_defs): Use it rather than assuming location of REG.
4956 (find_and_remove_re): Verify first operand of extension is
4957 a REG before adding the insns to the copy list.
4959 2014-02-14 Roland McGrath <mcgrathr@google.com>
4961 * configure.ac (HAVE_AS_IX86_UD2): New test for 'ud2' mnemonic.
4962 * configure: Regenerated.
4963 * config.in: Regenerated.
4964 * config/i386/i386.md (trap) [HAVE_AS_IX86_UD2]: Use the mnemonic
4965 instead of ASM_SHORT.
4967 2014-02-14 Vladimir Makarov <vmakarov@redhat.com>
4968 Richard Earnshaw <rearnsha@arm.com>
4970 PR rtl-optimization/59535
4971 * lra-constraints.c (process_alt_operands): Encourage alternative
4972 when unassigned pseudo class is superset of the alternative class.
4973 (inherit_reload_reg): Don't inherit when optimizing for code size.
4974 * config/arm/arm.h (MODE_BASE_REG_CLASS): Add version for LRA
4975 returning CORE_REGS for anything but Thumb1 and BASE_REGS for
4976 modes not less than 4 for Thumb1.
4978 2014-02-14 Kyle McMartin <kyle@redhat.com>
4981 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Define for AArch64.
4983 2014-02-14 Richard Biener <rguenther@suse.de>
4985 * cilk-common.c (cilk_arrow): Build a MEM_REF, not an INDIRECT_REF.
4986 (get_frame_arg): Drop the assert with langhook types_compatible_p.
4987 Do not strip INDIRECT_REFs.
4989 2014-02-14 Richard Biener <rguenther@suse.de>
4992 * lto-streamer-out.c (DFS_write_tree_body): Do not follow
4993 DECL_FUNCTION_SPECIFIC_TARGET.
4994 (hash_tree): Do not hash DECL_FUNCTION_SPECIFIC_TARGET.
4995 * tree-streamer-out.c (pack_ts_target_option): Remove.
4996 (streamer_pack_tree_bitfields): Do not stream TS_TARGET_OPTION.
4997 (write_ts_function_decl_tree_pointers): Do not stream
4998 DECL_FUNCTION_SPECIFIC_TARGET.
4999 * tree-streamer-in.c (unpack_ts_target_option): Remove.
5000 (unpack_value_fields): Do not stream TS_TARGET_OPTION.
5001 (lto_input_ts_function_decl_tree_pointers): Do not stream
5002 DECL_FUNCTION_SPECIFIC_TARGET.
5004 2014-02-14 Jakub Jelinek <jakub@redhat.com>
5006 * tree-vect-loop.c (vect_is_slp_reduction): Don't set use_stmt twice.
5007 (get_initial_def_for_induction, vectorizable_induction): Ignore
5008 debug stmts when looking for exit_phi.
5009 (vectorizable_live_operation): Fix up condition.
5011 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
5013 * config/nds32/nds32.c (nds32_asm_function_prologue): Do not use
5014 nreverse() because it changes the content of original tree list.
5016 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
5018 * config/nds32/t-mlibs (MULTILIB_OPTIONS): Fix typo in comment.
5019 * config/nds32/nds32.c (nds32_merge_decl_attributes): Likewise.
5021 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
5023 * config/nds32/nds32.c (nds32_naked_function_p): Follow the
5024 GNU coding standards.
5026 2014-02-13 Jakub Jelinek <jakub@redhat.com>
5029 * dwarf2out.c (gen_subprogram_die): Don't call
5030 add_calling_convention_attribute if subr_die is old_die.
5032 2014-02-13 Sharad Singhai <singhai@google.com>
5034 * doc/optinfo.texi: Fix order of nodes.
5036 2014-02-13 Uros Bizjak <ubizjak@gmail.com>
5038 * config/i386/sse.md (xop_vmfrcz<mode>2): Generate const0 in
5039 operands[2], not operands[3].
5041 2014-02-13 Richard Biener <rguenther@suse.de>
5044 * doc/install.texi (ISL): Update recommended version to 0.12.2,
5045 mention the possibility of an in-tree build.
5046 (CLooG): Update recommended version to 0.18.1, mention the
5047 possibility of an in-tree build and clarify that the ISL
5048 bundled with CLooG does not work.
5050 2014-02-13 Jakub Jelinek <jakub@redhat.com>
5053 * expr.c (compress_float_constant): If x is a hard register,
5054 extend into a pseudo and then move to x.
5056 2014-02-13 Dominik Vogt <vogt@linux.vnet.ibm.com>
5058 * config/s390/s390.c (s390_asm_output_function_label): Fix crash
5059 caused by bad second argument to warning_at() with -mhotpatch and
5060 nested functions (e.g. with gfortran).
5062 2014-02-13 Richard Sandiford <rdsandiford@googlemail.com>
5064 * opts.c (option_name): Remove "enabled by default" rider.
5066 2014-02-12 John David Anglin <danglin@gcc.gnu.org>
5068 * config/pa/pa.c (pa_option_override): Remove auto increment FIXME.
5070 2014-02-12 H.J. Lu <hongjiu.lu@intel.com>
5071 Uros Bizjak <ubizjak@gmail.com>
5074 * configure.ac (HAVE_AS_GOTOFF_IN_DATA): Pass --32 to GNU assembler.
5075 * configure: Regenerated.
5077 2014-02-12 Richard Biener <rguenther@suse.de>
5079 * vec.c (vec_prefix::calculate_allocation): Move as
5080 inline variant to vec.h.
5081 (vec_prefix::calculate_allocation_1): New out-of-line version.
5082 * vec.h (vec_prefix::calculate_allocation_1): Declare.
5083 (vec_prefix::m_has_auto_buf): Rename to ...
5084 (vec_prefix::m_using_auto_storage): ... this.
5085 (vec_prefix::calculate_allocation): Inline the easy cases
5086 and dispatch to calculate_allocation_1 which doesn't need the
5088 (va_heap::reserve): Use gcc_checking_assert.
5089 (vec<T, A, vl_embed>::embedded_init): Add argument to initialize
5090 m_using_auto_storage.
5091 (auto_vec): Change m_vecpfx member to a vec<T, va_heap, vl_embed>
5093 (vec<T, va_heap, vl_ptr>::reserve): Remove redundant check.
5094 (vec<T, va_heap, vl_ptr>::release): Avoid casting.
5095 (vec<T, va_heap, vl_ptr>::using_auto_storage): Simplify.
5097 2014-02-12 Richard Biener <rguenther@suse.de>
5099 * gcse.c (compute_transp): break from loop over canon_modify_mem_list
5100 when we found a dependence.
5102 2014-02-12 Thomas Schwinge <thomas@codesourcery.com>
5104 * gimplify.c (gimplify_call_expr, gimplify_modify_expr): Move
5106 (maybe_fold_stmt): ... into this new function.
5107 * omp-low.c (lower_omp): Update comment.
5109 * omp-low.c (lower_omp_target): Add clobber for sizes array, after
5112 * omp-low.c (diagnose_sb_0): Make sure label_ctx is valid to
5115 2014-02-12 James Greenhalgh <james.greenhalgh@arm.com>
5117 * config/arm/aarch-cost-tables.h (generic_extra_costs): Fix
5118 identifiers in comments.
5119 (cortexa53_extra_costs): Likewise.
5120 * config/arm/arm.c (cortexa9_extra_costs): Fix identifiers in comments.
5121 (cortexa7_extra_costs): Likewise.
5122 (cortexa12_extra_costs): Likewise.
5123 (cortexa15_extra_costs): Likewise.
5124 (v7m_extra_costs): Likewise.
5126 2014-02-12 Richard Biener <rguenther@suse.de>
5129 * gimple-low.c (lower_builtin_posix_memalign): Lower conditional
5130 of posix_memalign being successful.
5131 (lower_stmt): Restrict lowering of posix_memalign to when
5132 -ftree-bit-ccp is enabled.
5134 2014-02-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
5136 * config/avr/avr-c.c (avr_resolve_overloaded_builtin): Pass vNULL for
5138 * config/spu/spu-c.c (spu_resolve_overloaded_builtin): Likewise.
5140 2014-02-12 Eric Botcazou <ebotcazou@adacore.com>
5142 PR rtl-optimization/60116
5143 * combine.c (try_combine): Also remove dangling REG_DEAD notes on the
5144 other_insn once the combination has been validated.
5146 2014-02-11 Jan Hubicka <hubicka@ucw.cz>
5149 * ipa-utils.h (possible_polymorphic_call_targets): Update prototype
5151 * ipa-devirt.c: Include demangle.h
5152 (odr_violation_reported): New static variable.
5153 (add_type_duplicate): Update odr_violations.
5154 (maybe_record_node): Add completep parameter; update it.
5155 (record_target_from_binfo): Add COMPLETEP parameter;
5156 update it as needed.
5157 (possible_polymorphic_call_targets_1): Likewise.
5158 (struct polymorphic_call_target_d): Add nonconstruction_targets;
5159 rename FINAL to COMPLETE.
5160 (record_targets_from_bases): Sanity check we found the binfo;
5161 fix COMPLETEP updating.
5162 (possible_polymorphic_call_targets): Add NONCONSTRUTION_TARGETSP
5163 parameter, fix computing of COMPLETEP.
5164 (dump_possible_polymorphic_call_targets): Imrove readability of dump;
5165 at LTO time do demangling.
5166 (ipa_devirt): Use nonconstruction_targets; Improve dumps.
5167 * gimple-fold.c (gimple_get_virt_method_for_vtable): Add can_refer
5169 (gimple_get_virt_method_for_binfo): Likewise.
5170 * gimple-fold.h (gimple_get_virt_method_for_binfo,
5171 gimple_get_virt_method_for_vtable): Update prototypes.
5173 2014-02-11 Vladimir Makarov <vmakarov@redhat.com>
5176 * genautomata.c (add_presence_absence): Fix typo with
5177 {final_}presence_list.
5179 2014-02-11 Michael Meissner <meissner@linux.vnet.ibm.com>
5182 * config/rs6000/rs6000.md (128-bit GPR splitter): Add a splitter
5183 for VSX/Altivec vectors that land in GPR registers.
5185 2014-02-11 Richard Henderson <rth@redhat.com>
5186 Jakub Jelinek <jakub@redhat.com>
5189 * tree-sra.c (load_assign_lhs_subreplacements): Add VIEW_CONVERT_EXPR
5190 around drhs if type conversion to lacc->type is not useless.
5192 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5194 * config/aarch64/aarch64-cores.def (cortex-a57): Use cortexa57
5196 (cortex-a57.cortex-a53): Likewise.
5197 * config/aarch64/aarch64.c (cortexa57_tunings): New tuning struct.
5199 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5201 * config/arm/thumb2.md (*thumb2_movhi_insn): Add alternatives for
5204 2014-02-11 Renlin Li <Renlin.Li@arm.com>
5206 * doc/sourcebuild.texi: Document check_effective_target_arm_vfp3_ok and
5207 add_options_for_arm_vfp3.
5209 2014-02-11 Jeff Law <law@redhat.com>
5212 * expr.c (expand_expr_addr_expr_1): Handle expand_expr returning an
5213 object with an undesirable mode.
5215 2014-02-11 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5218 * config/i386/sol2-9.h: New file.
5219 * config.gcc (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*,
5220 *-*-solaris2.9*): Use it.
5222 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
5224 * config/microblaze/microblaze.md: Add movsi4_rev insn pattern.
5225 * config/microblaze/predicates.md: Add reg_or_mem_operand predicate.
5227 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
5229 * config/microblaze/microblaze.c: Extend mcpu version format
5231 2014-02-10 David Holsgrove <david.holsgrove@xilinx.com>
5233 * config/microblaze/microblaze.h: Define SIZE_TYPE and PTRDIFF_TYPE.
5235 2014-02-10 Richard Henderson <rth@redhat.com>
5238 * calls.c (expand_call): Don't double-push for reg_parm_stack_space.
5239 * config/i386/i386.c (init_cumulative_args): Remove sorry for 64-bit
5240 ms-abi vs -mno-accumulate-outgoing-args.
5241 (ix86_expand_prologue): Unconditionally call ix86_eax_live_at_start_p.
5242 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Fix comment with
5245 2014-02-10 Bernd Edlinger <bernd.edlinger@hotmail.de>
5248 * cfgexpand.c (expand_asm_operands): Attach source location to
5249 ASM_INPUT rtx objects.
5250 * print-rtl.c (print_rtx): Check for UNKNOWN_LOCATION.
5252 2014-02-10 Nick Clifton <nickc@redhat.com>
5254 * config/mn10300/mn10300.c (popcount): New function.
5255 (mn10300_expand_prologue): Include saved registers in stack usage
5258 2014-02-10 Jeff Law <law@redhat.com>
5261 * reload1.c (emit_input_reload_insns): Do not create invalid RTL
5262 when changing the SET_DEST of a prior insn to avoid an input reload.
5264 2014-02-10 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
5266 * config/rs6000/sysv4.h (ENDIAN_SELECT): Do not attempt to enforce
5267 big-endian mode for -mcall-aixdesc, -mcall-freebsd, -mcall-netbsd,
5268 -mcall-openbsd, or -mcall-linux.
5269 (CC1_ENDIAN_BIG_SPEC): Remove.
5270 (CC1_ENDIAN_LITTLE_SPEC): Remove.
5271 (CC1_ENDIAN_DEFAULT_SPEC): Remove.
5272 (CC1_SPEC): Remove (always empty) %cc1_endian_... spec.
5273 (SUBTARGET_EXTRA_SPECS): Remove %cc1_endian_big, %cc1_endian_little,
5274 and %cc1_endian_default.
5275 * config/rs6000/sysv4le.h (CC1_ENDIAN_DEFAULT_SPEC): Remove.
5277 2014-02-10 Richard Biener <rguenther@suse.de>
5279 PR tree-optimization/60115
5280 * tree-eh.c (tree_could_trap_p): Unify TARGET_MEM_REF and
5281 MEM_REF handling. Properly verify that the accesses are not
5282 out of the objects bound.
5284 2014-02-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5286 * config/aarch64/aarch64.c (aarch64_override_options): Fix typo from
5289 2014-02-10 Eric Botcazou <ebotcazou@adacore.com>
5291 * ipa-devirt.c (get_polymorphic_call_info_from_invariant): Return
5292 proper constants and fix formatting.
5293 (possible_polymorphic_call_targets): Fix formatting.
5295 2014-02-10 Kirill Yukhin <kirill.yukhin@intel.com>
5296 Ilya Tocar <ilya.tocar@intel.com>
5298 * config/i386/avx512fintrin.h (_mm512_storeu_epi64): Removed.
5299 (_mm512_loadu_epi32): Renamed into...
5300 (_mm512_loadu_si512): This.
5301 (_mm512_storeu_epi32): Renamed into...
5302 (_mm512_storeu_si512): This.
5303 (_mm512_maskz_ceil_ps): Removed.
5304 (_mm512_maskz_ceil_pd): Ditto.
5305 (_mm512_maskz_floor_ps): Ditto.
5306 (_mm512_maskz_floor_pd): Ditto.
5307 (_mm512_floor_round_ps): Ditto.
5308 (_mm512_floor_round_pd): Ditto.
5309 (_mm512_ceil_round_ps): Ditto.
5310 (_mm512_ceil_round_pd): Ditto.
5311 (_mm512_mask_floor_round_ps): Ditto.
5312 (_mm512_mask_floor_round_pd): Ditto.
5313 (_mm512_mask_ceil_round_ps): Ditto.
5314 (_mm512_mask_ceil_round_pd): Ditto.
5315 (_mm512_maskz_floor_round_ps): Ditto.
5316 (_mm512_maskz_floor_round_pd): Ditto.
5317 (_mm512_maskz_ceil_round_ps): Ditto.
5318 (_mm512_maskz_ceil_round_pd): Ditto.
5319 (_mm512_expand_pd): Ditto.
5320 (_mm512_expand_ps): Ditto.
5321 * config/i386/i386.c (ix86_builtins): Remove
5322 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK.
5323 (bdesc_args): Ditto.
5324 * config/i386/predicates.md (const1256_operand): New.
5325 (const_1_to_2_operand): Ditto.
5326 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
5327 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
5328 (*avx512pf_gatherpf<mode>sf): Ditto.
5329 (avx512pf_gatherpf<mode>df): Ditto.
5330 (*avx512pf_gatherpf<mode>df_mask): Ditto.
5331 (*avx512pf_gatherpf<mode>df): Ditto.
5332 (avx512pf_scatterpf<mode>sf): Ditto.
5333 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
5334 (*avx512pf_scatterpf<mode>sf): Ditto.
5335 (avx512pf_scatterpf<mode>df): Ditto.
5336 (*avx512pf_scatterpf<mode>df_mask): Ditto.
5337 (*avx512pf_scatterpf<mode>df): Ditto.
5338 (avx512f_expand<mode>): Removed.
5339 (<shift_insn><mode>3<mask_name>): Change predicate type.
5341 2014-02-08 Jakub Jelinek <jakub@redhat.com>
5343 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers
5344 not at the end of datarefs vector use ordered_remove to avoid
5345 reordering datarefs vector.
5348 * gimplify.c (gimplify_bind_expr): In ORT_SIMD region
5349 mark local addressable non-static vars as GOVD_PRIVATE
5350 instead of GOVD_LOCAL.
5351 * omp-low.c (lower_omp_for): Move gimple_bind_vars
5352 and BLOCK_VARS of gimple_bind_block to new_stmt rather
5356 * tree-ssa-ccp.c (surely_varying_stmt_p): Don't return true
5357 if TYPE_ATTRIBUTES (gimple_call_fntype ()) contain
5358 assume_aligned or alloc_align attributes.
5359 (bit_value_assume_aligned): Add ATTR, PTRVAL and ALLOC_ALIGN
5360 arguments. Handle also assume_aligned and alloc_align attributes.
5361 (evaluate_stmt): Adjust bit_value_assume_aligned caller. Handle
5362 calls to functions with assume_aligned or alloc_align attributes.
5363 * doc/extend.texi: Document assume_aligned and alloc_align attributes.
5365 2014-02-08 Terry Guo <terry.guo@arm.com>
5367 * doc/invoke.texi: Document ARM -march=armv7e-m.
5369 2014-02-08 Jakub Jelinek <jakub@redhat.com>
5371 * cilk-common.c (cilk_init_builtins): Clear TREE_NOTHROW
5372 flag on __cilkrts_rethrow builtin.
5375 * ipa-cp.c (determine_versionability): Fail at -O0
5376 or __attribute__((optimize (0))) or -fno-ipa-cp functions.
5377 * tree-sra.c (ipa_sra_preliminary_function_checks): Similarly.
5380 2014-02-04 Jakub Jelinek <jakub@redhat.com>
5383 * tree-inline.c (copy_forbidden): Fail for
5384 __attribute__((optimize (0))) functions.
5386 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
5388 * varpool.c: Include pointer-set.h.
5389 (varpool_remove_unreferenced_decls): Variables in other partitions
5390 will not be output; be however careful to not lose information
5393 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
5395 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do O(1)
5396 lookup in the vtable constructor.
5398 2014-02-07 Jeff Law <law@redhat.com>
5401 * config/m68k/m68k.md (ashldi_extsi): Turn into a
5402 define_insn_and_split.
5404 * ipa-inline.c (inline_small_functions): Fix typos.
5406 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5408 * config/s390/s390-protos.h (s390_can_use_simple_return_insn)
5409 (s390_can_use_return_insn): Declare.
5410 * config/s390/s390.h (EPILOGUE_USES): Define.
