1 2013-10-29 Eric Botcazou <ebotcazou@adacore.com>
3 * expr.c (expand_expr_real_1) <MEM_EXPR>: Eliminate small redundancy.
5 2013-10-29 David Malcolm <dmalcolm@redhat.com>
7 * doc/gty.texi ("Inheritance and GTY"): Make it clear that
8 to use autogenerated markers for a class-hierarchy, every class
9 must have a GTY marker.
10 * gengtype.h (struct type): Add linked list of subclasses to
11 the "s" member of the union.
12 (add_subclass): New decl.
13 * gengtype-state.c (read_state_struct_type): Set up subclass
15 * gengtype.c (get_ultimate_base_class): New.
17 (new_structure): Set up subclass linked list.
18 (set_gc_used_type): Propagate usage information to subclasses.
19 (output_mangled_typename): Use get_ultimate_base_class.
20 (walk_subclasses): Use the subclass linked list, avoiding an
21 O(N^2) when writing out all types.
22 (walk_type): Issue an error if the base class is missing a tag,
23 rather than generating bogus C code. Add a gcc_unreachable
24 default case, in case people omit tags from concrete subclasses,
25 or get the values wrong.
26 (write_func_for_structure): Issue an error for subclasses for
27 which the base doesn't have a "desc", since otherwise the
28 autogenerated routines for the base would silently fail to visit
30 (write_root): Use get_ultimate_base_class, tweaking constness of
31 tp to match that function's signature.
33 2013-10-29 David Malcolm <dmalcolm@redhat.com>
35 * doc/gty.texi (GTY Options): Add note about inheritance to
36 description of desc and tag.
37 (Inheritance and GTY): New.
39 2013-10-29 David Malcolm <dmalcolm@redhat.com>
41 * gengtype-parse.c (opts_have): Drop "static" so that
42 we can use this from gengtype.c.
43 * gengtype.c (set_gc_used_type): Mark any base class as used;
44 update field traversal to visit inherited fields.
45 (output_mangled_typename): Convert references to classes within
46 an inheritance hierarchy to reference the ultimate base class,
47 since only it will have gt_ functions.
48 (get_string_option): New.
49 (walk_subclasses): New.
50 (walk_type): Treat GTY structs that have a "desc" as being the
51 root of an inheritance hierarchy. Generate a switch on it
52 within the marking function which walks all subclasses, adding
53 cases for them via walk_subclasses. For subclasses, visit all
54 fields of the type (including inherited ones).
55 (write_func_for_structure): Don't write fns for subclasses, only
56 for the ultimate base class within an inheritance hierarchy.
57 Subclasses-marking will be handled by the base class marking
59 (write_types): Likewise.
60 (write_local_func_for_structure): Likewise.
61 (USED_BY_TYPED_GC_P): Emit allocators for subclasses that have
62 a "tag" option (and are thus concrete subclasses).
63 (write_root): Use the marker function for the ultimate base class.
64 * gengtype.h (FOR_ALL_INHERITED_FIELDS): New.
65 (opts_have): Add declaration.
67 2013-10-28 Vladimir Makarov <vmakarov@redhat.com>
69 * lra-spills.c (lra_final_code_change): Remove useless move insns
70 originated from moves of pseudos.
72 2013-10-28 Jeff Law <law@redhat.com>
74 * lower-subreg.c (resolve_simple_move): Fix comment typo.
76 2013-10-28 Trevor Saunders <tsaunders@mozilla.com>
78 * df-scan.c (df_collection_rec): Adjust.
79 (copy_defs): New constant.
80 (copy_uses): Likewise.
81 (copy_eq_uses): Likewise.
84 (df_insn_rescan): Adjust.
85 (df_notes_rescan): Likewise.
86 (df_swap_refs): Likewise.
87 (df_sort_and_compress_refs): Likewise.
88 (df_sort_and_compress_mws): Likewise.
89 (df_install_refs): Likewise.
90 (df_install_mws): Likewise.
91 (df_refs_add_to_chains): Add flags parameter controlling which vectors
93 (df_bb_refs_record): Adjust.
94 (df_record_entry_block_defs): Likewise.
95 (df_record_exit_block_defs): Likewise.
96 (df_refs_verify): Likewise.
97 (df_mws_verify): Likewise.
98 (df_insn_refs_verify): Likewise.
99 (df_bb_verify): Likewise.
100 * ipa-pure-const.c (finish_state): Remove.
102 * tree-data-ref.c tree-ssa-alias.c tree-ssa-loop-ivcanon.c
103 tree-ssa-threadedge.c tree-vect-loop-manip.c tree-vect-slp.c
104 var-tracking.c: Adjust.
105 * vec.c (stack_vecs): Remove.
106 (register_stack_vec): Likewise.
107 (stack_vec_register_index): Likewise.
108 (unregister_stack_vec): Likewise.
109 * vec.h (struct va_stack): Remove.
110 (struct vec<T, A, vl_ptr>): Specialize as
111 struct vec<T, va_heap, vl_ptr> instead since va_heap is the only
112 allocation strategy compatable with the vl_ptr layout.
113 (struct vec<T, va_gc, vl_ptr>): Remove because it now gets an empty
114 specialization anyway.
115 (class stack_vec): New class.
116 (vec_stack_alloc): Remove.
117 (vec<T, va_heap, vl_ptr>::using_auto_storage): New method.
119 2013-10-28 Alexander Ivchenko <alexander.ivchenko@intel.com>
120 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
121 Sergey Lega <sergey.s.lega@intel.com>
122 Anna Tikhonova <anna.tikhonova@intel.com>
123 Ilya Tocar <ilya.tocar@intel.com>
124 Andrey Turetskiy <andrey.turetskiy@intel.com>
125 Ilya Verbin <ilya.verbin@intel.com>
126 Kirill Yukhin <kirill.yukhin@intel.com>
127 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
129 * config/i386/i386.md (prefetch): Allow TARGET_AVX512PF.
130 (*prefetch_avx512pf_<mode>): New.
131 * config/i386/sse.md (avx512f_vmcmp<mode>3): Ditto.
132 (avx512f_maskcmp<mode>3): Ditto.
133 (vashrv16si3): Ditto.
135 2013-10-28 Alexander Ivchenko <alexander.ivchenko@intel.com>
136 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
137 Sergey Lega <sergey.s.lega@intel.com>
138 Anna Tikhonova <anna.tikhonova@intel.com>
139 Ilya Tocar <ilya.tocar@intel.com>
140 Andrey Turetskiy <andrey.turetskiy@intel.com>
141 Ilya Verbin <ilya.verbin@intel.com>
142 Kirill Yukhin <kirill.yukhin@intel.com>
143 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
145 * config/i386/i386.md (any_truncate): New.
147 * config/i386/predicates.md (const_8_to_9_operand): New.
148 (const_10_to_11_operand): Ditto.
149 (const_12_to_13_operand): Ditto.
150 (const_14_to_15_operand): Ditto.
151 (const_16_to_19_operand): Ditto.
152 (const_20_to_23_operand): Ditto.
153 (const_24_to_27_operand): Ditto.
154 (const_28_to_31_operand): Ditto.
155 * config/i386/sse.md (unspec): Add UNSPEC_UNSIGNED_FIX_NOTRUNC.
156 (cvtusi2<ssescalarmodesuffix>32): New.
157 (cvtusi2<ssescalarmodesuffix>64): Ditto.
158 (ufloatv16siv16sf2): Ditto.
159 (avx512f_fix_notruncv16sfv16si): Ditto.
160 (avx512f_ufix_notruncv16sfv16si): Ditto.
161 (avx512f_vcvtss2usi): Ditto.
162 (avx512f_vcvtss2usiq): Ditto.
163 (avx512f_vcvttss2usi): Ditto.
164 (avx512f_vcvttss2usiq): Ditto.
165 (avx512f_vcvtsd2usi): Ditto.
166 (avx512f_vcvtsd2usiq): Ditto.
167 (avx512f_vcvttsd2usi): Ditto.
168 (avx512f_vcvttsd2usiq): Ditto.
169 (ufloatv8siv8df): Ditto.
170 (avx512f_cvtdq2pd512_2): Ditto.
171 (avx512f_cvtpd2dq512): Ditto.
172 (avx512f_ufix_notruncv8dfv8si): Ditto.
173 (avx512f_cvtpd2ps512): Ditto.
174 (vec_unpacks_lo_v16sf): Ditto.
175 (vec_unpacks_hi_v16sf): Ditto.
176 (vec_unpacks_float_hi_v16si): Ditto.
177 (vec_unpacks_float_lo_v16si): Ditto.
178 (avx512f_unpckhps512): Ditto.
179 (avx512f_unpcklps512): Ditto.
180 (avx512f_movshdup512): Ditto.
181 (avx512f_movsldup512): Ditto.
182 (vec_extract_lo_v32hi): Ditto.
183 (vec_extract_hi_v32hi): Ditto.
184 (vec_extract_lo_v64qi): Ditto.
185 (vec_extract_hi_v64qi): Ditto.
186 (avx512f_unpckhpd512): Ditto.
187 (avx512f_movddup512): Ditto.
188 (avx512f_unpcklpd512): Ditto.
189 (*avx512f_unpcklpd512): Ditto.
190 (avx512f_shufps512_1): Ditto.
191 (avx512f_shufpd512_1): Ditto.
192 (avx512f_interleave_highv8di): Ditto.
193 (avx512f_interleave_lowv8di): Ditto.
194 (PMOV_DST_MODE): Ditto.
195 (pmov_src_mode): Ditto.
196 (pmov_src_lower): Ditto.
198 (*avx512f_<code><pmov_src_lower><mode>2): Ditto.
199 (*avx512f_<code>v8div16qi2): Ditto.
200 (*avx512f_<code>v8div16qi2_store): Ditto.
201 (vec_widen_umult_even_v16si): Ditto.
202 (*vec_widen_umult_even_v16si): Ditto.
203 (vec_widen_smult_even_v16si): Ditto.
204 (*vec_widen_smult_even_v16si): Ditto.
205 (avx512f_interleave_highv16si): Ditto.
206 (avx512f_interleave_lowv16si): Ditto.
207 (avx512f_<code>v16qiv16si2): Ditto.
208 (avx512f_<code>v16hiv16si2): Ditto.
209 (avx512f_<code>v8qiv8di2): Ditto.
210 (avx512f_<code>v8hiv8di2): Ditto.
211 (avx512f_<code>v8siv8di2): Ditto.
212 (avx512cd_maskb_vec_dupv8di): Ditto.
213 (avx512cd_maskw_vec_dupv16si): Ditto.
214 (avx512f_vcvtph2ps512): Ditto.
215 (avx512f_vcvtps2ph512): Ditto.
216 (VEC_EXTRACT_MODE): Extened with wider modes.
217 (VEC_PERM_AVX2): Ditto.
218 (VEC_PERM_CONST): Ditto.
221 2013-10-28 Joern Rennecke <joern.rennecke@embecosm.com>
223 * config/arc/arc.c (arc_ccfsm_post_advance):
224 Add comment about TYPE_RETURN.
226 2013-10-28 Bin Cheng <bin.cheng@arm.com>
228 * tree-ssa-loop-ivopts.c (strip_offset_1): Change parameter type.
229 Count DECL_FIELD_BIT_OFFSET for COMPONENT_REF.
230 (strip_offset): Convert offset to unsigned number.
232 2013-10-27 Tom de Vries <tom@codesourcery.com>
234 * cfgexpand.c (gimple_expand_cfg): Remove test for parm_birth_insn.
235 Don't commit insertions after NOTE_INSN_FUNCTION_BEG.
237 2013-10-27 Oleg Endo <olegendo@gcc.gnu.org>
239 * config/sh/sh.c (MSW, LSW): Move and rename macros to...
240 * config/sh/sh.h (SH_REG_MSW_OFFSET, SH_REG_LSW_OFFSET): ... here.
241 (TARGET_BIG_ENDIAN): New macro.
242 * config/sh/sh.md: Use it instead of !TARGET_LITTLE_ENDIAN.
243 Use SH_REG_MSW_OFFSET and SH_REG_LSW_OFFSET.
244 * config/sh/sh.c: Likewise.
245 * config/sh/sh.h: Likewise.
247 2013-10-27 Hans-Peter Nilsson <hp@axis.com>
249 * config/cris/cris.c (cris_emit_trap_for_misalignment): Replace the
250 removed PRED_MUDFLAP with PRED_NORETURN. Correct file-path in comment.
252 2013-10-26 Oleg Endo <olegendo@gcc.gnu.org>
254 * config/sh/sh.md (movmemsi): Remove empty constraints and predicates.
256 (cmpstr_t, cmpstrsi): Fix formatting.
258 2013-10-26 Oleg Endo <olegendo@gcc.gnu.org>
261 * config/sh/predicates.md (general_movdst_operand): Allow reg+reg
262 addressing, do not use general_operand for memory operands.
264 2013-10-26 Vladimir Makarov <vmakarov@redhat.com>
267 2013-10-25 Vladimir Makarov <vmakarov@redhat.com>
268 * lra-spills.c (lra_final_code_change): Remove useless move insns.
270 2013-10-26 Jeff Law <law@redhat.com>
272 * predict.c (PRED_MUDFLAP): Remove.
273 * targhooks.c (build_va_arg_indirect_ref): Remove mudflap support.
275 * Makefile.in (C_COMMON_OBJS): Remove tree-mudflap.
276 (OBJS): Remove tree-nomudflap.o
277 (GTFILES): Remove tree-mudflap.c
278 * builtins.c (expand_builtin_alloc): Remove mudflap support.
279 * gcc.c (MFWRAP_SPEC, MFLIB_SPEC): Likewise.
280 (mfwrap_spec, mflib_spec): Likewise.
281 (cpp_unique_options, cc1_options, static_specs): Likewise.
282 * gimplify (gimplify_vla_decl, build_va_arg_indirect_ref): Likewise.
283 * passes.def: Likewise.
284 * toplev.c (compile_file, process_options): Likewise.
285 * tree-inline.c (copy_tree_r): Likewise.
286 * tree-pass.,h (make_pass_mudflap_1, make_pass_mudflap_2): Likewise.
287 * varasm.c (make_decl_rtl, make_decl_rtl_for_debug): Likewise.
288 (build_constant_desc, output_constant_def_contents): Likewise.
289 (categorize_decl_for_section): Likewise.
290 * tree-mudflap.c: Removed.
291 * tree-mudflap.h: Removed.
292 * tree-nomudflap.c: Removed.
293 * bfin/uclinux.h (MFWRAP_SPEC): Remove.
294 * moxie/uclinux.h (MFWRAP_SPEC): Likewise.
295 * rs6000/aix.h (MFWRAP_SPEC, MFLIB_SPEC): Likewise.
296 * config/sol2.h (MFLIB_SPEC): Likewise.
297 * doc/install.texi: Remove mudflap references.
298 * doc/passes.texi: Similarly.
299 * doc/sourcebuild.texi: Similarly.
300 * doc/invoke.texi: Remove mudlfap related options.
302 2013-10-25 Vladimir Makarov <vmakarov@redhat.com>
304 PR rtl-optimization/58759
305 * lra-constraints.c (lra_constraints): Remove wrong condition to
306 remove insn setting up an equivalent pseudo.
308 2013-10-25 Vladimir Makarov <vmakarov@redhat.com>
310 * config/rs6000/rs6000-protos.h
311 (rs6000_secondary_memory_needed_mode): New prototype.
312 * config/rs6000/rs6000.c: Include ira.h.
313 (TARGET_LRA_P): Redefine.
314 (rs6000_legitimate_offset_address_p): Call
315 legitimate_constant_pool_address_p in strict mode for LRA.
316 (rs6000_legitimate_address_p): Ditto.
317 (legitimate_lo_sum_address_p): Add code for LRA.
319 (rs6000_emit_move): Add LRA version of code to generate load/store
321 (rs6000_secondary_memory_needed_mode): New.
322 (rs6000_alloc_sdmode_stack_slot): Do nothing for LRA.
323 (rs6000_secondary_reload_class): Return NO_REGS for LRA for
324 constants, memory, and FP registers.
326 * config/rs6000/rs6000.h (SECONDARY_MEMORY_NEEDED_MODE): New
328 * config/rs6000/rs6000.opt (mlra): New option.
329 * lra-spills.c (lra_final_code_change): Remove useless move insns.
331 2013-10-25 Yufeng Zhang <yufeng.zhang@arm.com>
333 * tree-ssa-math-opts.c (convert_plusminus_to_widen): Call
334 has_single_use () and not do the conversion if has_single_use ()
335 returns false for the multiplication result.
337 2013-10-25 David Malcolm <dmalcolm@redhat.com>
339 * tree.h (EXCEPTIONAL_CLASS_P): Rename parameter from "CODE"
340 to "NODE", since this works on a "tree", not an
342 (CONSTANT_CLASS_P): Likewise.
345 (INDIRECT_REF_P): Likewise.
346 (REFERENCE_CLASS_P): Likewise.
347 (COMPARISON_CLASS_P): Likewise.
348 (UNARY_CLASS_P): Likewise.
349 (BINARY_CLASS_P): Likewise.
350 (STATEMENT_CLASS_P): Likewise.
351 (VL_EXP_CLASS_P): Likewise.
352 (EXPRESSION_CLASS_P): Likewise.
353 (IS_TYPE_OR_DECL_P): Likewise.
355 2013-10-25 Marc Glisse <marc.glisse@inria.fr>
357 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Look for an
358 ADDR_EXPR in the defining statement.
360 2013-10-25 Richard Biener <rguenther@suse.de>
362 PR tree-optimization/58626
363 * tree-loop-distribution.c (enum rdg_dep_type): Remove
364 anti_dd, output_dd and input_dd.
365 (struct rdg_edge): Remove level and relation members.
366 (RDGE_LEVEL, RDGE_RELATION): Remove.
368 (create_rdg_edge_for_ddr): Remove.
369 (create_rdg_edges_for_scalar): Adjust.
370 (create_edge_for_control_dependence): Likewise.
371 (create_rdg_edges): Split into ...
372 (create_rdg_flow_edges): ... this
373 (create_rdg_cd_edges): ... and this.
375 (build_rdg): Likewise, do not compute data dependences or
377 (pg_add_dependence_edges): New function.
379 (distribute_loop): First apply all non-dependence based
380 partition mergings. Then compute dependences between partitions
381 and merge and order partitions according to them.
383 2013-10-25 Eric Botcazou <ebotcazou@adacore.com>
385 PR rtl-optimization/58831
386 * alias.c (init_alias_analysis): At the beginning of each iteration, set
387 the reg_seen[N] bit if static_reg_base_value[N] is non-null.
389 2013-10-25 Eric Botcazou <ebotcazou@adacore.com>
391 * recog.c (search_ofs): New static variable moved from...
392 (peep2_find_free_register): ...here.
393 (peephole2_optimize): Initialize it.
395 2013-10-25 Tobias Burnus <burnus@net-b.de>
397 * doc/invoke.texi (fopenmp): Change supported OpenMP version to 4.0.
399 2013-10-25 Uros Bizjak <ubizjak@gmail.com>
401 * config/i386/i386.h (TARGET_MPX): New define.
402 (TARGET_MPX_P): Ditto.
404 2013-10-24 Ilya Enkovich <ilya.enkovich@intel.com>
406 * config/i386/constraints.md (B): New.
409 * config/i386/i386-c.c (ix86_target_macros_internal): Add __MPX__.
410 * config/i386/i386-modes.def (BND32): New.
412 * config/i386/i386-protos.h (ix86_bnd_prefixed_insn_p): New.
413 * config/i386/i386.c (isa_opts): Add mmpx.
414 (regclass_map): Add bound registers.
415 (dbx_register_map): Likewise.
416 (dbx64_register_map): Likewise.
417 (svr4_dbx_register_map): Likewise.
419 (ix86_option_override_internal): Support MPX ISA.
420 (ix86_conditional_register_usage): Support bound registers.
421 (print_reg): Likewise.
422 (ix86_code_end): Add MPX bnd prefix.
423 (output_set_got): Likewise.
424 (ix86_output_call_insn): Likewise.
425 (ix86_print_operand): Add '!' (MPX bnd) print prefix support.
426 (ix86_print_operand_punct_valid_p): Likewise.
427 (ix86_print_operand_address): Support UNSPEC_BNDMK_ADDR and
429 (ix86_class_likely_spilled_p): Add bound regs support.
430 (ix86_hard_regno_mode_ok): Likewise.
431 (x86_order_regs_for_local_alloc): Likewise.
432 (ix86_bnd_prefixed_insn_p): New.
433 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Fix to new value.
434 (FIXED_REGISTERS): Add bound registers.
435 (CALL_USED_REGISTERS): Likewise.
436 (REG_ALLOC_ORDER): Likewise.
437 (HARD_REGNO_NREGS): Likewise.
439 (VALID_BND_REG_MODE): New.
440 (FIRST_BND_REG): New.
442 (reg_class): Add BND_REGS.
443 (REG_CLASS_NAMES): Likewise.
444 (REG_CLASS_CONTENTS): Likewise.
446 (ANY_BND_REG_P): New.
448 (HI_REGISTER_NAMES): Add bound registers.
449 * config/i386/i386.md (UNSPEC_BNDMK): New.
450 (UNSPEC_BNDMK_ADDR): New.
451 (UNSPEC_BNDSTX): New.
452 (UNSPEC_BNDLDX): New.
453 (UNSPEC_BNDLDX_ADDR): New.
457 (UNSPEC_MPX_FENCE): New.
460 (type): Add mpxmov, mpxmk, mpxchk, mpxld, mpxst.
461 (length_immediate): Likewise.
462 (prefix_0f): Likewise.
464 (prefix_rep): Check for bnd prefix.
466 (length): Use length_nobnd if specified.
471 (*jcc_1): Add bnd prefix and rename length attr to length_nobnd.
474 (simple_return_internal): Likewise.
475 (simple_return_pop_internal): Likewise.
476 (*indirect_jump): Add MPX bnd prefix.
477 (*tablejump_1): Likewise.
478 (simple_return_internal_long): Likewise.
479 (simple_return_indirect_internal): Likewise.
483 (*mov<mode>_internal_mpx): New.
484 (<mode>_<bndcheck>): New.
485 (*<mode>_<bndcheck>): New.
490 * config/i386/predicates.md (lea_address_operand): Rename to...
491 (address_no_seg_operand): ... this.
492 (address_mpx_no_base_operand): New.
493 (address_mpx_no_index_operand): New.
494 (bnd_mem_operator): New.
495 * config/i386/i386.opt (mmpx): New.
496 * doc/invoke.texi: Add documentation for the flags -mmpx, -mno-mpx.
497 * doc/rtl.texi Add documentation for BND32mode and BND64mode.
499 2013-10-24 Ilya Enkovich <ilya.enkovich@intel.com>
501 * mode-classes.def (MODE_POINTER_BOUNDS): New.
502 * tree.def (POINTER_BOUNDS_TYPE): New.
503 * genmodes.c (complete_mode): Support MODE_POINTER_BOUNDS.
504 (POINTER_BOUNDS_MODE): New.
505 (make_pointer_bounds_mode): New.
506 * machmode.h (POINTER_BOUNDS_MODE_P): New.
507 * stor-layout.c (int_mode_for_mode): Support MODE_POINTER_BOUNDS.
508 (layout_type): Support POINTER_BOUNDS_TYPE.
509 * tree-pretty-print.c (dump_generic_node): Support POINTER_BOUNDS_TYPE.
510 * tree.c (build_int_cst_wide): Support POINTER_BOUNDS_TYPE.
511 (type_contains_placeholder_1): Likewise.
512 * tree.h (POINTER_BOUNDS_TYPE_P): New.
513 * varasm.c (output_constant): Support POINTER_BOUNDS_TYPE.
514 * doc/rtl.texi (MODE_POINTER_BOUNDS): New.
516 2013-10-24 Igor Shevlyakov <igor.shevlyakov@gmail.com>
518 * expr.c (expand_expr_real_1): Use mode of memory reference rather than
519 mode of address computation when calling memory_address_addr_space.
521 2013-08-24 Richard Henderson <rth@twiddle.net>
524 * optabs.c (maybe_emit_atomic_exchange): Use create_input_operand
525 instead of create_convert_operand_to.
526 (maybe_emit_sync_lock_test_and_set): Likewise.
527 (expand_atomic_compare_and_swap): Likewise.
528 (maybe_emit_compare_and_swap_exchange_loop): Don't convert_modes.
530 2013-08-24 Sriraman Tallam <tmsriram@google.com>
532 * cgraph.c (cgraph_fnver_htab): Move GTY((...)) to be before htab_t.
533 Change param_is to use the struct and not the pointer to the struct.
535 2013-10-24 Andrew MacLeod <amacleod@redhat.com>
537 * builtins.c (dummy_object, gimplify_va_arg_expr): Move to gimplify.c.
538 * gimplify.c (build_va_arg_indirect_ref, std_gimplify_va_arg_expr):
540 (dummy_object, gimplify_va_arg_expr): Relocate from builtins.c.
541 * targhooks.c (build_va_arg_indirect_ref, std_gimplify_va_arg_expr):
542 Relocate from gimplify.c.
543 * targhooks.h: Add 2 prototypes.
544 * tree.h: Delete 2 prototypes.
546 2013-10-24 Igor Shevlyakov <igor.shevlyakov@gmail.com>
548 * tree-ssa-loop-ivopts.c (multiplier_allowed_in_address_p ): Check both
549 [reg+mult*reg] and [mult*reg] to determine if multiplier is allowed.
551 2013-10-24 Cong Hou <congh@google.com>
553 * convert.c (convert_to_real): Guard those unsafe math function
554 convertions with flag_unsafe_math_optimizations. Handle sqrt()
557 2013-10-24 Markus Trippelsdorf <markus@trippelsdorf.de>
560 * cgraph.c (cgraph_create_edge_1): Add indirect_unknown_callee
562 (cgraph_create_edge): Use the new argument.
563 (cgraph_create_indirect_edge): Likewise.
565 2013-10-24 Joern Rennecke <joern.rennecke@embecosm.com>
567 * gcc/config/arc/arc.c (arc_ccfsm_post_advance): Also handle
569 (arc_ifcvt) <case 1 and 2>: Check that arc_ccfsm_post_advance
570 changes statep->state.
572 2013-10-24 Tobias Burnus <burnus@net-b.de>
575 * tree-cfg.c (replace_loop_annotate): New function.
576 (execute_build_cfg): Call it.
577 * gimplify.c (gimple_boolify, gimplify_expr): Handle ANNOTATE_EXPR.
578 * internal-fn.c (expand_ANNOTATE): New function.
579 * internal-fn.def (ANNOTATE): Define as new internal function.
580 * tree-core.h (tree_node_kind): Add annot_expr_ivdep_kind.
581 * tree-pretty-print.c (dump_generic_node): Handle ANNOTATE_EXPR.
582 * tree.def (ANNOTATE_EXPR): New DEFTREECODE.
583 * doc/extend.texi (Pragmas): Document #pragma ivdep.
584 * doc/generic.texi (Expressions): Document ANNOTATE_EXPR.
586 2013-10-17 Ian Bolton <ian.bolton@arm.com>
587 Marcus Shawcroft <marcus.shawcroft@arm.com>
589 * config/aarch64/aarch64.c (aarch64_preferred_reload_class):
590 Special case reload SP+C into none GENERAL_REGS.
592 2013-10-24 Michael Matz <matz@suse.de>
594 * gengtype.c (is_file_equal): Check that files will be same length.
596 2013-10-25 Christian Bruel <christian.bruel@st.com>
598 * config.gcc (sh-*): Add sh-mem.o to extra_obj.
599 * gcc/config/sh/t-sh (sh-mem.o): New rule.
600 * gcc/config/sh/sh-mem.cc (expand_block_move): Moved here.
601 (sh_expand_cmpstr): New function.
602 * gcc/config/sh/sh.c (force_into, expand_block_move): Move to sh-mem.c.
603 * gcc/config/sh/sh-protos.h (sh_expand_cmpstr): Declare.
604 * gcc/config/sh/sh.md (cmpstrsi, cmpstr_t): New patterns.
607 2013-10-24 Jan-Benedict Glaw <jbglaw@lug-owl.de>
609 * configure.ac (ZW_PROG_COMPILER_DEPENDENCIES): Use CXX instead of CC.
610 * Makefile.in (CXXDEPMODE): Assign and change users.
612 * configure: Regenerate.
614 2013-10-23 David Malcolm <dmalcolm@redhat.com>
616 * gengtype-parse.c (require_without_advance): New.
617 (type): For GTY-marked types that are not GTY((user)), parse any
618 base classes, requiring them to be single-inheritance, and not
619 be templates. For non-GTY-marked types and GTY((user)),
620 continue to skip over any C++ inheritance specification.
621 * gengtype-state.c (state_writer::write_state_struct_type):
622 Write base class of type (if any).
623 (read_state_struct_type): Read base class of type (if any).
624 * gengtype.c (new_structure): Add a "base_class" parameter.
625 (create_optional_field_): Update for new parameter to new_structure.
626 (adjust_field_rtx_def): Likewise.
627 (adjust_field_tree_exp): Likewise.
628 * gengtype.h (struct type): Add "base_class" field to the s
630 (new_structure): Add "base" parameter.
632 2013-10-23 Sriraman Tallam <tmsriram@google.com>
635 * config/i386/i386.c (ix86_option_override_internal):
636 Change TARGET_SSE2 to TARGET_SSE2_P (opts->...)
637 (ix86_valid_target_attribute_tree):
638 Change TARGET_64BIT to TARGET_64BIT_P (opts->...)
639 Change TARGET_SSE to TARGET_SSE_P (opts->...)
641 2013-10-23 Andrew MacLeod <amacleod@redhat.com>
643 * tree-ssa-loop.h: Remove include files.
644 * gengtype.c (open_base_files): Adjust include list for gtype-desc.c.
645 * cfgloopmanip.c: Move required includes from tree-ssa-loop.h.
646 * graphite-clast-to-gimple.c: Likewise.
647 * graphite-scop-detection.c: Likewise.
648 * graphite-sese-to-poly.c: Likewise.
649 * ipa-inline-analysis.c: Likewise.
650 * ipa-pure-const.c: Likewise.
651 * loop-init.c: Likewise.
652 * passes.c: Likewise.
653 * predict.c: Likewise.
654 * tree-cfg.c: Likewise.
655 * tree-cfgcleanup.c: Likewise.
656 * tree-chrec.c: Likewise.
657 * tree-data-ref.c: Likewise.
658 * tree-loop-distribution.c: Likewise.
659 * tree-parloops.c: Likewise.
660 * tree-predcom.c: Likewise.
661 * tree-scalar-evolution.c: Likewise.
662 * tree-ssa-address.c: Likewise.
663 * tree-ssa.c: Likewise.
664 * tree-ssa-dce.c: Likewise.
665 * tree-ssa-loop.c: Likewise.
666 * tree-ssa-loop-im.c: Likewise.
667 * tree-ssa-loop-ivcanon.c: Likewise.
668 * tree-ssa-loop-ivopts.c: Likewise.
669 * tree-ssa-loop-manip.c: Likewise.
670 * tree-ssa-loop-niter.c: Likewise.
671 * tree-ssa-loop-prefetch.c: Likewise.
672 * tree-ssa-loop-unswitch.c: Likewise.
673 * tree-ssa-reassoc.c: Likewise.
674 * tree-vect-data-refs.c: Likewise.
675 * tree-vect-loop.c: Likewise.
676 * tree-vect-loop-manip.c: Likewise.
677 * tree-vectorizer.c: Likewise.
678 * tree-vect-stmts.c: Likewise.
679 * tree-vrp.c: Likewise.
681 2013-10-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
683 * config/rs6000/altivec.md (mulv8hi3): Adjust for little endian.
685 2013-10-23 Jakub Jelinek <jakub@redhat.com>
687 PR tree-optimization/58775
688 PR tree-optimization/58791
689 * tree-ssa-reassoc.c (reassoc_stmt_dominates_stmt_p): New function.
690 (insert_stmt_after): Rewritten, don't move the stmt, but really
692 (get_stmt_uid_with_default): Remove.
693 (build_and_add_sum): Use insert_stmt_after and
694 reassoc_stmt_dominates_stmt_p. Fix up uid if bb contains only labels.
695 (update_range_test): Set uid on stmts added by
696 force_gimple_operand_gsi. Don't immediately modify statements
697 in inter-bb optimization, just update oe->op values.
698 (optimize_range_tests): Return bool whether any changed have been made.
699 (update_ops): New function.
700 (struct inter_bb_range_test_entry): New type.
701 (maybe_optimize_range_tests): Perform statement changes here.
702 (not_dominated_by, appears_later_in_bb, get_def_stmt,
703 ensure_ops_are_available): Remove.
704 (find_insert_point): Rewritten.
705 (rewrite_expr_tree): Remove MOVED argument, add CHANGED argument,
706 return LHS of the (new resp. old) stmt. Don't call
707 ensure_ops_are_available, don't reuse SSA_NAMEs, recurse first
708 instead of last, move new stmt at the right place.
709 (linearize_expr, repropagate_negates): Don't reuse SSA_NAMEs.
710 (negate_value): Likewise. Set uids.
711 (break_up_subtract_bb): Initialize uids.
712 (reassociate_bb): Adjust rewrite_expr_tree caller.
713 (do_reassoc): Don't call renumber_gimple_stmt_uids.
715 2013-10-23 David Edelsohn <dje.gcc@gmail.com>
718 * config/rs6000/rs6000.md (mulsi3_internal1 and splitter): Add
719 TARGET_32BIT final condition.
720 (mulsi3_internal2 and splitter): Same.
722 2013-10-23 Jeff Law <law@redhat.com>
724 * tree-ssa-threadedge.c (thread_across_edge): Do not allow threading
725 through joiner blocks with abnormal outgoing edges.
727 * tree-ssa-threadupdate.c (thread_block_1): Renamed from thread_block.
728 Add parameter JOINERS, to allow/disallow threading through joiner
730 (thread_block): New. Call thread_block_1.
731 (mark_threaded_blocks): Remove code to filter out certain cases
732 of threading through joiner blocks.
733 (thread_through_all_blocks): Document how we can have a dangling
734 edge AUX field and clear it.
736 2013-10-23 Ian Lance Taylor <iant@google.com>
738 * doc/invoke.texi (Option Summary): Remove -fno-default-inline.
739 (C++ Dialect Options): Likewise.
740 (Optimize Options): Likewise.
742 2013-10-23 Tom de Vries <tom@codesourcery.com>
744 PR tree-optimization/58805
745 * tree-ssa-tail-merge.c (stmt_local_def): Add gimple_vdef check.
747 2013-10-23 Jakub Jelinek <jakub@redhat.com>
749 * tree-vect-patterns.c (vect_recog_divmod_pattern): Optimize
750 sequence based on get_range_info returned range.
752 2013-10-23 Andrew MacLeod <amacleod@redhat.com>
754 * tree-ssa.h: Remove all #include's
755 * gengtype.c (open_base_files): Adjust include list for gtype-desc.c.
756 * alias.c: Move required includes from tree-ssa.h.
758 * builtins.c: Likewise.
760 * cfgexpand.c: Likewise.
761 * cfghooks.c: Likewise.
762 * cfgloop.c: Likewise.
763 * cfgloopmanip.c: Likewise.
764 * cgraph.c: Likewise.
765 * cgraphbuild.c: Likewise.
766 * cgraphclones.c: Likewise.
767 * cgraphunit.c: Likewise.
769 * except.c: Likewise.
772 * fold-const.c: Likewise.
773 * ggc-page.c: Likewise.
774 * gimple-builder.c: Likewise.
775 * gimple-fold.c: Likewise.
776 * gimple-iterator.c: Likewise.
777 * gimple-low.c: Likewise.
778 * gimple-pretty-print.c: Likewise.
779 * gimple-ssa-strength-reduction.c: Likewise.
780 * gimple-streamer-in.c: Likewise.
781 * gimple-streamer-out.c: Likewise.
782 * gimplify.c: Likewise.
783 * graphite-blocking.c: Likewise.
784 * graphite-clast-to-gimple.c: Likewise.
785 * graphite-dependences.c: Likewise.
786 * graphite-interchange.c: Likewise.
787 * graphite-optimize-isl.c: Likewise.
788 * graphite-poly.c: Likewise.
789 * graphite-scop-detection.c: Likewise.
790 * graphite-sese-to-poly.c: Likewise.
791 * graphite.c: Likewise.
792 * ipa-cp.c: Likewise.
793 * ipa-inline-analysis.c: Likewise.
794 * ipa-inline-transform.c: Likewise.
795 * ipa-inline.c: Likewise.
796 * ipa-prop.c: Likewise.
797 * ipa-pure-const.c: Likewise.
798 * ipa-reference.c: Likewise.
799 * ipa-split.c: Likewise.
800 * ipa-utils.c: Likewise.
801 * loop-init.c: Likewise.
802 * lto-cgraph.c: Likewise.
803 * lto-section-in.c: Likewise.
804 * lto-section-out.c: Likewise.
805 * lto-streamer-in.c: Likewise.
806 * lto-streamer-out.c: Likewise.
807 * lto-streamer.c: Likewise.
808 * omp-low.c: Likewise.
809 * passes.c: Likewise.
810 * predict.c: Likewise.
811 * print-tree.c: Likewise.
812 * profile.c: Likewise.
814 * targhooks.c: Likewise.
815 * tracer.c: Likewise.
816 * trans-mem.c: Likewise.
817 * tree-call-cdce.c: Likewise.
818 * tree-cfg.c: Likewise.
819 * tree-cfgcleanup.c: Likewise.
820 * tree-chrec.c: Likewise.
821 * tree-complex.c: Likewise.
822 * tree-data-ref.c: Likewise.
823 * tree-dfa.c: Likewise.
824 * tree-eh.c: Likewise.
825 * tree-emutls.c: Likewise.
826 * tree-if-conv.c: Likewise.
827 * tree-inline.c: Likewise.
828 * tree-into-ssa.c: Likewise.
829 * tree-loop-distribution.c: Likewise.
830 * tree-mudflap.c: Likewise.
831 * tree-nested.c: Likewise.
832 * tree-nrv.c: Likewise.
833 * tree-object-size.c: Likewise.
834 * tree-outof-ssa.c: Likewise.
835 * tree-parloops.c: Likewise.
836 * tree-phinodes.c: Likewise.
837 * tree-predcom.c: Likewise.
838 * tree-pretty-print.c: Likewise.
839 * tree-profile.c: Likewise.
840 * tree-scalar-evolution.c: Likewise.
841 * tree-sra.c: Likewise.
842 * tree-ssa-address.c: Likewise.
843 * tree-ssa-alias.c: Likewise.
844 * tree-ssa-ccp.c: Likewise.
845 * tree-ssa-coalesce.c: Likewise.
846 * tree-ssa-copy.c: Likewise.
847 * tree-ssa-copyrename.c: Likewise.
848 * tree-ssa-dce.c: Likewise.
849 * tree-ssa-dom.c: Likewise.
850 * tree-ssa-dse.c: Likewise.
851 * tree-ssa-forwprop.c: Likewise.
852 * tree-ssa-ifcombine.c: Likewise.
853 * tree-ssa-live.c: Likewise.
854 * tree-ssa-loop-ch.c: Likewise.
855 * tree-ssa-loop-im.c: Likewise.
856 * tree-ssa-loop-ivcanon.c: Likewise.
857 * tree-ssa-loop-ivopts.c: Likewise.
858 * tree-ssa-loop-manip.c: Likewise.
859 * tree-ssa-loop-niter.c: Likewise.
860 * tree-ssa-loop-prefetch.c: Likewise.
861 * tree-ssa-loop-unswitch.c: Likewise.
862 * tree-ssa-loop.c: Likewise.
863 * tree-ssa-math-opts.c: Likewise.
864 * tree-ssa-operands.c: Likewise.
865 * tree-ssa-phiopt.c: Likewise.
866 * tree-ssa-phiprop.c: Likewise.
867 * tree-ssa-pre.c: Likewise.
868 * tree-ssa-propagate.c: Likewise.
869 * tree-ssa-reassoc.c: Likewise.
870 * tree-ssa-sccvn.c: Likewise.
871 * tree-ssa-sink.c: Likewise.
872 * tree-ssa-strlen.c: Likewise.
873 * tree-ssa-structalias.c: Likewise.
874 * tree-ssa-tail-merge.c: Likewise.
875 * tree-ssa-ter.c: Likewise.
876 * tree-ssa-threadedge.c: Likewise.
877 * tree-ssa-threadupdate.c: Likewise.
878 * tree-ssa-uncprop.c: Likewise.
879 * tree-ssa-uninit.c: Likewise.
880 * tree-ssa.c: Likewise.
881 * tree-ssanames.c: Likewise.
882 * tree-stdarg.c: Likewise.
883 * tree-streamer-in.c: Likewise.
884 * tree-switch-conversion.c: Likewise.
885 * tree-tailcall.c: Likewise.
886 * tree-vect-data-refs.c: Likewise.
887 * tree-vect-generic.c: Likewise.
888 * tree-vect-loop-manip.c: Likewise.
889 * tree-vect-loop.c: Likewise.
890 * tree-vect-patterns.c: Likewise.
891 * tree-vect-slp.c: Likewise.
892 * tree-vect-stmts.c: Likewise.
893 * tree-vectorizer.c: Likewise.
894 * tree-vrp.c: Likewise.
897 * value-prof.c: Likewise.
898 * var-tracking.c: Likewise.
899 * varpool.c: Likewise.
900 * vtable-verify.c: Likewise.
902 2013-10-23 Jan-Benedict Glaw <jbglaw@lug-owl.de>
904 * config/tilegx/tilegx.c: Include "tree.h".
906 2013-10-23 Jakub Jelinek <jakub@redhat.com>
908 * gimple-pretty-print.c (dump_ssaname_info): Always print "# " before
909 the info, not after it.
910 (gump_gimple_phi): Add COMMENT argument, if true, print "# " after
911 dump_ssaname_info call.
912 (pp_gimple_stmt_1): Adjust caller.
913 (dump_phi_nodes): Likewise. Don't print "# " here.
915 2013-10-22 Jan Hubicka <jh@suse.cz>
917 * i386.h (TARGET_MISALIGNED_MOVE_STRING_PROLOGUES_EPILOGUES): New
919 * x86-tune.def (TARGET_MISALIGNED_MOVE_STRING_PROLOGUES): Define it.
920 * i386.c (expand_small_movmem_or_setmem): New function.
921 (expand_set_or_movmem_prologue_epilogue_by_misaligned_moves): New
923 (alg_usable_p): Add support for value ranges; cleanup.
924 (ix86_expand_set_or_movmem): Add support for misaligned moves.
926 2013-10-22 Sterling Augustine <saugustine@google.com>
928 * doc/invoke.texi: Document -ggnu-pubnames.
929 * common.opt: Add new option -ggnu-pubnames and modify -gpubnames
931 * dwarf2out.c: Include gdb/gdb-index.h.
932 (DEBUG_PUBNAMES_SECTION, DEBUG_PUBTYPES_SECTION): Handle
933 debug_generate_pub_sections.
934 (is_java, output_pubtables, output_pubname): New functions.
935 (include_pubname_in_output): Handle debug_generate_pub_sections at
937 (size_of_pubnames): Use new local space_for_flags based on
938 debug_generate_pub_sections.
939 (output_pubnames): Unify pubnames and pubtypes output logic.
940 Genericize comments. Call output_pubname.
941 (dwarf2out_finish): Move logic to output_pubnames and call it.
943 2013-10-22 Uros Bizjak <ubizjak@gmail.com>
946 * config/i386/i386.c (put_condition_code) <case GTU, case LEU>:
947 Remove CCCmode handling.
948 <case LTU>: Return 'c' suffix for CCCmode.
949 <case GEU>: Return 'nc' suffix for CCCmode.
950 (ix86_cc_mode) <case GTU, case LEU>: Do not generate overflow checks.
951 * config/i386/i386.md (*sub<mode>3_cconly_overflow): Remove.
952 (*sub<mode>3_cc_overflow): Ditto.
953 (*subsi3_zext_cc_overflow): Ditto.
955 2013-10-22 Steve Ellcey <sellcey@mips.com>
957 * config/mips/mips.c (mips_rtx_costs): Fix cost estimate for nor
958 (AND (NOT OP1) (NOT OP2)).
960 2013-10-22 Bill Schmidt <wschmidt@vnet.ibm.com>
962 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
963 meaning of merge-high and merge-low masks for little endian; avoid
964 use of vector-pack masks for little endian for mismatched modes.
966 2013-10-22 Jan-Benedict Glaw <jbglaw@lug-owl.de>
968 * config/tilepro/tilepro.c: Include "tree.h".
970 2013-10-22 Andreas Schwab <schwab@suse.de>
972 * config/m68k/m68k.c (notice_update_cc): Handle register conflict
975 2013-10-22 Paolo Carlini <paolo.carlini@oracle.com>
977 * doc/contrib.texi ([Fran@,{c}ois Dumont], [Tim Shen],
978 [Ed Smith-Rowland]): New entries.
979 ([Stephen M. Webb]): Update.
981 2013-10-22 Andrew MacLeod <amacleod@redhat.com>
983 * tree-ssa-ter.h: Remove duplicate copy of file contents.
985 2013-10-21 Marek Polacek <polacek@redhat.com>
988 * fold-const.c (fold_range_test): Return 0 if the type is not
991 2013-10-21 Richard Sandiford <rdsandiford@googlemail.com>
993 * system.h: Move hwint.h include further down.
994 * hwint.h (sext_hwi, zext_hwi): Define unconditionally. Add
995 gcc_checking_asserts.
996 * hwint.c (sext_hwi, zext_hwi): Delete ENABLE_CHECKING versions.
998 2013-10-21 Bernd Edlinger <bernd.edlinger@hotmail.de>
1000 Fix volatile issues in optimize_bit_field_compare.
1001 * fold-const.c (optimize_bit_field_compare): Bail out if
1002 lvolatilep or rvolatilep.
1004 2013-10-21 Bernd Edlinger <bernd.edlinger@hotmail.de>
1006 Fix DECL_BIT_FIELD depencency on flag_strict_volatile_bitfields
1007 and get_inner_reference returning different pmode for non-volatile
1008 bit-field members dependent on flag_strict_volatile_bitfields.
1009 * stor-layout.c (layout_decl): Remove special handling of
1010 flag_strict_volatile_bitfields.
1011 * expr.c (get_inner_reference): Don't use DECL_BIT_FIELD
1012 if flag_strict_volatile_bitfields > 0 and TREE_THIS_VOLATILE.
1014 2013-10-21 Paulo Matos <pmatos@broadcom.com>
1016 * ipa-inline.c (edge_badness): Cap edge->count at max_count for badness
1019 2013-10-21 Jeff Law <law@redhat.com>
1021 * tree-ssa-threadedge.c (thread_through_normal_block): New
1022 argument VISITED. Remove VISISTED as a local variable. When we
1023 have a threadable jump, verify the destination of the jump has not
1025 (thread_across_edge): Allocate VISITED bitmap once at function
1026 scope and use it throughout. Make sure to set appropriate bits in
1027 VISITED for E (start of jump thread path).
1028 * tree-ssa-threadupdate.c (mark_threaded_blocks): Reject threading
1029 through a joiner if any edge on the path has a recorded jump thread.
1031 2013-10-21 Ian Lance Taylor <iant@google.com>
1033 * doc/invoke.texi (Optimize Options): For -fno-toplevel-reorder,
1034 don't imply that attributes can solve all problems.
1035 (Directory Options): Fix typo.
1037 2013-10-21 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1039 * config/arm/arm.c (cortexa9_extra_costs): Update mult costs for
1040 extend and extend_add.
1042 2013-10-21 Richard Biener <rguenther@suse.de>
1044 PR tree-optimization/58794
1045 * fold-const.c (operand_equal_p): Compare FIELD_DECL operand
1046 of COMPONENT_REFs with OEP_CONSTANT_ADDRESS_OF left in place.
1048 2013-10-21 Richard Biener <rguenther@suse.de>
1051 * fold-const.c (fold_binary_loc): Fold ((T) (X /[ex] C)) * C
1052 to (T) X for sign-changing conversions (or no conversion).
1054 2013-10-20 Uros Bizjak <ubizjak@gmail.com>
1056 * config/i386/i386.md (kxnor<mode>): Add FLAGS_REG clobber.
1058 2013-10-20 Jan Hubicka <jh@suse.cz>
1060 * config/i386/i386-tune.def: Add comment; organize into categories
1062 2013-10-21 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
1064 * config/i386/i386.c (expand_set_or_movmem_via_loop): Add issetmem
1065 argument. Update function comment.
1066 (expand_set_or_movmem_via_rep): New function combining
1067 expand_movmem_via_rep_mov and expand_setmem_via_rep_stos.
1068 (expand_movmem_via_rep_mov): Remove.
1069 expand_setmem_via_rep_stos): Remove.
1070 (expand_movmem_epilogue): Update calls correspondingly.
1071 (expand_setmem_epilogue_via_loop): Likewise.
1073 (expand_setmem_epilogue): Add VEC_VALUE argument, refactor.
1074 (expand_set_or_movmem_prologue): New function combining
1075 expand_movmem_prologue and expand_setmem_prologue.
1076 (expand_movmem_prologue): Remove.
1077 (expand_setmem_prologue): Remove.
1078 (expand_set_or_movmem_constant_prologue): New function combining
1079 expand_constant_movmem_prologue and expand_constant_setmem_prologue.
1080 (expand_constant_movmem_prologue): Remove.
1081 (expand_constant_setmem_prologue): Remove.
1082 (promote_duplicated_reg): Allow vector-const0 value.
1083 (ix86_expand_set_or_movmem): New function combining ix86_expand_movmem
1084 and ix86_expand_setmem.
1085 (ix86_expand_movmem): Call ix86_expand_set_or_movmem.
1086 (ix86_expand_setmem): Call ix86_expand_set_or_movmem.
1088 2013-10-21 Diego Novillo <dnovillo@google.com>
1090 * asan.c: Include tree.h
1091 * bb-reorder.c: Likewise.
1092 * cfgcleanup.c: Likewise.
1093 * cfgloopmanip.c: Likewise.
1094 * data-streamer-in.c: Likewise.
1095 * data-streamer-out.c: Likewise.
1096 * data-streamer.c: Likewise.
1097 * dwarf2cfi.c: Likewise.
1098 * graphite-blocking.c: Likewise.
1099 * graphite-clast-to-gimple.c: Likewise.
1100 * graphite-dependences.c: Likewise.
1101 * graphite-interchange.c: Likewise.
1102 * graphite-optimize-isl.c: Likewise.
1103 * graphite-poly.c: Likewise.
1104 * graphite-scop-detection.c: Likewise.
1105 * graphite-sese-to-poly.c: Likewise.
1106 * graphite.c: Likewise.
1107 * ipa-devirt.c: Likewise.
1108 * ipa-profile.c: Likewise.
1111 * loop-init.c: Likewise.
1112 * loop-unroll.c: Likewise.
1113 * lower-subreg.c: Likewise.
1114 * lto/lto-object.c: Likewise.
1115 * recog.c: Likewise.
1116 * reginfo.c: Likewise.
1117 * tree-loop-distribution.c: Likewise.
1118 * tree-parloops.c: Likewise.
1119 * tree-ssa-strlen.c: Likewise.
1120 * tree-streamer.c: Likewise.
1121 * value-prof.c: Likewise.
1122 * target-globals.c: Likewise.
1123 * expr.h: Include tree-core.h instead of tree.h.
1124 * gimple.h: Likewise.
1125 * ipa-prop.h: Likewise.
1126 * ipa-utils.h: Likewise.
1127 * lto-streamer.h: Likewise.
1128 * streamer-hooks.h: Likewise.
1129 * ipa-reference.h: Include cgraph.h instead of tree.h.
1130 * cgraph.h: Include basic-block.h instead of tree.h.
1131 * tree-streamer.h: Do not include tree.h.
1132 * genattrtab.c (write_header): Generate inclusion of
1134 * genautomata.c (main): Likewise.
1135 * genemit.c: Likewise.
1136 * genopinit.c: Likewise.
1137 * genoutput.c (output_prologue): Likewise.
1138 * genpeep.c: Likewise.
1140 2013-10-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1142 * config/rs6000/altivec.md (vec_unpacku_hi_v16qi): Adjust for
1144 (vec_unpacku_hi_v8hi): Likewise.
1145 (vec_unpacku_lo_v16qi): Likewise.
1146 (vec_unpacku_lo_v8hi): Likewise.
1148 2013-10-20 Jan Hubicka <jh@suse.cz>
1150 * config/i386/x86-tune.def (X86_TUNE_SLOW_IMUL_IMM32_MEM,
1151 X86_TUNE_SLOW_IMUL_IMM8): Keep enabled only for K8 and AMDFAM10.
1152 (X86_TUNE_USE_VECTOR_FP_CONVERTS): Disable for generic.
1154 2013-10-20 Richard Sandiford <rdsandiford@googlemail.com>
1156 * config/mips/mips.h (ISA_HAS_WSBH): Define.
1157 * config/mips/mips.md (UNSPEC_WSBH, UNSPEC_DSBH, UNSPEC_DSHD): New
1159 (bswaphi2, bswapsi2, bswapdi2, wsbh, dsbh, dshd): New patterns.
1161 2013-10-19 John David Anglin <danglin@gcc.gnu.org>
1164 * system.h: Undef m_slot.
1166 2013-10-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1168 * config/rs6000/rs6000.c (vspltis_constant): Make sure we check
1169 all elements for both endian flavors.
1171 2013-10-19 Uros Bizjak <ubizjak@gmail.com>
1174 * config/i386/i386.c (ix86_function_value_regno): Add DX_REG,
1175 ST1_REG and XMM1_REG for 32bit and 64bit targets. Also add DI_REG
1176 and SI_REG for 64bit SYSV ABI targets.
1178 2013-10-19 Uros Bizjak <ubizjak@gmail.com>
1180 * mode-switching.c (create_pre_exit): Rename maybe_builtin_apply
1181 to multi_reg_return. Clarify that we are skipping USEs of multiple
1182 return registers. Use bool type where appropriate.
1184 2013-10-18 Jan Hubicka <jh@suse.cz>
1186 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Disable accumulation
1188 * x86-tune.def (X86_TUNE_USE_LEAVE): Update comment.
1189 (X86_TUNE_PUSH_MEMORY): Likewise.
1190 (X86_TUNE_AVX256_UNALIGNED_LOAD_OPTIMAL,
1191 X86_TUNE_AVX256_UNALIGNED_STORE_OPTIMAL): New.
1192 (X86_TUNE_ACCUMULATE_OUTGOING_ARGS, X86_TUNE_ALWAYS_FANCY_MATH_387):
1194 * i386.c (x86_accumulate_outgoing_args, x86_arch_always_fancy_math_387,
1195 x86_avx256_split_unaligned_load, x86_avx256_split_unaligned_store):
1197 (ix86_option_override_internal): Update to use tune features instead
1200 2013-10-18 Cong Hou <congh@google.com>
1202 PR tree-optimization/58508
1203 * tree-vect-loop-manip.c (vect_loop_versioning): Hoist loop invariant
1204 statement that contains data refs with zero-step.
1206 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
1208 * tree-ssa.h: Don't include gimple-low.h, tree-ssa-address.h,
1209 sbitmap.h, tree-ssa-threadedge.h, tree-ssa-dom.h and tree-cfgcleanup.h.
1210 * gimple-low.c (gimple_check_call_arg,
1211 gimple_check_call_matching_types): Move to cgraph.c.
1212 * gimple-low.h: Remove prototype.
1213 * cgraph.c: (gimple_check_call_arg, gimple_check_call_matching_types):
1214 Relocate from gimple-low.c.
1215 * cgraph.h: Add prototype. Don't include basic-block.h.
1216 * gimplify.c: Add gimple-low to include list.
1217 * omp-low.c: Add gimple-low and tree-cfgcleanup.h to include list.
1218 * tree-eh.c: Add gimple-low to include list.
1219 * tree-nested.c: Likewise.
1220 * cfgexpand.c: Add tree-ssa-address.h to include list.
1222 * gimple-fold.c: Likewise.
1223 * gimple-ssa-strength-reduction.c: Likewise.
1224 * trans-mem.c: Likewise.
1225 * tree-mudflap.c: Likewise.
1226 * tree-ssa-loop-ivopts.c: Likewise.
1227 * tree-ssa-dom.c: Include tree-ssa-threadedge.h and tree-ssa-dom.h.
1228 (degenerate_phi_result): Move to tree-phinodes.c.
1229 * tree-ssa-loop-ch.c: Include tree-ssa-threadedge.h.
1230 * tree-ssa-threadedge.c: Likewise.
1231 * tree-vrp.c: Likewise.
1232 * tree-phinodes.c (degenerate_phi_result): Relocate here.
1233 * tree-ssa-dom.h (degenerate_phi_result): Remove Prototype.
1234 * tree-phinodes.h (degenerate_phi_result): Add prototype.
1235 * tree-ssa-copy.c: Include tree-ssa-dom.h.
1236 * tree-ssa-forwprop.c: Likewise.
1237 * tree-cfgcleanup.c (execute_cleanup_cfg_post_optimizing,
1238 pass_data_cleanup_cfg_post_optimizing,
1239 make_pass_cleanup_cfg_post_optimizing): Relocate from tree-optimize.c.
1240 * tree-optimize.c: Delete File.
1241 * graphite.c: Include tree-cfgcleanup.h.
1242 * passes.c: Likewise.
1243 * tree-cfg.c: Likewise.
1244 * tree-profile.c: Likewise.
1245 * tree-ssa-dse.c: Likewise.
1246 * tree-ssa-loop-ivcanon.c: Likewise.
1247 * tree-switch-conversion.c: Don't include tree-ssa-operands.h.
1248 * tree-outof-ssa.c: Include sbitmap.h.
1249 * tree-ssa-live.c: Likewise.
1250 * tree-ssa-propagate.c: Likewise.
1251 * tree-ssa-structalias.c: Likewise.
1252 * tree-stdarg.c: Likewise.
1253 * Makefile.in (OBJS): Delete tree-optimize.o.
1254 * basic-block.h (gcov_type, gcov_type_unsigned): Move to coretypes.h.
1255 * coretypes.h (gcov_type, gcov_type_unsigned): Relocate here.
1256 * varasm.c: Include basic-block.h.
1257 * cfgloop.h: Include function.h instead of basic-block.h
1258 (bb_loop_depth): Move to cfgloop.c.
1259 * cfgloop.c (bb_loop_depth): Relocate from cfgloop.h.
1261 2013-10-18 Teresa Johnson <tejohnson@google.com>
1263 * predict.c (probably_never_executed): Compare frequency-based
1264 count to number of training runs.
1265 * params.def (UNLIKELY_BB_COUNT_FRACTION): New parameter.
1267 2013-10-18 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1269 * config/arm/arm.c (cortexa9_extra_costs): New table.
1270 (arm_cortex_a9_tune): Use cortexa9_extra_costs.
1272 2013-10-18 Jeff Law <law@redhat.com>
1274 * tree-ssa-threadupdate.c: Do not include "tm.h" or "tm_p.h".
1276 * tree-ssa-threadupdate.c: Include "dbgcnt.h".
1277 (register_jump_thread): Add "registered_jump_thread" debug
1279 * dbgcnt.def (registered_jump_thread): New debug counter.
1281 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
1283 * config/rs6000/rs6000.c: Include cgraph.h.
1285 2013-10-18 Teresa Johnson <tejohnson@google.com>
1287 * tree-ssa-tail-merge.c (replace_block_by): Update edge
1288 weights during merging.
1290 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
1292 * tree-cfg.h: Rename from tree-flow.h. Remove #includes.
1293 * tree-ssa.h: Relocate required #includes from tree-cfg.h.
1294 * tree-ssa-operands.h: Remove prototype.
1295 * tree-ssa-operands.c (virtual_operand_p): Move to gimple.c.
1296 * gimple.c (virtual_operand_p): Relocate from gimple.c.
1297 * gimple.h: Add prototype.
1298 * gimple-ssa.h: Include tree-ssa-operands.h.
1299 * tree-dump.c: Add tree-cfg.h to include list.
1300 * tree-ssa-alias.c: Add ipa-reference.h to include list.
1301 * config/alpha/alpha.c: Include gimple-ssa.h instead of tree-flow.h.
1302 * config/i386/i386.c: Don't include tree-flow.h.
1303 * config/rs6000/rs6000.c: Likewise.
1305 2013-10-18 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1307 * config/frv/frv.c (frv_init_cumulative_args): Fix wrong cast.
1309 2013-10-18 Richard Biener <rguenther@suse.de>
1311 * stor-layout.c (layout_type): Do not change TYPE_PRECISION
1312 or TYPE_UNSIGNED of integral types.
1313 (set_min_and_max_values_for_integral_type): Leave TYPE_MIN/MAX_VALUE
1314 NULL_TREE for zero-precision integral types.
1316 2013-10-18 James Greenhalgh <james.greenhalgh@arm.com>
1318 * config/aarch64/arm_neon.h
1319 (vcvt<ds>_n_<fsu><32,64>_<fsu><32,64>): Correct argument types.
1321 2013-10-17 Sriraman Tallam <tmsriram@google.com>
1324 * opth-gen.awk: Define target_flags_explicit.
1326 2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
1328 * config/rs6000/rs6000.c (enum rs6000_reload_reg_type): Add new
1329 fields to the reg_addr array that describes the valid addressing
1330 mode for any register, general purpose registers, floating point
1331 registers, and Altivec registers.
1332 (FIRST_RELOAD_REG_CLASS): Likewise.
1333 (LAST_RELOAD_REG_CLASS): Likewise.
1334 (struct reload_reg_map_type): Likewise.
1335 (reload_reg_map_type): Likewise.
1336 (RELOAD_REG_VALID): Likewise.
1337 (RELOAD_REG_MULTIPLE): Likewise.
1338 (RELOAD_REG_INDEXED): Likewise.
1339 (RELOAD_REG_OFFSET): Likewise.
1340 (RELOAD_REG_PRE_INCDEC): Likewise.
1341 (RELOAD_REG_PRE_MODIFY): Likewise.
1342 (reg_addr): Likewise.
1343 (mode_supports_pre_incdec_p): New helper functions to say whether
1344 a given mode supports PRE_INC, PRE_DEC, and PRE_MODIFY.
1345 (mode_supports_pre_modify_p): Likewise.
1346 (rs6000_debug_vector_unit): Rearrange the -mdebug=reg output to
1347 print the valid address mode bits for each mode.
1348 (rs6000_debug_print_mode): Likewise.
1349 (rs6000_debug_reg_global): Likewise.
1350 (rs6000_setup_reg_addr_masks): New function to set up the address
1351 mask bits for each type.
1352 (rs6000_init_hard_regno_mode_ok): Use memset to clear arrays.
1353 Call rs6000_setup_reg_addr_masks to set up the address mask bits.
1354 (rs6000_legitimate_address_p): Use mode_supports_pre_incdec_p and
1355 mode_supports_pre_modify_p to determine if PRE_INC, PRE_DEC, and
1356 PRE_MODIFY are supported.
1357 (rs6000_output_move_128bit): Change to use {src,dest}_vmx_p for altivec
1358 registers, instead of {src,dest}_av_p.
1359 (rs6000_print_options_internal): Tweak the debug output slightly.
1361 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
1363 * config/i386/sse.md (*vec_widen_smult_even_v8si): Remove
1366 2013-10-17 Andrew MacLeod <amacleod@redhat.com>
1368 * tree-flow.h (struct omp_region): Move to omp-low.c.
1369 Remove omp_ prototypes and variables.
1370 * gimple.h (omp_reduction_init): Move prototype to omp-low.h.
1371 (copy_var_decl): Relocate prototype from tree-flow.h.
1372 * gimple.c (copy_var_decl): Relocate from omp-low.c.
1373 * tree.h: Move prototype to omp-low.h.
1374 * omp-low.h: New File. Relocate prototypes here.
1375 * omp-low.c (struct omp_region): Make local here.
1376 (root_omp_region): Make static.
1377 (copy_var_decl) Move to gimple.c.
1378 (new_omp_region): Make static.
1379 (make_gimple_omp_edges): New. Refactored from tree-cfg.c make_edges.
1380 * tree-cfg.c: Include omp-low.h.
1381 (make_edges): Factor out OMP specific bits to make_gimple_omp_edges.
1382 * gimplify.c: Include omp-low.h.
1383 * tree-parloops.c: Likewise.
1385 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
1387 * config/i386/i386.c (ix86_fixup_binary_operands): When both source
1388 operands are in memory, prefer to force non-matched operand 1 to
1391 2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
1394 * config/rs6000/rs6000.c (rs6000_legitimate_address_p): Only
1395 restrict TImode addresses to single indirect registers if both
1396 -mquad-memory and -mvsx-timode are used.
1397 (rs6000_output_move_128bit): Use quad_load_store_p to determine if
1398 we should emit load/store quad. Remove using %y for quad memory
1401 * config/rs6000/rs6000.md (mov<mode>_ppc64, TI/PTImode): Add
1402 constraints to allow load/store quad on machines where TImode is
1403 not allowed in VSX registers. Use 'n' instead of 'F' constraint
1404 for TImode to load integer constants.
1406 2013-10-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1408 * config/aarch64/aarch64.c (aarch64_print_operand): Handle 'c'.
1410 2013-10-17 Marcus Shawcroft <marcus.shawcroft@arm.com>
1412 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Adjust
1413 handling of STACK_REG.
1415 2013-10-17 Richard Biener <rguenther@suse.de>
1417 PR tree-optimization/58143
1418 * tree-ssa-loop-im.c (arith_code_with_undefined_signed_overflow):
1420 (rewrite_to_defined_overflow): Likewise.
1421 (move_computations_dom_walker::before_dom): Rewrite stmts
1422 with undefined signed overflow that are not always executed
1423 into unsigned arithmetic.
1425 2013-10-16 Michael Meissner <meissner@linux.vnet.ibm.com>
1428 * config/rs6000/rs6000.opt (rs6000_isa_flags_explicit): Move the
1429 explicit isa flag to be an options variable, instead of using
1430 global_options_set. Remove define from rs6000.h.
1431 * config/rs6000/rs6000.h (rs6000_isa_flags_explicit): Likewise.
1433 * config/rs6000/rs6000.c (rs6000_option_override_internal):
1434 Initialize rs6000_isa_flags_explicit.
1435 (rs6000_function_specific_save): Add gcc_options* parameter, so
1436 that the powerpc builds after the 2013-10-15 changes.
1437 (rs6000_function_specific_restore): Likewise.
1439 2013-10-16 DJ Delorie <dj@redhat.com>
1441 * config/rl78/rl78.c (rl78_alloc_address_registers_macax): Verify
1442 op is a REG before checking REGNO.
1443 (rl78_alloc_physical_registers): Verify pattern is a SET before
1446 2013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1448 * gcc/config/rs6000/vector.md (vec_unpacks_hi_v4sf): Correct for
1450 (vec_unpacks_lo_v4sf): Likewise.
1451 (vec_unpacks_float_hi_v4si): Likewise.
1452 (vec_unpacks_float_lo_v4si): Likewise.
1453 (vec_unpacku_float_hi_v4si): Likewise.
1454 (vec_unpacku_float_lo_v4si): Likewise.
1456 2013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1458 * config/rs6000/vsx.md (vsx_concat_<mode>): Adjust output for LE.
1459 (vsx_concat_v2sf): Likewise.
1461 2013-10-16 James Greenhalgh <james.greenhalgh@arm.com>
1463 * config/aarch64/aarch64.md
1464 (*mov<mode>_aarch64): Fix output template for DUP (element) Scalar.
1466 2013-10-16 Andrew MacLeod <amacleod@redhat.com>
1468 PR tree-optimization/58697
1469 * cfgloop.c (get_estimated_loop_iterations_int): Rename from
1470 estimated_loop_iterations_int.
1471 (max_stmt_executions_int): Call get_max_loop_iterations_int.
1472 (get_max_loop_iterations_int): New. HWINT version of
1473 get_max_loop_iterations.
1474 * cfgloop.h: Add prototypes.
1475 * loop-iv.c (find_simple_exit): call get_estimated_loop_iterations_int.
1476 * loop-unroll.c (decide_peel_once_rolling): Call
1477 get_estimated_loop_iterations_int.
1478 * tree-ssa-loop-niter.c (estimated_loop_iterations_int): Add back.
1479 * tree-ssa-loop-niter.h: Tweak prototypes.
1481 2013-10-16 David Malcolm <dmalcolm@redhat.com>
1483 * gengtype-parse.c (struct_field_seq): Ignore access-control
1484 keywords ("public:" etc).
1486 2013-10-16 Marcus Shawcroft <marcus.shawcroft@arm.com>
1488 * config/aarch64/aarch64.c (aarch64_regno_regclass): Classify
1489 FRAME_POINTER_REGNUM and ARG_POINTER_REGNUM as POINTER_REGS.
1491 2013-10-16 Yvan Roux <yvan.roux@linaro.org>
1493 * config/arm/arm.opt (mlra): New option.
1494 * config/arm/arm.c (arm_lra_p): New function.
1495 (TARGET_LRA_P): Define.
1497 2013-10-16 Paulo Matos <pmatos@broadcom.com>
1499 * tree-core.h (tree_code_name): Remove.
1500 * tree.h (get_tree_code_name): New prototype.
1501 * tree.c (tree_code_name): Make static.
1502 (get_tree_code_name): New function.
1503 (dump_tree_statistics, tree_check_failed, tree_not_check_failed,
1504 tree_class_check_failed, tree_range_check_failed,
1505 tree_not_class_check_failed, omp_clause_check_failed,
1506 tree_contains_struct_check_failed, tree_operand_check_failed): Use new
1507 wrapper get_tree_code_name instead of calling tree_code_name directly.
1508 * tree-vrp.c (dump_asserts_for): Likewise.
1509 * tree-dump.c (dequeue_and_dump): Likewise.
1510 * tree-pretty-print.c (do_niy, dump_generic_node): Likewise.
1511 * tree-pretty-print.h (pp_unsupported_tree): Likewise.
1512 * lto-streamer-out.c (lto_write_tree, DFS_write_tree): Likewise.
1513 * tree-ssa-dom.c (print_expr_hash_elt): Likewise.
1514 * gimple-pretty-print.c (dump_unary_rhs, dump_binary_rhs,
1515 dump_ternary_rhs, dump_gimple_assign, dump_gimple_cond,
1516 dump_gimple_omp_for): Likewise.
1517 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Likewise.
1518 * tree-ssa-pre.c (print_pre_expr): Likewise.
1519 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Likewise.
1520 * print-tree.c (print_node_brief, print_node): Likewise.
1521 * gimple.c (gimple_check_failed): Likewise.
1522 * lto-streamer.c (lto_tag_name, print_lto_report): Likewise.
1523 * config/frv/frv.c (frv_init_cumulative_args): Likewise.
1524 * config/mep/mep.c (mep_validate_vliw): Likewise.
1525 * config/iq2000/iq2000.c (init_cumulative_args): Likewise.
1526 * config/rs6000/rs6000.c (init_cumulative_args): Likewise.
1528 2013-10-16 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
1530 * config/i386/i386.c (ix86_option_override_internal): Enable FMA4
1533 2013-10-16 Hans-Peter Nilsson <hp@axis.com>
1535 * config/cris/t-elfmulti (MULTILIB_OPTIONS, MULTILIB_DIRNAMES)
1536 (MULTILIB_MATCHES): Add multilib for -march=v8.
1538 2013-10-15 Sriraman Tallam <tmsriram@google.com>
1541 * optc-save-gen.awk: Add extra parameter to the save and restore
1543 * opth-gen.awk: Generate new TARGET_* macros to accept a parameter.
1544 * tree.c (build_optimization_node): New parameter. Add extra parameter
1545 to call to cl_optimization_save.
1546 (build_target_option_node): New parameter. Add extra parameter
1547 to call to cl_target_option_save.
1548 * tree.h (build_optimization_node): New parameter.
1549 (build_target_option_node): New parameter.
1550 * c-family/c-common.c (handle_optimize_attribute): Fix calls to
1551 build_optimization_node and build_target_option_node.
1552 * c-family/c-pragma.c (handle_pragma_optimize): Ditto.
1553 (handle_pragma_push_options): Ditto.
1554 * toplev.c (process_options): Ditto.
1555 * opts.c (init_options_struct): Check for opts_set non-null.
1556 * target.def (target_option.save): New parameter.
1557 (target_option.restore): New parameter.
1558 * tm.texi: Generate.
1559 * config/i386/i386-c.c (ix86_target_macros_internal): Ditto.
1560 (ix86_pragma_target_parse): Ditto.
1561 * config/i386/i386-protos.h (ix86_valid_target_attribute_tree): New
1563 * config/rs6000/rs6000.c (rs6000_option_override_internal): Fix calls
1564 to build_optimization_node and build_target_option_node.
1565 (rs6000_valid_attribute_p): Ditto.
1566 (rs6000_pragma_target_parse): Ditto.
1567 * config/i386/i386.opt (x_ix86_target_flags_explicit): New TargetSave
1569 * config/i386/i386.h:
1570 TARGET_64BIT_P: New Macro
1571 TARGET_MMX_P: New Macro.
1572 TARGET_3DNOW_P: New Macro.
1573 TARGET_3DNOW_A_P: New Macro.
1574 TARGET_SSE_P: New Macro.
1575 TARGET_SSE2_P: New Macro.
1576 TARGET_SSE3_P: New Macro.
1577 TARGET_SSSE3_P: New Macro.
1578 TARGET_SSE4_1_P: New Macro.
1579 TARGET_SSE4_2_P: New Macro.
1580 TARGET_AVX_P: New Macro.
1581 TARGET_AVX2_P: New Macro.
1582 TARGET_AVX512F_P: New Macro.
1583 TARGET_AVX512PF_P: New Macro.
1584 TARGET_AVX512ER_P: New Macro.
1585 TARGET_AVX512CD_P: New Macro.
1586 TARGET_FMA_P: New Macro.
1587 TARGET_SSE4A_P: New Macro.
1588 TARGET_FMA4_P: New Macro.
1589 TARGET_XOP_P: New Macro.
1590 TARGET_LWP_P: New Macro.
1591 TARGET_ABM_P: New Macro.
1592 TARGET_BMI_P: New Macro.
1593 TARGET_BMI2_P: New Macro.
1594 TARGET_LZCNT_P: New Macro.
1595 TARGET_TBM_P: New Macro.
1596 TARGET_POPCNT_P: New Macro.
1597 TARGET_SAHF_P: New Macro.
1598 TARGET_MOVBE_P: New Macro.
1599 TARGET_CRC32_P: New Macro.
1600 TARGET_AES_P: New Macro.
1601 TARGET_PCLMUL_P: New Macro.
1602 TARGET_CMPXCHG16B_P: New Macro.
1603 TARGET_FSGSBASE_P: New Macro.
1604 TARGET_RDRND_P: New Macro.
1605 TARGET_F16C_P: New Macro.
1606 TARGET_RTM_P: New Macro.
1607 TARGET_HLE_P: New Macro.
1608 TARGET_RDSEED_P: New Macro.
1609 TARGET_PRFCHW_P: New Macro.
1610 TARGET_ADX_P: New Macro.
1611 TARGET_FXSR_P: New Macro.
1612 TARGET_XSAVE_P: New Macro.
1613 TARGET_XSAVEOPT_P: New Macro.
1614 TARGET_LP64_P: New Macro.
1615 TARGET_X32_P: New Macro.
1616 TARGET_FPMATH_DEFAULT_P: New Macro.
1617 TARGET_FLOAT_RETURNS_IN_80387_P: New Macro.
1618 * config/i386/i386.c (ix86_option_override_internal): New parameters.
1620 Change ix86_tune_string to access opts->x_ix86_tune_string.
1621 Change ix86_isa_flags to access opts->x_ix86_isa_flags.
1622 Change ix86_arch_string to access opts->x_ix86_arch_string.
1623 Change ix86_stringop_alg to access opts->x_ix86_stringop_alg.
1624 Change ix86_pmode to access opts->x_ix86_pmode.
1625 Change ix86_abi to access opts->x_ix86_abi.
1626 Change ix86_cmodel to access opts->x_ix86_cmodel.
1627 Change ix86_asm_dialect to access opts->x_ix86_asm_dialect.
1628 Change ix86_isa_flags_explicit to access
1629 opts->x_ix86_isa_flags_explicit.
1630 Change ix86_dump_tunes to access opts->x_ix86_dump_tunes.
1631 Change ix86_regparm to access opts->x_ix86_regparm.
1632 Change ix86_branch_cost to access opts->x_ix86_branch_cost.
1633 Change ix86_preferred_stack_boundary_arg to access
1634 opts->x_ix86_preferred_stack_boundary_arg.
1635 Change ix86_force_align_arg_pointer to access
1636 opts->x_ix86_force_align_arg_pointer.
1637 Change ix86_incoming_stack_boundar_arg to access
1638 opts->x_ix86_incoming_stack_boundar_arg.
1639 Change ix86_fpmath to access opts->x_ix86_fpmath.
1640 Change ix86_veclibabi_type to access opts->x_ix86_veclibabi_type.
1641 Change ix86_recip_name to access opts->x_ix86_recip_name.
1642 Change ix86_stack_protector_guard to access
1643 opts->x_ix86_stack_protector_guard.
1644 Change ix86_tune_memcpy_strategy to access
1645 opts->x_ix86_tune_memcpy_strategy.
1646 Change ix86_tune_memset_strategy to access
1647 opts->x_ix86_tune_memset_strategy.
1648 Change global_options to access opts.
1649 Change global_options_set to access opts_set.
1650 Change TARGET_64BIT to TARGET_64BIT_P (opts->...).
1651 Change TARGET_MMX to TARGET_MMX_P (opts->...).
1652 Change TARGET_3DNOW to TARGET_3DNOW_P (opts->...).
1653 Change TARGET_3DNOW_A to TARGET_3DNOW_A_P (opts->...).
1654 Change TARGET_SSE to TARGET_SSE_P (opts->...).
1655 Change TARGET_SSE2 to TARGET_SSE2_P (opts->...).
1656 Change TARGET_SSE3 to TARGET_SSE3_P (opts->...).
1657 Change TARGET_SSSE3 to TARGET_SSSE3_P (opts->...).
1658 Change TARGET_SSE4_1 to TARGET_SSE4_1_P (opts->...).
1659 Change TARGET_SSE4_2 to TARGET_SSE4_2_P (opts->...).
1660 Change TARGET_AVX to TARGET_AVX_P (opts->...).
1661 Change TARGET_AVX2 to TARGET_AVX2_P (opts->...).
1662 Change TARGET_AVX512F to TARGET_AVX512F_P (opts->...).
1663 Change TARGET_AVX512PF to TARGET_AVX512PF_P (opts->...).
1664 Change TARGET_AVX512ER to TARGET_AVX512ER_P (opts->...).
1665 Change TARGET_AVX512CD to TARGET_AVX512CD_P (opts->...).
1666 Change TARGET_FMA to TARGET_FMA_P (opts->...).
1667 Change TARGET_SSE4A to TARGET_SSE4A_P (opts->...).
1668 Change TARGET_FMA4 to TARGET_FMA4_P (opts->...).
1669 Change TARGET_XOP to TARGET_XOP_P (opts->...).
1670 Change TARGET_LWP to TARGET_LWP_P (opts->...).
1671 Change TARGET_ABM to TARGET_ABM_P (opts->...).
1672 Change TARGET_BMI to TARGET_BMI_P (opts->...).
1673 Change TARGET_BMI2 to TARGET_BMI2_P (opts->...).
1674 Change TARGET_LZCNT to TARGET_LZCNT_P (opts->...).
1675 Change TARGET_TBM to TARGET_TBM_P (opts->...).
1676 Change TARGET_POPCNT to TARGET_POPCNT_P (opts->...).
1677 Change TARGET_SAHF to TARGET_SAHF_P (opts->...).
1678 Change TARGET_MOVBE to TARGET_MOVBE_P (opts->...).
1679 Change TARGET_CRC32 to TARGET_CRC32_P (opts->...).
1680 Change TARGET_AES to TARGET_AES_P (opts->...).
1681 Change TARGET_PCLMUL to TARGET_PCLMUL_P (opts->...).
1682 Change TARGET_CMPXCHG16B to TARGET_CMPXCHG16B_P (opts->...).
1683 Change TARGET_FSGSBASE to TARGET_FSGSBASE_P (opts->...).
1684 Change TARGET_RDRND to TARGET_RDRND_P (opts->...).
1685 Change TARGET_F16C to TARGET_F16C_P (opts->...).
1686 Change TARGET_RTM to TARGET_RTM_P (opts->...).
1687 Change TARGET_HLE to TARGET_HLE_P (opts->...).
1688 Change TARGET_RDSEED to TARGET_RDSEED_P (opts->...).
1689 Change TARGET_PRFCHW to TARGET_PRFCHW_P (opts->...).
1690 Change TARGET_ADX to TARGET_ADX_P (opts->...).
1691 Change TARGET_FXSR to TARGET_FXSR_P (opts->...).
1692 Change TARGET_XSAVE to TARGET_XSAVE_P (opts->...).
1693 Change TARGET_XSAVEOPT to TARGET_XSAVEOPT_P (opts->...).
1694 Change TARGET_LP64 to TARGET_LP64_P (opts->...).
1695 Change TARGET_X32 to TARGET_X32_P (opts->...).
1696 Change TARGET_FPMATH_DEFAULT to TARGET_FPMATH_DEFAULT_P (opts->...).
1697 Change TARGET_FLOAT_RETURNS_IN_80387 to
1698 TARGET_FLOAT_RETURNS_IN_80387_P (opts->...).
1699 (ix86_function_specific_save): New parameter. Use opts-> fields
1700 to replace global fields.
1701 (ix86_function_specific_restore): Ditto.
1702 (ix86_valid_target_attribute_inner_p): New parameters.
1704 Fix call to ix86_handle_option and set_option.
1705 (ix86_valid_target_attribute_tree): New parameters.
1706 Change global_options to access opts.
1707 Change global_options_set to access opts_set.
1708 Fix call to ix86_valid_target_attribute_inner_p.
1709 Change ix86_tune_string to access opts->x_ix86_tune_string.
1710 Change ix86_arch_string to access opts->x_ix86_arch_string.
1711 Change ix86_fpmath to access opts->x_ix86_fpmath
1712 Fix call to ix86_option_override_internal.
1713 Fix call to ix86_add_new_builtins.
1714 Fix calls to build_optimization_node and build_target_option_node.
1715 (ix86_valid_target_attribute_p): Remove access to global_options.
1716 Use new gcc_options structure func_options.
1717 Fix call to ix86_valid_target_attribute_tree.
1718 Fix call to build_optimization_node.
1719 (get_builtin_code_for_version): Fix call to
1720 ix86_valid_target_attribute_tree.
1722 2013-10-15 David Malcolm <dmalcolm@redhat.com>
1724 * Makefile.in (PICFLAG): New.
1725 (enable_host_shared): New.
1726 (INTERNAL_CFLAGS): Use PICFLAG.
1727 (LIBIBERTY): Use pic build of libiberty.a if configured with
1728 --enable-host-shared.
1729 * configure.ac: Add --enable-host-shared, setting up new
1731 * configure: Regenerate.
1732 * doc/install.texi (--enable-shared): Add note contrasting it with ...
1733 (--enable-host-shared): New option.
1735 2013-10-15 Richard Biener <rguenther@suse.de>
1737 * tree-tailcall.c (find_tail_calls): Don't use tail-call recursion
1738 for built-in functions.
1740 2013-10-15 Zhenqiang Chen <zhenqiang.chen@arm.com>
1742 * tree-ssa-reassoc.c: Include rtl.h and tm_p.h.
1743 (optimize_range_tests_1): New function,
1744 extracted from optimize_range_tests.
1745 (optimize_range_tests_xor): Similarly.
1746 (optimize_range_tests_diff): New function.
1747 (optimize_range_tests): Use optimize_range_tests_1.
1749 2013-10-15 Cong Hou <congh@google.com>
1751 * tree-vect-loop.c (vect_is_simple_reduction_1): Relax the
1752 requirement of the reduction pattern so that one operand of the
1753 reduction operation can come from outside of the loop.
1755 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1757 * config/arm/neon-schedgen.ml: Remove.
1758 * config/arm/cortex-a9-neon.md: Remove comment regarding
1761 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1763 * config/arm/types: Remove old neon types.
1765 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1767 * config/arm/cortex-a7.md
1768 (cortex_a7_neon_type): New.
1769 (cortex_a7_neon_mul): Update for new types.
1770 (cortex_a7_neon_mla): Likewise.
1771 (cortex_a7_neon): Likewise.
1773 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1775 * config/arm/cortex-a15-neon.md
1776 (cortex_a15_neon_type): New,
1778 (cortex_a15_neon_int_1): Remove.
1779 (cortex_a15_neon_int_2): Likewise.
1780 (cortex_a15_neon_int_3): Likewise.
1781 (cortex_a15_neon_int_4): Likewise.
1782 (cortex_a15_neon_int_5): Likewise.
1783 (cortex_a15_neon_vqneg_vqabs): Likewise.
1784 (cortex_a15_neon_vmov): Likewise.
1785 (cortex_a15_neon_vaba): Likewise.
1786 (cortex_a15_neon_vaba_qqq): Likewise.
1787 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1788 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
1789 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
1791 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1792 (cortex_a15_neon_mla_qqq_8_16): Likewise.
1793 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar): Likewise.
1794 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
1795 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
1796 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
1797 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
1798 (cortex_a15_neon_shift_1): Likewise.
1799 (cortex_a15_neon_shift_2): Likewise.
1800 (cortex_a15_neon_shift_3): Likewise.
1801 (cortex_a15_neon_vshl_ddd): Likewise.
1802 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
1803 (cortex_a15_neon_vsra_vrsra): Likewise.
1804 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
1805 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
1806 (cortex_a15_neon_bp_3cycle): Likewise.
1807 (cortex_a15_neon_ldm_2): Likewise.
1808 (cortex_a15_neon_stm_2): Likewise.
1809 (cortex_a15_neon_mcr): Likewise.
1810 (cortex_a15_neon_mrc): Likewise.
1811 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
1812 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
1813 (cortex_a15_neon_fp_vmul_ddd): Likewise.
1814 (cortex_a15_neon_fp_vmul_qqd): Likewise.
1815 (cortex_a15_neon_fp_vmla_ddd): Likewise.
1816 (cortex_a15_neon_fp_vmla_qqq): Likewise.
1817 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
1818 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
1819 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
1820 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
1821 (cortex_a15_neon_bp_simple): Likewise.
1822 (cortex_a15_neon_bp_2cycle): Likewise.
1823 (cortex_a15_neon_bp_3cycle): Likewise.
1824 (cortex_a15_neon_vld1_1_2_regs): Likewise.
1825 (cortex_a15_neon_vld1_3_4_regs): Likewise.
1826 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
1827 (cortex_a15_neon_vld2_4_regs): Likewise.
1828 (cortex_a15_neon_vld3_vld4): Likewise.
1829 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
1830 (cortex_a15_neon_vst1_3_4_regs): Likewise.
1831 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Rename to...
1832 (cortex_a15_neon_vst2_4_regs_vst3): ...This, update for new attributes.
1833 (cortex_a15_neon_vst3_vst4): Rename to...
1834 (cortex_a15_neon_vst4): This, update for new attributes.
1835 (cortex_a15_neon_vld1_vld2_lane): Update for new attributes.
1836 (cortex_a15_neon_vld3_vld4_lane): Likewise.
1837 (cortex_a15_neon_vst1_vst2_lane): Likewise.
1838 (cortex_a15_neon_vst3_vst4_lane): Likewise.
1839 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
1840 (cortex_a15_neon_ldm_2): Likewise.
1841 (cortex_a15_neon_stm_2): Likewise.
1842 (cortex_a15_neon_mcr): Likewise.
1843 (cortex_a15_neon_mcr_2_mcrr): Likewise.
1844 (cortex_a15_neon_mrc): Likewise.
1845 (cortex_a15_neon_mrrc): Likewise.
1847 (cortex_a15_neon_abd): New.
1848 (cortex_a15_neon_abd_q): Likewise.
1849 (cortex_a15_neon_aba): Likewise.
1850 (cortex_a15_neon_aba_q): Likewise.
1851 (cortex_a15_neon_acc): Likewise.
1852 (cortex_a15_neon_acc_q): Likewise.
1853 (cortex_a15_neon_arith_basic): Likewise.
1854 (cortex_a15_neon_arith_complex): Likewise.
1855 (cortex_a15_neon_multiply): Likewise.
1856 (cortex_a15_neon_multiply_q): Likewise.
1857 (cortex_a15_neon_mla): Likewise.
1858 (cortex_a15_neon_mla_q): Likewise.
1859 (cortex_a15_neon_sat_mla_long): Likewise.
1860 (cortex_a15_neon_shift_acc): Likewise.
1861 (cortex_a15_neon_shift_imm_basic): Likewise.
1862 (cortex_a15_neon_shift_imm_complex): Likewise.
1863 (cortex_a15_neon_shift_reg_basic): Likewise.
1864 (cortex_a15_neon_shift_reg_basic_q): Likewise.
1865 (cortex_a15_neon_shift_reg_complex): Likewise.
1866 (cortex_a15_neon_shift_reg_complex_q): Likewise.
1867 (cortex_a15_neon_fp_negabs): Likewise
1868 (cortex_a15_neon_fp_arith): Likewise
1869 (cortex_a15_neon_fp_arith_q): Likewise
1870 (cortex_a15_neon_fp_cvt_int): Likewise
1871 (cortex_a15_neon_fp_cvt_int_q): Likewise
1872 (cortex_a15_neon_fp_cvt_16): Likewise
1873 (cortex_a15_neon_fp_mul): Likewise
1874 (cortex_a15_neon_fp_mul_q): Likewise
1875 (cortex_a15_neon_fp_mla): Likewise
1876 (cortex_a15_neon_fp_mla_q): Likewise
1877 (cortex_a15_neon_fp_recps_rsqrte): Likewise.
1878 (cortex_a15_neon_fp_recps_rsqrte_q): Likewise.
1879 (cortex_a15_neon_bitops): Likewise.
1880 (cortex_a15_neon_bitops_q): Likewise.
1881 (cortex_a15_neon_from_gp): Likewise.
1882 (cortex_a15_neon_from_gp_q): Likewise.
1883 (cortex_a15_neon_tbl3_tbl4): Likewise.
1884 (cortex_a15_neon_zip_q): Likewise.
1885 (cortex_a15_neon_to_gp): Likewise.
1886 (cortex_a15_neon_load_a): Likewise.
1887 (cortex_a15_neon_load_b): Likewise.
1888 (cortex_a15_neon_load_c): Likewise.
1889 (cortex_a15_neon_load_d): Likewise.
1890 (cortex_a15_neon_load_e): Likewise.
1891 (cortex_a15_neon_load_f): Likewise.
1892 (cortex_a15_neon_store_a): Likewise.
1893 (cortex_a15_neon_store_b): Likewise.
1894 (cortex_a15_neon_store_c): Likewise.
1895 (cortex_a15_neon_store_d): Likewise.
1896 (cortex_a15_neon_store_e): Likewise.
1897 (cortex_a15_neon_store_f): Likewise.
1898 (cortex_a15_neon_store_g): Likewise.
1899 (cortex_a15_neon_store_h): Likewise.
1900 (cortex_a15_vfp_to_from_gp): Likewise.
1902 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1904 * config/arm/cortex-a9-neon.md (cortex_a9_neon_type): New.
1906 (cortex_a9_neon_vshl_ddd): Remove.
1907 (cortex_a9_neon_vst3_vst4): Likewise.
1908 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
1910 (cortex_a9_neon_bit_ops_q): New.
1912 (cortex_a9_neon_int_1): Use cortex_a8_neon_type.
1913 (cortex_a9_neon_int_2): Likewise.
1914 (cortex_a9_neon_int_3): Likewise.
1915 (cortex_a9_neon_int_4): Likewise.
1916 (cortex_a9_neon_int_5): Likewise.
1917 (cortex_a9_neon_vqneg_vqabs): Likewise.
1918 (cortex_a9_neon_vmov): Likewise.
1919 (cortex_a9_neon_vaba): Likewise.
1920 (cortex_a9_neon_vaba_qqq): Likewise.
1921 (cortex_a9_neon_shift_1): Likewise.
1922 (cortex_a9_neon_shift_2): Likewise.
1923 (cortex_a9_neon_shift_3): Likewise.
1924 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
1925 (cortex_a9_neon_vsra_vrsra): Likewise.
1926 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1927 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
1928 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
1930 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1931 (cortex_a9_neon_mla_qqq_8_16): Likewise.
1932 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
1934 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
1935 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
1936 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
1937 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
1938 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
1939 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
1940 (cortex_a9_neon_fp_vsum): Likewise.
1941 (cortex_a9_neon_fp_vmul_ddd): Likewise.
1942 (cortex_a9_neon_fp_vmul_qqd): Likewise.
1943 (cortex_a9_neon_fp_vmla_ddd): Likewise.
1944 (cortex_a9_neon_fp_vmla_qqq): Likewise.
1945 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
1946 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
1947 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
1948 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
1949 (cortex_a9_neon_bp_simple): Likewise.
1950 (cortex_a9_neon_bp_2cycle): Likewise.
1951 (cortex_a9_neon_bp_3cycle): Likewise.
1952 (cortex_a9_neon_ldr): Likewise.
1953 (cortex_a9_neon_str): Likewise.
1954 (cortex_a9_neon_vld1_1_2_regs): Likewise.
1955 (cortex_a9_neon_vld1_3_4_regs): Likewise.
1956 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
1957 (cortex_a9_neon_vld2_4_regs): Likewise.
1958 (cortex_a9_neon_vld3_vld4): Likewise.
1959 (cortex_a9_neon_vld1_vld2_lane): Likewise.
1960 (cortex_a9_neon_vld3_vld4_lane): Likewise.
1961 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
1962 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
1963 (cortex_a9_neon_vst1_3_4_regs): Likewise.
1964 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
1965 (cortex_a9_neon_vst1_vst2_lane): Likewise.
1966 (cortex_a9_neon_vst3_vst4_lane): Likewise.
1967 (cortex_a9_neon_mcr): Likewise.
1968 (cortex_a9_neon_mcr_2_mcrr): Likewise.
1969 (cortex_a9_neon_mrc): Likewise.
1970 (cortex_a9_neon_mrrc): Likewise.
1972 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1974 * config/arm/cortex-a8-neon.md (cortex_a8_neon_type): New.
1976 (cortex_a8_neon_vshl_ddd): Remove.
1977 (cortex_a8_neon_vst3_vst4): Likewise.
1978 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
1980 (cortex_a8_neon_bit_ops_q): New.
1982 (cortex_a8_neon_int_1): Use cortex_a8_neon_type.
1983 (cortex_a8_neon_int_2): Likewise..
1984 (cortex_a8_neon_int_3): Likewise.
1985 (cortex_a8_neon_int_5): Likewise.
1986 (cortex_a8_neon_vqneg_vqabs): Likewise.
1987 (cortex_a8_neon_int_4): Likewise.
1988 (cortex_a8_neon_vaba): Likewise.
1989 (cortex_a8_neon_vaba_qqq): Likewise.
1990 (cortex_a8_neon_shift_1): Likewise.
1991 (cortex_a8_neon_shift_2): Likewise.
1992 (cortex_a8_neon_shift_3): Likewise.
1993 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
1994 (cortex_a8_neon_vsra_vrsra): Likewise.
1995 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1996 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
1997 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
1999 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
2000 (cortex_a8_neon_mla_qqq_8_16): Likewise.
2001 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
2003 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
2004 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
2005 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
2006 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
2007 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
2008 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
2009 (cortex_a8_neon_fp_vsum): Likewise.
2010 (cortex_a8_neon_fp_vmul_ddd): Likewise.
2011 (cortex_a8_neon_fp_vmul_qqd): Likewise.
2012 (cortex_a8_neon_fp_vmla_ddd): Likewise.
2013 (cortex_a8_neon_fp_vmla_qqq): Likewise.
2014 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
2015 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
2016 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
2017 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
2018 (cortex_a8_neon_bp_simple): Likewise.
2019 (cortex_a8_neon_bp_2cycle): Likewise.
2020 (cortex_a8_neon_bp_3cycle): Likewise.
2021 (cortex_a8_neon_ldr): Likewise.
2022 (cortex_a8_neon_str): Likewise.
2023 (cortex_a8_neon_vld1_1_2_regs): Likewise.
2024 (cortex_a8_neon_vld1_3_4_regs): Likewise.
2025 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
2026 (cortex_a8_neon_vld2_4_regs): Likewise.
2027 (cortex_a8_neon_vld3_vld4): Likewise.
2028 (cortex_a8_neon_vld1_vld2_lane): Likewise.
2029 (cortex_a8_neon_vld3_vld4_lane): Likewise.
2030 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
2031 (cortex_a8_neon_vst1_3_4_regs): Likewise.
2032 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
2033 (cortex_a8_neon_vst1_vst2_lane): Likewise.
2034 (cortex_a8_neon_vst3_vst4_lane): Likewise.
2035 (cortex_a8_neon_mcr): Likewise.
2036 (cortex_a8_neon_mcr_2_mcrr): Likewise.
2037 (cortex_a8_neon_mrc): Likewise.
2038 (cortex_a8_neon_mrrc): Likewise.
2040 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
2042 * config/aarch64/iterators.md (Vetype): Add SF and DF modes.
2044 * config/aarch64/aarch64-simd.md (neon_type): Remove.
2045 (aarch64_simd_dup<mode>): Add "type" attribute.
2046 (aarch64_dup_lane<mode>): Likewise.
2047 (aarch64_dup_lane_<vswap_width_name><mode>): Likewise.
2048 (*aarch64_simd_mov<mode>): Likewise.
2049 (aarch64_simd_mov_from_<mode>low): Likewise.
2050 (aarch64_simd_mov_from_<mode>high): Likewise.
2051 (orn<mode>3): Likewise.
2052 (bic<mode>3): Likewise.
2053 (add<mode>3): Likewise.
2054 (sub<mode>3): Likewise.
2055 (mul<mode>3): Likewise.
2056 (*aarch64_mul3_elt<mode>): Likewise.
2057 (*aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
2058 (*aarch64_mul3_elt_to_128df): Likewise.
2059 (*aarch64_mul3_elt_to_64v2df): Likewise.
2060 (neg<mode>2): Likewise.
2061 (abs<mode>2): Likewise.
2062 (abd<mode>_3): Likewise.
2063 (aba<mode>_3): Likewise.
2064 (fabd<mode>_3): Likewise.
2065 (*fabd_scalar<mode>3): Likewise.
2066 (and<mode>3): Likewise.
2067 (ior<mode>3): Likewise.
2068 (xor<mode>3): Likewise.
2069 (one_cmpl<mode>2): Likewise.
2070 (aarch64_simd_vec_set<mode>): Likewise.
2071 (aarch64_simd_lshr<mode>): Likewise.
2072 (aarch64_simd_ashr<mode>): Likewise.
2073 (aarch64_simd_imm_shl<mode>): Likewise.
2074 (aarch64_simd_reg_sshl<mode): Likewise.
2075 (aarch64_simd_reg_shl<mode>_unsigned): Likewise.
2076 (aarch64_simd_reg_shl<mode>_signed): Likewise.
2077 (aarch64_simd_vec_setv2di): Likewise.
2078 (aarch64_simd_vec_set<mode>): Likewise.
2079 (aarch64_mla<mode>): Likewise.
2080 (*aarch64_mla_elt<mode>): Likewise.
2081 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
2082 (aarch64_mls<mode>): Likewise.
2083 (*aarch64_mls_elt<mode>): Likewise.
2084 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
2085 (<su><maxmin><mode>3): Likewise.
2086 (move_lo_quad_<mode>): Likewise.
2087 (aarch64_simd_move_hi_quad_<mode>): Likewise.
2088 (aarch64_simd_vec_pack_trunc_<mode>): Likewise.
2089 (vec_pack_trunc_<mode>): Likewise.
2090 (aarch64_simd_vec_unpack<su>_lo_<mode>): Likewise.
2091 (aarch64_simd_vec_unpack<su>_hi_<mode>): Likewise.
2092 (*aarch64_<su>mlal_lo<mode>): Likewise.
2093 (*aarch64_<su>mlal_hi<mode>): Likewise.
2094 (*aarch64_<su>mlsl_lo<mode>): Likewise.
2095 (*aarch64_<su>mlsl_hi<mode>): Likewise.
2096 (*aarch64_<su>mlal<mode>): Likewise.
2097 (*aarch64_<su>mlsl<mode>): Likewise.
2098 (aarch64_simd_vec_<su>mult_lo_<mode>): Likewise.
2099 (aarch64_simd_vec_<su>mult_hi_<mode>): Likewise.
2100 (add<mode>3): Likewise.
2101 (sub<mode>3): Likewise.
2102 (mul<mode>3): Likewise.
2103 (div<mode>3): Likewise.
2104 (neg<mode>2): Likewise.
2105 (abs<mode>2): Likewise.
2106 (fma<mode>4): Likewise.
2107 (*aarch64_fma4_elt<mode>): Likewise.
2108 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
2109 (*aarch64_fma4_elt_to_128df): Likewise.
2110 (*aarch64_fma4_elt_to_64v2df): Likewise.
2111 (fnma<mode>4): Likewise.
2112 (*aarch64_fnma4_elt<mode>): Likewise.
2113 (*aarch64_fnma4_elt_<vswap_width_name><mode>
2114 (*aarch64_fnma4_elt_to_128df): Likewise.
2115 (*aarch64_fnma4_elt_to_64v2df): Likewise.
2116 (<frint_pattern><mode>2): Likewise.
2117 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): Likewise.
2118 (<optab><fcvt_target><VDQF:VDQF:mode>2): Likewise.
2119 (vec_unpacks_lo_v4sf): Likewise.
2120 (aarch64_float_extend_lo_v2df): Likewise.
2121 (vec_unpacks_hi_v4sf): Likewise.
2122 (aarch64_float_truncate_lo_v2sf): Likewise.
2123 (aarch64_float_truncate_hi_v4sf): Likewise.
2124 (aarch64_vmls<mode>): Likewise.
2125 (<su><maxmin><mode>3): Likewise.
2126 (<maxmin_uns><mode>3): Likewise.
2127 (reduc_<sur>plus_<mode>): Likewise.
2128 (reduc_<sur>plus_v2di): Likewise.
2129 (reduc_<sur>plus_v2si): Likewise.
2130 (reduc_<sur>plus_<mode>): Likewise.
2131 (aarch64_addpv4sf): Likewise.
2132 (clz<mode>2): Likewise.
2133 (reduc_<maxmin_uns>_<mode>): Likewise.
2134 (reduc_<maxmin_uns>_v2di): Likewise.
2135 (reduc_<maxmin_uns>_v2si): Likewise.
2136 (reduc_<maxmin_uns>_<mode>): Likewise.
2137 (reduc_<maxmin_uns>_v4sf): Likewise.
2138 (aarch64_simd_bsl<mode>_internal): Likewise.
2139 (*aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): Likewise.
2140 (*aarch64_get_lane_zero_extendsi<mode>): Likewise.
2141 (aarch64_get_lane<mode>): Likewise.
2142 (*aarch64_combinez<mode>): Likewise.
2143 (aarch64_combine<mode>): Likewise.
2144 (aarch64_simd_combine<mode>): Likewise.
2145 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): Likewise.
2146 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): Likewise.
2147 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>): Likewise.
2148 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w<mode>): Likewise.
2149 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w2<mode>_internal): Likewise.
2150 (aarch64_<sur>h<addsub><mode>): Likewise.
2151 (aarch64_<sur><addsub>hn<mode>): Likewise.
2152 (aarch64_<sur><addsub>hn2<mode>): Likewise.
2153 (aarch64_pmul<mode>): Likewise.
2154 (aarch64_<su_optab><optab><mode>): Likewise.
2155 (aarch64_<sur>qadd<mode>): Likewise.
2156 (aarch64_sqmovun<mode>): Likewise.
2157 (aarch64_<sur>qmovn<mode>): Likewise.
2158 (aarch64_s<optab><mode>): Likewise.
2159 (aarch64_sq<r>dmulh<mode>): Likewise.
2160 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
2161 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
2162 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
2163 (aarch64_sqdml<SBINQOPS:as>l<mode>): Likewise.
2164 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
2165 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
2166 (aarch64_sqdml<SBINQOPS:as>l_n<mode>): Likewise.
2167 (aarch64_sqdml<SBINQOPS:as>l2<mode>_internal): Likewise.
2168 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
2169 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
2170 (aarch64_sqdmull<mode>): Likewise.
2171 (aarch64_sqdmull_lane<mode>_internal): Likewise.
2172 (aarch64_sqdmull_n<mode>): Likewise.
2173 (aarch64_sqdmull2<mode>_internal): Likewise.
2174 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
2175 (aarch64_sqdmull2_n<mode>_internal): Likewise.
2176 (aarch64_<sur>shl<mode>): Likewise.
2177 (aarch64_<sur>q<r>shl<mode>
2178 (aarch64_<sur>shll_n<mode>): Likewise.
2179 (aarch64_<sur>shll2_n<mode>): Likewise.
2180 (aarch64_<sur>shr_n<mode>): Likewise.
2181 (aarch64_<sur>sra_n<mode>): Likewise.
2182 (aarch64_<sur>s<lr>i_n<mode>): Likewise.
2183 (aarch64_<sur>qshl<u>_n<mode>): Likewise.
2184 (aarch64_<sur>q<r>shr<u>n_n<mode>): Likewise.
2185 (aarch64_cm<optab><mode>): Likewise.
2186 (aarch64_cm<optab>di): Likewise.
2187 (aarch64_cm<optab><mode>): Likewise.
2188 (aarch64_cm<optab>di): Likewise.
2189 (aarch64_cmtst<mode>): Likewise.
2190 (aarch64_cmtstdi): Likewise.
2191 (aarch64_cm<optab><mode>): Likewise.
2192 (*aarch64_fac<optab><mode>): Likewise.
2193 (aarch64_addp<mode>): Likewise.
2194 (aarch64_addpdi): Likewise.
2195 (sqrt<mode>2): Likewise.
2196 (vec_load_lanesoi<mode>): Likewise.
2197 (vec_store_lanesoi<mode>): Likewise.
2198 (vec_load_lanesci<mode>): Likewise.
2199 (vec_store_lanesci<mode>): Likewise.
2200 (vec_load_lanesxi<mode>): Likewise.
2201 (vec_store_lanesxi<mode>): Likewise.
2202 (*aarch64_mov<mode>): Likewise.
2203 (aarch64_ld2<mode>_dreg): Likewise.
2204 (aarch64_ld2<mode>_dreg): Likewise.
2205 (aarch64_ld3<mode>_dreg): Likewise.
2206 (aarch64_ld3<mode>_dreg): Likewise.
2207 (aarch64_ld4<mode>_dreg): Likewise.
2208 (aarch64_ld4<mode>_dreg): Likewise.
2209 (aarch64_tbl1<mode>): Likewise.
2210 (aarch64_tbl2v16qi): Likewise.
2211 (aarch64_combinev16qi): Likewise.
2212 (aarch64_<PERMUTE:perm_insn><PERMUTE:perm_hilo><mode>): Likewise.
2213 (aarch64_st2<mode>_dreg): Likewise.
2214 (aarch64_st2<mode>_dreg): Likewise.
2215 (aarch64_st3<mode>_dreg): Likewise.
2216 (aarch64_st3<mode>_dreg): Likewise.
2217 (aarch64_st4<mode>_dreg): Likewise.
2218 (aarch64_st4<mode>_dreg): Likewise.
2219 (*aarch64_simd_ld1r<mode>): Likewise.
2220 (aarch64_frecpe<mode>): Likewise.
2221 (aarch64_frecp<FRECP:frecp_suffix><mode>): Likewise.
2222 (aarch64_frecps<mode>): Likewise.
2224 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
2226 * config/arm/iterators.md (V_elem_ch): New.
2228 (VQH_type): Likewise.
2229 * config/arm/arm.md (is_neon_type): New.
2230 (conds): Use is_neon_type.
2231 (anddi3_insn): Update type attribute.
2232 (xordi3_insn): Likewise.
2233 (one_cmpldi2): Likewise.
2234 * gcc/config/arm/vfp.md (movhf_vfp_neon): Update type attribute.
2235 * gcc/config/arm/neon.md (neon_mov): Update type attribute.
2236 (*movmisalign<mode>_neon_store): Likewise.
2237 (*movmisalign<mode>_neon_load): Likewise.
2238 (vec_set<mode>_internal): Likewise.
2239 (vec_set<mode>_internal): Likewise.
2240 (vec_setv2di_internal): Likewise.
2241 (vec_extract<mode>): Likewise.
2242 (vec_extract<mode>): Likewise.
2243 (vec_extractv2di): Likewise.
2244 (*add<mode>3_neon): Likewise.
2245 (adddi3_neon): Likewise.
2246 (*sub<mode>3_neon): Likewise.
2247 (subdi3_neon): Likewise.
2248 (fma<VCVTF:mode>4): Likewise.
2249 (fma<VCVTF:mode>4_intrinsic): Likewise.
2250 (*fmsub<VCVTF:mode>4): Likewise.
2251 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
2252 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
2253 (ior<mode>3): Likewise.
2254 (and<mode>3): Likewise.
2255 (orn<mode>3_neon): Likewise.
2256 (orndi3_neon): Likewise.
2257 (bic<mode>3_neon): Likewise.
2258 (bicdi3_neon): Likewise.
2259 (xor<mode>3): Likewise.
2260 (one_cmpl<mode>2): Likewise.
2261 (abs<mode>2): Likewise.
2262 (neg<mode>2): Likewise.
2263 (negdi2_neon): Likewise.
2264 (*umin<mode>3_neon): Likewise.
2265 (*umax<mode>3_neon): Likewise.
2266 (*smin<mode>3_neon): Likewise.
2267 (*smax<mode>3_neon): Likewise.
2268 (vashl<mode>3): Likewise.
2269 (vashr<mode>3_imm): Likewise.
2270 (vlshr<mode>3_imm): Likewise.
2271 (ashl<mode>3_signed): Likewise.
2272 (ashl<mode>3_unsigned): Likewise.
2273 (neon_load_count): Likewise.
2274 (ashldi3_neon_noclobber): Likewise.
2275 (ashldi3_neon): Likewise.
2276 (signed_shift_di3_neon): Likewise.
2277 (unsigned_shift_di3_neon): Likewise.
2278 (ashrdi3_neon_imm_noclobber): Likewise.
2279 (lshrdi3_neon_imm_noclobber): Likewise.
2280 (<shift>di3_neon): Likewise.
2281 (widen_ssum<mode>3): Likewise.
2282 (widen_usum<mode>3): Likewise.
2283 (quad_halves_<code>v4si): Likewise.
2284 (quad_halves_<code>v4sf): Likewise.
2285 (quad_halves_<code>v8hi): Likewise.
2286 (quad_halves_<code>v16qi): Likewise.
2287 (reduc_splus_v2di): Likewise.
2288 (neon_vpadd_internal<mode>): Likewise.
2289 (neon_vpsmin<mode>): Likewise.
2290 (neon_vpsmax<mode>): Likewise.
2291 (neon_vpumin<mode>): Likewise.
2292 (neon_vpumax<mode>): Likewise.
2293 (*ss_add<mode>_neon): Likewise.
2294 (*us_add<mode>_neon): Likewise.
2295 (*ss_sub<mode>_neon): Likewise.
2296 (*us_sub<mode>_neon): Likewise.
2297 (neon_vadd<mode>_unspec): Likewise.
2298 (neon_vaddl<mode>): Likewise.
2299 (neon_vaddw<mode>): Likewise.
2300 (neon_vhadd<mode>): Likewise.
2301 (neon_vqadd<mode>): Likewise.
2302 (neon_vaddhn<mode>): Likewise.
2303 (neon_vmul<mode>): Likewise.
2304 (neon_vfms<VCVTF:mode>): Likewise.
2305 (neon_vmlal<mode>): Likewise.
2306 (neon_vmls<mode>): Likewise.
2307 (neon_vmlsl<mode>): Likewise.
2308 (neon_vqdmulh<mode>): Likewise.
2309 (neon_vqdmlal<mode>): Likewise.
2310 (neon_vqdmlsl<mode>): Likewise.
2311 (neon_vmull<mode>): Likewise.
2312 (neon_vqdmull<mode>): Likewise.
2313 (neon_vsub<mode>_unspec): Likewise.
2314 (neon_vsubl<mode>): Likewise.
2315 (neon_vsubw<mode>): Likewise.
2316 (neon_vqsub<mode>): Likewise.
2317 (neon_vhsub<mode>): Likewise.
2318 (neon_vsubhn<mode>): Likewise.
2319 (neon_vceq<mode>): Likewise.
2320 (neon_vcge<mode>): Likewise.
2321 (neon_vcgeu<mode>): Likewise.
2322 (neon_vcgt<mode>): Likewise.
2323 (neon_vcgtu<mode>): Likewise.
2324 (neon_vcle<mode>): Likewise.
2325 (neon_vclt<mode>): Likewise.
2326 (neon_vcage<mode>): Likewise.
2327 (neon_vcagt<mode>): Likewise.
2328 (neon_vtst<mode>): Likewise.
2329 (neon_vabd<mode>): Likewise.
2330 (neon_vabdl<mode>): Likewise.
2331 (neon_vaba<mode>): Likewise.
2332 (neon_vabal<mode>): Likewise.
2333 (neon_vmax<mode>): Likewise.
2334 (neon_vmin<mode>): Likewise.
2335 (neon_vpaddl<mode>): Likewise.
2336 (neon_vpadal<mode>): Likewise.
2337 (neon_vpmax<mode>): Likewise.
2338 (neon_vpmin<mode>): Likewise.
2339 (neon_vrecps<mode>): Likewise.
2340 (neon_vrsqrts<mode>): Likewise.
2341 (neon_vqabs<mode>): Likewise.
2342 (neon_vqneg<mode>): Likewise.
2343 (neon_vcls<mode>): Likewise.
2344 (clz<mode>2): Likewise.
2345 (popcount<mode>2): Likewise.
2346 (neon_vrecpe<mode>): Likewise.
2347 (neon_vrsqrte<mode>): Likewise.
2348 (neon_vget_lane<mode>_sext_internal): Likewise.
2349 (neon_vget_lane<mode>_zext_internal): Likewise.
2350 (neon_vdup_n<mode>): Likewise.
2351 (neon_vdup_n<mode>): Likewise.
2352 (neon_vdup_nv2di): Likewise.
2353 (neon_vdup_lane<mode>_interal): Likewise.
2354 (*neon_vswp<mode>): Likewise.
2355 (neon_vcombine<mode>): Likewise.
2356 (float<mode><V_cvtto>2): Likewise.
2357 (floatuns<mode><V_cvtto>2): Likewise.
2358 (fix_trunc<mode><V_cvtto>2): Likewise.
2359 (fixuns_trunc<mode><V_cvtto>2
2360 (neon_vcvt<mode>): Likewise.
2361 (neon_vcvt<mode>): Likewise.
2362 (neon_vcvtv4sfv4hf): Likewise.
2363 (neon_vcvtv4hfv4sf): Likewise.
2364 (neon_vcvt_n<mode>): Likewise.
2365 (neon_vcvt_n<mode>): Likewise.
2366 (neon_vmovn<mode>): Likewise.
2367 (neon_vqmovn<mode>): Likewise.
2368 (neon_vqmovun<mode>): Likewise.
2369 (neon_vmovl<mode>): Likewise.
2370 (neon_vmul_lane<mode>): Likewise.
2371 (neon_vmul_lane<mode>): Likewise.
2372 (neon_vmull_lane<mode>): Likewise.
2373 (neon_vqdmull_lane<mode>): Likewise.
2374 (neon_vqdmulh_lane<mode>): Likewise.
2375 (neon_vqdmulh_lane<mode>): Likewise.
2376 (neon_vmla_lane<mode>): Likewise.
2377 (neon_vmla_lane<mode>): Likewise.
2378 (neon_vmlal_lane<mode>): Likewise.
2379 (neon_vqdmlal_lane<mode>): Likewise.
2380 (neon_vmls_lane<mode>): Likewise.
2381 (neon_vmls_lane<mode>): Likewise.
2382 (neon_vmlsl_lane<mode>): Likewise.
2383 (neon_vqdmlsl_lane<mode>): Likewise.
2384 (neon_vext<mode>): Likewise.
2385 (neon_vrev64<mode>): Likewise.
2386 (neon_vrev32<mode>): Likewise.
2387 (neon_vrev16<mode>): Likewise.
2388 (neon_vbsl<mode>_internal): Likewise.
2389 (neon_vshl<mode>): Likewise.
2390 (neon_vqshl<mode>): Likewise.
2391 (neon_vshr_n<mode>): Likewise.
2392 (neon_vshrn_n<mode>): Likewise.
2393 (neon_vqshrn_n<mode>): Likewise.
2394 (neon_vqshrun_n<mode>): Likewise.
2395 (neon_vshl_n<mode>): Likewise.
2396 (neon_vqshl_n<mode>): Likewise.
2397 (neon_vqshlu_n<mode>): Likewise.
2398 (neon_vshll_n<mode>): Likewise.
2399 (neon_vsra_n<mode>): Likewise.
2400 (neon_vsri_n<mode>): Likewise.
2401 (neon_vsli_n<mode>): Likewise.
2402 (neon_vtbl1v8qi): Likewise.
2403 (neon_vtbl2v8qi): Likewise.
2404 (neon_vtbl3v8qi): Likewise.
2405 (neon_vtbl4v8qi): Likewise.
2406 (neon_vtbl1v16qi): Likewise.
2407 (neon_vtbl2v16qi): Likewise.
2408 (neon_vcombinev16qi): Likewise.
2409 (neon_vtbx1v8qi): Likewise.
2410 (neon_vtbx2v8qi): Likewise.
2411 (neon_vtbx3v8qi): Likewise.
2412 (neon_vtbx4v8qi): Likewise.
2413 (*neon_vtrn<mode>_insn): Likewise.
2414 (*neon_vzip<mode>_insn): Likewise.
2415 (*neon_vuzp<mode>_insn): Likewise.
2416 (neon_vld1<mode>): Likewise.
2417 (neon_vld1_lane<mode>): Likewise.
2418 (neon_vld1_lane<mode>): Likewise.
2419 (neon_vld1_dup<mode>): Likewise.
2420 (neon_vld1_dup<mode>): Likewise.
2421 (neon_vld1_dupv2di): Likewise.
2422 (neon_vst1<mode>): Likewise.
2423 (neon_vst1_lane<mode>): Likewise.
2424 (neon_vst1_lane<mode>): Likewise.
2425 (neon_vld2<mode>): Likewise.
2426 (neon_vld2<mode>): Likewise.
2427 (neon_vld2_lane<mode>): Likewise.
2428 (neon_vld2_lane<mode>): Likewise.
2429 (neon_vld2_dup<mode>): Likewise.
2430 (neon_vst2<mode>): Likewise.
2431 (neon_vst2<mode>): Likewise.
2432 (neon_vst2_lane<mode>): Likewise.
2433 (neon_vst2_lane<mode>): Likewise.
2434 (neon_vld3<mode>): Likewise.
2435 (neon_vld3qa<mode>): Likewise.
2436 (neon_vld3qb<mode>): Likewise.
2437 (neon_vld3_lane<mode>): Likewise.
2438 (neon_vld3_lane<mode>): Likewise.
2439 (neon_vld3_dup<mode>): Likewise.
2440 (neon_vst3<mode>): Likewise.
2441 (neon_vst3qa<mode>): Likewise.
2442 (neon_vst3qb<mode>): Likewise.
2443 (neon_vst3_lane<mode>): Likewise.
2444 (neon_vst3_lane<mode>): Likewise.
2445 (neon_vld4<mode>): Likewise.
2446 (neon_vld4qa<mode>): Likewise.
2447 (neon_vld4qb<mode>): Likewise.
2448 (neon_vld4_lane<mode>): Likewise.
2449 (neon_vld4_lane<mode>): Likewise.
2450 (neon_vld4_dup<mode>): Likewise.
2451 (neon_vst4<mode>): Likewise.
2452 (neon_vst4qa<mode>): Likewise.
2453 (neon_vst4qb<mode>): Likewise.
2454 (neon_vst4_lane<mode>): Likewise.
2455 (neon_vst4_lane<mode>): Likewise.
2456 (neon_vec_unpack<US>_lo_<mode>): Likewise.
2457 (neon_vec_unpack<US>_hi_<mode>): Likewise.
2458 (neon_vec_<US>mult_lo_<mode>): Likewise.
2459 (neon_vec_<US>mult_hi_<mode>): Likewise.
2460 (neon_vec_<US>shiftl_<mode>): Likewise.
2461 (neon_unpack<US>_<mode>): Likewise.
2462 (neon_vec_<US>mult_<mode>): Likewise.
2463 (vec_pack_trunc_<mode>): Likewise.
2464 (neon_vec_pack_trunc_<mode>): Likewise.
2465 (neon_vabd<mode>_2): Likewise.
2466 (neon_vabd<mode>_3): Likewise.
2468 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
2470 * config/aarch64/aarch64.md (movtf_aarch64): Update type attribute.
2471 (load_pair): Update type attribute.
2472 (store_pair): Update type attribute.
2473 * config/aarch64/iterators.md (q): New.
2475 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
2477 * config/arm/types.md: Add new types for Neon insns.
2479 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
2480 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2481 Sergey Lega <sergey.s.lega@intel.com>
2482 Anna Tikhonova <anna.tikhonova@intel.com>
2483 Ilya Tocar <ilya.tocar@intel.com>
2484 Andrey Turetskiy <andrey.turetskiy@intel.com>
2485 Ilya Verbin <ilya.verbin@intel.com>
2486 Kirill Yukhin <kirill.yukhin@intel.com>
2487 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2489 * config/i386/sse.md (unspec): Add UNSPEC_RCP14, UNSPEC_RSQRT14,
2490 UNSPEC_FIXUPIMM, UNSPEC_SCALEF, UNSPEC_GETEXP, UNSPEC_GETMANT,
2491 UNSPEC_EXP2, UNSPEC_RCP28, UNSPEC_RSQRT28.
2493 (srcp14<mode>): Ditto.
2494 (rsqrt14<mode>): Ditto.
2495 (rsqrt14<mode>): Ditto.
2496 (avx512f_vmscalef<mode>): Ditto.
2497 (avx512f_scalef<mode>): Ditto.
2498 (avx512f_getexp<mode>): Ditto.
2499 (avx512f_sgetexp<mode>): Ditto.
2500 (avx512f_fixupimm<mode>): Ditto.
2501 (avx512f_sfixupimm<mode>): Ditto.
2502 (avx512f_rndscale<mode>): Ditto.
2503 (*avx512er_exp2<mode>): Ditto.
2504 (*avx512er_rcp28<mode>): Ditto.
2505 (avx512er_rsqrt28<mode>): Ditto.
2506 (avx512f_getmant<mode>): Ditto.
2507 (avx512f_getmant<mode>): Ditto.
2508 (avx512f_rndscale<mode>): Fix formatting.
2510 2013-10-15 Martin Jambor <mjambor@suse.cz>
2512 * ipa-utils.h (ipa_edge_within_scc): Declare.
2513 * ipa-cp.c (edge_within_scc): Moved...
2514 * ipa-utils.c (ipa_edge_within_scc): ...here. Updated all callers.
2516 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
2517 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2518 Sergey Lega <sergey.s.lega@intel.com>
2519 Anna Tikhonova <anna.tikhonova@intel.com>
2520 Ilya Tocar <ilya.tocar@intel.com>
2521 Andrey Turetskiy <andrey.turetskiy@intel.com>
2522 Ilya Verbin <ilya.verbin@intel.com>
2523 Kirill Yukhin <kirill.yukhin@intel.com>
2524 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2526 * config/i386/predicates.md (const_8_to_15_operand): New.
2527 (const_16_to_31_operand): Ditto.
2528 * config/i386/sse.md (V8FI): New.
2530 (reduc_splus_v8df): Ditto.
2531 (reduc_splus_v16sf): Ditto.
2532 (avx512f_vextract<shuffletype>32x4_1): Ditto.
2533 (vec_extract_hi_<mode>): Ditto.
2534 (avx512f_vinsert<shuffletype>32x4_1): Ditto.
2535 (vec_set_lo_<mode>): Ditto.
2536 (vec_set_hi_<mode>): Ditto.
2537 (avx512f_shuf_<shuffletype>64x2_1): Ditto.
2538 (avx512f_shuf_<shuffletype>32x4_1): Ditto.
2539 (avx512f_pshufd_1): Ditto.
2540 (avx512f_broadcast<mode>): Ditto.
2541 (avx512f_broadcast<mode>): Ditto.
2542 (define_split): Split vec_extract_lo into move.
2543 (ssequartermode): Ditto.
2544 (ssedoublemode): Extened with wider modes.
2545 (vec_extract_lo_<mode>): Ditto.
2547 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
2548 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2549 Sergey Lega <sergey.s.lega@intel.com>
2550 Anna Tikhonova <anna.tikhonova@intel.com>
2551 Ilya Tocar <ilya.tocar@intel.com>
2552 Andrey Turetskiy <andrey.turetskiy@intel.com>
2553 Ilya Verbin <ilya.verbin@intel.com>
2554 Kirill Yukhin <kirill.yukhin@intel.com>
2555 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2557 * config/i386/predicates.md (register_or_constm1_operand): New.
2558 * config/i386/sse.md (unspec): Add UNSPEC_UNSIGNED_PCMP, UNSPEC_TESTM,
2559 UNSPEC_TESTNM, UNSPEC_VTERNLOG, UNSPEC_ALIGN, UNSPEC_CONFLICT,
2560 UNSPEC_MASKED_EQ, UNSPEC_MASKED_GT, UNSPEC_GATHER_PREFETCH,
2561 UNSPEC_SCATTER_PREFETCH
2563 (avx512f_ucmp<mode>3): Ditto.
2564 (avx512f_vternlog<mode>): Ditto.
2565 (avx512f_align<mode>): Ditto.
2566 (<shift_insn><mode>3): Ditto.
2567 (avx512f_<rotate>v<mode>): Ditto.
2568 (avx512f_<rotate><mode>): Ditto.
2569 (avx512f_eq<mode>3): Ditto.
2570 (avx512f_eq<mode>3_1): Ditto.
2571 (avx512f_gt<mode>3): Ditto.
2572 (avx512f_testm<mode>3): Ditto.
2573 (avx512f_testnm<mode>3): Ditto.
2574 (avx512pf_gatherpf<mode>): Ditto.
2575 (*avx512pf_gatherpf<mode>_mask): Ditto.
2576 (*avx512pf_gatherpf<mode>): Ditto.
2577 (avx512pf_scatterpf<mode>): Ditto.
2578 (*avx512pf_scatterpf<mode>_mask): Ditto.
2579 (*avx512pf_scatterpf<mode>): Ditto.
2580 (avx512f_vec_dup_gpr<mode>): Ditto.
2581 (clz<mode>2): Ditto.
2582 (conflict<mode>): Ditto.
2583 (REDUC_SMINMAX_MODE): Extened with wider modes.
2584 (reduc_<code>_<mode>): Ditto.
2585 (vlshr<mode>3): Ditto.
2586 (vashl<mode>3): Ditto.
2588 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
2589 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2590 Sergey Lega <sergey.s.lega@intel.com>
2591 Anna Tikhonova <anna.tikhonova@intel.com>
2592 Ilya Tocar <ilya.tocar@intel.com>
2593 Andrey Turetskiy <andrey.turetskiy@intel.com>
2594 Ilya Verbin <ilya.verbin@intel.com>
2595 Kirill Yukhin <kirill.yukhin@intel.com>
2596 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2598 * config/i386/sse.md (unspec): Added UNSPEC_VPERMI2, UNSPEC_VPERMT2,
2601 (avx512fmaskmode): Ditto.
2602 (bcstscalarsuff): Ditto.
2603 (avx512f_blendm<mode>): Ditto.
2604 (cmp_imm_predicate): Ditto.
2605 (avx512f_cmp<mode>3): Ditto.
2606 (avx512f_vec_dup<mode>): Ditto.
2607 (avx512f_vec_dup_mem<mode>): Ditto.
2608 (avx512f_vpermi2var<mode>3): Ditto.
2609 (avx512f_vpermt2var<mode>3): Ditto.
2610 (vec_init<mode>): Ditto.
2611 (avx512f_gathersi<mode>): Ditto.
2612 (*avx512f_gathersi<mode>): Ditto.
2613 (*avx512f_gathersi<mode>_2): Ditto.
2614 (avx512f_gatherdi<mode>): Ditto.
2615 (*avx512f_gatherdi<mode>): Ditto.
2616 (*avx512f_gatherdi<mode>_2): Ditto.
2617 (avx512f_scattersi<mode>): Ditto.
2618 (*avx512f_scattersi<mode>): Ditto.
2619 (avx512f_scatterdi<mode>): Ditto.
2620 (*avx512f_scatterdi<mode>): Ditto.
2621 (sseintprefix): Extened with wider modes.
2622 (VEC_GATHER_IDXSI): Ditto.
2623 (VEC_GATHER_IDXDI): Ditto.
2624 (VEC_GATHER_SRCDI): Ditto.
2626 2013-10-15 Matthew Gretton-Dann <matthew.gretton-dann@arm.com>
2627 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2629 * config/arm/t-aprofile: New file.
2630 * config.gcc: Handle --with-multilib-list option.
2632 2013-10-15 Bernd Schmidt <bernds@codesourcery.com>
2634 * reload1.c (reloads_unique_chain_p): Ensure that r1 is
2637 2013-10-15 Richard Biener <rguenther@suse.de>
2639 * tree-loop-distribution.c (build_empty_rdg): Inline into
2641 (rdg_flag_vertex): Inline into single user.
2642 (rdg_flag_vertex_and_dependent): Likewise.
2643 (build_rdg_partition_for_vertex): Remove processed bitmap.
2644 (rdg_build_partitions): Simplify.
2646 2013-10-15 Richard Biener <rguenther@suse.de>
2648 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1):
2649 Restructure forwarding through conversions and copies to
2650 avoid performing copy-propagation the wrong way. Adjust
2651 recursion invocations.
2652 (forward_propagate_addr_expr): Add argument stating if we
2653 are recursing from a single-use.
2654 (ssa_forward_propagate_and_combine): Adjust.
2656 2013-10-14 David Malcolm <dmalcolm@redhat.com>
2658 * dumpfile.h (gcc::dump_manager): New class, to hold state
2659 relating to dumpfile management.
2660 (get_dump_file_name): Remove in favor of method of dump_manager.
2661 (dump_initialized_p): Likewise.
2662 (dump_start): Likewise.
2663 (dump_finish): Likewise.
2664 (dump_switch_p): Likewise.
2665 (dump_register): Likewise.
2666 (get_dump_file_info): Likewise.
2667 * context.c (gcc::context::context): Construct the dump_manager
2669 * context.h (gcc::context::get_dumps): New.
2670 (gcc::context::m_dumps): New.
2671 * coverage.c (coverage_init): Port to dump_manager API.
2672 * dumpfile.c (extra_dump_files): Convert to field of gcc::dump_manager.
2673 (extra_dump_files_in_use): Likewise.
2674 (extra_dump_files_alloced): Likewise.
2675 (gcc::dump_manager::dump_manager): New.
2676 (dump_register): Convert to...
2677 (gcc::dump_manager::dump_register): ...method, replacing
2678 function-static next_dump with m_next_dump field.
2679 (get_dump_file_info): Convert to...
2680 (gcc::dump_manager::get_dump_file_info): ...method.
2681 (get_dump_file_name): Convert to...
2682 (gcc::dump_manager::get_dump_file_name): ...method.
2683 (dump_start): Convert to...
2684 (gcc::dump_manager::dump_start): ...method.
2685 (dump_finish): Convert to...
2686 (gcc::dump_manager::dump_finish): ...method.
2687 (dump_begin): Replace body with...
2688 (gcc::dump_manager::dump_begin): ...new method.
2689 (dump_phase_enabled_p): Convert to...
2690 (gcc::dump_manager::dump_phase_enabled_p): ...method.
2691 (dump_phase_enabled_p): Convert to...
2692 (gcc::dump_manager::dump_phase_enabled_p): ...method.
2693 (dump_initialized_p): Convert to...
2694 (gcc::dump_manager::dump_initialized_p): ...method.
2695 (dump_flag_name): Replace body with...
2696 (gcc::dump_manager::dump_flag_name): ...new method.
2697 (dump_enable_all): Convert to...
2698 (gcc::dump_manager::dump_enable_all): ...new method.
2699 (opt_info_enable_passes): Convert to...
2700 (gcc::dump_manager::opt_info_enable_passes): ...new method.
2701 (dump_switch_p_1): Convert to...
2702 (gcc::dump_manager::dump_switch_p_1): ...new method.
2703 (dump_switch_p): Convert to...
2704 (gcc::dump_manager::dump_switch_p): ...new method.
2705 (opt_info_switch_p): Port to dump_manager API.
2706 (enable_rtl_dump_file): Likewise.
2707 * opts-global.c (handle_common_deferred_options): Port to new
2709 * passes.c (pass_manager::finish_optimization_passes): Likewise.
2710 (pass_manager::register_one_dump_file): Likewise.
2711 (pass_manager::register_pass): Likewise.
2712 (pass_init_dump_file): Likewise.
2713 (pass_fini_dump_file): Likewise.
2714 * statistics.c (statistics_early_init): Likewise.
2716 2013-10-14 Richard Biener <rguenther@suse.de>
2718 * gimple.c (gimple_canonical_types, canonical_type_hash_cache,
2719 iterative_hash_canonical_type, gimple_canonical_type_hash,
2720 gimple_canonical_types_compatible_p, gimple_canonical_type_eq,
2721 gimple_register_canonical_type, print_gimple_types_stats,
2722 free_gimple_type_tables): Move to lto/lto.c
2723 (gt-gimple.h): Do not include.
2724 * gimple.h (gimple_register_canonical_type,
2725 print_gimple_types_stats, free_gimple_type_tables): Remove.
2726 * Makefile.in (GTFILES): Remove gimple.c.
2728 2013-10-14 Travis Snoozy <quandary@remstate.com>
2731 * config/msp430/msp430.c (msp430_option_override): Correct thinko
2732 scanning for msp430x targets.
2734 2013-10-14 Eric Botcazou <ebotcazou@adacore.com>
2737 * config/sparc/sparc-protos.h (widen_mem_for_ldd_peep): Declare.
2738 (registers_ok_for_ldd_peep): Move around.
2739 * config/sparc/sparc.c (widen_mem_for_ldd_peep): New.
2740 * config/sparc/sparc.md (widening peepholes): Use it.
2742 2013-10-14 Richard Biener <rguenther@suse.de>
2746 * gimple.c (iterative_hash_canonical_type): Make sure to
2747 record the hash into the correct hashtable slot.
2749 2013-10-13 Eric Botcazou <ebotcazou@adacore.com>
2751 PR rtl-optimization/58662
2752 * combine.c (try_combine): Take into account death nodes on I2 when
2753 splitting a PARALLEL of two independent SETs. Fix dump message.
2755 2013-10-12 Oleg Endo <olegendo@gcc.gnu.org>
2758 * config/sh/sh_treg_combine.cc: New SH specific RTL pass.
2759 * config.gcc (SH extra_objs): Add sh_ifcvt.o.
2760 * config/sh/t-sh (sh_treg_combine.o): New entry.
2761 * config/sh/sh.c (sh_fixed_condition_code_regs): New function that
2762 implements the target hook TARGET_FIXED_CONDITION_CODE_REGS.
2763 (register_sh_passes): New function. Register sh_treg_combine pass.
2764 (sh_option_override): Invoke it.
2765 (sh_canonicalize_comparison): Handle op0_preserve_value.
2766 * sh.md (*cbranch_t"): Do not try to optimize missed test and branch
2767 opportunities. Canonicalize branch condition.
2768 (nott): Allow only if pseudos can be created for non-SH2A.
2770 2013-10-12 H.J. Lu <hongjiu.lu@intel.com>
2773 * config/i386/i386.c (ix86_copy_addr_to_reg): New function.
2774 (ix86_expand_movmem): Replace copy_addr_to_reg with
2775 ix86_copy_addr_to_reg.
2776 (ix86_expand_setmem): Likewise.
2778 2013-10-12 Alexander Monakov <amonakov@ispras.ru>
2780 * config/i386/i386.c (ix86_expand_sse_compare_and_jump): Use mode
2781 provided by ix86_fp_compare_mode instead of CCFPUmode.
2783 2013-10-12 James Greenhalgh <james.greenhalgh@arm.com>
2785 * config/aarch64/arm_neon.h
2786 (vtbx<1,3>_<psu>8): Fix register constriants.
2788 2013-10-11 Jeff Law <law@redhat.com>
2790 PR tree-optimization/58640
2791 * tree-ssa-threadupdate.c (mark_threaded_blocks): Truncate jump
2792 threading paths that cross over two loop entry points.
2794 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2796 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): Generalize to
2797 handle vector float as well.
2798 (*vsx_le_perm_load_v4si): Likewise.
2799 (*vsx_le_perm_store_v2di): Likewise.
2800 (*vsx_le_perm_store_v4si): Likewise.
2802 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2804 * config/rs6000/vector.md (vec_realign_load<mode>): Generate vperm
2805 directly to circumvent subtract from splat{31} workaround.
2806 * config/rs6000/rs6000-protos.h (altivec_expand_vec_perm_le): New
2808 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): New.
2809 * config/rs6000/altivec.md (define_c_enum "unspec"): Add
2810 UNSPEC_VPERM_X and UNSPEC_VPERM_UNS_X.
2811 (altivec_vperm_<mode>): Convert to define_insn_and_split to
2812 separate big and little endian logic.
2813 (*altivec_vperm_<mode>_internal): New define_insn.
2814 (altivec_vperm_<mode>_uns): Convert to define_insn_and_split to
2815 separate big and little endian logic.
2816 (*altivec_vperm_<mode>_uns_internal): New define_insn.
2817 (vec_permv16qi): Add little endian logic.
2819 2013-10-11 Marc Glisse <marc.glisse@inria.fr>
2821 * doc/extend.texi (returns_nonnull): Remove arguments.
