1 2013-10-27 Tom de Vries <tom@codesourcery.com>
3 * cfgexpand.c (gimple_expand_cfg): Remove test for parm_birth_insn.
4 Don't commit insertions after NOTE_INSN_FUNCTION_BEG.
6 2013-10-27 Oleg Endo <olegendo@gcc.gnu.org>
8 * config/sh/sh.c (MSW, LSW): Move and rename macros to...
9 * config/sh/sh.h (SH_REG_MSW_OFFSET, SH_REG_LSW_OFFSET): ... here.
10 (TARGET_BIG_ENDIAN): New macro.
11 * config/sh/sh.md: Use it instead of !TARGET_LITTLE_ENDIAN.
12 Use SH_REG_MSW_OFFSET and SH_REG_LSW_OFFSET.
13 * config/sh/sh.c: Likewise.
14 * config/sh/sh.h: Likewise.
16 2013-10-27 Hans-Peter Nilsson <hp@axis.com>
18 * config/cris/cris.c (cris_emit_trap_for_misalignment): Replace the
19 removed PRED_MUDFLAP with PRED_NORETURN. Correct file-path in comment.
21 2013-10-26 Oleg Endo <olegendo@gcc.gnu.org>
23 * config/sh/sh.md (movmemsi): Remove empty constraints and predicates.
25 (cmpstr_t, cmpstrsi): Fix formatting.
27 2013-10-26 Oleg Endo <olegendo@gcc.gnu.org>
30 * config/sh/predicates.md (general_movdst_operand): Allow reg+reg
31 addressing, do not use general_operand for memory operands.
33 2013-10-26 Vladimir Makarov <vmakarov@redhat.com>
36 2013-10-25 Vladimir Makarov <vmakarov@redhat.com>
37 * lra-spills.c (lra_final_code_change): Remove useless move insns.
39 2013-10-26 Jeff Law <law@redhat.com>
41 * predict.c (PRED_MUDFLAP): Remove.
42 * targhooks.c (build_va_arg_indirect_ref): Remove mudflap support.
44 * Makefile.in (C_COMMON_OBJS): Remove tree-mudflap.
45 (OBJS): Remove tree-nomudflap.o
46 (GTFILES): Remove tree-mudflap.c
47 * builtins.c (expand_builtin_alloc): Remove mudflap support.
48 * gcc.c (MFWRAP_SPEC, MFLIB_SPEC): Likewise.
49 (mfwrap_spec, mflib_spec): Likewise.
50 (cpp_unique_options, cc1_options, static_specs): Likewise.
51 * gimplify (gimplify_vla_decl, build_va_arg_indirect_ref): Likewise.
52 * passes.def: Likewise.
53 * toplev.c (compile_file, process_options): Likewise.
54 * tree-inline.c (copy_tree_r): Likewise.
55 * tree-pass.,h (make_pass_mudflap_1, make_pass_mudflap_2): Likewise.
56 * varasm.c (make_decl_rtl, make_decl_rtl_for_debug): Likewise.
57 (build_constant_desc, output_constant_def_contents): Likewise.
58 (categorize_decl_for_section): Likewise.
59 * tree-mudflap.c: Removed.
60 * tree-mudflap.h: Removed.
61 * tree-nomudflap.c: Removed.
62 * bfin/uclinux.h (MFWRAP_SPEC): Remove.
63 * moxie/uclinux.h (MFWRAP_SPEC): Likewise.
64 * rs6000/aix.h (MFWRAP_SPEC, MFLIB_SPEC): Likewise.
65 * config/sol2.h (MFLIB_SPEC): Likewise.
66 * doc/install.texi: Remove mudflap references.
67 * doc/passes.texi: Similarly.
68 * doc/sourcebuild.texi: Similarly.
69 * doc/invoke.texi: Remove mudlfap related options.
71 2013-10-25 Vladimir Makarov <vmakarov@redhat.com>
73 PR rtl-optimization/58759
74 * lra-constraints.c (lra_constraints): Remove wrong condition to
75 remove insn setting up an equivalent pseudo.
77 2013-10-25 Vladimir Makarov <vmakarov@redhat.com>
79 * config/rs6000/rs6000-protos.h
80 (rs6000_secondary_memory_needed_mode): New prototype.
81 * config/rs6000/rs6000.c: Include ira.h.
82 (TARGET_LRA_P): Redefine.
83 (rs6000_legitimate_offset_address_p): Call
84 legitimate_constant_pool_address_p in strict mode for LRA.
85 (rs6000_legitimate_address_p): Ditto.
86 (legitimate_lo_sum_address_p): Add code for LRA.
88 (rs6000_emit_move): Add LRA version of code to generate load/store
90 (rs6000_secondary_memory_needed_mode): New.
91 (rs6000_alloc_sdmode_stack_slot): Do nothing for LRA.
92 (rs6000_secondary_reload_class): Return NO_REGS for LRA for
93 constants, memory, and FP registers.
95 * config/rs6000/rs6000.h (SECONDARY_MEMORY_NEEDED_MODE): New
97 * config/rs6000/rs6000.opt (mlra): New option.
98 * lra-spills.c (lra_final_code_change): Remove useless move insns.
100 2013-10-25 Yufeng Zhang <yufeng.zhang@arm.com>
102 * tree-ssa-math-opts.c (convert_plusminus_to_widen): Call
103 has_single_use () and not do the conversion if has_single_use ()
104 returns false for the multiplication result.
106 2013-10-25 David Malcolm <dmalcolm@redhat.com>
108 * tree.h (EXCEPTIONAL_CLASS_P): Rename parameter from "CODE"
109 to "NODE", since this works on a "tree", not an
111 (CONSTANT_CLASS_P): Likewise.
114 (INDIRECT_REF_P): Likewise.
115 (REFERENCE_CLASS_P): Likewise.
116 (COMPARISON_CLASS_P): Likewise.
117 (UNARY_CLASS_P): Likewise.
118 (BINARY_CLASS_P): Likewise.
119 (STATEMENT_CLASS_P): Likewise.
120 (VL_EXP_CLASS_P): Likewise.
121 (EXPRESSION_CLASS_P): Likewise.
122 (IS_TYPE_OR_DECL_P): Likewise.
124 2013-10-25 Marc Glisse <marc.glisse@inria.fr>
126 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Look for an
127 ADDR_EXPR in the defining statement.
129 2013-10-25 Richard Biener <rguenther@suse.de>
131 PR tree-optimization/58626
132 * tree-loop-distribution.c (enum rdg_dep_type): Remove
133 anti_dd, output_dd and input_dd.
134 (struct rdg_edge): Remove level and relation members.
135 (RDGE_LEVEL, RDGE_RELATION): Remove.
137 (create_rdg_edge_for_ddr): Remove.
138 (create_rdg_edges_for_scalar): Adjust.
139 (create_edge_for_control_dependence): Likewise.
140 (create_rdg_edges): Split into ...
141 (create_rdg_flow_edges): ... this
142 (create_rdg_cd_edges): ... and this.
144 (build_rdg): Likewise, do not compute data dependences or
146 (pg_add_dependence_edges): New function.
148 (distribute_loop): First apply all non-dependence based
149 partition mergings. Then compute dependences between partitions
150 and merge and order partitions according to them.
152 2013-10-25 Eric Botcazou <ebotcazou@adacore.com>
154 PR rtl-optimization/58831
155 * alias.c (init_alias_analysis): At the beginning of each iteration, set
156 the reg_seen[N] bit if static_reg_base_value[N] is non-null.
158 2013-10-25 Eric Botcazou <ebotcazou@adacore.com>
160 * recog.c (search_ofs): New static variable moved from...
161 (peep2_find_free_register): ...here.
162 (peephole2_optimize): Initialize it.
164 2013-10-25 Tobias Burnus <burnus@net-b.de>
166 * doc/invoke.texi (fopenmp): Change supported OpenMP version to 4.0.
168 2013-10-25 Uros Bizjak <ubizjak@gmail.com>
170 * config/i386/i386.h (TARGET_MPX): New define.
171 (TARGET_MPX_P): Ditto.
173 2013-10-24 Ilya Enkovich <ilya.enkovich@intel.com>
175 * config/i386/constraints.md (B): New.
178 * config/i386/i386-c.c (ix86_target_macros_internal): Add __MPX__.
179 * config/i386/i386-modes.def (BND32): New.
181 * config/i386/i386-protos.h (ix86_bnd_prefixed_insn_p): New.
182 * config/i386/i386.c (isa_opts): Add mmpx.
183 (regclass_map): Add bound registers.
184 (dbx_register_map): Likewise.
185 (dbx64_register_map): Likewise.
186 (svr4_dbx_register_map): Likewise.
188 (ix86_option_override_internal): Support MPX ISA.
189 (ix86_conditional_register_usage): Support bound registers.
190 (print_reg): Likewise.
191 (ix86_code_end): Add MPX bnd prefix.
192 (output_set_got): Likewise.
193 (ix86_output_call_insn): Likewise.
194 (ix86_print_operand): Add '!' (MPX bnd) print prefix support.
195 (ix86_print_operand_punct_valid_p): Likewise.
196 (ix86_print_operand_address): Support UNSPEC_BNDMK_ADDR and
198 (ix86_class_likely_spilled_p): Add bound regs support.
199 (ix86_hard_regno_mode_ok): Likewise.
200 (x86_order_regs_for_local_alloc): Likewise.
201 (ix86_bnd_prefixed_insn_p): New.
202 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Fix to new value.
203 (FIXED_REGISTERS): Add bound registers.
204 (CALL_USED_REGISTERS): Likewise.
205 (REG_ALLOC_ORDER): Likewise.
206 (HARD_REGNO_NREGS): Likewise.
208 (VALID_BND_REG_MODE): New.
209 (FIRST_BND_REG): New.
211 (reg_class): Add BND_REGS.
212 (REG_CLASS_NAMES): Likewise.
213 (REG_CLASS_CONTENTS): Likewise.
215 (ANY_BND_REG_P): New.
217 (HI_REGISTER_NAMES): Add bound registers.
218 * config/i386/i386.md (UNSPEC_BNDMK): New.
219 (UNSPEC_BNDMK_ADDR): New.
220 (UNSPEC_BNDSTX): New.
221 (UNSPEC_BNDLDX): New.
222 (UNSPEC_BNDLDX_ADDR): New.
226 (UNSPEC_MPX_FENCE): New.
229 (type): Add mpxmov, mpxmk, mpxchk, mpxld, mpxst.
230 (length_immediate): Likewise.
231 (prefix_0f): Likewise.
233 (prefix_rep): Check for bnd prefix.
235 (length): Use length_nobnd if specified.
240 (*jcc_1): Add bnd prefix and rename length attr to length_nobnd.
243 (simple_return_internal): Likewise.
244 (simple_return_pop_internal): Likewise.
245 (*indirect_jump): Add MPX bnd prefix.
246 (*tablejump_1): Likewise.
247 (simple_return_internal_long): Likewise.
248 (simple_return_indirect_internal): Likewise.
252 (*mov<mode>_internal_mpx): New.
253 (<mode>_<bndcheck>): New.
254 (*<mode>_<bndcheck>): New.
259 * config/i386/predicates.md (lea_address_operand): Rename to...
260 (address_no_seg_operand): ... this.
261 (address_mpx_no_base_operand): New.
262 (address_mpx_no_index_operand): New.
263 (bnd_mem_operator): New.
264 * config/i386/i386.opt (mmpx): New.
265 * doc/invoke.texi: Add documentation for the flags -mmpx, -mno-mpx.
266 * doc/rtl.texi Add documentation for BND32mode and BND64mode.
268 2013-10-24 Ilya Enkovich <ilya.enkovich@intel.com>
270 * mode-classes.def (MODE_POINTER_BOUNDS): New.
271 * tree.def (POINTER_BOUNDS_TYPE): New.
272 * genmodes.c (complete_mode): Support MODE_POINTER_BOUNDS.
273 (POINTER_BOUNDS_MODE): New.
274 (make_pointer_bounds_mode): New.
275 * machmode.h (POINTER_BOUNDS_MODE_P): New.
276 * stor-layout.c (int_mode_for_mode): Support MODE_POINTER_BOUNDS.
277 (layout_type): Support POINTER_BOUNDS_TYPE.
278 * tree-pretty-print.c (dump_generic_node): Support POINTER_BOUNDS_TYPE.
279 * tree.c (build_int_cst_wide): Support POINTER_BOUNDS_TYPE.
280 (type_contains_placeholder_1): Likewise.
281 * tree.h (POINTER_BOUNDS_TYPE_P): New.
282 * varasm.c (output_constant): Support POINTER_BOUNDS_TYPE.
283 * doc/rtl.texi (MODE_POINTER_BOUNDS): New.
285 2013-10-24 Igor Shevlyakov <igor.shevlyakov@gmail.com>
287 * expr.c (expand_expr_real_1): Use mode of memory reference rather than
288 mode of address computation when calling memory_address_addr_space.
290 2013-08-24 Richard Henderson <rth@twiddle.net>
293 * optabs.c (maybe_emit_atomic_exchange): Use create_input_operand
294 instead of create_convert_operand_to.
295 (maybe_emit_sync_lock_test_and_set): Likewise.
296 (expand_atomic_compare_and_swap): Likewise.
297 (maybe_emit_compare_and_swap_exchange_loop): Don't convert_modes.
299 2013-08-24 Sriraman Tallam <tmsriram@google.com>
301 * cgraph.c (cgraph_fnver_htab): Move GTY((...)) to be before htab_t.
302 Change param_is to use the struct and not the pointer to the struct.
304 2013-10-24 Andrew MacLeod <amacleod@redhat.com>
306 * builtins.c (dummy_object, gimplify_va_arg_expr): Move to gimplify.c.
307 * gimplify.c (build_va_arg_indirect_ref, std_gimplify_va_arg_expr):
309 (dummy_object, gimplify_va_arg_expr): Relocate from builtins.c.
310 * targhooks.c (build_va_arg_indirect_ref, std_gimplify_va_arg_expr):
311 Relocate from gimplify.c.
312 * targhooks.h: Add 2 prototypes.
313 * tree.h: Delete 2 prototypes.
315 2013-10-24 Igor Shevlyakov <igor.shevlyakov@gmail.com>
317 * tree-ssa-loop-ivopts.c (multiplier_allowed_in_address_p ): Check both
318 [reg+mult*reg] and [mult*reg] to determine if multiplier is allowed.
320 2013-10-24 Cong Hou <congh@google.com>
322 * convert.c (convert_to_real): Guard those unsafe math function
323 convertions with flag_unsafe_math_optimizations. Handle sqrt()
326 2013-10-24 Markus Trippelsdorf <markus@trippelsdorf.de>
329 * cgraph.c (cgraph_create_edge_1): Add indirect_unknown_callee
331 (cgraph_create_edge): Use the new argument.
332 (cgraph_create_indirect_edge): Likewise.
334 2013-10-24 Joern Rennecke <joern.rennecke@embecosm.com>
336 * gcc/config/arc/arc.c (arc_ccfsm_post_advance): Also handle
338 (arc_ifcvt) <case 1 and 2>: Check that arc_ccfsm_post_advance
339 changes statep->state.
341 2013-10-24 Tobias Burnus <burnus@net-b.de>
344 * tree-cfg.c (replace_loop_annotate): New function.
345 (execute_build_cfg): Call it.
346 * gimplify.c (gimple_boolify, gimplify_expr): Handle ANNOTATE_EXPR.
347 * internal-fn.c (expand_ANNOTATE): New function.
348 * internal-fn.def (ANNOTATE): Define as new internal function.
349 * tree-core.h (tree_node_kind): Add annot_expr_ivdep_kind.
350 * tree-pretty-print.c (dump_generic_node): Handle ANNOTATE_EXPR.
351 * tree.def (ANNOTATE_EXPR): New DEFTREECODE.
352 * doc/extend.texi (Pragmas): Document #pragma ivdep.
353 * doc/generic.texi (Expressions): Document ANNOTATE_EXPR.
355 2013-10-17 Ian Bolton <ian.bolton@arm.com>
356 Marcus Shawcroft <marcus.shawcroft@arm.com>
358 * config/aarch64/aarch64.c (aarch64_preferred_reload_class):
359 Special case reload SP+C into none GENERAL_REGS.
361 2013-10-24 Michael Matz <matz@suse.de>
363 * gengtype.c (is_file_equal): Check that files will be same length.
365 2013-10-25 Christian Bruel <christian.bruel@st.com>
367 * config.gcc (sh-*): Add sh-mem.o to extra_obj.
368 * gcc/config/sh/t-sh (sh-mem.o): New rule.
369 * gcc/config/sh/sh-mem.cc (expand_block_move): Moved here.
370 (sh_expand_cmpstr): New function.
371 * gcc/config/sh/sh.c (force_into, expand_block_move): Move to sh-mem.c.
372 * gcc/config/sh/sh-protos.h (sh_expand_cmpstr): Declare.
373 * gcc/config/sh/sh.md (cmpstrsi, cmpstr_t): New patterns.
376 2013-10-24 Jan-Benedict Glaw <jbglaw@lug-owl.de>
378 * configure.ac (ZW_PROG_COMPILER_DEPENDENCIES): Use CXX instead of CC.
379 * Makefile.in (CXXDEPMODE): Assign and change users.
381 * configure: Regenerate.
383 2013-10-23 David Malcolm <dmalcolm@redhat.com>
385 * gengtype-parse.c (require_without_advance): New.
386 (type): For GTY-marked types that are not GTY((user)), parse any
387 base classes, requiring them to be single-inheritance, and not
388 be templates. For non-GTY-marked types and GTY((user)),
389 continue to skip over any C++ inheritance specification.
390 * gengtype-state.c (state_writer::write_state_struct_type):
391 Write base class of type (if any).
392 (read_state_struct_type): Read base class of type (if any).
393 * gengtype.c (new_structure): Add a "base_class" parameter.
394 (create_optional_field_): Update for new parameter to new_structure.
395 (adjust_field_rtx_def): Likewise.
396 (adjust_field_tree_exp): Likewise.
397 * gengtype.h (struct type): Add "base_class" field to the s
399 (new_structure): Add "base" parameter.
401 2013-10-23 Sriraman Tallam <tmsriram@google.com>
404 * config/i386/i386.c (ix86_option_override_internal):
405 Change TARGET_SSE2 to TARGET_SSE2_P (opts->...)
406 (ix86_valid_target_attribute_tree):
407 Change TARGET_64BIT to TARGET_64BIT_P (opts->...)
408 Change TARGET_SSE to TARGET_SSE_P (opts->...)
410 2013-10-23 Andrew MacLeod <amacleod@redhat.com>
412 * tree-ssa-loop.h: Remove include files.
413 * gengtype.c (open_base_files): Adjust include list for gtype-desc.c.
414 * cfgloopmanip.c: Move required includes from tree-ssa-loop.h.
415 * graphite-clast-to-gimple.c: Likewise.
416 * graphite-scop-detection.c: Likewise.
417 * graphite-sese-to-poly.c: Likewise.
418 * ipa-inline-analysis.c: Likewise.
419 * ipa-pure-const.c: Likewise.
420 * loop-init.c: Likewise.
421 * passes.c: Likewise.
422 * predict.c: Likewise.
423 * tree-cfg.c: Likewise.
424 * tree-cfgcleanup.c: Likewise.
425 * tree-chrec.c: Likewise.
426 * tree-data-ref.c: Likewise.
427 * tree-loop-distribution.c: Likewise.
428 * tree-parloops.c: Likewise.
429 * tree-predcom.c: Likewise.
430 * tree-scalar-evolution.c: Likewise.
431 * tree-ssa-address.c: Likewise.
432 * tree-ssa.c: Likewise.
433 * tree-ssa-dce.c: Likewise.
434 * tree-ssa-loop.c: Likewise.
435 * tree-ssa-loop-im.c: Likewise.
436 * tree-ssa-loop-ivcanon.c: Likewise.
437 * tree-ssa-loop-ivopts.c: Likewise.
438 * tree-ssa-loop-manip.c: Likewise.
439 * tree-ssa-loop-niter.c: Likewise.
440 * tree-ssa-loop-prefetch.c: Likewise.
441 * tree-ssa-loop-unswitch.c: Likewise.
442 * tree-ssa-reassoc.c: Likewise.
443 * tree-vect-data-refs.c: Likewise.
444 * tree-vect-loop.c: Likewise.
445 * tree-vect-loop-manip.c: Likewise.
446 * tree-vectorizer.c: Likewise.
447 * tree-vect-stmts.c: Likewise.
448 * tree-vrp.c: Likewise.
450 2013-10-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
452 * config/rs6000/altivec.md (mulv8hi3): Adjust for little endian.
454 2013-10-23 Jakub Jelinek <jakub@redhat.com>
456 PR tree-optimization/58775
457 PR tree-optimization/58791
458 * tree-ssa-reassoc.c (reassoc_stmt_dominates_stmt_p): New function.
459 (insert_stmt_after): Rewritten, don't move the stmt, but really
461 (get_stmt_uid_with_default): Remove.
462 (build_and_add_sum): Use insert_stmt_after and
463 reassoc_stmt_dominates_stmt_p. Fix up uid if bb contains only labels.
464 (update_range_test): Set uid on stmts added by
465 force_gimple_operand_gsi. Don't immediately modify statements
466 in inter-bb optimization, just update oe->op values.
467 (optimize_range_tests): Return bool whether any changed have been made.
468 (update_ops): New function.
469 (struct inter_bb_range_test_entry): New type.
470 (maybe_optimize_range_tests): Perform statement changes here.
471 (not_dominated_by, appears_later_in_bb, get_def_stmt,
472 ensure_ops_are_available): Remove.
473 (find_insert_point): Rewritten.
474 (rewrite_expr_tree): Remove MOVED argument, add CHANGED argument,
475 return LHS of the (new resp. old) stmt. Don't call
476 ensure_ops_are_available, don't reuse SSA_NAMEs, recurse first
477 instead of last, move new stmt at the right place.
478 (linearize_expr, repropagate_negates): Don't reuse SSA_NAMEs.
479 (negate_value): Likewise. Set uids.
480 (break_up_subtract_bb): Initialize uids.
481 (reassociate_bb): Adjust rewrite_expr_tree caller.
482 (do_reassoc): Don't call renumber_gimple_stmt_uids.
484 2013-10-23 David Edelsohn <dje.gcc@gmail.com>
487 * config/rs6000/rs6000.md (mulsi3_internal1 and splitter): Add
488 TARGET_32BIT final condition.
489 (mulsi3_internal2 and splitter): Same.
491 2013-10-23 Jeff Law <law@redhat.com>
493 * tree-ssa-threadedge.c (thread_across_edge): Do not allow threading
494 through joiner blocks with abnormal outgoing edges.
496 * tree-ssa-threadupdate.c (thread_block_1): Renamed from thread_block.
497 Add parameter JOINERS, to allow/disallow threading through joiner
499 (thread_block): New. Call thread_block_1.