5411 * config/s390/s390.c (s390_mainpool_start): Allow two main_pool
5413 (s390_chunkify_start): Handle return JUMP_LABELs.
5414 (s390_early_mach): Emit a main_pool instruction on the entry edge.
5415 (s300_set_up_by_prologue, s390_can_use_simple_return_insn)
5416 (s390_can_use_return_insn): New functions.
5417 (s390_fix_long_loop_prediction): Handle conditional returns.
5418 (TARGET_SET_UP_BY_PROLOGUE): Define.
5419 * config/s390/s390.md (ANY_RETURN): New code iterator.
5420 (*creturn, *csimple_return, return, simple_return): New patterns.
5422 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5424 * config/s390/s390.c (s390_restore_gprs_from_fprs): Add REG_CFA_RESTORE
5425 notes to each restore. Also add REG_CFA_DEF_CFA when restoring %r15.
5426 (s390_optimize_prologue): Don't clear RTX_FRAME_RELATED_P. Update the
5427 REG_CFA_RESTORE list when deciding not to restore a register.
5429 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5431 * config/s390/s390.c: Include tree-pass.h and context.h.
5432 (s390_early_mach): New function, split out from...
5433 (s390_emit_prologue): ...here.
5434 (pass_data_s390_early_mach): New pass structure.
5435 (pass_s390_early_mach): New class.
5436 (s390_option_override): Create and register early_mach pass.
5437 Move to end of file.
5439 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5441 * var-tracking.c (vt_stack_adjustments): Don't require stack_adjusts
5442 to match for the exit block.
5444 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5446 * config/s390/s390.md ("atomic_load<mode>", "atomic_store<mode>")
5447 ("atomic_compare_and_swap<mode>", "atomic_fetch_<atomic><mode>"):
5448 Reject misaligned operands.
5450 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5452 * optabs.c (expand_atomic_compare_and_swap): Allow expander to fail.
5454 2014-02-07 Richard Biener <rguenther@suse.de>
5457 * gimple-low.c (lower_builtin_posix_memalign): New function.
5458 (lower_stmt): Call it to lower posix_memalign in a way
5459 to make alignment info accessible.
5461 2014-02-07 Jakub Jelinek <jakub@redhat.com>
5464 * tree.c (build_common_builtin_nodes): Set ECF_LEAF for
5465 __builtin_setjmp_receiver.
5467 2014-02-07 Richard Biener <rguenther@suse.de>
5470 * builtin-types.def (BT_FN_INT_PTRPTR_SIZE_SIZE): Add.
5471 * builtins.def (BUILT_IN_POSIX_MEMALIGN): Likewise.
5472 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
5473 Handle BUILT_IN_POSIX_MEMALIGN.
5474 (find_func_clobbers): Likewise.
5475 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
5476 (call_may_clobber_ref_p_1): Likewise.
5478 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5481 * ipa-devirt.c (record_target_from_binfo): Remove overactive
5484 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5487 * lto-cgraph.c (lto_output_node): Use
5488 symtab_get_symbol_partitioning_class.
5489 (lto_output_varpool_node): likewise.
5490 (symtab_get_symbol_partitioning_class): Move here from
5492 * cgraph.h (symbol_partitioning_class): Likewise.
5493 (symtab_get_symbol_partitioning_class): Declare.
5495 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5497 * ggc.h (ggc_internal_cleared_alloc): New macro.
5498 * vec.h (vec_safe_copy): Handle memory stats.
5499 * omp-low.c (simd_clone_struct_alloc): Use ggc_internal_cleared_alloc.
5500 * target-globals.c (save_target_globals): Likewise.
5502 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5505 * expr.c (emit_move_resolve_push): Export; be bit more selective
5506 on when to clear alias set.
5507 * expr.h (emit_move_resolve_push): Declare.
5508 * function.h (struct function): Add tail_call_marked.
5509 * tree-tailcall.c (optimize_tail_call): Set tail_call_marked.
5510 * config/i386/i386-protos.h (ix86_expand_push): Remove.
5511 * config/i386/i386.md (TImode move expander): De not call
5513 (FP push expanders): Preserve memory attributes.
5514 * config/i386/sse.md (push<mode>1): Remove.
5515 * config/i386/i386.c (ix86_expand_vector_move): Handle push operation.
5516 (ix86_expand_push): Remove.
5517 * config/i386/mmx.md (push<mode>1): Remove.
5519 2014-02-06 Jakub Jelinek <jakub@redhat.com>
5521 PR rtl-optimization/60030
5522 * internal-fn.c (ubsan_expand_si_overflow_mul_check): Surround
5523 lopart with paradoxical subreg before shifting it up by hprec.
5525 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5527 * config/arm/aarch-cost-tables.h (cortexa57_extra_costs): New table.
5528 Remove extra newline at end of file.
5529 * config/arm/arm.c (arm_cortex_a57_tune): New tuning struct.
5530 (arm_issue_rate): Handle cortexa57.
5531 * config/arm/arm-cores.def (cortex-a57): Use cortex_a57 tuning.
5532 (cortex-a57.cortex-a53): Likewise.
5534 2014-02-06 Jakub Jelinek <jakub@redhat.com>
5537 * config/arm/arm.c (emit_multi_reg_push): Add dwarf_regs_mask argument,
5538 don't record in REG_FRAME_RELATED_EXPR registers not set in that
5540 (arm_expand_prologue): Adjust all callers.
5541 (arm_unwind_emit_sequence): Allow saved, but not important for unwind
5542 info, registers also at the lowest numbered registers side. Use
5543 gcc_assert instead of abort, and SET_SRC/SET_DEST macros instead of
5547 * var-tracking.c (adjust_mems): Before adding a SET to
5548 amd->side_effects, adjust it's SET_SRC using simplify_replace_fn_rtx.
5550 2014-02-06 Alan Modra <amodra@gmail.com>
5553 * config/rs6000/rs6000.c (rs6000_secondary_memory_needed_mode): Only
5554 change SDmode to DDmode when lra_in_progress.
5556 2014-02-06 Jakub Jelinek <jakub@redhat.com>
5559 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers, call
5560 free_data_ref on the dr first, and before goto again also set dr
5561 to the next dr. For simd_lane_access, free old datarefs[i] before
5562 overwriting it. For get_vectype_for_scalar_type failure, don't
5563 free_data_ref if simd_lane_access.
5565 * Makefile.in (prefix.o, cppbuiltin.o): Depend on $(BASEVER).
5568 * tree.h (opts_for_fn): New inline function.
5569 (opt_for_fn): Define.
5570 * config/i386/i386.c (ix86_function_regparm): Use
5571 opt_for_fn (decl, optimize) instead of optimize.
5573 2014-02-06 Marcus Shawcroft <marcus.shawcroft@arm.com>
5575 * config/aarch64/aarch64.c (aarch64_classify_symbol): Fix logic
5576 for SYMBOL_REF in large memory model.
5578 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5580 * config/aarch64/aarch64-cores.def (cortex-a53): Specify CRC32
5582 (cortex-a57): Likewise.
5583 (cortex-a57.cortex-a53): Likewise.
5585 2014-02-06 Yury Gribov <y.gribov@samsung.com>
5586 Kugan Vivekanandarajah <kuganv@linaro.org>
5588 * config/arm/arm.c (arm_vector_alignment_reachable): Check
5590 * config/arm/arm.c (arm_builtin_support_vector_misalignment): Likewise.
5592 2014-02-06 Richard Biener <rguenther@suse.de>
5594 * tree-cfg.c (gimple_duplicate_sese_region): Fix ordering of
5595 set_loop_copy and initialize_original_copy_tables.
5597 2014-02-06 Alex Velenko <Alex.Velenko@arm.com>
5599 * config/aarch64/aarch64-simd.md
5600 (aarch64_ashr_simddi): Change QI to SI.
5602 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
5603 Jakub Jelinek <jakub@redhat.com>
5606 * ipa-inline-analysis.c (compute_bb_predicates): Ensure monotonicity
5609 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5611 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Change
5612 CODE_FOR_altivec_vpku[hw]um to
5613 CODE_FOR_altivec_vpku[hw]um_direct.
5614 * config/rs6000/altivec.md (vec_unpacks_hi_<VP_small_lc>): Change
5615 UNSPEC_VUNPACK_HI_SIGN to UNSPEC_VUNPACK_HI_SIGN_DIRECT.
5616 (vec_unpacks_lo_<VP_small_lc>): Change UNSPEC_VUNPACK_LO_SIGN to
5617 UNSPEC_VUNPACK_LO_SIGN_DIRECT.
5619 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5621 * config/rs6000/altivec.md (altivec_vsum2sws): Adjust code
5622 generation for -maltivec=be.
5623 (altivec_vsumsws): Simplify redundant test.
5625 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5627 * altivec.md (UNSPEC_VPACK_UNS_UNS_MOD_DIRECT): New unspec.
5628 (UNSPEC_VUNPACK_HI_SIGN_DIRECT): Likewise.
5629 (UNSPEC_VUNPACK_LO_SIGN_DIRECT): Likewise.
5630 (mulv8hi3): Use gen_altivec_vpkuwum_direct instead of
5631 gen_altivec_vpkuwum.
5632 (altivec_vpkpx): Test for VECTOR_ELT_ORDER_BIG instead of for
5634 (altivec_vpks<VI_char>ss): Likewise.
5635 (altivec_vpks<VI_char>us): Likewise.
5636 (altivec_vpku<VI_char>us): Likewise.
5637 (altivec_vpku<VI_char>um): Likewise.
5638 (altivec_vpku<VI_char>um_direct): New (copy of
5639 altivec_vpku<VI_char>um that still relies on BYTES_BIG_ENDIAN, for
5641 (altivec_vupkhs<VU_char>): Emit vupkls* instead of vupkhs* when
5642 target is little endian and -maltivec=be is not specified.
5643 (*altivec_vupkhs<VU_char>_direct): New (copy of
5644 altivec_vupkhs<VU_char> that always emits vupkhs*, for internal use).
5645 (altivec_vupkls<VU_char>): Emit vupkhs* instead of vupkls* when
5646 target is little endian and -maltivec=be is not specified.
5647 (*altivec_vupkls<VU_char>_direct): New (copy of
5648 altivec_vupkls<VU_char> that always emits vupkls*, for internal use).
5649 (altivec_vupkhpx): Emit vupklpx instead of vupkhpx when target is
5650 little endian and -maltivec=be is not specified.
5651 (altivec_vupklpx): Emit vupkhpx instead of vupklpx when target is
5652 little endian and -maltivec=be is not specified.
5654 2014-02-05 Richard Henderson <rth@redhat.com>
5657 * combine-stack-adj.c: Revert r206943.
5658 * sched-int.h (struct deps_desc): Add last_args_size.
5659 * sched-deps.c (init_deps): Initialize it.
5660 (sched_analyze_insn): Add OUTPUT dependencies between insns that
5661 contain REG_ARGS_SIZE notes.
5663 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
5665 * lto-cgraph.c (asm_nodes_output): Make global.
5666 * lto-wrapper.c (run_gcc): Pass down paralelizm to WPA.
5667 * gcc.c (AS_NEEDS_DASH_FOR_PIPED_INPUT): Allow WPA parameter
5668 (driver_handle_option): Handle OPT_fwpa.
5670 2014-02-05 Jakub Jelinek <jakub@redhat.com>
5673 * ipa-devirt.c (possible_polymorphic_call_targets): Fix
5674 a comment typo and formatting issue. If odr_hash hasn't been
5675 created, return vNULL and set *completep to false.
5678 * tree-eh.c (cleanup_empty_eh): Bail out on totally empty
5679 bb with no successors.
5681 2014-02-05 James Greenhalgh <james.greenhalgh@arm.com>
5684 * doc/invoke.texi (-march): Clarify documentation for ARM.
5688 2014-02-05 Richard Biener <rguenther@suse.de>
5690 * tree-vect-loop.c (vect_analyze_loop_2): Be more informative
5691 when not vectorizing because of too many alias checks.
5692 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
5693 Add more verboseness, avoid duplicate MSG_MISSED_OPTIMIZATION.
5695 2014-02-05 Nick Clifton <nickc@redhat.com>
5697 * config/mn10300/mn10300.c (mn10300_hard_regno_mode_ok): Do not
5698 accept extended registers in any mode when compiling for the MN10300.
5700 2014-02-05 Yury Gribov <y.gribov@samsung.com>
5702 * cif-code.def (ATTRIBUTE_MISMATCH): New CIF code.
5703 * ipa-inline.c (report_inline_failed_reason): Handle mismatched
5704 sanitization attributes.
5705 (can_inline_edge_p): Likewise.
5706 (sanitize_attrs_match_for_inline_p): New function.
5708 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
5710 * ipa-prop.c (detect_type_change): Shor circuit testing of
5711 type changes on THIS pointer.
5713 2014-02-04 John David Anglin <danglin@gcc.gnu.org>
5716 * config/pa/pa.c (legitimize_tls_address): Return original address
5717 if not passed a SYMBOL_REF rtx.
5718 (hppa_legitimize_address): Call legitimize_tls_address for all TLS
5720 (pa_emit_move_sequence): Simplify TLS source operands.
5721 (pa_legitimate_constant_p): Reject all TLS constants.
5722 * config/pa/pa.h (PA_SYMBOL_REF_TLS_P): Correct comment.
5723 (CONSTANT_ADDRESS_P): Reject TLS CONST addresses.
5725 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
5727 * ipa.c (function_and_variable_visibility): Decompose DECL_ONE_ONLY
5728 groups when we know they are controlled by LTO.
5729 * varasm.c (default_binds_local_p_1): If object is in other partition,
5730 it will be resolved locally.
5732 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
5734 * config/host-linux.c (linux_gt_pch_use_address): Don't
5735 use SSIZE_MAX because it is not always defined.
5737 2014-02-04 Vladimir Makarov <vmakarov@redhat.com>
5740 * lra-constraints.c (need_for_split_p): Use more 3 reloads as
5741 threshold for pseudo splitting.
5742 (update_ebb_live_info): Process call argument hard registers and
5743 hard registers from insn definition too.
5744 (max_small_class_regs_num): New constant.
5745 (inherit_in_ebb): Update live hard regs through EBBs. Update
5746 reloads_num only for small register classes. Don't split for
5749 2014-02-04 Markus Trippelsdorf <markus@trippelsdorf.de>
5752 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that target
5755 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
5757 * gimple-fold.c (can_refer_decl_in_current_unit_p): Default
5760 2014-02-04 Marek Polacek <polacek@redhat.com>
5762 * gdbinit.in (pel): Define.
5764 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
5766 * doc/invoke.texi (fstrict-volatile-bitfields): Clarify current
5769 2014-02-04 Richard Biener <rguenther@suse.de>
5772 * lto-streamer-out.c (tree_is_indexable): Force NAMELIST_DECLs
5773 in function context local.
5774 (lto_output_tree_ref): Do not write trees from lto_output_tree_ref.
5775 * lto-streamer-in.c (lto_input_tree_ref): Handle LTO_namelist_decl_ref
5776 similar to LTO_imported_decl_ref.
5778 2014-02-04 Jakub Jelinek <jakub@redhat.com>
5780 PR tree-optimization/60002
5781 * cgraphclones.c (build_function_decl_skip_args): Clear
5784 PR tree-optimization/60023
5785 * tree-if-conv.c (predicate_mem_writes): Pass true instead of
5786 false to gsi_replace.
5787 * tree-vect-stmts.c (vect_finish_stmt_generation): If stmt
5788 has been in some EH region and vec_stmt could throw, add
5789 vec_stmt into the same EH region.
5790 * tree-data-ref.c (get_references_in_stmt): If IFN_MASK_LOAD
5791 has no lhs, ignore it.
5792 * internal-fn.c (expand_MASK_LOAD): Likewise.
5795 * tree-inline.c (copy_forbidden): Fail for
5796 __attribute__((optimize (0))) functions.
5799 * omp-low.c (simd_clone_struct_copy): If from->inbranch
5800 is set, copy one less argument.
5801 (expand_simd_clones): Don't subtract clone_info->inbranch
5802 from simd_clone_struct_alloc argument.
5804 PR rtl-optimization/57915
5805 * recog.c (simplify_while_replacing): If all unary/binary/relational
5806 operation arguments are constant, attempt to simplify those.
5809 * expmed.c (expand_mult): For MODE_VECTOR_INT multiplication
5810 if there is no vashl<mode>3 or ashl<mode>3 insn, skip_synth.
5812 2014-02-04 Richard Biener <rguenther@suse.de>
5814 PR tree-optimization/60012
5815 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Apply
5816 TBAA disambiguation to all DDRs.
5818 2014-02-04 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5821 * config/sol2.h (LINK_LIBGCC_MAPFILE_SPEC): Define.
5822 (LINK_SPEC): Use it for -shared, -shared-libgcc.
5824 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
5827 * tree.c (get_binfo_at_offset): Do not get confused by empty classes;
5829 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
5831 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Remove.
5832 * gimple-fold.h (gimple_extract_devirt_binfo_from_cst): Remove.
5834 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
5837 * ipa-cp.c (ipa_get_indirect_edge_target_1): Use ipa-devirt
5838 to figure out targets of polymorphic calls with known decl.
5839 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
5840 * ipa-utils.h (get_polymorphic_call_info_from_invariant): Declare.
5841 * ipa-devirt.c (get_polymorphic_call_info_for_decl): Break out from ...
5842 (get_polymorphic_call_info): ... here.
5843 (get_polymorphic_call_info_from_invariant): New function.
5845 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
5847 * ipa-cp.c (ipa_get_indirect_edge_target_1): Do direct
5848 lookup via vtable pointer; check for type consistency
5849 and turn inconsitent facts into UNREACHABLE.
5850 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
5851 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do not ICE on
5852 type inconsistent querries; return UNREACHABLE instead.
5854 2014-02-03 Richard Henderson <rth@twiddle.net>
5857 * tree-ssa-uninit.c (push_to_worklist): Don't re-push if we've
5858 already processed this node.
5859 (normalize_one_pred_1): Pass along mark_set.
5860 (normalize_one_pred): Create and destroy a pointer_set_t.
5861 (normalize_one_pred_chain): Likewise.
5863 2014-02-03 Laurent Aflonsi <laurent.alfonsi@st.com>
5865 PR gcov-profile/58602
5866 * gcc/gcov-io.c (gcov_open): Open with truncation when mode < 0.
5868 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
5871 * ipa-cp.c (ipa_get_indirect_edge_target_1): Give up on
5872 -fno-devirtualize; try to devirtualize by the knowledge of
5873 virtual table pointer given by aggregate propagation.
5874 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
5875 (ipa_print_node_jump_functions): Dump also offset that
5876 is relevant for polymorphic calls.
5877 (determine_known_aggregate_parts): Add arg_type parameter; use it
5878 instead of determining the type from pointer type.
5879 (ipa_compute_jump_functions_for_edge): Update call of
5880 determine_known_aggregate_parts.
5881 * gimple-fold.c (gimple_get_virt_method_for_vtable): Break out from ...
5882 (gimple_get_virt_method_for_binfo): ... here; simplify using
5883 vtable_pointer_value_to_vtable.
5884 * gimple-fold.h (gimple_get_virt_method_for_vtable): Declare.
5885 * ipa-devirt.c (subbinfo_with_vtable_at_offset): Turn OFFSET parameter
5886 to unsigned HOST_WIDE_INT; use vtable_pointer_value_to_vtable.
5887 (vtable_pointer_value_to_vtable): Break out from ...; handle also
5889 (vtable_pointer_value_to_binfo): ... here.
5890 * ipa-utils.h (vtable_pointer_value_to_vtable): Declare.