2823 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2824 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2825 Sergey Lega <sergey.s.lega@intel.com>
2826 Anna Tikhonova <anna.tikhonova@intel.com>
2827 Ilya Tocar <ilya.tocar@intel.com>
2828 Andrey Turetskiy <andrey.turetskiy@intel.com>
2829 Ilya Verbin <ilya.verbin@intel.com>
2830 Kirill Yukhin <kirill.yukhin@intel.com>
2831 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2833 * config/i386/sse.md (VI48F_256_512): New.
2834 (avx2_permvar<mode>): Change to ...
2835 (<avx2_avx512f>_permvar<mode>): This.
2837 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2838 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2839 Sergey Lega <sergey.s.lega@intel.com>
2840 Anna Tikhonova <anna.tikhonova@intel.com>
2841 Ilya Tocar <ilya.tocar@intel.com>
2842 Andrey Turetskiy <andrey.turetskiy@intel.com>
2843 Ilya Verbin <ilya.verbin@intel.com>
2844 Kirill Yukhin <kirill.yukhin@intel.com>
2845 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2847 * config/i386/i386.c (bdesc_args): Change corresponding pattern for
2848 __builtin_ia32_cvtps2dq, __builtin_ia32_cvtps2dq256.
2849 * config/i386/sse.md (VI4_AVX): New.
2850 (sf2simodelower): Ditto.
2851 (sse2_cvtps2dq): Change to ...
2852 (<sse2_avx_avx512f>_fix_notrunc<sf2simodelower><mode>): This.
2854 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2855 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2856 Sergey Lega <sergey.s.lega@intel.com>
2857 Anna Tikhonova <anna.tikhonova@intel.com>
2858 Ilya Tocar <ilya.tocar@intel.com>
2859 Andrey Turetskiy <andrey.turetskiy@intel.com>
2860 Ilya Verbin <ilya.verbin@intel.com>
2861 Kirill Yukhin <kirill.yukhin@intel.com>
2862 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2864 * config/i386/sse.md (V_512): New.
2866 (vcond<V_512:mode><VF_512:mode>): Ditto.
2867 (vcond<V_512:mode><VI_512:mode>): Ditto.
2868 (vcondu<V_512:mode><VI_512:mode>): Ditto.
2870 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2871 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2872 Sergey Lega <sergey.s.lega@intel.com>
2873 Anna Tikhonova <anna.tikhonova@intel.com>
2874 Ilya Tocar <ilya.tocar@intel.com>
2875 Andrey Turetskiy <andrey.turetskiy@intel.com>
2876 Ilya Verbin <ilya.verbin@intel.com>
2877 Kirill Yukhin <kirill.yukhin@intel.com>
2878 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2880 * config/i386/i386.c (ix86_rtx_costs): Enable fma for TARGET_AVX512F.
2881 * config/i386/sse.md (FMAMODEM): Changed modes and conditions.
2883 (fma<mode>4): Removed condition.
2884 (fms<mode>4): Ditto.
2885 (fnma<mode>4): Ditto.
2886 (fnms<mode>4): Ditto.
2887 (fma4i_fmadd_<mode>): Ditto.
2888 (*fma_fmadd_<mode>): Ditto.
2889 (*fma_fmsub_<mode>): Ditto.
2890 (*fma_fnmadd_<mode>): Ditto.
2891 (*fma_fnmsub_<mode>): Ditto.
2892 (fmaddsub_<mode>): Allow for TARGET_AVX512F.
2893 (*fma_fmaddsub_<mode>): Ditto.
2894 (*fma_fmsubadd_<mode>): Ditto.
2895 (*fmai_fmadd_<mode>): Ditto.
2896 (*fmai_fmsub_<mode>): Ditto.
2897 (*fmai_fnmadd_<mode>): Ditto.
2898 (*fmai_fnmsub_<mode>): Ditto.
2900 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2901 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2902 Sergey Lega <sergey.s.lega@intel.com>
2903 Anna Tikhonova <anna.tikhonova@intel.com>
2904 Ilya Tocar <ilya.tocar@intel.com>
2905 Andrey Turetskiy <andrey.turetskiy@intel.com>
2906 Ilya Verbin <ilya.verbin@intel.com>
2907 Kirill Yukhin <kirill.yukhin@intel.com>
2908 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2910 * config/i386/sse.md (VI248_AVX2_8_AVX512F): New.
2911 (VI124_256): Changed to ...
2912 (VI124_256_48_512): This.
2913 (ssepackmode): Extended with wider modes.
2914 (<code><mode>3): Changed iterator.
2915 (*avx2_<code><mode>3): Ditto.
2916 (vec_pack_trunc_<mode>): Ditto.
2918 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2919 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2920 Sergey Lega <sergey.s.lega@intel.com>
2921 Anna Tikhonova <anna.tikhonova@intel.com>
2922 Ilya Tocar <ilya.tocar@intel.com>
2923 Andrey Turetskiy <andrey.turetskiy@intel.com>
2924 Ilya Verbin <ilya.verbin@intel.com>
2925 Kirill Yukhin <kirill.yukhin@intel.com>
2926 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2928 * config/i386/sse.md (VI124_AVX2_48_AVX512F): New.
2929 (VI8F_256_512): Ditto.
2930 (abs<mode>2): Changed iterator.
2931 (avx2_perm<mode>): Changed to ...
2932 (<avx2_avx512f>_perm<mode>): This.
2933 (avx2_perm<mode>_1): Changed to ...
2934 (<avx2_avx512f>_perm<mode>_1): This.
2936 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2937 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2938 Sergey Lega <sergey.s.lega@intel.com>
2939 Anna Tikhonova <anna.tikhonova@intel.com>
2940 Ilya Tocar <ilya.tocar@intel.com>
2941 Andrey Turetskiy <andrey.turetskiy@intel.com>
2942 Ilya Verbin <ilya.verbin@intel.com>
2943 Kirill Yukhin <kirill.yukhin@intel.com>
2944 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2946 * config/i386/sse.md (VI48_AVX512F): New.
2947 (VI48_AVX2): Changed to ...
2948 (VI48_AVX2_48_AVX512F): This.
2949 (avx2_ashrv<mode>): Changed to ...
2950 (<avx2_avx512f>_ashrv<mode>): This.
2951 (avx2_<shift_insn>v<mode>): Changed to ...
2952 (<avx2_avx512f>_<shift_insn>v<mode>): This.
2954 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2955 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2956 Sergey Lega <sergey.s.lega@intel.com>
2957 Anna Tikhonova <anna.tikhonova@intel.com>
2958 Ilya Tocar <ilya.tocar@intel.com>
2959 Andrey Turetskiy <andrey.turetskiy@intel.com>
2960 Ilya Verbin <ilya.verbin@intel.com>
2961 Kirill Yukhin <kirill.yukhin@intel.com>
2962 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2964 * config/i386/sse.md (VI4_AVX512F): New.
2965 (VI8_AVX2_AVX512F): Ditto.
2966 (mul<mode>3): Extended with wider modes.
2967 (*<sse4_1_avx2>_mul<mode>3): Ditto.
2968 (mul<mode>3): Ditto.
2969 (vec_widen_<s>mult_odd_<mode>): Ditto.
2971 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2972 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2973 Sergey Lega <sergey.s.lega@intel.com>
2974 Anna Tikhonova <anna.tikhonova@intel.com>
2975 Ilya Tocar <ilya.tocar@intel.com>
2976 Andrey Turetskiy <andrey.turetskiy@intel.com>
2977 Ilya Verbin <ilya.verbin@intel.com>
2978 Kirill Yukhin <kirill.yukhin@intel.com>
2979 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2981 * config/i386/sse.md (VI2_AVX512F): New.
2982 (VI124_AVX512F): Ditto.
2983 (sseunpackmode): Extended with wider modes.
2984 (sseunpackfltmode): Ditto.
2985 (vec_unpacks_float_hi_<mode>): Ditto.
2986 (vec_unpacks_float_lo_<mode>): Ditto.
2987 (vec_unpacku_float_hi_<mode>): Ditto.
2988 (vec_unpacku_float_lo_<mode>): Ditto.
2989 (vec_unpacks_lo_<mode>): Ditto.
2990 (vec_unpacks_hi_<mode>): Ditto.
2991 (vec_unpacku_lo_<mode>): Ditto.
2992 (vec_unpacku_hi_<mode>): Ditto.
2994 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2995 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2996 Sergey Lega <sergey.s.lega@intel.com>
2997 Anna Tikhonova <anna.tikhonova@intel.com>
2998 Ilya Tocar <ilya.tocar@intel.com>
2999 Andrey Turetskiy <andrey.turetskiy@intel.com>
3000 Ilya Verbin <ilya.verbin@intel.com>
3001 Kirill Yukhin <kirill.yukhin@intel.com>
3002 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
3004 * config/i386/i386.md (multdiv): New.
3005 (multdiv_mnemonic): Ditto.
3006 * config/i386/sse.md (<sse>_vmmul<mode>3): Changed to...
3007 (<sse>_vm<multdiv_mnemonic><mode>3): This.
3008 (<sse>_vmdiv<mode>3): Removed.
3010 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
3011 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
3012 Sergey Lega <sergey.s.lega@intel.com>
3013 Anna Tikhonova <anna.tikhonova@intel.com>
3014 Ilya Tocar <ilya.tocar@intel.com>
3015 Andrey Turetskiy <andrey.turetskiy@intel.com>
3016 Ilya Verbin <ilya.verbin@intel.com>
3017 Kirill Yukhin <kirill.yukhin@intel.com>
3018 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
3020 * config/i386/sse.md (V): Extended with wider modes.
3022 (ssehalfvecmode): Ditto.
3024 (ssepackfltmode): Ditto.
3025 (avx_vec_concat<mode>): Ditto.
3026 (V_256_512): New iterator.
3027 (VF2_512_256): Ditto.
3028 (si2dfmode): New attribute.
3029 (si2dfmodelower): Ditto.
3031 (concat_tg_mode): Ditto.
3032 (floatv4siv4df2): Changed to ...
3033 (float<si2dfmodelower><mode>2): This.
3034 (avx_cvtps2pd256): Changed to ...
3035 (<sse2_avx_avx512f>_cvtps2pd<avxsizesuffix>): This.
3036 (vec_pack_trunc_v4df): Changed to ...
3037 (vec_pack_trunc_<mode>): This.
3038 (avx_vpermil<mode>): Changed to ...
3039 (<sse2_avx_avx512f>_vpermil<mode>): This.
3040 (<fixsuffix>fix_truncv8dfv8si2): New.
3041 (vec_pack_sfix_trunc_v8df): Ditto.
3042 (avx512f_rndscale<mode>): Ditto.
3043 (avx512f_roundpd512): Ditto.
3044 (vec_pack_ufix_trunc_<mode>): Updated iterator.
3046 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
3047 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
3048 Sergey Lega <sergey.s.lega@intel.com>
3049 Anna Tikhonova <anna.tikhonova@intel.com>
3050 Ilya Tocar <ilya.tocar@intel.com>
3051 Andrey Turetskiy <andrey.turetskiy@intel.com>
3052 Ilya Verbin <ilya.verbin@intel.com>
3053 Kirill Yukhin <kirill.yukhin@intel.com>
3054 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
3056 * config/i386/i386.md (any_fix): New iterator.
3057 (fixsuffix): New attribute.
3058 * config/i386/sse.md (VF1): Extened with wider modes.
3062 (sseintvecmodelower): Ditto.
3063 (ssescalarmode): Ditto.
3064 (ssescalarnum): Ditto.
3066 (ssexmmmode): Ditto.
3067 (<fixsuffix>fix_truncv16sfv16si2): Ditto.
3068 (<sse>_rcp<mode>2): Change iterator.
3069 (rsqrt<mode>2): Ditto.
3070 (<sse>_rsqrt<mode>2): Ditto.
3071 (avx2_vec_dup<mode>): Ditto.
3072 (<sse4_1>_round<ssemodesuffix>_sfix<avxsizesuffix>): Ditto.
3073 (round<mode>2_sfix): Ditto.
3074 (avx2_pbroadcast<mode>): Ditto.
3075 (*andnot<mode>3): Handle XI mode.
3076 (*<code><mode>3): Ditto.
3077 (AVXTOSSEMODE): Removed.
3078 (avx_vpermil<mode>): Changed to ...
3079 (<sse2_avx_avx512f>_vpermil<mode>): This.
3081 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
3082 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
3083 Sergey Lega <sergey.s.lega@intel.com>
3084 Anna Tikhonova <anna.tikhonova@intel.com>
3085 Ilya Tocar <ilya.tocar@intel.com>
3086 Andrey Turetskiy <andrey.turetskiy@intel.com>
3087 Ilya Verbin <ilya.verbin@intel.com>
3088 Kirill Yukhin <kirill.yukhin@intel.com>
3089 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
3091 * config/i386/sse.md (<sse2>_movnt<mode>): Update constraint to "v".
3092 (<sse>_comi): Ditto.
3093 (<sse>_ucomi): Ditto.
3094 (sse_cvtss2siq_2): Ditto.
3095 (sse2_cvtsd2si): Ditto.
3096 (sse2_cvtsd2siq): Ditto.
3097 (sse2_cvttsd2si): Ditto.
3098 (sse2_cvttsd2siq): Ditto.
3099 (<shift_insn><mode>3): Ditto.
3100 (sse2_cvtsi2sdq): Update constraint and prefix.
3101 (sse_cvtsi2ss): Update prefix.
3102 (sse_cvtsi2ssq): Ditto.
3104 2013-10-11 Jakub Jelinek <jakub@redhat.com>
3106 * tree-vrp.c (infer_nonnull_range): Use is_gimple_call,
3107 ignore internal calls.
3109 2013-10-11 Richard Biener <rguenther@suse.de>
3111 * tree-pretty-print.c (dump_generic_node): Allow to dump both (D)
3112 and (ab) for SSA_NAMEs. Mark INTEGER_CSTs with (OVF) if
3113 TREE_OVERFLOW is set.
3115 2013-10-11 Thomas Schwinge <thomas@codesourcery.com>
3117 * tree.h (OMP_CLAUSE_CODE): Remove duplicate definition.
3119 * gimple.c: GIMPLE statements have subcodes, not sub-codes.
3120 * gimple.h: Likewise.
3122 * doc/generic.texi (OpenMP): OMP_CLAUSE_* are subcodes, not sub-codes.
3124 * doc/generic.texi (Adding new DECL node types): Explain *_CHECK
3127 * doc/gimple.texi (is_gimple_omp): Move into the correct section.
3129 * acinclude.m4 (gcc_GAS_FLAGS): Add more gcc_cv_as_flags overrides.
3130 * configure: Regenerate.
3132 2013-10-11 Jakub Jelinek <jakub@redhat.com>
3134 * tree-pretty-print.c (dump_omp_clause): Handle OMP_CLAUSE__LOOPTEMP_
3135 and new OpenMP 4.0 clauses, handle UDR OMP_CLAUSE_REDUCTION,
3136 formatting fixes, use pp_colon instead of pp_character (..., ':'),
3137 similarly pp_right_paren.
3138 (dump_generic_node): Handle OMP_DISTRIBUTE, OMP_TEAMS,
3139 OMP_TARGET_DATA, OMP_TARGET, OMP_TARGET_UPDATE, OMP_TASKGROUP,
3140 allow OMP_FOR_INIT to be NULL, handle OMP_ATOMIC_SEQ_CST.
3141 * tree.c (omp_clause_num_ops, omp_clause_code_name): Add OpenMP 4.0
3143 (omp_declare_simd_clauses_equal,
3144 omp_remove_redundant_declare_simd_attrs): New functions.
3145 (attribute_value_equal): Use omp_declare_simd_clauses_equal.
3146 (walk_tree_1): Handle new OpenMP 4.0 clauses.
3147 * tree.h (OMP_LOOP_CHECK): Define.
3148 (OMP_FOR_BODY, OMP_FOR_CLAUSES, OMP_FOR_INIT, OMP_FOR_COND,
3149 OMP_FOR_INCR, OMP_FOR_PRE_BODY): Use it.
3150 (OMP_TASKGROUP_BODY, OMP_TEAMS_BODY, OMP_TEAMS_CLAUSES,
3151 OMP_TARGET_DATA_BODY, OMP_TARGET_DATA_CLAUSES, OMP_TARGET_BODY,
3152 OMP_TARGET_CLAUSES, OMP_TARGET_UPDATE_CLAUSES, OMP_CLAUSE_SIZE,
3153 OMP_ATOMIC_SEQ_CST, OMP_CLAUSE_DEPEND_KIND, OMP_CLAUSE_MAP_KIND,
3154 OMP_CLAUSE_MAP_ZERO_BIAS_ARRAY_SECTION, OMP_CLAUSE_PROC_BIND_KIND,
3155 OMP_CLAUSE_REDUCTION_OMP_ORIG_REF, OMP_CLAUSE_ALIGNED_ALIGNMENT,
3156 OMP_CLAUSE_NUM_TEAMS_EXPR, OMP_CLAUSE_THREAD_LIMIT_EXPR,
3157 OMP_CLAUSE_DEVICE_ID, OMP_CLAUSE_DIST_SCHEDULE_CHUNK_EXPR,
3158 OMP_CLAUSE_SIMDLEN_EXPR): Define.
3159 (OMP_CLAUSE_DECL): Change range up to OMP_CLAUSE__LOOPTEMP_.
3160 (omp_remove_redundant_declare_simd_attrs): New prototype.
3161 * gimple.def (GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET,
3162 GIMPLE_OMP_TEAMS): New codes.
3163 (GIMPLE_OMP_RETURN): Use GSS_OMP_ATOMIC_STORE instead of GSS_BASE.
3164 * omp-low.c (struct omp_context): Add cancel_label and cancellable
3166 (target_nesting_level): New variable.
3167 (extract_omp_for_data): Handle GF_OMP_FOR_KIND_DISTRIBUTE and
3168 OMP_CLAUSE_DIST_SCHEDULE. Don't fallback to library implementation
3169 for collapse > 1 static schedule unless ordered.
3170 (get_ws_args_for): Add par_stmt argument. Handle combined loops.
3171 (determine_parallel_type): Adjust get_ws_args_for caller.
3172 (install_var_field): Handle mask & 4 for double indirection.
3173 (scan_sharing_clauses): Ignore shared clause on teams construct.
3174 Handle OMP_CLAUSE__LOOPTEMP_ and new OpenMP 4.0 clauses.
3175 (create_omp_child_function): If inside target or declare target
3176 constructs, set "omp declare target" attribute on the child function.
3177 (find_combined_for): New function.
3178 (scan_omp_parallel): Handle combined loops.
3179 (scan_omp_target, scan_omp_teams): New functions.
3180 (check_omp_nesting_restrictions): Check new OpenMP 4.0 nesting
3181 restrictions and set ctx->cancellable for cancellable constructs.
3182 (scan_omp_1_stmt): Call check_omp_nesting_restrictions also on
3183 selected builtin calls. Handle GIMPLE_OMP_TASKGROUP,
3184 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS.
3185 (build_omp_barrier): Add lhs argument, return gimple rather than tree.
3186 (omp_clause_aligned_alignment): New function.
3187 (lower_rec_simd_input_clauses): Only call SET_DECL_VALUE_EXPR on decls.
3188 (lower_rec_input_clauses): Add FD argument. Ignore shared clauses
3189 on teams constructs. Handle user defined reductions and new
3191 (lower_reduction_clauses): Don't set placeholder to address of ref
3192 if it has already the right type.
3193 (lower_send_clauses): Handle OMP_CLAUSE__LOOPTEMP_.
3194 (expand_parallel_call): Use the new non-_start suffixed builtins,
3195 handle OMP_CLAUSE_PROC_BIND, don't call the outlined function
3196 and GOMP_parallel_end after the call.
3197 (expand_task_call): Handle OMP_CLAUSE_DEPEND.
3198 (expand_omp_for_init_counts): Handle combined loops.
3199 (expand_omp_for_init_vars): Add inner_stmt argument, handle combined
3201 (expand_omp_for_generic): Likewise. Use GOMP_loop_end_cancel at the
3202 end of cancellable loops.
3203 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
3204 Likewise. Handle collapse > 1 loops.
3205 (expand_omp_simd): Handle combined loops.
3206 (expand_omp_for): Add inner_stmt argument, adjust callers of
3207 expand_omp_for* functions, use expand_omp_for_static*chunk even
3208 for collapse > 1 unless ordered.
3209 (expand_omp_sections): Use GOMP_sections_end_cancel at the end
3210 of cancellable sections.
3211 (expand_omp_single): Remove need_barrier variable, just rely on
3212 gimple_omp_return_nowait_p. Adjust build_omp_barrier caller.
3213 (expand_omp_synch): Allow GIMPLE_OMP_TASKGROUP and GIMPLE_OMP_TEAMS.
3214 (expand_omp_atomic_load, expand_omp_atomic_store,
3215 expand_omp_atomic_fetch_op): Handle gimple_omp_atomic_seq_cst_p.
3216 (expand_omp_target): New function.
3217 (expand_omp): Handle combined loops. Handle GIMPLE_OMP_TASKGROUP,
3218 GIMPLE_OMP_TEAMS, GIMPLE_OMP_TARGET.
3219 (build_omp_regions_1): Immediately close region for
3220 GF_OMP_TARGET_KIND_UPDATE.
3221 (maybe_add_implicit_barrier_cancel): New function.
3222 (lower_omp_sections): Adjust lower_rec_input_clauses caller. Handle
3224 (lower_omp_single): Likewise. Add clobber after the barrier.
3225 (lower_omp_taskgroup): New function.
3226 (lower_omp_for): Handle combined loops. Adjust
3227 lower_rec_input_clauses caller. Handle cancellation.
3228 (lower_depend_clauses): New function.
3229 (lower_omp_taskreg): Lower depend clauses. Adjust
3230 lower_rec_input_clauses caller. Add clobber after the call. Handle
3232 (lower_omp_target, lower_omp_teams): New functions.
3233 (lower_omp_1): Handle cancellation. Handle GIMPLE_OMP_TASKGROUP,
3234 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GOMP_barrier, GOMP_cancel
3235 and GOMP_cancellation_point calls.
3236 (lower_omp): Fold stmts inside of target region.
3237 (diagnose_sb_1, diagnose_sb_2): Handle GIMPLE_OMP_TASKGROUP,
3238 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
3239 * builtin-types.def (DEF_FUNCTION_TYPE_8): Document.
3240 (BT_FN_VOID_OMPFN_PTR_UINT,
3241 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG,
3242 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG,
3243 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT): Remove.
3244 (BT_FN_VOID_OMPFN_PTR_UINT_UINT_UINT,
3245 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_UINT,
3246 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG_UINT,
3247 BT_FN_BOOL_INT, BT_FN_BOOL_INT_BOOL, BT_FN_VOID_UINT_UINT,
3248 BT_FN_VOID_INT_PTR_SIZE_PTR_PTR_PTR,
3249 BT_FN_VOID_INT_OMPFN_PTR_SIZE_PTR_PTR_PTR,
3250 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR): New.
3251 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1,
3252 call_may_clobber_ref_p_1): Handle BUILT_IN_GOMP_BARRIER_CANCEL,
3253 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_LOOP_END_CANCEL,
3254 BUILT_IN_GOMP_SECTIONS_END_CANCEL. Don't handle
3255 BUILT_IN_GOMP_PARALLEL_END.
3256 * gimple-low.c (lower_stmt): Handle GIMPLE_OMP_TASKGROUP,
3257 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
3258 * gimple-pretty-print.c (dump_gimple_omp_for): Handle
3259 GF_OMP_FOR_KIND_DISTRIBUTE.
3260 (dump_gimple_omp_target, dump_gimple_omp_teams): New functions.
3261 (dump_gimple_omp_block): Handle GIMPLE_OMP_TASKGROUP.
3262 (dump_gimple_omp_return): Print lhs if it has any.
3263 (dump_gimple_omp_atomic_load, dump_gimple_omp_atomic_store): Handle
3264 gimple_omp_atomic_seq_cst_p.
3265 (pp_gimple_stmt_1): Handle GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET
3266 and GIMPLE_OMP_TEAMS.
3267 * langhooks.c (lhd_omp_mappable_type): New function.
3268 * tree-vectorizer.c (struct simd_array_to_simduid): Fix up comment.
3269 * langhooks.h (struct lang_hooks_for_types): Add omp_mappable_type
3271 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_MAP,
3272 GOVD_ALIGNED and GOVD_MAP_TO_ONLY.
3273 (enum omp_region_type): Add ORT_TEAMS, ORT_TARGET_DATA and ORT_TARGET.
3274 (struct gimplify_omp_ctx): Add combined_loop field.
3275 (gimplify_call_expr, gimplify_modify_expr): Don't call fold_stmt
3276 on stmts inside of target region.
3277 (is_gimple_stmt): Return true for OMP_DISTRIBUTE and OMP_TASKGROUP.
3278 (omp_firstprivatize_variable): Handle GOVD_MAP, GOVD_ALIGNED,
3279 ORT_TARGET and ORT_TARGET_DATA.
3280 (omp_add_variable): Avoid checks on readding var for GOVD_ALIGNED.
3282 (omp_notice_threadprivate_variable): Complain about threadprivate
3283 variables in target region.
3284 (omp_notice_variable): Complain about vars with non-mappable type
3285 in target region. Handle ORT_TEAMS, ORT_TARGET and ORT_TARGET_DATA.
3286 (omp_check_private): Ignore ORT_TARGET* regions.
3287 (gimplify_scan_omp_clauses, gimplify_adjust_omp_clauses_1,
3288 gimplify_adjust_omp_clauses): Handle new OpenMP 4.0 clauses.
3289 (find_combined_omp_for): New function.
3290 (gimplify_omp_for): Handle gimplification of combined loops.
3291 (gimplify_omp_workshare): Gimplify also OMP_TARGET, OMP_TARGET_DATA,
3293 (gimplify_omp_target_update): New function.
3294 (gimplify_omp_atomic): Handle OMP_ATOMIC_SEQ_CST.
3295 (gimplify_expr): Handle OMP_DISTRIBUTE, OMP_TARGET, OMP_TARGET_DATA,
3296 OMP_TARGET_UPDATE, OMP_TEAMS, OMP_TASKGROUP.
3297 (gimplify_body): If fndecl has "omp declare target" attribute, add
3298 implicit ORT_TARGET context around it.
3299 * tree.def (OMP_DISTRIBUTE, OMP_TEAMS, OMP_TARGET_DATA, OMP_TARGET,
3300 OMP_TASKGROUP, OMP_TARGET_UPDATE): New tree codes.
3301 * tree-nested.c (convert_nonlocal_reference_stmt,
3302 convert_local_reference_stmt, convert_gimple_call): Handle
3303 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
3304 * omp-builtins.def (BUILT_IN_GOMP_TASK): Use
3305 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR
3306 instead of BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT.
3307 (BUILT_IN_GOMP_TARGET, BUILT_IN_GOMP_TARGET_DATA,
3308 BUILT_IN_GOMP_TARGET_END_DATA, BUILT_IN_GOMP_TARGET_UPDATE,
3309 BUILT_IN_GOMP_TEAMS, BUILT_IN_BARRIER_CANCEL,
3310 BUILT_IN_GOMP_LOOP_END_CANCEL,
3311 BUILT_IN_GOMP_SECTIONS_END_CANCEL, BUILT_IN_OMP_GET_TEAM_NUM,
3312 BUILT_IN_OMP_GET_NUM_TEAMS, BUILT_IN_GOMP_TASKGROUP_START,
3313 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_PARALLEL_LOOP_STATIC,
3314 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC,
3315 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED,
3316 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME, BUILT_IN_GOMP_PARALLEL,
3317 BUILT_IN_GOMP_PARALLEL_SECTIONS, BUILT_IN_GOMP_CANCEL,
3318 BUILT_IN_GOMP_CANCELLATION_POINT): New built-ins.
3319 (BUILT_IN_GOMP_PARALLEL_LOOP_STATIC_START,
3320 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC_START,
3321 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED_START,
3322 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME_START,
3323 BUILT_IN_GOMP_PARALLEL_START, BUILT_IN_GOMP_PARALLEL_END,
3324 BUILT_IN_GOMP_PARALLEL_SECTIONS_START): Remove.
3325 * tree-inline.c (remap_gimple_stmt, estimate_num_insns):
3326 Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
3327 * gimple.c (gimple_build_omp_taskgroup, gimple_build_omp_target,
3328 gimple_build_omp_teams): New functions.
3329 (walk_gimple_op): Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and
3330 GIMPLE_OMP_TASKGROUP. Walk optional lhs on GIMPLE_OMP_RETURN.
3331 (walk_gimple_stmt, gimple_copy): Handle GIMPLE_OMP_TARGET,
3332 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
3333 * gimple.h (enum gf_mask): GF_OMP_FOR_KIND_DISTRIBUTE,
3334 GF_OMP_FOR_COMBINED, GF_OMP_FOR_COMBINED_INTO,
3335 GF_OMP_TARGET_KIND_MASK, GF_OMP_TARGET_KIND_REGION,
3336 GF_OMP_TARGET_KIND_DATA, GF_OMP_TARGET_KIND_UPDATE,
3337 GF_OMP_ATOMIC_SEQ_CST): New.
3338 (gimple_build_omp_taskgroup, gimple_build_omp_target,
3339 gimple_build_omp_teams): New prototypes.
3340 (gimple_has_substatements): Handle GIMPLE_OMP_TARGET,
3341 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
3342 (gimple_omp_subcode): Use GIMPLE_OMP_TEAMS instead of
3343 GIMPLE_OMP_SINGLE as end of range.
3344 (gimple_omp_return_set_lhs, gimple_omp_return_lhs,
3345 gimple_omp_return_lhs_ptr, gimple_omp_atomic_seq_cst_p,
3346 gimple_omp_atomic_set_seq_cst, gimple_omp_for_combined_p,
3347 gimple_omp_for_set_combined_p, gimple_omp_for_combined_into_p,
3348 gimple_omp_for_set_combined_into_p, gimple_omp_target_clauses,
3349 gimple_omp_target_clauses_ptr, gimple_omp_target_set_clauses,
3350 gimple_omp_target_kind, gimple_omp_target_set_kind,
3351 gimple_omp_target_child_fn, gimple_omp_target_child_fn_ptr,
3352 gimple_omp_target_set_child_fn, gimple_omp_target_data_arg,
3353 gimple_omp_target_data_arg_ptr, gimple_omp_target_set_data_arg,
3354 gimple_omp_teams_clauses, gimple_omp_teams_clauses_ptr,
3355 gimple_omp_teams_set_clauses): New inlines.
3356 (CASE_GIMPLE_OMP): Add GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS
3357 and GIMPLE_OMP_TASKGROUP.
3358 * tree-core.h (enum omp_clause_code): Add new OpenMP 4.0 clause codes.
3359 (enum omp_clause_depend_kind, enum omp_clause_map_kind,
3360 enum omp_clause_proc_bind_kind): New.
3361 (union omp_clause_subcode): Add depend_kind, map_kind and
3362 proc_bind_kind fields.
3363 * tree-cfg.c (make_edges): Handle GIMPLE_OMP_TARGET,
3364 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
3365 * langhooks-def.h (lhd_omp_mappable_type): New prototype.
3366 (LANG_HOOKS_OMP_MAPPABLE_TYPE): Define.
3367 (LANG_HOOKS_FOR_TYPES_INITIALIZER): Add it.
3369 2013-10-10 Teresa Johnson <tejohnson@google.com>
3371 * predict.c (tree_estimate_probability): Add new parameter
3372 for estimate_bb_frequencies.
3373 (estimate_bb_frequencies): Add new parameter to force estimation.
3374 (rebuild_frequencies): When max frequency in function is small,
3375 recompute counts from frequencies.
3376 * predict.h (estimate_bb_frequencies): New parameter.
3378 2013-10-10 David Malcolm <dmalcolm@redhat.com>
3380 * ipa-inline.c (ipa_inline): Fix leak of "order" when
3381 optimizations are disabled.
3383 2013-10-10 David Malcolm <dmalcolm@redhat.com>
3385 * coverage.c (coverage_finish): Fix leak of da_file_name.
3387 2013-10-10 Jan Hubicka <jh@suse.cz>
3389 * config/i386/x86-tune.def: Enable X86_TUNE_SSE_TYPELESS_STORES
3390 for generic, enable X86_TUNE_SSE_LOAD0_BY_PXOR for Bulldozer,
3393 2013-10-10 Jakub Jelinek <jakub@redhat.com>
3396 * stmt.c (expand_asm_operands): Add FALLTHRU_BB argument,
3397 if any labels are in FALLTHRU_BB, use a special label emitted
3398 immediately after the asm goto insn rather than label_rtx
3400 (expand_asm_stmt): Adjust caller.
3401 * cfgrtl.c (commit_one_edge_insertion): Force splitting of
3402 edge if the last insn in predecessor is a jump with single successor,
3403 but it isn't simplejump_p.
3405 2013-10-10 Richard Biener <rguenther@suse.de>
3407 PR tree-optimization/58656
3408 * tree-ssa-pre.c (phi_translate): Do not cache failed translations.
3410 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
3412 * gimplify.c: Include expr.h and tm_p.h for targets with special
3413 va-arg padding requirements.
3415 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
3417 * tree-flow.h: Move some prototypes to gimple.h.
3418 (gimple_fold_indirect_ref): Move prototype to gimple-fold.h.
3419 * gimple.h: Relocate some prototypes from tree-flow.h
3420 * builtins.c (std_gimplify_va_arg_expr, build_va_arg_indirect_ref):
3422 * gimplify.c (gimple_fold_indirect_ref): Move to gimple-fold.c.
3423 (build_va_arg_indirect_ref): Relocate and make static.
3424 (std_gimplify_va_arg_expr): Relocate here.
3425 * gimple-fold.c (gimple_fold_indirect_ref): Relocate here.
3426 * gimple-fold.h (gimple_fold_indirect_ref): Add prototype.
3428 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3430 * doc/md.texi: Document the mnemonic attribute.
3432 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3435 * gensupport.c (gen_mnemonic_attr): Handle (set (attr x) y) and
3436 (set_attr_alternative x ...) when searching for user defined
3439 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
3441 * config/aplha/alpha.c: Add gimple-ssa.h to include list.
3443 2013-10-09 Easwaran Raman <eraman@google.com>
3445 * params.def (PARAM_MIN_SIZE_FOR_STACK_SHARING): New param...
3446 * cfgexpand.c (defer_stack_allocation): ...use here
3447 * doc/invoke.texi: Add documentation for min-size-for-stack-sharing.
3449 2013-10-09 Zhenqiang Chen <zhenqiang.chen@arm.com>
3451 * tree-ssa-phiopts.c (rhs_is_fed_for_value_replacement): New function.
3452 (operand_equal_for_value_replacement): New function, extracted from
3453 value_replacement and enhanced to catch more cases.
3454 (value_replacement): Use operand_equal_for_value_replacement.
3456 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
3458 * loop-doloop.c (doloop_modify, doloop_optimize): Use
3459 get_max_loop_iterations.
3461 2013-10-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3463 * config/arm/aarch-common.c (arm_early_load_addr_dep):
3464 Place comment above function.
3466 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
3468 * tree-flow.h: Remove all remaining prototypes, enums and structs that
3469 are not related to tree-cfg.c.
3470 * tree-ssa-address.h: New file. Relocate prototypes.
3471 * tree-ssa-address.c: (struct mem_address): Relocate from tree-flow.h.
3472 (addr_for_mem_ref): New. Combine call to get_address_description and
3473 return addr_for_mem_ref.
3474 * expr.c (expand_expr_real_1): Use new addr_for_mem_ref routine.
3475 * tree-ssa-live.h: Adjust prototypes.
3476 * passes.c: Include tree-ssa-live.h.
3477 * gimple-pretty-print.h (gimple_dump_bb): Add prototype.
3478 * graphite.c (graphite_transform_loops): Make static.
3479 (graphite_transforms, gate_graphite_transforms, pass_data_graphite,
3480 make_pass_graphite, pass_data_graphite_transforms,
3481 make_pass_graphite_transforms): Relocate here from tree-ssa-loop.c.
3482 * ipa-pure-const.c (warn_function_noreturn): Make static.
3483 (execute_warn_function_noreturn, gate_warn_function_noreturn,
3484 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
3485 Relocate from tree-cfg.c
3486 * tree-cfg.c (tree_node_can_be_shared, gimple_empty_block_p): Make
3488 (execute_warn_function_noreturn, gate_warn_function_noreturn,
3489 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
3490 Move to ipa-pure-const.c.
3491 (execute_fixup_cfg, class pass_fixup_cfg, make_pass_fixup_cfg):
3492 Relocate from tree-optimize.c.
3493 * tree-optimize.c (execute_fixup_cfg, class pass_fixup_cfg,
3494 make_pass_fixup_cfg): Move to tree-cfg.c.
3495 * tree-chrec.h: (enum ev_direction): Relocate here from tree-flow.h.
3496 Relocate some prototypes.
3497 * tree-data-ref.h (tree_check_data_deps) Add prototype.
3498 * tree-dump.c (dump_function_to_file): Remove prototype.
3499 Add tree-flow.h to the include file.
3500 * tree-dump.h: Remove prototype.
3501 * tree-parloops.h: New File. Add prototypes.
3502 * tree-parloops.c (gate_tree_parallelize_loops, tree_parallelize_loops,
3503 pass_data_parallelize_loops, make_pass_parallelize_loops): Relocate
3504 from tree-ssa-loop.c.
3505 * tree-predcom.c (run_tree_predictive_commoning,
3506 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom):
3507 Relocate here from tree-ssa-loop.c.
3508 * tree-ssa-dom.c (tree_ssa_dominator_optimize) Don't call
3509 ssa_name_values.release ().
3510 * tree-ssa-threadedge.h: New File. Relocate prototypes here.
3511 (ssa_name_values): Relocate from tree-flow.h.
3512 * tree-ssa.h: Include tree-ssa-threadedge.h and tree-ssa-address.h.
3513 * tree-ssa-loop.c (run_tree_predictive_commoning,
3514 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom,
3515 graphite_transforms, gate_graphite_transforms, pass_data_graphite,
3516 make_pass_graphite, pass_data_graphite_transforms,
3517 make_pass_graphite_transforms, gate_tree_parallelize_loops,
3518 tree_parallelize_loops, pass_data_parallelize_loops,
3519 make_pass_parallelize_loops): Move to other files.
3520 * tree-vectorizer.h (lpeel_tree_duplicate_loop_to_edge_cfg): Prototype
3522 * tree.h: Remove prototypes from tree-address.c.
3524 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
3526 * tree-flow.h (tm_restart_node, gimple_df): Move to gimple-ssa.h.
3527 (struct int_tree_map): Move to tree-hasher.h
3528 (SCALE, LABEL, PERCENT): Move to gimple.h
3529 * tree-flow-inline.h: Delete. Move functions to other files.
3530 (unmodifiable_var_p, ref_contains_array_ref): Unused, so delete.
3531 * gimple-ssa.h (tm_restart_node, gimple_df): Relocate from tree-flow.h.
3532 (gimple_in_ssa_p, gimple_vop): Relocate from tree-flow-inline.h
3533 * gimple.h (imple_stmt_max_uid, set_gimple_stmt_max_uid,
3534 inc_gimple_stmt_max_uid, get_lineno): Relocate from tree-flow-inline.h.
3535 (SCALE, LABEL, PERCENT): Relocate from tree-flow.h
3536 * tree-hasher.h: Don't include tree-flow.h.
3537 (struct int_tree_map): Relocate from tree-flow.h.
3538 * tree-sra.c (contains_view_convert_expr_p): Relocate from
3539 tree-flow-inline.h and make static.
3540 * tree-ssa-alias.h (ranges_overlap_p): Relocate from
3542 * tree-ssa-operands.c (gimple_ssa_operands): Relocate from
3543 tree-flow-inline.h and make static.
3544 * tree.h (is_global_var, may_be_aliased): Relocate from
3546 * Makefile.in (GTFILES): Remove tree-flow.h and add gimple-ssa.h.
3547 * value-prof.c: No longer include tree-flow-inline.h.
3548 * tree-switch-conversion.c: No longer include tree-flow-inline.h.
3550 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
3552 * tree-flow.h: Move some protoypes. Include new tree-ssa-loop.h.
3553 (struct affine_iv, struct tree_niter_desc): Move to tree-ssa-loop.h.
3554 (enum move_pos): Move to tree-ssa-loop-im.h
3555 * cfgloop.h: Move some prototypes.
3556 (gcov_type_to_double_int): relocate from tree-ssa-loop.niter.c.
3557 * tree-flow-inline.h (loop_containing_stmt): Move to tree-ssa-loop.h.
3558 * tree-ssa-loop.h: New File. Include other tree-ssa-loop-*.h files.
3559 (struct affine_iv, struct tree_niter_desc): Relocate from tree-flow.h.
3560 (loop_containing_stmt): Relocate from tree-flow-inline.h.
3561 * tree-ssa-loop-ch.c: (do_while_loop_p): Make static.
3562 * tree-ssa-loop-im.c (for_each_index): Move to tree-ssa-loop.c.
3563 (enum move_pos): Relocate here.
3564 (lsm_tmp_name_add, gen_lsm_tmp_name, get_lsm_tmp_name): Move to
3566 (execute_sm_if_changed_flag_set): Change get_lsm_tmp_name call.
3567 (tree_ssa_loop_im, gate_tree_ssa_loop_im, pass_data_lim,
3568 make_pass_lim): Relocate here from tree-ssa-loop.c.
3569 * tree-ssa-loop-ivcanon.c (tree_num_loop_insns): Move to
3571 (loop_edge_to_cancel, unloop_loops): Make static.
3572 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
3573 make_pass_iv_canon): Relocate from tree-ssa-loop.c.
3574 (tree_complete_unroll, gate_tree_complete_unroll,
3575 pass_data_complete_unroll, make_pass_complete_unroll): Relocate here.
3576 (tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
3577 pass_data_complete_unrolli, make_pass_complete_unrolli): Relocate here.
3578 * tree-ssa-loop-ivopts.c: Remove local prototypes.
3579 (stmt_invariant_in_loop_p): Remove unused function.
3580 * tree-ssa-loop-ivopts.h: New file. Add prototypes.
3581 * tree-ssa-loop-manip.h: New file. Add prototypes.
3582 * tree-ssa-loop-niter.c (record_niter_bound): Move to cfgloop.c.
3583 (gcov_type_to_double_int): Move to cfgloop.h.
3584 (double_int_cmp, bound_index,
3585 estimate_numbers_of_iterations_loop): Make static.
3586 (estimated_loop_iterations): Factor out get_estimated_loop_iterations.
3587 (max_loop_iterations): Factor out get_max_loop_iterations.
3588 (estimated_loop_iterations_int, max_stmt_executions_int): Move to
3590 * tree-ssa-loop-niter.h: New file. Add prototypes.
3591 * tree-ssa-loop-prefetch.c (tree_ssa_loop_prefetch,
3592 gate_tree_ssa_loop_prefetch, pass_data_loop_prefetch,
3593 make_pass_loop_prefetch): Relocate from tree-ssa-loop.c.
3594 * tree-ssa-loop-unswitch.c (tree_ssa_loop_unswitch,
3595 gate_tree_ssa_loop_unswitch, pass_data_tree_unswitch,
3596 make_pass_tree_unswitch): Relocate from tree-ssa-loop.c.
3597 * tree-ssa-loop.c (tree_ssa_loop_im, gate_tree_ssa_loop_im,
3598 pass_data_lim, make_pass_lim): Move to tree-ssa-loop-im.c.
3599 (tree_ssa_loop_unswitch, gate_tree_ssa_loop_unswitch,
3600 pass_data_tree_unswitch, make_pass_tree_unswitch): Move.
3601 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
3602 make_pass_iv_canon, tree_complete_unroll, gate_tree_complete_unroll,
3603 pass_data_complete_unroll, make_pass_complete_unroll,
3604 tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
3605 pass_data_complete_unrolli, make_pass_complete_unrolli): Move to
3606 tree-ssa-loop-ivcanon.c.
3607 (tree_ssa_loop_prefetch, gate_tree_ssa_loop_prefetch,
3608 pass_data_loop_prefetch, make_pass_loop_prefetch): Move to
3609 tree-ssa-loop-prefetch.c.
3610 (for_each_index, lsm_tmp_name_add, gen_lsm_tmp_name): Relocate from
3612 (get_lsm_tmp_name): Relocate and add suffix parameter.
3613 (tree_num_loop_insns): Relocate from tree-ssa-ivcanon.c.
3614 * tree-scalar-evolution.h (simple_iv): Don't use affive_iv typedef.
3615 * cfgloop.c (record_niter_bound, estimated_loop_iterations_int,
3616 max_stmt_executions_int): Move from tree-ssa-loop-niter.c.
3617 (get_estimated_loop_iterations): Factor out accessor from
3618 estimated_loop_iterations in tree-ssa-loop-niter.c.
3619 (get_max_loop_iterations): Factor out accessor from
3620 _max_loop_iterations in tree-ssa-niter.c.
3621 * loop-unroll.c (decide_unroll_constant_iterations,
3622 decide_unroll_runtime_iterations, decide_peel_simple,
3623 decide_unroll_stupid): Use new get_* accessors.
3625 2013-10-09 Marc Glisse <marc.glisse@inria.fr>
3627 PR tree-optimization/20318
3628 * doc/extend.texi (returns_nonnull): New function attribute.
3629 * fold-const.c (tree_expr_nonzero_warnv_p): Look for returns_nonnull
3631 * tree-vrp.c (gimple_stmt_nonzero_warnv_p): Likewise.
3632 (stmt_interesting_for_vrp): Accept all GIMPLE_CALL.
3634 2013-10-09 Eric Botcazou <ebotcazou@adacore.com>
3637 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): Return
3638 false if both components are bitfields.
3640 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
3642 * config/aarch64/arm_neon.h (vclz_s8, vclz_s16, vclz_s32)
3643 (vclzq_s8, vclzq_s16, vclzq_s32, vclz_u8, vclz_u16, vclz_u32)
3644 (vclzq_u8, vclzq_u16, vclzq_u32): Replace ASM with C.
3645 * config/aarch64/aarch64.h
3646 (CLZ_DEFINED_VALUE_AT_ZERO): Macro fixed for clz.
3647 * config/aarch64/aarch64-simd-builtins.def
3648 (VAR1 (UNOP, clz, 0, v4si)): Replaced with iterator.
3650 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
3652 * config/aarch64/arm_neon.h (vadd_f64, vsub_f64): Implementation added.
3654 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
3656 * config/aarch64/arm_neon.h (vdiv_f64): Added.
3658 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
3660 * config/aarch64/arm_neon.h (vneg_f32): Asm replaced with C.
3661 (vneg_f64): New intrinsic.
3662 (vneg_s8): Asm replaced with C.
3663 (vneg_s16): Likewise.
3664 (vneg_s32): Likewise.
3665 (vneg_s64): New intrinsic.
3666 (vnegq_f32): Asm replaced with C.
3667 (vnegq_f64): Likewise.
3668 (vnegq_s8): Likewise.
3669 (vnegq_s16): Likewise.
3670 (vnegq_s32): Likewise.
3671 (vnegq_s64): Likewise.
3673 2013-10-09 Renlin Li <Renlin.Li@arm.com>
3675 * config/arm/arm.c (arm_output_mi_thunk): Use plus_constant.
3677 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3679 * config/s390/s390.c (s390_register_info_stdarg_fpr): Remove
3680 packed stack special handling.
3681 (s390_frame_info, s390_emit_prologue, s390_emit_epilogue): Switch
3682 back to fixed stack slots for FPRs saved due to stdarg.
3684 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3686 * config/s390/s390.c (s390_frame_info): Restructure function.
3688 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3690 * config/s390/s390.c (struct s390_frame_layout): New field
3692 (cfun_save_arg_fprs_p, cfun_gpr_save_slot): New macros.
3693 (s390_reg_clobbered_rtx, s390_regs_ever_clobbered): Change type of
3694 regs_ever_clobbered to char*.
3695 (s390_regs_ever_clobbered): Check crtl->saves_all_registers instead
3696 of cfun->has_nonlocal_label. Ignore frame related restore INSNs.
3697 (s390_register_info): Enable FPR save slots. Move/Copy some
3698 functionality into ...
3699 (s390_register_info_gprtofpr, s390_register_info_stdarg_fpr)
3700 (s390_register_info_stdarg_gpr, s390_optimize_register_info): New
3702 (s390_frame_info): Do gpr slot allocation here now. stdarg does
3703 not imply a stack frame.
3704 (s390_init_frame_layout): Remove variable clobbered_regs.
3705 (s390_update_register_info): Remove function.
3706 (s390_hard_regno_rename_ok): Call-saved regs without a save slot
3707 cannot be used for register renaming.
3708 (s390_hard_regno_scratch_ok): New function.
3709 (TARGET_HARD_REGNO_SCRATCH_OK): Define target hook.
3710 (s390_initial_elimination_offset): Change offset calculation of
3711 the return address pointer.
3712 (save_gprs): Deal with only r6 being saved from the call-saved regs.
3713 (restore_gprs): Set frame related flag.
3714 (s390_save_gprs_to_fprs, s390_restore_gprs_from_fprs): New functions.
3715 (s390_emit_prologue): Call s390_register_info instead of
3716 s390_update_frame_layout. Call s390_save_gprs_to_fprs.
3717 (s390_emit_epilogue): Call s390_restore_gprs_from_fprs.
3718 (s390_optimize_prologue): Call s390_optimize_register_info.
3719 Try to remove also FPR slot save/restore INSNs. Remove frame
3720 related flags from restore INSNs.
3722 2013-10-08 DJ Delorie <dj@redhat.com>
3724 * config/rl78/rl78-expand.md (movqi): use operands[] not operandN.
3727 * config/rl78/rl78.c (rl78_print_operand_1): Change %c to %C to
3728 avoid conflict with the MI use of %c.
3729 * config/rl78/rl78-real.md: change %c to %C throughout.
3730 * config/rl78/rl78-virt.md: Likewise.
3732 2013-10-08 Jan Hubicka <jh@suse.cz>
3734 * config/i386/i386.c (ix86_option_override_internal): Switch
3735 to SSE math for -ffast-math when target ISA supports SSE2.
3737 2013-10-08 Andrew MacLeod <amacleod@redhat.com>
3739 * tree-flow.h: Remove some prototypes.
3740 * tree.h: Remove some protypes, add a couple.
3741 * tree.c (using_eh_for_cleanups_flag, using_eh_for_cleanups,
3742 using_eh_for_cleanups_p): Add interface routines for front ends.
3743 * tree-eh.h: New file. Add protoptyes.
3744 * tree-eh.c (using_eh_for_cleanups_p, using_eh_for_cleanups): Delete.
3745 (add_stmt_to_eh_lp_fn): Make static.
3746 (lower_try_finally): Use new using_eh_for_cleanups_p.
3747 * emit-rtl.c: Include tree-eh.h.
3748 * gimple.h: Include tree-eh.h.
3750 2013-10-08 Marc Glisse <marc.glisse@inria.fr>
3752 PR tree-optimization/58480
3753 * tree-vrp.c (infer_nonnull_range): New function.
3754 (infer_value_range): Call infer_nonnull_range.
3756 2013-10-08 Dehao Chen <dehao@google.com>
3758 PR tree-optimization/58619
3759 * tree-inline.c (copy_phis_for_bb): Combine location data
3762 2013-10-08 Zhenqiang Chen <zhenqiang.chen@linaro.org>
3765 * config/arm/arm.c (arm_emit_ldrd_pop): Attach
3766 RTX_FRAME_RELATED_P on INSN.
3768 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3770 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const_le): New.
3771 (altivec_expand_vec_perm_const): Call it.
3773 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3775 * config/rs6000/vector.md (mov<mode>): Emit permuted move
3776 sequences for LE VSX loads and stores at expand time.
3777 * config/rs6000/rs6000-protos.h (rs6000_emit_le_vsx_move): New
3779 * config/rs6000/rs6000.c (rs6000_const_vec): New.
3780 (rs6000_gen_le_vsx_permute): New.
3781 (rs6000_gen_le_vsx_load): New.
3782 (rs6000_gen_le_vsx_store): New.
3783 (rs6000_gen_le_vsx_move): New.
3784 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): New.
3785 (*vsx_le_perm_load_v4si): New.
3786 (*vsx_le_perm_load_v8hi): New.
3787 (*vsx_le_perm_load_v16qi): New.
3788 (*vsx_le_perm_store_v2di): New.
3789 (*vsx_le_perm_store_v4si): New.
3790 (*vsx_le_perm_store_v8hi): New.
3791 (*vsx_le_perm_store_v16qi): New.
3792 (*vsx_xxpermdi2_le_<mode>): New.
3793 (*vsx_xxpermdi4_le_<mode>): New.
3794 (*vsx_xxpermdi8_le_V8HI): New.
3795 (*vsx_xxpermdi16_le_V16QI): New.
3796 (*vsx_lxvd2x2_le_<mode>): New.
3797 (*vsx_lxvd2x4_le_<mode>): New.
3798 (*vsx_lxvd2x8_le_V8HI): New.
3799 (*vsx_lxvd2x16_le_V16QI): New.
3800 (*vsx_stxvd2x2_le_<mode>): New.
3801 (*vsx_stxvd2x4_le_<mode>): New.
3802 (*vsx_stxvd2x8_le_V8HI): New.
3803 (*vsx_stxvd2x16_le_V16QI): New.
3805 2013-10-07 Renlin Li <Renlin.Li@arm.com>
3807 * config/arm/arm-cores.def (cortex-a53): Use cortex tuning.
3809 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3811 * config/s390/s390.c (s390_register_info): Make the call-saved FPR
3812 loop to work also for 31bit ABI.
3813 Save the stack pointer for frame_size > 0.
3815 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3817 * config/s390/s390.md ("tbegin", "tbegin_nofloat", "tbegin_retry")
3818 ("tbegin_retry_nofloat", "tend", "tabort", "tx_assist"): Remove
3819 constraint letters from expanders.
3820 ("tbegin_retry", "tbegin_retry_nofloat"): Change predicate of the
3821 retry count to general_operand.
3822 ("tabort"): Give operand 0 a mode.
3823 ("tabort_1"): Add mode and constraint letter for operand 0.
3824 * doc/extend.texi: Fix protoype of __builtin_non_tx_store.
3826 2013-10-04 Jeff Law <law@redhat.com>
3828 * tree-ssa-threadedge.c: Fix some trailing whitespace problems.
3830 * tree-ssa-threadedge.c (thread_through_normal_block): Broken
3832 (thread_across_edge): Here. Call it.
3834 2013-10-04 Cary Coutant <ccoutant@google.com>
3836 * dwarf2out.c (dw_sra_loc_expr): Release addr_table entries when
3837 discarding a location list expression (or a piece of one).
3839 2013-10-03 Jan Hubicka <jh@suse.cz>
3841 * config/i386/i386.c (ix86_issue_rate): Pentium4, Nocona has issue
3842 rate of 2. Core2, Corei7 and Haswell has issue rate of 4.
3843 (ix86_adjust_cost): Remove Atom case; fix core2/corei7/Haswell case.
3845 2013-10-03 Jan Hubicka <jh@suse.cz>
3847 * config/i386/i386.c (ix86_option_override_internal): Do not enable
3848 accumulate-outgoing-args when producing unwind info.
3850 2013-10-03 Wei Mi <wmi@google.com>
3852 * lra-constraints.c (insert_move_for_subreg): New function
3853 extracted from simplify_operand_subreg.
3854 (simplify_operand_subreg): Add reload for paradoxical subreg.
3856 2013-10-03 Rong Xu <xur@google.com>
3858 * ipa-inline-analysis.c (find_foldable_builtin_expect): Find
3859 the candidate of builtin_expect such that we should fix the
3860 size/time estimation.
3861 (estimate_function_body_sizes): Do the acutally size/time fix-up
3864 2013-10-03 Rong Xu <xur@google.com>
3866 * predict.c (tree_predict_by_opcode): Get the probability
3867 for builtin_expect from param builtin_expect_probability.
3868 * params.def (BUILTIN_EXPECT_PROBABILITY): New parameter.
3869 * predict.def (PRED_BUILTIN_EXPECT_RELAXED): Fix comments.
3870 * doc/invoke.texi: Add documentation for builtin-expect-probability.
3872 2013-10-03 Marc Glisse <marc.glisse@inria.fr>
3875 * common.opt (fcheck-new): Moved from c.opt. Make it 'Common'.
3876 * calls.c (alloca_call_p): Use get_callee_fndecl.
3877 * fold-const.c (tree_expr_nonzero_warnv_p): Handle operator new.
3878 * tree-vrp.c (gimple_stmt_nonzero_warnv_p, stmt_interesting_for_vrp):
3880 (vrp_visit_stmt): Remove duplicated code.
3882 2013-10-03 Michael Meissner <meissner@linux.vnet.ibm.com>
3884 * config/rs6000/rs6000-builtin.def (XSRDPIM): Use floatdf2,
3885 ceildf2, btruncdf2, instead of vsx_* name.
3887 * config/rs6000/vsx.md (vsx_add<mode>3): Change arithmetic
3888 iterators to only do V2DF and V4SF here. Move the DF code to
3889 rs6000.md where it is combined with SF mode. Replace <VSv> with
3890 just 'v' since only vector operations are handled with these insns
3891 after moving the DF support to rs6000.md.
3892 (vsx_sub<mode>3): Likewise.
3893 (vsx_mul<mode>3): Likewise.
3894 (vsx_div<mode>3): Likewise.
3895 (vsx_fre<mode>2): Likewise.
3896 (vsx_neg<mode>2): Likewise.
3897 (vsx_abs<mode>2): Likewise.
3898 (vsx_nabs<mode>2): Likewise.
3899 (vsx_smax<mode>3): Likewise.
3900 (vsx_smin<mode>3): Likewise.
3901 (vsx_sqrt<mode>2): Likewise.
3902 (vsx_rsqrte<mode>2): Likewise.
3903 (vsx_fms<mode>4): Likewise.
3904 (vsx_nfma<mode>4): Likewise.
3905 (vsx_copysign<mode>3): Likewise.
3906 (vsx_btrunc<mode>2): Likewise.
3907 (vsx_floor<mode>2): Likewise.
3908 (vsx_ceil<mode>2): Likewise.
3909 (vsx_smaxsf3): Delete scalar ops that were moved to rs6000.md.
3910 (vsx_sminsf3): Likewise.
3911 (vsx_fmadf4): Likewise.
3912 (vsx_fmsdf4): Likewise.
3913 (vsx_nfmadf4): Likewise.
3914 (vsx_nfmsdf4): Likewise.
3915 (vsx_cmpdf_internal1): Likewise.
3917 * config/rs6000/rs6000.h (TARGET_SF_SPE): Define macros to make it
3918 simpler to select whether a target has SPE or traditional floating
3919 point support in iterators.
3920 (TARGET_DF_SPE): Likewise.
3921 (TARGET_SF_FPR): Likewise.
3922 (TARGET_DF_FPR): Likewise.
3923 (TARGET_SF_INSN): Macros to say whether floating point support
3924 exists for a given operation for expanders.
3925 (TARGET_DF_INSN): Likewise.
3927 * config/rs6000/rs6000.c (Ftrad): New mode attributes to allow
3928 combining of SF/DF mode operations, using both traditional and VSX
3936 (abs<mode>2): Combine SF/DF modes using traditional floating point
3937 instructions. Add support for using the upper DF registers with
3938 VSX support, and SF registers with power8-vector support. Update
3939 expanders for operations supported by both the SPE and traditional
3940 floating point units.
3941 (abs<mode>2_fpr): Likewise.
3942 (nabs<mode>2): Likewise.
3943 (nabs<mode>2_fpr): Likewise.
3944 (neg<mode>2): Likewise.
3945 (neg<mode>2_fpr): Likewise.
3946 (add<mode>3): Likewise.
3947 (add<mode>3_fpr): Likewise.
3948 (sub<mode>3): Likewise.
3949 (sub<mode>3_fpr): Likewise.
3950 (mul<mode>3): Likewise.
3951 (mul<mode>3_fpr): Likewise.
3952 (div<mode>3): Likewise.
3953 (div<mode>3_fpr): Likewise.
3954 (sqrt<mode>3): Likewise.
3955 (sqrt<mode>3_fpr): Likewise.
3956 (fre<Fs>): Likewise.
3957 (rsqrt<mode>2): Likewise.
3958 (cmp<mode>_fpr): Likewise.
3959 (smax<mode>3): Likewise.
3960 (smin<mode>3): Likewise.
3961 (smax<mode>3_vsx): Likewise.
3962 (smin<mode>3_vsx): Likewise.
3963 (negsf2): Delete SF operations that are merged with DF.
3970 (fmasf4_fpr): Likewise.
3971 (fmssf4_fpr): Likewise.
3972 (nfmasf4_fpr): Likewise.
3973 (nfmssf4_fpr): Likewise.
3974 (sqrtsf2): Likewise.
3975 (rsqrtsf_internal1): Likewise.
3976 (smaxsf3): Likewise.
3977 (sminsf3): Likewise.
3978 (cmpsf_internal1): Likewise.
3979 (copysign<mode>3_fcpsgn): Add VSX/power8-vector support.
3980 (negdf2): Delete DF operations that are merged with SF.
3982 (nabsdf2): Likewise.
3988 (rsqrtdf_internal1): Likewise.
3989 (fmadf4_fpr): Likewise.
3990 (fmsdf4_fpr): Likewise.
3991 (nfmadf4_fpr): Likewise.
3992 (nfmsdf4_fpr): Likewise.
3993 (sqrtdf2): Likewise.
3994 (smaxdf3): Likewise.
3995 (smindf3): Likewise.
3996 (cmpdf_internal1): Likewise.
3997 (lrint<mode>di2): Use TARGET_<MODE>_FPR macro.
3998 (btrunc<mode>2): Delete separate expander, and combine with the
3999 insn and add VSX instruction support. Use TARGET_<MODE>_FPR.
4000 (btrunc<mode>2_fpr): Likewise.
4001 (ceil<mode>2): Likewise.
4002 (ceil<mode>2_fpr): Likewise.
4003 (floor<mode>2): Likewise.
4004 (floor<mode>2_fpr): Likewise.
4005 (fma<mode>4_fpr): Combine SF and DF fused multiply/add support.
4006 Add support for using the upper registers with VSX and
4007 power8-vector. Move insns to be closer to the define_expands. On
4008 VSX systems, prefer the traditional form of FMA over the VSX
4009 version, since the traditional form allows the target not to
4010 overlap with the inputs.
4011 (fms<mode>4_fpr): Likewise.
4012 (nfma<mode>4_fpr): Likewise.
4013 (nfms<mode>4_fpr): Likewise.
4015 2013-10-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4016 Richard Earnshaw <richard.earnshaw@arm.com>
4018 * config/arm/aarch-common-protos.h (struct alu_cost_table): New.
4019 (struct mult_cost_table): Likewise.
4020 (struct mem_cost_table): Likewise.
4021 (struct fp_cost_table): Likewise.
4022 (struct vector_cost_table): Likewise.
4023 (cpu_cost_table): Likewise.
4024 * config/arm/arm.opt (mold-rts-costs): New option.
4025 (mnew-generic-costs): Likewise.
4026 * config/arm/arm.c (generic_extra_costs): New table.
4027 (cortexa15_extra_costs): Likewise.
4028 (arm_slowmul_tune): Use NULL as new costs.
4029 (arm_fastmul_tune): Likewise.
4030 (arm_strongarm_tune): Likewise.
4031 (arm_xscale_tune): Likewise.
4032 (arm_9e_tune): Likewise.
4033 (arm_v6t2_tune): Likewise.
4034 (arm_cortex_a5_tune): Likewise.
4035 (arm_cortex_a9_tune): Likewise.
4036 (arm_v6m_tune): Likewise.
4037 (arm_fa726te_tune): Likewise.
4038 (arm_cortex_a15_tune): Use cortex15_extra_costs.
4039 (arm_cortex_tune): Use generict_extra_costs.
4040 (shifter_op_p): New function.
4041 (arm_unspec_cost): Likewise.
4042 (LIBCALL_COST): Define.
4043 (arm_new_rtx_costs): New function.
4044 (arm_rtx_costs): Use arm_new_rtx_costs when core-specific
4045 table is available. Use old costs otherwise unless mnew-generic-costs
4047 * config/arm/arm-protos.h (tune_params): Add insn_extra_cost field.
4048 (cpu_cost_table): Declare.
4050 2013-10-03 Marcus Shawcroft <marcus.shawcroft@arm.com>
4053 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>)
4054 (*subs_mul_imm_<mode>)
4055 (*add_<shift>_<mode>, *add_<shift>_si_uxtw,*add_mul_imm_<mode>)
4056 (*sub_<shift>_<mode>)
4057 (*sub_<shift>_si_uxtw,*sub_mul_imm_<mode>, *sub_mul_imm_si_uxtw):
4058 Remove k constraint.
4060 2013-10-03 Ian Bolton <ian.bolton@arm.com>
4062 * config/aarch64/aarch64.c (aarch64_secondary_reload): Remove legacy
4064 * config/aarch64/aarch64.md (reload_sp_immediate): Likewise.
4066 2013-10-02 Teresa Johnson <tejohnson@google.com>
4068 * predict.c (probably_never_executed): New function.
4069 (probably_never_executed_bb_p): Invoke probably_never_executed.
4070 (probably_never_executed_edge_p): Ditto.
4071 * bb-reorder.c (find_rarely_executed_basic_blocks_and_crossing_edges):
4072 Treat profile insanities conservatively.
4074 2013-10-02 John David Anglin <danglin@gcc.gnu.org>
4076 * config.gcc (hppa*64*-*-linux*): Don't add pa/t-linux to tmake_file.
4078 2013-10-02 Vladimir Makarov <vmakarov@redhat.com>
4080 * lra-constraints.c (process_alt_operand): Calculate scratch_p and
4081 use it. Use smaller increase for scratch. Don't increase reject
4082 for early clobber scratch.
4083 * lra-eliminations.c (eliminate_regs_in_insn): Remove all insns
4084 setting eliminated regs except setting fp from hfp.
4085 (lra_eliminate): Check lra_insn_recog_data on NULL.
4087 2013-10-02 Michael Meissner <meissner@linux.vnet.ibm.com>
4090 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Turn off
4091 setting -mvsx-timode by default until the underlying problem is fixed.
4092 (RS6000_CPU, power7 defaults): Likewise.
4094 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
4096 * config/x-linux (host-linux.o): Remove header dependencies.
4097 Use $(COMPILE) and $(POSTCOMPILE).
4098 * config/t-linux-android (linux-android.o): Ditto.
4100 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
4102 * Makefile.in (expmed.o-warn): Remove.
4104 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
4106 * graphite-scop-detection.c: Include tree-ssa-propagate,h.
4107 * graphite-sese-to-poly.c: Include tree-ssa-propagate.h.
4109 2013-10-02 Teresa Johnson <tejohnson@google.com>
4111 * dojump.c (do_jump_1): Divide probability between
4112 both conditions of a TRUTH_ANDIF_EXPR/TRUTH_ORIF_EXPR.
4114 2013-10-02 Tom Tromey <tromey@redhat.com>
4116 * Makefile.in (DRIVER_DEFINES): Use $(if), not $(and).
4118 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
4120 * tree-flow.h: Remove some prototypes.
4121 * tree-ssa-dce.c (mark_virtual_operand_for_renaming,
4122 mark_virtual_phi_result_for_renaming): Move to tree-into-ssa.c.
4123 * tree-into-ssa.c (mark_virtual_operand_for_renaming,
4124 mark_virtual_phi_result_for_renaming): Relocate here.
4125 * tree-into-ssa.h: Add prototypes.
4126 * tree-ssa-phiopt.c: (tree_ssa_phiopt_worker) Use
4127 single_pred_before_succ_order.
4128 (blocks_in_phiopt_order): Rename and move to cfganal.c.
4129 (nonfreeing_call_p) Move to gimple.c.
4130 * cfganal.c (single_pred_before_succ_order): Move and renamed from
4132 * basic-block.h (single_pred_before_succ_order): Add prototype.
4133 * gimple.c (nonfreeing_call_p): Relocate here.
4134 * gimple.h: Add prototype.
4135 * tree-ssa-ifcombine.c: Include tree-ssa-phiopt.h.
4136 * tree-ssa-dom.h: New file. Relocate prototypes here.
4137 * tree-ssa.h: Include tree-ssa-dom.h.
4139 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
4141 * config/i386/x-i386 (driver-i386.o): Remove header dependencies.
4142 Use $(COMPILE) and $(POSTCOMPILE).
4144 * config/alpha/x-alpha (driver-alpha.o): Ditto.
4146 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
4148 * tree-flow.h: Remove some prototypes.
4149 * gimple-fold.h: Add prototypes from gimple.h and tree-flow.h.
4150 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
4151 * tree-ssa-copy.c (may_propagate*, propagate_value, replace_exp,
4152 propagate_tree_value*): Move from here to...
4153 * tree-ssa-propagate.c (may_propagate*, propagate_value, replace_exp,
4154 propagate_tree_value*): Relocate here.
4155 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
4156 * gimple.h: Include gimple-fold.h, move prototypes into gimple-fold.h.
4157 * gimple-fold.c: Remove gimple-fold.h from include list.
4158 * tree-vrp.c: Remove gimple-fold.h from include list.
4159 * tree-ssa-sccvn.c: Remove gimple-fold.h from include list.
4160 * tree-ssa-ccp.c: Remove gimple-fold.h from include list.
4161 * tree-scalar-evolution.c: Add tree-ssa-propagate.h to include list.
4162 * tree-ssa-pre.c: Add tree-ssa-propagate.h to include list.
4163 * sese.c: Add tree-ssa-propagate.h to include list.
4165 2013-10-02 Richard Biener <rguenther@suse.de>
4167 * tree-loop-distribution.c: Include tree-vectorizer.h for
4169 (enum partition_kind): Remove PKIND_REDUCTION.
4170 (struct partition_s): Remove has_writes member, add reduction_p member.
4171 (partition_alloc): Adjust.
4172 (partition_builtin_p): Likewise.
4173 (partition_has_writes): Remove.
4174 (partition_reduction_p): New function.
4175 (partition_merge_into): Likewise.
4176 (generate_code_for_partition): Commonize builtin partition
4178 (rdg_cannot_recompute_vertex_p): Remove.
4179 (already_processed_vertex_p): Likewise.
4180 (rdg_flag_vertex): Do not set has_writes.
4181 (classify_partition): Adjust.
4182 (rdg_build_partitions): Do not set has_writes, treat all
4183 partitions as useful.
4184 (distribute_loop): Record number of library calls generated. Adjust.
4185 (tree_loop_distribution): Report number of loops and library
4186 calls generated as opt-info.
4188 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
4190 * tree-flow.h: Include new .h files. Move prototypes.
4191 * tree-cfgcleanup.h: New file. Add prototypes from tree-flow.h.
4192 * tree-dfa.h: New File. Add prototypes from tree-flow.h.
4193 (get_addr_base_and_unit_offset_1) Move from tree-flow-inline.h.
4194 * tree-pretty-print.h: Add prototypes from tree-flow.h.
4195 * tree-into-ssa.h: New File. Add prototypes from tree-flow.h.
4196 ({debug|dump}*): Move debugging prototypes out of tree-into-ssa.c.
4197 * tree-into-ssa.c ({debug|dump}*): Move prototypes to header file.
4198 * tree.h (get_ref_base_and_extent): Move prototype out.
4199 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Move to
4201 * gimple-low.h: New File. Add prototypes from tree-flow.h.
4202 * gimple-low.c (try_catch_may_fallthru, block_may_fallthru): Move to...
4203 * tree.c (try_catch_may_fallthru, block_may_fallthru): Here.
4204 * tree-scalar-evolution.c: Include tree.h.
4205 * sese.c: Include tree.h.
4206 * dumpfile.c: Move gimple-pretty-print.h include after tree.h.
4207 * dwarf2out.c: Include tree-dfa.h.
4208 * tree-chrec.c: Include tree.h.
4209 * tree-data-ref.c: Include tree.h.
4211 2013-10-02 Yufeng Zhang <yufeng.zhang@arm.com>
4213 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref):
4216 2013-10-02 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4218 * config/t-sol2 (sol2-c.o): Remove header dependencies.
4219 Use $(COMPILE) and $(POSTCOMPILE).
4220 (sol2-cxx.o): Likewise.
4221 (sol2-stubs.o): Likewise.
4223 * config/x-solaris (host-solaris.o): Likewise.
4225 * config/sparc/t-sparc (sparc.o): Remove.
4226 (sparc-c.o): Remove header dependencies.
4227 Use $(COMPILE) and $(POSTCOMPILE).
4228 * config/sparc/x-sparc: Likewise.
4230 2013-10-02 Joern Rennecke <joern.rennecke@embecosm.com>
4232 * config/arc/arc-opts.h: Add 2013 to Copyright years.
4233 * config/arc/arc700.md: Likewise.
4234 * config/arc/arc-modes.def: Likewise.
4235 * config/arc/arc-simd.h: Likewise.
4236 * config/arc/t-arc-uClibc: Likewise.
4237 * config/arc/t-arc-newlib: Likewise.
4239 2013-10-02 Renlin Li <renlin.li@arm.com>
4241 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
4243 (aarch64_expand_epilogue): Likewise.
4245 2013-10-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4246 Yufeng Zhang <yufeng.zhang@arm.com>
4248 * gimple-ssa-strength-reduction.c (legal_cast_p_1): Forward
4250 (backtrace_base_for_ref): Call get_unwidened with 'base_in' if
4251 'base_in' represent a conversion and legal_cast_p_1 holds; set
4252 'base_in' with the returned value from get_unwidened.
4254 2013-10-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4256 * config/arm/arm.c (arm_legitimize_reload_address): Explain why
4257 plus_constant is not used.
4259 2013-10-01 Wei Mi <wmi@google.com>
4261 * config/i386/x86-tune.def (DEF_TUNE): Remove m_CORE_ALL.
4262 * config/i386/i386.md: Add define_peephole2 to
4263 break partial reg stall for cvtss2sd/cvtsd2ss.
4265 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
4267 * config/arc/arc.c (pass_arc_ifcvt::clone):
4268 Update for ctxt_ -> m_ctxt change.
4270 2013-10-01 Jeff Law <law@redhat.com>
4272 * tree-ssa-threadupdate.c (struct redirection_data): Delete
4273 outgoing_edge and intermediate_edge fields. Instead store the path.
4274 (redirection_data::hash): Hash on the last edge's destination index.
4275 (redirection_data::equal): Check the entire thread path.
4276 (lookup_redirectio_data): Corresponding changes.
4277 (create_edge_and_update_destination_phis): Likewise.
4278 (thread_single_edge): Likewise.
4280 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
4281 Diego Novillo <dnovillo@google.com>
4283 * config/arc/simdext.md (UNSPEC_ARC_SIMD_VLD32WH): Delete.
4284 (UNSPEC_ARC_SIMD_VLD32WL): Likewise.
4285 (vld32wh_insn, vld32wl_insn): Delete commented-out old
4286 versions of these patterns.
4288 * doc/extend.texi (long_call/medium_call/short_call): Typo fix.
4289 (__builtin_arc_aligned): Likewise.
4291 * config/arc/arc.md: Expand adc_0 comment stating the intended
4292 purpose and why it isn't ready.
4293 Replace commented out call_value_via_label_mixed with a
4294 plain comment about bl_s.
4296 * config/arc/arc.c (stdio.h): Don't include directly.
4297 (arc_expand_epilogue): Remove [0]: Remove fp_restored_p.
4298 Remove if (1) condition.
4299 (arc_encode_section_info): Fix comment.
4301 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
4303 * config/arc/arc.c (arc_conditional_register_usage):
4304 Use ARC_FIRST_SIMD_VR_REG / ARC_LAST_SIMD_VR_REG.
4305 Also set reg_alloc_order for DMA config regs.
4307 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
4308 Jeremy Bennett <jeremy.bennett@embecosm.com>
4310 * doc/install.texi (--with-cpu): Mention ARC.
4311 (arc-*-elf32): New paragraph.
4312 (arc-linux-uclibc): Likewise.
4313 * doc/md.texi (Machine Constraints): Add ARC part.
4314 * doc/invoke.texi: (menu): Add ARC Options.
4315 (Machine Dependent Options) <ARC Options>: Add synopsis.
4316 (node ARC Options): Add.
4317 * doc/extend.texi (long_call / short_call attribute): Add ARC.
4318 (ARC Built-in Functions): New section defining
4319 generic ARC built-in functions.
4320 (ARC SIMD Built-in Functions): New section defining SIMD specific
4322 (Declaring Attributes of Functions): Extended
4323 description of short_call and long_call attributes for ARC and
4324 added index entries.
4326 2013-10-01 Saurabh Verma <saurabh.verma@codito.com>
4327 Ramana Radhakrishnan <ramana.radhakrishnan@codito.com>
4328 Joern Rennecke <joern.rennecke@embecosm.com>
4329 Muhammad Khurram Riaz <khurram.riaz@arc.com>
4330 Brendan Kehoe <brendan@zen.org>
4331 Michael Eager <eager@eagercon.com>
4332 Simon Cook <simon.cook@embecosm.com>
4333 Jeremy Bennett <jeremy.bennett@embecosm.com>
4335 * config/arc, common/config/arc: New directories.
4337 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
4338 Brendan Kehoe <brendan@zen.org>
4339 Simon Cook <simon.cook@embecosm.com>
4341 * config.gcc (arc*-*-elf*, arc*-*-linux-uclibc*): New configurations.
4343 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
4345 * tree-ssa-live.h (coalesce_ssa_name): Move Prototype to...
4346 * tree-ssa-coalesce.h: New. Move prototype to here.
4347 * tree-outof-ssa.h: Include tree-ssa-coalesce.h.
4348 * tree-ssa-coalesce.c: Include tree-outof-ssa.h.
4349 (gimple_can_coalesce_p): Move to...
4350 * gimple.c (gimple_can_coalesce_p): Here.
4352 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
4354 * tree-into-ssa.c (enum need_phi_state): Relocate from tree-flow.h.
4355 (dump_decl_set): Move to gimple.c.
4356 * gimple.h: Don't include tree-ssa-operands.h.
4357 (dump_decl_set): Add prototype.
4358 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
4359 Move to gimple-ssa.h.
4360 (phi_ssa_name_p, phi_nodes, phi_nodes_ptr, gimple_phi_arg_def,
4361 gimple_phi_arg_def_ptr, gimple_phi_arg_edge, gimple_phi_arg_location,
4362 gimple_phi_arg_location_from_edge, gimple_phi_arg_set_location,
4363 gimple_phi_arg_has_location): Relocate from tree-flow-inline.h
4364 * gimple.c (walk_stmt_load_store_ops): Use gimple_phi_arg_def rather
4366 (dump_decl_set): Relocate here.
4367 * gimple-ssa.h: New file.
4368 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
4369 Relocate from gimple.h.
4370 * tree-cfg.c (has_zero_uses_1, single_imm_use_1): Move to...
4371 * tree-ssa-operands.c (swap_ssa_operands): Rename from
4372 swap_tree_operands and remove non-ssa path.
4373 (has_zero_uses_1, single_imm_use_1): Relocate from tree-cfg.c.
4374 * tree-ssa-reassoc.c (linearize_expr_tree, repropagate_negates): Use
4376 * tree-vect-loop.c (destroy_loop_vec_info, vect_is_slp_reduction,
4377 vect_is_simple_reduction_1): Use swap_ssa_operands.
4378 * tree-flow.h: Move various prototypes to tree-phinodes.h.
4379 (enum need_phi_state): Move to tree-into-ssa.c.
4380 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
4381 BREAK_FROM_IMM_USE_STMT): Move to ssa-iterators.h.
4382 (swap_tree_operands): Rename and move prototype to tree-ssa-operands.h.
4383 * tree-flow-inline.h (delink_imm_use, link_imm_use_to_list,
4384 link_imm_use, set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
4385 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
4386 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
4387 num_imm_uses): Move to ssa-iterators.h.
4388 (get_use_from_ptr, get_def_from_ptr): Move to tree-ssa-operands.h
4389 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Move to
4391 (op_iter_done, op_iter_next_def, op_iter_next_tree,
4392 clear_and_done_ssa_iter, op_iter_init, op_iter_init_use,
4393 op_iter_init_def, op_iter_init_tree, single_ssa_tree_operand,
4394 single_ssa_use_operand, single_ssa_def_operand, zero_ssa_operands,
4395 num_ssa_operands, delink_stmt_imm_use, single_phi_def,
4396 op_iter_init_phiuse, op_iter_init_phidef, end_imm_use_stmt_p,
4397 end_imm_use_stmt_traverse, move_use_after_head, link_use_stmts_after,
4398 first_imm_use_stmt, next_imm_use_stmt, first_imm_use_on_stmt,
4399 end_imm_use_on_stmt_p, next_imm_use_on_stmt): Move to ssa-iterators.h.
4400 (gimple_phi_arg_def, gimple_phi_arg_def_ptr, gimple_phi_arg_edge,
4401 gimple_phi_arg_location, gimple_phi_arg_location_from_edge,
4402 gimple_phi_arg_set_location, gimple_phi_arg_has_location, phi_nodes,
4403 phi_nodes_ptr, phi_ssa_name_p): Move to gimple.h.
4404 (set_phi_nodes): Move to tree-phinodes.h.
4405 * tree-ssa-operands.h (enum ssa_op_iter_type,
4406 struct ssa_operand_iterator_d, SSA_OP*, FOR_EACH_SSA*, SINGLE_SSA*,
4407 ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS): Move to ssa-iterators.h.
4408 (dump_decl_set): Remove prototype.
4409 (get_use_from_ptr, get_def_from_ptr): Relocate from tree-flow.h.
4410 * tree-phinodes.h: New file. Move some prototypes from tree-flow.h.
4411 (set_phi_nodes): Relocate from tree-flow-inline.h.
4412 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Relocate from
4414 * tree-ssa.h: Add tree-phinodes.h, gimple-ssa.h, ssa-iterators.h to
4415 include list. Temporarily add gimple.h to include list.
4416 * ssa-iterators.h: New file.
4417 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
4418 BREAK_FROM_IMM_USE_STMT): Relocate from tree-flow.h.
4419 (enum ssa_op_iter_type, struct ssa_operand_iterator_d, SSA_OP*,
4420 FOR_EACH_SSA*, SINGLE_SSA*, ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS):
4421 Relocate from tree-ssa-operands.h.
4422 (delink_imm_use, link_imm_use_to_list, link_imm_use,
4423 set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
4424 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
4425 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
4426 num_imm_uses, get_use_from_ptr, get_def_from_ptr,
4427 phi_arg_index_from_use, op_iter_done, op_iter_next_def,
4428 op_iter_next_tree, clear_and_done_ssa_iter, op_iter_init,
4429 op_iter_init_use, op_iter_init_def, op_iter_init_tree,
4430 single_ssa_tree_operand, single_ssa_use_operand, single_ssa_def_operand,
4431 zero_ssa_operands, num_ssa_operands, delink_stmt_imm_use,
4432 single_phi_def, op_iter_init_phiuse, op_iter_init_phidef,
4433 end_imm_use_stmt_p, end_imm_use_stmt_traverse, move_use_after_head,
4434 link_use_stmts_after, first_imm_use_stmt, next_imm_use_stmt,
4435 first_imm_use_on_stmt, end_imm_use_on_stmt_p, next_imm_use_on_stmt):
4436 Relocate from tree-flow-inline.h.
4437 * tree-outof-ssa.h: Change _SSAEXPAND_H macro to GCC_TREE_OUTOF_SSA_H.
4439 2013-10-01 Vidya Praveen <vidyapraveen@arm.com>
4442 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l2<mode>_internal): Rename to ...
4443 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): ... this;
4444 Insert '\t' to output template.
4445 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): New.
4446 (aarch64_saddl2<mode>, aarch64_uaddl2<mode>): Modify to call
4447 gen_aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal() instead.
4448 (aarch64_ssubl2<mode>, aarch64_usubl2<mode>): Ditto.
4450 2013-10-01 Uros Bizjak <ubizjak@gmail.com>
4452 * doc/install.texi (Host/target specific installation notes for GCC):
4453 Put @anchor before @heading.
4454 * doc/gcc.texi (titlepage): Use @uref and http:// prefix for website.
4455 Use @email for email addresses.
4457 2013-10-01 Jeff Law <law@redhat.com>
4459 * tree-ssa-threadedge.c (thread_across_edge): Make path a pointer to
4460 a vec. Only delete the path if we create one without successfully
4461 registering a jump thread.
4462 * tree-ssa-threadupdate.h (register_jump_thread): Pass in path vector
4464 * tree-ssa-threadupdate.c (threaded_edges): Remove. No longer used
4465 (paths): New vector of jump threading paths.
4466 (THREAD_TARGET, THREAD_TARGET2): Remove accessor macros.
4467 (THREAD_PATH): New accessor macro for the entire thread path.
4468 (lookup_redirection_data): Get intermediate and final outgoing edge
4469 from the thread path.
4470 (create_edge_and_update_destination_phis): Copy the threading path.
4471 (ssa_fix_duplicate_block_edges): Get edges and block types from the
4472 jump threading path.
4473 (ssa_redirect_edges): Get edges and block types from the jump threading
4474 path. Free the path vector.
4475 (thread_block): Get edges from the jump threading path. Look at the
4476 entire path to see if we thread to a loop exit. If we cancel a jump
4477 thread request, then free the path vector.
4478 (thread_single_edge): Get edges and block types from the jump threading
4479 path. Free the path vector.
4480 (thread_through_loop_header): Get edges and block types from the jump
4481 threading path. Free the path vector.
4482 (mark_threaded_blocks): Iterate over the vector of paths and store
4483 the path on the appropriate edge. Get edges and block types from the
4484 jump threading path.
4485 (mark_threaded_blocks): Get edges and block types from the jump
4486 threading path. Free the path vector.
4487 (thread_through_all_blocks): Use the vector of paths rather than
4488 a vector of 3-edge sets.
4489 (register_jump_thread): Accept pointer to a path vector rather
4490 than the path vector itself. Store the path vector for later use.
4493 2013-10-01 Jakub Jelinek <jakub@redhat.com>
4494 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4497 * config/s390/s390.c (s390_split_branches): Modify check for table
4499 (s390_chunkify_start): Rearrange table jump insn check in order to
4500 deal with compare and branch insns correctly.
4502 2013-10-01 Kugan Vivekanandarajah <kuganv@linaro.org>
4506 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
4507 * config/arm/arm.md (arm_ashldi3_1bit): define_insn into
4508 define_insn_and_split.
4509 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
4510 (shiftsi3_compare): New pattern.
4512 * config/arm/unspecs.md (UNSPEC_RRX): New.
4514 2013-10-01 Alan Modra <amodra@gmail.com>
4516 * stmt.c (expand_asm_operands): Revert part of 2013-09-24 special
4517 casing inout operands.
4519 2013-10-01 Richard Biener <rguenther@suse.de>
4521 PR tree-optimization/58553
4522 * tree-loop-distribution.c (struct partition_s): Add niter member.
4523 (classify_partition): Populate niter member for the partition
4524 and properly identify whether the relevant store happens before
4525 or after the loop exit.
4526 (generate_memset_builtin): Use niter member from the partition.
4527 (generate_memcpy_builtin): Likewise.
4529 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
4531 * vec.h (vec_prefix, vec): Prefix member names with "m_".
4532 * vec.c (vec_prefix::calculate_allocation): Update accordingly.
4534 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
4536 * basic-block.h (edge_list): Prefix member names with "m_".
4537 * context.h (context): Likewise.
4538 * domwalk.h (dom_walker): Likewise.
4539 * gengtype-state.c (s_expr_writer, state_writer): Likewise.
4540 * graphite-sese-to-poly.c (sese_dom_walker): Likewise.
4541 * hash-table.h (hash_table): Likewise.
4542 * machmode.h (bit_field_mode_iterator): Likewise.
4543 * pass_manager.h (pass_list): Likewise.
4544 * tree-into-ssa.c (mark_def_dom_walker): Likewise.
4545 * tree-pass.h (pass_data): Likewise.
4546 * tree-ssa-dom.c (dom_opt_dom_walker): Likewise.
4547 * tree-ssa-phiopt.c (nontrapping_dom_walker): Likewise,
4548 * tree-ssa-uncprop.c (uncprop_dom_walker): Likewise.
4549 * asan.c (pass_data_asan): Update accordingly.
4550 * cfganal.c (control_dependences::find_control_dependence): Likewise.
4551 (control_dependences::control_dependences): Likewise.
4552 (control_dependences::~control_dependences): Likewise.
4553 (control_dependences::~control_dependences): Likewise.
4554 (control_dependences::get_edges_dependent_on): Likewise.
4555 * cgraphbuild.c (pass_data_rebuild_cgraph_edges::clone): Likewise.
4556 (pass_data_remove_cgraph_callee_edges::clone): Likewise.
4557 * context.c (gcc::context::context): Likewise.
4558 * cprop.c (pass_rtl_cprop::clone): Likewise.
4559 * domwalk.c (dom_walker::walk): Likewise.
4560 * ipa-inline-analysis.c (pass_inline_parameters::clone): Likewise.
4561 * ipa-pure-const.c (pass_local_pure_const::clone): Likewise.
4562 * mode-switching.c (pass_mode_switching::clone): Likewise.
4563 * passes.c (opt_pass::opt_pass): Likewise.
4564 (pass_manager::pass_manager): Likewise.
4565 * predict.c (pass_strip_predict_hints::clone): Likewise.
4566 * recog.c (pass_data pass_data_peephole2::clone): Likewise.
4567 (pass_split_all_insns::clone): Likewise.
4568 * stor-layout.c (bit_field_mode_iterator::bit_field_mode_iterator):
4570 (bit_field_mode_iterator::next_mode): Likewise.
4571 (bit_field_mode_iterator::prefer_smaller_modes): Likewise.
4572 * tree-cfg.c (pass_split_crit_edges::clone): Likewise.
4573 * tree-cfgcleanup.c (pass_merge_phi::clone): Likewise.
4574 * tree-complex.c (pass_lower_complex::clone): Likewise.
4575 * tree-eh.c (pass_cleanup_eh::clone): Likewise.
4576 * tree-object-size.c (pass_object_sizes::clone): Likewise.
4577 * tree-optimize.c (pass_fixup_cfg::clone): Likewise.
4578 * tree-ssa-ccp.c (pass_data_ccp::clone): Likewise.
4579 (pass_fold_builtins::clone): Likewise.
4580 * tree-ssa-copy.c (pass_data_copy_prop::clone): Likewise.
4581 * tree-ssa-copyrename.c (pass_rename_ssa_copies::clone): Likewise.
4582 * tree-ssa-dce.c (pass_dce::clone, pass_dce_loop::clone): Likewise.
4583 (pass_cd_dce::clone): Likewise.
4584 * tree-ssa-dom.c (pass_dominator::clone): Likewise.
4585 (pass_phi_only_cprop::clone): Likewise.
4586 * tree-ssa-dse.c (pass_dse::clone): Likewise.
4587 * tree-ssa-forwprop.c (pass_forwprop::clone): Likewise.
4588 * tree-ssa-loop.c (pass_lim::clone): Likewise.
4589 * tree-ssa-phiopt.c (pass_phiopt::clone): Likewise.
4590 * tree-ssa-pre.c (pass_fre::clone): Likewise.
4591 * tree-ssa-reassoc.c (pass_reassoc::clone): Likewise.
4592 * tree-ssa-uninit.c (pass_late_warn_uninitialized::clone): Likewise.
4593 * tree-tailcall.c (pass_tail_recursion::clone): Likewise.
4594 * tree-vect-generic.c (pass_lower_vector_ssa::clone): Likewise.
4595 * tree-vrp.c (pass_vrp::clone): Likewise.
4596 * tsan.c (pass_tsan::clone): Likewise.
4598 2013-09-30 Jakub Jelinek <jakub@redhat.com>
4601 * fold-const.c (tree_unary_nonnegative_warnv_p): Use
4602 INTEGRAL_TYPE_P (t) instead of TREE_CODE (t) == INTEGER_TYPE.
4605 * fold-const.c (fold_ternary_loc): For A < 0 : <sign bit of A> : 0
4606 optimization, punt if sign_bit_p looked through any zero extension.
4608 2013-09-30 Teresa Johnson <tejohnson@google.com>
4610 * tree-ssa-threadupdate.c (ssa_fix_duplicate_block_edges):
4611 Update redirected out edge count in joiner case.
4612 (ssa_redirect_edges): Common the joiner and non-joiner cases
4613 so that joiner case gets profile updates.
4615 2013-09-30 Richard Biener <rguenther@suse.de>
4617 PR tree-optimization/58554
4618 * tree-loop-distribution.c (classify_partition): Require
4619 unconditionally executed stores for memcpy and memset recognition.
4620 (tree_loop_distribution): Calculate dominance info.
4622 2013-09-30 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
4624 * config/aarch64/aarch64.h (MCOUNT_NAME): Define.
4625 (NO_PROFILE_COUNTERS): Likewise.
4626 (PROFILE_HOOK): Likewise.
4627 (FUNCTION_PROFILER): Likewise.
4628 * config/aarch64/aarch64.c (aarch64_function_profiler): Remove.
4630 2013-09-30 Iain Sandoe <iain@codesourcery.com>
4632 * config/rs6000/darwin.md (load_macho_picbase_si): Wrap machopic
4633 calls and defines in TARGET_MACHO conditional.
4634 (load_macho_picbase_di): Likewise.
4635 (reload_macho_picbase): Likewise.
4636 (reload_macho_picbase_si): Likewise.
4637 (reload_macho_picbase_di): Likewise.
4638 (nonlocal_goto_receiver): Likewise.
4640 2013-09-30 Nick Clifton <nickc@redhat.com>
4642 * config/msp430/msp430.c (msp430x_names): New array. Lists MCUs
4643 that use the MSP430X ISA.
4644 (msp430_option_override): Scan -mmcu command line option for any
4645 MCU name that supports the MSP430X ISA.
4646 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add matches for known
4647 -mmcu options which enable the MSP430X ISA.
4649 2013-09-30 Richard Biener <rguenther@suse.de>
4652 * tree-cfg.c (make_abnormal_goto_edges): Skip debug statements
4653 before looking for setjmp-like calls.
4655 2013-09-29 Iain Sandoe <iain@codesourcery.com>
4658 * config/darwin-protos.h (machopic_get_function_picbase): New.
4659 * config/darwin.c (machopic_get_function_picbase): New.
4660 * config/rs6000/darwin.md (load_macho_picbase_si): Update picbase
4661 label for a new func. (load_macho_picbase_di): Likewise.
4662 (reload_macho_picbase): New expand.
4663 (reload_macho_picbase_si): New insn.
4664 (reload_macho_picbase_di): New insn.
4665 (nonlocal_goto_receiver): New define and split.
4666 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_RELD_MPIC.
4667 (unspecv enum): Add UNSPECV_NLGR.
4669 2013-09-29 Iain Sandoe <iain@codesourcery.com>
4671 * config/rs6000/rs6000.c (rs6000_init_dwarf_reg_sizes_extra): Ensure
4672 that altivec registers are correctly sized on Darwin.
4674 2013-09-29 Iain Sandoe <iain@codesourcery.com>
4676 * config/t-darwin (darwin.o, darwin-c.o, darwin-f.o,
4677 darwin-driver.o): Use COMPILE and POSTCOMPILE.
4678 * config/x-darwin (host-darwin.o): Likewise.
4679 * config/i386/x-darwin (host-i386-darwin.o): Likewise.
4680 * config/rs6000/x-darwin (host-ppc-darwin.o): Likewise.
4681 * config/rs6000/x-darwin64 (host-ppc64-darwin.o): Likewise.
4683 2013-09-29 Uros Bizjak <ubizjak@gmail.com>
4685 * doc/invoke.texi: Fix usage of @tie{} command.
4687 2013-09-29 Eric Botcazou <ebotcazou@adacore.com>
4689 * config/sparc/sync.md: Add peephole for consecutive memory barriers.
4691 2013-09-28 Jan Hubicka <jh@suse.cz>
4693 * config/i386/x86-tune.def: Add documentation for each of the options;
4696 2013-09-28 Jan Hubicka <jh@suse.cz>
4698 * x86-tune.def (X86_TUNE_SSE_UNALIGNED_LOAD_OPTIMAL): Enable for
4700 (X86_TUNE_SSE_UNALIGNED_STORE_OPTIMAL): Likewise.
4701 (X86_TUNE_FOUR_JUMP_LIMIT): Drop for generic and buldozer.
4702 (X86_TUNE_PAD_RETURNS): Drop for buldozer chips.
4703 (X86_TUNE_AVOID_VECTOR_DECODE): Drop for generic.
4704 (X86_TUNE_REASSOC_FP_TO_PARALLEL): Enable for generic.
4706 2013-09-28 Richard Sandiford <rdsandiford@googlemail.com>
4708 * alloc-pool.c, asan.c, auto-inc-dec.c, basic-block.h, bb-reorder.c,
4709 bitmap.c, bitmap.h, bt-load.c, builtins.c, calls.c, cfgcleanup.c,
4710 cfgexpand.c, cfghooks.c, cfgloop.c, cfgloopmanip.c, cfgrtl.c, cgraph.c,
4711 cgraph.h, cgraphbuild.c, cgraphclones.c, cgraphunit.c, collect2.c,
4712 combine-stack-adj.c, combine.c, compare-elim.c, context.c, context.h,
4713 cprop.c, cse.c, cselib.c, dbxout.c, dce.c, defaults.h, df-core.c,
4714 df-problems.c, df-scan.c, df.h, diagnostic.c, double-int.c, dse.c,
4715 dumpfile.c, dwarf2asm.c, dwarf2cfi.c, dwarf2out.c, emit-rtl.c,
4716 errors.c, except.c, expmed.c, expr.c, file-find.c, final.c,
4717 fixed-value.c, fold-const.c, function.c, fwprop.c, gcc-ar.c, gcc.c,
4718 gcov-io.c, gcov-io.h, gcov.c, gcse.c, genattr-common.c, genattr.c,
4719 genattrtab.c, genautomata.c, genconfig.c, genemit.c, genextract.c,
4720 genflags.c, gengenrtl.c, gengtype-state.c, gengtype.c, genmodes.c,
4721 genopinit.c, genoutput.c, genpeep.c, genpreds.c, genrecog.c,
4722 gensupport.c, ggc-common.c, ggc-page.c, gimple-fold.c, gimple-low.c,
4723 gimple-pretty-print.c, gimple-ssa-strength-reduction.c, gimple.c,
4724 gimple.h, godump.c, graphite-clast-to-gimple.c,
4725 graphite-optimize-isl.c, graphite-poly.h, graphite-sese-to-poly.c,
4726 graphite.c, haifa-sched.c, hash-table.c, hash-table.h, hwint.c,
4727 hwint.h, ifcvt.c, incpath.c, init-regs.c, input.h, intl.c, intl.h,
4728 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
4729 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c,
4730 ipa-utils.c, ipa.c, ira-build.c, ira.c, jump.c, loop-doloop.c,
4731 loop-init.c, loop-invariant.c, loop-iv.c, lower-subreg.c, lto-cgraph.c,
4732 lto-streamer-in.c, lto-streamer-out.c, lto-wrapper.c, mcf.c,
4733 mode-switching.c, modulo-sched.c, omp-low.c, optabs.c, opts.c,
4734 pass_manager.h, passes.c, plugin.c, postreload-gcse.c, postreload.c,
4735 predict.c, prefix.c, pretty-print.c, print-rtl.c, print-tree.c,
4736 profile.c, read-md.c, real.c, real.h, recog.c, ree.c, reg-stack.c,
4737 regcprop.c, reginfo.c, regmove.c, regrename.c, regs.h, regstat.c,
4738 reload1.c, reorg.c, rtl.c, rtl.h, rtlanal.c, sbitmap.c, sched-rgn.c,
4739 sdbout.c, sel-sched-ir.c, sel-sched.c, sparseset.c, stack-ptr-mod.c,
4740 statistics.c, stmt.c, stor-layout.c, store-motion.c, streamer-hooks.h,
4741 system.h, target-hooks-macros.h, targhooks.c, targhooks.h, toplev.c,
4742 tracer.c, trans-mem.c, tree-browser.c, tree-call-cdce.c, tree-cfg.c,
4743 tree-cfgcleanup.c, tree-complex.c, tree-data-ref.c, tree-data-ref.h,
4744 tree-eh.c, tree-emutls.c, tree-flow.h, tree-if-conv.c, tree-into-ssa.c,
4745 tree-iterator.c, tree-loop-distribution.c, tree-mudflap.c,
4746 tree-nested.c, tree-nomudflap.c, tree-nrv.c, tree-object-size.c,
4747 tree-optimize.c, tree-pass.h, tree-pretty-print.c, tree-profile.c,
4748 tree-scalar-evolution.c, tree-sra.c, tree-ssa-ccp.c,
4749 tree-ssa-coalesce.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
4750 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
4751 tree-ssa-ifcombine.c, tree-ssa-live.c, tree-ssa-loop-ch.c,
4752 tree-ssa-loop-im.c, tree-ssa-loop-ivopts.c, tree-ssa-loop-prefetch.c,
4753 tree-ssa-loop.c, tree-ssa-math-opts.c, tree-ssa-operands.c,
4754 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
4755 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
4756 tree-ssa-structalias.c, tree-ssa-threadedge.c, tree-ssa-threadupdate.c,
4757 tree-ssa-uncprop.c, tree-ssa-uninit.c, tree-ssa.c, tree-ssanames.c,
4758 tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
4759 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
4760 tree-vect-stmts.c, tree-vectorizer.c, tree-vectorizer.h, tree-vrp.c,
4761 tree.c, tree.h, tsan.c, tsystem.h, value-prof.c, var-tracking.c,
4762 varasm.c, vec.h, vmsdbgout.c, vtable-verify.c, web.c: Add missing
4763 whitespace before "(".
4765 2013-09-28 Sandra Loosemore <sandra@codesourcery.com>
4767 * expr.h (extract_bit_field): Remove packedp parameter.
4768 * expmed.c (extract_fixed_bit_field): Remove packedp parameter
4769 from forward declaration.
4770 (store_split_bit_field): Remove packedp arg from calls to
4771 extract_fixed_bit_field.
4772 (extract_bit_field_1): Remove packedp parameter and packedp
4773 argument from recursive calls and calls to extract_fixed_bit_field.
4774 (extract_bit_field): Remove packedp parameter and corresponding
4775 arg to extract_bit_field_1.
4776 (extract_fixed_bit_field): Remove packedp parameter. Remove code
4778 (extract_split_bit_field): Remove packedp arg from call to
4779 extract_fixed_bit_field.
4780 * expr.c (emit_group_load_1): Adjust calls to extract_bit_field.
4781 (copy_blkmode_from_reg): Likewise.
4782 (copy_blkmode_to_reg): Likewise.
4783 (read_complex_part): Likewise.
4784 (store_field): Likewise.
4785 (expand_expr_real_1): Likewise.
4786 * calls.c (store_unaligned_arguments_into_pseudos): Adjust call
4787 to extract_bit_field.
4788 * config/tilegx/tilegx.c (tilegx_expand_unaligned_load): Adjust
4789 call to extract_bit_field.
4790 * config/tilepro/tilepro.c (tilepro_expand_unaligned_load): Adjust
4791 call to extract_bit_field.
4792 * doc/invoke.texi (Code Gen Options): Remove mention of warnings
4793 and special packedp behavior from -fstrict-volatile-bitfields
4796 2013-09-27 Jan-Benedict Glaw <jbglaw@lug-owl.de>
4798 * lra-eliminations.c (init_elim_table): Guard value_p.
4800 2013-09-27 Michael Meissner <meissner@linux.vnet.ibm.com>
4802 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Allow
4803 DFmode, DImode, and SFmode in the upper VSX registers based on the
4804 -mupper-regs-{df,sf} flags. Fix wu constraint to be ALTIVEC_REGS
4805 if -mpower8-vector. Combine -mvsx-timode handling with the rest
4806 of the VSX register handling.
4808 * config/rs6000/rs6000.md (f32_lv): Use %x0 for VSX regsters.
4810 (zero_extendsidi2_lfiwzx): Add support for loading into the
4811 Altivec registers with -mpower8-vector. Use wu/wv constraints to
4812 only do VSX memory options on Altivec registers.
4813 (extendsidi2_lfiwax): Likewise.
4814 (extendsfdf2_fpr): Likewise.
4815 (mov<mode>_hardfloat, SF/SD modes): Likewise.
4816 (mov<mode>_hardfloat32, DF/DD modes): Likewise.
4817 (mov<mode>_hardfloat64, DF/DD modes): Likewise.
4818 (movdi_internal64): Likewise.
4820 2013-09-27 Xinliang David Li <davidxl@google.com>
4822 * opts.c (finish_options): Adjust parameters
4823 according to vect cost model.
4824 (common_handle_option): Set dynamic vect cost
4826 targhooks.c (default_add_stmt_cost): Compute stmt cost
4828 * tree-vect-loop.c (vect_estimate_min_profitable_iters):
4829 Use helper function.
4830 * tree-vectorizer.h (unlimited_cost_model): New function.
4831 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use helper function.
4832 * tree-vect-data-refs.c (vect_peeling_hash_insert): Use helper
4834 (vect_enhance_data_refs_alignment): Ditto.
4835 * flag-types.h: New enum.
4836 * common/config/i386/i386-common.c (ix86_option_init_struct):
4837 No need to initialize vect_cost_model flag.
4838 * config/i386/i386.c (ix86_add_stmt_cost): Compute stmt cost
4841 2013-09-27 Diego Novillo <dnovillo@google.com>
4843 * gimple.h (enum ssa_mode): Remove.
4845 2013-09-27 Paulo Matos <pmatos@broadcom.com>
4847 * cfgloop.h (number_of_loops): Fix typo in check for null.
4849 2013-09-27 Jakub Jelinek <jakub@redhat.com>
4852 * tree-cfg.c (move_sese_region_to_fn): Also move loops that
4853 are children of outermost saved_cfun's loop, and set it up to
4854 be moved to dest_cfun's outermost loop. Fix up num_nodes adjustments
4855 if loop != loop0 and SESE region contains bbs that belong to loop0.
4857 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
4859 * rtlanal.c (must_be_base_p, must_be_index_p): Delete.
4860 (binary_scale_code_p, get_base_term, get_index_term): New functions.
4861 (set_address_segment, set_address_base, set_address_index)
4862 (set_address_disp): Accept the argument unconditionally.
4863 (baseness): Remove must_be_base_p and must_be_index_p checks.
4864 (decompose_normal_address): Classify as much as possible in the
4867 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
4869 * cse.c (count_reg_usage): Handle INT_LIST.
4870 * lra-eliminations.c (lra_eliminate_regs_1): Likewise.
4871 * reginfo.c (reg_scan_mark_refs): Likewise.
4872 * reload1.c (eliminate_regs_1): Likewise.
4874 2013-09-27 Iain Sandoe <iain@codesourcery.com>
4877 * rtlanal.c (lsb_bitfield_op_p): Make both parts of the comparison
4880 2013-09-27 Richard Biener <rguenther@suse.de>
4882 PR tree-optimization/58459
4883 * tree-ssa-forwprop.c (forward_propagate_addr_expr): Remove
4884 restriction not propagating into loops.
4886 2013-09-26 Florian Weimer <fw@deneb.enyo.de>
4888 * tree-ssa.h (walk_use_def_chains_fn, walk_use_def_chains): Delete.
4889 * tree-ssa.c (walk_use_def_chains_1, walk_use_def_chains): Delete.
4890 * doc/tree-ssa.texi (Walking use-def chains): Delete.
4892 2013-09-26 Richard Biener <rguenther@suse.de>
4894 * tree-into-ssa.c (rewrite_into_ssa): Make more SSA names to anonymous.
4896 2013-09-26 Richard Biener <rguenther@suse.de>
4898 * alias.h (component_uses_parent_alias_set): Rename to ...
4899 (component_uses_parent_alias_set_from): ... this.
4900 * alias.c (component_uses_parent_alias_set): Rename to ...
4901 (component_uses_parent_alias_set_from): ... this and return
4903 (reference_alias_ptr_type_1): Use the result from
4904 component_uses_parent_alias_set_from instead of stripping
4905 components one at a time.
4906 * emit-rtl.c (set_mem_attributes_minus_bitpos): Adjust.
4908 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
4910 * tree-ssa-live.h (find_replaceable_exprs, dump_replaceable_exprs):
4911 Move prototypes to...
4912 * tree-ssa-ter.h: New File. Move prototypes here.
4913 * tree-flow.h (stmt_is_replaceable_p): Remove prototype.
4914 * tree-outof-ssa.h: New. Rename ssaexpand.h, include tree-ssa-ter.h.
4915 * tree-outof-ssa.c (ssa_is_replaceable_p): New. Refactor common bits
4916 from is_replaceable_p.
4917 * tree-ssa-ter.c (is_replaceable_p, stmt_is_replaceable_p): Delete.
4918 (ter_is_replaceable_p): New. Use new refactored ssa_is_replaceable_p.
4919 (process_replaceable): Use ter_is_replaceable_p.
4920 (find_replaceable_in_bb): Use ter_is_replaceable_p.
4921 * expr.c (stmt_is_replaceable_p): Relocate from tree-ssa-ter.c. Use
4922 newly refactored ssa_is_replaceable_p.
4923 * cfgexpand.c: Include tree-outof-ssa.h.
4924 * ssaexpand.h: Delete.
4926 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
4928 * gimple.c (gimple_replace_lhs): Move to tree-ssa.c and rename.
4929 (struct count_ptr_d, count_ptr_derefs, count_uses_and_derefs): Move to
4931 (create_gimple_tmp): Delete.
4932 (get_expr_type, build_assign, build_type_cast): Move to...
4933 * gimple-builder.c: New File.
4934 (get_expr_type): Relocate from gimple.c.
4935 (build_assign, build_type_cast): Change to only create ssanames.
4936 * gimple.h: Move prototypes to...
4937 * gimple-builder.h: New File. Here.
4938 * tree-ssa.h: And here.
4939 * tree-ssa.c (struct count_ptr_d, count_ptr_derefs,
4940 count_uses_and_derefs): Relocate from gimple.c.
4941 (gimple_replace_ssa_lhs): Renamed gimple_replace_ssa from gimple.c
4942 * tree-ssa-reassoc.c (repropagate_negates): Use gimple_replace_ssa_lhs.
4943 * tree-ssa-math-opts (execute_cse_reciprocals): Use
4944 gimple_replace_ssa_lhs.
4945 * asan.c: Include gimple-builder.h.
4946 * Makefile.in: Add gimple-builder.o.
4948 2013-09-26 Richard Biener <rguenther@suse.de>
4950 * tree-ssa-live.c (var_map_base_init): Handle SSA names with
4951 DECL_IGNORED_P base VAR_DECLs like anonymous SSA names.
4952 (loe_visit_block): Use gcc_checking_assert.
4953 * tree-ssa-coalesce.c (create_outofssa_var_map): Use
4954 gimple_assign_ssa_name_copy_p.
4955 (gimple_can_coalesce_p): Adjust according to the var_map_base_init
4958 2013-09-26 David Edelsohn <dje.gcc@gmail.com>
4960 * config/rs6000/t-rs6000 (rs6000.o): Remove.
4961 (rs6000-c.o): Use COMPILE and POSTCOMPILE.
4963 2013-09-26 Richard Biener <rguenther@suse.de>
4965 PR tree-optimization/58539
4966 * tree-vect-loop.c (vect_create_epilog_for_reduction): Honor
4967 the fact that debug statements are not taking part in loop-closed
4970 2013-09-26 Nick Clifton <nickc@redhat.com>
4972 * config/msp430/msp430.c (msp430_expand_epilogue): Fix compile
4973 time warning message.
4974 (msp430_print_operand_raw): Delete unused letter parameter.
4975 (TARGET_PRINT_OPERAND_ADDRESS): Define.
4976 (msp430_print_operand_address): New function.
4977 (msp430_print_operand): Move address printing code from here to
4979 * config/msp430/msp430.md (movsipsi2): Add comment in generated
4981 (zero_extendpsisi2): Likewise.
4982 (extendpsisi2): New pattern.
4983 (andneghi3): New pattern.
4985 2013-09-26 Yvan Roux <yvan.roux@linaro.org>
4987 * config/aarch64/aarch64.opt (mlra): New option.
4988 * config/aarch64/aarch64.c (aarch64_lra_p): New function.
4989 (TARGET_LRA_P): Define.
4991 2013-09-26 Eric Botcazou <ebotcazou@adacore.com>
4993 * expr.c (expand_assignment): Remove obsolete comment.
4995 2013-09-25 Jeff Law <law@redhat.com>
4997 * tree-flow.h (thread_through_all_blocks): Prototype moved into
4998 tree-ssa-threadupdate.h.
4999 (register_jump_thread): Similarly.
5000 * tree-ssa-threadupdate.h: New header file.
5001 * tree-ssa-dom.c: Include tree-ssa-threadupdate.h.
5002 * tree-vrp.c: Likewise.
5003 * tree-ssa-threadedge.c: Include tree-ssa-threadupdate.h.
5004 (thread_around_empty_blocks): Change type of path vector argument to
5005 an edge,type pair from just an edge. Initialize both elements when
5006 appending to a jump threading path. Tweak references to elements
5008 (thread_across_edge): Similarly. Release memory for the elements
5010 * tree-ssa-threadupdate.c: Include tree-ssa-threadupdate.h.
5011 (dump_jump_thread_path): New function broken out from
5012 register_jump_thread.
5013 (register_jump_thread): Use dump_jump_thread_path. Change type of
5014 path vector entries. Search the path for NULL edges and dump
5015 the path if one is found. Tweak the conversion of path to 3-edge
5016 form to use the block copy type information embedded in the path.
5018 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
5020 * lra.c (update_inc_notes): Remove all REG_DEAD and REG_UNUSED notes.
5022 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
5023 Vladimir Makarov <vmakarov@redhat.com>
5025 * rtlanal.c (lsb_bitfield_op_p): New predicate for bitfield operations
5026 from the least significant bit.
5027 (strip_address_mutations): Add bitfield operations handling.
5028 (must_be_index_p): Add shifting and rotate operations handling.
5029 (set_address_base): Use must_be_base_p predicate.
5030 (set_address_index): Use must_be_index_p predicate.
5032 2013-09-25 Alexander Ivchenko <alexander.ivchenko@intel.com>
5033 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5034 Sergey Lega <sergey.s.lega@intel.com>
5035 Anna Tikhonova <anna.tikhonova@intel.com>
5036 Ilya Tocar <ilya.tocar@intel.com>
5037 Andrey Turetskiy <andrey.turetskiy@intel.com>
5038 Ilya Verbin <ilya.verbin@intel.com>
5039 Kirill Yukhin <kirill.yukhin@intel.com>
5040 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5042 * config/i386/i386.c (ix86_avx256_split_vector_move_misalign):
5044 (ix86_expand_vector_move_misalign): Support new unaligned load and
5045 stores and use new names.
5046 (CODE_FOR_sse2_storedqu): Rename to ...
5047 (CODE_FOR_sse2_storedquv16qi): ... this.
5048 (CODE_FOR_sse2_loaddqu): Rename to ...
5049 (CODE_FOR_sse2_loaddquv16qi): ... this.
5050 (CODE_FOR_avx_loaddqu256): Rename to ...
5051 (CODE_FOR_avx_loaddquv32qi): ... this.
5052 (CODE_FOR_avx_storedqu256): Rename to ...
5053 (CODE_FOR_avx_storedquv32qi): ... this.
5054 * config/i386/i386.md (fpint_logic): New.
5055 * config/i386/sse.md (VMOVE): Extend for AVX512.
5059 (VI_UNALIGNED_LOADSTORE): Ditto.
5060 (sse2_avx_avx512f): Ditto.
5061 (sse2_avx2): Extend for AVX512.
5062 (sse4_1_avx2): Ditto.
5063 (avx2_avx512f): New.
5064 (sse): Extend for AVX512.
5067 (avxsizesuffix): Ditto.
5068 (sseintvecmode): Ditto.
5070 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Ditto.
5071 (<sse>_storeu<ssemodesuffix><avxsizesuffix>): Ditto.
5072 (<sse2>_loaddqu<avxsizesuffix>): Extend for AVX512 and rename to ...
5073 (<sse2_avx_avx512f>_loaddqu<mode>): ... this.
5074 (<sse2>_storedqu<avxsizesuffix>): Extend for AVX512 and rename to ...
5075 (<sse2_avx_avx512f>_storedqu<mode): ... this.
5076 (<sse>_movnt<mode>): Replace constraint "x" with "v".
5077 (STORENT_MODE): Extend for AVX512.
5078 (*absneg<mode>2): Replace constraint "x" with "v".
5079 (*mul<mode>3): Ditto.
5080 (*ieee_smin<mode>3): Ditto.
5081 (*ieee_smax<mode>3): Ditto.
5082 (avx_cmp<mode>3): Replace VF with VF_128_256.
5083 (*<sse>_maskcmp<mode>3_comm): Ditto.
5084 (<sse>_maskcmp<mode>3): Ditto.
5085 (<sse>_andnot<mode>3): Extend for AVX512.
5086 (<code><mode>3, anylogic): Replace VF with VF_128_256.
5087 (<code><mode>3, fpint_logic): New.
5088 (*<code><mode>3): Extend for AVX512.
5089 (avx512flogicsuff): New.
5090 (avx512f_<logic><mode>): Ditto.
5091 (<sse>_movmsk<ssemodesuffix><avxsizesuffix>): Replace VF with
5093 (<sse4_1>_blend<ssemodesuffix><avxsizesuffix>): Ditto.
5094 (<sse4_1>_blendv<ssemodesuffix><avxsizesuffix>): Ditto.
5095 (<sse4_1>_dp<ssemodesuffix><avxsizesuffix>): Ditto.
5096 (avx_vtest<ssemodesuffix><avxsizesuffix>): Ditto.
5097 (<sse4_1>_round<ssemodesuffix><avxsizesuffix>): Ditto.
5098 (xop_vpermil2<mode>3): Ditto.
5099 (*avx_vpermilp<mode>): Extend for AVX512 and rename to ...
5100 (*<sse2_avx_avx512f>_vpermilp<mode>): ... this.
5101 (avx_vpermilvar<mode>3): Extend for AVX512 and rename to ...
5102 (<sse2_avx_avx512f>_vpermilvar<mode>3): ... this.
5104 2013-09-25 Tom Tromey <tromey@redhat.com>
5106 * Makefile.in (PARTITION_H, LTO_SYMTAB_H, COMMON_TARGET_DEF_H)
5107 (RTL_ERROR_H, TRANS_MEM_H, COVERAGE_H, DEMANGLE_H, ALIAS_H)
5108 (SCHED_INT_H, SEL_SCHED_IR_H, SEL_SCHED_DUMP_H, VALTRACK_H, DDG_H)
5109 (GGC_INTERNAL_H, DECNUM_H, BACKTRACE_H, MKDEPS_H, TREE_HASHER_H)
5110 (TREE_SSA_LIVE_H, SSAEXPAND_H, DWARF2OUT_H, SCEV_H, OMEGA_H)
5111 (TREE_DATA_REF_H, IRA_INT_H, LRA_INT_H, DBGCNT_H, DATA_STREAMER_H)
5112 (GIMPLE_STREAMER_H, TREE_STREAMER_H, STREAMER_HOOKS_H)
5113 (TREE_VECTORIZER_H, IPA_INLINE_H, GSTAB_H, LIBFUNCS_H)
5114 (GRAPHITE_HTAB_H): Remove.
5116 2013-09-25 Tom Tromey <tromey@redhat.com>
5118 * config/mcore/t-mcore (CROSS_FLOAT_H): Remove.
5120 2013-09-25 Tom Tromey <tromey@redhat.com>
5122 * config/t-glibc (glibc-c.o): Use COMPILE and POSTCOMPILE.
5124 2013-09-25 Tom Tromey <tromey@redhat.com>
5126 * config/i386/t-i386 (i386.o): Remove.
5127 (i386-c.o): Use COMPILE and POSTCOMPILE.
5129 2013-09-25 Tom Tromey <tromey@redhat.com>
5131 * Makefile.in ($(out_object_file)): Use COMPILE and POSTCOMPILE.
5133 2013-09-25 Tom Tromey <tromey@redhat.com>
5135 * Makefile.in (graph.o, sbitmap.o, sparseset.o, gcc-ar.o)
5136 (gcc-ranlib.o, gcc-nm.o, collect2.o, collect2-aix.o, tlink.o)
5137 (lto-wrapper.o, default-c.o, attribs.o, incpath.o, prefix.o)
5138 (gcc.o, options.o, options-save.o, version.o, gtype-desc.o)
5139 (trans-mem.o, ggc-common.o, ggc-page.o, ggc-none.o, stringpool.o)
5140 (convert.o, double-int.o, lto-compress.o, data-streamer-in.o)
5141 (data-streamer-out.o, data-streamer.o, gimple-streamer-in.o)
5142 (gimple-streamer-out.o, tree-streamer.o, tree-streamer-in.o)
5143 (tree-streamer-out.o, streamer-hooks.o, lto-cgraph.o)
5144 (lto-streamer-in.o, lto-streamer-out.o, lto-section-in.o)
5145 (lto-section-out.o, lto-opts.o, lto-streamer.o, langhooks.o)
5146 (test-dump.o, tree.o, tree-dump.o, tree-inline.o, print-tree.o)
5147 (stor-layout.o, asan.o, tsan.o, ubsan.o, tree-ssa-tail-merge.o)
5148 (tree-ssa-structalias.o, tree-ssa-uninit.o, tree-ssa.o)
5149 (tree-into-ssa.o, tree-ssa-ter.o, tree-ssa-coalesce.o)
5150 (tree-outof-ssa.o, tree-ssa-dse.o, tree-ssa-forwprop.o)
5151 (tree-ssa-phiprop.o, tree-ssa-ifcombine.o, tree-ssa-phiopt.o)
5152 (tree-nrv.o, tree-ssa-copy.o, tree-ssa-propagate.o)
5153 (tree-ssa-dom.o, tree-ssa-uncprop.o, tree-ssa-threadedge.o)
5154 (tree-ssa-threadupdate.o, tree-ssanames.o, tree-phinodes.o)
5155 (domwalk.o, tree-ssa-live.o, tree-ssa-copyrename.o)
5156 (tree-ssa-pre.o, tree-ssa-sccvn.o)
5157 (gimple-ssa-strength-reduction.o, tree-vrp.o, tree-cfg.o)
5158 (tree-cfgcleanup.o, tree-tailcall.o, tree-ssa-sink.o)
5159 (tree-nested.o, tree-if-conv.o, tree-iterator.o, tree-dfa.o)
5160 (tree-ssa-operands.o, tree-eh.o, tree-ssa-loop.o)
5161 (tree-ssa-loop-unswitch.o, tree-ssa-address.o)
5162 (tree-ssa-loop-niter.o, tree-ssa-loop-ivcanon.o)
5163 (tree-ssa-loop-ch.o, tree-ssa-loop-prefetch.o, tree-predcom.o)
5164 (tree-ssa-loop-ivopts.o, tree-affine.o, tree-ssa-loop-manip.o)
5165 (tree-ssa-loop-im.o, tree-ssa-math-opts.o, tree-ssa-alias.o)
5166 (tree-ssa-reassoc.o, tree-optimize.o, gimplify.o)
5167 (gimple-iterator.o, gimple-fold.o, gimple-low.o, omp-low.o)
5168 (tree-browser.o, omega.o, tree-chrec.o, tree-scalar-evolution.o)
5169 (tree-data-ref.o, sese.o, graphite.o, graphite-blocking.o)
5170 (graphite-clast-to-gimple.o, graphite-dependences.o)
5171 (graphite-interchange.o, graphite-poly.o)
5172 (graphite-scop-detection.o, graphite-sese-to-poly.o)
5173 (graphite-optimize-isl.o, tree-vect-loop.o)
5174 (tree-vect-loop-manip.o, tree-vect-patterns.o, tree-vect-slp.o)
5175 (tree-vect-stmts.o, tree-vect-data-refs.o, tree-vectorizer.o)
5176 (vtable-verify.o, tree-loop-distribution.o, tree-parloops.o)
5177 (tree-stdarg.o, tree-object-size.o, internal-fn.o, gimple.o)
5178 (gimple-pretty-print.o, tree-mudflap.o, tree-nomudflap.o)
5179 (tree-pretty-print.o, tree-diagnostic.o, fold-const.o)
5180 (diagnostic.o, diagnostic-color.o, opts.o, opts-global.o)
5181 (opts-common.o, targhooks.o, common/common-targhooks.o, input.o)
5182 (toplev.o, hwint.o, passes.o, plugin.o, main.o, host-default.o)
5183 (rtl-error.o, rtl.o, print-rtl.o, rtlanal.o, varasm.o, function.o)
5184 (statistics.o, stmt.o, except.o, expr.o, dojump.o, builtins.o)
5185 (calls.o, expmed.o, explow.o, optabs.o, dbxout.o, debug.o)
5186 (sdbout.o, dwarf2out.o, dwarf2cfi.o, dwarf2asm.o, vmsdbgout.o)
5187 (xcoffout.o, godump.o, emit-rtl.o, real.o, realmpfr.o, dfp.o)
5188 (fixed-value.o, jump.o, simplify-rtx.o, symtab.o, cgraph.o)
5189 (cgraphunit.o, cgraphclones.o, cgraphbuild.o, varpool.o, ipa.o)
5190 (ipa-profile.o, ipa-devirt.o, ipa-prop.o, ipa-ref.o, ipa-cp.o)
5191 (ipa-split.o, ipa-inline.o, ipa-inline-analysis.o)
5192 (ipa-inline-transform.o, ipa-utils.o, ipa-reference.o)
5193 (ipa-pure-const.o, coverage.o, cselib.o, cse.o, dce.o, dumpfile.o)
5194 (dse.o, fwprop.o, web.o, ree.o, cprop.o, gcse.o, store-motion.o)
5195 (resource.o, lcm.o, mode-switching.o, tree-ssa-dce.o)
5196 (tree-call-cdce.o, tree-ssa-ccp.o, tree-ssa-strlen.o, tree-sra.o)
5197 (tree-switch-conversion.o, tree-complex.o, tree-emutls.o)
5198 (tree-vect-generic.o, df-core.o, df-problems.o, df-scan.o)
5199 (regstat.o, valtrack.o, var-tracking.o, profile.o, mcf.o)
5200 (tree-profile.o, value-prof.o, loop-doloop.o, alloc-pool.o)
5201 (auto-inc-dec.o, cfg.o, cfghooks.o, cfgexpand.o, cfgrtl.o)
5202 (cfganal.o, cfgbuild.o, cfgcleanup.o, cfgloop.o, cfgloopanal.o)
5203 (graphds.o, loop-iv.o, loop-invariant.o, cfgloopmanip.o)
5204 (loop-init.o, loop-unswitch.o, loop-unroll.o, dominance.o)
5205 (et-forest.o, combine.o, reginfo.o, bitmap.o, vec.o, hash-table.o)
5206 (reload.o, reload1.o, rtlhooks.o, postreload.o, postreload-gcse.o)
5207 (caller-save.o, bt-load.o, reorg.o, alias.o, stack-ptr-mod.o)
5208 (init-regs.o, ira-build.o, ira-costs.o, ira-conflicts.o)
5209 (ira-color.o, ira-emit.o, ira-lives.o, ira.o, lra.o)
5210 (lra-assigns.o, lra-coalesce.o, lra-constraints.o)
5211 (lra-eliminations.o, lra-lives.o, lra-spills.o, regmove.o)
5212 (combine-stack-adj.o, compare-elim.o, ddg.o, modulo-sched.o)
5213 (haifa-sched.o, sched-deps.o, sched-rgn.o, sched-ebb.o)
5214 (sched-vis.o, sel-sched.o, sel-sched-dump.o, sel-sched-ir.o)
5215 (final.o, recog.o, reg-stack.o, sreal.o, predict.o, lists.o)
5216 (bb-reorder.o, tracer.o, timevar.o, regcprop.o, regrename.o)
5217 (ifcvt.o, params.o, pointer-set.o, hooks.o, pretty-print.o)
5218 (errors.o, dbgcnt.o, lower-subreg.o, target-globals.o)
5219 (hw-doloop.o, file-find.o, context.o, $(common_out_object_file))
5220 (insn-attrtab.o, insn-automata.o, insn-dfatab.o, insn-emit.o)
5221 (insn-enums.o, insn-extract.o, insn-latencytab.o, insn-modes.o)
5222 (insn-opinit.o, insn-output.o, insn-peep.o, insn-preds.o)
5223 (insn-recog.o, intl.o, cppbuiltin.o, cppdefault.o, gcov.o)
5224 (gcov-dump.o): Remove.
5225 (default-c.o): Use COMPILE and POSTCOMPILE.
5226 (CFLAGS-gcc.o): New variable.
5227 ($(common_out_object_file)): Use COMPILE and POSTCOMPILE.
5229 2013-09-25 Tom Tromey <tromey@redhat.com>
5231 * Makefile.in (c-family/cppspec.o, c-family/c-common.o)
5232 (c-family/c-cppbuiltin.o, c-family/c-dump.o, c-family/c-format.o)
5233 (c-family/c-gimplify.o, c-family/c-lex.o, c-family/c-omp.o)
5234 (c-family/c-opts.o, c-family/c-pch.o, c-family/c-ppoutput.o)
5235 (c-family/c-pragma.o, c-family/c-pretty-print.o)
5236 (c-family/c-semantics.o, c-family/c-ada-spec.o)
5237 (c-family/array-notation-common.o, c-family/stub-objc.o)
5238 (c-family/c-ubsan.o): Remove.
5240 2013-09-25 Tom Tromey <tromey@redhat.com>
5242 * Makefile.in (C_TREE_H): Reference c/c-tree.h.
5244 2013-09-25 Tom Tromey <tromey@redhat.com>
5246 * Makefile.in (DRIVER_DEFINES): Use $(and), not shell code,
5247 to add -DENABLE_SHARED_LIBGCC.
5248 (gcc.o): Don't use subshell.
5250 2013-09-25 Tom Tromey <tromey@redhat.com>
5252 * Makefile.in (OUTPUT_OPTION): Define as "-o $@".
5253 * configure.ac: Don't invoke AM_PROG_CC_C_O.
5254 (NO_MINUS_C_MINUS_O, OUTPUT_OPTION): Don't subst.
5255 * configure, config.in: Rebuild.
5257 2013-09-25 Tom Tromey <tromey@redhat.com>
5259 * Makefile.in (CCDEPMODE, DEPDIR, depcomp, COMPILE.base)
5260 (COMPILE, POSTCOMPILE): New variables.
5261 (.cc.o .c.o): Use COMPILE, POSTCOMPILE.
5262 (DEPFILES): New variable.
5263 Include ".Po" files.
5264 * configure.ac: Add checks for dependency checking.
5265 * configure, aclocal.m4: Regenerate.
5267 2013-09-25 Tom Tromey <tromey@redhat.com>
5269 * Makefile.in (ALL_HOST_BACKEND_OBJS): Add lto-wrapper.o.
5270 ($(ALL_HOST_OBJS)): Move order-only dependency to end of file.
5272 2013-09-25 Tom Tromey <tromey@redhat.com>
5274 * Makefile.in (generated_files): Add options.h,
5275 target-hooks-def.h, insn-opinit.h,
5276 common/common-target-hooks-def.h, pass-instances.def,
5277 c-family/c-target-hooks-def.h.
5279 2013-09-25 Jeff Law <law@redhat.com>
5281 * tree-ssa-threadedge.c (thread_across_edge): Use foo.last () rather
5282 than foo[foo.length () - 1] to access last member in a vec.
5283 * tree-ssa-threadupdate.c (register_jump_thread): Similarly.
5285 2013-09-25 Richard Biener <rguenther@suse.de>
5288 * tree.c (iterative_hash_expr): Remove MEM_REF special handling.
5290 2013-09-25 Jan Hubicka <jh@suse.cz>
5292 * cgraph.c (cgraph_resolve_speculation): Use semantical equivalency
5295 2013-09-25 Marek Polacek <polacek@redhat.com>
5298 * ubsan.c (ubsan_type_descriptor): Handle IDENTIFIER_NODEs
5299 when determining the type name.
5301 2013-09-24 Oleg Endo <olegendo@gcc.gnu.org>
5303 * config/sh/sh.md: Fix formatting.
5305 2013-09-24 Xinliang David Li <davidxl@google.com>
5307 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Check
5308 max peel iterations parameter.
5309 * param.def: New parameter.
5310 * doc/invoke.texi: Document New parameter.
5312 2013-09-24 Christophe Lyon <christophe.lyon@linaro.org>
5314 * gimple-pretty-print.c: Various whitespace tweaks.
5315 * tree-core.h: Likewise.
5316 * tree-pretty-print.c: Likewise.
5317 * tree-ssa-alias.c: Likewise.
5318 * tree-ssa-copy.c: Likewise.
5319 * tree-ssanames.c: Likewise.
5320 * tree-ssanames.h: Likewise.
5321 * tree-vrp.c: Likewise.
5323 2013-09-24 Alan Modra <amodra@gmail.com>
5327 * stmt.c (expand_asm_operands): Call expand_expr with EXPAND_MEMORY
5328 for output operands that disallow regs. Don't use EXPAND_WRITE on
5331 2013-09-24 Richard Biener <rguenther@suse.de>
5334 * tree.c (reference_alias_ptr_type): Move ...
5335 * alias.c (reference_alias_ptr_type): ... here and implement
5336 in terms of the new reference_alias_ptr_type_1.
5337 (ref_all_alias_ptr_type_p): New helper.
5338 (get_deref_alias_set_1): Drop flag_strict_aliasing here,
5339 use ref_all_alias_ptr_type_p.
5340 (get_deref_alias_set): Add flag_strict_aliasing check here.
5341 (reference_alias_ptr_type_1): New function, split out from ...
5342 (get_alias_set): ... here.
5343 (alias_ptr_types_compatible_p): New function.
5344 * alias.h (reference_alias_ptr_type): Declare.
5345 (alias_ptr_types_compatible_p): Likewise.
5346 * tree.h (reference_alias_ptr_type): Remove.
5347 * fold-const.c (operand_equal_p): Use alias_ptr_types_compatible_p
5348 to compare MEM_REF alias types.
5350 2013-09-24 Richard Biener <rguenther@suse.de>
5352 * tree-vrp.c (vrp_finalize): Check for SSA name presence.
5354 2013-09-23 Michael Meissner <meissner@linux.vnet.ibm.com>
5356 * config/rs6000/rs6000.c (rs6000_vector_reload): Delete, combine
5357 reload helper function arrays into a single array reg_addr.
5358 (reload_fpr_gpr): Likewise.
5359 (reload_gpr_vsx): Likewise.
5360 (reload_vsx_gpr): Likewise.
5361 (struct rs6000_reg_addr): Likewise.
5362 (reg_addr): Likewise.
5363 (rs6000_debug_reg_global): Change rs6000_vector_reload,
5364 reload_fpr_gpr, reload_gpr_vsx, reload_vsx_gpr uses to reg_addr.
5365 (rs6000_init_hard_regno_mode_ok): Likewise.
5366 (rs6000_secondary_reload_direct_move): Likewise.
5367 (rs6000_secondary_reload): Likewise.
5369 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add new
5370 constraints: wu, ww, and wy. Repurpose wv constraint added during
5371 power8 changes. Put wg constraint in alphabetical order.
5373 * config/rs6000/rs6000.opt (-mvsx-scalar-float): New debug switch
5374 for future work to add ISA 2.07 VSX single precision support.
5375 (-mvsx-scalar-double): Change default from -1 to 1, update
5376 documentation comment.
5377 (-mvsx-scalar-memory): Rename debug switch to -mupper-regs-df.
5378 (-mupper-regs-df): New debug switch to control whether DF values
5379 can go in the traditional Altivec registers.
5380 (-mupper-regs-sf): New debug switch to control whether SF values
5381 can go in the traditional Altivec registers.
5383 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print wu, ww,
5385 (rs6000_init_hard_regno_mode_ok): Use ssize_t instead of int for
5386 loop variables. Rename -mvsx-scalar-memory to -mupper-regs-df.
5387 Add new constraints, wu/ww/wy. Repurpose wv constraint.
5388 (rs6000_debug_legitimate_address_p): Print if we are running
5389 before, during, or after reload.
5390 (rs6000_secondary_reload): Add a comment.
5391 (rs6000_opt_masks): Add -mupper-regs-df, -mupper-regs-sf.
5393 * config/rs6000/constraints.md (wa constraint): Sort w<x>
5394 constraints. Update documentation string.
5395 (wd constraint): Likewise.
5396 (wf constraint): Likewise.
5397 (wg constraint): Likewise.
5398 (wn constraint): Likewise.
5399 (ws constraint): Likewise.
5400 (wt constraint): Likewise.
5401 (wx constraint): Likewise.
5402 (wz constraint): Likewise.
5403 (wu constraint): New constraint for ISA 2.07 SFmode scalar
5405 (ww constraint): Likewise.
5406 (wy constraint): Likewise.
5407 (wv constraint): Repurpose ISA 2.07 constraint that we did not use
5408 in the previous submissions.
5409 * doc/md.texi (PowerPC and IBM RS6000): Likewise.
5411 2013-09-23 Richard Sandiford <rdsandiford@googlemail.com>
5413 * doc/rtl.texi (REG_NOTES): Say that int_list can also be used.
5414 (REG_BR_PROB): Say that the probability is stored in an int_list.
5415 * reg-notes.def: Update commentary to mention INT_LIST.
5416 * rtl.def (EXPR_LIST, INSN_LIST): Capitalize comments.
5417 (INT_LIST): New rtx.
5418 * rtl.h (add_int_reg_note, add_shallow_copy_of_reg_note): Declare.
5419 * rtlanal.c (int_reg_note_p): New function.
5420 (alloc_reg_note): Assert that the note does not have an int argument.
5421 (add_int_reg_note, add_shallow_copy_of_reg_note): New functions.
5422 * combine.c (distribute_notes): Use add_shallow_copy_of_rtx.
5423 * cse.c (cse_process_notes_1): Expect REG_EQUAL to be an EXPR_LIST
5424 rather than an INSN_LIST. Handle INT_LIST.
5425 * ifcvt.c (cond_exec_process_insns): Take the probability as an int
5426 rather than an rtx. Use gen_rtx_INT_LIST to create a REG_BR_PROB note.
5427 (cond_exec_process_if_block): Use XINT to extract REG_BR_PROB values.
5428 Manipulate them as ints rather than rtxes.
5429 * reg-stack.c (subst_asm_stack_regs): Only handle EXPR_LIST notes.