500 (mark_threaded_blocks): Remove code to filter out certain cases
501 of threading through joiner blocks.
502 (thread_through_all_blocks): Document how we can have a dangling
503 edge AUX field and clear it.
505 2013-10-23 Ian Lance Taylor <iant@google.com>
507 * doc/invoke.texi (Option Summary): Remove -fno-default-inline.
508 (C++ Dialect Options): Likewise.
509 (Optimize Options): Likewise.
511 2013-10-23 Tom de Vries <tom@codesourcery.com>
513 PR tree-optimization/58805
514 * tree-ssa-tail-merge.c (stmt_local_def): Add gimple_vdef check.
516 2013-10-23 Jakub Jelinek <jakub@redhat.com>
518 * tree-vect-patterns.c (vect_recog_divmod_pattern): Optimize
519 sequence based on get_range_info returned range.
521 2013-10-23 Andrew MacLeod <amacleod@redhat.com>
523 * tree-ssa.h: Remove all #include's
524 * gengtype.c (open_base_files): Adjust include list for gtype-desc.c.
525 * alias.c: Move required includes from tree-ssa.h.
527 * builtins.c: Likewise.
529 * cfgexpand.c: Likewise.
530 * cfghooks.c: Likewise.
531 * cfgloop.c: Likewise.
532 * cfgloopmanip.c: Likewise.
533 * cgraph.c: Likewise.
534 * cgraphbuild.c: Likewise.
535 * cgraphclones.c: Likewise.
536 * cgraphunit.c: Likewise.
538 * except.c: Likewise.
541 * fold-const.c: Likewise.
542 * ggc-page.c: Likewise.
543 * gimple-builder.c: Likewise.
544 * gimple-fold.c: Likewise.
545 * gimple-iterator.c: Likewise.
546 * gimple-low.c: Likewise.
547 * gimple-pretty-print.c: Likewise.
548 * gimple-ssa-strength-reduction.c: Likewise.
549 * gimple-streamer-in.c: Likewise.
550 * gimple-streamer-out.c: Likewise.
551 * gimplify.c: Likewise.
552 * graphite-blocking.c: Likewise.
553 * graphite-clast-to-gimple.c: Likewise.
554 * graphite-dependences.c: Likewise.
555 * graphite-interchange.c: Likewise.
556 * graphite-optimize-isl.c: Likewise.
557 * graphite-poly.c: Likewise.
558 * graphite-scop-detection.c: Likewise.
559 * graphite-sese-to-poly.c: Likewise.
560 * graphite.c: Likewise.
561 * ipa-cp.c: Likewise.
562 * ipa-inline-analysis.c: Likewise.
563 * ipa-inline-transform.c: Likewise.
564 * ipa-inline.c: Likewise.
565 * ipa-prop.c: Likewise.
566 * ipa-pure-const.c: Likewise.
567 * ipa-reference.c: Likewise.
568 * ipa-split.c: Likewise.
569 * ipa-utils.c: Likewise.
570 * loop-init.c: Likewise.
571 * lto-cgraph.c: Likewise.
572 * lto-section-in.c: Likewise.
573 * lto-section-out.c: Likewise.
574 * lto-streamer-in.c: Likewise.
575 * lto-streamer-out.c: Likewise.
576 * lto-streamer.c: Likewise.
577 * omp-low.c: Likewise.
578 * passes.c: Likewise.
579 * predict.c: Likewise.
580 * print-tree.c: Likewise.
581 * profile.c: Likewise.
583 * targhooks.c: Likewise.
584 * tracer.c: Likewise.
585 * trans-mem.c: Likewise.
586 * tree-call-cdce.c: Likewise.
587 * tree-cfg.c: Likewise.
588 * tree-cfgcleanup.c: Likewise.
589 * tree-chrec.c: Likewise.
590 * tree-complex.c: Likewise.
591 * tree-data-ref.c: Likewise.
592 * tree-dfa.c: Likewise.
593 * tree-eh.c: Likewise.
594 * tree-emutls.c: Likewise.
595 * tree-if-conv.c: Likewise.
596 * tree-inline.c: Likewise.
597 * tree-into-ssa.c: Likewise.
598 * tree-loop-distribution.c: Likewise.
599 * tree-mudflap.c: Likewise.
600 * tree-nested.c: Likewise.
601 * tree-nrv.c: Likewise.
602 * tree-object-size.c: Likewise.
603 * tree-outof-ssa.c: Likewise.
604 * tree-parloops.c: Likewise.
605 * tree-phinodes.c: Likewise.
606 * tree-predcom.c: Likewise.
607 * tree-pretty-print.c: Likewise.
608 * tree-profile.c: Likewise.
609 * tree-scalar-evolution.c: Likewise.
610 * tree-sra.c: Likewise.
611 * tree-ssa-address.c: Likewise.
612 * tree-ssa-alias.c: Likewise.
613 * tree-ssa-ccp.c: Likewise.
614 * tree-ssa-coalesce.c: Likewise.
615 * tree-ssa-copy.c: Likewise.
616 * tree-ssa-copyrename.c: Likewise.
617 * tree-ssa-dce.c: Likewise.
618 * tree-ssa-dom.c: Likewise.
619 * tree-ssa-dse.c: Likewise.
620 * tree-ssa-forwprop.c: Likewise.
621 * tree-ssa-ifcombine.c: Likewise.
622 * tree-ssa-live.c: Likewise.
623 * tree-ssa-loop-ch.c: Likewise.
624 * tree-ssa-loop-im.c: Likewise.
625 * tree-ssa-loop-ivcanon.c: Likewise.
626 * tree-ssa-loop-ivopts.c: Likewise.
627 * tree-ssa-loop-manip.c: Likewise.
628 * tree-ssa-loop-niter.c: Likewise.
629 * tree-ssa-loop-prefetch.c: Likewise.
630 * tree-ssa-loop-unswitch.c: Likewise.
631 * tree-ssa-loop.c: Likewise.
632 * tree-ssa-math-opts.c: Likewise.
633 * tree-ssa-operands.c: Likewise.
634 * tree-ssa-phiopt.c: Likewise.
635 * tree-ssa-phiprop.c: Likewise.
636 * tree-ssa-pre.c: Likewise.
637 * tree-ssa-propagate.c: Likewise.
638 * tree-ssa-reassoc.c: Likewise.
639 * tree-ssa-sccvn.c: Likewise.
640 * tree-ssa-sink.c: Likewise.
641 * tree-ssa-strlen.c: Likewise.
642 * tree-ssa-structalias.c: Likewise.
643 * tree-ssa-tail-merge.c: Likewise.
644 * tree-ssa-ter.c: Likewise.
645 * tree-ssa-threadedge.c: Likewise.
646 * tree-ssa-threadupdate.c: Likewise.
647 * tree-ssa-uncprop.c: Likewise.
648 * tree-ssa-uninit.c: Likewise.
649 * tree-ssa.c: Likewise.
650 * tree-ssanames.c: Likewise.
651 * tree-stdarg.c: Likewise.
652 * tree-streamer-in.c: Likewise.
653 * tree-switch-conversion.c: Likewise.
654 * tree-tailcall.c: Likewise.
655 * tree-vect-data-refs.c: Likewise.
656 * tree-vect-generic.c: Likewise.
657 * tree-vect-loop-manip.c: Likewise.
658 * tree-vect-loop.c: Likewise.
659 * tree-vect-patterns.c: Likewise.
660 * tree-vect-slp.c: Likewise.
661 * tree-vect-stmts.c: Likewise.
662 * tree-vectorizer.c: Likewise.
663 * tree-vrp.c: Likewise.
666 * value-prof.c: Likewise.
667 * var-tracking.c: Likewise.
668 * varpool.c: Likewise.
669 * vtable-verify.c: Likewise.
671 2013-10-23 Jan-Benedict Glaw <jbglaw@lug-owl.de>
673 * config/tilegx/tilegx.c: Include "tree.h".
675 2013-10-23 Jakub Jelinek <jakub@redhat.com>
677 * gimple-pretty-print.c (dump_ssaname_info): Always print "# " before
678 the info, not after it.
679 (gump_gimple_phi): Add COMMENT argument, if true, print "# " after
680 dump_ssaname_info call.
681 (pp_gimple_stmt_1): Adjust caller.
682 (dump_phi_nodes): Likewise. Don't print "# " here.
684 2013-10-22 Jan Hubicka <jh@suse.cz>
686 * i386.h (TARGET_MISALIGNED_MOVE_STRING_PROLOGUES_EPILOGUES): New
688 * x86-tune.def (TARGET_MISALIGNED_MOVE_STRING_PROLOGUES): Define it.
689 * i386.c (expand_small_movmem_or_setmem): New function.
690 (expand_set_or_movmem_prologue_epilogue_by_misaligned_moves): New
692 (alg_usable_p): Add support for value ranges; cleanup.
693 (ix86_expand_set_or_movmem): Add support for misaligned moves.
695 2013-10-22 Sterling Augustine <saugustine@google.com>
697 * doc/invoke.texi: Document -ggnu-pubnames.
698 * common.opt: Add new option -ggnu-pubnames and modify -gpubnames
700 * dwarf2out.c: Include gdb/gdb-index.h.
701 (DEBUG_PUBNAMES_SECTION, DEBUG_PUBTYPES_SECTION): Handle
702 debug_generate_pub_sections.
703 (is_java, output_pubtables, output_pubname): New functions.
704 (include_pubname_in_output): Handle debug_generate_pub_sections at
706 (size_of_pubnames): Use new local space_for_flags based on
707 debug_generate_pub_sections.
708 (output_pubnames): Unify pubnames and pubtypes output logic.
709 Genericize comments. Call output_pubname.
710 (dwarf2out_finish): Move logic to output_pubnames and call it.
712 2013-10-22 Uros Bizjak <ubizjak@gmail.com>
715 * config/i386/i386.c (put_condition_code) <case GTU, case LEU>:
716 Remove CCCmode handling.
717 <case LTU>: Return 'c' suffix for CCCmode.
718 <case GEU>: Return 'nc' suffix for CCCmode.
719 (ix86_cc_mode) <case GTU, case LEU>: Do not generate overflow checks.
720 * config/i386/i386.md (*sub<mode>3_cconly_overflow): Remove.
721 (*sub<mode>3_cc_overflow): Ditto.
722 (*subsi3_zext_cc_overflow): Ditto.
724 2013-10-22 Steve Ellcey <sellcey@mips.com>
726 * config/mips/mips.c (mips_rtx_costs): Fix cost estimate for nor
727 (AND (NOT OP1) (NOT OP2)).
729 2013-10-22 Bill Schmidt <wschmidt@vnet.ibm.com>
731 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
732 meaning of merge-high and merge-low masks for little endian; avoid
733 use of vector-pack masks for little endian for mismatched modes.
735 2013-10-22 Jan-Benedict Glaw <jbglaw@lug-owl.de>
737 * config/tilepro/tilepro.c: Include "tree.h".
739 2013-10-22 Andreas Schwab <schwab@suse.de>
741 * config/m68k/m68k.c (notice_update_cc): Handle register conflict
744 2013-10-22 Paolo Carlini <paolo.carlini@oracle.com>
746 * doc/contrib.texi ([Fran@,{c}ois Dumont], [Tim Shen],
747 [Ed Smith-Rowland]): New entries.
748 ([Stephen M. Webb]): Update.
750 2013-10-22 Andrew MacLeod <amacleod@redhat.com>
752 * tree-ssa-ter.h: Remove duplicate copy of file contents.
754 2013-10-21 Marek Polacek <polacek@redhat.com>
757 * fold-const.c (fold_range_test): Return 0 if the type is not
760 2013-10-21 Richard Sandiford <rdsandiford@googlemail.com>
762 * system.h: Move hwint.h include further down.
763 * hwint.h (sext_hwi, zext_hwi): Define unconditionally. Add
764 gcc_checking_asserts.
765 * hwint.c (sext_hwi, zext_hwi): Delete ENABLE_CHECKING versions.
767 2013-10-21 Bernd Edlinger <bernd.edlinger@hotmail.de>
769 Fix volatile issues in optimize_bit_field_compare.
770 * fold-const.c (optimize_bit_field_compare): Bail out if
771 lvolatilep or rvolatilep.
773 2013-10-21 Bernd Edlinger <bernd.edlinger@hotmail.de>
775 Fix DECL_BIT_FIELD depencency on flag_strict_volatile_bitfields
776 and get_inner_reference returning different pmode for non-volatile
777 bit-field members dependent on flag_strict_volatile_bitfields.
778 * stor-layout.c (layout_decl): Remove special handling of
779 flag_strict_volatile_bitfields.
780 * expr.c (get_inner_reference): Don't use DECL_BIT_FIELD
781 if flag_strict_volatile_bitfields > 0 and TREE_THIS_VOLATILE.
783 2013-10-21 Paulo Matos <pmatos@broadcom.com>
785 * ipa-inline.c (edge_badness): Cap edge->count at max_count for badness
788 2013-10-21 Jeff Law <law@redhat.com>
790 * tree-ssa-threadedge.c (thread_through_normal_block): New
791 argument VISITED. Remove VISISTED as a local variable. When we
792 have a threadable jump, verify the destination of the jump has not
794 (thread_across_edge): Allocate VISITED bitmap once at function
795 scope and use it throughout. Make sure to set appropriate bits in
796 VISITED for E (start of jump thread path).
797 * tree-ssa-threadupdate.c (mark_threaded_blocks): Reject threading
798 through a joiner if any edge on the path has a recorded jump thread.
800 2013-10-21 Ian Lance Taylor <iant@google.com>
802 * doc/invoke.texi (Optimize Options): For -fno-toplevel-reorder,
803 don't imply that attributes can solve all problems.
804 (Directory Options): Fix typo.
806 2013-10-21 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
808 * config/arm/arm.c (cortexa9_extra_costs): Update mult costs for
809 extend and extend_add.
811 2013-10-21 Richard Biener <rguenther@suse.de>
813 PR tree-optimization/58794
814 * fold-const.c (operand_equal_p): Compare FIELD_DECL operand
815 of COMPONENT_REFs with OEP_CONSTANT_ADDRESS_OF left in place.
817 2013-10-21 Richard Biener <rguenther@suse.de>
820 * fold-const.c (fold_binary_loc): Fold ((T) (X /[ex] C)) * C
821 to (T) X for sign-changing conversions (or no conversion).
823 2013-10-20 Uros Bizjak <ubizjak@gmail.com>
825 * config/i386/i386.md (kxnor<mode>): Add FLAGS_REG clobber.
827 2013-10-20 Jan Hubicka <jh@suse.cz>
829 * config/i386/i386-tune.def: Add comment; organize into categories
831 2013-10-21 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
833 * config/i386/i386.c (expand_set_or_movmem_via_loop): Add issetmem
834 argument. Update function comment.
835 (expand_set_or_movmem_via_rep): New function combining
836 expand_movmem_via_rep_mov and expand_setmem_via_rep_stos.
837 (expand_movmem_via_rep_mov): Remove.
838 expand_setmem_via_rep_stos): Remove.
839 (expand_movmem_epilogue): Update calls correspondingly.
840 (expand_setmem_epilogue_via_loop): Likewise.
842 (expand_setmem_epilogue): Add VEC_VALUE argument, refactor.
843 (expand_set_or_movmem_prologue): New function combining
844 expand_movmem_prologue and expand_setmem_prologue.
845 (expand_movmem_prologue): Remove.
846 (expand_setmem_prologue): Remove.
847 (expand_set_or_movmem_constant_prologue): New function combining
848 expand_constant_movmem_prologue and expand_constant_setmem_prologue.
849 (expand_constant_movmem_prologue): Remove.
850 (expand_constant_setmem_prologue): Remove.
851 (promote_duplicated_reg): Allow vector-const0 value.
852 (ix86_expand_set_or_movmem): New function combining ix86_expand_movmem
853 and ix86_expand_setmem.
854 (ix86_expand_movmem): Call ix86_expand_set_or_movmem.
855 (ix86_expand_setmem): Call ix86_expand_set_or_movmem.
857 2013-10-21 Diego Novillo <dnovillo@google.com>
859 * asan.c: Include tree.h
860 * bb-reorder.c: Likewise.
861 * cfgcleanup.c: Likewise.
862 * cfgloopmanip.c: Likewise.
863 * data-streamer-in.c: Likewise.
864 * data-streamer-out.c: Likewise.
865 * data-streamer.c: Likewise.
866 * dwarf2cfi.c: Likewise.
867 * graphite-blocking.c: Likewise.
868 * graphite-clast-to-gimple.c: Likewise.
869 * graphite-dependences.c: Likewise.
870 * graphite-interchange.c: Likewise.
871 * graphite-optimize-isl.c: Likewise.
872 * graphite-poly.c: Likewise.
873 * graphite-scop-detection.c: Likewise.
874 * graphite-sese-to-poly.c: Likewise.
875 * graphite.c: Likewise.
876 * ipa-devirt.c: Likewise.
877 * ipa-profile.c: Likewise.
880 * loop-init.c: Likewise.
881 * loop-unroll.c: Likewise.
882 * lower-subreg.c: Likewise.
883 * lto/lto-object.c: Likewise.
885 * reginfo.c: Likewise.
886 * tree-loop-distribution.c: Likewise.
887 * tree-parloops.c: Likewise.
888 * tree-ssa-strlen.c: Likewise.
889 * tree-streamer.c: Likewise.
890 * value-prof.c: Likewise.
891 * target-globals.c: Likewise.
892 * expr.h: Include tree-core.h instead of tree.h.
893 * gimple.h: Likewise.
894 * ipa-prop.h: Likewise.
895 * ipa-utils.h: Likewise.
896 * lto-streamer.h: Likewise.
897 * streamer-hooks.h: Likewise.
898 * ipa-reference.h: Include cgraph.h instead of tree.h.
899 * cgraph.h: Include basic-block.h instead of tree.h.
900 * tree-streamer.h: Do not include tree.h.
901 * genattrtab.c (write_header): Generate inclusion of
903 * genautomata.c (main): Likewise.
904 * genemit.c: Likewise.
905 * genopinit.c: Likewise.
906 * genoutput.c (output_prologue): Likewise.
907 * genpeep.c: Likewise.
909 2013-10-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
911 * config/rs6000/altivec.md (vec_unpacku_hi_v16qi): Adjust for
913 (vec_unpacku_hi_v8hi): Likewise.
914 (vec_unpacku_lo_v16qi): Likewise.
915 (vec_unpacku_lo_v8hi): Likewise.
917 2013-10-20 Jan Hubicka <jh@suse.cz>
919 * config/i386/x86-tune.def (X86_TUNE_SLOW_IMUL_IMM32_MEM,
920 X86_TUNE_SLOW_IMUL_IMM8): Keep enabled only for K8 and AMDFAM10.
921 (X86_TUNE_USE_VECTOR_FP_CONVERTS): Disable for generic.
923 2013-10-20 Richard Sandiford <rdsandiford@googlemail.com>
925 * config/mips/mips.h (ISA_HAS_WSBH): Define.
926 * config/mips/mips.md (UNSPEC_WSBH, UNSPEC_DSBH, UNSPEC_DSHD): New
928 (bswaphi2, bswapsi2, bswapdi2, wsbh, dsbh, dshd): New patterns.
930 2013-10-19 John David Anglin <danglin@gcc.gnu.org>
933 * system.h: Undef m_slot.
935 2013-10-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
937 * config/rs6000/rs6000.c (vspltis_constant): Make sure we check
938 all elements for both endian flavors.
940 2013-10-19 Uros Bizjak <ubizjak@gmail.com>
943 * config/i386/i386.c (ix86_function_value_regno): Add DX_REG,
944 ST1_REG and XMM1_REG for 32bit and 64bit targets. Also add DI_REG
945 and SI_REG for 64bit SYSV ABI targets.
947 2013-10-19 Uros Bizjak <ubizjak@gmail.com>
949 * mode-switching.c (create_pre_exit): Rename maybe_builtin_apply
950 to multi_reg_return. Clarify that we are skipping USEs of multiple
951 return registers. Use bool type where appropriate.
953 2013-10-18 Jan Hubicka <jh@suse.cz>
955 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Disable accumulation
957 * x86-tune.def (X86_TUNE_USE_LEAVE): Update comment.
958 (X86_TUNE_PUSH_MEMORY): Likewise.
959 (X86_TUNE_AVX256_UNALIGNED_LOAD_OPTIMAL,
960 X86_TUNE_AVX256_UNALIGNED_STORE_OPTIMAL): New.
961 (X86_TUNE_ACCUMULATE_OUTGOING_ARGS, X86_TUNE_ALWAYS_FANCY_MATH_387):
963 * i386.c (x86_accumulate_outgoing_args, x86_arch_always_fancy_math_387,
964 x86_avx256_split_unaligned_load, x86_avx256_split_unaligned_store):
966 (ix86_option_override_internal): Update to use tune features instead
969 2013-10-18 Cong Hou <congh@google.com>
971 PR tree-optimization/58508
972 * tree-vect-loop-manip.c (vect_loop_versioning): Hoist loop invariant
973 statement that contains data refs with zero-step.
975 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
977 * tree-ssa.h: Don't include gimple-low.h, tree-ssa-address.h,
978 sbitmap.h, tree-ssa-threadedge.h, tree-ssa-dom.h and tree-cfgcleanup.h.
979 * gimple-low.c (gimple_check_call_arg,
980 gimple_check_call_matching_types): Move to cgraph.c.
981 * gimple-low.h: Remove prototype.
982 * cgraph.c: (gimple_check_call_arg, gimple_check_call_matching_types):
983 Relocate from gimple-low.c.
984 * cgraph.h: Add prototype. Don't include basic-block.h.
985 * gimplify.c: Add gimple-low to include list.
986 * omp-low.c: Add gimple-low and tree-cfgcleanup.h to include list.
987 * tree-eh.c: Add gimple-low to include list.
988 * tree-nested.c: Likewise.
989 * cfgexpand.c: Add tree-ssa-address.h to include list.
991 * gimple-fold.c: Likewise.
992 * gimple-ssa-strength-reduction.c: Likewise.
993 * trans-mem.c: Likewise.
994 * tree-mudflap.c: Likewise.
995 * tree-ssa-loop-ivopts.c: Likewise.
996 * tree-ssa-dom.c: Include tree-ssa-threadedge.h and tree-ssa-dom.h.
997 (degenerate_phi_result): Move to tree-phinodes.c.
998 * tree-ssa-loop-ch.c: Include tree-ssa-threadedge.h.
999 * tree-ssa-threadedge.c: Likewise.
1000 * tree-vrp.c: Likewise.
1001 * tree-phinodes.c (degenerate_phi_result): Relocate here.
1002 * tree-ssa-dom.h (degenerate_phi_result): Remove Prototype.
1003 * tree-phinodes.h (degenerate_phi_result): Add prototype.
1004 * tree-ssa-copy.c: Include tree-ssa-dom.h.
1005 * tree-ssa-forwprop.c: Likewise.