5892 2014-02-03 Teresa Johnson <tejohnson@google.com>
5894 * tree-vect-slp.c (vect_supported_load_permutation_p): Avoid
5895 redef of outer loop index variable.
5897 2014-02-03 Marc Glisse <marc.glisse@inria.fr>
5901 * doc/extend.texi (Function Attributes): Typo.
5903 2014-02-03 Cong Hou <congh@google.com>
5905 PR tree-optimization/60000
5906 * tree-vect-loop.c (vect_transform_loop): Set pattern_def_seq to NULL
5907 if the vectorized statement is a store. A store statement can only
5908 appear at the end of pattern statements.
5910 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
5912 * config/i386/i386.c (flag_opts): Add -mlong-double-128.
5913 (ix86_option_override_internal): Default long double to 64-bit for
5914 32-bit Bionic and to 128-bit for 64-bit Bionic.
5916 * config/i386/i386.h (LONG_DOUBLE_TYPE_SIZE): Use 128 if
5917 TARGET_LONG_DOUBLE_128 is true.
5918 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Likewise.
5920 * config/i386/i386.opt (mlong-double-80): Negate -mlong-double-64.
5921 (mlong-double-64): Negate -mlong-double-128.
5922 (mlong-double-128): New option.
5924 * config/i386/i386-c.c (ix86_target_macros): Define
5925 __LONG_DOUBLE_128__ for TARGET_LONG_DOUBLE_128.
5927 * doc/invoke.texi: Document -mlong-double-128.
5929 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
5931 PR rtl-optimization/60024
5932 * sel-sched.c (init_regs_for_mode): Check if mode is OK first.
5934 2014-02-03 Markus Trippelsdorf <markus@trippelsdorf.de>
5936 * doc/invoke.texi (fprofile-reorder-functions): Fix typo.
5938 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
5940 PR rtl-optimization/57662
5941 * sel-sched.c (code_motion_path_driver): Do not mark already not
5942 existing blocks in the visiting bitmap.
5944 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
5946 * sel-sched-ir.c (sel_gen_insn_from_expr_after): Reset INSN_DELETED_P
5947 on the insn being emitted.
5949 2014-02-03 James Greenhalgh <james.greenhalgh@arm.com>
5950 Will Deacon <will.deacon@arm.com>
5952 * doc/gimple.texi (gimple_asm_clear_volatile): Remove.
5954 2014-02-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5956 * config/arm/arm-tables.opt: Regenerate.
5958 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5960 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): Generalize
5961 for vector types other than V16QImode.
5962 * config/rs6000/altivec.md (altivec_vperm_<mode>): Change to a
5963 define_expand, and call altivec_expand_vec_perm_le when producing
5964 code with little endian element order.
5965 (*altivec_vperm_<mode>_internal): New insn having previous
5966 behavior of altivec_vperm_<mode>.
5967 (altivec_vperm_<mode>_uns): Change to a define_expand, and call
5968 altivec_expand_vec_perm_le when producing code with little endian
5970 (*altivec_vperm_<mode>_uns_internal): New insn having previous
5971 behavior of altivec_vperm_<mode>_uns.
5973 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5975 * config/rs6000/altivec.md (UNSPEC_VSUMSWS_DIRECT): New unspec.
5976 (altivec_vsumsws): Add handling for -maltivec=be with a little
5978 (altivec_vsumsws_direct): New.
5979 (reduc_splus_<mode>): Call gen_altivec_vsumsws_direct instead of
5980 gen_altivec_vsumsws.
5982 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
5984 * ipa-devirt.c (subbinfo_with_vtable_at_offset,
5985 vtable_pointer_value_to_binfo): New functions.
5986 * ipa-utils.h (vtable_pointer_value_to_binfo): Declare.
5987 * ipa-prop.c (extr_type_from_vtbl_ptr_store): Use it.
5989 2014-02-02 Sandra Loosemore <sandra@codesourcery.com>
5991 * config/nios2/nios2.md (load_got_register): Initialize GOT
5992 pointer from _gp_got instead of _GLOBAL_OFFSET_TABLE_.
5993 * config/nios2/nios2.c (nios2_function_profiler): Likewise.
5995 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
5997 * ipa-prop.c (update_jump_functions_after_inlining): When type is not
5998 preserverd by passthrough, do not propagate the type.
6000 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
6002 * config/mips/mips.c (MIPS_GET_FCSR, MIPS_SET_FCSR): New macros.
6003 (mips_atomic_assign_expand_fenv): New function.
6004 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): Define.
6006 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
6008 * doc/extend.texi (__builtin_mips_get_fcsr): Document.
6009 (__builtin_mips_set_fcsr): Likewise.
6010 * config/mips/mips-ftypes.def: Add MIPS_VOID_FTYPE_USI and
6011 MIPS_USI_FTYPE_VOID.
6012 * config/mips/mips-protos.h (mips16_expand_get_fcsr): Declare
6013 (mips16_expand_set_fcsr): Likewise.
6014 * config/mips/mips.c (mips16_get_fcsr_stub): New variable.
6015 (mips16_set_fcsr_stub): Likewise.
6016 (mips16_get_fcsr_one_only_stub): New class.
6017 (mips16_set_fcsr_one_only_stub): Likewise.
6018 (mips16_expand_get_fcsr, mips16_expand_set_fcsr): New functions.
6019 (mips_code_end): Output the get_fcsr and set_fcsr stubs, if needed.
6020 (BUILTIN_AVAIL_MIPS16, AVAIL_ALL): New macros.
6021 (hard_float): New availability predicate.
6022 (mips_builtins): Add get_fcsr and set_fcsr.
6023 (mips_expand_builtin): Check BUILTIN_AVAIL_MIPS16.
6024 * config/mips/mips.md (UNSPEC_GET_FCSR, UNSPEC_SET_FCSR): New unspecs.
6025 (GET_FCSR_REGNUM, SET_FCSR_REGNUM): New constants.
6026 (mips_get_fcsr, *mips_get_fcsr, mips_get_fcsr_mips16_<mode>)
6027 (mips_set_fcsr, *mips_set_fcsr, mips_set_fcsr_mips16_<mode>): New
6030 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
6032 * config/mips/mips.c (mips_one_only_stub): New class.
6033 (mips_need_mips16_rdhwr_p): Replace with...
6034 (mips16_rdhwr_stub): ...this new variable.
6035 (mips16_stub_call_address): New function.
6036 (mips16_rdhwr_one_only_stub): New class.
6037 (mips_expand_thread_pointer): Use mips16_stub_call_address.
6038 (mips_output_mips16_rdhwr): Delete.
6039 (mips_finish_stub): New function.
6040 (mips_code_end): Use it to handle rdhwr stubs.
6042 2014-02-02 Uros Bizjak <ubizjak@gmail.com>
6045 * config/i386/i386.c (classify_argument): Fix handling of bit_offset
6046 when calculating size of integer atomic types.
6048 2014-02-02 H.J. Lu <hongjiu.lu@intel.com>
6050 * ipa-inline-analysis.c (true_predicate_p): Fix a typo in comments.
6052 2014-02-01 Jakub Jelinek <jakub@redhat.com>
6054 PR tree-optimization/60003
6055 * gimple-low.c (lower_builtin_setjmp): Set cfun->has_nonlocal_label.
6056 * profile.c (branch_prob): Use gimple_call_builtin_p
6057 to check for BUILT_IN_SETJMP_RECEIVER.
6058 * tree-inline.c (copy_bb): Call notice_special_calls.
6060 2014-01-31 Vladimir Makarov <vmakarov@redhat.com>
6063 * lra-constraints.c (process_alt_operands): Update reload_sum only
6066 2014-01-31 Richard Henderson <rth@redhat.com>
6069 * tree-eh.c (lower_try_finally_switch): Delay lowering finally block
6070 until after else_eh is processed.
6072 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
6074 * config/i386/avx512fintrin.h (_MM_FROUND_TO_NEAREST_INT),
6075 (_MM_FROUND_TO_NEG_INF), (_MM_FROUND_TO_POS_INF),
6076 (_MM_FROUND_TO_ZERO), (_MM_FROUND_CUR_DIRECTION): Are already defined
6077 in smmintrin.h, remove them.
6078 (_MM_FROUND_NO_EXC): Same as above, bit also wrong value.
6079 * config/i386/i386.c (ix86_print_operand): Split sae and rounding.
6080 * config/i386/i386.md (ROUND_SAE): Fix value.
6081 * config/i386/predicates.md (const_4_or_8_to_11_operand): New.
6082 (const48_operand): New.
6083 * config/i386/subst.md (round), (round_expand): Use
6084 const_4_or_8_to_11_operand.
6085 (round_saeonly), (round_saeonly_expand): Use const48_operand.
6087 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
6089 * config/i386/constraints.md (Yk): Swap meaning with k.
6090 * config/i386/i386.md (movhi_internal): Change Yk to k.
6091 (movqi_internal): Ditto.
6092 (*k<logic><mode>): Ditto.
6095 (kandn<mode>): Ditto.
6096 (*<code>hi_1): Ditto.
6097 (*<code>qi_1): Ditto.
6098 (kxnor<mode>): Ditto.
6099 (kortestzhi): Ditto.
6100 (kortestchi): Ditto.
6102 (*one_cmplhi2_1): Ditto.
6103 (*one_cmplqi2_1): Ditto.
6104 * config/i386/sse.md (): Change k to Yk.
6105 (avx512f_load<mode>_mask): Ditto.
6106 (avx512f_blendm<mode>): Ditto.
6107 (avx512f_store<mode>_mask): Ditto.
6108 (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
6109 (avx512f_storedqu<mode>_mask): Ditto.
6110 (avx512f_cmp<mode>3<mask_scalar_merge_name><round_saeonly_name>):
6112 (avx512f_ucmp<mode>3<mask_scalar_merge_name>): Ditto.
6113 (avx512f_vmcmp<mode>3<round_saeonly_name>): Ditto.
6114 (avx512f_vmcmp<mode>3_mask<round_saeonly_name>): Ditto.
6115 (avx512f_maskcmp<mode>3): Ditto.
6116 (avx512f_fmadd_<mode>_mask<round_name>): Ditto.
6117 (avx512f_fmadd_<mode>_mask3<round_name>): Ditto.
6118 (avx512f_fmsub_<mode>_mask<round_name>): Ditto.
6119 (avx512f_fmsub_<mode>_mask3<round_name>): Ditto.
6120 (avx512f_fnmadd_<mode>_mask<round_name>): Ditto.
6121 (avx512f_fnmadd_<mode>_mask3<round_name>): Ditto.
6122 (avx512f_fnmsub_<mode>_mask<round_name>): Ditto.
6123 (avx512f_fnmsub_<mode>_mask3<round_name>): Ditto.
6124 (avx512f_fmaddsub_<mode>_mask<round_name>): Ditto.
6125 (avx512f_fmaddsub_<mode>_mask3<round_name>): Ditto.
6126 (avx512f_fmsubadd_<mode>_mask<round_name>): Ditto.
6127 (avx512f_fmsubadd_<mode>_mask3<round_name>): Ditto.
6128 (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
6129 (vec_extract_lo_<mode>_maskm): Ditto.
6130 (vec_extract_hi_<mode>_maskm): Ditto.
6131 (avx512f_vternlog<mode>_mask): Ditto.
6132 (avx512f_fixupimm<mode>_mask<round_saeonly_name>): Ditto.
6133 (avx512f_sfixupimm<mode>_mask<round_saeonly_name>): Ditto.
6134 (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
6135 (avx512f_<code>v8div16qi2_mask): Ditto.
6136 (avx512f_<code>v8div16qi2_mask_store): Ditto.
6137 (avx512f_eq<mode>3<mask_scalar_merge_name>_1): Ditto.
6138 (avx512f_gt<mode>3<mask_scalar_merge_name>): Ditto.
6139 (avx512f_testm<mode>3<mask_scalar_merge_name>): Ditto.
6140 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Ditto.
6141 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
6142 (*avx512pf_gatherpf<mode>df_mask): Ditto.
6143 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
6144 (*avx512pf_scatterpf<mode>df_mask): Ditto.
6145 (avx512cd_maskb_vec_dupv8di): Ditto.
6146 (avx512cd_maskw_vec_dupv16si): Ditto.
6147 (avx512f_vpermi2var<mode>3_maskz): Ditto.
6148 (avx512f_vpermi2var<mode>3_mask): Ditto.
6149 (avx512f_vpermi2var<mode>3_mask): Ditto.
6150 (avx512f_vpermt2var<mode>3_maskz): Ditto.
6151 (*avx512f_gathersi<mode>): Ditto.
6152 (*avx512f_gathersi<mode>_2): Ditto.
6153 (*avx512f_gatherdi<mode>): Ditto.
6154 (*avx512f_gatherdi<mode>_2): Ditto.
6155 (*avx512f_scattersi<mode>): Ditto.
6156 (*avx512f_scatterdi<mode>): Ditto.
6157 (avx512f_compress<mode>_mask): Ditto.
6158 (avx512f_compressstore<mode>_mask): Ditto.
6159 (avx512f_expand<mode>_mask): Ditto.
6160 * config/i386/subst.md (mask): Change k to Yk.
6161 (mask_scalar_merge): Ditto.
6164 2014-01-31 Marc Glisse <marc.glisse@inria.fr>
6166 * doc/extend.texi (Vector Extensions): Document ?: in C++.
6168 2014-01-31 Richard Biener <rguenther@suse.de>
6171 * builtins.c (fold_builtin_memory_op): Make sure to not
6172 use a floating-point mode or a boolean or enumeral type for
6175 2014-01-30 DJ Delorie <dj@redhat.com>
6177 * config/msp430/msp430.h (LIB_SPEC): Add -lcrt
6178 * config/msp430/msp430.md (msp430_refsym_need_exit): New.
6179 * config/msp430/msp430.c (msp430_expand_epilogue): Call it
6180 whenever main() has an epilogue.
6182 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6184 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Remove
6185 unused variable "field".
6186 * config/rs6000/vsx.md (vsx_mergel_<mode>): Add missing DONE.
6187 (vsx_mergeh_<mode>): Likewise.
6188 * config/rs6000/altivec.md (altivec_vmrghb): Likewise.
6189 (altivec_vmrghh): Likewise.
6190 (altivec_vmrghw): Likewise.
6191 (altivec_vmrglb): Likewise.
6192 (altivec_vmrglh): Likewise.
6193 (altivec_vmrglw): Likewise.
6194 (altivec_vspltb): Add missing uses.
6195 (altivec_vsplth): Likewise.
6196 (altivec_vspltw): Likewise.
6197 (altivec_vspltsf): Likewise.
6199 2014-01-30 Jakub Jelinek <jakub@redhat.com>
6202 * ifcvt.c (cond_exec_process_insns): Don't conditionalize
6203 frame related instructions.
6205 2014-01-30 Vladimir Makarov <vmakarov@redhat.com>
6207 PR rtl-optimization/59959
6208 * lra-constrains.c (simplify_operand_subreg): Assign NO_REGS to
6209 any reload of register whose subreg is invalid.
6211 2014-01-30 Jakub Jelinek <jakub@redhat.com>
6213 * config/i386/f16cintrin.h (_cvtsh_ss): Avoid -Wnarrowing warning.
6214 * config/i386/avx512fintrin.h (_mm512_mask_cvtusepi64_storeu_epi32):
6215 Add missing return type - void.
6217 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6219 * gcc/config/rs6000/rs6000.c (rs6000_expand_vector_init): Use
6220 gen_vsx_xxspltw_v4sf_direct instead of gen_vsx_xxspltw_v4sf;
6221 remove element index adjustment for endian (now handled in vsx.md
6223 (altivec_expand_vec_perm_const): Use
6224 gen_altivec_vsplt[bhw]_direct instead of gen_altivec_vsplt[bhw].
6225 * gcc/config/rs6000/vsx.md (UNSPEC_VSX_XXSPLTW): New unspec.
6226 (vsx_xxspltw_<mode>): Adjust element index for little endian.
6227 * gcc/config/rs6000/altivec.md (altivec_vspltb): Divide into a
6228 define_expand and a new define_insn *altivec_vspltb_internal;
6229 adjust for -maltivec=be on a little endian target.
6230 (altivec_vspltb_direct): New.
6231 (altivec_vsplth): Divide into a define_expand and a new
6232 define_insn *altivec_vsplth_internal; adjust for -maltivec=be on a
6233 little endian target.
6234 (altivec_vsplth_direct): New.
6235 (altivec_vspltw): Divide into a define_expand and a new
6236 define_insn *altivec_vspltw_internal; adjust for -maltivec=be on a
6237 little endian target.
6238 (altivec_vspltw_direct): New.
6239 (altivec_vspltsf): Divide into a define_expand and a new
6240 define_insn *altivec_vspltsf_internal; adjust for -maltivec=be on
6241 a little endian target.
6243 2014-01-30 Richard Biener <rguenther@suse.de>
6245 PR tree-optimization/59993
6246 * tree-ssa-forwprop.c (associate_pointerplus): Check we
6247 can propagate form the earlier stmt and avoid the transform
6248 when the intermediate result is needed.
6250 2014-01-30 Alangi Derick <alangiderick@gmail.com>
6252 * README.Portability: Fix typo.
6254 2014-01-30 David Holsgrove <david.holsgrove@xilinx.com>
6256 * config/microblaze/microblaze.md(cstoresf4, cbranchsf4): Replace
6257 comparison_operator with ordered_comparison_operator.
6259 2014-01-30 Nick Clifton <nickc@redhat.com>
6261 * config/mn10300/mn10300-protos.h (mn10300_store_multiple_operation_p):
6262 Rename to mn10300_store_multiple_regs.
6263 * config/mn10300/mn10300.c: Likewise.
6264 * config/mn10300/mn10300.md (store_movm): Fix typo: call
6265 store_multiple_regs.
6266 * config/mn10300/predicates.md (mn10300_store_multiple_operation):
6267 Call mn10300_store_multiple_regs.
6269 2014-01-30 Nick Clifton <nickc@redhat.com>
6270 DJ Delorie <dj@redhat.com>
6272 * config/rl78/rl78.c (register_sizes): Make the "upper half" of
6273 %fp 2 to keep registers after it properly word-aligned.
6274 (rl78_alloc_physical_registers_umul): Handle the case where both
6275 input operands are the same.
6277 2014-01-30 Richard Biener <rguenther@suse.de>
6279 PR tree-optimization/59903
6280 * tree-vect-loop.c (vect_transform_loop): Guard multiple-types
6283 2014-01-30 Jason Merrill <jason@redhat.com>
6286 * tree.c (walk_type_fields): Handle VECTOR_TYPE.
6289 * cgraphunit.c (expand_thunk): Copy volatile arg to a temporary.
6291 2014-01-30 Richard Biener <rguenther@suse.de>
6293 PR tree-optimization/59951
6294 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Skip uses in debug insns.
6296 2014-01-30 Savin Zlobec <savin.zlobec@gmail.com>
6299 * config/nios2/nios2.c (nios2_fpu_insn_asm): Fix asm output of
6300 SFmode to DFmode case.
6302 2014-01-29 DJ Delorie <dj@redhat.com>
6304 * config/msp430/msp430.opt (-minrt): New.
6305 * config/msp430/msp430.h (STARTFILE_SPEC): Link alternate runtime
6307 (ENDFILE_SPEC): Likewise.
6309 2014-01-29 Jan Hubicka <hubicka@ucw.cz>
6311 * ipa-inline-analysis.c (clobber_only_eh_bb_p): New function.
6312 (estimate_function_body_sizes): Use it.
6314 2014-01-29 Paolo Carlini <paolo.carlini@oracle.com>
6317 * dwarf2out.c (is_cxx_auto): New.