5430 * regmove.c (copy_src_to_dest): Likewise.
5431 * sched-vis.c (print_insn_with_notes): Handle INT_LIST.
5433 * config/i386/winnt.c (i386_pe_seh_unwind_emit): Sink pat assignment
5434 into the cases that need it.
5435 * config/arm/arm.c (arm_unwind_emit): Likewise.
5437 * asan.c (asan_clear_shadow): Use add_int_reg_note for REG_BR_PROB.
5438 * emit-rtl.c (try_split, emit_copy_of_insn_after): Likewise.
5439 * loop-doloop.c (add_test, doloop_modify): Likewise.
5440 * loop-unswitch.c (compare_and_jump_seq): Likewise.
5441 * optabs.c (emit_cmp_and_jump_insn_1): Likewise.
5442 * predict.c (combine_predictions_for_insn): Likewise.
5443 * print-rtl.c (print_rtx): Handle INT_LIST.
5444 * config/aarch64/aarch64.c (aarch64_emit_unlikely_jump): Likewise.
5445 * config/alpha/alpha.c (emit_unlikely_jump): Likewise.
5446 * config/arm/arm.c (emit_unlikely_jump): Likewise.
5447 * config/i386/i386.c (ix86_expand_split_stack_prologue): Likewise.
5448 (ix86_split_fp_branch, predict_jump): Likewise.
5449 * config/rs6000/rs6000.c (emit_unlikely_jump): Likewise.
5450 * config/sh/sh.c (expand_cbranchsi4): Likewise.
5451 * config/spu/spu.c (ea_load_store_inline): Likewise.
5453 * cfgbuild.c (compute_outgoing_frequencies): Use XINT to access the
5454 value of a REG_BR_PROB note.
5455 * cfgrtl.c (force_nonfallthru_and_redirect): Likewise.
5456 (update_br_prob_note, rtl_verify_edges, purge_dead_edges): Likewise.
5457 * emit-rtl.c (try_split): Likewise.
5458 * predict.c (br_prob_note_reliable_p): Likewise.
5459 (invert_br_probabilities, combine_predictions_for_insn): Likewise.
5460 * reorg.c (mostly_true_jump): Likewise.
5461 * config/bfin/bfin.c (cbranch_predicted_taken_p): Likewise.
5462 * config/frv/frv.c (frv_print_operand_jump_hint): Likewise.
5463 * config/i386/i386.c (ix86_print_operand): Likewise.
5464 * config/ia64/ia64.c (ia64_print_operand): Likewise.
5465 * config/mmix/mmix.c (mmix_print_operand): Likewise.
5466 * config/rs6000/rs6000.c (output_cbranch): Likewise.
5467 * config/s390/s390.c (s390_expand_tbegin): Likewise.
5468 * config/sh/sh.c (sh_print_operand, sh_adjust_cost): Likewise.
5469 * config/sparc/sparc.c (output_cbranch): Likewise.
5470 * config/spu/spu.c (get_branch_target): Likewise.
5471 * config/tilegx/tilegx.c (cbranch_predicted_p): Likewise.
5472 * config/tilepro/tilepro.c (cbranch_predicted_p): Likewise.
5474 2013-09-23 Jan Hubicka <jh@suse.cz>
5476 * ipa-cp.c (ipa_get_indirect_edge_target_1): Add sanity check
5478 * ipa-utils.h (possible_polymorphic_call_target_p): New function.
5479 * ipa-devirt.c (possible_polymorphic_call_target_p): Be tolerant
5481 * gimple-fold.c: Include ipa-utils.h and gimple-pretty-print.h
5482 (gimple_fold_call): Dump inconsistent devirtualizations; add
5483 sanity check for type based devirtualizations.
5484 * ipa-prop.c: Include ipa-utils.h
5485 (ipa_intraprocedural_devirtualization): Add sanity check.
5486 (try_make_edge_direct_virtual_call): Likewise.
5488 2013-09-23 Eric Botcazou <ebotcazou@adacore.com>
5490 * tree-ssa-ccp.c (insert_clobber_before_stack_restore): Recurse on copy
5491 assignment statements.
5493 2013-09-23 Kugan Vivekanandarajah <kuganv@linaro.org>
5495 * gimple-pretty-print.c (dump_ssaname_info): New function.
5496 (dump_gimple_phi): Call it.
5497 (pp_gimple_stmt_1): Likewise.
5498 * tree-core.h (tree_ssa_name): New union ssa_name_info_type field.
5499 (range_info_def): Declare.
5500 * tree-pretty-print.c (pp_double_int): New function.
5501 (dump_generic_node): Call it.
5502 * tree-pretty-print.h (pp_double_int): Declare.
5503 * tree-ssa-alias.c (dump_alias_info): Check pointer type.
5504 * tree-ssanames.h (range_info_def): New structure.
5505 (value_range_type): Move definition here.
5506 (set_range_info, value_range_type, duplicate_ssa_name_range_info):
5508 * tree-ssanames.c (make_ssa_name_fn): Check pointer type at
5510 (set_range_info): New function.
5511 (get_range_info): Likewise.
5512 (duplicate_ssa_name_range_info): Likewise.
5513 (duplicate_ssa_name_fn): Check pointer type and call
5514 duplicate_ssa_name_range_info.
5515 * tree-ssa-copy.c (fini_copy_prop): Likewise.
5516 * tree-vrp.c (value_range_type): Remove definition, now in
5518 (vrp_finalize): Call set_range_info to update value range of SSA_NAMEs.
5519 * tree.h (SSA_NAME_PTR_INFO): Macro changed to access via union.
5520 (SSA_NAME_RANGE_INFO): New macro.
5522 2013-09-23 Richard Biener <rguenther@suse.de>
5524 PR tree-optimization/58464
5525 * tree-ssa-pre.c (phi_trans_lookup): Remove.
5526 (phi_trans_add): Change to add conditionally on being not present.
5527 (phi_translate_1): Remove recursion detection here.
5528 (phi_translate): Pre-seed the cache with NULL to catch
5529 recursion here in a more generic way.
5530 (bitmap_find_leader): Adjust comment.
5531 (get_representative_for): Dump value-numbers.
5532 (create_expression_by_pieces): Likewise.
5533 (insert_into_preds_of_block): Likewise.
5535 2013-09-23 Christian Bruel <christian.bruel@st.com>
5538 * config/sh/sh.md (movsf_ie): Allow fpul_operand.
5539 * config/sh/predicate.md (arith_reg_operand): Disallow FPUL_REG.
5541 2013-09-23 James Greenhalgh <james.greenhalgh@arm.com>
5544 2013-09-20 Renlin Li <renlin.li@arm.com>
5546 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
5548 (aarch64_expand_epilogue): Likewise.
5549 (aarch64_legitimize_reload_address): Likewise.
5551 2013-09-22 Eric Botcazou <ebotcazou@adacore.com>
5553 * gimplify.c (gimplify_asm_expr): Reset the TREE_CHAIN of clobbers to
5554 NULL_TREE before pushing them onto the vector. Likewise for labels.
5556 2013-09-21 Eric Botcazou <ebotcazou@adacore.com>
5558 * config/ia64/predicates.md (ia64_cbranch_operator): Accept unordered
5559 comparison operators when -fno-trapping-math is in effect.
5560 * config/ia64/ia64.c (ia64_expand_compare): Add support for unordered
5561 comparison operators in TFmode and assert that unsupported operators
5563 (ia64_print_operand): Likewise.
5565 2013-09-21 Jan Hubicka <jh@suse.cz>
5567 * x86-tune.def (partial_reg_stall): Disable for CoreI7 and newer.
5568 (sse_typeless_stores): Enable for core
5569 (sse_load0_by_pxor): Likewise.
5570 (four_jump_limit): Disable for core.
5571 (pad_returns): Likewise.
5572 (avoid_vector_decode): Likewise.
5573 (fuse_cmp_and_branch): Enable for cores.
5574 * i386.c (x86_accumulate_outgoing_args): Disable for cores.
5576 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
5579 * config/pa/pa.c (pa_option_override): Disable auto increment and
5580 decrement instructions until reload is completed.
5582 * config/pa/pa-linux.h (TARGET_OS_CPP_BUILTINS): Define
5583 __GCC_HAVE_SYNC_COMPARE_AND_SWAP_1, __GCC_HAVE_SYNC_COMPARE_AND_SWAP_2,
5584 and __GCC_HAVE_SYNC_COMPARE_AND_SWAP_4.
5586 2013-09-20 DJ Delorie <dj@redhat.com>
5587 Nick Clifton <nickc@redhat.com>
5589 * config/rl78/rl78.c: Various whitespace and comment tweaks.
5590 (need_to_save): Save bank 0 on interrupts.
5591 (characterize_address): Strip far address wrappers.
5592 (rl78_as_legitimate_address): Likewise.
5593 (transcode_memory_rtx): Likewise.
5594 (rl78_peep_movhi_p): Disable this peephole after devirt.
5595 (rl78_propogate_register_origins): Forget all origins when a
5597 * config/rl78/rl78-virt.md: Various whitespace tweaks.
5598 * config/rl78/rl78-real.md: Various whitespace tweaks. Additional
5600 * config/rl78/rl78.md (sel_rb): Disable for G10 just in case.
5601 * config/rl78/rl78-expand.md (movqi): Check for subregs of consts.
5602 * config/rl78/rl78.h (LINK_SPEC): Pass -gc-sections unless
5604 * config/rl78/constraints.md: Various whitespace and paren tweaks.
5606 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
5608 * config/pa/pa.md: In "scc" insn patterns, change output template to
5609 handle const0_rtx in reg_or_0_operand operands.
5611 2013-09-20 Martin Husemann <martin@NetBSD.org>
5614 * config/vax/vax.c (vax_output_int_move): Use D format specifier.
5615 * config/vax/vax.md (ashldi3, <unnamed>): Ditto.
5617 2013-09-20 Richard Biener <rguenther@suse.de>
5620 * tree-scalar-evolution.c (struct scev_info_str): Shrink by
5621 remembering SSA name version and block index.
5622 (new_scev_info_str): Adjust.
5623 (hash_scev_info): Likewise. Also hash the block index.
5624 (eq_scev_info): Adjust.
5625 (find_var_scev_info): Likewise.
5626 (struct instantiate_cache_entry): Remove.
5627 (struct instantiate_cache_type): Use a htab to map name, block
5629 (instantiate_cache_type::~instantiate_cache_type): Adjust.
5630 (get_instantiated_value_entry): Likewise.
5631 (hash_idx_scev_info, eq_idx_scev_info): New functions.
5632 (instantiate_scev_name): Adjust.
5634 2013-09-20 Jeff Law <law@redhat.com>
5636 * tree-ssa-dom.c (record_temporary_equivalences): Add comment.
5638 2013-09-20 Yufeng Zhang <yufeng.zhang@arm.com>
5640 * config/aarch64/aarch64-builtins.c (aarch64_simd_expand_args):
5641 Call aarch64_simd_expand_args to update op[argc].
5643 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
5645 * plugin.c (parse_plugin_arg_opt): Accept equal sign inside
5648 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
5650 * gengtype.c (file_rules): Added rule for *.cc files.
5651 (get_output_file_with_visibility): Give fatal message when no
5654 2013-09-20 Renlin Li <renlin.li@arm.com>
5656 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use plus_constant.
5657 (aarch64_expand_epilogue): Likewise.
5658 (aarch64_legitimize_reload_address): Likewise.
5660 2013-09-20 Bernd Edlinger <bernd.edlinger@hotmail.de>
5663 * expr.c (expand_assignment): Remove misalignp code path.
5665 2013-09-20 Marek Polacek <polacek@redhat.com>
5668 * ubsan.c (get_ubsan_type_info_for_type): Use TYPE_SIZE instead of
5669 TYPE_PRECISION. Add asserts.
5671 2013-09-20 Richard Biener <rguenther@suse.de>
5673 PR tree-optimization/58453
5674 * tree-loop-distribution.c (distribute_loop): Apply the cost
5675 model for -ftree-loop-distribute-patterns, too.
5677 2013-09-20 Richard Biener <rguenther@suse.de>
5680 * tree-chrec.h (build_polynomial_chrec): Use gcc_checking_assert,
5681 make type comparison less strict.
5683 2013-09-20 Alan Modra <amodra@gmail.com>
5685 * configure: Regenerate.
5686 * aclocal.m4: Regenerate.
5688 2013-09-20 Marek Polacek <polacek@redhat.com>
5691 * doc/extend.texi: Document that attribute used is meant to be used
5692 on variables with static storage duration.
5694 2013-09-19 Jakub Jelinek <jakub@redhat.com>
5696 PR tree-optimization/58472
5697 * tree-vect-stmts.c (vectorizable_store, vectorizable_load): For
5698 simd_lane_access set inv_p = false.
5699 * omp-low.c (lower_rec_input_clauses): Set TREE_NO_WARNING on
5700 the simduid magic VAR_DECL.
5702 2013-09-19 Jan Hubicka <jh@suse.cz>
5704 * i386.c (generic_memcpy, generic_memset): Fix 32bit template.
5706 2013-09-17 Jeff Law <law@redhat.com>
5708 * tree-ssa-dom.c (record_temporary_equivalences): New function
5709 split out of dom_opt_dom_walker::after_dom_children.
5710 (dom_opt_dom_walker::thread_across_edge): Move common code
5711 in here from dom_opt_dom_walker::after_dom_children.
5712 (dom_opt_dom_walker::after_dom_children): Corresponding simplifictions.
5714 2013-09-19 Jan Hubicka <jh@suse.cz>
5716 * i386.h (TARGET_GENERIC32, TARGET_GENERIC64): Remove.
5717 (TARGET_GENERIC): Use PROCESOR_GENERIC
5718 (enum processor_type): Unify generic32 and 64.
5719 * i386.md (cpu): Likewise.
5720 * x86-tune.def (use_leave): Enable for generic32.
5721 (avoid_vector_decode, slow_imul_imm32_mem, slow_imul_imm8): Likewise.
5722 * athlon.md: Change generic64 to generic in all occurences.
5723 * i386-c.c (ix86_target_macros_internal): Unify generic64 and 32.
5724 (ix86_target_macros_internal): Likewise.
5725 * driver-i386.c (host_detect_local_cpu): Likewise.
5726 * i386.c (generic64_memcpy, generic64_memset, generic64_cost): Rename
5728 (generic_memcpy, generic_memset, generic_cost): This one.
5729 (generic32_memcpy, generic32_memset, generic32_cost): Remove.
5730 (m_GENERIC32, m_GENERIC64): Remove.
5731 (m_GENERIC): Turn into one flag.
5732 (processor_target): Unify generic tunnings.
5733 (ix86_option_override_internal): Replace generic32/64 by generic.
5734 (ix86_issue_rate): Likewise.
5735 (ix86_adjust_cost): Likewise.
5737 2013-09-19 Jan Hubicka <jh@suse.cz>
5739 * cgraph.c (cgraph_create_edge_1): Avoid uninitialized read
5740 of speculative flag.
5742 2013-09-19 Jakub Jelinek <jakub@redhat.com>
5744 * omp-low.c (expand_omp_sections): Always pass len - 1 to
5745 GOMP_sections_start, even if !exit_reachable.
5747 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
5749 * lra-constraints.c (need_for_all_save_p): Use macro
5750 HARD_REGNO_CALL_PART_CLOBBERED.
5751 * lra-lives.c (check_pseudos_live_through_calls): Use the macro to
5752 set up pseudo conflict hard regs.
5754 2013-09-18 Michael Meissner <meissner@linux.vnet.ibm.com>
5757 * config/rs6000/paired.md (movmisalignv2sf): Fix to allow memory
5760 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
5762 PR rtl-optimization/58438
5763 * lra.c (lra): Clear lra_optional_reload_pseudos in upper loop.
5764 * lra-constraints.c (undo_optional_reloads): Keep optional reloads
5765 from previous subpasses.
5767 2013-09-18 Richard Earnshaw <rearnsha@arm.com>
5769 * arm.c (arm_get_frame_offsets): Validate architecture supports
5770 LDRD/STRD before accepting the tuning preference.
5771 (arm_expand_prologue): Likewise.
5772 (arm_expand_epilogue): Likewise.
5774 2013-09-18 Richard Biener <rguenther@suse.de>
5776 PR tree-optimization/58417
5777 * tree-chrec.c (chrec_fold_plus_1): Assert that we do not
5778 have chrecs with symbols defined in the loop as operands.
5779 (chrec_fold_multiply): Likewise.
5780 * tree-scalar-evolution.c (interpret_rhs_expr): Instantiate
5781 parameters before folding binary operations.
5782 (struct instantiate_cache_entry_hasher): Remove.
5783 (struct instantiate_cache_type): Use a pointer-map.
5784 (instantiate_cache_type::instantiate_cache_type): New function.
5785 (instantiate_cache_type::get): Likewise.
5786 (instantiate_cache_type::set): Likewise.
5787 (instantiate_cache_type::~instantiate_cache_type): Adjust.
5788 (get_instantiated_value_entry): Likewise.
5789 (global_cache): New global.
5790 (instantiate_scev_r, instantiate_scev_poly, instantiate_scev_binary,
5791 instantiate_array_ref, instantiate_scev_convert, instantiate_scev_3,
5792 instantiate_scev_2, instantiate_scev_1): Do not pass along cache.
5793 (instantiate_scev_name): Adjust.
5794 (instantiate_scev): Construct global instead of local cache.
5795 (resolve_mixers): Likewise.
5797 2013-09-18 Daniel Morris <danielm@ecoscentric.com>
5798 Paolo Carlini <paolo.carlini@oracle.com>
5801 * doc/implement-cxx.texi: Fix references to the C++ standards.
5803 2013-09-18 Jakub Jelinek <jakub@redhat.com>
5805 * omp-low.c (copy_var_decl): Copy DECL_ATTRIBUTES.
5806 * tree-vect-data-refs.c (vect_analyze_data_refs): For
5807 simd_lane_access drs, update also DR_ALIGNED_TO.
5809 2013-09-18 Marek Polacek <polacek@redhat.com>
5812 * doc/extend.texi: Document no_sanitize_undefined attribute.
5813 * builtins.c (fold_builtin_0): Don't sanitize function if it has the
5814 no_sanitize_undefined attribute.
5816 2013-09-18 Nick Clifton <nickc@redhat.com>
5818 * config/msp430/msp430.h (ASM_SPEC): Pass -md on to the assembler.
5819 (ASM_DECLARE_FUNCTION_NAME): Define.
5821 2013-09-17 Trevor Saunders <tsaunders@mozilla.com>
5823 * compare-elim.c (find_comparison_dom_walker): New class
5824 (find_comparisons_in_bb): Rename to
5825 find_comparison_dom_walker::before_dom_children
5826 (find_comparisons): Adjust
5827 * domwalk.c (walk_dominator_tree): Rename to dom_walker::walk, and
5829 (init_walk_dominator_tree, fini_walk_dominator_tree): Remove
5830 * domwalk.h (dom_walk_data): Convert it To a class dom_walker.
5831 (init_walk_dominator_tree): Remove declaration.
5832 (fini_walk_dominator_tree): Remove declaration.
5833 * fwprop.c (single_def_use_dom_walker): New class
5834 (single_def_use_enter_block): Convert to
5835 single_def_use_dom_walker::before_dom_children.
5836 (single_def_use_leave_block): Convert to
5837 single_def_use_dom_walker::after_dom_children.
5838 (build_single_def_use_links): Adjust.
5839 * gimple-ssa-strength-reduction.c (find_candidates_dom_walker): New
5841 (find_candidates_in_block): Convert to
5842 find_candidates_dom_walker::before_dom_children.
5843 (execute_strength_reduction): Adjust.
5844 * graphite-sese-to-poly.c (struct bsc, build_sese_conditions): Remove.
5845 (sese_dom_walker): New class.
5846 (sese_dom_walker::sese_dom_walker): New constructor.
5847 (sese_dom_walker::~sese_dom_walker): New destructor.
5848 (build_sese_conditions_before): Convert to
5849 sese_dom_walker::before_dom_children.
5850 (build_sese_conditions_after): Convert to
5851 sese_dom_walker::after_dom_children.
5852 (build_poly_scop): Adjust
5853 * tree-into-ssa.c (rewrite_dom_walker): New class
5854 (rewrite_enter_block): Convert to
5855 rewrite_dom_walker::before_dom_children.
5856 (rewrite_leave_block): Convert to
5857 rewrite_dom_walker::after_dom_children.
5858 (rewrite_update_dom_walker): New class.
5859 (rewrite_update_enter_block): Convert to
5860 rewrite_update_dom_walker::before_dom_children.
5861 (rewrite_update_leave_block): Convert to
5862 rewrite_update_dom_walker::after_dom_children.
5863 (rewrite_blocks, rewrite_into_ssa): Adjust.
5864 (mark_def_dom_walker): New class.
5865 (mark_def_dom_walker::mark_def_dom_walker): New constructor.
5866 (mark_def_dom_walker::~mark_def_dom_walker): New destructor.
5867 (mark_def_sites_blocks): Convert to
5868 mark_def_dom_walker::before_dom_children.
5869 (mark_def_site_blocks): Remove.
5870 * tree-ssa-dom.c (dom_opt_dom_walker): New class.
5871 (tree_ssa_dominator_optimize): Adjust.
5872 (dom_thread_across_edge): Convert to method
5873 dom_opt_dom_walker::thread_across_edge.
5874 (dom_opt_enter_block): Convert to member function
5875 dom_opt_dom_walker::before_dom_children.
5876 (dom_opt_leave_block): Convert to member function
5877 dom_opt_dom_walker::after_dom_children.
5878 * tree-ssa-dse.c (dse_dom_walker): New class.
5879 (dse_enter_block): Convert to member function
5880 dse_dom_walker::before_dom_children.
5881 (tree_ssa_dse): Adjust.
5882 * tree-ssa-loop-im.c (invariantness_dom_walker): New class.
5883 (determine_invariantness_stmt): Convert to method
5884 invariantness_dom_walker::before_dom_children.
5885 (determine_invariantness): Remove
5886 (move_computations_dom_walker): New class.
5887 (move_computations_stmt): Convert to method
5888 move_computations_dom_walker::before_dom_children.
5889 (move_computations, tree_ssa_lim): Adjust.
5890 * tree-ssa-phiopt.c (nontrapping_dom_walker): New class.
5891 (nt_init_block): Convert to method
5892 notrappping_dom_walker::before_dom_children.
5893 (nt_fini_block): Convert to method
5894 method nontrapping_dom_walker::after_dom_children.
5895 (get_non_trapping): Adjust.
5896 * tree-ssa-pre.c (eliminate_dom_walker): New class.
5897 (eliminate_bb): Convert to method
5898 eliminate_dom_walker::before_dom_children.
5899 (eliminate_leave_block): Convert to method
5900 eliminate_dom_walker::after_dom_children.
5901 (eliminate): Adjust.
5902 * tree-ssa-strlen.c (strlen_dom_walker): New class.
5903 (strlen_enter_block): Convert to method
5904 strlen_dom_walker::before_dom_children.
5905 (strlen_leave_block): Convert to method
5906 method strlen_dom_walker::after_dom_children.
5907 (tree_ssa_strlen): Adjust.
5908 * tree-ssa-uncprop.c (uncprop_dom_walker): New class.
5909 (tree_ssa_uncprop): Adjust.
5910 (uncprop_leave_block): Convert to method
5911 uncprop_dom_walker::after_dom_children.
5912 (uncprop_leave_block): Convert to method
5913 uncprop_dom_walker::before_dom_children.
5915 2013-09-18 Bin Cheng <bin.cheng@arm.com>
5917 * config/arm/arm.c (thumb1_reorg): Search for flag setting insn before
5918 branch in same basic block. Check both src and dest of the move insn.
5920 2013-09-17 Nick Clifton <nickc@redhat.com>
5922 * config/rl78/rl78-real.md (bf): New pattern.
5924 * config/rl78/rl78.c (rl78_print_operand_1): Handle %B.
5925 (rl78_print_operand): Do not put a # before a %B.
5926 * config/rl78/rl78.opt: Tweak doc strings.
5928 2013-09-17 DJ Delorie <dj@redhat.com>
5930 * config/rl78/constraints.md (Wcv): Allow up to $r31.
5931 * config/rl78/rl78.c (rl78_asm_file_start: Likewise.
5932 (rl78_option_override): Likewise, if -mallregs.
5933 (is_virtual_register): Likewise.
5934 * config/rl78/rl78.h (reg_class): Extend VREGS to $r31.
5935 (REGNO_OK_FOR_BASE_P): Likewise.
5936 * config/rl78/rl78.opt (-mallregs): New.
5938 2013-09-17 Nick Clifton <nickc@redhat.com>
5940 * config/rl78/rl78.c (need_to_save): Change return type to bool.
5941 For interrupt functions: save all call clobbered registers if the
5942 interrupt handler is not a leaf function.
5943 (rl78_expand_prologue): Always recompute the frame information.
5944 For interrupt functions: only select bank 0 if one of the bank 0
5945 registers is going to be psuhed.
5947 2013-09-17 DJ Delorie <dj@redhat.com>
5949 * config/rl78/constraints.md: For each W* constraint, rename to C*
5950 and create a W* constraint that checks for an optional ES: prefix
5952 * config/rl78/rl78.md (UNS_ES_ADDR): New.
5953 (es_addr): New. Used to wrap far addresses.
5954 * config/rl78/rl78-protos.h (rl78_es_addr): New.
5955 (rl78_es_base): New.
5956 * config/rl78/rl78.c (rl78_as_legitimate_address): Accept "unspec"
5957 wrapped far addresses.
5958 (rl78_print_operand_1): Unwrap far addresses before processing.
5959 (rl78_lo16): Wrap far addresses in unspecs.
5960 (rl78_es_addr): New.
5961 (rl78_es_base): New.
5962 (insn_ok_now): Check for not-yet-wrapped far addresses.
5963 (transcode_memory_rtx): Properly re-wrap far addresses.
5965 2013-09-17 Sebastian Huber <sebastian.huber@embedded-brains.de>
5967 * config/sparc/t-rtems: Add leon3 multilibs.
5969 2013-09-17 Cong Hou <congh@google.com>
5971 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Fix a bug
5972 when checking the dot production pattern. The type of rhs operand
5973 of multiply is now checked correctly.
5975 2013-09-17 Jeff Law <law@redhat.com>
5977 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
5978 edge implied equivalences into successor phis.
5979 * tree-ssa-threadupdate.c (phi_args_equal_on_edges): Moved into
5980 here from tree-ssa-threadedge.c.
5981 (mark_threaded_blocks): When threading through a joiner, if both
5982 successors of the joiner's clone reach the same block, verify the
5983 PHI arguments are equal. If not, cancel the jump threading request.
5984 * tree-ssa-threadedge.c (phi_args_equal_on_edges): Moved into
5985 tree-ssa-threadupdate.c
5986 (thread_across_edge): Don't check PHI argument equality when
5987 threading through joiner block here.
5989 2013-09-17 Andrew MacLeod <amacleod@redhat.com>
5991 * tree-flow.h (ssa_undefined_value_p): Remove prototype.
5992 * tree-ssa.c (ssa_undefined_value_p): Move pass independent parts here.
5993 (warn_uninit, warn_uninitialized_vars,
5994 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
5995 Move to tree-ssa-uninit.c.
5996 * tree-ssa-uninit.c (ssa_undefined_value_p): Move to tree-ssa.c.
5997 (has_undefined_value_p): New. Pass dependant parts of
5998 ssa_undefined_value_p.
5999 (uninit_undefined_value_p): Use has_undefined_value_p.
6000 (warn_uninit, warn_uninitialized_vars,
6001 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
6002 Move from tree-ssa.c.
6003 * tree-ssa.h: Adjust prototypes.
6005 2013-09-17 Jan Hubicka <jh@suse.cz>
6008 * cif-code.def (FUNCTION_NOT_OPTIMIZED): New CIF code.
6009 * ipa-inline.c (can_inline_edge_p): Do not downgrade
6010 FUNCTION_NOT_OPTIMIZED.
6011 * ipa-inline-analysis.c (compute_inline_parameters): Function
6012 not optimized is not inlinable unless it is alwaysinline.
6013 (inline_analyze_function): Force calls in not optimized
6014 function not inlinable.
6016 2013-09-17 Jan Hubicka <jh@suse.cz>
6019 * ipa-devirt.c (ipa_devirt): Be ready for symtab_nonoverwritable_alias
6021 * ipa.c (function_and_variable_visibility): Likewise.
6022 * ipa-profile.c (ipa_profile): Likewise.
6024 2013-09-17 Bernd Edlinger <bernd.edlinger@hotmail.de>
6027 * cgraph.c (cgraph_function_body_availability): Check for ifunc
6028 attribute, and don't inline the resolver in this case.
6030 2013-09-17 Teresa Johnson <tejohnson@google.com>
6032 * coverage.c (get_coverage_counts): Add missing newline.
6034 2013-09-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6036 PR tree-optimization/58088
6037 * fold-const.c (mask_with_trailing_zeros): New function.
6038 (fold_binary_loc): Make sure we don't recurse infinitely
6039 when the X in (X & C1) | C2 is a tree of the form (Y * K1) & K2.
6040 Use mask_with_trailing_zeros where appropriate.
6042 2013-09-17 Yuri Rumyantsev <ysrumyan@gmail.com>
6044 * config/i386/i386.c (distance_agu_use_in_bb) : Proper initialization
6045 of 'prev' var to get better distance estimation.
6047 2013-09-17 Eric Botcazou <ebotcazou@adacore.com>
6049 * tree-inline.h (struct copy_body_data): Add transform_parameter.
6050 * tree-inline.c (is_parameter_of): New predicate.
6051 (remap_gimple_op_r): Do not propagate TREE_THIS_NOTRAP on MEM_REF if
6052 a parameter has been remapped.
6053 (copy_tree_body_r): Likewise on INDIRECT_REF and MEM_REF.
6054 (optimize_inline_calls): Initialize transform_parameter.
6055 (copy_gimple_seq_and_replace_locals): Likewise.
6056 (tree_function_versioning): Likewise.
6057 (maybe_inline_call_in_expr): Likewise.
6059 2013-09-17 Nick Clifton <nickc@redhat.com>
6061 * config/msp430/msp430-protos.h: Add prototypes for new functions.
6062 * config/msp430/msp430.c (msp430_preserve_reg_p): Add support for
6064 (is_attr_func): New function.
6065 (msp430_is_interrupt_func): New function.
6066 (is_naked_func): New function.
6067 (is_reentrant_func): New function.
6068 (is_critical_func): New function.
6069 (msp430_start_function): Add annotations for function attributes.
6070 (msp430_attr): New function.
6071 (msp430_attribute_table): New.
6072 (msp430_function_section): New function.
6073 (TARGET_ASM_FUNCTION_SECTION): Define.
6074 (msp430_builtin): New enum.
6075 (msp430_init_builtins): New function.
6076 (msp430_builtin_devl): New function.
6077 (msp430_expand_builtin): New function.
6078 (TARGET_INIT_BUILTINS): Define.
6079 (TARGET_EXPAND_BUILTINS): Define.
6080 (TARGET_BUILTIN_DECL): Define.
6081 (msp430_expand_prologue): Add support for naked, interrupt,
6082 critical and reentrant functions.
6083 (msp430_expand_epilogue): Likewise.
6084 (msp430_print_operand): Handle 'O' character.
6085 * config/msp430/msp430.h (TARGET_CPU_CPP_BUILTINS): Define
6087 * config/msp430/msp430.md (unspec): Add UNS_DINT, UNS_EINT,
6088 UNS_PUSH_INTR, UNS_POP_INTR, UNS_BIC_SR, UNS_BIS_SR.
6089 (pushm): Use a 'n' rather than an 'i' constraint.
6090 (msp_return): Add generation of the interrupt return instruction.
6091 (disable_interrupts): New pattern.
6092 (enable_interrupts): New pattern.
6093 (push_intr_state): New pattern.
6094 (pop_intr_state): New pattern.
6095 (bic_SR): New pattern.
6096 (bis_SR): New pattern.
6097 * doc/extend.texi: Document MSP430 function attributes and builtin
6100 2013-09-17 Richard Biener <rguenther@suse.de>
6102 PR tree-optimization/58432
6103 * tree-loop-distribution.c (tree_loop_distribution): Also
6104 scan PHIs for outside loop uses and seed a partition from them.
6106 2013-09-17 Bin Cheng <bin.cheng@arm.com>
6108 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref): New.
6109 (restructure_reference): Call backtrace_base_for_ref.
6111 2013-09-17 Alan Modra <amodra@gmail.com>
6114 * config/rs6000/driver-rs6000.c (elf_platform): Revert 2013-06-11
6117 2013-09-16 DJ Delorie <dj@redhat.com>
6119 * config/rl78/rl78.c (rl78_asm_file_start): Specify alternate
6120 vregs location for RL78/G10.
6121 (rl78_expand_prologue): Avoid SEL on G10.
6122 (rl78_expand_epilogue): Likewise.
6123 (rl78_peep_movhi_p): Can't move a constant to memory in HImode.
6124 * config/rl78/rl78.h (TARGET_CPU_CPP_BUILTINS): Define
6125 __RL78_G10__ when appropriate.
6126 (ASM_SPEC): Pass -mg10 along to the assembler.
6127 * config/rl78/rl78.md (sel_rb): Disable for G10.
6128 * config/rl78/rl78.opt: Add -mg10 option.
6129 * config/rl78/t-rl78: Add -mg10 multilib.
6131 2013-09-16 Xinliang David Li <davidxl@google.com>
6133 * tree-if-conv.c (main_tree_if_conversion): Check new flag.
6134 * omp-low.c (omp_max_vf): Ditto.
6135 (expand_omp_simd): Ditto.
6136 * tree-vectorizer.c (vectorize_loops): Ditto.
6137 (gate_vect_slp): Ditto.
6138 (gate_increase_alignment): Ditto.
6139 * tree-ssa-pre.c (inhibit_phi_insertion): Ditto.
6140 * tree-ssa-loop.c (gate_tree_vectorize): Ditto.
6141 (gate_tree_vectorize): Name change.
6142 (tree_vectorize): Ditto.
6143 (pass_vectorize::gate): Call new function.
6144 (pass_vectorize::execute): Ditto.
6145 * opts.c: O3 default setting change.
6146 (finish_options): Check new flag.
6147 * doc/invoke.texi: Document new flags.
6148 * common.opt: New flags.
6150 2013-09-16 Andreas Schwab <schwab@linux-m68k.org>
6152 * doc/tm.texi.in (Cond Exec Macros): Remove node.
6153 (Condition Code): Don't reference it.
6154 * doc/tm.texi: Regenerate.
6156 2013-09-16 Vladimir Makarov <vmakarov@redhat.com>
6159 * lra-constraints.c (undo_optional_reloads): Consider all optional
6160 reload even if it did not get a hard reg.
6162 2013-09-16 Teresa Johnson <tejohnson@google.com>
6164 * dumpfile.c (dump_loc): Remove newline emission.
6165 * tree-vect-data-refs.c (vect_lanes_optab_supported_p): Add newline
6166 emission to dump_printf_loc calls where missing.
6167 (vect_mark_for_runtime_alias_test): Ditto.
6168 (vect_analyze_data_ref_dependence): Ditto.
6169 (vect_analyze_data_ref_dependences): Ditto.
6170 (vect_slp_analyze_data_ref_dependence): Ditto.
6171 (vect_slp_analyze_data_ref_dependences): Ditto.
6172 (vect_compute_data_ref_alignment): Ditto.
6173 (vect_update_misalignment_for_peel): Ditto.
6174 (vect_verify_datarefs_alignment): Ditto.
6175 (vector_alignment_reachable_p): Ditto.
6176 (vect_get_data_access_cost): Ditto.
6177 (vect_enhance_data_refs_alignment): Ditto.
6178 (vect_find_same_alignment_drs): Ditto.
6179 (vect_analyze_data_refs_alignment): Ditto.
6180 (vect_analyze_group_access): Ditto.
6181 (vect_analyze_data_ref_access): Ditto.
6182 (vect_analyze_data_ref_accesses): Ditto.
6183 (vect_prune_runtime_alias_test_list): Ditto.
6184 (vect_analyze_data_refs): Ditto.
6185 (vect_create_addr_base_for_vector_ref): Ditto.
6186 (vect_create_data_ref_ptr): Ditto.
6187 (vect_grouped_store_supported): Ditto.
6188 (vect_grouped_load_supported): Ditto.
6189 * value-prof.c (check_counter): Ditto.
6190 (check_ic_target): Ditto.
6191 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Ditto.
6192 (vect_recog_widen_mult_pattern): Ditto.
6193 (vect_recog_widen_sum_pattern): Ditto.
6194 (vect_recog_over_widening_pattern): Ditto.
6195 (vect_recog_widen_shift_pattern): Ditto.
6196 (vect_recog_rotate_pattern): Ditto.
6197 (vect_recog_vector_vector_shift_pattern): Ditto.
6198 (vect_recog_divmod_pattern): Ditto.
6199 (vect_recog_mixed_size_cond_pattern): Ditto.
6200 (vect_recog_bool_pattern): Ditto.
6201 (vect_pattern_recog_1): Ditto.
6202 (vect_pattern_recog): Ditto.
6203 * tree-vect-loop.c (vect_determine_vectorization_factor): Ditto.
6204 (vect_is_simple_iv_evolution): Ditto.
6205 (vect_analyze_scalar_cycles_1): Ditto.
6206 (vect_get_loop_niters): Ditto.
6207 (vect_analyze_loop_1): Ditto.
6208 (vect_analyze_loop_form): Ditto.
6209 (vect_analyze_loop_operations): Ditto.
6210 (vect_analyze_loop_2): Ditto.
6211 (vect_analyze_loop): Ditto.
6212 (report_vect_op): Ditto.
6213 (vect_is_slp_reduction): Ditto.
6214 (vect_is_simple_reduction_1): Ditto.
6215 (vect_get_known_peeling_cost): Ditto.
6216 (vect_estimate_min_profitable_iters): Ditto.
6217 (vect_model_reduction_cost): Ditto.
6218 (vect_model_induction_cost): Ditto.
6219 (get_initial_def_for_induction): Ditto.
6220 (vect_create_epilog_for_reduction): Ditto.
6221 (vectorizable_reduction): Ditto.
6222 (vectorizable_induction): Ditto.
6223 (vectorizable_live_operation): Ditto.
6224 (vect_loop_kill_debug_uses): Ditto.
6225 (vect_transform_loop): Ditto.
6226 * tree-vect-stmts.c (vect_mark_relevant): Ditto.
6227 (vect_stmt_relevant_p): Ditto.
6228 (process_use): Ditto.
6229 (vect_mark_stmts_to_be_vectorized): Ditto.
6230 (vect_model_simple_cost): Ditto.
6231 (vect_model_promotion_demotion_cost): Ditto.
6232 (vect_model_store_cost): Ditto.
6233 (vect_get_store_cost): Ditto.
6234 (vect_model_load_cost): Ditto.
6235 (vect_get_load_cost): Ditto.
6236 (vect_init_vector_1): Ditto.
6237 (vect_get_vec_def_for_operand): Ditto.
6238 (vect_finish_stmt_generation): Ditto.
6239 (vectorizable_call): Ditto.
6240 (vectorizable_conversion): Ditto.
6241 (vectorizable_assignment): Ditto.
6242 (vectorizable_shift): Ditto.
6243 (vectorizable_operation): Ditto.
6244 (vectorizable_store): Ditto.
6245 (vectorizable_load): Ditto.
6246 (vectorizable_condition): Ditto.
6247 (vect_analyze_stmt): Ditto.
6248 (vect_transform_stmt): Ditto.
6249 (vect_is_simple_use): Ditto.
6250 * tree-vect-loop-manip.c (slpeel_make_loop_iterate_ntimes): Ditto.
6251 (vect_can_advance_ivs_p): Ditto.
6252 (vect_update_ivs_after_vectorizer): Ditto.
6253 (vect_do_peeling_for_loop_bound): Ditto.
6254 (vect_gen_niters_for_prolog_loop): Ditto.
6255 (vect_update_inits_of_drs): Ditto.
6256 (vect_create_cond_for_alias_checks): Ditto.
6257 * tree-vect-slp.c (vect_get_and_check_slp_defs): Ditto.
6258 (vect_build_slp_tree_1): Ditto.
6259 (vect_supported_load_permutation_p): Ditto.
6260 (vect_analyze_slp_instance): Ditto.
6261 (vect_analyze_slp): Ditto.
6262 (vect_make_slp_decision): Ditto.
6263 (vect_detect_hybrid_slp): Ditto.
6264 (vect_bb_vectorization_profitable_p): Ditto.
6265 (vect_slp_analyze_bb_1): Ditto.
6266 (vect_update_slp_costs_according_to_vf): Ditto.
6267 (vect_get_mask_element): Ditto.
6268 (vect_transform_slp_perm_load): Ditto.
6269 (vect_schedule_slp_instance): Ditto.
6270 (vect_schedule_slp): Ditto.
6271 (vect_slp_transform_bb): Ditto.
6272 * profile.c (read_profile_edge_counts): Ditto.
6273 (compute_branch_probabilities): Ditto.
6274 * coverage.c (get_coverage_counts): Ditto.
6276 2013-09-16 Diego Novillo <dnovillo@google.com>
6278 * tree-core.h: Add missing comment lines from refactoring of tree.h.
6280 2013-09-16 Jan Hubicka <jh@suse.cz>
6282 * gimple-fold.c (can_refer_decl_in_current_unit_p): Do not accept
6283 abstract functions; for static functions check the presence of body.
6285 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
6287 * config/aarch64/aarch64-simd-builtins.def (fma): New.
6288 * config/aarch64/aarch64-simd.md
6289 (aarch64_mla_elt<mode>): New.
6290 (aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
6291 (aarch64_mls_elt<mode>): Likewise.
6292 (aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
6293 (aarch64_fma4_elt<mode>): Likewise.
6294 (aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
6295 (aarch64_fma4_elt_to_128v2df): Likewise.
6296 (aarch64_fma4_elt_to_64df): Likewise.
6297 (fnma<mode>4): Likewise.
6298 (aarch64_fnma4_elt<mode>): Likewise.
6299 (aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
6300 (aarch64_fnma4_elt_to_128v2df): Likewise.
6301 (aarch64_fnma4_elt_to_64df): Likewise.
6302 * config/aarch64/iterators.md (VDQSF): New.
6303 * config/aarch64/arm_neon.h
6304 (vfm<as><sdq>_lane<q>_f<32, 64>): Convert to C implementation.
6305 (vml<sa><q>_lane<q>_<fsu><16, 32, 64>): Likewise.
6307 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
6309 * config/aarch64/aarch64-simd.md (aarch64_mul3_elt<mode>): New.
6310 (aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
6311 (aarch64_mul3_elt_to_128df): Likewise.
6312 (aarch64_mul3_elt_to_64v2df): Likewise.
6313 * config/aarch64/iterators.md (VEL): Also handle DFmode.
6315 (VMUL_CHANGE_NLANES) Likewise.
6318 * config/aarch64/arm_neon.h
6319 (vmul<q>_lane<q>_<suf><16,32,64>): Convert to C implementation.
6321 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
6323 * config/aarch64/arm_neon.h
6324 (vcvtx_high_f32_f64): Fix parameters.
6326 2013-09-16 Jan-Benedict Glaw <jbglaw@lug-owl.de>
6327 Uros Bizjak <ubizjak@gmail.com>
6329 * config/alpha.c: Include tree-ssanames.h.
6331 2013-09-16 Richard Biener <rguenther@suse.de>
6333 * tree-loop-distribution.c (enum rdg_dep_type): Add control_dd.
6334 (dot_rdg_1): Handle control_dd.
6335 (create_edge_for_control_dependence): New function.
6336 (create_rdg_edges): Add control dependences if asked for.
6337 (build_rdg): Likewise.
6338 (generate_loops_for_partition): If there are not necessary
6339 control stmts remove all their dependencies.
6340 (collect_condition_stmts, rdg_flag_loop_exits): Remove.
6341 (distribute_loop): Pass on control dependences.
6342 (tree_loop_distribution): Compute control dependences and remove
6343 restriction on number of loop nodes.
6345 2013-09-16 Jakub Jelinek <jakub@redhat.com>
6347 * ipa-prop.c (ipa_compute_jump_functions_for_edge): Return early
6350 2013-09-16 Richard Sandiford <rdsandiford@googlemail.com>
6352 * cse.c (try_const_anchors): Punt on CC modes.
6354 2013-09-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
6356 * config/vax/constraints.md (T): Add missing CONSTANT_P check.
6358 2013-09-14 John David Anglin <danglin@gcc.gnu.org>
6361 * config/pa/pa.c (pa_expand_prologue): Change mode in gen_rtx_POST_INC
6364 2013-09-14 Iain Sandoe <iain@codesourcery.com>
6367 * config/darwin.c (darwin_objc2_section): Note if ObjC Metadata is
6369 (darwin_objc1_section): Likewise.
6370 (darwin_file_end): Emit Image Info section when required.
6372 2013-09-14 Jan Hubicka <jh@suse.cz>
6374 * tree-into-ssa.c (gate_into_ssa): New.
6375 (pass_data_build_ssa): Use it.
6376 * cgraph.h (expand_thunk): Update prototype.
6377 * cgraphunit.c (analyze_function): Expand thunks early.
6378 (expand_thunk): Fix DECL_CONTEXT of reust_decl;
6379 build proper cgraph; set in_ssa_p; clear bogus TREE_ASM_WRITTEN;
6380 set lowered flag; do not add new function.
6381 (assemble_thunks_and_aliases): Update.
6382 * tree-ssa.c (gate_init_datastructures): New gate.
6383 (pass_data_init_datastructures): Use it.
6385 2013-09-14 Iain Sandoe <iain@codesourcery.com>
6388 * config/i386/i386.c (ix86_function_arg_regno_p): Make Darwin use the
6389 xmm register set described in the psABI.
6391 2013-09-13 Evgeny Gavrin <e.gavrin@samsung.com>
6393 * dwarf2out.c (should_emit_struct_debug): Add check
6394 for type_decl variable is not NULL.
6396 2013-09-13 Jacek Caban <jacek@codeweavers.com>
6398 * config.gcc: Use new winnt-c.c target hooks
6399 * config/t-winnt: New file
6400 * config/winnt-c.c: New file
6401 * doc/tm.texi.in: Document new hook
6402 * doc/tm.texi: Regenerated
6404 2013-09-13 Jan Hubicka <jh@suse.cz>
6407 * ipa-inline.c (check_callers): New function.
6408 (check_caller_edge): Remove.
6409 (want_inline_function_to_all_callers_p): Also permit alises that are
6411 (inline_to_all_callers): Terminate the walk when devirtualization
6412 introduce new calls.
6414 2013-09-13 Jan Hubicka <jh@suse.cz>
6416 * ipa-inline-analysis.c (struct growth_data): Add node.
6417 (do_estimate_growth_1): Fix detection of recursion.
6419 2013-09-13 Jakub Jelinek <jakub@redhat.com>
6421 PR tree-optimization/58392
6422 * tree-cfg.c (move_sese_region_to_fn): Rename loop variable
6423 to avoid shadowing of outer loop variable. If
6424 saved_cfun->has_simduid_loops or saved_cfun->has_force_vect_loops,
6425 replace_by_duplicate_decl simduid of loops that have it set and
6426 set dest_cfun->has_simduid_loops and/or
6427 dest_cfun->has_force_vect_loops.
6428 * omp-low.c (build_outer_var_ref): Call maybe_lookup_decl_in_outer_ctx
6429 instead of maybe_lookup_decl.
6430 * tree-inline.c (copy_loops): Change blocks_to_copy argument to id.
6431 Use id->blocks_to_copy instead of blocks_to_copy. Adjust recursive
6432 call. Copy over force_vect and copy and remap simduid. Set
6433 cfun->has_simduid_loops and/or cfun->has_force_vect_loops.
6434 (copy_cfg_body): Remove blocks_to_copy argument. Use
6435 id->blocks_to_copy instead of blocks_to_copy. Adjust copy_loops
6436 caller. Don't set cfun->has_simduid_loops and/or
6437 cfun->has_force_vect_loops here.
6438 (copy_body): Remove blocks_to_copy argument. Adjust copy_cfg_body
6440 (expand_call_inline, tree_function_versioning): Adjust copy_body
6443 2013-09-13 Martin Jambor <mjambor@suse.cz>
6446 * ipa-prop.c (try_make_edge_direct_simple_call): Be less strict in
6447 the assert if the edge was a speculative one.
6449 2013-09-13 Richard Biener <rguenther@suse.de>
6451 * tree-data-ref.h (known_dependences_p): Move here ...
6452 * tree-loop-distribution.c (known_dependences_p): ... from here.
6453 (dump_rdg_component, debug_rdg_component): Remove.
6455 (generate_loops_for_partition): Use gimple_uid instead of
6456 relying on matching stmt visit order.
6457 (rdg_build_partitions): Take starting stmt vector.
6458 (ldist_gen): Merge into ...
6459 (distribute_loop): ... this function. Do not compute starting
6461 * tree-cfg.c (gimple_duplicate_bb): Copy UID for PHIs.
6463 2013-09-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6465 * config/arm/arm.md (arm_cmpsi_insn): Split rI alternative.
6466 Set type attribute correctly. Set predicable_short_it attribute.
6467 (cmpsi_shiftsi): Remove %? from output template.
6469 2013-09-13 Richard Biener <rguenther@suse.de>
6471 * tree-loop-distribution.c (struct rdg_component,
6472 rdg_defs_used_in_other_loops_p, free_rdg_components,
6473 rdg_build_components): Remove.
6474 (stmts_from_loop): Do not record virtual PHIs.
6475 (generate_loops_for_partition): Skip virtual PHIs.
6476 (build_rdg_partition_for_component): Rename to ...
6477 (build_rdg_partition_for_vertex): ... this and adjust.
6478 (rdg_build_partitions): Take a vector of starting vertices
6479 instead of components. Remove unnecessary leftover handling.
6480 (ldist_gen): Do not build components or record other stores.
6481 (distribute_loop): Do not distribute loops containing stmts
6484 2013-09-13 Christian Bruel <christian.bruel@st.com>
6487 * config/sh/sh.md (mov<mode>_reg_reg): Allow memory reloads.
6489 2013-09-13 Kai Tietz <ktietz@redhat.com>
6491 * config.gcc: Separate cases for mingw and cygwin targets,
6492 and add 64-bit cygwin target case.
6494 * config/i386/winnt-cxx.c (i386_pe_type_dllexport_p): Don't
6495 dll-export inline-functions.
6496 * config/i386/winnt.c (i386_pe_determine_dllexport_p): Likewise.
6499 2013-09-13 Jeff Law <law@redhat.com>
6503 2013-09-06 Jeff Law <law@redhat.com>
6505 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
6506 edge implied equivalences into successor phis.
6508 2013-09-12 DJ Delorie <dj@redhat.com>
6510 * config/rl78/rl78-virt.md: Change from | to \; for asm line
6513 2013-09-12 Brooks Moses <bmoses@google.com>
6516 * Makefile.in: Do not install driver binaries in $(target)/bin.
6518 2013-09-12 DJ Delorie <dj@redhat.com>
6520 * config/rl78/rl78.opt (mrelax): New.
6521 * config/rl78/rl78.h (ASM_SPEC): New, pass on -mrelax to gas.
6522 * config/rl78/rl78.h (LINK_SPEC): New, pass on -mrelax to ld.
6524 * config/rl78/rl78.c (rl78_expand_prologue): Use AX to copy
6526 (rl78_expand_epilogue): Likewise.
6528 2013-09-12 Vladimir Makarov <vmakarov@redhat.com>
6531 * lra-eliminations.c (remove_reg_equal_offset_note): New.
6532 (eliminate_regs_in_insn): Rewrite frame pointer to hard frame
6533 pointer elimination with using remove_reg_equal_offset_note.
6535 2013-09-12 DJ Delorie <dj@redhat.com>
6537 * config/msp430/: New port.
6538 * config.gcc (msp430): Added.
6539 * doc/invoke.texi: Document MSP430 options.
6540 * doc/install.texi: Document msp430-elf
6541 * doc/md.texi: Document msp430-elf
6542 * doc/contrib.texi: Document msp430-elf
6544 * cfgexpand.c (expand_debug_expr): Avoid sign-extending SImode to
6547 2013-09-12 Martin Jambor <mjambor@suse.cz>
6550 * ipa-prop.c (remove_described_reference): Give up if the edge in the
6551 reference descriptor is NULL.
6552 (ipa_edge_removal_hook): If owning a reference descriptor, set its
6555 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
6557 * tree-flow.h (FREE_SSANAMES): Move to tree-ssanames.c
6558 (SSANAMES, MODIFIED_NORETURN_CALLS, DEFAULT_DEFS, ptr_info_def,
6559 num_ssa_names, ssa_name): Move to tree-ssanames.h + prototypes.
6560 * tree-flow-inline.h (make_ssa_name, copy_ssa_name, duplicate_ssa_name,
6561 make_temp_ssa_name): move to tree-ssanames.h
6562 * tree-ssa-alias.h: Move prototype.
6563 * tree-ssa.h: Include tree-ssanames.h.
6564 * tree-ssanames.c (FREE_SSANAMES): Move to here.
6565 * tree-ssanames.h: New. Move items from tree-flow*.h
6566 * Makefile.in (tree-ssanames.h): Add to tree-ssanames.o and GTFILES.
6568 2013-09-12 Richard Biener <rguenther@suse.de>
6570 PR tree-optimization/58404
6571 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Also
6572 propagate non-invariant addresses into dereferences wrapped
6573 in component references.
6575 2013-09-12 Richard Biener <rguenther@suse.de>
6577 PR tree-optimization/58402
6578 * passes.def: Move pass_late_warn_uninitialized later.
6580 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
6582 * tree-ssa.h: New. Move content from tree-flow.h and
6584 * tree-flow.h (_edge_var_map, edge_var_map_vector): Move to tree-ssa.h.
6585 Move prototypes belonging to tree-ssa.c.
6586 * tree-flow-inline.h (redirect_edge_var_map_def,
6587 redirect_edge_var_map_result, redirect_edge_var_map_location): Move to
6589 * gimple.h: Adjust prototypes.
6590 * tree-ssa.c (useless_type_conversion_p, types_compatible_p): Move
6592 * gimple.c (useless_type_conversion_p, types_compatible_p): Here.
6593 * tree.h: Move prototype to tree-ssa.h.
6594 * gengtype.c (open_base_files): Replace tree-flow.h with tree-ssa.h.
6595 * Makefile.in: (TREE_SSA_H, TREE_FLOW_H): Adjust dependencies.
6596 * alias.c, asan.c, builtins.c, calls.c, cfgexpand.c, cfghooks.c,
6597 cfgloop.c, cfgloopmanip.c, cgraph.c, cgraphbuild.c, cgraphclones.c,
6598 cgraphunit.c, dse.c, except.c, expr.c, final.c, fold-const.c,
6599 ggc-page.c, gimple-fold.c, gimple-iterator.c, gimple-low.c,
6600 gimple-pretty-print.c, gimple-ssa-strength-reduction.c,
6601 gimple-streamer-in.c, gimple-streamer-out.c, gimple.c, gimplify.c,
6602 graphite-blocking.c, graphite-clast-to-gimple.c,
6603 graphite-dependences.c, graphite-interchange.c,
6604 graphite-optimize-isl.c, graphite-poly.c, graphite-scop-detection.c,
6605 graphite-sese-to-poly.c, graphite.c, ipa-cp.c, ipa-inline-analysis.c,
6606 ipa-inline-transform.c, ipa-inline.c, ipa-prop.c, ipa-pure-const.c,
6607 ipa-reference.c, ipa-split.c, ipa-utils.c,
6608 loop-init.c, lto-cgraph.c, lto-section-in.c, lto-section-out.c,
6609 lto-streamer-in.c, lto-streamer-out.c, lto-streamer.c, omp-low.c,
6610 passes.c, predict.c, print-tree.c, profile.c, sese.c, targhooks.c,
6611 tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
6612 tree-chrec.c, tree-complex.c, tree-data-ref.c, tree-dfa.c, tree-eh.c,
6613 tree-emutls.c, tree-if-conv.c, tree-inline.c, tree-into-ssa.c,
6614 tree-loop-distribution.c, tree-mudflap.c, tree-nested.c, tree-nrv.c,
6615 tree-object-size.c, tree-optimize.c, tree-outof-ssa.c, tree-parloops.c,
6616 tree-phinodes.c, tree-predcom.c, tree-pretty-print.c, tree-profile.c,
6617 tree-scalar-evolution.c, tree-sra.c, tree-ssa*.c, tree-stdarg.c,
6618 tree-streamer-in.c, tree-switch-conversion.c, tree-tailcall.c,
6619 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
6620 tree-vect-loop.c, tree-vect-patterns.c, tree-vect-slp.c,
6621 tree-vect-stmts.c, tree-vectorizer.c, tree-vrp.c, tsan.c,
6622 value-prof.c, var-tracking.c,
6623 varpool.c, vtable-verify.c: Replace tree-flow.h with tree-ssa.h
6625 2013-09-12 Richard Biener <rguenther@suse.de>
6627 PR tree-optimization/58396
6628 * tree-loop-distribution.c (create_rdg_edges): Free unused DDRs.
6629 (build_rdg): Take a loop-nest parameter, fix memleaks.
6630 (distribute_loop): Compute loop-nest here and pass it to build_rdg.
6632 2013-09-12 Yuri Rumyantsev <ysrumyan@gmail.com>
6634 * config/i386/x86-tune.def: Turn on X86_TUNE_AVOID_MEM_OPND_FOR_CMOVE
6637 2013-09-12 Cameron McInally <cameron.mcinally@nyu.edu>
6639 * doc/extend.texi: Fix errors in x86 FMA builtin naming.
6640 The FMA instruction names should have a 'v' prefix.
6642 2013-09-12 Richard Biener <rguenther@suse.de>
6644 * tree-loop-distribution.c (dot_rdg_1): Make graph prettier.
6645 (dot_rdg): Use popen instead of system in optional code.
6646 (remaining_stmts, upstream_mem_writes): Remove global bitmaps.
6647 (already_processed_vertex_p): Adjust.
6648 (has_anti_or_output_dependence, predecessor_has_mem_write,
6649 mark_nodes_having_upstream_mem_writes, has_upstream_mem_writes,
6650 rdg_flag_uses): Remove.
6651 (rdg_flag_vertex): Simplify.
6652 (rdg_flag_vertex_and_dependent): Rely on a correct RDG and
6654 (build_rdg_partition_for_component): Process the first vertex
6655 of a component only.
6656 (ldist_gen): Do not compute remaining_stmts or upstream_mem_writes.
6658 2013-09-12 Alan Modra <amodra@gmail.com>
6660 * config/rs6000/rs6000.c (toc_relative_expr_p): Use add_cint_operand.
6662 2013-09-11 DJ Delorie <dj@redhat.com>
6663 Nick Clifton <nickc@redhat.com>
6665 * config/rl78/predicates.md (rl78_cmp_operator_signed): New.
6666 (rl78_stack_based_mem): New.
6667 * config/rl78/constraints.md (Iv08): New.
6676 * config/rl78/rl78-expand.md (movqi): Reject more SUBREG operands.
6678 (movsi): Change from expand to insn-and-split.
6679 (ashrsi3): Clobber AX.
6683 * config/rl78/rl78.md (CC_REG): Fix.
6684 (addsi3): Allow memory and immediate operands.
6685 (addsi3_internal): Split into...
6686 (addsi3_internal_virt): ...new, and ...
6687 (addsi3_internal_real): ...new.
6689 (subsi3_internal_virt): New.
6690 (subsi3_internal_real): New.
6691 (mulsi3): Add memory operand.
6692 (mulsi3_rl78): Likewise.
6693 (mulsi3_g13): Likewise.
6694 * config/rl78/rl78-real.md (cbranchqi4_real_signed): New.
6695 (cbranchqi4_real): Add more constraint options.
6696 (cbranchhi4_real): Expand pattern.
6697 (cbranchhi4_real_signed): New.
6698 (cbranchhi4_real_inverted): New.
6699 (cbranchsi4_real_lt): New.
6700 (cbranchsi4_real_ge): New.
6701 (cbranchsi4_real_signed): New.
6702 (cbranchsi4_real): New.
6704 * config/rl78/rl78-virt.md (ashrsi3_virt): Add custom cases for
6706 (lshrsi3_virt): Likewise.
6707 (ashlsi3_virt): Likewise.
6708 (cbranchqi4_virt_signed): New.
6709 (cbranchhi4_virt_signed): New.
6710 (cbranchsi4_virt): New.
6711 * config/rl78/rl78.c: Whitespace fixes throughout.
6712 (move_elim_pass): New.
6713 (pass_data_rl78_move_elim): New.
6714 (pass_rl78_move_elim): New.
6715 (make_pass_rl78_move_elim): New.
6716 (rl78_devirt_info): Run devirt earlier.
6717 (rl78_move_elim_info): New.
6718 (rl78_asm_file_start): Register it.
6719 (rl78_split_movsi): New.
6720 (rl78_as_legitimate_address): Allow virtual base registers when
6722 (rl78_addr_space_convert): Remove spurious debug stuff.
6723 (rl78_print_operand_1): Add z,s,S,r,E modifiers.
6724 (rl78_print_operand): More cases for not printing '#'.
6725 (rl78_expand_compare): Remove most of the logic.
6726 (content_memory): New.
6727 (clear_content_memory): New.
6728 (get_content_index): New.
6729 (get_content_name): New.
6730 (display_content_memory): New.
6731 (update_content): New.
6732 (record_content): New.
6733 (already_contains): New.
6734 (insn_ok_now): Re-recog insns with virtual registers.
6735 (add_postponed_content_update): New.
6736 (process_postponed_content_update): New.
6737 (gen_and_emit_move): New.
6738 (transcode_memory_rtx): Record new location content.
6739 Use gen_and_emit_move.
6740 (force_into_acc): New.
6741 (move_to_acc): Use gen_and_emit_move.
6742 (move_from_acc): Likewise.
6743 (move_acc_to_reg): Likewise.
6744 (move_to_x): Likewise.
6745 (move_to_hl): Likewise.
6746 (move_to_de): Likewise.
6747 (rl78_alloc_physical_registers_op1): Record location content.
6748 (has_constraint): New.
6749 (rl78_alloc_physical_registers_op2): Record location content.
6751 (rl78_alloc_physical_registers_ro1): Likewise.
6752 (rl78_alloc_physical_registers_cmp): Likewise.
6753 (rl78_alloc_physical_registers_umul): Likewise.
6754 (rl78_alloc_address_registers_macax): New.
6755 (rl78_alloc_physical_registers): Initialize and set location
6756 content memory as needed.
6757 (rl78_reorg): Make sure split2 is called.
6758 (rl78_rtx_costs): New.
6760 2013-09-11 Richard Sandiford <rdsandiford@googlemail.com>
6762 * simplify-rtx.c (simplify_unary_operation_1): Use simplify_gen_binary
6763 for (not (neg ...)) and (neg (not ...)) cases.
6765 2013-09-11 Richard Biener <rguenther@suse.de>
6768 * passes.def: Split critical edges before late uninit warning passes.
6769 * tree-cfg.c (pass_split_crit_edges): Implement clone method.
6771 2013-09-11 Jakub Jelinek <jakub@redhat.com>
6773 PR tree-optimization/58385
6774 * fold-const.c (build_range_check): If both low and high are NULL,
6775 use omit_one_operand_loc to preserve exp side-effects.
6777 2013-09-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6779 * config/arm/arm.md (arm_shiftsi3): New alternative l/l/M.
6781 2013-09-11 Richard Biener <rguenther@suse.de>
6783 * tree-data-ref.c (dump_rdg_vertex, debug_rdg_vertex,
6784 dump_rdg_component, debug_rdg_component, dump_rdg, debug_rdg,
6785 dot_rdg_1, dot_rdg, rdg_vertex_for_stmt, create_rdg_edge_for_ddr,
6786 create_rdg_edges_for_scalar, create_rdg_edges, create_rdg_vertices,
6787 stmts_from_loop, known_dependences_p, build_empty_rdg,
6788 build_rdg, free_rdg, rdg_defs_used_in_other_loops_p): Move ...
6789 * tree-loop-distribution.c: ... here.
6790 * tree-data-ref.h (struct rdg_vertex, RDGV_STMT, RDGV_DATAREFS,
6791 RDGV_HAS_MEM_WRITE, RDGV_HAS_MEM_READS, RDG_STMT, RDG_DATAREFS,
6792 RDG_MEM_WRITE_STMT, RDG_MEM_READS_STMT, enum rdg_dep_type,
6793 struct rdg_edge, RDGE_TYPE, RDGE_LEVEL, RDGE_RELATION): Move ...
6794 * tree-loop-distribution.c: ... here.
6795 * tree-loop-distribution.c: Include gimple-pretty-print.h.
6796 (struct partition_s): Add loops member.
6797 (partition_alloc, partition_free, rdg_flag_uses, rdg_flag_vertex,
6798 rdg_flag_vertex_and_dependent, rdg_flag_loop_exits,
6799 build_rdg_partition_for_component, rdg_build_partitions): Adjust.
6801 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
6802 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
6803 Sergey Lega <sergey.s.lega@intel.com>
6804 Anna Tikhonova <anna.tikhonova@intel.com>
6805 Ilya Tocar <ilya.tocar@intel.com>
6806 Andrey Turetskiy <andrey.turetskiy@intel.com>
6807 Ilya Verbin <ilya.verbin@intel.com>
6808 Kirill Yukhin <kirill.yukhin@intel.com>
6809 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
6811 * config/i386/constraints.md (k): New.
6813 * config/i386/i386.c (const regclass_map): Add new mask registers.
6814 (dbx_register_map): Ditto.
6815 (dbx64_register_map): Ditto.
6816 (svr4_dbx_register_map): Ditto.
6817 (ix86_conditional_register_usage): Squash mask registers if AVX512F is
6819 (ix86_preferred_reload_class): Disable constants for mask registers.
6820 (ix86_secondary_reload): Do spill of mask register using 32-bit insn.
6821 (ix86_hard_regno_mode_ok): Support new mask registers.
6822 (x86_order_regs_for_local_alloc): Ditto.
6823 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Update.
6824 (FIXED_REGISTERS): Add new mask registers.
6825 (CALL_USED_REGISTERS): Ditto.
6826 (REG_ALLOC_ORDER): Ditto.
6827 (VALID_MASK_REG_MODE): New.
6828 (FIRST_MASK_REG): Ditto.
6829 (LAST_MASK_REG): Ditto.
6830 (reg_class): Add MASK_EVEX_REGS, MASK_REGS.
6831 (MAYBE_MASK_CLASS_P): New.
6832 (REG_CLASS_NAMES): Add MASK_EVEX_REGS, MASK_REGS.
6833 (REG_CLASS_CONTENTS): Ditto.
6834 (MASK_REGNO_P): New.
6835 (ANY_MASK_REG_P): Ditto.
6836 (HI_REGISTER_NAMES): Add new mask registers.
6837 * config/i386/i386.md (MASK0_REG, MASK1_REG, MASK2_REG, MASK3_REG,
6838 MASK4_REG, MASK5_REG, MASK6_REG, MASK7_REG): Constants for new
6840 (attribute "type"): Add mskmov, msklog.
6841 (attribute "length_immediate"): Support them.
6842 (attribute "memory"): Ditto.
6843 (attribute "prefix_0f"): Ditto.
6844 (*movhi_internal): Support new mask registers.
6845 (*movqi_internal): Ditto.
6846 (define_split): Split out clobber pattern is a logic
6847 insn on mask registers.
6848 (*k<logic><mode>): New.
6849 (*andhi_1): Extend to support mask regs.
6850 (*andqi_1): Extend to support mask regs.
6852 (define_split): Split and-not to and and not if operands
6854 (*<code><mode>_1): Separate HI mode to new pattern...
6855 (*<code>hi_1): This.
6856 (*<code>qi_1): Extend to support mask regs.
6858 (kortestzhi): Ditto.
6859 (kortestchi): Ditto.
6861 (*one_cmpl<mode>2_1): Remove HImode and handle it...
6862 (*one_cmplhi2_1): ...Here, now with mask registers support.
6863 (*one_cmplqi2_1): Support new mask registers.
6864 (HI/QImode arithmetics splitter): Don't split if mask registers
6866 (HI/QImode not splitter): Ditto.
6867 * config/i386/predicated.md (mask_reg_operand): New.
6868 (general_reg_operand): Ditto.
6870 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
6872 * doc/invoke.texi: Document fxsr, xsave and xsaveopt options.
6873 * doc/extend.texi: Document fxsr, xsave and xsaveopt builtins.
6875 2013-09-10 Jeff Law <law@redhat.com>
6877 PR tree-optimization/58380
6878 * tree-ssa-threadupdate.c (thread_block): Recognize another case
6879 of threading through a buried loop header.
6881 * tree-ssa-threadedge.c (thread_around_empty_blocks): Correct
6882 return value for single successor case.
6884 2013-09-10 Jan Hubicka <jh@suse.cz>
6886 * ipa-devirt.c (ipa_devirt): Enable with LTO.
6888 2013-09-10 Richard Earnshaw <rearnsha@arm.com>
6891 * arm/vfp.md (combine_vcvt_f32_<FCVTI32typename>): Fix pattern to
6892 support conditional execution.
6893 (combine_vcvt_f64_<FCVTI32typename>): Likewise.
6895 2013-09-10 Vladimir Makarov <vmakarov@redhat.com>
6897 * lra.c (lra): Clear lra_optional_reload_pseudos before every
6899 * lra-constraints.c (curr_insn_transform): Switch on optional reloads.
6900 Check destination too to check move insn.
6901 (undo_optional_reloads): Add check that the original peudo did not
6902 changed its allocation and the optional reload was inherited on last
6903 inheritance pass. Break loop after deciding to keep optional reload.
6904 (lra_undo_inheritance): Add check that inherited pseudo still in
6907 2013-09-10 James Greenhalgh <james.greenhalgh@arm.com>
6909 * config/aarch64/aarch64.md (generic_sched): New.
6910 * config/aarch64/aarch64-generic.md (load): Make conditional
6911 on generic_sched attribute.
6912 (nonload): Likewise.
6914 2013-09-10 Jan Hubicka <jh@suse.cz>
6916 * lto-cgraph.c: Include ipa-utils.h.
6917 (compute_ltrans_boundary): Also add possible targets into the boundary.
6919 2013-09-10 Jan Hubicka <jh@suse.cz>
6921 * gimple-fold.c (gimple_get_virt_method_for_binfo): Pass real
6922 VAR_DECL of vtable rather than full expression.
6924 2013-09-10 Jan Hubicka <jh@suse.cz>
6925 Paolo Carlini <paolo.carlini@oracle.com>
6927 * cgraphunit.c (analyze_functions): Save input_location, set it
6928 to UNKNOWN_LOCATION and restore it at the end.
6930 2013-09-10 Martin Jambor <mjambor@suse.cz>
6932 * ipa-cp.c (propagate_constants_topo): Do not ignore SCC
6933 represented by a thunk.
6935 2013-09-10 Jeff Law <law@redhat.com>
6937 PR tree-optimization/58343
6938 * tree-ssa-threadupdate.c (thread_block): Identify and disable
6939 jump threading requests through loop headers buried in the middle
6940 of a jump threading path.
6942 * tree-ssa-threadedge.c (thread_around_empty_blocks): Fix thinko
6943 in return value/type.
6945 2013-09-10 Jakub Jelinek <jakub@redhat.com>
6947 PR rtl-optimization/58365
6948 * cfgcleanup.c (merge_memattrs): Also clear MEM_READONLY_P
6949 resp. MEM_NOTRAP_P if they differ, or set MEM_VOLATILE_P if
6952 2013-09-10 Richard Biener <rguenther@suse.de>
6954 * tree-data-ref.h (build_rdg): Drop all parameters but loop.
6955 * tree-data-ref.c (create_rdg_vertices): Collect all data
6956 references, signal failure to the caller, use data-ref API.
6957 (build_rdg): Compute data references only once. Maintain lifetime
6958 of data references and data dependences from within RDG.
6959 (free_rdg): Free dependence relations.
6960 * tree-loop-distribution.c (rdg_flag_uses): Drop weird code
6961 inventing extra dependences.
6962 (distribute_loop): Update for RDG API changes.
6964 2013-09-10 Kai Tietz <ktietz@redhat.com>
6966 * doc/invoke.texi (fms-extensions): Document changed
6967 behavior for ms-abi targets.
6968 * config/i386/i386.c (ix86_option_override_internal):
6969 Set default value of option -fms-extension for ms-abi targets.
6971 2013-09-10 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
6973 * config/i386/i386.c (ix86_expand_movmem): Fix epilogue generation.
6975 2013-09-10 Alan Modra <amodra@gmail.com>
6978 * config/rs6000/rs6000.md (bswapdi2_64bit): Disable for volatile mems.
6980 2013-09-10 Alan Modra <amodra@gmail.com>
6982 * config/rs6000/predicates.md (add_cint_operand): New.
6983 (reg_or_add_cint_operand, small_toc_ref): Use add_cint_operand.
6984 * config/rs6000/rs6000.md (largetoc_high_plus): Restrict offset
6985 using add_cint_operand.
6986 (largetoc_high_plus_aix): Likewise.
6988 2013-09-09 Jakub Jelinek <jakub@redhat.com>
6990 PR tree-optimization/58364
6991 * tree-ssa-reassoc.c (init_range_entry): For BIT_NOT_EXPR on
6992 BOOLEAN_TYPE, only invert in_p and continue with arg0 if
6993 the current range can't be an unconditional true or false.
6995 2013-09-09 James Greenhalgh <james.greenhalgh@arm.com>
6997 * config/aarch64/arm_neon.h (vrsqrte_f64): Fix parameter type.
6999 2013-09-09 Uros Bizjak <ubizjak@gmail.com>
7001 * ipa-prop.c (ipa_modify_call_arguments): Initialize deref_align.
7003 2013-09-09 Paolo Carlini <paolo.carlini@oracle.com>
7006 * doc/invoke.texi (-Wdelete-incomplete): Document it.
7008 2013-09-09 Ian Bolton <ian.bolton@arm.com>
7010 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Return
7011 NO_REGS for immediate that can't be moved directly into FP_REGS.
7013 2013-09-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7015 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return CC_SWP for
7016 comparison with negated operand.
7017 * config/aarch64/aarch64.md (compare_neg<mode>): Match canonical
7020 2013-09-09 Richard Biener <rguenther@suse.de>
7023 * cfgloopmanip.c (fix_bb_placements): When fixing the placement
7024 of a subloop record all its block as affecting loop-closed SSA form.
7026 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
7028 * expmed.c (lshift_value): Take an unsigned HOST_WIDE_INT instead
7029 of an rtx/bitpos pair.
7030 (store_fixed_bit_field): Update accordingly.
7032 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
7034 * asan.c (asan_emit_stack_protection): Use gen_int_mode instead of
7036 * builtins.c (expand_errno_check): Likewise.
7037 * dwarf2cfi.c (init_return_column_size): Likewise.
7038 * except.c (sjlj_mark_call_sites): Likewise.
7039 * expr.c (move_by_pieces_1, store_by_pieces_2): Likewise.
7040 * lra-constraints.c (emit_inc): Likewise.
7041 * ree.c (combine_set_extension): Likewise.
7042 * regmove.c (fixup_match_2): Likewise.
7043 * reload1.c (inc_for_reload): Likewise.
7045 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
7047 * combine.c (simplify_set, expand_field_assignment, extract_left_shift)
7048 (force_to_mode, simplify_shift_const_1, simplify_comparison):
7049 Use gen_int_mode with the mode of the associated simplify_* call.
7050 * explow.c (probe_stack_range, anti_adjust_stack_and_probe): Likewise.
7051 * expmed.c (expand_shift_1): Likewise.
7052 * function.c (instantiate_virtual_regs_in_insn): Likewise.
7053 * loop-iv.c (iv_number_of_iterations): Likewise.
7054 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
7055 * simplify-rtx.c (simplify_binary_operation_1): Likewise.
7057 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
7059 * asan.c (asan_clear_shadow): Use gen_int_mode with the mode
7060 of the associated expand_* call.
7061 (asan_emit_stack_protection): Likewise.
7062 * builtins.c (round_trampoline_addr): Likewise.
7063 * explow.c (allocate_dynamic_stack_space, probe_stack_range): Likewise.
7064 * expmed.c (expand_smod_pow2, expand_sdiv_pow2, expand_divmod)
7065 (emit_store_flag): Likewise.
7066 * expr.c (emit_move_resolve_push, push_block, emit_single_push_insn_1)
7067 (emit_push_insn, optimize_bitfield_assignment_op, expand_expr_real_1):
7069 * function.c (instantiate_virtual_regs_in_insn): Likewise.
7070 * ifcvt.c (noce_try_store_flag_constants): Likewise.
7071 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
7072 * modulo-sched.c (generate_prolog_epilog): Likewise.
7073 * optabs.c (expand_binop, widen_leading, expand_doubleword_clz)
7074 (expand_ctz, expand_ffs, expand_unop): Likewise.
7076 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
7078 * alias.c (addr_side_effect_eval): Use gen_int_mode with the mode
7079 of the associated gen_rtx_* call.
7080 * caller-save.c (init_caller_save): Likewise.
7081 * combine.c (find_split_point, make_extraction): Likewise.
7082 (make_compound_operation): Likewise.
7083 * dwarf2out.c (mem_loc_descriptor): Likewise.
7084 * explow.c (plus_constant, probe_stack_range): Likewise.
7085 * expmed.c (expand_mult_const): Likewise.
7086 * expr.c (emit_single_push_insn_1, do_tablejump): Likewise.
7087 * reload1.c (init_reload): Likewise.
7088 * valtrack.c (cleanup_auto_inc_dec): Likewise.
7089 * var-tracking.c (adjust_mems): Likewise.
7090 * modulo-sched.c (sms_schedule): Likewise, but use gen_rtx_GT
7091 rather than gen_rtx_fmt_ee.
7093 2013-09-09 Jan Hubicka <jh@suse.cz>
7096 * value-prof.c (gimple_ic): Copy also abnormal edges.
7098 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
7100 * asan.c (asan_shadow_cst): Use gen_int_mode.
7102 2013-09-08 Jan Hubicka <jh@suse.cz>
7104 * ipa-profile.c: Add toplevel comment.
7105 (ipa_propagate_frequency_1): Be more conservative when profile is read.
7106 (contains_hot_call_p): New function.
7107 (ipa_propagate_frequency): Set frequencies based on counts when
7109 * predict.c (compute_function_frequency): Use PROFILE_READ gueard for
7110 profile; do not tamper with profile after inlining if it is read.
7112 2013-09-08 Jan Hubicka <jh@suse.cz>
7114 * ipa-prop.c (try_make_edge_direct_simple_call): Do not special case
7117 2013-09-08 Jan Hubicka <jh@suse.cz>
7119 * ipa.c (walk_polymorphic_call_targets): Fix redirection before IPA
7122 2013-09-08 Jeff Law <law@redhat.com>
7125 * tree-ssa-threadedge.c (thread_across_edge): Fix initialization
7128 2013-09-08 Andi Kleen <ak@linux.intel.com>
7130 * tree-inline.c (estimate_num_insns): Limit asm cost to 1000.
7132 2013-09-08 Jan Hubicka <jh@suse.cz>
7134 * ipa.c (walk_polymorphic_call_targets): Fix inliner summary update.
7136 2013-09-08 Richard Sandiford <rdsandiford@googlemail.com>
7138 * ira.c (update_equiv_regs): Only call set_paradoxical_subreg
7139 for non-debug insns.
7140 * lra.c (new_insn_reg): Take the containing insn as a parameter.
7141 Only modify lra_reg_info[].biggest_mode if it's non-debug insn.
7142 (collect_non_operand_hard_regs, add_regs_to_insn_regno_info): Update
7145 2013-09-08 Jan Hubicka <jh@suse.cz>
7147 * cgraphunit.c (walk_polymorphic_call_targets): Permit 0 possible
7148 targets and devirtualize to BUILT_IN_UNREACHABLE.
7149 * timevar.def (TV_IPA_UNREACHABLE): New timevar.
7150 * ipa.c (walk_polymorphic_call_targets): New function.
7151 (symtab_remove_unreachable_nodes): Use it; do not keep all virtual
7152 functions; use the new timevar.
7153 * ipa-devirt.c (maybe_record_node): Do not insert static nodes that
7154 was removed from the program.
7155 (record_binfo): If BINFO corresponds to an anonymous namespace, we may
7156 not consider it in the walk when its vtable is dead.
7157 (possible_polymorphic_call_targets_1): Pass anonymous flag to
7159 (devirt_variable_node_removal_hook): New function.
7160 (possible_polymorphic_call_targets): Also register
7161 devirt_variable_node_removal_hook.
7162 (ipa_devirt): Do not do non-speculative devirtualization.
7163 (gate_ipa_devirt): One execute if devirtualizing speculatively.
7165 2013-09-08 Jan Hubicka <jh@suse.cz>
7167 * cgraph.h (varpool_node_hook, varpool_node_hook_list,
7168 varpool_add_node_removal_hook, varpool_add_variable_insertion_hook,
7169 varpool_remove_variable_insertion_hook): Declare.
7170 * varpool.c (varpool_node_hook_list): New structure.
7171 (first_varpool_node_removal_hook,
7172 first_varpool_variable_insertion_hook): New variables.
7173 (varpool_add_node_removal_hook, varpool_remove_node_removal_hook,
7174 varpool_call_node_removal_hooks, varpool_add_variable_insertion_hook,
7175 varpool_remove_variable_insertion_hook,
7176 varpool_call_variable_insertion_hooks): New functions.
7177 (varpool_remove_node): Use it.
7179 2013-09-08 Paolo Carlini <paolo.carlini@oracle.com>
7182 * diagnostic.c (diagnostic_build_prefix): When s.file is
7183 "<built-in>" don't output line and column numbers.
7185 2013-09-06 Jan Hubicka <jh@suse.cz>
7187 * cgraphunit.c (expand_thunk): Get body before touching arguments.
7188 * lto-streamer-out.c: Stream thunks, too.
7189 * lto-streamer-in.c (input_function): Pop cfun here
7190 (lto_read_body): Instead of here.
7192 2013-09-06 Caroline Tice <cmtice@google.com>
7194 * doc/install.texi: Add documentation for the --enable-vtable-verify
7195 and the --disable-libvtv configure options.
7197 2013-09-06 Jeff Law <law@redhat.com>
7199 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
7200 edge implied equivalences into successor phis.
7202 2013-09-06 Joern Rennecke <joern.rennecke@embecosm.com>
7204 * resource.c (mark_referenced_resources): Handle COND_EXEC.
7206 2013-09-06 Claudiu Zissulescu <claziss@synopsys.com>
7208 * resource.c (mark_target_live_regs): Compute resources taking
7209 into account if a call is predicated or not.
7211 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
7213 * toplev.c (output_stack_usage): Be prepared for suffixes created by
7214 the compiler in the function names.
7216 2013-09-06 Jan Hubicka <jh@suse.cz>
7219 * ipa-inline.c (has_caller_p): New function.
7220 (want_inline_function_to_all_callers_p): Use it.
7221 (sum_callers, inline_to_all_callers): Break out from ...
7222 (ipa_inline): ... here.
7224 2013-09-06 Jan Hubicka <jh@suse.cz>
7226 * config/i386/i386.c (ix86_hard_regno_mode_ok): AVX modes are valid
7227 only when AVX is enabled.
7229 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
7231 * config/aarch64/aarch64.md
7232 (*movtf_aarch64): Use neon_<ls>dm_2 as type where v8type
7233 is fpsimd_<load/store>2.
7234 (load_pair<mode>): Likewise.
7235 (store_pair<mode>): Likewise.
7237 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
7239 * config/arm/types.md (type): Add "mrs" type.
7240 * config/aarch64/aarch64.md
7241 (aarch64_load_tp_hard): Make type "mrs".
7243 (load_tp_hard): Make type "mrs".
7244 * config/arm/cortex-a15.md: Update with new attributes.
7245 * config/arm/cortex-a5.md: Update with new attributes.
7246 * config/arm/cortex-a53.md: Update with new attributes.
7247 * config/arm/cortex-a7.md: Update with new attributes.
7248 * config/arm/cortex-a8.md: Update with new attributes.
7249 * config/arm/cortex-a9.md: Update with new attributes.
7250 * config/arm/cortex-m4.md: Update with new attributes.
7251 * config/arm/cortex-r4.md: Update with new attributes.
7252 * config/arm/fa526.md: Update with new attributes.
7253 * config/arm/fa606te.md: Update with new attributes.
7254 * config/arm/fa626te.md: Update with new attributes.
7255 * config/arm/fa726te.md: Update with new attributes.
7257 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
7259 * config/aarch64/aarch64.md
7260 (*movti_aarch64): Use "multiple" for type where v8type is "move2".
7261 (*movtf_aarch64): Likewise.
7263 (thumb1_movdi_insn): Use "multiple" for type where more than one
7264 instruction is used for a move.
7265 (*arm32_movhf): Likewise.
7266 (*thumb_movdf_insn): Likewise.
7268 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
7270 * config/arm/types.md (type): Rename fcpys to fmov.
7272 (*arm_movsi_vfp): Rename type fcpys as fmov.
7273 (*thumb2_movsi_vfp): Likewise
7274 (*movhf_vfp_neon): Likewise
7275 (*movhf_vfp): Likewise
7276 (*movsf_vfp): Likewise
7277 (*thumb2_movsf_vfp): Likewise
7278 (*movsfcc_vfp): Likewise
7279 (*thumb2_movsfcc_vfp): Likewise
7280 * config/aarch64/aarch64-simd.md
7281 (move_lo_quad_<mode>): Replace type mov_reg with fmovs.
7282 * config/aarch64/aarch64.md
7283 (*movsi_aarch64): Replace type mov_reg with fmovs.
7284 (*movdi_aarch64): Likewise
7285 (*movsf_aarch64): Likewise
7286 (*movdf_aarch64): Likewise
7288 (cortexa7_older_only): Rename TYPE_FCPYS to TYPE_FMOV.
7289 * config/arm/iwmmxt.md
7290 (*iwmmxt_movsi_insn): Rename type fcpys as fmov.
7291 * config/arm/arm1020e.md: Update with new attributes.
7292 * config/arm/cortex-a15-neon.md: Update with new attributes.
7293 * config/arm/cortex-a5.md: Update with new attributes.
7294 * config/arm/cortex-a53.md: Update with new attributes.
7295 * config/arm/cortex-a7.md: Update with new attributes.
7296 * config/arm/cortex-a8-neon.md: Update with new attributes.
7297 * config/arm/cortex-a9.md: Update with new attributes.
7298 * config/arm/cortex-m4-fpu.md: Update with new attributes.
7299 * config/arm/cortex-r4f.md: Update with new attributes.
7300 * config/arm/marvell-pj4.md: Update with new attributes.
7301 * config/arm/vfp11.md: Update with new attributes.
7303 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
7305 * config/aarch64/aarch64.md
7306 (*madd<mode>): Fix type attribute.
7307 (*maddsi_uxtw): Likewise.
7308 (*msub<mode>): Likewise.
7309 (*msubsi_uxtw): Likewise.
7310 (<su_optab>maddsidi4): Likewise.
7311 (<su_optab>msubsidi4): Likewise.
7313 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
7315 * config/arm/types.md: Split fdiv<sd> as fsqrt<sd>, fdiv<sd>.
7316 * config/arm/arm.md (core_cycles): Remove fdiv.
7317 * config/arm/vfp.md:
7318 (*sqrtsf2_vfp): Update for attribute changes.
7319 (*sqrtdf2_vfp): Likewise.
7320 * config/aarch64/aarch64.md:
7321 (sqrt<mode>2): Update for attribute changes.
7322 * config/arm/arm1020e.md: Update with new attributes.
7323 * config/arm/cortex-a15-neon.md: Update with new attributes.
7324 * config/arm/cortex-a5.md: Update with new attributes.
7325 * config/arm/cortex-a53.md: Update with new attributes.
7326 * config/arm/cortex-a7.md: Update with new attributes.
7327 * config/arm/cortex-a8-neon.md: Update with new attributes.
7328 * config/arm/cortex-a9.md: Update with new attributes.
7329 * config/arm/cortex-m4-fpu.md: Update with new attributes.
7330 * config/arm/cortex-r4f.md: Update with new attributes.
7331 * config/arm/marvell-pj4.md: Update with new attributes.
7332 * config/arm/vfp11.md: Update with new attributes.
7334 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
7336 * config/arm/types.md
7337 (type): Split f_cvt as f_cvt, f_cvtf2i, f_cvti2f.
7338 * config/aarch64/aarch64.md
7339 (l<fcvt_pattern><su_optab><GPF:mode><GPI:mode>2): Update with
7341 (fix_trunc<GPF:mode><GPI:mode>2): Likewise.
7342 (fixuns_trunc<GPF:mode><GPI:mode>2): Likewise.
7343 (float<GPI:mode><GPF:mode>2): Likewise.
7345 (*truncsisf2_vfp): Update with new attributes.
7346 (*truncsidf2_vfp): Likewise.
7347 (fixuns_truncsfsi2): Likewise.
7348 (fixuns_truncdfsi2): Likewise.
7349 (*floatsisf2_vfp): Likewise.
7350 (*floatsidf2_vfp): Likewise.
7351 (floatunssisf2): Likewise.
7352 (floatunssidf2): Likewise.
7353 (*combine_vcvt_f32_<FCVTI32typename>): Likewise.
7354 (*combine_vcvt_f64_<FCVTI32typename>): Likewise.
7355 * config/arm/arm1020e.md: Update with new attributes.
7356 * config/arm/cortex-a15-neon.md: Update with new attributes.
7357 * config/arm/cortex-a5.md: Update with new attributes.
7358 * config/arm/cortex-a53.md: Update with new attributes.
7359 * config/arm/cortex-a7.md: Update with new attributes.
7360 * config/arm/cortex-a8-neon.md: Update with new attributes.
7361 * config/arm/cortex-a9.md: Update with new attributes.
7362 * config/arm/cortex-m4-fpu.md: Update with new attributes.
7363 * config/arm/cortex-r4f.md: Update with new attributes.
7364 * config/arm/marvell-pj4.md: Update with new attributes.
7365 * config/arm/vfp11.md: Update with new attributes.
7367 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
7369 * config/aarch64/arm_neon.h
7370 (vqtbl<1,2,3,4><q>_s8): Fix control vector parameter type.
7371 (vqtbx<1,2,3,4><q>_s8): Likewise.
7373 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
7375 * config/arm/types.md: Add "no_insn", "multiple" and "untyped" types.
7376 * config/arm/arm-fixed.md: Add type attribute to all insn patterns.
7377 (add<mode>3): Add type attribute.
7378 (add<mode>3): Likewise.
7379 (usadd<mode>3): Likewise.
7380 (ssadd<mode>3): Likewise.
7381 (sub<mode>3): Likewise.
7382 (sub<mode>3): Likewise.
7383 (ussub<mode>3): Likewise.
7384 (sssub<mode>3): Likewise.
7385 (ssmulsa3): Likewise.
7386 (usmulusa3): Likewise.
7387 (arm_usatsihi): Likewise.
7389 (*movdi_vfp): Add types for all instructions.
7390 (*movdi_vfp_cortexa8): Likewise.
7391 (*movhf_vfp_neon): Likewise.
7392 (*movhf_vfp): Likewise.
7393 (*movdf_vfp): Likewise.
7394 (*thumb2_movdf_vfp): Likewise.
7395 (*thumb2_movdfcc_vfp): Likewise.
7396 * config/arm/arm.md: Add type attribute to all insn patterns.
7397 (*thumb1_adddi3): Add type attribute.
7398 (*arm_adddi3): Likewise.
7399 (*adddi_sesidi_di): Likewise.
7400 (*adddi_zesidi_di): Likewise.
7401 (*thumb1_addsi3): Likewise.
7402 (addsi3_compare0): Likewise.
7403 (*addsi3_compare0_scratch): Likewise.
7404 (*compare_negsi_si): Likewise.
7405 (cmpsi2_addneg): Likewise.
7406 (*addsi3_carryin_<optab>): Likewise.
7407 (*addsi3_carryin_alt2_<optab>): Likewise.
7408 (*addsi3_carryin_clobercc_<optab>): Likewise.
7409 (*subsi3_carryin): Likewise.
7410 (*subsi3_carryin_const): Likewise.
7411 (*subsi3_carryin_compare): Likewise.
7412 (*subsi3_carryin_compare_const): Likewise.
7413 (*arm_subdi3): Likewise.
7414 (*thumb_subdi3): Likewise.
7415 (*subdi_di_zesidi): Likewise.
7416 (*subdi_di_sesidi): Likewise.
7417 (*subdi_zesidi_di): Likewise.
7418 (*subdi_sesidi_di): Likewise.
7419 (*subdi_zesidi_ze): Likewise.
7420 (thumb1_subsi3_insn): Likewise.
7421 (*arm_subsi3_insn): Likewise.
7422 (*anddi3_insn): Likewise.
7423 (*anddi_zesidi_di): Likewise.
7424 (*anddi_sesdi_di): Likewise.
7425 (*ne_zeroextracts): Likewise.
7426 (*ne_zeroextracts): Likewise.
7427 (*ite_ne_zeroextr): Likewise.
7428 (*ite_ne_zeroextr): Likewise.
7429 (*anddi_notdi_di): Likewise.
7430 (*anddi_notzesidi): Likewise.
7431 (*anddi_notsesidi): Likewise.
7432 (andsi_notsi_si): Likewise.
7433 (thumb1_bicsi3): Likewise.
7434 (*iordi3_insn): Likewise.
7435 (*iordi_zesidi_di): Likewise.
7436 (*iordi_sesidi_di): Likewise.
7437 (*thumb1_iorsi3_insn): Likewise.
7438 (*xordi3_insn): Likewise.
7439 (*xordi_zesidi_di): Likewise.
7440 (*xordi_sesidi_di): Likewise.
7441 (*arm_xorsi3): Likewise.
7442 (*andsi_iorsi3_no): Likewise.
7443 (*smax_0): Likewise.
7444 (*smax_m1): Likewise.
7445 (*arm_smax_insn): Likewise.
7446 (*smin_0): Likewise.
7447 (*arm_smin_insn): Likewise.
7448 (*arm_umaxsi3): Likewise.
7449 (*arm_uminsi3): Likewise.
7450 (*minmax_arithsi): Likewise.
7451 (*minmax_arithsi_): Likewise.
7452 (*satsi_<SAT:code>): Likewise.
7453 (arm_ashldi3_1bit): Likewise.
7454 (arm_ashrdi3_1bit): Likewise.
7455 (arm_lshrdi3_1bit): Likewise.
7456 (*arm_negdi2): Likewise.
7457 (*thumb1_negdi2): Likewise.
7458 (*arm_negsi2): Likewise.
7459 (*thumb1_negsi2): Likewise.
7460 (*negdi_extendsid): Likewise.
7461 (*negdi_zero_extend): Likewise.
7462 (*arm_abssi2): Likewise.
7463 (*thumb1_abssi2): Likewise.
7464 (*arm_neg_abssi2): Likewise.
7465 (*thumb1_neg_abss): Likewise.
7466 (one_cmpldi2): Likewise.
7467 (extend<mode>di2): Likewise.
7468 (*compareqi_eq0): Likewise.
7469 (*arm_extendhisi2addsi): Likewise.
7470 (*arm_movdi): Likewise.
7471 (*thumb1_movdi_insn): Likewise.
7472 (*arm_movt): Likewise.
7473 (*thumb1_movsi_insn): Likewise.
7474 (pic_add_dot_plus_four): Likewise.
7475 (pic_add_dot_plus_eight): Likewise.
7476 (tls_load_dot_plus_eight): Likewise.
7477 (*thumb1_movhi_insn): Likewise.
7478 (*thumb1_movsf_insn): Likewise.
7479 (*movdf_soft_insn): Likewise.
7480 (*thumb_movdf_insn): Likewise.
7481 (cbranchsi4_insn): Likewise.
7482 (cbranchsi4_scratch): Likewise.
7483 (*negated_cbranchsi4): Likewise.
7484 (*tbit_cbranch): Likewise.
7485 (*tlobits_cbranch): Likewise.
7486 (*tstsi3_cbranch): Likewise.
7487 (*cbranchne_decr1): Likewise.
7488 (*addsi3_cbranch): Likewise.
7489 (*addsi3_cbranch_scratch): Likewise.
7490 (*arm_cmpdi_insn): Likewise.
7491 (*arm_cmpdi_unsig): Likewise.
7492 (*arm_cmpdi_zero): Likewise.
7493 (*thumb_cmpdi_zero): Likewise.
7494 (*deleted_compare): Likewise.
7495 (*mov_scc): Likewise.
7496 (*mov_negscc): Likewise.
7497 (*mov_notscc): Likewise.
7498 (*cstoresi_eq0_thumb1_insn): Likewise.
7499 (cstoresi_nltu_thumb1): Likewise.
7500 (cstoresi_ltu_thu): Likewise.
7501 (thumb1_addsi3_addgeu): Likewise.
7502 (*arm_jump): Likewise.
7503 (*thumb_jump): Likewise.
7504 (*check_arch2): Likewise.
7505 (arm_casesi_internal): Likewise.
7506 (thumb1_casesi_dispatch): Likewise.
7507 (*arm_indirect_jump): Likewise.
7508 (*thumb1_indirect_jump): Likewise.
7510 (*and_scc): Likewise.
7511 (*ior_scc): Likewise.
7512 (*compare_scc): Likewise.
7513 (*cond_move): Likewise.
7514 (*cond_arith): Likewise.
7515 (*cond_sub): Likewise.
7516 (*cmp_ite0): Likewise.
7517 (*cmp_ite1): Likewise.
7518 (*cmp_and): Likewise.
7519 (*cmp_ior): Likewise.
7520 (*ior_scc_scc): Likewise.
7521 (*ior_scc_scc_cmp): Likewise.
7522 (*and_scc_scc): Likewise.
7523 (*and_scc_scc_cmp): Likewise.
7524 (*and_scc_scc_nod): Likewise.
7525 (*negscc): Likewise.
7526 (movcond_addsi): Likewise.
7527 (movcond): Likewise.
7528 (*ifcompare_plus_move): Likewise.
7529 (*if_plus_move): Likewise.
7530 (*ifcompare_move_plus): Likewise.
7531 (*if_move_plus): Likewise.
7532 (*ifcompare_arith_arith): Likewise.
7533 (*if_arith_arith): Likewise.
7534 (*ifcompare_arith_move): Likewise.
7535 (*if_arith_move): Likewise.
7536 (*ifcompare_move_arith): Likewise.
7537 (*if_move_arith): Likewise.
7538 (*ifcompare_move_not): Likewise.
7539 (*if_move_not): Likewise.
7540 (*ifcompare_not_move): Likewise.
7541 (*if_not_move): Likewise.
7542 (*ifcompare_shift_move): Likewise.
7543 (*if_shift_move): Likewise.
7544 (*ifcompare_move_shift): Likewise.
7545 (*if_move_shift): Likewise.
7546 (*ifcompare_shift_shift): Likewise.
7547 (*ifcompare_not_arith): Likewise.
7548 (*ifcompare_arith_not): Likewise.
7549 (*if_arith_not): Likewise.
7550 (*ifcompare_neg_move): Likewise.
7551 (*if_neg_move): Likewise.
7552 (*ifcompare_move_neg): Likewise.
7553 (*if_move_neg): Likewise.
7554 (prologue_thumb1_interwork): Likewise.
7555 (*cond_move_not): Likewise.
7556 (*sign_extract_onebit): Likewise.
7557 (*not_signextract_onebit): Likewise.
7558 (stack_tie): Likewise.
7559 (align_4): Likewise.
7560 (align_8): Likewise.
7561 (consttable_end): Likewise.
7562 (consttable_1): Likewise.
7563 (consttable_2): Likewise.
7564 (consttable_4): Likewise.
7565 (consttable_8): Likewise.
7566 (consttable_16): Likewise.
7567 (*thumb1_tablejump): Likewise.
7568 (prefetch): Likewise.
7569 (force_register_use): Likewise.
7570 (thumb_eh_return): Likewise.
7571 (load_tp_hard): Likewise.
7572 (load_tp_soft): Likewise.
7573 (tlscall): Likewise.
7574 (*arm_movtas_ze): Likewise.
7575 (*arm_rev): Likewise.
7576 (*arm_revsh): Likewise.
7577 (*arm_rev16): Likewise.
7578 * config/arm/thumb2.md
7579 (*thumb2_smaxsi3): Likewise.
7580 (*thumb2_sminsi3): Likewise.
7581 (*thumb32_umaxsi3): Likewise.
7582 (*thumb2_uminsi3): Likewise.
7583 (*thumb2_negdi2): Likewise.
7584 (*thumb2_abssi2): Likewise.
7585 (*thumb2_neg_abss): Likewise.
7586 (*thumb2_movsi_insn): Likewise.
7587 (tls_load_dot_plus_four): Likewise.
7588 (*thumb2_movhi_insn): Likewise.
7589 (*thumb2_mov_scc): Likewise.
7590 (*thumb2_mov_negs): Likewise.
7591 (*thumb2_mov_negs): Likewise.
7592 (*thumb2_mov_nots): Likewise.
7593 (*thumb2_mov_nots): Likewise.
7594 (*thumb2_movsicc_): Likewise.
7595 (*thumb2_movsfcc_soft_insn): Likewise.
7596 (*thumb2_indirect_jump): Likewise.
7597 (*thumb2_and_scc): Likewise.
7598 (*thumb2_ior_scc): Likewise.
7599 (*thumb2_ior_scc_strict_it): Likewise.
7600 (*thumb2_cond_move): Likewise.
7601 (*thumb2_cond_arith): Likewise.
7602 (*thumb2_cond_ari): Likewise.
7603 (*thumb2_cond_sub): Likewise.
7604 (*thumb2_negscc): Likewise.
7605 (*thumb2_movcond): Likewise.
7606 (thumb2_casesi_internal): Likewise.
7607 (thumb2_casesi_internal_pic): Likewise.
7608 (*thumb2_alusi3_short): Likewise.
7609 (*thumb2_mov<mode>_shortim): Likewise.
7610 (*thumb2_addsi_short): Likewise.
7611 (*thumb2_subsi_short): Likewise.
7612 (thumb2_addsi3_compare0): Likewise.
7613 (*thumb2_cbz): Likewise.
7614 (*thumb2_cbnz): Likewise.
7615 (*thumb2_one_cmplsi2_short): Likewise.
7616 (*thumb2_negsi2_short): Likewise.
7617 (*orsi_notsi_si): Likewise.
7618 * config/arm/arm1020e.md: Update with new attributes.
7619 * config/arm/arm1026ejs.md: Update with new attributes.
7620 * config/arm/arm1136jfs.md: Update with new attributes.
7621 * config/arm/arm926ejs.md: Update with new attributes.
7622 * config/arm/cortex-a15.md: Update with new attributes.
7623 * config/arm/cortex-a5.md: Update with new attributes.
7624 * config/arm/cortex-a53.md: Update with new attributes.
7625 * config/arm/cortex-a7.md: Update with new attributes.
7626 * config/arm/cortex-a8.md: Update with new attributes.
7627 * config/arm/cortex-a9.md: Update with new attributes.
7628 * config/arm/cortex-m4.md: Update with new attributes.
7629 * config/arm/cortex-r4.md: Update with new attributes.
7630 * config/arm/fa526.md: Update with new attributes.
7631 * config/arm/fa606te.md: Update with new attributes.
7632 * config/arm/fa626te.md: Update with new attributes.
7633 * config/arm/fa726te.md: Update with new attributes.
7635 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
7637 * config/aarch64/aarch64-simd.md
7638 (aarch64_sqdml<SBINQOPS:as>l_n<mode>_internal): Use
7639 <vwx> iterator to ensure correct register choice.
7640 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
7641 (aarch64_sqdmull_n<mode>): Likewise.
7642 (aarch64_sqdmull2_n<mode>_internal): Likewise.
7643 * config/aarch64/arm_neon.h
7644 (vml<as><q>_lane<q>_<su>16): Use 'x' constraint for element vector.
7645 (vml<as><q>_n_<su>16): Likewise.
7646 (vml<as>l_high_lane<q>_<su>16): Likewise.
7647 (vml<as>l_high_n_<su>16): Likewise.
7648 (vml<as>l_lane<q>_<su>16): Likewise.
7649 (vml<as>l_n_<su>16): Likewise.
7650 (vmul<q>_lane<q>_<su>16): Likewise.
7651 (vmul<q>_n_<su>16): Likewise.
7652 (vmull_lane<q>_<su>16): Likewise.
7653 (vmull_n_<su>16): Likewise.
7654 (vmull_high_lane<q>_<su>16): Likewise.
7655 (vmull_high_n_<su>16): Likewise.
7656 (vqrdmulh<q>_n_s16): Likewise.
7658 2013-09-06 Tejas Belagod <tejas.belagod@arm.com>
7660 * config/aarch64/arm_neon.h: Fix all vdup<bhsd_lane<q> intrinsics to
7661 have the correct lane parameter.
7663 2013-09-06 Richard Biener <rguenther@suse.de>
7665 * cfganal.c (control_dependences::~control_dependences):
7666 Properly free all of the vector.
7668 2013-09-06 Kirill Yukhin <kirill.yukhin@intel.com>
7671 * config/i386/i386.c (ix86_conditional_register_usage):
7672 Proper initialize extended SSE registers.
7674 2013-09-06 Jan Hubicka <jh@suse.cz>
7676 PR tree-optimization/58311
7677 * ipa-devirt.c (gate_ipa_devirt): Only execute when optimizing.
7679 2013-09-06 Jan Hubicka <jh@suse.cz>
7681 * Makefile.in (tree-sra.o): Update dependencies.
7682 * tree-sra.c: Include ipa-utils.h
7683 (scan_function): Use recursive_call_p.
7684 (has_caller_p): New function.
7685 (cgraph_for_node_and_aliases): Count also callers of aliases.
7687 2013-09-06 Jan Hubicka <jh@suse.cz>
7690 * cgraph.h (symtab_semantically_equivalent_p): Declare.
7691 * tree-tailcall.c: Include ipa-utils.h.
7692 (find_tail_calls): Use it.
7693 * ipa-pure-const.c (check_call): Likewise.
7694 * ipa-utils.c (recursive_call_p): New function.
7695 * ipa-utils.h (recursive_call_p): Dclare.
7696 * symtab.c (symtab_nonoverwritable_alias): Fix formatting.
7697 (symtab_semantically_equivalent_p): New function.
7698 * Makefile.in (tree-tailcall.o): Update dependencies.
7700 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
7702 * ipa-split.c (split_function): Set DECL_NO_INLINE_WARNING_P on the
7705 2013-09-06 Richard Biener <rguenther@suse.de>
7707 * lto-streamer.h (lto_global_var_decls): Remove.
7708 * Makefile.in (OBJS): Remove lto-symtab.o.
7709 (lto-symtab.o): Remove.
7710 (GTFILES): Remove lto-symtab.c
7711 * lto-symtab.c: Move to lto/
7713 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7715 * config/s390/s390.md (UNSPEC_FPINT_FLOOR, UNSPEC_FPINT_BTRUNC)
7716 (UNSPEC_FPINT_ROUND, UNSPEC_FPINT_CEIL, UNSPEC_FPINT_NEARBYINT)
7717 (UNSPEC_FPINT_RINT): New constant definitions.
7718 (FPINT, fpint_name, fpint_roundingmode): New integer iterator
7719 definition with 2 attributes.
7720 ("<FPINT:fpint_name><BFP:mode>2", "rint<BFP:mode>2")
7721 ("<FPINT:fpint_name><DFP:mode>2", "rint<DFP:mode>2"): New pattern
7724 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7726 * config/s390/s390.md: Add "bcr_flush" value to mnemonic attribute.
7727 ("mem_thread_fence_1"): Use bcr 14,0 for z196 and later.
7728 Set the mnemonic attribute to "bcr_flush". Set the "z196prop"
7729 attribute to "z196_alone".
7730 * config/s390/2827.md: Add "bcr_flush" to "ooo_groupalone" and
7733 2013-09-06 Richard Biener <rguenther@suse.de>
7735 * basic-block.h (class control_dependences): New.
7736 * tree-ssa-dce.c (control_dependence_map): Remove.
7738 (EXECUTE_IF_CONTROL_DEPENDENT): Remove.
7739 (set_control_dependence_map_bit, clear_control_dependence_bitmap,
7740 find_pdom, find_control_dependence, find_all_control_dependences):
7742 (mark_control_dependent_edges_necessary,
7743 find_obviously_necessary_stmts, propagate_necessity, tree_dce_init,
7744 tree_dce_done, perform_tree_ssa_dce): Adjust.
7745 * cfganal.c (set_control_dependence_map_bit,
7746 clear_control_dependence_bitmap, find_pdom, find_control_dependence,
7747 find_all_control_dependences): Move from tree-ssa-dce.c and
7748 implement as methods of control_dependences class.
7749 (control_dependences::control_dependences): New.
7750 (control_dependences::~control_dependences): Likewise.
7751 (control_dependences::get_edges_dependent_on): Likewise.
7752 (control_dependences::get_edge): Likewise.
7754 2013-09-04 Jan Hubicka <jh@suse.cz>
7756 * tree.c (types_same_for_odr): Drop overactive check.
7757 * ipa-devirt.c (hash_type_name): Likewise.
7759 2013-09-04 Jan Hubicka <jh@suse.cz>
7761 * cgraphunit.c (walk_polymorphic_call_targets): Break out from ...
7762 (analyze_functions): ... here.
7764 2013-09-04 Jan Hubicka <jh@suse.cz>
7767 * cgraphunit.c (analyze_functions): Clear AUX fields
7768 after processing; initialize assembler name has.
7770 2013-09-05 Jeff Law <law@redhat.com>
7772 * tree-ssa-threadedge.c (thread_around_empty_blocks): Renamed
7773 from thread_around_empty_block. Record threading path into PATH.
7774 Recurse if threading through the initial block is successful.
7775 (thread_across_edge): Corresponding changes to slightly simplify.
7777 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
7779 * config/aarch64/aarch64.md
7780 (type): Remove frecpe, frecps, frecpx.
7781 (aarch64_frecp<FRECP:frecp_suffix><mode>): Move to aarch64-simd.md,
7782 fix to be a TARGET_SIMD instruction.
7783 (aarch64_frecps): Remove.
7784 * config/aarch64/aarch64-simd.md
7785 (aarch64_frecp<FRECP:frecp_suffix><mode>): New, moved from aarch64.md
7786 (aarch64_frecps<mode>): Handle all float/vector of float modes.
7788 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
7789 Sofiane Naci <sofiane.naci@arm.com>
7791 * config/arm/types.md (define_attr "type"): Expand "arlo_imm"
7792 into "adr", "alu_imm", "alus_imm", "logic_imm", "logics_imm".
7793 Expand "arlo_reg" into "adc_reg", "adc_imm", "adcs_reg", "adcs_imm",
7794 "alu_ext", "alu_reg", "alus_ext", "alus_reg", "bfm", "csel",
7795 "logic_reg", "logics_reg", "rev". Expand "arlo_shift" into
7796 "alu_shift_imm", "alus_shift_imm", "logic_shift_imm",
7797 "logics_shift_imm". Expand "arlo_shift_reg" into "alu_shift_reg",
7798 "alus_shift_reg", "logic_shift_reg", "logics_shift_reg". Expand "clz"
7799 into "clz, "rbit". Rename "shift" to "shift_imm".
7800 * config/arm/arm.md (define_attr "core_cycles"): Update for attribute
7801 changes. Update for attribute changes all occurrences of arlo_* and
7803 * config/arm/arm-fixed.md: Update for attribute changes
7804 all occurrences of arlo_* types.
7805 * config/arm/thumb2.md: Update for attribute changes all occurrences
7807 * config/arm/arm.c (xscale_sched_adjust_cost): (rtx insn, rtx
7808 (cortexa7_older_only): Likewise.
7809 (cortexa7_younger): Likewise.
7810 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
7811 (1020alu_shift_op): Likewise.
7812 (1020alu_shift_reg_op): Likewise.
7813 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
7814 (alu_shift_op): Likewise.
7815 (alu_shift_reg_op): Likewise.
7816 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
7817 (11_alu_shift_op): Likewise.
7818 (11_alu_shift_reg_op): Likewise.
7819 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
7820 (9_alu_shift_reg_op): Likewise.
7821 * config/arm/cortex-a15.md (cortex_a15_alu): Update for
7823 (cortex_a15_alu_shift): Likewise.
7824 (cortex_a15_alu_shift_reg): Likewise.
7825 * config/arm/cortex-a5.md (cortex_a5_alu): Update for
7827 (cortex_a5_alu_shift): Likewise.
7828 * config/arm/cortex-a53.md (cortex_a53_alu): Update for
7830 (cortex_a53_alu_shift): Likewise.
7831 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for
7833 (cortex_a7_alu_reg): Likewise.
7834 (cortex_a7_alu_shift): Likewise.
7835 * config/arm/cortex-a8.md (cortex_a8_alu): Update for
7837 (cortex_a8_alu_shift): Likewise.
7838 (cortex_a8_alu_shift_reg): Likewise.
7839 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
7840 (cortex_a9_dp_shift): Likewise.
7841 * config/arm/cortex-m4.md (cortex_m4_alu): Update for
7843 * config/arm/cortex-r4.md
7844 (cortex_r4_alu): Update for attribute changes.
7845 (cortex_r4_mov): Likewise.
7846 (cortex_r4_alu_shift_reg): Likewise.
7847 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
7848 (526_alu_shift_op): Likewise.
7849 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
7850 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
7851 (626te_alu_shift_op): Likewise.
7852 * config/arm/fa726te.md (726te_alu_op): Update for attribute changes.
7853 (726te_alu_shift_op): Likewise.
7854 (726te_alu_shift_reg_op): Likewise.
7855 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
7856 (mp626_alu_shift_op): Likewise.
7857 * config/arm/marvell-pj4.md (pj4_alu): Update for attribute changes.
7858 (pj4_alu_conds): Likewise.
7859 (pj4_shift): Likewise.
7860 (pj4_shift_conds): Likewise.
7861 (pj4_alu_shift): Likewise.
7862 (pj4_alu_shift_conds): Likewise.
7863 * config/aarch64/aarch64.md: Update for attribute change
7864 all occurrences of arlo_* and shift* types.
7866 2013-09-05 Mike Stump <mikestump@comcast.net>
7868 * tree.h: Move documentation for tree_function_decl to tree-core.h
7869 with the declaration.
7871 2013-09-05 Peter Bergner <bergner@vnet.ibm.com>
7874 * reginfo.c (choose_hard_reg_mode): Scan through all mode classes
7875 looking for widest mode.
7877 2013-09-05 Eric Botcazou <ebotcazou@adacore.com>
7879 * config.gcc (*-*-vxworks*): Do not override an existing extra_objs.
7881 2013-09-05 Richard Biener <rguenther@suse.de>
7883 PR tree-optimization/58137
7884 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size):
7885 Do not create vectors of pointers.
7886 * tree-vect-loop.c (get_initial_def_for_induction): Use proper
7887 types for the components of the vector initializer.
7888 * tree-cfg.c (verify_gimple_assign_binary): Remove special-casing
7889 allowing pointer vectors with PLUS_EXPR/MINUS_EXPR.
7891 2013-09-05 Martin Jambor <mjambor@suse.cz>
7893 * ipa-prop.c (remove_described_reference): Accept missing references,
7894 return false if that hppens, otherwise return true.
7895 (cgraph_node_for_jfunc): New function.
7896 (try_decrement_rdesc_refcount): Likewise.
7897 (try_make_edge_direct_simple_call): Use them.
7898 (ipa_edge_removal_hook): Remove references from rdescs.
7899 (ipa_edge_duplication_hook): Clone rdescs and their references
7900 when the new edge has the same caller as the old one.
7901 * cgraph.c (cgraph_resolve_speculation): Remove speculative
7902 reference before removing any edges.
7904 2013-09-05 Richard Earnshaw <rearnsha@arm.com>
7906 * arm.c (thumb2_emit_strd_push): Rewrite to use pre-decrement on
7908 * thumb2.md (thumb2_storewb_parisi): New pattern.
7910 2013-09-05 Yufeng Zhang <yufeng.zhang@arm.com>
7912 * config/aarch64/aarch64-option-extensions.def: Add
7913 AARCH64_OPT_EXTENSION of 'crc'.
7914 * config/aarch64/aarch64.h (AARCH64_FL_CRC): New define.
7915 (AARCH64_ISA_CRC): Ditto.
7916 * doc/invoke.texi (-march and -mcpu feature modifiers): Add
7917 description of the CRC extension.
7919 2013-09-05 Alexander Ivchenko <alexander.ivchenko@intel.com>
7921 * config/rs6000/linux64.h: Define OPTION_BIONIC and OPTION_UCLIBC.
7922 * config/rs6000/linux.h: Ditto.
7923 * alpha/linux.h: Ditto.
7924 * config/bfin/uclinux.h: Define TARGET_LIBC_HAS_FUNCTION as
7925 no_c99_libc_has_function.
7926 * config/c6x/uclinux-elf.h: Ditto.
7927 * config/lm32/uclinux-elf.h: Ditto.
7928 * config/m68k/uclinux.h: Ditto.
7929 * config/moxie/uclinux.h: Ditto.
7930 * config.gcc (bfin*-linux-uclibc*): Add t-linux-android to tmake_file.
7931 (crisv32-*-linux*, cris-*-linux*): Ditto.
7932 * config/bfin/bfin.c: Include "tm_p.h".
7934 2013-09-05 Richard Biener <rguenther@suse.de>
7936 * tree-vect-loop.c (vect_analyze_loop_operations): Properly
7937 check for a definition without a basic-block.
7939 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
7940 Sofiane Naci <sofiane.naci@arm.com>
7942 * config/aarch64/aarch64.md
7943 (*movti_aarch64): Rename r_2_f and f_2_r.
7944 (*movsf_aarch64): Likewise.
7945 (*movdf_aarch64): Likewise.
7946 (*movtf_aarch64): Likewise.
7947 (aarch64_movdi_<mode>low): Likewise.
7948 (aarch64_movdi_<mode>high): Likewise.
7949 (aarch64_mov<mode>high_di): Likewise.
7950 (aarch64_mov<mode>low_di): Likewise.
7951 (aarch64_movtilow_tilow): Likewise.
7952 * config/arm/arm.md (attribute "neon_type"): Delete. Move attribute
7953 values to config/arm/types.md
7954 (attribute "conds"): Update for attribute change.
7955 (anddi3_insn): Likewise.
7956 (iordi3_insn): Likewise.
7957 (xordi3_insn): Likewise.
7958 (one_cmpldi2): Likewise.
7959 * config/arm/types.md (type): Add Neon types.
7960 * config/arm/neon.md (neon_mov<mode>): Remove "neon_type" attribute,
7961 use "type" attribute.
7962 (movmisalign<mode>_neon_store): Likewise.
7963 (movmisalign<mode>_neon_load): Likewise.
7964 (vec_set<mode>_internal): Likewise.
7965 (vec_setv2di_internal): Likewise.
7966 (vec_extract<mode>): Likewise.
7967 (vec_extractv2di): Likewise.
7968 (add<mode>3_neon): Likewise.
7969 (adddi3_neon): Likewise.
7970 (sub<mode>3_neon): Likewise.
7971 (subdi3_neon): Likewise.
7972 (mul<mode>3_neon): Likewise.
7973 (mul<mode>3add<mode>_neon): Likewise.
7974 (mul<mode>3neg<mode>add<mode>_neon): Likewise.
7975 (fma<VCVTF:mode>4)): Likewise.
7976 (fma<VCVTF:mode>4_intrinsic): Likewise.
7977 (fmsub<VCVTF:mode>4)): Likewise.
7978 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
7979 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
7980 (ior<mode>3): Likewise.
7981 (and<mode>3): Likewise.
7982 (anddi3_neon): Likewise.
7983 (orn<mode>3_neon): Likewise.
7984 (orndi3_neon): Likewise.
7985 (bic<mode>3_neon): Likewise.
7986 (bicdi3_neon): Likewise.
7987 (xor<mode>3): Likewise.
7988 (one_cmpl<mode>2): Likewise.
7989 (abs<mode>2): Likewise.
7990 (neg<mode>2): Likewise.
7991 (umin<mode>3_neon): Likewise.
7992 (umax<mode>3_neon): Likewise.
7993 (smin<mode>3_neon): Likewise.
7994 (smax<mode>3_neon): Likewise.
7995 (vashl<mode>3): Likewise.
7996 (vashr<mode>3_imm): Likewise.
7997 (vlshr<mode>3_imm): Likewise.
7998 (ashl<mode>3_signed): Likewise.
7999 (ashl<mode>3_unsigned): Likewise.
8000 (neon_load_count): Likewise.
8001 (ashldi3_neon_noclobber): Likewise.
8002 (signed_shift_di3_neon): Likewise.
8003 (unsigned_shift_di3_neon): Likewise.
8004 (ashrdi3_neon_imm_noclobber): Likewise.
8005 (lshrdi3_neon_imm_noclobber): Likewise.
8006 (widen_ssum<mode>3): Likewise.
8007 (widen_usum<mode>3): Likewise.
8008 (quad_halves_<code>v4si): Likewise.
8009 (quad_halves_<code>v4sf): Likewise.
8010 (quad_halves_<code>v8hi): Likewise.
8011 (quad_halves_<code>v16qi): Likewise.
8012 (reduc_splus_v2di): Likewise.
8013 (neon_vpadd_internal<mode>): Likewise.
8014 (neon_vpsmin<mode>): Likewise.
8015 (neon_vpsmax<mode>): Likewise.
8016 (neon_vpumin<mode>): Likewise.
8017 (neon_vpumax<mode>): Likewise.
8018 (ss_add<mode>_neon): Likewise.
8019 (us_add<mode>_neon): Likewise.
8020 (ss_sub<mode>_neon): Likewise.
8021 (us_sub<mode>_neon): Likewise.
8022 (neon_vadd<mode>_unspec): Likewise.
8023 (neon_vaddl<mode>): Likewise.
8024 (neon_vaddw<mode>): Likewise.
8025 (neon_vhadd<mode>): Likewise.
8026 (neon_vqadd<mode>): Likewise.
8027 (neon_vaddhn<mode>): Likewise.
8028 (neon_vmul<mode>): Likewise.
8029 (neon_vmla<mode>): Likewise.
8030 (neon_vmlal<mode>): Likewise.
8031 (neon_vmls<mode>): Likewise.
8032 (neon_vmlsl<mode>): Likewise.
8033 (neon_vqdmulh<mode>): Likewise.
8034 (neon_vqdmlal<mode>): Likewise.
8035 (neon_vqdmlsl<mode>): Likewise.
8036 (neon_vmull<mode>): Likewise.
8037 (neon_vqdmull<mode>): Likewise.
8038 (neon_vsub<mode>_unspec): Likewise.
8039 (neon_vsubl<mode>): Likewise.
8040 (neon_vsubw<mode>): Likewise.
8041 (neon_vqsub<mode>): Likewise.
8042 (neon_vhsub<mode>): Likewise.
8043 (neon_vsubhn<mode>): Likewise.
8044 (neon_vceq<mode>): Likewise.
8045 (neon_vcge<mode>): Likewise.
8046 (neon_vcgeu<mode>): Likewise.
8047 (neon_vcgt<mode>): Likewise.
8048 (neon_vcgtu<mode>): Likewise.
8049 (neon_vcle<mode>): Likewise.
8050 (neon_vclt<mode>): Likewise.
8051 (neon_vcage<mode>): Likewise.
8052 (neon_vcagt<mode>): Likewise.
8053 (neon_vtst<mode>): Likewise.
8054 (neon_vabd<mode>): Likewise.
8055 (neon_vabdl<mode>): Likewise.
8056 (neon_vaba<mode>): Likewise.
8057 (neon_vabal<mode>): Likewise.
8058 (neon_vmax<mode>): Likewise.
8059 (neon_vmin<mode>): Likewise.
8060 (neon_vpaddl<mode>): Likewise.
8061 (neon_vpadal<mode>): Likewise.
8062 (neon_vpmax<mode>): Likewise.
8063 (neon_vpmin<mode>): Likewise.
8064 (neon_vrecps<mode>): Likewise.
8065 (neon_vrsqrts<mode>): Likewise.
8066 (neon_vqabs<mode>): Likewise.
8067 (neon_vqneg<mode>): Likewise.
8068 (neon_vcls<mode>): Likewise.
8069 (clz<mode>2): Likewise.
8070 (popcount<mode>2): Likewise.
8071 (neon_vrecpe): Likewise.
8072 (neon_vrsqrte): Likewise.
8073 (neon_vget_lane<mode>_sext_internal): Likewise.
8074 (neon_vget_lane<mode>_zext_internal): Likewise.
8075 (neon_vdup_n<mode>): Likewise.
8076 (neon_vdup_nv2di): Likewise.
8077 (neon_vdpu_lane<mode>_internal): Likewise.
8078 (neon_vswp<mode>): Likewise.
8079 (float<mode><V_cvtto>2): Likewise.
8080 (floatuns<mode><V_cvtto>2): Likewise.
8081 (fix_trunc<mode><V_cvtto>)2): Likewise
8082 (fixuns_trunc<mode><V_cvtto)2): Likewise.
8083 (neon_vcvt<mode>): Likewise.
8084 (neon_vcvtv4sfv4hf): Likewise.
8085 (neon_vcvtv4hfv4sf): Likewise.
8086 (neon_vcvt_n<mode>): Likewise.
8087 (neon_vmovn<mode>): Likewise.
8088 (neon_vqmovn<mode>): Likewise.
8089 (neon_vqmovun<mode>): Likewise.
8090 (neon_vmovl<mode>): Likewise.
8091 (neon_vmul_lane<mode>): Likewise.
8092 (neon_vmull_lane<mode>): Likewise.
8093 (neon_vqdmull_lane<mode>): Likewise.
8094 (neon_vqdmulh_lane<mode>): Likewise.
8095 (neon_vmla_lane<mode>): Likewise.
8096 (neon_vmlal_lane<mode>): Likewise.
8097 (neon_vqdmlal_lane<mode>): Likewise.
8098 (neon_vmls_lane<mode>): Likewise.
8099 (neon_vmlsl_lane<mode>): Likewise.
8100 (neon_vqdmlsl_lane<mode>): Likewise.
8101 (neon_vext<mode>): Likewise.
8102 (neon_vrev64<mode>): Likewise.
8103 (neon_vrev32<mode>): Likewise.
8104 (neon_vrev16<mode>): Likewise.
8105 (neon_vbsl<mode>_internal): Likewise.
8106 (neon_vshl<mode>): Likewise.
8107 (neon_vqshl<mode>): Likewise.
8108 (neon_vshr_n<mode>): Likewise.
8109 (neon_vshrn_n<mode>): Likewise.
8110 (neon_vqshrn_n<mode>): Likewise.
8111 (neon_vqshrun_n<mode>): Likewise.
8112 (neon_vshl_n<mode>): Likewise.
8113 (neon_vqshl_n<mode>): Likewise.
8114 (neon_vqshlu_n<mode>): Likewise.
8115 (neon_vshll_n<mode>): Likewise.
8116 (neon_vsra_n<mode>): Likewise.
8117 (neon_vsri_n<mode>): Likewise.
8118 (neon_vsli_n<mode>): Likewise.
8119 (neon_vtbl1v8qi): Likewise.
8120 (neon_vtbl2v8qi): Likewise.
8121 (neon_vtbl3v8qi): Likewise.
8122 (neon_vtbl4v8qi): Likewise.
8123 (neon_vtbx1v8qi): Likewise.
8124 (neon_vtbx2v8qi): Likewise.
8125 (neon_vtbx3v8qi): Likewise.
8126 (neon_vtbx4v8qi): Likewise.
8127 (neon_vtrn<mode>_internal): Likewise.
8128 (neon_vzip<mode>_internal): Likewise.
8129 (neon_vuzp<mode>_internal): Likewise.
8130 (neon_vld1<mode>): Likewise.
8131 (neon_vld1_lane<mode>): Likewise.
8132 (neon_vld1_dup<mode>): Likewise.
8133 (neon_vld1_dupv2di): Likewise.
8134 (neon_vst1<mode>): Likewise.
8135 (neon_vst1_lane<mode>): Likewise.
8136 (neon_vld2<mode>): Likewise.
8137 (neon_vld2_lane<mode>): Likewise.
8138 (neon_vld2_dup<mode>): Likewise.
8139 (neon_vst2<mode>): Likewise.
8140 (neon_vst2_lane<mode>): Likewise.
8141 (neon_vld3<mode>): Likewise.
8142 (neon_vld3qa<mode>): Likewise.
8143 (neon_vld3qb<mode>): Likewise.
8144 (neon_vld3_lane<mode>): Likewise.
8145 (neon_vld3_dup<mode>): Likewise.
8146 (neon_vst3<mode>): Likewise.
8147 (neon_vst3qa<mode>): Likewise.
8148 (neon_vst3qb<mode>): Likewise.
8149 (neon_vst3_lane<mode>): Likewise.
8150 (neon_vld4<mode>): Likewise.
8151 (neon_vld4qa<mode>): Likewise.
8152 (neon_vld4qb<mode>): Likewise.
8153 (neon_vld4_lane<mode>): Likewise.
8154 (neon_vld4_dup<mode>): Likewise.
8155 (neon_vst4<mode>): Likewise.
8156 (neon_vst4qa<mode>): Likewise.
8157 (neon_vst4qb<mode>): Likewise.
8158 (neon_vst4_lane<mode>): Likewise.
8159 (neon_vec_unpack<US>_lo_<mode>): Likewise.
8160 (neon_vec_unpack<US>_hi_<mode>): Likewise.
8161 (neon_vec_<US>mult_lo_<mode>): Likewise.
8162 (neon_vec_<US>mult_hi_<mode>): Likewise.
8163 (neon_vec_<US>shiftl_<mode>): Likewise.
8164 (neon_unpack<US>_<mode>): Likewise.
8165 (neon_vec_<US>mult_<mode>): Likewise.
8166 (vec_pack_trunc_<mode>): Likewise.
8167 (neon_vec_pack_trunk_<mode>): Likewise.
8168 (neon_vabd<mode>_2): Likewise.
8169 (neon_vabd<mode>_3): Likewise.
8170 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
8171 (thumb2_movsi_vfp): Likewise.
8172 (movdi_vfp): Likewise.
8173 (movdi_vfp_cortexa8): Likewise.
8174 (movhf_vfp_neon): Likewise.
8175 (movhf_vfp): Likewiwse.
8176 (movsf_vfp): Likewiwse.
8177 (thumb2_movsf_vfp): Likewiwse.
8178 (movdf_vfp): Likewise.
8179 (thumb2_movdf_vfp): Likewise.
8180 (movsfcc_vfp): Likewise.
8181 (thumb2_movsfcc_vfp): Likewise.
8182 (movdfcc_vfp): Likewise.
8183 (thumb2_movdfcc_vfp): Likewise.
8184 * config/arm/arm.c (cortexa7_older_only): Update for attribute change.
8185 * config/arm/arm1020e.md (v10_c2v): Update for attribute change.
8186 (v10_v2c): Likewise.
8187 * config/arm/cortex-a15-neon.md (cortex_a15_neon_int_1): Update for
8189 (cortex_a15_neon_int_2): Likewise.
8190 (cortex_a15_neon_int_3): Likewise.
8191 (cortex_a15_neon_int_4): Likewise.
8192 (cortex_a15_neon_int_5): Likewise.
8193 (cortex_a15_neon_vqneg_vqabs): Likewise.
8194 (cortex_a15_neon_vmov): Likewise.
8195 (cortex_a15_neon_vaba): Likewise.
8196 (cortex_a15_neon_vaba_qqq): Likewise.
8197 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
8198 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
8199 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_\
8200 scalar_64_32_long_scalar): Likewise.
8201 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
8202 (cortex_a15_neon_mla_qqq_8_16): Likewise.
8203 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
8204 lotype_qdd_64_32_long): Likewise.
8205 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
8206 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
8207 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
8208 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
8209 (cortex_a15_neon_shift_1): Likewise.
8210 (cortex_a15_neon_shift_2): Likewise.
8211 (cortex_a15_neon_shift_3): Likewise.
8212 (cortex_a15_neon_vshl_ddd): Likewise.
8213 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
8214 (cortex_a15_neon_vsra_vrsra): Likewise.
8215 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
8216 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
8217 (cortex_a15_neon_fp_vmul_ddd): Likewise.
8218 (cortex_a15_neon_fp_vmul_qqd): Likewise.
8219 (cortex_a15_neon_fp_vmla_ddd): Likewise.
8220 (cortex_a15_neon_fp_vmla_qqq): Likewise.
8221 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
8222 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
8223 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
8224 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
8225 (cortex_a15_neon_bp_simple): Likewise.
8226 (cortex_a15_neon_bp_2cycle): Likewise.
8227 (cortex_a15_neon_bp_3cycle): Likewise.
8228 (cortex_a15_neon_vld1_1_2_regs): Likewise.
8229 (cortex_a15_neon_vld1_3_4_regs): Likewise.
8230 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
8231 (cortex_a15_neon_vld2_4_regs): Likewise.
8232 (cortex_a15_neon_vld3_vld4): Likewise.
8233 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
8234 (cortex_a15_neon_vst1_3_4_regs): Likewise.
8235 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Likewise.
8236 (cortex_a15_neon_vst3_vst4): Likewise.
8237 (cortex_a15_neon_vld1_vld2_lane): Likewise.
8238 (cortex_a15_neon_vld3_vld4_lane" 10
8239 (cortex_a15_neon_vst1_vst2_lane): Likewise.
8240 (cortex_a15_neon_vst3_vst4_lane): Likewise.
8241 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
8242 (cortex_a15_neon_ldm_2): Likewise.0
8243 (cortex_a15_neon_stm_2): Likewise.
8244 (cortex_a15_neon_mcr): Likewise.
8245 (cortex_a15_neon_mcr_2_mcrr): Likewise.
8246 (cortex_a15_neon_mrc): Likewise.
8247 (cortex_a15_neon_mrrc): Likewise.
8248 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
8250 (cortex_a15_alu_shift): Likewise.
8251 (cortex_a15_alu_shift_reg): Likewise.
8252 (cortex_a15_mult32): Likewise.
8253 (cortex_a15_mult64): Likewise.
8254 (cortex_a15_block): Likewise.
8255 (cortex_a15_branch): Likewise.
8256 (cortex_a15_load1): Likewise.
8257 (cortex_a15_load3): Likewise.
8258 (cortex_a15_store1): Likewise.
8259 (cortex_a15_store3): Likewise.
8260 (cortex_a15_call): Likewise.
8261 * config/arm/cortex-a5.md (cortex_a5_r2f): Update for attribute change.
8262 (cortex_a5_f2r): Likewise.
8263 * config/arm/cortex-a53.md (cortex_a53_r2f): Update for attribute
8265 (cortex_a53_f2r): Likewise.
8266 * config/arm/cortex-a7.md
8267 (cortex_a7_branch): Update for attribute change.
8268 (cortex_a7_call): Likewise.
8269 (cortex_a7_alu_imm): Likewise.
8270 (cortex_a7_alu_reg): Likewise.
8271 (cortex_a7_alu_shift): Likewise.
8272 (cortex_a7_mul): Likewise.
8273 (cortex_a7_load1): Likewise.
8274 (cortex_a7_store1): Likewise.
8275 (cortex_a7_load2): Likewise.
8276 (cortex_a7_store2): Likewise.
8277 (cortex_a7_load3): Likewise.
8278 (cortex_a7_store3): Likewise.
8279 (cortex_a7_load4): Likewise.
8280 (cortex_a7_store4): Likewise.
8281 (cortex_a7_fpalu): Likewise.
8282 (cortex_a7_fconst): Likewise.
8283 (cortex_a7_fpmuls): Likewise.
8284 (cortex_a7_neon_mul): Likewise.
8285 (cortex_a7_fpmacs): Likewise.
8286 (cortex_a7_neon_mla: Likewise.
8287 (cortex_a7_fpmuld: Likewise.
8288 (cortex_a7_fpmacd: Likewise.
8289 (cortex_a7_fpfmad: Likewise.
8290 (cortex_a7_fdivs: Likewise.
8291 (cortex_a7_fdivd: Likewise.
8292 (cortex_a7_r2f: Likewise.
8293 (cortex_a7_f2r: Likewise.
8294 (cortex_a7_f_flags: Likewise.
8295 (cortex_a7_f_loads: Likewise.
8296 (cortex_a7_f_loadd: Likewise.
8297 (cortex_a7_f_stores: Likewise.
8298 (cortex_a7_f_stored: Likewise.
8299 (cortex_a7_neon): Likewise.
8300 * config/arm/cortex-a8-neon.md
8301 (cortex_a8_neon_mrc): Update for attribute change.
8302 (cortex_a8_neon_mrrc): Likewise.
8303 (cortex_a8_neon_int_1): Likewise.
8304 (cortex_a8_neon_int_2): Likewise.
8305 (cortex_a8_neon_int_3): Likewise.
8306 (cortex_a8_neon_int_4): Likewise.
8307 (cortex_a8_neon_int_5): Likewise.
8308 (cortex_a8_neon_vqneg_vqabs): Likewise.
8309 (cortex_a8_neon_vmov): Likewise.
8310 (cortex_a8_neon_vaba): Likewise.
8311 (cortex_a8_neon_vaba_qqq): Likewise.
8312 (cortex_a8_neon_vsma): Likewise.
8313 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
8314 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
8315 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
8316 long_scalar): Likewise.
8317 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
8318 (cortex_a8_neon_mla_qqq_8_16): Likewise.
8319 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
8320 long_scalar_qdd_64_32_long): Likewise.
8321 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
8322 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
8323 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
8324 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
8325 (cortex_a8_neon_shift_1): Likewise.
8326 (cortex_a8_neon_shift_2): Likewise.
8327 (cortex_a8_neon_shift_3): Likewise.
8328 (cortex_a8_neon_vshl_ddd): Likewise.
8329 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
8330 (cortex_a8_neon_vsra_vrsra): Likewise.
8331 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
8332 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
8333 (cortex_a8_neon_fp_vsum): Likewise.
8334 (cortex_a8_neon_fp_vmul_ddd): Likewise.
8335 (cortex_a8_neon_fp_vmul_qqd): Likewise.