1006 * tree-cfgcleanup.c (execute_cleanup_cfg_post_optimizing,
1007 pass_data_cleanup_cfg_post_optimizing,
1008 make_pass_cleanup_cfg_post_optimizing): Relocate from tree-optimize.c.
1009 * tree-optimize.c: Delete File.
1010 * graphite.c: Include tree-cfgcleanup.h.
1011 * passes.c: Likewise.
1012 * tree-cfg.c: Likewise.
1013 * tree-profile.c: Likewise.
1014 * tree-ssa-dse.c: Likewise.
1015 * tree-ssa-loop-ivcanon.c: Likewise.
1016 * tree-switch-conversion.c: Don't include tree-ssa-operands.h.
1017 * tree-outof-ssa.c: Include sbitmap.h.
1018 * tree-ssa-live.c: Likewise.
1019 * tree-ssa-propagate.c: Likewise.
1020 * tree-ssa-structalias.c: Likewise.
1021 * tree-stdarg.c: Likewise.
1022 * Makefile.in (OBJS): Delete tree-optimize.o.
1023 * basic-block.h (gcov_type, gcov_type_unsigned): Move to coretypes.h.
1024 * coretypes.h (gcov_type, gcov_type_unsigned): Relocate here.
1025 * varasm.c: Include basic-block.h.
1026 * cfgloop.h: Include function.h instead of basic-block.h
1027 (bb_loop_depth): Move to cfgloop.c.
1028 * cfgloop.c (bb_loop_depth): Relocate from cfgloop.h.
1030 2013-10-18 Teresa Johnson <tejohnson@google.com>
1032 * predict.c (probably_never_executed): Compare frequency-based
1033 count to number of training runs.
1034 * params.def (UNLIKELY_BB_COUNT_FRACTION): New parameter.
1036 2013-10-18 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1038 * config/arm/arm.c (cortexa9_extra_costs): New table.
1039 (arm_cortex_a9_tune): Use cortexa9_extra_costs.
1041 2013-10-18 Jeff Law <law@redhat.com>
1043 * tree-ssa-threadupdate.c: Do not include "tm.h" or "tm_p.h".
1045 * tree-ssa-threadupdate.c: Include "dbgcnt.h".
1046 (register_jump_thread): Add "registered_jump_thread" debug
1048 * dbgcnt.def (registered_jump_thread): New debug counter.
1050 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
1052 * config/rs6000/rs6000.c: Include cgraph.h.
1054 2013-10-18 Teresa Johnson <tejohnson@google.com>
1056 * tree-ssa-tail-merge.c (replace_block_by): Update edge
1057 weights during merging.
1059 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
1061 * tree-cfg.h: Rename from tree-flow.h. Remove #includes.
1062 * tree-ssa.h: Relocate required #includes from tree-cfg.h.
1063 * tree-ssa-operands.h: Remove prototype.
1064 * tree-ssa-operands.c (virtual_operand_p): Move to gimple.c.
1065 * gimple.c (virtual_operand_p): Relocate from gimple.c.
1066 * gimple.h: Add prototype.
1067 * gimple-ssa.h: Include tree-ssa-operands.h.
1068 * tree-dump.c: Add tree-cfg.h to include list.
1069 * tree-ssa-alias.c: Add ipa-reference.h to include list.
1070 * config/alpha/alpha.c: Include gimple-ssa.h instead of tree-flow.h.
1071 * config/i386/i386.c: Don't include tree-flow.h.
1072 * config/rs6000/rs6000.c: Likewise.
1074 2013-10-18 Jan-Benedict Glaw <jbglaw@lug-owl.de>
1076 * config/frv/frv.c (frv_init_cumulative_args): Fix wrong cast.
1078 2013-10-18 Richard Biener <rguenther@suse.de>
1080 * stor-layout.c (layout_type): Do not change TYPE_PRECISION
1081 or TYPE_UNSIGNED of integral types.
1082 (set_min_and_max_values_for_integral_type): Leave TYPE_MIN/MAX_VALUE
1083 NULL_TREE for zero-precision integral types.
1085 2013-10-18 James Greenhalgh <james.greenhalgh@arm.com>
1087 * config/aarch64/arm_neon.h
1088 (vcvt<ds>_n_<fsu><32,64>_<fsu><32,64>): Correct argument types.
1090 2013-10-17 Sriraman Tallam <tmsriram@google.com>
1093 * opth-gen.awk: Define target_flags_explicit.
1095 2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
1097 * config/rs6000/rs6000.c (enum rs6000_reload_reg_type): Add new
1098 fields to the reg_addr array that describes the valid addressing
1099 mode for any register, general purpose registers, floating point
1100 registers, and Altivec registers.
1101 (FIRST_RELOAD_REG_CLASS): Likewise.
1102 (LAST_RELOAD_REG_CLASS): Likewise.
1103 (struct reload_reg_map_type): Likewise.
1104 (reload_reg_map_type): Likewise.
1105 (RELOAD_REG_VALID): Likewise.
1106 (RELOAD_REG_MULTIPLE): Likewise.
1107 (RELOAD_REG_INDEXED): Likewise.
1108 (RELOAD_REG_OFFSET): Likewise.
1109 (RELOAD_REG_PRE_INCDEC): Likewise.
1110 (RELOAD_REG_PRE_MODIFY): Likewise.
1111 (reg_addr): Likewise.
1112 (mode_supports_pre_incdec_p): New helper functions to say whether
1113 a given mode supports PRE_INC, PRE_DEC, and PRE_MODIFY.
1114 (mode_supports_pre_modify_p): Likewise.
1115 (rs6000_debug_vector_unit): Rearrange the -mdebug=reg output to
1116 print the valid address mode bits for each mode.
1117 (rs6000_debug_print_mode): Likewise.
1118 (rs6000_debug_reg_global): Likewise.
1119 (rs6000_setup_reg_addr_masks): New function to set up the address
1120 mask bits for each type.
1121 (rs6000_init_hard_regno_mode_ok): Use memset to clear arrays.
1122 Call rs6000_setup_reg_addr_masks to set up the address mask bits.
1123 (rs6000_legitimate_address_p): Use mode_supports_pre_incdec_p and
1124 mode_supports_pre_modify_p to determine if PRE_INC, PRE_DEC, and
1125 PRE_MODIFY are supported.
1126 (rs6000_output_move_128bit): Change to use {src,dest}_vmx_p for altivec
1127 registers, instead of {src,dest}_av_p.
1128 (rs6000_print_options_internal): Tweak the debug output slightly.
1130 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
1132 * config/i386/sse.md (*vec_widen_smult_even_v8si): Remove
1135 2013-10-17 Andrew MacLeod <amacleod@redhat.com>
1137 * tree-flow.h (struct omp_region): Move to omp-low.c.
1138 Remove omp_ prototypes and variables.
1139 * gimple.h (omp_reduction_init): Move prototype to omp-low.h.
1140 (copy_var_decl): Relocate prototype from tree-flow.h.
1141 * gimple.c (copy_var_decl): Relocate from omp-low.c.
1142 * tree.h: Move prototype to omp-low.h.
1143 * omp-low.h: New File. Relocate prototypes here.
1144 * omp-low.c (struct omp_region): Make local here.
1145 (root_omp_region): Make static.
1146 (copy_var_decl) Move to gimple.c.
1147 (new_omp_region): Make static.
1148 (make_gimple_omp_edges): New. Refactored from tree-cfg.c make_edges.
1149 * tree-cfg.c: Include omp-low.h.
1150 (make_edges): Factor out OMP specific bits to make_gimple_omp_edges.
1151 * gimplify.c: Include omp-low.h.
1152 * tree-parloops.c: Likewise.
1154 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
1156 * config/i386/i386.c (ix86_fixup_binary_operands): When both source
1157 operands are in memory, prefer to force non-matched operand 1 to
1160 2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
1163 * config/rs6000/rs6000.c (rs6000_legitimate_address_p): Only
1164 restrict TImode addresses to single indirect registers if both
1165 -mquad-memory and -mvsx-timode are used.
1166 (rs6000_output_move_128bit): Use quad_load_store_p to determine if
1167 we should emit load/store quad. Remove using %y for quad memory
1170 * config/rs6000/rs6000.md (mov<mode>_ppc64, TI/PTImode): Add
1171 constraints to allow load/store quad on machines where TImode is
1172 not allowed in VSX registers. Use 'n' instead of 'F' constraint
1173 for TImode to load integer constants.
1175 2013-10-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1177 * config/aarch64/aarch64.c (aarch64_print_operand): Handle 'c'.
1179 2013-10-17 Marcus Shawcroft <marcus.shawcroft@arm.com>
1181 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Adjust
1182 handling of STACK_REG.
1184 2013-10-17 Richard Biener <rguenther@suse.de>
1186 PR tree-optimization/58143
1187 * tree-ssa-loop-im.c (arith_code_with_undefined_signed_overflow):
1189 (rewrite_to_defined_overflow): Likewise.
1190 (move_computations_dom_walker::before_dom): Rewrite stmts
1191 with undefined signed overflow that are not always executed
1192 into unsigned arithmetic.
1194 2013-10-16 Michael Meissner <meissner@linux.vnet.ibm.com>
1197 * config/rs6000/rs6000.opt (rs6000_isa_flags_explicit): Move the
1198 explicit isa flag to be an options variable, instead of using
1199 global_options_set. Remove define from rs6000.h.
1200 * config/rs6000/rs6000.h (rs6000_isa_flags_explicit): Likewise.
1202 * config/rs6000/rs6000.c (rs6000_option_override_internal):
1203 Initialize rs6000_isa_flags_explicit.
1204 (rs6000_function_specific_save): Add gcc_options* parameter, so
1205 that the powerpc builds after the 2013-10-15 changes.
1206 (rs6000_function_specific_restore): Likewise.
1208 2013-10-16 DJ Delorie <dj@redhat.com>
1210 * config/rl78/rl78.c (rl78_alloc_address_registers_macax): Verify
1211 op is a REG before checking REGNO.
1212 (rl78_alloc_physical_registers): Verify pattern is a SET before
1215 2013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1217 * gcc/config/rs6000/vector.md (vec_unpacks_hi_v4sf): Correct for
1219 (vec_unpacks_lo_v4sf): Likewise.
1220 (vec_unpacks_float_hi_v4si): Likewise.
1221 (vec_unpacks_float_lo_v4si): Likewise.
1222 (vec_unpacku_float_hi_v4si): Likewise.
1223 (vec_unpacku_float_lo_v4si): Likewise.
1225 2013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1227 * config/rs6000/vsx.md (vsx_concat_<mode>): Adjust output for LE.
1228 (vsx_concat_v2sf): Likewise.
1230 2013-10-16 James Greenhalgh <james.greenhalgh@arm.com>
1232 * config/aarch64/aarch64.md
1233 (*mov<mode>_aarch64): Fix output template for DUP (element) Scalar.
1235 2013-10-16 Andrew MacLeod <amacleod@redhat.com>
1237 PR tree-optimization/58697
1238 * cfgloop.c (get_estimated_loop_iterations_int): Rename from
1239 estimated_loop_iterations_int.
1240 (max_stmt_executions_int): Call get_max_loop_iterations_int.
1241 (get_max_loop_iterations_int): New. HWINT version of
1242 get_max_loop_iterations.
1243 * cfgloop.h: Add prototypes.
1244 * loop-iv.c (find_simple_exit): call get_estimated_loop_iterations_int.
1245 * loop-unroll.c (decide_peel_once_rolling): Call
1246 get_estimated_loop_iterations_int.
1247 * tree-ssa-loop-niter.c (estimated_loop_iterations_int): Add back.
1248 * tree-ssa-loop-niter.h: Tweak prototypes.
1250 2013-10-16 David Malcolm <dmalcolm@redhat.com>
1252 * gengtype-parse.c (struct_field_seq): Ignore access-control
1253 keywords ("public:" etc).
1255 2013-10-16 Marcus Shawcroft <marcus.shawcroft@arm.com>
1257 * config/aarch64/aarch64.c (aarch64_regno_regclass): Classify
1258 FRAME_POINTER_REGNUM and ARG_POINTER_REGNUM as POINTER_REGS.
1260 2013-10-16 Yvan Roux <yvan.roux@linaro.org>
1262 * config/arm/arm.opt (mlra): New option.
1263 * config/arm/arm.c (arm_lra_p): New function.
1264 (TARGET_LRA_P): Define.
1266 2013-10-16 Paulo Matos <pmatos@broadcom.com>
1268 * tree-core.h (tree_code_name): Remove.
1269 * tree.h (get_tree_code_name): New prototype.
1270 * tree.c (tree_code_name): Make static.
1271 (get_tree_code_name): New function.
1272 (dump_tree_statistics, tree_check_failed, tree_not_check_failed,
1273 tree_class_check_failed, tree_range_check_failed,
1274 tree_not_class_check_failed, omp_clause_check_failed,
1275 tree_contains_struct_check_failed, tree_operand_check_failed): Use new
1276 wrapper get_tree_code_name instead of calling tree_code_name directly.
1277 * tree-vrp.c (dump_asserts_for): Likewise.
1278 * tree-dump.c (dequeue_and_dump): Likewise.
1279 * tree-pretty-print.c (do_niy, dump_generic_node): Likewise.
1280 * tree-pretty-print.h (pp_unsupported_tree): Likewise.
1281 * lto-streamer-out.c (lto_write_tree, DFS_write_tree): Likewise.
1282 * tree-ssa-dom.c (print_expr_hash_elt): Likewise.
1283 * gimple-pretty-print.c (dump_unary_rhs, dump_binary_rhs,
1284 dump_ternary_rhs, dump_gimple_assign, dump_gimple_cond,
1285 dump_gimple_omp_for): Likewise.
1286 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Likewise.
1287 * tree-ssa-pre.c (print_pre_expr): Likewise.
1288 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Likewise.
1289 * print-tree.c (print_node_brief, print_node): Likewise.
1290 * gimple.c (gimple_check_failed): Likewise.
1291 * lto-streamer.c (lto_tag_name, print_lto_report): Likewise.
1292 * config/frv/frv.c (frv_init_cumulative_args): Likewise.
1293 * config/mep/mep.c (mep_validate_vliw): Likewise.
1294 * config/iq2000/iq2000.c (init_cumulative_args): Likewise.
1295 * config/rs6000/rs6000.c (init_cumulative_args): Likewise.
1297 2013-10-16 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
1299 * config/i386/i386.c (ix86_option_override_internal): Enable FMA4
1302 2013-10-16 Hans-Peter Nilsson <hp@axis.com>
1304 * config/cris/t-elfmulti (MULTILIB_OPTIONS, MULTILIB_DIRNAMES)
1305 (MULTILIB_MATCHES): Add multilib for -march=v8.
1307 2013-10-15 Sriraman Tallam <tmsriram@google.com>
1310 * optc-save-gen.awk: Add extra parameter to the save and restore
1312 * opth-gen.awk: Generate new TARGET_* macros to accept a parameter.
1313 * tree.c (build_optimization_node): New parameter. Add extra parameter
1314 to call to cl_optimization_save.
1315 (build_target_option_node): New parameter. Add extra parameter
1316 to call to cl_target_option_save.
1317 * tree.h (build_optimization_node): New parameter.
1318 (build_target_option_node): New parameter.
1319 * c-family/c-common.c (handle_optimize_attribute): Fix calls to
1320 build_optimization_node and build_target_option_node.
1321 * c-family/c-pragma.c (handle_pragma_optimize): Ditto.
1322 (handle_pragma_push_options): Ditto.
1323 * toplev.c (process_options): Ditto.
1324 * opts.c (init_options_struct): Check for opts_set non-null.
1325 * target.def (target_option.save): New parameter.
1326 (target_option.restore): New parameter.
1327 * tm.texi: Generate.
1328 * config/i386/i386-c.c (ix86_target_macros_internal): Ditto.
1329 (ix86_pragma_target_parse): Ditto.
1330 * config/i386/i386-protos.h (ix86_valid_target_attribute_tree): New
1332 * config/rs6000/rs6000.c (rs6000_option_override_internal): Fix calls
1333 to build_optimization_node and build_target_option_node.
1334 (rs6000_valid_attribute_p): Ditto.
1335 (rs6000_pragma_target_parse): Ditto.
1336 * config/i386/i386.opt (x_ix86_target_flags_explicit): New TargetSave
1338 * config/i386/i386.h:
1339 TARGET_64BIT_P: New Macro
1340 TARGET_MMX_P: New Macro.
1341 TARGET_3DNOW_P: New Macro.
1342 TARGET_3DNOW_A_P: New Macro.
1343 TARGET_SSE_P: New Macro.
1344 TARGET_SSE2_P: New Macro.
1345 TARGET_SSE3_P: New Macro.
1346 TARGET_SSSE3_P: New Macro.
1347 TARGET_SSE4_1_P: New Macro.
1348 TARGET_SSE4_2_P: New Macro.
1349 TARGET_AVX_P: New Macro.
1350 TARGET_AVX2_P: New Macro.
1351 TARGET_AVX512F_P: New Macro.
1352 TARGET_AVX512PF_P: New Macro.
1353 TARGET_AVX512ER_P: New Macro.
1354 TARGET_AVX512CD_P: New Macro.
1355 TARGET_FMA_P: New Macro.
1356 TARGET_SSE4A_P: New Macro.
1357 TARGET_FMA4_P: New Macro.
1358 TARGET_XOP_P: New Macro.
1359 TARGET_LWP_P: New Macro.
1360 TARGET_ABM_P: New Macro.
1361 TARGET_BMI_P: New Macro.
1362 TARGET_BMI2_P: New Macro.
1363 TARGET_LZCNT_P: New Macro.
1364 TARGET_TBM_P: New Macro.
1365 TARGET_POPCNT_P: New Macro.
1366 TARGET_SAHF_P: New Macro.
1367 TARGET_MOVBE_P: New Macro.
1368 TARGET_CRC32_P: New Macro.
1369 TARGET_AES_P: New Macro.
1370 TARGET_PCLMUL_P: New Macro.
1371 TARGET_CMPXCHG16B_P: New Macro.
1372 TARGET_FSGSBASE_P: New Macro.
1373 TARGET_RDRND_P: New Macro.
1374 TARGET_F16C_P: New Macro.
1375 TARGET_RTM_P: New Macro.
1376 TARGET_HLE_P: New Macro.
1377 TARGET_RDSEED_P: New Macro.
1378 TARGET_PRFCHW_P: New Macro.
1379 TARGET_ADX_P: New Macro.
1380 TARGET_FXSR_P: New Macro.
1381 TARGET_XSAVE_P: New Macro.
1382 TARGET_XSAVEOPT_P: New Macro.
1383 TARGET_LP64_P: New Macro.
1384 TARGET_X32_P: New Macro.
1385 TARGET_FPMATH_DEFAULT_P: New Macro.
1386 TARGET_FLOAT_RETURNS_IN_80387_P: New Macro.
1387 * config/i386/i386.c (ix86_option_override_internal): New parameters.
1389 Change ix86_tune_string to access opts->x_ix86_tune_string.
1390 Change ix86_isa_flags to access opts->x_ix86_isa_flags.
1391 Change ix86_arch_string to access opts->x_ix86_arch_string.
1392 Change ix86_stringop_alg to access opts->x_ix86_stringop_alg.
1393 Change ix86_pmode to access opts->x_ix86_pmode.
1394 Change ix86_abi to access opts->x_ix86_abi.
1395 Change ix86_cmodel to access opts->x_ix86_cmodel.
1396 Change ix86_asm_dialect to access opts->x_ix86_asm_dialect.
1397 Change ix86_isa_flags_explicit to access
1398 opts->x_ix86_isa_flags_explicit.
1399 Change ix86_dump_tunes to access opts->x_ix86_dump_tunes.
1400 Change ix86_regparm to access opts->x_ix86_regparm.
1401 Change ix86_branch_cost to access opts->x_ix86_branch_cost.
1402 Change ix86_preferred_stack_boundary_arg to access
1403 opts->x_ix86_preferred_stack_boundary_arg.
1404 Change ix86_force_align_arg_pointer to access
1405 opts->x_ix86_force_align_arg_pointer.
1406 Change ix86_incoming_stack_boundar_arg to access
1407 opts->x_ix86_incoming_stack_boundar_arg.
1408 Change ix86_fpmath to access opts->x_ix86_fpmath.
1409 Change ix86_veclibabi_type to access opts->x_ix86_veclibabi_type.
1410 Change ix86_recip_name to access opts->x_ix86_recip_name.
1411 Change ix86_stack_protector_guard to access
1412 opts->x_ix86_stack_protector_guard.
1413 Change ix86_tune_memcpy_strategy to access
1414 opts->x_ix86_tune_memcpy_strategy.
1415 Change ix86_tune_memset_strategy to access
1416 opts->x_ix86_tune_memset_strategy.
1417 Change global_options to access opts.
1418 Change global_options_set to access opts_set.
1419 Change TARGET_64BIT to TARGET_64BIT_P (opts->...).
1420 Change TARGET_MMX to TARGET_MMX_P (opts->...).
1421 Change TARGET_3DNOW to TARGET_3DNOW_P (opts->...).
1422 Change TARGET_3DNOW_A to TARGET_3DNOW_A_P (opts->...).
1423 Change TARGET_SSE to TARGET_SSE_P (opts->...).
1424 Change TARGET_SSE2 to TARGET_SSE2_P (opts->...).
1425 Change TARGET_SSE3 to TARGET_SSE3_P (opts->...).
1426 Change TARGET_SSSE3 to TARGET_SSSE3_P (opts->...).
1427 Change TARGET_SSE4_1 to TARGET_SSE4_1_P (opts->...).
1428 Change TARGET_SSE4_2 to TARGET_SSE4_2_P (opts->...).
1429 Change TARGET_AVX to TARGET_AVX_P (opts->...).
1430 Change TARGET_AVX2 to TARGET_AVX2_P (opts->...).
1431 Change TARGET_AVX512F to TARGET_AVX512F_P (opts->...).
1432 Change TARGET_AVX512PF to TARGET_AVX512PF_P (opts->...).
1433 Change TARGET_AVX512ER to TARGET_AVX512ER_P (opts->...).
1434 Change TARGET_AVX512CD to TARGET_AVX512CD_P (opts->...).
1435 Change TARGET_FMA to TARGET_FMA_P (opts->...).
1436 Change TARGET_SSE4A to TARGET_SSE4A_P (opts->...).
1437 Change TARGET_FMA4 to TARGET_FMA4_P (opts->...).
1438 Change TARGET_XOP to TARGET_XOP_P (opts->...).
1439 Change TARGET_LWP to TARGET_LWP_P (opts->...).
1440 Change TARGET_ABM to TARGET_ABM_P (opts->...).
1441 Change TARGET_BMI to TARGET_BMI_P (opts->...).
1442 Change TARGET_BMI2 to TARGET_BMI2_P (opts->...).
1443 Change TARGET_LZCNT to TARGET_LZCNT_P (opts->...).
1444 Change TARGET_TBM to TARGET_TBM_P (opts->...).
1445 Change TARGET_POPCNT to TARGET_POPCNT_P (opts->...).
1446 Change TARGET_SAHF to TARGET_SAHF_P (opts->...).
1447 Change TARGET_MOVBE to TARGET_MOVBE_P (opts->...).
1448 Change TARGET_CRC32 to TARGET_CRC32_P (opts->...).
1449 Change TARGET_AES to TARGET_AES_P (opts->...).
1450 Change TARGET_PCLMUL to TARGET_PCLMUL_P (opts->...).
1451 Change TARGET_CMPXCHG16B to TARGET_CMPXCHG16B_P (opts->...).
1452 Change TARGET_FSGSBASE to TARGET_FSGSBASE_P (opts->...).
1453 Change TARGET_RDRND to TARGET_RDRND_P (opts->...).
1454 Change TARGET_F16C to TARGET_F16C_P (opts->...).
1455 Change TARGET_RTM to TARGET_RTM_P (opts->...).
1456 Change TARGET_HLE to TARGET_HLE_P (opts->...).
1457 Change TARGET_RDSEED to TARGET_RDSEED_P (opts->...).
1458 Change TARGET_PRFCHW to TARGET_PRFCHW_P (opts->...).
1459 Change TARGET_ADX to TARGET_ADX_P (opts->...).
1460 Change TARGET_FXSR to TARGET_FXSR_P (opts->...).
1461 Change TARGET_XSAVE to TARGET_XSAVE_P (opts->...).
1462 Change TARGET_XSAVEOPT to TARGET_XSAVEOPT_P (opts->...).
1463 Change TARGET_LP64 to TARGET_LP64_P (opts->...).
1464 Change TARGET_X32 to TARGET_X32_P (opts->...).
1465 Change TARGET_FPMATH_DEFAULT to TARGET_FPMATH_DEFAULT_P (opts->...).
1466 Change TARGET_FLOAT_RETURNS_IN_80387 to
1467 TARGET_FLOAT_RETURNS_IN_80387_P (opts->...).
1468 (ix86_function_specific_save): New parameter. Use opts-> fields
1469 to replace global fields.