6318 (is_base_type): Use it.
6319 (gen_type_die_with_usage): Likewise.
6321 2014-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6323 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Use
6324 CODE_FOR_altivec_vmrg*_direct rather than CODE_FOR_altivec_vmrg*.
6325 * config/rs6000/vsx.md (vsx_mergel_<mode>): Adjust for
6326 -maltivec=be with LE targets.
6327 (vsx_mergeh_<mode>): Likewise.
6328 * config/rs6000/altivec.md (UNSPEC_VMRG[HL]_DIRECT): New unspecs.
6329 (mulv8hi3): Use gen_altivec_vmrg[hl]w_direct.
6330 (altivec_vmrghb): Replace with define_expand and new
6331 *altivec_vmrghb_internal insn; adjust for -maltivec=be with LE targets.
6332 (altivec_vmrghb_direct): New define_insn.
6333 (altivec_vmrghh): Replace with define_expand and new
6334 *altivec_vmrghh_internal insn; adjust for -maltivec=be with LE targets.
6335 (altivec_vmrghh_direct): New define_insn.
6336 (altivec_vmrghw): Replace with define_expand and new
6337 *altivec_vmrghw_internal insn; adjust for -maltivec=be with LE targets.
6338 (altivec_vmrghw_direct): New define_insn.
6339 (*altivec_vmrghsf): Adjust for endianness.
6340 (altivec_vmrglb): Replace with define_expand and new
6341 *altivec_vmrglb_internal insn; adjust for -maltivec=be with LE targets.
6342 (altivec_vmrglb_direct): New define_insn.
6343 (altivec_vmrglh): Replace with define_expand and new
6344 *altivec_vmrglh_internal insn; adjust for -maltivec=be with LE targets.
6345 (altivec_vmrglh_direct): New define_insn.
6346 (altivec_vmrglw): Replace with define_expand and new
6347 *altivec_vmrglw_internal insn; adjust for -maltivec=be with LE targets.
6348 (altivec_vmrglw_direct): New define_insn.
6349 (*altivec_vmrglsf): Adjust for endianness.
6350 (vec_widen_umult_hi_v16qi): Use gen_altivec_vmrghh_direct.
6351 (vec_widen_umult_lo_v16qi): Use gen_altivec_vmrglh_direct.
6352 (vec_widen_smult_hi_v16qi): Use gen_altivec_vmrghh_direct.
6353 (vec_widen_smult_lo_v16qi): Use gen_altivec_vmrglh_direct.
6354 (vec_widen_umult_hi_v8hi): Use gen_altivec_vmrghw_direct.
6355 (vec_widen_umult_lo_v8hi): Use gen_altivec_vmrglw_direct.
6356 (vec_widen_smult_hi_v8hi): Use gen_altivec_vmrghw_direct.
6357 (vec_widen_smult_lo_v8hi): Use gen_altivec_vmrglw_direct.
6359 2014-01-29 Marcus Shawcroft <marcus.shawcroft@arm.com>
6361 * config/aarch64/aarch64.c (aarch64_expand_mov_immediate)
6362 (aarch64_legitimate_address_p, aarch64_class_max_nregs): Adjust
6365 2014-01-29 Richard Biener <rguenther@suse.de>
6367 PR tree-optimization/58742
6368 * tree-ssa-forwprop.c (associate_pointerplus): Rename to
6369 associate_pointerplus_align.
6370 (associate_pointerplus_diff): New function.
6371 (associate_pointerplus): Likewise. Call associate_pointerplus_align
6372 and associate_pointerplus_diff.
6374 2014-01-29 Richard Biener <rguenther@suse.de>
6376 * lto-streamer.h (LTO_major_version): Bump to 3.
6377 (LTO_minor_version): Reset to 0.
6379 2014-01-29 Renlin Li <Renlin.Li@arm.com>
6381 * config/arm/arm-arches.def (ARM_ARCH): Add armv7ve arch.
6382 * config/arm/arm.c (FL_FOR_ARCH7VE): New.
6383 (arm_file_start): Generate correct asm header for armv7ve.
6384 * config/arm/bpabi.h: Add multilib support for armv7ve.
6385 * config/arm/driver-arm.c: Change the architectures of cortex-a7
6386 and cortex-a15 to armv7ve.
6387 * config/arm/t-aprofile: Add multilib support for armv7ve.
6388 * doc/invoke.texi: Document -march=armv7ve.
6390 2014-01-29 Richard Biener <rguenther@suse.de>
6392 PR tree-optimization/58742
6393 * tree-ssa-forwprop.c (associate_plusminus): Return true
6394 if we changed sth, defer EH cleanup to ...
6395 (ssa_forward_propagate_and_combine): ... here. Call simplify_mult.
6396 (simplify_mult): New function.
6398 2014-01-29 Jakub Jelinek <jakub@redhat.com>
6401 PR tree-optimization/59920
6402 * tree.c (build_common_builtin_nodes): Remove
6403 __builtin_setjmp_dispatcher initialization.
6404 * omp-low.h (make_gimple_omp_edges): Add a new int * argument.
6405 * profile.c (branch_prob): Use gsi_start_nondebug_after_labels_bb
6406 instead of gsi_after_labels + manually skipping debug stmts.
6407 Don't ignore bbs with BUILT_IN_SETJMP_DISPATCHER, instead
6408 ignore bbs with IFN_ABNORMAL_DISPATCHER.
6409 * tree-inline.c (copy_edges_for_bb): Remove
6410 can_make_abnormal_goto argument, instead add abnormal_goto_dest
6411 argument. Ignore computed_goto_p stmts. Don't call
6412 make_abnormal_goto_edges. If a call might need abnormal edges
6413 for non-local gotos, see if it already has an edge to
6414 IFN_ABNORMAL_DISPATCHER or if it is IFN_ABNORMAL_DISPATCHER
6415 with true argument, don't do anything then, otherwise add
6416 EDGE_ABNORMAL from the call's bb to abnormal_goto_dest.
6417 (copy_cfg_body): Compute abnormal_goto_dest, adjust copy_edges_for_bb
6419 * gimple-low.c (struct lower_data): Remove calls_builtin_setjmp.
6420 (lower_function_body): Don't emit __builtin_setjmp_dispatcher.
6421 (lower_stmt): Don't set data->calls_builtin_setjmp.
6422 (lower_builtin_setjmp): Adjust comment.
6423 * builtins.def (BUILT_IN_SETJMP_DISPATCHER): Remove.
6424 * tree-cfg.c (found_computed_goto): Remove.
6425 (factor_computed_gotos): Remove.
6426 (make_goto_expr_edges): Return bool, true for computed gotos.
6427 Don't call make_abnormal_goto_edges.
6428 (build_gimple_cfg): Don't set found_computed_goto, don't call
6429 factor_computed_gotos.
6430 (computed_goto_p): No longer static.
6431 (make_blocks): Don't set found_computed_goto.
6432 (get_abnormal_succ_dispatcher, handle_abnormal_edges): New functions.
6433 (make_edges): If make_goto_expr_edges returns true, push bb
6434 into ab_edge_goto vector, for stmt_can_make_abnormal_goto calls
6435 instead of calling make_abnormal_goto_edges push bb into ab_edge_call
6436 vector. Record mapping between bbs and OpenMP regions if there
6437 are any, adjust make_gimple_omp_edges caller. Call
6438 handle_abnormal_edges.
6439 (make_abnormal_goto_edges): Remove.
6440 * tree-cfg.h (make_abnormal_goto_edges): Remove.
6441 (computed_goto_p, get_abnormal_succ_dispatcher): New prototypes.
6442 * internal-fn.c (expand_ABNORMAL_DISPATCHER): New function.
6443 * builtins.c (expand_builtin): Don't handle BUILT_IN_SETJMP_DISPATCHER.
6444 * internal-fn.def (ABNORMAL_DISPATCHER): New.
6445 * omp-low.c (make_gimple_omp_edges): Add region_idx argument, when
6446 filling *region also set *region_idx to (*region)->entry->index.
6449 * read-rtl.c (read_rtx_code): Clear all of RTX_CODE_SIZE (code).
6450 For REGs set ORIGINAL_REGNO.
6452 2014-01-29 Bingfeng Mei <bmei@broadcom.com>
6454 * doc/md.texi: Mention that a target shouldn't implement
6455 vec_widen_(s|u)mul_even/odd pair if it is less efficient
6458 2014-01-29 Jakub Jelinek <jakub@redhat.com>
6460 PR tree-optimization/59594
6461 * tree-vect-data-refs.c (vect_analyze_data_ref_accesses): Sort
6462 a copy of the datarefs vector rather than the vector itself.
6464 2014-01-28 Jason Merrill <jason@redhat.com>
6467 * dwarf2out.c (auto_die): New static.
6468 (gen_type_die_with_usage): Handle C++1y 'auto'.
6469 (gen_subprogram_die): If in-class DIE had 'auto', emit type again
6472 2014-01-28 H.J. Lu <hongjiu.lu@intel.com>
6475 * config/i386/gnu-user64.h (SPEC_32): Add "m16|" to "m32".
6476 (SPEC_X32): Likewise.
6477 (SPEC_64): Likewise.
6478 * config/i386/i386.c (ix86_option_override_internal): Turn off
6479 OPTION_MASK_ISA_64BIT, OPTION_MASK_ABI_X32 and OPTION_MASK_ABI_64
6481 (x86_file_start): Output .code16gcc for TARGET_16BIT.
6482 * config/i386/i386.h (TARGET_16BIT): New macro.
6483 (TARGET_16BIT_P): Likewise.
6484 * config/i386/i386.opt: Add m16.
6485 * doc/invoke.texi: Document -m16.
6487 2014-01-28 Jakub Jelinek <jakub@redhat.com>
6489 PR preprocessor/59935
6490 * input.c (location_get_source_line): Bail out on when line number
6491 is zero, and test the return value of lookup_or_add_file_to_cache_tab.
6493 2014-01-28 Richard Biener <rguenther@suse.de>
6495 PR tree-optimization/58742
6496 * tree-ssa-forwprop.c (associate_plusminus): Handle
6497 pointer subtraction of the form (T)(P + A) - (T)P.
6499 2014-01-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6501 * config/arm/arm.c (arm_new_rtx_costs): Remove useless statement
6504 2014-01-28 Richard Biener <rguenther@suse.de>
6507 2014-01-28 Richard Biener <rguenther@suse.de>
6509 PR rtl-optimization/45364
6510 PR rtl-optimization/59890
6511 * var-tracking.c (local_get_addr_clear_given_value): Handle
6512 already cleared slot.
6513 (val_reset): Handle not allocated local_get_addr_cache.
6514 (vt_find_locations): Use post-order on the inverted CFG.
6516 2014-01-28 Richard Biener <rguenther@suse.de>
6518 * tree-data-ref.h (ddr_is_anti_dependent, ddrs_have_anti_deps): Remove.
6520 2014-01-28 Richard Biener <rguenther@suse.de>
6522 PR rtl-optimization/45364
6523 PR rtl-optimization/59890
6524 * var-tracking.c (local_get_addr_clear_given_value): Handle
6525 already cleared slot.
6526 (val_reset): Handle not allocated local_get_addr_cache.
6527 (vt_find_locations): Use post-order on the inverted CFG.
6529 2014-01-28 Alan Modra <amodra@gmail.com>
6531 * Makefile.in (BUILD_CPPFLAGS): Do not use ALL_CPPFLAGS.
6532 * configure.ac <recursive call for build != host>: Define
6533 GENERATOR_FILE. Comment. Use CXX_FOR_BUILD, CXXFLAGS_FOR_BUILD
6534 and LD_FOR_BUILD too.
6535 * configure: Regenerate.
6537 2014-01-27 Allan Sandfeld Jensen <sandfeld@kde.org>
6539 * config/i386/i386.c (get_builtin_code_for_version): Separate
6540 Westmere from Nehalem, Ivy Bridge from Sandy Bridge and
6541 Broadwell from Haswell.
6543 2014-01-27 Steve Ellcey <sellcey@mips.com>
6545 * common/config/mips/mips-common.c (TARGET_DEFAULT_TARGET_FLAGS):
6546 Remove TARGET_FP_EXCEPTIONS_DEFAULT and MASK_FUSED_MADD.
6547 * config/mips/mips.c (mips_option_override): Change setting
6549 * config/mips/mips.h (TARGET_FP_EXCEPTIONS_DEFAULT): Remove.
6550 * config/mips/mips.opt (DSP, DSPR2, FP_EXCEPTIONS, FUSED_MADD, MIPS3D):
6551 Change from Mask to Var.
6553 2014-01-27 Jeff Law <law@redhat.com>
6555 * ipa-inline.c (inline_small_functions): Fix typo.
6557 2014-01-27 Ilya Tocar <ilya.tocar@intel.com>
6559 * config/i386/avx512fintrin.h (_mm512_mask_cvtepi32_storeu_epi8): New.
6560 (_mm512_mask_cvtsepi32_storeu_epi8): Ditto.
6561 (_mm512_mask_cvtusepi32_storeu_epi8): Ditto.
6562 (_mm512_mask_cvtepi32_storeu_epi16): Ditto.
6563 (_mm512_mask_cvtsepi32_storeu_epi16): Ditto.
6564 (_mm512_mask_cvtusepi32_storeu_epi16): Ditto.
6565 (_mm512_mask_cvtepi64_storeu_epi32): Ditto.
6566 (_mm512_mask_cvtsepi64_storeu_epi32): Ditto.
6567 (_mm512_mask_cvtusepi64_storeu_epi32): Ditto.
6568 (_mm512_mask_cvtepi64_storeu_epi16): Ditto.
6569 (_mm512_mask_cvtsepi64_storeu_epi16): Ditto.
6570 (_mm512_mask_cvtusepi64_storeu_epi16): Ditto.
6571 (_mm512_mask_cvtepi64_storeu_epi8): Ditto.
6572 (_mm512_mask_cvtsepi64_storeu_epi8): Ditto.
6573 (_mm512_mask_cvtusepi64_storeu_epi8): Ditto.
6574 (_mm512_storeu_epi64): Ditto.
6575 (_mm512_cmpge_epi32_mask): Ditto.
6576 (_mm512_cmpge_epu32_mask): Ditto.
6577 (_mm512_cmpge_epi64_mask): Ditto.
6578 (_mm512_cmpge_epu64_mask): Ditto.
6579 (_mm512_cmple_epi32_mask): Ditto.
6580 (_mm512_cmple_epu32_mask): Ditto.
6581 (_mm512_cmple_epi64_mask): Ditto.
6582 (_mm512_cmple_epu64_mask): Ditto.
6583 (_mm512_cmplt_epi32_mask): Ditto.
6584 (_mm512_cmplt_epu32_mask): Ditto.
6585 (_mm512_cmplt_epi64_mask): Ditto.
6586 (_mm512_cmplt_epu64_mask): Ditto.
6587 (_mm512_cmpneq_epi32_mask): Ditto.
6588 (_mm512_cmpneq_epu32_mask): Ditto.
6589 (_mm512_cmpneq_epi64_mask): Ditto.
6590 (_mm512_cmpneq_epu64_mask): Ditto.
6591 (_mm512_expand_pd): Ditto.
6592 (_mm512_expand_ps): Ditto.
6593 * config/i386/i386-builtin-types.def: Add PV16QI, PV16QI, PV16HI,
6594 VOID_PV8SI_V8DI_QI, VOID_PV8HI_V8DI_QI, VOID_PV16QI_V8DI_QI,
6595 VOID_PV16QI_V16SI_HI, VOID_PV16HI_V16SI_HI.
6596 * config/i386/i386.c (ix86_builtins): Add
6597 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK,
6598 IX86_BUILTIN_PMOVDB512_MEM, IX86_BUILTIN_PMOVDW512_MEM,
6599 IX86_BUILTIN_PMOVQB512_MEM, IX86_BUILTIN_PMOVQD512_MEM,
6600 IX86_BUILTIN_PMOVQW512_MEM, IX86_BUILTIN_PMOVSDB512_MEM,
6601 IX86_BUILTIN_PMOVSDW512_MEM, IX86_BUILTIN_PMOVSQB512_MEM,
6602 IX86_BUILTIN_PMOVSQD512_MEM, IX86_BUILTIN_PMOVSQW512_MEM,
6603 IX86_BUILTIN_PMOVUSDB512_MEM, IX86_BUILTIN_PMOVUSDW512_MEM,
6604 IX86_BUILTIN_PMOVUSQB512_MEM, IX86_BUILTIN_PMOVUSQD512_MEM,
6605 IX86_BUILTIN_PMOVUSQW512_MEM.
6606 (bdesc_special_args): Add __builtin_ia32_pmovusqd512mem_mask,
6607 __builtin_ia32_pmovsqd512mem_mask,
6608 __builtin_ia32_pmovqd512mem_mask,
6609 __builtin_ia32_pmovusqw512mem_mask,
6610 __builtin_ia32_pmovsqw512mem_mask,
6611 __builtin_ia32_pmovqw512mem_mask,
6612 __builtin_ia32_pmovusdw512mem_mask,
6613 __builtin_ia32_pmovsdw512mem_mask,
6614 __builtin_ia32_pmovdw512mem_mask,
6615 __builtin_ia32_pmovqb512mem_mask,
6616 __builtin_ia32_pmovusqb512mem_mask,
6617 __builtin_ia32_pmovsqb512mem_mask,
6618 __builtin_ia32_pmovusdb512mem_mask,
6619 __builtin_ia32_pmovsdb512mem_mask,
6620 __builtin_ia32_pmovdb512mem_mask.
6621 (bdesc_args): Add __builtin_ia32_expanddf512,
6622 __builtin_ia32_expandsf512.
6623 (ix86_expand_special_args_builtin): Handle VOID_FTYPE_PV8SI_V8DI_QI,
6624 VOID_FTYPE_PV8HI_V8DI_QI, VOID_FTYPE_PV16HI_V16SI_HI,
6625 VOID_FTYPE_PV16QI_V8DI_QI, VOID_FTYPE_PV16QI_V16SI_HI.
6626 * config/i386/sse.md (unspec): Add UNSPEC_EXPAND_NOMASK.
6627 (avx512f_<code><pmov_src_lower><mode>2_mask_store): New.
6628 (*avx512f_<code>v8div16qi2_store_mask): Renamed to ...
6629 (avx512f_<code>v8div16qi2_mask_store): This.
6630 (avx512f_expand<mode>): New.
6632 2014-01-27 Kirill Yukhin <kirill.yukhin@intel.com>
6634 * config/i386/avx512pfintrin.h (_mm512_mask_prefetch_i32gather_pd):
6636 (_mm512_mask_prefetch_i64gather_pd): Ditto.
6637 (_mm512_prefetch_i32scatter_pd): Ditto.
6638 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
6639 (_mm512_prefetch_i64scatter_pd): Ditto.
6640 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
6641 (_mm512_mask_prefetch_i32gather_ps): Fix operand type.
6642 (_mm512_mask_prefetch_i64gather_ps): Ditto.
6643 (_mm512_prefetch_i32scatter_ps): Ditto.
6644 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
6645 (_mm512_prefetch_i64scatter_ps): Ditto.
6646 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
6647 * config/i386/i386-builtin-types.def: Define
6648 VOID_FTYPE_QI_V8SI_PCINT64_INT_INT
6649 and VOID_FTYPE_QI_V8DI_PCINT64_INT_INT.