8336 (cortex_a8_neon_fp_vmla_ddd): Likewise.
8337 (cortex_a8_neon_fp_vmla_qqq): Likewise.
8338 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
8339 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
8340 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
8341 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
8342 (cortex_a8_neon_bp_simple): Likewise.
8343 (cortex_a8_neon_bp_2cycle): Likewise.
8344 (cortex_a8_neon_bp_3cycle): Likewise.
8345 (cortex_a8_neon_ldr): Likewise.
8346 (cortex_a8_neon_str): Likewise.
8347 (cortex_a8_neon_vld1_1_2_regs): Likewise.
8348 (cortex_a8_neon_vld1_3_4_regs): Likewise.
8349 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
8350 (cortex_a8_neon_vld2_4_regs): Likewise.
8351 (cortex_a8_neon_vld3_vld4): Likewise.
8352 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
8353 (cortex_a8_neon_vst1_3_4_regs): Likewise.
8354 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
8355 (cortex_a8_neon_vst3_vst4): Likewise.
8356 (cortex_a8_neon_vld1_vld2_lane): Likewise.
8357 (cortex_a8_neon_vld3_vld4_lane): Likewise.
8358 (cortex_a8_neon_vst1_vst2_lane): Likewise.
8359 (cortex_a8_neon_vst3_vst4_lane): Likewise.
8360 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
8361 (cortex_a8_neon_mcr): Likewise.
8362 (cortex_a8_neon_mcr_2_mcrr): Likewise.
8363 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
8364 * config/arm/cortex-a9-neon.md (ca9_neon_mrc): Update for attribute
8366 (ca9_neon_mrrc): Likewise.
8367 (cortex_a9_neon_int_1): Likewise.
8368 (cortex_a9_neon_int_2): Likewise.
8369 (cortex_a9_neon_int_3): Likewise.
8370 (cortex_a9_neon_int_4): Likewise.
8371 (cortex_a9_neon_int_5): Likewise.
8372 (cortex_a9_neon_vqneg_vqabs): Likewise.
8373 (cortex_a9_neon_vmov): Likewise.
8374 (cortex_a9_neon_vaba): Likewise.
8375 (cortex_a9_neon_vaba_qqq): Likewise.
8376 (cortex_a9_neon_vsma): Likewise.
8377 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
8378 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
8379 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
8380 long_scalar): Likewise.
8381 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
8382 (cortex_a9_neon_mla_qqq_8_16): Likewise.
8383 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
8384 long_scalar_qdd_64_32_long): Likewise.
8385 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
8386 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
8387 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
8388 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
8389 (cortex_a9_neon_shift_1): Likewise.
8390 (cortex_a9_neon_shift_2): Likewise.
8391 (cortex_a9_neon_shift_3): Likewise.
8392 (cortex_a9_neon_vshl_ddd): Likewise.
8393 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
8394 (cortex_a9_neon_vsra_vrsra): Likewise.
8395 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
8396 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
8397 (cortex_a9_neon_fp_vsum): Likewise.
8398 (cortex_a9_neon_fp_vmul_ddd): Likewise.
8399 (cortex_a9_neon_fp_vmul_qqd): Likewise.
8400 (cortex_a9_neon_fp_vmla_ddd): Likewise.
8401 (cortex_a9_neon_fp_vmla_qqq): Likewise.
8402 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
8403 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
8404 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
8405 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
8406 (cortex_a9_neon_bp_simple): Likewise.
8407 (cortex_a9_neon_bp_2cycle): Likewise.
8408 (cortex_a9_neon_bp_3cycle): Likewise.
8409 (cortex_a9_neon_ldr): Likewise.
8410 (cortex_a9_neon_str): Likewise.
8411 (cortex_a9_neon_vld1_1_2_regs): Likewise.
8412 (cortex_a9_neon_vld1_3_4_regs): Likewise.
8413 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
8414 (cortex_a9_neon_vld2_4_regs): Likewise.
8415 (cortex_a9_neon_vld3_vld4): Likewise.
8416 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
8417 (cortex_a9_neon_vst1_3_4_regs): Likewise.
8418 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
8419 (cortex_a9_neon_vst3_vst4): Likewise.
8420 (cortex_a9_neon_vld1_vld2_lane): Likewise.
8421 (cortex_a9_neon_vld3_vld4_lane): Likewise.
8422 (cortex_a9_neon_vst1_vst2_lane): Likewise.
8423 (cortex_a9_neon_vst3_vst4_lane): Likewise.
8424 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
8425 (cortex_a9_neon_mcr): Likewise.
8426 (cortex_a9_neon_mcr_2_mcrr): Likewise.
8427 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
8428 (cortex_a9_fps): Likewise.
8429 * config/arm/cortex-m4-fpu.md (cortex_m4_vmov_2): Update for attribute
8431 (cortex_m4_fmuls): Likewise.
8432 * config/arm/cortex-r4f.md (cortex_r4_mcr): Update for attribute
8434 (cortex_r4_mrc): Likewise.
8435 * config/arm/iterators.md: Update comment referring to neon_type.
8436 * config/arm/iwmmxt.md (iwmmxt_arm_movdi): Update for attribute change.
8437 (iwmmxt_movsi_insn): Likewise.
8438 * config/arm/marvell-pj4.md (pj4_vfp_to_core): Update for
8440 (pj4_core_to_vfp): Likewise.
8441 * config/arm/neon-schedgen.ml (emit_insn_reservations): Update for
8443 * config/arm/vfp11.md (vfp_fload): Update for attribute change.
8444 (vfp_fstore): Likewise.
8445 * doc/md.texi: Change references to neon_type to refer to type.
8447 2013-09-04 Dodji Seketeli <dodji@redhat.com>
8449 * tree.h (DECL_BUILT_IN): Fix typo in comment.
8451 2013-09-04 David Edelsohn <dje.gcc@gmail.com>
8453 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Only emit
8456 2013-09-04 Easwaran Raman <eraman@google.com>
8459 * tree-ssa-reassoc.c (get_stmt_uid_with_default): New function,
8460 (build_and_add_sum): Use it.
8461 (appears_later_in_bb): Simplify code.
8463 2013-09-04 Teresa Johnson <tejohnson@google.com>
8465 * dumpfile.c (dump_finish): Don't close stderr/stdout.
8467 2013-09-04 James Greenhalgh <james.greenhalgh@arm.com>
8469 * config/aarch64/arm_neon.h (vaddvq_<su>64): Fix return types.
8471 2013-09-04 Jan Hubicka <jh@suse.cz>
8473 * Makefile.in (ipa-devirt.o): Add dependency on diagnostic.h
8474 * ipa-devirt.c: Include diganostic.h
8475 (odr_type_d): Add types and types_set.
8476 (hash_type_name): Work for types with vtables during LTO.
8477 (odr_hasher::remove): Fix comment; destroy types_set.
8478 (add_type_duplicate): New function,
8479 (get_odr_type): Use it.
8480 (dump_type_inheritance_graph): Dump type duplicates.
8481 * ipa.c (symtab_remove_unreachable_nodes): Build type inheritance
8483 * tree.c (types_same_for_odr): Give exact answers on types with
8486 2013-09-04 Dodji Seketeli <dodji@redhat.com>
8488 * tree.h (DECL_BUILT_IN, DECL_IS_BUILTIN): Add more comments
8489 explaining their differences.
8491 2013-09-04 Sandeep Kumar Singh<Sandeep.Singh2@kpitcummins.com>
8493 * config/rx/rx.h: Add option -mcpu for target variants RX100 and RX200.
8495 2013-09-03 Jeff Law <law@redhat.com>
8497 * tree-ssa-threadedge.c (thread_across_edge): Record entire path
8498 when not threading through a joiner block. Pass joiner/no joiner
8499 state to register_jump_thread.
8500 * tree-ssa-threadupdate.c (register_jump_thread): Get joiner/no joiner
8501 state from argument rather than implying on path length.
8502 Dump the entire jump thread path into debugging dump.
8503 * tree-flow.h (register_jump_thread): Update prototype.
8505 2013-08-29 Xinliang David Li <davidxl@google.com>
8507 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
8508 Remove a trivial gcc_assert.
8510 2013-08-29 Xinliang David Li <davidxl@google.com>
8512 * tree-vect-slp.c (destroy_bb_vec_info): Data ref cleanup.
8513 * tree-vect-loop.c (destroy_bb_vec_info): Ditto.
8514 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
8515 Delay base decl alignment adjustment.
8516 * tree-vectorizer.c (vect_destroy_datarefs): New function.
8517 * tree-vectorizer.h: New data structure.
8518 (set_dr_misalignment): New function.
8519 (dr_misalignment): Ditto.
8520 * tree-vect-stmts.c (vectorizable_store): Ensure alignment.
8521 (vectorizable_load): Ditto.
8522 (ensure_base_align): New function.
8523 (vectorize_loops): Add dbg_cnt support.
8524 (execute_vect_slp): Ditto.
8525 * dbgcnt.def: New debug counter.
8526 * Makefile: New dependency.
8528 2013-09-03 Meador Inge <meadori@codesourcery.com>
8532 2013-08-30 Meador Inge <meadori@codesourcery.com>
8534 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
8536 2013-09-03 David Edelsohn <dje.gcc@gmail.com>
8538 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Emit lglobl for
8539 function descriptor.
8541 2013-09-03 Richard Biener <rguenther@suse.de>
8543 * tree-affine.c (add_elt_to_tree): Fix association issue,
8544 avoid useless converts and make sure to always return a
8545 properly typed result.
8547 2013-09-03 Richard Biener <rguenther@suse.de>
8550 * fold-const.c (negate_expr_p): Fix division case.
8551 (negate_expr): Likewise.
8553 2013-09-03 Richard Biener <rguenther@suse.de>
8556 * tree-streamer-out.c: Include tm.h.
8557 * Makefile.in (tree-streamer-out.o): Depend on $(TM_H).
8559 2013-09-03 Jan Hubicka <jh@suse.cz>
8561 * tree-profile.c (tree_profiling): Cleanup CFG when done.
8563 2013-09-03 Alan Modra <amodra@gmail.com>
8565 * config.gcc (powerpc*-*-linux*): Add support for little-endian
8566 multilibs to big-endian target and vice versa.
8567 * config/rs6000/t-linux64: Use := assignment on all vars.
8568 (MULTILIB_EXTRA_OPTS): Remove fPIC.
8569 (MULTILIB_OSDIRNAMES): Specify using mapping from multilib_options.
8570 * config/rs6000/t-linux64le: New file.
8571 * config/rs6000/t-linux64bele: New file.
8572 * config/rs6000/t-linux64lebe: New file.
8574 2013-09-02 Jan Hubicka <jh@suse.cz>
8576 * ipa-inline-transform.c (inline_transform): Do not
8577 optimize_inline_calls when not optimizing.
8579 2013-09-02 Jan Hubicka <jh@suse.cz>
8581 * lto-symtab.c (lto_symtab_merge_symbols): Add comments; merge
8582 duplicated nodes for assembler names.
8583 * symtab.c (symtab_unregister_node): Do not attempt to unlink
8584 hard registers from assembler name hash.
8586 2013-09-02 Jan Hubicka <jh@suse.cz>
8588 * ipa-split.c (execute_split_functions): Split externally visible
8589 functions called once.
8591 2013-09-02 Martin Jambor <mjambor@suse.cz>
8594 * ipa-prop.c (ipa_edge_duplication_hook): Always put new rdesc to the
8595 linked list. When finding the correct duplicate, also consider also
8596 the caller in additon to its inlined_to node.
8598 2013-09-02 James Greenhalgh <james.greenhalgh@arm.com>
8600 * config/aarch64/aarch64-simd-builtins.def
8601 (dup_lane_scalar): Remove.
8602 * config/aarch64/aarch64-simd.md
8603 (aarch64_simd_dup): Add 'w->w' alternative.
8604 (aarch64_dup_lane<mode>): Allow for VALL.
8605 (aarch64_dup_lane_scalar<mode>): Remove.
8606 (aarch64_dup_lane_<vswap_width_name><mode>): New.
8607 (aarch64_get_lane_signed<mode>): Add w->w altenative.
8608 (aarch64_get_lane_unsigned<mode>): Likewise.
8609 (aarch64_get_lane<mode>): Likewise.
8610 * config/aarch64/aarch64.c (aarch64_evpc_dup): New.
8611 (aarch64_expand_vec_perm_const_1): Use aarch64_evpc_dup.
8612 * config/aarch64/iterators.md (VSWAP_WIDTH): New.
8613 (VCON): Change container of V2SF.
8614 (vswap_width_name): Likewise.
8615 * config/aarch64/arm_neon.h
8616 (__aarch64_vdup_lane_any): New.
8617 (__aarch64_vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
8618 (vdup<q>_n_<psuf><8,16,32,64>): Convert to C implementation.
8619 (vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
8621 2013-09-02 Eric Botcazou <ebotcazou@adacore.com>
8624 * expr.c (emit_move_complex): Do not move complex FP values as parts if
8625 the source or the destination is a single hard register.
8627 2013-09-02 Richard Biener <rguenther@suse.de>
8630 * tree-scalar-evolution.c (instantiate_scev_name): Allow
8633 2013-09-02 Richard Biener <rguenther@suse.de>
8635 * tree-affine.c (add_elt_to_tree): Avoid converting all pointer
8636 arithmetic to sizetype.
8638 2013-09-02 Bin Cheng <bin.cheng@arm.com>
8640 * tree-ssa-loop-ivopts.c (set_autoinc_for_original_candidates):
8641 Find auto-increment use both before and after candidate.
8643 2013-09-02 Marek Polacek <polacek@redhat.com>
8645 * Makefile.in (ubsan.o): Add $(TM_P_H) dependency.
8647 2013-09-01 Jan Hubicka <jh@suse.cz>
8649 * Makefile.in: Add ipa-profile.o
8650 (ipa.o, ipa-devrit.o, ipa-inline-analysis.o): Adjust dependencies.
8651 * cgraph.c (struct cgraph_propagate_frequency_data,
8652 cgraph_propagate_frequency_1, cgraph_propagate_frequency): Move to
8653 ipa-profile.c; replace cgraph_ by ipa_ prefix.
8654 * cgraph.h (cgraph_propagate_frequency): Remove.
8655 * ipa-inline-analysis.c: Include ipa-utils.h;
8656 drop duplicated cfgloop.h.
8657 (inline_update_callee_summaries): Update.
8658 * ipa-profile.c: New file.
8659 * ipa-utils.h (ipa_propagate_frequency): Declare.
8660 * ipa.c: Do not include pointer-set.h, hash-table.h, lto-streamer.h,
8661 data-streamer.h, value-prof.h.
8662 (symtab_remove_unreachable_nodes): Update profile.
8663 (struct histogram_entry, histogram, histogram_pool, histogram_hash,
8664 account_time_size, cmp_counts, dump_histogram,
8665 ipa_profile_generate_summary, ipa_profile_write_summary,
8666 ipa_profile_read_summary, ipa_profile, gate_ipa_profile,
8667 pass_data_ipa_profile, pass_ipa_profile, make_pass_ipa_profile):
8668 Move to ipa-profile.c.
8670 2013-09-01 John David Anglin <danglin@gcc.gnu.org>
8672 * config/pa/pa.md: Allow "const 0" operand 1 in "scc" insns.
8674 2013-09-01 Jan Hubicka <jh@suse.cz>
8676 * common.opt (fdevirtualize-speculatively): New function.
8677 * invoke.texi (fdevirtualize-speculatively): Document.
8678 * ipa-devirt.c: Include ipa-inline.h
8679 (likely_target_p): New function.
8680 (ipa_devirt): New function.
8681 (gate_ipa_devirt): New function.
8682 (pass_data_ipa_devirt): New static var.
8683 (pass_ipa_devirt): Likewise.
8684 (make_pass_ipa_devirt): New function.
8685 * opts.c (default_options): Add OPT_fdevirtualize_speculatively.
8686 (common_handle_option): Disable devirtualization when
8687 value range profiling is available.
8688 * passes.def (pass_ipa_devirt): Add.
8689 * timever.def (TV_IPA_DEVIRT): New timevar.
8690 * tree-pass.h (make_pass_ipa_devirt):
8692 2013-09-01 Iain Sandoe <iain@codesourcery.com>
8694 * config/darwin.h (LINK_COMMAND_SPEC_A): Revise sanitizer specs to
8695 include sanitize(undefined).
8697 2013-08-31 Diego Novillo <dnovillo@google.com>
8699 * Makefile.in (TREE_CORE_H): Define.
8701 (GTFILES): Add tree-core.h.
8702 * builtins.c (built_in_class_names): Use BUILT_IN_LAST to
8704 * tree-core.h: New file.
8705 Move all data structures, enum, typedefs, global
8706 declarations and constants from ...
8709 2013-08-31 Jan Hubicka <jh@suse.cz>
8711 * bulitins.c (expand_builtin): Do not early exit for gcov
8712 instrumented functions.
8714 2013-08-31 Marek Polacek <polacek@redhat.com>
8716 * ubsan.c: Include tm_p.h.
8718 2013-08-31 Jan Hubicka <jh@suse.cz>
8720 * gimple-streamer-in.c (input_gimple_stmt): Silence parameter unused
8723 * cgraph.c (cgraph_get_body): Update call of lto_input_function_body.
8724 * gimple-streamer-in.c (input_gimple_stmt): Move sanity check to ...
8725 * tree-cfg.c (verify_gimple_label): ... here.
8726 * ipa-utils.c: Include lto-streamer.h, ipa-inline.h
8727 (ipa_merge_profiles): New function.
8728 * lto-streamer-in.c (lto_read_body): Take node instead of fn_decl.
8729 (lto_input_function_body): Likewise.
8730 * ipa-utils.h (ipa_merge_profiles): Declare.
8731 * lto-streamer.h (lto_input_function_body): Update prototype.
8732 (emit_label_in_global_context_p): Remove.
8733 * lto-symtab.c: Include ipa-utils.h
8734 (lto_cgraph_replace_node): Use ipa_merge_profiles.
8736 2013-08-31 Jan Hubicka <jh@suse.cz>
8738 * cgraph.c (cgraph_speculative_call_info): Fix ref lookup
8740 2013-08-31 Jan Hubicka <jh@suse.cz>
8742 * basic-block.h (apply_scale): Make scale parmeter gcov_type.
8744 2013-08-31 Uros Bizjak <ubizjak@gmail.com>
8746 * config/alpha/alpha.c (alpha_emit_conditional_move): Update
8747 "cmp" RTX before signed_comparison_operator check to account
8750 2013-08-30 Jan Hubicka <jh@suse.cz>
8752 * ipa-prop.c (ipa_set_jf_known_type): Check that we add only records.
8753 (detect_type_change_1): Rename to ...
8754 (detect_type_change): ... this one; early return on non-polymorphic
8756 (detect_type_change_ssa): Add comp_type parameter; update
8757 use of detect_type_change.
8758 (compute_complex_assign_jump_func): Add param_type parameter;
8759 update use of detect_type_change_ssa.
8760 (compute_complex_ancestor_jump_func): Likewise.
8761 (ipa_get_callee_param_type): New function.
8762 (ipa_compute_jump_functions_for_edge): Compute parameter type;
8763 update calls to the jump function computation functions.
8765 2013-08-30 Teresa Johnson <tejohnson@google.com>
8766 Steven Bosscher <steven@gcc.gnu.org>
8768 * cfgrtl.c (fixup_new_cold_bb): New routine.
8769 (commit_edge_insertions): Invoke fixup_partitions.
8770 (find_partition_fixes): New routine.
8771 (fixup_partitions): Ditto.
8772 (verify_hot_cold_block_grouping): Update comments.
8773 (rtl_verify_edges): Invoke find_partition_fixes.
8774 (rtl_verify_bb_pointers): Update comments.
8775 (rtl_verify_bb_layout): Ditto.
8776 * basic-block.h (probably_never_executed_edge_p): Declare.
8777 (fixup_partitions): Ditto.
8778 * cfgcleanup.c (try_optimize_cfg): Invoke fixup_partitions.
8779 * bb-reorder.c (sanitize_hot_paths): New function.
8780 (find_rarely_executed_basic_blocks_and_crossing_edges): Invoke
8782 * predict.c (probably_never_executed_edge_p): New routine.
8783 * cfg.c (check_bb_profile): Add partition insanity warnings.
8785 2013-08-30 Meador Inge <meadori@codesourcery.com>
8787 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
8789 2013-08-30 Marek Polacek <polacek@redhat.com>
8791 * Makefile.in (ubsan.o): Add.
8792 (c-family/c-ubsan.o): Add.
8793 (builtins.o): Add ubsan.h dependency.
8794 * ubsan.h: New file.
8795 * ubsan.c: New file.
8796 * common.opt: Add -fsanitize=undefined option.
8797 (flag_sanitize): Add variable.
8798 (fsanitize=): Add option. Add Driver.
8799 (fsanitize=thread): Remove option.
8800 (fsanitize=address): Likewise.
8801 (static-libubsan): New option.
8802 * doc/invoke.texi: Document the new flag and -static-libubsan.
8803 * sanitizer.def (DEF_SANITIZER_BUILTIN): Define.
8804 (BUILT_IN_UBSAN_HANDLE_BUILTIN_UNREACHABLE): Define.
8805 * builtin-attrs.def (ATTR_COLD): Define.
8806 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
8807 * builtins.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW,
8808 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS): Define.
8809 * flag-types.h (sanitize_code): New enum.
8810 * opts.c (common_handle_option): Parse command line arguments
8811 of -fsanitize=. Add -fsanitize=unreachable option.
8812 * varasm.c (get_variable_section): Adjust.
8813 (assemble_noswitch_variable): Likewise.
8814 (assemble_variable): Likewise.
8815 (output_constant_def_contents): Likewise.
8816 (categorize_decl_for_section): Likewise.
8817 (place_block_symbol): Likewise.
8818 (output_object_block): Likewise.
8819 * builtins.def: Likewise.
8820 * toplev.c (compile_file): Likewise.
8821 (process_options): Likewise.
8822 * cppbuiltin.c: Likewise.
8823 * tsan.c (tsan_pass): Likewise.
8824 (tsan_gate): Likewise.
8825 (tsan_gate_O0): Likewise.
8826 * cfgexpand.c (partition_stack_vars): Likewise.
8827 (expand_stack_vars): Likewise.
8828 (defer_stack_allocation): Likewise.
8829 (expand_used_vars): Likewise.
8830 * cfgcleanup.c (old_insns_match_p): Likewise.
8831 * asan.c (asan_finish_file): Likewise.
8832 (asan_instrument): Likewise.
8833 (gate_asan): Likewise.
8834 (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR_PTR.
8835 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
8836 (asan_global_struct): Use pointer_sized_int_node instead
8837 calling build_nonstandard_integer_type.
8838 (initialize_sanitizer_builtins): Likewise.
8839 (asan_finish_file): Likewise.
8840 * gcc.c: Document %{%:function(args):X}.
8841 (static_spec_functions): Add sanitize.
8842 (handle_spec_function): Add retval_nonnull argument and if non-NULL,
8843 store funcval != NULL there.
8844 (do_spec_1): Adjust handle_spec_function caller.
8845 (handle_braces): Allow %:function(args) as condition.
8846 (sanitize_spec_function): New function.
8847 (ADD_STATIC_LIBUBSAN_LIBS): Define.
8848 (LIBUBSAN_SPEC): Likewise.
8849 (LIBUBSAN_EARLY_SPEC): Likewise.
8850 (SANITIZER_SPEC): Handle libubsan.
8851 (SANITIZER_EARLY_SPEC): Likewise.
8852 * config/darwin.h (LINK_COMMAND_SPEC_A): Use %:sanitize(address)
8853 instead of fsanitize=address.
8854 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Use %:sanitize(address)
8855 instead of fsanitize=address*.
8856 * builtins.c: Include ubsan.h.
8857 (fold_builtin_0): Instrument __builtin_unreachable.
8858 * config/rs6000/rs6000.h (FRAME_GROWS_DOWNWARD): Use flag_sanitize
8859 instead of flag_asan.
8860 * tree.h (enum tree_index): Add TI_POINTER_SIZED_TYPE.
8861 (pointer_sized_int_node): Define.
8862 * tree.c (build_common_tree_nodes): Initialize pointer_sized_int_node.
8864 2013-08-30 Mike Stump <mikestump@comcast.net>
8866 * doc/install.texi (Prerequisites): Note regression in Tcl 8.6
8869 2013-08-29 Jan Hubicka <jh@suse.cz>
8871 * cgraph.c (cgraph_function_body_availability): Handle weakref
8873 * passes.def: Remove pass_fixup_cfg.
8874 * ipa-inline.c (ipa_inline): When not optimizing, do not inline;
8875 track when we need to remove functions.
8876 (gate_ipa_inline): Execute inlining always; add comment why.
8877 (pass_data_ipa_inline): Remove TODO_remove_functions.
8878 * ipa-inline-analysis.c (inline_generate_summary): When not optimizing
8879 do not produce summaries.
8880 * symtab.c (change_decl_assembler_name): Handle renaming of weakrefs.
8881 (symtab_nonoverwritable_alias): Assert we are not called on weakref.
8882 * varpool.c (cgraph_variable_initializer_availability): Fix weakrefs,
8883 constant pool and vtable.
8885 2013-08-30 Tejas Belagod <tejas.belagod@arm.com>
8887 * config/aarch64/arm_neon.h (__AARCH64_UINT64_C, __AARCH64_INT64_C):
8888 New arm_neon.h's internal macros to specify 64-bit constants.
8889 Avoid using stdint.h's macros.
8891 2013-08-30 Joern Rennecke <joern.rennecke@embecosm.com>
8893 * recog.c (verify_changes): Verify that changes[i].old is non-zero
8894 before applying REG_P.
8896 2013-08-30 Jakub Jelinek <jakub@redhat.com>
8898 PR tree-optimization/58277
8899 * tree-ssa-strlen.c (strlen_enter_block): If do_invalidate gave up
8900 after seeing too many stmts with vdef in between dombb and current
8901 bb, invalidate everything.
8903 2013-08-30 Richard Biener <rguenther@suse.de>
8905 * fold-const.c (fold_single_bit_test): Fix overflow test.
8907 2013-08-30 Eric Botcazou <ebotcazou@adacore.com>
8909 * function.c (assign_parm_setup_reg): For a parameter passed by pointer
8910 and which can live in a register, always retrieve the value on entry.
8911 * var-tracking.c (add_stores): Treat the copy on entry for a parameter
8912 passed by invisible reference specially.
8913 (emit_notes_in_bb) <MO_VAL_USE>: Emit notes before the instruction.
8914 (vt_add_function_parameter): Correctly deal with a parameter passed by
8915 invisible reference.
8917 2013-08-30 Jan Hubicka <jh@suse.cz>
8919 * tree.c (set_call_expr_flags): Fix handling of TM_PURE.
8921 2013-08-30 Richard Biener <rguenther@suse.de>
8923 PR tree-optimization/58228
8924 * tree-vect-data-refs.c (vect_analyze_data_ref_access): Do not
8925 allow invariant loads in nested loop vectorization.
8927 2013-08-30 Richard Biener <rguenther@suse.de>
8929 PR tree-optimization/58223
8930 * tree-loop-distribution.c (has_anti_dependence): Rename to ...
8931 (has_anti_or_output_dependence): ... this and adjust to also
8932 look for output dependences.
8933 (mark_nodes_having_upstream_mem_writes): Adjust.
8934 (rdg_flag_uses): Likewise.
8936 2013-08-30 Richard Biener <rguenther@suse.de>
8938 PR tree-optimization/58010
8939 * tree-vect-loop.c (vect_create_epilog_for_reduction): Remove
8940 assert that we have a loop-closed PHI.
8942 2013-08-29 Jan Hubicka <jh@suse.cz>
8944 * lto-symtab.c (lto_cgraph_replace_node): Free decl_in_state.
8945 * cgraph.c (cgraph_release_function_body): Free decl_in_state.
8946 * lto-section-in.c (lto_free_function_in_decl_state): New function.
8947 (lto_free_function_in_decl_state_for_node): New function.
8949 2013-08-29 Xinliang David Li <davidxl@google.com>
8951 * loop-unroll.c (report_unroll_peel): Minor message change.
8952 * tree-vect-loop-manip.c (vect_do_peeling_for_alignment):
8953 Emit alignment peeling message with default -fopt-info.
8954 (vect_loop_versioning): Emit loop version info message.
8955 * tree-vectorizer.c (vectorize_loops): Minor message change.
8956 (execute_vect_slp): Ditto.
8958 2013-08-29 Eric Botcazou <ebotcazou@adacore.com>
8960 * cgraphclones.c (cgraph_create_virtual_clone): Compute the DECL_NAME
8961 of the clone from the DECL_NAME of the original function.
8963 2013-08-29 Oleg Endo <olegendo@gcc.gnu.org>
8965 * passes.c (register_pass): Add overload.
8966 * tree-pass.h (register_pass): Forward declare it. Add comment.
8968 2013-08-29 Jan Hubicka <jh@suse.cz>
8970 * lto-streamer-out.c (hash_tree): Stream DECL_FINAL_P,
8971 DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P and TYPE_FINAL_P.
8972 * lto-streamer-in.c (unpack_ts_decl_with_vis_value_fields): Stream
8973 DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
8974 (unpack_ts_type_common_value_fields): Stream TYPE_FINAL_P.
8975 * tree-streamer-out.c (pack_ts_decl_with_vis_value_fields):
8976 Add DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
8977 (pack_ts_type_common_value_fields): Add TYPE_FINAL_P.
8979 2013-08-29 Teresa Johnson <tejohnson@google.com>
8981 * dumpfile.c (dump_loc): Output column number.
8982 * dumpfile.h (OPTGROUP_OTHER): Add and enable under OPTGROUP_ALL.
8983 * doc/invoke.texi: Document optall -fopt-info flag.
8984 * profile.c (read_profile_edge_counts): Use new dump framework.
8985 (compute_branch_probabilities): Ditto.
8986 * passes.c (pass_manager::register_one_dump_file): Use OPTGROUP_OTHER
8987 when pass not in any opt group.
8988 * pass_manager.h (pass_manager::get_pass_profile): New method.
8989 * value-prof.c (check_counter): Use new dump framework.
8990 (check_ic_target): Ditto.
8991 * coverage.c (get_coverage_counts): Ditto.
8992 (coverage_init): Setup new dump framework.
8994 2013-08-29 Richard Biener <rguenther@suse.de>
8996 PR tree-optimization/58246
8997 * tree-ssa-dce.c (mark_aliased_reaching_defs_necessary_1): Properly
8998 handle the dominance check inside a basic-block.
9000 2013-08-29 Richard Biener <rguenther@suse.de>
9003 * tree-ssa-copy.c (may_propagate_copy): Allow propagating
9004 of default defs that appear in abnormal PHI nodes.
9006 2013-08-29 Richard Biener <rguenther@suse.de>
9008 PR tree-optimization/57685
9009 * tree-vrp.c (register_edge_assert_for_1): Recurse only for
9010 single-use operands to avoid exponential complexity.
9012 2013-08-28 Dehao Chen <dehao@google.com>
9014 * ipa-inline.c (edge_badness): Fix integer underflow.
9016 2013-08-28 Uros Bizjak <ubizjak@gmail.com>
9018 * gtm-builtins.def (_ITM_free): Declare leaf.
9020 2013-08-28 Jakub Jelinek <jakub@redhat.com>
9023 * config/i386/i386.md (*tls_global_dynamic_64_largepic): New insn.
9024 (*tls_local_dynamic_base_64_largepic): Likewise.
9025 (tls_global_dynamic_64_<mode>, tls_local_dynamic_base_64_<mode>):
9026 Remove predicate from call operand.
9027 * config/i386/i386.c (ix86_tls_get_addr): For -mcmodel=large -fpic
9028 return sum of pic_offset_table_rtx and UNSPEC_PLTOFF of the symbol.
9030 2013-08-28 Jeff Law <law@redhat.com>
9032 * tree-ssa-threadedge.c (thread_around_empty_block): Remove
9033 checks for the number of predecessors and successors allowed.
9034 * tree-ssa-threadupdate.c (mark_threaded_blocks): Ignore requests
9035 which require copying a joiner block if there is a request which
9036 is a subpath that requires no joiner block copying.
9038 2013-08-28 Jan Hubicka <jh@suse.cz>
9040 * lto-streamer-out.c (DFS_write_tree_body): Drop
9041 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX and BINFO_VPTR_INDEX.
9042 (hash_tree): Do not hash DECL_DEFER_OUTPUT, BINFO_INHERITANCE_CHAIN,
9043 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX, DECL_IN_TEXT_SECTION.
9044 * tree-streamer-in.c (unpack_ts_decl_common_value_fields):
9045 Do not read DECL_ERROR_ISSUED.
9046 (unpack_ts_decl_with_vis_value_fields): Do not read
9048 (lto_input_ts_binfo_tree_pointers): Do not read
9049 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX
9050 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Do not
9051 write DECL_ERROR_ISSUED..
9052 (pack_ts_decl_with_vis_value_fields): Do not write
9054 (write_ts_binfo_tree_pointers): Do not read BINFO_INHERITANCE_CHAIN,
9055 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX.
9056 * print-tree.c (print_node): Do not print DECL_ERROR_ISSUED.
9057 * tree.h (tree_decl_common): Update comment.
9058 (DECL_ERROR_ISSUED): Remove.
9060 2013-08-28 Jakub Jelinek <jakub@redhat.com>
9063 * omp-low.c (copy_var_decl): Copy over TREE_NO_WARNING flag.
9065 2013-08-28 Jan Hubicka <jh@suse.cz>
9067 * builtins.def (free): Declare leaf.
9069 2013-08-27 David Malcolm <dmalcolm@redhat.com>
9072 * configure.ac (gdbinit.in): Add import of gcc/gdbhooks.py.
9073 * configure: Regenerate.
9075 2013-08-27 Martin Jambor <mjambor@suse.cz>
9077 * ipa-prop.h (ipa_pass_through_data): New field type_preserved.
9078 (ipa_ancestor_jf_data): Likewise.
9079 (ipa_get_jf_pass_through_agg_preserved): Fix comment typo.
9080 (ipa_get_jf_pass_through_type_preserved): New function.
9081 (ipa_get_jf_ancestor_agg_preserved): Fix comment typo.
9082 (ipa_get_jf_ancestor_type_preserved): New function.
9083 * ipa-cp.c (ipa_get_jf_pass_through_result): Honor type_preserved flag.
9084 (ipa_get_jf_ancestor_result): Likewise.
9085 (propagate_vals_accross_pass_through): Use
9086 ipa_get_jf_pass_through_result to do all the value mappings.
9087 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Dump the
9088 type_preserved flag.
9089 (ipa_set_jf_cst_copy): New function.
9090 (ipa_set_jf_simple_pass_through): Set the type_preserved flag.
9091 (ipa_set_jf_arith_pass_through): Likewise.
9092 (ipa_set_ancestor_jf): Likewise.
9093 (compute_complex_assign_jump_func): Set type_preserved instead of
9095 (ipa_compute_jump_functions_for_edge): Likewise.
9096 (combine_known_type_and_ancestor_jfs): Honor type_preserved.
9097 (update_jump_functions_after_inlining): Update type_preserved.
9098 Explicitely create jump functions when combining one with pass_through.
9099 (ipa_write_jump_function): Stream the type_preserved flags.
9100 (ipa_read_jump_function): Likewise.
9102 2013-08-27 Jakub Jelinek <jakub@redhat.com>
9103 Aldy Hernandez <aldyh@redhat.com>
9105 * Makefile.in (omp-low.o): Depend on $(TARGET_H).
9106 * cfgloop.h (struct loop): Add safelen, force_vect, simduid.
9107 * function.h (struct function): Add has_force_vect_loops and
9109 * gimple-pretty-print.c (dump_gimple_omp_for): Handle GF_OMP_FOR_KIND*.
9110 * gimple.c (gimple_build_omp_critical): Add KIND argument and
9112 * gimple.def: Update CLAUSES comments.
9113 * gimple.h (enum gf_mask): Add GF_OMP_FOR_KIND_{FOR,SIMD}.
9114 (gimple_build_omp_for): Add argument to prototype.
9115 (gimple_omp_for_kind): New.
9116 (gimple_omp_for_set_kind): New.
9117 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_LINEAR to
9118 GOVD_DATA_SHARE_CLASS.
9119 (enum omp_region_type): Add ORT_SIMD.
9120 (gimple_add_tmp_var): Handle ORT_SIMD.
9121 (gimplify_var_or_parm_decl): Same.
9122 (is_gimple_stmt): Same.
9123 (omp_firstprivatize_variable): Same.
9124 (omp_add_variable): Only use splay_tree_insert if lookup failed.
9125 (omp_notice_variable): Handle ORT_SIMD.
9126 (omp_is_private): Add SIMD argument and handle it as well as ORT_SIMD.
9127 (omp_check_private): Handle ORT_SIMD.
9128 (gimplify_scan_omp_clauses): Handle OMP_CLAUSE_LINEAR and
9130 (gimplify_adjust_omp_clauses_1): Handle GOVD_LINEAR.
9131 Handle OMP_CLAUSE_LASTPRIVATE.
9132 (gimplify_adjust_omp_clauses): Handle OMP_CLAUSE_LINEAR and
9134 (gimplify_omp_for): Handle OMP_SIMD and OMP_CLAUSE_LINEAR.
9135 (gimplify_expr): Handle OMP_SIMD.
9136 * internal-fn.c (expand_GOMP_SIMD_LANE): New.
9137 (expand_GOMP_SIMD_VF): New.
9138 (expand_GOMP_SIMD_LAST_LANE): New.
9139 * internal-fn.def (GOMP_SIMD_LANE): New.
9140 (GOMP_SIMD_VF): New.
9141 (GOMP_SIMD_LAST_LANE): New.
9142 * omp-low.c: Include target.h.
9143 (extract_omp_for_data): Handle OMP_SIMD, OMP_CLAUSE_LINEAR,
9145 (check_omp_nesting_restrictions): Same.
9147 (lower_rec_simd_input_clauses): New.
9148 (lower_rec_input_clauses): Handle OMP_SIMD, GF_OMP_FOR_KIND_SIMD,
9150 (lower_lastprivate_clauses): Handle OMP_CLAUSE_LINEAR,
9151 GF_OMP_FOR_KIND_SIMD, OMP_SIMD.
9152 (expand_omp_build_assign): New.
9153 (expand_omp_for_init_counts): New.
9154 (expand_omp_for_init_vars): New.
9155 (extract_omp_for_update_vars): New.
9156 (expand_omp_for_generic): Use expand_omp_for_{init,update}_vars
9157 and rewrite accordingly.
9158 (expand_omp_simd): New.
9159 (expand_omp_for): Use expand_omp_simd.
9160 (lower_omp_for_lastprivate): Unshare vinit when appropriate.
9161 (lower_omp_for): Do not lower the body.
9162 * tree-data-ref (get_references_in_stmt): Allow IFN_GOMP_SIMD_LANE
9164 * tree-flow.h (find_omp_clause): Remove prototype.
9165 * tree-if-conv.c (main_tree_if_conversion): Run if doing if conversion,
9166 forcing vectorization of the loop, or if flag_tree_vectorize.
9167 (gate_tree_if_conversion): Similarly.
9168 * tree-inline.c (remap_gimple_stmt): Pass for kind argument to
9169 gimple_build_omp_for.
9170 (copy_cfg_body): set has_force_vect_loops and has_simduid_loops.
9171 * tree-parloops (create_parallel_loop): Pass kind argument to
9172 gimple_build_omp_for.
9173 * tree-pretty-print.c (dump_omp_clause): Add cases for
9174 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_LINEAR, OMP_CLAUSE_SAFELEN,
9175 OMP_CLAUSE__SIMDUID_.
9176 (dump_generic_node): Handle OMP_SIMD.
9177 * tree-ssa-ccp.c (likely_value): Handle IFN_GOMP_SIMD*.
9178 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Do not
9179 unroll OMP_SIMD loops here.
9180 * tree-ssa-loop.c (gate_tree_vectorize): Run if has_force_vect_loops.
9181 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Handle
9183 (vect_analyze_data_refs): Handle simd loops.
9184 * tree-vect-loop.c (vectorizable_live_operation): Handle
9186 * tree-vect-stmts.c (vectorizable_call): Handle IFN_GOMP_SIMD_LANE.
9187 (vectorizable_store): Handle STMT_VINFO_SIMD_LANE_ACCESS_P.
9188 (vectorizable_load): Same.
9189 * tree-vectorizer.c: Include hash-table.h and tree-ssa-propagate.h.
9190 (struct simduid_to_vf): New.
9191 (simduid_to_vf::hash): New.
9192 (simduid_to-vf::equal): New.
9193 (struct simd_array_to_simduid): New.
9194 (simd_array_to_simduid::hash): New.
9195 (simd_array_to_simduid::equal): New.
9196 (adjust_simduid_builtins): New.
9197 (struct note_simd_array_uses_struct): New.
9198 (note_simd_array_uses_cb): New.
9199 (note_simd_array_uses): New.
9200 (vectorize_loops): Handle simd hints and adjust simd builtins
9202 * tree-vectorizer.h (struct _stmt_vec_info): Add
9203 simd_lane_access_p field.
9204 (STMT_VINFO_SIMD_LANE_ACCESS_P): New macro.
9205 * tree.c (omp_clause_num_ops): Add entries for OMP_CLAUSE_LINEAR,
9206 OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_UNIFORM.
9207 (omp_clause_code_name): Same.
9208 (walk_tree_1): Handle OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN,
9209 OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_LINEAR.
9210 * tree.def (OMP_SIMD): New entry.
9211 * tree.h (enum omp_clause_code): Add entries for OMP_CLAUSE_LINEAR,
9212 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_.
9213 (OMP_CLAUSE_DECL): Adjust range for new clauses.
9214 (OMP_CLAUSE_LINEAR_NO_COPYIN): New.
9215 (OMP_CLAUSE_LINEAR_NO_COPYOUT): New.
9216 (OMP_CLAUSE_LINEAR_STEP): New.
9217 (OMP_CLAUSE_SAFELEN_EXPR): New.
9218 (OMP_CLAUSE__SIMDUID__DECL): New.
9219 (find_omp_clause): New prototype.
9221 2013-08-27 H.J. Lu <hongjiu.lu@intel.com>
9223 * config/i386/driver-i386.c (host_detect_local_cpu): Update
9224 Haswell processor detection.
9226 2013-08-27 Christian Widmer <shadow@umbrox.de>
9229 * config/i386/driver-i386.c (host_detect_local_cpu): Add detection
9230 of Ivy Bridge and Haswell processors. Assume core-avx2 for unknown
9231 AVX2 capable processors.
9233 2013-08-27 Tejas Belagod <tejas.belagod@arm.com>
9235 * config/aarch64/arm_neon.h: Replace all inline asm implementations
9236 of vget_low_* with implementations in terms of other intrinsics.
9238 2013-08-27 Marc Glisse <marc.glisse@inria.fr>
9241 * doc/extend.texi (__builtin_isinf_sign): Restrict the return
9242 values to -1, 0 and 1.
9244 2013-08-27 Vidya Praveen <vidyapraveen@arm.com>
9246 * config/aarch64/aarch64.md (unspec): Add UNSPEC_SISD_SSHL,
9247 UNSPEC_SISD_USHL, UNSPEC_USHL_2S, UNSPEC_SSHL_2S, UNSPEC_SISD_NEG.
9248 (<optab><mode>3_insn): Remove.
9249 (aarch64_ashl_sisd_or_int_<mode>3): New Pattern.
9250 (aarch64_lshr_sisd_or_int_<mode>3): Likewise.
9251 (aarch64_ashr_sisd_or_int_<mode>3): Likewise.
9252 (define_split for aarch64_lshr_sisd_or_int_di3): Likewise.
9253 (define_split for aarch64_lshr_sisd_or_int_si3): Likewise.
9254 (define_split for aarch64_ashr_sisd_or_int_di3): Likewise.
9255 (define_split for aarch64_ashr_sisd_or_int_si3): Likewise.
9256 (aarch64_sisd_ushl, aarch64_sisd_sshl): Likewise.
9257 (aarch64_ushl_2s, aarch64_sshl_2s, aarch64_sisd_neg_qi): Likewise.
9258 (ror<mode>3_insn): Likewise.
9259 * config/aarch64/predicates.md (aarch64_simd_register): New.
9261 2013-08-27 Richard Biener <rguenther@suse.de>
9263 PR tree-optimization/57521
9264 * tree-if-conv.c (if_convertible_bb_p): Verify that at least
9265 one edge is non-critical.
9266 (find_phi_replacement_condition): Make sure to use a non-critical
9267 edge. Cleanup and remove old bug workarounds.
9268 (bb_postdominates_preds): Remove.
9269 (if_convertible_loop_p_1): Do not compute post-dominators.
9270 (combine_blocks): Do not free post-dominators.
9271 (main_tree_if_conversion): Likewise.
9272 (pass_data_if_conversion): Add TODO_verify_ssa.
9274 2013-08-27 DJ Delorie <dj@redhat.com>
9276 * config/i386/djgpp.h (ASM_DECLARE_FUNCTION_NAME): New.
9278 2013-08-27 Yufeng Zhang <yufeng.zhang@arm.com>
9280 * function.c (assign_parm_find_data_types): Set passed_mode and
9281 nominal_mode to the TYPE_MODE of nominal_type for the built
9282 pointer type in case of the struct-pass-by-reference.
9284 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
9286 * config/avr/avr-stdint.h (INT16_TYPE): Change default to "int".
9287 (UINT16_TYPE): Change default to "unsigned int".
9289 * config/avr/avr.opt (mfract-convert-truncate): New option.
9290 * config/avr/avr.c (avr_out_fract): Unless TARGET_FRACT_CONV_TRUNC
9291 is set, round negative fractional integers according to n1169
9292 when converting to integer types.
9294 2013-08-26 Jan Hubicka <jh@suse.cz>
9296 * cgraph.c (cgraph_propagate_frequency): Do not assume that virtual
9297 methods can not be called indirectly when their address is not taken.
9299 2013-08-26 Jan Hubicka <jh@suse.cz>
9301 * gimple-fold.c (gimple_get_virt_method_for_binfo): Use
9304 2013-08-26 Jan Hubicka <jh@suse.cz>
9306 * ipa.c (comdat_can_be_unshared_p_1): C++ constructors and destructors
9309 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
9311 * reload.c (find_valid_class): Allow classes that do not include
9312 FIRST_PSEUDO_REGISTER - 1.
9314 2013-08-26 Jan Hubicka <jh@suse.cz>
9316 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix formatting;
9317 fix edge count/frequency when speculation failed; fix type check
9318 for the direct call.
9320 2013-08-26 Jan Hubicka <jh@suse.cz>
9322 * ipa-prop.c (ipa_print_node_params): Do not ICE during WPA.
9324 2013-08-26 Jan Hubicka <jh@suse.cz>
9326 * ipa-inline-transform.c (inline_transform): Be ready for basic block
9327 to be changed by edge redirection.
9329 2013-08-26 Jan Hubicka <jh@suse.cz>
9331 * cgraph.c (cgraph_speculative_call_info): Fix parameter order and
9332 formating; add sanity check.
9333 (cgraph_resolve_speculation): Add FIXME about scaling profiles.
9334 (cgraph_redirect_edge_call_stmt_to_callee): Fix ICE in debug dump.
9335 * ipa-inline.c (heap_edge_removal_hook): Reset node growth cache.
9336 (resolve_noninline_speculation): Update callee keys, too.
9338 2013-08-26 Jan Hubicka <jh@suse.cz>
9340 * tree.h (tree_decl_with_vis): Add cxx_constructor, cxx_destructor.
9341 (DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P): New macros.
9343 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
9345 * config/i386/i386.c (x86_64_elf_select_section): Put ATTRIBUTE_UNUSED
9348 2013-08-26 Uros Bizjak <ubizjak@gmail.com>
9350 * config/i386/i386.c (ix86_debug_options): Remove prototype.
9351 (x86_64_elf_select_section): Ditto.
9352 (ix86_handle_tm_regparm_attribute): Remove ATTRIBUTE_UNUSED on used
9354 (ix86_pass_by_reference): Ditto.
9355 (output_set_got): Ditto.
9356 (ix86_unary_operator_ok): Ditto.
9357 (ix86_expand_builtin): Ditto.
9359 2013-08-23 Jan Hubicka <jh@suse.cz>
9361 * cgraph.c (cgraph_turn_edge_to_speculative): Fix debug output.
9363 2013-08-23 Jan Hubicka <jh@suse.cz>
9365 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
9366 (tree_decl_with_vis): Add FINAL field.
9368 2013-08-23 Jeff Law <law@redhat.com>
9370 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
9371 the debugging dump when the expression is fully redundant.
9373 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
9375 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
9376 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
9377 * pretty-print.c (pp_formatted_text_data): Likewise.
9378 (pp_write_text_to_stream): Likewise.
9379 (pp_write_text_as_dot_label_to_stream): Likewise.
9380 (pp_append_r): Likewise.
9381 (pp_format): Likewise.
9382 (pp_flush): Likewise.
9383 (pp_clear_output_area): Likewise.
9384 (pp_append_text): Likewise.
9385 (pp_formatted_text): Likewise.
9386 (pp_remaining_character_count_for_line): Likewise.
9387 (pp_newline): Likewise.
9388 (pp_character): Likewise.
9389 (output_buffer::~output_buffer): Define.
9390 (pretty_printer::~pretty_printer): Destruct output buffer.
9391 * pretty-print.h (output_buffer::~output_buffer): Declare.
9392 (pretty_printer::~pretty_printer): Declare virtual.
9394 2013-08-24 Marc Glisse <marc.glisse@inria.fr>
9397 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
9398 HOST_WIDE_INT_M1U): New macros.
9399 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
9400 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
9401 unsigned -1 for lshift.
9402 * cse.c (cse_insn): Likewise.
9403 * double-int.c (rshift_double, lshift_double): Likewise.
9404 * builtins.c (fold_builtin_bitop): Likewise.
9405 * combine.c (force_to_mode): Likewise.
9406 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
9407 * simplify-rtx.c (simplify_const_unary_operation,
9408 simplify_const_binary_operation): Likewise.
9409 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
9410 check_va_list_escapes): Likewise.
9411 * rtlanal.c (nonzero_bits1): Likewise.
9412 * expmed.c (expand_smod_pow2): Likewise.
9413 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
9415 2013-08-23 Jan Hubicka <jh@suse.cz>
9417 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
9418 as having address taken.
9420 2013-08-23 Jan Hubicka <jh@suse.cz>
9422 * ipa-utils.h (method_class_type): Declare.
9423 * ipa-devirt.c (method_class_type): Export.
9425 * cgraphunit.c (analyze_functions): Do basic devirtualization;
9426 do not walk base classes of anonymous types.
9428 2013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
9430 PR rtl-optimization/58220
9432 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
9433 handle SEQUENCE insns properly.
9435 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
9437 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
9439 (pp_newline_and_indent): Likewise.
9440 (pp_separate_with): Likewise.
9441 * pretty-print.c (pp_newline_and_flush): Define.
9442 (pp_newline_and_indent): Likewise.
9443 (pp_separate_with): Likewise.
9445 2013-08-23 Jakub Jelinek <jakub@redhat.com>
9448 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
9449 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
9451 2013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
9453 * config/i386/predicates.md (ext_sse_reg_operand): New.
9454 * config/i386/i386.md (*movti_internal): Use
9455 predicate to determine if EVEX is needed.
9456 (*movsi_internal): Ditto.
9457 (*movdf_internal): Ditto.
9458 (*movsf_internal): Ditto.
9459 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
9461 2013-08-23 Jakub Jelinek <jakub@redhat.com>
9463 PR tree-optimization/58209
9464 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
9465 (find_tail_calls): Give up for pointer result types if m is non-NULL.
9466 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
9467 emit POINTER_PLUS_EXPR.
9468 (create_tailcall_accumulator): For pointer result type accumulate in
9471 2013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
9473 * configure.ac: Add backslashes missing from the last change.
9474 * configure: Regenerate.
9476 2013-08-22 Jan Hubicka <jh@susue.cz>
9478 * ipa.c (function_and_variable_visibility): First remember function
9479 was global and then make it local.
9481 2013-08-22 Julian Brown <julian@codesourcery.com>
9483 * configure.ac: Add aarch64 to list of arches which use "nop" in
9485 * configure: Regenerate.
9487 2013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9489 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
9490 gnu_libc_has_function.
9491 * config/s390/tpf.h: Likewise.
9493 2013-08-22 Jan Hubicka <jh@susue.cz>
9495 * timevar.c (validate_phases): Add cast.
9497 2013-08-22 Jan Hubicka <jh@susue.cz>
9499 * timevar.c (validate_phases): Use size_t for memory.
9500 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
9502 2013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
9504 * pretty-print.h (output_buffer::output_buffer): Declare.
9505 (pretty_printer::pretty_printer): Likewise.
9506 (pp_construct): Remove.
9507 * pretty-print.c (output_buffer::output_buffer): Define.
9508 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
9509 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
9510 (print_gimple_expr): Likewise.
9511 (print_gimple_seq): Likewise.
9512 (gimple_dump_bb): Likewise.
9513 * sched-vis.c (dump_value_slim): Likewise.
9514 (dump_insn_slim): Likewise.
9515 (dump_rtl_slim): Likewise.
9516 (str_pattern_slim): Likewise.
9517 * tree-mudflap.c (mf_varname_tree): Likewise.
9518 * graph.c (print_graph_cfg): Likewise.
9519 (start_graph_dump): Likewise.
9520 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
9522 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
9523 pretty printer initialization.
9524 * coretypes.h (diagnostic_context): Remove superflous type alias
9526 (pretty_printer): Likewise. Declare directly as a class.
9527 (pretty_print_info): Remove declaration as class.
9528 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
9529 and pp_clear_output_area.
9530 (asan_add_global): Likewise.
9532 2013-08-22 Jan Hubicka <jh@suse.cz>
9534 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
9535 * ipa-utils.h (update_type_inheritance_graph): Declare.
9536 (possible_polymorphic_call_target_p): Declare.
9537 (possible_polymorphic_call_target_p): New.
9538 * ipa-devirt.c: Update toplevel comments.
9539 (cached_polymorphic_call_targets): Move up.
9540 (odr_type_d): Move ID down.
9541 (polymorphic_type_binfo_p): Update comment.
9542 (odr_hasher::remove): Likewise;
9543 (get_odr_type): Set anonymous_namespace.
9544 (dump_odr_type): Dump it.
9545 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
9546 (maybe_record_node): Record node in cached_polymorphic_call_targets.
9547 (record_binfo): Add comment.
9548 (free_polymorphic_call_targets_hash): Do not ICE when cache is not
9550 (devirt_node_removal_hook): Do not iCE when cache is freed.
9551 (possible_polymorphic_call_target_p): New predicate.
9552 (update_type_inheritance_graph): New function.
9554 2013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
9555 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
9556 Sergey Lega <sergey.s.lega@intel.com>
9557 Anna Tikhonova <anna.tikhonova@intel.com>
9558 Ilya Tocar <ilya.tocar@intel.com>
9559 Andrey Turetskiy <andrey.turetskiy@intel.com>
9560 Ilya Verbin <ilya.verbin@intel.com>
9561 Kirill Yukhin <kirill.yukhin@intel.com>
9562 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
9564 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
9565 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
9566 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
9567 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
9568 (OPTION_MASK_ISA_AVX2_UNSET): Update.
9569 (OPTION_MASK_ISA_AVX512F_UNSET): New.
9570 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
9571 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
9572 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
9573 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
9574 OPT_mavx512pf, OPT_mavx512er cases.
9575 * config/i386/constraints.md (v): New constraint.
9576 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
9577 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
9578 (bit_AVX512CD): New.
9579 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
9580 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
9581 * config/i386/i386-c.c (ix86_target_macros_internal):
9582 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
9584 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
9585 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
9586 * config/i386/i386.c (regclass_map, dbx_register_map)
9587 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
9588 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
9589 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
9591 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
9592 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
9593 -mavx512pf options. Fix formatting.
9594 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
9595 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
9596 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
9597 -mavx512cd, -mavx512pf options.
9598 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
9599 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
9600 (ix86_preferred_output_reload_class): Replace SSE_REGS with
9602 (ix86_hard_regno_mode_ok): Support 512-bit registers.
9603 (ix86_set_reg_reg_cost): Ditto.
9604 (x86_order_regs_for_local_alloc): Ditto.
9605 (MAX_VECT_LEN): Extend to 64-byte.
9606 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
9607 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
9608 (TARGET_AVX512ER, TARGET_AVX512CD): New.
9609 (BIGGEST_ALIGNMENT): Extend to 512-bits.
9610 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
9611 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
9612 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
9613 (SSE_REG_MODE_P): Support new modes.
9614 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
9615 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
9616 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
9617 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
9618 (REG_CLASS_CONTENTS): Add new registers.
9619 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
9620 (EXT_REX_SSE_REGNO_P): New.
9621 (HI_REGISTER_NAMES): Add new registers.
9622 * config/i386/i386.md: Define constants for new registers.
9623 (mode): Add new 512-bit modes.
9624 (prefix): Support evex prefix.
9625 (isa): Support avx512f, noavx512f, fma_avx512f.
9626 (ssemodesuffix): Add new 512-bit modes.
9628 (*movxi_internal_avx512f): Ditto.
9629 (*movdi_internal): Replace constraint "x" with the new constraint "v".
9631 (*movsi_internal): Likewise.
9632 (*movdf_internal): Likewise.
9633 (*movsf_internal): Likewise.
9634 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
9635 (<code><mode>3): Likewise.
9636 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd):
9638 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
9639 with the new constraint "v".
9640 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
9642 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
9643 with the new constraint "v".
9644 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
9645 (<sse2>_storedqu<avxsizesuffix>): Likewise.
9646 (*<plusminus_insn><mode>3): Likewise.
9647 (<sse>_vm<plusminus_insn><mode>3): Likewise.
9648 (*mul<mode>3): Likewise.
9649 (<sse>_vmmul<mode>3): Likewise.
9650 (<sse>_div<mode>3): Likewise.
9651 (<sse>_vmdiv<mode>3): Likewise.
9652 (<sse>_sqrt<mode>2): Likewise.
9653 (<sse>_vmsqrt<mode>2): Likewise.
9654 (*<code><mode>3_finite): Likewise.
9655 (*<code><mode>3) <smaxmin>: Likewise.
9656 (<sse>_vm<code><mode>3): Likewise.
9657 (*<code><mode>3) <any_logic>: Likewise.
9658 (*fma_fmadd_<mode>): Likewise.
9659 (*fma_fmsub_<mode>): Likewise.
9660 (*fma_fnmadd_<mode>): Likewise.
9661 (*fma_fnmsub_<mode>): Likewise.
9662 (*fma_fmaddsub_<mode>): Likewise.
9663 (*fma_fmsubadd_<mode>): Likewise.
9664 (*fmai_fmadd_<mode>): Likewise.
9665 (*fmai_fmsub_<mode>): Likewise.
9666 (*fmai_fnmadd_<mode>): Likewise.
9667 (*fmai_fnmsub_<mode>): Likewise.
9668 (sse_cvtsi2ss): Likewise.
9669 (sse_cvtsi2ssq): Likewise.
9670 (sse_cvtss2si): Likewise.
9671 (sse_cvtss2si_2): Likewise.
9672 (sse_cvtss2siq): Likewise.
9673 (sse_cvtss2siq_2): Likewise.
9674 (sse_cvttss2si): Likewise.
9675 (sse_cvtss2siq_2): Likewise.
9676 (float<sseintvecmodelower><mode>2): Likewise.
9677 (sse2_cvtsd2si_2): Likewise.
9678 (sse2_cvtsd2siq_2): Likewise.
9679 (*<plusminus_insn><mode>3): Likewise.
9680 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
9681 (*<sse4_1_avx2>_mul<mode>3): Likewise.
9682 (ashr<mode>3): Likewise.
9683 (<shift_insn><mode>3): Likewise.
9684 (avx2_<code><mode>3): Likewise.
9685 (*avx2_<code><mode>3): Likewise.
9686 (*andnot<mode>3): Likewise.
9687 (*<code><mode>3) <any_logic>: Likewise.
9688 (abs<mode>2): Likewise.
9689 (avx2_permvar<mode>): Likewise.
9690 (avx2_perm<mode>_1): Likewise.
9691 (*avx_vpermilp<mode>): Likewise.
9692 (avx_vpermilvar<mode>3): Likewise.
9693 (avx2_ashrv<mode>): Likewise.
9694 (avx2_<shift_insn>v<mode>): Likewise.
9695 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
9697 * doc/rtl.texi: Document XImode.
9699 2013-08-21 Jeff Law <law@redhat.com>
9701 * tree-flow.h (register_jump_thread): Pass vector of edges
9702 instead of each important edge.
9703 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
9704 thread path into a vector and pass that to register_jump_thread.
9705 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
9706 passed in edge vector to the current 3-edge form.
9709 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
9711 * dce.c (fini_dce): Call df_analyze again just in case
9712 delete_unmarked_insns removed anything.
9714 2013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
9716 * reload.h (struct reg_equivs): Rename to ..
9717 (struct reg_equivs_s): .. this.
9719 2013-08-20 Martin Liska <marxin.liska@gmail.com>
9721 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
9723 2013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
9725 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest comment.
9727 2013-08-21 Jeff Law <law@redhat.com>
9729 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
9730 simplify assignments too. If the RHS collapses to a singleton
9731 range, then return the value for the range.
9733 2013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
9735 * config/i386/sse.md (V16): Rename to...
9737 (mov<mode>): Update iterator name.
9738 (*mov<mode>_internal): Ditto.
9739 (push<mode>1): Ditto.
9740 (movmisalign<mode>): Ditto.
9742 2013-08-20 Jan Hubicka <jh@suse.cz>
9745 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
9746 entry for direct edges.
9747 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
9749 2013-08-20 David Malcolm <dmalcolm@redhat.com>
9751 Revert my last two changes, r201865 and r201864:
9754 2013-08-20 David Malcolm <dmalcolm@redhat.com>
9756 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
9757 instances can own GC refs.
9759 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
9760 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
9761 (gcc::context::gt_pch_nx): Likewise.
9762 (gcc::context::gt_pch_nx): Likewise.
9763 * ggc.h (gt_ggc_mx <T>): New.
9764 (gt_pch_nx_with_op <T>): New.
9765 (gt_pch_nx <T>): New.
9766 * passes.c (opt_pass::gt_ggc_mx): New.
9767 (opt_pass::gt_pch_nx): New.
9768 (opt_pass::gt_pch_nx_with_op): New.
9769 (pass_manager::gt_ggc_mx): New.
9770 (pass_manager::gt_pch_nx): New.
9771 (pass_manager::gt_pch_nx_with_op): New.
9772 (pass_manager::operator new): Use
9773 ggc_internal_cleared_alloc_stat rather than xcalloc.
9774 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
9775 (pass_manager::gt_ggc_mx): New.
9776 (pass_manager::gt_pch_nx): New.
9777 (pass_manager::gt_pch_nx_with_op): New.
9778 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
9779 (opt_pass::operator new): New.
9780 (opt_pass::gt_ggc_mx): New.
9781 (opt_pass::gt_pch_nx): New.
9782 (opt_pass::gt_pch_nx_with_op): New.
9785 2013-08-20 David Malcolm <dmalcolm@redhat.com>
9787 * Makefile.in (GTFILES): Add context.h.
9788 * context.c (gcc::context::operator new): New.
9789 (gcc::context::gt_ggc_mx): New.
9790 (gcc::context::gt_pch_nx): New.
9791 (gcc::context::gt_pch_nx): New.
9792 * context.h (gcc::context): Add GTY((user)) marking.
9793 (gcc::context::operator new): New.
9794 (gcc::context::gt_ggc_mx): New.
9795 (gcc::context::gt_pch_nx): New.
9796 (gcc::context::gt_pch_nx): New.
9797 (g): Add GTY marking.
9798 (gt_ggc_mx (gcc::context *)): New.
9799 (gt_pch_nx (gcc::context *)): New.
9800 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
9801 void *cookie)): New.
9802 * gengtype.c (open_base_files) <ifiles>: Add context.h.
9804 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
9806 * dce.c (fini_dce): Call df_analyze again just in case
9807 delete_unmarked_insns removed anything.
9809 2013-08-20 Teresa Johnson <tejohnson@google.com>
9811 PR rtl-optimizations/57451
9812 * final.c (reemit_insn_block_notes): Prevent lexical blocks
9813 from crossing split section boundaries.
9815 2013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
9817 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
9818 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
9819 with MULTLIB_DEFAULTS.
9821 2013-08-20 Nick Clifton <nickc@redhat.com>
9823 * target.def (narrow_volatile_bitfield): Note that the default
9824 value is false, not !TARGET_STRICT_ALIGN.
9825 * doc/tm.texi: Regenerate.
9827 2013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
9829 Fix LIB_SPEC for systems without libpthread.
9831 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
9832 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
9834 * config/i386/linux-common.h: Likewise.
9835 * config/mips/linux-common.h: Likewise.
9837 2013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
9839 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
9842 2013-08-20 David Malcolm <dmalcolm@redhat.com>
9844 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
9845 instances can own GC refs.
9847 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
9848 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
9849 (gcc::context::gt_pch_nx): Likewise.
9850 (gcc::context::gt_pch_nx): Likewise.
9851 * ggc.h (gt_ggc_mx <T>): New.
9852 (gt_pch_nx_with_op <T>): New.
9853 (gt_pch_nx <T>): New.
9854 * passes.c (opt_pass::gt_ggc_mx): New.
9855 (opt_pass::gt_pch_nx): New.
9856 (opt_pass::gt_pch_nx_with_op): New.
9857 (pass_manager::gt_ggc_mx): New.
9858 (pass_manager::gt_pch_nx): New.
9859 (pass_manager::gt_pch_nx_with_op): New.
9860 (pass_manager::operator new): Use
9861 ggc_internal_cleared_alloc_stat rather than xcalloc.
9862 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
9863 (pass_manager::gt_ggc_mx): New.
9864 (pass_manager::gt_pch_nx): New.
9865 (pass_manager::gt_pch_nx_with_op): New.
9866 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
9867 (opt_pass::operator new): New.
9868 (opt_pass::gt_ggc_mx): New.
9869 (opt_pass::gt_pch_nx): New.
9870 (opt_pass::gt_pch_nx_with_op): New.
9872 2013-08-20 David Malcolm <dmalcolm@redhat.com>
9874 * Makefile.in (GTFILES): Add context.h.
9875 * context.c (gcc::context::operator new): New.
9876 (gcc::context::gt_ggc_mx): New.
9877 (gcc::context::gt_pch_nx): New.
9878 (gcc::context::gt_pch_nx): New.
9879 * context.h (gcc::context): Add GTY((user)) marking.
9880 (gcc::context::operator new): New.
9881 (gcc::context::gt_ggc_mx): New.
9882 (gcc::context::gt_pch_nx): New.
9883 (gcc::context::gt_pch_nx): New.
9884 (g): Add GTY marking.
9885 (gt_ggc_mx (gcc::context *)): New.
9886 (gt_pch_nx (gcc::context *)): New.
9887 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
9888 void *cookie)): New.
9889 * gengtype.c (open_base_files) <ifiles>: Add context.h.
9891 2013-08-20 Alan Modra <amodra@gmail.com>
9894 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
9895 (rs6000_emit_epilogue): Likewise.
9897 2013-08-19 Dehao Chen <dehao@google.com>
9899 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
9901 2013-08-19 Peter Bergner <bergner@vnet.ibm.com>
9902 Jakub Jelinek <jakub@redhat.com>
9904 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
9905 (BUILT_IN_FABSD64): Likewise.
9906 (BUILT_IN_FABSD128): Likewise.
9907 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
9908 (fold_builtin_1): Likewise.
9909 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping
9910 destination and source operands.
9911 (*abstd2_fpr): Likewise.
9912 (*nabstd2_fpr): Likewise.
9914 2013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
9916 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
9919 2013-08-19 Aldy Hernandez <aldyh@redhat.com>
9921 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
9924 2013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
9926 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
9927 * builtins.c (default_libc_has_function): New.
9928 (gnu_libc_has_function): Ditto.
9929 (no_c99_libc_has_function): Ditto.
9930 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
9931 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
9932 (fold_builtin_sincos): Likewise.
9933 (fold_builtin_cexp): Likewise.
9934 * builtins.def (DEF_C94_BUILTIN): Likewise.
9935 (DEF_C99_BUILTIN): Likewise.
9936 (DEF_C99_C90RES_BUILTIN): Likewise.
9937 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
9938 definitions to using this define.
9939 * config/darwin-protos.h (darwin_libc_has_function): New.
9940 * config/darwin.c (darwin_libc_has_function): Ditto.
9941 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
9942 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
9943 * config/darwin.h: Ditto.
9944 * config/elfos.h: Ditto.
9945 * config/freebsd.h: Ditto.
9946 * config/i386/cygming.h: Ditto.
9947 * config/i386/djgpp.h: Ditto.
9948 * config/i386/i386-interix.h: Ditto.
9949 * config/microblaze/microblaze.h: Ditto.
9950 * config/mmix/mmix.h: Ditto.
9951 * config/gnu-user.h: Ditto.
9952 * config/ia64/hpux.h: Ditto.
9953 * config/pa/pa-hpux.h: Ditto.
9954 * config/pdp11/pdp11.h: Ditto.
9955 * config/picochip/picochip.h: Ditto.
9956 * config/linux.h: Ditto.
9957 * config/netbsd.h: Ditto.
9958 * config/openbsd.h: Ditto.
9959 * config/rs6000/aix43.h: Ditto.
9960 * config/rs6000/aix51.h: Ditto.
9961 * config/rs6000/aix52.h: Ditto.
9962 * config/rs6000/aix53.h: Ditto.
9963 * config/rs6000/aix61.h: Ditto.
9964 * config/rs6000/darwin.h: Ditto.
9965 * config/rs6000/linux.h: Ditto.
9966 * config/rs6000/linux64.h: Ditto.
9967 * config/s390/tpf.h: Ditto.
9968 * config/sol2-10.h: Ditto.
9969 * config/sol2.h: Ditto.
9970 * config/vms/vms.h: Ditto.
9971 * config/vxworks.h: Ditto.
9972 * config/linux-android.c (linux_android_libc_has_function):
9973 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
9974 * config/linux-protos.h (linux_android_libc_has_function):
9976 * config/i386/i386.c (ix86_libc_has_function): New.
9977 * config/i386/i386-protos.h
9978 (ix86_libc_has_function): New declaration.
9979 * config/i386/i386.md
9980 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
9981 ("isinf<mode>2): Likewise.
9982 * convert.c (convert_to_integer): Using new target hook
9983 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
9984 TARGET_C99_FUNCTIONS.
9985 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
9986 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
9987 * coretypes.h (function_class): New enum for different
9988 classes of functions.
9989 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
9990 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
9991 (TARGET_HAS_SINCOS): Likewise.
9992 (TARGET_LIBC_HAS_FUNCTION): New.
9993 * doc/tm.texi: Regenerated.
9994 * targhooks.h (default_libc_has_function): New declaration.
9995 (no_c99_libc_has_function): Ditto.
9996 (gnu_libc_has_function): Ditto.
9997 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
9998 and TARGET_HAS_SINCOS.
10000 2013-08-18 Jan Hubicka <jh@suse.cz>
10002 * Makeifle-in (ipa-devirt.o): New.
10003 (GTFILES): Add ipa-utils.h and ipa-devirt.c
10004 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
10005 (analyze_functions): Look into possible targets of polymorphic call.
10006 * dumpfile.c (dump_files): Add type-inheritance dump.
10007 * dumpfile.h (TDI_inheritance): New.
10008 * ipa-devirt.c: New file.
10009 * ipa-utils.h (odr_type_d): Forward declare.
10010 (odr_type): New type.
10011 (build_type_inheritance_graph): Declare.
10012 (possible_polymorphic_call_targets): Declare and introduce inline
10013 variant when only edge is pased.
10014 (dump_possible_polymorphic_call_targets): Likewise.
10015 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
10016 * tree.c (type_in_anonymous_namespace_p): Break out from ...
10017 (types_same_for_odr): ... here.
10018 * tree.h (type_in_anonymous_namespace_p): Declare.
10020 2013-08-18 Jakub Jelinek <jakub@redhat.com>
10022 PR tree-optimization/58006
10023 * tree-parloops.c (take_address_of): Don't ICE if get_name
10025 (eliminate_local_variables_stmt): Remove clobber stmts.
10027 2013-08-18 Eric Botcazou <ebotcazou@adacore.com>
10029 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
10030 error message is issued for an alias to undefined symbol.
10032 2013-08-18 Jan Hubicka <jh@suse.cz>
10034 * cgraph.c (cgraph_create_indirect_edge): Discover
10035 polymorphic calls and record basic info into indirect_info.
10036 * gimple-fold.c (gimple_fold_call): When doing BINFO based
10037 devirtualization, ignore objc function calls.
10038 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
10039 call with no parm index info.
10040 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
10041 * tree.c (virtual_method_call_p): New function.
10042 * tree.h (virtual_method_call_p): Declare.
10044 2013-08-16 Jan Hubicka <jh@suse.cz>
10046 PR middle-end/58179
10047 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
10049 2013-08-16 David Edelsohn <dje.gcc@gmail.com>
10051 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
10054 2013-08-16 David Malcolm <dmalcolm@redhat.com>
10056 * gengtype.c (type_for_name): Add special-case support for
10057 locating types within the "gcc::" namespace.
10058 (open_base_files): Emit a "using namespace gcc" directive.
10060 2013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
10063 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
10064 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
10066 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
10067 array instead of each individual operand as a separate argument.
10068 (emit_fusion_gpr_load): Likewise.
10069 (expand_fusion_gpr_load): Add new function declaration.
10071 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
10072 signature to have the operands passed as an array, instead of as
10073 separate arguments. Allow ZERO_EXTEND to be in the memory
10074 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
10075 depend on the register live/dead flags when peepholes are run.
10076 (expand_fusion_gpr_load): New function to be called from the
10077 peephole2 pass, to change the register that addis sets to be the
10079 (emit_fusion_gpr_load): Change the calling signature to have the
10080 operands passed as an array, instead of as separate arguments.
10081 Allow ZERO_EXTEND to be in the memory address, and also
10082 SIGN_EXTEND if -mpower8-fusion-sign.
10084 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
10085 unspec enumeration.
10086 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
10087 adjust the register addis loads up in the peephole2 pass. Do not
10088 depend on the register live/dead state when the peephole pass is done.
10090 2013-08-16 David Malcolm <dmalcolm@redhat.com>
10092 * gengtype.c (create_user_defined_type): Ensure that the kind
10093 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
10094 declaration is seen before the GTY((user)) marking.
10096 2013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
10099 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
10101 2013-08-16 Jan Hubicka <jh@suse.cz>
10103 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
10104 arugment expected_type.
10105 (gimple_fold_call): Use it.
10106 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
10107 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
10108 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
10109 (try_make_edge_direct_virtual_call): Likewise.
10110 * tree.c (obj_type_ref_class): New.
10111 * tree.h (obj_type_ref_class): Use it.
10113 2013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
10115 * sched-vis.c (rtl_slim_pp_initialized): Remove.
10116 (rtl_slim_pp): Likewise.
10117 (init_rtl_slim_pretty_print): Likewise.
10118 (dump_value_slim): Don't call it. Use local pretty printer.
10119 (dump_insn_slim): Likewise.
10120 (dump_rtl_slim): Likewise.
10121 (str_pattern_slim): Likewise.
10122 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
10125 2013-08-16 Jakub Jelinek <jakub@redhat.com>
10127 PR tree-optimization/58164
10128 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
10129 walk gimple_goto_dest of GIMPLE_GOTO.
10131 PR tree-optimization/58165
10132 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
10133 bi_call must be the last stmt in a bb, don't split_block, instead
10134 use fallthru edge from it and give up if there is none.
10135 Release conds vector when returning early.
10137 2013-08-14 Xinliang David Li <davidxl@google.com>
10139 * config/i386/i386.c (ix86_option_override_internal):
10140 Remove unused variable and field.
10142 2013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10145 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
10146 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
10147 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
10148 and Linux, correct BLKmode alignment when 128-bit alignment is
10149 required and compatibility flag is not set.
10150 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
10151 for zero-size arguments when compatibility flag is not set.
10153 2013-08-14 Jakub Jelinek <jakub@redhat.com>
10155 PR tree-optimization/58145
10156 * tree-sra.c (build_ref_for_offset): If prev_base has
10157 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
10159 2013-08-14 Xinliang David Li <davidxl@google.com>
10161 * config/i386/i386.c (ix86_option_override_internal):
10162 Fix uninitialized variable error.
10164 2013-08-14 Xinliang David Li <davidxl@google.com>
10166 * config/i386/i386.opt: Define two new options.
10167 * config/i386/x86-tune.def: Add arch selector field in macros.
10168 * config/i386/i386.h: Adjust macro definition.
10169 * config/i386/i386.c (ix86_option_override_internal):
10171 (parse_mtune_ctrl_str): New function.
10172 (set_ix86_tune_features): New function.
10173 (ix86_function_specific_restore): Call the new helper function.
10175 2013-08-14 Andrey Belevantsev <abel@ispras.ru>
10177 PR rtl-optimization/57662
10178 * sel-sched.c (code_motion_process_successors): When the current insn
10179 is removed after the recursive traversal, break from the loop.
10180 Add comments and debug printouts.
10182 2013-08-14 Jakub Jelinek <jakub@redhat.com>
10183 Alexandre Oliva <aoliva@redhat.com>
10186 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
10187 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
10188 there also UNSPEC_PLTOFF.
10190 2013-08-14 Marek Polacek <polacek@redhat.com>
10192 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
10193 NUM_CONDITIONS bit positions.
10195 2013-08-13 Cary Coutant <ccoutant@google.com>
10197 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
10198 (attr_checksum): Hash vector contents instead of pointer.
10199 (attr_checksum_ordered): Likewise.
10201 2013-08-13 Uros Bizjak <ubizjak@gmail.com>
10203 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
10204 when Pmode != word_mode. Add length_address attribute.
10205 (sse3_monitor_<mode>): Merge from sse3_monitor and
10206 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
10207 Pmode != word_mode. Update insn length attribute.
10208 * config/i386/i386.c (ix86_option_override_internal): Update
10209 ix86_gen_monitor selection for merged sse3_monitor insn.
10211 2013-08-13 Julian Brown <julian@codesourcery.com>
10213 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
10214 perform invalid legitimization on greater-than-word-size modes for
10215 TARGET_E500_DOUBLE.
10217 2013-08-13 Vladimir Makarov <vmakarov@redhat.com>
10219 * ira.c (setup_class_translate_array): Use aclass instead of cl
10220 for classes not fully covered by allocno classes.
10222 2013-08-13 Jakub Jelinek <jakub@redhat.com>
10224 PR tree-optimization/57661
10225 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
10226 * tree-inline.c (tree_function_versioning): Initialize it.
10227 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
10228 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
10229 that is not being copied.
10232 * asan.c (instrument_strlen_call): Fix typo in comment.
10233 Use char * type even for the lhs of POINTER_PLUS_EXPR.
10235 2013-08-13 Steve Ellcey <sellcey@mips.com>
10237 * config/mips/mips.md (prefetch): Use lw instead of ld on
10238 loongson in 32bit mode.
10240 2013-08-13 Nick Clifton <nickc@redhat.com>
10242 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
10244 2013-08-13 Jan Hubicka <jh@suse.cz>
10246 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
10247 introduced edge; fix typo in sanity check.
10248 (cgraph_resolve_speculation): Export; improve diagnostic.
10249 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
10250 speculation at type mismatch.
10251 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
10252 (cgraph_resolve_speculation): Declare.
10253 (symtab_can_be_discarded): New function.
10254 * value-prof.c (gimple_ic_transform): Remove actual transform code.
10255 * ipa-inline-transform.c (speculation_removed): New global var.
10256 (clone_inlined_nodes): See if speculation can be removed.
10257 (inline_call): If speculations was removed, we growths may not match.
10258 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
10259 (speculation_useful_p): New function.
10260 (resolve_noninline_speculation): New function.
10261 (inline_small_functions): Resolve useless speculations.
10262 * ipa-inline.h (speculation_useful_p): Declare
10263 * ipa.c (can_replace_by_local_alias): Simplify.
10264 (ipa_profile): Produce speculative calls in non-lto, too;
10265 add simple cost model; produce local aliases.
10267 2013-08-13 David Malcolm <dmalcolm@redhat.com>
10269 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
10272 2013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
10274 * config/i386/i386.c (ix86_function_versions): Use error + inform.
10276 2013-08-12 Uros Bizjak <ubizjak@gmail.com>
10278 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
10279 iterator instead of X87MODEF.
10281 2013-08-12 Perez Read <netfirewall@gmail.com>
10284 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
10285 operand 0 for intel asm alternative.
10286 (*movabs<mode>_2): Ditto for operand 1.
10288 2013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
10290 * config/aarch64/arm_none.h
10291 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
10293 2013-08-12 Nick Clifton <nickc@redhat.com>
10295 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
10297 2013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
10299 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
10300 expand for QI/HImode operand to produce more effictive code for
10301 unsigned char(short) --> float(double) conversion.
10303 2013-08-12 Alexander Monakov <amonakov@ispras.ru>
10305 * doc/invoke.texi: Mention that -ftls-model does not force the final
10308 2013-08-12 Marek Polacek <polacek@redhat.com>
10309 Marc Glisse <marc.glisse@inria.fr>
10311 PR tree-optimization/57980
10312 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
10313 when creating -1 constant.
10315 2013-08-10 Jan Hubicka <jh@suse.cz>
10317 Workaround binutils PR14342.
10318 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
10319 (gimple_init_edge_profiler): Likewise.
10320 (gimple_gen_ic_func_profiler): Likewise.
10322 2013-08-09 Jan Hubicka <jh@suse.cz>
10324 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
10326 2013-08-09 Xinliang David Li <davidxl@google.com>
10328 * config/i386/stringop.def: New file.
10329 * config/i386/stringop.opt: New file.
10330 * config/i386/i386-opts.h: Include stringopt.def.
10331 * config/i386/i386.opt: Include stringopt.opt.
10332 * config/i386/i386.c (ix86_option_override_internal):
10333 Override default size based stringop inline strategies with options.
10334 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
10337 2013-08-09 Jan Hubicka <jh@suse.cz>
10339 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
10341 2013-08-09 Jan Hubicka <jh@suse.cz>
10343 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
10345 (dump_cgraph_node): Dump profile-id.
10346 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
10347 and common_target_probability.
10348 * lto-cgraph.c (lto_output_edge): Stream common targets.
10349 (lto_output_node): Stream profile ids.
10350 (input_node): Stream profile ids.
10351 (input_edge): Stream common targets.
10352 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
10353 * ipa.c: Include value-prof.h
10354 (ipa_profile_generate_summary): Turn indirect call statement histograms
10355 into common targets.
10356 (ipa_profile): Turn common targets into speculative edges.
10358 2013-08-09 Jan Hubicka <jh@suse.cz>
10360 * cgraph.h (cgraph_node): Add profile_id.
10361 * value-prof.c (cgraph_node_map): Turn into pointer_map.
10362 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
10363 (del_node_map): Update.
10364 (find_func_by_funcdef_no): Replace by ...
10365 (find_func_by_profile_id): ... this one.
10366 (gimple_ic_transform): Do not remove useful histograms when
10367 speculation is not done; dump info when indirect call removal
10369 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
10370 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
10371 (__gcov_indirect_call_profiler_v2): .. this one.
10372 * profile.h (init_node_map): Update.
10373 * coverage.c (coverage_compute_profile_id): New function.
10374 * coverage.h (coverage_compute_profile_id): Declare.
10375 * tree-profile.c (init_ic_make_global_vars): Make
10376 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
10377 (gimple_init_edge_profiler): Update prototype of
10378 __gcov_indirect_call_profiler.
10379 (gimple_gen_ic_func_profiler): Simplify.
10380 (tree_profiling): Use init_node_map
10382 2013-08-09 Jan Hubicka <jh@suse.cz>
10384 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
10385 non-speculative refs.
10386 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
10387 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
10388 (cgraph_set_call_stmt): Likewise.
10389 (cgraph_create_edge_1): Fix release checking compilatoin;
10390 clear lto_stmt_uid.
10391 (cgraph_free_edge): Free indirect info.
10392 (cgraph_turn_edge_to_speculative): New function.
10393 (cgraph_speculative_call_info): New function.
10394 (cgraph_make_edge_direct): Return direct edge; handle speculation.
10395 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
10396 (dump_cgraph_node): Dump speculation.
10397 (verify_edge_count_and_frequency): Accept speculative edges.
10398 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
10399 (verify_cgraph_node): Handle speculation.
10400 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
10401 (cgraph_set_call_stmt): Update prototype.
10402 (cgraph_make_edge_direct): Update prototype.
10403 (cgraph_speculative_call_info): Declare.
10404 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
10405 to change; update call of ipa_find_references.
10406 * ipa-ref.c (ipa_record_reference): Fix return value; clear
10407 lto_stmt_uid and speculative flags.
10408 (ipa_dump_references): Dump speculation.
10409 (ipa_clone_references): Clone speculative flag.
10410 (ipa_clone_referring): Likewise.
10411 (ipa_clone_ref): New function.
10412 (ipa_find_reference): Look into lto_stmt_uids
10413 (ipa_clear_stmts_in_references): Do not clear speculative calls.
10414 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
10415 (ipa_find_reference): Update declaration.
10416 (ipa_clone_ref): Declare.
10417 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
10418 stream speculative flag.
10419 (lto_output_ref): Stream statements uids and speculation.
10420 (input_ref): Likewise.
10421 (input_edge): Stream speuclation.
10422 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
10423 (cgraph_set_call_stmt_including_clones): Handle speculation.
10424 * ipa-inline.c (heap_edge_removal_hook): New function.
10425 (inline_small_functions): Register it.
10426 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
10427 also initialize refs.
10428 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
10430 (try_make_edge_direct_simple_call): Likewise.
10431 (try_make_edge_direct_simple_call): Likewise.
10432 (update_indirect_edges_after_inlining): Likewise.
10433 (remove_described_reference): Look proper lto_stmt_uid.
10434 (propagate_controlled_uses): Likewise.
10435 (propagate_controlled_uses): Liekwise.
10436 * tree-inline.c (copy_bb): Copy speculative edges.
10437 (redirect_all_calls): New function.
10438 (copy_cfg_body): Do redirection after loop info is updated.
10439 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
10441 2013-08-09 Jan Hubicka <jh@suse.cz>
10443 * lto-streamer-out.c (output_function): Renumber PHIs.
10444 * lto-streamer-in.c (input_function): Likewise.
10446 2013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
10448 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
10449 (get_lane_unsigned): Likewise.
10450 (dup_lane_scalar): Likewise.
10451 (get_lane): enable for VALL.
10452 * config/aarch64/aarch64-simd.md
10453 (aarch64_dup_lane_scalar<mode>): Remove.
10454 (aarch64_get_lane_signed<mode>): Likewise.
10455 (aarch64_get_lane_unsigned<mode>): Likewise.
10456 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
10457 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
10458 (aarch64_get_lane<mode>): Enable for all vector modes.
10459 (aarch64_get_lanedi): Remove misleading constraints.
10460 * config/aarch64/arm_neon.h
10461 (__aarch64_vget_lane_any): Define.
10462 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
10463 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
10464 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
10465 * config/aarch64/iterators.md (VDQQH): New.
10466 (VDQQHS): Likewise.
10467 (vwcore): Likewise.
10469 2013-08-09 Eric Botcazou <ebotcazou@adacore.com>
10471 * configure.ac: Add GAS check for LEON instructions on SPARC.
10472 * configure: Regenerate.
10473 * config.in: Likewise.
10474 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
10476 * config/sparc/sparc.opt (LEON, LEON3): New masks.
10477 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
10479 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
10480 (AS_LEON_FLAG): New macro.
10481 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
10482 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
10483 Deal with LEON and LEON3 for the memory model.
10484 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
10485 (atomic_compare_and_swap<mode>_1): Likewise.
10486 (*atomic_compare_and_swap<mode>_1): Likewise.
10488 2013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
10490 * config/arm/neon.md (vcond): Fix floating-point vector
10491 comparisons against 0.
10493 2013-08-08 Vladimir Makarov <vmakarov@redhat.com>
10495 * lra-constraints.c (emit_spill_move): Remove assert.
10496 (process_alt_operands): Add more debugging
10497 output. Increase reject for spilling into memory. Decrease
10498 reject for reloading scratch.
10499 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
10501 2013-08-08 Steve Ellcey <sellcey@mips.com>
10503 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
10504 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
10505 micromips incompatible. Add nan2008.
10506 (MULTILIB_DIRNAMES): Add nan2008.
10507 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
10508 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
10509 and micromips incompatible. Add nan2008.
10510 (MULTILIB_DIRNAMES): Add nan2008.
10511 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
10513 2013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
10515 PR rtl-optimization/58079
10516 * combine.c (combine_simplify_rtx): Avoid using SUBST if
10517 simplify_comparison has widened a comparison with an integer.
10519 2013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10521 * config/arm/neon.md (movmisalign<mode>): Disable when we
10522 don't allow unaligned accesses.
10523 (*movmisalign<mode>_neon_store): Likewise.
10524 (*movmisalign<mode>_neon_load): Likewise.
10525 (*movmisalign<mode>_neon_store): Likewise.
10526 (*movmisalign<mode>_neon_load): Likewise.
10528 2013-08-08 Jan Hubicka <jh@suse.cz>
10530 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
10531 (make_pass_rebuild_cgraph_edges): Also clear references.
10532 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
10533 * ipa-inline-transform.c (inline_transform): Remove all references
10535 * cgraphunit.c (expand_function): Remove all references after
10537 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
10538 (ipa_find_reference): Rewrite to iterator.
10539 (remove_stmt_references): Likewise.
10540 (ipa_clear_stmts_in_references): New function.
10541 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
10542 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
10544 * ipa-split.c (split_function): Remove references in split function.
10546 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
10549 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
10550 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
10552 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
10555 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
10556 suggested mode for the assignment isn't compatible with the
10557 registers required.
10559 2013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
10562 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
10564 2013-08-07 Xinliang David Li <davidxl@google.com>
10566 * config/i386/i386.opt: New option -mtune-ctrl=.
10567 * config/i386/x86-tune.def: New file.
10568 * config/i386/i386.h: include x86-tune.def.
10569 * config/i386/i386.c (ix86_option_override_internal):
10570 Parsing -mtune-ctrl= option and set tune features.
10572 2013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
10575 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
10576 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
10579 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
10581 * rtl.h (update_alignments): Declare.
10582 * final.c (grow_label_align): New function extracted from...
10583 (shorten_branches): ...here. Call it.
10584 (update_alignments): New function.
10585 * reorg.c (sibling_labels): New variable.
10586 (get_label_before): Add SIBLING parameter. If it is non-zero, push
10587 the new label along with it onto the sibling_labels vector.
10588 (fill_simple_delay_slots): Adjust call to get_label_before.
10589 (fill_slots_from_thread): Likewise.
10590 (relax_delay_slots): Likewise.
10591 (make_return_insns): Likewise.
10592 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
10594 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
10596 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
10597 document its semantics.
10598 (diagnostic_report_diagnostic): Adjust accordingly.
10600 2013-08-07 David Malcolm <dmalcolm@redhat.com>
10602 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
10603 (sparc_option_override): ...and port to new C++ pass API.
10604 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
10606 2013-08-07 Peter Bergner <bergner@vnet.ibm.com>
10608 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
10610 2013-08-06 Caroline Tice <cmtice@google.com>
10612 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
10613 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
10614 * tree-pass.h: Add pass_vtable_verify.
10615 * varasm.c (assemble_variable): Add code to properly set the comdat
10616 section and name for the .vtable_map_vars section.
10617 (assemble_vtyv_preinit_initializer): New function.
10618 (default_sectin_type_flags): Make sure .vtable_map_vars section has
10620 * output.h: Add function decl for assemble_vtv_preinit_initializer.
10621 * vtable-verify.c: New file.
10622 * vtable-verify.h: New file.
10623 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
10624 initialiation levels.
10625 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
10626 * passes.def: Insert pass_vtable_verify.
10627 * aclocal.m4: Reorder includes.
10628 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
10629 -fvtv-counts options.
10630 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
10631 as appropriate, if -fvtable-verify=... is used.
10632 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
10633 -fvtable-verify=... is used.
10634 * Makefile.in (OBJS): Add vtable-verify.o to list.
10635 (vtable-verify.o): Add new build rule.
10636 (GTFILES): Add vtable-verify.c to list.
10637 * common.opt (fvtable-verify=): New flag.
10638 (vtv_priority): Values for fvtable-verify= flag.
10639 (fvtv-counts): New flag.
10640 (fvtv-debug): New flag.
10641 * tree.h (save_vtable_map_decl): New extern function decl.
10643 2013-08-07 David Malcolm <dmalcolm@redhat.com>
10645 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
10646 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
10647 (pass_data_rl78_devirt): ...new pass_data instance and...
10648 (make_pass_rl78_devirt): ...new function.
10649 (rl78_asm_file_start): Port pass registration to new C++ API.
10651 2013-08-07 David Malcolm <dmalcolm@redhat.com>
10653 * coretypes.h (rtl_opt_pass): Add.
10654 (gcc::context): Add.
10655 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
10656 (epiphany_init): Port to new C++ pass API.
10657 (epiphany_optimize_mode_switching): Likewise.
10658 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
10659 (pass_manager::get_pass_mode_switching): New.
10660 (pass_manager::get_pass_peephole2): New.
10661 * mode-switching.c (pass_mode_switching): Add clone method.
10662 * recog.c (pass_peephole2): Add clone method.
10663 (pass_split_all_insns): Add clone method.
10665 2013-08-06 David Malcolm <dmalcolm@redhat.com>
10667 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
10668 (mips_option_override): ...here, porting to new C++ API for passes.
10670 2013-08-06 Jan Hubicka <jh@suse.cz>
10672 * cgraph.c (cgraph_get_body): New function based on lto.c
10674 * cgraph.h (cgraph_get_body): Declare.
10675 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
10677 * cgraphunit.c (expand_function): Get body prior expanding.
10678 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
10679 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
10681 * passes.c (do_per_function_toporder): Get body.
10682 * tree-inline.c (expand_call_inline): Get body prior inlining it.
10683 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
10685 2013-08-06 Martin Jambor <mjambor@suse.cz>
10688 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
10689 is not re-finalized. Rename second parameter to no_collect.
10691 2013-08-06 Martin Jambor <mjambor@suse.cz>
10693 PR middle-end/58041
10694 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
10695 MEM_REF has proper alignment information.
10697 2013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
10700 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
10702 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
10703 rtx (*) (rtx, ...) with insn_gen_fn.
10704 * genoutput.c (output_insn_data): Cast gen_? function pointers to
10705 insn_gen_fn::stored_funcptr. Add initializer braces.
10707 2013-08-05 David Malcolm <dmalcolm@redhat.com>
10709 Rewrite how instances of passes are cloned to remove assumptions
10710 about their sizes (thus allowing pass subclasses to have
10711 additional data fields, albeit non-GC-managed ones at this point).
10713 * passes.c (make_pass_instance): Now that passes have clone
10714 methods, rewrite this function to eliminate XNEW and memcpy
10715 calls that used hardcoded sizes. Since this function no longer
10716 creates pass instances, rename it to...
10717 (add_pass_instance): ...this. Document the old way that passes were
10718 numbered and flagged, and rework this function to continue using it.
10719 (next_pass_1): Add an initial_pass argument for use by
10721 (position_pass): When adding multiple instances of a pass, use
10722 the pass's clone method, rather than relying on the XNEW/memcpy
10723 within the former make_pass_instance (now add_pass_instance).
10724 (pass_manager::pass_manager): When invoking next_pass_1, also supply
10725 the initial instance of the current pass within the pass manager.
10727 2013-08-05 David Malcolm <dmalcolm@redhat.com>
10729 This is the automated part of the conversion of passes from C
10730 structs to C++ classes.
10732 Patch autogenerated by refactor_passes.py from
10733 https://github.com/davidmalcolm/gcc-refactoring-scripts
10734 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
10736 * asan.c (pass_asan): Convert from a global struct to a subclass of
10737 gimple_opt_pass along with...
10738 (pass_data_asan): ...new pass_data instance and...
10739 (make_pass_asan): ...new function.
10740 (pass_asan_O0): Convert from a global struct to a subclass of
10741 gimple_opt_pass along with...
10742 (pass_data_asan_O0): ...new pass_data instance and...
10743 (make_pass_asan_O0): ...new function.
10744 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
10745 subclass of rtl_opt_pass along with...
10746 (pass_data_inc_dec): ...new pass_data instance and...
10747 (make_pass_inc_dec): ...new function.
10748 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
10749 a subclass of rtl_opt_pass along with...
10750 (pass_data_reorder_blocks): ...new pass_data instance and...
10751 (make_pass_reorder_blocks): ...new function.
10752 (pass_duplicate_computed_gotos): Convert from a global struct to a
10753 subclass of rtl_opt_pass along with...
10754 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
10755 (make_pass_duplicate_computed_gotos): ...new function.
10756 (pass_partition_blocks): Convert from a global struct to a subclass of
10757 rtl_opt_pass along with...
10758 (pass_data_partition_blocks): ...new pass_data instance and...
10759 (make_pass_partition_blocks): ...new function.
10760 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
10761 struct to a subclass of rtl_opt_pass along with...
10762 (pass_data_branch_target_load_optimize1): ...new pass_data instance
10764 (make_pass_branch_target_load_optimize1): ...new function.
10765 (pass_branch_target_load_optimize2): Convert from a global struct to a
10766 subclass of rtl_opt_pass along with...
10767 (pass_data_branch_target_load_optimize2): ...new pass_data instance
10769 (make_pass_branch_target_load_optimize2): ...new function.
10770 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
10771 of rtl_opt_pass along with...
10772 (pass_data_jump): ...new pass_data instance and...
10773 (make_pass_jump): ...new function.
10774 (pass_jump2): Convert from a global struct to a subclass of
10775 rtl_opt_pass along with...
10776 (pass_data_jump2): ...new pass_data instance and...
10777 (make_pass_jump2): ...new function.
10778 * cfgexpand.c (pass_expand): Convert from a global struct to a
10779 subclass of rtl_opt_pass along with...
10780 (pass_data_expand): ...new pass_data instance and...
10781 (make_pass_expand): ...new function.
10782 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
10783 of rtl_opt_pass along with...
10784 (pass_data_free_cfg): ...new pass_data instance and...
10785 (make_pass_free_cfg): ...new function.
10786 (pass_into_cfg_layout_mode): Convert from a global struct to a
10787 subclass of rtl_opt_pass along with...
10788 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
10789 (make_pass_into_cfg_layout_mode): ...new function.
10790 (pass_outof_cfg_layout_mode): Convert from a global struct to a
10791 subclass of rtl_opt_pass along with...
10792 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
10793 (make_pass_outof_cfg_layout_mode): ...new function.
10794 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
10795 struct to a subclass of gimple_opt_pass along with...
10796 (pass_data_build_cgraph_edges): ...new pass_data instance and...
10797 (make_pass_build_cgraph_edges): ...new function.
10798 (pass_rebuild_cgraph_edges): Convert from a global struct to a
10799 subclass of gimple_opt_pass along with...
10800 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
10801 (make_pass_rebuild_cgraph_edges): ...new function.
10802 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
10803 subclass of gimple_opt_pass along with...
10804 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
10806 (make_pass_remove_cgraph_callee_edges): ...new function.
10807 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
10808 struct to a subclass of rtl_opt_pass along with...
10809 (pass_data_stack_adjustments): ...new pass_data instance and...
10810 (make_pass_stack_adjustments): ...new function.
10811 * combine.c (pass_combine): Convert from a global struct to a subclass
10812 of rtl_opt_pass along with...
10813 (pass_data_combine): ...new pass_data instance and...
10814 (make_pass_combine): ...new function.
10815 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
10816 global struct to a subclass of rtl_opt_pass along with...
10817 (pass_data_compare_elim_after_reload): ...new pass_data instance
10819 (make_pass_compare_elim_after_reload): ...new function.
10820 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
10821 of rtl_opt_pass along with...
10822 (pass_data_rtl_cprop): ...new pass_data instance and...
10823 (make_pass_rtl_cprop): ...new function.
10824 * cse.c (pass_cse): Convert from a global struct to a subclass of
10825 rtl_opt_pass along with...
10826 (pass_data_cse): ...new pass_data instance and...
10827 (make_pass_cse): ...new function.
10828 (pass_cse2): Convert from a global struct to a subclass of
10829 rtl_opt_pass along with...
10830 (pass_data_cse2): ...new pass_data instance and...
10831 (make_pass_cse2): ...new function.
10832 (pass_cse_after_global_opts): Convert from a global struct to a
10833 subclass of rtl_opt_pass along with...
10834 (pass_data_cse_after_global_opts): ...new pass_data instance and...
10835 (make_pass_cse_after_global_opts): ...new function.
10836 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
10837 of rtl_opt_pass along with...
10838 (pass_data_ud_rtl_dce): ...new pass_data instance and...
10839 (make_pass_ud_rtl_dce): ...new function.
10840 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
10841 rtl_opt_pass along with...
10842 (pass_data_fast_rtl_dce): ...new pass_data instance and...
10843 (make_pass_fast_rtl_dce): ...new function.
10844 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
10845 a subclass of rtl_opt_pass along with...
10846 (pass_data_df_initialize_opt): ...new pass_data instance and...
10847 (make_pass_df_initialize_opt): ...new function.
10848 (pass_df_initialize_no_opt): Convert from a global struct to a
10849 subclass of rtl_opt_pass along with...
10850 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
10851 (make_pass_df_initialize_no_opt): ...new function.
10852 (pass_df_finish): Convert from a global struct to a subclass of
10853 rtl_opt_pass along with...
10854 (pass_data_df_finish): ...new pass_data instance and...
10855 (make_pass_df_finish): ...new function.
10856 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
10857 rtl_opt_pass along with...
10858 (pass_data_rtl_dse1): ...new pass_data instance and...
10859 (make_pass_rtl_dse1): ...new function.
10860 (pass_rtl_dse2): Convert from a global struct to a subclass of
10861 rtl_opt_pass along with...
10862 (pass_data_rtl_dse2): ...new pass_data instance and...
10863 (make_pass_rtl_dse2): ...new function.
10864 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
10865 subclass of rtl_opt_pass along with...
10866 (pass_data_dwarf2_frame): ...new pass_data instance and...
10867 (make_pass_dwarf2_frame): ...new function.
10868 * except.c (pass_set_nothrow_function_flags): Convert from a global
10869 struct to a subclass of rtl_opt_pass along with...
10870 (pass_data_set_nothrow_function_flags): ...new pass_data instance
10872 (make_pass_set_nothrow_function_flags): ...new function.
10873 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
10874 subclass of rtl_opt_pass along with...
10875 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
10877 (make_pass_convert_to_eh_region_ranges): ...new function.
10878 * final.c (pass_compute_alignments): Convert from a global struct to a
10879 subclass of rtl_opt_pass along with...
10880 (pass_data_compute_alignments): ...new pass_data instance and...
10881 (make_pass_compute_alignments): ...new function.
10882 (pass_final): Convert from a global struct to a subclass of
10883 rtl_opt_pass along with...
10884 (pass_data_final): ...new pass_data instance and...
10885 (make_pass_final): ...new function.
10886 (pass_shorten_branches): Convert from a global struct to a subclass of
10887 rtl_opt_pass along with...
10888 (pass_data_shorten_branches): ...new pass_data instance and...
10889 (make_pass_shorten_branches): ...new function.
10890 (pass_clean_state): Convert from a global struct to a subclass of
10891 rtl_opt_pass along with...
10892 (pass_data_clean_state): ...new pass_data instance and...
10893 (make_pass_clean_state): ...new function.
10894 * function.c (pass_instantiate_virtual_regs): Convert from a global
10895 struct to a subclass of rtl_opt_pass along with...
10896 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
10897 (make_pass_instantiate_virtual_regs): ...new function.
10898 (pass_leaf_regs): Convert from a global struct to a subclass of
10899 rtl_opt_pass along with...
10900 (pass_data_leaf_regs): ...new pass_data instance and...
10901 (make_pass_leaf_regs): ...new function.
10902 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
10903 subclass of rtl_opt_pass along with...
10904 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
10906 (make_pass_thread_prologue_and_epilogue): ...new function.
10907 (pass_match_asm_constraints): Convert from a global struct to a
10908 subclass of rtl_opt_pass along with...
10909 (pass_data_match_asm_constraints): ...new pass_data instance and...
10910 (make_pass_match_asm_constraints): ...new function.
10911 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
10912 subclass of rtl_opt_pass along with...
10913 (pass_data_rtl_fwprop): ...new pass_data instance and...
10914 (make_pass_rtl_fwprop): ...new function.
10915 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
10916 rtl_opt_pass along with...
10917 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
10918 (make_pass_rtl_fwprop_addr): ...new function.
10919 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
10920 rtl_opt_pass along with...
10921 (pass_data_rtl_pre): ...new pass_data instance and...
10922 (make_pass_rtl_pre): ...new function.
10923 (pass_rtl_hoist): Convert from a global struct to a subclass of
10924 rtl_opt_pass along with...
10925 (pass_data_rtl_hoist): ...new pass_data instance and...
10926 (make_pass_rtl_hoist): ...new function.
10927 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
10928 subclass of gimple_opt_pass along with...
10929 (pass_data_lower_cf): ...new pass_data instance and...
10930 (make_pass_lower_cf): ...new function.
10931 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
10932 from a global struct to a subclass of gimple_opt_pass along with...
10933 (pass_data_strength_reduction): ...new pass_data instance and...
10934 (make_pass_strength_reduction): ...new function.
10935 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
10936 of rtl_opt_pass along with...
10937 (pass_data_rtl_ifcvt): ...new pass_data instance and...
10938 (make_pass_rtl_ifcvt): ...new function.
10939 (pass_if_after_combine): Convert from a global struct to a subclass of
10940 rtl_opt_pass along with...
10941 (pass_data_if_after_combine): ...new pass_data instance and...
10942 (make_pass_if_after_combine): ...new function.
10943 (pass_if_after_reload): Convert from a global struct to a subclass of
10944 rtl_opt_pass along with...
10945 (pass_data_if_after_reload): ...new pass_data instance and...
10946 (make_pass_if_after_reload): ...new function.
10947 * init-regs.c (pass_initialize_regs): Convert from a global struct to
10948 a subclass of rtl_opt_pass along with...
10949 (pass_data_initialize_regs): ...new pass_data instance and...
10950 (make_pass_initialize_regs): ...new function.
10951 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
10952 of ipa_opt_pass_d along with...
10953 (pass_data_ipa_cp): ...new pass_data instance and...
10954 (make_pass_ipa_cp): ...new function.
10955 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
10956 global struct to a subclass of gimple_opt_pass along with...
10957 (pass_data_inline_parameters): ...new pass_data instance and...
10958 (make_pass_inline_parameters): ...new function.
10959 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
10960 subclass of gimple_opt_pass along with...
10961 (pass_data_early_inline): ...new pass_data instance and...
10962 (make_pass_early_inline): ...new function.
10963 (pass_ipa_inline): Convert from a global struct to a subclass of
10964 ipa_opt_pass_d along with...
10965 (pass_data_ipa_inline): ...new pass_data instance and...
10966 (make_pass_ipa_inline): ...new function.
10967 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
10968 struct to a subclass of gimple_opt_pass along with...
10969 (pass_data_local_pure_const): ...new pass_data instance and...
10970 (make_pass_local_pure_const): ...new function.
10971 (pass_ipa_pure_const): Convert from a global struct to a subclass of
10972 ipa_opt_pass_d along with...
10973 (pass_data_ipa_pure_const): ...new pass_data instance and...
10974 (make_pass_ipa_pure_const): ...new function.
10975 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
10976 to a subclass of ipa_opt_pass_d along with...
10977 (pass_data_ipa_reference): ...new pass_data instance and...
10978 (make_pass_ipa_reference): ...new function.
10979 * ipa-split.c (pass_split_functions): Convert from a global struct to
10980 a subclass of gimple_opt_pass along with...
10981 (pass_data_split_functions): ...new pass_data instance and...
10982 (make_pass_split_functions): ...new function.
10983 (pass_feedback_split_functions): Convert from a global struct to a
10984 subclass of gimple_opt_pass along with...
10985 (pass_data_feedback_split_functions): ...new pass_data instance and...
10986 (make_pass_feedback_split_functions): ...new function.
10987 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
10988 global struct to a subclass of simple_ipa_opt_pass along with...
10989 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
10991 (make_pass_ipa_function_and_variable_visibility): ...new function.
10992 (pass_ipa_free_inline_summary): Convert from a global struct to a
10993 subclass of simple_ipa_opt_pass along with...
10994 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
10995 (make_pass_ipa_free_inline_summary): ...new function.
10996 (pass_ipa_whole_program_visibility): Convert from a global struct to a
10997 subclass of ipa_opt_pass_d along with...
10998 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
11000 (make_pass_ipa_whole_program_visibility): ...new function.
11001 (pass_ipa_profile): Convert from a global struct to a subclass of
11002 ipa_opt_pass_d along with...
11003 (pass_data_ipa_profile): ...new pass_data instance and...
11004 (make_pass_ipa_profile): ...new function.
11005 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
11006 ipa_opt_pass_d along with...
11007 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
11008 (make_pass_ipa_cdtor_merge): ...new function.
11009 * ira.c (pass_ira): Convert from a global struct to a subclass of
11010 rtl_opt_pass along with...
11011 (pass_data_ira): ...new pass_data instance and...
11012 (make_pass_ira): ...new function.
11013 (pass_reload): Convert from a global struct to a subclass of
11014 rtl_opt_pass along with...
11015 (pass_data_reload): ...new pass_data instance and...
11016 (make_pass_reload): ...new function.
11017 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
11018 subclass of rtl_opt_pass along with...
11019 (pass_data_cleanup_barriers): ...new pass_data instance and...
11020 (make_pass_cleanup_barriers): ...new function.
11021 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
11022 of rtl_opt_pass along with...
11023 (pass_data_loop2): ...new pass_data instance and...
11024 (make_pass_loop2): ...new function.
11025 (pass_rtl_loop_init): Convert from a global struct to a subclass of
11026 rtl_opt_pass along with...
11027 (pass_data_rtl_loop_init): ...new pass_data instance and...
11028 (make_pass_rtl_loop_init): ...new function.
11029 (pass_rtl_loop_done): Convert from a global struct to a subclass of
11030 rtl_opt_pass along with...
11031 (pass_data_rtl_loop_done): ...new pass_data instance and...
11032 (make_pass_rtl_loop_done): ...new function.
11033 (pass_rtl_move_loop_invariants): Convert from a global struct to a
11034 subclass of rtl_opt_pass along with...
11035 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
11036 (make_pass_rtl_move_loop_invariants): ...new function.
11037 (pass_rtl_unswitch): Convert from a global struct to a subclass of
11038 rtl_opt_pass along with...
11039 (pass_data_rtl_unswitch): ...new pass_data instance and...
11040 (make_pass_rtl_unswitch): ...new function.
11041 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
11042 subclass of rtl_opt_pass along with...
11043 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
11045 (make_pass_rtl_unroll_and_peel_loops): ...new function.
11046 (pass_rtl_doloop): Convert from a global struct to a subclass of
11047 rtl_opt_pass along with...
11048 (pass_data_rtl_doloop): ...new pass_data instance and...
11049 (make_pass_rtl_doloop): ...new function.
11050 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
11051 a subclass of rtl_opt_pass along with...
11052 (pass_data_lower_subreg): ...new pass_data instance and...
11053 (make_pass_lower_subreg): ...new function.
11054 (pass_lower_subreg2): Convert from a global struct to a subclass of
11055 rtl_opt_pass along with...
11056 (pass_data_lower_subreg2): ...new pass_data instance and...
11057 (make_pass_lower_subreg2): ...new function.
11058 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
11059 struct to a subclass of ipa_opt_pass_d along with...
11060 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
11061 (make_pass_ipa_lto_gimple_out): ...new function.
11062 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
11063 of ipa_opt_pass_d along with...
11064 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
11065 (make_pass_ipa_lto_finish_out): ...new function.
11066 * mode-switching.c (pass_mode_switching): Convert from a global struct
11067 to a subclass of rtl_opt_pass along with...
11068 (pass_data_mode_switching): ...new pass_data instance and...
11069 (make_pass_mode_switching): ...new function.
11070 * modulo-sched.c (pass_sms): Convert from a global struct to a
11071 subclass of rtl_opt_pass along with...
11072 (pass_data_sms): ...new pass_data instance and...
11073 (make_pass_sms): ...new function.
11074 * omp-low.c (pass_expand_omp): Convert from a global struct to a
11075 subclass of gimple_opt_pass along with...
11076 (pass_data_expand_omp): ...new pass_data instance and...
11077 (make_pass_expand_omp): ...new function.
11078 (pass_lower_omp): Convert from a global struct to a subclass of
11079 gimple_opt_pass along with...
11080 (pass_data_lower_omp): ...new pass_data instance and...
11081 (make_pass_lower_omp): ...new function.
11082 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
11083 of gimple_opt_pass along with...
11084 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
11085 (make_pass_diagnose_omp_blocks): ...new function.
11086 * passes.c (pass_early_local_passes): Convert from a global struct to
11087 a subclass of simple_ipa_opt_pass along with...
11088 (pass_data_early_local_passes): ...new pass_data instance and...
11089 (make_pass_early_local_passes): ...new function.
11090 (pass_all_early_optimizations): Convert from a global struct to a
11091 subclass of gimple_opt_pass along with...
11092 (pass_data_all_early_optimizations): ...new pass_data instance and...
11093 (make_pass_all_early_optimizations): ...new function.
11094 (pass_all_optimizations): Convert from a global struct to a subclass
11095 of gimple_opt_pass along with...
11096 (pass_data_all_optimizations): ...new pass_data instance and...
11097 (make_pass_all_optimizations): ...new function.
11098 (pass_all_optimizations_g): Convert from a global struct to a subclass
11099 of gimple_opt_pass along with...
11100 (pass_data_all_optimizations_g): ...new pass_data instance and...
11101 (make_pass_all_optimizations_g): ...new function.
11102 (pass_rest_of_compilation): Convert from a global struct to a subclass
11103 of rtl_opt_pass along with...
11104 (pass_data_rest_of_compilation): ...new pass_data instance and...
11105 (make_pass_rest_of_compilation): ...new function.
11106 (pass_postreload): Convert from a global struct to a subclass of
11107 rtl_opt_pass along with...
11108 (pass_data_postreload): ...new pass_data instance and...
11109 (make_pass_postreload): ...new function.
11110 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
11111 subclass of rtl_opt_pass along with...
11112 (pass_data_gcse2): ...new pass_data instance and...
11113 (make_pass_gcse2): ...new function.
11114 * postreload.c (pass_postreload_cse): Convert from a global struct to
11115 a subclass of rtl_opt_pass along with...
11116 (pass_data_postreload_cse): ...new pass_data instance and...
11117 (make_pass_postreload_cse): ...new function.
11118 * predict.c (pass_profile): Convert from a global struct to a subclass
11119 of gimple_opt_pass along with...
11120 (pass_data_profile): ...new pass_data instance and...
11121 (make_pass_profile): ...new function.
11122 (pass_strip_predict_hints): Convert from a global struct to a subclass
11123 of gimple_opt_pass along with...
11124 (pass_data_strip_predict_hints): ...new pass_data instance and...
11125 (make_pass_strip_predict_hints): ...new function.
11126 * recog.c (pass_peephole2): Convert from a global struct to a subclass
11127 of rtl_opt_pass along with...
11128 (pass_data_peephole2): ...new pass_data instance and...
11129 (make_pass_peephole2): ...new function.
11130 (pass_split_all_insns): Convert from a global struct to a subclass of
11131 rtl_opt_pass along with...
11132 (pass_data_split_all_insns): ...new pass_data instance and...
11133 (make_pass_split_all_insns): ...new function.
11134 (pass_split_after_reload): Convert from a global struct to a subclass
11135 of rtl_opt_pass along with...
11136 (pass_data_split_after_reload): ...new pass_data instance and...
11137 (make_pass_split_after_reload): ...new function.
11138 (pass_split_before_regstack): Convert from a global struct to a
11139 subclass of rtl_opt_pass along with...
11140 (pass_data_split_before_regstack): ...new pass_data instance and...
11141 (make_pass_split_before_regstack): ...new function.
11142 (pass_split_before_sched2): Convert from a global struct to a subclass
11143 of rtl_opt_pass along with...
11144 (pass_data_split_before_sched2): ...new pass_data instance and...
11145 (make_pass_split_before_sched2): ...new function.
11146 (pass_split_for_shorten_branches): Convert from a global struct to a
11147 subclass of rtl_opt_pass along with...
11148 (pass_data_split_for_shorten_branches): ...new pass_data instance
11150 (make_pass_split_for_shorten_branches): ...new function.
11151 * ree.c (pass_ree): Convert from a global struct to a subclass of
11152 rtl_opt_pass along with...
11153 (pass_data_ree): ...new pass_data instance and...
11154 (make_pass_ree): ...new function.
11155 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
11156 subclass of rtl_opt_pass along with...
11157 (pass_data_stack_regs): ...new pass_data instance and...
11158 (make_pass_stack_regs): ...new function.
11159 (pass_stack_regs_run): Convert from a global struct to a subclass of
11160 rtl_opt_pass along with...
11161 (pass_data_stack_regs_run): ...new pass_data instance and...
11162 (make_pass_stack_regs_run): ...new function.
11163 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
11164 subclass of rtl_opt_pass along with...
11165 (pass_data_cprop_hardreg): ...new pass_data instance and...
11166 (make_pass_cprop_hardreg): ...new function.
11167 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
11168 subclass of rtl_opt_pass along with...
11169 (pass_data_reginfo_init): ...new pass_data instance and...
11170 (make_pass_reginfo_init): ...new function.
11171 * regmove.c (pass_regmove): Convert from a global struct to a subclass
11172 of rtl_opt_pass along with...
11173 (pass_data_regmove): ...new pass_data instance and...
11174 (make_pass_regmove): ...new function.
11175 * regrename.c (pass_regrename): Convert from a global struct to a
11176 subclass of rtl_opt_pass along with...
11177 (pass_data_regrename): ...new pass_data instance and...
11178 (make_pass_regrename): ...new function.
11179 * reorg.c (pass_delay_slots): Convert from a global struct to a
11180 subclass of rtl_opt_pass along with...
11181 (pass_data_delay_slots): ...new pass_data instance and...
11182 (make_pass_delay_slots): ...new function.
11183 (pass_machine_reorg): Convert from a global struct to a subclass of
11184 rtl_opt_pass along with...
11185 (pass_data_machine_reorg): ...new pass_data instance and...
11186 (make_pass_machine_reorg): ...new function.
11187 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
11188 of rtl_opt_pass along with...
11189 (pass_data_sched): ...new pass_data instance and...
11190 (make_pass_sched): ...new function.
11191 (pass_sched2): Convert from a global struct to a subclass of
11192 rtl_opt_pass along with...
11193 (pass_data_sched2): ...new pass_data instance and...
11194 (make_pass_sched2): ...new function.
11195 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
11196 to a subclass of rtl_opt_pass along with...
11197 (pass_data_stack_ptr_mod): ...new pass_data instance and...
11198 (make_pass_stack_ptr_mod): ...new function.
11199 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
11200 to a subclass of rtl_opt_pass along with...
11201 (pass_data_rtl_store_motion): ...new pass_data instance and...
11202 (make_pass_rtl_store_motion): ...new function.
11203 * tracer.c (pass_tracer): Convert from a global struct to a subclass
11204 of gimple_opt_pass along with...
11205 (pass_data_tracer): ...new pass_data instance and...
11206 (make_pass_tracer): ...new function.
11207 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
11208 to a subclass of gimple_opt_pass along with...
11209 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
11210 (make_pass_diagnose_tm_blocks): ...new function.
11211 (pass_lower_tm): Convert from a global struct to a subclass of
11212 gimple_opt_pass along with...
11213 (pass_data_lower_tm): ...new pass_data instance and...
11214 (make_pass_lower_tm): ...new function.
11215 (pass_tm_init): Convert from a global struct to a subclass of
11216 gimple_opt_pass along with...
11217 (pass_data_tm_init): ...new pass_data instance and...
11218 (make_pass_tm_init): ...new function.
11219 (pass_tm_mark): Convert from a global struct to a subclass of
11220 gimple_opt_pass along with...
11221 (pass_data_tm_mark): ...new pass_data instance and...
11222 (make_pass_tm_mark): ...new function.
11223 (pass_tm_edges): Convert from a global struct to a subclass of
11224 gimple_opt_pass along with...
11225 (pass_data_tm_edges): ...new pass_data instance and...
11226 (make_pass_tm_edges): ...new function.
11227 (pass_tm_memopt): Convert from a global struct to a subclass of
11228 gimple_opt_pass along with...
11229 (pass_data_tm_memopt): ...new pass_data instance and...
11230 (make_pass_tm_memopt): ...new function.
11231 (pass_ipa_tm): Convert from a global struct to a subclass of
11232 simple_ipa_opt_pass along with...
11233 (pass_data_ipa_tm): ...new pass_data instance and...
11234 (make_pass_ipa_tm): ...new function.
11235 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
11236 subclass of gimple_opt_pass along with...
11237 (pass_data_call_cdce): ...new pass_data instance and...
11238 (make_pass_call_cdce): ...new function.
11239 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
11240 subclass of gimple_opt_pass along with...
11241 (pass_data_build_cfg): ...new pass_data instance and...
11242 (make_pass_build_cfg): ...new function.
11243 (pass_split_crit_edges): Convert from a global struct to a subclass of
11244 gimple_opt_pass along with...
11245 (pass_data_split_crit_edges): ...new pass_data instance and...
11246 (make_pass_split_crit_edges): ...new function.
11247 (pass_warn_function_return): Convert from a global struct to a
11248 subclass of gimple_opt_pass along with...
11249 (pass_data_warn_function_return): ...new pass_data instance and...
11250 (make_pass_warn_function_return): ...new function.
11251 (pass_warn_function_noreturn): Convert from a global struct to a
11252 subclass of gimple_opt_pass along with...
11253 (pass_data_warn_function_noreturn): ...new pass_data instance and...
11254 (make_pass_warn_function_noreturn): ...new function.
11255 (pass_warn_unused_result): Convert from a global struct to a subclass
11256 of gimple_opt_pass along with...
11257 (pass_data_warn_unused_result): ...new pass_data instance and...
11258 (make_pass_warn_unused_result): ...new function.
11259 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
11260 a subclass of gimple_opt_pass along with...
11261 (pass_data_merge_phi): ...new pass_data instance and...
11262 (make_pass_merge_phi): ...new function.
11263 * tree-complex.c (pass_lower_complex): Convert from a global struct to
11264 a subclass of gimple_opt_pass along with...
11265 (pass_data_lower_complex): ...new pass_data instance and...
11266 (make_pass_lower_complex): ...new function.
11267 (pass_lower_complex_O0): Convert from a global struct to a subclass of
11268 gimple_opt_pass along with...
11269 (pass_data_lower_complex_O0): ...new pass_data instance and...
11270 (make_pass_lower_complex_O0): ...new function.
11271 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
11272 subclass of gimple_opt_pass along with...
11273 (pass_data_lower_eh): ...new pass_data instance and...
11274 (make_pass_lower_eh): ...new function.
11275 (pass_refactor_eh): Convert from a global struct to a subclass of
11276 gimple_opt_pass along with...
11277 (pass_data_refactor_eh): ...new pass_data instance and...
11278 (make_pass_refactor_eh): ...new function.
11279 (pass_lower_resx): Convert from a global struct to a subclass of
11280 gimple_opt_pass along with...
11281 (pass_data_lower_resx): ...new pass_data instance and...
11282 (make_pass_lower_resx): ...new function.
11283 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
11284 of gimple_opt_pass along with...
11285 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
11286 (make_pass_lower_eh_dispatch): ...new function.
11287 (pass_cleanup_eh): Convert from a global struct to a subclass of
11288 gimple_opt_pass along with...
11289 (pass_data_cleanup_eh): ...new pass_data instance and...
11290 (make_pass_cleanup_eh): ...new function.
11291 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
11292 to a subclass of simple_ipa_opt_pass along with...
11293 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
11294 (make_pass_ipa_lower_emutls): ...new function.
11295 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
11296 a subclass of gimple_opt_pass along with...
11297 (pass_data_if_conversion): ...new pass_data instance and...
11298 (make_pass_if_conversion): ...new function.
11299 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
11300 subclass of gimple_opt_pass along with...
11301 (pass_data_build_ssa): ...new pass_data instance and...
11302 (make_pass_build_ssa): ...new function.
11303 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
11304 global struct to a subclass of gimple_opt_pass along with...
11305 (pass_data_loop_distribution): ...new pass_data instance and...
11306 (make_pass_loop_distribution): ...new function.
11307 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
11308 subclass of gimple_opt_pass along with...
11309 (pass_data_mudflap_1): ...new pass_data instance and...
11310 (make_pass_mudflap_1): ...new function.
11311 (pass_mudflap_2): Convert from a global struct to a subclass of
11312 gimple_opt_pass along with...
11313 (pass_data_mudflap_2): ...new pass_data instance and...
11314 (make_pass_mudflap_2): ...new function.
11315 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
11316 subclass of gimple_opt_pass along with...
11317 (pass_data_mudflap_1): ...new pass_data instance and...
11318 (make_pass_mudflap_1): ...new function.
11319 (pass_mudflap_2): Convert from a global struct to a subclass of
11320 gimple_opt_pass along with...
11321 (pass_data_mudflap_2): ...new pass_data instance and...
11322 (make_pass_mudflap_2): ...new function.
11323 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
11324 gimple_opt_pass along with...
11325 (pass_data_nrv): ...new pass_data instance and...
11326 (make_pass_nrv): ...new function.
11327 (pass_return_slot): Convert from a global struct to a subclass of
11328 gimple_opt_pass along with...
11329 (pass_data_return_slot): ...new pass_data instance and...
11330 (make_pass_return_slot): ...new function.
11331 * tree-object-size.c (pass_object_sizes): Convert from a global struct
11332 to a subclass of gimple_opt_pass along with...
11333 (pass_data_object_sizes): ...new pass_data instance and...
11334 (make_pass_object_sizes): ...new function.
11335 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
11336 global struct to a subclass of gimple_opt_pass along with...
11337 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
11339 (make_pass_cleanup_cfg_post_optimizing): ...new function.
11340 (pass_fixup_cfg): Convert from a global struct to a subclass of
11341 gimple_opt_pass along with...
11342 (pass_data_fixup_cfg): ...new pass_data instance and...
11343 (make_pass_fixup_cfg): ...new function.
11344 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
11345 (make_pass_mudflap_1): ...new function.
11346 (pass_mudflap_2): Replace declaration with that of...
11347 (make_pass_mudflap_2): ...new function.
11348 (pass_asan): Replace declaration with that of...
11349 (make_pass_asan): ...new function.
11350 (pass_asan_O0): Replace declaration with that of...
11351 (make_pass_asan_O0): ...new function.
11352 (pass_tsan): Replace declaration with that of...
11353 (make_pass_tsan): ...new function.
11354 (pass_tsan_O0): Replace declaration with that of...
11355 (make_pass_tsan_O0): ...new function.
11356 (pass_lower_cf): Replace declaration with that of...
11357 (make_pass_lower_cf): ...new function.
11358 (pass_refactor_eh): Replace declaration with that of...
11359 (make_pass_refactor_eh): ...new function.
11360 (pass_lower_eh): Replace declaration with that of...
11361 (make_pass_lower_eh): ...new function.
11362 (pass_lower_eh_dispatch): Replace declaration with that of...
11363 (make_pass_lower_eh_dispatch): ...new function.
11364 (pass_lower_resx): Replace declaration with that of...
11365 (make_pass_lower_resx): ...new function.
11366 (pass_build_cfg): Replace declaration with that of...
11367 (make_pass_build_cfg): ...new function.
11368 (pass_early_tree_profile): Replace declaration with that of...
11369 (make_pass_early_tree_profile): ...new function.
11370 (pass_cleanup_eh): Replace declaration with that of...
11371 (make_pass_cleanup_eh): ...new function.
11372 (pass_sra): Replace declaration with that of...
11373 (make_pass_sra): ...new function.
11374 (pass_sra_early): Replace declaration with that of...
11375 (make_pass_sra_early): ...new function.
11376 (pass_early_ipa_sra): Replace declaration with that of...
11377 (make_pass_early_ipa_sra): ...new function.
11378 (pass_tail_recursion): Replace declaration with that of...
11379 (make_pass_tail_recursion): ...new function.
11380 (pass_tail_calls): Replace declaration with that of...
11381 (make_pass_tail_calls): ...new function.
11382 (pass_tree_loop): Replace declaration with that of...
11383 (make_pass_tree_loop): ...new function.
11384 (pass_tree_loop_init): Replace declaration with that of...
11385 (make_pass_tree_loop_init): ...new function.
11386 (pass_lim): Replace declaration with that of...
11387 (make_pass_lim): ...new function.
11388 (pass_tree_unswitch): Replace declaration with that of...
11389 (make_pass_tree_unswitch): ...new function.
11390 (pass_predcom): Replace declaration with that of...
11391 (make_pass_predcom): ...new function.
11392 (pass_iv_canon): Replace declaration with that of...
11393 (make_pass_iv_canon): ...new function.
11394 (pass_scev_cprop): Replace declaration with that of...
11395 (make_pass_scev_cprop): ...new function.
11396 (pass_empty_loop): Replace declaration with that of...
11397 (make_pass_empty_loop): ...new function.
11398 (pass_record_bounds): Replace declaration with that of...
11399 (make_pass_record_bounds): ...new function.
11400 (pass_graphite): Replace declaration with that of...
11401 (make_pass_graphite): ...new function.
11402 (pass_graphite_transforms): Replace declaration with that of...
11403 (make_pass_graphite_transforms): ...new function.
11404 (pass_if_conversion): Replace declaration with that of...
11405 (make_pass_if_conversion): ...new function.
11406 (pass_loop_distribution): Replace declaration with that of...
11407 (make_pass_loop_distribution): ...new function.
11408 (pass_vectorize): Replace declaration with that of...
11409 (make_pass_vectorize): ...new function.
11410 (pass_slp_vectorize): Replace declaration with that of...
11411 (make_pass_slp_vectorize): ...new function.
11412 (pass_complete_unroll): Replace declaration with that of...
11413 (make_pass_complete_unroll): ...new function.
11414 (pass_complete_unrolli): Replace declaration with that of...
11415 (make_pass_complete_unrolli): ...new function.
11416 (pass_parallelize_loops): Replace declaration with that of...
11417 (make_pass_parallelize_loops): ...new function.
11418 (pass_loop_prefetch): Replace declaration with that of...
11419 (make_pass_loop_prefetch): ...new function.
11420 (pass_iv_optimize): Replace declaration with that of...
11421 (make_pass_iv_optimize): ...new function.
11422 (pass_tree_loop_done): Replace declaration with that of...
11423 (make_pass_tree_loop_done): ...new function.
11424 (pass_ch): Replace declaration with that of...
11425 (make_pass_ch): ...new function.
11426 (pass_ccp): Replace declaration with that of...
11427 (make_pass_ccp): ...new function.
11428 (pass_phi_only_cprop): Replace declaration with that of...
11429 (make_pass_phi_only_cprop): ...new function.
11430 (pass_build_ssa): Replace declaration with that of...
11431 (make_pass_build_ssa): ...new function.
11432 (pass_build_alias): Replace declaration with that of...
11433 (make_pass_build_alias): ...new function.
11434 (pass_build_ealias): Replace declaration with that of...
11435 (make_pass_build_ealias): ...new function.
11436 (pass_dominator): Replace declaration with that of...
11437 (make_pass_dominator): ...new function.
11438 (pass_dce): Replace declaration with that of...
11439 (make_pass_dce): ...new function.
11440 (pass_dce_loop): Replace declaration with that of...
11441 (make_pass_dce_loop): ...new function.
11442 (pass_cd_dce): Replace declaration with that of...
11443 (make_pass_cd_dce): ...new function.
11444 (pass_call_cdce): Replace declaration with that of...
11445 (make_pass_call_cdce): ...new function.
11446 (pass_merge_phi): Replace declaration with that of...
11447 (make_pass_merge_phi): ...new function.
11448 (pass_split_crit_edges): Replace declaration with that of...
11449 (make_pass_split_crit_edges): ...new function.
11450 (pass_pre): Replace declaration with that of...
11451 (make_pass_pre): ...new function.
11452 (pass_profile): Replace declaration with that of...
11453 (make_pass_profile): ...new function.
11454 (pass_strip_predict_hints): Replace declaration with that of...
11455 (make_pass_strip_predict_hints): ...new function.
11456 (pass_lower_complex_O0): Replace declaration with that of...
11457 (make_pass_lower_complex_O0): ...new function.
11458 (pass_lower_complex): Replace declaration with that of...
11459 (make_pass_lower_complex): ...new function.
11460 (pass_lower_vector): Replace declaration with that of...
11461 (make_pass_lower_vector): ...new function.
11462 (pass_lower_vector_ssa): Replace declaration with that of...
11463 (make_pass_lower_vector_ssa): ...new function.
11464 (pass_lower_omp): Replace declaration with that of...
11465 (make_pass_lower_omp): ...new function.
11466 (pass_diagnose_omp_blocks): Replace declaration with that of...
11467 (make_pass_diagnose_omp_blocks): ...new function.
11468 (pass_expand_omp): Replace declaration with that of...
11469 (make_pass_expand_omp): ...new function.
11470 (pass_expand_omp_ssa): Replace declaration with that of...
11471 (make_pass_expand_omp_ssa): ...new function.
11472 (pass_object_sizes): Replace declaration with that of...
11473 (make_pass_object_sizes): ...new function.
11474 (pass_strlen): Replace declaration with that of...
11475 (make_pass_strlen): ...new function.
11476 (pass_fold_builtins): Replace declaration with that of...
11477 (make_pass_fold_builtins): ...new function.
11478 (pass_stdarg): Replace declaration with that of...
11479 (make_pass_stdarg): ...new function.
11480 (pass_early_warn_uninitialized): Replace declaration with that of...
11481 (make_pass_early_warn_uninitialized): ...new function.
11482 (pass_late_warn_uninitialized): Replace declaration with that of...
11483 (make_pass_late_warn_uninitialized): ...new function.
11484 (pass_cse_reciprocals): Replace declaration with that of...
11485 (make_pass_cse_reciprocals): ...new function.
11486 (pass_cse_sincos): Replace declaration with that of...
11487 (make_pass_cse_sincos): ...new function.
11488 (pass_optimize_bswap): Replace declaration with that of...
11489 (make_pass_optimize_bswap): ...new function.
11490 (pass_optimize_widening_mul): Replace declaration with that of...
11491 (make_pass_optimize_widening_mul): ...new function.
11492 (pass_warn_function_return): Replace declaration with that of...
11493 (make_pass_warn_function_return): ...new function.
11494 (pass_warn_function_noreturn): Replace declaration with that of...
11495 (make_pass_warn_function_noreturn): ...new function.
11496 (pass_cselim): Replace declaration with that of...
11497 (make_pass_cselim): ...new function.
11498 (pass_phiopt): Replace declaration with that of...
11499 (make_pass_phiopt): ...new function.
11500 (pass_forwprop): Replace declaration with that of...
11501 (make_pass_forwprop): ...new function.
11502 (pass_phiprop): Replace declaration with that of...
11503 (make_pass_phiprop): ...new function.
11504 (pass_tree_ifcombine): Replace declaration with that of...
11505 (make_pass_tree_ifcombine): ...new function.
11506 (pass_dse): Replace declaration with that of...
11507 (make_pass_dse): ...new function.
11508 (pass_nrv): Replace declaration with that of...
11509 (make_pass_nrv): ...new function.
11510 (pass_rename_ssa_copies): Replace declaration with that of...
11511 (make_pass_rename_ssa_copies): ...new function.
11512 (pass_sink_code): Replace declaration with that of...
11513 (make_pass_sink_code): ...new function.
11514 (pass_fre): Replace declaration with that of...
11515 (make_pass_fre): ...new function.
11516 (pass_check_data_deps): Replace declaration with that of...
11517 (make_pass_check_data_deps): ...new function.
11518 (pass_copy_prop): Replace declaration with that of...
11519 (make_pass_copy_prop): ...new function.
11520 (pass_vrp): Replace declaration with that of...
11521 (make_pass_vrp): ...new function.
11522 (pass_uncprop): Replace declaration with that of...
11523 (make_pass_uncprop): ...new function.
11524 (pass_return_slot): Replace declaration with that of...
11525 (make_pass_return_slot): ...new function.
11526 (pass_reassoc): Replace declaration with that of...
11527 (make_pass_reassoc): ...new function.
11528 (pass_rebuild_cgraph_edges): Replace declaration with that of...
11529 (make_pass_rebuild_cgraph_edges): ...new function.
11530 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
11531 (make_pass_remove_cgraph_callee_edges): ...new function.
11532 (pass_build_cgraph_edges): Replace declaration with that of...
11533 (make_pass_build_cgraph_edges): ...new function.
11534 (pass_local_pure_const): Replace declaration with that of...
11535 (make_pass_local_pure_const): ...new function.
11536 (pass_tracer): Replace declaration with that of...
11537 (make_pass_tracer): ...new function.
11538 (pass_warn_unused_result): Replace declaration with that of...
11539 (make_pass_warn_unused_result): ...new function.
11540 (pass_diagnose_tm_blocks): Replace declaration with that of...
11541 (make_pass_diagnose_tm_blocks): ...new function.
11542 (pass_lower_tm): Replace declaration with that of...
11543 (make_pass_lower_tm): ...new function.
11544 (pass_tm_init): Replace declaration with that of...
11545 (make_pass_tm_init): ...new function.
11546 (pass_tm_mark): Replace declaration with that of...
11547 (make_pass_tm_mark): ...new function.
11548 (pass_tm_memopt): Replace declaration with that of...
11549 (make_pass_tm_memopt): ...new function.
11550 (pass_tm_edges): Replace declaration with that of...
11551 (make_pass_tm_edges): ...new function.
11552 (pass_split_functions): Replace declaration with that of...
11553 (make_pass_split_functions): ...new function.