1470 (ix86_function_specific_restore): Ditto.
1471 (ix86_valid_target_attribute_inner_p): New parameters.
1473 Fix call to ix86_handle_option and set_option.
1474 (ix86_valid_target_attribute_tree): New parameters.
1475 Change global_options to access opts.
1476 Change global_options_set to access opts_set.
1477 Fix call to ix86_valid_target_attribute_inner_p.
1478 Change ix86_tune_string to access opts->x_ix86_tune_string.
1479 Change ix86_arch_string to access opts->x_ix86_arch_string.
1480 Change ix86_fpmath to access opts->x_ix86_fpmath
1481 Fix call to ix86_option_override_internal.
1482 Fix call to ix86_add_new_builtins.
1483 Fix calls to build_optimization_node and build_target_option_node.
1484 (ix86_valid_target_attribute_p): Remove access to global_options.
1485 Use new gcc_options structure func_options.
1486 Fix call to ix86_valid_target_attribute_tree.
1487 Fix call to build_optimization_node.
1488 (get_builtin_code_for_version): Fix call to
1489 ix86_valid_target_attribute_tree.
1491 2013-10-15 David Malcolm <dmalcolm@redhat.com>
1493 * Makefile.in (PICFLAG): New.
1494 (enable_host_shared): New.
1495 (INTERNAL_CFLAGS): Use PICFLAG.
1496 (LIBIBERTY): Use pic build of libiberty.a if configured with
1497 --enable-host-shared.
1498 * configure.ac: Add --enable-host-shared, setting up new
1500 * configure: Regenerate.
1501 * doc/install.texi (--enable-shared): Add note contrasting it with ...
1502 (--enable-host-shared): New option.
1504 2013-10-15 Richard Biener <rguenther@suse.de>
1506 * tree-tailcall.c (find_tail_calls): Don't use tail-call recursion
1507 for built-in functions.
1509 2013-10-15 Zhenqiang Chen <zhenqiang.chen@arm.com>
1511 * tree-ssa-reassoc.c: Include rtl.h and tm_p.h.
1512 (optimize_range_tests_1): New function,
1513 extracted from optimize_range_tests.
1514 (optimize_range_tests_xor): Similarly.
1515 (optimize_range_tests_diff): New function.
1516 (optimize_range_tests): Use optimize_range_tests_1.
1518 2013-10-15 Cong Hou <congh@google.com>
1520 * tree-vect-loop.c (vect_is_simple_reduction_1): Relax the
1521 requirement of the reduction pattern so that one operand of the
1522 reduction operation can come from outside of the loop.
1524 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1526 * config/arm/neon-schedgen.ml: Remove.
1527 * config/arm/cortex-a9-neon.md: Remove comment regarding
1530 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1532 * config/arm/types: Remove old neon types.
1534 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1536 * config/arm/cortex-a7.md
1537 (cortex_a7_neon_type): New.
1538 (cortex_a7_neon_mul): Update for new types.
1539 (cortex_a7_neon_mla): Likewise.
1540 (cortex_a7_neon): Likewise.
1542 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1544 * config/arm/cortex-a15-neon.md
1545 (cortex_a15_neon_type): New,
1547 (cortex_a15_neon_int_1): Remove.
1548 (cortex_a15_neon_int_2): Likewise.
1549 (cortex_a15_neon_int_3): Likewise.
1550 (cortex_a15_neon_int_4): Likewise.
1551 (cortex_a15_neon_int_5): Likewise.
1552 (cortex_a15_neon_vqneg_vqabs): Likewise.
1553 (cortex_a15_neon_vmov): Likewise.
1554 (cortex_a15_neon_vaba): Likewise.
1555 (cortex_a15_neon_vaba_qqq): Likewise.
1556 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1557 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
1558 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
1560 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1561 (cortex_a15_neon_mla_qqq_8_16): Likewise.
1562 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar): Likewise.
1563 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
1564 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
1565 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
1566 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
1567 (cortex_a15_neon_shift_1): Likewise.
1568 (cortex_a15_neon_shift_2): Likewise.
1569 (cortex_a15_neon_shift_3): Likewise.
1570 (cortex_a15_neon_vshl_ddd): Likewise.
1571 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
1572 (cortex_a15_neon_vsra_vrsra): Likewise.
1573 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
1574 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
1575 (cortex_a15_neon_bp_3cycle): Likewise.
1576 (cortex_a15_neon_ldm_2): Likewise.
1577 (cortex_a15_neon_stm_2): Likewise.
1578 (cortex_a15_neon_mcr): Likewise.
1579 (cortex_a15_neon_mrc): Likewise.
1580 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
1581 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
1582 (cortex_a15_neon_fp_vmul_ddd): Likewise.
1583 (cortex_a15_neon_fp_vmul_qqd): Likewise.
1584 (cortex_a15_neon_fp_vmla_ddd): Likewise.
1585 (cortex_a15_neon_fp_vmla_qqq): Likewise.
1586 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
1587 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
1588 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
1589 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
1590 (cortex_a15_neon_bp_simple): Likewise.
1591 (cortex_a15_neon_bp_2cycle): Likewise.
1592 (cortex_a15_neon_bp_3cycle): Likewise.
1593 (cortex_a15_neon_vld1_1_2_regs): Likewise.
1594 (cortex_a15_neon_vld1_3_4_regs): Likewise.
1595 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
1596 (cortex_a15_neon_vld2_4_regs): Likewise.
1597 (cortex_a15_neon_vld3_vld4): Likewise.
1598 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
1599 (cortex_a15_neon_vst1_3_4_regs): Likewise.
1600 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Rename to...
1601 (cortex_a15_neon_vst2_4_regs_vst3): ...This, update for new attributes.
1602 (cortex_a15_neon_vst3_vst4): Rename to...
1603 (cortex_a15_neon_vst4): This, update for new attributes.
1604 (cortex_a15_neon_vld1_vld2_lane): Update for new attributes.
1605 (cortex_a15_neon_vld3_vld4_lane): Likewise.
1606 (cortex_a15_neon_vst1_vst2_lane): Likewise.
1607 (cortex_a15_neon_vst3_vst4_lane): Likewise.
1608 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
1609 (cortex_a15_neon_ldm_2): Likewise.
1610 (cortex_a15_neon_stm_2): Likewise.
1611 (cortex_a15_neon_mcr): Likewise.
1612 (cortex_a15_neon_mcr_2_mcrr): Likewise.
1613 (cortex_a15_neon_mrc): Likewise.
1614 (cortex_a15_neon_mrrc): Likewise.
1616 (cortex_a15_neon_abd): New.
1617 (cortex_a15_neon_abd_q): Likewise.
1618 (cortex_a15_neon_aba): Likewise.
1619 (cortex_a15_neon_aba_q): Likewise.
1620 (cortex_a15_neon_acc): Likewise.
1621 (cortex_a15_neon_acc_q): Likewise.
1622 (cortex_a15_neon_arith_basic): Likewise.
1623 (cortex_a15_neon_arith_complex): Likewise.
1624 (cortex_a15_neon_multiply): Likewise.
1625 (cortex_a15_neon_multiply_q): Likewise.
1626 (cortex_a15_neon_mla): Likewise.
1627 (cortex_a15_neon_mla_q): Likewise.
1628 (cortex_a15_neon_sat_mla_long): Likewise.
1629 (cortex_a15_neon_shift_acc): Likewise.
1630 (cortex_a15_neon_shift_imm_basic): Likewise.
1631 (cortex_a15_neon_shift_imm_complex): Likewise.
1632 (cortex_a15_neon_shift_reg_basic): Likewise.
1633 (cortex_a15_neon_shift_reg_basic_q): Likewise.
1634 (cortex_a15_neon_shift_reg_complex): Likewise.
1635 (cortex_a15_neon_shift_reg_complex_q): Likewise.
1636 (cortex_a15_neon_fp_negabs): Likewise
1637 (cortex_a15_neon_fp_arith): Likewise
1638 (cortex_a15_neon_fp_arith_q): Likewise
1639 (cortex_a15_neon_fp_cvt_int): Likewise
1640 (cortex_a15_neon_fp_cvt_int_q): Likewise
1641 (cortex_a15_neon_fp_cvt_16): Likewise
1642 (cortex_a15_neon_fp_mul): Likewise
1643 (cortex_a15_neon_fp_mul_q): Likewise
1644 (cortex_a15_neon_fp_mla): Likewise
1645 (cortex_a15_neon_fp_mla_q): Likewise
1646 (cortex_a15_neon_fp_recps_rsqrte): Likewise.
1647 (cortex_a15_neon_fp_recps_rsqrte_q): Likewise.
1648 (cortex_a15_neon_bitops): Likewise.
1649 (cortex_a15_neon_bitops_q): Likewise.
1650 (cortex_a15_neon_from_gp): Likewise.
1651 (cortex_a15_neon_from_gp_q): Likewise.
1652 (cortex_a15_neon_tbl3_tbl4): Likewise.
1653 (cortex_a15_neon_zip_q): Likewise.
1654 (cortex_a15_neon_to_gp): Likewise.
1655 (cortex_a15_neon_load_a): Likewise.
1656 (cortex_a15_neon_load_b): Likewise.
1657 (cortex_a15_neon_load_c): Likewise.
1658 (cortex_a15_neon_load_d): Likewise.
1659 (cortex_a15_neon_load_e): Likewise.
1660 (cortex_a15_neon_load_f): Likewise.
1661 (cortex_a15_neon_store_a): Likewise.
1662 (cortex_a15_neon_store_b): Likewise.
1663 (cortex_a15_neon_store_c): Likewise.
1664 (cortex_a15_neon_store_d): Likewise.
1665 (cortex_a15_neon_store_e): Likewise.
1666 (cortex_a15_neon_store_f): Likewise.
1667 (cortex_a15_neon_store_g): Likewise.
1668 (cortex_a15_neon_store_h): Likewise.
1669 (cortex_a15_vfp_to_from_gp): Likewise.
1671 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1673 * config/arm/cortex-a9-neon.md (cortex_a9_neon_type): New.
1675 (cortex_a9_neon_vshl_ddd): Remove.
1676 (cortex_a9_neon_vst3_vst4): Likewise.
1677 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
1679 (cortex_a9_neon_bit_ops_q): New.
1681 (cortex_a9_neon_int_1): Use cortex_a8_neon_type.
1682 (cortex_a9_neon_int_2): Likewise.
1683 (cortex_a9_neon_int_3): Likewise.
1684 (cortex_a9_neon_int_4): Likewise.
1685 (cortex_a9_neon_int_5): Likewise.
1686 (cortex_a9_neon_vqneg_vqabs): Likewise.
1687 (cortex_a9_neon_vmov): Likewise.
1688 (cortex_a9_neon_vaba): Likewise.
1689 (cortex_a9_neon_vaba_qqq): Likewise.
1690 (cortex_a9_neon_shift_1): Likewise.
1691 (cortex_a9_neon_shift_2): Likewise.
1692 (cortex_a9_neon_shift_3): Likewise.
1693 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
1694 (cortex_a9_neon_vsra_vrsra): Likewise.
1695 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1696 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
1697 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
1699 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1700 (cortex_a9_neon_mla_qqq_8_16): Likewise.
1701 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
1703 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
1704 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
1705 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
1706 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
1707 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
1708 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
1709 (cortex_a9_neon_fp_vsum): Likewise.
1710 (cortex_a9_neon_fp_vmul_ddd): Likewise.
1711 (cortex_a9_neon_fp_vmul_qqd): Likewise.
1712 (cortex_a9_neon_fp_vmla_ddd): Likewise.
1713 (cortex_a9_neon_fp_vmla_qqq): Likewise.
1714 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
1715 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
1716 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
1717 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
1718 (cortex_a9_neon_bp_simple): Likewise.
1719 (cortex_a9_neon_bp_2cycle): Likewise.
1720 (cortex_a9_neon_bp_3cycle): Likewise.
1721 (cortex_a9_neon_ldr): Likewise.
1722 (cortex_a9_neon_str): Likewise.
1723 (cortex_a9_neon_vld1_1_2_regs): Likewise.
1724 (cortex_a9_neon_vld1_3_4_regs): Likewise.
1725 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
1726 (cortex_a9_neon_vld2_4_regs): Likewise.
1727 (cortex_a9_neon_vld3_vld4): Likewise.
1728 (cortex_a9_neon_vld1_vld2_lane): Likewise.
1729 (cortex_a9_neon_vld3_vld4_lane): Likewise.
1730 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
1731 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
1732 (cortex_a9_neon_vst1_3_4_regs): Likewise.
1733 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
1734 (cortex_a9_neon_vst1_vst2_lane): Likewise.
1735 (cortex_a9_neon_vst3_vst4_lane): Likewise.
1736 (cortex_a9_neon_mcr): Likewise.
1737 (cortex_a9_neon_mcr_2_mcrr): Likewise.
1738 (cortex_a9_neon_mrc): Likewise.
1739 (cortex_a9_neon_mrrc): Likewise.
1741 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1743 * config/arm/cortex-a8-neon.md (cortex_a8_neon_type): New.
1745 (cortex_a8_neon_vshl_ddd): Remove.
1746 (cortex_a8_neon_vst3_vst4): Likewise.
1747 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
1749 (cortex_a8_neon_bit_ops_q): New.
1751 (cortex_a8_neon_int_1): Use cortex_a8_neon_type.
1752 (cortex_a8_neon_int_2): Likewise..
1753 (cortex_a8_neon_int_3): Likewise.
1754 (cortex_a8_neon_int_5): Likewise.
1755 (cortex_a8_neon_vqneg_vqabs): Likewise.
1756 (cortex_a8_neon_int_4): Likewise.
1757 (cortex_a8_neon_vaba): Likewise.
1758 (cortex_a8_neon_vaba_qqq): Likewise.
1759 (cortex_a8_neon_shift_1): Likewise.
1760 (cortex_a8_neon_shift_2): Likewise.
1761 (cortex_a8_neon_shift_3): Likewise.
1762 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
1763 (cortex_a8_neon_vsra_vrsra): Likewise.
1764 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1765 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
1766 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
1768 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1769 (cortex_a8_neon_mla_qqq_8_16): Likewise.
1770 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
1772 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
1773 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
1774 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
1775 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
1776 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
1777 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
1778 (cortex_a8_neon_fp_vsum): Likewise.
1779 (cortex_a8_neon_fp_vmul_ddd): Likewise.
1780 (cortex_a8_neon_fp_vmul_qqd): Likewise.
1781 (cortex_a8_neon_fp_vmla_ddd): Likewise.
1782 (cortex_a8_neon_fp_vmla_qqq): Likewise.
1783 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
1784 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
1785 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
1786 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
1787 (cortex_a8_neon_bp_simple): Likewise.
1788 (cortex_a8_neon_bp_2cycle): Likewise.
1789 (cortex_a8_neon_bp_3cycle): Likewise.
1790 (cortex_a8_neon_ldr): Likewise.
1791 (cortex_a8_neon_str): Likewise.
1792 (cortex_a8_neon_vld1_1_2_regs): Likewise.
1793 (cortex_a8_neon_vld1_3_4_regs): Likewise.
1794 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
1795 (cortex_a8_neon_vld2_4_regs): Likewise.
1796 (cortex_a8_neon_vld3_vld4): Likewise.
1797 (cortex_a8_neon_vld1_vld2_lane): Likewise.
1798 (cortex_a8_neon_vld3_vld4_lane): Likewise.
1799 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
1800 (cortex_a8_neon_vst1_3_4_regs): Likewise.
1801 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
1802 (cortex_a8_neon_vst1_vst2_lane): Likewise.
1803 (cortex_a8_neon_vst3_vst4_lane): Likewise.
1804 (cortex_a8_neon_mcr): Likewise.
1805 (cortex_a8_neon_mcr_2_mcrr): Likewise.
1806 (cortex_a8_neon_mrc): Likewise.
1807 (cortex_a8_neon_mrrc): Likewise.
1809 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1811 * config/aarch64/iterators.md (Vetype): Add SF and DF modes.
1813 * config/aarch64/aarch64-simd.md (neon_type): Remove.
1814 (aarch64_simd_dup<mode>): Add "type" attribute.
1815 (aarch64_dup_lane<mode>): Likewise.
1816 (aarch64_dup_lane_<vswap_width_name><mode>): Likewise.
1817 (*aarch64_simd_mov<mode>): Likewise.
1818 (aarch64_simd_mov_from_<mode>low): Likewise.
1819 (aarch64_simd_mov_from_<mode>high): Likewise.
1820 (orn<mode>3): Likewise.
1821 (bic<mode>3): Likewise.
1822 (add<mode>3): Likewise.
1823 (sub<mode>3): Likewise.
1824 (mul<mode>3): Likewise.
1825 (*aarch64_mul3_elt<mode>): Likewise.
1826 (*aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
1827 (*aarch64_mul3_elt_to_128df): Likewise.
1828 (*aarch64_mul3_elt_to_64v2df): Likewise.
1829 (neg<mode>2): Likewise.
1830 (abs<mode>2): Likewise.
1831 (abd<mode>_3): Likewise.
1832 (aba<mode>_3): Likewise.
1833 (fabd<mode>_3): Likewise.
1834 (*fabd_scalar<mode>3): Likewise.
1835 (and<mode>3): Likewise.
1836 (ior<mode>3): Likewise.
1837 (xor<mode>3): Likewise.
1838 (one_cmpl<mode>2): Likewise.
1839 (aarch64_simd_vec_set<mode>): Likewise.
1840 (aarch64_simd_lshr<mode>): Likewise.
1841 (aarch64_simd_ashr<mode>): Likewise.
1842 (aarch64_simd_imm_shl<mode>): Likewise.
1843 (aarch64_simd_reg_sshl<mode): Likewise.
1844 (aarch64_simd_reg_shl<mode>_unsigned): Likewise.
1845 (aarch64_simd_reg_shl<mode>_signed): Likewise.
1846 (aarch64_simd_vec_setv2di): Likewise.
1847 (aarch64_simd_vec_set<mode>): Likewise.
1848 (aarch64_mla<mode>): Likewise.
1849 (*aarch64_mla_elt<mode>): Likewise.
1850 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
1851 (aarch64_mls<mode>): Likewise.
1852 (*aarch64_mls_elt<mode>): Likewise.
1853 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
1854 (<su><maxmin><mode>3): Likewise.
1855 (move_lo_quad_<mode>): Likewise.
1856 (aarch64_simd_move_hi_quad_<mode>): Likewise.
1857 (aarch64_simd_vec_pack_trunc_<mode>): Likewise.
1858 (vec_pack_trunc_<mode>): Likewise.
1859 (aarch64_simd_vec_unpack<su>_lo_<mode>): Likewise.
1860 (aarch64_simd_vec_unpack<su>_hi_<mode>): Likewise.
1861 (*aarch64_<su>mlal_lo<mode>): Likewise.
1862 (*aarch64_<su>mlal_hi<mode>): Likewise.
1863 (*aarch64_<su>mlsl_lo<mode>): Likewise.
1864 (*aarch64_<su>mlsl_hi<mode>): Likewise.
1865 (*aarch64_<su>mlal<mode>): Likewise.
1866 (*aarch64_<su>mlsl<mode>): Likewise.
1867 (aarch64_simd_vec_<su>mult_lo_<mode>): Likewise.
1868 (aarch64_simd_vec_<su>mult_hi_<mode>): Likewise.
1869 (add<mode>3): Likewise.
1870 (sub<mode>3): Likewise.
1871 (mul<mode>3): Likewise.
1872 (div<mode>3): Likewise.
1873 (neg<mode>2): Likewise.
1874 (abs<mode>2): Likewise.
1875 (fma<mode>4): Likewise.
1876 (*aarch64_fma4_elt<mode>): Likewise.
1877 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
1878 (*aarch64_fma4_elt_to_128df): Likewise.
1879 (*aarch64_fma4_elt_to_64v2df): Likewise.
1880 (fnma<mode>4): Likewise.
1881 (*aarch64_fnma4_elt<mode>): Likewise.
1882 (*aarch64_fnma4_elt_<vswap_width_name><mode>
1883 (*aarch64_fnma4_elt_to_128df): Likewise.
1884 (*aarch64_fnma4_elt_to_64v2df): Likewise.
1885 (<frint_pattern><mode>2): Likewise.
1886 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): Likewise.
1887 (<optab><fcvt_target><VDQF:VDQF:mode>2): Likewise.
1888 (vec_unpacks_lo_v4sf): Likewise.
1889 (aarch64_float_extend_lo_v2df): Likewise.
1890 (vec_unpacks_hi_v4sf): Likewise.
1891 (aarch64_float_truncate_lo_v2sf): Likewise.
1892 (aarch64_float_truncate_hi_v4sf): Likewise.
1893 (aarch64_vmls<mode>): Likewise.
1894 (<su><maxmin><mode>3): Likewise.
1895 (<maxmin_uns><mode>3): Likewise.
1896 (reduc_<sur>plus_<mode>): Likewise.