6650 * config/i386/i386.c (ix86_builtins): Define IX86_BUILTIN_GATHERPFQPD,
6651 IX86_BUILTIN_GATHERPFDPD, IX86_BUILTIN_SCATTERPFDPD,
6652 IX86_BUILTIN_SCATTERPFQPD.
6653 (ix86_init_mmx_sse_builtins): Define __builtin_ia32_gatherpfdpd,
6654 __builtin_ia32_gatherpfdps, __builtin_ia32_gatherpfqpd,
6655 __builtin_ia32_gatherpfqps, __builtin_ia32_scatterpfdpd,
6656 __builtin_ia32_scatterpfdps, __builtin_ia32_scatterpfqpd,
6657 __builtin_ia32_scatterpfqps.
6658 (ix86_expand_builtin): Expand new built-ins.
6659 * config/i386/sse.md (avx512pf_gatherpf<mode>): Add SF suffix,
6660 fix memory access data type.
6661 (*avx512pf_gatherpf<mode>_mask): Ditto.
6662 (*avx512pf_gatherpf<mode>): Ditto.
6663 (avx512pf_scatterpf<mode>): Ditto.
6664 (*avx512pf_scatterpf<mode>_mask): Ditto.
6665 (*avx512pf_scatterpf<mode>): Ditto.
6666 (GATHER_SCATTER_SF_MEM_MODE): New.
6667 (avx512pf_gatherpf<mode>df): Ditto.
6668 (*avx512pf_gatherpf<mode>df_mask): Ditto.
6669 (*avx512pf_scatterpf<mode>df): Ditto.
6671 2014-01-27 Jakub Jelinek <jakub@redhat.com>
6674 * expmed.h (expmed_mode_index): Rework so that analysis and optimziers
6675 know when the MODE_PARTIAL_INT and MODE_VECTOR_INT cases can never be
6678 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
6680 * common/config/arm/arm-common.c
6681 (arm_rewrite_mcpu): Handle multiple names.
6683 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
6685 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
6687 * gimple-builder.h (create_gimple_tmp): Delete.
6689 2014-01-27 Christian Bruel <christian.bruel@st.com>
6691 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Fix remaining bytes after
6694 2014-01-26 John David Anglin <danglin@gcc.gnu.org>
6696 * config/pa/pa.md (call): Generate indirect long calls to non-local
6697 functions when outputing 32-bit code.
6698 (call_value): Likewise except for special call to buggy powf function.
6700 * config/pa/pa.c (pa_attr_length_indirect_call): Adjust length of
6701 portable runtime and PIC indirect calls.
6702 (pa_output_indirect_call): Remove unnecessary nop from portable runtime
6703 and PIC call sequences. Use ldo instead of blr to set return register
6704 in PIC call sequence.
6706 2014-01-25 Walter Lee <walt@tilera.com>
6708 * config/tilegx/sync.md (atomic_fetch_sub): Fix negation and
6709 avoid clobbering a live register.
6711 2014-01-25 Walter Lee <walt@tilera.com>
6713 * config/tilegx/tilegx-c.c (tilegx_cpu_cpp_builtins):
6714 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2}.
6715 * config/tilegx/tilepro-c.c (tilepro_cpu_cpp_builtins):
6716 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2,4,8}.
6718 2014-01-25 Walter Lee <walt@tilera.com>
6720 * config/tilegx/tilegx.c (tilegx_function_arg): Start 16-byte
6721 arguments on even registers.
6722 (tilegx_gimplify_va_arg_expr): Align 16-byte var args to
6724 * config/tilegx/tilegx.h (STACK_BOUNDARY): Change to 16 bytes.
6725 (BIGGEST_ALIGNMENT): Ditto.
6726 (BIGGEST_FIELD_ALIGNMENT): Ditto.
6728 2014-01-25 Walter Lee <walt@tilera.com>
6730 * config/tilegx/tilegx.c (tilegx_gen_bundles): Delete barrier
6731 insns before bundling.
6732 * config/tilegx/tilegx.md (tile_network_barrier): Update comment.
6734 2014-01-25 Walter Lee <walt@tilera.com>
6736 * config/tilegx/tilegx.c (tilegx_expand_builtin): Set
6737 PREFETCH_SCHEDULE_BARRIER_P to true for prefetches.
6738 * config/tilepro/tilepro.c (tilepro_expand_builtin): Ditto.
6740 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
6742 * config/mips/constraints.md (kl): Delete.
6743 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Turn into
6744 define expands, using...
6745 (divmod<mode>4_mips16, udivmod<mode>4_mips16): ...these new
6746 instructions for MIPS16.
6747 (*divmod<mode>4, *udivmod<mode>4): New patterns, taken from the
6748 non-MIPS16 version of the old divmod<mode>4 and udivmod<mode>4.
6750 2014-01-25 Walter Lee <walt@tilera.com>
6752 * config/tilepro/tilepro.md (ctzdi2): Use register_operand predicate.
6756 2014-01-25 Walter Lee <walt@tilera.com>
6758 * config/tilegx/tilegx.c (tilegx_expand_to_rtl_hook): New.
6759 (TARGET_EXPAND_TO_RTL_HOOK): Define.
6761 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
6763 * rtlanal.c (canonicalize_condition): Split out duplicated mode check.
6766 2014-01-25 Jakub Jelinek <jakub@redhat.com>
6768 * print-rtl.c (in_call_function_usage): New var.
6769 (print_rtx): When in CALL_INSN_FUNCTION_USAGE, always print
6770 EXPR_LIST mode as mode and not as reg note name.
6773 * cfgloopmanip.c (copy_loop_info): If
6774 loop->warned_aggressive_loop_optimizations, make sure
6775 the flag is set in target loop too.
6777 2014-01-24 Balaji V. Iyer <balaji.v.iyer@intel.com>
6779 * builtins.c (is_builtin_name): Renamed flag_enable_cilkplus to
6781 * builtins.def: Likewise.
6782 * cilk.h (fn_contains_cilk_spawn_p): Likewise.
6783 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Likewise.
6784 * ira.c (ira_setup_eliminable_regset): Likewise.
6785 * omp-low.c (gate_expand_omp): Likewise.
6786 (execute_lower_omp): Likewise.
6787 (diagnose_sb_0): Likewise.
6788 (gate_diagnose_omp_blocks): Likewise.
6789 (simd_clone_clauses_extract): Likewise.
6792 2014-01-24 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6794 * config/rs6000/rs6000.c (rs6000_expand_vec_perm_const_1): Remove
6795 correction for little endian...
6796 * config/rs6000/vsx.md (vsx_xxpermdi2_<mode>_1): ...and move it to
6799 2014-01-24 Jeff Law <law@redhat.com>
6801 PR tree-optimization/59919
6802 * tree-vrp.c (find_assert_locations_1): Do not register asserts
6803 for non-returning calls.
6805 2014-01-24 James Greenhalgh <james.greenhalgh@arm.com>
6807 * common/config/aarch64/aarch64-common.c
6808 (aarch64_rewrite_mcpu): Handle multiple names.
6809 * config/aarch64/aarch64.h
6810 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
6812 2014-01-24 Dodji Seketeli <dodji@redhat.com>
6814 * input.c (add_file_to_cache_tab): Handle the case where fopen
6817 2014-01-23 H.J. Lu <hongjiu.lu@intel.com>
6820 * config/i386/i386.md (pushsf splitter): Get stack adjustment
6821 from push operand if code of push isn't PRE_DEC.
6823 2014-01-23 Michael Meissner <meissner@linux.vnet.ibm.com>
6826 * doc/invoke.texi (RS/6000 and PowerPC Options): Document
6827 -mquad-memory-atomic. Update -mquad-memory documentation to say
6828 it is only used for non-atomic loads/stores.
6830 * config/rs6000/predicates.md (quad_int_reg_operand): Allow either
6831 -mquad-memory or -mquad-memory-atomic switches.
6833 * config/rs6000/rs6000-cpus.def (ISA_2_7_MASKS_SERVER): Add
6834 -mquad-memory-atomic to ISA 2.07 support.
6836 * config/rs6000/rs6000.opt (-mquad-memory-atomic): Add new switch
6837 to separate support of normal quad word memory operations (ldq, stq)
6838 from the atomic quad word memory operations.
6840 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
6841 support to separate non-atomic quad word operations from atomic
6842 quad word operations. Disable non-atomic quad word operations in
6843 little endian mode so that we don't have to swap words after the
6844 load and before the store.
6845 (quad_load_store_p): Add comment about atomic quad word support.
6846 (rs6000_opt_masks): Add -mquad-memory-atomic to the list of
6847 options printed with -mdebug=reg.
6849 * config/rs6000/rs6000.h (TARGET_SYNC_TI): Use
6850 -mquad-memory-atomic as the test for whether we have quad word
6851 atomic instructions.
6852 (TARGET_SYNC_HI_QI): If either -mquad-memory-atomic, -mquad-memory,
6853 or -mp8-vector are used, allow byte/half-word atomic operations.
6855 * config/rs6000/sync.md (load_lockedti): Insure that the address
6856 is a proper indexed or indirect address for the lqarx instruction.
6857 On little endian systems, swap the hi/lo registers after the lqarx
6859 (load_lockedpti): Use indexed_or_indirect_operand predicate to
6860 insure the address is valid for the lqarx instruction.
6861 (store_conditionalti): Insure that the address is a proper indexed
6862 or indirect address for the stqcrx. instruction. On little endian
6863 systems, swap the hi/lo registers before doing the stqcrx.
6865 (store_conditionalpti): Use indexed_or_indirect_operand predicate to
6866 insure the address is valid for the stqcrx. instruction.
6868 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
6869 Define __QUAD_MEMORY__ and __QUAD_MEMORY_ATOMIC__ based on what
6870 type of quad memory support is available.
6872 2014-01-23 Vladimir Makarov <vmakarov@redhat.com>
6875 * lra-constraints.c (simplify_operand_subreg): Spill pseudo if
6876 there is a danger of looping.
6878 2014-01-23 Pat Haugen <pthaugen@us.ibm.com>
6880 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
6881 force flag_ira_loop_pressure if set via command line.
6883 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
6885 * config/aarch64/aarch64-simd-builtins.def (ashr): DI mode removed.
6886 (ashr_simd): New builtin handling DI mode.
6887 * config/aarch64/aarch64-simd.md (aarch64_ashr_simddi): New pattern.
6888 (aarch64_sshr_simddi): New match pattern.
6889 * config/aarch64/arm_neon.h (vshr_n_s32): Builtin call modified.
6890 (vshrd_n_s64): Likewise.
6891 * config/aarch64/predicates.md (aarch64_shift_imm64_di): New predicate.
6893 2014-01-23 Nick Clifton <nickc@redhat.com>
6895 * config/msp430/msp430.h (ASM_SPEC): Pass the -mcpu as -mcpu.
6896 (LIB_SPEC): Drop use of memory.ld and peripherals.ld scripts in
6897 favour of mcu specific scripts.
6898 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more matches for
6901 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
6902 Alex Velenko <Alex.Velenko@arm.com>
6904 * config/aarch64/arm_neon.h (vaddv_s8): __LANE0 cleanup.
6905 (vaddv_s16): Likewise.
6906 (vaddv_s32): Likewise.
6907 (vaddv_u8): Likewise.
6908 (vaddv_u16): Likewise.
6909 (vaddv_u32): Likewise.
6910 (vaddvq_s8): Likewise.
6911 (vaddvq_s16): Likewise.
6912 (vaddvq_s32): Likewise.
6913 (vaddvq_s64): Likewise.
6914 (vaddvq_u8): Likewise.
6915 (vaddvq_u16): Likewise.
6916 (vaddvq_u32): Likewise.
6917 (vaddvq_u64): Likewise.
6918 (vaddv_f32): Likewise.
6919 (vaddvq_f32): Likewise.
6920 (vaddvq_f64): Likewise.
6921 (vmaxv_f32): Likewise.
6922 (vmaxv_s8): Likewise.
6923 (vmaxv_s16): Likewise.
6924 (vmaxv_s32): Likewise.
6925 (vmaxv_u8): Likewise.
6926 (vmaxv_u16): Likewise.
6927 (vmaxv_u32): Likewise.
6928 (vmaxvq_f32): Likewise.
6929 (vmaxvq_f64): Likewise.
6930 (vmaxvq_s8): Likewise.
6931 (vmaxvq_s16): Likewise.
6932 (vmaxvq_s32): Likewise.
6933 (vmaxvq_u8): Likewise.
6934 (vmaxvq_u16): Likewise.
6935 (vmaxvq_u32): Likewise.
6936 (vmaxnmv_f32): Likewise.
6937 (vmaxnmvq_f32): Likewise.
6938 (vmaxnmvq_f64): Likewise.
6939 (vminv_f32): Likewise.
6940 (vminv_s8): Likewise.
6941 (vminv_s16): Likewise.
6942 (vminv_s32): Likewise.
6943 (vminv_u8): Likewise.
6944 (vminv_u16): Likewise.
6945 (vminv_u32): Likewise.
6946 (vminvq_f32): Likewise.
6947 (vminvq_f64): Likewise.
6948 (vminvq_s8): Likewise.
6949 (vminvq_s16): Likewise.
6950 (vminvq_s32): Likewise.
6951 (vminvq_u8): Likewise.
6952 (vminvq_u16): Likewise.
6953 (vminvq_u32): Likewise.
6954 (vminnmv_f32): Likewise.
6955 (vminnmvq_f32): Likewise.
6956 (vminnmvq_f64): Likewise.
6958 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
6960 * config/aarch64/aarch64-simd.md
6961 (aarch64_dup_lane<mode>): Correct lane number on big-endian.
6962 (aarch64_dup_lane_<vswap_widthi_name><mode>): Likewise.
6963 (*aarch64_mul3_elt<mode>): Likewise.
6964 (*aarch64_mul3_elt<vswap_width_name><mode>): Likewise.
6965 (*aarch64_mul3_elt_to_64v2df): Likewise.
6966 (*aarch64_mla_elt<mode>): Likewise.
6967 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
6968 (*aarch64_mls_elt<mode>): Likewise.
6969 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
6970 (*aarch64_fma4_elt<mode>): Likewise.
6971 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
6972 (*aarch64_fma4_elt_to_64v2df): Likewise.
6973 (*aarch64_fnma4_elt<mode>): Likewise.
6974 (*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
6975 (*aarch64_fnma4_elt_to_64v2df): Likewise.
6976 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
6977 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
6978 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
6979 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
6980 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
6981 (aarch64_sqdmull_lane<mode>_internal): Likewise.
6982 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
6984 2013-01-23 Alex Velenko <Alex.Velenko@arm.com>
6986 * config/aarch64/aarch64-simd.md
6987 (aarch64_be_checked_get_lane<mode>): New define_expand.
6988 * config/aarch64/aarch64-simd-builtins.def
6989 (BUILTIN_VALL (GETLANE, be_checked_get_lane, 0)):
6990 New builtin definition.
6991 * config/aarch64/arm_neon.h: (__aarch64_vget_lane_any):
6992 Use new safe be builtin.
6994 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
6996 * config/aarch64/aarch64-simd.md (aarch64_be_ld1<mode>):
6998 (aarch64_be_st1<mode>): Likewise.
6999 (aarch_ld1<VALL:mode>): Define_expand modified.
7000 (aarch_st1<VALL:mode>): Likewise.
7001 * config/aarch64/aarch64.md (UNSPEC_LD1): New unspec definition.
7002 (UNSPEC_ST1): Likewise.
7004 2014-01-23 David Holsgrove <david.holsgrove@xilinx.com>
7006 * config/microblaze/microblaze.md: Add trap insn and attribute
7008 2014-01-23 Dodji Seketeli <dodji@redhat.com>
7010 PR preprocessor/58580
7011 * input.h (location_get_source_line): Take an additional line_size
7013 (void diagnostics_file_cache_fini): Declare new function.
7014 * input.c (struct fcache): New type.
7015 (fcache_tab_size, fcache_buffer_size, fcache_line_record_size):
7016 New static constants.
7017 (diagnostic_file_cache_init, total_lines_num)
7018 (lookup_file_in_cache_tab, evicted_cache_tab_entry)
7019 (add_file_to_cache_tab, lookup_or_add_file_to_cache_tab)
7020 (needs_read, needs_grow, maybe_grow, read_data, maybe_read_data)
7021 (get_next_line, read_next_line, goto_next_line, read_line_num):
7022 New static function definitions.
7023 (diagnostic_file_cache_fini): New function.
7024 (location_get_source_line): Take an additional output line_len
7025 parameter. Re-write using lookup_or_add_file_to_cache_tab and
7027 * diagnostic.c (diagnostic_finish): Call
7028 diagnostic_file_cache_fini.
7029 (adjust_line): Take an additional input parameter for the length
7030 of the line, rather than calculating it with strlen.
7031 (diagnostic_show_locus): Adjust the use of
7032 location_get_source_line and adjust_line with respect to their new
7033 signature. While displaying a line now, do not stop at the first
7034 null byte. Rather, display the zero byte as a space and keep
7035 going until we reach the size of the line.
7036 * Makefile.in: Add vec.o to OBJS-libcommon
7038 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
7039 Ilya Tocar <ilya.tocar@intel.com>
7041 * config/i386/avx512fintrin.h (_mm512_kmov): New.
7042 * config/i386/i386.c (IX86_BUILTIN_KMOV16): Ditto.
7043 (__builtin_ia32_kmov16): Ditto.
7044 * config/i386/i386.md (UNSPEC_KMOV): New.
7047 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
7049 * config/i386/avx512fintrin.h (_mm512_loadu_si512): Rename.
7050 (_mm512_storeu_si512): Ditto.
7052 2014-01-23 Richard Sandiford <rdsandiford@googlemail.com>
7055 * rtl.h (get_referenced_operands): Declare.
7056 * recog.c (get_referenced_operands): New function.
7057 * config/mips/mips.c (mips_reorg_process_insns): Check which asm
7058 operands have been referenced when recording LO_SUM references.
7060 2014-01-22 David Holsgrove <david.holsgrove@xilinx.com>
7062 * config/microblaze/microblaze.md: Correct bswaphi2 insn.
7064 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
7066 * config/i386/x86-tune.def (X86_TUNE_ACCUMULATE_OUTGOING_ARGS):
7067 Enable for generic and recent AMD targets.
7069 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
7071 * combine-stack-adj.c (combine_stack_adjustments_for_block): Remove
7072 ARG_SIZE note when adjustment was eliminated.
7074 2014-01-22 Jeff Law <law@redhat.com>
7076 PR tree-optimization/59597
7077 * tree-ssa-threadupdate.c (dump_jump_thread_path): Move to earlier
7078 in file. Accept new argument REGISTERING and use it to modify
7079 dump output appropriately.
7080 (register_jump_thread): Corresponding changes.
7081 (mark_threaded_blocks): Reinstate code to cancel unprofitable
7082 thread paths involving joiner blocks. Add code to dump cancelled
7083 jump threading paths.
7085 2014-01-22 Vladimir Makarov <vmakarov@redhat.com>
7087 PR rtl-optimization/59477
7088 * lra-constraints.c (inherit_in_ebb): Process call for living hard
7089 regs. Update reloads_num and potential_reload_hard_regs for all insns.
7091 2014-01-22 Tom Tromey <tromey@redhat.com>
7093 * config/i386/i386-interix.h (i386_pe_unique_section): Don't use
7095 * config/cr16/cr16-protos.h (notice_update_cc): Don't use PARAMS.
7097 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
7099 PR rtl-optimization/59896
7100 * lra-constraints.c (process_alt_operands): Check unused note for
7101 matched operands of insn with no output reloads.
7103 2014-01-21 Richard Sandiford <rdsandiford@googlemail.com>
7105 * config/mips/mips.c (mips_move_to_gpr_cost): Add M16_REGS case.