11554 (pass_feedback_split_functions): Replace declaration with that of...
11555 (make_pass_feedback_split_functions): ...new function.
11556 (pass_strength_reduction): Replace declaration with that of...
11557 (make_pass_strength_reduction): ...new function.
11558 (pass_ipa_lower_emutls): Replace declaration with that of...
11559 (make_pass_ipa_lower_emutls): ...new function.
11560 (pass_ipa_function_and_variable_visibility): Replace declaration with
11562 (make_pass_ipa_function_and_variable_visibility): ...new function.
11563 (pass_ipa_tree_profile): Replace declaration with that of...
11564 (make_pass_ipa_tree_profile): ...new function.
11565 (pass_early_local_passes): Replace declaration with that of...
11566 (make_pass_early_local_passes): ...new function.
11567 (pass_ipa_whole_program_visibility): Replace declaration with that
11569 (make_pass_ipa_whole_program_visibility): ...new function.
11570 (pass_ipa_lto_gimple_out): Replace declaration with that of...
11571 (make_pass_ipa_lto_gimple_out): ...new function.
11572 (pass_ipa_increase_alignment): Replace declaration with that of...
11573 (make_pass_ipa_increase_alignment): ...new function.
11574 (pass_ipa_inline): Replace declaration with that of...
11575 (make_pass_ipa_inline): ...new function.
11576 (pass_ipa_free_lang_data): Replace declaration with that of...
11577 (make_pass_ipa_free_lang_data): ...new function.
11578 (pass_ipa_free_inline_summary): Replace declaration with that of...
11579 (make_pass_ipa_free_inline_summary): ...new function.
11580 (pass_ipa_cp): Replace declaration with that of...
11581 (make_pass_ipa_cp): ...new function.
11582 (pass_ipa_reference): Replace declaration with that of...
11583 (make_pass_ipa_reference): ...new function.
11584 (pass_ipa_pure_const): Replace declaration with that of...
11585 (make_pass_ipa_pure_const): ...new function.
11586 (pass_ipa_pta): Replace declaration with that of...
11587 (make_pass_ipa_pta): ...new function.
11588 (pass_ipa_lto_finish_out): Replace declaration with that of...
11589 (make_pass_ipa_lto_finish_out): ...new function.
11590 (pass_ipa_tm): Replace declaration with that of...
11591 (make_pass_ipa_tm): ...new function.
11592 (pass_ipa_profile): Replace declaration with that of...
11593 (make_pass_ipa_profile): ...new function.
11594 (pass_ipa_cdtor_merge): Replace declaration with that of...
11595 (make_pass_ipa_cdtor_merge): ...new function.
11596 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
11598 (make_pass_cleanup_cfg_post_optimizing): ...new function.
11599 (pass_init_datastructures): Replace declaration with that of...
11600 (make_pass_init_datastructures): ...new function.
11601 (pass_fixup_cfg): Replace declaration with that of...
11602 (make_pass_fixup_cfg): ...new function.
11603 (pass_expand): Replace declaration with that of...
11604 (make_pass_expand): ...new function.
11605 (pass_instantiate_virtual_regs): Replace declaration with that of...
11606 (make_pass_instantiate_virtual_regs): ...new function.
11607 (pass_rtl_fwprop): Replace declaration with that of...
11608 (make_pass_rtl_fwprop): ...new function.
11609 (pass_rtl_fwprop_addr): Replace declaration with that of...
11610 (make_pass_rtl_fwprop_addr): ...new function.
11611 (pass_jump): Replace declaration with that of...
11612 (make_pass_jump): ...new function.
11613 (pass_jump2): Replace declaration with that of...
11614 (make_pass_jump2): ...new function.
11615 (pass_lower_subreg): Replace declaration with that of...
11616 (make_pass_lower_subreg): ...new function.
11617 (pass_cse): Replace declaration with that of...
11618 (make_pass_cse): ...new function.
11619 (pass_fast_rtl_dce): Replace declaration with that of...
11620 (make_pass_fast_rtl_dce): ...new function.
11621 (pass_ud_rtl_dce): Replace declaration with that of...
11622 (make_pass_ud_rtl_dce): ...new function.
11623 (pass_rtl_dce): Replace declaration with that of...
11624 (make_pass_rtl_dce): ...new function.
11625 (pass_rtl_dse1): Replace declaration with that of...
11626 (make_pass_rtl_dse1): ...new function.
11627 (pass_rtl_dse2): Replace declaration with that of...
11628 (make_pass_rtl_dse2): ...new function.
11629 (pass_rtl_dse3): Replace declaration with that of...
11630 (make_pass_rtl_dse3): ...new function.
11631 (pass_rtl_cprop): Replace declaration with that of...
11632 (make_pass_rtl_cprop): ...new function.
11633 (pass_rtl_pre): Replace declaration with that of...
11634 (make_pass_rtl_pre): ...new function.
11635 (pass_rtl_hoist): Replace declaration with that of...
11636 (make_pass_rtl_hoist): ...new function.
11637 (pass_rtl_store_motion): Replace declaration with that of...
11638 (make_pass_rtl_store_motion): ...new function.
11639 (pass_cse_after_global_opts): Replace declaration with that of...
11640 (make_pass_cse_after_global_opts): ...new function.
11641 (pass_rtl_ifcvt): Replace declaration with that of...
11642 (make_pass_rtl_ifcvt): ...new function.
11643 (pass_into_cfg_layout_mode): Replace declaration with that of...
11644 (make_pass_into_cfg_layout_mode): ...new function.
11645 (pass_outof_cfg_layout_mode): Replace declaration with that of...
11646 (make_pass_outof_cfg_layout_mode): ...new function.
11647 (pass_loop2): Replace declaration with that of...
11648 (make_pass_loop2): ...new function.
11649 (pass_rtl_loop_init): Replace declaration with that of...
11650 (make_pass_rtl_loop_init): ...new function.
11651 (pass_rtl_move_loop_invariants): Replace declaration with that of...
11652 (make_pass_rtl_move_loop_invariants): ...new function.
11653 (pass_rtl_unswitch): Replace declaration with that of...
11654 (make_pass_rtl_unswitch): ...new function.
11655 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
11656 (make_pass_rtl_unroll_and_peel_loops): ...new function.
11657 (pass_rtl_doloop): Replace declaration with that of...
11658 (make_pass_rtl_doloop): ...new function.
11659 (pass_rtl_loop_done): Replace declaration with that of...
11660 (make_pass_rtl_loop_done): ...new function.
11661 (pass_web): Replace declaration with that of...
11662 (make_pass_web): ...new function.
11663 (pass_cse2): Replace declaration with that of...
11664 (make_pass_cse2): ...new function.
11665 (pass_df_initialize_opt): Replace declaration with that of...
11666 (make_pass_df_initialize_opt): ...new function.
11667 (pass_df_initialize_no_opt): Replace declaration with that of...
11668 (make_pass_df_initialize_no_opt): ...new function.
11669 (pass_reginfo_init): Replace declaration with that of...
11670 (make_pass_reginfo_init): ...new function.
11671 (pass_inc_dec): Replace declaration with that of...
11672 (make_pass_inc_dec): ...new function.
11673 (pass_stack_ptr_mod): Replace declaration with that of...
11674 (make_pass_stack_ptr_mod): ...new function.
11675 (pass_initialize_regs): Replace declaration with that of...
11676 (make_pass_initialize_regs): ...new function.
11677 (pass_combine): Replace declaration with that of...
11678 (make_pass_combine): ...new function.
11679 (pass_if_after_combine): Replace declaration with that of...
11680 (make_pass_if_after_combine): ...new function.
11681 (pass_ree): Replace declaration with that of...
11682 (make_pass_ree): ...new function.
11683 (pass_partition_blocks): Replace declaration with that of...
11684 (make_pass_partition_blocks): ...new function.
11685 (pass_match_asm_constraints): Replace declaration with that of...
11686 (make_pass_match_asm_constraints): ...new function.
11687 (pass_regmove): Replace declaration with that of...
11688 (make_pass_regmove): ...new function.
11689 (pass_split_all_insns): Replace declaration with that of...
11690 (make_pass_split_all_insns): ...new function.
11691 (pass_fast_rtl_byte_dce): Replace declaration with that of...
11692 (make_pass_fast_rtl_byte_dce): ...new function.
11693 (pass_lower_subreg2): Replace declaration with that of...
11694 (make_pass_lower_subreg2): ...new function.
11695 (pass_mode_switching): Replace declaration with that of...
11696 (make_pass_mode_switching): ...new function.
11697 (pass_sms): Replace declaration with that of...
11698 (make_pass_sms): ...new function.
11699 (pass_sched): Replace declaration with that of...
11700 (make_pass_sched): ...new function.
11701 (pass_ira): Replace declaration with that of...
11702 (make_pass_ira): ...new function.
11703 (pass_reload): Replace declaration with that of...
11704 (make_pass_reload): ...new function.
11705 (pass_clean_state): Replace declaration with that of...
11706 (make_pass_clean_state): ...new function.
11707 (pass_branch_prob): Replace declaration with that of...
11708 (make_pass_branch_prob): ...new function.
11709 (pass_value_profile_transformations): Replace declaration with that
11711 (make_pass_value_profile_transformations): ...new function.
11712 (pass_postreload_cse): Replace declaration with that of...
11713 (make_pass_postreload_cse): ...new function.
11714 (pass_gcse2): Replace declaration with that of...
11715 (make_pass_gcse2): ...new function.
11716 (pass_split_after_reload): Replace declaration with that of...
11717 (make_pass_split_after_reload): ...new function.
11718 (pass_branch_target_load_optimize1): Replace declaration with that
11720 (make_pass_branch_target_load_optimize1): ...new function.
11721 (pass_thread_prologue_and_epilogue): Replace declaration with that
11723 (make_pass_thread_prologue_and_epilogue): ...new function.
11724 (pass_stack_adjustments): Replace declaration with that of...
11725 (make_pass_stack_adjustments): ...new function.
11726 (pass_peephole2): Replace declaration with that of...
11727 (make_pass_peephole2): ...new function.
11728 (pass_if_after_reload): Replace declaration with that of...
11729 (make_pass_if_after_reload): ...new function.
11730 (pass_regrename): Replace declaration with that of...
11731 (make_pass_regrename): ...new function.
11732 (pass_cprop_hardreg): Replace declaration with that of...
11733 (make_pass_cprop_hardreg): ...new function.
11734 (pass_reorder_blocks): Replace declaration with that of...
11735 (make_pass_reorder_blocks): ...new function.
11736 (pass_branch_target_load_optimize2): Replace declaration with that
11738 (make_pass_branch_target_load_optimize2): ...new function.
11739 (pass_leaf_regs): Replace declaration with that of...
11740 (make_pass_leaf_regs): ...new function.
11741 (pass_split_before_sched2): Replace declaration with that of...
11742 (make_pass_split_before_sched2): ...new function.
11743 (pass_compare_elim_after_reload): Replace declaration with that of...
11744 (make_pass_compare_elim_after_reload): ...new function.
11745 (pass_sched2): Replace declaration with that of...
11746 (make_pass_sched2): ...new function.
11747 (pass_stack_regs): Replace declaration with that of...
11748 (make_pass_stack_regs): ...new function.
11749 (pass_stack_regs_run): Replace declaration with that of...
11750 (make_pass_stack_regs_run): ...new function.
11751 (pass_df_finish): Replace declaration with that of...
11752 (make_pass_df_finish): ...new function.
11753 (pass_compute_alignments): Replace declaration with that of...
11754 (make_pass_compute_alignments): ...new function.
11755 (pass_duplicate_computed_gotos): Replace declaration with that of...
11756 (make_pass_duplicate_computed_gotos): ...new function.
11757 (pass_variable_tracking): Replace declaration with that of...
11758 (make_pass_variable_tracking): ...new function.
11759 (pass_free_cfg): Replace declaration with that of...
11760 (make_pass_free_cfg): ...new function.
11761 (pass_machine_reorg): Replace declaration with that of...
11762 (make_pass_machine_reorg): ...new function.
11763 (pass_cleanup_barriers): Replace declaration with that of...
11764 (make_pass_cleanup_barriers): ...new function.
11765 (pass_delay_slots): Replace declaration with that of...
11766 (make_pass_delay_slots): ...new function.
11767 (pass_split_for_shorten_branches): Replace declaration with that of...
11768 (make_pass_split_for_shorten_branches): ...new function.
11769 (pass_split_before_regstack): Replace declaration with that of...
11770 (make_pass_split_before_regstack): ...new function.
11771 (pass_convert_to_eh_region_ranges): Replace declaration with that
11773 (make_pass_convert_to_eh_region_ranges): ...new function.
11774 (pass_shorten_branches): Replace declaration with that of...
11775 (make_pass_shorten_branches): ...new function.
11776 (pass_set_nothrow_function_flags): Replace declaration with that of...
11777 (make_pass_set_nothrow_function_flags): ...new function.
11778 (pass_dwarf2_frame): Replace declaration with that of...
11779 (make_pass_dwarf2_frame): ...new function.
11780 (pass_final): Replace declaration with that of...
11781 (make_pass_final): ...new function.
11782 (pass_rtl_seqabstr): Replace declaration with that of...
11783 (make_pass_rtl_seqabstr): ...new function.
11784 (pass_release_ssa_names): Replace declaration with that of...
11785 (make_pass_release_ssa_names): ...new function.
11786 (pass_early_inline): Replace declaration with that of...
11787 (make_pass_early_inline): ...new function.
11788 (pass_inline_parameters): Replace declaration with that of...
11789 (make_pass_inline_parameters): ...new function.
11790 (pass_update_address_taken): Replace declaration with that of...
11791 (make_pass_update_address_taken): ...new function.
11792 (pass_convert_switch): Replace declaration with that of...
11793 (make_pass_convert_switch): ...new function.
11794 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
11795 to a subclass of simple_ipa_opt_pass along with...
11796 (pass_data_ipa_tree_profile): ...new pass_data instance and...
11797 (make_pass_ipa_tree_profile): ...new function.
11798 * tree-sra.c (pass_sra_early): Convert from a global struct to a
11799 subclass of gimple_opt_pass along with...
11800 (pass_data_sra_early): ...new pass_data instance and...
11801 (make_pass_sra_early): ...new function.
11802 (pass_sra): Convert from a global struct to a subclass of
11803 gimple_opt_pass along with...
11804 (pass_data_sra): ...new pass_data instance and...
11805 (make_pass_sra): ...new function.
11806 (pass_early_ipa_sra): Convert from a global struct to a subclass of
11807 gimple_opt_pass along with...
11808 (pass_data_early_ipa_sra): ...new pass_data instance and...
11809 (make_pass_early_ipa_sra): ...new function.
11810 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
11811 subclass of gimple_opt_pass along with...
11812 (pass_data_ccp): ...new pass_data instance and...
11813 (make_pass_ccp): ...new function.
11814 (pass_fold_builtins): Convert from a global struct to a subclass of
11815 gimple_opt_pass along with...
11816 (pass_data_fold_builtins): ...new pass_data instance and...
11817 (make_pass_fold_builtins): ...new function.
11818 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
11819 subclass of gimple_opt_pass along with...
11820 (pass_data_copy_prop): ...new pass_data instance and...
11821 (make_pass_copy_prop): ...new function.
11822 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
11823 global struct to a subclass of gimple_opt_pass along with...
11824 (pass_data_rename_ssa_copies): ...new pass_data instance and...
11825 (make_pass_rename_ssa_copies): ...new function.
11826 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
11827 subclass of gimple_opt_pass along with...
11828 (pass_data_dce): ...new pass_data instance and...
11829 (make_pass_dce): ...new function.
11830 (pass_dce_loop): Convert from a global struct to a subclass of
11831 gimple_opt_pass along with...
11832 (pass_data_dce_loop): ...new pass_data instance and...
11833 (make_pass_dce_loop): ...new function.
11834 (pass_cd_dce): Convert from a global struct to a subclass of
11835 gimple_opt_pass along with...
11836 (pass_data_cd_dce): ...new pass_data instance and...
11837 (make_pass_cd_dce): ...new function.
11838 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
11839 subclass of gimple_opt_pass along with...
11840 (pass_data_dominator): ...new pass_data instance and...
11841 (make_pass_dominator): ...new function.
11842 (pass_phi_only_cprop): Convert from a global struct to a subclass of
11843 gimple_opt_pass along with...
11844 (pass_data_phi_only_cprop): ...new pass_data instance and...
11845 (make_pass_phi_only_cprop): ...new function.
11846 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
11847 subclass of gimple_opt_pass along with...
11848 (pass_data_dse): ...new pass_data instance and...
11849 (make_pass_dse): ...new function.
11850 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
11851 a subclass of gimple_opt_pass along with...
11852 (pass_data_forwprop): ...new pass_data instance and...
11853 (make_pass_forwprop): ...new function.
11854 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
11855 struct to a subclass of gimple_opt_pass along with...
11856 (pass_data_tree_ifcombine): ...new pass_data instance and...
11857 (make_pass_tree_ifcombine): ...new function.
11858 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
11859 subclass of gimple_opt_pass along with...
11860 (pass_data_ch): ...new pass_data instance and...
11861 (make_pass_ch): ...new function.
11862 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
11863 subclass of gimple_opt_pass along with...
11864 (pass_data_tree_loop): ...new pass_data instance and...
11865 (make_pass_tree_loop): ...new function.
11866 (pass_tree_loop_init): Convert from a global struct to a subclass of
11867 gimple_opt_pass along with...
11868 (pass_data_tree_loop_init): ...new pass_data instance and...
11869 (make_pass_tree_loop_init): ...new function.
11870 (pass_lim): Convert from a global struct to a subclass of
11871 gimple_opt_pass along with...
11872 (pass_data_lim): ...new pass_data instance and...
11873 (make_pass_lim): ...new function.
11874 (pass_tree_unswitch): Convert from a global struct to a subclass of
11875 gimple_opt_pass along with...
11876 (pass_data_tree_unswitch): ...new pass_data instance and...
11877 (make_pass_tree_unswitch): ...new function.
11878 (pass_predcom): Convert from a global struct to a subclass of
11879 gimple_opt_pass along with...
11880 (pass_data_predcom): ...new pass_data instance and...
11881 (make_pass_predcom): ...new function.
11882 (pass_vectorize): Convert from a global struct to a subclass of
11883 gimple_opt_pass along with...
11884 (pass_data_vectorize): ...new pass_data instance and...
11885 (make_pass_vectorize): ...new function.
11886 (pass_graphite): Convert from a global struct to a subclass of
11887 gimple_opt_pass along with...
11888 (pass_data_graphite): ...new pass_data instance and...
11889 (make_pass_graphite): ...new function.
11890 (pass_graphite_transforms): Convert from a global struct to a subclass
11891 of gimple_opt_pass along with...
11892 (pass_data_graphite_transforms): ...new pass_data instance and...
11893 (make_pass_graphite_transforms): ...new function.
11894 (pass_check_data_deps): Convert from a global struct to a subclass of
11895 gimple_opt_pass along with...
11896 (pass_data_check_data_deps): ...new pass_data instance and...
11897 (make_pass_check_data_deps): ...new function.
11898 (pass_iv_canon): Convert from a global struct to a subclass of
11899 gimple_opt_pass along with...
11900 (pass_data_iv_canon): ...new pass_data instance and...
11901 (make_pass_iv_canon): ...new function.
11902 (pass_scev_cprop): Convert from a global struct to a subclass of
11903 gimple_opt_pass along with...
11904 (pass_data_scev_cprop): ...new pass_data instance and...
11905 (make_pass_scev_cprop): ...new function.
11906 (pass_record_bounds): Convert from a global struct to a subclass of
11907 gimple_opt_pass along with...
11908 (pass_data_record_bounds): ...new pass_data instance and...
11909 (make_pass_record_bounds): ...new function.
11910 (pass_complete_unroll): Convert from a global struct to a subclass of
11911 gimple_opt_pass along with...
11912 (pass_data_complete_unroll): ...new pass_data instance and...
11913 (make_pass_complete_unroll): ...new function.
11914 (pass_complete_unrolli): Convert from a global struct to a subclass of
11915 gimple_opt_pass along with...
11916 (pass_data_complete_unrolli): ...new pass_data instance and...
11917 (make_pass_complete_unrolli): ...new function.
11918 (pass_parallelize_loops): Convert from a global struct to a subclass
11919 of gimple_opt_pass along with...
11920 (pass_data_parallelize_loops): ...new pass_data instance and...
11921 (make_pass_parallelize_loops): ...new function.
11922 (pass_loop_prefetch): Convert from a global struct to a subclass of
11923 gimple_opt_pass along with...
11924 (pass_data_loop_prefetch): ...new pass_data instance and...
11925 (make_pass_loop_prefetch): ...new function.
11926 (pass_iv_optimize): Convert from a global struct to a subclass of
11927 gimple_opt_pass along with...
11928 (pass_data_iv_optimize): ...new pass_data instance and...
11929 (make_pass_iv_optimize): ...new function.
11930 (pass_tree_loop_done): Convert from a global struct to a subclass of
11931 gimple_opt_pass along with...
11932 (pass_data_tree_loop_done): ...new pass_data instance and...
11933 (make_pass_tree_loop_done): ...new function.
11934 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
11935 struct to a subclass of gimple_opt_pass along with...
11936 (pass_data_cse_reciprocals): ...new pass_data instance and...
11937 (make_pass_cse_reciprocals): ...new function.
11938 (pass_cse_sincos): Convert from a global struct to a subclass of
11939 gimple_opt_pass along with...
11940 (pass_data_cse_sincos): ...new pass_data instance and...
11941 (make_pass_cse_sincos): ...new function.
11942 (pass_optimize_bswap): Convert from a global struct to a subclass of
11943 gimple_opt_pass along with...
11944 (pass_data_optimize_bswap): ...new pass_data instance and...
11945 (make_pass_optimize_bswap): ...new function.
11946 (pass_optimize_widening_mul): Convert from a global struct to a
11947 subclass of gimple_opt_pass along with...
11948 (pass_data_optimize_widening_mul): ...new pass_data instance and...
11949 (make_pass_optimize_widening_mul): ...new function.
11950 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
11951 subclass of gimple_opt_pass along with...
11952 (pass_data_phiopt): ...new pass_data instance and...
11953 (make_pass_phiopt): ...new function.
11954 (pass_cselim): Convert from a global struct to a subclass of
11955 gimple_opt_pass along with...
11956 (pass_data_cselim): ...new pass_data instance and...
11957 (make_pass_cselim): ...new function.
11958 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
11959 subclass of gimple_opt_pass along with...
11960 (pass_data_phiprop): ...new pass_data instance and...
11961 (make_pass_phiprop): ...new function.
11962 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
11963 subclass of gimple_opt_pass along with...
11964 (pass_data_pre): ...new pass_data instance and...
11965 (make_pass_pre): ...new function.
11966 (pass_fre): Convert from a global struct to a subclass of
11967 gimple_opt_pass along with...
11968 (pass_data_fre): ...new pass_data instance and...
11969 (make_pass_fre): ...new function.
11970 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
11971 subclass of gimple_opt_pass along with...
11972 (pass_data_reassoc): ...new pass_data instance and...
11973 (make_pass_reassoc): ...new function.
11974 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
11975 subclass of gimple_opt_pass along with...
11976 (pass_data_sink_code): ...new pass_data instance and...
11977 (make_pass_sink_code): ...new function.
11978 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
11979 subclass of gimple_opt_pass along with...
11980 (pass_data_strlen): ...new pass_data instance and...
11981 (make_pass_strlen): ...new function.
11982 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
11983 struct to a subclass of gimple_opt_pass along with...
11984 (pass_data_build_alias): ...new pass_data instance and...
11985 (make_pass_build_alias): ...new function.
11986 (pass_build_ealias): Convert from a global struct to a subclass of
11987 gimple_opt_pass along with...
11988 (pass_data_build_ealias): ...new pass_data instance and...
11989 (make_pass_build_ealias): ...new function.
11990 (pass_ipa_pta): Convert from a global struct to a subclass of
11991 simple_ipa_opt_pass along with...
11992 (pass_data_ipa_pta): ...new pass_data instance and...
11993 (make_pass_ipa_pta): ...new function.
11994 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
11995 subclass of gimple_opt_pass along with...
11996 (pass_data_uncprop): ...new pass_data instance and...
11997 (make_pass_uncprop): ...new function.
11998 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
11999 global struct to a subclass of gimple_opt_pass along with...
12000 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
12001 (make_pass_late_warn_uninitialized): ...new function.
12002 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
12003 to a subclass of gimple_opt_pass along with...
12004 (pass_data_init_datastructures): ...new pass_data instance and...
12005 (make_pass_init_datastructures): ...new function.
12006 (pass_early_warn_uninitialized): Convert from a global struct to a
12007 subclass of gimple_opt_pass along with...
12008 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
12009 (make_pass_early_warn_uninitialized): ...new function.
12010 (pass_update_address_taken): Convert from a global struct to a
12011 subclass of gimple_opt_pass along with...
12012 (pass_data_update_address_taken): ...new pass_data instance and...
12013 (make_pass_update_address_taken): ...new function.
12014 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
12015 struct to a subclass of gimple_opt_pass along with...
12016 (pass_data_release_ssa_names): ...new pass_data instance and...
12017 (make_pass_release_ssa_names): ...new function.
12018 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
12019 subclass of gimple_opt_pass along with...
12020 (pass_data_stdarg): ...new pass_data instance and...
12021 (make_pass_stdarg): ...new function.
12022 * tree-switch-conversion.c (pass_convert_switch): Convert from a
12023 global struct to a subclass of gimple_opt_pass along with...
12024 (pass_data_convert_switch): ...new pass_data instance and...
12025 (make_pass_convert_switch): ...new function.
12026 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
12027 to a subclass of gimple_opt_pass along with...
12028 (pass_data_tail_recursion): ...new pass_data instance and...
12029 (make_pass_tail_recursion): ...new function.
12030 (pass_tail_calls): Convert from a global struct to a subclass of
12031 gimple_opt_pass along with...
12032 (pass_data_tail_calls): ...new pass_data instance and...
12033 (make_pass_tail_calls): ...new function.
12034 * tree-vect-generic.c (pass_lower_vector): Convert from a global
12035 struct to a subclass of gimple_opt_pass along with...
12036 (pass_data_lower_vector): ...new pass_data instance and...
12037 (make_pass_lower_vector): ...new function.
12038 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
12039 gimple_opt_pass along with...
12040 (pass_data_lower_vector_ssa): ...new pass_data instance and...
12041 (make_pass_lower_vector_ssa): ...new function.
12042 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
12043 to a subclass of gimple_opt_pass along with...
12044 (pass_data_slp_vectorize): ...new pass_data instance and...
12045 (make_pass_slp_vectorize): ...new function.
12046 (pass_ipa_increase_alignment): Convert from a global struct to a
12047 subclass of simple_ipa_opt_pass along with...
12048 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
12049 (make_pass_ipa_increase_alignment): ...new function.
12050 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
12051 gimple_opt_pass along with...
12052 (pass_data_vrp): ...new pass_data instance and...
12053 (make_pass_vrp): ...new function.
12054 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
12055 subclass of simple_ipa_opt_pass along with...
12056 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
12057 (make_pass_ipa_free_lang_data): ...new function.
12058 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
12059 gimple_opt_pass along with...
12060 (pass_data_tsan): ...new pass_data instance and...
12061 (make_pass_tsan): ...new function.
12062 (pass_tsan_O0): Convert from a global struct to a subclass of
12063 gimple_opt_pass along with...
12064 (pass_data_tsan_O0): ...new pass_data instance and...
12065 (make_pass_tsan_O0): ...new function.
12066 * var-tracking.c (pass_variable_tracking): Convert from a global
12067 struct to a subclass of rtl_opt_pass along with...
12068 (pass_data_variable_tracking): ...new pass_data instance and...
12069 (make_pass_variable_tracking): ...new function.
12070 * web.c (pass_web): Convert from a global struct to a subclass of
12071 rtl_opt_pass along with...
12072 (pass_data_web): ...new pass_data instance and...
12073 (make_pass_web): ...new function.
12074 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
12075 declaration with that of...
12076 (make_pass_mode_switch_use): ...new function.
12077 (pass_resolve_sw_modes): Replace declaration with that of...
12078 (make_pass_resolve_sw_modes): ...new function.
12079 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
12080 from a global struct to a subclass of rtl_opt_pass along with...
12081 (pass_data_mode_switch_use): ...new pass_data instance and...
12082 (make_pass_mode_switch_use): ...new function.
12083 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
12084 from a global struct to a subclass of rtl_opt_pass along with...
12085 (pass_data_resolve_sw_modes): ...new pass_data instance and...
12086 (make_pass_resolve_sw_modes): ...new function.
12087 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
12088 struct to a subclass of rtl_opt_pass along with...
12089 (pass_data_insert_vzeroupper): ...new pass_data instance and...
12090 (make_pass_insert_vzeroupper): ...new function.
12091 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
12092 global struct to a subclass of rtl_opt_pass along with...
12093 (pass_data_work_around_errata): ...new pass_data instance and...
12094 (make_pass_work_around_errata): ...new function.
12095 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
12096 struct to a subclass of rtl_opt_pass along with...
12097 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
12098 (make_pass_mips_machine_reorg2): ...new function.
12100 2013-08-05 David Malcolm <dmalcolm@redhat.com>
12102 * passes.c (pass_manager::operator new): New.
12104 2013-08-05 David Malcolm <dmalcolm@redhat.com>
12106 Handwritten part of conversion of passes to C++ classes.
12108 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
12109 (toplev.o): Add dep on PASS_MANAGER_H.
12110 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
12111 of early local pases to reflect this moving from a global to a
12112 member of gcc::pass_manager.
12113 (cgraph_add_new_function): Likewise.
12114 * lto-cgraph.c (lto_output_node): Update for conversion of
12115 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
12116 * passes.c (opt_pass::clone): New.
12117 (opt_pass::gate): New.
12118 (opt_pass::execute): New.
12119 (opt_pass::opt_pass): New.
12120 (pass_manager::execute_early_local_passes): New.
12121 (pass_manager::execute_pass_mode_switching): new.
12122 (finish_optimization_passes): Convert to...
12123 (pass_manager::finish_optimization_passes): ...this.
12124 (finish_optimization_passes): Update for conversion of passes to
12126 (register_dump_files_1): Use has_gate since we cannot portably
12127 check a vtable entry against NULL.
12128 (dump_one_pass): Likewise.
12129 (ipa_write_summaries_2): Likewise.
12130 (ipa_write_optimization_summaries_1): Likewise.
12131 (ipa_read_summaries_1): Likewise.
12132 (ipa_read_optimization_summaries_1): Likewise.
12133 (execute_ipa_stmt_fixups): Likewise.
12134 (pass_manager::pass_manager): Rewrite pass-creation, invoking
12135 pass-creation functions rather than wiring up globals, and
12136 storing the results in fields of pass_manager generated using
12137 pass-instances.def.
12138 (pass_manager::dump_profile_report): Update for conversion of
12139 passes to C++ classes.
12140 (pass_manager::execute_ipa_summary_passes): Likewise.
12141 (execute_one_ipa_transform_pass): Likewise.
12142 (execute_one_pass): Use has_gate and has_execute since we cannot
12143 portably check a vtable entry against NULL.
12144 * pass_manager.h (pass_manager::finish_optimization_passes): New.
12145 (pass_manager): Use pass-instances.def to add fields for the
12146 various pass instances.
12147 * toplev.c (finalize): Update for move of
12148 finish_optimization_passes to a method of gcc::pass_manager.
12149 * toplev.h (finish_optimization_passes): Move to method of class
12151 * tree-pass.h (struct pass_data): New.
12152 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
12153 (opt_pass::gate): Convert to virtual function.
12154 (opt_pass::~opt_pass): New.
12155 (opt_pass::clone): New.
12156 (opt_pass::execute): Convert to virtual function.
12157 (opt_pass::opt_pass): New.
12158 (opt_pass::ctxt_): new.
12159 (gimple_opt_pass): Convert to subclass of opt_pass.
12160 (gimple_opt_pass::gimple_opt_pass): New.
12161 (rtl_opt_pass): Convert to subclass of opt_pass.
12162 (rtl_opt_pass::rtl_opt_pass): New.
12163 (ipa_opt_pass_d): Convert to subclass of opt_pass.
12164 (ipa_opt_pass_d::ipa_opt_pass_d): New.
12165 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
12166 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
12167 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
12168 invocation of pass_mode_switching to reflect this moving from a
12169 global to a member of gcc::pass_manager.
12170 (ix86_option_override): Rework how pass_insert_vzeroupper is
12171 added to the pass_manager to reflect autogenerated changes.
12172 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
12174 2013-08-05 Richard Earnshaw <rearnsha@arm.com>
12176 PR rtl-optimization/57708
12177 * recog.c (peep2_find_free_register): Validate all regs in a
12180 2013-08-05 Jan Hubicka <jh@suse.cz>
12183 * cgraph.c (verify_cgraph_node): Accept local flags from other
12185 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
12186 (function_and_variable_visibility): Likewise.
12187 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
12189 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
12191 * graph.c (init_graph_slim_pretty_print): Remove.
12192 (print_graph_cfg): Do not call it. Use local pretty printer.
12193 (start_graph_dump): Likewise.
12195 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
12197 * gimple-pretty-print.c (buffer): Remove.
12198 (initialized): Likewise.
12199 (maybe_init_pretty_print): Likewise.
12200 (print_gimple_stmt): Do not call it. Use non-static local
12201 pretty_printer variable.
12202 (print_gimple_expr): Likewise.
12203 (print_gimple_seq): Likewise.
12204 (gimple_dump_bb): Likewise.
12206 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
12208 * asan.c (asan_pp): Remove.
12209 (asan_pp_initialized): Likewise.
12210 (asan_pp_initialize): Likewise.
12211 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
12212 (asan_emit_stack_protection): Tidy. Use local pretty printer.
12213 (asan_add_global): Likewise.
12215 2013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
12217 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
12218 * diagnostic.h (diagnostic_flush_buffer): Adjust.
12219 * pretty-print.c (pp_formatted_text_data): Likewise.
12220 (pp_indent): Rename from pp_base_indent.
12221 (pp_format): Rename from pp_base_format.
12222 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
12223 (pp_format_verbatim): Rename from pp_base_format_verbatim.
12224 (pp_flush): Rename from pp_base_flush.
12225 (pp_set_line_maximum_length): Rename from
12226 pp_base_set_line_maximum_length.
12227 (pp_clear_output_area): Rename from pp_base_clear_output_area.
12228 (pp_set_prefix): Rename from pp_base_set_prefix.
12229 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
12230 (pp_emit_prefix): Rename from pp_base_emit_prefix.
12231 (pp_append_text): Rename from pp_base_append_text.
12232 (pp_formatted_text): Rename from pp_base_formatted_text.
12233 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
12234 (pp_remaining_character_count_for_line): Rename from
12235 pp_base_remaining_character_count_for_line.
12236 (pp_newline): Rename from pp_base_newline.
12237 (pp_character): Rename from pp_base_character.
12238 (pp_string): Rename from pp_base_string.
12239 (pp_maybe_space): Rename from pp_base_maybe_space.
12240 * asan.c (asan_pp_string): Adjust.
12241 (asan_emit_stack_protection): Likewise.
12242 (asan_add_global): Likewise.
12243 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
12244 * tree-mudflap.c (mf_varname_tree): Likewise.
12245 * tree-pretty-print.c (pp_tree_identifier): Rename from
12246 pp_base_tree_identifier.
12247 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
12248 Declare as function.
12250 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
12252 * pretty-print.h (pp_bar_bar): New.
12253 (pp_ampersand_ampersand): Likewise.
12254 (pp_less_equal): Likewise.
12255 (pp_greater_equal): Likewise.
12256 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
12257 printer functions instead of pp_string or operators and punctuators.
12258 (dump_gimple_call): Likewise.
12259 (dump_gimple_omp_for): Likewise.
12260 (dump_gimple_transaction): Likewise.
12261 (dump_gimple_phi): Likewise.
12262 (pp_gimple_stmt_1): Likewise.
12263 * sched-vis.c (print_insn): Likewise.
12264 * tree-mudflap.c (mf_varname_tree): Likewise.
12265 * tree-pretty-print.c (dump_block_node): Likewise.
12266 (dump_generic_node): Likewise.
12268 2013-08-02 Jan Hubicka <jh@suse.cz>
12270 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
12272 * lto-streamer-out.c (tree_is_indexable): Results decls and
12273 parm decls are not indexable.
12274 (DFS_write_tree_body): Do not follow args and results.
12275 (hash_tree): Likewise.
12276 (output_functions): Rearrange so struct function is needed
12277 only when real body is output; be able to also ouptut abstract
12278 functions; output DECL_ARGUMENTS and DECL_RESULT.
12279 (lto_output): When not in WPA, ale store abstract functions.
12280 (write_symbol): Do not care about RESULT_DECL.
12281 (output_symbol_p): Handle correctly sbtract decls.
12282 * lto-streamer-in.c (input_function): Rearrange so struct
12283 function can be NULL at entry; allow streaming of
12284 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
12285 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
12286 sanity check during LTO.
12287 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
12288 RESULT_DECl and DECL_ARGUMENTS.
12289 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
12292 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
12294 * pretty-print.h (pp_underscore): New.
12296 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
12297 printer functions instead of pp_character.
12298 (dump_binary_rhs): Likewise.
12299 (dump_ternary_rhs): Likewise.
12300 (dump_gimple_call_args): Likewise.
12301 (pp_points_to_solution): Likewise.
12302 (dump_gimple_call): Likewise.
12303 (dump_gimple_switch): Likewise.
12304 (dump_gimple_cond): Likewise.
12305 (dump_gimple_bind): Likewise.
12306 (dump_gimple_try): Likewise.
12307 (dump_gimple_omp_for): Likewise.
12308 (dump_gimple_omp_continue): Likewise.
12309 (dump_gimple_omp_single): Likewise.
12310 (dump_gimple_omp_sections): Likewise.
12311 (dump_gimple_omp_block): Likewise.
12312 (dump_gimple_omp_critical): Likewise.
12313 (dump_gimple_transaction): Likewise.
12314 (dump_gimple_asm): Likewise.
12315 (dump_gimple_phi): Likewise.
12316 (dump_gimple_omp_parallel): Likewise.
12317 (dump_gimple_omp_task): Likewise.
12318 (dump_gimple_omp_atomic_load): Likewise.
12319 (dump_gimple_omp_atomic_store): Likewise.
12320 (dump_gimple_mem_ops): Likewise.
12321 (pp_gimple_stmt_1): Likewise.
12322 (pp_cfg_jump): Likewise.
12323 (dump_implicit_edges): Likewise.
12324 (gimple_dump_bb_for_graph): Likewise.
12325 * graph.c (draw_cfg_node): Likewise.
12326 * langhooks.c (lhd_print_error_function): Likewise.
12327 * sched-vis.c (print_exp): Likewise.
12328 (print_value): Likewise.
12329 (print_pattern): Likewise.
12330 (print_insn): Likewise.
12331 (rtl_dump_bb_for_graph): Likewise.
12332 * tree-pretty-print.c (dump_function_declaration): Likewise.
12333 (dump_array_domain): Likewise.
12334 (dump_omp_clause): Likewise.
12335 (dump_location): Likewise.
12336 (dump_generic_node): Likewise.
12337 (print_struct_decl): Likewise.
12338 * diagnostic.c (diagnostic_show_locus): Use pp_space.
12340 2013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
12342 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
12343 candidate table when replacing a candidate statement.
12344 (replace_rhs_if_not_dup): Likewise.
12345 (replace_one_candidate): Likewise.
12347 2013-08-02 Jan Hubicka <jh@suse.cz>
12348 Martin Liska <marxin.liska@gmail.com>
12350 * cgraphunit.c (add_new_function): Fix logic when adding from
12352 (assemble_thunk): Rename to ...
12353 (expand_thunk); .. this one; export; get it working with
12354 general functions; make produced gimple valid.
12355 * cgraph.h (expand_thunk): Declare.
12357 2013-08-02 Jan Hubicka <jh@suse.cz>
12359 * ipa-cp.c (gather_context_independent_values): Use
12360 ipa_get_param_move_cost.
12361 (get_replacement_map): Remove PARAM; move parameter folding
12363 (create_specialized_node): Update.
12364 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
12365 assert that we have gimple body; update move_cost.
12366 (count_formal_params): Assert that we have gimple body.
12367 (ipa_dump_param): New function.
12368 (ipa_alloc_node_params): Break out from ...
12369 (ipa_initialize_node_params): ... here.
12370 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
12371 (ipa_write_node_info): Stream move costs.
12372 (ipa_read_node_info): Read move costs.
12373 (ipa_update_after_lto_read): Do not recompute node params.
12374 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
12375 (ipa_get_param): Check we are not in WPA.
12376 (ipa_get_param_move_cost): New.
12377 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
12378 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
12379 parm numbers to be present.
12381 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
12383 PR rtl-optimization/58048
12384 * lra-constraints.c (process_alt_operands): Don't check asm
12385 operand on register.
12387 2013-08-02 Eric Botcazou <ebotcazou@adacore.com>
12389 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
12390 the implied StoreLoad barrier for atomic operations if before.
12392 2013-08-02 Jan Hubicka <jh@suse.cz>
12393 Martin Liska <marxin.liska@gmail.com>
12395 * cgraph.c (cgraph_function_body_availability): Do not check
12397 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
12398 symtab_node_availability): Declare.
12399 * ipa.c (can_replace_by_local_alias): New.
12400 (function_and_variable_visibility): Use it.
12401 * symtab.c (symtab_for_node_and_aliases,
12402 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
12404 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
12406 PR rtl-optimization/57963
12407 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
12408 (lra_constraints): Use them.
12410 2013-08-02 Sofiane Naci <sofiane.naci@arm.com>
12412 * config/arm/types.md (define_attr "type"): Add "load_acq"
12414 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
12416 (cortex_a53_store1): Likewise.
12418 2013-08-01 Jan Hubicka <jh@suse.cz>
12420 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
12421 partitions are not needed.
12423 2013-08-01 Uros Bizjak <ubizjak@gmail.com>
12425 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
12426 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
12427 MAYBE_NON_Q_CLASS_P where appropriate.
12429 2013-08-01 Jan Hubicka <jh@suse.cz>
12431 * cgraph.h (release_function_body): Declare.
12432 * tree.c (free_lang_data_in_decl): Free, parameters and return values
12433 of unused delcarations.
12435 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12437 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
12438 RTL form when subtracting a constant.
12440 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12442 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
12443 Generate canonical plus rtx with negated immediate instead of minus
12445 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
12447 2013-08-01 Jan Hubicka <jh@suse.cz>
12449 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
12450 (cgraph_release_function_body): Likewise.
12451 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
12452 * cgraph.h (cgrpah_node): Rename abstract_and_needed
12453 to used_as_abstract_origin.
12454 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
12455 symbols used as abstract origins.
12456 * cgraphunit.c (analyze_functions): Update.
12457 * ipa.c (symtab_remove_unreachable_nodes): Recompute
12458 used_as_abstract_origin.
12459 * tree-inline.c (tree_function_versioning): Update
12460 used_as_abstract_origin; be ready for DECL_RESULT and
12461 DECL_ARGUMENTS to be NULL.
12463 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
12464 for abstract functions.
12465 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
12468 2013-08-01 Jan Hubicka <jh@suse.cz>
12470 * profile.c (compute_value_histograms): Fix thinko.
12472 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
12474 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
12475 aarch-common-protos.h to extra_headers.
12476 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
12477 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
12478 * config/aarch64/t-aarch64 (aarch-common.o): Define.
12480 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
12482 * config/aarch64/aarch64.md (define_attr "type"): Delete.
12483 Include "../arm/types.md". Define "type" attribute for all patterns.
12484 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
12487 2013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
12489 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
12490 to support power8 load fusion.
12491 (fusion_gpr_mem_load): Likewise.
12493 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
12495 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
12496 declarations for power8 load fusion.
12497 (emit_fusion_gpr_load): Likewise.
12499 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
12500 tuning for power8, turn on fusion mode by default. Turn on sign
12501 extending fusion mode if normal fusion mode is on, and we are at
12503 (fusion_gpr_load_p): New function, return true if we can fuse an
12504 addis instruction with a dependent load to a GPR.
12505 (emit_fusion_gpr_load): Emit the instructions for power8 load
12508 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
12509 (VSX load fusion peepholes): New peepholes to fuse together an
12510 addi instruction with a VSX load instruction.
12512 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
12513 peepholes to fuse an addis instruction with a load to a GPR base
12514 register. If we are supporting sign extending fusions, convert
12515 sign extending loads to zero extending loads and add an explicit
12518 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
12520 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
12521 aarch-common-protos.h to extra_headers.
12522 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
12523 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
12524 (arm_early_store_addr_dep): Likewise.
12525 (arm_no_early_alu_shift_dep): Likewise.
12526 (arm_no_early_alu_shift_value_dep): Likewise.
12527 (arm_no_early_mul_dep): Likewise.
12528 (arm_no_early_store_addr_dep): Likewise.
12529 (arm_mac_accumulator_is_mul_result): Likewise.
12530 (arm_mac_accumulator_is_result): Likewise.
12531 * config/arm/aarch-common.c: ... here. New file.
12532 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
12534 (arm_early_store_addr_dep): Likewise.
12535 (arm_no_early_alu_shift_dep): Likewise.
12536 (arm_no_early_alu_shift_value_dep): Likewise.
12537 (arm_no_early_mul_dep): Likewise.
12538 (arm_no_early_store_addr_dep): Likewise.
12539 (arm_mac_accumulator_is_mul_result): Likewise.
12540 (arm_mac_accumulator_is_result): Likewise.
12541 * config/arm/aarch-common-protos.h: ... here. New file.
12542 * config/arm/t-arm (aarch-common.o): Define.
12544 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
12546 * config/arm/arm.md: Include new file "types.md".
12547 (define_attr "type"): Move from here to ...
12548 (define_attr "mul32"): Likewise.
12549 (define_attr "mul64"): Likewise.
12550 * config/arm/types.md: ... here. New file.
12552 2013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
12554 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
12555 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
12557 2013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
12559 * gen-pass-instances.awk: Fix offset of substr().
12561 2013-07-31 David Malcolm <dmalcolm@redhat.com>
12563 * Makefile.in (pass-instances.def): New.
12564 (passes.o): Replace dependency on passes.def with one on
12567 * gen-pass-instances.awk: New.
12569 * passes.c (pass_manager::pass_manager): Use pass-instances.def
12570 rather than passes.def, updating local definition of NEXT_PASS
12571 macro to add an extra NUM parameter (currently unused).
12573 2013-07-30 David Malcolm <dmalcolm@redhat.com>
12575 * Makefile.in (PASS_MANAGER_H): New.
12576 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
12577 (passes.o): Likewise.
12578 (statistics.o): Likewise.
12579 (cgraphunit.o): Likewise.
12580 (context.o): Depend on PASS_MANAGER_H.
12582 * pass_manager.h: New.
12584 * cgraphunit.c (cgraph_add_new_function): Update for moves
12585 of globals to fields of pass_manager.
12586 (analyze_function): Likewise.
12587 (expand_function): Likewise.
12588 (ipa_passes): Likewise.
12589 (compile): Likewise.
12591 * context.c (context::context): New.
12592 * context.h (context::context): New.
12593 (context::get_passes): New.
12594 (context::passes_): New.
12596 * lto-cgraph.c (input_node): Update for moves of globals to
12597 fields of pass_manager.
12599 * passes.c (all_passes): Remove, in favor of a field of the
12600 same name within the new class pass_manager.
12601 (all_small_ipa_passes): Likewise.
12602 (all_lowering_passes): Likewise.
12603 (all_regular_ipa_passes): Likewise.
12604 (all_late_ipa_passes): Likewise.
12605 (all_lto_gen_passes): Likewise.
12606 (passes_by_id): Likewise.
12607 (passes_by_id_size): Likewise.
12608 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
12609 the new class pass_manager.
12610 (set_pass_for_id): Convert to...
12611 (pass_manager::set_pass_for_id): ...method.
12612 (get_pass_for_id): Convert to...
12613 (pass_manager::get_pass_for_id): ...method.
12614 (register_one_dump_file): Move body of implementation into...
12615 (pass_manager::register_one_dump_file): ...here.
12616 (register_dump_files_1): Convert to...
12617 (pass_manager::register_dump_files_1): ...method.
12618 (register_dump_files): Convert to...
12619 (pass_manager::register_dump_files): ...method.
12620 (create_pass_tab): Update for moves of globals to fields of
12622 (dump_passes): Move body of implementation into...
12623 (pass_manager::dump_passes): ...here.
12624 (register_pass): Move body of implementation into...
12625 (pass_manager::register_pass): ...here.
12626 (init_optimization_passes): Convert into...
12627 (pass_manager::pass_manager): ...constructor for new
12628 pass_manager class, and initialize the pass_lists array.
12629 (check_profile_consistency): Update for moves of globals to
12630 fields of pass_manager.
12631 (dump_profile_report): Move body of implementation into...
12632 (pass_manager::dump_profile_report): ...here.
12633 (ipa_write_summaries_1): Update for moves of pass lists from
12634 being globals to fields of pass_manager.
12635 (ipa_write_optimization_summaries): Likewise.
12636 (ipa_read_summaries): Likewise.
12637 (ipa_read_optimization_summaries): Likewise.
12638 (execute_all_ipa_stmt_fixups): Likewise.
12640 * statistics.c (statistics_fini): Update for moves of globals to
12641 fields of pass_manager.
12643 * toplev.c (general_init): Replace call to
12644 init_optimization_passes with construction of the pass_manager
12647 * tree-pass.h (all_passes): Remove, in favor of a field of the
12648 same name within the new class pass_manager.
12649 (all_small_ipa_passes): Likewise.
12650 (all_lowering_passes): Likewise.
12651 (all_regular_ipa_passes): Likewise.
12652 (all_lto_gen_passes): Likewise.
12653 (all_late_ipa_passes): Likewise.
12654 (passes_by_id): Likewise.
12655 (passes_by_id_size): Likewise.
12656 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
12657 the new class pass_manager.
12658 (get_pass_for_id): Remove.
12660 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
12662 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
12665 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
12667 * arm.md (mulhi3): New expand pattern.
12669 2013-07-30 Jan Hubicka <jh@suse.cz>
12670 Martin Liska <marxin.liska@gmail.com>
12672 * profile.c (compute_value_histograms): Do not ICE when
12673 there is mismatch only on some counters.
12675 2013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
12677 PR rtl-optimization/57637
12678 * function.c (move_insn_for_shrink_wrap): Also check the
12679 GEN set of the LIVE problem for the liveness analysis
12680 if it exists, otherwise give up.
12682 2013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
12684 PR tree-optimization/57993
12685 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
12686 replaced statement in the candidate table.
12687 (phi_add_costs): Return infinite cost when the hidden basis does
12688 not dominate all phis on which the candidate is dependent.
12689 (replace_one_candidate): Record replaced statement in the
12692 2013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
12694 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
12695 (ashlv2si3): New expander.
12696 (*ashlv2si3_i): New define_insn_and_split.
12697 * predicates.md (float_operation): Allow patterns with three
12698 basic sub-patterns.
12700 PR rtl-optimization/58021
12701 * mode-switching.c (create_pre_exit): Always split off preceding
12702 insns if we are not at the basic block head.
12704 2013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
12706 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
12707 (UCLIBC_DYNAMIC_LINKER): New macro.
12708 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
12710 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
12711 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
12713 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
12714 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
12715 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
12716 for SF and DF modes. Use ieee_quad_format for TF mode.
12717 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
12718 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
12719 (mips_option_override): Handle `-mnan=legacy'.
12720 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
12721 `-mabs=2008' and `-mnan=2008'.
12722 (OPTION_DEFAULT_SPECS): Add "nan" default.
12723 (ASM_SPEC): Handle `-mnan='.
12724 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
12725 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
12726 comment accordingly.
12727 (neg<mode>2): Likewise.
12728 * config/mips/mips.opt (mabs, mnan): New options.
12729 * doc/install.texi (Configuration): Document `--with-nan=' option.
12730 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
12732 (MIPS Options): Document them.
12733 * config.gcc <mips*-*-*>: Handle `--with-nan='.
12734 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
12735 * configure: Regenerate.
12736 * config.in: Regenerate.
12738 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
12740 * config/i386/i386.md (float post-reload splitters): Do not check
12741 for subregs of SSE registers.
12743 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
12744 H.J. Lu <hongjiu.lu@intel.com>
12748 * config/i386/i386.md (post-reload splitter
12749 to avoid partial SSE reg dependency stalls): New pattern.
12751 2013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
12753 * config/s390/s390.md ("movcc"): Swap load and store instructions.
12755 2013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
12757 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
12758 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
12760 2013-07-26 Cary Coutant <ccoutant@google.com>
12762 * dwarf2out.c (die_checksum_ordered): Don't include template
12763 instantiations in signature.
12764 (is_template_parameter): New function.
12765 (is_template_instantiation): New function.
12766 (generate_skeleton_bottom_up): Don't include template instantiations
12768 (generate_skeleton): Likewise.
12769 (break_out_comdat_types): Move recursive call to break out nested
12771 (prune_unused_types_mark_generic_parms_dies): Call
12772 is_template_parameter.
12774 2013-07-26 Ian Bolton <ian.bolton@arm.com>
12776 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
12777 uses vector registers.
12778 * config/aarch64/iterators.md: Add attributes rtn and vas.
12780 2013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12781 Richard Earnshaw <richard.earnshaw@arm.com>
12783 * combine.c (simplify_comparison): Re-canonicalize operands
12785 * config/arm/arm.md (movcond_addsi): New splitter.
12787 2013-07-25 Sterling Augustine <saugustine@google.com>
12789 * dwarf2out.c (size_of_pubnames): Move code to...
12790 (include_pubname_in_output): ...here. New.
12791 (want_pubnames): Rearrange.
12792 (output_pubnames): Call include_pubname_in_output. Move assertion.
12794 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
12796 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
12798 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
12801 * doc/extend.texi: Remove obsolete builtins. Fix
12802 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
12804 2013-07-25 Jan Hubicka <jh@suse.cz>
12806 * cgraph.c (release_function_body): Break out from ...
12807 (cgraph_release_function_body): ... this one; also release DECL_RESULT
12808 and DECL_ARGUMENTS.
12809 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
12810 old_tree in the map.
12811 (create_specialized_node): Update.
12812 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
12814 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
12815 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
12816 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
12817 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
12820 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12822 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
12823 addsi3_carryin_alt2_<optab>): Correct output template.
12825 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12827 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
12828 Adjust for arm_restrict_it.
12829 Remove trailing whitespace.
12831 2013-07-25  Mark Kettenis  <kettenis@openbsd.org>
12833 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
12834 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
12836 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
12838 2013-07-25 Vladimir Makarov <vmakarov@redhat.com>
12840 PR rtl-optimization/57960
12841 * lra-constraints.c (process_alt_operands): Use the right mode
12842 when checking strict_low.
12844 2013-07-25 Jan Hubicka <jh@suse.cz>
12846 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
12847 * cgraph.c (cgraph_remove_node): Do not release function body
12848 when in cgraph streaming.
12849 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
12850 in other partitions are not considered reachable; fix handling of
12853 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
12855 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
12857 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
12862 * config/arm/arm.md ("*sibcall_insn): Replace use of
12863 Ss with US. Adjust output for v5 and v4t.
12864 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
12866 * config/arm/constraints.md ("Ss"): Rename to US.
12868 2013-07-25 Terry Guo <terry.guo@arm.com>
12870 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
12871 shift_add/shift_sub0/shift_sub1 RTXs.
12873 2013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
12874 Anton Blanchard <anton@au1.ibm.com>
12876 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
12877 (altivec_vpks<VI_char>ss): Likewise.
12878 (altivec_vpks<VI_char>us): Likewise.
12879 (altivec_vpku<VI_char>us): Likewise.
12880 (altivec_vpku<VI_char>um): Likewise.
12882 2013-07-24 David Malcolm <dmalcolm@redhat.com>
12884 Introduce context class.
12886 * Makefile.in (CONTEXT_H): New.
12887 (OBJS): Add context.o.
12888 (toplev.o): Add CONTEXT_H to dependencies.
12891 * toplev.c (general_init): Create the singleton gcc::context instance.
12897 2013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
12899 PR rtl-optimization/57968
12900 * mode-switching.c (create_pre_exit): Allow instructions that
12901 don't set a return register to need a non-exit mode.
12903 2013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
12904 Anton Blanchard <anton@au1.ibm.com>
12906 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
12907 operands to vperm for little endian.
12908 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
12909 of lvsl to create the control mask for a vperm for little endian.
12911 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12912 Anton Blanchard <anton@au1.ibm.com>
12914 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
12915 two operands for little-endian.
12917 2013-07-23 Steve Ellcey <sellcey@mips.com>
12919 * config/mips/mips.c (mips_case_values_threshold): New.
12920 (TARGET_CASE_VALUES_THRESHOLD): Define.
12922 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12923 Anton Blanchard <anton@au1.ibm.com>
12925 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
12926 selection of field for vector splat in little endian mode.
12928 2013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
12930 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
12931 expanders to rs6000.md.
12932 (ior<mode>3): Likewise.
12933 (and<mode>3): Likewise.
12934 (one_cmpl<mode>2): Likewise.
12935 (nor<mode>3): Likewise.
12936 (andc<mode>3): Likewise.
12937 (eqv<mode>3): Likewise.
12938 (nand<mode>3): Likewise.
12939 (orc<mode>3): Likewise.
12941 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
12944 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
12945 to split multi-word logical operations.
12946 (rs6000_split_logical_di): Likewise.
12947 (rs6000_split_logical): Likewise.
12949 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
12950 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
12951 and allow TImode operations in 32-bit.
12952 (vsx_and<mode>3_64bit): Likewise.
12953 (vsx_ior<mode>3_32bit): Likewise.
12954 (vsx_ior<mode>3_64bit): Likewise.
12955 (vsx_xor<mode>3_32bit): Likewise.
12956 (vsx_xor<mode>3_64bit): Likewise.
12957 (vsx_one_cmpl<mode>2_32bit): Likewise.
12958 (vsx_one_cmpl<mode>2_64bit): Likewise.
12959 (vsx_nor<mode>3_32bit): Likewise.
12960 (vsx_nor<mode>3_64bit): Likewise.
12961 (vsx_andc<mode>3_32bit): Likewise.
12962 (vsx_andc<mode>3_64bit): Likewise.
12963 (vsx_eqv<mode>3_32bit): Likewise.
12964 (vsx_eqv<mode>3_64bit): Likewise.
12965 (vsx_nand<mode>3_32bit): Likewise.
12966 (vsx_nand<mode>3_64bit): Likewise.
12967 (vsx_orc<mode>3_32bit): Likewise.
12968 (vsx_orc<mode>3_64bit): Likewise.
12970 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
12971 logical types in GPRs.
12973 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
12974 logical insns to rs6000.md, and allow TImode operations in
12976 (altivec_ior<mode>3): Likewise.
12977 (altivec_xor<mode>3): Likewise.
12978 (altivec_one_cmpl<mode>2): Likewise.
12979 (altivec_nor<mode>3): Likewise.
12980 (altivec_andc<mode>3): Likewise.
12982 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
12983 attributes for moving the 128-bit logical operations into
12985 (BOOL_REGS_OUTPUT): Likewise.
12986 (BOOL_REGS_OP1): Likewise.
12987 (BOOL_REGS_OP2): Likewise.
12988 (BOOL_REGS_UNARY): Likewise.
12989 (BOOL_REGS_AND_CR0): Likewise.
12990 (one_cmpl<mode>2): Add support for DI logical operations on
12991 32-bit, splitting the operations to 32-bit.
12992 (anddi3): Likewise.
12993 (iordi3): Likewise.
12994 (xordi3): Likewise.
12995 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
12996 changes to combine the 32/64-bit code, allow logical operations on
12997 TI mode in 32-bit, and to use similar match_operator patterns like
12998 scalar mode uses. Combine the Altivec and VSX code for logical
12999 operations, and move it here.
13000 (ior<mode>3, 128-bit types): Likewise.
13001 (xor<mode>3, 128-bit types): Likewise.
13002 (one_cmpl<mode>3, 128-bit types): Likewise.
13003 (nor<mode>3, 128-bit types): Likewise.
13004 (andc<mode>3, 128-bit types): Likewise.
13005 (eqv<mode>3, 128-bit types): Likewise.
13006 (nand<mode>3, 128-bit types): Likewise.
13007 (orc<mode>3, 128-bit types): Likewise.
13008 (and<mode>3_internal): Likewise.
13009 (bool<mode>3_internal): Likewise.
13010 (boolc<mode>3_internal1): Likewise.
13011 (boolc<mode>3_internal2): Likewise.
13012 (boolcc<mode>3_internal1): Likewise.
13013 (boolcc<mode>3_internal2): Likewise.
13014 (eqv<mode>3_internal1): Likewise.
13015 (eqv<mode>3_internal2): Likewise.
13016 (one_cmpl1<mode>3_internal): Likewise.
13018 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
13020 * config/microblaze/microblaze.c (microblaze_expand_prologue):
13021 Rename flag_stack_usage to flag_stack_usage_info.
13023 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
13025 * config/microblaze/sync.md: New file.
13026 * config/microblaze/microblaze.md: Include sync.md
13027 * config/microblaze/microblaze.c: Add print_operand 'y'.
13028 * config/microblaze/constraints.md: Add memory_contraint
13029 'Q' which is a single register.
13031 2013-07-23 Eric Botcazou <ebotcazou@adacore.com>
13033 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
13035 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
13037 * reload.c (find_reloads): Exit loop once we find this operand
13038 cannot be reloaded somehow for this alternative.
13040 * reload.c (find_reloads): Exit loop once we find a hard register.
13042 * rtlanal.c (computed_jump_p): Exit loop once we find label
13045 * i386.c (ix86_pad_returns): Exit loop after setting replace.
13047 * cfgloopmanip.c (remove_path): Exit loop after setting
13050 * gensupport.c (subst_dup): Avoid loop if code is not
13051 MATCH_DUP nor MATCH_OP_DUP.
13053 2013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
13055 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
13057 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
13059 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
13060 true for SP_REGNUM if mode == ptr_mode.
13061 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
13062 with value R0_REGNUM + 31.
13064 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
13066 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
13067 pad pointer-typed argument downward.
13069 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
13071 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
13072 and __ILP32__ when the ILP32 model is in use.
13074 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
13076 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
13077 (aarch64_load_symref_appropriately): In the case of
13078 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
13079 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
13080 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
13081 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
13082 if 'mode' doesn't equal to 'ptr_mode'.
13083 (aarch64_output_mi_thunk): Add an assertion on the alignment of
13084 'vcall_offset'; change to call aarch64_emit_move differently depending
13085 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
13086 to calculate the upper bound of 'vcall_offset'.
13087 (aarch64_cannot_force_const_mem): Change to also return true if
13089 (aarch64_legitimize_reload_address): In the case of large
13090 displacements, add new local variable 'xmode' and an assertion
13091 based on it; change to use 'xmode' to generate the new rtx and
13093 (aarch64_asm_trampoline_template): Change to generate the template
13094 differently depending on TARGET_ILP32 or not; change to use
13095 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
13096 (aarch64_trampoline_size): Removed.
13097 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
13098 and replace immediate literals with it. Change to use 'ptr_mode'
13099 instead of 'DImode' and call convert_memory_address if the mode
13100 of 'fnaddr' doesn't equal to 'ptr_mode'.
13101 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
13103 (aarch64_elf_asm_destructor): Likewise.
13104 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
13105 on TARGET_ILP32 instead of aarch64_trampoline_size.
13106 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
13107 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
13108 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
13109 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
13110 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
13111 (storewb_pair<GPI:mode>_<P:mode>): ... this.
13112 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
13113 depending on the value of 'mode'.
13114 (add_losym_<mode>): New.
13115 (ldr_got_small_<mode>): New, based on ldr_got_small.
13116 (ldr_got_small): Remove.
13117 (ldr_got_small_sidi): New.
13118 * config/aarch64/iterators.md (P): New.
13119 (PTR): Change to 'ptr_mode' in the condition.
13121 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
13123 * config.gcc (aarch64*-*-*): Support --with-abi.
13124 (aarch64*-*-elf): Support --with-multilib-list.
13125 (aarch64*-*-linux*): Likewise.
13126 (supported_defaults): Add abi to aarch64*-*-*.
13127 * configure.ac: Mention AArch64 for --with-multilib-list.
13128 * configure: Re-generated.
13129 * config/aarch64/biarchilp32.h: New file.
13130 * config/aarch64/biarchlp64.h: New file.
13131 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
13133 (MULTILIB_DEFAULTS): Ditto.
13134 (DRIVER_SELF_SPECS): Ditto.
13135 (ASM_SPEC): Update to also substitute -mabi.
13136 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
13137 file whose name depends on -mabi= and -mbig-endian.
13138 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
13140 (POINTER_SIZE): New define.
13141 (POINTERS_EXTEND_UNSIGNED): Ditto.
13142 (enum aarch64_abi_type): New enumeration tag.
13143 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
13144 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
13145 (TARGET_ILP32): New define.
13146 * config/aarch64/aarch64.opt (mabi): New.
13147 (aarch64_abi): New.
13148 (ilp32, lp64): New values for -mabi.
13149 * config/aarch64/t-aarch64 (comma): New define.
13150 (MULTILIB_OPTIONS): Ditto.
13151 (MULTILIB_DIRNAMES): Ditto.
13152 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
13153 * doc/invoke.texi: Document -mabi for AArch64.
13155 2013-07-23 Georg-Johann Lay <avr@gjlay.de>
13157 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
13159 2013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
13160 Anton Blanchard <anton@au1.ibm.com>
13162 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
13163 endianness when selecting field to splat.
13165 2013-07-22 Eric Christopher <echristo@gmail.com>
13167 * dwarf2out.c (die_odr_checksum): New function to use
13168 CHECKSUM_ macros and ULEB128 for DIE tag.
13169 (generate_type_signature): Use.
13171 2013-07-22 Eric Botcazou <ebotcazou@adacore.com>
13173 * config.gcc (sparc*-*-*): Accept leon3 processor.
13174 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
13175 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
13176 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
13177 * config/sparc/sparc.opt (enum processor_type): Add leon3.
13178 (mfix-ut699): Adjust comment.
13179 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
13180 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
13181 (CPP_CPU_SPEC): Likewise.
13182 (ASM_CPU_SPEC): Likewise.
13183 * config/sparc/sparc.c (leon3_cost): New constant.
13184 (sparc_option_override): Add leon3 support.
13185 (mem_ref): New function.
13186 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
13187 (sparc_do_work_around_errata): Look into the instruction in the delay
13188 slot and adjust accordingly. Add fix for the data cache nullify issues
13189 of the UT699. Change insertion position for the NOP.
13190 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
13191 (leon3_load): New reservation.
13192 (leon_store): Bump latency to 2.
13193 (grfpu): New automaton.
13194 (grfpu_alu): New unit.
13195 (grfpu_ds): Likewise.
13196 (leon_fp_alu): Adjust.
13197 (leon_fp_mult): Delete.
13198 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
13199 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
13200 * config/sparc/sparc.md (cpu): Add leon3.
13201 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
13202 (swapsi): Likewise.
13203 (atomic_test_and_set): Likewise.
13204 (ldstub): Likewise.
13206 2013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
13208 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
13209 default for R5900 targets.
13210 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
13211 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
13212 * config/mips/mips.c (mips_option_override): Report an error for
13213 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
13214 for -march=r5900 -mhard-float.
13216 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
13218 * df-problems.c (can_move_insns_across): Exit loop once we
13219 find a non-fixed, non-global register.
13221 * ipa-pure-const.c (propagate_nothrow): Exit loop after
13224 * omega.c (omega_eliminate_red): Break after setting red_found.
13225 (omega_problem_has_red_equations): Similarly after setting found.
13226 (omega_query_variable): Similarly after setting coupled.
13228 2013-07-22 Marek Polacek <polacek@redhat.com>
13230 * gimplify.c: Don't include gimple.h twice.
13232 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13234 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
13235 instead of TARGET_THUMB1.
13236 (Pz): New constraint.
13237 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
13239 (compare_negsi_si): Likewise.
13240 (compare_addsi2_op0): Likewise.
13241 (compare_addsi2_op1): Likewise.
13242 (addsi3_carryin_<optab>): Likewise.
13243 (addsi3_carryin_alt2_<optab>): Likewise.
13244 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
13245 for arm_restrict_it.
13246 (subsi3_carryin): Likewise.
13247 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
13248 (minmax_arithsi): Disable for arm_restrict_it.
13249 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
13250 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
13251 (satsi_<SAT:code>_shift): Likewise.
13252 (arm_shiftsi3): Add alternative for 16-bit encoding.
13253 (arm32_movhf): Disable for arm_restrict_it.
13254 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
13255 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
13257 2013-07-22 Sofiane Naci <sofiane.naci@arm.com>
13259 * config/arm/arm.md (attribute "insn"): Delete.
13260 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
13261 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
13262 (not_shiftsi): Update for attribute change.
13263 (not_shiftsi_compare0): Likewise.
13264 (not_shiftsi_compare0_scratch): Likewise.
13265 (arm_one_cmplsi2): Likewise.
13266 (thumb1_one_cmplsi2): Likewise.
13267 (notsi_compare0): Likewise.
13268 (notsi_compare0_scratch): Likewise.
13269 (thumb1_movdi_insn): Likewise.
13270 (arm_movsi_insn): Likewise.
13271 (movhi_insn_arch4): Likewise.
13272 (movhi_bytes): Likewise.
13273 (arm_movqi_insn): Likewise.
13274 (thumb1_movqi_insn): Likewise.
13275 (arm32_movhf): Likewise.
13276 (thumb1_movhf): Likewise.
13277 (arm_movsf_soft_insn): Likewise.
13278 (thumb1_movsf_insn): Likewise.
13279 (thumb_movdf_insn): Likewise.
13280 (movsicc_insn): Likewise.
13281 (movsfcc_soft_insn): Likewise.
13282 (and_scc): Likewise.
13283 (cond_move): Likewise.
13284 (if_move_not): Likewise.
13285 (if_not_move): Likewise.
13286 (if_shift_move): Likewise.
13287 (if_move_shift): Likewise.
13288 (if_shift_shift): Likewise.
13289 (if_not_arith): Likewise.
13290 (if_arith_not): Likewise.
13291 (cond_move_not): Likewise.
13292 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
13293 (neon_mov<mode>): Likewise.
13294 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
13295 (thumb2_movsi_vfp): Likewise.
13296 (movsf_vfp): Likewise.
13297 (thumb2_movsf_vfp): Likewise.
13298 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
13300 (cortexa7_older_only): Likewise.
13301 (cortexa7_younger): Likewise.
13302 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
13303 (1020alu_shift_op): Likewise.
13304 (1020alu_shift_reg_op): Likewise.
13305 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
13306 (alu_shift_op): Likewise.
13307 (alu_shift_reg_op): Likewise.
13308 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
13309 (11_alu_shift_op): Likewise.
13310 (11_alu_shift_reg_op): Likewise.
13311 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
13312 (9_alu_shift_reg_op): Likewise.
13313 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
13315 (cortex_a15_alu_shift): Likewise.
13316 (cortex_a15_alu_shift_reg): Likewise.
13317 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
13318 (cortex_a5_alu_shift): Likewise.
13319 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
13321 (cortex_a53_alu_shift): Likewise.
13322 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
13324 (cortex_a7_alu_reg): Likewise.
13325 (cortex_a7_alu_shift): Likewise.
13326 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
13327 (cortex_a8_alu_shift): Likewise.
13328 (cortex_a8_alu_shift_reg): Likewise.
13329 (cortex_a8_mov): Likewise.
13330 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
13331 (cortex_a9_dp_shift): Likewise.
13332 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
13333 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
13334 (cortex_r4_mov): Likewise.
13335 (cortex_r4_alu_shift): Likewise.
13336 (cortex_r4_alu_shift_reg): Likewise.
13337 * config/arm/fa526.md (526_alu_op): Update for attribute change.
13338 (526_alu_shift_op): Likewise.
13339 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
13340 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
13341 (626te_alu_shift_op): Likewise.
13342 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
13343 (726te_alu_op): Likewise.
13344 (726te_alu_shift_op): Likewise.
13345 (726te_alu_shift_reg_op): Likewise.
13346 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
13347 (mp626_alu_shift_op): Likewise.
13348 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
13349 (pj4_alu_e1_conds): Likewise.
13350 (pj4_alu): Likewise.
13351 (pj4_alu_conds): Likewise.
13352 (pj4_shift): Likewise.
13353 (pj4_shift_conds): Likewise.
13354 (pj4_alu_shift): Likewise.
13355 (pj4_alu_shift_conds): Likewise.
13357 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13359 * config/arm/predicates.md (shiftable_operator_strict_it):
13361 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
13362 Disable cond_exec version for arm_restrict_it.
13363 (thumb2_smaxsi3): Convert to generate cond_exec.
13364 (thumb2_sminsi3): Likewise.
13365 (thumb32_umaxsi3): Likewise.
13366 (thumb2_uminsi3): Likewise.
13367 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
13368 (thumb2_neg_abssi2): Likewise.
13369 (thumb2_mov_scc): Add alternative for 16-bit encoding.
13370 (thumb2_movsicc_insn): Adjust alternatives.
13371 (thumb2_mov_negscc): Disable for arm_restrict_it.
13372 (thumb2_mov_negscc_strict_it): New pattern.
13373 (thumb2_mov_notscc_strict_it): New pattern.
13374 (thumb2_mov_notscc): Disable for arm_restrict_it.
13375 (thumb2_ior_scc): Likewise.
13376 (thumb2_ior_scc_strict_it): New pattern.
13377 (thumb2_cond_move): Adjust for arm_restrict_it.
13378 (thumb2_cond_arith): Disable for arm_restrict_it.
13379 (thumb2_cond_arith_strict_it): New pattern.
13380 (thumb2_cond_sub): Adjust for arm_restrict_it.
13381 (thumb2_movcond): Likewise.
13382 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
13383 (thumb2_zero_extendhisi2_v6): Likewise.
13384 (thumb2_zero_extendqisi2_v6): Likewise.
13385 (orsi_notsi_si): Likewise.
13386 (orsi_not_shiftsi_si): Likewise.
13388 2013-07-22 Georg-Johann Lay <avr@gjlay.de>
13390 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
13391 instruction sequence is 1 byte shorter.
13393 2013-07-22 Uros Bizjak <ubizjak@gmail.com>
13395 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
13396 it is not needed after split.
13398 2013-07-20 Iain Sandoe <iain@codesourcery.com>
13401 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
13402 second label for nonlocal goto receivers. Don't output pic base labels
13403 unless we're producing PIC; mark that action unreachable().
13404 (ix86_save_reg): If the function contains a nonlocal label, save the
13406 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
13407 * config/darwin.c (emitted_pic_label_num): New GTY.
13408 (update_pic_label_number_if_needed): New.
13409 (machopic_output_function_base_name): Adjust for nonlocal receiver
13411 (machopic_should_output_picbase_label): New.
13412 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
13413 (nonlocal_goto_receiver): New insn and split.
13415 2013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
13417 * config/aarch64/aarch64-builtins.c
13418 (aarch64_fold_builtin): Fold abs in all modes.
13419 * config/aarch64/aarch64-simd-builtins.def
13420 (abs): Enable for all modes.
13421 * config/aarch64/arm_neon.h
13422 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
13423 (vabs_f64): Add missing intrinsic.
13425 2013-07-19 Ian Bolton <ian.bolton@arm.com>
13427 * config/aarch64/arm_neon.h (vabs_s64): New function
13429 2013-07-19 Georg-Johann Lay <avr@gjlay.de>
13432 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
13433 * config/avr/avr.md (adjust_len): Add `round'.
13434 * config/avr/avr-protos.h (avr_out_round): New prototype.
13435 (avr_out_plus): Add `out_label' argument.
13436 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
13437 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
13438 Handle the case where `insn' is just a pattern.
13439 (avr_out_bitop): Handle the case where `insn' is just a pattern.
13440 (avr_out_round): New function.
13441 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
13443 2013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
13445 * config/microblaze/microblaze.c (microblaze_expand_prologue):
13446 Add check for flag_stack_usage to handle -fstack-usage support
13448 2013-07-18 Pat Haugen <pthaugen@us.ibm.com>
13450 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
13451 interaction for new Power8 flags and VSX.
13453 2013-07-18 Sriraman Tallam <tmsriram@google.com>
13455 PR middle-end/57698
13456 * tree-inline.c (expand_call_inline): Emit errors during
13457 early_inlining only if optimization is not turned on.
13459 2013-07-18 David Malcolm <dmalcolm@redhat.com>
13463 * passes.c (init_optimization_passes): Move the construction of
13464 the pass hierarchy into a new passes.def file.
13466 * Makefile.in (passes.o): Add dependency on passes.def.
13468 2013-07-18 David Malcolm <dmalcolm@redhat.com>
13470 * passes.c (init_optimization_passes): Introduce macros for
13471 constructing the tree of passes (INSERT_PASSES_AFTER,
13472 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
13473 TERMINATE_PASS_LIST).
13475 2013-07-18 Vladimir Makarov <vmakarov@redhat.com>
13476 Wei Mi <wmi@google.com>
13478 PR rtl-optimization/57878
13479 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
13481 (reload_pseudo_compare_func): Check nregs first for reload
13484 2013-07-18 David Malcolm <dmalcolm@redhat.com>
13486 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
13488 2013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
13490 * read-rtl.c (validate_const_int): Once an invalid character is
13491 seen, quit the loop.
13493 * gengtype.c (write_roots): Similarly once we find the "deletable"
13494 or "if_marked" option.
13496 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
13498 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
13499 "xtab" and "sat". Move value "clz" from here to ...
13500 (attriubte "type"): ... here.
13501 (satsi_<SAT:code>): Delete "insn" attribute.
13502 (satsi_<SAT:code>_shift): Likewise.
13503 (arm_zero_extendqisi2addsi): Likewise.
13504 (arm_extendqisi2addsi): Likewise.
13505 (clzsi2): Update for attribute changes.
13506 (rbitsi2): Likewise.
13507 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
13509 (arm_usatsihi): Likewise.
13510 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
13512 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
13514 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
13515 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
13516 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
13517 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
13518 in alphabetical order.
13519 (attribute "core_cycles"): Update for attribute changes.
13520 (arm_addsi3): Likewise.
13521 (addsi3_compare0): Likewise.
13522 (addsi3_compare0_scratch): Likewise.
13523 (addsi3_compare_op1): Likewise.
13524 (addsi3_compare_op2): Likewise.
13525 (compare_addsi2_op0): Likewise.
13526 (compare_addsi2_op1): Likewise.
13527 (addsi3_carryin_shift_<optab>): Likewise.
13528 (subsi3_carryin_shift): Likewise.
13529 (rsbsi3_carryin_shift): Likewise.
13530 (arm_subsi3_insn): Likewise.
13531 (subsi3_compare0): Likewise.
13532 (subsi3_compare): Likewise.
13533 (arm_andsi3_insn): Likewise.
13534 (thumb1_andsi3_insn): Likewise.
13535 (andsi3_compare0): Likewise.
13536 (andsi3_compare0_scratch): Likewise.
13537 (zeroextractsi_compare0_scratch
13538 (andsi_not_shiftsi_si): Likewise.
13539 (iorsi3_insn): Likewise.
13540 (iorsi3_compare0): Likewise.
13541 (iorsi3_compare0_scratch): Likewise.
13542 (arm_xorsi3): Likewise.
13543 (thumb1_xorsi3_insn): Likewise.
13544 (xorsi3_compare0): Likewise.
13545 (xorsi3_compare0_scratch): Likewise.
13546 (satsi_<SAT:code>_shift): Likewise.
13548 (arm_shiftsi3): Likewise.
13549 (shiftsi3_compare0): Likewise.
13550 (not_shiftsi): Likewise.
13551 (not_shiftsi_compare0): Likewise.
13552 (not_shiftsi_compare0_scratch): Likewise.
13553 (arm_one_cmplsi2): Likewise.
13554 (thumb_one_complsi2): Likewise.
13555 (notsi_compare0): Likewise.
13556 (notsi_compare0_scratch): Likewise.
13557 (thumb1_zero_extendhisi2): Likewise.
13558 (arm_zero_extendhisi2): Likewise.
13559 (arm_zero_extendhisi2_v6): Likewise.
13560 (arm_zero_extendhisi2addsi): Likewise.
13561 (thumb1_zero_extendqisi2): Likewise.
13562 (thumb1_zero_extendqisi2_v6): Likewise.
13563 (arm_zero_extendqisi2): Likewise.
13564 (arm_zero_extendqisi2_v6): Likewise.
13565 (arm_zero_extendqisi2addsi): Likewise.
13566 (thumb1_extendhisi2): Likewise.
13567 (arm_extendhisi2): Likewise.
13568 (arm_extendhisi2_v6): Likewise.
13569 (arm_extendqisi): Likewise.
13570 (arm_extendqisi_v6): Likewise.
13571 (arm_extendqisi2addsi): Likewise.
13572 (thumb1_extendqisi2): Likewise.
13573 (thumb1_movdi_insn): Likewise.
13574 (arm_movsi_insn): Likewise.
13575 (movsi_compare0): Likewise.
13576 (movhi_insn_arch4): Likewise.
13577 (movhi_bytes): Likewise.
13578 (arm_movqi_insn): Likewise.
13579 (thumb1_movqi_insn): Likewise.
13580 (arm32_movhf): Likewise.
13581 (thumb1_movhf): Likewise.
13582 (arm_movsf_soft_insn): Likewise.
13583 (thumb1_movsf_insn): Likewise.
13584 (movdf_soft_insn): Likewise.
13585 (thumb_movdf_insn): Likewise.
13586 (arm_cmpsi_insn): Likewise.
13587 (cmpsi_shiftsi): Likewise.
13588 (cmpsi_shiftsi_swp): Likewise.
13589 (arm_cmpsi_negshiftsi_si): Likewise.
13590 (movsicc_insn): Likewise.
13591 (movsfcc_soft_insn): Likewise.
13592 (arith_shiftsi): Likewise.
13593 (arith_shiftsi_compare0
13594 (arith_shiftsi_compare0_scratch
13595 (sub_shiftsi): Likewise.
13596 (sub_shiftsi_compare0
13597 (sub_shiftsi_compare0_scratch
13598 (and_scc): Likewise.
13599 (cond_move): Likewise.
13600 (if_plus_move): Likewise.
13601 (if_move_plus): Likewise.
13602 (if_move_not): Likewise.
13603 (if_not_move): Likewise.
13604 (if_shift_move): Likewise.
13605 (if_move_shift): Likewise.
13606 (if_shift_shift): Likewise.
13607 (if_not_arith): Likewise.
13608 (if_arith_not): Likewise.
13609 (cond_move_not): Likewise.
13610 (thumb1_ashlsi3): Set type attribute.
13611 (thumb1_ashrsi3): Likewise.
13612 (thumb1_lshrsi3): Likewise.
13613 (thumb1_rotrsi3): Likewise.
13614 (shiftsi3_compare0_scratch): Likewise.
13615 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
13616 (neon_mov<mode>): Likewise.
13617 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
13619 (thumb2_movsi_insn): Likewise.
13620 (thumb2_cmpsi_neg_shiftsi): Likewise.
13621 (thumb2_extendqisi_v6): Likewise.
13622 (thumb2_zero_extendhisi2_v6): Likewise.
13623 (thumb2_zero_extendqisi2_v6): Likewise.
13624 (thumb2_shiftsi3_short): Likewise.
13625 (thumb2_addsi3_compare0_scratch): Likewise.
13626 (orsi_not_shiftsi_si): Likewise.
13627 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
13628 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
13630 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
13631 (1020alu_shift_op): Likewise.
13632 (1020alu_shift_reg_op): Likewise.
13633 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
13634 (alu_shift_op): Likewise.
13635 (alu_shift_reg_op): Likewise.
13636 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
13637 (11_alu_shift_op): Likewise.
13638 (11_alu_shift_reg_op): Likewise.
13639 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
13640 (9_alu_shift_reg_op): Likewise.
13641 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
13643 (cortex_a15_alu_shift): Likewise.
13644 (cortex_a15_alu_shift_reg): Likewise.
13645 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
13647 (cortex_a5_alu_shift): Likewise.
13648 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
13650 (cortex_a53_alu_shift): Likewise.
13651 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
13653 (cortex_a7_alu_reg): Likewise.
13654 (cortex_a7_alu_shift): Likewise.
13655 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
13657 (cortex_a8_alu_shift): Likewise.
13658 (cortex_a8_alu_shift_reg): Likewise.
13659 (cortex_a8_mov): Likewise.
13660 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
13661 (cortex_a9_dp_shift): Likewise.
13662 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
13664 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
13666 (cortex_r4_mov): Likewise.
13667 (cortex_r4_alu_shift): Likewise.
13668 (cortex_r4_alu_shift_reg): Likewise.
13669 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
13670 (526_alu_shift_op): Likewise.
13671 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
13672 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
13673 (626te_alu_shift_op): Likewise.
13674 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
13675 (726te_alu_op): Likewise.
13676 (726te_alu_shift_op): Likewise.
13677 (726te_alu_shift_reg_op): Likewise.
13678 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
13679 (mp626_alu_shift_op): Likewise.
13680 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
13681 (pj4_alu_e1_conds): Likewise.
13682 (pj4_alu): Likewise.
13683 (pj4_alu_conds): Likewise.
13684 (pj4_shift): Likewise.
13685 (pj4_shift_conds): Likewise.
13686 (pj4_alu_shift): Likewise.
13687 (pj4_alu_shift_conds): Likewise.
13688 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
13690 (cortexa7_older_only): Likewise.
13691 (cortexa7_younger): Likewise.
13693 2013-07-18 David Malcolm <dmalcolm@redhat.com>
13695 * ipa-pure-const.c (generate_summary): Rename to...
13696 (pure_const_generate_summary): ... this.
13698 2013-07-17 Iain Sandoe <iain@codesourcery.com>
13700 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
13702 2013-07-17 Yvan Roux <yvan.roux@linaro.org>
13705 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
13708 2013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13710 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
13711 enabled without -march=zEC12.
13712 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
13715 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
13717 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
13718 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
13719 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
13720 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
13721 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
13722 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
13723 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
13725 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
13726 and ISA_HAS_NMADD3_NMSUB3.
13727 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
13728 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
13729 (nmsub4<mode>, nmsub3<mode>): Likewise.
13730 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
13732 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
13734 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
13735 TARGET_MIPS5400 checking.
13737 2013-07-16 Jakub Jelinek <jakub@redhat.com>
13738 Peter Bergner <bergner@vnet.ibm.com>
13740 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
13741 registers in the comment.
13742 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
13743 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
13744 rather than FIRST_PSEUDO_REGISTERS.
13746 2013-07-16 Peter Bergner <bergner@vnet.ibm.com>
13748 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
13749 enable extra ISA flags with TARGET_HTM.
13751 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
13753 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
13756 2013-07-15 Cong Hou <congh@google.com>
13758 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
13759 in compare function for sorting.
13761 2013-07-15 Peter Bergner <bergner@vnet.ibm.com>
13763 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
13764 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
13765 * config/rs6000/rs6000.opt: Add -mhtm option.
13766 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
13767 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
13768 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
13769 __HTM__ if the HTM instructions are available.
13770 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
13771 htm_spr_reg_operand): New define_predicates.
13772 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
13773 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
13775 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
13776 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
13777 HTM builtin functions.
13778 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
13779 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
13780 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
13781 (rs6000_builtin_mask_calculate): Likewise.
13782 (rs6000_option_override_internal): Likewise.
13783 (bdesc_htm): Add new HTM builtin support.
13784 (htm_spr_num): New function.
13785 (htm_spr_regno): Likewise.
13786 (rs6000_htm_spr_icode): Likewise.
13787 (htm_expand_builtin): Likewise.
13788 (htm_init_builtins): Likewise.
13789 (rs6000_expand_builtin): Add support for HTM builtin functions.
13790 (rs6000_init_builtins): Likewise.
13791 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
13793 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
13794 (TARGET_HTM, MASK_HTM): Define macros.
13795 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
13796 (FIXED_REGISTERS): Likewise.
13797 (CALL_USED_REGISTERS): Likewise.
13798 (CALL_REALLY_USED_REGISTERS): Likewise.
13799 (REG_ALLOC_ORDER): Likewise.
13800 (enum reg_class): Likewise.
13801 (REG_CLASS_NAMES): Likewise.
13802 (REG_CLASS_CONTENTS): Likewise.
13803 (REGISTER_NAMES): Likewise.
13804 (ADDITIONAL_REGISTER_NAMES): Likewise.
13805 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
13806 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
13807 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
13808 * config/rs6000/htm.md: New file.
13809 * config/rs6000/htmintrin.h: New file.
13810 * config/rs6000/htmxlintrin.h: New file.
13812 2013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
13814 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
13815 Define SYMBOL_TINY_GOT, update comment.
13816 * config/aarch64/aarch64.c
13817 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
13818 (aarch64_expand_mov_immediate): Likewise.
13819 (aarch64_print_operand): Likewise.
13820 (aarch64_classify_symbol): Likewise.
13821 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
13822 (ldr_got_tiny): Define.
13824 2013-07-13 Tobias Grosser <tobias@grosser.es>
13826 PR tree-optimization/54094
13827 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
13828 scheduling dimension for the parallelism check from the polyhedral
13829 information in the AST.
13830 * graphite-dependences.c (carries_deps): Do not assume the schedule is
13833 2013-07-13 Jason Merrill <jason@redhat.com>
13835 * print-tree.c (debug_vec_tree): Use debug_raw.
13836 (debug_raw (vec<tree, va_gc> &)): New.
13837 (debug_raw (vec<tree, va_gc> *)): New.
13838 * tree.h: Declare them.
13840 2013-07-13 Bin Cheng <bin.cheng@arm.com>
13842 * ifcvt.c (ifcvt_after_combine): New static variable.
13843 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
13845 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
13846 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
13847 rest_of_handle_if_after_reload): Pass new argument for if_convert.
13849 2013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
13851 * config/mips/mips.c (mips_expand_call): Remove empty statement.
13853 2013-07-12 Michael Matz <matz@suse.de>
13855 PR middle-end/55771
13856 * convert.c (convert_to_real): Reject non-float inner types.
13858 2013-07-12 Tejas Belagod <tejas.belagod@arm.com>
13860 * config/aarch64/aarch64-protos.h
13861 (aarch64_simd_immediate_valid_for_move): Remove.
13862 * config/aarch64/aarch64.c (simd_immediate_info): New member.
13863 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
13865 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
13867 2013-07-11 Steve Ellcey <sellcey@mips.com>
13869 * config/mips/mips.c (mips_conditional_register_usage): Do not
13870 use t[0-7] registers in MIPS16 mode when optimizing for size.
13872 2013-07-11 Sriraman Tallam <tmsriram@google.com>
13874 * config/i386/i386.c (dispatch_function_versions): Fix array
13875 indexing of function_version_info to match actual_versions.
13877 2013-07-11 Teresa Johnson <tejohnson@google.com>
13879 * vec.h (struct va_gc): Move release out-of-line.
13880 (va_gc::release): Call ggc_free on released vec.
13882 2013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
13884 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
13885 Require GOT register as additional operand in UNSPEC.
13886 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
13887 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
13888 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
13889 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
13890 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
13891 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
13892 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
13894 2013-07-11 Georg-Johann Lay <avr@gjlay.de>
13897 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
13898 name seen by assembler/linker rather if available.
13900 2013-07-11 Andreas Schwab <schwab@suse.de>
13902 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
13904 2013-07-10 Vladimir Makarov <vmakarov@redhat.com>
13906 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
13908 2013-07-10 Joseph Myers <joseph@codesourcery.com>
13910 * doc/tm.texi.in: Move hook documentation to ....
13911 * target.def: ... here.
13913 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
13914 text on @hook line.
13915 * doc/tm.texi: Regenerate.
13917 2013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
13920 * doc/invoke.texi: Document -Wconditionally-supported.
13922 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
13925 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
13928 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
13931 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
13932 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
13933 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
13934 Remove duplicate devices.
13935 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
13936 * config/avr/t-multilib: Regenerate.
13937 * config/avr/avr-tables.opt: Regenerate.
13938 * doc/avr-mmcu.texi: Regenerate.
13940 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
13943 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
13945 2013-07-10 Graham Stott <graham.stott@btinternet.com>
13947 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
13948 the cost of MULT when optimizing for size.
13950 2013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
13952 * config/cr16/cr16-protos.h: Don't include target.h.
13954 2013-07-09 Joseph Myers <joseph@codesourcery.com>
13956 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
13957 adjust register size for TDmode and TFmode for VSX registers.
13959 2013-07-08 Kai Tietz <ktietz@redhat.com>
13962 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
13963 hook_bool_const_tree_true.
13965 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13967 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
13968 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
13969 * config/s390/s390.md: Define FPR*_REGNUM constants.
13970 Fix FPR2_REGNUM constant (18 -> 17).
13971 ("*trunc<BFP:mode><DFP_ALL:mode>2")
13972 ("*trunc<DFP_ALL:mode><BFP:mode>2")
13973 ("trunc<BFP:mode><DFP_ALL:mode>2")
13974 ("trunc<DFP_ALL:mode><BFP:mode>2")
13975 ("*extend<BFP:mode><DFP_ALL:mode>2")
13976 ("*extend<DFP_ALL:mode><BFP:mode>2")
13977 ("extend<BFP:mode><DFP_ALL:mode>2")
13978 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
13981 2013-07-08 Graham Stott <graham.stott@btinternet.com>
13983 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
13985 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13987 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
13988 and cfun_fpr_bit_p to cfun_fpr_save_p.
13989 (s390_frame_area, s390_register_info, s390_frame_info)
13990 (s390_emit_prologue, s390_emit_epilogue)
13991 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
13993 * config/s390/s390.h: Define *_REGNUM macros for floating point
13996 2013-07-08 Eric Botcazou <ebotcazou@adacore.com>
13998 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
14000 2013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
14002 PR rtl-optimization/57786
14003 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
14004 and break out of the loop when it is set to false.
14006 2013-07-08 Jakub Jelinek <jakub@redhat.com>
14009 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
14010 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
14011 (const_int 63)) 0)).
14012 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
14013 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
14014 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
14016 PR rtl-optimization/57829
14017 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
14018 mask bits outside of mode are just sign-extension from mode to HWI.
14020 2013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
14022 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
14023 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
14024 adjust_address instead of change_address to keep info about alignment.
14025 (emit_strmov): Remove.
14026 (emit_memmov): New function.
14027 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
14028 (expand_movmem_epilogue): Likewise and return updated rtx for
14030 (expand_constant_movmem_prologue): Likewise and return updated rtx for
14031 destination and source.
14032 (decide_alignment): Refactor, handle vector_loop.
14033 (ix86_expand_movmem): Likewise.
14034 (ix86_expand_setmem): Likewise.
14035 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
14037 2013-07-07 Uros Bizjak <ubizjak@gmail.com>
14039 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
14040 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
14042 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
14044 * config/i386/sse.md (sse_movlhps): Change alternative 3
14045 of operand 2 to "m".
14047 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
14050 * config/i386/sse.md (iptr): New mode attribute.
14051 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
14052 (<sse>_vm<plusminus_insn><mode>3): Ditto.
14053 (<sse>_vmmul<mode>3): Ditto.
14054 (<sse>_vmdiv<mode>3): Ditto.
14055 (sse_vmrcpv4sf2): Ditto.
14056 (<sse>_vmsqrt<mode>2): Ditto.
14057 (sse_vmrsqrtv4sf2): Ditto.
14058 (<sse>_vm<code><mode>3): Ditto.
14059 (avx_vmcmp<mode>3): Ditto.
14060 (<sse>_vmmaskcmp<mode>3): Ditto.
14061 (<sse>_comi): Ditto.
14062 (<sse>_ucomi): Ditto.
14063 (*xop_vmfrcz_<mode>): Ditto.
14064 (*fmai_fmadd_<mode>): Ditto.
14065 (*fmai_fmsub_<mode>): Ditto.
14066 (*fmai_fnmadd_<mode>): Ditto.
14067 (*fmai_fnmsub_<mode>): Ditto.
14068 (*fma4i_vmfmadd_<mode>): Ditto.
14069 (*fma4i_vmfmsub_<mode>): Ditto.
14070 (*fma4i_vmfnmadd_<mode>): Ditto.
14071 (*fma4i_vmfnmsub_<mode>): Ditto.
14072 (*xop_vmfrcz_<mode>): Ditto.
14073 (sse_cvtps2pi): Ditto.
14074 (sse_cvttps2pi): Ditto.
14075 (sse_cvtss2si): Ditto.
14076 (sse_cvtss2si_2): Ditto.
14077 (sse_cvtss2siq_2): Ditto.
14078 (sse_cvttss2si): Ditto.
14079 (sse_cvttss2siq): Ditto.
14080 (sse_cvtsd2si): Ditto.
14081 (sse_cvtsd2si_2): Ditto.
14082 (sse_cvtsd2siq_2): Ditto.
14083 (sse_cvttsd2si): Ditto.
14084 (sse_cvttsd2siq): Ditto.
14085 (sse_cvtsd2ss): Ditto.
14086 (sse_cvtss2sd): Ditto.
14087 (avx2_pbroadcast<mode>): Ditto.
14088 (avx2_pbroadcast<mode>_1): Ditto.
14089 (*avx_vperm_broadcast_v4sf): Ditto.
14091 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
14092 (sse_movlhps): Ditto.
14093 (sse_storehps): Ditto.
14094 (sse_loadhps): Ditto.
14095 (sse_storelps): Ditto.
14096 (sse_loadlps): Ditto.
14097 (*vec_concatv4sf): Ditto.
14098 (*vec_interleave_highv2df): Ditto.
14099 (*vec_interleave_lowv2df): Ditto.
14100 (*vec_extractv2df_1_sse): Ditto.
14101 (*vec_extractv2df_0_sse): Ditto.
14102 (sse2_storelpd): Ditto.
14103 (sse2_loadlpd): Ditto.
14104 (sse2_movsd): Ditto.
14105 (*vec_concatv4si): Ditto.
14106 (vec_concatv2di): Ditto.
14108 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
14109 for Intel asm dialect.
14110 (mmx_punpcklwd): Ditto.
14111 (mmx_punpckldq): Ditto.
14113 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
14114 for intel assembler dialect.
14116 2013-07-06 Jakub Jelinek <jakub@redhat.com>
14119 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
14120 for BUILT_IN_C{LZ,LRSB}*.
14121 * tree.h (CASE_INT_FN): Add FN##IMAX case.
14122 * tree-vrp.c (extract_range_basic): Handle
14123 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
14124 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
14125 fall thru to code calling set_value*.
14126 * builtins.c (expand_builtin): Remove *IMAX cases.
14127 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
14128 if width is bigger than 2*HWI.
14130 2013-07-05 Vladimir Makarov <vmakarov@redhat.com>
14132 PR rtl-optimization/55342
14133 * lra-int.h (lra_subreg_reload_pseudos): New.
14134 * lra.c: Add undoing optional reloads to the block diagram.
14135 (lra_subreg_reload_pseudos): New.
14136 (lra_optional_reload_pseudos): Change comments.
14137 (lra): Init and clear lra_subreg_reload_pseudos. Clear
14138 lra_optional_reload_pseudos after undo transformations.
14139 * lra-assigns.c (pseudo_prefix_title): New.
14140 (lra_setup_reg_renumber): Use it.
14141 (spill_for): Ditto. Check subreg reload pseudos too.
14142 (assign_by_spills): Consider subreg reload pseudos too.
14143 * lra-constraints.c (simplify_operand_subreg): Use
14144 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
14145 (curr_insn_transform): Recognize and do optional reloads.
14146 (undo_optional_reloads): New.
14147 (lra_undo_inheritance): Call undo_optional_reloads.
14149 2013-07-05 Thomas Quinot <quinot@adacore.com>
14151 * tree-complex.c (expand_complex_operations_1): Fix typo.
14153 2013-07-04 Tejas Belagod <tejas.belagod@arm.com>
14155 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
14156 (tune_params): New member 'const vec_costs'.
14157 * config/aarch64/aarch64.c (generic_vector_cost): New.
14158 (generic_tunings): New member 'generic_vector_cost'.
14159 (aarch64_builtin_vectorization_cost): New.
14160 (aarch64_add_stmt_cost): New.
14161 (TARGET_VECTORIZE_ADD_STMT_COST): New.
14162 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
14164 2013-07-03 Jakub Jelinek <jakub@redhat.com>
14167 * config/i386/predicates.md (vsib_address_operand): Disallow
14168 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
14170 2013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
14172 PR middle-end/55030
14173 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
14174 expand_builtin_setjmp_receiver.
14175 (expand_label): Adjust, call expand_builtin_setjmp_receiver
14176 with NULL for the label parameter.
14177 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
14178 the frame-pointer. Adjust comments.
14179 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
14180 only if LABEL is non-NULL.
14182 2013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
14184 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
14185 (ARM_ABI_AAPCS64): Ditto.
14187 (ARM_DEFAULT_ABI): Ditto.
14189 2013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
14191 * config/aarch64/aarch64-builtins.c
14192 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
14193 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
14195 * config/aarch64/aarch64-simd.md
14196 (aarch64_ld1<VALL:mode>): New.
14197 (aarch64_st1<VALL:mode>): Likewise.
14198 * config/aarch64/arm_neon.h
14199 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
14201 2013-07-02 Sriraman Tallam <tmsriram@google.com>
14203 * config/i386/i386.c (gate_insert_vzeroupper): Check if
14205 (ix86_option_override_internal):Turn on all -mavx target flags by
14206 default as they are dependent on AVX anyway.
14208 2013-07-02 Cary Coutant <ccoutant@google.com>
14210 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
14211 deterministic hash.
14212 (loc_checksum_ordered): Likewise.
14213 (hash_loc_operands): Remove inline keyword.
14215 2013-07-02 Jakub Jelinek <jakub@redhat.com>
14217 PR tree-optimization/57741
14218 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
14219 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
14220 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
14221 Allow REAL_CST step_exprs if flag_associative_math.
14222 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
14224 2013-07-02 Ian Bolton <ian.bolton@arm.com>
14226 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
14228 2013-07-02 Ian Bolton <ian.bolton@arm.com>
14230 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
14232 2013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14234 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
14236 (iorsi3_insn): Likewise.
14237 (arm_xorsi3): Likewise.
14239 2013-07-01 Sofiane Naci <sofiane.naci@arm.com>
14241 * arm.md (attribute "wtype"): Delete. Move attribute values from here
14243 (attribute "type"): ... here, and prefix with "wmmx_".
14244 (attribute "core_cycles"): Update for attribute changes.
14245 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
14246 (tbcstv4hi): Likewise.
14247 (tbcstv2si): Likewise.
14248 (iwmmxt_iordi3): Likewise.
14249 (iwmmxt_xordi3): Likewise.
14250 (iwmmxt_anddi3): Likewise.
14251 (iwmmxt_nanddi3): Likewise.
14252 (iwmmxt_arm_movdi): Likewise.
14253 (iwmmxt_movsi_insn): Likewise.
14254 (mov<mode>_internal): Likewise.
14255 (and<mode>3_iwmmxt): Likewise.
14256 (ior<mode>3_iwmmxt): Likewise.
14257 (xor<mode>3_iwmmxt): Likewise.
14258 (add<mode>3_iwmmxt): Likewise.
14259 (ssaddv8qi3): Likewise.
14260 (ssaddv4hi3): Likewise.
14261 (ssaddv2si3): Likewise.
14262 (usaddv8qi3): Likewise.
14263 (usaddv4hi3): Likewise.
14264 (usaddv2si3): Likewise.
14265 (sub<mode>3_iwmmxt): Likewise.
14266 (sssubv8qi3): Likewise.
14267 (sssubv4hi3): Likewise.
14268 (sssubv2si3): Likewise.
14269 (ussubv8qi3): Likewise.
14270 (ussubv4hi3): Likewise.
14271 (ussubv2si3): Likewise.
14272 (mulv4hi3_iwmmxt): Likewise.
14273 (smulv4hi3_highpart): Likewise.
14274 (umulv4hi3_highpart): Likewise.
14275 (iwmmxt_wmacs): Likewise.
14276 (iwmmxt_wmacsz): Likewise.
14277 (iwmmxt_wmacu): Likewise.
14278 (iwmmxt_wmacuz): Likewise.
14279 (iwmmxt_clrdi): Likewise.
14280 (iwmmxt_clrv8qi): Likewise.
14281 (iwmmxt_clr4hi): Likewise.
14282 (iwmmxt_clr2si): Likewise.
14283 (iwmmxt_uavgrndv8qi3): Likewise.
14284 (iwmmxt_uavgrndv4hi3): Likewise.
14285 (iwmmxt_uavgv8qi3): Likewise.
14286 (iwmmxt_uavgv4hi3): Likewise.
14287 (iwmmxt_tinsrb): Likewise.
14288 (iwmmxt_tinsrh): Likewise.
14289 (iwmmxt_tinsrw): Likewise.
14290 (iwmmxt_textrmub): Likewise.
14291 (iwmmxt_textrmsb): Likewise.
14292 (iwmmxt_textrmuh): Likewise.
14293 (iwmmxt_textrmsh): Likewise.
14294 (iwmmxt_textrmw): Likewise.
14295 (iwmxxt_wshufh): Likewise.
14296 (eqv8qi3): Likewise.
14297 (eqv4hi3): Likewise.
14298 (eqv2si3): Likewise.
14299 (gtuv8qi3): Likewise.
14300 (gtuv4hi3): Likewise.
14301 (gtuv2si3): Likewise.
14302 (gtv8qi3): Likewise.
14303 (gtv4hi3): Likewise.
14304 (gtv2si3): Likewise.
14305 (smax<mode>3_iwmmxt): Likewise.
14306 (umax<mode>3_iwmmxt): Likewise.
14307 (smin<mode>3_iwmmxt): Likewise.
14308 (umin<mode>3_iwmmxt): Likewise.
14309 (iwmmxt_wpackhss): Likewise.
14310 (iwmmxt_wpackwss): Likewise.
14311 (iwmmxt_wpackdss): Likewise.
14312 (iwmmxt_wpackhus): Likewise.
14313 (iwmmxt_wpackwus): Likewise.
14314 (iwmmxt_wpackdus): Likewise.
14315 (iwmmxt_wunpckihb): Likewise.
14316 (iwmmxt_wunpckihh): Likewise.
14317 (iwmmxt_wunpckihw): Likewise.
14318 (iwmmxt_wunpckilb): Likewise.
14319 (iwmmxt_wunpckilh): Likewise.
14320 (iwmmxt_wunpckilw): Likewise.
14321 (iwmmxt_wunpckehub): Likewise.
14322 (iwmmxt_wunpckehuh): Likewise.
14323 (iwmmxt_wunpckehuw): Likewise.
14324 (iwmmxt_wunpckehsb): Likewise.
14325 (iwmmxt_wunpckehsh): Likewise.
14326 (iwmmxt_wunpckehsw): Likewise.
14327 (iwmmxt_wunpckelub): Likewise.
14328 (iwmmxt_wunpckeluh): Likewise.
14329 (iwmmxt_wunpckeluw): Likewise.
14330 (iwmmxt_wunpckelsb): Likewise.
14331 (iwmmxt_wunpckelsh): Likewise.
14332 (iwmmxt_wunpckelsw): Likewise.
14333 (ror<mode>3): Likewise.
14334 (ashr<mode>3_iwmmxt): Likewise.
14335 (lshr<mode>3_iwmmxt): Likewise.
14336 (ashl<mode>3_iwmmxt): Likewise.
14337 (ror<mode>3_di): Likewise.
14338 (ashr<mode>3_di): Likewise.
14339 (lshr<mode>3_di): Likewise.
14340 (ashl<mode>3_di): Likewise.
14341 (iwmmxt_wmadds): Likewise.
14342 (iwmmxt_wmaddu): Likewise.
14343 (iwmmxt_tmia): Likewise.
14344 (iwmmxt_tmiaph): Likewise.
14345 (iwmmxt_tmiabb): Likewise.
14346 (iwmmxt_tmiatb): Likewise.
14347 (iwmmxt_tmiabt): Likewise.
14348 (iwmmxt_tmiatt): Likewise.
14349 (iwmmxt_tmovmskb): Likewise.
14350 (iwmmxt_tmovmskh): Likewise.
14351 (iwmmxt_tmovmskw): Likewise.
14352 (iwmmxt_waccb): Likewise.
14353 (iwmmxt_wacch): Likewise.
14354 (iwmmxt_waccw): Likewise.
14355 (iwmmxt_waligni): Likewise.
14356 (iwmmxt_walignr): Likewise.
14357 (iwmmxt_walignr0): Likewise.
14358 (iwmmxt_walignr1): Likewise.
14359 (iwmmxt_walignr2): Likewise.
14360 (iwmmxt_walignr3): Likewise.
14361 (iwmmxt_wsadb): Likewise.
14362 (iwmmxt_wsadh): Likewise.
14363 (iwmmxt_wsadbz): Likewise.
14364 (iwmmxt_wsadhz): Likewise.
14365 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
14366 (iwmmxt_wabsdiffb): Likewise.
14367 (iwmmxt_wabsdiffh): Likewise.
14368 (iwmmxt_wabsdiffw): Likewise.
14369 (iwmmxt_waddsubhx): Likewise
14370 (iwmmxt_wsubaddhx): Likewise.
14371 (addc<mode>3): Likewise.
14372 (iwmmxt_avg4): Likewise.
14373 (iwmmxt_avg4r): Likewise.
14374 (iwmmxt_wmaddsx): Likewise.
14375 (iwmmxt_wmaddux): Likewise.
14376 (iwmmxt_wmaddsn): Likewise.
14377 (iwmmxt_wmaddun): Likewise.
14378 (iwmmxt_wmulwsm): Likewise.
14379 (iwmmxt_wmulwum): Likewise.
14380 (iwmmxt_wmulsmr): Likewise.
14381 (iwmmxt_wmulumr): Likewise.
14382 (iwmmxt_wmulwsmr): Likewise.
14383 (iwmmxt_wmulwumr): Likewise.
14384 (iwmmxt_wmulwl): Likewise.
14385 (iwmmxt_wqmulm): Likewise.
14386 (iwmmxt_wqmulwm): Likewise.
14387 (iwmmxt_wqmulmr): Likewise.
14388 (iwmmxt_wqmulwmr): Likewise.
14389 (iwmmxt_waddbhusm): Likewise.
14390 (iwmmxt_waddbhusl): Likewise.
14391 (iwmmxt_wqmiabb): Likewise.
14392 (iwmmxt_wqmiabt): Likewise.
14393 (iwmmxt_wqmiatb): Likewise.
14394 (iwmmxt_wqmiatt): Likewise.
14395 (iwmmxt_wqmiabbn): Likewise.
14396 (iwmmxt_wqmiabtn): Likewise.
14397 (iwmmxt_wqmiatbn): Likewise.
14398 (iwmmxt_wqmiattn): Likewise.
14399 (iwmmxt_wmiabb): Likewise.
14400 (iwmmxt_wmiabt): Likewise.
14401 (iwmmxt_wmiatb): Likewise.
14402 (iwmmxt_wmiatt): Likewise.
14403 (iwmmxt_wmiabbn): Likewise.
14404 (iwmmxt_wmiabtn): Likewise.
14405 (iwmmxt_wmiatbn): Likewise.
14406 (iwmmxt_wmiattn): Likewise.
14407 (iwmmxt_wmiawbb): Likewise.
14408 (iwmmxt_wmiawbt): Likewise.
14409 (iwmmxt_wmiawtb): Likewise.
14410 (iwmmxt_wmiawtt): Likewise.
14411 (iwmmxt_wmiawbbn): Likewise.
14412 (iwmmxt_wmiawbtn): Likewise.
14413 (iwmmxt_wmiawtbn): Likewise.
14414 (iwmmxt_wmiawttn): Likewise.
14415 (iwmmxt_wmerge): Likewise.
14416 (iwmmxt_tandc<mode>3): Likewise.
14417 (iwmmxt_torc<mode>3): Likewise.
14418 (iwmmxt_torvsc<mode>3): Likewise.
14419 (iwmmxt_textrc<mode>3): Likewise.
14420 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
14421 (wmmxt_pack): Likewise.
14422 (wmmxt_mult_c1): Likewise.
14423 (wmmxt_mult_c2): Likewise.
14424 (wmmxt_alu_c1): Likewise.
14425 (wmmxt_alu_c2): Likewise.
14426 (wmmxt_alu_c3): Likewise.
14427 (wmmxt_transfer_c1): Likewise.
14428 (wmmxt_transfer_c2): Likewise.
14429 (wmmxt_transfer_c3): Likewise.
14430 (marvell_f_iwmmxt_wstr): Likewise.
14431 (marvell_f_iwmmxt_wldr): Likewise.
14433 2013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
14435 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
14437 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
14440 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
14441 * lra-constraints.c (need_for_split_p): Check call used hard regs
14442 living through calls.
14444 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
14445 call used regs for call insn.
14447 2013-06-28 Jakub Jelinek <jakub@redhat.com>
14450 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
14451 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
14453 2013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
14455 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
14457 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
14459 * lra-constraints.c (need_for_split_p): Check call used hard regs
14460 living through calls.
14462 2013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
14465 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
14466 to tie with any other modes. Eliminate Altivec vector mode tests,
14467 since these are a subset of ALTIVEC or VSX vector modes. Simplify
14468 code, to return 0 if testing MODE2 for a condition, if we've
14469 already tested MODE1 for the same condition.
14471 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
14473 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
14476 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
14478 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
14479 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
14481 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
14483 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
14485 (aarch64_symbolic_constant_p): Remove.
14486 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
14487 static. Fix line length and white space.
14488 (aarch64_symbolic_constant_p): Remove.
14489 * config/aarch64/predicates.md (aarch64_valid_symref):
14490 Use aarch64_classify_symbol_expression.
14492 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14494 * config/arm/constraints.md (Ts): New constraint.
14495 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
14497 (compare_scc): Use "Ts" constraint for operand 0.
14498 (ior_scc_scc): Likewise.
14499 (and_scc_scc): Likewise.
14500 (and_scc_scc_nodom): Likewise.
14501 (ior_scc_scc_cmp): Likewise for operand 7.
14502 (and_scc_scc_cmp): Likewise.
14503 * config/arm/thumb2.md (thumb2_movsi_insn):
14504 Add alternatives for 16-bit encodings.
14505 (thumb2_movhi_insn): Likewise.
14506 (thumb2_movsicc_insn): Likewise.
14507 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
14508 (thumb2_negscc): Use "Ts" constraint.
14509 Move mvn instruction outside cond_exec block.
14510 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
14511 for 16-bit encodings.
14513 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14515 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
14517 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
14518 (mulsi3subsi): Likewise.
14519 (mulsidi3adddi): Likewise.
14520 (mulsidi3_v6): Likewise.
14521 (umulsidi3_v6): Likewise.
14522 (umulsidi3adddi_v6): Likewise.
14523 (smulsi3_highpart_v6): Likewise.
14524 (umulsi3_highpart_v6): Likewise.
14525 (mulhisi3tb): Likewise.
14526 (mulhisi3bt): Likewise.
14527 (mulhisi3tt): Likewise.
14528 (maddhisi4): Likewise.
14529 (maddhisi4tb): Likewise.
14530 (maddhisi4tt): Likewise.
14531 (maddhidi4): Likewise.
14532 (maddhidi4tb): Likewise.
14533 (maddhidi4tt): Likewise.
14534 (zeroextractsi_compare0_scratch): Likewise.
14535 (insv_zero): Likewise.
14536 (insv_t2): Likewise.
14537 (anddi_notzesidi_di): Likewise.
14538 (anddi_notsesidi_di): Likewise.
14539 (andsi_notsi_si): Likewise.
14540 (iordi_zesidi_di): Likewise.
14541 (xordi_zesidi_di): Likewise.
14542 (andsi_iorsi3_notsi): Likewise.
14543 (smax_0): Likewise.
14544 (smax_m1): Likewise.
14545 (smin_0): Likewise.
14546 (not_shiftsi): Likewise.
14547 (unaligned_loadsi): Likewise.
14548 (unaligned_loadhis): Likewise.
14549 (unaligned_loadhiu): Likewise.
14550 (unaligned_storesi): Likewise.
14551 (unaligned_storehi): Likewise.
14552 (extv_reg): Likewise.
14553 (extzv_t2): Likewise.
14554 (divsi3): Likewise.
14555 (udivsi3): Likewise.
14556 (arm_zero_extendhisi2addsi): Likewise.
14557 (arm_zero_extendqisi2addsi): Likewise.
14558 (compareqi_eq0): Likewise.
14559 (arm_extendhisi2_v6): Likewise.
14560 (arm_extendqisi2addsi): Likewise.
14561 (arm_movt): Likewise.
14562 (thumb2_ldrd): Likewise.
14563 (thumb2_ldrd_base): Likewise.
14564 (thumb2_ldrd_base_neg): Likewise.
14565 (thumb2_strd): Likewise.
14566 (thumb2_strd_base): Likewise.
14567 (thumb2_strd_base_neg): Likewise.
14568 (arm_negsi2): Add alternative for 16-bit encoding.
14569 (arm_one_cmplsi2): Likewise.
14571 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14573 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
14574 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
14575 (movdfcc): Likewise.
14576 * config/arm/vfp.md (*thumb2_movsf_vfp):
14577 Disable predication for arm_restrict_it.
14578 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
14579 (*thumb2_movdfcc_vfp): Likewise.
14580 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
14581 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
14582 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
14583 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
14584 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
14585 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
14586 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
14587 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
14588 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
14589 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
14590 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
14591 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
14592 Disable predication for arm_restrict_it.
14594 2013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
14596 * config/i386/bmiintrin.h (_bextr_u32): New.
14597 (_bextr_u64): Ditto.
14599 2013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
14601 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
14602 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
14603 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
14604 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
14605 * config/mips/n32-elf.h: ...this new file.
14607 2013-06-27 Marc Glisse <marc.glisse@inria.fr>
14610 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
14611 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
14613 2013-06-27 Catherine Moore <clm@codesourcery.com>
14615 * config/mips/mips-tables.opt: Regenerate.
14616 * config/mips/mips-cpus.def: Add m14ke and m14kec.
14617 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
14618 * doc/invoke.texi: Add -m14kc.
14620 2013-06-27 Jakub Jelinek <jakub@redhat.com>
14623 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
14624 constraints of operand 1 and 2.
14627 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
14628 to match RTL canonicalization. Swap predicates and
14629 constraints of operand 1 and 2.
14631 2013-06-27 Vladimir Makarov <vmakarov@redhat.com>
14633 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
14634 Process OP_INOUT regs for splitting too.
14636 2013-06-27 Jakub Jelinek <jakub@redhat.com>
14638 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
14639 decl before the loop, initialize to NULL.
14640 (vectorizable_load): Initialize ptr_incr to NULL.
14642 2013-06-27 Martin Jambor <mjambor@suse.cz>
14645 * ipa-ref.h (ipa_maybe_record_reference): Declare.
14646 * ipa-ref.c (ipa_maybe_record_reference): New function.
14647 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
14648 * ipa-cp.c (create_specialized_node): Record potential references from
14650 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
14652 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
14654 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
14655 parameter 'mode' of type 'enum machine_mode mode'; change to pass
14656 'mode' to force_reg.
14657 (aarch64_add_offset): Update calls to aarch64_force_temporary.
14658 (aarch64_expand_mov_immediate): Likewise.
14660 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
14662 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
14663 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
14665 2013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14667 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
14668 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
14669 (struct machine_function): Add tbegin_p.
14670 (s390_canonicalize_comparison): Fold CC mode compares to
14671 conditional jump if possible.
14672 (s390_emit_jump): Return the emitted jump.
14673 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
14674 Handle CCRAWmode compares.
14675 (s390_option_override): Default to -mhtm if available.
14676 (s390_reg_clobbered_rtx): Handle floating point regs as well.
14677 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
14678 FPRs instead of df_regs_ever_live_p.
14679 (s390_optimize_nonescaping_tx): New function.
14680 (s390_init_frame_layout): Extend clobbered_regs array to cover
14682 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
14683 (s390_expand_tbegin): New function.
14684 (enum s390_builtin): New enum definition.
14685 (code_for_builtin): New array definition.
14686 (s390_init_builtins): New function.
14687 (s390_expand_builtin): New function.
14688 (TARGET_INIT_BUILTINS): Define.
14689 (TARGET_EXPAND_BUILTIN): Define.
14690 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
14691 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
14692 (s390_alc_comparison): Likewise.
14693 * config/s390/s390-modes.def: Add CCRAWmode.
14694 * config/s390/s390.h (processor_flags): Add PF_TX.
14695 (TARGET_CPU_HTM): Define macro.
14696 (TARGET_HTM): Define macro.
14697 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
14698 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
14699 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
14700 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
14701 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
14702 (TBEGIN_MASK, TBEGINC_MASK): New constants.
14703 ("*cc_to_int"): Move up.
14704 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
14705 constants other than 0.
14706 ("*ccraw_to_int"): New insn and splitter definition.
14707 ("tbegin", "tbegin_nofloat", "tbegin_retry")
14708 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
14709 ("tx_assist"): New expander.
14710 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
14711 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
14712 * config/s390/s390.opt: Add -mhtm option.
14713 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
14714 * config/s390/htmxlintrin.h: New file.
14715 * config/s390/htmintrin.h: New file.
14716 * config/s390/s390intrin.h: New file.
14717 * doc/extend.texi: Document htm builtins.
14718 * config.gcc: Add the new header files to extra_headers.
14720 2013-06-26 Thomas Schwinge <thomas@codesourcery.com>
14722 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
14723 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
14725 2013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
14726 Pat Haugen <pthaugen@us.ibm.com>
14727 Peter Bergner <bergner@vnet.ibm.com>
14729 * config/rs6000/power8.md: New.
14730 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
14731 setting for power8 entry.
14732 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
14733 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
14734 test for Power4/Power5 only.
14735 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
14737 (force_new_group): Adjust comment.
14738 * config/rs6000/rs6000.md: Include power8.md.
14740 2013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
14742 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
14743 * config/arm/arm-protos.h (arm_max_conditional_execute): New
14745 (tune_params): Update comment.
14746 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
14747 (arm_max_conditional_execute): New function.
14748 (thumb2_final_prescan_insn): Use max_insn_skipped and
14749 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
14751 2013-06-25 Jakub Jelinek <jakub@redhat.com>
14753 PR tree-optimization/57705
14754 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
14755 SSA_NAME step, provided that it is not defined inside the loop.
14756 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
14757 (get_initial_def_for_induction): Handle SSA_NAME IV step.
14759 2013-06-25 Martin Jambor <mjambor@suse.cz>
14761 PR middle-end/57670
14762 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
14763 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
14765 (ipa_note_param_call): Initialize member_ptr flag.
14766 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
14767 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
14768 (ipa_write_indirect_edge_info): Stream member_ptr flag.
14769 (ipa_read_indirect_edge_info): Likewise.
14771 2013-06-25 Richard Biener <rguenther@suse.de>
14773 PR middle-end/56977
14774 * passes.c (init_optimization_passes): Move pass_fold_builtins
14775 and pass_dce earlier with -Og.
14777 2013-06-25 Eric Botcazou <ebotcazou@adacore.com>
14779 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
14780 <BIT_FIELD_REF>: Remove trailing TAB.
14781 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
14784 2013-06-24 Martin Jambor <mjambor@suse.cz>
14786 PR tree-optimization/57358
14787 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
14788 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
14789 (ipa_analyze_params_uses): Generate pessimistic info when true.
14791 2013-06-24 Martin Jambor <mjambor@suse.cz>
14793 PR tree-optimization/57539
14794 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
14795 global.inlined_to of the new node to it. All callers changed.
14796 * ipa-inline-transform.c (clone_inlined_nodes): New variable
14797 inlining_into, pass it to cgraph_clone_node.
14798 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
14799 ipa_free_edge_args_substructures.
14800 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
14801 rdesc linked list. Do not assert rdesc edges have inlined caller.
14802 Assert we have found an rdesc in the rdesc list.
14804 2013-06-24 Richard Biener <rguenther@suse.de>
14806 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
14807 (pointer_set_lookup): Declare.
14808 (class pointer_map): New template class implementing a
14809 generic pointer to T map.
14810 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
14811 pointer_map<T>::contains, pointer_map<T>::insert,
14812 pointer_map<T>::traverse): New functions.
14813 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
14814 (pointer_set_lookup): New function.
14815 (pointer_set_contains): Use pointer_set_lookup.
14816 (pointer_set_insert): Likewise.
14817 (insert_aux): Remove.
14818 (struct pointer_map_t): Embed a pointer_set_t.
14819 (pointer_map_create): Adjust.
14820 (pointer_map_destroy): Likewise.
14821 (pointer_map_contains): Likewise.
14822 (pointer_map_insert): Likewise.
14823 (pointer_map_traverse): Likewise.
14824 * tree-streamer.h (struct streamer_tree_cache_d): Use a
14825 pointer_map<unsigned> instead of a pointer_map_t.
14826 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
14827 (streamer_tree_cache_lookup): Likewise.
14828 (streamer_tree_cache_create): Likewise.
14829 (streamer_tree_cache_delete): Likewise.
14830 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
14831 pointer_map<unsigned> instead of a pointer_map_t.
14832 (lto_init_tree_ref_encoder): Adjust.
14833 (lto_destroy_tree_ref_encoder): Likewise.
14834 * lto-section-out.c (lto_output_decl_index): Likewise.
14835 (lto_record_function_out_decl_state): Likewise.
14836 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
14838 2013-06-24 Richard Biener <rguenther@suse.de>
14840 PR tree-optimization/57488
14841 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
14843 2013-06-24 Alan Modra <amodra@gmail.com>
14845 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
14846 (gen_easy_altivec_constant): Likewise.
14847 * config/rs6000/predicates.md (easy_vector_constant_add_self,
14848 easy_vector_constant_msb): Likewise.
14850 2013-06-23 Jakub Jelinek <jakub@redhat.com>
14853 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
14854 add missing return true.
14856 2013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
14859 * config/sh/predicates.md (general_extend_operand): Invoke
14860 general_movsrc_operand for memory operands.
14861 (general_movsrc_operand): Allow reg+reg addressing, do not use
14862 general_operand for memory operands.
14864 2013-06-23 Sriraman Tallam <tmsriram@google.com>
14866 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
14867 when current target options does not apply.
14868 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
14869 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
14870 * config/i386/bmiintrin.h: Pass appropriate target
14871 attributes to header.
14872 * config/i386/mmintrin.h: Ditto.
14873 * config/i386/nmmintrin.h: Ditto.
14874 * config/i386/avx2intrin.h: Ditto.
14875 * config/i386/fxsrintrin.h: Ditto.
14876 * config/i386/tbmintrin.h: Ditto.
14877 * config/i386/xsaveintrin.h: Ditto.
14878 * config/i386/f16cintrin.h: Ditto.
14879 * config/i386/xtestintrin.h: Ditto.
14880 * config/i386/xsaveoptintrin.h: Ditto.
14881 * config/i386/bmi2intrin.h: Ditto.
14882 * config/i386/lzcntintrin.h: Ditto.
14883 * config/i386/smmintrin.h: Ditto.
14884 * config/i386/wmmintrin.h: Ditto.
14885 * config/i386/x86intrin.h: Remove all header include guards.
14886 * config/i386/prfchwintrin.h: Ditto.
14887 * config/i386/pmmintrin.h: Ditto.
14888 * config/i386/tmmintrin.h: Ditto.
14889 * config/i386/xmmintrin.h: Ditto.
14890 * config/i386/popcntintrin.h: Ditto.
14891 * config/i386/rdseedintrin.h: Ditto.
14892 * config/i386/ammintrin.h: Ditto.
14893 * config/i386/emmintrin.h: Ditto.
14894 * config/i386/immintrin.h: Remove all header include guards.
14895 * config/i386/fma4intrin.h: Ditto.
14896 * config/i386/lwpintrin.h: Ditto.
14897 * config/i386/xopintrin.h: Ditto.
14898 * config/i386/ia32intrin.h: Ditto.
14899 * config/i386/avxintrin.h: Ditto.
14900 * config/i386/rtmintrin.h: Ditto.
14901 * config/i386/fmaintrin.h: Ditto.
14902 * config/i386/mm3dnow.h: Ditto.
14904 2013-06-22 Sriraman Tallam <tmsriram@google.com>
14906 * common/config/i386/i386-common.c: Handle LZCNT.
14908 2013-06-22 Andi Kleen <ak@linux.intel.com>
14910 * doc/extend.texi: Use __atomic_store_n instead of
14911 __atomic_store in HLE example.
14913 2013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
14915 * config/sh/sh.c: Remove <cstdlib> workaround.
14917 2013-06-21 Andi Kleen <ak@linux.intel.com>
14919 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
14921 2013-06-21 Andi Kleen <ak@linux.intel.com>
14923 * doc/extend.texi: Document that __atomic_clear and
14924 __atomic_test_and_set should only be used with bool.
14926 2013-06-20 Jan Hubicka <jh@suse.cz>
14928 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
14929 types_same_for_odr.
14930 * tree.c (decls_same_for_odr): New function.
14931 (same_for_edr): New function.
14932 (types_same_for_odr): New function.
14933 (get_binfo_at_offset): Use it.
14934 * tree.h (types_same_for_odr): Declare.
14936 2013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
14937 Jason Merrill <jason@redhat.com>
14939 * system.h: Include <cstdlib> as well as <stdlib.h>.
14941 2013-06-20 Uros Bizjak <ubizjak@gmail.com>
14944 * config/i386/i386.c (construct_container): Report error if
14945 long double is used with disabled x87 float returns.
14947 2013-06-20 Jan Hubicka <jh@suse.cz>
14949 * lto-cgraph.c (input_symtab): Do not set cgraph state.
14951 2013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
14953 PR rtl-optimization/57425
14954 PR rtl-optimization/57569
14955 * alias.c (write_dependence_p): Remove parameters mem_mode and
14956 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
14957 Changed all callers.
14958 (canon_anti_dependence): Get comments and semantics in sync.
14959 Add parameter mem_canonicalized. Changed all callers.
14960 * rtl.h (canon_anti_dependence): Update prototype.
14962 2013-06-20 Richard Biener <rguenther@suse.de>
14964 * data-streamer-in.c (streamer_read_uhwi): Optimize single
14965 byte case, inline streamer_read_uchar and defer section
14968 2013-06-20 Richard Biener <rguenther@suse.de>
14970 PR tree-optimization/57584
14971 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
14972 SSA names into the expanded expression that take part in
14973 abnormal coalescing.
14975 2013-06-19 Sharad Singhai <singhai@google.com>
14977 * gcov.c (print_usage): Handle new option.
14978 (process_args): Ditto.
14979 (get_gcov_intermediate_filename): New function.
14980 (output_intermediate_file): New function.
14981 (output_gcov_file): New function
14982 (generate_results): Handle new option.
14983 (release_function): Relase demangled name.
14984 (read_graph_file): Handle demangled name.
14985 (output_lines): Ditto.
14986 * doc/gcov.texi: Document gcov intermediate format.
14988 2013-06-19 Vladimir Makarov <vmakarov@redhat.com>
14991 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
14992 (lra_emit_add): Use the functions. Add comment about Y as an
14995 2013-06-19 David Edelsohn <dje.gcc@gmail.com>
14998 * collect2.c (collect_atexit): New.
14999 (collect_exit): Delete.
15000 (main): Register collect_atexit with atexit.
15001 (collect_wait): Change collect_exit to exit.
15003 * collect2.h (collect_exit): Delete.
15004 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
15006 2013-06-19 Wei Mi <wmi@google.com>
15008 PR rtl-optimization/57518
15009 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
15010 if regno is used in paradoxical subreg.
15011 (update_equiv_regs): Check pdx_subregs[regno] before
15012 set a reg to be equivalent with a mem.
15014 2013-06-19 Matthias Klose <doko@ubuntu.com>
15017 * file-find.h (find_a_file): Add a mode parameter.
15018 * file-find.c (find_a_file): Likewise.
15019 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
15020 with X_OK for the executables.
15021 * collect2.c (main): Call find_a_file with X_OK.
15023 2013-06-19 Steve Ellcey <sellcey@mips.com>
15026 * config/mips/mips.md (casesi_internal_mips16_<mode>):
15027 Use NEXT_INSN instead of next_real_insn.
15029 2013-06-19 Jan Hubicka <jh@suse.cz>
15031 * cgraph.h (const_value_known_p): Replace by ...
15032 (ctor_for_folding): .. this one.
15033 * cgraphunit.c (process_function_and_variable_attributes): Use it.
15034 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
15035 * expr.c (expand_expr_real_1): Likewise.
15036 (string_constant): Likewise.
15037 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
15038 * ipa.c (process_references): Likewise.
15039 (symtab_remove_unreachable_nodes): Likewise.
15040 * ipa-inline-analysis.c (param_change_prob): Likewise.
15041 * gimple-fold.c (canonicalize_constructor_val): Likewise.
15042 (get_base_constructor): Likwise.
15043 * varpool.c (varpool_remove_node): Likewise.
15044 (varpool_remove_initializer): LIkewise.
15045 (dump_varpool_node): LIkwise.
15046 (const_value_known_p): Rewrite to ...
15047 (ctor_for_folding): ... this one.
15049 2013-06-19 Jakub Jelinek <jakub@redhat.com>
15052 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
15053 PERSONALITY in $PATH derived prefixes.
15055 2013-06-19 Jeff Law <law@redhat.com>
15057 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
15060 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
15061 (simplify_bitwise_binary): Use it to simpify certain binary ops on
15064 2013-06-19 Sofiane Naci <sofiane.naci@arm.com>
15066 * config/arm/vfp.md: Move VFP instruction classification documentation
15068 * config/arm/arm.md: ... here. Update instruction classification
15071 2013-06-19 Richard Earnshaw <rearnsha@arm.com>
15073 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
15074 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
15075 pattern. Use more efficient sequences on ARMv5 and Thumb2.
15077 2013-06-19 Steven Bosscher <steven@gcc.gnu.org>
15080 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
15081 with NEXT_INSN. Use tablejump_p to check for jump table data
15084 2013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
15087 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
15088 that now in C++ the value is correct per the C++ standards.
15090 2013-06-19 Richard Biener <rguenther@suse.de>
15092 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
15093 for global context.
15095 2013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
15098 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
15101 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
15102 with next_active_insn.
15104 2013-06-18 Sriraman Tallam <tmsriram@google.com>
15106 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
15107 functions are inlined during failures to flag an error.
15108 * tree-inline.c (expand_call_inline): Allow the error to be flagged
15109 in early inline pass.
15111 2013-06-18 H.J. Lu <hongjiu.lu@intel.com>
15113 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
15116 2013-06-18 Julian Brown <julian@codesourcery.com>
15118 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
15119 Permit virtual register pre-reload if !strict.
15120 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
15122 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
15124 * config/arm/neon.md (movmisalign<mode>): Use
15125 neon_perm_struct_or_reg_operand instead of
15126 neon_struct_or_register_operand.
15127 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
15128 neon_permissive_struct_operand instead of neon_struct_operand.
15129 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
15130 neon_vector_mem_operand.
15131 * config/arm/predicates.md (neon_struct_operand): Adjust call to
15132 neon_vector_mem_operand.
15133 (neon_permissive_struct_operand): New.
15134 (neon_struct_or_register_operand): Rename to...
15135 (neon_perm_struct_or_reg_operand): This. Adjust call to
15136 neon_vector_mem_operand.
15138 2013-06-18 Richard Biener <rguenther@suse.de>
15140 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
15141 * lto-streamer.h: Include pointer-set.h.
15142 (struct lto_decl_slot): Remove.
15143 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
15144 Remove next_index entry.
15145 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
15146 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
15147 (lto_init_tree_ref_encoder): Adjust.
15148 (lto_destroy_tree_ref_encoder): Likewise.
15149 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
15150 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
15151 (lto_output_decl_index): Adjust.
15152 (lto_new_out_decl_state): Likewise.
15153 (lto_record_function_out_decl_state): Likewise.
15154 * lto-streamer-out.c (copy_function): Likewise.
15156 2013-06-18 Richard Biener <rguenther@suse.de>
15158 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
15159 * cgraphunit.c: Include cfgloop.h.
15160 (init_lowered_empty_function): Initialize the loop tree.
15161 (assemble_thunk): Insert new BBs into loops.
15163 2013-06-18 Richard Biener <rguenther@suse.de>
15165 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
15166 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
15167 the map from cache entry to cache index optional.
15168 (streamer_tree_cache_replace_tree): Adjust accordingly.
15169 (streamer_tree_cache_append): Likewise.
15170 (streamer_tree_cache_delete): Likewise.
15171 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
15172 streamer cache map from cache entry to cache index.
15173 * lto-streamer-out.c (create_output_block): Adjust.
15175 2013-06-18 Sofiane Naci <sofiane.naci@arm.com>
15177 * config/arm/arm.md (attribute "insn"): Move multiplication and
15178 division attributes to...
15179 (attribute "type"): ... here. Remove mult.
15180 (attribute "mul32"): New attribute.
15181 (attribute "mul64"): Add umaal.
15182 (*arm_mulsi3): Update attributes.
15183 (*arm_mulsi3_v6): Likewise.
15184 (*thumb_mulsi3): Likewise.
15185 (*thumb_mulsi3_v6): Likewise.
15186 (*mulsi3_compare0): Likewise.
15187 (*mulsi3_compare0_v6): Likewise.
15188 (*mulsi_compare0_scratch): Likewise.
15189 (*mulsi_compare0_scratch_v6): Likewise.
15190 (*mulsi3addsi): Likewise.
15191 (*mulsi3addsi_v6): Likewise.
15192 (*mulsi3addsi_compare0): Likewise.
15193 (*mulsi3addsi_compare0_v6): Likewise.
15194 (*mulsi3addsi_compare0_scratch): Likewise.
15195 (*mulsi3addsi_compare0_scratch_v6): Likewise.
15196 (*mulsi3subsi): Likewise.
15197 (*mulsidi3adddi): Likewise.
15198 (*mulsi3addsi_v6): Likewise.
15199 (*mulsidi3adddi_v6): Likewise.
15200 (*mulsidi3_nov6): Likewise.
15201 (*mulsidi3_v6): Likewise.
15202 (*umulsidi3_nov6): Likewise.
15203 (*umulsidi3_v6): Likewise.