1897 (reduc_<sur>plus_v2di): Likewise.
1898 (reduc_<sur>plus_v2si): Likewise.
1899 (reduc_<sur>plus_<mode>): Likewise.
1900 (aarch64_addpv4sf): Likewise.
1901 (clz<mode>2): Likewise.
1902 (reduc_<maxmin_uns>_<mode>): Likewise.
1903 (reduc_<maxmin_uns>_v2di): Likewise.
1904 (reduc_<maxmin_uns>_v2si): Likewise.
1905 (reduc_<maxmin_uns>_<mode>): Likewise.
1906 (reduc_<maxmin_uns>_v4sf): Likewise.
1907 (aarch64_simd_bsl<mode>_internal): Likewise.
1908 (*aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): Likewise.
1909 (*aarch64_get_lane_zero_extendsi<mode>): Likewise.
1910 (aarch64_get_lane<mode>): Likewise.
1911 (*aarch64_combinez<mode>): Likewise.
1912 (aarch64_combine<mode>): Likewise.
1913 (aarch64_simd_combine<mode>): Likewise.
1914 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): Likewise.
1915 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): Likewise.
1916 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>): Likewise.
1917 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w<mode>): Likewise.
1918 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w2<mode>_internal): Likewise.
1919 (aarch64_<sur>h<addsub><mode>): Likewise.
1920 (aarch64_<sur><addsub>hn<mode>): Likewise.
1921 (aarch64_<sur><addsub>hn2<mode>): Likewise.
1922 (aarch64_pmul<mode>): Likewise.
1923 (aarch64_<su_optab><optab><mode>): Likewise.
1924 (aarch64_<sur>qadd<mode>): Likewise.
1925 (aarch64_sqmovun<mode>): Likewise.
1926 (aarch64_<sur>qmovn<mode>): Likewise.
1927 (aarch64_s<optab><mode>): Likewise.
1928 (aarch64_sq<r>dmulh<mode>): Likewise.
1929 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
1930 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
1931 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
1932 (aarch64_sqdml<SBINQOPS:as>l<mode>): Likewise.
1933 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
1934 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
1935 (aarch64_sqdml<SBINQOPS:as>l_n<mode>): Likewise.
1936 (aarch64_sqdml<SBINQOPS:as>l2<mode>_internal): Likewise.
1937 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
1938 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
1939 (aarch64_sqdmull<mode>): Likewise.
1940 (aarch64_sqdmull_lane<mode>_internal): Likewise.
1941 (aarch64_sqdmull_n<mode>): Likewise.
1942 (aarch64_sqdmull2<mode>_internal): Likewise.
1943 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
1944 (aarch64_sqdmull2_n<mode>_internal): Likewise.
1945 (aarch64_<sur>shl<mode>): Likewise.
1946 (aarch64_<sur>q<r>shl<mode>
1947 (aarch64_<sur>shll_n<mode>): Likewise.
1948 (aarch64_<sur>shll2_n<mode>): Likewise.
1949 (aarch64_<sur>shr_n<mode>): Likewise.
1950 (aarch64_<sur>sra_n<mode>): Likewise.
1951 (aarch64_<sur>s<lr>i_n<mode>): Likewise.
1952 (aarch64_<sur>qshl<u>_n<mode>): Likewise.
1953 (aarch64_<sur>q<r>shr<u>n_n<mode>): Likewise.
1954 (aarch64_cm<optab><mode>): Likewise.
1955 (aarch64_cm<optab>di): Likewise.
1956 (aarch64_cm<optab><mode>): Likewise.
1957 (aarch64_cm<optab>di): Likewise.
1958 (aarch64_cmtst<mode>): Likewise.
1959 (aarch64_cmtstdi): Likewise.
1960 (aarch64_cm<optab><mode>): Likewise.
1961 (*aarch64_fac<optab><mode>): Likewise.
1962 (aarch64_addp<mode>): Likewise.
1963 (aarch64_addpdi): Likewise.
1964 (sqrt<mode>2): Likewise.
1965 (vec_load_lanesoi<mode>): Likewise.
1966 (vec_store_lanesoi<mode>): Likewise.
1967 (vec_load_lanesci<mode>): Likewise.
1968 (vec_store_lanesci<mode>): Likewise.
1969 (vec_load_lanesxi<mode>): Likewise.
1970 (vec_store_lanesxi<mode>): Likewise.
1971 (*aarch64_mov<mode>): Likewise.
1972 (aarch64_ld2<mode>_dreg): Likewise.
1973 (aarch64_ld2<mode>_dreg): Likewise.
1974 (aarch64_ld3<mode>_dreg): Likewise.
1975 (aarch64_ld3<mode>_dreg): Likewise.
1976 (aarch64_ld4<mode>_dreg): Likewise.
1977 (aarch64_ld4<mode>_dreg): Likewise.
1978 (aarch64_tbl1<mode>): Likewise.
1979 (aarch64_tbl2v16qi): Likewise.
1980 (aarch64_combinev16qi): Likewise.
1981 (aarch64_<PERMUTE:perm_insn><PERMUTE:perm_hilo><mode>): Likewise.
1982 (aarch64_st2<mode>_dreg): Likewise.
1983 (aarch64_st2<mode>_dreg): Likewise.
1984 (aarch64_st3<mode>_dreg): Likewise.
1985 (aarch64_st3<mode>_dreg): Likewise.
1986 (aarch64_st4<mode>_dreg): Likewise.
1987 (aarch64_st4<mode>_dreg): Likewise.
1988 (*aarch64_simd_ld1r<mode>): Likewise.
1989 (aarch64_frecpe<mode>): Likewise.
1990 (aarch64_frecp<FRECP:frecp_suffix><mode>): Likewise.
1991 (aarch64_frecps<mode>): Likewise.
1993 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1995 * config/arm/iterators.md (V_elem_ch): New.
1997 (VQH_type): Likewise.
1998 * config/arm/arm.md (is_neon_type): New.
1999 (conds): Use is_neon_type.
2000 (anddi3_insn): Update type attribute.
2001 (xordi3_insn): Likewise.
2002 (one_cmpldi2): Likewise.
2003 * gcc/config/arm/vfp.md (movhf_vfp_neon): Update type attribute.
2004 * gcc/config/arm/neon.md (neon_mov): Update type attribute.
2005 (*movmisalign<mode>_neon_store): Likewise.
2006 (*movmisalign<mode>_neon_load): Likewise.
2007 (vec_set<mode>_internal): Likewise.
2008 (vec_set<mode>_internal): Likewise.
2009 (vec_setv2di_internal): Likewise.
2010 (vec_extract<mode>): Likewise.
2011 (vec_extract<mode>): Likewise.
2012 (vec_extractv2di): Likewise.
2013 (*add<mode>3_neon): Likewise.
2014 (adddi3_neon): Likewise.
2015 (*sub<mode>3_neon): Likewise.
2016 (subdi3_neon): Likewise.
2017 (fma<VCVTF:mode>4): Likewise.
2018 (fma<VCVTF:mode>4_intrinsic): Likewise.
2019 (*fmsub<VCVTF:mode>4): Likewise.
2020 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
2021 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
2022 (ior<mode>3): Likewise.
2023 (and<mode>3): Likewise.
2024 (orn<mode>3_neon): Likewise.
2025 (orndi3_neon): Likewise.
2026 (bic<mode>3_neon): Likewise.
2027 (bicdi3_neon): Likewise.
2028 (xor<mode>3): Likewise.
2029 (one_cmpl<mode>2): Likewise.
2030 (abs<mode>2): Likewise.
2031 (neg<mode>2): Likewise.
2032 (negdi2_neon): Likewise.
2033 (*umin<mode>3_neon): Likewise.
2034 (*umax<mode>3_neon): Likewise.
2035 (*smin<mode>3_neon): Likewise.
2036 (*smax<mode>3_neon): Likewise.
2037 (vashl<mode>3): Likewise.
2038 (vashr<mode>3_imm): Likewise.
2039 (vlshr<mode>3_imm): Likewise.
2040 (ashl<mode>3_signed): Likewise.
2041 (ashl<mode>3_unsigned): Likewise.
2042 (neon_load_count): Likewise.
2043 (ashldi3_neon_noclobber): Likewise.
2044 (ashldi3_neon): Likewise.
2045 (signed_shift_di3_neon): Likewise.
2046 (unsigned_shift_di3_neon): Likewise.
2047 (ashrdi3_neon_imm_noclobber): Likewise.
2048 (lshrdi3_neon_imm_noclobber): Likewise.
2049 (<shift>di3_neon): Likewise.
2050 (widen_ssum<mode>3): Likewise.
2051 (widen_usum<mode>3): Likewise.
2052 (quad_halves_<code>v4si): Likewise.
2053 (quad_halves_<code>v4sf): Likewise.
2054 (quad_halves_<code>v8hi): Likewise.
2055 (quad_halves_<code>v16qi): Likewise.
2056 (reduc_splus_v2di): Likewise.
2057 (neon_vpadd_internal<mode>): Likewise.
2058 (neon_vpsmin<mode>): Likewise.
2059 (neon_vpsmax<mode>): Likewise.
2060 (neon_vpumin<mode>): Likewise.
2061 (neon_vpumax<mode>): Likewise.
2062 (*ss_add<mode>_neon): Likewise.
2063 (*us_add<mode>_neon): Likewise.
2064 (*ss_sub<mode>_neon): Likewise.
2065 (*us_sub<mode>_neon): Likewise.
2066 (neon_vadd<mode>_unspec): Likewise.
2067 (neon_vaddl<mode>): Likewise.
2068 (neon_vaddw<mode>): Likewise.
2069 (neon_vhadd<mode>): Likewise.
2070 (neon_vqadd<mode>): Likewise.
2071 (neon_vaddhn<mode>): Likewise.
2072 (neon_vmul<mode>): Likewise.
2073 (neon_vfms<VCVTF:mode>): Likewise.
2074 (neon_vmlal<mode>): Likewise.
2075 (neon_vmls<mode>): Likewise.
2076 (neon_vmlsl<mode>): Likewise.
2077 (neon_vqdmulh<mode>): Likewise.
2078 (neon_vqdmlal<mode>): Likewise.
2079 (neon_vqdmlsl<mode>): Likewise.
2080 (neon_vmull<mode>): Likewise.
2081 (neon_vqdmull<mode>): Likewise.
2082 (neon_vsub<mode>_unspec): Likewise.
2083 (neon_vsubl<mode>): Likewise.
2084 (neon_vsubw<mode>): Likewise.
2085 (neon_vqsub<mode>): Likewise.
2086 (neon_vhsub<mode>): Likewise.
2087 (neon_vsubhn<mode>): Likewise.
2088 (neon_vceq<mode>): Likewise.
2089 (neon_vcge<mode>): Likewise.
2090 (neon_vcgeu<mode>): Likewise.
2091 (neon_vcgt<mode>): Likewise.
2092 (neon_vcgtu<mode>): Likewise.
2093 (neon_vcle<mode>): Likewise.
2094 (neon_vclt<mode>): Likewise.
2095 (neon_vcage<mode>): Likewise.
2096 (neon_vcagt<mode>): Likewise.
2097 (neon_vtst<mode>): Likewise.
2098 (neon_vabd<mode>): Likewise.
2099 (neon_vabdl<mode>): Likewise.
2100 (neon_vaba<mode>): Likewise.
2101 (neon_vabal<mode>): Likewise.
2102 (neon_vmax<mode>): Likewise.
2103 (neon_vmin<mode>): Likewise.
2104 (neon_vpaddl<mode>): Likewise.
2105 (neon_vpadal<mode>): Likewise.
2106 (neon_vpmax<mode>): Likewise.
2107 (neon_vpmin<mode>): Likewise.
2108 (neon_vrecps<mode>): Likewise.
2109 (neon_vrsqrts<mode>): Likewise.
2110 (neon_vqabs<mode>): Likewise.
2111 (neon_vqneg<mode>): Likewise.
2112 (neon_vcls<mode>): Likewise.
2113 (clz<mode>2): Likewise.
2114 (popcount<mode>2): Likewise.
2115 (neon_vrecpe<mode>): Likewise.
2116 (neon_vrsqrte<mode>): Likewise.
2117 (neon_vget_lane<mode>_sext_internal): Likewise.
2118 (neon_vget_lane<mode>_zext_internal): Likewise.
2119 (neon_vdup_n<mode>): Likewise.
2120 (neon_vdup_n<mode>): Likewise.
2121 (neon_vdup_nv2di): Likewise.
2122 (neon_vdup_lane<mode>_interal): Likewise.
2123 (*neon_vswp<mode>): Likewise.
2124 (neon_vcombine<mode>): Likewise.
2125 (float<mode><V_cvtto>2): Likewise.
2126 (floatuns<mode><V_cvtto>2): Likewise.
2127 (fix_trunc<mode><V_cvtto>2): Likewise.
2128 (fixuns_trunc<mode><V_cvtto>2
2129 (neon_vcvt<mode>): Likewise.
2130 (neon_vcvt<mode>): Likewise.
2131 (neon_vcvtv4sfv4hf): Likewise.
2132 (neon_vcvtv4hfv4sf): Likewise.
2133 (neon_vcvt_n<mode>): Likewise.
2134 (neon_vcvt_n<mode>): Likewise.
2135 (neon_vmovn<mode>): Likewise.
2136 (neon_vqmovn<mode>): Likewise.
2137 (neon_vqmovun<mode>): Likewise.
2138 (neon_vmovl<mode>): Likewise.
2139 (neon_vmul_lane<mode>): Likewise.
2140 (neon_vmul_lane<mode>): Likewise.
2141 (neon_vmull_lane<mode>): Likewise.
2142 (neon_vqdmull_lane<mode>): Likewise.
2143 (neon_vqdmulh_lane<mode>): Likewise.
2144 (neon_vqdmulh_lane<mode>): Likewise.
2145 (neon_vmla_lane<mode>): Likewise.
2146 (neon_vmla_lane<mode>): Likewise.
2147 (neon_vmlal_lane<mode>): Likewise.
2148 (neon_vqdmlal_lane<mode>): Likewise.
2149 (neon_vmls_lane<mode>): Likewise.
2150 (neon_vmls_lane<mode>): Likewise.
2151 (neon_vmlsl_lane<mode>): Likewise.
2152 (neon_vqdmlsl_lane<mode>): Likewise.
2153 (neon_vext<mode>): Likewise.
2154 (neon_vrev64<mode>): Likewise.
2155 (neon_vrev32<mode>): Likewise.
2156 (neon_vrev16<mode>): Likewise.
2157 (neon_vbsl<mode>_internal): Likewise.
2158 (neon_vshl<mode>): Likewise.
2159 (neon_vqshl<mode>): Likewise.
2160 (neon_vshr_n<mode>): Likewise.
2161 (neon_vshrn_n<mode>): Likewise.
2162 (neon_vqshrn_n<mode>): Likewise.
2163 (neon_vqshrun_n<mode>): Likewise.
2164 (neon_vshl_n<mode>): Likewise.
2165 (neon_vqshl_n<mode>): Likewise.
2166 (neon_vqshlu_n<mode>): Likewise.
2167 (neon_vshll_n<mode>): Likewise.
2168 (neon_vsra_n<mode>): Likewise.
2169 (neon_vsri_n<mode>): Likewise.
2170 (neon_vsli_n<mode>): Likewise.
2171 (neon_vtbl1v8qi): Likewise.
2172 (neon_vtbl2v8qi): Likewise.
2173 (neon_vtbl3v8qi): Likewise.
2174 (neon_vtbl4v8qi): Likewise.
2175 (neon_vtbl1v16qi): Likewise.
2176 (neon_vtbl2v16qi): Likewise.
2177 (neon_vcombinev16qi): Likewise.
2178 (neon_vtbx1v8qi): Likewise.
2179 (neon_vtbx2v8qi): Likewise.
2180 (neon_vtbx3v8qi): Likewise.
2181 (neon_vtbx4v8qi): Likewise.
2182 (*neon_vtrn<mode>_insn): Likewise.
2183 (*neon_vzip<mode>_insn): Likewise.
2184 (*neon_vuzp<mode>_insn): Likewise.
2185 (neon_vld1<mode>): Likewise.
2186 (neon_vld1_lane<mode>): Likewise.
2187 (neon_vld1_lane<mode>): Likewise.
2188 (neon_vld1_dup<mode>): Likewise.
2189 (neon_vld1_dup<mode>): Likewise.
2190 (neon_vld1_dupv2di): Likewise.
2191 (neon_vst1<mode>): Likewise.
2192 (neon_vst1_lane<mode>): Likewise.
2193 (neon_vst1_lane<mode>): Likewise.
2194 (neon_vld2<mode>): Likewise.
2195 (neon_vld2<mode>): Likewise.
2196 (neon_vld2_lane<mode>): Likewise.
2197 (neon_vld2_lane<mode>): Likewise.
2198 (neon_vld2_dup<mode>): Likewise.
2199 (neon_vst2<mode>): Likewise.
2200 (neon_vst2<mode>): Likewise.
2201 (neon_vst2_lane<mode>): Likewise.
2202 (neon_vst2_lane<mode>): Likewise.
2203 (neon_vld3<mode>): Likewise.
2204 (neon_vld3qa<mode>): Likewise.
2205 (neon_vld3qb<mode>): Likewise.
2206 (neon_vld3_lane<mode>): Likewise.
2207 (neon_vld3_lane<mode>): Likewise.
2208 (neon_vld3_dup<mode>): Likewise.
2209 (neon_vst3<mode>): Likewise.
2210 (neon_vst3qa<mode>): Likewise.
2211 (neon_vst3qb<mode>): Likewise.
2212 (neon_vst3_lane<mode>): Likewise.
2213 (neon_vst3_lane<mode>): Likewise.
2214 (neon_vld4<mode>): Likewise.
2215 (neon_vld4qa<mode>): Likewise.
2216 (neon_vld4qb<mode>): Likewise.
2217 (neon_vld4_lane<mode>): Likewise.
2218 (neon_vld4_lane<mode>): Likewise.
2219 (neon_vld4_dup<mode>): Likewise.
2220 (neon_vst4<mode>): Likewise.
2221 (neon_vst4qa<mode>): Likewise.
2222 (neon_vst4qb<mode>): Likewise.
2223 (neon_vst4_lane<mode>): Likewise.
2224 (neon_vst4_lane<mode>): Likewise.
2225 (neon_vec_unpack<US>_lo_<mode>): Likewise.
2226 (neon_vec_unpack<US>_hi_<mode>): Likewise.
2227 (neon_vec_<US>mult_lo_<mode>): Likewise.
2228 (neon_vec_<US>mult_hi_<mode>): Likewise.
2229 (neon_vec_<US>shiftl_<mode>): Likewise.
2230 (neon_unpack<US>_<mode>): Likewise.
2231 (neon_vec_<US>mult_<mode>): Likewise.
2232 (vec_pack_trunc_<mode>): Likewise.
2233 (neon_vec_pack_trunc_<mode>): Likewise.
2234 (neon_vabd<mode>_2): Likewise.
2235 (neon_vabd<mode>_3): Likewise.
2237 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
2239 * config/aarch64/aarch64.md (movtf_aarch64): Update type attribute.
2240 (load_pair): Update type attribute.
2241 (store_pair): Update type attribute.
2242 * config/aarch64/iterators.md (q): New.
2244 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
2246 * config/arm/types.md: Add new types for Neon insns.
2248 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
2249 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2250 Sergey Lega <sergey.s.lega@intel.com>
2251 Anna Tikhonova <anna.tikhonova@intel.com>
2252 Ilya Tocar <ilya.tocar@intel.com>
2253 Andrey Turetskiy <andrey.turetskiy@intel.com>
2254 Ilya Verbin <ilya.verbin@intel.com>
2255 Kirill Yukhin <kirill.yukhin@intel.com>
2256 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2258 * config/i386/sse.md (unspec): Add UNSPEC_RCP14, UNSPEC_RSQRT14,
2259 UNSPEC_FIXUPIMM, UNSPEC_SCALEF, UNSPEC_GETEXP, UNSPEC_GETMANT,
2260 UNSPEC_EXP2, UNSPEC_RCP28, UNSPEC_RSQRT28.
2262 (srcp14<mode>): Ditto.
2263 (rsqrt14<mode>): Ditto.
2264 (rsqrt14<mode>): Ditto.
2265 (avx512f_vmscalef<mode>): Ditto.
2266 (avx512f_scalef<mode>): Ditto.
2267 (avx512f_getexp<mode>): Ditto.
2268 (avx512f_sgetexp<mode>): Ditto.
2269 (avx512f_fixupimm<mode>): Ditto.
2270 (avx512f_sfixupimm<mode>): Ditto.
2271 (avx512f_rndscale<mode>): Ditto.
2272 (*avx512er_exp2<mode>): Ditto.
2273 (*avx512er_rcp28<mode>): Ditto.
2274 (avx512er_rsqrt28<mode>): Ditto.
2275 (avx512f_getmant<mode>): Ditto.
2276 (avx512f_getmant<mode>): Ditto.
2277 (avx512f_rndscale<mode>): Fix formatting.
2279 2013-10-15 Martin Jambor <mjambor@suse.cz>
2281 * ipa-utils.h (ipa_edge_within_scc): Declare.
2282 * ipa-cp.c (edge_within_scc): Moved...
2283 * ipa-utils.c (ipa_edge_within_scc): ...here. Updated all callers.
2285 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
2286 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2287 Sergey Lega <sergey.s.lega@intel.com>
2288 Anna Tikhonova <anna.tikhonova@intel.com>
2289 Ilya Tocar <ilya.tocar@intel.com>
2290 Andrey Turetskiy <andrey.turetskiy@intel.com>
2291 Ilya Verbin <ilya.verbin@intel.com>
2292 Kirill Yukhin <kirill.yukhin@intel.com>
2293 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2295 * config/i386/predicates.md (const_8_to_15_operand): New.
2296 (const_16_to_31_operand): Ditto.
2297 * config/i386/sse.md (V8FI): New.
2299 (reduc_splus_v8df): Ditto.
2300 (reduc_splus_v16sf): Ditto.
2301 (avx512f_vextract<shuffletype>32x4_1): Ditto.
2302 (vec_extract_hi_<mode>): Ditto.
2303 (avx512f_vinsert<shuffletype>32x4_1): Ditto.
2304 (vec_set_lo_<mode>): Ditto.
2305 (vec_set_hi_<mode>): Ditto.
2306 (avx512f_shuf_<shuffletype>64x2_1): Ditto.
2307 (avx512f_shuf_<shuffletype>32x4_1): Ditto.
2308 (avx512f_pshufd_1): Ditto.
2309 (avx512f_broadcast<mode>): Ditto.
2310 (avx512f_broadcast<mode>): Ditto.
2311 (define_split): Split vec_extract_lo into move.
2312 (ssequartermode): Ditto.
2313 (ssedoublemode): Extened with wider modes.
2314 (vec_extract_lo_<mode>): Ditto.