7106 (mips_move_from_gpr_cost): Likewise.
7108 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
7110 PR rtl-optimization/59858
7111 * lra-constraints.c (SMALL_REGISTER_CLASS_P): Use
7112 ira_class_hard_regs_num.
7113 (process_alt_operands): Increase reject for dying matched operand.
7115 2014-01-21 Jakub Jelinek <jakub@redhat.com>
7118 * config/i386/i386.c (expand_small_movmem_or_setmem): If mode is
7119 smaller than size, perform several stores or loads and stores
7120 at dst + count - size to store or copy all of size bytes, rather
7121 than just last modesize bytes.
7123 2014-01-20 DJ Delorie <dj@redhat.com>
7125 * config/rl78/rl78.c (rl78_propogate_register_origins): Verify
7126 that CLOBBERs are REGs before propogating their values.
7128 2014-01-20 H.J. Lu <hongjiu.lu@intel.com>
7131 * cgraph.c (cgraph_inline_failed_string): Add type to DEFCIFCODE.
7132 (cgraph_inline_failed_type): New function.
7133 * cgraph.h (DEFCIFCODE): Add type.
7134 (cgraph_inline_failed_type_t): New enum.
7135 (cgraph_inline_failed_type): New prototype.
7136 * cif-code.def: Add CIF_FINAL_NORMAL to OK, FUNCTION_NOT_CONSIDERED,
7137 FUNCTION_NOT_OPTIMIZED, REDEFINED_EXTERN_INLINE,
7138 FUNCTION_NOT_INLINE_CANDIDATE, LARGE_FUNCTION_GROWTH_LIMIT,
7139 LARGE_STACK_FRAME_GROWTH_LIMIT, MAX_INLINE_INSNS_SINGLE_LIMIT,
7140 MAX_INLINE_INSNS_AUTO_LIMIT, INLINE_UNIT_GROWTH_LIMIT,
7141 RECURSIVE_INLINING, UNLIKELY_CALL, NOT_DECLARED_INLINED,
7142 OPTIMIZING_FOR_SIZE, ORIGINALLY_INDIRECT_CALL,
7143 INDIRECT_UNKNOWN_CALL, USES_COMDAT_LOCAL.
7144 Add CIF_FINAL_ERROR to UNSPECIFIED, BODY_NOT_AVAILABLE,
7145 FUNCTION_NOT_INLINABLE, OVERWRITABLE, MISMATCHED_ARGUMENTS,
7146 EH_PERSONALITY, NON_CALL_EXCEPTIONS, TARGET_OPTION_MISMATCH,
7147 OPTIMIZATION_MISMATCH.
7148 * tree-inline.c (expand_call_inline): Emit errors during
7149 early_inlining if cgraph_inline_failed_type returns CIF_FINAL_ERROR.
7151 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
7154 * config/i386/sse.md (*andnot<mode>3<mask_name>): Handle MODE_V16SF
7155 mode attribute in insn output.
7157 2014-01-20 Eric Botcazou <ebotcazou@adacore.com>
7159 * output.h (output_constant): Delete.
7160 * varasm.c (output_constant): Make private.
7162 2014-01-20 Alex Velenko <Alex.Velenko@arm.com>
7164 * config/aarch64/aarch64-simd.md (vec_perm<mode>): Add BE check.
7166 2014-01-20 Jakub Jelinek <jakub@redhat.com>
7169 * tree.h (fold_builtin_strcat): New prototype.
7170 * builtins.c (fold_builtin_strcat): No longer static. Add len
7171 argument, if non-NULL, don't call c_strlen. Optimize
7172 directly into __builtin_memcpy instead of __builtin_strcpy.
7173 (fold_builtin_2): Adjust fold_builtin_strcat caller.
7174 * gimple-fold.c (gimple_fold_builtin): Handle BUILT_IN_STRCAT.
7176 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
7178 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
7179 for SImode_address_operand operands, having only a REG argument.
7181 2014-01-20 Marcus Shawcroft <marcus.shawcroft@arm.com>
7183 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER): Expand
7184 loader name using mbig-endian.
7185 (LINUX_TARGET_LINK_SPEC): Pass linker -m flag.
7187 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
7189 * doc/invoke.texi (-march): Clarify documentation for AArch64.
7193 2014-01-20 Tejas Belagod <tejas.belagod@arm.com>
7195 * config/aarch64/aarch64-protos.h
7196 (aarch64_cannot_change_mode_class_ptr): Declare.
7197 * config/aarch64/aarch64.c (aarch64_cannot_change_mode_class,
7198 aarch64_cannot_change_mode_class_ptr): New.
7199 * config/aarch64/aarch64.h (CANNOT_CHANGE_MODE_CLASS): Change to call
7200 backend hook aarch64_cannot_change_mode_class.
7202 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
7204 * common/config/aarch64/aarch64-common.c
7205 (aarch64_handle_option): Don't handle any option order logic here.
7206 * config/aarch64/aarch64.c (aarch64_parse_arch): Do not override
7207 selected_cpu, warn on architecture version mismatch.
7208 (aarch64_override_options): Fix parsing order for option strings.
7210 2014-01-20 Jan-Benedict Glaw <jbglaw@lug-owl.de>
7211 Iain Sandoe <iain@codesourcery.com>
7214 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Fix unused variable
7215 warning. Amend comment to reflect current functionality.
7217 2014-01-20 Richard Biener <rguenther@suse.de>
7220 * builtins.c (fold_builtin_strcat): Remove case better handled
7221 by tree-ssa-strlen.c.
7223 2014-01-20 Alan Lawrence <alan.lawrence@arm.com>
7225 * config/aarch64/aarch64.opt
7226 (mcpu, march, mtune): Make case-insensitive.
7228 2014-01-20 Jakub Jelinek <jakub@redhat.com>
7231 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
7232 if operands[1] is a REG or ZERO_EXTEND of a REG.
7234 2014-01-19 Jan Hubicka <hubicka@ucw.cz>
7236 * varasm.c (compute_reloc_for_constant): Use targetm.binds_local_p.
7238 2014-01-19 John David Anglin <danglin@gcc.gnu.org>
7240 * config/pa/pa.c (pa_attr_length_millicode_call): Correct length of
7241 long non-pic millicode calls.
7243 2014-01-19 Jan-Benedict Glaw <jbglaw@lug-owl.de>
7245 * config/vax/vax.h (FUNCTION_ARG_REGNO_P): Fix unused variable warning.
7247 2014-01-19 Kito Cheng <kito@0xlab.org>
7249 * builtins.c (expand_movstr): Check movstr expand done or fail.
7251 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
7252 H.J. Lu <hongjiu.lu@intel.com>
7255 * config/i386/i386.md (*lea<mode>): Zero-extend return register
7256 to DImode for zero-extended addresses.
7258 2014-01-19 Jakub Jelinek <jakub@redhat.com>
7260 PR rtl-optimization/57763
7261 * bb-reorder.c (fix_crossing_unconditional_branches): Set JUMP_LABEL
7262 on the new indirect jump_insn and increment LABEL_NUSES (label).
7264 2014-01-18 H.J. Lu <hongjiu.lu@intel.com>
7268 * config.gcc (x86_archs): New variable.
7269 (x86_64_archs): Likewise.
7270 (x86_cpus): Likewise.
7271 Use $x86_archs, $x86_64_archs and $x86_cpus to check valid
7272 --with-arch/--with-cpu= options.
7273 Support --with-arch=/--with-cpu={nehalem,westmere,
7274 sandybridge,ivybridge,haswell,broadwell,bonnell,silvermont}.
7276 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
7278 * config/i386/i386.c (ix86_adjust_cost): Reorder PROCESSOR_K8
7279 and PROCESSOR_ATHLON to simplify code. Move "memory" calculation.
7281 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
7283 * config/i386/i386.md (*swap<mode>): Rename from swap<mode>.
7285 2014-01-18 Jakub Jelinek <jakub@redhat.com>
7288 * config/i386/i386-c.c (ix86_pragma_target_parse): Temporarily
7289 clear cpp_get_options (parse_in)->warn_unused_macros for
7290 ix86_target_macros_internal with cpp_define.
7292 2014-01-18 Richard Sandiford <rdsandiford@googlemail.com>
7294 * jump.c (delete_related_insns): Keep (use (insn))s.
7295 * reorg.c (redundant_insn): Check for barriers too.
7297 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
7299 * config/i386/i386.c (ix86_split_lea_for_addr): Fix a comment typo.
7301 2014-01-17 John David Anglin <danglin@gcc.gnu.org>
7303 * config/pa/pa.c (pa_attr_length_indirect_call): Don't output a short
7304 call to $$dyncall when TARGET_LONG_CALLS is true.
7306 2014-01-17 Jeff Law <law@redhat.com>
7308 * ree.c (combine_set_extension): Temporarily disable test for
7309 changing number of hard registers.
7311 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
7314 * ipa-inline-analysis.c (inline_free_summary):
7315 Do not free summary of aliases.
7317 2014-01-17 Jakub Jelinek <jakub@redhat.com>
7320 * gimplify.c (gimplify_expr): Use create_tmp_var
7321 instead of create_tmp_var_raw. If cond doesn't have
7322 integral type, don't add the IFN_ANNOTATE builtin at all.
7324 2014-01-17 Martin Jambor <mjambor@suse.cz>
7327 * ipa-cp.c (prev_edge_clone): New variable.
7328 (grow_next_edge_clone_vector): Renamed to grow_edge_clone_vectors.
7329 Also resize prev_edge_clone vector.
7330 (ipcp_edge_duplication_hook): Also update prev_edge_clone.
7331 (ipcp_edge_removal_hook): New function.
7332 (ipcp_driver): Register ipcp_edge_removal_hook.
7334 2014-01-17 Andrew Pinski <apinski@cavium.com>
7335 Steve Ellcey <sellcey@mips.com>
7338 * config/mips/mips.c (mips_print_operand): Check operand mode instead
7341 2014-01-17 Jeff Law <law@redhat.com>
7344 * passes.def: Reorder pass_copy_prop, pass_unrolli, pass_ccp sequence
7345 so that pass_ccp runs first.
7347 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
7349 * config/i386/i386.c (ix86_lea_outperforms): Use TARGET_XXX.
7350 (ix86_adjust_cost): Use !TARGET_XXX.
7351 (do_reorder_for_imul): Likewise.
7352 (swap_top_of_ready_list): Likewise.
7353 (ix86_sched_reorder): Likewise.
7355 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
7357 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
7358 PROCESSOR_INTEL. Treat like PROCESSOR_GENERIC.
7359 * config/i386/i386.c (intel_memcpy): New. Duplicate slm_memcpy.
7360 (intel_memset): New. Duplicate slm_memset.
7361 (intel_cost): New. Duplicate slm_cost.
7362 (m_INTEL): New macro.
7363 (processor_target_table): Add "intel".
7364 (ix86_option_override_internal): Replace PROCESSOR_SILVERMONT
7365 with PROCESSOR_INTEL for "intel".
7366 (ix86_lea_outperforms): Support PROCESSOR_INTEL. Duplicate
7367 PROCESSOR_SILVERMONT.
7368 (ix86_issue_rate): Likewise.
7369 (ix86_adjust_cost): Likewise.
7370 (ia32_multipass_dfa_lookahead): Likewise.
7371 (swap_top_of_ready_list): Likewise.
7372 (ix86_sched_reorder): Likewise.
7373 (ix86_avoid_lea_for_addr): Check TARGET_AVOID_LEA_FOR_ADDR
7374 instead of TARGET_OPT_AGU.
7375 * config/i386/i386.h (TARGET_INTEL): New.
7376 (TARGET_AVOID_LEA_FOR_ADDR): Likewise.
7377 (processor_type): Add PROCESSOR_INTEL.
7378 * config/i386/x86-tune.def: Support m_INTEL. Duplicate m_SILVERMONT.
7379 Add X86_TUNE_AVOID_LEA_FOR_ADDR.
7381 2014-01-17 Marek Polacek <polacek@redhat.com>
7384 * gimple-fold.c (fold_array_ctor_reference): Don't fold if element
7387 2014-01-17 Richard Biener <rguenther@suse.de>
7389 PR tree-optimization/46590
7390 * opts.c (default_options_table): Add entries for
7391 OPT_fbranch_count_reg, OPT_fmove_loop_invariants and OPT_ftree_pta,
7392 all enabled at -O1 but not for -Og.
7393 * common.opt (fbranch-count-reg): Remove Init(1).
7394 (fmove-loop-invariants): Likewise.
7395 (ftree-pta): Likewise.
7397 2014-01-17 Jakub Jelinek <jakub@redhat.com>
7399 * config/i386/i386.c (ix86_data_alignment): For compatibility with
7400 (incorrect) GCC 4.8 and earlier alignment assumptions ensure we align
7401 decls to at least the GCC 4.8 used alignments.
7404 * tree-nested.c (convert_nonlocal_reference_stmt,
7405 convert_local_reference_stmt): For NAMELIST_DECLs in gimple_bind_vars
7406 of GIMPLE_BIND stmts, adjust associated decls.
7408 2014-01-17 Richard Biener <rguenther@suse.de>
7410 PR tree-optimization/46590
7411 * vec.h (vec<>::bseach): New member function implementing
7412 binary search according to C89 bsearch.
7413 (vec<>::qsort): Avoid calling ::qsort for vectors with sizes 0 or 1.
7414 * tree-ssa-loop-im.c (struct mem_ref): Make stored member a
7415 bitmap pointer again. Make accesses_in_loop a flat array.
7416 (mem_ref_obstack): New global.
7417 (outermost_indep_loop): Adjust for mem_ref->stored changes.
7418 (mark_ref_stored): Likewise.
7419 (ref_indep_loop_p_2): Likewise.
7420 (set_ref_stored_in_loop): New helper function.
7421 (mem_ref_alloc): Allocate mem_refs on the mem_ref_obstack obstack.
7422 (memref_free): Adjust.
7423 (record_mem_ref_loc): Simplify.
7424 (gather_mem_refs_stmt): Adjust.
7425 (sort_locs_in_loop_postorder_cmp): New function.
7426 (analyze_memory_references): Sort accesses_in_loop after
7427 loop postorder number.
7428 (find_ref_loc_in_loop_cmp): New function.
7429 (for_all_locs_in_loop): Find relevant cluster of locs in
7430 accesses_in_loop and iterate without recursion.
7431 (execute_sm): Avoid uninit warning.
7432 (struct ref_always_accessed): Simplify.
7433 (ref_always_accessed::operator ()): Likewise.
7434 (ref_always_accessed_p): Likewise.
7435 (tree_ssa_lim_initialize): Initialize mem_ref_obstack, compute
7436 loop postorder numbers here.
7437 (tree_ssa_lim_finalize): Free mem_ref_obstack and loop postorder
7440 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
7443 * passes.c (rest_of_decl_compilation): Don't call varpool_finalize_decl
7444 on decls for which assemble_alias has been called.
7446 2014-01-17 Nick Clifton <nickc@redhat.com>
7448 * config/msp430/msp430.opt: (mcpu): New option.
7449 * config/msp430/msp430.c (msp430_mcu_name): Use target_mcu.
7450 (msp430_option_override): Parse target_cpu. If the MCU name
7451 matches a generic string, clear target_mcu.
7452 (msp430_attr): Allow numeric interrupt values up to 63.
7453 (msp430_expand_epilogue): No longer invert operand 1 of gen_popm.
7454 * config/msp430/msp430.h (ASM_SPEC): Convert -mcpu into a -mmcu
7456 * config/msp430/t-msp430: (MULTILIB_MATCHES): Remove mcu matches.
7458 * config/msp430/msp430.md (popm): Use %J rather than %I.
7459 (addsi3): Use msp430_nonimmediate_operand for operand 2.
7460 (addhi_cy_i): Use immediate_operand for operand 2.
7461 * doc/invoke.texi: Document -mcpu option.
7463 2014-01-17 Richard Biener <rguenther@suse.de>
7465 PR rtl-optimization/38518
7466 * df.h (df_analyze_loop): Declare.
7467 * df-core.c: Include cfgloop.h.
7468 (df_analyze_1): Split out main part of df_analyze.
7469 (df_analyze): Adjust.
7470 (loop_inverted_post_order_compute): New function.
7471 (loop_post_order_compute): Likewise.
7472 (df_analyze_loop): New function avoiding whole-function
7474 * loop-invariant.c (find_defs): Use df_analyze_loop.
7475 (find_invariants): Adjust.
7476 * loop-iv.c (iv_analysis_loop_init): Use df_analyze_loop.
7478 2014-01-17 Zhenqiang Chen <zhenqiang.chen@arm.com>
7480 * config/arm/arm.c (arm_v7m_tune): Set max_insns_skipped to 2.
7481 (thumb2_final_prescan_insn): Set max to MAX_INSN_PER_IT_BLOCK.
7483 2014-01-16 Ilya Enkovich <ilya.enkovich@intel.com>
7485 * ipa-ref.c (ipa_remove_stmt_references): Fix references
7486 traversal when removing references.
7488 2014-01-16 Jan Hubicka <hubicka@ucw.cz>
7491 * tree.c (get_binfo_at_offset): Look harder for virtual bases.
7493 2014-01-16 Bernd Schmidt <bernds@codesourcery.com>
7496 * reload.c (find_reloads_address_1): Do not use RELOAD_OTHER when
7497 pushing a reload for an autoinc when we had previously reloaded an
7498 inner part of the address.
7500 2014-01-16 Jakub Jelinek <jakub@redhat.com>
7502 * tree-vectorizer.h (struct _loop_vec_info): Add no_data_dependencies
7504 (LOOP_VINFO_NO_DATA_DEPENDENCIES): Define.
7505 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Clear it
7506 when not giving up or versioning for alias only because of
7508 (vect_analyze_data_ref_dependences): Set to true.
7509 * tree-vect-stmts.c (hoist_defs_of_uses): Return false if def_stmt
7511 (vectorizable_load): Use LOOP_VINFO_NO_DATA_DEPENDENCIES instead of
7512 LOOP_REQUIRES_VERSIONING_FOR_ALIAS, add && !nested_in_vect_loop
7516 * expr.c (expand_expr_real_1): Handle init == NULL_TREE.
7519 * config/i386/i386.c (ix86_expand_builtin): If target doesn't satisfy
7520 operand 0 predicate for gathers, use a new pseudo as subtarget.
7522 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
7525 * lra-constraints.c (process_alt_operands): Add printing debug info.
7526 Check absence of input/output reloads for matched operands too.
7528 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
7530 PR rtl-optimization/59835
7531 * ira.c (ira_init_register_move_cost): Increase cost for
7534 2014-01-16 Alan Lawrence <alan.lawrence@arm.com>
7536 * config/arm/arm.opt (mcpu, march, mtune): Make case-insensitive.
7538 2014-01-16 Richard Earnshaw <rearnsha@arm.com>
7541 * aarch64.c (aarch64_split_128bit_move): Don't lookup REGNO on
7542 non-register objects. Use gen_(high/low)part more consistently.
7545 2014-01-16 Michael Meissner <meissner@linux.vnet.ibm.com>
7548 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Add little
7549 endian support, remove tests for WORDS_BIG_ENDIAN.
7550 (p8_mfvsrd_3_<mode>): Likewise.
7551 (reload_gpr_from_vsx<mode>): Likewise.
7552 (reload_gpr_from_vsxsf): Likewise.
7553 (p8_mfvsrd_4_disf): Likewise.
7555 2014-01-16 Richard Biener <rguenther@suse.de>
7557 PR rtl-optimization/46590
7558 * lcm.c (compute_antinout_edge): Use postorder iteration.
7559 (compute_laterin): Use inverted postorder iteration.