15204 (*umulsidi3adddi): Likewise.
15205 (*umulsidi3adddi_v6): Likewise.
15206 (*smulsi3_highpart_nov6): Likewise.
15207 (*smulsi3_highpart_v6): Likewise.
15208 (*umulsi3_highpart_nov6): Likewise.
15209 (*umulsi3_highpart_v6): Likewise.
15210 (mulhisi3): Likewise.
15211 (*mulhisi3tb): Likewise.
15212 (*mulhisi3bt): Likewise.
15213 (*mulhisi3tt): Likewise.
15214 (maddhisi4): Likewise.
15215 (*maddhisi4tb): Likewise.
15216 (*maddhisi4tt): Likewise.
15217 (maddhidi4): Likewise.
15218 (*maddhidi4tb): Likewise.
15219 (*maddhidi4tt): Likewise.
15220 (divsi3): Likewise.
15221 (udivsi3): Likewise.
15222 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
15223 (thumb2_mulsi_short_compare0): Likewise.
15224 (thumb2_mulsi_short_compare0_scratch): Likewise.
15225 * config/arm/arm1020e.md (1020mult1): Update attribute change.
15226 (1020mult2): Likewise.
15227 (1020mult3): Likewise.
15228 (1020mult4): Likewise.
15229 (1020mult5): Likewise.
15230 (1020mult6): Likewise.
15231 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
15233 (cortex_a15_mult64): Likewise.
15234 (cortex_a15_sdiv): Likewise.
15235 (cortex_a15_udiv): Likewise.
15236 * config/arm/arm1026ejs.md (mult1): Update attribute change.
15242 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
15243 (pj4_ir_div): Likewise.
15244 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
15245 (11_mult2): Likewise.
15246 (11_mult3): Likewise.
15247 (11_mult4): Likewise.
15248 (11_mult5): Likewise.
15249 (11_mult6): Likewise.
15250 (11_mult7): Likewise.
15251 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
15252 (cortex_a8_mla): Likewise.
15253 (cortex_a8_mull): Likewise.
15254 (cortex_a8_smulwy): Likewise.
15255 (cortex_a8_smlald): Likewise.
15256 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
15257 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
15258 (cortex_r4_mul_3): Likewise.
15259 (cortex_r4_mla_4): Likewise.
15260 (cortex_r4_mla_3): Likewise.
15261 (cortex_r4_smlald): Likewise.
15262 (cortex_r4_mull): Likewise.
15263 (cortex_r4_sdiv): Likewise.
15264 (cortex_r4_udiv): Likewise.
15265 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
15266 (cortex_a7_idiv): Likewise.
15267 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
15268 (9_mult2): Likewise.
15269 (9_mult3): Likewise.
15270 (9_mult4): Likewise.
15271 (9_mult5): Likewise.
15272 (9_mult6): Likewise.
15273 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
15274 (cortex_a53_sdiv): Likewise.
15275 (cortex_a53_udiv): Likewise.
15276 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
15277 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
15278 (mp626_mult2): Likewise.
15279 (mp626_mult3): Likewise.
15280 (mp626_mult4): Likewise.
15281 * config/arm/fa526.md (526_mult1): Update attribute change.
15282 (526_mult2): Likewise.
15283 * config/arm/arm-generic.md (mult): Update attribute change.
15284 (mult_ldsched_strongarm): Likewise.
15285 (mult_ldsched): Likewise.
15286 (multi_cycle): Likewise.
15287 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
15288 * config/arm/fa606te.md (606te_mult1): Update attribute change.
15289 (606te_mult2): Likewise.
15290 (606te_mult3): Likewise.
15291 (606te_mult4): Likewise.
15292 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
15293 (cortex_a9_mac16): Likewise.
15294 (cortex_a9_multiply): Likewise.
15295 (cortex_a9_mac): Likewise.
15296 (cortex_a9_multiply_long): Likewise.
15297 * config/arm/fa626te.md (626te_mult1): Update attribute change.
15298 (626te_mult2): Likewise.
15299 (626te_mult3): Likewise.
15300 (626te_mult4): Likewise.
15302 2013-06-18 Richard Biener <rguenther@suse.de>
15305 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
15307 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
15310 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
15311 with next_active_insn.
15313 2013-06-18 Alan Modra <amodra@gmail.com>
15315 * config/rs6000/rs6000.h (enum data_align): New.
15316 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
15317 (DATA_ABI_ALIGNMENT): Define.
15318 (CONSTANT_ALIGNMENT): Correct comment.
15319 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
15320 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
15322 2013-06-17 David Malcolm <dmalcolm@redhat.com>
15324 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
15325 ATTRIBUTE_UNUSED marking.
15327 2013-06-17 Sofiane Naci <sofiane.naci@arm.com>
15329 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
15330 alternative and update.
15331 (aarch64_dup_lanedi): Delete.
15332 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
15333 * config/aarch64/aarch64-simd-builtins.def: Update.
15335 2013-06-17 Richard Biener <rguenther@suse.de>
15337 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
15338 (lto_input_scc): Declare.
15339 (lto_input_tree_1): Likewise.
15340 (struct lto_stats_d): Add num_tree_bodies_output and
15341 num_pickle_refs_output.
15342 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
15343 (lto_read_tree_1): Split out from ...
15344 (lto_read_tree): ... this.
15345 (lto_input_scc): New function.
15346 (lto_input_tree_1): Split out from ...
15347 (lto_input_tree): ... this. Handle LTO_tree_scc.
15348 (lto_data_in_create): Create the streamer cache without hashes.
15349 * lto-streamer-out.c (create_output_block): Create the streamer
15350 cache with hashes when not doing WPA.
15351 (lto_write_tree_1): Split out from ...
15352 (lto_write_tree): ... this.
15353 (get_symbol_initial_value): New function.
15354 (lto_output_tree_1): Split out from ...
15355 (lto_output_tree): ... this. Write trees as series of SCCs
15356 using a DFS walk via DFS_write_tree.
15357 (struct sccs, struct scc_entry): New types.
15358 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
15359 (DFS_write_tree_body): New function.
15360 (DFS_write_tree): Likewise.
15361 (hash_tree): Likewise.
15362 (scc_entry_compare): Likewise.
15363 (hash_scc): Likewise.
15364 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
15365 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
15366 TREE_CHAIN as regular reference.
15367 (streamer_read_integer_cst): Remove.
15368 (streamer_get_pickled_tree): Adjust.
15369 * tree-streamer-out.c (streamer_write_chain): Disable streaming
15370 of DECL_EXTERNALs in BLOCK_VARS for now.
15371 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
15373 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
15374 Add hash value argument and record that if hashes are recorded
15376 (streamer_tree_cache_insert_1): Adjust.
15377 (streamer_tree_cache_insert): Likewise.
15378 (streamer_tree_cache_insert_at): Rename to ...
15379 (streamer_tree_cache_replace_tree): ... this and adjust.
15380 (streamer_tree_cache_append): Adjust.
15381 (record_common_node): Likewise.
15382 (streamer_tree_cache_create): Add argument whether to
15383 record hash values together with trees.
15384 (streamer_tree_cache_delete): Adjust.
15385 * tree-streamer.h (struct streamer_tree_cache_d): Add
15387 (streamer_read_integer_cst): Remove.
15388 (streamer_tree_cache_insert): Adjust.
15389 (streamer_tree_cache_append): Likewise.
15390 (streamer_tree_cache_insert_at): Rename to ...
15391 (streamer_tree_cache_replace_tree): ... this and adjust.
15392 (streamer_tree_cache_create): Add argument whether to record hashes.
15393 (streamer_tree_cache_get): Rename to ...
15394 (streamer_tree_cache_get_tree): ... this.
15395 (streamer_tree_cache_get_hash): New function.
15396 * tree.c (cache_integer_cst): New function.
15397 * tree.h (cache_integer_cst): Declare.
15398 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
15399 * lto-symtab.c (lto_varpool_replace_node): Only release
15400 DECL_INITIAL of non-prevailing decls.
15401 * varpool.c (varpool_remove_initializer): Do not release
15402 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
15404 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
15406 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
15407 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
15408 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
15409 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
15410 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
15411 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
15412 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
15413 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
15414 instead of TARGET_64BIT.
15415 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
15416 Require ISA_HAS_<D>DIV.
15418 2013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
15420 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
15421 (mips*-*-linux*): Move default with_llsc setting to where other
15423 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
15425 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
15426 Likewise. Remove default with_tune setting. Move default float
15427 setting to its own block. Handle with_llsc in the same block as above.
15429 2013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
15431 PR rtl-optimization/57425
15432 PR rtl-optimization/57569
15433 * alias.c (write_dependence_p): Add new parameters mem_mode,
15434 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
15435 Changed all callers.
15436 (canon_anti_dependence): New function.
15437 * cse.c (check_dependence): Use canon_anti_dependence.
15438 * cselib.c (cselib_invalidate_mem): Likewise.
15439 * rtl.h (canon_anti_dependence): Declare.
15441 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
15443 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
15444 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
15445 ".set mips3" for 64-bit targets.
15447 2013-06-15 Dehao Chen <dehao@google.com>
15449 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
15450 * gimple-low.c (gimple_check_call_matching_types): Likewise.
15451 (gimple_check_call_args): Likewise.
15452 * value-prof.c (check_ic_target): Likewise.
15453 * ipa-inline.c (early_inliner): Likewise.
15454 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
15455 * cgraph.c (cgraph_create_edge_1): Likewise.
15456 (cgraph_make_edge_direct): Likewise.
15458 2013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
15461 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
15462 rs6000_output_move_128bit to handle emitting quad memory
15463 operations. Set attribute length to 8 bytes.
15465 2013-06-14 Vidya Praveen <vidyapraveen@arm.com>
15467 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
15469 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
15470 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
15471 (aarch64_<su>mlsl<mode>): Likewise.
15473 2013-06-14 Mike Stump <mikestump@comcast.net>
15475 * Makefile.in (TARGET_H): Add insn-codes.h.
15477 2013-06-14 Alan Modra <amodra@gmail.com>
15479 PR middle-end/57134
15480 PR middle-end/57586
15481 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
15482 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
15483 bitfield expansion when EXPAND_MEMORY.
15484 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
15486 2013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
15488 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
15489 test for clearing quad memory on 32-bit later.
15491 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
15493 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
15494 (fold_negate_expr): Likewise.
15495 (fold_real_zero_addition_p): Handle vectors.
15496 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
15498 2013-06-14 Alan Modra <amodra@gmail.com>
15500 * varasm.c (force_const_mem): Revert 2013-06-07 change.
15502 2013-06-13 Jan Hubicka <jh@suse.cz>
15504 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
15505 Local comdats are not externally visible.
15506 * symtab.c (dump_symtab_base): Dump externally visible.
15507 (verify_symtab_base): Verify back links in the symtab hash.
15509 2013-06-13 Bin Cheng <bin.cheng@arm.com>
15511 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
15512 CONVERT_EXPR as equal nodes.
15514 2013-06-13 Bin Cheng <bin.cheng@arm.com>
15516 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
15518 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
15520 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
15521 Generalize to complex and vector.
15522 * tree.c (build_all_ones_cst): New function.
15523 * tree.h (build_all_ones_cst): Declare it.
15525 2013-06-13 Alan Modra <amodra@gmail.com>
15527 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
15528 * config/rs6000/rs6000.md (signbittf2): New insn.
15529 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
15530 (abstf2_internal, cmptf_internal2): Likewise.
15531 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
15533 2013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
15534 Pat Haugen <pthaugen@us.ibm.com>
15535 Peter Bergner <bergner@vnet.ibm.com>
15537 * config/rs6000/rs6000.c (emit_load_locked): Add support for
15538 power8 byte, half-word, and quad-word atomic instructions.
15539 (emit_store_conditional): Likewise.
15540 (rs6000_expand_atomic_compare_and_swap): Likewise.
15541 (rs6000_expand_atomic_op): Likewise.
15543 * config/rs6000/sync.md (larx): Add new modes for power8.
15545 (AINT): New mode iterator to include TImode as well as normal
15546 integer modes on power8.
15547 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
15548 that VSX registers are not considered. Use AINT mode iterator
15549 instead of INT1 to allow inclusion of quad word atomic operations
15551 (load_locked<mode>): Likewise.
15552 (store_conditional<mode>): Likewise.
15553 (atomic_compare_and_swap<mode>): Likewise.
15554 (atomic_exchange<mode>): Likewise.
15555 (atomic_nand<mode>): Likewise.
15556 (atomic_fetch_<fetchop_name><mode>): Likewise.
15557 (atomic_nand_fetch<mode>): Likewise.
15558 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
15560 (ATOMIC): On power8, add QImode, HImode modes.
15561 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
15562 modes that promote to SImode.
15563 (load_lockedti): Convert TImode arguments to PTImode, so that we
15564 get a guaranteed even/odd register pair.
15565 (load_lockedpti): Likewise.
15566 (store_conditionalti): Likewise.
15567 (store_conditionalpti): Likewise.
15569 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
15570 atomic load/store instructions.
15573 2013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
15575 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
15577 (insn_count): New attribute, with most cases extracted from...
15578 (length): ...here. Redefine most cases in terms of insn_count.
15579 (single_insn): Delete.
15580 (can_delay): Use insn_count to check for single instructions.
15581 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
15582 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
15583 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
15584 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
15585 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
15586 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
15587 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
15588 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
15589 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
15590 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
15591 rather than "length".
15592 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
15593 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
15594 Use "insn_count" rather than "length".
15595 * config/mips/mips-dsp.md
15596 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
15597 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
15600 2013-06-12 Marc Glisse <marc.glisse@inria.fr>
15602 PR tree-optimization/57361
15603 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
15605 2013-06-12 Sofiane Naci <sofiane.naci@arm.com>
15607 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
15609 (aarch64_simd_combine<mode>): New instruction expansion.
15610 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
15611 function prototype.
15612 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
15613 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
15615 2013-06-12 Jan Hubicka <jh@suse.cz>
15617 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
15618 decl has when in streaming stage.
15619 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
15620 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
15622 2013-06-12 Roland Stigge <stigge@antcom.de>
15625 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
15627 2013-06-12 Jakub Jelinek <jakub@redhat.com>
15629 PR tree-optimization/57537
15630 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
15631 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
15633 2013-06-12 Richard Biener <rguenther@suse.de>
15635 * data-streamer.h (streamer_write_char_stream): CSE
15636 obs->current_pointer.
15637 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
15638 streamer_write_char_stream manually and optimize the resulting loop.
15639 (streamer_write_hwi_stream): Likewise.
15641 2013-06-12 Jan Hubicka <jh@suse.cz>
15643 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
15644 * cgraph.h (varpool_create_empty_node): Declare.
15645 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
15647 * symtab.c (symtab_unregister_node): Be lax about missin entries
15649 (symtab_get_node): Update comment.
15650 * varpool.c (varpool_create_empty_node): Break out from ...
15651 (varpool_node_for_decl): ... here.
15652 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
15654 2013-06-12 Eric Botcazou <ebotcazou@adacore.com>
15656 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
15657 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
15658 part. Use straight-line flow at the end.
15659 <COMPONENT_REF>: Remove superfluous else.
15660 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
15662 2013-06-12 Jakub Jelinek <jakub@redhat.com>
15665 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
15666 target hook even for !TREE_PUBLIC decls. If no resolution info
15667 is available, return false for common and external decls.
15669 2013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
15671 * config/rl78/constraints.md (U): New constraint.
15672 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
15675 2013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
15678 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
15679 to allow returning address to AT_PLATFORM name.
15681 2013-06-11 Jan Hubicka <jh@suse.cz>
15683 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
15684 * cgraph.h (symtab_node_base): Add weakref flag.
15685 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
15686 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
15687 (output_weakrefs): Use weakref flag.
15688 * fold-const.c (simple_operand_p): Handle WEAK.
15689 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
15690 * ipa.c (varpool_externally_visible_p): Drop weakref.
15691 (function_and_variable_visibility): Update comment; fix weakref
15692 sanity checks; do not clear DECL_WEAK on them.
15693 * lto-cgraph.c (lto_output_node): update.
15694 (lto_output_varpool_node): Update.
15695 (input_overwrite_node): Update.
15696 (input_node): Update.
15697 (input_varpool_node): Update.
15698 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
15699 (lto_symtab_merge_symbols): Add sanity check.
15700 (lto_symtab_prevailing_decl): Do not special case weakrefs.
15701 * passes.c (rest_of_decl_compilation): Set static flag, too.
15702 * symtab.c (dump_symtab_base): Dump weakref.
15703 (verify_symtab_base): Sanity check weakrefs.
15704 (symtab_make_decl_local): Remove duplicated code.
15705 (symtab_alias_ultimate_target): Simplify.
15706 * varpool.c (varpool_create_variable_alias): Set weakref flag.
15708 2013-06-11 Tom de Vries <tom@codesourcery.com>
15710 * genautomata.c (gen_regexp_sequence): Handle els_num == -1. Handle
15711 sequence_vect == NULL.
15713 2013-06-11 DJ Delorie <dj@redhat.com>
15715 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
15716 (rl78_unwind_word_mode): New.
15718 2013-06-11 David Malcolm <dmalcolm@redhat.com>
15720 * final.c (debug_prefix_maps): Make static.
15722 2013-06-11 David Malcolm <dmalcolm@redhat.com>
15724 * function.c (initial_trampoline): Remove stray copy.
15726 2013-06-11 Sofiane Naci <sofiane.naci@arm.com>
15728 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
15730 2013-06-11 Martin Jambor <mjambor@suse.cz>
15732 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
15733 within bounds at the beginning of the function.
15735 2013-06-11 Alan Modra <amodra@gmail.com>
15737 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
15739 (get_named_section): Don't NULL !DECL_P decl.
15741 2013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
15743 * doc/invoke.texi (core-avx2): Document.
15745 (atom): Updated with MOVBE.
15747 2013-06-11 Richard Biener <rguenther@suse.de>
15749 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
15751 2013-06-11 Anton Blanchard <anton@samba.org>
15753 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
15754 correct shift value in little-endian mode.
15756 2013-06-11 Jakub Jelinek <jakub@redhat.com>
15759 * varasm.c (get_variable_align): Move #endif to the right place.
15761 2013-06-10 Cary Coutant <ccoutant@google.com>
15763 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
15764 for hash so that hash table traversal order is deterministic.
15766 2013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
15767 Pat Haugen <pthaugen@us.ibm.com>
15768 Peter Bergner <bergner@vnet.ibm.com>
15770 * config/rs6000/vector.md (GPR move splitter): Do not split moves
15771 of vectors in GPRS if they are direct moves or quad word load or
15774 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
15776 (direct_move_p): Likewise.
15777 (quad_load_store_p): Likewise.
15779 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
15780 classes into bins based on the physical register type.
15781 (reg_class_to_reg_type): Likewise.
15782 (IS_STD_REG_TYPE): Likewise.
15783 (IS_FP_VECT_REG_TYPE): Likewise.
15784 (reload_fpr_gpr): Arrays to determine what insn to use if we can
15785 use direct move instructions.
15786 (reload_gpr_vsx): Likewise.
15787 (reload_vsx_gpr): Likewise.
15788 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
15789 information that is a simplification of register classes. Also
15790 precalculate direct move reload helpers.
15791 (direct_move_p): New function to return true if the operation can
15792 be done as a direct move instruciton.
15793 (quad_load_store_p): New function to return true if the operation
15794 is a quad memory operation.
15795 (rs6000_legitimize_address): If quad memory, only allow register
15796 indirect for TImode addresses.
15797 (rs6000_legitimate_address_p): Likewise.
15798 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
15799 (rs6000_reload_register_type): Likewise.
15800 (register_to_reg_type): Return register type.
15801 (rs6000_secondary_reload_simple_move): New helper function for
15802 secondary reload and secondary memory needed to identify anything
15803 that is a simple move, and does not need reloading.
15804 (rs6000_secondary_reload_direct_move): New helper function for
15805 secondary reload to identify cases that can be done with several
15806 instructions via the direct move instructions.
15807 (rs6000_secondary_reload_move): New helper function for secondary
15808 reload to identify moves between register types that can be done.
15809 (rs6000_secondary_reload): Add support for quad memory operations
15810 and for direct move.
15811 (rs6000_secondary_memory_needed): Likewise.
15812 (rs6000_debug_secondary_memory_needed): Change argument names.
15813 (rs6000_output_move_128bit): New function to return the move to
15814 use for 128-bit moves, including knowing about the various
15815 limitations of quad memory operations.
15817 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
15818 memory operations. call rs6000_output_move_128bit for the actual
15819 instruciton(s) to generate.
15820 (vsx_movti_64bit): Likewise.
15822 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
15823 (UNSPEC_P8V_MTVSRWZ): Likewise.
15824 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
15825 (UNSPEC_P8V_MTVSRD): Likewise.
15826 (UNSPEC_P8V_XXPERMDI): Likewise.
15827 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
15828 (UNSPEC_FUSION_GPR): Likewise.
15829 (FMOVE128_GPR): New iterator for direct move.
15830 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
15831 (f32_sv): Likewise.
15832 (f32_dm): Likewise.
15833 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
15834 loads and direct move instructions.
15835 (zero_extendsidi2_lfiwzx): Likewise.
15836 (extendsidi2_lfiwax): Likewise.
15837 (extendsidi2_nocell): Likewise.
15838 (floatsi<mode>2_lfiwax): Likewise.
15839 (lfiwax): Likewise.
15840 (floatunssi<mode>2_lfiwzx): Likewise.
15841 (lfiwzx): Likewise.
15842 (fix_trunc<mode>_stfiwx): Likewise.
15843 (fixuns_trunc<mode>_stfiwx): Likewise.
15844 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
15845 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
15846 (parity<mode>2_cmpb): Set length/type attr.
15847 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
15848 for 'mr.' to fast_compare.
15849 (bpermd_<mode>): Change type attr to popcnt.
15850 (p8_fmrgow_<mode>): New insns for power8 direct move support.
15851 (p8_mtvsrwz_1): Likewise.
15852 (p8_mtvsrwz_2): Likewise.
15853 (reload_fpr_from_gpr<mode>): Likewise.
15854 (p8_mtvsrd_1): Likewise.
15855 (p8_mtvsrd_2): Likewise.
15856 (p8_xxpermdi_<mode>): Likewise.
15857 (reload_vsx_from_gpr<mode>): Likewise.
15858 (reload_vsx_from_gprsf): Likewise.
15859 (p8_mfvsrd_3_<mode>): LIkewise.
15860 (reload_gpr_from_vsx<mode>): Likewise.
15861 (reload_gpr_from_vsxsf): Likewise.
15862 (p8_mfvsrd_4_disf): Likewise.
15863 (multi-word GPR splits): Do not split direct moves or quad memory
15866 2013-06-10 David Malcolm <dmalcolm@redhat.com>
15868 * tree-into-ssa.c (interesting_blocks): Make static.
15870 2013-06-10 Jakub Jelinek <jakub@redhat.com>
15873 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
15874 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
15875 Use DATA_ABI_ALIGNMENT for that case instead if defined.
15876 (get_variable_align): New function.
15877 (get_variable_section, emit_bss, emit_common,
15878 assemble_variable_contents, place_block_symbol): Use
15879 get_variable_align instead of DECL_ALIGN.
15880 (assemble_noswitch_variable): Add align argument, use it
15881 instead of DECL_ALIGN.
15882 (assemble_variable): Adjust caller. Use get_variable_align
15883 instead of DECL_ALIGN.
15884 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
15886 (DATA_ABI_ALIGNMENT): Define.
15887 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
15888 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
15889 opt is false, only return the psABI mandated alignment increase.
15890 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
15891 (DATA_ABI_ALIGNMENT): ... this.
15892 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
15893 (DATA_ABI_ALIGNMENT): ... this.
15894 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
15895 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
15896 (DATA_ABI_ALIGNMENT): ... this.
15897 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
15898 * doc/tm.texi: Regenerated.
15900 2013-06-10 Uros Bizjak <ubizjak@gmail.com>
15902 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
15903 cmp_code to construct REG_EQUAL note.
15905 2013-06-09 Jakub Jelinek <jakub@redhat.com>
15908 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
15909 that operands[2] doesn't overlap with operands[0].
15911 2013-06-09 David Edelsohn <dje.gcc@gmail.com>
15912 Jan Hubicka <jh@suse.cz>
15914 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
15915 hack to mark symbols as used.
15917 2013-06-08 Vladimir Makarov <vmakarov@redhat.com>
15919 PR rtl-optimization/57559
15920 * lra-constraints.c (process_alt_operands): Don't discourage
15921 memory with known offset for offsetable memory constraint.
15922 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
15924 2013-06-08 Eric Botcazou <ebotcazou@adacore.com>
15926 * varasm.c (struct oc_local_state): Reorder fields.
15927 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
15928 and adjust accordingly.
15929 (output_constructor): Reorder initialization code and adjust call to
15930 output_constructor_bitfield.
15932 2013-06-07 Jan Hubicka <jh@suse.cz>
15934 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
15936 2013-06-07 David Malcolm <dmalcolm@redhat.com>
15938 * tree-object-size.c (unknown): Make const.
15940 2013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
15942 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
15943 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
15944 for last alternative in the cpu_facility attribute.
15946 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15949 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
15950 (xordi3): Change operand 2 constraint to arm_xordi_operand.
15951 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
15952 * config/arm/constraints.md (Dg): New constraint.
15953 * config/arm/neon.md (xordi3_neon): Remove.
15954 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
15955 * config/arm/predicates.md (arm_xordi_operand): New predicate.
15957 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15959 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
15960 Clean up alternatives.
15962 2013-06-07 Alan Modra <amodra@gmail.com>
15964 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
15967 2013-06-07 Alan Modra <amodra@gmail.com>
15969 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
15971 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15973 * config/arm/constraints.md (Df): New constraint.
15974 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
15975 Correct length attribute for last two alternatives.
15977 2013-06-07 Alan Modra <amodra@gmail.com>
15979 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
15980 override user -mfp-in-toc.
15981 (offsettable_ok_by_alignment): Consider just the current access
15982 rather than the whole object, unless BLKmode. Handle
15983 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
15984 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
15985 for -mcmodel=medium.
15986 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
15987 override user -mfp-in-toc or -msum-in-toc. Default to
15988 -mno-fp-in-toc for -mcmodel=medium.
15990 2013-06-06 DJ Delorie <dj@redhat.com>
15992 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
15993 TARGET_VALID_POINTER_MODE.
15995 2013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
15996 Pat Haugen <pthaugen@us.ibm.com>
15997 Peter Bergner <bergner@vnet.ibm.com>
15999 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
16000 Document new power8 builtins.
16002 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
16003 condition code register, to allow 128-bit logical operations to be
16004 done in the VSX or GPR registers.
16005 (nor<mode>3): Use the canonical form for nor.
16006 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
16007 vclz*, and vpopcnt* vector instructions.
16008 (nand<mode>3): Likewise.
16009 (orc<mode>3): Likewise.
16010 (clz<mode>2): LIkewise.
16011 (popcount<mode>2): Likewise.
16013 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
16014 that only the GPRs are recognized.
16016 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
16017 support for new power8 builtins.
16019 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
16021 (xscvdpspn): Likewise.
16027 (vpopcnt): Likewise.
16028 (vpopcntb): Likewise.
16029 (vpopcnth): Likewise.
16030 (vpopcntw): Likewise.
16031 (vpopcntd): Likewise.
16033 (vmrgew): Likewise.
16034 (vmrgow): Likewise.
16036 (eqv_v16qi3): Likewise.
16037 (eqv_v8hi3): Likewise.
16038 (eqv_v4si3): Likewise.
16039 (eqv_v2di3): Likewise.
16040 (eqv_v4sf3): Likewise.
16041 (eqv_v2df3): Likewise.
16043 (nand_v16qi3): Likewise.
16044 (nand_v8hi3): Likewise.
16045 (nand_v4si3): Likewise.
16046 (nand_v2di3): Likewise.
16047 (nand_v4sf3): Likewise.
16048 (nand_v2df3): Likewise.
16050 (orc_v16qi3): Likewise.
16051 (orc_v8hi3): Likewise.
16052 (orc_v4si3): Likewise.
16053 (orc_v2di3): Likewise.
16054 (orc_v4sf3): Likewise.
16055 (orc_v2df3): Likewise.
16057 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
16058 allow power8 quad mode in 64-bit.
16059 (rs6000_builtin_vectorized_function): Add support to vectorize
16060 ISA 2.07 count leading zeros, population count builtins.
16061 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
16062 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
16063 (builtin_function_type): Add vgbbd builtin function which takes an
16065 (altivec_expand_vec_perm_const): Add support for new power8 merge
16068 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
16069 that does not include TImdoe for use with 32-bit.
16070 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
16072 (UNSPEC_VSX_CVDPSPN): Likewise.
16073 (vsx_xscvdpspn): Likewise.
16074 (vsx_xscvspdpn): Likewise.
16075 (vsx_xscvdpspn_scalar): Likewise.
16076 (vsx_xscvspdpn_directmove): Likewise.
16077 (vsx_and<mode>3): Split logical operations into 32-bit and
16078 64-bit. Add support to do logical operations on TImode as well as
16079 VSX vector types. Allow logical operations to be done in either
16080 VSX registers or in general purpose registers in 64-bit mode. Add
16081 splitters if GPRs were used. For AND, add clobber of CCmode to
16082 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
16084 (vsx_and<mode>3_32bit): Likewise.
16085 (vsx_and<mode>3_64bit): Likewise.
16086 (vsx_ior<mode>3): Likewise.
16087 (vsx_ior<mode>3_32bit): Likewise.
16088 (vsx_ior<mode>3_64bit): Likewise.
16089 (vsx_xor<mode>3): Likewise.
16090 (vsx_xor<mode>3_32bit): Likewise.
16091 (vsx_xor<mode>3_64bit): Likewise.
16092 (vsx_one_cmpl<mode>2): Likewise.
16093 (vsx_one_cmpl<mode>2_32bit): Likewise.
16094 (vsx_one_cmpl<mode>2_64bit): Likewise.
16095 (vsx_nor<mode>3): Likewise.
16096 (vsx_nor<mode>3_32bit): Likewise.
16097 (vsx_nor<mode>3_64bit): Likewise.
16098 (vsx_andc<mode>3): Likewise.
16099 (vsx_andc<mode>3_32bit): Likewise.
16100 (vsx_andc<mode>3_64bit): Likewise.
16101 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
16102 and xxlorc instructions.
16103 (vsx_eqv<mode>3_64bit): Likewise.
16104 (vsx_nand<mode>3_32bit): Likewise.
16105 (vsx_nand<mode>3_64bit): Likewise.
16106 (vsx_orc<mode>3_32bit): Likewise.
16107 (vsx_orc<mode>3_64bit): Likewise.
16109 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
16111 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
16113 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
16114 (p8_vmrgow): Likewise.
16115 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
16116 GPRs to be split under VSX.
16117 (p8v_clz<mode>2): Add power8 count leading zero support.
16118 (p8v_popcount<mode>2): Add power8 population count support.
16119 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
16122 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
16125 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
16127 (vec_nand): Likewise.
16128 (vec_vclz): Likewise.
16129 (vec_vclzb): Likewise.
16130 (vec_vclzd): Likewise.
16131 (vec_vclzh): Likewise.
16132 (vec_vclzw): Likewise.
16133 (vec_vgbbd): Likewise.
16134 (vec_vmrgew): Likewise.
16135 (vec_vmrgow): Likewise.
16136 (vec_vpopcnt): Likewise.
16137 (vec_vpopcntb): Likewise.
16138 (vec_vpopcntd): Likewise.
16139 (vec_vpopcnth): Likewise.
16140 (vec_vpopcntw): Likewise.
16142 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
16144 PR rtl-optimization/57468
16145 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
16148 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
16150 PR rtl-optimization/57459
16151 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
16152 type when setting live regs.
16154 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
16156 * config/s390/s390.opt (mlra): New option.
16157 * config/s390/s390.c (s390_decompose_address): Check displacement
16158 for all registers for LRA.
16159 (s390_secondary_reload): Don't used secondary reloads for LRA.
16160 (s390_lra_p): New function.
16161 (TARGET_LRA_P): Define.
16162 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
16163 of attribute cpu_facility to zarch for the last alternative.
16164 (*cmpmem_short): Ditto.
16166 2013-06-06 Eric Botcazou <ebotcazou@adacore.com>
16168 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
16169 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
16170 (arm_expand_prologue): Likewise.
16172 2013-06-06 Teresa Johnson <tejohnson@google.com>
16175 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
16176 as this is now done by redirect_edge_and_branch_force.
16177 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
16178 barriers, and fix interaction with splitting.
16179 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
16180 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
16181 reflect changes made in the routine.
16182 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
16183 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
16184 since this is called in cfglayout mode, and replace partition fixup
16185 with assert as that is now done by force_nonfallthru_and_redirect.
16186 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
16187 already be marked with region crossing note.
16188 (insert_section_boundary_note): Make non-static, gate on flag
16189 has_bb_partition, rewrite to also check for multiple partitions.
16190 (rest_of_handle_reorder_blocks): Remove call to
16191 insert_section_boundary_note, now done later during free_cfg.
16192 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
16193 * bb-reorder.h (insert_section_boundary_note): Declare.
16194 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
16195 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
16196 invoke insert_section_boundary_note.
16197 (try_redirect_by_replacing_jump): Remove unnecessary
16198 check for region crossing note.
16199 (fixup_partition_crossing): New function.
16200 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
16201 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
16202 in non-cfglayout mode.
16203 (force_nonfallthru_and_redirect): Fixup partition boundaries,
16204 remove old code that tried to do this. Emit barrier correctly
16205 when we are in cfglayout mode.
16206 (last_bb_in_partition): New function.
16207 (rtl_split_edge): Correctly fixup partition boundaries.
16208 (commit_one_edge_insertion): Remove old code that tried to
16209 fixup region crossing edge since this is now handled in
16210 split_block, and set up insertion point correctly since
16211 block may now end in a jump.
16212 (verify_hot_cold_block_grouping): Guard against checking when not in
16213 linearized RTL mode.
16214 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
16216 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
16217 rtl_verify_flow_info, so not called in cfglayout mode.
16218 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
16219 (fixup_reorder_chain): Remove old code that attempted to fixup region
16220 crossing note as this is now handled in force_nonfallthru_and_redirect.
16221 (duplicate_insn_chain): Don't duplicate switch section notes.
16222 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
16224 * basic-block.h (emit_barrier_after_bb): Declare.
16226 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16228 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
16229 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
16230 arm_usatsihi): Adjust alternatives for arm_restrict_it.
16232 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16234 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
16236 * config/arm/ldmstm.md: Regenerate.
16238 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16240 * config/arm/sync.md (atomic_loaddi_1):
16241 Disable predication for arm_restrict_it.
16242 (arm_load_exclusive<mode>): Likewise.
16243 (arm_load_exclusivesi): Likewise.
16244 (arm_load_exclusivedi): Likewise.
16245 (arm_load_acquire_exclusive<mode>): Likewise.
16246 (arm_load_acquire_exclusivesi): Likewise.
16247 (arm_load_acquire_exclusivedi): Likewise.
16248 (arm_store_exclusive<mode>): Likewise.
16249 (arm_store_exclusive<mode>): Likewise.
16250 (arm_store_release_exclusivedi): Likewise.
16251 (arm_store_release_exclusive<mode>): Likewise.
16253 2013-06-06 Richard Biener <rguenther@suse.de>
16255 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
16257 (lto_tag_is_tree_code_p): Adjust.
16258 (lto_tag_is_gimple_code_p): Likewise.
16259 (lto_gimple_code_to_tag): Likewise.
16260 (lto_tag_to_gimple_code): Likewise.
16261 (lto_tree_code_to_tag): Likewise.
16262 (lto_tag_to_tree_code): Likewise.
16263 * data-streamer.h (streamer_write_hwi_in_range): Use
16264 uhwi streaming to stream the normalized range.
16265 (streamer_read_hwi_in_range): Likewise.
16267 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16269 * config/arm/arm.md (enabled_for_depr_it): New attribute.
16270 (predicable_short_it): Likewise.
16271 (predicated): Likewise.
16272 (enabled): Handle above.
16273 (define_cond_exec): Set predicated attribute to yes.
16275 2013-06-05 Mike Stump <mikestump@comcast.net>
16277 * gdbinit.in (__FUNCTION__): Add.
16279 2013-06-05 Uros Bizjak <ubizjak@gmail.com>
16281 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
16282 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
16284 2013-06-05 Jan Hubicka <jh@suse.cz>
16286 * varasm.c (mark_decl_referenced): Revert the removal until targets
16289 2013-06-05 David Edelsohn <dje.gcc@gmail.com>
16291 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
16292 instead of mark_decl_referenced.
16294 2013-06-05 Jan Hubicka <jh@suse.cz>
16296 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
16297 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
16298 and symtab_used_from_object_file_p.
16299 (cgraph_make_node_local_1): Clear forced_by_abi.
16300 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
16301 * cgraph.h (symtab_node_base): Add forced_by_abi.
16302 (decide_is_variable_needed): Remove.
16303 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
16304 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
16305 (decide_is_symbol_needed): ... this one; handle symbols in general;
16306 always analyze virtuals; honnor forced_by_abi.
16307 (cgraph_finalize_function): Update.
16308 (varpool_finalize_decl): Update.
16309 (symbol_defined_and_needed): Remove.
16310 (analyze_functions): Update.
16311 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
16312 output_refs, input_overwrite_node): Handle forced_by_abi.
16313 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
16314 (address_taken_from_non_vtable_p): ... this one.
16315 (comdat_can_be_unshared_p_1): New function.
16316 (cgraph_comdat_can_be_unshared_p): Rename to ...
16317 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
16318 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
16319 (function_and_variable_visibility): Clear forced_by_abi as needed.
16320 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
16321 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
16322 * symtab.c (dump_symtab_base): Dump forced_by_abi.
16323 * varpool.c (decide_is_variable_needed): Remove.
16325 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16327 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
16328 (arm_option_override): Override arm_restrict_it where appropriate.
16329 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
16330 * config/arm/arm.opt (mrestrict-it): New command-line option.
16331 * doc/invoke.texi: Document -mrestrict-it.
16333 2013-06-05 David Malcolm <dmalcolm@redhat.com>
16335 * tsan.c (tsan_atomic_table): Make const.
16337 2013-06-05 Richard Biener <rguenther@suse.de>
16339 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
16340 index associated with the tree we are supposed to replace.
16341 * tree-streamer-out.c (pack_ts_base_value_fields): Output
16342 TREE_ASM_WRITTEN as zero for everything but SSA names.
16344 2013-06-05 David Malcolm <dmalcolm@redhat.com>
16346 * tree-ssa-structalias.c (call_stmt_vars): Make static.
16348 2013-06-04 Jan Hubicka <jh@suse.cz>
16350 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
16351 (input_node, input_varpool_node): Handle correctly external same
16353 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
16354 nodes at ltrans stage.
16356 2013-06-04 Jan Hubicka <jh@suse.cz>
16358 * ipa-inline.c (update_caller_keys): Fix availability test.
16359 (update_callee_keys): Likewise.
16360 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
16361 to follow ELF standard.
16363 2013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
16365 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
16366 (mips64r5900el-*-elf*): New configurations.
16367 * config/mips/mips-cpus.def (r5900): New processor.
16368 * config/mips/mips-tables.opt: Regenerate.
16369 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
16370 (mips_issue_rate): Handle PROCESSOR_R5900.
16371 (mips_reorg_process_insns): Force reorder mode for the R5900.
16372 * config/mips/mips.h (TARGET_MIPS5900): Define.
16373 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
16375 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
16377 * config/mips/mips.md (processor): Add r5900.
16378 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
16380 2013-06-04 Ian Bolton <ian.bolton@arm.com>
16382 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
16383 into function to generate MOVI instruction.
16384 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
16385 (aarch64_preferred_simd_mode): Turn into wrapper.
16386 (aarch64_output_scalar_simd_mov_immediate): New function.
16387 * config/aarch64/aarch64-protos.h: Add prototype for above.
16389 2013-06-04 Ian Bolton <ian.bolton@arm.com>
16391 * config/aarch64/aarch64.c (simd_immediate_info): Remove
16392 element_char member.
16393 (sizetochar): Return signed char.
16394 (aarch64_simd_valid_immediate): Remove elchar and other
16395 unnecessary variables.
16396 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
16397 Calculate element_char as required.
16398 * config/aarch64/aarch64-protos.h: Update and move prototype
16399 for aarch64_output_simd_mov_immediate.
16400 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
16403 2013-06-04 Ian Bolton <ian.bolton@arm.com>
16405 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
16406 information completed by aarch64_simd_valid_immediate.
16407 (aarch64_legitimate_constant_p): Update arguments.
16408 (aarch64_simd_valid_immediate): Work with struct rather than many
16410 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
16411 (aarch64_simd_make_constant): Update arguments.
16412 (aarch64_output_simd_mov_immediate): Work with struct rather than
16413 many pointers. Output immediate directly rather than as operand.
16414 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
16416 * config/aarch64/constraints.md (Dn): Update arguments.
16418 2013-06-04 Ian Bolton <ian.bolton@arm.com>
16420 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
16422 (aarch64_simd_immediate_valid_for_move): Remove.
16423 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
16424 (aarch64_simd_make_constant): Update call.
16425 (aarch64_output_simd_mov_immediate): Update call.
16426 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
16428 * config/aarch64/constraints.md (Dn): Update call.
16430 2013-06-04 Ian Bolton <ian.bolton@arm.com>
16432 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
16433 return type to bool for prototype.
16434 (aarch64_legitimate_constant_p): Check for true instead of not -1.
16435 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
16436 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
16438 2013-06-04 Catherine Moore <clm@codesourcery.com>
16440 * config/mips/mips.opt (meva): New.
16441 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
16442 (ASM_SPEC): Handle -meva.
16443 * doc/invoke.texi (meva): Document.
16445 2013-06-04 Alan Modra <amodra@gmail.com>
16447 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
16450 2013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16452 * rtl.def: Add extra fourth optional field to define_cond_exec.
16453 * gensupport.c (process_one_cond_exec): Process attributes from
16455 * doc/md.texi: Document fourth field in define_cond_exec.
16457 2013-06-04 Eric Botcazou <ebotcazou@adacore.com>
16459 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
16460 out the processing order as in store_bit_field_1.
16462 2013-06-04 Jan Hubicka <jh@suse.cz>
16464 PR middle-end/57500
16465 * cgraphunit.c (cgraph_process_same_body_aliases): Create
16466 non-VAR_DECL node if it does not exist yet.
16468 2013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
16470 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
16471 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
16472 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
16473 target_cpu_default setting.
16475 2013-06-03 Teresa Johnson <tejohnson@google.com>
16477 * dumpfile.c (opt_info_switch_p): Change -fopt-info
16478 default to -fopt-info=optimized instead of all.
16479 * doc/invoke.texi: Ditto.
16480 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
16481 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
16482 (execute_vect_slp): Emit BB vectorization success under
16483 MSG_OPTIMIZED_LOCATIONS.
16484 * tree-vect-slp.c (vect_slp_transform_bb): Change
16485 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
16486 * tree-vect-loop.c (vect_transform_loop): Ditto.
16488 2013-06-03 Jason Merrill <jason@redhat.com>
16491 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
16492 Use TARGET_EXPR for C++.
16494 2013-06-03 Jakub Jelinek <jakub@redhat.com>
16496 PR rtl-optimization/57268
16497 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
16498 if DEBUG_INSN_P (insn).
16501 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
16503 PR rtl-optimization/57268
16504 * sched-deps.c (sched_analyze_2): Flush dependence lists if
16505 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
16507 2013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
16509 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
16510 (ix86_avoid_lea_for_addr): Likewise.
16511 (exact_dependency_1): Likewise.
16512 (ix86_adjust_cost): Likewise.
16513 (swap_top_of_ready_list): Fix formatting and !reload_completed check
16515 (do_reorder_for_imul): Fix typo, formatting and
16516 !reload_completed check removed.
16517 (ix86_sched_reorder): Fix typo and formatting.
16518 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
16521 2013-06-03 Sofiane Naci <sofiane.naci@arm.com>
16523 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
16525 2013-06-03 Eric Botcazou <ebotcazou@adacore.com>
16527 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
16528 <STRING_CST>: Likewise.
16529 <VECTOR_CST>: Likewise.
16531 2013-06-01 Janus Weil <janus@gcc.gnu.org>
16532 Mikael Morin <mikael@gcc.gnu.org>
16534 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
16535 * config.in: Regenerated.
16536 * configure: Regenerated.
16538 2013-06-01 Jan Hubicka <jh@suse.cz>
16540 PR middle-end/57366
16541 * cgraphunit.c (compile): When weakref is not supported,
16542 set up transparent aliases before final output pass.
16543 * varasm.c (assemble_alias): Do not try to do it here.
16545 2013-06-01 Jan Hubicka <jh@suse.cz>
16547 PR middle-end/57467
16548 * passes.c (for_per_function): Skip unanalyzed functions.
16550 2013-06-01 Jan Hubicka <jh@suse.cz>
16552 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
16553 (lto_symtab_merge_symbols_1): ... this one.
16554 (lto_symtab_merge_cgraph_nodes): Rename to ...
16555 (lto_symtab_merge_symbols): ... this one; simplify.
16556 * cgraph.c (same_body_aliases_done): Rename to ...
16557 (cpp_implicit_aliases_done): ... this one.
16558 (cgraph_create_function_alias): Update.
16559 (cgraph_same_body_alias): Update.
16560 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
16561 (verify_edge_corresponds_to_fndecl): Simplify.
16562 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
16563 (cgraph_node): Remove same_body_alias.
16564 (varpool_node): Remove alias_of and extra_name_alias.
16565 (same_body_aliases_done): Rename to ..
16566 (cpp_implicit_aliases_done): ... this one.
16567 (symtab_alias_ultimate_target): Add default parameter.
16568 (symtab_resolve_alias): New function.
16569 (fixup_same_cpp_alias_visibility): Declare.
16570 (cgraph_function_node): Add default parameter.
16571 (cgraph_node_asm_name): Likewise.
16572 (cgraph_function_or_thunk_node): Add default parameter; do
16573 not ICE when it is NULL.
16574 (varpool_variable_node): Likewise.
16575 * tree-emutls.c (create_emultls_var): Update.
16576 (ipa_lower_emutls): Update.
16577 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
16578 (cgraph_reset_node): Reset alias info.
16579 (cgraph_finalize_function): Update.
16580 (fixup_same_cpp_alias_visibility): Move to symtab.c.
16581 (analyze_function): Simplify.
16582 (cgraph_process_same_body_aliases): Simplify.
16583 (analyze_functions): Fixup same body aliases.
16584 (handle_alias_pairs): Simplify.
16585 (assemble_thunk): Update.
16586 (assemble_thunks_and_aliases): Update.
16587 (output_weakrefs): Rewrite.
16588 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
16589 (lto_output_varpool_node): Likewise.
16590 (compute_ltrans_boundary): Remve assert.
16591 (get_alias_symbol): New functoin.
16592 (input_node): Rewrite alias handling.
16593 (input_varpool_node): Likewise.
16594 * ipa-pure-const.c (propagate_pure_const): Fix formating.
16595 * ipa.c (process_references): Handle weakrefs correctly.
16596 (symtab_remove_unreachable_nodes): Likewise.
16597 * trans-mem.c (get_cg_data): Update.
16598 (ipa_tm_create_version_alias): Update.
16599 (ipa_tm_execute): Update.
16600 * symtab.c (dump_symtab_base): Dump aliases.
16601 (verify_symtab_base): Verify aliases.
16602 (symtab_node_availability): New function.
16603 (symtab_alias_ultimate_target): Simplify.
16604 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
16605 handle all the fixup cases.
16606 (symtab_resolve_alias): New function.
16607 * passes.c (ipa_write_summaries): Handle weakrefs.
16608 * varpool.c (varpool_analyze_node): Simplify.
16609 (assemble_aliases): Update.
16610 (varpool_create_variable_alias): Simplify.
16611 (varpool_extra_name_alias): Simplify.
16612 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
16613 (lto_symtab_merge_symbols): ... this one.
16615 2013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
16618 PR rtl-optimization/57268
16619 * sched-deps.c (sched_analyze_2): Flush dependence lists if
16620 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
16622 2013-06-01 Tobias Burnus <burnus@net-b.de>
16624 Partially reverted:
16625 2013-05-31 Tobias Burnus <burnus@net-b.de>
16627 PR middle-end/57073
16628 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
16631 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
16633 PR rtl-optimization/57268
16634 * sched-deps.c (sched_analyze_2): Flush dependence lists if
16635 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
16637 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
16639 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
16640 unordered comparison operators when -fno-trapping-math is in effect
16642 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
16643 and implement unordered comparison operators properly on the e500.
16645 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
16647 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
16648 for constant scalar integers.
16649 (simplify_relational_operation_1): Likewise.
16651 2013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
16653 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
16654 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
16657 2013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
16658 Igor Zamyatin <igor.zamyatin@intel.com>
16660 Silvermont (SLM) architecture performance tuning.
16661 * config/i386/i386.h (enum ix86_tune_indices): Add
16662 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
16663 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
16665 * config/i386/i386.c (initial_ix86_tune_features)
16666 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
16667 (ix86_lea_outperforms): Handle Silvermont tuning.
16668 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
16670 (ix86_use_lea_for_mov): Likewise.
16671 (ix86_avoid_lea_for_addr): Likewise.
16672 (ix86_lea_for_add_ok): Likewise.
16673 (exact_dependency_1): New function.
16674 (exact_store_load_dependency): Likewise.
16675 (ix86_adjust_cost): Handle Silvermont tuning.
16676 (do_reoder_for_imul): Likewise.
16677 (swap_top_of_ready_list): New function.
16678 (ix86_sched_reorder): Changed to handle Silvermont tuning.
16680 * config/i386/i386.md (peepholes that split memory operand in fp
16683 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
16685 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
16686 Remove un-necessary braces.
16688 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
16690 * config/aarch64/aarch64.c (aarch64_classify_symbol):
16691 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
16693 2013-05-31 Tobias Burnus <burnus@net-b.de>
16695 PR middle-end/57073
16696 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
16698 2013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16701 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
16702 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
16703 * config/arm/neon.md (iordi3_neon): Remove.
16704 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
16705 * config/arm/predicates.md (imm_for_neon_logic_operand):
16706 Move to earlier in the file.
16707 (neon_logic_op2): Likewise.
16708 (arm_iordi_operand_neon): New predicate.
16710 2013-05-31 Richard Biener <rguenther@suse.de>
16712 PR tree-optimization/57478
16713 PR tree-optimization/57453
16714 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
16717 2013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
16719 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
16720 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
16722 2013-05-30 Tobias Burnus <burnus@net-b.de>
16723 Thomas Koenig <tkoenig@gcc.gnu.org>
16725 PR middle-end/57073
16726 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
16727 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
16729 2013-05-30 Steven Bosscher <steven@gcc.gnu.org>
16731 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
16733 2013-05-30 Vladimir Makarov <vmakarov@redhat.com>
16735 * target.def (register_usage_leveling_p): New hook.
16736 * targhooks.c (default_register_usage_leveling_p): New.
16737 * targhooks.h (default_register_usage_leveling_p): New prototype.
16738 * lra-assigns.c (register_usage_leveling_p): Use the hook.
16739 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
16740 * doc/tm.texi: Update.
16741 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
16743 2013-05-30 Ian Bolton <ian.bolton@arm.com>
16745 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
16746 (*insv_reg<mode>): New define_insn.
16748 2013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
16750 PR rtl-optimization/57439
16751 * postreload.c (move2add_valid_value_p): Check that we have
16752 a zero subreg_regno_offset when accessing the register in
16753 the requested mode.
16755 2013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
16756 Igor Zamyatin <igor.zamyatin@intel.com>
16758 Silvermont (SLM) architecture pipeline model, tuning and
16760 * config.gcc: Add slm config options and target.
16762 * config/i386/slm.md: New.
16764 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
16766 * config/i386/i386-c.c (ix86_target_macros_internal): New case
16768 (ix86_target_macros_internal): Likewise.
16770 * config/i386/i386.c (slm_cost): New cost.
16771 (m_SLM): New macro flag.
16772 (initial_ix86_tune_features): Set m_SLM.
16773 (x86_accumulate_outgoing_args): Likewise.
16774 (x86_arch_always_fancy_math_387): Likewise.
16775 (processor_target_table): Add slm cost.
16776 (cpu_names): Add slm cpu name.
16777 (x86_option_override_internal): Set SLM ISA.
16778 (ix86_issue_rate): New case PROCESSOR_SLM.
16779 (ia32_multipass_dfa_lookahead): Likewise.
16780 (fold_builtin_cpu): Add slm.
16782 * config/i386/i386.h (TARGET_SLM): New target macro.
16783 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
16784 (processor_type): Add PROCESSOR_SLM.
16786 * config/i386/i386.md (cpu): Add new value "slm".
16787 (slm.md): Include slm.md.
16789 2013-05-30 Bernd Schmidt <bernds@codesourcery.com>
16790 Zhenqiang Chen <zhenqiang.chen@linaro.org>
16792 * config/arm/arm-protos.h: Add and update function protos.
16793 * config/arm/arm.c (use_simple_return_p): New added.
16794 (thumb2_expand_return): Check simple_return flag.
16795 * config/arm/arm.md: Add simple_return and conditional simple_return.
16796 * config/arm/iterators.md: Add iterator for return and simple_return.
16798 2013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
16800 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
16801 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
16802 (arm_emit_vfp_multi_reg_pop): Likewise.
16803 (thumb2_emit_ldrd_pop): Likewise.
16804 (arm_expand_epilogue): Add misc REG_CFA notes.
16805 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
16807 2013-05-29 Lawrence Crowl <crowl@google.com>
16809 * config/arm/t-arm: Update for below.
16811 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
16812 Change type to hash_table. Update dependent calls and types.
16814 * config/i386/t-cygming: Update for below.
16816 * config/i386/t-interix: Update for below.
16818 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
16819 Change type to hash_table. Update dependent calls and types.
16820 (i386_find_on_wrapper_list::wrappers): Likewise.
16822 * config/ia64/t-ia64: Update for below.
16824 * config/ia64/ia64.c (bundle_state_table):
16825 Change type to hash_table. Update dependent calls and types.
16827 * config/mips/mips.c (mips_reorg_process_insns::htab):
16828 Change type to hash_table. Update dependent calls and types.
16830 * config/sol2.c (solaris_comdat_htab):
16831 Change type to hash_table. Update dependent calls and types.
16833 * config/t-sol2: Update for above.
16835 2013-05-29 Teresa Johnson <tejohnson@google.com>
16837 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
16838 functions are not yet marked as defined.
16840 2013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
16841 Pat Haugen <pthaugen@us.ibm.com>
16842 Peter Bergner <bergner@vnet.ibm.com>
16844 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
16848 (vrotl<mode>3): Likewise.
16849 (vashl<mode>3): Likewise.
16850 (vlshr<mode>3): Likewise.
16851 (vashr<mode>3): Likewise.
16853 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
16854 support for power8 V2DI builtins.
16856 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
16857 power8 V2DI builtins.
16858 (vupkhsw): Likewise.
16859 (vupklsw): Likewise.
16860 (vaddudm): Likewise.
16861 (vminsd): Likewise.
16862 (vmaxsd): Likewise.
16863 (vminud): Likewise.
16864 (vmaxud): Likewise.
16865 (vpkudum): Likewise.
16866 (vpksdss): Likewise.
16867 (vpkudus): Likewise.
16868 (vpksdus): Likewise.
16873 (vsubudm): Likewise.
16874 (vcmpequd): Likewise.
16875 (vcmpgtsd): Likewise.
16876 (vcmpgtud): Likewise.
16877 (vcmpequd_p): Likewise.
16878 (vcmpgtsd_p): Likewise.
16879 (vcmpgtud_p): Likewise.
16880 (vupkhsw): Likewise.
16881 (vupklsw): Likewise.
16882 (vaddudm): Likewise.
16883 (vmaxsd): Likewise.
16884 (vmaxud): Likewise.
16885 (vminsd): Likewise.
16886 (vminud): Likewise.
16887 (vpksdss): Likewise.
16888 (vpksdus): Likewise.
16889 (vpkudum): Likewise.
16890 (vpkudus): Likewise.
16895 (vsubudm): Likewise.
16897 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
16898 support for power8 V2DI instructions.
16900 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
16901 power8 V2DI instructions. Combine pack and unpack insns to use an
16902 iterator for each mode. Check whether a particular mode supports
16903 Altivec instructions instead of just checking TARGET_ALTIVEC.
16904 (UNSPEC_VPKUWUM): Likewise.
16905 (UNSPEC_VPKSHSS): Likewise.
16906 (UNSPEC_VPKSWSS): Likewise.
16907 (UNSPEC_VPKUHUS): Likewise.
16908 (UNSPEC_VPKSHUS): Likewise.
16909 (UNSPEC_VPKUWUS): Likewise.
16910 (UNSPEC_VPKSWUS): Likewise.
16911 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
16912 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
16913 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
16914 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
16915 (UNSPEC_VUPKHSB): Likewise.
16916 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
16917 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
16918 (UNSPEC_VUPKHSH): Likewise.
16919 (UNSPEC_VUPKLSB): Likewise.
16920 (UNSPEC_VUPKLSH): Likewise.
16922 (VI_char): Likewise.
16923 (VI_scalar): Likewise.
16924 (VI_unit): Likewise.
16926 (VP_small): Likewise.
16927 (VP_small_lc): Likewise.
16928 (VU_char): Likewise.
16929 (add<mode>3): Likewise.
16930 (altivec_vaddcuw): Likewise.
16931 (altivec_vaddu<VI_char>s): Likewise.
16932 (altivec_vadds<VI_char>s): Likewise.
16933 (sub<mode>3): Likewise.
16934 (altivec_vsubcuw): Likewise.
16935 (altivec_vsubu<VI_char>s): Likewise.
16936 (altivec_vsubs<VI_char>s): Likewise.
16937 (altivec_vavgs<VI_char>): Likewise.
16938 (altivec_vcmpbfp): Likewise.
16939 (altivec_eq<mode>): Likewise.
16940 (altivec_gt<mode>): Likewise.
16941 (altivec_gtu<mode>): Likewise.
16942 (umax<mode>3): Likewise.
16943 (smax<mode>3): Likewise.
16944 (umin<mode>3): Likewise.
16945 (smin<mode>3): Likewise.
16946 (altivec_vpkuhum): Likewise.
16947 (altivec_vpkuwum): Likewise.
16948 (altivec_vpkshss): Likewise.
16949 (altivec_vpkswss): Likewise.
16950 (altivec_vpkuhus): Likewise.
16951 (altivec_vpkshus): Likewise.
16952 (altivec_vpkuwus): Likewise.
16953 (altivec_vpkswus): Likewise.
16954 (altivec_vpks<VI_char>ss): Likewise.
16955 (altivec_vpks<VI_char>us): Likewise.
16956 (altivec_vpku<VI_char>us): Likewise.
16957 (altivec_vpku<VI_char>um): Likewise.
16958 (altivec_vrl<VI_char>): Likewise.
16959 (altivec_vsl<VI_char>): Likewise.
16960 (altivec_vsr<VI_char>): Likewise.
16961 (altivec_vsra<VI_char>): Likewise.
16962 (altivec_vsldoi_<mode>): Likewise.
16963 (altivec_vupkhsb): Likewise.
16964 (altivec_vupkhs<VU_char>): Likewise.
16965 (altivec_vupkls<VU_char>): Likewise.
16966 (altivec_vupkhsh): Likewise.
16967 (altivec_vupklsb): Likewise.
16968 (altivec_vupklsh): Likewise.
16969 (altivec_vcmpequ<VI_char>_p): Likewise.
16970 (altivec_vcmpgts<VI_char>_p): Likewise.
16971 (altivec_vcmpgtu<VI_char>_p): Likewise.
16972 (abs<mode>2): Likewise.
16973 (vec_unpacks_hi_v16qi): Likewise.
16974 (vec_unpacks_hi_v8hi): Likewise.
16975 (vec_unpacks_lo_v16qi): Likewise.
16976 (vec_unpacks_hi_<VP_small_lc>): Likewise.
16977 (vec_unpacks_lo_v8hi): Likewise.
16978 (vec_unpacks_lo_<VP_small_lc>): Likewise.
16979 (vec_pack_trunc_v8h): Likewise.
16980 (vec_pack_trunc_v4si): Likewise.
16981 (vec_pack_trunc_<mode>): Likewise.
16983 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
16985 (vec_vmaxsd): Likewise.
16986 (vec_vmaxud): Likewise.
16987 (vec_vminsd): Likewise.
16988 (vec_vminud): Likewise.
16989 (vec_vpksdss): Likewise.
16990 (vec_vpksdus): Likewise.
16991 (vec_vpkudum): Likewise.
16992 (vec_vpkudus): Likewise.
16993 (vec_vrld): Likewise.
16994 (vec_vsld): Likewise.
16995 (vec_vsrad): Likewise.
16996 (vec_vsrd): Likewise.
16997 (vec_vsubudm): Likewise.
16998 (vec_vupkhsw): Likewise.
16999 (vec_vupklsw): Likewise.
17001 2013-05-29 Jan Hubicka <jh@suse.cz>
17003 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
17004 flags; reorder rest of fields in more consistent way.
17005 (varpool_node): Remove analyzed, finalized and alias.
17006 (cgraph_ndoe): Likewise.
17007 (symtab_alias_ultimate_target): New function.
17008 (cgraph_function_node): Move offline.
17009 (cgraph_reset_node): Declare.
17010 (cgraph_comdat_can_be_unshared_p): Remove.
17011 (varpool_remove_initializer): Declare.
17012 (varpool_first_defined_variable, varpool_next_defined_variable
17013 cgraph_first_defined_function, cgraph_next_defined_function): Update.
17014 (cgraph_function_with_gimple_body_p): Update.
17015 (varpool_all_refs_explicit_p): Update.
17016 (symtab_alias_target): New function.
17017 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
17018 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
17019 (cgraph_function_or_thunk_node): Simplify using
17020 symtab_alias_ultimate_target.
17021 (varpool_variable_node): Likewise.
17022 * cgraph.c (cgraph_create_function_alias): Update.
17023 (cgraph_add_thunk): Update.
17024 (cgraph_remove_node): Update.
17025 (dump_cgraph_node): Do not dump removed flags.
17026 (cgraph_function_body_availability): Update.
17027 (cgraph_propagate_frequency): Update.
17028 (verify_cgraph_node): Check sanity of local flag.
17029 (cgraph_function_node): Move here from cgraph.h; revamp for
17030 cgraph_function_or_thunk_node.
17031 * lto-symtab.c (lto_varpool_replace_node): Update.
17032 (lto_symtab_resolve_can_prevail_p): Update.
17033 (lto_symtab_merge_cgraph_nodes): Update.
17034 * ipa-cp.c (determine_versionability, initialize_node_lattices,
17035 propagate_constants_accross_call, devirtualization_time_bonus,
17036 ipcp_propagate_stage): Update.
17037 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
17038 * ipa-inline-transform.c (clone_inlined_nodes,
17039 preserve_function_body_p): Update.
17040 * ipa-reference.c (propagate): Update.
17041 (write_node_summary_p): Update.
17042 * toplev.c (wrapup_global_declaration_2): Update.
17043 * cgraphunit.c (cgraph_analyze_function): Rename to ...
17044 (analyze_function) ... this one.
17045 (cgraph_process_new_functions): Update.
17046 (cgraph_reset_node): Export.
17047 (cgraph_finalize_function): Update.
17048 (cgraph_add_new_function): Update.
17049 (process_function_and_variable_attributes): Update.
17050 (varpool_finalize_decl): Update.
17051 (symbol_finalized): Remove.
17052 (symbol_finalized_and_needed): Rename to ...
17053 (symbol_defined_and_needed): ... update.
17054 (cgraph_analyze_functions): Update.
17055 (handle_alias_pairs): Update.
17056 (mark_functions_to_output): Update.
17057 (assemble_thunk): Update.
17058 (output_in_order): Update.
17059 (output_weakrefs): Update.
17060 (finalize_compilation_unit): Update.
17061 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
17062 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
17063 input_node, input_varpool_node): Update.
17064 * dbxout.c (dbxout_expand_expr): Update.
17065 * cgraphclones.c (cgraph_clone_node): Update.
17066 (cgraph_copy_node_for_versioning): Update.
17067 (cgraph_materialize_clone): Update.
17068 (cgraph_materialize_all_clones): Update.
17069 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
17070 propagate_pure_const, propagate_nothrow): Update.
17071 * lto-streamer-out.c (lto_output, write_symbol): Update.
17072 * ipa-utils.c (ipa_reverse_postorder): Update.
17073 * ipa-inline.c (can_inline_edge_p): Update.
17074 (update_caller_keys, ipa_inline): Update.
17075 * dwarf2out.c (reference_to_unused,
17076 premark_types_used_by_global_vars_helper): Update.
17077 * tree-eh.c (tree_could_trap_p): Update.
17078 * ipa-split.c (consider_split, execute_split_functions): Update.
17079 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
17080 has_addr_references_p): Update; move ahead in file for better
17082 (process_references): Simplify.
17083 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
17084 bodies are removed.
17085 (cgraph_comdat_can_be_unshared_p): Make static.
17086 (cgraph_externally_visible_p): Update.
17087 (varpool_externally_visible_p): Update.
17088 (function_and_variable_visibility): Update.
17089 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
17090 ipa_tm_mark_force_output_node): Update.
17091 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
17092 estimate_edge_devirt_benefit, inline_generate_summary,
17093 inline_write_summary): Update.
17094 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
17095 * ipa-prop.c (ipa_compute_jump_functions): Update.
17096 (ipa_print_node_params, ipa_prop_read_section,
17097 ipa_update_after_lto_read, read_replacements_section): Update.
17098 * varasm.c (mark_decl_referenced): Update.
17099 (assemble_alias, dump_tm_clone_pairs): Update.
17100 * tree-inline.c (copy_bb): Update.
17101 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
17103 * symtab.c (dump_symtab_base): Print new flags.
17104 (verify_symtab_base): Verify new flags.
17105 (symtab_alias_ultimate_target): New function.
17106 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
17107 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
17109 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
17111 * i386.c (ix86_get_function_versions_dispatcher,
17112 ix86_generate_version_dispatcher_body): Update.
17113 (fold_builtin_cpu): Use varpool_add_new_variable.
17114 * varpool.c (varpool_remove_initializer): Break out from ...
17115 (varpool_remove_node): ... this one.
17116 (dump_varpool_node, varpool_node_for_asm,
17117 cgraph_variable_initializer_availability, varpool_analyze_node,
17118 varpool_assemble_decl, varpool_remove_unreferenced_decls,
17119 varpool_finalize_named_section_flags, varpool_create_variable_alias):
17122 2013-05-29 Jan Hubicka <jh@suse.cz>
17124 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
17126 2013-05-29 Easwaran Raman <eraman@google.com>
17128 PR tree-optimization/57442
17129 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
17130 when control exits the main loop.
17132 2013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
17134 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
17136 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
17137 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
17138 * rx/t-rx: Add rx100 under multi library matches option for nofpu
17141 2013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17143 PR tree-optimization/57441
17144 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
17145 Don't limit size of incr_vec to number of candidates.
17147 2013-05-29 Steve Ellcey <sellcey@imgtec.com>
17149 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
17150 and mips16 directories.
17151 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
17152 (MULTILIB_DIRNAMES): Ditto.
17153 (MULTILIB_EXCEPTIONS): Add new exceptions.
17154 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
17155 (MULTILIB_DIRNAMES): Ditto.
17156 (MULTILIB_EXCEPTIONS): Add new exceptions.
17158 2012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
17159 Marcus Shawcroft <marcus.shawcroft@arm.com>
17161 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
17162 SYMBOL_TINY_ABSOLUTE.
17163 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
17164 SYMBOL_TINY_ABSOLUTE.
17165 (aarch64_expand_mov_immediate): Likewise.
17166 (aarch64_classify_symbol): Likewise.
17167 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
17168 Permit SYMBOL_TINY_ABSOLUTE.
17169 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
17171 2013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
17172 Marcus Shawcroft <marcus.shawcroft@arm.com>
17174 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
17175 Refactor if/switch. Replace gcc_assert with if.
17177 2013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
17179 * config/i386/i386.c (initial_ix86_tune_features): Enable
17180 FP Reassociation for AMD bdver1 and bdver2.
17182 2013-05-29 Martin Jambor <mjambor@suse.cz>
17184 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
17185 and IMAGPART_EXPR do not occur within other handled_components.
17187 2013-05-29 Richard Biener <rguenther@suse.de>
17189 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
17190 access on whether the use is in the BB we currently try to
17192 (vect_bb_vectorization_profitable_p): Pass the BB we currently
17193 vectorize to vect_bb_slp_scalar_cost.
17195 2013-05-29 Richard Biener <rguenther@suse.de>
17197 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
17198 computing scalar cost offsetted by stmts that are kept live
17200 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
17201 for computation of scalar cost.
17203 2013-05-28 Steve Ellcey <sellcey@mips.com>
17205 * config/mips/mips-cpus.def (mips32r2): Change processor type.
17207 2013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
17209 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
17210 array notation built-in reduction functions.
17211 * doc/passes.texi (Passes): Added documentation about changes done
17213 * doc/invoke.texi (C Dialect Options): Added documentation about
17214 the -fcilkplus flag.
17215 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
17216 (BUILTINS_DEF): Depend on cilkplus.def.
17217 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
17218 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
17219 * cilkplus.def: New file.
17221 2013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
17223 PR rtl-optimization/57439
17224 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
17226 2013-05-28 Easwaran Raman <eraman@google.com>
17228 PR tree-optimization/57337
17229 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
17230 (find_insert_point): Correctly identify the insertion point
17231 when two statements with the same UID is compared.
17233 2013-05-28 Richard Biener <rguenther@suse.de>
17235 PR tree-optimization/56787
17236 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
17237 from the list of data references.
17238 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
17240 (vect_analyze_loop_operations): Likewise.
17241 (vect_transform_loop): Remove clobbers.
17243 2013-05-28 Martin Jambor <mjambor@suse.cz>
17245 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
17246 and REALPART_EXPRs have scalar type.
17248 2013-05-28 Richard Biener <rguenther@suse.de>
17250 PR tree-optimization/57411
17251 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
17253 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
17254 virtual operand propagation.
17256 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
17258 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
17259 destination register for bmasksi_vis.
17260 (vector_init_bshuffle): Likewise.
17261 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
17263 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
17265 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
17266 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
17267 mode if the instruction isn't available in the original mode.
17268 * config/sparc/sparc.opt (mfix-ut699): New option.
17269 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
17270 (divdf3): Turn into expander.
17271 (divdf3_nofix): New insn.
17272 (divdf3_fix): Likewise.
17273 (divsf3): Disable if -mfix-ut699.
17274 (sqrtdf2): Turn into expander.
17275 (sqrtdf2_nofix): New insn.
17276 (sqrtdf2_fix): Likewise.
17277 (sqrtsf2): Disable if -mfix-ut699.
17279 2013-05-27 Richard Biener <rguenther@suse.de>
17281 PR middle-end/57412
17282 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
17283 block for the new loop.
17285 2013-05-27 Richard Biener <rguenther@suse.de>
17287 PR tree-optimization/57343
17288 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
17289 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
17290 (number_of_iterations_cond): Do not build the folded tree.
17292 2013-05-27 Richard Biener <rguenther@suse.de>
17295 PR middle-end/57381
17296 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
17297 OEP_CONSTANT_ADDRESS_OF retained.
17299 PR tree-optimization/57417
17300 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
17301 for unchanged base.
17302 (set_ssa_val_to): Compare addresses using
17303 get_addr_base_and_unit_offset.
17305 2013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
17307 PR rtl-optimization/56833
17308 * postreload.c (move2add_record_mode): New function.
17309 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
17310 (move2add_use_add2_insn): Use move2add_record_sym_value.
17311 (move2add_use_add3_insn): Likewise.
17312 (reload_cse_move2add): Use move2add_valid_value_p and
17313 move2add_record_mode. Invalidate call-clobbered and REG_INC
17314 affected regs by setting reg_mode to VOIDmode.
17315 (move2add_note_store): Don't pretend the inside of a SUBREG is
17316 the actual destination. Invalidate single/leading registers by
17317 setting reg_mode to VOIDmode.
17318 Use move2add_record_sym_value, move2add_valid_value_p and
17319 move2add_record_mode.
17321 2013-05-27 Richard Biener <rguenther@suse.de>
17323 PR tree-optimization/57396
17324 * tree-affine.c (double_int_constant_multiple_p): Properly
17325 return false for val == 0 and div != 0.
17327 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
17329 * config/mips/mips.h: Use #elif in preprocessor conditions.
17331 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
17334 * config/mips/constraints.md (kl): New constraint.
17335 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
17336 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
17337 constraint for operand 0. Split after CSE for MIPS16. Emit a move
17338 from LO for MIPS16.
17339 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
17341 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
17344 * config/mips/mips.c (mips_can_inline_p): New function.
17345 (TARGET_CAN_INLINE_P): Define.
17347 2013-05-25 Steven Bosscher <steven@gcc.gnu.org>
17349 * sched-int.h (ds_t, dw_t): Make unsigned int.
17350 Fix documentation that describes how all the ds_t bits are used.
17351 Reserve the last bit for delayed-branch scheduling.
17352 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
17353 (BITS_PER_DEP_WEAK): Fix definition and documentation.
17354 (gen_dep_weak_1): Remove prototype.
17355 * sched-deps.c (get_dep_weak_1): Make static.
17356 * target.def (speculate_insn, needs_block_p, gen_spec_check,
17357 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
17358 * doc/tm.texi: Regenerate.
17359 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
17361 2013-05-24 Steven Bosscher <steven@gcc.gnu.org>
17364 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
17366 2013-05-24 Nathan Sidwell <nathan@codesourcery.com>
17367 Sandra Loosemore <sandra@codesourcery.com>
17369 * config.gcc (powerpc-*): Allow native for with-cpu.
17371 2013-05-24 Jeff Law <law@redhat.com>
17373 PR tree-optimization/57124
17374 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
17375 conversion feeding a condition if the range has an overflow
17376 if -fstrict-overflow. Add warnings for when we do make the
17379 2013-05-24 Dehao Chen <dehao@google.com>
17381 * tree-cfg.c (locus_discrim_map): Fix the typo.
17382 (locus_discrim_hasher): Likewise.
17383 (locus_discrim_hasher::hash): Likewise.
17384 (locus_discrim_hasher::equal): Likewise.
17386 2013-05-24 Martin Jambor <mjambor@suse.cz>
17388 PR tree-optimization/57294
17389 * cgraph.h (ipa_record_stmt_references): Declare.
17390 * cgraphbuild.c (ipa_record_stmt_references): New function.
17391 (build_cgraph_edges): Use ipa_record_stmt_references.
17392 (rebuild_cgraph_edges): Likewise.
17393 (cgraph_rebuild_references): Likewise.
17394 * ipa-prop.c (ipa_modify_call_arguments): Discard references
17395 associated with the old statement and build references from the
17396 newly built statements.
17397 * ipa-ref.c (ipa_remove_stmt_references): New function.
17398 * ipa-ref.h (ipa_remove_stmt_references): Declare.
17400 2013-05-24 Vladimir Makarov <vmakarov@redhat.com>
17402 * lra-constraints.c (emit_spill_move): Use smaller mode for
17404 (check_and_process_move): Consider mem-reg moves for secondary
17406 (curr_insn_transform): Don't lose insns emitted before for
17407 secondary memory moves.
17408 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
17409 reg set up in the current insn.
17411 2013-05-24 Dehao Chen <dehao@google.com>
17413 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
17415 (locus_descrim_hasher::equal): Likewise.
17416 (build_gimple_cfg): New discriminator assignment algorithm.
17417 (make_edges): Likewise.
17418 (next_discriminator_for_locus): Likewise.
17419 (same_line_p): Likewise.
17420 (assign_discriminators): Likewise.
17421 (make_cond_expr_edges): Likewise.
17422 (make_gimple_switch_edges): Likewise.
17423 (make_goto_expr_edges): Likewise.
17424 (make_gimple_asm_edges): Likewise.
17426 2013-05-24 Ian Bolton <ian.bolton@arm.com>
17428 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
17429 X format specifier to only display bottom 16 bits.
17430 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
17431 immediate to match for operand 2, since it will be masked.
17433 2013-05-24 Richard Biener <rguenther@suse.de>
17435 PR tree-optimization/57287
17436 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
17437 all SSA names that occur in abnormal PHIs.
17439 2013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
17442 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
17443 that index is not negative.
17445 2013-05-24 Eric Botcazou <ebotcazou@adacore.com>
17447 PR rtl-optimization/55177
17448 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
17449 (simplify_byte_swapping_operation): New.
17450 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
17451 (simplify_relational_operation_1): Deal with BSWAP.
17453 2013-05-23 Richard Henderson <rth@redhat.com>
17456 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
17457 (ix86_reorg): Call it.
17459 2013-05-23 Uros Bizjak <ubizjak@gmail.com>
17462 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
17463 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
17464 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
17466 2013-05-23 Christian Bruel <christian.bruel@st.com>
17469 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
17471 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
17472 Marcus Shawcroft <marcus.shawcroft@arm.com>
17474 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
17475 * config/aarch64/constraints.md (Usa): Remove.
17476 * doc/md.texi (AArch64 Usa): Remove.
17478 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
17479 Marcus Shawcroft <marcus.shawcroft@arm.com>
17481 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
17482 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
17483 * config/aarch64/predicates.md (aarch64_const_address): Remove.
17484 (aarch64_mov_operand): Use aarch64_mov_operand_p.
17486 2013-05-23 Vidya Praveen <vidyapraveen@arm.com>
17488 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
17489 instruction (AdvSIMD).
17490 * config/aarch64/aarch64-builtins.c
17491 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
17492 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
17494 2013-05-23 Martin Jambor <mjambor@suse.cz>
17496 PR middle-end/57347
17497 * tree.h (contains_bitfld_component_ref_p): Declare.
17498 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
17499 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
17501 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
17502 not access a bit-field. Assert all final offsets are byte-aligned.
17504 2013-05-23 Richard Biener <rguenther@suse.de>
17506 PR tree-optimization/57380
17507 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
17508 least one invariant or re-used load.
17509 * passes.c (init_optimization_passes): Move pass_phiprop before
17512 2013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
17514 * config/aarch64/aarch64-simd.md
17515 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
17517 2013-05-23 Richard Biener <rguenther@suse.de>
17519 PR middle-end/57381
17520 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
17521 OEP_CONSTANT_ADDRESS_OF retained.
17523 2013-05-23 Jakub Jelinek <jakub@redhat.com>
17525 PR middle-end/57344
17526 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
17527 don't lower unit. Handle unit not being always BITS_PER_WORD.
17529 2013-05-23 Richard Biener <rguenther@suse.de>
17531 PR rtl-optimization/57341
17532 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
17533 instead of true_dependence.
17535 2013-05-22 David Malcolm <dmalcolm@redhat.com>
17537 * bb-reorder.c (branch_threshold): Make const.
17538 (exec_threshold): Ditto.
17540 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
17541 Pat Haugen <pthaugen@us.ibm.com>
17542 Peter Bergner <bergner@vnet.ibm.com>
17544 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
17545 documentation for the power8 crypto builtins.
17547 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
17549 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
17550 macros for defining power8 builtin functions.
17551 (BU_P8V_AV_2): Likewise.
17552 (BU_P8V_AV_P): Likewise.
17553 (BU_P8V_VSX_1): Likewise.
17554 (BU_P8V_OVERLOAD_1): Likewise.
17555 (BU_P8V_OVERLOAD_2): Likewise.
17556 (BU_CRYPTO_1): Likewise.
17557 (BU_CRYPTO_2): Likewise.
17558 (BU_CRYPTO_3): Likewise.
17559 (BU_CRYPTO_OVERLOAD_1): Likewise.
17560 (BU_CRYPTO_OVERLOAD_2): Likewise.
17561 (XSCVSPDP): Fix typo, point to the correct instruction.
17562 (VCIPHER): Add power8 crypto builtins.
17563 (VCIPHERLAST): Likewise.
17564 (VNCIPHER): Likewise.
17565 (VNCIPHERLAST): Likewise.
17566 (VPMSUMB): Likewise.
17567 (VPMSUMH): Likewise.
17568 (VPMSUMW): Likewise.
17569 (VPERMXOR_V2DI): Likewise.
17570 (VPERMXOR_V4SI: Likewise.
17571 (VPERMXOR_V8HI: Likewise.
17572 (VPERMXOR_V16QI: Likewise.
17573 (VSHASIGMAW): Likewise.
17574 (VSHASIGMAD): Likewise.
17575 (VPMSUM): Likewise.
17576 (VPERMXOR): Likewise.
17577 (VSHASIGMA): Likewise.
17579 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
17580 __CRYPTO__ if the crypto instructions are available.
17581 (altivec_overloaded_builtins): Add support for overloaded power8
17584 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
17585 support for power8 crypto builtins.
17586 (builtin_function_type): Likewise.
17587 (altivec_init_builtins): Add support for builtins that take vector
17588 long long (V2DI) arguments.
17590 * config/rs6000/crypto.md: New file, define power8 crypto
17593 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
17594 Pat Haugen <pthaugen@us.ibm.com>
17595 Peter Bergner <bergner@vnet.ibm.com>
17597 * doc/invoke.texi (Option Summary): Add power8 options.
17598 (RS/6000 and PowerPC Options): Likewise.
17600 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
17601 constraints.md instead of rs6000.h. Reorder w* constraints. Add
17602 wm, wn, wr documentation.
17604 * config/rs6000/constraints.md (wm): New constraint for VSX
17605 registers if direct move instructions are enabled.
17606 (wn): New constraint for no registers.
17607 (wq): New constraint for quad word even GPR registers.
17608 (wr): New constraint if 64-bit instructions are enabled.
17609 (wv): New constraint if power8 vector instructions are enabled.
17610 (wQ): New constraint for quad word memory locations.
17612 * config/rs6000/predicates.md (const_0_to_15_operand): New
17613 constraint for 0..15 for crypto instructions.
17614 (gpc_reg_operand): If VSX allow registers in VSX registers as well
17615 as GPR and floating point registers.
17616 (int_reg_operand): New predicate to match only GPR registers.
17617 (base_reg_operand): New predicate to match base registers.
17618 (quad_int_reg_operand): New predicate to match even GPR registers
17619 for quad memory operations.
17620 (vsx_reg_or_cint_operand): New predicate to allow vector logical
17621 operations in both GPR and VSX registers.
17622 (quad_memory_operand): New predicate for quad memory operations.
17623 (reg_or_indexed_operand): New predicate for direct move support.
17625 * config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
17626 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
17627 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
17628 (POWERPC_MASKS): Add power8 options.
17629 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
17632 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
17633 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
17635 * config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
17636 (-mpower8-fusion): New power8 options.
17637 (-mpower8-fusion-sign): Likewise.
17638 (-mpower8-vector): Likewise.
17639 (-mcrypto): Likewise.
17640 (-mdirect-move): Likewise.
17641 (-mquad-memory): Likewise.
17643 * config/rs6000/rs6000.c (power8_cost): Initial definition for power8.
17644 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
17646 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
17647 (rs6000_debug_vector_unit): Add p8_vector.
17648 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
17649 definitions. Also print fusion state.
17650 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
17651 (rs6000_builtin_mask_calculate): Add power8 builtin support.
17652 (rs6000_option_override_internal): Add support for power8.
17653 (rs6000_common_init_builtins): Add debugging for skipped builtins
17654 if -mdebug=builtin.
17655 (rs6000_adjust_cost): Add power8 support.
17656 (rs6000_issue_rate): Likewise.
17657 (insn_must_be_first_in_group): Likewise.
17658 (insn_must_be_last_in_group): Likewise.
17659 (force_new_group): Likewise.
17660 (rs6000_register_move_cost): Likewise.
17661 (rs6000_opt_masks): Likewise.
17663 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
17664 power8 capable assembler, default to power7 options.
17665 (TARGET_DIRECT_MOVE): Likewise.
17666 (TARGET_CRYPTO): Likewise.
17667 (TARGET_P8_VECTOR): Likewise.
17668 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
17669 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
17670 (VECTOR_MEM_P8_VECTOR_P): Likewise.
17671 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
17672 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
17673 (TARGET_XSCVDPSPN): Likewise.
17674 (TARGET_XSCVSPDPN): Likewsie.
17675 (TARGET_SYNC_HI_QI): Likewise.
17676 (TARGET_SYNC_TI): Likewise.
17677 (MASK_CRYPTO): Likewise.
17678 (MASK_DIRECT_MOVE): Likewise.
17679 (MASK_P8_FUSION): Likewise.
17680 (MASK_P8_VECTOR): Likewise.
17681 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
17682 temporary used by some of the direct move instructions to get two FP
17683 temporary registers does not force creation of a stack frame.
17684 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
17685 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
17686 that any VSX registers are tieable, even if they are also an
17687 Altivec vector mode.
17688 (r6000_reg_class_enum): Add wm, wr, wv constraints.
17689 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
17690 (RS6000_BTM_CRYPTO): Likewise.
17691 (RS6000_BTM_COMMON): Likewise.
17693 * config/rs6000/rs6000.md (cpu attribute): Add power8.
17694 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
17695 (enum rs6000_vector): Add power8 vector support.
17697 2013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
17701 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
17702 (any_sibcall_could_use_r3): this and handle indirect calls.
17703 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
17705 2013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17707 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
17709 2013-05-22 Richard Biener <rguenther@suse.de>
17711 PR middle-end/57349
17712 * profile.c (branch_prob): Do not split blocks that are
17713 abnormally receiving from ECF_RETURNS_TWICE functions.
17715 2013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
17717 * recog.c (offsettable_address_addr_space_p): Fix calculation of
17718 address mode. Move pointer mode initialization to the same place.
17720 2013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
17722 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
17723 while it has any effect.
17725 2013-05-21 Easwaran Raman <eraman@google.com>
17727 PR tree-optimization/57322
17728 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
17729 UID of the statement added to the BB to be 1.
17731 2013-05-21 Jakub Jelinek <jakub@redhat.com>
17733 PR tree-optimization/57331
17734 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
17735 of conversion from pointer type to integral type with integer.
17737 2013-05-21 Martin Jambor <mjambor@suse.cz>
17740 * ipa-prop.c (ipa_read_node_info): Process param_used and
17741 controlled_uses in the same order as when writing.
17743 2013-05-21 Magnus Granberg <baldrick@free.fr>
17746 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
17748 2013-05-21 Richard Biener <rguenther@suse.de>
17750 PR tree-optimization/57318
17751 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
17752 estimate stmts with side-effects as likely eliminated.
17754 2013-05-21 Richard Biener <rguenther@suse.de>
17756 PR tree-optimization/57330
17757 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
17758 preserve the call stmts fntype.
17760 2013-05-21 Richard Biener <rguenther@suse.de>
17762 PR tree-optimization/57303
17763 * tree-ssa-sink.c (statement_sink_location): Improve killing
17764 stmt detection and properly handle self-assignments.
17766 2013-05-21 Christian Bruel <christian.bruel@st.com>
17768 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
17769 spanning registers. LEAF_REG_REMAP is supported only for contiguous
17770 registers. Set register size out of the PARALLEL loop.
17772 2013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
17775 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
17776 (*fmasf4, *fmasf4_media): New insns.
17778 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
17780 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
17781 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
17782 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
17783 (mips_idiv_insns): Update the comments to say that the returned
17784 instruction counts are in units of BASE_INSN_LENGTH.
17785 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
17786 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
17787 using 2 rather than 4 as the length of indirect MIPS16 and
17788 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
17789 length of a NOP. Don't divide MIPS16 lengths by 2.
17790 (mips16_split_long_branches): Assume a branch is long if the
17791 length is greater than 4 rather than 8.
17792 * config/mips/mips.md (length): Give MIPS16 lengths directly,
17793 rather than multiplying them by 2. Multiply instruction counts
17794 by BASE_INSN_LENGTH rather than 4.
17795 (*jump_mips16, tls_get_tp_mips16_<mode>)
17796 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
17798 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
17800 * config/mips/mips.md (extended_mips16): Remove branch case.
17801 (length): Remove duplicated extended_mips16 test.
17803 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
17805 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
17807 2013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
17809 * recog.h (Recog_data): Rename to...
17810 (recog_data_d): ...this.
17811 (recog_data): Update accordingly.
17812 * recog.c (recog_data): Likewise.
17813 * reload.c (save_recog_data): Likewise.
17814 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
17815 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
17817 2013-05-17 Julian Brown <julian@codesourcery.com>
17819 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
17820 found in a REG_EQUAL note, invalidate it.
17822 2013-05-17 Easwaran Raman <eraman@google.com>
17824 * tree-ssa-reassoc.c (find_insert_point): New function.
17825 (insert_stmt_after): Likewise.
17826 (get_def_stmt): Likewise.
17827 (ensure_ops_are_available): Likewise.
17828 (not_dominated_by): Likewise.
17829 (rewrite_expr_tree): Do not move statements beyond what is
17830 necessary. Remove call to swap_ops_for_binary_stmt...
17831 (reassociate_bb): ... and move it here.
17832 (build_and_add_sum): Assign UIDs for new statements.
17833 (linearize_expr): Likewise.
17834 (do_reassoc): Renumber gimple statement UIDs.
17836 2013-05-17 Jan Hubicka <jh@suse.cz>
17838 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
17840 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
17841 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
17843 (output_weakrefs): Update.
17845 2013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
17846 Martin Jambor <mjambor@suse.cz>
17848 PR middle-end/57276
17849 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
17850 value that corresponds to the given aggval is found in values vector.
17852 2013-05-17 Uros Bizjak <ubizjak@gmail.com>
17854 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
17855 sse, sse2, sse3, ssse3 and sse4a flags to options.
17857 2013-05-17 David Malcolm <dmalcolm@redhat.com>
17859 * gengtype-state.c: (s_expr_writer): New class, to handle
17860 prettifying of output layout of s-expressions.
17861 (state_writer): New class, to write out gtype.state.
17862 (state_written_type_count): Move this variable into member data of
17864 (s_expr_writer::s_expr_writer): New code: constructor for new class
17865 (state_writer::state_writer(): ditto
17866 (s_expr_writer::write_new_line): New function
17867 (s_expr_writer::write_any_indent): ditto
17868 (s_expr_writer::begin_s_expr): ditto
17869 (s_expr_writer::end_s_expr): ditto
17870 (write_state_fileloc): convert to method of state_writer...
17871 (state_writer:: write_state_fileloc): ...and use methods of
17872 s_expr_writer to write indentation into the gtype.state output file
17873 to visually represent the hierarchical structure of the list
17875 (write_state_fields): ditto, renaming to...
17876 (state_writer::write_state_fields)
17877 (write_state_a_string): ditto, renaming to...
17878 (state_writer::write_state_a_string)
17879 (write_state_string_option): ditto, renaming to...
17880 (state_writer::write_state_string_option)
17881 (write_state_type_option): ditto, renaming to...
17882 (state_writer::write_state_type_option)
17883 (write_state_nested_option): ditto, renaming to...
17884 (state_writer::write_state_nested_option)
17885 (write_state_option): ditto, renaming to...
17886 (state_writer::write_state_option)
17887 (write_state_options): ditto, renaming to...
17888 (state_writer::write_state_options)
17889 (write_state_lang_bitmap): ditto, renaming to...
17890 (state_writer::write_state_lang_bitmap)
17891 (write_state_version): ditto, renaming to...
17892 (state_writer::write_state_version)
17893 (write_state_scalar_type): ditto, renaming to...
17894 (state_writer::write_state_scalar_type)
17895 (write_state_string_type): ditto, renaming to...
17896 (state_writer::write_state_string_type)
17897 (write_state_undefined_type): ditto, renaming to...
17898 (state_writer::write_state_undefined_type)
17899 (write_state_struct_union_type): ditto, renaming to...
17900 (state_writer::write_state_struct_union_type)
17901 (write_state_struct_type): ditto, renaming to...
17902 (state_writer::write_state_struct_type)
17903 (write_state_user_struct_type): ditto, renaming to...
17904 (state_writer::write_state_user_struct_type)
17905 (write_state_lang_struct_type): ditto, renaming to...
17906 (state_writer::write_state_lang_struct_type)
17907 (write_state_param_struct_type): ditto, renaming to...
17908 (state_writer::write_state_param_struct_type)
17909 (write_state_pointer_type): ditto, renaming to...
17910 (state_writer::write_state_pointer_type)
17911 (write_state_array_type): ditto, renaming to...
17912 (state_writer::write_state_array_type)
17913 (write_state_gc_used): ditto, renaming to...
17914 (state_writer::write_state_gc_used)
17915 (write_state_common_type_content): ditto, renaming to...
17916 (state_writer::write_state_common_type_content)
17917 (write_state_type): ditto, renaming to...
17918 (state_writer::write_state_type)
17919 (write_state_pair_list): ditto, renaming to...
17920 (state_writer::write_state_pair_list)
17921 (write_state_pair): ditto, renaming to...
17922 (state_writer::write_state_pair)
17923 (write_state_typedefs): ditto, renaming to...
17924 (state_writer::write_state_typedefs)
17925 (write_state_structures): ditto, renaming to...
17926 (state_writer::write_state_structures)
17927 (write_state_param_structs): ditto, renaming to...
17928 (state_writer::write_state_param_structs)
17929 (write_state_variables): ditto, renaming to...
17930 (state_writer::write_state_variables)
17931 (write_state_srcdir): ditto, renaming to...
17932 (state_writer::write_state_srcdir)
17933 (write_state_files_list): ditto, renaming to...
17934 (state_writer::write_state_files_list)
17935 (write_state_languages): ditto, renaming to...
17936 (state_writer::write_state_languages)
17937 (write_state): create a state_writer instance and use it when
17938 writing out the state file
17940 2013-05-17 Mike Stump <mikestump@comcast.net>
17942 PR rtl-optimization/57304
17943 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
17944 accessing DF_REF_REAL_LOC.
17946 2013-05-17 Jakub Jelinek <jakub@redhat.com>
17948 PR rtl-optimization/57281
17949 PR rtl-optimization/57300
17950 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
17951 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
17952 what the other splitter did if the registers are dead.
17954 2013-05-17 Richard Biener <rguenther@suse.de>
17956 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
17959 2013-05-17 Jakub Jelinek <jakub@redhat.com>
17961 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
17964 2013-05-17 Marek Polacek <polacek@redhat.com>
17966 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
17967 length when doing non-zero store of storing '\0' to '\0'.
17969 2013-05-17 Jakub Jelinek <jakub@redhat.com>
17971 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
17972 vect_external_def oprnd1 with loop_vinfo, try to emit
17973 optional cast, negation and and stmts on the loop preheader
17974 edge instead of into the pattern def seq.
17976 PR tree-optimization/57051
17977 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
17978 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
17980 2013-05-16 Nick Clifton <nickc@redhat.com>
17982 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
17983 (rl78_is_naked_func): New function.
17984 (rl78_expand_prologue): Skip prologue generation for naked functions.
17985 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
17986 * doc/extend.texi (naked): Add RL78 to the list of processors
17987 that supports this attribute.
17989 2013-05-16 Jeff Law <law@redhat.com>
17991 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
17993 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
17995 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
17996 cache parameters using detect_caches_amd also for CYRIX,
17997 NSC and TM2 signatures.
17999 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
18000 Dzianis Kahanovich <mahatma@eu.by>
18004 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
18005 VIA/Centaur processors and determine their cache parameters
18006 using detect_caches_amd.
18008 2013-05-16 Teresa Johnson <tejohnson@google.com>
18010 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
18011 (rtl_verify_edges): New function.
18012 (rtl_verify_bb_insns): Ditto.
18013 (rtl_verify_bb_pointers): Ditto.
18014 (rtl_verify_bb_insn_chain): Ditto.
18015 (rtl_verify_fallthru): Ditto.
18016 (rtl_verify_bb_layout): Ditto.
18017 (rtl_verify_flow_info_1): Outline checks into new functions.
18018 (rtl_verify_flow_info): Ditto.
18020 2013-05-16 Steve Ellcey <sellcey@imgtec.com>
18022 * cfghooks.c (copy_bbs): Add update_dominance argument.
18023 * cfghooks.h (copy_bbs): Update prototype.
18024 * tree-cfg.c (gimple_duplicate_sese_region):
18025 Add update_dominance argument.
18026 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
18027 * tree-ssa-loop-ch.c (copy_loop_headers): Update
18028 gimple_duplicate_sese_region call.
18029 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
18030 Update copy_bbs call.
18031 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
18032 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
18034 2013-05-16 Jakub Jelinek <jakub@redhat.com>
18036 * tree-vectorizer.h (NUM_PATTERNS): Increment.
18037 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
18038 vect_recog_rotate_pattern.
18039 (vect_recog_rotate_pattern): New function.
18041 2013-05-16 Jason Merrill <jason@redhat.com>
18043 * Makefile.in (LLINKER): New variable.
18044 (mostlyclean): Remove link mutex.
18045 * configure.ac: Handle --enable-link-mutex.
18046 * lock-and-run.sh: New script.
18048 2013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
18051 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
18054 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
18056 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
18058 2013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
18060 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
18061 * config/arm/arm.c (next_consecutive_mem): New function.
18062 (gen_movmem_ldrd_strd): Likewise.
18063 * config/arm/arm.md (movmemqi): Update condition and code.
18064 (unaligned_loaddi, unaligned_storedi): New patterns.
18066 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
18068 * config.gcc: Obsolete *-*-solaris2.9*.
18069 * doc/install.texi (Specific, *-*-solaris2*): Document it.
18071 2013-05-16 Richard Biener <rguenther@suse.de>
18073 * passes.c (init_optimization_passes): Move pass_parallelize_loops
18074 earlier, after GRAPHITE transforms and IV canonicalization.
18076 2013-05-16 Jakub Jelinek <jakub@redhat.com>
18078 * omp-low.c (extract_omp_for_data): For collapsed loops,
18079 if at least one of the loops is known at compile time to
18080 iterate zero times, set count to 0.
18081 (expand_omp_regimplify_p): New function.
18082 (expand_omp_for_generic): For collapsed loops, if at least
18083 one of the loops isn't known to iterate at least once,
18084 add runtime check with setting count to 0.
18085 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
18086 For unsigned types if it isn't known at compile time that
18087 the loop will iterate at least once, add runtime check to bypass
18088 the whole loop if initial condition isn't true.
18090 2013-05-16 Nathan Sidwell <nathan@codesourcery.com>
18092 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
18094 2013-05-16 Marc Glisse <marc.glisse@inria.fr>
18096 PR middle-end/57286
18097 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
18098 transformations to avoid an infinite loop.
18100 2013-05-16 Marek Polacek <polacek@redhat.com>
18102 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
18104 2013-05-15 Leif Ekblad <leif@rdos.net>
18106 * config/i386/i386.c (ix86_decompose_address): Use
18107 DEFAULT_TLS_SEG_REG to access TLS segment register.
18108 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
18109 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
18110 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
18112 2013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
18115 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
18116 sibling calls to functions that would normally be lazily bound,
18117 unless $gp is call-clobbered.
18119 2013-05-15 Uros Bizjak <ubizjak@gmail.com>
18121 * config/i386/i386.c (ix86_option_override_internal): Update
18122 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
18123 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
18124 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
18125 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
18127 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
18129 2013-05-15 Andreas Schwab <schwab@suse.de>
18131 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
18132 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
18135 2013-05-15 Teresa Johnson <tejohnson@google.com>
18137 * loop-unroll.c (report_unroll_peel): Check decision before
18138 emitting unroll/peel message.
18140 2013-05-15 Teresa Johnson <tejohnson@google.com>
18142 * function.h (has_bb_partition): New rtl_data flag.
18143 (bb_reorder_complete): Ditto.
18144 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
18145 instead of flag_reorder_blocks_and_partition.
18146 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
18147 with some enhancements.
18148 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
18149 * bb-reorder.c (connect_traces): Check for has_bb_partition
18150 instead of flag_reorder_blocks_and_partition.
18151 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
18152 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
18153 verify_hot_cold_block_grouping.
18154 (partition_hot_cold_basic_blocks): Set has_bb_partition.
18156 2013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
18159 * config/arm/predicates.md (call_insn_operand): New predicate.
18160 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
18161 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
18162 if insn is not a tail call.
18163 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
18165 * config/arm/arm.h (enum reg_class): New caller save register class.
18166 (REG_CLASS_NAMES): Likewise.
18167 (REG_CLASS_CONTENTS): Likewise.
18168 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
18171 2013-05-15 Richard Biener <rguenther@suse.de>
18173 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
18174 of MSG_OPTIMIZED_LOCATIONS.
18175 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
18176 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
18178 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
18179 of MSG_OPTIMIZED_LOCATIONS.
18180 (execute_vect_slp): Likewise.
18181 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
18182 (vect_create_cond_for_alias_checks): Likewise.
18183 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
18184 (vect_recog_widen_mult_pattern): Likewise.
18185 (vect_recog_widen_sum_pattern): Likewise.
18186 (vect_recog_over_widening_pattern): Likewise.
18187 (vect_recog_widen_shift_pattern): Likewise.
18188 (vect_recog_vector_vector_shift_pattern): Likewise.
18189 (vect_recog_divmod_pattern): Likewise.
18190 (vect_recog_mixed_size_cond_pattern): Likewise.
18191 (vect_recog_bool_pattern): Likewise.
18192 (vect_pattern_recog_1): Likewise.
18194 2013-05-15 Martin Jambor <mjambor@suse.cz>
18196 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
18197 non-functions to builtin_unreachable.
18198 * ipa-inline-transform.c (inline_call): Do not assert estimates were
18199 correct when new direct edges were discovered.
18201 2013-05-15 Martin Jambor <mjambor@suse.cz>
18203 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
18204 header, print symbol order instead of node uid, print more information
18205 about indirect edge targets.
18206 (ipa_make_edge_direct_to_target): Print symbol order instead of node
18208 (ipa_make_edge_direct_to_target): Likewise.
18209 (remove_described_reference): Likewise.
18210 (propagate_controlled_uses): Likewise.
18211 (ipa_print_node_params): Also print symbol order.
18212 (ipcp_transform_function): Print symbol order instead of node uids.
18213 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
18214 (cgraph_get_create_real_symbol_node): Likewise.
18215 * ipa-cp.c (print_lattice): Likewise.
18216 (print_all_lattices): Likewise.
18217 (determine_versionability): Likewise.
18218 (initialize_node_lattices): Likewise.
18219 (estimate_local_effects): Likewise.
18220 (update_profiling_info): Likewise.
18221 (create_specialized_node): Likewise.
18222 (perhaps_add_new_callers): Likewise.
18223 (decide_about_value): Likewise.
18224 (decide_whether_version_node): Likewise.
18225 (identify_dead_nodes): Likewise.
18226 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
18227 (dump_inline_summary): Likewise.
18228 (estimate_node_size_and_time): Likewise.
18229 (inline_analyze_function): Likewise.
18230 * ipa-inline.c (report_inline_failed_reason): Likewise.
18231 (want_early_inline_function_p): Likewise.
18232 (edge_badness): Likewise.
18233 (update_edge_key): Likewise.
18234 (inline_small_functions): Likewise. Add dumping of order to two other
18236 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
18237 instead of node uids.
18238 (propagate_pure_const): Likewise.
18239 (propagate_pure_const): Likewise.
18240 * ipa-utils.c (dump_cgraph_node_set): Likewise.
18241 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
18242 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
18244 * tree-pretty-print.c (dump_function_header): Likewise.
18245 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
18246 Print symbol order instead of node uids.
18248 2013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
18250 * config/s390/s390.c (s390_register_move_cost): Don't impose the
18251 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
18253 2013-05-15 Richard Biener <rguenther@suse.de>
18255 PR tree-optimization/57275
18256 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
18257 return value for fail to do runtime alias checks for gather loads.
18259 2013-05-15 Jan Hubicka <jh@suse.cz>
18263 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
18264 weakrefs are not external.
18265 (lto_symtab_merge_decls): Fix thinko when dealing with
18266 non-lto_symtab decls.
18267 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
18268 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
18269 * varpool.c (dump_varpool_node): Dump more flags.
18271 2013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
18273 * config/i386/i386.c (processor_alias_table): Add instruction
18274 FSGSBASE for AMD bdver3 architecture.
18276 2013-05-14 Jakub Jelinek <jakub@redhat.com>
18278 * tree.c (warn_deprecated_use): Print file:line using locus color.
18279 * diagnostic.c (diagnostic_report_current_module): Print file:line
18280 and file:line:column using locus color.
18282 2013-05-14 Mike Stump <mikestump@comcast.net>
18284 * gdbinit.in: Add __null.
18286 2013-05-14 Mike Stump <mikestump@comcast.net>
18288 * recog.h: Rename struct recog_data to Recog_data.
18289 * recog.c: Likewise.
18290 * reload.c (can_reload_into): Likewise.
18291 * config/picochip/picochip.c: Likewise.
18293 2013-05-14 Mike Stump <mikestump@comcast.net>
18295 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
18297 2013-05-14 Steven Bosscher <steven@gcc.gnu.org>
18299 * resource.h (struct resources): Remove unch_memory member.
18300 (CLEAR_RESOURCE): Don't clear unch_memory.
18301 * resource.c (mark_referenced_resources): Don't set it.
18302 (mark_set_resources): Likewise.
18303 (mark_target_live_regs): Don't clear it.
18304 (init_resource_info): Likewise.
18305 * reorg.c (resource_conflicts_p): Don't compare it.
18306 (redundant_insn): Don't set it.
18308 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
18310 * emit-rtl.c (next_label): Remove unused function.
18311 (skip_consecutive_labels, link_cc0_insns): Move to ...
18312 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
18313 only place where these functions are used, and make them static.
18315 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
18317 * fold-const.c (fold_negate_expr): Handle vectors.
18318 (fold_truth_not_expr): Make it static.
18319 (fold_invert_truthvalue): New static function.
18320 (invert_truthvalue_loc): Handle vectors. Do not call
18321 fold_truth_not_expr directly.
18322 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
18323 <TRUTH_NOT_EXPR>: Do not cast to boolean.
18324 (fold_comparison): Handle vector constants.
18325 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
18326 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
18327 * tree.h (fold_truth_not_expr): Remove declaration.
18329 2013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
18331 * config/aarch64/aarch64-simd.md
18332 (aarch64_vcond_internal<mode>): Rename to...
18333 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
18334 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
18335 float modes. Clarify all iterator modes.
18336 (vcond<mode><mode>): Use new name for vcond expanders.
18337 (vcond<v_cmp_result><mode>): Likewise.
18338 (vcondu<mode><mode>: Likewise.
18339 * config/aarch64/iterators.md (VDQF_COND): New.
18341 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
18344 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
18345 variable for the shift amount. Check that we shift by non-negative
18348 2013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
18351 * config/arm/arm.h (EPILOGUE_USES): Only return true
18352 for LR_REGNUM after epilogue_completed.
18354 2013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
18356 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
18357 is error_mark_node.
18359 2013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
18362 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
18363 and Solaris 11+/x86 with gld.
18364 * configure: Regenerate.
18366 2013-05-14 Jakub Jelinek <jakub@redhat.com>
18368 * expmed.c (expand_shift_1): Canonicalize rotates by
18369 constant bitsize / 2 to bitsize - 1.
18370 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
18371 case ROTATERT>: Likewise.
18374 2013-05-10 Jakub Jelinek <jakub@redhat.com>
18376 * config/i386/i386.md (rotateinv): New code attr.
18377 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
18378 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
18379 roll $31, %eax, etc.
18381 2013-05-14 Richard Biener <rguenther@suse.de>
18383 PR middle-end/57235
18384 * tree-eh.c (sink_clobbers): Give up for successors with
18385 multiple predecessors and no virtual uses.
18387 2013-05-14 Eric Botcazou <ebotcazou@adacore.com>
18389 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
18390 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
18392 2013-05-14 Jakub Jelinek <jakub@redhat.com>
18394 PR middle-end/57251
18395 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
18396 the case when both op0 and op1 have VOIDmode.
18398 2013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
18400 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
18401 in multiply-accumulate mode.
18403 2013-05-13 Guozhi Wei <carrot@google.com>
18405 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
18407 2013-05-13 Kai Tietz <ktietz@redhat.com>
18410 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
18411 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
18412 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
18413 * config/i386/i386.c (ix86_option_override_internal): Likewise.
18414 (ix86_expand_prologue): Likewise.
18415 (ix86_expand_split_stack_prologue): Likewise.
18416 (legitimate_pic_address_disp_p): Likewise.
18417 (legitimize_pic_address): Likewise.
18418 (legitimize_tls_address): Likewise.
18419 (legitimize_pe_coff_symbol): Likewise.
18420 (output_pic_addr_const): Likewise.
18421 (construct_plt_address): Likewise.
18422 (ix86_expand_call): Likewise.
18423 (x86_output_mi_thunk): Likewise.
18424 (x86_function_profiler): Likewise.
18426 2013-05-13 Sofiane Naci <sofiane.naci@arm.com>
18428 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
18429 similar switch cases.
18430 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
18431 (aarch64_simd_mov_to_<mode>low): Delete.
18432 (aarch64_simd_mov_to_<mode>high): Delete.
18433 (move_lo_quad_<mode>): Add w<-r alternative.
18434 (aarch64_simd_move_hi_quad_<mode>): Likewise.
18435 (aarch64_simd_mov_from_*): Update type attribute.
18436 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
18439 2013-05-13 Jan Hubicka <jh@suse.cz>
18441 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
18442 * config/i386/i386.c (ix86_compute_frame_layout,
18443 ix86_expand_epilogue, emit_i387_cw_initialization,
18444 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
18445 ix86_local_alignment): Fix use of size/speed predicates.
18447 2013-05-13 Jakub Jelinek <jakub@redhat.com>
18449 PR tree-optimization/45216
18450 PR tree-optimization/57157
18451 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
18452 the (-Y) & (B - 1) variant if OP is |.
18453 * expmed.c (expand_shift_1): For rotations by const0_rtx just
18454 return shifted. Use (-op1) & (prec - 1) as other_amount
18455 instead of prec - op1.
18457 2013-05-13 Martin Jambor <mjambor@suse.cz>
18459 PR middle-end/42371
18460 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
18461 (ipa_constant_data): New type.
18462 (ipa_jump_func): Use ipa_constant_data to hold information about
18463 constant jump functions.
18464 (ipa_get_jf_constant): Adjust to jump function type changes.
18465 (ipa_get_jf_constant_rdesc): New function.
18466 (ipa_param_descriptor): New field controlled_uses.
18467 (ipa_get_controlled_uses): New function.
18468 (ipa_set_controlled_uses): Likewise.
18469 * ipa-ref.h (ipa_find_reference): Declare.
18470 * ipa-prop.c (ipa_cst_ref_desc): New type.
18471 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
18473 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
18474 New parameter cs. Adjust all callers.
18475 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
18476 (remove_described_reference): New function.
18477 (jfunc_rdesc_usable): Likewise.
18478 (try_make_edge_direct_simple_call): Decrement controlled use count,
18479 attempt to remove reference if it hits zero.
18480 (combine_controlled_uses_counters): New function.
18481 (propagate_controlled_uses): Likewise.
18482 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
18483 (ipa_edge_duplication_hook): Duplicate reference descriptions.
18484 (ipa_print_node_params): Print described use counter.
18485 (ipa_write_jump_function): Adjust to jump function type changes.
18486 (ipa_read_jump_function): New parameter CS, pass it to
18487 ipa_set_jf_constant. Adjust caller.
18488 (ipa_write_node_info): Stream controlled use count
18489 (ipa_read_node_info): Likewise.
18490 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
18492 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
18493 count. Remove cloning-added reference if it reaches zero.
18494 * ipa-ref.c (ipa_find_reference): New function.
18496 2013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
18498 * config/i386/i386.c (processor_target_table): Modified default
18499 alignment values for AMD BD and BT architectures.
18501 2013-05-13 Marc Glisse <marc.glisse@inria.fr>
18503 * tree-vect-generic.c (uniform_vector_p): Move ...
18504 * tree.c (uniform_vector_p): ... here.
18505 * tree.h (uniform_vector_p): Declare it.
18506 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
18509 2013-05-13 Jakub Jelinek <jakub@redhat.com>
18511 PR tree-optimization/57230
18512 * tree-ssa-strlen.c (handle_char_store): Record length for
18513 array store from STRING_CST.
18515 PR tree-optimization/57230
18516 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
18519 2013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
18521 * config/epiphany/epiphany.c (epiphany_init): Check size of
18522 NUM_MODES_FOR_MODE_SWITCHING.
18523 (epiphany_expand_prologue):
18524 Remove CONFIG_REGNUM initial value handling code.
18525 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
18526 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
18527 (emit_set_fp_mode, epiphany_mode_after): Likewise.
18528 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
18529 Don't return 1 for FP_MODE_NONE.
18530 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
18531 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
18532 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
18533 * config/epiphany/epiphany.md (save_config): New pattern.
18535 2013-05-12 Uros Bizjak <ubizjak@gmail.com>
18537 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
18539 2013-05-10 Uros Bizjak <ubizjak@gmail.com>
18541 * config/i386/i386.md (memory): Handle sseishft1.
18542 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
18543 (*vec_extractv2di_1): Ditto.
18545 2013-05-10 Vladimir Makarov <vmakarov@redhat.com>
18547 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
18550 2013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
18552 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
18553 Add mthumb/march=armv7-a multilib.
18554 Add mthumb/march=armv7-r multilib.
18555 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
18557 2013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
18559 * config/v850/t-rtems: Add more multilibs.
18561 2013-05-10 Richard Biener <rguenther@suse.de>
18563 PR tree-optimization/57214
18564 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
18565 not propagate from SSA names that occur in abnormal PHI nodes.
18567 2013-05-10 Marc Glisse <marc.glisse@inria.fr>
18569 * stor-layout.c (element_precision): New function.
18570 * machmode.h (element_precision): Declare it.
18571 * tree.c (build_minus_one_cst): New function.
18572 (element_precision): Likewise.
18573 * tree.h (build_minus_one_cst): Declare new function.
18574 (element_precision): Likewise.
18575 * fold-const.c (operand_equal_p): Use element_precision.
18576 (fold_binary_loc): Handle vector types.
18577 * convert.c (convert_to_integer): Use element_precision.
18578 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
18581 2013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
18583 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
18584 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
18585 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
18586 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
18587 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
18588 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
18589 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
18590 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
18591 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
18592 (Uuw8): New constraints.
18593 (Usb4): Move into alphabetical order.
18594 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
18595 (sd8_operand, ub8_operand, uw8_operand): New predicates.
18596 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
18597 previously unnamed patterns.
18598 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
18599 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
18600 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
18601 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
18602 of set_attr_alternative/if_then_else. Use extended_mips16 instead
18603 of specific lengths.
18605 2013-05-10 Jakub Jelinek <jakub@redhat.com>
18607 * config/i386/i386.md (rotateinv): New code attr.
18608 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
18609 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
18610 roll $31, %eax, etc.
18612 PR tree-optimization/45216
18613 PR tree-optimization/57157
18614 * tree-ssa-forwprop.c (simplify_rotate): New function.
18615 (ssa_forward_propagate_and_combine): Call it.
18617 2013-05-10 Richard Biener <rguenther@suse.de>
18619 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
18620 disable peeling when we version for aliasing.
18621 (vector_alignment_reachable_p): Honor explicit user alignment.
18622 (vect_supportable_dr_alignment): Likewise.
18623 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
18624 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
18625 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
18626 then peeling to arrange for the cost-model check to come first.
18628 2013-05-10 Alan Modra <amodra@gmail.com>
18630 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
18631 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
18632 * configure: Regenerate.
18634 2013-05-10 Alan Modra <amodra@gmail.com>
18637 * varasm.c (default_elf_select_section): Move !DECL_P check..
18638 (get_named_section): ..to here before calling get_section_name.
18640 (default_section_type_flags): Add DECL_P check.
18641 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
18642 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
18644 2013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
18646 * config/epiphany/epiphany.c (epiphany_expand_prologue):
18647 When using gen_stack_adjust_str with a register offset, add a
18648 REG_FRAME_RELATED_EXPR note.
18650 2013-05-09 Uros Bizjak <ubizjak@gmail.com>
18652 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
18653 (*vec_extractv4si_zext_mem): Ditto.
18654 (*vec_extractv2di): Add 0->x and x->x alternatives.
18655 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
18656 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
18658 2013-05-09 Jason Merrill <jason@redhat.com>
18660 N3639 C++1y VLA support
18661 * gimplify.c (gimplify_vla_decl): Don't touch an existing
18664 * tree.c (build_constructor_va): New.
18665 * tree.h: Declare it.
18667 2013-05-09 Martin Jambor <mjambor@suse.cz>
18670 * gimple-fold.c (canonicalize_constructor_val): Call
18671 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
18673 2013-05-09 Jan Hubicka <jh@suse.cz>
18674 Richard Biener <rguenther@suse.de>
18677 * symtab.c (symtab_make_decl_local): Do not add private names.
18679 2013-05-09 Jan Hubicka <jh@suse.cz>
18682 * symtab.c (insert_to_assembler_name_hash): Handle clones.
18683 (unlink_from_assembler_name_hash): Likewise.
18684 (symtab_prevail_in_asm_name_hash, symtab_register_node,
18685 symtab_unregister_node, symtab_initialize_asm_name_hash,
18686 change_decl_assembler_name): Update.
18688 2013-05-09 Sofiane Naci <sofiane.naci@arm.com>
18690 * config/aarch64/aarch64.md: New movtf split.
18691 (*movtf_aarch64): Update.
18692 (aarch64_movdi_tilow): Handle TF modes and rename to
18693 aarch64_movdi_<mode>low.
18694 (aarch64_movdi_tihigh): Handle TF modes and rename to
18695 aarch64_movdi_<mode>high
18696 (aarch64_movtihigh_di): Handle TF modes and rename to
18697 aarch64_mov<mode>high_di
18698 (aarch64_movtilow_di): Handle TF modes and rename to
18699 aarch64_mov<mode>low_di
18700 (aarch64_movtilow_tilow): Remove spurious whitespace.
18701 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
18703 (aarch64_print_operand): Update.
18705 2013-05-09 Alan Modra <amodra@gmail.com>
18707 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
18709 * configure: Regenerate.
18711 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
18713 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
18714 splitter preparation statements.
18715 * config/i386/sse.md (*vec_extract* splitters): Ditto.
18716 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
18719 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18721 * gimple-ssa-strength-reduction.c (count_candidates): Change
18722 return value to int.
18723 (analyze_candidates_and_replace): Change type of length to int.
18725 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
18727 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
18728 (*vec_extract<mode>): Use VI12_128 mode iterator.
18729 (*vec_extract<mode>_mem): Ditto.
18730 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
18733 2013-05-08 Diego Novillo <dnovillo@google.com>
18738 2012-08-17 Diego Novillo <dnovillo@google.com>
18741 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
18742 * config.in: Regenerate.
18743 * configure: Regenerate.
18744 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
18746 2013-05-08 Jan Hubicka <jh@suse.cz>
18749 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
18750 * cgraph.h (symtab_node_base): Add unique_name.
18751 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
18752 input_overwrite_node, input_varpool_node): Stream unique_name.
18753 * cgraphclones.c (cgraph_create_virtual_clone,
18754 cgraph_function_versioning): Set unique_name.
18755 * ipa.c (function_and_variable_visibility): Set unique_name.
18757 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18759 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
18760 (alloc_cand_and_find_basis): Restrict conditional candidate
18761 processing to CAND_MULTs.
18763 2013-05-08 Jan Hubicka <jh@suse.cz>
18766 lto-symtab.c (lto_symtab_symbol_p): New function.
18767 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
18768 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
18769 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
18770 Skip static symbols.
18772 2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
18774 PR tree-optimization/57200
18775 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
18776 Only call inform if the preceding warning_at returns true.
18778 2013-05-07 Han Shen <shenhan@google.com>
18780 * cfgexpand.c (record_or_union_type_has_array_p): New function.
18781 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
18782 * common.opt (fstack-protector-strong): New option.
18783 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
18784 * doc/invoke.texi (Optimization Options): Document
18785 "-fstack-protector-strong".
18786 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
18788 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
18790 * config/mips/mips.c (mips_machine_reorg2): Return 0.
18792 2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
18794 * ira.c (update_equiv_regs): Add insn having equiv memory even if
18795 it is not lhs of the insn.
18796 (setup_reg_equiv): Remove insn having equiv memory which it is not
18798 * lra-constraints.c (process_address): Try to improve generation
18799 code for address base + disp.
18800 (lra_constraints): Make correct the code for checking insn setting
18801 up backward equivalence. Remove insn only if it is in the init
18803 * lra-eliminations.c (update_reg_eliminate): Change return value.
18804 (lra_eliminate): Use the result.
18806 2013-05-07 Uros Bizjak <ubizjak@gmail.com>
18808 * config/i386/sse.md (ssescalarnummask): New mode attribute.
18809 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
18810 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
18811 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
18812 register target operands.
18813 (*vec_extractv8hi_sse2): New pattern.
18814 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
18815 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
18816 (*vec_extract<mode>_mem): New insn and split pattern.
18818 2013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
18820 * config/arm/arm.c (arm_asan_shadow_offset): New function.
18821 (TARGET_ASAN_SHADOW_OFFSET): Define.
18822 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
18823 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
18825 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18827 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
18828 (incr_vec_index): Return -1 if increment not found.
18829 (create_add_on_incoming_edge): Assert if increment not found.
18830 (record_increment): Limit number of increments recorded.
18831 (all_phi_incrs_profitable): Return false if an increment not found.
18832 (replace_profitable_candidates): Don't process increments that were
18834 (analyze_candidates_and_replace): Limit size of incr_vec.
18836 2013-05-07 Richard Biener <rguenther@suse.de>
18838 * calls.c (special_function_p): setjmp-like functions are leaf.
18839 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
18840 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
18842 2013-05-07 Sofiane Naci <sofiane.naci@arm.com>
18844 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
18845 (aarch64_simd_mov<mode>): New expander.
18846 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
18847 (aarch64_simd_mov_to_<mode>high): Likewise.
18848 (aarch64_simd_mov_from_<mode>low): Likewise.
18849 (aarch64_simd_mov_from_<mode>high): Likewise.
18850 (aarch64_dup_lane<mode>): Update.
18851 (aarch64_dup_lanedi): New instruction pattern.
18852 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
18853 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
18855 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18857 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
18858 (replace_mult_candidate): Remove unnecessary argument; remove
18859 unnecessary parameter from call to introduce_cast_before_cand.
18860 (replace_unconditional_candidate): Remove unnecessary parameter
18861 from call to replace_mult_candidate.
18862 (replace_conditional_candidate): Likewise.
18863 (insert_initializers): Use make_temp_ssa_name.
18864 (introduce_cast_before_cand): Remove unnecessary argument; use
18865 make_temp_ssa_name.
18866 (replace_one_candidate): Remove unnecessary argument; remove
18867 unnecessary parameter from calls to introduce_cast_before_cand.
18868 (replace_profitable_candidates): Remove unnecessary parameters
18869 from calls to replace_one_candidate.
18871 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18873 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
18874 phi def as possibly hiding a basis for a CAND_ADD whose operands
18875 have been commuted in the analysis.
18876 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
18878 2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
18880 * config/aarch64/aarch64.md
18881 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
18882 shift value between 0-4.
18884 2013-05-07 Richard Biener <rguenther@suse.de>
18886 * double-int.h (rshift): New overload.
18887 * double-int.c (rshift): New function.
18888 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
18889 (create_reference_ops_from_ref): Remove.
18890 (vn_reference_insert): Use shared ops for constructing the
18891 reference and copy it.
18893 2013-05-07 Richard Biener <rguenther@suse.de>
18895 PR middle-end/57190
18896 * tree-eh.c (sink_clobbers): Properly propagate
18897 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
18899 2013-05-07 Jakub Jelinek <jakub@redhat.com>
18901 PR tree-optimization/57149
18902 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
18903 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
18904 collect_phi_def_edges, execute_late_warn_uninitialized): Use
18905 uninit_undefined_value_p instead of ssa_undefined_value_p.
18908 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
18909 for modifier == EXPAND_INITIALIZER.
18911 2013-05-07 Anton Blanchard <anton@samba.org>
18913 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
18914 for powerpc64 little endian.
18915 * configure: Regenerate.
18917 2013-05-06 Graham Stott <grahams@btinternet.com>
18919 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
18920 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
18921 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
18922 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
18924 2013-05-06 Graham Stott <grahams@btinternet.com>
18926 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
18927 codes which allow non-lvalues.
18929 2013-05-06 Marc Glisse <marc.glisse@inria.fr>
18931 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
18932 components are all 1s.
18933 (integer_minus_onep): New function.
18934 * tree.h (integer_minus_onep): Declare it.
18935 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
18936 integer_minus_onep instead of integer_all_onesp.
18938 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
18941 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
18942 variations of these patterns.
18944 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
18946 * config/i386/i386.md (isa): Add x64_sse4 member.
18947 (enabled): Handle x64_sse4.
18948 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
18949 instruction for 64bit SSE4_1 targets. Update insn attributes.
18950 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
18951 instruction for SSE4_1 targets. Update insn attributes.
18952 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
18953 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
18954 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
18956 (*vec_extractv4si): Rename from *sse4_1_pextrd.
18957 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
18958 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
18960 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
18963 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
18965 2013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
18967 * final.c (do_assembler_dialects): Don't handle curly braces and
18968 vertical bar escaped by % as dialect delimiters.
18969 (output_asm_insn): Print curly braces and vertical bar if escaped
18970 by % and ASSEMBLER_DIALECT defined.
18971 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
18972 * doc/tm.texi: Regenerated.
18974 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
18976 * config/mips/mips.c: Include tree-pass.h.
18977 (mips_reorg): Split in pre- and post-dbr_schedule parts.
18978 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
18979 (pass_mips_machine_reorg2): New machine specific pass.
18980 (insert_pass_mips_machine_reorg2): New pass plugin definition.
18981 (mips_option_override): Register the new pass.
18982 * rtl.h (cleanup_barriers): Remove prototype.
18983 (dbr_schedule): Likewise.
18984 * jump.c (cleanup_barriers): Make static.
18985 * reorg.c (dbr_schedule): Likewise.
18987 2013-05-06 Richard Biener <rguenther@suse.de>
18989 PR tree-optimization/57185
18990 * tree-parloops.c (add_field_for_reduction): Handle anonymous
18991 SSA names properly.
18993 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
18996 * config/i386/i386.c (add_parameter_dependencies): Add dependence
18997 between "first_arg" and "insn", not "last" and "insn".
18999 2013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
19001 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
19002 (find_candidates_in_block): Re-enable slsr_process_phi.
19003 (create_phi_basis): Fix double counting of candidate adjustment.
19005 2013-05-06 Richard Biener <rguenther@suse.de>
19007 PR middle-end/57147
19008 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
19009 the edge is also fallthru, preserve it and just clear the
19011 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
19012 also complex, preserve that and just clear the fallthru flag.
19013 * tree-inline.c (update_ssa_across_abnormal_edges): Also
19014 update virtual operands.
19016 2013-05-06 Alan Modra <amodra@gmail.com>
19018 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
19019 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
19020 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
19021 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
19022 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
19023 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
19025 2013-05-06 Alan Modra <amodra@gmail.com>
19027 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
19028 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
19029 (DEFAULT_ASM_ENDIAN): Define.
19030 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
19031 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
19032 Update -K PIC clause from sysv4.h.
19033 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
19034 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
19036 2013-05-06 Alan Modra <amodra@gmail.com>
19038 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
19039 twice for little-endian.
19040 (ashrdi3_no_power, ashrdi3): Support little-endian.
19042 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
19045 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
19046 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
19048 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
19049 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
19050 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
19053 2013-05-05 Steven Bosscher <steven@gcc.gnu.org>
19054 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
19056 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
19057 * config/pa/pa.opt: Make mbig-switch a no-op.
19058 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
19059 (CASE_VECTOR_MODE): Always return SImode.
19060 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
19061 for the !TARGET_BIG_SWITCH case.
19062 * config/pa/pa-linux.h: Likewise.
19063 * config/pa/pa-openbsd.h: Likewise.
19064 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
19065 * config/pa/pa.md (short_jump): Remove define_insn.
19066 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
19067 (casesi0): Remove define_insn.
19068 (type): Remove btable_branch.
19069 (pa_combine_type): Likewise.
19070 (in_nullified_branch_delay): Likewise.
19071 (in_call_delay): Likewise.
19072 (define_delay): Likewise.
19073 (define_insn_reservation "Z3"): Likewise.
19074 (define_insn_reservation "Z4"): Likewise.
19075 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
19076 (pa_adjust_insn_length): Remove adjustment for btable branches.
19077 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
19080 2013-05-05 Uros Bizjak <ubizjak@gmail.com>
19082 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
19083 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
19084 Add m->r,x alternatives.
19085 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
19086 splitters using SWI48x mode iterator.
19087 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
19088 TARGET_64BIT. Add m->x alternative.
19089 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
19090 Add o->x alternative. Enable for TARGET_SSE.
19091 (sse_storeq): Remove expander.
19092 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
19093 with memory input operand.
19094 (*vec_extractv2di_1 splitter): New.
19095 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
19096 * config/i386/i386.md (ssevecmodelower): New mode attribute.
19098 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
19100 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
19101 (INT_LOWPART): Delete.
19102 (extract_MB): Adjust.
19103 (extract_ME): Adjust.
19104 (print_operand): Adjust.
19106 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
19108 * config/rs6000/predicates.md (reg_or_add_cint_operand,
19109 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
19110 (reg_or_logical_cint_operand, easy_fp_constant,
19111 logical_const_operand): Delete "CONST_DOUBLE" case.
19112 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
19113 "HOST_BITS_PER_WIDE_INT == 64" test.
19114 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
19115 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
19116 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
19117 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
19119 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
19120 CONST_DOUBLE DImode/VOIDmode case.
19121 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
19122 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
19123 CONST_DOUBLE VOIDmode case.
19124 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
19125 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
19126 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
19127 Delete CONST_DOUBLE case.
19128 (splitters for mov FMOVE64 const_double): Delete
19129 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
19130 "HOST_BITS_PER_WIDE_INT >= 64" test.
19131 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
19133 (mov DI const_double): Delete.
19135 2013-05-04 Jakub Jelinek <jakub@redhat.com>
19137 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
19138 on op shows all bits zero in mode of a lowpart subreg, return zero.
19140 2013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
19143 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
19144 to save TFmode registers and DImode to save TImode registers for
19145 caller save operations.
19146 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
19147 mark being partially clobbered since they only use the first
19150 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
19151 and TDmode only use the upper 64-bits of each VSX register.
19153 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
19155 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
19156 (find_candidates_in_block): Disable slsr_process_phi.
19158 2013-05-03 Guozhi Wei <carrot@google.com>
19160 * coverage.c (coverage_obj_init): Move the construction of gcov
19162 (build_init_ctor): ... here.
19164 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
19166 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
19167 (slsr_cand_d): Redefine def_phi.
19168 (stride_status, phi_adjust_status, count_phis_status): New enums.
19169 (find_phi_def): New.
19170 (find_basis_for_base_expr): New.
19171 (find_basis_for_candidate): Handle hidden bases.
19172 (alloc_cand_and_find_basis): Handle phi candidates.
19173 (slsr_process_phi): New.
19174 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
19175 (create_mul_imm_cand): Likewise.
19176 (create_add_ssa_cand): Exclude phi base candidates.
19177 (create_add_imm_cand): Likewise.
19178 (slsr_process_cast): Likewise.
19179 (slsr_process_copy): Likewise.
19180 (find_candidates_in_block): Handle phi candidates.
19181 (dump_candidate): Likewise.
19182 (unconditional_cands): Delete.
19183 (unconditional_cands_with_known_stride_p): Delete.
19184 (phi_dependent_cand_p): New.
19185 (cand_increment): Handle phi-dependent candidates.
19186 (replace_dependent): Delete.
19187 (replace_mult_candidate): New.
19188 (replace_unconditional_candidate): New.
19189 (incr_vec_index): Move to avoid forward reference.
19190 (create_add_on_incoming_edge): New.
19191 (create_phi_basis): New.
19192 (replace_dependents): Delete.
19193 (replace_conditional_candidate): New.
19194 (phi_add_costs): New.
19195 (replace_uncond_cands_and_profitable_phis): New.
19196 (record_increment): Handle phi adjustments.
19197 (record_phi_increments): New.
19198 (record_increments): Handle phi adjustments.
19199 (phi_incr_cost): New.
19200 (lowest_cost_path): Handle phis.
19201 (total_savings): Likewise.
19202 (analyze_increments): Likewise.
19203 (ncd_with_phi): New.
19204 (ncd_of_cand_and_phis): New.
19205 (nearest_common_dominator_for_cands): Handle phi increments.
19206 (all_phi_incrs_profitable): New.
19207 (replace_profitable_candidates): Handle phi-dependent candidates.
19208 (analyze_candidates_and_replace): Likewise.
19210 2013-05-03 Teresa Johnson <tejohnson@google.com>
19213 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
19214 do not exceed REG_BR_PROB_BASE.
19216 2013-05-03 Jeff Law <law@redhat.com>
19218 PR tree-optimization/57144
19219 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
19220 operand of the condition will bit into the new type when eliminating
19221 a cast feeding a condition.
19223 2013-05-03 Jakub Jelinek <jakub@redhat.com>
19225 PR rtl-optimization/57130
19226 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
19227 of COMPARE as in_code to the recursive call if needed.
19229 2013-05-03 Uros Bizjak <ubizjak@gmail.com>
19231 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
19232 (enabled): Handle new members.
19233 * config/i386/sse.md (*vec_concatv2si): Merge from
19234 *vec_concatv2si_sse2 and vec_concatv2si_sse.
19235 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
19237 2013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
19239 PR tree-optimization/57027
19240 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
19241 for fnms opportunity, check we got the prerequisite kind
19242 of tree / gimple before using accessor functions.
19244 2013-05-03 Richard Biener <rguenther@suse.de>
19246 * double-int.h (lshift): New overload without precision
19247 and arith argument.
19248 (operator *=, operator +=, operator -=): Move ...
19249 * double-int.c (operator *=, operator +=, operator -=): ... here
19250 and implement more efficiently.
19251 (mul_double_with_sign): Remove.
19252 (lshift_double): Adjust to take unsinged shift argument, push
19253 dispatching code to callers.
19254 (mul_double_wide_with_sign): Add early out for callers that
19255 are not interested in high parts or overflow.
19256 (lshift): New function.
19257 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
19258 dispatch code here.
19259 (lrotate, rrotate): Use logical shifts.
19260 * expr.c (get_inner_reference): Use lshift.
19261 * fixed-value.c (do_fixed_divide): Likewise.
19262 * tree-dfa.c (get_ref_base_and_extent): Likewise.
19263 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
19264 (indirect_refs_may_alias_p): Likewise.
19265 (stmt_kills_ref_p_1): Likewise.
19267 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
19269 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
19271 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
19273 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
19274 scalar form of FABD instruction.
19276 2013-05-02 Vladimir Makarov <vmakarov@redhat.com>
19278 * lra-constraints.c (process_alt_operands): Add checking alt
19279 number to choose the best alternative.
19281 2013-05-02 Richard Biener <rguenther@suse.de>
19283 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
19284 bitmap and its handling.
19285 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
19287 2013-05-02 Richard Biener <rguenther@suse.de>
19289 PR middle-end/57140
19290 * tree-inline.c (copy_loops): Properly handle removed loops.
19291 (copy_cfg_body): Mark destination loops for fixup if source
19292 loops needed fixup.
19294 2013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
19297 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
19298 generating simple_return for naked functions.
19300 2013-05-02 Martin Jambor <mjambor@suse.cz>
19302 PR middle-end/56988
19303 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
19304 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
19306 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
19307 ipa_agg_replacement_value structures.
19308 (known_aggs_to_agg_replacement_list): Likewise.
19309 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
19310 (read_agg_replacement_chain): Likewise.
19311 (ipcp_transform_function): Also check that by_ref flags match.
19313 2013-05-02 Richard Biener <rguenther@suse.de>
19315 * graphds.h (struct graph): Add obstack member.
19316 * graphds.c (new_graph): Initialize obstack and allocate
19318 (add_edge): Allocate edge from the obstack.
19319 (free_graph): Free the obstack instead of all edges and vertices.
19321 2013-05-02 Teresa Johnson <tejohnson@google.com>
19323 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
19325 * cfg.c (update_bb_profile_for_threading): Ditto.
19326 * tree-inline.c (copy_bb): Ditto.
19327 (copy_edges_for_bb): Ditto.
19328 (initialize_cfun): Ditto.
19329 (copy_cfg_body): Ditto.
19330 (expand_call_inline): Ditto.
19331 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
19332 (estimate_node_size_and_time): Ditto.
19333 (inline_merge_summary): Ditto.
19334 * cgraphclones.c (cgraph_clone_edge): Ditto.
19335 (cgraph_clone_node): Ditto.
19336 * sched-rgn.c (compute_dom_prob_ps): Ditto.
19337 (compute_trg_info): Ditto.
19339 2013-05-02 Ian Bolton <ian.bolton@arm.com>
19341 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
19342 S reg when fp attribute set.
19343 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
19345 2013-05-02 Ian Bolton <ian.bolton@arm.com>
19347 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
19349 (*and_one_cmplsi3_compare0_uxtw): Likewise.
19350 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
19351 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
19353 2013-05-02 Richard Biener <rguenther@suse.de>
19355 * tree-scalar-evolution.c (scev_info_hasher): Remove.
19356 (struct instantiate_cache_entry): New type.
19357 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
19358 (struct instantiate_cache_type): New type.
19359 (set_instantiated_value, get_instantiated_value): Remove.
19360 (get_instantiated_value_entry): New function.
19361 (instantiate_scev_name): Use the new cache and adjust.
19362 (instantiate_scev_poly): Adjust.
19363 (instantiate_scev_binary): Likewise.
19364 (instantiate_array_ref): Likewise.
19365 (instantiate_scev_convert): Likewise.
19366 (instantiate_scev_not): Likewise.
19367 (instantiate_scev_3): Likewise.
19368 (instantiate_scev_2): Likewise.
19369 (instantiate_scev_r): Likewise.
19370 (instantiate_scev): Likewise.
19371 (resolve_mixers): Likewise.
19373 2013-05-01 Vladimir Makarov <vmakarov@redhat.com>
19376 * lra-constraints.c (best_small_class_operands_num): Remove.