2316 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
2317 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2318 Sergey Lega <sergey.s.lega@intel.com>
2319 Anna Tikhonova <anna.tikhonova@intel.com>
2320 Ilya Tocar <ilya.tocar@intel.com>
2321 Andrey Turetskiy <andrey.turetskiy@intel.com>
2322 Ilya Verbin <ilya.verbin@intel.com>
2323 Kirill Yukhin <kirill.yukhin@intel.com>
2324 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2326 * config/i386/predicates.md (register_or_constm1_operand): New.
2327 * config/i386/sse.md (unspec): Add UNSPEC_UNSIGNED_PCMP, UNSPEC_TESTM,
2328 UNSPEC_TESTNM, UNSPEC_VTERNLOG, UNSPEC_ALIGN, UNSPEC_CONFLICT,
2329 UNSPEC_MASKED_EQ, UNSPEC_MASKED_GT, UNSPEC_GATHER_PREFETCH,
2330 UNSPEC_SCATTER_PREFETCH
2332 (avx512f_ucmp<mode>3): Ditto.
2333 (avx512f_vternlog<mode>): Ditto.
2334 (avx512f_align<mode>): Ditto.
2335 (<shift_insn><mode>3): Ditto.
2336 (avx512f_<rotate>v<mode>): Ditto.
2337 (avx512f_<rotate><mode>): Ditto.
2338 (avx512f_eq<mode>3): Ditto.
2339 (avx512f_eq<mode>3_1): Ditto.
2340 (avx512f_gt<mode>3): Ditto.
2341 (avx512f_testm<mode>3): Ditto.
2342 (avx512f_testnm<mode>3): Ditto.
2343 (avx512pf_gatherpf<mode>): Ditto.
2344 (*avx512pf_gatherpf<mode>_mask): Ditto.
2345 (*avx512pf_gatherpf<mode>): Ditto.
2346 (avx512pf_scatterpf<mode>): Ditto.
2347 (*avx512pf_scatterpf<mode>_mask): Ditto.
2348 (*avx512pf_scatterpf<mode>): Ditto.
2349 (avx512f_vec_dup_gpr<mode>): Ditto.
2350 (clz<mode>2): Ditto.
2351 (conflict<mode>): Ditto.
2352 (REDUC_SMINMAX_MODE): Extened with wider modes.
2353 (reduc_<code>_<mode>): Ditto.
2354 (vlshr<mode>3): Ditto.
2355 (vashl<mode>3): Ditto.
2357 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
2358 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2359 Sergey Lega <sergey.s.lega@intel.com>
2360 Anna Tikhonova <anna.tikhonova@intel.com>
2361 Ilya Tocar <ilya.tocar@intel.com>
2362 Andrey Turetskiy <andrey.turetskiy@intel.com>
2363 Ilya Verbin <ilya.verbin@intel.com>
2364 Kirill Yukhin <kirill.yukhin@intel.com>
2365 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2367 * config/i386/sse.md (unspec): Added UNSPEC_VPERMI2, UNSPEC_VPERMT2,
2370 (avx512fmaskmode): Ditto.
2371 (bcstscalarsuff): Ditto.
2372 (avx512f_blendm<mode>): Ditto.
2373 (cmp_imm_predicate): Ditto.
2374 (avx512f_cmp<mode>3): Ditto.
2375 (avx512f_vec_dup<mode>): Ditto.
2376 (avx512f_vec_dup_mem<mode>): Ditto.
2377 (avx512f_vpermi2var<mode>3): Ditto.
2378 (avx512f_vpermt2var<mode>3): Ditto.
2379 (vec_init<mode>): Ditto.
2380 (avx512f_gathersi<mode>): Ditto.
2381 (*avx512f_gathersi<mode>): Ditto.
2382 (*avx512f_gathersi<mode>_2): Ditto.
2383 (avx512f_gatherdi<mode>): Ditto.
2384 (*avx512f_gatherdi<mode>): Ditto.
2385 (*avx512f_gatherdi<mode>_2): Ditto.
2386 (avx512f_scattersi<mode>): Ditto.
2387 (*avx512f_scattersi<mode>): Ditto.
2388 (avx512f_scatterdi<mode>): Ditto.
2389 (*avx512f_scatterdi<mode>): Ditto.
2390 (sseintprefix): Extened with wider modes.
2391 (VEC_GATHER_IDXSI): Ditto.
2392 (VEC_GATHER_IDXDI): Ditto.
2393 (VEC_GATHER_SRCDI): Ditto.
2395 2013-10-15 Matthew Gretton-Dann <matthew.gretton-dann@arm.com>
2396 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2398 * config/arm/t-aprofile: New file.
2399 * config.gcc: Handle --with-multilib-list option.
2401 2013-10-15 Bernd Schmidt <bernds@codesourcery.com>
2403 * reload1.c (reloads_unique_chain_p): Ensure that r1 is
2406 2013-10-15 Richard Biener <rguenther@suse.de>
2408 * tree-loop-distribution.c (build_empty_rdg): Inline into
2410 (rdg_flag_vertex): Inline into single user.
2411 (rdg_flag_vertex_and_dependent): Likewise.
2412 (build_rdg_partition_for_vertex): Remove processed bitmap.
2413 (rdg_build_partitions): Simplify.
2415 2013-10-15 Richard Biener <rguenther@suse.de>
2417 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1):
2418 Restructure forwarding through conversions and copies to
2419 avoid performing copy-propagation the wrong way. Adjust
2420 recursion invocations.
2421 (forward_propagate_addr_expr): Add argument stating if we
2422 are recursing from a single-use.
2423 (ssa_forward_propagate_and_combine): Adjust.
2425 2013-10-14 David Malcolm <dmalcolm@redhat.com>
2427 * dumpfile.h (gcc::dump_manager): New class, to hold state
2428 relating to dumpfile management.
2429 (get_dump_file_name): Remove in favor of method of dump_manager.
2430 (dump_initialized_p): Likewise.
2431 (dump_start): Likewise.
2432 (dump_finish): Likewise.
2433 (dump_switch_p): Likewise.
2434 (dump_register): Likewise.
2435 (get_dump_file_info): Likewise.
2436 * context.c (gcc::context::context): Construct the dump_manager
2438 * context.h (gcc::context::get_dumps): New.
2439 (gcc::context::m_dumps): New.
2440 * coverage.c (coverage_init): Port to dump_manager API.
2441 * dumpfile.c (extra_dump_files): Convert to field of gcc::dump_manager.
2442 (extra_dump_files_in_use): Likewise.
2443 (extra_dump_files_alloced): Likewise.
2444 (gcc::dump_manager::dump_manager): New.
2445 (dump_register): Convert to...
2446 (gcc::dump_manager::dump_register): ...method, replacing
2447 function-static next_dump with m_next_dump field.
2448 (get_dump_file_info): Convert to...
2449 (gcc::dump_manager::get_dump_file_info): ...method.
2450 (get_dump_file_name): Convert to...
2451 (gcc::dump_manager::get_dump_file_name): ...method.
2452 (dump_start): Convert to...
2453 (gcc::dump_manager::dump_start): ...method.
2454 (dump_finish): Convert to...
2455 (gcc::dump_manager::dump_finish): ...method.
2456 (dump_begin): Replace body with...
2457 (gcc::dump_manager::dump_begin): ...new method.
2458 (dump_phase_enabled_p): Convert to...
2459 (gcc::dump_manager::dump_phase_enabled_p): ...method.
2460 (dump_phase_enabled_p): Convert to...
2461 (gcc::dump_manager::dump_phase_enabled_p): ...method.
2462 (dump_initialized_p): Convert to...
2463 (gcc::dump_manager::dump_initialized_p): ...method.
2464 (dump_flag_name): Replace body with...
2465 (gcc::dump_manager::dump_flag_name): ...new method.
2466 (dump_enable_all): Convert to...
2467 (gcc::dump_manager::dump_enable_all): ...new method.
2468 (opt_info_enable_passes): Convert to...
2469 (gcc::dump_manager::opt_info_enable_passes): ...new method.
2470 (dump_switch_p_1): Convert to...
2471 (gcc::dump_manager::dump_switch_p_1): ...new method.
2472 (dump_switch_p): Convert to...
2473 (gcc::dump_manager::dump_switch_p): ...new method.
2474 (opt_info_switch_p): Port to dump_manager API.
2475 (enable_rtl_dump_file): Likewise.
2476 * opts-global.c (handle_common_deferred_options): Port to new
2478 * passes.c (pass_manager::finish_optimization_passes): Likewise.
2479 (pass_manager::register_one_dump_file): Likewise.
2480 (pass_manager::register_pass): Likewise.
2481 (pass_init_dump_file): Likewise.
2482 (pass_fini_dump_file): Likewise.
2483 * statistics.c (statistics_early_init): Likewise.
2485 2013-10-14 Richard Biener <rguenther@suse.de>
2487 * gimple.c (gimple_canonical_types, canonical_type_hash_cache,
2488 iterative_hash_canonical_type, gimple_canonical_type_hash,
2489 gimple_canonical_types_compatible_p, gimple_canonical_type_eq,
2490 gimple_register_canonical_type, print_gimple_types_stats,
2491 free_gimple_type_tables): Move to lto/lto.c
2492 (gt-gimple.h): Do not include.
2493 * gimple.h (gimple_register_canonical_type,
2494 print_gimple_types_stats, free_gimple_type_tables): Remove.
2495 * Makefile.in (GTFILES): Remove gimple.c.
2497 2013-10-14 Travis Snoozy <quandary@remstate.com>
2500 * config/msp430/msp430.c (msp430_option_override): Correct thinko
2501 scanning for msp430x targets.
2503 2013-10-14 Eric Botcazou <ebotcazou@adacore.com>
2506 * config/sparc/sparc-protos.h (widen_mem_for_ldd_peep): Declare.
2507 (registers_ok_for_ldd_peep): Move around.
2508 * config/sparc/sparc.c (widen_mem_for_ldd_peep): New.
2509 * config/sparc/sparc.md (widening peepholes): Use it.
2511 2013-10-14 Richard Biener <rguenther@suse.de>
2515 * gimple.c (iterative_hash_canonical_type): Make sure to
2516 record the hash into the correct hashtable slot.
2518 2013-10-13 Eric Botcazou <ebotcazou@adacore.com>
2520 PR rtl-optimization/58662
2521 * combine.c (try_combine): Take into account death nodes on I2 when
2522 splitting a PARALLEL of two independent SETs. Fix dump message.
2524 2013-10-12 Oleg Endo <olegendo@gcc.gnu.org>
2527 * config/sh/sh_treg_combine.cc: New SH specific RTL pass.
2528 * config.gcc (SH extra_objs): Add sh_ifcvt.o.
2529 * config/sh/t-sh (sh_treg_combine.o): New entry.
2530 * config/sh/sh.c (sh_fixed_condition_code_regs): New function that
2531 implements the target hook TARGET_FIXED_CONDITION_CODE_REGS.
2532 (register_sh_passes): New function. Register sh_treg_combine pass.
2533 (sh_option_override): Invoke it.
2534 (sh_canonicalize_comparison): Handle op0_preserve_value.
2535 * sh.md (*cbranch_t"): Do not try to optimize missed test and branch
2536 opportunities. Canonicalize branch condition.
2537 (nott): Allow only if pseudos can be created for non-SH2A.
2539 2013-10-12 H.J. Lu <hongjiu.lu@intel.com>
2542 * config/i386/i386.c (ix86_copy_addr_to_reg): New function.
2543 (ix86_expand_movmem): Replace copy_addr_to_reg with
2544 ix86_copy_addr_to_reg.
2545 (ix86_expand_setmem): Likewise.
2547 2013-10-12 Alexander Monakov <amonakov@ispras.ru>
2549 * config/i386/i386.c (ix86_expand_sse_compare_and_jump): Use mode
2550 provided by ix86_fp_compare_mode instead of CCFPUmode.
2552 2013-10-12 James Greenhalgh <james.greenhalgh@arm.com>
2554 * config/aarch64/arm_neon.h
2555 (vtbx<1,3>_<psu>8): Fix register constriants.
2557 2013-10-11 Jeff Law <law@redhat.com>
2559 PR tree-optimization/58640
2560 * tree-ssa-threadupdate.c (mark_threaded_blocks): Truncate jump
2561 threading paths that cross over two loop entry points.
2563 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2565 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): Generalize to
2566 handle vector float as well.
2567 (*vsx_le_perm_load_v4si): Likewise.
2568 (*vsx_le_perm_store_v2di): Likewise.
2569 (*vsx_le_perm_store_v4si): Likewise.
2571 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2573 * config/rs6000/vector.md (vec_realign_load<mode>): Generate vperm
2574 directly to circumvent subtract from splat{31} workaround.
2575 * config/rs6000/rs6000-protos.h (altivec_expand_vec_perm_le): New
2577 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): New.
2578 * config/rs6000/altivec.md (define_c_enum "unspec"): Add
2579 UNSPEC_VPERM_X and UNSPEC_VPERM_UNS_X.
2580 (altivec_vperm_<mode>): Convert to define_insn_and_split to
2581 separate big and little endian logic.
2582 (*altivec_vperm_<mode>_internal): New define_insn.
2583 (altivec_vperm_<mode>_uns): Convert to define_insn_and_split to
2584 separate big and little endian logic.
2585 (*altivec_vperm_<mode>_uns_internal): New define_insn.
2586 (vec_permv16qi): Add little endian logic.
2588 2013-10-11 Marc Glisse <marc.glisse@inria.fr>
2590 * doc/extend.texi (returns_nonnull): Remove arguments.
2592 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2593 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2594 Sergey Lega <sergey.s.lega@intel.com>
2595 Anna Tikhonova <anna.tikhonova@intel.com>
2596 Ilya Tocar <ilya.tocar@intel.com>
2597 Andrey Turetskiy <andrey.turetskiy@intel.com>
2598 Ilya Verbin <ilya.verbin@intel.com>
2599 Kirill Yukhin <kirill.yukhin@intel.com>
2600 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2602 * config/i386/sse.md (VI48F_256_512): New.
2603 (avx2_permvar<mode>): Change to ...
2604 (<avx2_avx512f>_permvar<mode>): This.
2606 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2607 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2608 Sergey Lega <sergey.s.lega@intel.com>
2609 Anna Tikhonova <anna.tikhonova@intel.com>
2610 Ilya Tocar <ilya.tocar@intel.com>
2611 Andrey Turetskiy <andrey.turetskiy@intel.com>
2612 Ilya Verbin <ilya.verbin@intel.com>
2613 Kirill Yukhin <kirill.yukhin@intel.com>
2614 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2616 * config/i386/i386.c (bdesc_args): Change corresponding pattern for
2617 __builtin_ia32_cvtps2dq, __builtin_ia32_cvtps2dq256.
2618 * config/i386/sse.md (VI4_AVX): New.
2619 (sf2simodelower): Ditto.
2620 (sse2_cvtps2dq): Change to ...
2621 (<sse2_avx_avx512f>_fix_notrunc<sf2simodelower><mode>): This.
2623 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2624 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2625 Sergey Lega <sergey.s.lega@intel.com>
2626 Anna Tikhonova <anna.tikhonova@intel.com>
2627 Ilya Tocar <ilya.tocar@intel.com>
2628 Andrey Turetskiy <andrey.turetskiy@intel.com>
2629 Ilya Verbin <ilya.verbin@intel.com>
2630 Kirill Yukhin <kirill.yukhin@intel.com>
2631 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2633 * config/i386/sse.md (V_512): New.
2635 (vcond<V_512:mode><VF_512:mode>): Ditto.
2636 (vcond<V_512:mode><VI_512:mode>): Ditto.
2637 (vcondu<V_512:mode><VI_512:mode>): Ditto.
2639 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2640 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2641 Sergey Lega <sergey.s.lega@intel.com>
2642 Anna Tikhonova <anna.tikhonova@intel.com>
2643 Ilya Tocar <ilya.tocar@intel.com>
2644 Andrey Turetskiy <andrey.turetskiy@intel.com>
2645 Ilya Verbin <ilya.verbin@intel.com>
2646 Kirill Yukhin <kirill.yukhin@intel.com>
2647 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2649 * config/i386/i386.c (ix86_rtx_costs): Enable fma for TARGET_AVX512F.
2650 * config/i386/sse.md (FMAMODEM): Changed modes and conditions.
2652 (fma<mode>4): Removed condition.
2653 (fms<mode>4): Ditto.
2654 (fnma<mode>4): Ditto.
2655 (fnms<mode>4): Ditto.
2656 (fma4i_fmadd_<mode>): Ditto.
2657 (*fma_fmadd_<mode>): Ditto.
2658 (*fma_fmsub_<mode>): Ditto.
2659 (*fma_fnmadd_<mode>): Ditto.
2660 (*fma_fnmsub_<mode>): Ditto.
2661 (fmaddsub_<mode>): Allow for TARGET_AVX512F.
2662 (*fma_fmaddsub_<mode>): Ditto.
2663 (*fma_fmsubadd_<mode>): Ditto.
2664 (*fmai_fmadd_<mode>): Ditto.
2665 (*fmai_fmsub_<mode>): Ditto.
2666 (*fmai_fnmadd_<mode>): Ditto.
2667 (*fmai_fnmsub_<mode>): Ditto.
2669 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2670 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2671 Sergey Lega <sergey.s.lega@intel.com>
2672 Anna Tikhonova <anna.tikhonova@intel.com>
2673 Ilya Tocar <ilya.tocar@intel.com>
2674 Andrey Turetskiy <andrey.turetskiy@intel.com>
2675 Ilya Verbin <ilya.verbin@intel.com>
2676 Kirill Yukhin <kirill.yukhin@intel.com>
2677 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2679 * config/i386/sse.md (VI248_AVX2_8_AVX512F): New.
2680 (VI124_256): Changed to ...
2681 (VI124_256_48_512): This.
2682 (ssepackmode): Extended with wider modes.
2683 (<code><mode>3): Changed iterator.
2684 (*avx2_<code><mode>3): Ditto.
2685 (vec_pack_trunc_<mode>): Ditto.
2687 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2688 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2689 Sergey Lega <sergey.s.lega@intel.com>
2690 Anna Tikhonova <anna.tikhonova@intel.com>
2691 Ilya Tocar <ilya.tocar@intel.com>
2692 Andrey Turetskiy <andrey.turetskiy@intel.com>
2693 Ilya Verbin <ilya.verbin@intel.com>
2694 Kirill Yukhin <kirill.yukhin@intel.com>
2695 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2697 * config/i386/sse.md (VI124_AVX2_48_AVX512F): New.
2698 (VI8F_256_512): Ditto.
2699 (abs<mode>2): Changed iterator.
2700 (avx2_perm<mode>): Changed to ...
2701 (<avx2_avx512f>_perm<mode>): This.
2702 (avx2_perm<mode>_1): Changed to ...
2703 (<avx2_avx512f>_perm<mode>_1): This.
2705 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2706 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2707 Sergey Lega <sergey.s.lega@intel.com>
2708 Anna Tikhonova <anna.tikhonova@intel.com>
2709 Ilya Tocar <ilya.tocar@intel.com>
2710 Andrey Turetskiy <andrey.turetskiy@intel.com>
2711 Ilya Verbin <ilya.verbin@intel.com>
2712 Kirill Yukhin <kirill.yukhin@intel.com>
2713 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2715 * config/i386/sse.md (VI48_AVX512F): New.
2716 (VI48_AVX2): Changed to ...
2717 (VI48_AVX2_48_AVX512F): This.
2718 (avx2_ashrv<mode>): Changed to ...
2719 (<avx2_avx512f>_ashrv<mode>): This.
2720 (avx2_<shift_insn>v<mode>): Changed to ...
2721 (<avx2_avx512f>_<shift_insn>v<mode>): This.
2723 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2724 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2725 Sergey Lega <sergey.s.lega@intel.com>
2726 Anna Tikhonova <anna.tikhonova@intel.com>
2727 Ilya Tocar <ilya.tocar@intel.com>
2728 Andrey Turetskiy <andrey.turetskiy@intel.com>
2729 Ilya Verbin <ilya.verbin@intel.com>
2730 Kirill Yukhin <kirill.yukhin@intel.com>
2731 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2733 * config/i386/sse.md (VI4_AVX512F): New.
2734 (VI8_AVX2_AVX512F): Ditto.
2735 (mul<mode>3): Extended with wider modes.
2736 (*<sse4_1_avx2>_mul<mode>3): Ditto.
2737 (mul<mode>3): Ditto.
2738 (vec_widen_<s>mult_odd_<mode>): Ditto.
2740 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2741 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2742 Sergey Lega <sergey.s.lega@intel.com>
2743 Anna Tikhonova <anna.tikhonova@intel.com>
2744 Ilya Tocar <ilya.tocar@intel.com>
2745 Andrey Turetskiy <andrey.turetskiy@intel.com>
2746 Ilya Verbin <ilya.verbin@intel.com>
2747 Kirill Yukhin <kirill.yukhin@intel.com>
2748 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2750 * config/i386/sse.md (VI2_AVX512F): New.
2751 (VI124_AVX512F): Ditto.
2752 (sseunpackmode): Extended with wider modes.
2753 (sseunpackfltmode): Ditto.
2754 (vec_unpacks_float_hi_<mode>): Ditto.
2755 (vec_unpacks_float_lo_<mode>): Ditto.
2756 (vec_unpacku_float_hi_<mode>): Ditto.
2757 (vec_unpacku_float_lo_<mode>): Ditto.
2758 (vec_unpacks_lo_<mode>): Ditto.
2759 (vec_unpacks_hi_<mode>): Ditto.
2760 (vec_unpacku_lo_<mode>): Ditto.
2761 (vec_unpacku_hi_<mode>): Ditto.
2763 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2764 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2765 Sergey Lega <sergey.s.lega@intel.com>
2766 Anna Tikhonova <anna.tikhonova@intel.com>
2767 Ilya Tocar <ilya.tocar@intel.com>
2768 Andrey Turetskiy <andrey.turetskiy@intel.com>
2769 Ilya Verbin <ilya.verbin@intel.com>
2770 Kirill Yukhin <kirill.yukhin@intel.com>
2771 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2773 * config/i386/i386.md (multdiv): New.
2774 (multdiv_mnemonic): Ditto.
2775 * config/i386/sse.md (<sse>_vmmul<mode>3): Changed to...
2776 (<sse>_vm<multdiv_mnemonic><mode>3): This.
2777 (<sse>_vmdiv<mode>3): Removed.
2779 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2780 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2781 Sergey Lega <sergey.s.lega@intel.com>
2782 Anna Tikhonova <anna.tikhonova@intel.com>
2783 Ilya Tocar <ilya.tocar@intel.com>
2784 Andrey Turetskiy <andrey.turetskiy@intel.com>
2785 Ilya Verbin <ilya.verbin@intel.com>
2786 Kirill Yukhin <kirill.yukhin@intel.com>
2787 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2789 * config/i386/sse.md (V): Extended with wider modes.
2791 (ssehalfvecmode): Ditto.
2793 (ssepackfltmode): Ditto.