7561 2014-01-16 Nick Clifton <nickc@redhat.com>
7564 * varasm.c (output_constant): Return the number of bytes actually
7566 (output_constructor_array_range): Update the field size with the
7567 number of bytes emitted by output_constant.
7568 (output_constructor_regular_field): Likewise. Also do not
7569 complain if the total number of bytes emitted is now greater
7570 than the expected fieldpos.
7571 * output.h (output_constant): Update prototype and descriptive comment.
7573 2014-01-16 Marek Polacek <polacek@redhat.com>
7576 * cgraph.c (gimple_check_call_args): Don't use DECL_ARG_TYPE if
7577 it is error_mark_node.
7579 2014-01-15 Uros Bizjak <ubizjak@gmail.com>
7581 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use
7582 VALID_AVX256_REG_OR_OI_MODE.
7584 2014-01-15 Pat Haugen <pthaugen@us.ibm.com>
7586 * config/rs6000/rs6000.c (rs6000_output_function_prologue): Check if
7587 current procedure should be profiled.
7589 2014-01-15 Andrew Pinski <apinski@cavium.com>
7591 * config/aarch64/aarch64.c (aarch64_register_move_cost): Correct cost
7592 of moving from/to the STACK_REG register class.
7594 2014-01-15 Richard Henderson <rth@redhat.com>
7597 * reginfo.c (global_regs_decl): Globalize.
7598 * rtl.h (global_regs_decl): Declare.
7599 * ira.c (do_reload): Diagnose frame_pointer_needed and it
7600 reserved via global_regs.
7602 2014-01-15 Teresa Johnson <tejohnson@google.com>
7604 * tree-ssa-sccvn.c (visit_reference_op_call): Handle NULL vdef.
7606 2014-01-15 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
7608 * config/rs6000/altivec.md (mulv8hi3): Explicitly generate vmulesh
7609 and vmulosh rather than call gen_vec_widen_smult_*.
7610 (vec_widen_umult_even_v16qi): Test VECTOR_ELT_ORDER_BIG rather
7611 than BYTES_BIG_ENDIAN to determine use of even or odd instruction.
7612 (vec_widen_smult_even_v16qi): Likewise.
7613 (vec_widen_umult_even_v8hi): Likewise.
7614 (vec_widen_smult_even_v8hi): Likewise.
7615 (vec_widen_umult_odd_v16qi): Likewise.
7616 (vec_widen_smult_odd_v16qi): Likewise.
7617 (vec_widen_umult_odd_v8hi): Likewise.
7618 (vec_widen_smult_odd_v8hi): Likewise.
7619 (vec_widen_umult_hi_v16qi): Explicitly generate vmuleub and
7620 vmuloub rather than call gen_vec_widen_umult_*.
7621 (vec_widen_umult_lo_v16qi): Likewise.
7622 (vec_widen_smult_hi_v16qi): Explicitly generate vmulesb and
7623 vmulosb rather than call gen_vec_widen_smult_*.
7624 (vec_widen_smult_lo_v16qi): Likewise.
7625 (vec_widen_umult_hi_v8hi): Explicitly generate vmuleuh and vmulouh
7626 rather than call gen_vec_widen_umult_*.
7627 (vec_widen_umult_lo_v8hi): Likewise.
7628 (vec_widen_smult_hi_v8hi): Explicitly gnerate vmulesh and vmulosh
7629 rather than call gen_vec_widen_smult_*.
7630 (vec_widen_smult_lo_v8hi): Likewise.
7632 2014-01-15 Jeff Law <law@redhat.com>
7634 PR tree-optimization/59747
7635 * ree.c (find_and_remove_re): Properly handle case where a second
7636 eliminated extension requires widening a copy created for elimination
7637 of a prior extension.
7638 (combine_set_extension): Ensure that the number of hard regs needed
7639 for a destination register does not change when we widen it.
7641 2014-01-15 Sebastian Huber <sebastian.huber@embedded-brains.de>
7643 * config.gcc (*-*-rtems*): Add t-rtems to tmake_file.
7644 (arm*-*-uclinux*eabi*): Do not override an existing tmake_file.
7645 (arm*-*-eabi* | arm*-*-symbianelf* | arm*-*-rtems*): Likwise.
7646 (arm*-*-rtems*): Use t-rtems from existing tmake_file.
7647 (avr-*-rtems*): Likewise.
7648 (bfin*-rtems*): Likewise.
7649 (moxie-*-rtems*): Likewise.
7650 (h8300-*-rtems*): Likewise.
7651 (i[34567]86-*-rtems*): Likewise.
7652 (lm32-*-rtems*): Likewise.
7653 (m32r-*-rtems*): Likewise.
7654 (m68k-*-rtems*): Likewise.
7655 (microblaze*-*-rtems*): Likewise.
7656 (mips*-*-rtems*): Likewise.
7657 (powerpc-*-rtems*): Likewise.
7658 (sh-*-rtems*): Likewise.
7659 (sparc-*-rtems*): Likewise.
7660 (sparc64-*-rtems*): Likewise.
7661 (v850-*-rtems*): Likewise.
7662 (m32c-*-rtems*): Likewise.
7664 2014-01-15 Vladimir Makarov <vmakarov@redhat.com>
7666 PR rtl-optimization/59511
7667 * ira.c (ira_init_register_move_cost): Use memory costs for some
7668 cases of register move cost calculations.
7669 * lra-constraints.c (lra_constraints): Use REG_FREQ_FROM_BB
7670 instead of BB frequency.
7671 * lra-coalesce.c (move_freq_compare_func, lra_coalesce): Ditto.
7672 * lra-assigns.c (find_hard_regno_for): Ditto.
7674 2014-01-15 Richard Biener <rguenther@suse.de>
7676 PR tree-optimization/59822
7677 * tree-vect-stmts.c (hoist_defs_of_uses): New function.
7678 (vectorizable_load): Use it to hoist defs of uses of invariant
7679 loads out of the loop.
7681 2014-01-15 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
7682 Kugan Vivekanandarajah <kuganv@linaro.org>
7685 * config/aarch64/aarch64.c (aarch64_build_constant): Fix incorrect
7688 2014-01-15 Richard Biener <rguenther@suse.de>
7690 PR rtl-optimization/59802
7691 * lcm.c (compute_available): Use inverted postorder to seed
7692 the initial worklist.
7694 2014-01-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7697 * config/s390/s390.c (s390_preferred_reload_class): Don't return
7698 ADDR_REGS for invalid symrefs in non-PIC code.
7700 2014-01-15 Jakub Jelinek <jakub@redhat.com>
7703 * builtins.c (determine_block_size): Initialize *probable_max_size
7704 even if len_rtx is CONST_INT.
7706 2014-01-14 Andrew Pinski <apinski@cavium.com>
7708 * config/aarch64/aarch64-protos.h (tune_params): Add issue_rate.
7709 * config/aarch64/aarch64.c (generic_tunings): Add issue rate of 2.
7710 (cortexa53_tunings): Likewise.
7711 (aarch64_sched_issue_rate): New function.
7712 (TARGET_SCHED_ISSUE_RATE): Define.
7714 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
7716 * ira-costs.c (find_costs_and_classes): Add missed
7717 ira_init_register_move_cost_if_necessary.
7719 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
7722 * config/arm/arm.c (arm_coproc_mem_operand): Add lra_in_progress.
7724 2014-01-14 H.J. Lu <hongjiu.lu@intel.com>
7727 * config/i386/i386.c (type_natural_mode): Add a bool parameter
7728 to indicate if type is used for function return value. Warn ABI
7729 change if the vector mode isn't available for function return value.
7730 (ix86_function_arg_advance): Pass false to type_natural_mode.
7731 (ix86_function_arg): Likewise.
7732 (ix86_gimplify_va_arg): Likewise.
7733 (function_arg_32): Don't warn ABI change.
7734 (ix86_function_value): Pass true to type_natural_mode.
7735 (ix86_return_in_memory): Likewise.
7736 (ix86_struct_value_rtx): Removed.
7737 (TARGET_STRUCT_VALUE_RTX): Likewise.
7739 2014-01-14 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7741 * jump.c (redirect_jump_2): Remove REG_CROSSING_JUMP notes when
7742 converting a conditional jump into a conditional return.
7744 2014-01-14 Richard Biener <rguenther@suse.de>
7746 PR tree-optimization/58921
7747 PR tree-optimization/59006
7748 * tree-vect-loop-manip.c (vect_loop_versioning): Remove code
7749 hoisting invariant stmts.
7750 * tree-vect-stmts.c (vectorizable_load): Insert the splat of
7751 invariant loads on the preheader edge if possible.
7753 2014-01-14 Joey Ye <joey.ye@arm.com>
7755 * doc/plugin.texi (Building GCC plugins): Update to C++.
7757 2014-01-14 Kirill Yukhin <kirill.yukhin@intel.com>
7759 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): New.
7760 (_mm_rcp28_round_ss): Ditto.
7761 (_mm_rsqrt28_round_sd): Ditto.
7762 (_mm_rsqrt28_round_ss): Ditto.
7763 (_mm_rcp28_sd): Ditto.
7764 (_mm_rcp28_ss): Ditto.
7765 (_mm_rsqrt28_sd): Ditto.
7766 (_mm_rsqrt28_ss): Ditto.
7767 * config/i386/avx512fintrin.h (_mm512_stream_load_si512): Ditto.
7768 * config/i386/i386-builtin-types.def (V8DI_FTYPE_PV8DI): Ditto.
7769 * config/i386/i386.c (IX86_BUILTIN_MOVNTDQA512): Ditto.
7770 (IX86_BUILTIN_RCP28SD): Ditto.
7771 (IX86_BUILTIN_RCP28SS): Ditto.
7772 (IX86_BUILTIN_RSQRT28SD): Ditto.
7773 (IX86_BUILTIN_RSQRT28SS): Ditto.
7774 (bdesc_special_args): Define __builtin_ia32_movntdqa512,
7775 __builtin_ia32_rcp28sd_round, __builtin_ia32_rcp28ss_round,
7776 __builtin_ia32_rsqrt28sd_round, __builtin_ia32_rsqrt28ss_round.
7777 (ix86_expand_special_args_builtin): Expand new FTYPE.
7778 * config/i386/sse.md (define_mode_attr "sse4_1_avx2"): Expand to V8DI.
7779 (srcp14<mode>): Make insn unary.
7780 (avx512f_vmscalef<mode><round_name>): Use substed predicate.
7781 (avx512f_sgetexp<mode><round_saeonly_name>): Ditto.
7782 (avx512f_rndscale<mode><round_saeonly_name>): Ditto.
7783 (<sse4_1_avx2>_movntdqa): Extend to 512 bits.
7784 (avx512er_exp2<mode><mask_name><round_saeonly_name>):
7785 Fix rounding: make it SAE only.
7786 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
7788 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
7790 (avx512er_vmrcp28<mode><round_saeonly_name>): Ditto.
7791 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Ditto.
7792 (avx512f_getmant<mode><mask_name><round_saeonly_name>): Ditto.
7793 * config/i386/subst.md (round_saeonly_mask_scalar_operand3): Remove.
7794 (round_saeonly_mask_scalar_operand4): Ditto.
7795 (round_saeonly_mask_scalar_op3): Ditto.
7796 (round_saeonly_mask_scalar_op4): Ditto.
7798 2014-01-13 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7800 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
7801 Implement -maltivec=be for vec_insert and vec_extract.
7803 2014-01-10 DJ Delorie <dj@redhat.com>
7805 * config/msp430/msp430.md (call_internal): Don't allow memory
7806 references with SP as the base register.
7807 (call_value_internal): Likewise.
7808 * config/msp430/constraints.md (Yc): New. For memory references
7809 that don't use SP as a base register.
7811 * config/msp430/msp430.c (msp430_print_operand): Add 'J' to mean
7812 "an integer without a # prefix"
7813 * config/msp430/msp430.md (epilogue_helper): Use it.
7815 2014-01-13 Jakub Jelinek <jakub@redhat.com>
7818 * config/i386/i386.c (ix86_vectorize_builtin_gather): Uncomment
7819 AVX512F gather builtins.
7820 * tree-vect-stmts.c (vectorizable_mask_load_store): For now punt
7821 on gather decls with INTEGER_TYPE masktype.
7822 (vectorizable_load): For INTEGER_TYPE masktype, put the INTEGER_CST
7823 directly into the builtin rather than hoisting it before loop.
7825 PR tree-optimization/59387
7826 * tree-scalar-evolution.c: Include gimple-fold.h and gimplify-me.h.
7827 (scev_const_prop): If folded_casts and type has undefined overflow,
7828 use force_gimple_operand instead of force_gimple_operand_gsi and
7829 for each added stmt if it is assign with
7830 arith_code_with_undefined_signed_overflow, call
7831 rewrite_to_defined_overflow.
7832 * tree-ssa-loop-im.c: Don't include gimplify-me.h, include
7833 gimple-fold.h instead.
7834 (arith_code_with_undefined_signed_overflow,
7835 rewrite_to_defined_overflow): Moved to ...
7836 * gimple-fold.c (arith_code_with_undefined_signed_overflow,
7837 rewrite_to_defined_overflow): ... here. No longer static.
7838 Include gimplify-me.h.
7839 * gimple-fold.h (arith_code_with_undefined_signed_overflow,
7840 rewrite_to_defined_overflow): New prototypes.
7842 2014-01-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7844 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Fix typo in description.
7846 2014-01-13 Eric Botcazou <ebotcazou@adacore.com>
7848 * builtins.c (get_object_alignment_2): Minor tweak.
7849 * tree-ssa-loop-ivopts.c (may_be_unaligned_p): Rewrite.
7851 2014-01-13 Christian Bruel <christian.bruel@st.com>
7853 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Unroll small sizes and
7854 optimized non constant lengths.
7856 2014-01-13 Jakub Jelinek <jakub@redhat.com>
7859 * omp-low.c (expand_omp_atomic_pipeline): Expand the initial
7860 load as __atomic_load_N if possible.
7862 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
7864 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): Remove
7866 (rs6000_expand_builtin): Adjust call.
7868 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
7871 * config/rs6000/rs6000.h (SWITCHABLE_TARGET): Define.
7872 * config/rs6000/rs6000.c: Include target-globals.h.
7873 (rs6000_set_current_function): Instead of doing target_reinit
7874 unconditionally, use save_target_globals_default_opts and
7875 restore_target_globals.
7877 * config/rs6000/rs6000-builtin.def (mffs, mtfsf): Add builtins for
7879 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): New.
7880 (rs6000_expand_builtin): Handle mffs and mtfsf.
7881 (rs6000_init_builtins): Define mffs and mtfsf.
7882 * config/rs6000/rs6000.md (UNSPECV_MFFS, UNSPECV_MTFSF): New constants.
7883 (rs6000_mffs): New pattern.
7884 (rs6000_mtfsf): New pattern.
7886 2014-01-11 Bin Cheng <bin.cheng@arm.com>
7888 * tree-ssa-loop-ivopts.c (iv_ca_narrow): New parameter.
7889 Start narrowing with START. Apply candidate-use pair
7890 and check overall cost in narrowing.
7891 (iv_ca_prune): Pass new argument.
7893 2014-01-10 Jeff Law <law@redhat.com>
7896 * ree.c (combine_reaching_defs): Ensure the defining statement
7897 occurs before the extension when optimizing extensions with
7898 different source and destination hard registers.
7900 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
7903 * ipa-devirt.c (build_type_inheritance_graph): Also add types of
7904 vtables into the type inheritance graph.
7906 2014-01-10 Jakub Jelinek <jakub@redhat.com>
7908 PR rtl-optimization/59754
7909 * ree.c (combine_reaching_defs): Disallow !SCALAR_INT_MODE_P
7910 modes in the REGNO != REGNO case.
7912 2014-01-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7914 * config/rs6000/rs6000-builtin.def: Fix pasto for VPKSDUS.
7916 2014-01-10 Jakub Jelinek <jakub@redhat.com>
7918 PR tree-optimization/59745
7919 * tree-predcom.c (tree_predictive_commoning_loop): Call
7920 free_affine_expand_cache if giving up because components is NULL.
7922 * target-globals.c (save_target_globals): Allocate < 4KB structs using
7923 GC in payload of target_globals struct instead of allocating them on
7924 the heap and the larger structs separately using GC.
7925 * target-globals.h (struct target_globals): Make regs, hard_regs,
7926 reload, expmed, ira, ira_int and lra_fields GTY((atomic)) instead
7927 of GTY((skip)) and change type to void *.
7928 (reset_target_globals): Cast loads from those fields to corresponding
7931 2014-01-10 Steve Ellcey <sellcey@mips.com>
7934 * Makefile.in (PLUGIN_HEADERS): Add gimplify.h, gimple-iterator.h,
7935 gimple-ssa.h, fold-const.h, tree-cfg.h, tree-into-ssa.h,
7936 tree-ssanames.h, print-tree.h, varasm.h, and context.h.
7938 2014-01-10 Richard Earnshaw <rearnsha@arm.com>
7941 * aarch64-modes.def (CC_Zmode): New flags mode.
7942 * aarch64.c (aarch64_select_cc_mode): Only allow NEG when the condition
7943 represents an equality.
7944 (aarch64_get_condition_code): Handle CC_Zmode.
7945 * aarch64.md (compare_neg<mode>): Restrict to equality operations.
7947 2014-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7949 * config/s390/s390.c (s390_expand_tbegin): Remove jump over CC
7950 extraction in good case.
7952 2014-01-10 Richard Biener <rguenther@suse.de>
7954 PR tree-optimization/59374
7955 * tree-vect-slp.c (vect_slp_analyze_bb_1): Move dependence
7956 checking after SLP discovery. Mark stmts not participating
7957 in any SLP instance properly.
7959 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7961 * config/arm/arm.c (arm_new_rtx_costs): Use destination mode
7962 when handling a SET rtx.
7964 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7966 * config/arm/arm-cores.def (cortex-a53): Specify FL_CRC32.
7967 (cortex-a57): Likewise.
7968 (cortex-a57.cortex-a53): Likewise. Remove redundant flags.
7970 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7972 * config/arm/arm.c (arm_init_iwmmxt_builtins): Skip
7973 non-iwmmxt builtins.
7975 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
7979 * ipa-devirt.c record_target_from_binfo): Take as argument
7980 stack of binfos and lookup matching one for virtual inheritance.
7981 (possible_polymorphic_call_targets_1): Update.
7983 2014-01-10 Huacai Chen <chenhc@lemote.com>
7985 * config/mips/driver-native.c (host_detect_local_cpu): Handle new
7986 kernel strings for Loongson-2E/2F/3A.
7988 2014-01-10 Jakub Jelinek <jakub@redhat.com>
7991 * tree-vect-data-refs.c (vect_analyze_data_refs): Check
7992 is_gimple_call before calling gimple_call_internal_p.
7994 2014-01-09 Steve Ellcey <sellcey@mips.com>
7996 * Makefile.in (TREE_FLOW_H): Remove.
7997 (TREE_SSA_H): Add file names from tree-flow.h.
7998 * doc/tree-ssa.texi (Annotations): Remove reference to tree-flow.h
7999 * tree.h: Remove tree-flow.h reference.
8000 * hash-table.h: Remove tree-flow.h reference.
8001 * tree-ssa-loop-niter.c (dump_affine_iv): Replace tree-flow.h
8002 reference with tree-ssa-loop.h.
8004 2014-01-09 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8006 * doc/invoke.texi: Add -maltivec={be,le} options, and document
8007 default element-order behavior for -maltivec.
8008 * config/rs6000/rs6000.opt: Add -maltivec={be,le} options.