19377 (process_alt_operands): Remove small_class_operands_num. Take
19378 small classes operands into losers and only if the operand is not
19379 matched. Modify debugging output.
19380 (curr_insn_transform): Remove best_small_class_operands_num.
19383 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
19385 * config/aarch64/aarch64-builtins.c
19386 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
19387 * config/aarch64/aarch64-simd-builtins.def
19388 (reduc_splus_): Add new modes.
19389 (reduc_uplus_): New.
19390 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
19391 (reduc_uplus_v4sf): Likewise.
19392 (reduc_splus_v4sf): Likewise.
19393 (aarch64_addv<mode>): Likewise.
19394 (reduc_uplus_<mode>): Likewise.
19395 (reduc_splus_<mode>): Likewise.
19396 (aarch64_addvv2di): Likewise.
19397 (reduc_uplus_v2di): Likewise.
19398 (reduc_splus_v2di): Likewise.
19399 (aarch64_addvv2si): Likewise.
19400 (reduc_uplus_v2si): Likewise.
19401 (reduc_splus_v2si): Likewise.
19402 (reduc_<sur>plus_<mode>): New.
19403 (reduc_<sur>plus_v2di): Likewise.
19404 (reduc_<sur>plus_v2si): Likewise.
19405 (reduc_<sur>plus_v4sf): Likewise.
19406 (aarch64_addpv4sf): Likewise.
19407 * config/aarch64/arm_neon.h
19408 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
19409 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
19410 add UNSPEC_SADDV, UNSPEC_UADDV.
19412 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
19414 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
19416 * config/aarch64/arm_neon.h
19417 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
19419 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
19421 * config/aarch64/aarch64-builtins
19422 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
19424 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
19426 * config/aarch64/aarch64-simd-builtins.def
19427 (reduc_smax_): New.
19428 (reduc_smin_): Likewise.
19429 (reduc_umax_): Likewise.
19430 (reduc_umin_): Likewise.
19431 (reduc_smax_nan_): Likewise.
19432 (reduc_smin_nan_): Likewise.
19435 (smax): Update for V2SF, V4SF and V2DF modes.
19438 (smin_nan): Likewise.
19439 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
19440 (<su><maxmin><mode>3): ...This, refactor.
19441 (s<maxmin><mode>3): New.
19442 (<maxmin_uns><mode>3): Likewise.
19443 (reduc_<maxmin_uns>_<mode>): Refactor.
19444 (reduc_<maxmin_uns>_v4sf): Likewise.
19445 (reduc_<maxmin_uns>_v2si): Likewise.
19446 (aarch64_<fmaxmin><mode>: Remove.
19447 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
19449 (vmin<q>_f<32,64>): Likewise.
19450 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
19452 (su): Add mappings for smax, smin, umax, umin.
19454 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
19455 (FMAXMIN): Rename as...
19456 (FMAXMIN_UNS): ...This.
19458 (fmaxminv): Likewise.
19459 (fmaxmin): Likewise.
19461 (maxmin_uns_op): Likewise.
19463 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
19465 * config/aarch64/arm_neon.h
19466 (vac<ge, gt><sd>_f<32, 64>): Rename to...
19467 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
19468 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
19470 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
19472 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
19473 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
19475 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
19477 * config/aarch64/aarch64-simd.md
19478 (vcond<mode>_internal): Handle special cases for constant masks.
19479 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
19480 (vcondu<mode><mode>): Likewise.
19481 (vcond<v_cmp_result><mode>): New.
19483 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
19485 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
19486 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
19487 * config/aarch64/aarch64-simd-builtins.def
19488 (cmeq): Update to BUILTIN_VALLDI.
19493 * config/aarch64/arm_neon.h
19494 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
19495 to builtins or C as appropriate.
19497 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
19499 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
19501 (cmhi): Rename to...
19503 * config/aarch64/aarch64-simd.md
19504 (simd_mode): Add SF.
19505 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
19506 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
19507 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
19508 (cstore<mode>_neg): ...This.
19509 * config/aarch64/iterators.md
19511 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
19512 (COMPARISONS): New.
19513 (UCOMPARISONS): Likewise.
19514 (optab): Add missing comparisons.
19520 (VCMP_S): Likewise.
19521 (VCMP_U): Likewise.
19522 (V_cmp_result): Add DF, SF modes.
19523 (v_cmp_result): Likewise.
19525 (vmtype): Likewise.
19526 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
19528 2013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
19530 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
19531 define_insn to define_insn_and_split.
19532 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
19533 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
19534 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
19535 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
19536 (thumb2_negscc): Likewise.
19538 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
19540 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
19542 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
19544 * config/arm/thumb2.md: Remove trailing whitespaces.
19546 2013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
19548 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
19549 Use gen_int_mode rather than GEN_INT.
19551 2013-04-30 H.J. Lu <hongjiu.lu@intel.com>
19553 * value-prof.c (stream_in_histogram_value): Remove the strayed
19556 2013-04-30 Richard Biener <rguenther@suse.de>
19558 PR middle-end/57122
19559 * cfghooks.c (split_edge): Properly check for the loop latch edge.
19561 2013-04-30 Richard Biener <rguenther@suse.de>
19563 PR middle-end/57107
19564 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
19566 2013-04-30 Andrey Belevantsev <abel@ispras.ru>
19568 PR rtl-optimization/56957
19569 PR rtl-optimization/57105
19570 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
19571 variable. Use just INSN_UID for determining whether an insn
19572 should be only disconnected from the insn stream.
19573 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
19575 2013-04-30 Jakub Jelinek <jakub@redhat.com>
19577 PR tree-optimization/57104
19578 * tsan.c (instrument_expr): Don't instrument accesses to
19579 DECL_HARD_REGISTER VAR_DECLs.
19581 2013-04-30 Richard Biener <rguenther@suse.de>
19583 * function.h (loops_for_fn): New inline function.
19584 (set_loops_for_fn): Likewise.
19585 * cfgloop.h (place_new_loop): Add struct function parameter.
19586 (get_loop): Likewise.
19587 (get_loops): Likewise.
19588 (number_of_loops): Likewise.
19589 (fel_next): Adjust.
19590 (fel_init): Likewise.
19591 * cfg.c (get_loop_copy): Adjust.
19592 * cfgloop.c (flow_loops_dump): Likewise.
19593 (record_loop_exits): Likewise.
19594 (verify_loop_structure): Likewise.
19595 * cfgloopanal.c (mark_irreducible_loops): Likewise.
19596 (estimate_reg_pressure_cost): Likewise.
19597 (mark_loop_exit_edges): Likewise.
19598 * cfgloopmanip.c (place_new_loop): Likewise.
19599 (add_loop): Likewise.
19600 (duplicate_loop): Likewise.
19601 * graph.c (draw_cfg_nodes): Likewise.
19602 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
19603 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
19604 (extract_affine_chrec): Likewise.
19605 (build_scop_iteration_domain): Likewise.
19606 * graphite.c (graphite_initialize): Likewise.
19607 * ira-build.c (create_loop_tree_nodes): Likewise.
19608 (more_one_region_p): Likewise.
19609 (rebuild_regno_allocno_maps): Likewise.
19610 (mark_loops_for_removal): Likewise.
19611 (mark_all_loops_for_removal): Likewise.
19612 (remove_unnecessary_regions): Likewise.
19613 (ira_build): Likewise.
19614 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
19615 * loop-init.c (fix_loop_structure): Likewise.
19616 (gate_rtl_move_loop_invariants): Likewise.
19617 (gate_rtl_unswitch): Likewise.
19618 (gate_rtl_unroll_and_peel_loops): Likewise.
19619 (rtl_doloop): Likewise.
19620 * lto-streamer-in.c (input_cfg): Likewise.
19621 * lto-streamer-out.c (output_cfg): Likewise.
19622 * modulo-sched.c (sms_schedule): Likewise.
19623 * predict.c (tree_estimate_probability): Likewise.
19624 (tree_estimate_probability_driver): Likewise.
19625 (estimate_loops): Likewise.
19626 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
19627 (move_sese_region_to_fn): Likewise.
19628 (debug_loop_num): Likewise.
19629 * tree-chrec.c (chrec_evaluate): Likewise.
19630 (hide_evolution_in_other_loops_than_loop): Likewise.
19631 (chrec_component_in_loop_num): Likewise.
19632 (reset_evolution_in_loop): Likewise.
19633 (evolution_function_is_invariant_rec_p): Likewise.
19634 * tree-if-conv.c (main_tree_if_conversion): Likewise.
19635 * tree-inline.c (copy_loops): Likewise.
19636 (copy_cfg_body): Likewise.
19637 (tree_function_versioning): Likewise.
19638 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
19639 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
19641 (add_to_evolution_1): Likewise.
19642 (scev_const_prop): Likewise.
19643 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
19644 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
19645 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
19646 (tree_ssa_lim_initialize): Likewise.
19647 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
19648 (verify_loop_closed_ssa): Likewise.
19649 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
19650 (tree_ssa_loop_im): Likewise.
19651 (tree_ssa_loop_unswitch): Likewise.
19652 (tree_vectorize): Likewise.
19653 (check_data_deps): Likewise.
19654 (tree_ssa_loop_ivcanon): Likewise.
19655 (tree_ssa_loop_bounds): Likewise.
19656 (tree_complete_unroll): Likewise.
19657 (tree_complete_unroll_inner): Likewise.
19658 (tree_parallelize_loops): Likewise.
19659 (tree_ssa_loop_prefetch): Likewise.
19660 (tree_ssa_loop_ivopts): Likewise.
19661 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
19662 * tree-vectorizer.c (vectorize_loops): Likewise.
19664 2013-04-29 Mike Frysinger <vapier@gentoo.org>
19666 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
19667 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
19668 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
19669 with EABI_LINK_SPEC.
19671 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
19674 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
19677 2013-04-29 Vladimir Makarov <vmakarov@redhat.com>
19680 * lra-constraints.c (process_alt_operands): Discourage a bit more
19681 using memory for pseudos. Print cost dump for alternatives.
19682 Modify cost values for conflicts with early clobbers.
19683 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
19685 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
19688 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
19690 2013-04-29 Ian Bolton <ian.bolton@arm.com>
19692 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
19693 from/to S register.
19694 (movdi_aarch64): Support LDR/STR from/to D register.
19696 2013-04-29 Ian Bolton <ian.bolton@arm.com>
19698 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
19699 or higher by default.
19701 2013-04-29 Richard Biener <rguenther@suse.de>
19703 PR middle-end/57075
19704 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
19705 even if not adding abnormal edges for calls that can make
19708 2013-04-29 Richard Biener <rguenther@suse.de>
19710 PR middle-end/57103
19711 * tree-cfg.c (move_stmt_op): Fix condition under which to update
19713 (move_stmt_r): Remove redundant checking.
19715 2013-04-29 Teresa Johnson <tejohnson@google.com>
19718 * basic-block.h (apply_scale): New function.
19719 (apply_probability): Use apply_scale.
19720 * gimple-streamer-in.c (input_bb): Ditto.
19721 * lto-streamer-in.c (input_cfg): Ditto.
19722 * lto-cgraph.c (merge_profile_summaries): Ditto.
19723 * tree-optimize.c (execute_fixup_cfg): Ditto.
19724 * tree-inline.c (copy_bb): Update comment to use apply_scale.
19725 (copy_edges_for_bb): Ditto.
19726 (copy_cfg_body): Ditto.
19728 2013-04-29 Tom de Vries <tom@codesourcery.com>
19730 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
19731 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
19732 (tail_merge_optimize): Handle current_loops == NULL.
19734 2013-04-26 Jeff Law <law@redhat.com>
19736 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
19737 (simplify_cond_using_ranges): Generalize code to simplify
19738 COND_EXPRs where one argument is a constant and the other
19739 is an SSA_NAME created by an integral type conversion.
19741 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19743 * config/arm/arm.md (store_minmaxsi): Use only when
19744 optimize_insn_for_size_p.
19746 2013-04-29 Christian Bruel <christian.bruel@st.com>
19749 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
19751 2013-04-29 Richard Biener <rguenther@suse.de>
19753 PR middle-end/57089
19754 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
19755 loop tree make sure to schedule a fixup for the child as well.
19756 (expand_omp_for_generic): Properly add loops.
19757 (expand_omp_for_static_nochunk): Likewise.
19758 (expand_omp_for_static_chunk): Likewise.
19759 (expand_omp_for): For the degenerate case fixup loops.
19760 (expand_omp_sections): Fix default bb placement in loops.
19761 (expand_omp_atomic_pipeline): Properly add loops.
19763 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19765 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
19767 2013-04-29 Tom de Vries <tom@codesourcery.com>
19769 * tree-ssa-tail-merge.c: Update header comment.
19771 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
19773 * config/aarch64/arm_neon.h
19774 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
19775 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
19776 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
19777 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
19778 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
19779 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
19780 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
19781 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
19783 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
19785 * config/aarch64/aarch64-simd.md
19786 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
19787 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
19788 fix_trunc, fixuns_trunc.
19789 (ftrunc<VDQF:mode>2): New.
19790 * config/aarch64/iterators.md (optab): Add fix, fixuns.
19791 (fix_trunc_optab): New.
19793 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
19795 * config/aarch64/aarch64-builtins.c
19796 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
19797 iceilf, lround, iroundf.
19799 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
19802 * config/i386/i386.h (enum ix86_tune_indices)
19803 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
19804 New, split from X86_TUNE_INTER_UNIT_MOVES.
19805 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
19806 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
19807 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
19808 (TARGET_INTER_UNIT_MOVES): Remove.
19809 * config/i386/i386.c (initial_ix86_tune_features): Update.
19810 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
19811 (ix86_expand_convert_uns_didf_sse): Use
19812 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
19813 (ix86_expand_vector_init_one_nonzero): Ditto.
19814 (ix86_expand_vector_init_interleave): Ditto.
19815 (inline_secondary_memory_needed): Return true for moves from SSE class
19816 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
19817 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
19818 * config/i386/constraints.md (Yi, Ym): Depend on
19819 TARGET_INTER_UNIT_MOVES_TO_VEC.
19820 (Yj, Yn): New constraints.
19821 * config/i386/i386.md (*movdi_internal): Change constraints of
19822 operand 1 from Yi to Yj and from Ym to Yn.
19823 (*movsi_internal): Ditto.
19824 (*movdf_internal): Ditto.
19825 (*movsf_internal): Ditto.
19826 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
19827 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
19828 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
19829 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
19830 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
19831 * config/i386/sse.md (movdi_to_sse): Ditto.
19832 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
19833 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
19834 TARGET_INTER_UNIT_MOVES.
19835 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
19836 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
19837 instead of TARGET_INTER_UNIT_MOVES.
19838 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
19839 operand 1 from Yi to Yj and from Ym to Yn.
19841 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
19843 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
19844 (float_truncate_hi_): Likewise.
19845 (float_extend_lo_): Likewise.
19846 (float_truncate_lo_): Likewise.
19847 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
19848 (aarch64_float_extend_lo_v2df): Likewise.
19849 (vec_unpacks_hi_v4sf): Likewise.
19850 (aarch64_float_truncate_lo_v2sf): Likewise.
19851 (aarch64_float_truncate_hi_v4sf): Likewise.
19852 (vec_pack_trunc_v2df): Likewise.
19853 (vec_pack_trunc_df): Likewise.
19855 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
19857 * config/aarch64/aarch64-builtins.c
19858 (aarch64_fold_builtin): Fold float conversions.
19859 * config/aarch64/aarch64-simd-builtins.def
19860 (floatv2si, floatv4si, floatv2di): New.
19861 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
19862 * config/aarch64/aarch64-simd.md
19863 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
19864 * config/aarch64/iterators.md (FLOATUORS): New.
19865 (optab): Add float, floatuns.
19866 (su_optab): Likewise.
19868 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
19870 * config/aarch64/aarch64-builtins.c
19871 (aarch64_builtin_vectorized_function): Use new names for
19873 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
19874 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
19875 (fcvtzu): Split as...
19876 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
19877 (fcvtas): Split as...
19878 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
19879 (fcvtau): Split as...
19880 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
19881 (fcvtps): Split as...
19882 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
19883 (fcvtpu): Split as...
19884 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
19885 (fcvtms): Split as...
19886 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
19887 (fcvtmu): Split as...
19888 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
19889 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
19890 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
19891 (lfrintnusf, lfrintnudf): Likewise.
19892 * config/aarch64/aarch64-simd.md
19893 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
19895 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
19896 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
19897 (fcvt_pattern): Likewise.
19899 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
19901 * config/aarch64/aarch64-simd.md
19902 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
19903 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
19905 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
19907 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
19908 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
19909 (vrnd<a,m,n,p>_f32): Implement using builtins.
19910 (vrnd<i,x><q>_f<32, 64>): New.
19912 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
19914 * config/aarch64/aarch64-builtins.c
19915 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
19916 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
19917 (frintz): Rename to...
19919 (frintp): Rename to...
19921 (frintm): Rename to...
19923 (frinti): Rename to...
19924 (nearbyint): ...this.
19925 (frintx): Rename to...
19927 (frinta): Rename to...
19929 * config/aarch64/aarch64-simd.md
19930 (aarch64_frint<frint_suffix><mode>): Delete.
19931 (<frint_pattern><mode>2): Convert to insn.
19932 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
19933 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
19934 (frint_pattern): Likewise.
19935 (frint_suffix): Likewise.
19937 2013-04-29 Richard Biener <rguenther@suse.de>
19939 PR tree-optimization/57081
19940 * loop-init.c: Include tree-flow.h.
19941 (loop_optimizer_finalize): Free number of iteration estimates.
19942 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
19944 2013-04-29 Jakub Jelinek <jakub@redhat.com>
19946 PR tree-optimization/57083
19947 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
19948 non-singleton shift count range, zero extend low_bound for uns case.
19950 * config/i386/predicates.md (general_vector_operand): New predicate.
19951 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
19952 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
19953 if they aren't nonimmediate operands. If their original values
19954 satisfy const_vector_equal_evenodd_p, don't shift them.
19955 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
19956 predicates. For the SSE4.1 case force operands[{1,2}] into registers
19957 if not nonimmediate_operand.
19958 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
19959 instead of register_operand.
19960 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
19962 2013-04-28 Eric Botcazou <ebotcazou@adacore.com>
19964 * stor-layout.c (finalize_size_functions): Allocate a structure and
19965 reset cfun before dumping the functions.
19967 2013-04-27 Jakub Jelinek <jakub@redhat.com>
19969 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
19972 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
19973 use xop_pmacsdqh if uns_p.
19974 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
19975 the immediate rotate count.
19977 2013-04-26 Vladimir Makarov <vmakarov@redhat.com>
19979 * rtl.h (struct rtx_def): Add comment for field jump.
19980 (LRA_SUBREG_P): New macro.
19981 * recog.c (register_operand): Check LRA_SUBREG_P.
19982 * lra.c (lra): Add note at the end of RTL code. Align non-empty
19984 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
19985 (lra_final_code_change): Skip subreg change for operators.
19986 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
19987 if there are no operand changes.
19988 * lra-constraints.c (curr_insn_set): New.
19989 (match_reload): Set LRA_SUBREG_P.
19990 (emit_spill_move): Ditto.
19991 (check_and_process_move): Use curr_insn_set. Process only single
19992 set insns. Don't initialize sec_mem_p and change_p.
19993 (simplify_operand_subreg): Use LRA_SUBREG_P.
19994 (reg_in_class_p): New function.
19995 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
19996 of #ifdef. Add code to remove cycling.
19997 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
19998 non-null disp. Reload inner instead of disp when base and index
19999 are null. Try to put lo_sum into register.
20000 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
20001 (check_and_process_move): Move code for move cost check to
20002 simple_move_p. Remove equiv_substitution.
20003 (simple_move_p): New function.
20004 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
20005 curr_insn_set. Call check_and_process_move only for single set
20006 insns. Use the new function. Move call of check_and_process_move
20007 after operand equiv substitution and address process.
20009 2013-04-26 Jakub Jelinek <jakub@redhat.com>
20012 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
20013 with nonlocal goto receivers or returns twice calls, ignore
20014 unininitialized values from abnormal edges to nl goto receiver
20015 or returns twice call.
20017 2013-04-26 Jakub Jelinek <jakub@redhat.com>
20019 PR tree-optimization/57051
20020 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
20021 and VEC_RSHIFT_EXPR if shift count is a multiple of element
20024 2013-04-26 Richard Biener <rguenther@suse.de>
20026 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
20027 (expand_omp_taskreg): Likewise. Mark loops for fixup.
20028 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
20029 (fixup_loop_arrays_after_move): New function.
20030 (move_sese_region_to_fn): Properly outline the loop tree parts
20031 of the SESE region.
20033 2013-04-26 Uros Bizjak <ubizjak@gmail.com>
20035 * config/i386/i386.md (type, unit): Fix long lines.
20037 2013-04-26 Richard Biener <rguenther@suse.de>
20039 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
20040 (lto-streamer-out.o): Likewise.
20041 * cfgloop.c (init_loops_structure): Export, add struct function
20042 argument and adjust.
20043 (flow_loops_find): Adjust.
20044 * cfgloop.h (enum loop_estimation): Add EST_LAST.
20045 (init_loops_structure): Declare.
20046 * lto-streamer-in.c: Include cfgloop.h.
20047 (input_cfg): Input the loop tree.
20048 * lto-streamer-out.c: Include cfgloop.h.
20049 (output_cfg): Output the loop tree.
20050 (output_struct_function_base): Do not drop PROP_loops.
20052 2013-03-26 Richard Biener <rguenther@suse.de>
20054 * tree-cfg.c (execute_build_cfg): Build the loop tree.
20055 (pass_build_cfg): Provide PROP_loops.
20056 (move_sese_region_to_fn): Remove loops that are outlined into fn
20058 * tree-inline.c: Include cfgloop.h.
20059 (initialize_cfun): Do not drop PROP_loops.
20060 (copy_loops): New function.
20061 (copy_cfg_body): Copy loop structure.
20062 (tree_function_versioning): Initialize destination loop tree.
20063 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
20064 (pass_parallelize_loops): Do IL verification.
20065 * loop-init.c (loop_optimizer_init): Fixup loops if required.
20066 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
20067 the CFG make sure we fixup loops as well.
20068 * tree-ssa-tail-merge.c: Include cfgloop.h.
20069 (replace_block_by): When merging loop latches mark loops for fixup.
20070 * lto-streamer-out.c (output_struct_function_base): Drop
20071 PROP_loops for now.
20072 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
20073 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
20074 * ipa-split.c: Include cfgloop.h.
20075 (split_function): Add the new return block to the loop tree root.
20076 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
20077 whether we have removed the forwarder block.
20078 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
20079 * cfgloop.h (place_new_loop): Declare.
20080 * cfgloopmanip.c (place_new_loop): Export.
20081 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
20082 (tree-switch-conversion.o): Likewise.
20083 (tree-complex.o): Likewise.
20084 (tree-inline.o): Likewise.
20085 (tree-ssa-tailmerge.o): Likewise.
20086 (ipa-split.o): Likewise.
20087 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
20088 (tree-ssa-copy.o): Likewise.
20089 * tree-switch-conversion.c: Include cfgloop.h
20090 (process_switch): If we emit a bit-test cascade, schedule loops
20092 * tree-complex.c: Include cfgloop.h.
20093 (expand_complex_div_wide): Properly add new basic-blocks to loops.
20094 * asan.c: Include cfgloop.h.
20095 (create_cond_insert_point): Properly add new basic-blocks to
20096 loops, schedule loop fixup.
20097 * cfgloop.c (verify_loop_structure): Check that looks are not
20099 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
20101 (expand_omp_for_generic): Likewise.
20102 (expand_omp_sections): Likewise.
20103 (expand_omp_atomic_pipeline): Schedule loops for fixup.
20104 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
20105 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
20106 is initialized, not when loops are present.
20107 * tree-parloops.c (parallelize_loops): Remove checking here.
20108 * passes.c (init_optimization_passes): Schedule a copy-propagation
20109 pass before complete unrolling of inner loops.
20111 2013-04-26 Jakub Jelinek <jakub@redhat.com>
20113 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
20114 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
20115 (colorize_init): Add argument to _WIN32 version.
20116 * toplev.c: Include diagnostic-color.h.
20117 (process_options): Default to -fdiagnostics-color=auto if
20118 GCC_COLORS env var is in the environment.
20119 * common.opt (fdiagnostics-color=): Add Var and Init.
20120 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
20121 env var is in the environment, the default is auto rather than never.
20123 * diagnostic.h (file_name_as_prefix): Add context argument.
20124 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
20125 the string as locus.
20126 * langhooks.c (lhd_print_error_function): Adjust caller.
20128 2013-04-25 Lawrence Crowl <crowl@google.com>
20130 * var-tracking.c (shared_hash_def::htab):
20131 Change type to hash_table. Update dependent calls and types.
20133 2013-04-25 Lawrence Crowl <crowl@google.com>
20135 * Makefile.in: Update as needed below.
20137 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
20138 Move declaration to after the type's method definitons.
20140 * attribs.c (htab_t scoped_attributes::attribute_hash):
20141 Change type to hash_table. Update dependent calls and types.
20143 * bitmap.c (htab_t bitmap_desc_hash):
20144 Change type to hash_table. Update dependent calls and types.
20146 * cselib.c (htab_t cselib_hash_table):
20147 Change type to hash_table. Update dependent calls and types.
20149 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
20150 (hash_string_slot_node): Move implementation into lto-streamer.h
20151 struct string_slot_hasher.
20152 (eq_string_slot_node): Likewise.
20154 * data-streamer-out.c: Update output_block::string_hash_table
20155 dependent calls and types.
20157 * dwarf2cfi.c (htab_t trace_index):
20158 Change type to hash_table. Update dependent calls and types.
20160 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
20161 Change type to hash_table. Update dependent calls and types.
20162 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
20163 (htab_t optimize_external_refs::map): Likewise.
20164 (htab_t output_comp_unit::extern_map): Likewise.
20165 (htab_t output_comdat_type_unit::extern_map): Likewise.
20166 (htab_t output_macinfo::macinfo_htab): Likewise.
20167 (htab_t optimize_location_lists::htab): Likewise.
20168 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
20170 * except.c (htab_t ehspec_hash_type):
20171 Change type to hash_table. Update dependent calls and types.
20172 (assign_filter_values::ttypes): Likewise.
20173 (assign_filter_values::ehspec): Likewise.
20174 (sjlj_assign_call_site_values::ar_hash): Likewise.
20175 (convert_to_eh_region_ranges::ar_hash): Likewise.
20177 * gcse.c (htab_t pre_ldst_table):
20178 Change type to hash_table. Update dependent calls and types.
20180 * ggc-common.c (htab_t saving_htab):
20181 Change type to hash_table. Update dependent calls and types.
20182 (htab_t loc_hash): Likewise.
20183 (htab_t ptr_hash): Likewise.
20184 (call_count): Rename ggc_call_count.
20185 (call_alloc): Rename ggc_call_alloc.
20186 (loc_descriptor): Rename make_loc_descriptor.
20187 (add_statistics): Rename ggc_add_statistics.
20189 * ggc-common.c (saving_htab):
20190 Change type to hash_table. Update dependent calls and types.
20192 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
20193 (push_gimplify_context): Likewise.
20194 (pop_gimplify_context): Likewise.
20195 (struct gimple_temp_hash_elt): Added.
20196 (struct gimplify_hasher): Likewise.
20197 (struct gimplify_ctx.temp_htab):
20198 Change type to hash_table. Update dependent calls and types.
20200 * gimple-fold.c: Include gimplify-ctx.h.
20202 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
20203 Change type to hash_table. Update dependent calls and types.
20204 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
20205 avoid potential global name collision.
20207 * gimplify.c: Include gimplify-ctx.h.
20208 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
20209 (htab_t gimplify_ctx::temp_htab):
20210 Update dependent calls and types for new type hash_table.
20211 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
20212 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
20214 * gimplify-ctx.h: New.
20215 (struct gimple_temp_hash_elt): Move from gimplify.c.
20216 (class gimplify_hasher): New.
20217 (struct gimplify_ctx): Move from gimple.h.
20218 (htab_t gimplify_ctx::temp_htab):
20219 Change type to hash_table. Update dependent calls and types.
20221 * graphite-clast-to-gimple.c: Include graphite-htab.h.
20222 (htab_t ivs_params::newivs_index):
20223 Change type to hash_table. Update dependent calls and types.
20224 (htab_t ivs_params::params_index): Likewise.
20225 (htab_t print_generated_program::params_index): Likewise.
20226 (htab_t gloog::newivs_index): Likewise.
20227 (htab_t gloog::params_index): Likewise.
20229 * graphite.c: Include graphite-htab.h.
20230 4htab_t graphite_transform_loops::bb_pbb_mapping):
20231 Change type to hash_table. Update dependent calls and types.
20233 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
20234 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
20235 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
20237 * graphite-dependences.c: Include graphite-htab.h.
20238 (loop_is_parallel_p): Change hash table type of parameter.
20240 * graphite-htab.h: New.
20241 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
20242 (extern find_pbb_via_hash): Move from graphite-poly.h.
20243 (extern loop_is_parallel_p): Move from graphite-poly.h.
20244 (extern get_loop_body_pbbs): Move from graphite-poly.h.
20246 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
20247 (extern loop_is_parallel_p): Move to graphite-htab.h.
20248 (extern get_loop_body_pbbs): Move to graphite-htab.h.
20250 * haifa-sched.c (htab_t delay_htab):
20251 Change type to hash_table. Update dependent calls and types.
20252 (htab_t delay_htab_i2): Likewise.
20254 * ira-color.c (htab_t allocno_hard_regs_htab):
20255 Change type to hash_table. Update dependent calls and types.
20257 * ira-costs.c (htab_t cost_classes_htab):
20258 Change type to hash_table. Update dependent calls and types.
20260 * loop-invariant.c (htab_t merge_identical_invariants::eq):
20261 Change type to hash_table. Update dependent calls and types.
20263 * loop-iv.c (htab_t bivs):
20264 Change type to hash_table. Update dependent calls and types.
20266 * loop-unroll.c (htab_t opt_info::insns_to_split):
20267 Change type to hash_table. Update dependent calls and types.
20268 (htab_t opt_info::insns_with_var_to_expand): Likewise.
20270 * lto-streamer.h (struct string_slot): Move from data-streamer.h
20271 (struct string_slot_hasher): New.
20272 (htab_t output_block::string_hash_table):
20273 Change type to hash_table. Update dependent calls and types.
20275 * lto-streamer-in.c (freeing_string_slot_hasher): New.
20276 (htab_t file_name_hash_table):
20277 Change type to hash_table. Update dependent calls and types.
20279 * lto-streamer-out.c: Update output_block::string_hash_table dependent
20282 * lto-streamer.c (htab_t tree_htab):
20283 Change type to hash_table. Update dependent calls and types.
20285 * omp-low.c: Include gimplify-ctx.h.
20287 * passes.c (htab_t name_to_pass_map):
20288 Change type to hash_table. Update dependent calls and types.
20289 (pass_traverse): Rename to passes_pass_traverse.
20291 * plugin.c (htab_t event_tab):
20292 Change type to hash_table. Update dependent calls and types.
20294 * postreload-gcse.c (htab_t expr_table):
20295 Change type to hash_table. Update dependent calls and types.
20296 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
20298 * sese.c (debug_rename_map_1): Make extern.
20299 (htab_t copy_bb_and_scalar_dependences::rename_map):
20300 Change type to hash_table. Update dependent calls and types.
20302 * sese.h (extern debug_rename_map): Move to .c file.
20304 * store-motion.c (htab_t store_motion_mems_table):
20305 Change type to hash_table. Update dependent calls and types.
20307 * trans-mem.c (htab_t tm_new_mem_hash):
20308 Change type to hash_table. Update dependent calls and types.
20310 * tree-browser.c (htab_t TB_up_ht):
20311 Change type to hash_table. Update dependent calls and types.
20313 * tree-cfg.c (htab_t discriminator_per_locus):
20314 Change type to hash_table. Update dependent calls and types.
20316 * tree-complex.c: Include tree-hasher.h
20317 (htab_t complex_variable_components):
20318 Change type to hash_table. Update dependent calls and types.
20320 * tree-eh.c (htab_t finally_tree):
20321 Change type to hash_table. Update dependent calls and types.
20323 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
20324 struct int_tree_hasher.
20325 (extern int_tree_map_eq): Likewise.
20326 (uid_decl_map_hash): Removed.
20327 (extern decl_tree_map_eq): Likewise.
20329 * tree-hasher.h: New.
20330 (struct int_tree_hasher): New.
20331 (typedef int_tree_htab_type): New.
20333 * tree-inline.c: Include gimplify-ctx.h.
20335 * tree-mudflap.c: Include gimplify-ctx.h.
20337 * tree-parloops.c: Include tree-hasher.h.
20338 (htab_t eliminate_local_variables_stmt::decl_address):
20339 Change type to hash_table. Update dependent calls and types.
20340 (htab_t separate_decls_in_region::decl_copies): Likewise.
20342 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
20343 Change type to hash_table. Update dependent calls and types.
20345 * tree-sra.c (candidates):
20346 Change type to hash_table. Update dependent calls and types.
20348 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
20350 (int_tree_map_hash): Likewise.
20352 * tree-ssa-dom.c (htab_t avail_exprs):
20353 Change type to hash_table. Update dependent calls and types.
20355 * tree-ssa-live.c (var_map_base_init::tree_to_index):
20356 Change type to hash_table. Update dependent calls and types.
20358 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
20359 Change type to hash_table. Update dependent calls and types.
20361 * tree-ssa-phiopt.c (seen_ssa_names):
20362 Change type to hash_table. Update dependent calls and types.
20364 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
20365 Change type to hash_table. Update dependent calls and types.
20367 * tree-ssa-uncprop.c (equiv):
20368 Change type to hash_table. Update dependent calls and types.
20370 2013-04-25 Jakub Jelinek <jakub@redhat.com>
20372 PR rtl-optimization/57003
20373 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
20374 call note_stores with kill_clobbered_value callback again after
20375 killing regs_invalidated_by_call.
20377 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
20379 * config/aarch64/aarch64-simd.md
20380 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
20381 (aarch64_simd_bsl<mode>): Likewise.
20382 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
20384 2013-04-25 Marek Polacek <polacek@redhat.com>
20386 PR tree-optimization/57066
20387 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
20389 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
20391 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
20393 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
20395 * config/aarch64/aarch64-builtins.c
20396 (aarch64_fold_builtin): New.
20397 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
20398 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
20399 * config/aarch64/aarch64-simd-builtins.def (abs): New.
20400 * config/aarch64/arm_neon.h
20401 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
20403 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
20404 Tejas Belagod <tejas.belagod@arm.com>
20406 * config/aarch64/aarch64-builtins.c
20407 (aarch64_gimple_fold_builtin): New.
20408 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
20409 * config/aarch64/aarch64-simd-builtins.def (addv): New.
20410 * config/aarch64/aarch64-simd.md (addpv4sf): New.
20411 (addvv4sf): Update.
20412 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
20414 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
20416 * config/aarch64/aarch64.md
20417 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
20419 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
20421 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
20422 (*ngcsi_uxtw): New pattern.
20424 2013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20425 Julian Brown <julian@codesourcery.com>
20427 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
20428 (TB_DREG): Add T_V4HF.
20429 (v4hf_UP): New macro.
20430 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
20431 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
20432 Handle initialisation of V4HF. Adjust initialisation of reinterpret
20434 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
20435 (arm_vector_mode_supported_p): Handle V4HF.
20436 (arm_mangle_map): Handle V4HFmode.
20437 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
20438 * config/arm/arm_neon_builtins.def: Add entries for
20439 vcvtv4hfv4sf, vcvtv4sfv4hf.
20440 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
20441 (neon_vcvtv4hfv4sf): Likewise.
20442 * config/arm/neon-gen.ml: Handle half-precision floating point
20444 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
20445 * config/arm/arm_neon.h: Regenerate.
20446 * config/arm/neon.ml (type elts): Add F16.
20447 (type vectype): Add T_float16x4, T_floatHF.
20448 (type vecmode): Add V4HF.
20449 (type features): Add Requires_FP_bit feature.
20450 (elt_width): Handle F16.
20451 (elt_class): Likewise.
20452 (elt_of_class_width): Likewise.
20453 (mode_of_elt): Refactor.
20454 (type_for_elt): Handle F16, fix error messages.
20455 (vectype_size): Handle T_float16x4.
20456 (vcvt_sh): New function.
20457 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
20458 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
20459 (string_of_mode): Handle V4HF.
20460 * doc/arm-neon-intrinsics.texi: Regenerate.
20462 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
20464 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
20465 format specifier in 'X' case.
20467 2013-04-25 Alan Modra <amodra@gmail.com>
20470 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
20471 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
20472 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
20473 Repeat for many other rotate/shift and mask patterns using subregs.
20474 Name lshiftrt insns.
20475 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
20476 on WORDS_BIG_ENDIAN.
20478 2013-04-25 Alan Modra <amodra@gmail.com>
20480 * config.gcc: Support little-endian powerpc-linux targets.
20481 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
20482 (LINK_OS_LINUX_SPEC): Define.
20483 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
20484 Preserve MASK_LITTLE_ENDIAN.
20485 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
20486 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
20487 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
20488 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
20489 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
20490 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
20491 Correct fp word order for little-endian. Don't shift toc entries
20492 smaller than a word for little-endian.
20493 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
20494 (bswapdi2 splits): Correct low-part subreg for little-endian.
20495 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
20496 low/high where such is correct only for be.
20497 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
20498 little-endian for -mcall-aixdesc.
20500 2013-04-25 Alan Modra <amodra@gmail.com>
20502 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
20503 replace_equiv_address_nv.
20505 2013-04-25 Alan Modra <amodra@gmail.com>
20507 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
20509 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
20512 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
20513 * rtl.h (struct rtx_def): ...
20515 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
20517 PR rtl-optimizations/57046
20518 * lra-constraints (split_reg): Set up lra_risky_transformations_p
20519 for multi-reg splits.
20521 2013-04-24 H.J. Lu <hongjiu.lu@intel.com>
20523 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
20525 2013-04-24 Sterling Augustine <saugustine@google.com>
20527 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
20528 (comp_dir_string, debug_str_dwo_section): New.
20529 (DEBUG_STR_DWO_SECTION): Rename to ...
20530 (DEBUG_DWO_STR_SECTION): ... this.
20531 (DEBUG_NORM_STR_SECTION): Delete.
20532 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
20533 (DEBUG_STR_DWO_SECTION_FLAGS): New.
20534 (find_AT_string): Move most logic to ...
20535 (find_AT_string_in_table): ... here. New.
20536 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
20537 add_skeleton_AT_string. Delete logic.
20538 (output_skeleton_debug_sections): Remove call to
20539 add_top_level_skeleton_die_attrs.
20540 (add_comp_dir_attribute): Move logic to comp_dir_string.
20541 (dwarf2out_init): Initialize debug_str_dwo_section.
20542 (output_indirect_string): Call find_string_form.
20543 (output_indirect_strings): Rewrite.
20544 (prune_unused_types): Empty skeleton_debug_str_hash.
20545 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
20546 (dwarf2out_finish): Call output_indirect_strings.
20548 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
20550 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
20552 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
20554 * rtl.h (struct rtx_def): Add comment for field jump.
20555 (LRA_SUBREG_P): New macro.
20556 * recog.c (register_operand): Check LRA_SUBREG_P.
20557 * lra.c (lra): Add note at the end of RTL code. Align non-empty
20559 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
20560 (lra_final_code_change): Skip subreg change for operators.
20561 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
20562 if there are no operand changes.
20563 * lra-constraints.c (curr_insn_set): New.
20564 (match_reload): Set LRA_SUBREG_P.
20565 (emit_spill_move): Ditto.
20566 (check_and_process_move): Use curr_insn_set. Process only single
20567 set insns. Don't initialize sec_mem_p and change_p.
20568 (simplify_operand_subreg): Use LRA_SUBREG_P.
20569 (reg_in_class_p): New function.
20570 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
20571 of #ifdef. Add code to remove cycling.
20572 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
20573 non-null disp. Reload inner instead of disp when base and index
20574 are null. Try to put lo_sum into register.
20575 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
20576 (check_and_process_move): Move code for move cost check to
20577 simple_move_p. Remove equiv_substitution.
20578 (simple_move_p): New function.
20579 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
20580 curr_insn_set. Call check_and_process_move only for single set
20581 insns. Use the new function. Move call of check_and_process_move
20582 after operand equiv substitution and address process.
20584 2013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
20586 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
20587 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
20588 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
20590 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
20592 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
20594 2013-04-24 Marek Polacek <polacek@redhat.com>
20596 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
20597 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
20598 (select_loops_exit_conditions): Likewise.
20599 (number_of_iterations_for_all_loops): Likewise.
20600 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
20601 (scev_analysis): Likewise.
20603 2013-04-02 Catherine Moore <clm@codesourcery.com>
20604 Chao-ying Fu <fu@mips.com>
20606 * config/mips/micromips.md (jraddiusp): New pattern.
20607 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
20608 instruction if possible.
20610 2013-04-24 Alan Modra <amodra@gmail.com>
20612 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
20614 2013-04-24 Julian Brown <julian@codesourcery.com>
20615 Chung-Lin Tang <cltang@codesourcery.com>
20617 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
20618 dependency behavior in enumeration type DIE generation. Add TODO note
20619 to comments about future DW_FORM_sdata/udata re-work of related code.
20621 2013-04-23 Lawrence Crowl <crowl@google.com>
20623 * Makefile.in: Update as needed below.
20625 * hash-table.h (class hash_table):
20626 Correct many methods with parameter types compare_type to the correct
20627 value_type. (Correct code was unlikely to notice the change.)
20628 (hash_table::elements_with_deleted) New.
20629 (class hashtable::iterator): New.
20630 (hashtable::begin()): New.
20631 (hashtable::end()): New.
20632 (FOR_EACH_HASH_TABLE_ELEMENT): New.
20634 * statistics.c (statistics_hashes):
20635 Change type to hash_table. Update dependent calls and types.
20637 * tree-into-ssa.c (var_infos):
20638 Change type to hash_table. Update dependent calls and types.
20640 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
20641 Change type to hash_table. Update dependent calls and types.
20643 * tree-ssa-loop-im.c (struct mem_ref.refs):
20644 Change type to hash_table. Update dependent calls and types.
20646 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
20647 Change type to hash_table. Update dependent calls and types.
20649 * tree-ssa-sccvn.c (vn_tables_s::nary):
20650 Change type to hash_table. Update dependent calls and types.
20651 (vn_tables_s::phis): Likewise.
20652 (vn_tables_s::references): Likewise.
20654 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
20655 (vn_reference_eq): Update parameter and return types.
20657 * tree-ssa-structalias.c (pointer_equiv_class_table):
20658 Change type to hash_table. Update dependent calls and types.
20659 (location_equiv_class_table): Likewise.
20661 * tree-vect-data-refs.c: Consequential changes for making
20662 peeling a hash_table.
20664 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
20665 (destroy_loop_vec_info): Dependent hash_table update.
20667 * tree-vectorizer.h (peeling_htab):
20668 Change type to hash_table. Update dependent calls and types.
20670 2013-04-23 Shiva Chen <shiva0217@gmail.com>
20672 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
20673 to check the register content is equal or not.
20674 * lra-constraints.c (match_reload): Use lra_assign_reg_val
20675 to assign register content record.
20676 * lra-eliminations.c (update_reg_eliminate): Use
20677 lra_update_reg_val_offset to update register content offset.
20678 * lra-int.h (struct lra_reg): Add offset member.
20679 (lra_reg_val_equal_p): New static inline function.
20680 (lra_update_reg_val_offset): New static inline function.
20681 (lra_assign_reg_val): New static inline function.
20682 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
20683 to assign register content record.
20684 (initialize_lra_reg_info_element): Initial offset to zero.
20686 2013-04-23 Catherine Moore <clm@codesourcery.com>
20688 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
20689 operands. Record compression.
20691 2013-04-23 Xinliang David Li <davidxl@google.com>
20693 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
20695 2013-04-23 Richard Biener <rguenther@suse.de>
20697 PR middle-end/57036
20698 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
20699 parameter, only add abnormal goto edges from the copied body
20700 if the call could perform abnormal gotos.
20701 (copy_cfg_body): Adjust.
20703 2013-04-23 Sofiane Naci <sofiane.naci@arm.com>
20705 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
20707 2013-04-23 Andreas Schwab <schwab@linux-m68k.org>
20709 * coretypes.h (gimple_stmt_iterator): Add struct to make
20712 2013-04-23 Richard Biener <rguenther@suse.de>
20714 PR tree-optimization/57026
20715 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
20716 from SSA names occuring in abnormal PHI nodes.
20718 2013-04-22 Andi Kleen <ak@linux.intel.com>
20720 * lto/lto.c (print_lto_report_1): Fix LTO report names.
20722 2013-04-22 Andi Kleen <ak@linux.intel.com>
20724 * lto/lto.c (print_lto_report_1): Declare early.
20725 (read_cgraph_and_symbols): Call print_lto_report_1 early.
20727 2013-04-22 Andi Kleen <ak@linux.intel.com>
20729 * common.opt (-flto-report-wpa): Add.
20730 * doc/invoke.texi (-flto-report-wpa): Add.
20731 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
20734 2013-04-22 Xinliang David Li <davidxl@google.com>
20736 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
20737 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
20738 * Makefile.in: New dependency
20740 David Daney <ddaney.cavm@gmail.com>
20742 * configure.ac (gcc_cv_as_micromips_support): Use the
20743 --fatal-warnings option.
20744 * configure: Regenerate.
20746 2013-04-22 Marek Polacek <polacek@redhat.com>
20749 * tsan.c (instrument_expr): Don't instrument expression
20750 in case its size is zero.
20752 2013-04-22 Uros Bizjak <ubizjak@gmail.com>
20756 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
20758 * config/alpha/alpha.c (TARGET_LRA_P): New define.
20760 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
20762 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
20763 (gimple_stmt_iterator): New typedef.
20764 * gimple.h (gimple_stmt_iterator): Rename to...
20765 (gimple_stmt_iterator_d): ... This.
20766 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
20767 trees be valid for GIMPLE and GENERIC.
20768 (TARGET_GIMPLE_FOLD_BUILTIN): New.
20769 * gimple-fold.c (gimple_fold_call): Call target hook
20770 gimple_fold_builtin.
20771 * hooks.c (hook_bool_gsiptr_false): New.
20772 * hooks.h (hook_bool_gsiptr_false): New.
20773 * target.def (fold_stmt): New.
20774 * doc/tm.texi: Regenerate.
20776 2013-04-22 Vladimir Makarov <vmakarov@redhat.com>
20779 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
20780 a set sp if no stack realignment.
20782 2013-04-22 Nick Clifton <nickc@redhat.com>
20784 * config.gcc (tilegx-linux): Extend extra_objs rather than
20786 (tilepro-linux): Likewise.
20788 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
20790 * config/aarch64/aarch64-builtins.c
20792 (CF0, CF1, CF2, CF3, CF4, CF10): New.
20793 (VAR<1-12>): Add MAP parameter.
20794 (BUILTIN_*): Likewise.
20795 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
20796 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
20797 (aarch64_ushl_n<mode>): Likewise.
20798 (aarch64_sshr_n<mode>): Likewise.
20799 (aarch64_ushr_n<mode>): Likewise.
20800 (aarch64_<maxmin><mode>): Likewise.
20801 (aarch64_sqrt<mode>): Likewise.
20802 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
20803 (vshr<q>_n_*): Likewise.
20805 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
20807 * config/aarch64/aarch64-builtins.c
20808 (aarch64_simd_builtin_type_mode): Handle SF types.
20810 (BUILTIN_GPF): Define.
20811 (aarch64_init_simd_builtins): Handle SF types.
20812 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
20813 (frecps): Likewise.
20814 (frecpx): Likewise.
20815 * config/aarch64/aarch64-simd.md
20816 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
20817 (aarch64_frecpe<mode>): New.
20818 (aarch64_frecps<mode>): Likewise.
20819 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
20820 (v8type): Add frecp<esx>.
20821 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
20822 (aarch64_frecps<mode>): Likewise.
20823 * config/aarch64/iterators.md (FRECP): New.
20824 (frecp_suffix): Likewise.
20825 * config/aarch64/arm_neon.h
20826 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
20828 2013-04-22 Christian Bruel <christian.bruel@st.com>
20831 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
20832 (REG_CLASS_NAMES): Idem.
20833 (REG_CLASS_CONTENTS): Idem.
20834 (REGCLASS_HAS_FP_REG): Idem.
20835 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
20836 (sh_conditional_register_usage): Idem.
20838 2013-04-21 Jeff Law <law@redhat.com>
20840 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
20841 (ssa_forward_propagate_and_combine): Use it.
20843 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
20845 * lra.c: Update the flow chart diagram.
20847 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
20849 PR rtl-optimization/56847
20850 * lra-constraints.c (process_alt_operands): Discourage alternative
20851 with non-matche doffsettable memory constraint fro memory with
20854 2013-04-19 Richard Biener <rguenther@suse.de>
20856 PR tree-optimization/56982
20857 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
20859 * gimplify.c (gimplify_call_expr): Notice special calls.
20860 (gimplify_modify_expr): Likewise.
20861 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
20862 abnormal control flow receivers.
20863 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
20864 in the same way as cfun->has_nonlocal_labels.
20865 (gimple_purge_dead_abnormal_call_edges): Likewise.
20866 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
20867 receivers start a basic-block.
20869 2013-04-19 Richard Biener <rguenther@suse.de>
20871 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
20873 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
20874 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
20875 (SLP_TREE_LOAD_PERMUTATION): Add.
20876 (vect_transform_slp_perm_load): Adjust prototype.
20877 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
20878 (vect_free_slp_instance): Likewise.
20879 (vect_create_new_slp_node): Likewise.
20880 (vect_supported_slp_permutation_p): Remove.
20881 (vect_slp_rearrange_stmts): Adjust.
20882 (vect_supported_load_permutation_p): Likewise. Inline
20883 vect_supported_slp_permutation_p here.
20884 (vect_analyze_slp_instance): Compute load permutations per
20885 slp node instead of per instance.
20886 (vect_get_slp_defs): Adjust.
20887 (vect_transform_slp_perm_load): Likewise.
20888 (vect_schedule_slp_instance): Remove redundant code.
20889 (vect_schedule_slp): Remove hack for PR56270, add it ...
20890 * tree-vect-stmts.c (vectorizable_load): ... here, do not
20891 CSE loads for SLP. Adjust.
20893 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
20895 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
20896 spelling in two comments.
20898 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
20901 * config/arm/arm.c (load_multiple_sequence): Require SP
20902 as base register for loads if SP is in the register list.
20904 2013-04-19 Martin Jambor <mjambor@suse.cz>
20906 PR tree-optimization/56718
20907 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
20908 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
20909 and made public. Adjusted all callers.
20910 (ipa_intraprocedural_devirtualization): New function.
20911 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
20912 (ipa_intraprocedural_devirtualization): Likewise.
20913 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
20915 2013-04-19 Richard Biener <rguenther@suse.de>
20917 PR tree-optimization/57000
20918 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
20920 2013-04-19 Terry Guo <terry.guo@arm.com>
20922 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
20924 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
20925 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
20926 (cortex_m4_fmacs): Use new reservations.
20927 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
20929 2013-04-18 Vladimir Makarov <vmakarov@redhat.com>
20931 PR rtl-optimization/56999
20932 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
20934 (lra_coalesce): Remove split_origin_bitmap and related code.
20935 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
20936 ranges if necessary.
20938 2013-04-18 Uros Bizjak <ubizjak@gmail.com>
20940 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
20942 (ix86_expand_call): Remove clobbered_registers array and use
20943 x86_64_ms_sysv_extra_clobbered_registers instead.
20944 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
20946 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
20948 * config/i386/i386.md (*call_rex64_ms_sysv): Use
20949 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
20950 (*call_value_rex64_ms_sysv): Ditto.
20952 2013-04-18 Cary Coutant <ccoutant@google.com>
20954 * dwarf2out.c (output_pubnames): Check die_perennial_p of
20955 parent instead of die_mark.
20957 2013-04-18 Diego Novillo <dnovillo@google.com>
20959 * gimple.c (create_gimple_tmp): New.
20960 (get_expr_type): New.
20961 (build_assign): New.
20962 (build_type_cast): New.
20963 * gimple.h (enum ssa_mode): Define.
20964 (gimple_seq_set_location): New.
20965 * asan.c (build_check_stmt): Change some gimple_build_* calls
20966 to use build_assign and build_type_cast.
20968 2013-04-18 Richard Biener <rguenther@suse.de>
20970 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
20971 handle negative step. Remove redundant checks.
20972 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
20973 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
20974 for negative step and grouped loads fail to vectorize.
20976 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
20978 * emit-rtl.c (reset_insn_used_flags): New function.
20979 (reset_all_used_flags): Use it.
20980 (verify_insn_sharing): New function.
20981 (verify_rtl_sharing): Fix verification for SEQUENCEs.
20983 2013-04-18 Jakub Jelinek <jakub@redhat.com>
20985 PR tree-optimization/56984
20986 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
20987 and (x >> M) >= N don't register any assertion if N << M is the
20990 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
20992 * lower-subreg.c (resolve_simple_move): If called self-recursive,
20993 do not delete_insn insns that have not yet been emitted, only
20994 unlink them with remove_insn.
20995 * df-scan.c (df_insn_delete): Revert r197492.
20997 2013-04-17 Steven Bosscher <steven@gcc.gnu.org>
20999 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
21000 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
21002 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
21004 * config/arm/arm.md (movsicc_insn): Convert define_insn into
21005 define_insn_and_split.
21006 (and_scc,ior_scc,negscc): Likewise.
21007 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
21009 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
21011 * config/arm/arm.c (use_return_insn): Return 0 for targets that
21012 can benefit from using a sequence of LDRD instructions in epilogue
21013 instead of a single LDM instruction.
21015 2013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
21018 * doc/extend.texi: Fix typo.
21020 2013-04-17 Richard Biener <rguenther@suse.de>
21022 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
21023 (vect_build_slp_tree): ... here.
21024 (vect_build_slp_tree_1): Compute which stmts of the SLP group
21025 match. Remove special-casing of mismatched complex loads.
21026 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
21027 re-try the match with swapped commutative operands.
21028 (vect_supported_load_permutation_p): Remove special-casing of
21029 mismatched complex loads.
21030 (vect_analyze_slp_instance): Adjust.
21032 2013-04-17 Richard Biener <rguenther@suse.de>
21034 PR rtl-optimization/56921
21035 * cfgloop.h (struct loop): Add simple_loop_desc member.
21036 (struct niter_desc): Mark with GTY(()).
21037 (simple_loop_desc): Do not use aux field but simple_loop_desc.
21038 * loop-iv.c (get_simple_loop_desc): Likewise.
21039 (free_simple_loop_desc): Likewise.
21042 2013-04-16 Richard Biener <rguenther@suse.de>
21044 PR rtl-optimization/56921
21045 * loop-init.c (pass_rtl_move_loop_invariants): Add
21046 TODO_do_not_ggc_collect to todo_flags_finish.
21047 (pass_rtl_unswitch): Same.
21048 (pass_rtl_unroll_and_peel_loops): Same.
21049 (pass_rtl_doloop): Same.
21051 2013-04-17 Eric Botcazou <ebotcazou@adacore.com>
21053 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
21054 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
21055 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
21057 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
21058 * tree-streamer.c (record_common_node): Adjust reference in comment.
21060 2013-04-17 Terry Guo <terry.guo@arm.com>
21062 * config/arm/cortex-m4.md: Add a new bypass.
21064 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
21066 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
21068 (*subs_<optab><mode>_multp2): New pattern.
21069 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
21070 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
21072 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
21074 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
21075 (*subs_mul_imm_<mode>): New pattern.
21077 2013-04-16 David Edelsohn <dje.gcc@gmail.com>
21080 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
21081 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
21082 (vsx_movti_32bit): Change j->wa to O->wa.
21084 2013-04-16 Richard Biener <rguenther@suse.de>
21086 PR rtl-optimization/56921
21087 * loop-init.c (pass_rtl_move_loop_invariants): Add
21088 TODO_do_not_ggc_collect to todo_flags_finish.
21089 (pass_rtl_unswitch): Same.
21090 (pass_rtl_unroll_and_peel_loops): Same.
21091 (pass_rtl_doloop): Same.
21093 2013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
21095 * config/arm/arm.c (emit_multi_reg_push): New declaration
21096 for an existing function.
21097 (arm_emit_strd_push): New function.
21098 (arm_expand_prologue): Used here.
21099 (arm_emit_ldrd_pop): New function.
21100 (arm_expand_epilogue): Used here.
21101 (arm_get_frame_offsets): Update condition.
21102 (arm_emit_multi_reg_pop): Add a special case for load of a single
21103 register with writeback.
21105 2013-04-16 Uros Bizjak <ubizjak@gmail.com>
21107 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
21110 2013-04-16 Richard Biener <rguenther@suse.de>
21112 PR tree-optimization/56756
21113 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
21114 (first_mem_ref_loc): New.
21115 (execute_sm): Place the load temporarily before a previous
21116 access instead of in the latch edge to ensure its SSA dependencies
21117 are defined at points dominating the load.
21119 2013-04-16 Steven Bosscher <steven@gcc.gnu.org>
21121 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
21122 correct fix by moving header and footer insn to the footer of
21123 the merged basic block. Clear BB_END of the merged-away block.
21125 PR middle-end/43631
21126 * emit-rtl.c (make_note_raw): New function.
21127 (link_insn_into_chain): New static inline function.
21128 (add_insn): Use it.
21129 (add_insn_before, add_insn_after): Factor insn chain linking code...
21130 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
21131 using link_insn_into_chain.
21132 (note_outside_basic_block_p): New helper function for emit_note_after
21133 and emit_note_before.
21134 (emit_note_after): Use nobb variant of add_insn_after if the note
21135 should not be contained in a basic block.
21136 (emit_note_before): Use nobb variant of add_insn_before if the note
21137 should not be contained in a basic block.
21138 (emit_note_copy): Use make_note_raw.
21139 (emit_note): Likewise.
21140 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
21141 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
21142 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
21143 the moved barrier the tail of the basic block it follows.
21144 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
21146 2013-04-15 Jakub Jelinek <jakub@redhat.com>
21148 PR tree-optimization/56962
21149 * gimple-ssa-strength-reduction.c (record_increment): Only set
21150 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
21151 either rhs1 or rhs2 is equal to c->base_expr.
21153 2013-04-15 Richard Biener <rguenther@suse.de>
21155 PR tree-optimization/56933
21156 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
21158 (GROUP_READ_WRITE_DEPENDENCE): Remove.
21159 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
21160 * tree-vect-data-refs.c (vect_analyze_group_access): Move
21161 dependence check ...
21162 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
21164 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
21165 GROUP_READ_WRITE_DEPENDENCE.
21167 2013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
21169 * emit-rtl.c (reset_all_used_flags): New function.
21170 (verify_rtl_sharing): Call reset_all_used_flags before and after
21171 performing the checks.
21173 2013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21175 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
21176 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
21177 * config/arm/constraints.md (De): New constraint.
21178 * config/arm/neon.md (anddi3_neon): Delete.
21179 (neon_vand<mode>): Expand to standard anddi3 pattern.
21180 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
21181 Move earlier in the file.
21182 (neon_inv_logic_op2): Likewise.
21183 (arm_anddi_operand_neon): New predicate.
21185 2013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
21187 * configure.ac (gcc_cv_ld_as_needed): Set
21188 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
21189 Use -z ignore, -z record on *-*-solaris2*.
21190 (HAVE_LD_AS_NEEDED): Update comment.
21191 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
21192 * configure: Regenerate.
21193 * config.in: Regenerate.
21194 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
21195 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
21196 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
21197 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
21198 equivalents. Fix markup.
21199 * doc/tm.texi: Regenerate.
21201 2013-04-15 Andrew Hsieh <andrewhsieh.google.com>
21203 * config/i386/i386.opt: New option mstack-protector-guard=.
21204 * config/i386/i386-opts.h: Add enum stack_protector_guard.
21205 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
21206 TARGET_SSP_TLS_GUARD.
21207 * config/i386/i386.c (ix86_option_override_internal): Set
21208 ix86_stack_protector_guard.
21209 * config/i386/i386.md (stack_protect_set): Enable for
21210 TARGET_SSP_TLS_GUARD only.
21211 (stack_protect_set_<mode>): Ditto.
21212 (stack_protect_test): Ditto.
21213 (stack_protect_test_<mode>): Ditto.
21214 * doc/invoke.texi (i386 Option): Document.
21216 2013-04-15 Eric Botcazou <ebotcazou@adacore.com>
21219 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
21220 (S_MODES): Set H_MODE bit.
21221 (SF_MODES): Set only S_MODE and SF_MODE bits.
21222 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
21223 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
21224 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
21225 <MODE_FLOAT>: Likewise.
21227 2013-04-15 Joey Ye <joey.ye@arm.com>
21229 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
21231 2013-04-15 Joey Ye <joey.ye@arm.com>
21233 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
21235 (thumb_far_jump_used_p): Count instruction size and set
21238 2013-04-14 Eric Botcazou <ebotcazou@adacore.com>
21240 * reorg.c (fill_simple_delay_slots): Reindent block of code.
21241 * resource.c (mark_target_live_regs): Reformat conditional block.
21243 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
21245 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
21246 notes, they are emitted only just before final.
21247 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
21249 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
21251 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
21252 * cfgrtl.c (delete_insn): Call it here instead.
21253 * lra-spills.c (lra_final_code_change): Use delete_insn.
21254 * haifa-sched.c (sched_remove_insn): Likewise.
21255 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
21256 returning to the nop pool.
21257 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
21258 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
21260 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
21262 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
21263 * doc/tm.texi: Regenerated.
21265 2013-04-12 Uros Bizjak <ubizjak@gmail.com>
21267 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
21270 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
21272 * df-core.c (df_find_def): Compare register numbers.
21273 (df_find_use): Likewise.
21275 2013-04-12 Vladimir Makarov <vmakarov@redhat.com>
21278 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
21279 lra_in_progress for return.
21281 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
21283 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
21284 define_insn into define_insn_and_split and emit movsicc patterns.
21286 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
21288 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
21290 2013-04-12 Richard Biener <rguenther@suse.de>
21292 * tree-pass.h (TODO_do_not_ggc_collect): New.
21293 * passes.c (execute_one_ipa_transform_pass): Honor
21294 TODO_do_not_ggc_collect.
21295 (execute_one_pass): Likewise.
21298 2013-04-10 Richard Biener <rguenther@suse.de>
21300 * passes.c (init_optimization_passes): Remove reload pass.
21301 * ira.c (do_reload): Merge into ...
21303 (rest_of_handle_reload): Remove.
21304 (pass_reload): Likewise.
21305 * config/i386/i386.c (ix86_option_override): Refer to ira instead
21306 of reload for vzeroupper pass placement.
21308 2013-04-12 Jakub Jelinek <jakub@redhat.com>
21310 PR tree-optimization/56918
21311 PR tree-optimization/56920
21312 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
21313 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
21314 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
21315 use wide_mul_with_sign method.
21317 2013-04-12 Richard Biener <rguenther@suse.de>
21319 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
21320 not be considered a gimple constant.
21322 2013-04-12 Marc Glisse <marc.glisse@inria.fr>
21324 * fold-const.c (const_binop): Handle vector shifts by a scalar.
21325 (fold_binary_loc): Call const_binop also for mixed vector-scalar
21328 2013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
21329 Jakub Jelinek <jakub@redhat.com>
21331 * opts.c: Include diagnostic-color.h.
21332 (common_handle_option): Handle OPT_fdiagnostics_color_.
21333 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
21334 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
21335 (diagnostic-color.o): New.
21336 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
21337 (diagnostic_color_rule): New enum.
21338 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
21339 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
21340 the location string.
21341 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
21342 either NULL, or color kind.
21343 * diagnostic-color.c: New file.
21344 * diagnostic-color.h: New file.
21345 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
21347 * doc/invoke.texi (-fdiagnostics-color): Document.
21348 * pretty-print.h (pp_show_color): Define.
21349 (struct pretty_print_info): Add show_color field.
21350 * diagnostic.c: Include diagnostic-color.h.
21351 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
21352 macros. Colorize error:, warning: etc. strings and also the location
21354 (diagnostic_show_locus): Colorize the caret line.
21355 * pretty-print.c: Include diagnostic-color.h.
21356 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
21357 inside of %< %> quotes or quoted through q format modifier.
21359 2013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
21361 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
21363 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
21365 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
21366 code in CC_NZ mode.
21367 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
21370 2013-04-11 Marek Polacek <polacek@redhat.com>
21372 PR tree-optimization/48184
21373 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
21375 2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
21377 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
21378 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
21379 (skip_simple_arithmetic): Tidy up.
21380 * tree.h (skip_simple_constant_arithmetic): Declare.
21382 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
21384 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
21386 2013-04-11 Richard Biener <rguenther@suse.de>
21388 * tree-vect-loop.c (get_initial_def_for_induction): Properly
21389 generate vector constants.
21391 2013-04-11 Richard Biener <rguenther@suse.de>
21393 PR tree-optimization/56878
21394 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
21395 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
21397 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
21398 Prefer to align the DR with the most invariant base address.
21400 2013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
21402 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
21405 2013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
21407 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
21408 floating-point vector comparisons against 0.
21410 2013-04-11 Jakub Jelinek <jakub@redhat.com>
21412 PR tree-optimization/56899
21413 * fold-const.c (extract_muldiv_1): Apply distributive law
21414 only if TYPE_OVERFLOW_WRAPS (ctype).
21416 2013-04-11 Bin Cheng <bin.cheng@arm.com>
21419 * ira-costs.c (scan_one_insn): Check whether the source rtx of
21420 loading has side effect.
21422 2013-04-10 Steven Bosscher <steven@gcc.gnu.org>
21424 * config/sparc/sparc.c: Include tree-pass.h.
21425 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
21426 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
21427 head of file. Change return type. Split off gate function.
21428 (sparc_gate_work_around_errata): New function.
21429 (pass_work_around_errata): New pass definition.
21430 (insert_pass_work_around_errata) New pass insert definition to
21431 insert pass_work_around_errata just after delayed-branch scheduling.
21432 (sparc_option_override): Insert the pass.
21433 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
21435 2013-04-10 David S. Miller <davem@davemloft.net>
21437 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
21438 or -mcpu=hypersparc.
21440 * target.def (cstore_mode): New hook.
21441 * target.h: Include insn-codes.h
21442 * targhooks.c: Likewise.
21443 (default_cstore_mode): New function.
21444 * targhooks.h: Declare it.
21445 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
21446 * doc/tm.texi: Rebuild.
21447 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
21448 target hook, rather than inspecting the insn_data.
21449 * config/sparc/sparc.c (sparc_cstore_mode): New function.
21450 (TARGET_CSTORE_MODE): Redefine.
21451 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
21453 * config/sparc/predicates.md (cstore_result_operand): New special
21455 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
21456 Use it for operand 0.
21457 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
21458 (*snesi_special): Likewise.
21459 (*snesi_zero): Likewise.
21460 (*seqsi_zero): Likewise.
21461 (*sltu_insn): Likewise.
21462 (*sgeu_insn): Likewise.
21463 (*seqdi_special): Make operand 0 and comparison operation be of
21465 (*snedi_special): Likewise.
21466 (*snedi_special_vis3): Likewise.
21467 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
21468 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
21469 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
21470 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
21471 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
21472 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
21473 (*sltu_extend_sp64): Likewise.
21474 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
21475 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
21476 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
21477 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
21478 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
21480 2013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
21482 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
21483 (aarch64_start_file): Use the new function.
21485 2013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
21486 Jason Merrill <jason@redhat.com>
21488 * common.opt: Add -gdwarf.
21489 * opts.c (common_handle_option): Handle it.
21490 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
21492 2013-04-10 Richard Biener <rguenther@suse.de>
21494 * passes.c (execute_todo): Do not call ggc_collect conditional here.
21495 (execute_one_ipa_transform_pass): But unconditionally here.
21496 (execute_one_pass): And here.
21497 (init_optimization_passes): Remove reload pass.
21498 * tree-pass.h (TODO_ggc_collect): Remove.
21499 (pass_reload): Likewise.
21500 * ira.c (do_reload): Merge into ...
21502 (rest_of_handle_reload): Remove.
21503 (pass_reload): Likewise.
21504 * config/i386/i386.c (ix86_option_override): Refer to ira instead
21505 of reload for vzeroupper pass placement.
21506 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
21507 and todo_flags_finish of all passes.
21509 2013-04-10 Richard Biener <rguenther@suse.de>
21511 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
21512 first_const_oprnd field, rename first_def_type to first_op_type.
21513 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
21514 (vect_get_and_check_slp_defs): Always use the type of the
21515 operand. Allow mixed vect_external_def, vect_constant_def types.
21516 (vect_get_constant_vectors): Handle mixed vect_external_def,
21517 vect_constant_def types.
21519 2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
21521 PR tree-optimization/55524
21522 * tree-ssa-math-opts.c
21523 (convert_mult_to_fma): Don't use an fms construct
21524 when we don't have an fms operation, but fnma, and it looks
21525 likely that we'll be able to use the latter.
21527 2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
21529 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
21531 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
21532 inline fail caused by overwritable functions.
21534 2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
21536 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
21537 unnecessary bits in the constant power of two case.
21539 2013-04-10 Richard Biener <rguenther@suse.de>
21541 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
21542 broken code swapping operands.
21543 (vect_build_slp_tree): Do not compute load permutations here.
21544 (vect_analyze_slp_instance): Compute load permutations here,
21545 after building the SLP tree.
21547 2013-04-09 Christian Bruel <christian.bruel@st.com>
21549 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
21550 of next/prev_real_insn.
21552 2013-04-09 Jan Hubicka <jh@suse.cz>
21554 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
21555 Drop aliased parameter.
21556 (function_and_variable_visibility): Do not handle alias pairs.
21557 * cgraph.c (varpool_externally_visible_p): Update prototype.
21558 * varpool.c (varpool_add_new_variable): Update.
21560 2013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21562 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
21564 2013-04-09 Steven Bosscher <steven@gcc.gnu.org>
21566 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
21568 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
21570 2013-04-09 Marek Polacek <polacek@redhat.com>
21572 PR tree-optimization/48762
21573 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
21575 2013-04-09 Richard Biener <rguenther@suse.de>
21577 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
21579 (vect_build_slp_tree): Likewise.
21580 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
21581 calculating the cost of a SLP instance.
21582 (vect_analyze_slp_instance): Use it from here, after building
21585 2013-04-09 Jakub Jelinek <jakub@redhat.com>
21587 PR middle-end/56883
21588 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
21589 expand_omp_for_static_chunk): Use simple_p = true in
21590 force_gimple_operand_gsi calls when assigning to addressable decls.
21592 2013-04-09 Jeff Law <law@redhat.com>
21594 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
21595 when the boolean was created by converting a wider object which
21596 had a boolean range.
21598 2013-04-09 Richard Biener <rguenther@suse.de>
21600 * tree-vectorizer.h (slp_void_p): Remove.
21601 (slp_tree): Typedef before _slp_tree declaration.
21602 (struct _slp_tree): Use a vector of slp_tree as children.
21603 (vect_get_place_in_interleaving_chain): Remove.
21604 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
21606 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
21608 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
21609 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
21610 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
21611 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
21612 Use slp_node instead of slp_void_p and adjust.
21614 2013-04-09 Richard Biener <rguenther@suse.de>
21616 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
21617 work that is not necessary.
21619 2013-04-09 Jakub Jelinek <jakub@redhat.com>
21621 PR tree-optimization/56854
21622 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
21623 forward into clobber stmts if it would change MEM_REF lhs into
21626 2013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
21628 * tree.c (type_hash_lookup, type_hash_add): Make static.
21629 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
21631 2013-04-09 Richard Biener <rguenther@suse.de>
21633 * tree.h (unsave_expr_now): Remove.
21634 * tree-inline.c (mark_local_for_remap_r): Remove.
21635 (unsave_expr_1): Likewise.
21636 (unsave_r): Likewise.
21637 (unsave_expr_now): Likewise.
21638 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
21639 (propagate_tree_value): Likewise.
21641 2013-04-08 Steven Bosscher <steven@gcc.gnu.org>
21643 * doc/rtl.texi (sequence): Rewrite documentation to match the
21644 current use of SEQUENCE rtl objects.
21645 * rtl.def (SEQUENCE): Likewise.
21647 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
21648 Update documentation.
21649 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
21650 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
21652 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
21654 2013-04-08 Teresa Johnson <tejohnson@google.com>
21656 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
21657 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
21659 (estimate_edge_size_and_time): Add comment to suggest using rounding
21661 (estimate_node_size_and_time): Ditto.
21662 (remap_edge_change_prob): Use helper rounding divide methods.
21663 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
21664 (gimple_mod_pow2_value_transform): Ditto.
21665 (gimple_mod_subtract_transform): Ditto.
21666 (gimple_ic_transform): Ditto.
21667 (gimple_stringops_transform): Ditto.
21668 * stmt.c (conditional_probability): Ditto.
21669 (emit_case_dispatch_table): Ditto.
21670 * lto-cgraph.c (merge_profile_summaries): Ditto.
21671 * tree-optimize.c (execute_fixup_cfg): Ditto.
21672 * cfgcleanup.c (try_forward_edges): Ditto.
21673 * cfgloopmanip.c (scale_loop_profile): Ditto.
21675 (duplicate_loop_to_header_edge): Ditto.
21676 (lv_adjust_loop_entry_edge): Ditto.
21677 * tree-vect-loop.c (vect_transform_loop): Ditto.
21678 * profile.c (compute_branch_probabilities): Ditto.
21679 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
21680 * lto-streamer-in.c (input_cfg): Ditto.
21681 * gimple-streamer-in.c (input_bb): Ditto.
21682 * ipa-cp.c (update_profiling_info): Ditto.
21683 (update_specialized_profile): Ditto.
21684 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
21685 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
21687 * sched-rgn.c (compute_dom_prob_ps): Ditto.
21688 (compute_trg_info): Ditto.
21689 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
21690 (purge_dead_edges): Ditto.
21691 * loop-unswitch.c (unswitch_loop): Ditto.
21692 * cgraphclones.c (cgraph_clone_edge): Ditto.
21693 (cgraph_clone_node): Ditto.
21694 * tree-inline.c (copy_bb): Ditto.
21695 (copy_edges_for_bb): Ditto.
21696 (initialize_cfun): Ditto.
21697 (copy_cfg_body): Ditto.
21698 (expand_call_inline): Ditto.
21700 2013-04-08 Kai Tietz <ktietz@redhat.com>
21702 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
21703 TARGET_CYGWIN64 by TARGET_64BIT.
21705 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
21707 * config/epiphany/epiphany.md (GPR_1): New constant.
21708 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
21709 * config/epiphany/epiphany.c (gen_compare_reg):
21710 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
21711 is already in place.
21712 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
21713 Don't require being called during rtl expansion; If y operlaps r0,
21715 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
21716 (epiphany_expand_epilogue): Likewise.
21718 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
21719 Don't use CC_FPmode for ORDERED / UNORDERED.
21720 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
21722 * config/epiphany/constraints.md (CnL): New constraint.
21723 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
21724 * config/epiphany/predicates.md (add_operand): Allow 1024.
21726 * config/epiphany/epiphany.md (logical_op): New code iterator.
21727 (op_mnc): New code attribute.
21728 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
21729 (mov_f+1, mov_f+2): New peephole2 patterns.
21731 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
21732 (cstoresi4): Also allow re-use of zero result when doing a NE
21733 comparison to a non-zero operand.
21734 Use (clobber (scratch)) for first insn if the gpr output is not needed.
21736 * config/epiphany/epiphany.md (<insn_opname>v2si3):
21737 Use gen_addsi3_i / gen_subsi3_i.
21739 2013-04-08 Jakub Jelinek <jakub@redhat.com>
21743 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
21744 contain anything but clobbers, at most one __builtin_stack_restore,
21745 optionally debug stmts and final resx, and if it has at least one
21746 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
21747 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
21748 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
21749 which isn't defaut definition, remove them.
21750 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
21751 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
21752 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
21753 with MEM_REF LHS with SSA_NAME address.
21755 2013-04-08 Jeff Law <law@redhat.com>
21757 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
21759 2013-04-08 Richard Biener <rguenther@suse.de>
21761 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
21763 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
21764 determined vector type.
21765 (vect_analyze_data_refs): Likewise.
21766 (vect_get_new_vect_var): Adjust.
21767 (vect_create_destination_var): Preserve SSA name versions.
21768 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
21769 not dump anything here.
21771 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
21773 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
21774 Add member lr_slot_known.
21775 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
21777 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
21778 Remove code that sets lr_slot_offset according to what a previous
21779 version of epiphany_emit_save_restore used to do.
21780 (epiphany_emit_save_restore): When doing an lr save or restore,
21781 set/verify lr_slot_known and lr_slot_offset.
21783 2013-04-08 Xinyu Qi <xyqi@marvell.com>
21786 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
21789 2013-04-08 Richard Biener <rguenther@suse.de>
21791 * alias.c (find_base_term): Fix thinko in previous change.
21793 2013-04-08 Jakub Jelinek <jakub@redhat.com>
21795 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
21796 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
21797 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
21798 if possible to compute val.
21799 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
21800 For QImode integers don't require anything about precision. Use
21801 const_with_all_bytes_same to find out if the constant doesn't have
21802 repeated bytes in it.
21804 2013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
21806 * config/s390/s390.c (s390_expand_insv): Only accept insertions
21809 2013-04-08 Marek Polacek <polacek@redhat.com>
21811 PR rtl-optimization/48182
21812 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
21815 2013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
21818 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
21819 nuses, make sure we have a label.
21821 2013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
21824 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
21825 (rs6000_emit_swdiv_low_precision): Remove.
21826 (rs6000_emit_swdiv): Rewrite to handle between one and four
21827 iterations of Newton-Raphson generally; modify required number of
21828 iterations for some cases.
21829 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
21831 2013-04-05 Steven Bosscher <steven@gcc.gnu.org>
21833 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
21834 set-but-unused variable.
21836 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
21837 basic blocks of released function bodies garbage-collectable.
21839 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
21840 (struct rtl_opt_pass): Add TODO_df_finish.
21842 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
21844 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
21846 * config/arm/constraints.md (q): New constraint.
21847 * config/arm/ldrdstrd.md: New file.
21848 * config/arm/arm.md (ldrdstrd.md) New include.
21849 (arm_movdi): Use "q" instead of "r" constraint
21850 for double-word memory access.
21851 (movdf_soft_insn): Likewise.
21852 * config/arm/vfp.md (movdi_vfp): Likewise.
21853 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
21854 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
21855 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
21856 (mem_ok_for_ldrd_strd): Likewise.
21857 (output_move_double): Update assertion.
21859 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
21861 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
21863 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
21865 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
21866 define_insn_and_split.
21867 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
21869 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
21871 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
21872 define_insn_and_split.
21873 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
21874 (shiftsi3_compare): New pattern.
21875 (rrx): New pattern.
21876 * config/arm/unspecs.md (UNSPEC_RRX): New.
21878 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
21880 * config/arm/arm.md (negdi_extendsidi): New pattern.
21881 (negdi_zero_extendsidi): Likewise.
21883 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
21885 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
21886 define_insn_and_split.
21887 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
21888 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
21890 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
21892 * config/arm/arm.md (arm_subdi3): Convert define_insn into
21893 define_insn_and_split.
21894 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
21895 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
21897 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
21899 * config/arm/arm.md (subsi3_carryin): New pattern.
21900 (subsi3_carryin_const): Likewise.
21901 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
21902 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
21904 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
21906 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
21908 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
21910 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
21911 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
21913 2013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21915 * config/arm/arm.c (arm_expand_builtin): Change fcode
21916 type to unsigned int.
21918 2013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
21920 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
21922 2013-04-04 Ian Lance Taylor <iant@google.com>
21924 * doc/standards.texi (Standards): The Go frontend supports the Go 1
21927 2013-04-04 Steven Bosscher <steven@gcc.gnu.org>
21929 PR middle-end/56729
21930 * df-scan.c (df_insn_delete): Disable failing assert.
21932 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21934 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
21935 New function prototype.
21936 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
21937 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
21938 (arm_builtin_vectorized_function): New function.
21940 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21942 * config/arm/arm_neon_builtins.def: New file.
21943 * config/arm/arm.c (neon_builtin_data): Move contents to
21944 arm_neon_builtins.def.
21945 (enum arm_builtins): Include neon builtin definitions.
21946 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
21947 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
21949 2013-04-04 Marek Polacek <polacek@redhat.com>
21951 PR tree-optimization/48186
21952 * predict.c (maybe_hot_frequency_p): Return false if
21953 HOT_BB_FREQUENCY_FRACTION is 0.
21954 (cgraph_maybe_hot_edge_p): Likewise.
21956 2013-04-04 Richard Biener <rguenther@suse.de>
21958 PR tree-optimization/56826
21959 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
21962 2013-04-04 Richard Biener <rguenther@suse.de>
21964 PR tree-optimization/56213
21965 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
21966 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
21968 2013-04-04 Richard Biener <rguenther@suse.de>
21970 PR tree-optimization/56837
21971 * tree-loop-distribution.c (classify_partition): For non-zero
21972 values require that the value has the same precision as its
21973 mode to be useful as memset value.
21975 2013-04-03 Nick Clifton <nickc@redhat.com>
21977 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
21978 (fmssf4): Use fmsf.s on E3V5 architectures.
21979 (fnmasf4): Use fnmaf.s on E3V5 architectures.
21980 (fnmssf4): Use fnmsf.s on E3V5 architectures.
21982 2013-04-03 Jeff Law <law@redhat.com>
21984 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
21985 (lra-eliminations.o): Likewise.
21987 2013-04-03 Teresa Johnson <tejohnson@google.com>
21989 * gcov-io.c (compute_working_sets): Moved most of body of old
21990 compute_working_sets here from profile.c.
21991 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
21992 (gcov_working_set_t): Moved typedef here from basic-block.h
21993 (compute_working_set): Declare.
21994 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
21995 (get_working_sets): Renamed from compute_working_set,
21996 replace most of body with call to new compute_working_sets.
21997 (get_exec_counts): Replace call to compute_working_sets
21998 to get_working_sets.
21999 * profile.h (get_working_sets): Renamed from compute_working_set.
22000 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
22001 to get_working_sets.
22002 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
22003 * gcov-dump.c (dump_working_sets): New function.
22005 2013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
22007 * hwint.c (sext_hwi, zext_hwi): New functions.
22008 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
22009 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
22010 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
22011 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
22012 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
22013 (sext_hwi, zext_hwi): New functions.
22015 2013-04-03 Jeff Law <law@redhat.com>
22017 PR tree-optimization/56799
22018 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
22019 back test for widening conversion erroneously dropped in prior change.
22021 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22024 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
22025 instead of next_real_insn.
22027 2013-04-03 Marek Polacek <polacek@redhat.com>
22030 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
22032 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22035 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
22037 (thumb1_output_casesi): Likewise.
22038 (thumb2_output_casesi): Likewise.
22040 2013-04-03 Richard Biener <rguenther@suse.de>
22042 PR tree-optimization/56817
22043 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
22045 (tree_unroll_loops_completely_1): ... new function to manually
22046 walk the loop tree, properly defering outer loops of unrolled
22047 loops to later iterations.
22049 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
22051 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
22052 (vectorizable_load): Likewise.
22053 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
22054 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
22056 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
22058 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
22061 2013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
22063 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
22065 2013-04-03 Bin Cheng <bin.cheng@arm.com>
22067 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
22069 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
22071 PR tree-optimization/56790
22072 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
22075 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
22077 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
22079 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
22080 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
22083 2013-04-03 Jakub Jelinek <jakub@redhat.com>
22086 * tree.h (force_folding_builtin_constant_p): New decl.
22087 * builtins.c (force_folding_builtin_constant_p): New variable.
22088 (fold_builtin_constant_p): Fold immediately also if
22089 force_folding_builtin_constant_p.
22091 2013-04-03 Richard Biener <rguenther@suse.de>
22093 PR tree-optimization/56812
22094 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
22095 DRs of the same interleaving chain are independent.
22097 2013-04-02 Jason Merrill <jason@redhat.com>
22099 * gdbinit.in (pbb): Use debug fn.
22101 2013-04-02 Lawrence Crowl <crowl@google.com>
22103 * sese.h (struct ivtype_map_elt_s): Remove unused.
22104 (extern debug_ivtype_map): Remove unused.
22105 (extern eq_ivtype_map_elts): Remove unused.
22106 * sese.c (debug_ivtype_map): Removed unused.
22107 (debug_ivtype_map_1): Removed unused.
22108 (debug_ivtype_elt): Remove unused.
22109 (eq_ivtype_map_elts): Remove unused.
22112 2013-04-02 Kai Tietz <ktietz@redhat.com>
22115 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
22116 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
22117 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
22119 (legitimize_pe_coff_symbol): Likewise.
22120 (is_imported_p): New helper-function.
22121 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
22122 for Windows x64 targets.
22123 (ix86_expand_prologue): Optimize for pe-coff targets.
22124 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
22125 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
22126 medium/large code-model.
22127 (legitimize_pic_address): Likewise.
22128 (legitimize_tls_address): Likewise.
22129 (ix86_expand_call): Likewise.
22130 (x86_output_mi_thunk): Likewise.
22131 (get_dllimport_decl): Add new beimport argument.
22132 (construct_plt_address): Don't assert for x64 pe-coff targets.
22133 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
22135 (SYMBOL_FLAG_STUBVAR): New macro.
22136 (SYMBOL_REF_STUBVAR_P): Likewise.
22137 * config/i386/winnt.c (stub_list): New structure.
22138 (stub_head): New local variable.
22139 (i386_pe_record_stub): New function.
22140 (i386_pe_file_end): Emit refptr-stubs.
22142 2013-04-02 Jakub Jelinek <jakub@redhat.com>
22144 PR rtl-optimization/56745
22145 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
22146 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
22149 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
22150 and both of them are MEM_REFs, just compare first argument for
22151 equality and attempt to deal even with differing offsets.
22154 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
22155 of gimple_clobber_p to be MEM_REF.
22156 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
22157 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
22158 after gimplification.
22159 * asan.c (get_mem_ref_of_assignment): Don't instrument
22160 gimple_clobber_p stmts.
22161 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
22162 gimple_clobber_p stmt if they have MEM_REF lhs and
22163 are dead because of another gimple_clobber_p stmt.
22164 * tree-ssa-live.c (clear_unused_block_pointer): Treat
22165 gimple_clobber_p stmts like debug stmts.
22166 (remove_unused_locals): Remove clobbers with MEM_REF lhs
22167 that refer to unused VAR_DECLs or uninitialized values.
22168 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
22169 gimple_clobber_p stmts if they refer to removed parameters.
22170 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
22173 2013-04-02 Uros Bizjak <ubizjak@gmail.com>
22175 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
22176 using SWI48 mode attribute.
22178 2013-04-02 Wei Mi <wmi@google.com>
22180 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
22181 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
22182 *<rotate_insn><mode>3_mask in i386.md.
22184 2013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
22186 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
22188 2013-04-02 Richard Biener <rguenther@suse.de>
22190 PR tree-optimization/56778
22191 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
22192 Runtime alias tests are not supported for gather loads.
22193 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
22194 stmts referenced from SSA operands before updating SSA form.
22196 2013-04-02 Ian Caulfield <ian.caulfield@arm.com>
22197 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
22199 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
22200 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
22201 * config/arm/cortex-a53.md: New file.
22202 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
22203 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
22204 * config/arm/arm.c (arm_issue_rate): Likewise.
22205 * config/arm/arm-tune.md: Regenerate
22206 * config/arm/arm-tables.opt: Regenerate.
22207 * config/arm/arm-cores.def: Add cortex-a53.
22209 2013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
22211 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
22214 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
22216 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
22217 scalar load/store operations using B/H registers.
22218 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
22220 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
22222 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
22224 * config/aarch64/aarch64.c
22225 (aarch64_simd_scalar_immediate_valid_for_move): New.
22226 * config/aarch64/aarch64-protos.h
22227 (aarch64_simd_scalar_immediate_valid_for_move): New.
22228 * config/aarch64/constraints.md (Dh, Dq): New.
22229 * config/aarch64/iterators.md (hq): New.
22231 2013-04-02 Eric Botcazou <ebotcazou@adacore.com>
22233 * reorg.c (get_branch_condition): Deal with conditional returns.
22234 (fill_simple_delay_slots): Remove dead code dealing with jumps.
22236 2013-04-01 Wei Mi <wmi@google.com>
22238 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
22239 Truncate operand 2 using %b asm operand modifier.
22240 (*<shift_insn><mode>3_mask): Ditto.
22241 (*<rotate_insn><mode>3_mask): Ditto.
22243 2013-04-01 Steven Bosscher <steven@gcc.gnu.org>
22245 PR middle-end/56798
22246 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
22248 2013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
22250 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
22252 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
22254 2013-03-30 Lawrence Crowl <crowl@google.com>
22256 * dse.c (clear_alias_sets): Remove never set.
22257 (disqualified_clear_alias_sets): Remove never set.
22258 (clear_alias_mode_pool): Remove never set.
22259 (dse_step0): Remove condition that is never true.
22260 (canon_address): Remove condition that is never true.
22261 (dse_step7): Remove condition that is never true.
22262 (rest_of_handle_dse): Remove condition that is never true.
22263 (rest_of_handle_dse::did_global): Remove never read from above.
22264 (dse_step2_spill): Remove never called from above.
22265 (dse_step5_spill): Remove never called from above.
22267 2013-03-30 Steven Bosscher <steven@gcc.gnu.org>
22269 * doc/md.texi (Standard Names) <casesi>: Update documentation for
22270 JUMP_TABLE_DATA changes.
22271 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
22272 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
22273 (Insns) <jump_table_data>: New entry.
22274 * doc/tm.texi: Regenerate.
22276 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
22278 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
22279 for table jump at the end of a basic block using tablejump_p.
22280 * targhooks.c (default_invalid_within_doloop): Likewise.
22281 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
22282 target hook implementation that is identical to the default hook.
22283 (rs6000_invalid_within_doloop): Remove.
22285 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
22286 unused variable from tablejump_p call.
22288 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
22289 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
22290 (INSN_DELETED_P): Likewise.
22291 (emit_jump_table_data): New prototype.
22292 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
22293 after 4th as unused.
22294 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
22295 * sched-vis.c (print_insn): Likewise.
22296 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
22297 insn for compatibility with back ends that use next_active_insn to
22298 identify jump table data.
22299 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
22300 (remove_insn): Likewise.
22301 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
22303 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
22304 (emit_jump_table_data): New function.
22306 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
22307 basic block, a JUMP_TABLE_DATA never is.
22308 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
22309 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
22310 off from code handling real insns.
22311 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
22312 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
22313 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
22314 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
22315 is not a NONDEBUG_INSN_P.
22316 * ira-costs.c (scan_one_insn): Likewise.
22317 * jump.c (mark_all_labels): Likewise.
22318 (mark_jump_label_1): Likewise.
22319 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
22320 * lra.c (get_insn_freq): Expect all insns reaching here to be in
22322 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
22323 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
22324 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
22325 JUMP_TABLE_DATA_P insns.
22326 (calculate_elim_costs_all_insns): Likewise.
22327 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
22328 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
22329 (delete_output_reload): Code style fixups.
22330 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
22331 insn flags on this non-insn.
22332 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
22333 as scheduling barriers, for pre-change compatibility.
22334 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
22335 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
22337 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
22338 redundant JUMP_TABLE_DATA_P test.
22339 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
22340 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
22341 (frv_for_each_packet): Likewise.
22342 * config/i386/i386.c (min_insn_size): Likewise.
22343 (ix86_avoid_jump_mispredicts): Likewise.
22344 * config/m32r/m32r.c (m32r_is_insn): Likewise.
22345 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
22346 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
22347 (mips16_insn_length): Robustify.
22348 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
22349 (mips16_split_long_branches): Likewise.
22350 * config/pa/pa.c (pa_combine_instructions): Likewise.
22351 * config/rs6000/rs6000.c (get_next_active_insn): Treat
22352 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
22353 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
22354 as contributing to pool range lengths.
22355 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
22356 Remove redundant JUMP_TABLE_DATA_P test.
22357 (sh_loop_align): Likewise.
22358 (split_branches): Likewise.
22359 (sh_insn_length_adjustment): Likewise.
22360 * config/spu/spu.c (get_branch_target): Likewise.
22362 2013-03-29 Jan Hubicka <jh@suse.cz>
22364 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
22365 gcov streaming; stream hot bb threshold to ltrans.
22366 * predict.c (get_hot_bb_threshold): Break out from ....
22367 (maybe_hot_count_p): ... here.
22368 (set_hot_bb_threshold): New function.
22369 * lto-section-in.c (lto_section_name): Add profile.
22370 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
22371 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
22372 and data-streamer.h
22373 (histogram_entry): New structure.
22374 (histogram, histogram_pool): New global vars.
22375 (histogram_hash): New structure.
22376 (histogram_hash::hash): New method.
22377 (histogram_hash::equal): Likewise.
22378 (account_time_size): New function.
22379 (cmp_counts): New function.
22380 (dump_histogram): New function.
22381 (ipa_profile_generate_summary): New function.
22382 (ipa_profile_write_summary): New function.
22383 (ipa_profile_read_summary): New function.
22384 (ipa_profile): Decide on threshold.
22385 (pass_ipa_profile): Add ipa_profile_write_summary and
22386 ipa_profile_read_summary.
22387 * Makefile.in (ipa.o): Update dependencies.
22388 * lto-streamer.h (LTO_section_ipa_profile): New section.
22390 2013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
22392 * tree.h (VAR_P): New.
22394 2013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
22397 * doc/invoke.texi ([-fwhole-program]): Fix typo.
22399 2013-03-29 Steven Bosscher <steven@gcc.gnu.org>
22401 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
22402 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
22403 (control_flow_insn_p): Likewise.
22404 * cfgrtl.c (duplicate_insn_chain): Likewise.
22405 * final.c (get_attr_length_1): Likewise.
22406 (shorten_branches): Likewise.
22407 (final_scan_insn): Likewise.
22408 * function.c (instantiate_virtual_regs): Likewise.
22409 * gcse.c (insert_insn_end_basic_block): Likewise.
22410 * ira-costs.c (scan_one_insn): Likewise.
22411 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
22412 * lra.c (check_rtl): Likewise.
22413 * reload1.c (elimination_costs_in_insn): Likewise.
22414 * reorg.c (follow_jumps): Likewise.
22416 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
22417 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
22418 (thumb_far_jump_used_p): Likewise.
22419 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
22420 (workaround_speculation): Likewise.
22421 (add_sched_insns_for_speculation): Likewise.
22422 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
22423 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
22424 (frv_for_each_packet): Likewise.
22425 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
22426 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
22427 (final_emit_insn_group_barriers): Likewise.
22428 * config/m32r/m32r.c (m32r_is_insn): Likewise.
22429 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
22430 (mips16_insn_length): Likewise.
22431 * config/pa/pa.c (pa_reorg): Likewise.
22432 (pa_combine_instructions): Likewise.
22433 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
22434 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
22435 (sh_reorg): Likewise.
22436 (split_branches): Likewise.
22437 * config/spu/spu.c (get_branch_target): Likewise.
22439 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
22442 2013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
22444 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
22445 Fix declaration name.
22447 2013-03-28 Lawrence Crowl <crowl@google.com>
22449 * graphds.h (struct graph.indicies): Remove unused.
22450 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
22451 (SCOP_ORIGINAL_PDDRS): Remove unused.
22452 * sese.h (extern insert_loop_close_phis): Removed unused.
22453 (extern insert_guard_phis): Removed unused.
22454 (extern ivtype_map_elt_info): Removed unused.
22455 (new_ivtype_map_elt): Removed unused.
22456 * sese.c (ivtype_map_elt_info): Removed unused.
22458 2013-03-28 Lawrence Crowl <crowl@google.com>
22460 * Makefile.in: Add several missing include dependences.
22462 (test-dump.o): New. This object is not added to any executable,
22463 but is present for ad-hoc testing.
22465 (debug (const bitmap_head_def &)): New.
22466 (debug (const bitmap_head_def *)): New.
22468 (extern debug (const bitmap_head_def &)): New.
22469 (extern debug (const bitmap_head_def *)): New.
22471 (debug (edge_def &)): New.
22472 (debug (edge_def *)): New.
22474 (debug (basic_block_def &)): New.
22475 (debug (basic_block_def *)): New.
22477 (dump_node (const_tree, int, FILE *)): Correct source file.
22479 (debug (die_struct &)): New.
22480 (debug (die_struct *)): New.
22482 (extern debug (die_struct &)): New.
22483 (extern debug (die_struct *)): New.
22484 * gimple-pretty-print.c
22485 (debug (gimple_statement_d &)): New.
22486 (debug (gimple_statement_d *)): New.
22487 * gimple-pretty-print.h
22488 (extern debug (gimple_statement_d &)): New.
22489 (extern debug (gimple_statement_d *)): New.
22491 (debug (ira_allocno_copy &)): New.
22492 (debug (ira_allocno_copy *)): New.
22493 (debug (ira_allocno &)): New.
22494 (debug (ira_allocno *)): New.
22496 (extern debug (ira_allocno_copy &)): New.
22497 (extern debug (ira_allocno_copy *)): New.
22498 (extern debug (ira_allocno &)): New.
22499 (extern debug (ira_allocno *)): New.
22501 (debug (live_range &)): New.
22502 (debug (live_range *)): New.
22504 (debug (lra_live_range &)): New.
22505 (debug (lra_live_range *)): New.
22507 (debug (lra_live_range &)): New.
22508 (debug (lra_live_range *)): New.
22510 (debug (omega_pb_d &)): New.
22511 (debug (omega_pb_d *)): New.
22513 (extern debug (omega_pb_d &)): New.
22514 (extern debug (omega_pb_d *)): New.
22516 (debug (const rtx_def &)): New.
22517 (debug (const rtx_def *)): New.
22519 (debug_tree (tree): Move within file.
22520 (debug_raw (const tree_node &)): New.
22521 (debug_raw (const tree_node *)): New.
22522 (dump_tree_via_hooks (const tree_node *, int)): New.
22523 (debug (const tree_node &)): New.
22524 (debug (const tree_node *)): New.
22525 (debug_verbose (const tree_node &)): New.
22526 (debug_verbose (const tree_node *)): New.
22527 (debug_head (const tree_node &)): New.
22528 (debug_head (const tree_node *)): New.
22529 (debug_body (const tree_node &)): New.
22530 (debug_body (const tree_node *)): New.
22531 (debug_vec_tree (tree): Move and reimplement in terms of dump.
22532 (debug (vec<tree, va_gc> &)): New.
22533 (debug (vec<tree, va_gc> *)): New.
22535 (extern debug (const rtx_def &)): New.
22536 (extern debug (const rtx_def *)): New.
22538 (debug_raw (simple_bitmap_def &)): New.
22539 (debug_raw (simple_bitmap_def *)): New.
22540 (debug (simple_bitmap_def &)): New.
22541 (debug (simple_bitmap_def *)): New.
22543 (extern debug (simple_bitmap_def &)): New.
22544 (extern debug (simple_bitmap_def *)): New.
22545 (extern debug_raw (simple_bitmap_def &)): New.
22546 (extern debug_raw (simple_bitmap_def *)): New.
22548 (debug (vinsn_def &)): New.
22549 (debug (vinsn_def *)): New.
22550 (debug_verbose (vinsn_def &)): New.
22551 (debug_verbose (vinsn_def *)): New.
22552 (debug (expr_def &)): New.
22553 (debug (expr_def *)): New.
22554 (debug_verbose (expr_def &)): New.
22555 (debug_verbose (expr_def *)): New.
22556 (debug (vec<rtx> &)): New.
22557 (debug (vec<rtx> *)): New.
22559 (extern debug (vinsn_def &)): New.
22560 (extern debug (vinsn_def *)): New.
22561 (extern debug_verbose (vinsn_def &)): New.
22562 (extern debug_verbose (vinsn_def *)): New.
22563 (extern debug (expr_def &)): New.
22564 (extern debug (expr_def *)): New.
22565 (extern debug_verbose (expr_def &)): New.
22566 (extern debug_verbose (expr_def *)): New.
22567 (extern debug (vec<rtx> &)): New.
22568 (extern debug (vec<rtx> *)): New.
22570 (_list_iter_cond_expr): Make inline instead of static.
22572 (debug (sreal &)): New.
22573 (debug (sreal *)): New.
22575 (extern debug (sreal &)): New.
22576 (extern debug (sreal *)): New.
22578 (extern debug_raw (const tree_node &)): New.
22579 (extern debug_raw (const tree_node *)): New.
22580 (extern debug (const tree_node &)): New.
22581 (extern debug (const tree_node *)): New.
22582 (extern debug_verbose (const tree_node &)): New.
22583 (extern debug_verbose (const tree_node *)): New.
22584 (extern debug_head (const tree_node &)): New.
22585 (extern debug_head (const tree_node *)): New.
22586 (extern debug_body (const tree_node &)): New.
22587 (extern debug_body (const tree_node *)): New.
22588 (extern debug (vec<tree, va_gc> &)): New.
22589 (extern debug (vec<tree, va_gc> *)): New.
22591 (debug (struct loop &)): New.
22592 (debug (struct loop *)): New.
22593 (debug_verbose (struct loop &)): New.
22594 (debug_verbose (struct loop *)): New.
22595 * tree-dump.c: Add header dependence.
22597 (extern debug (struct loop &)): New.
22598 (extern debug (struct loop *)): New.
22599 (extern debug_verbose (struct loop &)): New.
22600 (extern debug_verbose (struct loop *)): New.
22602 (debug (data_reference &)): New.
22603 (debug (data_reference *)): New.
22604 (debug (vec<data_reference_p> &)): New.
22605 (debug (vec<data_reference_p> *)): New.
22606 (debug (vec<ddr_p> &)): New.
22607 (debug (vec<ddr_p> *)): New.
22609 (extern debug (data_reference &)): New.
22610 (extern debug (data_reference *)): New.
22611 (extern debug (vec<data_reference_p> &)): New.
22612 (extern debug (vec<data_reference_p> *)): New.
22613 (extern debug (vec<ddr_p> &)): New.
22614 (extern debug (vec<ddr_p> *)): New.
22616 (debug (pt_solution &)): New.
22617 (debug (pt_solution *)): New.
22619 (extern debug (pt_solution &)): New.
22620 (extern debug (pt_solution *)): New.
22622 (debug (_var_map &)): New.
22623 (debug (_var_map *)): New.
22624 (debug (tree_live_info_d &)): New.
22625 (debug (tree_live_info_d *)): New.
22627 (extern debug (_var_map &)): New.
22628 (extern debug (_var_map *)): New.
22629 (extern debug (tree_live_info_d &)): New.
22630 (extern debug (tree_live_info_d *)): New.
22632 2013-03-28 Jan Hubicka <jh@suse.cz>
22634 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
22636 2013-03-28 Ian Bolton <ian.bolton@arm.com>
22638 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
22639 record only when desired or required.
22641 2013-03-28 Uros Bizjak <ubizjak@gmail.com>
22643 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
22644 *vec_extractv2di_1_rex64. Use x64 isa attribute.
22646 2013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22648 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
22649 (*andsi3_compare0_uxtw): New pattern.
22650 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
22651 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
22653 2013-03-28 Jan Hubicka <jh@suse.cz>
22655 * data-streamer-in.c (streamer_read_gcov_count): New function.
22656 * gimple-streamer-out.c: Include value-prof.h.
22657 (output_gimple_stmt): Output histogram.
22658 (output_bb): Use streamer_write_gcov_count.
22659 * value-prof.c: Include data-streamer.h
22660 (dump_histogram_value): Add HIST_TYPE_MAX.
22661 (stream_out_histogram_value): New function.
22662 (stream_in_histogram_value): New function.
22663 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
22664 (stream_out_histogram_value, stream_in_histogram_value): Declare.
22665 * data-streamer-out.c (streamer_write_gcov_count): New function.
22666 (streamer_write_gcov_count_stream): New function.
22667 * lto-cgraph.c (lto_output_edge): Update counter streaming.
22668 (lto_output_node): Likewise.
22669 (input_node, input_edge): Likewise.
22670 * lto-streamer-out.c (output_cfg): Update streaming.
22671 * lto-streamer-in.c (input_cfg): Likewise.
22672 * data-streamer.h (streamer_write_gcov_count,
22673 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
22674 * gimple-streamer-in.c: Include value-prof.h
22675 (input_gimple_stmt): Input histograms.
22676 (input_bb): Update profile streaming.
22678 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
22680 * genmodes.c (emit_max_int): New function.
22681 (emit_insn_modes_h): Added call to emit_max_function.
22682 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
22684 * machmode.def: Fixed comment.
22686 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
22688 * combine.c (try_combine): Removed useless assert.
22689 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
22691 2013-03-28 Marek Polacek <polacek@redhat.com>
22692 Richard Biener <rguenther@suse.de>
22694 PR tree-optimization/56695
22695 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
22696 build signed result of a vector comparison.
22697 * tree-cfg.c (verify_gimple_comparison): Check that a result
22698 of a vector comparison has signed type.
22700 2013-03-28 Richard Biener <rguenther@suse.de>
22702 PR tree-optimization/37021
22703 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
22704 do not restrict gaps between groups.
22705 * tree-vect-stmts.c (vectorizable_load): Properly account for
22706 a gap between groups.
22708 2013-03-28 Eric Botcazou <ebotcazou@adacore.com>
22710 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
22711 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
22714 2013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
22716 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
22717 * doc/extend.texi (Named Address Spaces): Ditto.
22718 (Variable Attributes): Ditto.
22720 2013-03-27 Kai Tietz <ktietz@redhat.com>
22722 * config.build: Add support for cygwin x64 target.
22723 * config.gcc: Likewise.
22724 * config.host: Likewise.
22725 * configure.ac: Likewise
22726 * configure: Regenerated.
22728 2013-03-27 Kai Tietz <ktietz@redhat.com>
22730 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
22731 * config/i386/t-cygwin-w64: New file.
22732 * config/i386/cygwin-w64.h: New file.
22733 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
22734 and add support for x64-cygwin target.
22735 (CPP_SPEC): Likewise.
22736 (CXX_WRAP_SPEC_LIST): Undefine before define.
22737 (LIBGCJ_SONAME): Use 15 as version.
22739 2013-03-27 Richard Biener <rguenther@suse.de>
22741 PR tree-optimization/56716
22742 * tree-ssa-structalias.c (perform_var_substitution): Adjust
22743 dumping for ref nodes.
22745 2013-03-27 Martin Jambor <mjambor@suse.cz>
22747 PR tree-optimization/55334
22748 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
22749 restricted pointers to arrays.
22751 2013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
22753 * Makefile.in (.SUFFIXES): Add .cc.
22754 (.c.o): Apply same recipe for implicit rule .cc.o.
22756 2013-03-27 Richard Biener <rguenther@suse.de>
22758 PR tree-optimization/37021
22759 * tree-vect-data-refs.c (vect_check_strided_load): Allow
22760 REALPART/IMAGPART_EXPRs around the supported refs.
22761 * tree-ssa-structalias.c (find_func_aliases): Assume that
22762 floating-point values are not used to transfer pointers.
22764 2013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
22766 * target.def (TARGET_HAS_IFUNC_P): New target hook.
22767 * doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
22768 * doc/tm.texi: Regenerate.
22769 * targhooks.h (default_has_ifunc_p): New.
22770 * targhooks.c (default_has_ifunc_p): Ditto.
22771 * config/linux-protos.h: New file.
22772 * config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of this
22773 hook for linux which disables support of indirect functions in android.
22774 * config/linux-android.c: New file.
22775 * config/t-linux-android.c: Ditto.
22776 * config.gcc: Added new object file linux-android.o.
22777 * config/i386/i386.c (ix86_get_function_versions_dispatcher):
22778 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
22779 * varasm.c (do_assemble_alias): Likewise.
22780 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
22781 doesn't support indirect functions.
22782 * configure: Regenerate.
22784 2013-03-27 Bin Cheng <bin.cheng@arm.com>
22787 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
22788 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
22791 2013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
22793 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
22795 2013-03-27 Terry Guo <terry.guo@arm.com>
22797 * config/arm/arm-cores.def: Added core cortex-r7.
22798 * config/arm/arm-tune.md: Regenerated.
22799 * config/arm/arm-tables.opt: Regenerated.
22800 * doc/invoke.texi: Added entry for core cortex-r7.
22802 2013-03-27 Walter Lee <walt@tilera.com>
22804 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
22805 double-decrement of next_scratch_regno.
22807 2013-03-27 Walter Lee <walt@tilera.com>
22809 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
22811 (insn_v1mulus): Ditto.
22812 (insn_v2muls): Ditto.
22814 2013-03-27 Walter Lee <walt@tilera.com>
22816 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
22817 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
22819 2013-03-27 Walter Lee <walt@tilera.com>
22821 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
22822 (*sibcall_value): Ditto.
22824 2013-03-27 Walter Lee <walt@tilera.com>
22826 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
22827 (insn_mnz_v8qi): ... this ...
22828 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
22830 (insn_v<n>mnz): Replaced by ...
22831 (insn_v1mnz): ... this ...
22832 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
22834 (insn_mz_<mode>): Replaced by ...
22835 (insn_mz_v8qi): ... this ...
22836 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
22838 (insn_v<n>mz): Replaced by ...
22839 (insn_v1mz): ... this ...
22840 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
22843 2013-03-26 Eric Botcazou <ebotcazou@adacore.com>
22845 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
22847 2013-03-26 Roland McGrath <mcgrathr@google.com>
22849 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
22850 than fprintf with a non-constant, non-format string.
22852 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
22854 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
22855 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
22856 operand 0 predicate.
22857 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
22858 attribute. Use general_x64nomem_operand as operand 1 predicate.
22859 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
22860 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
22861 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
22862 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
22863 (mov<mode>_insv_1): Remove expander. Merge insn with
22864 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
22865 Use general_x64nomem_operand as operand 1 predicate.
22866 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
22867 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
22868 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
22869 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
22870 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
22871 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
22872 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
22873 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
22874 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
22875 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
22876 (general_x64nomem_operand): Ditto.
22878 2013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
22880 * config/rtems.opt: Add -pthread option.
22882 2013-03-26 Richard Biener <rguenther@suse.de>
22884 * alias.c (find_base_term): Avoid redundant and not used recursion.
22885 (base_alias_check): Get the initial base term from the caller.
22886 (true_dependence_1): Compute and pass base terms to base_alias_check.
22887 (write_dependence_p): Likewise.
22888 (may_alias_p): Likewise.
22890 2013-03-26 Sofiane Naci <sofiane.naci@arm.com>
22892 * config/aarch64/aarch64.c (aarch64_classify_address): Support
22893 PC-relative load in SI modes and above only.
22895 2013-03-26 Xinyu Qi <xyqi@marvell.com>
22897 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
22898 * config/arm/iwmmxt.md (WCGR0): Update.
22899 (WCGR1, WCGR2, WCGR3): Likewise.
22901 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
22903 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
22904 Use x64 and nox64 isa attributes.
22906 2013-03-26 Richard Biener <rguenther@suse.de>
22908 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
22909 alignment computations and rely on get_object_alignment_1
22910 for the !TYPE_P case.
22911 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
22913 2013-03-26 Walter Lee <walt@tilera.com>
22915 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
22916 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
22918 2013-03-25 Jeff Law <law@redhat.com>
22920 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
22921 check for INTEGRAL_TYPE_P that was missing due to checking in
22922 wrong version of prior patch.
22924 2013-03-25 Walter Lee <walt@tilera.com>
22926 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
22927 TILEGX_INSN_SHUFFLEBYTES1.
22928 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
22930 (tilegx_builtins): Ditto.
22931 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
22933 2013-03-25 Walter Lee <walt@tilera.com>
22935 * config/tilegx/tilegx.md (floatsisf2): New pattern.
22936 (floatunssisf2): New pattern.
22937 (floatsidf2): New pattern.
22938 (floatunssidf2): New pattern.
22940 2013-03-25 Walter Lee <walt@tilera.com>
22942 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
22943 tests for constraint J, K, N, P.
22945 2013-03-25 Walter Lee <walt@tilera.com>
22947 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
22948 Use indirect/pcrel encoding.
22949 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
22952 2013-03-25 Steve Ellcey <sellcey@mips.com>
22954 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
22955 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
22956 * config/mips/mips.c (mips_option_override): Set IMADD default.
22957 * config/mips/mips.h (PTF_AVOID_IMADD): New.
22958 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
22959 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
22960 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
22961 * doc/invoke.texi (-mimadd/-mno-imadd): New.
22963 2013-03-25 Jeff Law <law@redhat.com>
22965 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
22966 slightly to avoid creating and folding useless trees. Simplify
22967 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
22969 2013-03-25 Uros Bizjak <ubizjak@gmail.com>
22971 * config/i386/i386.md (*zero_extendsidi2): Merge with
22972 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
22973 * config/i386/predicates.md (x86_64_zext_operand): Rename from
22974 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
22975 targets. Clarify comment.
22977 2013-03-25 Martin Jambor <mjambor@suse.cz>
22979 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
22980 pass-through jump functions differently.
22981 (ipa_read_jump_function): Likewise. Also use setter functions to set
22984 2013-03-25 Martin Jambor <mjambor@suse.cz>
22986 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
22987 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
22989 (ipa_get_indirect_edge_target): New function.
22990 (devirtualization_time_bonus): New parameter known_aggs, pass it to
22991 ipa_get_indirect_edge_target. Update all callers.
22992 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
22993 ipa_get_indirect_edge_target_1 instead of calling
22994 ipa_get_indirect_edge_target.
22995 (create_specialized_node): Pass aggvlas to
22996 ipcp_discover_new_direct_edges.
22998 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23000 * config/arm/arm.md (f_sels, f_seld): New types.
23001 (*cmov<mode>): New pattern.
23002 * config/arm/predicates.md (arm_vsel_comparison_operator): New
23005 2013-03-25 Kai Tietz <ktietz@redhat.com>
23007 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
23008 POSIX-printf for mingw-hosted builds.
23010 2013-03-25 Richard Biener <rguenther@suse.de>
23012 PR middle-end/56694
23013 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
23014 must-not-throw stmt location.
23016 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23018 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
23019 Emit load-acquire versions when acq is true.
23020 (arm_emit_store_exclusive): Add rel parameter.
23021 Emit store-release versions when rel is true.
23022 (arm_split_compare_and_swap): Use acquire-release instructions
23024 of barriers when appropriate.
23025 (arm_split_atomic_op): Likewise.
23026 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
23027 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
23028 (VUNSPEC_SLX): Likewise.
23029 (VUNSPEC_LDA): Likewise.
23030 (VUNSPEC_STL): Likewise.
23031 * config/arm/sync.md (atomic_load<mode>): New pattern.
23032 (atomic_store<mode>): Likewise.
23033 (arm_load_acquire_exclusive<mode>): Likewise.
23034 (arm_load_acquire_exclusivesi): Likewise.
23035 (arm_load_acquire_exclusivedi): Likewise.
23036 (arm_store_release_exclusive<mode>): Likewise.
23038 2013-03-25 Catherine Moore <clm@codesourcery.com>
23040 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
23041 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
23042 * config/mip/predicates.md (lwsp_swsp_operand,
23043 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
23044 sb16_operand, db4_operand, db7_operand, ib3_operand,
23045 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
23046 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
23047 andi16_operand): New predicates.
23048 * config/mips/mips.md (compression): New attribute.
23049 (enabled): New attribute.
23050 (length): Consider compression in computing length.
23051 (shift_compression): New code attribute.
23052 (*add<mode>3): New operands. Record compression.
23053 (sub<mode>3): Likewise.
23054 (one_cmpl<mode>2): Likewise.
23055 (*and<mode>3): Likewise.
23056 (*ior<mode>3): Likewise.
23057 (unnamed pattern for xor): Likewise.
23058 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
23059 (*<optab><mode>3): Likewise.
23060 (*mov<mode>_internal: Likewise.
23061 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
23062 (mips_unsigned_immediate_p): New.
23063 (umips_lwsp_swsp_address_p): New.
23064 (m16_based_address_p): New.
23065 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
23066 (mips_unsigned_immediate_p): New prototype.
23067 (lwsp_swsp_address_p): New prototype.
23068 (m16_based_address_p): New prototype.
23069 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
23070 (mips_signed_immediate_p): New function.
23071 (m16_based_address_p): New function.
23072 (lwsp_swsp_address_p): New function.
23073 (mips_print_operand_punctuation): Recognize short delay slot insns
23074 for microMIPS.add<mode>3"
23076 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23079 * config/arm/iterators.md (v_cmp_result): New mode attribute.
23080 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
23082 2013-03-25 Richard Biener <rguenther@suse.de>
23084 PR tree-optimization/56689
23085 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
23088 2013-03-25 Richard Biener <rguenther@suse.de>
23090 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
23092 (memory_references): Likewise.
23093 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
23094 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
23095 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
23096 (gather_mem_refs_in_loops): Fold into ...
23097 (analyze_memory_references): ... this. Move initialization
23098 to tree_ssa_lim_initialize.
23099 (fill_always_executed_in): Rename to ...
23100 (fill_always_executed_in_1): ... this.
23101 (fill_always_executed_in): Move contains_call computation to
23102 this new function from ...
23103 (tree_ssa_lim_initialize): ... here.
23104 (tree_ssa_lim): Call fill_always_executed_in.
23106 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
23108 * postreload.c (reload_combine): Fix code detecting returns.
23110 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
23112 * function.c (emit_use_return_register_into_block): On cc0 targets,
23113 do not emit the sequence between cc0 setter and user.
23115 2013-03-25 Kai Tietz <ktietz@redhat.com>
23117 * config/i386/predicates.md (local_symbolic_operand): Interpret
23118 dll-imported symbols as none-local.
23120 2013-03-25 Richard Biener <rguenther@suse.de>
23122 * tree-ssa-loop-im.c (struct depend): Remove.
23123 (struct lim_aux_data): Make depends a vec of gimples.
23124 (free_lim_aux_data): Adjust.
23125 (add_dependency): Likewise.
23126 (set_level): Likewise.
23128 2013-03-25 Richard Biener <rguenther@suse.de>
23130 PR middle-end/56434
23131 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
23132 the pointer returned by calls with ECF_MALLOC set.
23134 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
23136 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
23138 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
23140 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
23141 using MMXMODE mode iterator.
23142 (*move<mode>_internal): Merge with *movv2sf_internal and
23143 *movv2sf_internal_rex64 using MMXMODE mode iterator.
23145 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
23147 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
23148 (record_last_mem_set_info): Likewise.
23150 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
23151 of XNEWVEC followed by memset.
23152 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
23154 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
23156 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
23157 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
23158 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
23159 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
23160 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
23161 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
23162 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
23163 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
23164 BARRIER_P instead of GET_CODE.
23166 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
23168 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
23169 inaccuracy in the probing code.
23171 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
23172 (ctrapdi4): Likewise.
23174 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
23176 * calls.c (expand_call): Add missing guard to code handling return
23177 of non-BLKmode structures in MSB.
23178 * function.c (expand_function_end): Likewise.
23180 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
23182 * combine.c (try_combine): Adjust comment. Do not add the set of
23183 insn #0 if the destination indirectly is set or dies in insn #2.
23184 Tidy up code to distribute a new note.
23186 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
23188 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
23189 also for alternatives 16 and 17.
23191 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
23193 * config/i386/sse.md (*mov<mode>_internal): Merge with
23194 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
23195 Emit insn template depending on type attribute. Use
23196 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
23197 movd instead of movq mnemonic for interunit moves. Rewrite mode
23198 attribute calculation. Remove unit attribute calculation.
23199 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
23200 Set prefix_data16 attribute for DImode ssemov types.
23201 Use Ym instead of y for SSE-MMX conversion alternatives.
23202 Reorder operand constraints.
23204 2013-03-22 Steven Bosscher <steven@gcc.gnu.org>
23206 * df.h (df_insn_delete): Adjust prototype.
23207 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
23208 and let it decide whether mark the basic block dirty.
23209 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
23210 * df-scan.c (df_insn_info_delete): New helper function, split
23211 off from df_insn_delete.
23212 (df_scan_free_bb_info): Use it.
23213 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
23215 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
23216 that the insn is actually an insn and it has a non-NULL basic block.
23217 Do not mark basic block dirty if only deleting a DEBUG_INSN.
23219 2013-03-22 Richard Biener <rguenther@suse.de>
23221 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
23223 (mem_ref_alloc): Do not allocate them.
23224 (refs_independent_p): Do not query or maintain a cache.
23226 2013-03-22 Richard Biener <rguenther@suse.de>
23228 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
23229 (gather_mem_refs_in_loops): Do not compute it.
23230 (analyze_memory_references): Do not allocate it.
23231 (tree_ssa_lim_finalize): Do not free it.
23232 (for_all_locs_in_loop): Do not query all_refs_in_loop.
23234 2013-03-22 Richard Biener <rguenther@suse.de>
23236 * is-a.h (as_a): Use gcc_checking_assert.
23238 2013-03-22 Ian Bolton <ian.bolton@arm.com>
23240 * config/aarch64/aarch64.c (aarch64_print_operand): New
23241 format specifier for printing a constant in hex.
23242 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
23243 format specifier for printing second operand.
23245 2013-03-22 Richard Biener <rguenther@suse.de>
23247 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
23249 (gather_mem_refs_in_loops): Perform store accumulation here.
23250 (create_vop_ref_mapping_loop): Remove.
23251 (create_vop_ref_mapping): Likewise.
23252 (analyze_memory_references): Initialize refs_stored_in_loop.
23253 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
23254 (record_indep_loop): Remove.
23255 (record_dep_loop): New function.
23256 (ref_indep_loop_p_1): Adjust to only walk over references
23257 in the loop, not its subloops.
23258 (ref_indep_loop_p): Rename to ...
23259 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
23260 maintaining a more fine-grained cache.
23261 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
23262 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
23264 2013-03-22 Richard Biener <rguenther@suse.de>
23266 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
23267 (struct mem_ref): Make accesses_in_loop a vec of a vec of
23268 aggregate mem_ref_loc.
23269 (free_mem_ref_locs): Inline into ...
23270 (memref_free): ... this and adjust.
23271 (mem_ref_alloc): Adjust.
23272 (mem_ref_locs_alloc): Remove.
23273 (record_mem_ref_loc): Adjust.
23274 (get_all_locs_in_loop): Rewrite into ...
23275 (for_all_locs_in_loop): ... this iterator.
23276 (rewrite_mem_ref_loc): New functor.
23277 (rewrite_mem_refs): Use for_all_locs_in_loop.
23278 (sm_set_flag_if_changed): New functor.
23279 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
23280 (ref_always_accessed): New functor.
23281 (ref_always_accessed_p): Use for_all_locs_in_loop.
23283 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
23285 * tree-pass.h (PROP_gimple_lvec): New.
23286 * passes.c (dump_properties): Handle PROP_gimple_lvec.
23287 (init_optimization_passes): Move pass_lower_vector.
23288 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
23290 (pass_lower_vector): Provide PROP_gimple_lvec.
23291 (pass_lower_vector_ssa): Likewise.
23292 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
23294 2013-03-21 Mark Wielaard <mjw@redhat.com>
23296 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
23298 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
23300 * config/i386/i386.md (*movdi_internal): Disparage slightly
23301 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
23302 conversion alternatives.
23304 2013-03-21 Jakub Jelinek <jakub@redhat.com>
23306 PR middle-end/48087
23307 * diagnostic.def (DK_WERROR): New kind.
23308 * diagnostic.h (werrorcount): Define.
23309 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
23310 promoted to DK_ERROR, increment DK_WERROR counter instead of
23312 * toplev.c (toplev_main): Call print_ignored_options even if
23313 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
23314 even if just werrorcount is non-zero.
23317 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
23319 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
23320 (string_cst_pool_decl): New function.
23321 (optimize_one_addr_into_implicit_ptr): New function.
23322 (resolve_addr_in_expr): Optimize DWARF location expression
23323 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
23324 which doesn't live in memory, but has DW_AT_location or
23325 DW_AT_const_value, or refers to a string literal, into
23326 DW_OP_GNU_implicit_pointer.
23327 (optimize_location_into_implicit_ptr): New function.
23328 (resolve_addr): If removing DW_AT_location of a variable because
23329 it was DW_OP_addr of address of the variable, but the variable doesn't
23330 live in memory, try to emit const value attribute for the initializer.
23332 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
23334 * tree.h (VECTOR_TYPE_P): New macro.
23335 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
23336 TYPE_MODE): Use it.
23337 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
23338 VEC_COND_EXPR cannot be lvalues.
23339 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
23341 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
23343 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
23344 Restrict the transformation to equal modes.
23346 2013-03-21 Richard Biener <rguenther@suse.de>
23348 PR tree-optimization/39326
23349 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
23350 (MEM_ANALYZABLE): Adjust.
23351 (record_mem_ref_loc): Move bitmap ops ...
23352 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
23353 unanalyzable refs, do not record locations for it.
23354 (analyze_memory_references): Allocate ref zero as shared
23356 (refs_independent_p): Do not test for unanalyzed mems here.
23357 (ref_indep_loop_p_1): Special-case disambiguation against
23358 the unanalyzed ref.
23359 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
23361 2013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
23363 * config/arm/arm-protos.h (tune_params): Add
23364 prefer_neon_for_64bits field.
23365 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
23366 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
23367 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
23368 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
23369 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
23370 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
23371 (arm_option_override): Handle -mneon-for-64bits new option.
23372 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
23373 (prefer_neon_for_64bits): Declare new variable.
23374 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
23375 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
23376 (arch_enabled): Handle new arch types. Remove support for onlya8
23378 (one_cmpldi2): Use new arch names.
23379 (zero_extend<mode>di2, extend<mode>di2): Ditto.
23380 * config/arm/arm.opt (mneon-for-64bits): Add option.
23381 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
23382 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
23383 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
23385 * doc/invoke.texi (-mneon-for-64bits): Document.
23387 2013-03-21 Richard Biener <rguenther@suse.de>
23389 PR tree-optimization/39326
23390 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
23391 (sort_bbs_in_loop_postorder_cmp): New function.
23392 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
23394 2013-03-21 Richard Biener <rguenther@suse.de>
23396 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
23397 (vect_insert_into_interleaving_chain): Likewise.
23398 (vect_drs_dependent_in_basic_block): Inline ...
23399 (vect_slp_analyze_data_ref_dependence): ... here. New function,
23401 (vect_analyze_data_ref_dependence): ... here. Simplify.
23402 (vect_check_interleaving): Simplify.
23403 (vect_analyze_data_ref_dependences): Likewise. Split out ...
23404 (vect_slp_analyze_data_ref_dependences): ... this new function.
23405 (dr_group_sort_cmp): New function.
23406 (vect_analyze_data_ref_accesses): Compute data-reference groups
23407 here instead of in vect_analyze_data_ref_dependence. Use
23408 a more efficient algorithm.
23409 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
23410 vect_slp_analyze_data_ref_dependences. Call
23411 vect_analyze_data_ref_accesses earlier.
23412 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
23413 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
23414 (vect_slp_analyze_data_ref_dependences): New prototype.
23416 2013-03-21 Richard Biener <rguenther@suse.de>
23418 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
23419 ref is stored in the loop.
23420 (find_refs_for_sm): Walk only over all stores.
23421 (store_motion_loop): Allocate from lim_bitmap_obstack.
23422 (store_motion): Likewise.
23424 2013-03-21 Richard Biener <rguenther@suse.de>
23426 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
23427 Update virtual SSA form.
23429 2013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
23431 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
23432 * configure: Regenerate.
23433 * config.in: Regenerate.
23434 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
23435 if !HAVE_LD_EH_FRAME_CIEV3.
23437 2013-03-21 Richard Biener <rguenther@suse.de>
23439 * tree-cfg.c (verify_expr_no_block): New function.
23440 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
23441 nor DECL_VALUE_EXPR have locations with associated blocks.
23442 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
23443 (clear_unused_block_pointer): Remove code dealing with
23444 blocks in DECL_DEBUG_EXPR locations.
23446 2013-03-21 Richard Biener <rguenther@suse.de>
23448 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
23449 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
23450 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
23451 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
23452 instead of DECL_DEBUG_EXPR_IS_FROM.
23453 * gimplify.c (gimplify_modify_expr): Likewise.
23454 * tree-cfg.c (verify_expr_location_1): Likewise.
23455 * tree-complex.c (create_one_component_var): Likewise.
23456 * tree-sra.c (create_access_replacement): Likewise.
23457 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
23458 (clear_unused_block_pointer): Likewise.
23459 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
23460 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
23461 * var-tracking.c (var_debug_decl): Likewise.
23462 (track_expr_p): Likewise.
23463 * tree-inline.c (add_local_variables): Likewise. Set
23464 DECL_HAS_DEBUG_EXPR_P after copying it.
23465 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
23466 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
23468 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
23471 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
23472 * configure: Regenerate.
23473 * config.in: Regenerate.
23474 * config/i386/i386.md (*movdf_internal): Use
23475 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
23476 movd instead of movq mnemonic for interunit moves.
23477 (*movdi_internal): Ditto.
23479 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
23481 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
23482 (abd<mode>_3): New pattern.
23483 (aba<mode>_3): New pattern.
23484 (fabd<mode>_3): New pattern.
23486 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
23488 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
23489 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
23490 occurrence of REGISTER_PREFIX as its empty string.
23492 2013-03-20 Jeff Law <law@redhat.com>
23494 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
23495 addititional equivalences for equality comparisons between an SSA_NAME
23496 and a constant where the SSA_NAME was set from a widening conversion.
23498 2013-03-20 Walter Lee <walt@tilera.com>
23500 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
23502 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
23504 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
23505 depending on type attribute.
23506 (*movti_internal): Ditto.
23507 (*movtf_internal): Ditto.
23508 (*movxf_internal): Ditto.
23509 (*movdf_internal): Ditto.
23510 (*movsf_internal): Ditto.
23512 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
23514 * config/i386/i386.md (*movti_internal): Set prefix attribute to
23515 maybe_vex for sselog1 and ssemov types.
23516 (*movdi_internal): Reorder operand constraints.
23517 (*movsi_internal): Ditto. Set prefix attribute to
23518 maybe_vex for sselog1 and ssemov types.
23519 (*movtf_internal): Set prefix attribute to maybe_vex
23520 for sselog1 and ssemov types.
23521 (*movdf_internal): Ditto. Set prefix_data16 attribute for
23522 DImode ssemov types. Reorder operand constraints.
23523 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
23524 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
23525 attribute for SImode ssemov types. Reorder operand constraints.
23527 2013-03-20 Martin Jambor <mjambor@suse.cz>
23529 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
23530 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
23532 2013-03-20 Pat Haugen <pthaugen@us.ibm.com>
23534 * config/rs6000/predicates.md (indexed_address, update_address_mem
23535 update_indexed_address_mem): New predicates.
23536 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
23537 attribute for load/store instructions.
23538 * config/rs6000/dfp.md (movsd_store): Likewise.
23539 (movsd_load): Likewise.
23540 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
23541 (unnamed HI->DI extend define_insn): Likewise.
23542 (unnamed SI->DI extend define_insn): Likewise.
23543 (unnamed QI->SI extend define_insn): Likewise.
23544 (unnamed QI->HI extend define_insn): Likewise.
23545 (unnamed HI->SI extend define_insn): Likewise.
23546 (unnamed HI->SI extend define_insn): Likewise.
23547 (extendsfdf2_fpr): Likewise.
23548 (movsi_internal1): Likewise.
23549 (movsi_internal1_single): Likewise.
23550 (movhi_internal): Likewise.
23551 (movqi_internal): Likewise.
23552 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
23553 attribute for load/store instructions.
23554 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
23556 (mov<mode>_softfloat): Likewise.
23557 (mov<mode>_hardfloat32): Likewise.
23558 (mov<mode>_hardfloat64): Likewise.
23559 (mov<mode>_softfloat64): Likewise.
23560 (movdi_internal32): Likewise.
23561 (movdi_internal64): Likewise.
23562 (probe_stack_<mode>): Likewise.
23564 2013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
23566 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
23567 floating point, and decimal floating point to reload iterator.
23569 * config/rs6000/constraints.md (wl constraint): New constraints to
23570 return FLOAT_REGS if certain options are used to reduce the number
23571 of separate patterns that exist in the file.
23572 (wx constraint): Likewise.
23573 (wz constraint): Likewise.
23575 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
23576 -mdebug=reg, print wg, wl, wx, and wz constraints.
23577 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
23578 Initialize the reload functions for 64-bit binary/decimal floating
23580 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
23581 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
23582 create the buffer on the stack to overcome not having a 32-bit
23584 (rs6000_emit_move): Likewise.
23585 (rs6000_secondary_memory_needed_rtx): Likewise.
23586 (rs6000_alloc_sdmode_stack_slot): Likewise.
23587 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
23588 via xxlxor, just like DFmode 0.0.
23590 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
23591 define as 1 if we are running on a power7 or newer.
23592 (enum r6000_reg_class_enum): Add new constraints.
23594 * config/rs6000/dfp.md (movsd): Delete, combine with binary
23595 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
23596 with other moves by using conditional constraits (wg). Use LFIWZX
23597 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
23598 (movsd splitter): Likewise.
23599 (movsd_hardfloat): Likewise.
23600 (movsd_softfloat): Likewise.
23602 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
23603 binary and decimal floating point moves.
23604 (fmove_ok): New attributes to combine binary and decimal floating
23605 point moves, and to combine power6x (mfpgpr) moves along normal
23607 (real_value_to_target): Likewise.
23608 (f32_lr): Likewise.
23609 (f32_lm): Likewise.
23610 (f32_li): Likewise.
23611 (f32_sr): Likewise.
23612 (f32_sm): Likewise.
23613 (f32_si): Likewise.
23614 (movsf): Combine binary and decimal floating point moves. Combine
23615 power6x (mfpgpr) moves with other moves by using conditional
23616 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
23617 (mov<mode> for SFmode/SDmode); Likewise.
23618 (SFmode/SDmode splitters): Likewise.
23619 (movsf_hardfloat): Likewise.
23620 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
23621 (movsf_softfloat): Likewise.
23622 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
23624 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
23625 wx and wz constraints.
23627 * config/rs6000/constraints.md (wg constraint): New constraint to
23628 return FLOAT_REGS if -mmfpgpr (power6x) was used.
23630 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
23633 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
23634 -mdebug=reg, print wg, wl, wx, and wz constraints.
23635 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
23636 Initialize the reload functions for 64-bit binary/decimal floating
23638 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
23639 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
23640 create the buffer on the stack to overcome not having a 32-bit
23642 (rs6000_emit_move): Likewise.
23643 (rs6000_secondary_memory_needed_rtx): Likewise.
23644 (rs6000_alloc_sdmode_stack_slot): Likewise.
23645 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
23646 via xxlxor, just like DFmode 0.0.
23648 * config/rs6000/dfp.md (movdd): Delete, combine with binary
23649 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
23650 with other moves by using conditional constraits (wg). Use LFIWZX
23651 and STFIWX for loading SDmode on power7.
23652 (movdd splitters): Likewise.
23653 (movdd_hardfloat32): Likewise.
23654 (movdd_softfloat32): Likewise.
23655 (movdd_hardfloat64_mfpgpr): Likewise.
23656 (movdd_hardfloat64): Likewise.
23657 (movdd_softfloat64): Likewise.
23659 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
23660 64-bit binary and decimal floating point moves.
23661 (FMOVE64X): Likewise.
23662 (movdf): Combine 64-bit binary and decimal floating point moves.
23663 Combine power6x (mfpgpr) moves with other moves by using
23664 conditional constraits (wg).
23665 (mov<mode> for DFmode/DDmode): Likewise.
23666 (DFmode/DDmode splitters): Likewise.
23667 (movdf_hardfloat32): Likewise.
23668 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
23669 (movdf_softfloat32): Likewise.
23670 (movdf_hardfloat64_mfpgpr): Likewise.
23671 (movdf_hardfloat64): Likewise.
23672 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
23673 (movdf_softfloat64): Likewise.
23674 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
23675 (reload_<mode>_load): Move to later in the file so they aren't in
23676 the middle of the floating point move insns.
23677 (reload_<mode>_store): Likewise.
23679 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
23682 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
23683 constraint if -mdebug=reg.
23684 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
23685 Enable using dd reload support if needed.
23687 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
23688 binary and decimal floating point moves in rs6000.md.
23689 (movtd_internal): Likewise.
23691 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
23692 decimal floating point moves.
23694 (movtf_internal): Likewise.
23695 (mov<mode>_internal, TDmode/TFmode): Likewise.
23696 (movtf_softfloat): Likewise.
23697 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
23699 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
23700 movdi_internal64, using wg constraint for move direct operations.
23701 (movdi_internal64): Likewise.
23703 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
23704 MODES_TIEABLE_P for selected modes. Print the numerical value of
23705 the various virtual registers. Use GPR/FPR first/last values,
23706 instead of hard coding the register numbers. Print which modes
23707 have reload functions registered.
23708 (rs6000_option_override_internal): If -mdebug=reg, trace the options
23709 settings before/after setting cpu, target and subtarget settings.
23710 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
23711 and for secondary reload failures in rs6000_secondary_reload_inner.
23712 (rs6000_secondary_reload_fail): Likewise.
23713 (rs6000_secondary_reload_inner): Likewise.
23715 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
23716 macros for first/last GPR and FPR registers.
23717 (LAST_GPR_REGNO): Likewise.
23718 (FIRST_FPR_REGNO): Likewise.
23719 (LAST_FPR_REGNO): Likewise.
23721 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
23722 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
23723 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
23724 (vcond<mode><mode>): Likewise.
23725 (vcondu<mode><mode>): Likewise.
23726 (vector_gtu<mode>): Likewise.
23727 (vector_gte<mode>): Likewise.
23728 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
23729 to prevent the compiler from converting DImode operations to TImode.
23730 (ior<mode>3): Likewise.
23731 (and<mode>3): Likewise.
23732 (one_cmpl<mode>2): Likewise.
23733 (nor<mode>3): Likewise.
23734 (andc<mode>3): Likewise.
23736 * config/rs6000/constraints.md (wt constraint): New constraint
23737 that returns VSX_REGS if TImode is allowed in VSX registers.
23739 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
23740 constant under VSX.