2794 (avx_vec_concat<mode>): Ditto.
2795 (V_256_512): New iterator.
2796 (VF2_512_256): Ditto.
2797 (si2dfmode): New attribute.
2798 (si2dfmodelower): Ditto.
2800 (concat_tg_mode): Ditto.
2801 (floatv4siv4df2): Changed to ...
2802 (float<si2dfmodelower><mode>2): This.
2803 (avx_cvtps2pd256): Changed to ...
2804 (<sse2_avx_avx512f>_cvtps2pd<avxsizesuffix>): This.
2805 (vec_pack_trunc_v4df): Changed to ...
2806 (vec_pack_trunc_<mode>): This.
2807 (avx_vpermil<mode>): Changed to ...
2808 (<sse2_avx_avx512f>_vpermil<mode>): This.
2809 (<fixsuffix>fix_truncv8dfv8si2): New.
2810 (vec_pack_sfix_trunc_v8df): Ditto.
2811 (avx512f_rndscale<mode>): Ditto.
2812 (avx512f_roundpd512): Ditto.
2813 (vec_pack_ufix_trunc_<mode>): Updated iterator.
2815 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2816 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2817 Sergey Lega <sergey.s.lega@intel.com>
2818 Anna Tikhonova <anna.tikhonova@intel.com>
2819 Ilya Tocar <ilya.tocar@intel.com>
2820 Andrey Turetskiy <andrey.turetskiy@intel.com>
2821 Ilya Verbin <ilya.verbin@intel.com>
2822 Kirill Yukhin <kirill.yukhin@intel.com>
2823 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2825 * config/i386/i386.md (any_fix): New iterator.
2826 (fixsuffix): New attribute.
2827 * config/i386/sse.md (VF1): Extened with wider modes.
2831 (sseintvecmodelower): Ditto.
2832 (ssescalarmode): Ditto.
2833 (ssescalarnum): Ditto.
2835 (ssexmmmode): Ditto.
2836 (<fixsuffix>fix_truncv16sfv16si2): Ditto.
2837 (<sse>_rcp<mode>2): Change iterator.
2838 (rsqrt<mode>2): Ditto.
2839 (<sse>_rsqrt<mode>2): Ditto.
2840 (avx2_vec_dup<mode>): Ditto.
2841 (<sse4_1>_round<ssemodesuffix>_sfix<avxsizesuffix>): Ditto.
2842 (round<mode>2_sfix): Ditto.
2843 (avx2_pbroadcast<mode>): Ditto.
2844 (*andnot<mode>3): Handle XI mode.
2845 (*<code><mode>3): Ditto.
2846 (AVXTOSSEMODE): Removed.
2847 (avx_vpermil<mode>): Changed to ...
2848 (<sse2_avx_avx512f>_vpermil<mode>): This.
2850 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2851 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2852 Sergey Lega <sergey.s.lega@intel.com>
2853 Anna Tikhonova <anna.tikhonova@intel.com>
2854 Ilya Tocar <ilya.tocar@intel.com>
2855 Andrey Turetskiy <andrey.turetskiy@intel.com>
2856 Ilya Verbin <ilya.verbin@intel.com>
2857 Kirill Yukhin <kirill.yukhin@intel.com>
2858 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2860 * config/i386/sse.md (<sse2>_movnt<mode>): Update constraint to "v".
2861 (<sse>_comi): Ditto.
2862 (<sse>_ucomi): Ditto.
2863 (sse_cvtss2siq_2): Ditto.
2864 (sse2_cvtsd2si): Ditto.
2865 (sse2_cvtsd2siq): Ditto.
2866 (sse2_cvttsd2si): Ditto.
2867 (sse2_cvttsd2siq): Ditto.
2868 (<shift_insn><mode>3): Ditto.
2869 (sse2_cvtsi2sdq): Update constraint and prefix.
2870 (sse_cvtsi2ss): Update prefix.
2871 (sse_cvtsi2ssq): Ditto.
2873 2013-10-11 Jakub Jelinek <jakub@redhat.com>
2875 * tree-vrp.c (infer_nonnull_range): Use is_gimple_call,
2876 ignore internal calls.
2878 2013-10-11 Richard Biener <rguenther@suse.de>
2880 * tree-pretty-print.c (dump_generic_node): Allow to dump both (D)
2881 and (ab) for SSA_NAMEs. Mark INTEGER_CSTs with (OVF) if
2882 TREE_OVERFLOW is set.
2884 2013-10-11 Thomas Schwinge <thomas@codesourcery.com>
2886 * tree.h (OMP_CLAUSE_CODE): Remove duplicate definition.
2888 * gimple.c: GIMPLE statements have subcodes, not sub-codes.
2889 * gimple.h: Likewise.
2891 * doc/generic.texi (OpenMP): OMP_CLAUSE_* are subcodes, not sub-codes.
2893 * doc/generic.texi (Adding new DECL node types): Explain *_CHECK
2896 * doc/gimple.texi (is_gimple_omp): Move into the correct section.
2898 * acinclude.m4 (gcc_GAS_FLAGS): Add more gcc_cv_as_flags overrides.
2899 * configure: Regenerate.
2901 2013-10-11 Jakub Jelinek <jakub@redhat.com>
2903 * tree-pretty-print.c (dump_omp_clause): Handle OMP_CLAUSE__LOOPTEMP_
2904 and new OpenMP 4.0 clauses, handle UDR OMP_CLAUSE_REDUCTION,
2905 formatting fixes, use pp_colon instead of pp_character (..., ':'),
2906 similarly pp_right_paren.
2907 (dump_generic_node): Handle OMP_DISTRIBUTE, OMP_TEAMS,
2908 OMP_TARGET_DATA, OMP_TARGET, OMP_TARGET_UPDATE, OMP_TASKGROUP,
2909 allow OMP_FOR_INIT to be NULL, handle OMP_ATOMIC_SEQ_CST.
2910 * tree.c (omp_clause_num_ops, omp_clause_code_name): Add OpenMP 4.0
2912 (omp_declare_simd_clauses_equal,
2913 omp_remove_redundant_declare_simd_attrs): New functions.
2914 (attribute_value_equal): Use omp_declare_simd_clauses_equal.
2915 (walk_tree_1): Handle new OpenMP 4.0 clauses.
2916 * tree.h (OMP_LOOP_CHECK): Define.
2917 (OMP_FOR_BODY, OMP_FOR_CLAUSES, OMP_FOR_INIT, OMP_FOR_COND,
2918 OMP_FOR_INCR, OMP_FOR_PRE_BODY): Use it.
2919 (OMP_TASKGROUP_BODY, OMP_TEAMS_BODY, OMP_TEAMS_CLAUSES,
2920 OMP_TARGET_DATA_BODY, OMP_TARGET_DATA_CLAUSES, OMP_TARGET_BODY,
2921 OMP_TARGET_CLAUSES, OMP_TARGET_UPDATE_CLAUSES, OMP_CLAUSE_SIZE,
2922 OMP_ATOMIC_SEQ_CST, OMP_CLAUSE_DEPEND_KIND, OMP_CLAUSE_MAP_KIND,
2923 OMP_CLAUSE_MAP_ZERO_BIAS_ARRAY_SECTION, OMP_CLAUSE_PROC_BIND_KIND,
2924 OMP_CLAUSE_REDUCTION_OMP_ORIG_REF, OMP_CLAUSE_ALIGNED_ALIGNMENT,
2925 OMP_CLAUSE_NUM_TEAMS_EXPR, OMP_CLAUSE_THREAD_LIMIT_EXPR,
2926 OMP_CLAUSE_DEVICE_ID, OMP_CLAUSE_DIST_SCHEDULE_CHUNK_EXPR,
2927 OMP_CLAUSE_SIMDLEN_EXPR): Define.
2928 (OMP_CLAUSE_DECL): Change range up to OMP_CLAUSE__LOOPTEMP_.
2929 (omp_remove_redundant_declare_simd_attrs): New prototype.
2930 * gimple.def (GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET,
2931 GIMPLE_OMP_TEAMS): New codes.
2932 (GIMPLE_OMP_RETURN): Use GSS_OMP_ATOMIC_STORE instead of GSS_BASE.
2933 * omp-low.c (struct omp_context): Add cancel_label and cancellable
2935 (target_nesting_level): New variable.
2936 (extract_omp_for_data): Handle GF_OMP_FOR_KIND_DISTRIBUTE and
2937 OMP_CLAUSE_DIST_SCHEDULE. Don't fallback to library implementation
2938 for collapse > 1 static schedule unless ordered.
2939 (get_ws_args_for): Add par_stmt argument. Handle combined loops.
2940 (determine_parallel_type): Adjust get_ws_args_for caller.
2941 (install_var_field): Handle mask & 4 for double indirection.
2942 (scan_sharing_clauses): Ignore shared clause on teams construct.
2943 Handle OMP_CLAUSE__LOOPTEMP_ and new OpenMP 4.0 clauses.
2944 (create_omp_child_function): If inside target or declare target
2945 constructs, set "omp declare target" attribute on the child function.
2946 (find_combined_for): New function.
2947 (scan_omp_parallel): Handle combined loops.
2948 (scan_omp_target, scan_omp_teams): New functions.
2949 (check_omp_nesting_restrictions): Check new OpenMP 4.0 nesting
2950 restrictions and set ctx->cancellable for cancellable constructs.
2951 (scan_omp_1_stmt): Call check_omp_nesting_restrictions also on
2952 selected builtin calls. Handle GIMPLE_OMP_TASKGROUP,
2953 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS.
2954 (build_omp_barrier): Add lhs argument, return gimple rather than tree.
2955 (omp_clause_aligned_alignment): New function.
2956 (lower_rec_simd_input_clauses): Only call SET_DECL_VALUE_EXPR on decls.
2957 (lower_rec_input_clauses): Add FD argument. Ignore shared clauses
2958 on teams constructs. Handle user defined reductions and new
2960 (lower_reduction_clauses): Don't set placeholder to address of ref
2961 if it has already the right type.
2962 (lower_send_clauses): Handle OMP_CLAUSE__LOOPTEMP_.
2963 (expand_parallel_call): Use the new non-_start suffixed builtins,
2964 handle OMP_CLAUSE_PROC_BIND, don't call the outlined function
2965 and GOMP_parallel_end after the call.
2966 (expand_task_call): Handle OMP_CLAUSE_DEPEND.
2967 (expand_omp_for_init_counts): Handle combined loops.
2968 (expand_omp_for_init_vars): Add inner_stmt argument, handle combined
2970 (expand_omp_for_generic): Likewise. Use GOMP_loop_end_cancel at the
2971 end of cancellable loops.
2972 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
2973 Likewise. Handle collapse > 1 loops.
2974 (expand_omp_simd): Handle combined loops.
2975 (expand_omp_for): Add inner_stmt argument, adjust callers of
2976 expand_omp_for* functions, use expand_omp_for_static*chunk even
2977 for collapse > 1 unless ordered.
2978 (expand_omp_sections): Use GOMP_sections_end_cancel at the end
2979 of cancellable sections.
2980 (expand_omp_single): Remove need_barrier variable, just rely on
2981 gimple_omp_return_nowait_p. Adjust build_omp_barrier caller.
2982 (expand_omp_synch): Allow GIMPLE_OMP_TASKGROUP and GIMPLE_OMP_TEAMS.
2983 (expand_omp_atomic_load, expand_omp_atomic_store,
2984 expand_omp_atomic_fetch_op): Handle gimple_omp_atomic_seq_cst_p.
2985 (expand_omp_target): New function.
2986 (expand_omp): Handle combined loops. Handle GIMPLE_OMP_TASKGROUP,
2987 GIMPLE_OMP_TEAMS, GIMPLE_OMP_TARGET.
2988 (build_omp_regions_1): Immediately close region for
2989 GF_OMP_TARGET_KIND_UPDATE.
2990 (maybe_add_implicit_barrier_cancel): New function.
2991 (lower_omp_sections): Adjust lower_rec_input_clauses caller. Handle
2993 (lower_omp_single): Likewise. Add clobber after the barrier.
2994 (lower_omp_taskgroup): New function.
2995 (lower_omp_for): Handle combined loops. Adjust
2996 lower_rec_input_clauses caller. Handle cancellation.
2997 (lower_depend_clauses): New function.
2998 (lower_omp_taskreg): Lower depend clauses. Adjust
2999 lower_rec_input_clauses caller. Add clobber after the call. Handle
3001 (lower_omp_target, lower_omp_teams): New functions.
3002 (lower_omp_1): Handle cancellation. Handle GIMPLE_OMP_TASKGROUP,
3003 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GOMP_barrier, GOMP_cancel
3004 and GOMP_cancellation_point calls.
3005 (lower_omp): Fold stmts inside of target region.
3006 (diagnose_sb_1, diagnose_sb_2): Handle GIMPLE_OMP_TASKGROUP,
3007 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
3008 * builtin-types.def (DEF_FUNCTION_TYPE_8): Document.
3009 (BT_FN_VOID_OMPFN_PTR_UINT,
3010 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG,
3011 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG,
3012 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT): Remove.
3013 (BT_FN_VOID_OMPFN_PTR_UINT_UINT_UINT,
3014 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_UINT,
3015 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG_UINT,
3016 BT_FN_BOOL_INT, BT_FN_BOOL_INT_BOOL, BT_FN_VOID_UINT_UINT,
3017 BT_FN_VOID_INT_PTR_SIZE_PTR_PTR_PTR,
3018 BT_FN_VOID_INT_OMPFN_PTR_SIZE_PTR_PTR_PTR,
3019 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR): New.
3020 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1,
3021 call_may_clobber_ref_p_1): Handle BUILT_IN_GOMP_BARRIER_CANCEL,
3022 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_LOOP_END_CANCEL,
3023 BUILT_IN_GOMP_SECTIONS_END_CANCEL. Don't handle
3024 BUILT_IN_GOMP_PARALLEL_END.
3025 * gimple-low.c (lower_stmt): Handle GIMPLE_OMP_TASKGROUP,
3026 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
3027 * gimple-pretty-print.c (dump_gimple_omp_for): Handle
3028 GF_OMP_FOR_KIND_DISTRIBUTE.
3029 (dump_gimple_omp_target, dump_gimple_omp_teams): New functions.
3030 (dump_gimple_omp_block): Handle GIMPLE_OMP_TASKGROUP.
3031 (dump_gimple_omp_return): Print lhs if it has any.
3032 (dump_gimple_omp_atomic_load, dump_gimple_omp_atomic_store): Handle
3033 gimple_omp_atomic_seq_cst_p.
3034 (pp_gimple_stmt_1): Handle GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET
3035 and GIMPLE_OMP_TEAMS.
3036 * langhooks.c (lhd_omp_mappable_type): New function.
3037 * tree-vectorizer.c (struct simd_array_to_simduid): Fix up comment.
3038 * langhooks.h (struct lang_hooks_for_types): Add omp_mappable_type
3040 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_MAP,
3041 GOVD_ALIGNED and GOVD_MAP_TO_ONLY.
3042 (enum omp_region_type): Add ORT_TEAMS, ORT_TARGET_DATA and ORT_TARGET.
3043 (struct gimplify_omp_ctx): Add combined_loop field.
3044 (gimplify_call_expr, gimplify_modify_expr): Don't call fold_stmt
3045 on stmts inside of target region.
3046 (is_gimple_stmt): Return true for OMP_DISTRIBUTE and OMP_TASKGROUP.
3047 (omp_firstprivatize_variable): Handle GOVD_MAP, GOVD_ALIGNED,
3048 ORT_TARGET and ORT_TARGET_DATA.
3049 (omp_add_variable): Avoid checks on readding var for GOVD_ALIGNED.
3051 (omp_notice_threadprivate_variable): Complain about threadprivate
3052 variables in target region.
3053 (omp_notice_variable): Complain about vars with non-mappable type
3054 in target region. Handle ORT_TEAMS, ORT_TARGET and ORT_TARGET_DATA.
3055 (omp_check_private): Ignore ORT_TARGET* regions.
3056 (gimplify_scan_omp_clauses, gimplify_adjust_omp_clauses_1,
3057 gimplify_adjust_omp_clauses): Handle new OpenMP 4.0 clauses.
3058 (find_combined_omp_for): New function.
3059 (gimplify_omp_for): Handle gimplification of combined loops.
3060 (gimplify_omp_workshare): Gimplify also OMP_TARGET, OMP_TARGET_DATA,
3062 (gimplify_omp_target_update): New function.
3063 (gimplify_omp_atomic): Handle OMP_ATOMIC_SEQ_CST.
3064 (gimplify_expr): Handle OMP_DISTRIBUTE, OMP_TARGET, OMP_TARGET_DATA,
3065 OMP_TARGET_UPDATE, OMP_TEAMS, OMP_TASKGROUP.
3066 (gimplify_body): If fndecl has "omp declare target" attribute, add
3067 implicit ORT_TARGET context around it.
3068 * tree.def (OMP_DISTRIBUTE, OMP_TEAMS, OMP_TARGET_DATA, OMP_TARGET,
3069 OMP_TASKGROUP, OMP_TARGET_UPDATE): New tree codes.
3070 * tree-nested.c (convert_nonlocal_reference_stmt,
3071 convert_local_reference_stmt, convert_gimple_call): Handle
3072 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
3073 * omp-builtins.def (BUILT_IN_GOMP_TASK): Use
3074 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR
3075 instead of BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT.
3076 (BUILT_IN_GOMP_TARGET, BUILT_IN_GOMP_TARGET_DATA,
3077 BUILT_IN_GOMP_TARGET_END_DATA, BUILT_IN_GOMP_TARGET_UPDATE,
3078 BUILT_IN_GOMP_TEAMS, BUILT_IN_BARRIER_CANCEL,
3079 BUILT_IN_GOMP_LOOP_END_CANCEL,
3080 BUILT_IN_GOMP_SECTIONS_END_CANCEL, BUILT_IN_OMP_GET_TEAM_NUM,
3081 BUILT_IN_OMP_GET_NUM_TEAMS, BUILT_IN_GOMP_TASKGROUP_START,
3082 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_PARALLEL_LOOP_STATIC,
3083 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC,
3084 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED,
3085 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME, BUILT_IN_GOMP_PARALLEL,
3086 BUILT_IN_GOMP_PARALLEL_SECTIONS, BUILT_IN_GOMP_CANCEL,
3087 BUILT_IN_GOMP_CANCELLATION_POINT): New built-ins.
3088 (BUILT_IN_GOMP_PARALLEL_LOOP_STATIC_START,
3089 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC_START,
3090 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED_START,
3091 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME_START,
3092 BUILT_IN_GOMP_PARALLEL_START, BUILT_IN_GOMP_PARALLEL_END,
3093 BUILT_IN_GOMP_PARALLEL_SECTIONS_START): Remove.
3094 * tree-inline.c (remap_gimple_stmt, estimate_num_insns):
3095 Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
3096 * gimple.c (gimple_build_omp_taskgroup, gimple_build_omp_target,
3097 gimple_build_omp_teams): New functions.
3098 (walk_gimple_op): Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and
3099 GIMPLE_OMP_TASKGROUP. Walk optional lhs on GIMPLE_OMP_RETURN.
3100 (walk_gimple_stmt, gimple_copy): Handle GIMPLE_OMP_TARGET,
3101 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
3102 * gimple.h (enum gf_mask): GF_OMP_FOR_KIND_DISTRIBUTE,
3103 GF_OMP_FOR_COMBINED, GF_OMP_FOR_COMBINED_INTO,
3104 GF_OMP_TARGET_KIND_MASK, GF_OMP_TARGET_KIND_REGION,
3105 GF_OMP_TARGET_KIND_DATA, GF_OMP_TARGET_KIND_UPDATE,
3106 GF_OMP_ATOMIC_SEQ_CST): New.
3107 (gimple_build_omp_taskgroup, gimple_build_omp_target,
3108 gimple_build_omp_teams): New prototypes.
3109 (gimple_has_substatements): Handle GIMPLE_OMP_TARGET,
3110 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
3111 (gimple_omp_subcode): Use GIMPLE_OMP_TEAMS instead of
3112 GIMPLE_OMP_SINGLE as end of range.
3113 (gimple_omp_return_set_lhs, gimple_omp_return_lhs,
3114 gimple_omp_return_lhs_ptr, gimple_omp_atomic_seq_cst_p,
3115 gimple_omp_atomic_set_seq_cst, gimple_omp_for_combined_p,
3116 gimple_omp_for_set_combined_p, gimple_omp_for_combined_into_p,
3117 gimple_omp_for_set_combined_into_p, gimple_omp_target_clauses,
3118 gimple_omp_target_clauses_ptr, gimple_omp_target_set_clauses,
3119 gimple_omp_target_kind, gimple_omp_target_set_kind,
3120 gimple_omp_target_child_fn, gimple_omp_target_child_fn_ptr,
3121 gimple_omp_target_set_child_fn, gimple_omp_target_data_arg,
3122 gimple_omp_target_data_arg_ptr, gimple_omp_target_set_data_arg,
3123 gimple_omp_teams_clauses, gimple_omp_teams_clauses_ptr,
3124 gimple_omp_teams_set_clauses): New inlines.
3125 (CASE_GIMPLE_OMP): Add GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS
3126 and GIMPLE_OMP_TASKGROUP.
3127 * tree-core.h (enum omp_clause_code): Add new OpenMP 4.0 clause codes.
3128 (enum omp_clause_depend_kind, enum omp_clause_map_kind,
3129 enum omp_clause_proc_bind_kind): New.
3130 (union omp_clause_subcode): Add depend_kind, map_kind and
3131 proc_bind_kind fields.
3132 * tree-cfg.c (make_edges): Handle GIMPLE_OMP_TARGET,
3133 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
3134 * langhooks-def.h (lhd_omp_mappable_type): New prototype.
3135 (LANG_HOOKS_OMP_MAPPABLE_TYPE): Define.
3136 (LANG_HOOKS_FOR_TYPES_INITIALIZER): Add it.
3138 2013-10-10 Teresa Johnson <tejohnson@google.com>
3140 * predict.c (tree_estimate_probability): Add new parameter
3141 for estimate_bb_frequencies.
3142 (estimate_bb_frequencies): Add new parameter to force estimation.
3143 (rebuild_frequencies): When max frequency in function is small,
3144 recompute counts from frequencies.
3145 * predict.h (estimate_bb_frequencies): New parameter.
3147 2013-10-10 David Malcolm <dmalcolm@redhat.com>
3149 * ipa-inline.c (ipa_inline): Fix leak of "order" when
3150 optimizations are disabled.
3152 2013-10-10 David Malcolm <dmalcolm@redhat.com>
3154 * coverage.c (coverage_finish): Fix leak of da_file_name.