8009 * config/rs6000/rs6000.c (rs6000_option_override_internal): Ensure
8010 that -maltivec={le,be} implies -maltivec; disallow -maltivec=le
8011 when targeting big endian, at least for now.
8012 * config/rs6000/rs6000.h: Add #define of VECTOR_ELT_ORDER_BIG.
8014 2014-01-09 Jakub Jelinek <jakub@redhat.com>
8017 * cfgexpand.c (expand_one_var): For SSA_NAMEs, if the underlying
8018 var satisfies use_register_for_decl, just take into account type
8019 alignment, rather than decl alignment.
8021 PR tree-optimization/59622
8022 * gimple-fold.c (gimple_fold_call): Fix a typo in message. For
8023 __builtin_unreachable replace the OBJ_TYPE_REF call with a call to
8024 __builtin_unreachable and add if needed a setter of the lhs SSA_NAME.
8025 Don't devirtualize for inplace at all. For targets.length () == 1,
8026 if the call is noreturn and cfun isn't in SSA form yet, clear lhs.
8028 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
8030 * config/i386/i386.md (cpu): Remove the unused btver1.
8032 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
8034 * gdbasan.in: Put a breakpoint on __sanitizer::Report.
8036 2014-01-09 Jakub Jelinek <jakub@redhat.com>
8039 * tree-core.h (struct target_globals): New forward declaration.
8040 (struct tree_target_option): Add globals field.
8041 * tree.h (TREE_TARGET_GLOBALS): Define.
8042 (prepare_target_option_nodes_for_pch): New prototype.
8043 * target-globals.h (struct target_globals): Define even if
8045 * tree.c (prepare_target_option_node_for_pch,
8046 prepare_target_option_nodes_for_pch): New functions.
8047 * config/i386/i386.h (SWITCHABLE_TARGET): Define.
8048 * config/i386/i386.c: Include target-globals.h.
8049 (ix86_set_current_function): Instead of doing target_reinit
8050 unconditionally, use save_target_globals_default_opts and
8051 restore_target_globals.
8053 2014-01-09 Richard Biener <rguenther@suse.de>
8055 PR tree-optimization/59715
8056 * tree-cfg.h (split_critical_edges): Declare.
8057 * tree-cfg.c (split_critical_edges): Export.
8058 * tree-ssa-sink.c (execute_sink_code): Split critical edges.
8060 2014-01-09 Max Ostapenko <m.ostapenko@partner.samsung.com>
8062 * cfgexpand.c (expand_stack_vars): Optionally disable
8063 asan stack protection.
8064 (expand_used_vars): Likewise.
8065 (partition_stack_vars): Likewise.
8066 * asan.c (asan_emit_stack_protection): Optionally disable
8067 after return stack usage.
8068 (instrument_derefs): Optionally disable memory access instrumentation.
8069 (instrument_builtin_call): Likewise.
8070 (instrument_strlen_call): Likewise.
8071 (asan_protect_global): Optionally disable global variables protection.
8072 * doc/invoke.texi: Added doc for new options.
8073 * params.def: Added new options.
8074 * params.h: Likewise.
8076 2014-01-09 Jakub Jelinek <jakub@redhat.com>
8078 PR rtl-optimization/59724
8079 * ifcvt.c (cond_exec_process_if_block): Don't call
8080 flow_find_head_matching_sequence with 0 longest_match.
8081 * cfgcleanup.c (flow_find_head_matching_sequence): Count even
8082 non-active insns if !stop_after.
8083 (try_head_merge_bb): Revert 2014-01-07 changes.
8085 2014-01-08 Jeff Law <law@redhat.com>
8087 * ree.c (get_sub_rtx): New function, extracted from...
8088 (merge_def_and_ext): Here.
8089 (combine_reaching_defs): Use get_sub_rtx.
8091 2014-01-08 Eric Botcazou <ebotcazou@adacore.com>
8093 * cgraph.h (varpool_variable_node): Do not choke on null node.
8095 2014-01-08 Catherine Moore <clm@codesourcery.com>
8097 * config/mips/mips.md (simple_return): Attempt to use JRC
8099 * config/mips/mips.h (MIPS_CALL): Attempt to use JALS for microMIPS.
8101 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
8103 PR rtl-optimization/59137
8104 * reorg.c (steal_delay_list_from_target): Call update_block for
8106 (steal_delay_list_from_fallthrough, relax_delay_slots): Likewise.
8108 2014-01-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8110 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Remove
8111 two duplicate entries.
8113 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
8116 2012-10-07 Richard Sandiford <rdsandiford@googlemail.com>
8118 * config/mips/mips.c (mips_truncated_op_cost): New function.
8119 (mips_rtx_costs): Adjust test for BADDU.
8120 * config/mips/mips.md (*baddu_di<mode>): Push truncates to operands.
8122 2012-10-02 Richard Sandiford <rdsandiford@googlemail.com>
8124 * config/mips/mips.md (*baddu_si_eb, *baddu_si_el): Merge into...
8125 (*baddu_si): ...this new pattern.
8127 2014-01-08 Jakub Jelinek <jakub@redhat.com>
8130 * ipa-prop.c (ipa_analyze_params_uses): Ignore uses in debug stmts.
8132 2014-01-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
8135 * expr.h (expand_expr_real, expand_expr_real_1): Add new parameter
8137 (expand_expr, expand_normal): Adjust.
8138 * expr.c (expand_expr_real, expand_expr_real_1): Add new parameter
8139 inner_reference_p. Use inner_reference_p to expand inner references.
8140 (store_expr): Adjust.
8141 * cfgexpand.c (expand_call_stmt): Adjust.
8143 2014-01-08 Rong Xu <xur@google.com>
8145 * gcov-io.c (gcov_var): Move from gcov-io.h.
8146 (gcov_position): Ditto.
8147 (gcov_is_error): Ditto.
8148 (gcov_rewrite): Ditto.
8149 * gcov-io.h: Refactor. Move gcov_var to gcov-io.h, and libgcov
8150 only part to libgcc/libgcov.h.
8152 2014-01-08 Marek Polacek <polacek@redhat.com>
8155 * omp-low.c (simd_clone_adjust): Don't crash if def is NULL.
8157 2014-01-08 Marek Polacek <polacek@redhat.com>
8160 * ubsan.c (ubsan_type_descriptor): Call strip_array_types on type2.
8162 2014-01-08 Jakub Jelinek <jakub@redhat.com>
8164 PR rtl-optimization/59649
8165 * stor-layout.c (get_mode_bounds): For BImode return
8166 0 and STORE_FLAG_VALUE.
8168 2014-01-08 Richard Biener <rguenther@suse.de>
8171 * gimple.h (is_gimple_builtin_call): Remove.
8172 (gimple_builtin_call_types_compatible_p): New.
8173 (gimple_call_builtin_p): New overload.
8174 * gimple.c (is_gimple_builtin_call): Remove.
8175 (validate_call): Rename to ...
8176 (gimple_builtin_call_types_compatible_p): ... this and export. Also
8178 (validate_type): New static function.
8179 (gimple_call_builtin_p): New overload and adjust.
8180 * gimple-fold.c (gimple_fold_builtin): Fold the return value.
8181 (gimple_fold_call): Likewise. Use gimple_call_builtin_p.
8182 (gimple_fold_stmt_to_constant_1): Likewise.
8183 * tsan.c (instrument_gimple): Use gimple_call_builtin_p.
8185 2014-01-08 Richard Biener <rguenther@suse.de>
8188 * gimplify.c (gimplify_expr): Gimplify register-register type
8189 VIEW_CONVERT_EXPRs to separate stmts.
8191 2014-01-07 Jeff Law <law@redhat.com>
8194 * ree.c (combine_set_extension): Handle case where source
8195 and destination registers in an extension insn are different.
8196 (combine_reaching_defs): Allow source and destination registers
8197 in extension to be different under limited circumstances.
8198 (add_removable_extension): Remove restriction that the
8199 source and destination registers in the extension are the same.
8200 (find_and_remove_re): Emit a copy from the extension's
8201 destination to its source after the defining insn if
8202 the source and destination registers are different.
8205 * ifcvt.c (merge_if_block): If we are merging a block with more than
8206 one successor with a block with no successors, remove any BARRIER
8207 after the second block.
8209 2014-01-07 Dan Xio Qiang <ziyan01@163.com>
8211 * hw-doloop.c (reorg_loops): Release the bitmap obstack.
8213 2014-01-07 John David Anglin <danglin@gcc.gnu.org>
8216 * config/pa/pa.c (pa_legitimate_address_p): Return false before reload
8217 for 14-bit register offsets when INT14_OK_STRICT is false.
8219 2014-01-07 Roland Stigge <stigge@antcom.de>
8220 Michael Meissner <meissner@linux.vnet.ibm.com>
8223 * config/rs6000/rs6000.c (rs6000_legitimate_offset_address_p):
8224 Only check TFmode for SPE constants. Don't check TImode or TDmode.
8226 2014-01-07 James Greenhalgh <james.greenhalgh@arm.com>
8228 * config/aarch64/aarch64-elf.h (ASM_SPEC): Remove identity spec for
8231 2014-01-07 Yufeng Zhang <yufeng.zhang@arm.com>
8233 * config/arm/arm.c (arm_expand_neon_args): Call expand_expr
8234 with EXPAND_MEMORY for NEON_ARG_MEMORY; check if the returned
8235 rtx is const0_rtx or not.
8237 2014-01-07 Richard Sandiford <rdsandiford@googlemail.com>
8240 * target-globals.c (save_target_globals): Remove this_fn_optab
8242 * toplev.c: Include optabs.h.
8243 (target_reinit): Temporarily restore the global options if another
8244 set of options are in force.
8246 2014-01-07 Jakub Jelinek <jakub@redhat.com>
8248 PR rtl-optimization/58668
8249 * cfgcleanup.c (flow_find_cross_jump): Don't count
8250 any jumps if dir_p is NULL. Remove p1 variable, use active_insn_p
8251 to determine what is counted.
8252 (flow_find_head_matching_sequence): Use active_insn_p to determine
8254 (try_head_merge_bb): Adjust for the flow_find_head_matching_sequence
8256 * ifcvt.c (count_bb_insns): Use active_insn_p && !JUMP_P to
8257 determine what is counted.
8259 PR tree-optimization/59643
8260 * tree-predcom.c (split_data_refs_to_components): If one dr is
8261 read and one write, determine_offset fails and the write isn't
8262 in the bad component, just put the read into the bad component.
8264 2014-01-07 Mike Stump <mikestump@comcast.net>
8265 Jakub Jelinek <jakub@redhat.com>
8268 * tree-core.h (struct tree_optimization_option): Change optabs
8269 type from unsigned char * to void *.
8270 * optabs.c (init_tree_optimization_optabs): Adjust
8271 TREE_OPTIMIZATION_OPTABS initialization.
8273 2014-01-06 Jakub Jelinek <jakub@redhat.com>
8276 * config/i386/i386.h (struct machine_function): Add
8277 no_drap_save_restore field.
8278 * config/i386/i386.c (ix86_save_reg): Use
8279 !cfun->machine->no_drap_save_restore instead of
8280 crtl->stack_realign_needed.
8281 (ix86_finalize_stack_realign_flags): Don't clear drap_reg unless
8282 this function clears frame_pointer_needed. Set
8283 cfun->machine->no_drap_save_restore if clearing frame_pointer_needed
8284 and DRAP reg is needed.
8286 2014-01-06 Marek Polacek <polacek@redhat.com>
8289 * doc/implement-c.texi: Mention that other integer types are
8290 permitted as bit-field types in strictly conforming mode.
8292 2014-01-06 Felix Yang <fei.yang0953@gmail.com>
8294 * modulo-sched.c (schedule_reg_moves): Clear distance1_uses if it
8297 2014-01-06 Richard Earnshaw <rearnsha@arm.com>
8299 * aarch64.c (aarch64_rtx_costs): Fix cost calculation for MADD.
8301 2014-01-06 Martin Jambor <mjambor@suse.cz>
8304 * ipa-cp.c (ipcp_discover_new_direct_edges): Changed param_index type
8306 * ipa-prop.c (ipa_print_node_params): Fix indentation.
8308 2014-01-06 Eric Botcazou <ebotcazou@adacore.com>
8312 * var-tracking.c (add_stores): Preserve the value of the source even if
8313 we don't record the store.
8315 2014-01-06 Terry Guo <terry.guo@arm.com>
8317 * config.gcc (arm*-*-*): Check --with-arch against arm-arches.def.
8319 2014-01-05 Iain Sandoe <iain@codesourcery.com>
8322 * config/darwin.c (darwin_function_section): Adjust return values to
8323 correspond to optimisation changes made in r206070.
8325 2014-01-05 Uros Bizjak <ubizjak@gmail.com>
8327 * config/i386/i386.c (ix86_data_alignment): Calculate max_align
8328 from prefetch_block tune setting.
8329 (nocona_cost): Correct size of prefetch block to 64.
8331 2014-01-04 Eric Botcazou <ebotcazou@adacore.com>
8333 * config/arm/arm.c (arm_get_frame_offsets): Revamp long lines.
8334 (arm_expand_epilogue_apcs_frame): Take into account the number of bytes
8335 used to save the static chain register in the computation of the offset
8336 from which the FP registers need to be restored.
8338 2014-01-04 Jakub Jelinek <jakub@redhat.com>
8340 PR tree-optimization/59519
8341 * tree-vect-loop-manip.c (slpeel_update_phi_nodes_for_guard1): Don't
8342 ICE if get_current_def (current_new_name) is already non-NULL, as long
8343 as it is a phi result of some other phi in *new_exit_bb that has
8346 * config/i386/sse.md (avx512f_load<mode>_mask): Emit vmovup{s,d}
8347 or vmovdqu* for misaligned_operand.
8348 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>,
8349 <sse2_avx_avx512f>_loaddqu<mode><mask_name>): Handle <mask_applied>.
8350 * config/i386/i386.c (ix86_expand_special_args_builtin): Set
8351 aligned_mem for AVX512F masked aligned load and store builtins and for
8354 2014-01-03 Bingfeng Mei <bmei@broadcom.com>
8356 PR tree-optimization/59651
8357 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
8358 Address range for negative step should be added by TYPE_SIZE_UNIT.
8360 2014-01-03 Andreas Schwab <schwab@linux-m68k.org>
8362 * config/m68k/m68k.c (handle_move_double): Handle pushes with
8363 overlapping registers also for registers other than the stack pointer.
8365 2014-01-03 Marek Polacek <polacek@redhat.com>
8368 * doc/extend.texi: Fix the return value of __builtin_FUNCTION and
8371 2014-01-03 Jakub Jelinek <jakub@redhat.com>
8374 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Don't consider
8377 * config/i386/i386.md (MODE_SIZE): New mode attribute.
8378 (push splitter): Use <P:MODE_SIZE> instead of
8379 GET_MODE_SIZE (<P:MODE>mode).
8380 (lea splitter): Use <MODE_SIZE> instead of GET_MODE_SIZE (<MODE>mode).
8381 (mov -1, reg peephole2): Likewise.
8382 * config/i386/sse.md (*mov<mode>_internal,
8383 <sse>_storeu<ssemodesuffix><avxsizesuffix>,
8384 <sse2_avx_avx512f>_storedqu<mode>, <sse>_andnot<mode>3,
8385 *<code><mode>3, *andnot<mode>3<mask_name>,
8386 <mask_codefor><code><mode>3<mask_name>): Likewise.
8387 * config/i386/subst.md (mask_mode512bit_condition,
8388 sd_mask_mode512bit_condition): Likewise.
8390 2014-01-02 Xinliang David Li <davidxl@google.com>
8392 PR tree-optimization/59303
8393 * tree-ssa-uninit.c (is_use_properly_guarded): Main cleanup.
8394 (dump_predicates): Better output format.
8395 (pred_equal_p): New function.
8396 (is_neq_relop_p): Ditto.
8397 (is_neq_zero_form_p): Ditto.
8398 (pred_expr_equal_p): Ditto.
8399 (pred_neg_p): Ditto.
8400 (simplify_pred): Ditto.
8401 (simplify_preds_2): Ditto.
8402 (simplify_preds_3): Ditto.
8403 (simplify_preds_4): Ditto.
8404 (simplify_preds): Ditto.
8406 (push_to_worklist): Ditto.
8407 (get_pred_info_from_cmp): Ditto.
8408 (is_degenerated_phi): Ditto.
8409 (normalize_one_pred_1): Ditto.
8410 (normalize_one_pred): Ditto.
8411 (normalize_one_pred_chain): Ditto.
8412 (normalize_preds): Ditto.
8413 (normalize_cond_1): Remove function.
8414 (normalize_cond): Ditto.
8415 (is_gcond_subset_of): Ditto.
8416 (is_subset_of_any): Ditto.
8417 (is_or_set_subset_of): Ditto.
8418 (is_and_set_subset_of): Ditto.
8419 (is_norm_cond_subset_of): Ditto.
8420 (pred_chain_length_cmp): Ditto.
8421 (convert_control_dep_chain_into_preds): Type change.
8422 (find_predicates): Ditto.
8423 (find_def_preds): Ditto.
8424 (destroy_predicates_vecs): Ditto.
8425 (find_matching_predicates_in_rest_chains): Ditto.
8426 (use_pred_not_overlap_with_undef_path_pred): Ditto.
8427 (is_pred_expr_subset): Ditto.
8428 (is_pred_chain_subset_of): Ditto.
8429 (is_included_in): Ditto.
8430 (is_superset_of): Ditto.
8432 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
8434 Update copyright years.
8436 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
8438 * common/config/arc/arc-common.c, config/arc/arc-modes.def,
8439 config/arc/arc-protos.h, config/arc/arc.c, config/arc/arc.h,
8440 config/arc/arc.md, config/arc/arc.opt,
8441 config/arm/arm_neon_builtins.def, config/arm/crypto.def,
8442 config/i386/avx512cdintrin.h, config/i386/avx512erintrin.h,
8443 config/i386/avx512fintrin.h, config/i386/avx512pfintrin.h,
8444 config/i386/btver2.md, config/i386/shaintrin.h, config/i386/slm.md,
8445 config/linux-protos.h, config/linux.c, config/winnt-c.c,
8446 diagnostic-color.c, diagnostic-color.h, gimple-ssa-isolate-paths.c,
8447 vtable-verify.c, vtable-verify.h: Use the standard form for the
8450 2014-01-02 Tobias Burnus <burnus@net-b.de>
8452 * gcc.c (process_command): Update copyright notice dates.
8453 * gcov-dump.c: Ditto.
8455 * doc/cpp.texi: Bump @copying's copyright year.
8456 * doc/cppinternals.texi: Ditto.
8457 * doc/gcc.texi: Ditto.
8458 * doc/gccint.texi: Ditto.
8459 * doc/gcov.texi: Ditto.
8460 * doc/install.texi: Ditto.
8461 * doc/invoke.texi: Ditto.
8463 2014-01-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8465 * config/nios2/nios2.h (BITS_PER_UNIT): Don't define it.
8467 2014-01-01 Jakub Jelinek <jakub@redhat.com>
8469 * config/i386/sse.md (*mov<mode>_internal): Guard
8470 EXT_REX_SSE_REGNO_P (REGNO ()) uses with REG_P.
8472 PR rtl-optimization/59647
8473 * cse.c (cse_process_notes_1): Don't substitute negative VOIDmode
8474 new_rtx into UNSIGNED_FLOAT rtxes.
8476 Copyright (C) 2014 Free Software Foundation, Inc.
8478 Copying and distribution of this file, with or without modification,
8479 are permitted in any medium without royalty provided the copyright
8480 notice and this notice are preserved.