23742 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
23743 similar to TImode, but it is restricted to being in the GPRs.
23745 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
23746 TImode to occupy a single VSX register.
23748 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
23749 -mvsx-timode for power7/power8.
23750 (power7 cpu): Likewise.
23751 (power8 cpu): Likewise.
23753 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
23754 sure that TFmode/TDmode take up two registers if they are ever
23755 allowed in the upper VSX registers.
23756 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
23758 (rs6000_init_hard_regno_mode_ok): Likewise.
23759 (rs6000_debug_reg_global): Add debugging for PTImode and wt
23760 constraint. Print if LRA is turned on.
23761 (rs6000_option_override_internal): Give an error if -mvsx-timode
23762 and VSX is not enabled.
23763 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
23764 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
23765 to reg+offset addressing. Use PTImode when checking offset
23766 addresses for validity.
23767 (reg_offset_addressing_ok_p): Likewise.
23768 (rs6000_legitimate_offset_address_p): Likewise.
23769 (rs6000_legitimize_address): Likewise.
23770 (rs6000_legitimize_reload_address): Likewise.
23771 (rs6000_legitimate_address_p): Likewise.
23772 (rs6000_eliminate_indexed_memrefs): Likewise.
23773 (rs6000_emit_move): Likewise.
23774 (rs6000_secondary_reload): Likewise.
23775 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
23776 reloads to fpr registers to continue to use reg+offset addressing,
23777 but 64-bit reloads to altivec registers need reg+reg addressing.
23778 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
23779 it. Treat LO_SUM like a PLUS operation.
23780 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
23781 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
23782 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
23783 registers to share a register with a smaller sized type, since VSX
23784 puts scalars in the upper 64-bits.
23785 (print_operand): Add support for PTImode.
23786 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
23787 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
23788 registers, but don't have arithmetic support.
23789 (rs6000_memory_move_cost): Add test for VSX.
23790 (rs6000_opt_masks): Add -mvsx-timode.
23792 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
23795 (VSr): Use wt constraint for TImode.
23796 (VSv): Drop TImode support.
23797 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
23798 (vsx_movti_64bit): Likewise.
23799 (vsx_movti_32bit): Likewise.
23800 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
23801 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
23802 one '?' on the appropriate output constraint. Do not allow TImode
23803 logical operations on 32-bit systems.
23804 (vsx_ior<mode>3): Likewise.
23805 (vsx_xor<mode>3): Likewise.
23806 (vsx_one_cmpl<mode>2): Likewise.
23807 (vsx_nor<mode>3): Likewise.
23808 (vsx_andc<mode>3): Likewise.
23809 (vsx_concat_<mode>): Likewise.
23810 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
23812 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
23813 OPTION_MASK_VSX_TIMODE.
23814 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
23815 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
23817 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
23818 (TI2 iterator): New iterator for TImode, PTImode.
23819 (wd mode attribute): Add values for vector types.
23820 (movti_string): Replace TI move operations with operations for TImode
23821 and PTImode. Add support for TImode being allowed in VSX registers.
23822 (mov<mode>_string, TImode/PTImode): Likewise.
23823 (movti_ppc64): Likewise.
23824 (mov<mode>_ppc64, TImode/PTImode): Likewise.
23825 (TI mode splitters): Likewise.
23827 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
23830 2013-03-20 Marc Glisse <marc.glisse@inria.fr>
23832 PR tree-optimization/56355
23833 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
23834 Also handle integers with undefined overflow.
23836 2013-03-20 Catherine Moore <clm@codesourcery.com>
23837 Maciej W. Rozycki <macro@codesourcery.com>
23838 Tom de Vries <tom@codesourcery.com>
23839 Nathan Sidwell <nathan@codesourcery.com>
23840 Iain Sandoe <iain@codesourcery.com>
23841 Nathan Froyd <froydnj@codesourcery.com>
23842 Chao-ying Fu <fu@mips.com>
23844 * doc/extend.texi: (micromips, nomicromips, nocompression):
23845 Document new function attributes.
23846 * doc/invoke.texi (minterlink-compressed, mmicromips,
23847 m14k, m14ke, m14kec): Document new options.
23848 (minterlink-mips16): Update documentation.
23849 * doc/md.texi (ZC, ZD): Document new constraints.
23850 * configure.ac (gcc_cv_as_micromips): Check if linker
23851 supports the .set micromips directive.
23852 * configure: Regenerate.
23853 * config.in: Regenerate.
23854 * config/mips/mips-tables.opt: Regenerate.
23855 * config/mips/micromips.md: New file.
23856 * constraints.md (ZC, ZD): New constraints.
23857 * config/mips/predicates.md (movep_src_register): New predicate.
23858 (movep_src_operand): New predicate.
23859 (non_volatile_mem_operand): New predicate.
23860 * config/mips/mips.md (multimem): New type.
23861 (length): Differentiate between 17-bit and 18-bit branch offsets.
23862 (MOVEP1, MOVEP2): New mode iterator.
23863 (mov_<load>l): Use ZC constraint.
23864 (mov_<load>r): Likewise.
23865 (mov_<store>l): Likewise.
23866 (mov_<store>r): Likewise.
23867 (*branch_equality<mode>_inverted): Add microMIPS support.
23868 (*branch_equality<mode>): Likewise.
23869 (*jump_absolute): Likewise.
23870 (indirect_jump_<mode>): Likewise.
23871 (tablejump_<mode>): Likewise.
23872 (<optab>_internal): Likewise.
23873 (sibcall_internal): Likewise.
23874 (sibcall_value_internal): Likewise.
23875 (prefetch): Use constraint ZD.
23876 * config/mips/mips.opt (minterlink-compressed): New option.
23877 (minterlink-mips16): Now an alias for minterlink-compressed.
23878 (mmicromips): New option.
23879 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
23880 (compare_and_swap_12): Likewise.
23881 (sync_add<mode>): Likewise.
23882 (sync_<optab>_12): Likewise.
23883 (sync_old_<optab>_12): Likewise.
23884 (sync_new_<optab>_12): Likewise.
23885 (sync_nand_12): Likewise.
23886 (sync_old_nand_12): Likewise.
23887 (sync_new_nand_12): Likewise.
23888 (sync_sub<mode>): Likewise.
23889 (sync_old_add<mode>): Likewise.
23890 (sync_old_sub<mode>): Likewise.
23891 (sync_new_add<mode>): Likewise.
23892 (sync_new_sub<mode>): Likewise.
23893 (sync_<optab><mode>): Likewise.
23894 (sync_old_<optab><mode>): Likewise.
23895 (sync_new_<optab><mode>): Likewise.
23896 (sync_nand<mode>): Likewise.
23897 (sync_old_nand<mode>): Likewise.
23898 (sync_new_nand<mode>): Likewise.
23899 (sync_lock_test_and_set<mode>): Likewise.
23900 (test_and_set_12): Likewise.
23901 (atomic_compare_and_swap<mode>): Likewise.
23902 (atomic_exchange<mode>_llsc): Likewise.
23903 (atomic_fetch_add<mode>_llsc): Likewise.
23904 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
23905 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
23906 (umips_save_restore_pattern_p): Likewise.
23907 (umips_load_store_pair_p): Likewise.
23908 (umips_output_load_store_pair): Likewise.
23909 (umips_movep_target_p): Likewise.
23910 (umips_12bit_offset_address_p): Likewise.
23911 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
23912 (mips_base_mips16): Rename this...
23913 (mips_base_compression_flags): ...to this. Update all uses.
23914 (mips_attribute_table): Add micromips, nomicromips and nocompression.
23915 (mips_mips16_decl_p): Delete.
23916 (mips_nomips16_decl_p): Delete.
23917 (mips_get_compress_on_flags): New function.
23918 (mips_get_compress_off_flags): New function.
23919 (mips_get_compress_mode): New function.
23920 (mips_get_compress_on_name): New function.
23921 (mips_get_compress_off_name): New function.
23922 (mips_insert_attributes): Support multiple compression types.
23923 (mips_merge_decl_attributes): Likewise.
23924 (umips_12bit_offset_address_p): New function.
23925 (mips_start_function_definition): Emit .set micromips directive.
23926 (mips_call_may_need_jalx_p): New function.
23927 (mips_function_ok_for_sibcall): Add microMIPS support.
23928 (mips_print_operand_punctuation): Support short delay slots and
23930 (umips_swm_mask, umips_swm_encoding): New.
23931 (umips_build_save_restore): New function.
23932 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
23933 (was_mips16_p): Remove.
23934 (old_compression_mode): New.
23935 (mips_set_compression_mode): New function.
23936 (mips_set_current_function): Add microMIPS support.
23937 (mips_option_override): Likewise.
23938 (umips_save_restore_pattern_p): New function.
23939 (umips_output_save_restore): New function.
23940 (umips_load_store_pair_p_1): New function.
23941 (umips_load_store_pair_p): New function.
23942 (umips_output_load_store_pair_1): New function.
23943 (umips_output_load_store_pair): New function.
23944 (umips_movep_target_p) New function.
23945 (mips_prepare_pch_save): Add microMIPS support.
23946 * config/mips/mips.h (TARGET_COMPRESSION): New.
23947 (TARGET_CPU_CPP_BUILTINS): Update macro
23948 to use new compression flags and to support microMIPS.
23949 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
23950 (MIPS_ARCH_FLOAT_SPEC): Likewise.
23951 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
23952 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
23953 (ASM_SPEC): Support mmicromips and mno-micromips.
23954 (M16STORE_REG_P): New macro.
23955 (MIPS_CALL): Support TARGET_MICROMIPS.
23956 (MICROMIPS_J): New macro.
23957 (mips_base_mips16): Rename this...
23958 (mips_base_compression_flags): ...to this.
23959 (UMIPS_12BIT_OFFSET_P): New macro.
23960 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
23961 (MULTILIB_DIRNAMES): Likewise.
23962 2013-03-20 Richard Biener <rguenther@suse.de>
23964 PR tree-optimization/56661
23965 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
23966 the result does not have to be distinct.
23968 2013-03-20 Richard Biener <rguenther@suse.de>
23970 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
23973 2013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
23974 Steven Bosscher <steven@gcc.gnu.org>
23976 PR rtl-optimization/56605
23977 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
23979 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
23982 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
23983 that require movd instead of movq.
23985 2013-03-20 Richard Biener <rguenther@suse.de>
23987 * tree-ssa-structalias.c (struct variable_info): Add pointer
23988 to the first field of an aggregate with sub-vars. Make
23989 this and the pointer to the next subfield its ID.
23990 (vi_next): New function.
23991 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
23992 storedanything_id, integer_id): Increment by one.
23993 (new_var_info, get_call_vi, lookup_call_clobber_vi,
23994 get_call_clobber_vi): Adjust.
23995 (solution_set_expand): Simplify and speedup.
23996 (solution_set_add): Inline into ...
23997 (set_union_with_increment): ... this. Adjust accordingly.
23998 (do_sd_constraint): Likewise.
23999 (do_ds_constraint): Likewise.
24000 (do_complex_constraint): Simplify.
24001 (build_pred_graph): Adjust.
24002 (solve_graph): Likewise. Simplify and speedup.
24003 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
24004 get_constraint_for_component_ref, get_constraint_for_1,
24005 first_vi_for_offset, first_or_preceding_vi_for_offset,
24006 create_function_info_for, create_variable_info_for_1,
24007 create_variable_info_for, intra_create_variable_infos): Adjust.
24008 (init_base_vars): Push NULL for ID zero.
24009 (compute_points_to_sets): Adjust.
24011 2013-03-20 Richard Biener <rguenther@suse.de>
24013 * cfgloop.c (verify_loop_structure): Streamline and avoid
24014 ICEing on corrupt loop tree.
24015 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
24018 2013-03-20 Richard Biener <rguenther@suse.de>
24020 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
24021 check whether an SSA update is needed.
24023 2013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
24025 * config/mips/constraints.md (T): Rename to...
24029 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
24030 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
24032 2013-03-19 Ian Bolton <ian.bolton@arm.com>
24034 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
24035 (*subsi3_carryin_uxtw): Likewise.
24037 2013-03-19 Ian Bolton <ian.bolton@arm.com>
24039 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
24040 (*rorsi3_insn_uxtw): Likewise.
24042 2013-03-19 Ian Bolton <ian.bolton@arm.com>
24044 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
24045 (*extrsi5_insn_uxtw): Likewise.
24047 2013-03-19 Richard Biener <rguenther@suse.de>
24049 PR tree-optimization/56273
24050 * passes.c (init_optimization_passes): Move second VRP after DOM.
24052 2013-03-19 Uros Bizjak <ubizjak@gmail.com>
24054 * config/i386/i386.md (*movti_internal): Merge from
24055 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
24056 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
24057 nox64 isa attributes.
24059 2013-03-18 Richard Biener <rguenther@suse.de>
24061 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
24063 (merge_node_constraints): Likewise.
24064 (build_succ_graph): Likewise.
24065 (valid_graph_edge): Inline into single caller.
24066 (unify_nodes): Likewise. Use bitmap_set_bit return value
24068 (scc_visit): Fix formatting and variable use.
24069 (do_sd_constraint): Use gcc_checking_assert.
24070 (do_ds_constraint): Likewise.
24071 (do_complex_constraint): Likewise.
24072 (condense_visit): Likewise. Cleanup.
24073 (dump_pred_graph): New function.
24074 (perform_var_substitution): Dump the pred-graph before
24075 variable substitution.
24076 (find_equivalent_node): Use gcc_checking_assert.
24077 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
24079 2013-03-18 Richard Biener <rguenther@suse.de>
24081 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
24082 Remove cond_expr_stmt_list argument and do not gimplify the
24084 (vect_loop_versioning): Adjust.
24085 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
24086 Cleanup to use less temporaries.
24087 (vect_create_data_ref_ptr): Cleanup.
24089 2013-03-18 Jakub Jelinek <jakub@redhat.com>
24091 PR tree-optimization/56635
24092 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
24093 require types_compatible_p types.
24095 2013-03-18 Nick Clifton <nickc@redhat.com>
24097 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
24098 spurious backslash.
24100 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
24101 Add missing line to comment describing function.
24103 2013-03-18 Richard Biener <rguenther@suse.de>
24105 PR tree-optimization/56210
24106 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
24107 Handle string / character search functions.
24108 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
24110 2013-03-18 Richard Biener <rguenther@suse.de>
24112 PR middle-end/56483
24113 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
24114 and implement properly.
24115 * gimple.h (gimple_cond_single_var_p): Remove.
24117 2013-03-18 Richard Biener <rguenther@suse.de>
24119 * tree-data-ref.h (find_data_references_in_loop): Declare.
24120 * tree-data-ref.c (get_references_in_stmt): Use a stack
24121 vector pre-allocated in the callers.
24122 (find_data_references_in_stmt): Adjust.
24123 (graphite_find_data_references_in_stmt): Likewise.
24124 (create_rdg_vertices): Likewise.
24125 (find_data_references_in_loop): Export.
24126 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
24127 Compute dependences here...
24128 (vect_analyze_data_refs): ...not here. When we encounter
24129 a non-vectorizable data reference in basic-block vectorization
24130 truncate the data reference vector. Do not bother to
24131 fixup data-dependence information for gather loads.
24132 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
24133 of data references, as reported.
24135 2013-03-18 Richard Biener <rguenther@suse.de>
24137 PR tree-optimization/3713
24138 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
24139 has_constants and expr.
24140 (stmt_has_constants): Properly valueize SSA names when deciding
24141 whether the stmt has constants.
24143 2013-03-18 Richard Biener <rguenther@suse.de>
24145 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
24146 whole function when there is nothing to do.
24147 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
24148 * tree-vectorizer.c (vectorize_loops): Update virtual and
24149 loop-closed SSA once.
24150 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
24152 2013-03-18 Richard Biener <rguenther@suse.de>
24154 PR middle-end/56113
24155 * domwalk.c (bb_postorder): New global static.
24156 (cmp_bb_postorder): New function.
24157 (walk_dominator_tree): Replace scheme imposing an order for
24158 visiting dominator sons by one sorting them at the time they
24159 are pushed on the stack.
24161 2013-03-18 Richard Biener <rguenther@suse.de>
24163 PR tree-optimization/39326
24164 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
24165 (struct mem_ref): Replace mem member with ao_ref typed member.
24166 (MEM_ANALYZABLE): Adjust.
24167 (memref_eq): Likewise.
24168 (mem_ref_alloc): Likewise.
24169 (gather_mem_refs_stmt): Likewise.
24170 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
24171 (execute_sm_if_changed_flag_set): Adjust.
24172 (execute_sm): Likewise.
24173 (ref_always_accessed_p): Likewise.
24174 (refs_independent_p): Likewise.
24175 (can_sm_ref_p): Likewise.
24177 2013-03-18 Jakub Jelinek <jakub@redhat.com>
24180 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
24181 return 1 even for !unsignedp.
24183 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
24185 * config/i386/i386.md (isa): Add x64 and nox64.
24186 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
24187 (*pushtf): Enable *roF alternative for x64 isa only.
24188 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
24189 mode attribute of integer alternatives to DImode for TARGET_64BIT.
24190 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
24191 (*movtf_internal): Merge from *movtf_internal_rex64 and
24192 *movtf_internal_sse. Use x64 and nox64 isa attributes.
24193 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
24194 nox64 isa attributes.
24195 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
24196 nox64 isa attributes.
24197 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
24199 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
24201 * config/alpha/alpha.c (TARGET_LRA_P): New define.
24203 2013-03-17 Jakub Jelinek <jakub@redhat.com>
24206 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
24207 class names. Remove trailing comma after "ALL_REGS".
24209 2013-03-16 Jan Hubicka <jh@suse.cz>
24211 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
24212 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
24213 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
24214 of cgraph_get_create_node.
24215 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
24217 2013-03-16 Jason Merrill <jason@redhat.com>
24220 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
24221 with DW_AT_default_value.
24223 2013-03-16 Jakub Jelinek <jakub@redhat.com>
24225 * BASE-VER: Set to 4.9.0.
24227 2013-03-14 Andi Kleen <ak@linux.intel.com>
24230 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
24231 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
24232 Document _x* TSX intrinsics.
24234 2013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
24235 David Holsgrove <david.holsgrove@xilinx.com>
24237 * configure.ac: Add MicroBlaze TLS support detection.
24238 * configure: Regenerate.
24239 * config/microblaze/microblaze-protos.h
24240 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
24241 symbol_mentioned_p, label_mentioned_p): Add prototypes.
24242 * config/microblaze/microblaze.c (microblaze_address_type): Add
24243 ADDRESS_TLS and tls_reloc address types.
24244 (microblaze_address_info): Add tls_reloc.
24245 (TARGET_HAVE_TLS): Define.
24246 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
24247 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
24248 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
24249 load_tls_operand, microblaze_call_tls_get_addr,
24250 microblaze_legitimize_tls_address): New functions.
24251 (microblaze_classify_unspec): Handle UNSPEC_TLS.
24252 (get_base_reg): Use microblaze_tls_symbol_p.
24253 (microblaze_classify_address): Handle TLS.
24254 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
24255 label_mentioned_p and microblaze_tls_referenced_p.
24256 (microblaze_legitimize_address): Handle TLS.
24257 (microblaze_address_insns): Handle ADDRESS_TLS.
24258 (pic_address_needs_scratch): Handle TLS.
24259 (print_operand_address): Handle TLS.
24260 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
24261 (microblaze_expand_move): Handle TLS.
24262 (microblaze_legitimate_constant_p): Check
24263 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
24264 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
24265 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
24266 (PIC_OFFSET_TABLE_REGNUM): Set.
24267 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
24268 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
24269 (addsi3, movsi_internal2, movdf_internal): Update constraints
24270 * config/microblaze/predicates.md (arith_plus_operand): Define
24271 (move_operand): Redefine as move_src_operand,
24272 check microblaze_tls_referenced_p.
24274 2013-03-14 Ian Bolton <ian.bolton@arm.com>
24276 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
24277 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
24279 2013-03-14 Ian Bolton <ian.bolton@arm.com>
24281 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
24284 2013-03-14 Jakub Jelinek <jakub@redhat.com>
24286 PR tree-optimization/53265
24287 * common.opt (Waggressive-loop-optimizations): New option.
24288 * tree-ssa-loop-niter.c: Include tree-pass.h.
24289 (do_warn_aggressive_loop_optimizations): New function.
24290 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
24291 if number_of_latch_executions returned constant.
24292 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
24293 early. If number_of_latch_executions returned constant, set
24294 nb_iterations_upper_bound back to it.
24295 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
24297 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
24298 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
24300 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
24301 (MULTILIB_OSDIRNAMES): Set.
24302 * genmultilib: If defaultosdirname doesn't start with :: , set
24303 defaultosdirname2 instead, clear it and emit two . multilib_raw
24304 entries instead of just one.
24306 2013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
24308 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
24309 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
24310 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
24311 (SUBTARGET_OVERRIDE_OPTIONS): New.
24313 2013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
24316 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
24317 (musermode): Convert to Var(TARGET_USERMODE).
24318 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
24319 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
24320 * config/sh/sh.c (sh_option_override): Use
24321 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
24322 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
24324 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
24326 (udivsi3_i4_single, divsi3_i4_single): Use
24327 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
24329 2013-03-13 Dave Korn <dave.korn.cygwin@gmail.com>
24331 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
24334 2013-03-13 Richard Biener <rguenther@suse.de>
24336 PR tree-optimization/56608
24337 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
24338 calls when vectorizing basic-blocks.
24340 2013-03-13 Jakub Jelinek <jakub@redhat.com>
24343 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
24346 2013-03-12 Jakub Jelinek <jakub@redhat.com>
24348 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
24350 2013-03-11 Jan Hubicka <jh@suse.cz>
24353 * lto-streamer-out.c (output_symbol_p): Skip references from
24354 constructors of external variables.
24356 2013-03-11 Jan Hubicka <jh@suse.cz>
24358 PR middle-end/56571
24359 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
24361 * emit-rtl.c (verify_rtx_sharing): Likewise.
24362 (copy_insn_1): Likewise.
24363 * rtl.c (copy_rtx): Likewise.
24365 2013-03-11 Georg-Johann Lay <avr@gjlay.de>
24368 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
24369 output_operand_lossage message.
24371 2013-03-11 Richard Earnshaw <rearnsha@arm.com>
24374 * arm.c (shift_op): Validate RTL pattern on the fly.
24375 (arm_print_operand, case 'S'): Don't use shift_operator to validate
24378 2013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
24381 * config/pa/pa.md (call_value): Check for calls to powf and direct to
24382 new call patterns that clobber %fr12.
24383 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
24384 split and postreload patterns.
24385 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
24386 registers %fr12 and %fr12R as call used.
24388 2013-03-09 Steven Bosscher <steven@gcc.gnu.org>
24390 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
24391 (canon_address, record_store, replace_read, check_mem_read_rtx,
24392 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
24393 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
24394 rest_of_handle_dse): Likewise.
24396 2013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
24398 PR middle-end/56524
24399 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
24401 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
24402 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
24403 (save_optabs_if_changed): Replace with...
24404 (init_tree_optimization_optabs): ...this.
24405 * optabs.c (save_optabs_if_changed): Rename to...
24406 (init_tree_optimization_optabs): ...this. Take the optimization node
24407 as argument. Do nothing if the base optabs are already correct.
24408 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
24409 to recompute optabs.
24410 * function.h (function): Remove optabs field.
24411 * function.c (invoke_set_current_function_hook): Call
24412 init_tree_optimization_optabs. Use the result to initialize
24415 2013-02-27 Aldy Hernandez <aldyh@redhat.com>
24417 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
24418 if GTMA_HAS_NO_INSTRUMENTATION.
24419 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
24420 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
24421 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
24422 * gimple-pretty-print.c (dump_gimple_transaction): Handle
24423 GTMA_HAS_NO_INSTRUMENTATION.
24425 2013-03-08 Jakub Jelinek <jakub@redhat.com>
24427 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
24430 2013-03-08 Marek Polacek <polacek@redhat.com>
24431 Jakub Jelinek <jakub@redhat.com>
24433 PR tree-optimization/56478
24434 * predict.c (is_comparison_with_loop_invariant_p): Change the
24435 type of loop_step to tree.
24436 (predict_loops): Adjust.
24437 (predict_iv_comparison): Perform the computations on double_ints.
24439 2013-03-08 Richard Biener <rguenther@suse.de>
24441 PR tree-optimization/56570
24442 * tree-cfg.c (verify_expr_location_1): Verify locations for
24444 * tree-sra.c (create_access_replacement): Strip locations
24445 from DECL_DEBUG_EXPRs.
24447 2013-03-08 Richard Biener <rguenther@suse.de>
24449 * tree-inline.c (expand_call_inline): Do not associate
24450 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
24451 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
24453 2013-03-08 Richard Biener <rguenther@suse.de>
24455 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
24456 or block changes with -Og. Fix for location / block encoding
24457 changes and PHI arguments with locations.
24459 2013-03-07 Steven Bosscher <steven@gcc.gnu.org>
24461 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
24463 (struct output_info): Likewise.
24464 (register_overhead): Remove bad gcc_assert.
24465 (bitmap_find_bit): If there is only a single bitmap element, do not
24466 count a miss as a search.
24467 (print_statistics): Update for counter type changes.
24468 (dump_bitmap_statistics): Likewise. Print headers such that they
24469 are properly lined up with the printed counters.
24471 2013-03-07 Jakub Jelinek <jakub@redhat.com>
24473 PR tree-optimization/56559
24474 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
24475 check that it has only a single use.
24477 2013-03-07 Richard Biener <rguenther@suse.de>
24479 * doc/invoke.texi (fwhole-program): Discourage use in combination
24482 2013-03-06 Jakub Jelinek <jakub@redhat.com>
24484 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
24486 PR tree-optimization/56539
24487 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
24488 instead of GSI_CONTINUE_LINKING as last argument to
24489 force_gimple_operand_gsi. Adjust function comment.
24491 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
24494 PR middle-end/56548
24495 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
24496 promoted mode, convert the result back to the original mode.
24498 2013-03-06 Richard Biener <rguenther@suse.de>
24500 PR middle-end/56294
24501 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
24502 (insert_updated_phi_nodes_compare_uids): New function.
24503 (update_ssa): Sort symbols_to_rename after UID before
24504 traversing it to insert PHI nodes.
24506 2013-03-06 Richard Biener <rguenther@suse.de>
24508 PR middle-end/50494
24509 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
24510 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
24513 2013-02-13 Richard Biener <rguenther@suse.de>
24516 * varasm.c (output_constant_def_1): Get the decl representing
24517 the constant as argument.
24518 (output_constant_def): Wrap output_constant_def_1.
24519 (make_decl_rtl): Use output_constant_def_1 with the decl
24520 representing the constant.
24521 (build_constant_desc): Optionally re-use a decl already
24522 representing the constant.
24523 (tree_output_constant_def): Adjust.
24525 2013-03-06 Joey Ye <joey.ye@arm.com>
24528 * gcc.c (convert_white_space): New function.
24529 (main): Handles white space in function name.
24531 2013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
24534 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
24535 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
24536 to SH_DIV_CALL_TABLE for TARGET_SH2.
24537 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
24539 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
24540 call-table options.
24542 2013-03-05 Sterling Augustine <saugustine@google.com>
24543 Cary Coutant <ccoutant@google.com>
24546 * dwarf2out.c (resolve_addr): Don't call
24547 remove_loc_list_addr_table_entries a second time for the same
24550 2013-03-05 Jakub Jelinek <jakub@redhat.com>
24553 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
24554 (avoid_complex_debug_insns): New function.
24555 (expand_debug_locations): Call it.
24557 PR rtl-optimization/56484
24558 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
24559 lifetimes of hard registers on small register class machines.
24561 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
24563 * config/microblaze/microblaze-protos.h: Rename
24564 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
24565 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
24567 (microblaze_fast_interrupt_function_p): New function.
24568 (microblaze_is_interrupt_handler): Rename to
24569 microblaze_is_interrupt_variant and add fast_interrupt check.
24570 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
24571 (save_restore_insns): Likewise.
24572 (compute_frame_size): Likewise.
24573 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
24574 (microblaze_globalize_label): Likewise.
24575 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
24576 * config/microblaze/microblaze.md: Use wrapper
24577 microblaze_is_interrupt_variant.
24579 2013-03-05 Kai Tietz <ktietz@redhat.com>
24581 * sdbout.c (sdbout_one_type): Switch to current function's section
24582 supporting cold/hot.
24584 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
24586 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
24589 2013-03-05 Jakub Jelinek <jakub@redhat.com>
24591 PR middle-end/56461
24592 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
24593 if VALGRIND_GET_VBITS is defined, temporarily make object
24594 memory all defined, and restore previous valgrind addressability
24595 and definability afterwards. Free this_object at the end.
24597 PR middle-end/56461
24598 * lra.c (lra): Call lra_clear_live_ranges if live_p,
24599 right before calling lra_create_live_ranges, also call it
24600 when clearing live_p. Only call lra_clear_live_ranges
24601 at the end if live_p.
24603 PR middle-end/56461
24604 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
24606 2013-03-05 Richard Biener <rguenther@suse.de>
24608 PR tree-optimization/56521
24609 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
24612 2013-03-05 Steven Bosscher <steven@gcc.gnu.org>
24615 * except.h (remove_unreachable_eh_regions): New prototype.
24616 * except.c (remove_eh_handler_splicer): New function, split out
24617 of remove_eh_handler.
24618 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
24619 warning about running it on many EH regions one at a time.
24620 (remove_unreachable_eh_regions_worker): New function, walk the
24621 EH tree in depth-first order and remove non-marked regions.
24622 (remove_unreachable_eh_regions): New function.
24623 * tree-eh.c (mark_reachable_handlers): New function, split out
24624 from remove_unreachable_handlers.
24625 (remove_unreachable_handlers): Use mark_reachable_handlers and
24626 remove_unreachable_eh_regions.
24627 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
24628 and remove_unreachable_eh_regions.
24630 2013-03-05 Richard Biener <rguenther@suse.de>
24632 PR middle-end/56525
24633 * loop-init.c (fix_loop_structure): Remove loops in two stages,
24634 not freeing them until the end.
24636 2013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
24638 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
24640 2013-03-05 Richard Biener <rguenther@suse.de>
24642 PR tree-optimization/56270
24643 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
24644 of loads after scheduling an SLP instance.
24646 2013-03-05 Jakub Jelinek <jakub@redhat.com>
24648 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
24650 (check_gcc_parallelize): Run guality.exp as a separate job from
24651 vect.exp with unsorted.exp and $(dg_target_exps) separately from
24652 struct-layout-1.exp with stackalign.exp.
24654 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
24656 PR middle-end/56461
24657 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
24658 load_index sbitmap even if some bit in it isn't set.
24660 PR middle-end/56461
24661 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
24662 (discover_iteration_bound_by_body_walk): Change queues to
24663 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
24664 spelling in comment. Call safe_push on queues[bound_index] directly.
24665 Release queues[queue_index] in every iteration unconditionally.
24666 Release bounds vector.
24668 PR middle-end/56461
24669 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
24670 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
24671 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
24674 2013-03-05 Richard Biener <rguenther@suse.de>
24677 * tree-inline.c (remap_blocks_to_null): New function.
24678 (expand_call_inline): When expanding a call stmt without
24679 an associated block inline remap all callee blocks to NULL.
24681 2013-03-05 Jakub Jelinek <jakub@redhat.com>
24683 PR rtl-optimization/56494
24684 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
24685 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
24686 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
24688 PR middle-end/56461
24689 * sel-sched-ir.c (free_sched_pools): Release
24690 succs_info_pool.stack[succs_info_pool.max_top] vectors too
24691 if succs_info_pool.max_top isn't -1.
24694 * opts.c (opts_obstack, opts_concat): Moved to...
24695 * opts-common.c (opts_obstack, opts_concat): ... here.
24697 2013-03-04 Jakub Jelinek <jakub@redhat.com>
24699 PR middle-end/56461
24700 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
24702 2013-03-04 Martin Jambor <mjambor@suse.cz>
24704 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
24705 all appropriate places.
24707 2013-01-04 Eric Botcazou <ebotcazou@adacore.com>
24709 PR tree-optimization/56424
24710 * ipa-split.c (split_function): Do not set the RSO flag if result is
24711 not by reference and its type is a register type.
24713 2013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
24715 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
24716 (microblaze_legitimate_pic_operand): Likewise
24717 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
24718 new function microblaze_legitimate_pic_operand
24719 * config/microblaze/microblaze-protos.h
24720 (microblaze_legitimate_pic_operand): Declare.
24722 2013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24724 * config/microblaze/predicates.md (call_insn_simple_operand):
24725 New predicate for supported rtx code types.
24726 * config/microblaze/microblaze.md (call_internal1): Use
24727 call_insn_simple_operand predicate.
24729 2013-03-04 Jakub Jelinek <jakub@redhat.com>
24731 PR middle-end/56461
24732 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
24733 partitions.ordered_remove.
24735 PR middle-end/56461
24736 * tree-vect-stmts.c (vectorizable_conversion): Don't call
24737 vec_oprnds0.create (1) for modifier == NONE.
24739 PR middle-end/56461
24740 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
24741 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
24742 vec_oprnds1 right before pushing anything to it for
24745 PR middle-end/56461
24746 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
24747 set nbbs to 0 instead of having separate code path.
24748 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
24749 instead of false as last argument if returning NULL.
24751 2013-03-03 Sandra Loosemore <sandra@codesourcery.com>
24753 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
24754 the attribute is now called "target" instead of "option".
24755 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
24756 * doc/tm.texi.in (Target Attributes): Likewise document the correct
24757 attribute/pragma name for TARGET_OPTION_VALID_P and
24758 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
24759 * doc/tm.texi: Regenerated.
24761 2013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
24763 * config/microblaze/microblaze.c:
24764 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
24765 * config/microblaze/microblaze.h: Add -mxl-reorder to
24767 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
24768 instructions emitted if TARGET_REORDER.
24769 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
24770 or 0 for -m/-mno case, but initialises as 2 to detect default use case
24773 2013-03-01 Xinliang David Li <davidxl@google.com>
24775 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
24778 2013-03-01 Jakub Jelinek <jakub@redhat.com>
24780 PR middle-end/56461
24781 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
24782 vector even when returning true. Fix up function comment formatting.
24784 PR middle-end/56461
24785 * ira-build.c (ira_loop_nodes_count): New variable.
24786 (create_loop_tree_nodes): Initialize it.
24787 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
24789 PR middle-end/56461
24790 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
24791 method on dr_chain and result_chain.
24792 * tree-vect-stmts.c (vectorizable_store): Only call
24793 result_chain.create if j == 0.
24795 PR middle-end/56461
24796 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
24797 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
24798 before overwriting it.
24800 2013-03-01 Tobias Burnus <burnus@net-b.de>
24802 * doc/extended.texi (C Extensions): Change order in @menu
24804 (Other MIPS Built-in Functions): Move last MIPS entry before
24805 "picoChip Built-in Functions".
24806 (SH Built-in Functions): Move after RX Built-in Functions.
24807 * doc/gcc.texi (Introduction): Change order in @menu
24809 * doc/md.texi (Constraints): Ditto.
24810 * gty.texi (Type Information): Ditto.
24811 (User-provided marking routines for template types): Make
24813 * doc/invoke.texi (AArch64 Options): Move before
24814 "Adapteva Epiphany Options".
24816 2013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
24817 Jakub Jelinek <jakub@redhat.com>
24820 * asan.c (gate_asan): Lookup no_sanitize_address instead of
24821 no_address_safety_analysis attribute.
24822 * doc/extend.texi (no_address_safety_attribute): Rename to
24823 no_sanitize_address attribute, mention no_address_safety_analysis
24824 attribute as deprecated alias.
24826 2013-02-28 Jakub Jelinek <jakub@redhat.com>
24828 PR middle-end/56461
24829 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
24830 type to vec<vec<tree> > *.
24831 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
24832 to be vec<tree> instead of vec<tree> *, set vec_defs
24833 to vNULL and call vec_defs.create (number_of_vects), adjust other
24835 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
24836 vectorizable_condition): Adjust vect_get_slp_defs callers.
24838 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
24840 * config/aarch64/aarch64.c
24841 (aarch64_float_const_representable): Remove unused variable.
24843 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
24845 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
24847 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
24849 * config/aarch64/aarch64-builtins.c
24850 (aarch64_init_simd_builtins): Make static.
24852 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
24854 * config/aarch64/aarch64.c
24855 (aarch64_simd_make_constant): Make static.
24857 2013-02-28 Martin Jambor <mjambor@suse.cz>
24859 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
24860 with no initialization to the RHS of debug statements.
24862 2013-02-28 Martin Jambor <mjambor@suse.cz>
24864 PR tree-optimization/56294
24865 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
24867 (get_access_replacement): Do not call create_access_replacement.
24868 Assert a replacement exists.
24869 (get_repl_default_def_ssa_name): Create the replacement declaration
24872 2013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
24874 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
24875 final_end_function.
24877 2013-02-28 Marek Polacek <polacek@redhat.com>
24879 PR rtl-optimization/56466
24880 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
24881 if we're changing a loop.
24882 (peel_loops_completely): Likewise.
24884 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
24887 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
24889 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
24892 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
24893 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
24894 INTX_FTYPE_FX, FX_FTYPE_INTX.
24895 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
24897 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
24899 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
24900 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
24901 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
24902 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
24903 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
24904 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
24905 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
24906 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
24907 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
24908 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
24909 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
24910 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
24911 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
24912 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
24913 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
24914 (avrxmega6): Increase max flash segments from 5 to 6.
24915 * config/avr/t-multilib: Regenerate.
24916 * config/avr/avr-tables.opt: Regenerate.
24917 * doc/avr-mmcu.texi: Regenerate.
24919 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
24921 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
24922 (avr_device_to_arch): Rename to avr_device_to_ld.
24923 (avr_device_to_as): New prototype.
24924 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
24925 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
24926 * config/avr/driver-avr.c (avr_device_to_as): New.
24927 (avr_device_to_arch): Rename to avr_device_to_ld.
24929 2013-02-27 Jakub Jelinek <jakub@redhat.com>
24931 PR middle-end/56461
24932 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
24933 method on dr_chain and result_chain.
24935 PR middle-end/56461
24936 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
24937 pointer_set_destroy on not_executed_last_iteration.
24939 PR middle-end/56461
24940 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
24942 PR middle-end/56461
24943 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
24944 FOR_EACH_DEFINED_FUNCTION when freeing state.
24946 PR middle-end/56461
24947 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
24949 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
24952 PR middle-end/56461
24953 * ipa-cp.c (decide_whether_version_node): Call vec_free on
24954 known_aggs[i].items and release known_aggs vector.
24956 PR middle-end/56461
24957 * ipa-reference.c (propagate): Free node_info even for alias nodes.
24959 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24961 * config/microblaze/microblaze.c (microblaze_emit_compare):
24962 Use xor for EQ/NE comparisions.
24963 * config/microblaze/microblaze.md (cstoresf4): Add constraints
24964 (cbranchsf4): Adjust operator to comparison_operator.
24966 2013-02-27 Jakub Jelinek <jakub@redhat.com>
24968 PR middle-end/56461
24969 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
24971 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
24972 vec_safe_push, always update *slot.
24973 (redirect_edge_var_map_clear): Use vec_free.
24974 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
24975 (free_var_map_entry): Use vec_free.
24976 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
24977 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
24979 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
24981 PR middle-end/45472
24982 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
24983 when the may_trap_p bit of the exprs being merged differs.
24984 Reorder tests for speculativeness in the logical and operator.
24986 2013-02-27 Jakub Jelinek <jakub@redhat.com>
24988 * incpath.c (add_standard_paths): Use reconcat instead of concat
24989 where appropriate and avoid leaking memory.
24991 * opts.h: Include obstack.h.
24992 (opts_concat): New prototype.
24993 (opts_obstack): New declaration.
24994 * opts.c (opts_concat): New function.
24995 (opts_obstack): New variable.
24996 (init_options_struct): Call gcc_init_obstack on opts_obstack.
24997 (finish_options): Use opts_concat instead of concat
24998 and XOBNEWVEC instead of XNEWVEC.
24999 * opts-common.c (generate_canonical_option, decode_cmdline_option,
25000 generate_option): Likewise.
25001 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
25002 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
25005 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
25006 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
25008 2013-02-26 Jakub Jelinek <jakub@redhat.com>
25010 PR middle-end/56461
25011 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
25013 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
25015 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
25016 (arm_block_move_unaligned_straight): Likewise.
25017 (arm_adjust_block_mem): Likewise.
25019 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
25022 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
25023 temp, cond and label.
25024 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
25027 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
25028 * config/c6x/c6x.h (dbx_register_map): Update declaration.
25031 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
25032 of prologue/epilogue functions.
25035 * config/tilegx/tilegx.c (tilegx_expand_prologue):
25036 Remove unused variable cfa_offset.
25037 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
25040 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
25041 type promotion to unsigned.
25044 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
25045 for HOST_WIDE_INT of 32 bit / same size as int.
25046 (arm_block_move_unaligned_straight): Likewise.
25047 (arm_adjust_block_mem): Likewise.
25050 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
25053 2013-02-26 Marek Polacek <polacek@redhat.com>
25055 PR tree-optimization/56426
25056 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
25058 2013-02-26 Richard Biener <rguenther@suse.de>
25061 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
25062 unused variable loops.
25064 2013-02-26 Jakub Jelinek <jakub@redhat.com>
25066 PR tree-optimization/56448
25067 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
25068 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
25069 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
25070 later operands of the references, or even first operand for
25071 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
25073 PR tree-optimization/56443
25074 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
25075 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
25076 to type_for_mode langhook.
25078 2013-02-25 Matt Turner <mattst88@gmail.com>
25080 * doc/invoke.texi: Document r4700.
25082 2013-02-25 Richard Biener <rguenther@suse.de>
25084 PR tree-optimization/56175
25085 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
25087 (simplify_bitwise_binary): ... here. Also guard the conversion
25088 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
25090 2013-02-25 Catherine Moore <clm@codesourcery.com>
25093 2013-02-24 Catherine Moore <clm@codesourcery.com>
25094 Maciej W. Rozycki <macro@codesourcery.com>
25095 Tom de Vries <tom@codesourcery.com>
25096 Nathan Sidwell <nathan@codesourcery.com>
25097 Iain Sandoe <iain@codesourcery.com>
25098 Nathan Froyd <froydnj@codesourcery.com>
25099 Chao-ying Fu <fu@mips.com>
25101 * doc/extend.texi: (micromips, nomicromips, nocompression):
25102 Document new function attributes.
25103 * doc/invoke.texi (minterlink-compressed, mmicromips,
25104 m14k, m14ke, m14kec): Document new options.
25105 (minterlink-mips16): Update documentation.
25106 * doc/md.texi (ZC, ZD): Document new constraints.
25107 * configure.ac (gcc_cv_as_micromips): Check if linker
25108 supports the .set micromips directive.
25109 * configure: Regenerate.
25110 * config.in: Regenerate.
25111 * config/mips/mips-tables.opt: Regenerate.
25112 * config/mips/micromips.md: New file.
25113 * constraints.md (ZC, AD): New constraints.
25114 * config/mips/predicates.md (movep_src_register): New predicate.
25115 (movep_src_operand): New predicate.
25116 (non_volatile_mem_operand): New predicate.
25117 * config/mips/mips.md (multimem): New type.
25118 (length): Differentiate between 17-bit and 18-bit branch offsets.
25119 (MOVEP1, MOVEP2): New mode iterator.
25120 (mov_<load>l): Use ZC constraint.
25121 (mov_<load>r): Likewise.
25122 (mov_<store>l): Likewise.
25123 (mov_<store>r): Likewise.
25124 (*branch_equality<mode>_inverted): Add microMIPS support.
25125 (*branch_equality<mode>): Likewise.
25126 (*jump_absolute): Likewise.
25127 (indirect_jump_<mode>): Likewise.
25128 (tablejump_<mode>): Likewise.
25129 (<optab>_internal): Likewise.
25130 (sibcall_internal): Likewise.
25131 (sibcall_value_internal): Likewise.
25132 (prefetch): Use constraint ZD.
25133 * config/mips/mips.opt (minterlink-compressed): New option.
25134 (minterlink-mips16): Now an alias for minterlink-compressed.
25135 (mmicromips): New option.
25136 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
25137 (compare_and_swap_12): Likewise.
25138 (sync_add<mode>): Likewise.
25139 (sync_<optab>_12): Likewise.
25140 (sync_old_<optab>_12): Likewise.
25141 (sync_new_<optab>_12): Likewise.
25142 (sync_nand_12): Likewise.
25143 (sync_old_nand_12): Likewise.
25144 (sync_new_nand_12): Likewise.
25145 (sync_sub<mode>): Likewise.
25146 (sync_old_add<mode>): Likewise.
25147 (sync_old_sub<mode>): Likewise.
25148 (sync_new_add<mode>): Likewise.
25149 (sync_new_sub<mode>): Likewise.
25150 (sync_<optab><mode>): Likewise.
25151 (sync_old_<optab><mode>): Likewise.
25152 (sync_new_<optab><mode>): Likewise.
25153 (sync_nand<mode>): Likewise.
25154 (sync_old_nand<mode>): Likewise.
25155 (sync_new_nand<mode>): Likewise.
25156 (sync_lock_test_and_set<mode>): Likewise.
25157 (test_and_set_12): Likewise.
25158 (atomic_compare_and_swap<mode>): Likewise.
25159 (atomic_exchange<mode>_llsc): Likewise.
25160 (atomic_fetch_add<mode>_llsc): Likewise.
25161 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
25162 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
25163 (umips_save_restore_pattern_p): Likewise.
25164 (umips_load_store_pair_p): Likewise.
25165 (umips_output_load_store_pair): Likewise.
25166 (umips_movep_target_p): Likewise.
25167 (umips_12bit_offset_address_p): Likewise.
25168 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
25169 (mips_base_mips16): Rename this...
25170 (mips_base_compression_flags): ...to this. Update all uses.
25171 (mips_attribute_table): Add micromips, nomicromips and nocompression.
25172 (mips_mips16_decl_p): Delete.
25173 (mips_nomips16_decl_p): Delete.
25174 (mips_get_compress_on_flags): New function.
25175 (mips_get_compress_off_flags): New function.
25176 (mips_get_compress_mode): New function.
25177 (mips_get_compress_on_name): New function.
25178 (mips_get_compress_off_name): New function.
25179 (mips_insert_attributes): Support multiple compression types.
25180 (mips_merge_decl_attributes): Likewise.
25181 (umips_12bit_offset_address_p): New function.
25182 (mips_start_function_definition): Emit .set micromips directive.
25183 (mips_call_may_need_jalx_p): New function.
25184 (mips_function_ok_for_sibcall): Add microMIPS support.
25185 (mips_print_operand_punctuation): Support short delay slots and
25187 (umips_swm_mask, umips_swm_encoding): New.
25188 (umips_build_save_restore): New function.
25189 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
25190 (was_mips16_p): Remove.
25191 (old_compression_mode): New.
25192 (mips_set_compression_mode): New function.
25193 (mips_set_current_function): Add microMIPS support.
25194 (mips_option_override): Likewise.
25195 (umips_save_restore_pattern_p): New function.
25196 (umips_output_save_restore): New function.
25197 (umips_load_store_pair_p_1): New function.
25198 (umips_load_store_pair_p): New function.
25199 (umips_output_load_store_pair_1): New function.
25200 (umips_output_load_store_pair): New function.
25201 (umips_movep_target_p) New function.
25202 (mips_prepare_pch_save): Add microMIPS support.
25203 * config/mips/mips.h (TARGET_COMPRESSION): New.
25204 (TARGET_CPU_CPP_BUILTINS): Update macro
25205 to use new compression flags and to support microMIPS.
25206 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
25207 (MIPS_ARCH_FLOAT_SPEC): Likewise.
25208 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
25209 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
25210 (ASM_SPEC): Support mmicromips and mno-micromips.
25211 (M16STORE_REG_P): New macro.
25212 (MIPS_CALL): Support TARGET_MICROMIPS.
25213 (MICROMIPS_J): New macro.
25214 (mips_base_mips16): Rename this...
25215 (mips_base_compression_flags): ...to this.
25216 (UMIPS_12BIT_OFFSET_P): New macro.
25217 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
25218 (MULTILIB_DIRNAMES): Likewise.
25220 2013-02-25 Tom de Vries <tom@codesourcery.com>
25222 PR rtl-optimization/56131
25223 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
25224 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
25225 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
25227 2013-02-25 Tobias Burnus <burnus@net-b.de>
25229 * doc/invoke.texi (-fsanitize=): Move from optimization
25230 to debugging options.
25232 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
25234 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
25236 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
25237 Alexander Monakov <amonakov@ispras.ru>
25239 PR middle-end/56077
25240 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
25241 flush pending lists also on non-jumps. Adjust comment.
25243 2013-02-24 Catherine Moore <clm@codesourcery.com>
25244 Maciej W. Rozycki <macro@codesourcery.com>
25245 Tom de Vries <tom@codesourcery.com>
25246 Nathan Sidwell <nathan@codesourcery.com>
25247 Iain Sandoe <iain@codesourcery.com>
25248 Nathan Froyd <froydnj@codesourcery.com>
25249 Chao-ying Fu <fu@mips.com>
25251 * doc/extend.texi: (micromips, nomicromips, nocompression):
25252 Document new function attributes.
25253 * doc/invoke.texi (minterlink-compressed, mmicromips,
25254 m14k, m14ke, m14kec): Document new options.
25255 (minterlink-mips16): Update documentation.
25256 * doc/md.texi (ZC, ZD): Document new constraints.
25257 * configure.ac (gcc_cv_as_micromips): Check if linker
25258 supports the .set micromips directive.
25259 * configure: Regenerate.
25260 * config.in: Regenerate.
25261 * config/mips/mips-tables.opt: Regenerate.
25262 * config/mips/micromips.md: New file.
25263 * constraints.md (ZC, AD): New constraints.
25264 * config/mips/predicates.md (movep_src_register): New predicate.
25265 (movep_src_operand): New predicate.
25266 (non_volatile_mem_operand): New predicate.
25267 * config/mips/mips.md (multimem): New type.
25268 (length): Differentiate between 17-bit and 18-bit branch offsets.
25269 (MOVEP1, MOVEP2): New mode iterator.
25270 (mov_<load>l): Use ZC constraint.
25271 (mov_<load>r): Likewise.
25272 (mov_<store>l): Likewise.
25273 (mov_<store>r): Likewise.
25274 (*branch_equality<mode>_inverted): Add microMIPS support.
25275 (*branch_equality<mode>): Likewise.
25276 (*jump_absolute): Likewise.
25277 (indirect_jump_<mode>): Likewise.
25278 (tablejump_<mode>): Likewise.
25279 (<optab>_internal): Likewise.
25280 (sibcall_internal): Likewise.
25281 (sibcall_value_internal): Likewise.
25282 (prefetch): Use constraint ZD.
25283 * config/mips/mips.opt (minterlink-compressed): New option.
25284 (minterlink-mips16): Now an alias for minterlink-compressed.
25285 (mmicromips): New option.
25286 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
25287 (compare_and_swap_12): Likewise.
25288 (sync_add<mode>): Likewise.
25289 (sync_<optab>_12): Likewise.
25290 (sync_old_<optab>_12): Likewise.
25291 (sync_new_<optab>_12): Likewise.
25292 (sync_nand_12): Likewise.
25293 (sync_old_nand_12): Likewise.
25294 (sync_new_nand_12): Likewise.
25295 (sync_sub<mode>): Likewise.
25296 (sync_old_add<mode>): Likewise.
25297 (sync_old_sub<mode>): Likewise.
25298 (sync_new_add<mode>): Likewise.
25299 (sync_new_sub<mode>): Likewise.
25300 (sync_<optab><mode>): Likewise.
25301 (sync_old_<optab><mode>): Likewise.
25302 (sync_new_<optab><mode>): Likewise.
25303 (sync_nand<mode>): Likewise.
25304 (sync_old_nand<mode>): Likewise.
25305 (sync_new_nand<mode>): Likewise.
25306 (sync_lock_test_and_set<mode>): Likewise.
25307 (test_and_set_12): Likewise.
25308 (atomic_compare_and_swap<mode>): Likewise.
25309 (atomic_exchange<mode>_llsc): Likewise.
25310 (atomic_fetch_add<mode>_llsc): Likewise.
25311 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
25312 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
25313 (umips_save_restore_pattern_p): Likewise.
25314 (umips_load_store_pair_p): Likewise.
25315 (umips_output_load_store_pair): Likewise.
25316 (umips_movep_target_p): Likewise.
25317 (umips_12bit_offset_address_p): Likewise.
25318 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
25319 (mips_base_mips16): Rename this...
25320 (mips_base_compression_flags): ...to this. Update all uses.
25321 (mips_attribute_table): Add micromips, nomicromips and nocompression.
25322 (mips_mips16_decl_p): Delete.
25323 (mips_nomips16_decl_p): Delete.
25324 (mips_get_compress_on_flags): New function.
25325 (mips_get_compress_off_flags): New function.
25326 (mips_get_compress_mode): New function.
25327 (mips_get_compress_on_name): New function.
25328 (mips_get_compress_off_name): New function.
25329 (mips_insert_attributes): Support multiple compression types.
25330 (mips_merge_decl_attributes): Likewise.
25331 (umips_12bit_offset_address_p): New function.
25332 (mips_start_function_definition): Emit .set micromips directive.
25333 (mips_call_may_need_jalx_p): New function.
25334 (mips_function_ok_for_sibcall): Add microMIPS support.
25335 (mips_print_operand_punctuation): Support short delay slots and
25337 (umips_swm_mask, umips_swm_encoding): New.
25338 (umips_build_save_restore): New function.
25339 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
25340 (was_mips16_p): Remove.
25341 (old_compression_mode): New.
25342 (mips_set_compression_mode): New function.
25343 (mips_set_current_function): Add microMIPS support.
25344 (mips_option_override): Likewise.
25345 (umips_save_restore_pattern_p): New function.
25346 (umips_output_save_restore): New function.
25347 (umips_load_store_pair_p_1): New function.
25348 (umips_load_store_pair_p): New function.
25349 (umips_output_load_store_pair_1): New function.
25350 (umips_output_load_store_pair): New function.
25351 (umips_movep_target_p) New function.
25352 (mips_prepare_pch_save): Add microMIPS support.
25353 * config/mips/mips.h (TARGET_COMPRESSION): New.
25354 (TARGET_CPU_CPP_BUILTINS): Update macro
25355 to use new compression flags and to support microMIPS.
25356 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
25357 (MIPS_ARCH_FLOAT_SPEC): Likewise.
25358 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
25359 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
25360 (ASM_SPEC): Support mmicromips and mno-micromips.
25361 (M16STORE_REG_P): New macro.
25362 (MIPS_CALL): Support TARGET_MICROMIPS.
25363 (MICROMIPS_J): New macro.
25364 (mips_base_mips16): Rename this...
25365 (mips_base_compression_flags): ...to this.
25366 (UMIPS_12BIT_OFFSET_P): New macro.
25367 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
25368 (MULTILIB_DIRNAMES): Likewise.
25370 2013-02-24 Jakub Jelinek <jakub@redhat.com>
25373 * target-globals.c (save_target_globals): For init_reg_sets and
25374 target_reinit remporarily set this_fn_optabs to this_target_optabs.
25376 2013-02-22 James Greenhalgh <james.greenhalgh@arm.com>
25378 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
25379 * config/aarch64/t-aarch64
25380 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
25382 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
25384 PR inline-asm/56148
25385 * lra-constraints.c (process_alt_operands): Reload operand
25386 conflicting with earlier clobber only if no more other conflicting
25389 2013-02-22 Jakub Jelinek <jakub@redhat.com>
25392 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
25393 if not linking a shared library.
25395 2013-02-22 Seth LaForge <sethml@google.com>
25397 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
25399 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
25401 * config/arm/arm.md (split for extendsidi): Update condition.
25402 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
25403 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
25404 (qhs_zextenddi_cstr): Likewise.
25406 2013-02-21 Jakub Jelinek <jakub@redhat.com>
25408 PR middle-end/56420
25409 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
25410 avoid signed wrapping.
25411 (expand_mult): Handle properly multiplication by
25412 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
25413 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
25414 in the compiler if coeff is HOST_WIDE_INT_MIN.
25415 (expand_divmod): Don't make ext_op1 static, change it's type to
25416 uhwi. Avoid undefined behavior in -INTVAL (op1).
25418 PR rtl-optimization/50339
25419 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
25421 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
25422 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
25423 into splitting_ashiftrt field.
25424 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
25426 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
25429 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
25431 PR middle-end/56108
25432 * trans-mem.c (execute_tm_mark): Do not expand transactions that
25433 are sure to go irrevocable.
25435 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
25437 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
25438 scalars are valid operands.
25440 2013-02-21 Martin Jambor <mjambor@suse.cz>
25442 PR tree-optimization/56310
25443 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
25444 only matching indices and non-negative final offsets.
25445 (intersect_aggregates_with_edge): Pass src_idx to
25446 agg_replacements_to_vector. Pass src_idx insstead of index to
25447 intersect_with_agg_replacements.
25449 2013-02-21 Martin Jambor <mjambor@suse.cz>
25451 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
25452 instead of hard-wired defaults.
25454 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
25456 * doc/invoke.texi (MIPS Options): Update documentation of the
25457 floating-point multiply-accumulate instruction restrictions.
25459 2013-02-21 Kostya Serebryany <kcc@google.com>
25461 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
25462 asan_shadow_offset on x86_64 linux.
25464 2013-02-21 Richard Biener <rguenther@suse.de>
25466 PR tree-optimization/56415
25468 2013-02-11 Richard Biener <rguenther@suse.de>
25470 PR tree-optimization/56273
25471 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
25474 2013-02-21 Jakub Jelinek <jakub@redhat.com>
25477 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
25480 PR inline-asm/56405
25481 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
25482 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
25484 2013-02-20 Jan Hubicka <jh@suse.cz>
25486 PR tree-optimization/56265
25487 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
25488 when target is referenced for first time.
25490 2013-02-20 Richard Biener <rguenther@suse.de>
25492 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
25493 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
25494 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
25495 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
25496 not return anything.
25497 (rename_ssa_copies): Do not remove unused locals.
25498 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
25499 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
25500 * passes.c (execute_function_todo): Do not schedule unused locals
25501 removal if cleanup_tree_cfg did something.
25502 * tree-ssa-live.c (remove_unused_locals): Dump statistics
25503 about the number of removed locals.
25505 2013-02-20 Richard Biener <rguenther@suse.de>
25507 PR tree-optimization/56398
25508 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
25510 2013-02-20 Martin Jambor <mjambor@suse.cz>
25512 PR tree-optimization/55334
25513 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
25514 restricted pointers to arrays.
25516 2013-02-20 Richard Biener <rguenther@suse.de>
25517 Jakub Jelinek <jakub@redhat.com>
25519 PR tree-optimization/56396
25520 * tree-ssa-ccp.c (n_const_val): New static variable.
25521 (get_value): Return NULL for SSA names we don't have a lattice
25523 (ccp_initialize): Initialize n_const_val.
25524 * tree-ssa-copy.c (n_copy_of): New static variable.
25525 (init_copy_prop): Initialize n_copy_of.
25526 (get_value): Return NULL_TREE for SSA names we don't have a
25529 2013-02-20 Martin Jambor <mjambor@suse.cz>
25531 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
25533 2013-02-20 Richard Biener <rguenther@suse.de>
25535 * genpreds.c (write_lookup_constraint): Do not compare first
25536 letter of the constraint again.
25538 2013-02-20 Richard Biener <rguenther@suse.de>
25540 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
25542 (get_use_iv_cost): Terminate hashtable walk when coming across
25545 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
25547 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
25548 reassociation for avx2 targets.
25550 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
25552 * config/microblaze/microblaze.c: microblaze_has_clz = 0
25553 Add version check for v8.10.a to enable microblaze_has_clz
25554 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
25555 version and TARGET_PATTERN_COMPARE check
25556 * config/microblaze/microblaze.md: New clzsi2 instruction
25558 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
25560 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
25561 function before branching.
25563 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
25565 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
25567 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
25569 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
25571 PR middle-end/55889
25572 * sel-sched.c: Include ira.h.
25573 (implicit_clobber_conflict_p): New function.
25574 (moveup_expr): Use it.
25575 * Makefile.in (sel-sched.o): Depend on ira.h.
25577 2013-02-19 Richard Biener <rguenther@suse.de>
25579 PR tree-optimization/56384
25580 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
25581 (vn_hash_type): Split out from ...
25582 (vn_hash_constant_with_type): ... here.
25583 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
25584 (vn_phi_eq): Compare types from vn_phi_s structure.
25585 (vn_phi_lookup): Populate vn_phi_s type.
25586 (vn_phi_insert): Likewise.
25588 2013-02-19 Jakub Jelinek <jakub@redhat.com>
25590 PR tree-optimization/56350
25591 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
25592 if haven't found reduction or nested cycle operand, rather than
25593 asserting we must find it.
25595 PR tree-optimization/56381
25596 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
25599 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
25600 Jakub Jelinek <jakub@redhat.com>
25603 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
25604 (swap_optab_enable): Same.
25605 (init_all_optabs): Use argument instead of global.
25606 * tree.h (struct tree_optimization_option): New field target_optabs.
25607 * expr.h (init_all_optabs): Add argument to prototype.
25608 (TREE_OPTIMIZATION_OPTABS): New.
25609 (save_optabs_if_changed): Protoize.
25610 * optabs.h: Declare this_fn_optabs.
25611 * optabs.c (save_optabs_if_changed): New.
25612 Declare this_fn_optabs.
25613 (init_optabs): Add argument to init_all_optabs() call.
25614 * function.c (invoke_set_current_function_hook): Handle per
25616 * function.h (struct function): New field optabs.
25617 * config/mips/mips.c (mips_set_mips16_mode): Handle when
25618 optimization_current_node has changed.
25619 * target-globals.h (save_target_globals_default_opts): Protoize.
25620 * target-globals.c (save_target_globals_default_opts): New.
25622 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
25625 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
25626 registers %fr12 and %fr12R as call used.
25629 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
25630 and HImode, require all displacements to be an integer multiple of
25632 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
25633 only allow QImode and HImode when reload is in progress and strict is
25634 true. Likewise for symbolic addresses. Use base14_operand to check
25635 displacements in REG+BASE addresses.
25637 2013-02-18 Richard Biener <rguenther@suse.de>
25639 PR tree-optimization/56366
25640 * tree-vect-loop.c (get_initial_def_for_induction): Properly
25641 handle sign-conversion of outer-loop initial induction value.
25643 2013-02-18 Richard Biener <rguenther@suse.de>
25645 PR middle-end/56349
25646 * cfghooks.c (merge_blocks): If we merge a latch into another
25647 block adjust references to it.
25648 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
25649 (verify_loop_structure): Verify that a recorded latch is in fact
25652 2013-02-18 Richard Biener <rguenther@suse.de>
25654 PR tree-optimization/56321
25655 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
25656 order SSA name release and virtual operand unlinking.
25658 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
25660 * config/microblaze/microblaze.md (save_stack_block): Define.
25661 (restore_stack_block): Likewise.
25663 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
25665 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
25666 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
25667 * config/microblaze/microblaze.c (microblaze_option_override):
25668 Bail out early for PIC modes when target does not support PIC.
25670 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
25672 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
25673 Replace with a microblaze version.
25674 (microblaze_trampoline_init): Adapt for microblaze.
25675 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
25678 2013-02-16 Jakub Jelinek <jakub@redhat.com>
25679 Dodji Seketeli <dodji@redhat.com>
25682 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
25683 (instrument_mem_region_access): Do not forget to always put
25684 instrumentation of the of 'base' and 'base + len' in a "if (len !=
25685 0) statement, even for cases where either 'base' or 'base + len'
25686 are not instrumented -- because they have been previously
25687 instrumented. Simplify the logic by putting all the statements
25688 instrument 'base + len' inside a sequence, and then insert that
25689 sequence right before the current insertion point. Then, to
25690 instrument 'base + len', just get an iterator on that statement.
25691 And do not forget to update the pointer to iterator the function
25692 received as argument.
25694 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
25696 PR rtl-optimization/56348
25697 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
25699 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
25701 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
25702 (clean_graph_dump_file): Pass base to start_graph_dump.
25704 2013-02-14 Richard Henderson <rth@redhat.com>
25707 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
25709 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
25711 * collect2-aix.h: Define F_LOADONLY.
25713 2013-02-14 Richard Biener <rguenther@suse.de>
25716 * varasm.c (output_constant_def_1): Get the decl representing
25717 the constant as argument.
25718 (output_constant_def): Wrap output_constant_def_1.
25719 (make_decl_rtl): Use output_constant_def_1 with the decl
25720 representing the constant.
25721 (build_constant_desc): Optionally re-use a decl already
25722 representing the constant.
25723 (tree_output_constant_def): Adjust.
25725 2013-02-14 Dodji Seketeli <dodji@redhat.com>
25728 * asan.c (instrument_builtin_call): Really put the length of the
25729 second source argument into src1_len.
25731 2013-02-13 Jakub Jelinek <jakub@redhat.com>
25733 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
25734 argument. If it is false, don't create edge from then_bb to
25736 (insert_if_then_before_iter): Pass true to it.
25737 (build_check_stmt): Pass false to it.
25738 (transform_statements): Flush hash table only on extended basic
25739 block boundaries, rather than at the beginning of every bb.
25740 Don't flush hash table on nonfreeing_call_p calls.
25741 * tree-flow.h (nonfreeing_call_p): New prototype.
25742 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
25744 2013-02-13 David S. Miller <davem@davemloft.net>
25746 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
25748 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
25751 * ira.c (max_regno_before_ira): Move from ...
25753 (fix_reg_equiv_init): Use max_regno_before_ira instead of
25756 2013-02-13 Jakub Jelinek <jakub@redhat.com>
25758 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
25760 2013-02-13 Richard Biener <rguenther@suse.de>
25763 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
25764 globals in MEM_REFs.
25766 2013-02-13 Richard Biener <rguenther@suse.de>
25768 * loop-init.c (loop_optimizer_init): Clear loop state when
25769 re-initializing preserved loops.
25770 * loop-unswitch.c (unswitch_single_loop): Return whether
25771 we unswitched the loop. Do not verify loop state here.
25772 (unswitch_loops): When we unswitched a loop discover new loops.
25774 2013-02-13 Kostya Serebryany <kcc@google.com>
25776 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
25778 * sanitizer.def: Rename __asan_init to __asan_init_v1.
25780 2013-02-12 Dodji Seketeli <dodji@redhat.com>
25782 Avoid instrumenting duplicated memory access in the same basic block
25783 * Makefile.in (asan.o): Add new dependency on hash-table.h
25784 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
25785 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
25786 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
25787 (free_mem_ref_resources, has_mem_ref_been_instrumented)
25788 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
25789 (get_mem_ref_of_assignment): New functions.
25790 (get_mem_refs_of_builtin_call): Extract from
25791 instrument_builtin_call and tweak a little bit to make it fit with
25793 (instrument_builtin_call): Use the new
25794 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
25795 of is_gimple_builtin_call.
25796 (instrument_derefs, instrument_mem_region_access): Insert the
25797 instrumented memory reference into the hash table.
25798 (maybe_instrument_assignment): Renamed instrument_assignment into
25799 this, and change it to advance the iterator when instrumentation
25800 actually happened and return true in that case. This makes it
25801 homogeneous with maybe_instrument_assignment, and thus give a
25802 chance to callers to be more 'regular'.
25803 (transform_statements): Clear the memory reference hash table
25804 whenever we enter a new BB, when we cross a function call, or when
25805 we are done transforming statements. Use
25806 maybe_instrument_assignment instead of instrumentation. No more
25807 need to special case maybe_instrument_assignment and advance the
25808 iterator after calling it; it's now handled just like
25809 maybe_instrument_call. Update comment.
25811 2013-02-13 Richard Biener <rguenther@suse.de>
25813 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
25814 Fix loop discovery code.
25816 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
25818 PR inline-asm/56148
25819 * lra-constraints.c (process_alt_operands): Match early clobber
25820 operand with itself. Check conflicts with earlyclobber only if
25821 the operand is not reloaded. Prefer to reload conflicting operand
25822 if earlyclobber and matching operands are the same.
25824 2013-02-12 Richard Biener <rguenther@suse.de>
25827 * lto-streamer-out.c (write_symbol): Do not output symbols
25828 for hard register variables.
25830 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
25833 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
25834 (umulsidi3_insn, mulsidi3_insn): New insns.
25836 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
25838 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
25839 (struct tune_params): Add vec_costs field.
25840 * config/arm/arm.c (arm_builtin_vectorization_cost)
25841 (arm_add_stmt_cost): New functions.
25842 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
25843 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
25844 (arm_default_vec_cost): New struct of type cpu_vec_costs.
25845 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
25846 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
25847 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
25848 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
25850 2013-02-12 Richard Biener <rguenther@suse.de>
25853 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
25854 decls again if possible.
25856 2013-02-12 Richard Biener <rguenther@suse.de>
25858 PR middle-end/56288
25859 * tree-ssa.c (verify_ssa_name): Fix check, move
25860 SSA_NAME_IN_FREE_LIST check up.
25862 2013-02-12 Jakub Jelinek <jakub@redhat.com>
25863 Steven Bosscher <steven@gcc.gnu.org>
25865 PR rtl-optimization/56151
25866 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
25867 equal to op0 or op1, and last_insn pattern is CODE operation
25868 with MEM dest and one of the operands matches that MEM.
25870 2013-02-11 Sriraman Tallam <tmsriram@google.com>
25872 * doc/extend.texi: Document Function Multiversioning and "default"
25873 parameter string to target attribute.
25874 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
25875 target attribute parameter is "default".
25876 (ix86_compare_version_priority): Remove checks for target attribute.
25877 (ix86_mangle_function_version_assembler_name): Change error to sorry.
25878 Remove check for target attribute equal to NULL. Add assert.
25879 (ix86_generate_version_dispatcher_body): Change error to sorry.
25881 2013-02-11 Iain Sandoe <iain@codesourcery.com>
25882 Jack Howarth <howarth@bromo.med.uc.edu>
25883 Patrick Marlier <patrick.marlier@gmail.com>
25886 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
25887 define ENDFILE_SPEC as TM_DESTRUCTOR.
25888 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
25890 2013-02-11 Alexander Potapenko <glider@google.com>
25891 Jack Howarth <howarth@bromo.med.uc.edu>
25892 Jakub Jelinek <jakub@redhat.com>
25895 * config/darwin.c (cdtor_record): Rename ctor_record.
25896 (sort_cdtor_records): Rename sort_ctor_records.
25897 (finalize_dtors): New routine to sort destructors by
25898 priority before use in assemble_integer.
25899 (machopic_asm_out_destructor): Use finalize_dtors if needed.
25901 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
25903 PR rtl-optimization/56275
25904 * simplify-rtx.c (avoid_constant_pool_reference): Check that
25905 offset is non-negative and less than cmode size before
25906 calling simplify_subreg.
25908 2013-02-11 Richard Biener <rguenther@suse.de>
25910 PR tree-optimization/56264
25911 * cfgloop.h (fix_loop_structure): Adjust prototype.
25912 * loop-init.c (fix_loop_structure): Return the number of
25913 newly discovered loops.
25914 * tree-cfgcleanup.c (repair_loop_structures): When new loops
25915 are discovered, do a full loop-closed SSA rewrite.
25917 2013-02-11 Richard Biener <rguenther@suse.de>
25919 PR tree-optimization/56273
25920 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
25922 (check_array_ref): Fix missing newline in dumps.
25923 (search_for_addr_array): Likewise.
25925 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
25927 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
25929 2013-02-09 Jakub Jelinek <jakub@redhat.com>
25932 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
25934 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
25936 PR rtl-optimization/56246
25937 * lra-constraints.c (simplify_operand_subreg): Try to reuse
25939 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
25940 constraints are satisfied.
25942 2013-02-08 Jeff Law <law@redhat.com>
25945 * emit-rtl.c (reg_is_parm_p): New function.
25946 * regs.h (reg_is_parm_p): New prototype.
25947 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
25948 callee-clobbered registers.
25950 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
25953 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
25954 If there is no implicit builtin declaration, just return NULL.
25956 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
25958 * config/i386/sse.md (FMAMODEM): New mode iterator.
25959 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
25960 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
25962 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
25964 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
25965 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
25966 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
25968 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
25970 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
25971 (microblaze*-*-elf): Likewise.
25972 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
25974 * config/microblaze/microblaze-c.c: Add builtin defines for
25975 _LITTLE_ENDIAN and _BIG_ENDIAN.
25976 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
25977 add to TARGET_DEFAULT flags.
25978 Expand ASM_SPEC and LINK_SPEC.
25979 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
25980 * config/microblaze/microblaze.md: Update extendsidi2 and
25981 movdi_internal instructions to use low-order / high-order reg
25983 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
25984 options and inversemask / mask of LITTLE_ENDIAN.
25985 * config/microblaze/t-microblaze: Expand multilib options to
25986 include mlittle-endian (le) and update exceptions patterns.
25988 2013-02-08 Jakub Jelinek <jakub@redhat.com>
25990 PR rtl-optimization/56195
25991 * lra-constraints.c (get_reload_reg): Don't reuse regs
25992 if they have smaller mode than requested, if they have
25993 wider mode than requested, try to return a SUBREG.
25995 PR tree-optimization/56250
25996 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
25997 if type is unsigned and code isn't MULT_EXPR.
25999 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
26001 PR tree-optimization/56064
26002 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
26003 bits according to mode.
26004 * fixed-value.h (fixed_from_double_int)
26005 (const_fixed_from_double_int): Adjust comments.
26007 2013-02-08 Richard Biener <rguenther@suse.de>
26010 * lto-streamer.h (struct data_in): Remove current_file, current_line
26011 and current_col members.
26012 * lto-streamer-out.c (lto_output_location): Stream changed bits
26013 en-block for efficiency.
26014 * lto-streamer-in.c (clear_line_info): Remove.
26015 (lto_input_location): Cache current file, line and column
26016 globally via local statics. Read changed bits en-block.
26017 (input_function): Do not call clear_line_info.
26018 (lto_read_body): Likewise.
26019 (lto_input_toplevel_asms): Likewise.
26021 2013-02-08 Michael Matz <matz@suse.de>
26023 PR tree-optimization/52448
26024 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
26025 (nt_call_phase): New static.
26026 (add_or_mark_expr): Only mark accesses with newer phase than any
26028 (nonfreeing_call_p): New.
26029 (nt_init_block): Update nt_call_phase, mark blocks as visited.
26030 (nt_fini_block): Keep blocks marked as visited.
26031 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
26033 2013-02-08 Richard Biener <rguenther@suse.de>
26035 * ira.c (ira): Free broken dominator information.
26037 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
26039 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
26041 2013-02-08 Marek Polacek <polacek@redhat.com>
26043 * cfgloop.c (verify_loop_structure): Add more checking of headers.
26045 2013-02-08 Richard Biener <rguenther@suse.de>
26047 PR middle-end/56181
26048 * cfgloop.h (flow_loops_find): Adjust.
26049 (bb_loop_header_p): Declare.
26050 * cfgloop.c (bb_loop_header_p): New function split out from ...
26051 (flow_loops_find): ... here. Adjust function signature,
26052 support incremental loop structure update.
26053 (verify_loop_structure): Cleanup. Verify a loop is a loop.
26054 * cfgloopmanip.c (fix_loop_structure): Move ...
26055 * loop-init.c (fix_loop_structure): ... here.
26056 (apply_loop_flags): Split out from ...
26057 (loop_optimizer_init): ... here.
26058 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
26059 in incremental mode, only remove dead loops here.
26061 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
26064 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
26065 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
26066 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
26067 (*round<mode>3.libgcc): New insns for fixed-modes.
26068 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
26069 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
26070 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
26071 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
26072 implementations. Define to __builtin_avr_absFX,
26073 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
26074 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
26075 __builtin_avr_countlsFX, respectively.
26076 * config/avr/avr-c.c (target.h): Include it.
26077 (enum avr_builtin_id): New enum.
26078 (avr_resolve_overloaded_builtin): New static function.
26079 (avr_register_target_pragmas): Use it to set
26080 targetm.resolve_overloaded_builtin.
26081 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
26082 tree nodes used by DEF_BUILTIN.
26083 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
26084 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
26085 <AVR_BUILTIN_xxBITS>: Same.
26087 2013-02-08 Richard Biener <rguenther@suse.de>
26089 * cfgloop.c (verify_loop_structure): Properly handle
26090 a loop exiting to another loop header.
26091 * ira-int.h (ira_loops): Remove.
26092 * ira.c (ira_loops): Remove.
26093 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
26094 (do_reload): Use loop_optimizer_finalize.
26095 * ira-build.c (create_loop_tree_nodes): Use get_loops and
26096 number_of_loops to access the loop tree.
26097 (more_one_region_p): Likewise.
26098 (finish_loop_tree_nodes): Likewise.
26099 (rebuild_regno_allocno_maps): Likewise.
26100 (mark_loops_for_removal): Likewise.
26101 (mark_all_loops_for_removal): Likewise.
26102 (remove_unnecessary_regions): Likewise.
26103 (ira_build): Likewise.
26104 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
26106 2013-02-08 Richard Biener <rguenther@suse.de>
26108 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
26109 * ipa-pure-const.c (analyze_function): Avoid calling
26110 mark_irreducible_loops twice.
26111 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
26113 2013-02-07 David S. Miller <davem@davemloft.net>
26115 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
26117 * var-tracking.c (vt_add_function_parameter): Test the presence of
26118 HAVE_window_save properly and do not remap argument registers when
26119 we have a leaf function.
26121 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
26124 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
26126 * config/i386/i386.c (ix86_print_operand): Ditto.
26128 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
26130 * lra-constraints.c (process_alt_operands): Fix recently added comment.
26132 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
26134 PR rtl-optimization/56225
26135 * lra-constraints.c (process_alt_operands): Check that reload hard
26136 reg can hold value for strict_low_part.
26138 2013-02-07 Jakub Jelinek <jakub@redhat.com>
26141 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
26142 dwarf2out_end_function.
26143 (in_first_function_p, maybe_at_text_label_p,
26144 first_loclabel_num_not_at_text_label): New variables.
26145 (dwarf2out_var_location): In the first function find out
26146 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
26147 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
26150 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
26152 PR rtl-optimization/56178
26153 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
26154 SUBREG of a register. Tidy up related block of code.
26155 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
26156 note if the source is a register or a SUBREG of a register.
26158 2013-02-07 Jakub Jelinek <jakub@redhat.com>
26161 * config/rs6000/rs6000.md (ptrm): New mode attr.
26162 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
26163 call_value_indirect_aix<pttrsize>,
26164 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
26167 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
26169 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
26170 if -bnortl. Convert to strcmp and strncmp.
26172 2013-02-07 Alan Modra <amodra@gmail.com>
26175 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
26176 addresses won't wrap when offsetting.
26177 (rs6000_secondary_reload): Provide secondary reloads needed for
26178 wrapping LO_SUM addresses.
26180 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
26182 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
26183 MACH, just __MACH__.
26185 2013-02-06 Richard Biener <rguenther@suse.de>
26187 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
26188 instead of calling fix_loop_structure.
26190 2013-02-06 Jakub Jelinek <jakub@redhat.com>
26192 PR middle-end/56217
26193 * omp-low.c (use_pointer_for_field): Return false if
26194 lower_send_shared_vars doesn't generate any copy-out code.
26196 2013-02-06 Tom de Vries <tom@codesourcery.com>
26198 PR rtl-optimization/56131
26199 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
26200 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
26201 of the label is NULL. Add comment.
26203 2013-02-05 Jakub Jelinek <jakub@redhat.com>
26205 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
26208 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
26209 (STATIC_LIBTSAN_LIBS): Likewise.
26210 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
26211 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
26212 is defined, don't add anything else beyond that.
26213 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
26214 (LINK_COMMAND_SPEC): Use them.
26216 PR tree-optimization/56205
26217 * tree-stdarg.c (check_all_va_list_escapes): Return true if
26218 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
26219 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
26221 2013-02-05 Richard Biener <rguenther@suse.de>
26223 PR tree-optimization/53342
26224 PR tree-optimization/53185
26225 * tree-vectorizer.h (vect_check_strided_load): Remove.
26226 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
26227 not disallow peeling for vectorized strided loads.
26228 (vect_check_strided_load): Make static and simplify.
26229 (vect_analyze_data_refs): Adjust.
26230 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
26231 correctly when vectorizing strided loads.
26233 2013-02-05 Richard Biener <rguenther@suse.de>
26235 * doc/install.texi: Refer to ISL, not PPL.
26237 2013-02-05 Jan Hubicka <jh@suse.cz>
26239 PR tree-optimization/55789
26240 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
26242 2013-02-05 Jan Hubicka <jh@suse.cz>
26244 PR tree-optimization/55789
26245 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
26246 the dead call anyway.
26248 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
26251 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
26253 2013-02-04 Alexander Potapenko <glider@google.com>
26254 Jack Howarth <howarth@bromo.med.uc.edu>
26255 Jakub Jelinek <jakub@redhat.com>
26258 * config/darwin.c (sort_ctor_records): Stabilized qsort
26259 on constructor priority by using original position.
26260 (finalize_ctors): New routine to sort constructors by
26261 priority before use in assemble_integer.
26262 (machopic_asm_out_constructor): Use finalize_ctors if needed.
26264 2013-02-04 Jakub Jelinek <jakub@redhat.com>
26267 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
26268 about visibility on artificial decls.
26269 * config/sol2.c (solaris_assemble_visibility): Likewise.
26271 2013-02-04 Kai Tietz <ktietz@redhat.com>
26274 * config/i386/i386.c (function_value_ms_64): Add additional valtype
26275 argument and improve checking of return-argument types for 16-byte
26277 (ix86_function_value_1): Add additional valtype argument on call
26278 of function_value_64.
26279 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
26280 handling infunction_value_64 function.
26282 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
26284 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
26286 2013-02-04 Richard Biener <rguenther@suse.de>
26288 PR tree-optimization/56188
26289 * tree-ssa-structalias.c (label_visit): Consider case with
26290 initially non-empty points-to set.
26291 (perform_var_substitution): Dump node mapping and clean up.
26293 2013-02-04 Richard Guenther <rguenther@suse.de>
26296 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
26297 node prevail as last resort.
26298 (lto_symtab_merge_decls): Remove guard on LTRANS here.
26299 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
26301 2013-02-04 Richard Biener <rguenther@suse.de>
26303 PR tree-optimization/56113
26304 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
26306 (equiv_class_lookup_or_add): ... this.
26307 (label_visit): Adjust and fix error in previous patch.
26308 (perform_var_substitution): Adjust.
26310 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
26312 * config/sh/divtab.c: Fix formatting and comments throughout the file.
26313 * config/sh/sh4-300.md: Likewise.
26314 * config/sh/sh4a.md: Likewise.
26315 * config/sh/constraints.md: Likewise.
26316 * config/sh/sh.md: Likewise.
26317 * config/sh/netbsd-elf.h: Likewise.
26318 * config/sh/predicates.md: Likewise.
26319 * config/sh/sh-protos.h: Likewise.
26320 * config/sh/ushmedia.h: Likewise.
26321 * config/sh/linux.h: Likewise.
26322 * config/sh/sh.c: Likewise.
26323 * config/sh/superh.h: Likewise.
26324 * config/sh/elf.h: Likewise.
26325 * config/sh/sh4.md: Likewise.
26326 * config/sh/sh.h: Likewise.
26328 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26330 * config/pa/constraints.md: Adjust unused letters. Change "T"
26331 constraint to match_test floating_point_store_memory_operand().
26332 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
26333 (base14_operand): New.
26334 (floating_point_store_memory_operand): New.
26335 (integer_store_memory_operand): Revise to use base14_operand and
26336 reg_plus_base_memory_operand.
26337 (move_dest_operand): Allow symbolic_memory_operands.
26338 (symbolic_memory_operand): Check for LO_SOM.
26339 (symbolic_operand): Change default case to break.
26340 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
26341 CONST_DOUBLE values to be reloaded by putting them into memory when
26342 the destination is a floating point register.
26343 (movdf): Remove code to handle CONST_DOUBLE.
26345 (reload_indf_r1): New.
26346 (reload_insf_r1): New.
26347 Consistently use "Q" and "T" constraints with integer and floating
26348 point move instructions, respectively.
26349 (movdi): Remove FAIL.
26350 Change predicate for source operand unamed DImode move from
26351 general_operand to move_src_operand.
26352 (umulsidi3): Change predicate for destination operand to
26354 Likewise for similar unamed patterns.
26355 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
26356 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
26357 (hppa_legitimize_address): Simplify mask calculation.
26358 (pa_emit_move_sequence): Revised handling of secondary reloads from
26359 REG+D addresses for floating point loads and stores. Directly handle
26360 loading CONST0_RTX (mode) to a floating point register.
26361 (pa_secondary_reload): Handle reloading DF and SFmode constant values
26362 to floating point registers. Don't restrict secondary reloads to
26363 floating point registers to integer modes. Revise some comments and
26365 (TARGET_LEGITIMATE_ADDRESS_P): Define.
26366 (pa_legitimate_address_p): New.
26367 (pa_legitimize_reload_address): New.
26368 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
26369 (STRICT_REG_OK_FOR_BASE_P): New.
26370 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
26371 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
26373 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
26374 Andrew Dixie <andrewd@gentrack.com>
26376 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
26379 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
26381 * expmed.c (extract_bit_field_1): Pass the full width of the
26382 structure to get_best_reg_extraction_insn.
26384 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
26387 * configure.ac (use_cxa_atexit): Add AIX.
26388 * configure: Regenerate.
26390 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
26392 2013-02-01 Jakub Jelinek <jakub@redhat.com>
26395 * final.c (need_profile_function): New variable.
26396 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
26397 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
26398 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
26399 notes, targetm.asm_out.function_prologue doesn't emit anything,
26400 HAVE_prologue and profiler should be emitted before prologue,
26401 set need_profile_function instead of emitting it.
26402 (final_scan_insn): If need_profile_function, emit
26403 profile_function on the first NOTE_INSN_BASIC_BLOCK or
26404 NOTE_INSN_FUNCTION_BEG note.
26406 2013-02-01 Richard Henderson <rth@redhat.com>
26408 * config/rs6000/rs6000.md (smulditi3): New.
26411 * config/alpha/alpha.md (umulditi3): New.
26413 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
26415 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
26416 (ASM_OUTPUT_ALIGNED_LOCAL): New.
26418 2013-02-01 Richard Biener <rguenther@suse.de>
26420 PR tree-optimization/56113
26421 * tree-ssa-structalias.c (label_visit): Reduce work for
26422 single-predecessor nodes.
26424 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
26426 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
26427 range isn't testing for zero.
26429 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
26431 PR middle-end/56113
26432 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
26434 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
26435 Nick Clifton <nickc@redhat.com>
26437 * config/v850/constraints.md (Q): Define as a memory constraint.
26438 * config/v850/predicates.md (label_ref_operand): New predicate.
26439 (e3v5_shift_operand): New predicate.
26440 (ior_operator): New predicate.
26441 * config/v850/t-v850: Add e3v5 multilib.
26442 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
26443 (v850_gen_movdi): Prototype.
26444 * config/v850/v850.c: Add support for e3v5 architecture.
26445 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
26447 (construct_save_jarl): Add e3v5 long JARL support.
26448 (v850_adjust_insn_length): New function. Adjust length of call
26449 insns when using e3v5 instructions.
26450 (v850_gen_movdi): New function: Generate instructions to move a
26452 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
26453 (CPP_SPEC): Define __v850e3v5__ as appropriate.
26454 (TARGET_USE_FPU): Enable for e3v5.
26455 (CONST_OK_FOR_W): New macro.
26456 (ADJUST_INSN_LENGTH): Define.
26457 * config/v850/v850.md (UNSPEC_LOOP): Define.
26458 (attr cpu): Add v850e3v5.
26459 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
26460 (movdi): New pattern.
26461 (movdi_internal): New pattern.
26462 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
26463 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
26464 (cstoresf4): Likewise.
26465 (cstoredf4): Likewise.
26466 (insv): New pattern.
26467 (rotlso3_a): New pattern.
26468 (rotlsi3_b): New pattern
26469 (rotlsi3_v850e3v5): New pattern.
26470 (doloop_begin): New pattern.
26471 (fix_loop_counter): New pattern.
26472 (doloop_end): New pattern.
26473 (branch_normal): Add e3v5 long branch support.
26474 (branch_invert): Likewise.
26475 (branch_z_normal): Likewise.
26476 (branch_z_invert): Likewise.
26477 (branch_nz_normal): Likewise.
26478 (branch_nz_invert): Likewise.
26479 (call_internal_short): Add e3v5 register-indirect JARL support.
26480 (call_internal_long): Likewise.
26481 (call_value_internal_short): Likewise.
26482 (call_value_internal_long): Likewise.
26483 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
26484 (mloop): New option.
26485 * config.gcc: Add support for configuring v840e3v5 target.
26486 * doc/invoke.texi: Document new v850 specific command line options.
26488 2013-01-31 Paul Koning <ni1d@arrl.net>
26492 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
26493 children if parent is a class.
26494 (prune_unused_types_prune): Don't add DW_AT_declaration.
26496 2013-01-31 Richard Biener <rguenther@suse.de>
26498 PR tree-optimization/56157
26499 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
26500 match up operand with SLP child.
26502 2013-01-31 Jason Merrill <jason@redhat.com>
26505 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
26506 parameters the first time.
26507 (gen_scheduled_generic_parms_dies): Check completeness here.
26509 2013-01-31 Richard Biener <rguenther@suse.de>
26511 PR middle-end/53073
26512 * common.opt (faggressive-loop-optimizations): New flag,
26513 enabled by default.
26514 * doc/invoke.texi (faggressive-loop-optimizations): Document.
26515 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
26516 infer_loop_bounds_from_undefined by it.
26518 2013-01-31 Richard Biener <rguenther@suse.de>
26520 PR tree-optimization/56150
26521 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
26522 visit virtual operands.
26523 (find_uses_to_rename_bb): Likewise.
26525 2013-01-31 Richard Biener <rguenther@suse.de>
26527 PR tree-optimization/56150
26528 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
26529 mixed store non-store stmts.
26531 2013-01-30 Jakub Jelinek <jakub@redhat.com>
26534 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
26535 LIBASAN_EARLY_SPEC is defined.
26536 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
26537 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
26539 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
26542 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
26543 invalid args instead of ICEing on it.
26544 (ix86_valid_target_attribute_tree): Return error_mark_node if
26545 ix86_valid_target_attribute_inner_p failed.
26546 (ix86_valid_target_attribute_p): Return false only if
26547 ix86_valid_target_attribute_tree returned error_mark_node. Allow
26548 target("default") attribute.
26549 (sorted_attr_string): Change argument from const char * to tree,
26550 merge in all target attribute arguments rather than just one.
26551 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
26552 instead of free. Avoid using strcat.
26553 (ix86_mangle_function_version_assembler_name): Mangle
26554 target("default") as if no target attribute is present. Adjust
26555 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
26556 instead of xmalloc and XDELETEVEC instead of free.
26557 (ix86_function_versions): Don't return true if one of the decls
26558 doesn't have target attribute. If they don't and one of the decls
26559 is DECL_FUNCTION_VERSIONED, report an error. Adjust
26560 sorted_attr_string caller. Use XDELETEVEC instead of free.
26561 (ix86_supports_function_versions): Remove.
26562 (make_name): Fix up formatting.
26563 (make_dispatcher_decl): Remove resolver_name and its initialization.
26564 Avoid leaking memory.
26565 (is_function_default_version): Return true if there is
26566 target("default") attribute rather than no target attribute at all.
26567 (make_resolver_func): Avoid leaking memory.
26568 (ix86_generate_version_dispatcher_body): Likewise.
26569 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
26570 * target.def (supports_function_versions): Remove.
26571 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
26572 * doc/tm.texi: Regenerated.
26574 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
26576 PR rtl-optimization/56144
26577 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
26578 for values with side effects.
26580 2013-01-30 Richard Biener <rguenther@suse.de>
26582 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
26583 (sparseset_pop): Likewise.
26584 * cfganal.c (compute_idf): Likewise. Increase work-stack size
26585 to be able to use quick_push in the worker loop.
26587 2013-01-30 Marek Polacek <polacek@redhat.com>
26589 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
26591 2013-01-30 Richard Biener <rguenther@suse.de>
26594 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
26596 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
26598 PR tree-optimization/56064
26599 * fixed-value.c (fixed_from_double_int): New function.
26600 * fixed-value.h (fixed_from_double_int): New prototype.
26601 (const_fixed_from_double_int): New static inline function.
26602 * fold-const.c (native_interpret_fixed): New static function.
26603 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
26604 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
26605 (native_encode_fixed): New static function.
26606 (native_encode_expr) <FIXED_CST>: Use it.
26607 (native_interpret_int): Move double_int worker code to...
26608 * double-int.c (double_int::from_buffer): ...this new static method.
26609 * double-int.h (double_int::from_buffer): Prototype it.
26611 2013-01-30 Richard Biener <rguenther@suse.de>
26613 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
26614 New pointer-map and obstack.
26615 (init_alias_vars): Allocate pointer-map and obstack.
26616 (delete_points_to_sets): Free them.
26617 (find_what_var_points_to): Cache result.
26618 (find_what_p_points_to): Adjust for changed interface of
26619 find_what_var_points_to.
26620 (compute_points_to_sets): Likewise.
26621 (ipa_pta_execute): Likewise.
26623 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
26625 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
26626 * configure: Regenerate.
26627 * config.in: Regenerate.
26628 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
26629 #nobits/#progbits if supported.
26631 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
26634 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
26635 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
26636 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
26638 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
26640 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
26641 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
26643 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
26645 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
26646 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
26648 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
26650 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
26652 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
26653 * config/arm/cortex-a7.md: New bypasses using
26654 arm_mac_accumulator_is_result.
26656 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
26658 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
26659 (cortex_a7_neon_mla): Likewise.
26660 (cortex_a7_fpfmad): New reservation.
26661 (cortex_a7_fpmacs): Use ffmas and update required units.
26662 (cortex_a7_fpmuld): Update required units and latency.
26663 (cortex_a7_fpmacd): Likewise.
26664 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
26665 (cortex_a7_neon). Likewise.
26666 (bypass) Update participating units.
26668 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
26670 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
26671 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
26673 * config/arm/vfp11.md (vfp_farith): Use ffmas.
26674 (vfp_fmul): Use ffmad.
26675 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
26676 (cortex_r4_fmacd): Use ffmad.
26677 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
26678 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
26679 (cortex_a9_fmacd): Use ffmad.
26680 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
26681 (cortex_a8_vfp_macd): Use ffmad.
26682 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
26683 (cortex_a5_fpmacd): Use ffmad.
26684 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
26685 (cortex_a15_vfp_macd): Use ffmad.
26686 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
26688 2013-01-29 Jason Merrill <jason@redhat.com>
26691 * varasm.c (default_assemble_visibility): Don't warn about
26692 visibility on artificial decls.
26694 2013-01-29 Richard Biener <rguenther@suse.de>
26696 PR tree-optimization/56113
26697 * tree-ssa-structalias.c (equiv_class_lookup): Also return
26699 (label_visit): Free duplicate bitmaps and record the leader instead.
26700 (perform_var_substitution): Adjust.
26702 2013-01-29 Richard Biener <rguenther@suse.de>
26704 PR tree-optimization/55270
26705 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
26706 the CFG, schedule loops for fixup.
26708 2013-01-29 Nick Clifton <nickc@redhat.com>
26710 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
26713 2013-01-28 Leif Ekblad <leif@rdos.net>
26715 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
26716 * config/i386/i386.h (TARGET_RDOS): New macro.
26717 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
26718 * config/i386/i386.c (ix86_option_override_internal): For 64bit
26719 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
26720 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
26721 DEFAULT_LARGE_SECTION_THRESHOLD.
26722 * config/i386/i386.md (R14_REG, R15_REG): New constants.
26723 * config/i386/rdos.h: New file.
26724 * config/i386/rdos64.h: New file.
26726 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
26729 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
26732 2013-01-28 Jakub Jelinek <jakub@redhat.com>
26734 PR rtl-optimization/56117
26735 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
26736 call cselib_lookup_from_insn on the MEM before calling
26737 add_insn_mem_dependence.
26739 2013-01-28 Richard Biener <rguenther@suse.de>
26741 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
26742 to a stmt that didn't have one.
26743 (copy_phis_for_bb): Likewise for PHI arguments.
26744 (copy_debug_stmt): Likewise for debug stmts.
26746 2013-01-28 Richard Biener <rguenther@suse.de>
26748 PR tree-optimization/56034
26749 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
26750 (partition_builtin_p): Adjust.
26751 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
26752 it is the last partition.
26753 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
26754 up the vertex for the definition.
26755 (classify_partition): Classify whether a partition is a
26756 PKIND_REDUCTION, thus has uses outside of the loop.
26757 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
26758 Merge all PKIND_REDUCTION partitions into the last partition.
26759 (tree_loop_distribution): Seed partitions from reductions as well.
26761 2013-01-28 Jakub Jelinek <jakub@redhat.com>
26763 PR tree-optimization/56125
26764 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
26765 pow(x,c) into sqrt(x) * powi(x, n/2) or
26766 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
26767 optimizing for size.
26768 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
26769 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
26772 PR tree-optimization/56094
26773 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
26774 to UNKNOWN_LOCATION while gimplifying expr.
26776 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
26779 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
26780 operand 0 in movabs insn template for -masm=intel asm alternative.
26781 (*movabs<mode>_2): Ditto for operand 1.
26783 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
26786 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
26789 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
26791 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
26794 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
26796 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
26797 Undef to avoid warning.
26799 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
26801 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
26802 * configure: Regenerate.
26804 2013-01-25 Jakub Jelinek <jakub@redhat.com>
26806 PR tree-optimization/56098
26807 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
26808 for stmts with volatile ops.
26809 (cond_store_replacement): Don't optimize if assign has volatile ops.
26810 (cond_if_else_store_replacement_1): Don't optimize if either
26811 then_assign or else_assign have volatile ops.
26812 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
26815 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
26817 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
26819 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
26821 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
26822 missing ':' in asm example.
26824 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
26826 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
26827 entries into lane and laneq entries.
26828 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
26829 Remove AdvSIMD scalar modes.
26830 (aarch64_sq<r>dmulh_laneq<mode>): New.
26831 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
26833 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
26834 builtin implementations to relfect changes in RTL in aarch64-simd.md.
26835 * config/aarch64/iterators.md (VCOND): New.
26838 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
26841 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
26842 Add NULL LIBNAME argument to existing definitions.
26843 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
26844 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
26845 * config/avr/avr.c (DEF_BUILTIN): Same.
26846 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
26847 (avr_expand_builtin): Expand to a vanilla call if a libgcc
26848 implementation is available (DECL_ASSEMBLER_NAME is set).
26849 (avr_fold_absfx): New static function.
26850 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
26851 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
26852 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
26853 AVR_BUILTIN_ABSLLK.
26854 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
26855 (abshk, absk, abslk, absllk): Provide as static inline functions.
26857 2013-01-25 Marek Polacek <polacek@redhat.com>
26859 PR tree-optimization/56035
26860 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
26862 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
26864 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
26865 (*movtf_internal_rex64): Add (!o,C) alternative
26866 (*movxf_internal_rex64): Ditto.
26867 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
26869 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
26871 * doc/invoke.texi: fix typo.
26872 * doc/objc.texi: fix typo.
26874 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
26876 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
26877 for the first two alternatives.
26879 2013-01-24 Diego Novillo <dnovillo@google.com>
26881 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
26882 (ggc-zone.o): Remove.
26883 * configure.ac: Remove option --with-gc.
26884 * configure: Re-generate.
26885 * doc/install.texi: Remove documentation for --with-gc.
26886 * gengtype.c (write_enum_defn): Remove. Update all users.
26887 (write_Types_process_field): Remove generation of gt_e_* argument.
26888 (output_type_enum): Remove. Update all users.
26889 (write_enum_defn): Remove. Update all users.
26890 (enum alloc_zone): Remove. Update all users.
26891 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
26892 * ggc-common.c (ggc_splay_alloc): Remove first argument.
26893 Update all callers.
26894 (struct ptr_data): Remove field TYPE. Update all users.
26895 (gt_pch_note_object): Remove argument TYPE. Update all users.
26896 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
26898 * ggc-none.c (ggc_alloc_typed_stat): Remove.
26899 (struct alloc_zone): Remove.
26900 (ggc_internal_alloc_zone_stat): Remove.
26901 (ggc_internal_cleared_alloc_zone_stat): Remove.
26902 * ggc-page.c (ggc_alloc_typed_stat): Remove.
26903 (ggc_pch_count_object): Remove last argument. Update all users.
26904 (ggc_pch_alloc_object): Remove last argument. Update all users.
26905 (struct alloc_zone): Remove.
26906 * ggc-zone.c: Remove.
26907 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
26908 (struct alloc_zone): Remove.
26909 (ggc_alloc_typed_stat): Remove.
26910 (ggc_alloc_typed): Remove.
26911 (ggc_splay_alloc): Remove first argument.
26912 (rtl_zone): Remove. Update all users.
26913 (tree_zone): Remove. Update all users.
26914 (tree_id_zone): Remove. Update all users.
26915 (ggc_internal_zone_alloc_stat): Remove. Update all users.
26916 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
26917 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
26918 * tree-ssanames.c: Remove references to zone allocator in comments.
26920 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
26922 * config/avr/avr.c (avr_out_fract): Make register numbers that
26923 might be outside of source operand signed.
26925 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
26927 * config/i386/constraints.md (Yf): New constraint.
26928 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
26929 of f constraint to conditionaly disable x87 register preferences.
26930 (*movdf_internal): Ditto.
26931 (*movsf_internal): Ditto.
26933 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
26935 PR inline-asm/55934
26936 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
26937 that have operands with impossible constraints.
26938 Add a FIXME for a speed-up opportunity.
26939 * lra-constraints.c (process_alt_operands): Verify that a class
26940 selected from constraints on asms is valid for the operand mode.
26941 (curr_insn_transform): Remove incorrect comment.
26943 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
26945 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
26946 TOC operand is a valid symbol ref in the constant pool.
26948 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
26950 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
26952 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
26955 * config/avr/stdfix.h: New file.
26956 * t-avr (stdfix-gcc.h): New rule to build it.
26957 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
26959 2013-01-23 Kostya Serebryany <kcc@google.com>
26961 * config/darwin.h: remove dependency on
26962 CoreFoundation (asan on Mac OS).
26964 2013-01-23 Jakub Jelinek <jakub@redhat.com>
26967 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
26968 instead of cmpdi_operand for first comparison operand.
26969 Don't assert that comparison operands aren't both constants.
26971 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
26973 * doc/install.texi (Downloading the Source): Update references to
26974 downloading separate components.
26976 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
26978 * doc/extend.texi (__int128): Improve grammar.
26980 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
26983 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
26984 alternative to (o,r).
26985 (*movdi_internal_rex64): Remove (!o,n) alternative.
26986 (DImode immediate->memory splitter): Remove.
26987 (DImode immediate->memory peephole2): Remove.
26988 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
26989 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
26990 alternative to (!o,*r).
26991 (*movtf_internal_sse): New pattern.
26992 (*movxf_internal_rex64): New pattern.
26993 (*movxf_internal): Disable for TARGET_64BIT.
26994 (*movdf_internal_rex64): Remove (!o,F) alternative.
26996 2013-01-22 Jakub Jelinek <jakub@redhat.com>
26998 PR middle-end/56074
26999 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
27000 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
27001 * tree-vect-loop-manip.c (find_loop_location): Also ignore
27002 stmt locations where LOCATION_LOCUS of the stmt location is
27003 UNKNOWN_LOCATION or BUILTINS_LOCATION.
27006 * config/i386/i386.md (UNSPEC_STOS): New.
27007 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
27008 *strsetqi_1): Add UNSPEC_STOS.
27010 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
27013 * doc/invoke.texi: Remove left over -Wsynth example.
27015 2013-01-21 Jakub Jelinek <jakub@redhat.com>
27017 PR tree-optimization/56051
27018 * fold-const.c (fold_binary_loc): Don't fold
27019 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
27020 a narrowing conversion, or widening conversion from signed
27023 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
27025 PR rtl-optimization/56023
27026 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
27027 dependent on debug instruction.
27029 2013-01-21 Martin Jambor <mjambor@suse.cz>
27031 PR middle-end/56022
27032 * function.c (allocate_struct_function): Call
27033 invoke_set_current_function_hook earlier.
27035 2013-01-21 Jakub Jelinek <jakub@redhat.com>
27037 * reload1.c (init_reload): Only initialize reload_obstack
27038 during the first call.
27040 2013-01-21 Marek Polacek <polacek@redhat.com>
27042 * cfgloop.c (verify_loop_structure): Fix up grammar.
27044 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
27046 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
27047 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
27049 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
27052 * config/arm/marvell-pj4.md: Update copyright year.
27053 Fix up use of alu to alu_reg and simple_alu_imm.
27055 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
27057 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
27059 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
27062 * lra-constraints.c (curr_insn_transform): Don't reuse original
27063 insn for secondary memory move when memory mode should be different.
27065 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
27067 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
27068 atomic_storedi_1): New patterns.
27070 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
27072 btver2 pipeline descriptions.
27073 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
27075 * config/i386/i386.md (btver2_decode): New type attributes.
27076 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
27078 * config/i386/btver2.md: New file describing btver2 pipelines.
27080 2013-01-19 Andrew Pinski <apinski@cavium.com>
27082 PR tree-optimization/52631
27083 * tree-ssa-sccvn (visit_use): Before looking up the original
27084 statement, try looking up the simplified expression.
27086 2013-01-19 Anthony Green <green@moxielogic.com>
27088 * config/moxie/moxie.c (moxie_expand_prologue): Set
27089 current_function_static_stack_size.
27091 2013-01-18 Jakub Jelinek <jakub@redhat.com>
27093 PR tree-optimization/56029
27094 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
27095 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
27097 2013-01-18 Sharad Singhai <singhai@google.com>
27099 PR tree-optimization/55995
27100 * dumpfile.c (dump_loc): Print location only if available.
27101 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
27103 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
27106 * lra-constraints.c (curr_insn_transform): Reuse original insn for
27107 secondary memory move.
27108 (inherit_reload_reg): Use rclass instead of cl for
27109 check_secondary_memory_needed_p.
27111 2013-01-18 Jakub Jelinek <jakub@redhat.com>
27113 PR middle-end/56015
27114 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
27115 the case where writing real complex part of target modifies op1.
27117 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
27119 * config/aarch64/aarch64-simd.md
27120 (aarch64_vcond_internal<mode>): Handle unordered cases.
27121 * config/aarch64/iterators.md (v_cmp_result): New.
27123 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
27124 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
27126 * config/arm/marvell-pj4.md: New file.
27127 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
27128 * config/arm/arm.md (generic_sched): Add marvell_pj4.
27129 (generic_vfp): Likewise.
27130 * config/arm/arm-cores.def: Add marvell-pj4.
27131 * config/arm/arm-tune.md: Regenerate.
27132 * config/arm/arm-tables.opt: Regenerate.
27133 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
27134 * doc/invoke.texi: Document marvell-pj4.
27136 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
27138 * config/aarch64/arm_neon.h: Map scalar types to standard types.
27140 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
27145 * var-tracking.c (negative_power_of_two_p): New.
27146 (global_get_addr_cache, local_get_addr_cache): New.
27147 (get_addr_from_global_cache, get_addr_from_local_cache): New.
27148 (vt_canonicalize_addr): Rewrite using the above. Adjust the
27150 (vt_stack_offset_p): Remove.
27151 (vt_canon_true_dep): Always canonicalize loc's address.
27152 (clobber_overlapping_mems): Make sure we have a MEM.
27153 (local_get_addr_clear_given_value): New.
27154 (val_reset): Clear local cached entries.
27155 (compute_bb_dataflow): Create and release the local cache.
27156 Disable duplicate MEMs clobbering.
27157 (emit_notes_in_bb): Clobber MEMs likewise.
27158 (vt_emit_notes): Create and release the local cache.
27159 (vt_initialize, vt_finalize): Create and release the global
27160 cache, respectively.
27161 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
27163 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
27165 PR libmudflap/53359
27166 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
27167 not found in the symtab.
27169 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
27172 PR rtl-optimization/55547
27173 PR rtl-optimization/53827
27176 * alias.c (offset_overlap_p): New, factored out of...
27177 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
27178 the conservative special case for symbolic constants. Don't
27179 adjust zero sizes on alignment.
27181 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
27183 PR rtl-optimization/52573
27184 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
27185 REG_UNUSED for the same register.
27187 2013-01-17 Richard Biener <rguenther@suse.de>
27188 Marek Polacek <polacek@redhat.com>
27190 PR rtl-optimization/55833
27191 * loop-unswitch.c (unswitch_loops): Move loop verification...
27192 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
27193 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
27194 Set it to true when we're removing a loop from hierarchy tree in
27195 an irreducible region.
27196 (fix_bb_placements): Adjust caller.
27197 (fix_loop_placements): Likewise.
27199 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
27201 * config/avr/builtins.def (DEF_BUILTIN): Factor out
27202 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
27203 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
27204 Remove ID. Adjust comments.
27205 * config/avr/avr-c.c (avr_builtin_name): Remove.
27206 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
27207 * config/avr/avr.c (avr_tolower): New static function.
27208 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
27209 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
27210 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
27213 2013-01-17 Jan Hubicka <jh@suse.cz>
27215 PR tree-optimization/55273
27216 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
27218 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
27221 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
27222 store through atomic_store<mode>_1.
27223 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
27225 2013-01-17 Martin Jambor <mjambor@suse.cz>
27227 PR tree-optimizations/55264
27228 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
27229 for virtual methods.
27230 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
27231 virtual methods before inlining is over.
27232 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
27234 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
27237 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
27239 PR rtl-optimization/56005
27240 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
27241 pending reads for prefetch.
27243 2013-01-16 Ian Bolton <ian.bolton@arm.com>
27245 * config/aarch64/aarch64.md
27246 (*cstoresi_neg_uxtw): New pattern.
27247 (*cmovsi_insn_uxtw): New pattern.
27248 (*<optab>si3_uxtw): New pattern.
27249 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
27250 (*<optab>si3_insn_uxtw): New pattern.
27251 (*bswapsi2_uxtw): New pattern.
27253 2013-01-16 Richard Biener <rguenther@suse.de>
27255 * tree-inline.c (tree_function_versioning): Remove set but
27256 never used variable.
27258 2013-01-16 Richard Biener <rguenther@suse.de>
27260 PR tree-optimization/55964
27261 * tree-flow.h (rename_variables_in_loop): Remove.
27262 (rename_variables_in_bb): Likewise.
27263 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
27264 (copy_loop_before): Adjust and delete update-ssa status.
27265 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
27266 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
27267 (rename_variables_in_loop): Remove.
27268 (slpeel_update_phis_for_duplicate_loop): Likewise.
27269 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
27270 use available cfg machinery instead of duplicating it.
27271 Update PHI nodes and perform poor-mans SSA update here.
27272 (slpeel_tree_peel_loop_to_edge): Adjust.
27274 2013-01-16 Richard Biener <rguenther@suse.de>
27276 PR tree-optimization/54767
27277 PR tree-optimization/53465
27278 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
27279 (vrp_visit_phi_node): For PHI arguments coming via backedges
27280 drop all symbolical range information.
27281 (execute_vrp): Compute backedges.
27283 2013-01-16 Richard Biener <rguenther@suse.de>
27285 * doc/install.texi: Update CLooG and ISL requirements to
27288 2013-01-16 Christian Bruel <christian.bruel@st.com>
27291 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
27292 (broken_move): Handle UNSPECV_SP_SWITCH_B.
27293 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
27295 2013-01-16 DJ Delorie <dj@redhat.com>
27297 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
27298 (UNSPECV_SP_SWITCH_E): New.
27299 (sp_switch_1): Change to an unspec.
27300 (sp_switch_2): Change to an unspec. Don't use post-inc when we
27303 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
27305 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
27306 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
27307 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
27308 (expand_mem_thread_fence): Ditto.
27309 (expand_mem_signal_fence): Ditto.
27310 (expand_atomic_load): Ditto.
27311 (expand_atomic_store): Ditto.
27313 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
27315 PR rtl-optimization/55547
27316 PR rtl-optimization/53827
27319 * alias.c (memrefs_conflict_p): Set sizes to negative after
27322 2013-01-15 Jakub Jelinek <jakub@redhat.com>
27325 * function.c (thread_prologue_and_epilogue_insns): Always
27326 add crtl->drap_reg to set_up_by_prologue.set, even if
27327 stack_realign_drap is false.
27329 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
27331 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
27332 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
27333 *call): Fix indention.
27335 2013-01-15 Tom de Vries <tom@codesourcery.com>
27338 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
27341 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
27343 PR rtl-optimization/55153
27344 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
27346 2013-01-15 Martin Jambor <mjambor@suse.cz>
27348 PR tree-optimization/55920
27349 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
27350 accesses as grp_to_be_debug_replaced.
27352 2013-01-15 Jakub Jelinek <jakub@redhat.com>
27354 PR tree-optimization/55920
27355 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
27356 there is non-useless type conversion needed from debug rhs to lhs,
27357 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
27359 2013-01-15 Joseph Myers <joseph@codesourcery.com>
27360 Mikael Pettersson <mikpe@it.uu.se>
27363 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
27365 (ASM_OUTPUT_CASE_LABEL): Remove.
27366 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
27367 * final.c (shorten_branches): Update alignment of labels before
27368 jump tables if CASE_VECTOR_SHORTEN_MODE.
27370 2013-01-15 Richard Biener <rguenther@suse.de>
27373 * system.h: Do not include gmp.h for building host tools.
27375 2013-01-15 Richard Biener <rguenther@suse.de>
27377 PR middle-end/55882
27378 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
27379 account for bitpos when computing alignment.
27381 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
27383 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
27384 (ix86_target_macros_internal): Likewise.
27386 * config/i386/i386.c (m_CORE2I7): Removed.
27387 (m_CORE_HASWELL): New macro.
27388 (m_CORE_ALL): Likewise.
27389 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
27390 (initial_ix86_arch_features): Likewise.
27391 (processor_target_table): Initializations for Core avx2.
27392 (cpu_names): New names "core-avx2".
27393 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
27394 PROCESSOR_CORE_HASWELL.
27395 (ix86_issue_rate): New case.
27396 (ia32_multipass_dfa_lookahead): Likewise.
27397 (ix86_sched_init_global): Likewise.
27399 * config/i386/i386.h (TARGET_HASWELL): New macro.
27400 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
27401 (processor_type): New PROCESSOR_HASWELL.
27403 2013-01-15 Jakub Jelinek <jakub@redhat.com>
27405 PR tree-optimization/55955
27406 * tree-vect-loop.c (vectorizable_reduction): Give up early on
27407 *SHIFT_EXPR and *ROTATE_EXPR codes.
27409 PR tree-optimization/48766
27410 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
27411 -ftrapv disable -fwrapv.
27413 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
27416 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
27417 etc. to 1 and not to __flash.
27418 Use LL suffix for __INT24_MAX__ with -mint8.
27419 Use ULL suffix for __UINT24_MAX__ with -mint8.
27421 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
27423 * config/avr/avr-arch.h
27424 (struct base_arch_s): Use typedef avr_arch_t instead.
27425 (struct arch_info_s): Use typedef avr_arch_info_t instead.
27426 (struct mcu_type_s): Use typedef avr_mcu_t instead.
27427 * config/avr/avr.c: Same.
27428 * config/avr/avr-devices.c: Same.
27429 * config/avr/driver-avr.c: Same.
27430 * config/avr/gen-avr-mmcu-texi.c: Same.
27431 * config/avr/avr-mcus.def: Adjust comment.
27433 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
27435 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
27436 * config/aarch64/iterators.md (VALLDI): New.
27438 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
27439 Andi Kleen <ak@linux.intel.com>
27442 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
27443 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
27446 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
27448 * config/avr/avr-stdint.h: Remove trailing blanks.
27449 * config/avr/avr-log.h: Same.
27450 * config/avr/avr-arch.h: Same.
27451 * config/avr/avr-devices.c: Same.
27452 * config/avr/avr-dimode.md: Same.
27453 * config/avr/predicates.md: Same.
27454 * config/avr/avr-c.c: Same. And fix typo.
27456 * config/avr/avr-protos.h: Same. And:
27457 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
27458 (init_cumulative_args): Rename to avr_init_cumulative_args.
27459 (expand_prologue): Rename to avr_expand_prologue.
27460 (expand_epilogue): Rename to avr_expand_epilogue.
27461 (adjust_insn_length): Rename to avr_adjust_insn_length.
27462 (notice_update_cc): Rename to avr_notice_update_cc.
27463 (final_prescan_insn): Rename to avr_final_prescan_insn.
27464 * config/avr/avr.c: Same.
27465 * config/avr/avr.h: Same.
27466 * config/avr/avr.md: Remove trailing blanks.
27467 (prologue): Use avr_expand_prologue.
27468 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
27470 2013-01-14 Richard Biener <rguenther@suse.de>
27472 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
27473 verify_location, collect_subblocks): New functions.
27474 (verify_gimple_in_cfg): Verify that locations only reference
27475 BLOCKs in the functions BLOCK tree.
27477 2013-01-14 Richard Biener <rguenther@suse.de>
27479 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
27481 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
27483 (insert_out_of_ssa_copy_on_edge): Likewise.
27484 (rewrite_close_phi_out_of_ssa): Likewise.
27485 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
27487 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
27488 propagated constants.
27489 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
27492 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
27494 * config/avr/avr-modes.def: Add GPL copyright notice.
27496 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
27498 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
27499 MEMMODEL_MASK to determine memory model.
27500 (atomic_store<mode>): Ditto from operands[2].
27501 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
27503 2013-01-13 Jakub Jelinek <jakub@redhat.com>
27506 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
27507 (fold_gimple_assign): Don't call unshare_expr here.
27508 (fold_ctor_reference): Call unshare_expr.
27510 2013-01-13 Terry Guo <terry.guo@arm.com>
27512 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
27513 * doc/fragments.texi: Document MULTILIB_REUSE.
27514 * gcc.c (multilib_reuse): New internal spec.
27515 (set_multilib_dir): Also search multilib from multilib_reuse.
27516 * genmultilib (tmpmultilib3): Refactor code.
27517 (tmpmultilib4): Ditto.
27518 (multilib_reuse): New multilib argument.
27520 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
27522 * Makefile.in: Update copyright.
27524 2013-01-12 Tom de Vries <tom@codesourcery.com>
27526 PR middle-end/55890
27527 * calls.c (expand_call): Check if arg_nr is valid.
27529 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
27531 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
27532 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
27533 documentation. Add missing '__' in front of
27534 __builtin_ia32_packssdw256.
27536 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
27539 * config/s390/s390.c (s390_preferred_reload_class): Do not return
27540 NO_REGS for larl operands.
27541 (s390_reload_larl_operand): Use s390_load_address instead of
27544 2013-01-11 Richard Biener <rguenther@suse.de>
27546 * tree-cfg.c (verify_node_sharing_1): Split out from ...
27547 (verify_node_sharing): ... here.
27548 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
27550 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
27552 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
27553 Substitute TREECHECKING.
27554 * configure: Regenerate.
27555 * Makefile.in (TREECHECKING): New.
27557 2013-01-11 Richard Guenther <rguenther@suse.de>
27559 PR tree-optimization/44061
27560 * tree-vrp.c (extract_range_basic): Compute zero as
27561 value-range for __builtin_constant_p of function parameters.
27563 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
27565 Update copyright years.
27567 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
27569 PR rtl-optimization/55672
27570 * lra-eliminations.c (mark_not_eliminable): Permit addition with
27571 const to be eliminable.
27573 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
27575 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
27576 * configure: Regenerate.
27578 2013-01-10 Richard Biener <rguenther@suse.de>
27580 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
27582 2013-01-10 Richard Biener <rguenther@suse.de>
27585 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
27586 locations for virtual PHI arguments.
27587 (rewrite_update_phi_arguments): Likewise.
27589 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
27591 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
27594 2013-01-10 Jakub Jelinek <jakub@redhat.com>
27596 PR tree-optimization/55921
27597 * tree-complex.c (expand_complex_asm): New function.
27598 (expand_complex_operations_1): Call it for GIMPLE_ASM.
27600 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
27603 * config/s390/s390.c (s390_symref_operand_p)
27604 (s390_loadrelative_operand_p): Merge the two functions.
27605 (s390_check_qrst_address, print_operand_address): Add parameters
27606 to s390_loadrelative_operand_p invokation.
27607 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
27608 (s390_reload_larl_operand, s390_secondary_reload): Use
27609 s390_loadrelative_operand_p instead of s390_symref_operand_p.
27610 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
27612 2013-01-09 Mike Stump <mikestump@comcast.net>
27614 * dse.c (record_store): Remove unnecessary assert.
27616 2013-01-09 Jan Hubicka <jh@suse.cz>
27618 PR tree-optimization/55569
27619 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
27620 * cfgloop.h (scale_loop_profile): Likewise.
27622 2013-01-09 Jan Hubicka <jh@suse.cz>
27625 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
27627 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
27629 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
27631 PR middle-end/55114
27632 * expr.h (maybe_emit_group_store): Declare.
27633 * expr.c (maybe_emit_group_store): New function.
27634 * builtins.c (expand_builtin_int_roundingfn): Call it.
27635 (expand_builtin_int_roundingfn_2): Likewise.
27637 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
27639 PR rtl-optimization/55829
27640 * lra-constraints.c (match_reload): Add code for absent output.
27641 (curr_insn_transform): Add code for reloads of matched inputs
27644 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
27646 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
27647 attribute of movddup insn to DF.
27648 (*vec_interleave_lowv2df): Ditto.
27649 (vec_dupv2df): Ditto.
27651 2013-01-09 Jan Hubicka <jh@suse.cz>
27653 PR tree-optimiation/55875
27654 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
27655 EVERY_ITERATION parameter.
27656 (number_of_iterations_exit): Check if exit is executed every iteration.
27657 (idx_infer_loop_bounds): Similarly here.
27658 (n_of_executions_at_most): Simplify
27659 to only test for cases where statement is dominated by the
27660 particular bound; handle correctly the "postdominance" test.
27661 (scev_probably_wraps_p): Use max loop iterations info
27662 as a global bound first.
27664 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
27665 Nick Clifton <nickc@redhat.com>
27667 * config/v850/v850.md (cbranchsf4): New pattern.
27668 (cstoresf4): New pattern.
27669 (cbranchdf4): New pattern.
27670 (cstoredf4): New pattern.
27671 (movsicc): Disallow floating point comparisons.
27672 (cmpsf_le_insn): Fix order of operators.
27673 (cmpsf_lt_insn): Likewise.
27674 (cmpsf_eq_insn): Likewise.
27675 (cmpdf_le_insn): Likewise.
27676 (cmpdf_lt_insn): Likewise.
27677 (cmpdf_eq_insn): Likewise.
27678 (cmpsf_ge_insn): Use LE comparison.
27679 (cmpdf_ge_insn): Likewise.
27680 (cmpsf_gt_insn): Use LT comparison.
27681 (cmpdf_gt_insn): Likewise.
27682 (cmpsf_ne_insn): Delete pattern.
27683 (cmpdf_ne_insn): Delete pattern.
27684 * config/v850/v850.c (v850_gen_float_compare): Use
27685 gen_cmpdf_eq_insn for NE comparison.
27686 (v850_float_z_comparison_operator)
27687 (v850_float_nz_comparison_operator): Move from here ...
27688 * config/v850/predicates.md: ... to here. Move GT and GE
27689 comparisons into v850_float_z_comparison_operator.
27690 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
27692 (v850_float_nz_comparison_operator): Likewise.
27694 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
27696 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
27697 with calls to gen_insvsi/gen_insvdi.
27699 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
27701 * config/i386/i386.c (initial_ix86_tune_features): Set up
27702 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
27704 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
27705 Jakub Jelinek <jakub@redhat.com>
27707 PR tree-optimization/48189
27708 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
27709 If nitercst is 0, don't predict the exit edge.
27711 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
27713 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
27714 in asm_fprintf with reg_names.
27715 (aarch64_print_operand_address): Likewise.
27716 (aarch64_return_addr): Likewise.
27717 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
27719 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
27721 * config/pa/pa.h (VAL_U6_BITS_P): Define.
27722 (INT_U6_BITS): Likewise.
27723 * config/pa/predicates.md (uint6_operand): New predicate.
27724 (shift5_operand, shift6_operand): Likewise.
27725 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
27727 (lshrdi3): Use shift6_operand.
27728 (shrpsi4, shrpdi4): New insn patterns.
27729 (extzv): Delete expander.
27730 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
27731 predicates in unamed zero extract patterns. Tighten common constraint.
27732 (extv): Delete expander.
27733 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
27734 predicates in unamed sign extract patterns. Tighten common constraint.
27735 (insv): Delete expander.
27736 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
27737 predicates in unamed insert patterns. Tighten common constraint.
27738 Change uint32_operand predicate to uint6_operand predicate in unamed
27739 DImode pattern to insert constant values of type 1...1xxxx.
27741 2013-01-04 Jan Hubicka <jh@suse.cz>
27743 PR tree-optimization/55823
27744 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
27747 2013-01-08 Jakub Jelinek <jakub@redhat.com>
27748 Uros Bizjak <ubizjak@gmail.com>
27750 PR rtl-optimization/55845
27751 * df-problems.c (can_move_insns_across): Stop scanning at
27752 volatile_insn_p source instruction or give up if
27753 across_from .. across_to range contains any volatile_insn_p
27756 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
27758 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
27759 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
27761 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
27762 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
27764 2013-01-08 Jakub Jelinek <jakub@redhat.com>
27767 * asan.c (asan_clear_shadow): New function.
27768 (asan_emit_stack_protection): Use it.
27770 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
27772 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
27773 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
27774 with tab instead of space.
27776 2013-01-08 Nick Clifton <nickc@redhat.com>
27778 * config/rl78/rl78.c (rl78_expand_prologue): Always select
27779 register bank 0 at the start of an interrupt handler.
27780 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
27783 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
27785 * config/aarch64/aarch64-simd.md
27786 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
27787 (aarch64_simd_bsl): Likewise.
27788 (aarch64_vcond_internal<mode>): Likewise.
27789 (vcond<mode><mode>): Likewise.
27790 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
27791 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
27793 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
27795 * config/aarch64/aarch64-builtins.c
27796 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
27798 2013-01-08 Martin Jambor <mjambor@suse.cz>
27801 * tree-sra.c (analyze_access_subtree): Return true also after
27802 potentially creating a debug-only replacement.
27804 2013-01-08 Jakub Jelinek <jakub@redhat.com>
27806 PR middle-end/55890
27807 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
27809 PR tree-optimization/54120
27810 * tree-vrp.c (range_fits_type_p): Don't allow
27811 src_precision < precision from signed vr to unsigned_p
27812 if vr->min or vr->max is negative.
27813 (simplify_float_conversion_using_ranges): Test can_float_p
27814 against CODE_FOR_nothing.
27816 2013-01-08 Jakub Jelinek <jakub@redhat.com>
27817 Richard Biener <rguenther@suse.de>
27819 PR middle-end/55851
27820 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
27821 types instead of just INTEGER_TYPE types.
27823 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
27825 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
27828 2013-01-07 Steve Ellcey <sellcey@mips.com>
27831 * config/mips/mips.opt: Change mad to mmad to match documentation.
27833 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
27836 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
27837 .progmemx.data now.
27839 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
27842 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
27843 (avr_addrspace_t): Add .section_name field.
27844 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
27846 (avr_addrspace): Same. Initialize .section_name. Remove last
27847 NULL entry. Put __memx into .progmemx.data.
27848 (progmem_section_prefix): Remove.
27849 (avr_asm_init_sections): No need to initialize progmem_section.
27850 (avr_asm_named_section): Use avr_addrspace[].section_name to get
27851 section name prefix.
27852 (avr_asm_select_section): Ditto. And use get_unnamed_section to
27853 retrieve the progmem section.
27854 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
27855 boundary to run over avr_addrspace[].
27856 (avr_register_target_pragmas): Ditto.
27858 2013-01-06 Jakub Jelinek <jakub@redhat.com>
27860 * varasm.c (output_constant_def_contents): For asan_protect_global
27861 protected strings, adjust DECL_ALIGN if needed, before testing for
27863 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
27864 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
27866 (output_object_block): For asan protected decls, emit asan padding
27867 after their contents.
27868 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
27869 (asan_finish_file): Test it here instead.
27871 2013-01-07 Nick Clifton <nickc@redhat.com>
27872 Matthias Klose <doko@debian.org>
27873 Doug Kwan <dougkwan@google.com>
27874 H.J. Lu <hongjiu.lu@intel.com>
27877 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
27879 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
27881 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
27883 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
27885 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
27887 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
27890 * doc/install.texi (Cross-Compiler-Specific Options): Document
27893 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
27895 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
27896 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
27897 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
27898 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
27899 vqmovun_high_s64): Fix source operand number and update copyright.
27901 2013-01-07 Richard Biener <rguenther@suse.de>
27903 PR middle-end/55890
27904 * gimple.h (gimple_call_builtin_p): New overload.
27905 * gimple.c (validate_call): New function.
27906 (gimple_call_builtin_p): Likewise.
27907 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
27908 Use gimple_call_builtin_p.
27909 (find_func_clobbers): Likewise.
27910 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
27911 (strlen_optimize_stmt): Likewise.
27913 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
27915 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
27916 (vld1q_dup_*): Likewise.
27917 (vld1_*): Likewise.
27918 (vld1q_*): Likewise.
27919 (vld1_lane_*): Likewise.
27920 (vld1q_lane_*): Likewise.
27922 2013-01-07 Richard Biener <rguenther@suse.de>
27924 * lto-streamer.h (LTO_minor_version): Bump to 2.
27926 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
27928 * config/aarch64/aarch64-protos.h
27929 (aarch64_const_double_zero_rtx_p): Rename to...
27930 (aarch64_float_const_zero_rtx_p): ...this.
27931 (aarch64_float_const_representable_p): New.
27932 (aarch64_output_simd_mov_immediate): Likewise.
27933 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
27934 move immediate case.
27935 * config/aarch64/aarch64.c
27936 (aarch64_const_double_zero_rtx_p): Rename to...
27937 (aarch64_float_const_zero_rtx_p): ...this.
27938 (aarch64_print_operand): Allow printing of new constants.
27939 (aarch64_valid_floating_const): New.
27940 (aarch64_legitimate_constant_p): Check for valid floating-point
27942 (aarch64_simd_valid_immediate): Likewise.
27943 (aarch64_vect_float_const_representable_p): New.
27944 (aarch64_float_const_representable_p): Likewise.
27945 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
27946 (aarch64_output_simd_mov_immediate): New.
27947 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
27948 (*movdf_aarch64): Likewise.
27949 * config/aarch64/constraints.md (Ufc): New.
27950 (Y): call aarch64_float_const_zero_rtx.
27951 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
27953 2013-01-07 Richard Biener <rguenther@suse.de>
27955 PR tree-optimization/55888
27956 PR tree-optimization/55862
27957 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
27958 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
27959 not if it is contained therein.
27961 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
27963 * config/avr/t-avr: Typo.
27965 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
27968 * config/avr/t-avr: Don't automatically rebuild
27969 $(srcdir)/config/avr/t-multilib
27970 $(srcdir)/config/avr/avr-tables.opt
27971 $(srcdir)/doc/avr-mmcu.texi
27972 (avr-mcus): New phony target to build them on request.
27973 (s-avr-mlib, s-avr-mmcu-texi): Remove.
27974 * avr/avr-mcus.def: Adjust comments.
27976 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
27978 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
27980 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
27982 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
27984 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
27986 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
27988 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
27990 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
27991 to generate profiling.
27992 * config/rs6000/aix64.h (LIB_SPEC): Same.
27994 2013-01-04 Andrew Pinski <apinski@cavium.com>
27996 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
27998 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
28000 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
28002 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
28004 (ix86_expand_move): Ditto.
28005 (ix86_zero_extend_to_Pmode): Ditto.
28006 (ix86_expand_call): Ditto.
28007 (ix86_expand_special_args_builtin): Ditto.
28008 (ix86_expand_builtin): Ditto.
28010 2013-01-04 Richard Biener <rguenther@suse.de>
28012 PR tree-optimization/55862
28013 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
28014 translating them through PHI nodes.
28016 2013-01-04 Martin Jambor <mjambor@suse.cz>
28018 PR tree-optimization/55755
28019 * tree-sra.c (sra_modify_assign): Do not check that an access has no
28020 children when trying to avoid producing a VIEW_CONVERT_EXPR.
28022 2013-01-04 Marek Polacek <polacek@redhat.com>
28024 PR middle-end/55859
28025 * opts.c (default_options_optimization): Clarify error message.
28027 2013-01-04 Richard Biener <rguenther@suse.de>
28029 PR middle-end/55863
28030 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
28033 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
28036 * config/pa/pa.md (movsi): Revert previous change.
28037 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
28040 2013-01-03 Richard Henderson <rth@redhat.com>
28042 * config/i386/i386.c (ix86_expand_move): Always assign to op1
28043 after eliminating TLS symbols.
28045 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
28048 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
28049 * graphite-poly.c (debug_gmp_value): Likewise.
28051 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
28054 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
28055 selected code model, define __code_mode_small__, __code_model_medium__,
28056 __code_model_large__, __code_model_32__ or __code_model_kernel__.
28057 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
28058 xchg temporary register with %k. Declare temporary register as
28060 [__x86_64__]: For medium and large code models, preserve %rbx register.
28062 2013-01-03 Richard Biener <rguenther@suse.de>
28064 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
28065 (dump_subscript): Adjust.
28066 (finalize_ddr_dependent): Do not dump redundant info.
28067 (analyze_siv_subscript): Adjust.
28068 (subscript_dependence_tester): Likewise.
28069 (compute_affine_dependence): Likewise.
28071 2013-01-03 Richard Biener <rguenther@suse.de>
28074 2013-01-03 Richard Biener <rguenther@suse.de>
28076 PR tree-optimization/55857
28077 * tree-vect-stmts.c (vectorizable_load): Do not setup
28078 re-alignment for invariant loads.
28080 2013-01-02 Richard Biener <rguenther@suse.de>
28082 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
28083 invariant load do not generate a vector load from the scalar location.
28085 2013-01-03 Richard Biener <rguenther@suse.de>
28087 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
28088 for not vectorizing.
28089 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
28090 not build INDIRECT_REFs, call get_name once only.
28091 (vect_create_data_ref_ptr): Likewise. Dump base object kind
28092 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
28094 2013-01-03 Richard Biener <rguenther@suse.de>
28096 PR tree-optimization/55857
28097 * tree-vect-stmts.c (vectorizable_load): Do not setup
28098 re-alignment for invariant loads.
28100 2013-01-03 Richard Biener <rguenther@suse.de>
28103 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
28104 prefer a built-in decl.
28106 2013-01-03 Jakub Jelinek <jakub@redhat.com>
28108 * gcc.c (process_command): Update copyright notice dates.
28109 * gcov.c (print_version): Likewise.
28110 * gcov-dump.c (print_version): Likewise.
28112 PR rtl-optimization/55838
28113 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
28114 iv0.step, iv1.step and step.
28116 2013-01-03 Jakub Jelinek <jakub@redhat.com>
28117 Marc Glisse <marc.glisse@inria.fr>
28119 PR tree-optimization/55832
28120 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
28121 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
28122 integer_{one,zero}_node.
28124 2013-01-03 Jakub Jelinek <jakub@redhat.com>
28127 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
28128 * var-tracking.c (reverse_op): Don't add reverse ops to
28129 VALUEs that have already
28130 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
28132 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
28134 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
28136 2013-01-02 Teresa Johnson <tejohnson@google.com>
28138 * dumpfile.c (dump_loc): Print filename with location.
28139 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
28140 new location_t parameter to emit complete unroll message with
28141 new dump framework.
28142 (canonicalize_loop_induction_variables): Compute loops location
28143 and pass to try_unroll_loop_completely.
28144 * loop-unroll.c (report_unroll_peel): New function.
28145 (peel_loops_completely): Use new dump format with location
28146 for main dumpfile message, and invoke report_unroll_peel on success.
28147 (decide_unrolling_and_peeling): Ditto.
28148 (decide_peel_once_rolling): Remove old dumpfile message subsumed
28149 by report_unroll_peel.
28150 (decide_peel_completely): Ditto.
28151 (decide_unroll_constant_iterations): Ditto.
28152 (decide_unroll_runtime_iterations): Ditto.
28153 (decide_peel_simple): Ditto.
28154 (decide_unroll_stupid): Ditto.
28155 * cfgloop.c (get_loop_location): New function.
28156 * cfgloop.h (get_loop_location): Declare.
28158 2013-01-02 Sriraman Tallam <tmsriram@google.com>
28160 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
28163 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
28165 PR middle-end/55198
28166 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
28167 BLKmode objects when EXPAND_MEMORY is specified.
28169 2013-01-02 Sriraman Tallam <tmsriram@google.com>
28171 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
28173 (fold_builtin_cpu): Do not share cpu model decls across statements.
28175 2013-01-02 Jason Merrill <jason@redhat.com>
28178 * tree.c (build_array_type_1): Revert earlier change.
28180 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
28182 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
28184 * config/aarch64/aarch64-tune.md: Re-generate.
28186 2013-01-02 Richard Biener <rguenther@suse.de>
28188 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
28189 invariant load do not generate a vector load from the scalar location.
28191 2013-01-02 Richard Biener <rguenther@suse.de>
28194 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
28195 * configure: Regenerate.
28197 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
28199 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
28200 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
28201 (expand_builtin_int_roundingfn_2): Keep the original target around
28202 for the fallback case.
28204 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
28206 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
28207 to be clear for sign changes.
28209 2013-01-01 Jan Hubicka <jh@suse.cz>
28211 * ipa-inline-analysis.c: Fix formatting.
28213 2013-01-01 Jakub Jelinek <jakub@redhat.com>
28215 PR tree-optimization/55831
28216 * tree-vect-loop.c (get_initial_def_for_induction): Use
28217 gsi_after_labels instead of gsi_start_bb.
28219 Copyright (C) 2013 Free Software Foundation, Inc.
28221 Copying and distribution of this file, with or without modification,
28222 are permitted in any medium without royalty provided the copyright
28223 notice and this notice are preserved.