3156 2013-10-10 Jan Hubicka <jh@suse.cz>
3158 * config/i386/x86-tune.def: Enable X86_TUNE_SSE_TYPELESS_STORES
3159 for generic, enable X86_TUNE_SSE_LOAD0_BY_PXOR for Bulldozer,
3162 2013-10-10 Jakub Jelinek <jakub@redhat.com>
3165 * stmt.c (expand_asm_operands): Add FALLTHRU_BB argument,
3166 if any labels are in FALLTHRU_BB, use a special label emitted
3167 immediately after the asm goto insn rather than label_rtx
3169 (expand_asm_stmt): Adjust caller.
3170 * cfgrtl.c (commit_one_edge_insertion): Force splitting of
3171 edge if the last insn in predecessor is a jump with single successor,
3172 but it isn't simplejump_p.
3174 2013-10-10 Richard Biener <rguenther@suse.de>
3176 PR tree-optimization/58656
3177 * tree-ssa-pre.c (phi_translate): Do not cache failed translations.
3179 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
3181 * gimplify.c: Include expr.h and tm_p.h for targets with special
3182 va-arg padding requirements.
3184 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
3186 * tree-flow.h: Move some prototypes to gimple.h.
3187 (gimple_fold_indirect_ref): Move prototype to gimple-fold.h.
3188 * gimple.h: Relocate some prototypes from tree-flow.h
3189 * builtins.c (std_gimplify_va_arg_expr, build_va_arg_indirect_ref):
3191 * gimplify.c (gimple_fold_indirect_ref): Move to gimple-fold.c.
3192 (build_va_arg_indirect_ref): Relocate and make static.
3193 (std_gimplify_va_arg_expr): Relocate here.
3194 * gimple-fold.c (gimple_fold_indirect_ref): Relocate here.
3195 * gimple-fold.h (gimple_fold_indirect_ref): Add prototype.
3197 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3199 * doc/md.texi: Document the mnemonic attribute.
3201 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3204 * gensupport.c (gen_mnemonic_attr): Handle (set (attr x) y) and
3205 (set_attr_alternative x ...) when searching for user defined
3208 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
3210 * config/aplha/alpha.c: Add gimple-ssa.h to include list.
3212 2013-10-09 Easwaran Raman <eraman@google.com>
3214 * params.def (PARAM_MIN_SIZE_FOR_STACK_SHARING): New param...
3215 * cfgexpand.c (defer_stack_allocation): ...use here
3216 * doc/invoke.texi: Add documentation for min-size-for-stack-sharing.
3218 2013-10-09 Zhenqiang Chen <zhenqiang.chen@arm.com>
3220 * tree-ssa-phiopts.c (rhs_is_fed_for_value_replacement): New function.
3221 (operand_equal_for_value_replacement): New function, extracted from
3222 value_replacement and enhanced to catch more cases.
3223 (value_replacement): Use operand_equal_for_value_replacement.
3225 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
3227 * loop-doloop.c (doloop_modify, doloop_optimize): Use
3228 get_max_loop_iterations.
3230 2013-10-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3232 * config/arm/aarch-common.c (arm_early_load_addr_dep):
3233 Place comment above function.
3235 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
3237 * tree-flow.h: Remove all remaining prototypes, enums and structs that
3238 are not related to tree-cfg.c.
3239 * tree-ssa-address.h: New file. Relocate prototypes.
3240 * tree-ssa-address.c: (struct mem_address): Relocate from tree-flow.h.
3241 (addr_for_mem_ref): New. Combine call to get_address_description and
3242 return addr_for_mem_ref.
3243 * expr.c (expand_expr_real_1): Use new addr_for_mem_ref routine.
3244 * tree-ssa-live.h: Adjust prototypes.
3245 * passes.c: Include tree-ssa-live.h.
3246 * gimple-pretty-print.h (gimple_dump_bb): Add prototype.
3247 * graphite.c (graphite_transform_loops): Make static.
3248 (graphite_transforms, gate_graphite_transforms, pass_data_graphite,
3249 make_pass_graphite, pass_data_graphite_transforms,
3250 make_pass_graphite_transforms): Relocate here from tree-ssa-loop.c.
3251 * ipa-pure-const.c (warn_function_noreturn): Make static.
3252 (execute_warn_function_noreturn, gate_warn_function_noreturn,
3253 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
3254 Relocate from tree-cfg.c
3255 * tree-cfg.c (tree_node_can_be_shared, gimple_empty_block_p): Make
3257 (execute_warn_function_noreturn, gate_warn_function_noreturn,
3258 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
3259 Move to ipa-pure-const.c.
3260 (execute_fixup_cfg, class pass_fixup_cfg, make_pass_fixup_cfg):
3261 Relocate from tree-optimize.c.
3262 * tree-optimize.c (execute_fixup_cfg, class pass_fixup_cfg,
3263 make_pass_fixup_cfg): Move to tree-cfg.c.
3264 * tree-chrec.h: (enum ev_direction): Relocate here from tree-flow.h.
3265 Relocate some prototypes.
3266 * tree-data-ref.h (tree_check_data_deps) Add prototype.
3267 * tree-dump.c (dump_function_to_file): Remove prototype.
3268 Add tree-flow.h to the include file.
3269 * tree-dump.h: Remove prototype.
3270 * tree-parloops.h: New File. Add prototypes.
3271 * tree-parloops.c (gate_tree_parallelize_loops, tree_parallelize_loops,
3272 pass_data_parallelize_loops, make_pass_parallelize_loops): Relocate
3273 from tree-ssa-loop.c.
3274 * tree-predcom.c (run_tree_predictive_commoning,
3275 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom):
3276 Relocate here from tree-ssa-loop.c.
3277 * tree-ssa-dom.c (tree_ssa_dominator_optimize) Don't call
3278 ssa_name_values.release ().
3279 * tree-ssa-threadedge.h: New File. Relocate prototypes here.
3280 (ssa_name_values): Relocate from tree-flow.h.
3281 * tree-ssa.h: Include tree-ssa-threadedge.h and tree-ssa-address.h.
3282 * tree-ssa-loop.c (run_tree_predictive_commoning,
3283 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom,
3284 graphite_transforms, gate_graphite_transforms, pass_data_graphite,
3285 make_pass_graphite, pass_data_graphite_transforms,
3286 make_pass_graphite_transforms, gate_tree_parallelize_loops,
3287 tree_parallelize_loops, pass_data_parallelize_loops,
3288 make_pass_parallelize_loops): Move to other files.
3289 * tree-vectorizer.h (lpeel_tree_duplicate_loop_to_edge_cfg): Prototype
3291 * tree.h: Remove prototypes from tree-address.c.
3293 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
3295 * tree-flow.h (tm_restart_node, gimple_df): Move to gimple-ssa.h.
3296 (struct int_tree_map): Move to tree-hasher.h
3297 (SCALE, LABEL, PERCENT): Move to gimple.h
3298 * tree-flow-inline.h: Delete. Move functions to other files.
3299 (unmodifiable_var_p, ref_contains_array_ref): Unused, so delete.
3300 * gimple-ssa.h (tm_restart_node, gimple_df): Relocate from tree-flow.h.
3301 (gimple_in_ssa_p, gimple_vop): Relocate from tree-flow-inline.h
3302 * gimple.h (imple_stmt_max_uid, set_gimple_stmt_max_uid,
3303 inc_gimple_stmt_max_uid, get_lineno): Relocate from tree-flow-inline.h.
3304 (SCALE, LABEL, PERCENT): Relocate from tree-flow.h
3305 * tree-hasher.h: Don't include tree-flow.h.
3306 (struct int_tree_map): Relocate from tree-flow.h.
3307 * tree-sra.c (contains_view_convert_expr_p): Relocate from
3308 tree-flow-inline.h and make static.
3309 * tree-ssa-alias.h (ranges_overlap_p): Relocate from
3311 * tree-ssa-operands.c (gimple_ssa_operands): Relocate from
3312 tree-flow-inline.h and make static.
3313 * tree.h (is_global_var, may_be_aliased): Relocate from
3315 * Makefile.in (GTFILES): Remove tree-flow.h and add gimple-ssa.h.
3316 * value-prof.c: No longer include tree-flow-inline.h.
3317 * tree-switch-conversion.c: No longer include tree-flow-inline.h.
3319 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
3321 * tree-flow.h: Move some protoypes. Include new tree-ssa-loop.h.
3322 (struct affine_iv, struct tree_niter_desc): Move to tree-ssa-loop.h.
3323 (enum move_pos): Move to tree-ssa-loop-im.h
3324 * cfgloop.h: Move some prototypes.
3325 (gcov_type_to_double_int): relocate from tree-ssa-loop.niter.c.
3326 * tree-flow-inline.h (loop_containing_stmt): Move to tree-ssa-loop.h.
3327 * tree-ssa-loop.h: New File. Include other tree-ssa-loop-*.h files.
3328 (struct affine_iv, struct tree_niter_desc): Relocate from tree-flow.h.
3329 (loop_containing_stmt): Relocate from tree-flow-inline.h.
3330 * tree-ssa-loop-ch.c: (do_while_loop_p): Make static.
3331 * tree-ssa-loop-im.c (for_each_index): Move to tree-ssa-loop.c.
3332 (enum move_pos): Relocate here.
3333 (lsm_tmp_name_add, gen_lsm_tmp_name, get_lsm_tmp_name): Move to
3335 (execute_sm_if_changed_flag_set): Change get_lsm_tmp_name call.
3336 (tree_ssa_loop_im, gate_tree_ssa_loop_im, pass_data_lim,
3337 make_pass_lim): Relocate here from tree-ssa-loop.c.
3338 * tree-ssa-loop-ivcanon.c (tree_num_loop_insns): Move to
3340 (loop_edge_to_cancel, unloop_loops): Make static.
3341 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
3342 make_pass_iv_canon): Relocate from tree-ssa-loop.c.
3343 (tree_complete_unroll, gate_tree_complete_unroll,
3344 pass_data_complete_unroll, make_pass_complete_unroll): Relocate here.
3345 (tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
3346 pass_data_complete_unrolli, make_pass_complete_unrolli): Relocate here.
3347 * tree-ssa-loop-ivopts.c: Remove local prototypes.
3348 (stmt_invariant_in_loop_p): Remove unused function.
3349 * tree-ssa-loop-ivopts.h: New file. Add prototypes.
3350 * tree-ssa-loop-manip.h: New file. Add prototypes.
3351 * tree-ssa-loop-niter.c (record_niter_bound): Move to cfgloop.c.
3352 (gcov_type_to_double_int): Move to cfgloop.h.
3353 (double_int_cmp, bound_index,
3354 estimate_numbers_of_iterations_loop): Make static.
3355 (estimated_loop_iterations): Factor out get_estimated_loop_iterations.
3356 (max_loop_iterations): Factor out get_max_loop_iterations.
3357 (estimated_loop_iterations_int, max_stmt_executions_int): Move to
3359 * tree-ssa-loop-niter.h: New file. Add prototypes.
3360 * tree-ssa-loop-prefetch.c (tree_ssa_loop_prefetch,
3361 gate_tree_ssa_loop_prefetch, pass_data_loop_prefetch,
3362 make_pass_loop_prefetch): Relocate from tree-ssa-loop.c.
3363 * tree-ssa-loop-unswitch.c (tree_ssa_loop_unswitch,
3364 gate_tree_ssa_loop_unswitch, pass_data_tree_unswitch,
3365 make_pass_tree_unswitch): Relocate from tree-ssa-loop.c.
3366 * tree-ssa-loop.c (tree_ssa_loop_im, gate_tree_ssa_loop_im,
3367 pass_data_lim, make_pass_lim): Move to tree-ssa-loop-im.c.
3368 (tree_ssa_loop_unswitch, gate_tree_ssa_loop_unswitch,
3369 pass_data_tree_unswitch, make_pass_tree_unswitch): Move.
3370 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
3371 make_pass_iv_canon, tree_complete_unroll, gate_tree_complete_unroll,
3372 pass_data_complete_unroll, make_pass_complete_unroll,
3373 tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
3374 pass_data_complete_unrolli, make_pass_complete_unrolli): Move to
3375 tree-ssa-loop-ivcanon.c.
3376 (tree_ssa_loop_prefetch, gate_tree_ssa_loop_prefetch,
3377 pass_data_loop_prefetch, make_pass_loop_prefetch): Move to
3378 tree-ssa-loop-prefetch.c.
3379 (for_each_index, lsm_tmp_name_add, gen_lsm_tmp_name): Relocate from
3381 (get_lsm_tmp_name): Relocate and add suffix parameter.
3382 (tree_num_loop_insns): Relocate from tree-ssa-ivcanon.c.
3383 * tree-scalar-evolution.h (simple_iv): Don't use affive_iv typedef.
3384 * cfgloop.c (record_niter_bound, estimated_loop_iterations_int,
3385 max_stmt_executions_int): Move from tree-ssa-loop-niter.c.
3386 (get_estimated_loop_iterations): Factor out accessor from
3387 estimated_loop_iterations in tree-ssa-loop-niter.c.
3388 (get_max_loop_iterations): Factor out accessor from
3389 _max_loop_iterations in tree-ssa-niter.c.
3390 * loop-unroll.c (decide_unroll_constant_iterations,
3391 decide_unroll_runtime_iterations, decide_peel_simple,
3392 decide_unroll_stupid): Use new get_* accessors.
3394 2013-10-09 Marc Glisse <marc.glisse@inria.fr>
3396 PR tree-optimization/20318
3397 * doc/extend.texi (returns_nonnull): New function attribute.
3398 * fold-const.c (tree_expr_nonzero_warnv_p): Look for returns_nonnull
3400 * tree-vrp.c (gimple_stmt_nonzero_warnv_p): Likewise.
3401 (stmt_interesting_for_vrp): Accept all GIMPLE_CALL.
3403 2013-10-09 Eric Botcazou <ebotcazou@adacore.com>
3406 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): Return
3407 false if both components are bitfields.
3409 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
3411 * config/aarch64/arm_neon.h (vclz_s8, vclz_s16, vclz_s32)
3412 (vclzq_s8, vclzq_s16, vclzq_s32, vclz_u8, vclz_u16, vclz_u32)
3413 (vclzq_u8, vclzq_u16, vclzq_u32): Replace ASM with C.
3414 * config/aarch64/aarch64.h
3415 (CLZ_DEFINED_VALUE_AT_ZERO): Macro fixed for clz.
3416 * config/aarch64/aarch64-simd-builtins.def
3417 (VAR1 (UNOP, clz, 0, v4si)): Replaced with iterator.
3419 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
3421 * config/aarch64/arm_neon.h (vadd_f64, vsub_f64): Implementation added.
3423 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
3425 * config/aarch64/arm_neon.h (vdiv_f64): Added.
3427 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
3429 * config/aarch64/arm_neon.h (vneg_f32): Asm replaced with C.
3430 (vneg_f64): New intrinsic.
3431 (vneg_s8): Asm replaced with C.
3432 (vneg_s16): Likewise.
3433 (vneg_s32): Likewise.
3434 (vneg_s64): New intrinsic.
3435 (vnegq_f32): Asm replaced with C.
3436 (vnegq_f64): Likewise.
3437 (vnegq_s8): Likewise.
3438 (vnegq_s16): Likewise.
3439 (vnegq_s32): Likewise.
3440 (vnegq_s64): Likewise.
3442 2013-10-09 Renlin Li <Renlin.Li@arm.com>
3444 * config/arm/arm.c (arm_output_mi_thunk): Use plus_constant.
3446 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3448 * config/s390/s390.c (s390_register_info_stdarg_fpr): Remove
3449 packed stack special handling.
3450 (s390_frame_info, s390_emit_prologue, s390_emit_epilogue): Switch
3451 back to fixed stack slots for FPRs saved due to stdarg.
3453 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3455 * config/s390/s390.c (s390_frame_info): Restructure function.
3457 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3459 * config/s390/s390.c (struct s390_frame_layout): New field
3461 (cfun_save_arg_fprs_p, cfun_gpr_save_slot): New macros.
3462 (s390_reg_clobbered_rtx, s390_regs_ever_clobbered): Change type of
3463 regs_ever_clobbered to char*.
3464 (s390_regs_ever_clobbered): Check crtl->saves_all_registers instead
3465 of cfun->has_nonlocal_label. Ignore frame related restore INSNs.
3466 (s390_register_info): Enable FPR save slots. Move/Copy some
3467 functionality into ...
3468 (s390_register_info_gprtofpr, s390_register_info_stdarg_fpr)
3469 (s390_register_info_stdarg_gpr, s390_optimize_register_info): New
3471 (s390_frame_info): Do gpr slot allocation here now. stdarg does
3472 not imply a stack frame.
3473 (s390_init_frame_layout): Remove variable clobbered_regs.
3474 (s390_update_register_info): Remove function.
3475 (s390_hard_regno_rename_ok): Call-saved regs without a save slot
3476 cannot be used for register renaming.
3477 (s390_hard_regno_scratch_ok): New function.
3478 (TARGET_HARD_REGNO_SCRATCH_OK): Define target hook.
3479 (s390_initial_elimination_offset): Change offset calculation of
3480 the return address pointer.
3481 (save_gprs): Deal with only r6 being saved from the call-saved regs.
3482 (restore_gprs): Set frame related flag.
3483 (s390_save_gprs_to_fprs, s390_restore_gprs_from_fprs): New functions.
3484 (s390_emit_prologue): Call s390_register_info instead of
3485 s390_update_frame_layout. Call s390_save_gprs_to_fprs.
3486 (s390_emit_epilogue): Call s390_restore_gprs_from_fprs.
3487 (s390_optimize_prologue): Call s390_optimize_register_info.
3488 Try to remove also FPR slot save/restore INSNs. Remove frame
3489 related flags from restore INSNs.
3491 2013-10-08 DJ Delorie <dj@redhat.com>
3493 * config/rl78/rl78-expand.md (movqi): use operands[] not operandN.
3496 * config/rl78/rl78.c (rl78_print_operand_1): Change %c to %C to
3497 avoid conflict with the MI use of %c.
3498 * config/rl78/rl78-real.md: change %c to %C throughout.
3499 * config/rl78/rl78-virt.md: Likewise.
3501 2013-10-08 Jan Hubicka <jh@suse.cz>
3503 * config/i386/i386.c (ix86_option_override_internal): Switch
3504 to SSE math for -ffast-math when target ISA supports SSE2.
3506 2013-10-08 Andrew MacLeod <amacleod@redhat.com>
3508 * tree-flow.h: Remove some prototypes.
3509 * tree.h: Remove some protypes, add a couple.
3510 * tree.c (using_eh_for_cleanups_flag, using_eh_for_cleanups,
3511 using_eh_for_cleanups_p): Add interface routines for front ends.
3512 * tree-eh.h: New file. Add protoptyes.
3513 * tree-eh.c (using_eh_for_cleanups_p, using_eh_for_cleanups): Delete.
3514 (add_stmt_to_eh_lp_fn): Make static.
3515 (lower_try_finally): Use new using_eh_for_cleanups_p.
3516 * emit-rtl.c: Include tree-eh.h.
3517 * gimple.h: Include tree-eh.h.
3519 2013-10-08 Marc Glisse <marc.glisse@inria.fr>
3521 PR tree-optimization/58480
3522 * tree-vrp.c (infer_nonnull_range): New function.
3523 (infer_value_range): Call infer_nonnull_range.
3525 2013-10-08 Dehao Chen <dehao@google.com>
3527 PR tree-optimization/58619
3528 * tree-inline.c (copy_phis_for_bb): Combine location data
3531 2013-10-08 Zhenqiang Chen <zhenqiang.chen@linaro.org>
3534 * config/arm/arm.c (arm_emit_ldrd_pop): Attach
3535 RTX_FRAME_RELATED_P on INSN.
3537 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3539 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const_le): New.
3540 (altivec_expand_vec_perm_const): Call it.
3542 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3544 * config/rs6000/vector.md (mov<mode>): Emit permuted move
3545 sequences for LE VSX loads and stores at expand time.
3546 * config/rs6000/rs6000-protos.h (rs6000_emit_le_vsx_move): New
3548 * config/rs6000/rs6000.c (rs6000_const_vec): New.
3549 (rs6000_gen_le_vsx_permute): New.
3550 (rs6000_gen_le_vsx_load): New.
3551 (rs6000_gen_le_vsx_store): New.
3552 (rs6000_gen_le_vsx_move): New.
3553 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): New.
3554 (*vsx_le_perm_load_v4si): New.
3555 (*vsx_le_perm_load_v8hi): New.
3556 (*vsx_le_perm_load_v16qi): New.
3557 (*vsx_le_perm_store_v2di): New.
3558 (*vsx_le_perm_store_v4si): New.
3559 (*vsx_le_perm_store_v8hi): New.
3560 (*vsx_le_perm_store_v16qi): New.
3561 (*vsx_xxpermdi2_le_<mode>): New.
3562 (*vsx_xxpermdi4_le_<mode>): New.
3563 (*vsx_xxpermdi8_le_V8HI): New.
3564 (*vsx_xxpermdi16_le_V16QI): New.
3565 (*vsx_lxvd2x2_le_<mode>): New.
3566 (*vsx_lxvd2x4_le_<mode>): New.
3567 (*vsx_lxvd2x8_le_V8HI): New.
3568 (*vsx_lxvd2x16_le_V16QI): New.
3569 (*vsx_stxvd2x2_le_<mode>): New.
3570 (*vsx_stxvd2x4_le_<mode>): New.
3571 (*vsx_stxvd2x8_le_V8HI): New.
3572 (*vsx_stxvd2x16_le_V16QI): New.
3574 2013-10-07 Renlin Li <Renlin.Li@arm.com>
3576 * config/arm/arm-cores.def (cortex-a53): Use cortex tuning.
3578 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3580 * config/s390/s390.c (s390_register_info): Make the call-saved FPR
3581 loop to work also for 31bit ABI.
3582 Save the stack pointer for frame_size > 0.
3584 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3586 * config/s390/s390.md ("tbegin", "tbegin_nofloat", "tbegin_retry")
3587 ("tbegin_retry_nofloat", "tend", "tabort", "tx_assist"): Remove
3588 constraint letters from expanders.
3589 ("tbegin_retry", "tbegin_retry_nofloat"): Change predicate of the
3590 retry count to general_operand.
3591 ("tabort"): Give operand 0 a mode.
3592 ("tabort_1"): Add mode and constraint letter for operand 0.
3593 * doc/extend.texi: Fix protoype of __builtin_non_tx_store.
3595 2013-10-04 Jeff Law <law@redhat.com>
3597 * tree-ssa-threadedge.c: Fix some trailing whitespace problems.
3599 * tree-ssa-threadedge.c (thread_through_normal_block): Broken
3601 (thread_across_edge): Here. Call it.
3603 2013-10-04 Cary Coutant <ccoutant@google.com>
3605 * dwarf2out.c (dw_sra_loc_expr): Release addr_table entries when
3606 discarding a location list expression (or a piece of one).
3608 2013-10-03 Jan Hubicka <jh@suse.cz>
3610 * config/i386/i386.c (ix86_issue_rate): Pentium4, Nocona has issue
3611 rate of 2. Core2, Corei7 and Haswell has issue rate of 4.
3612 (ix86_adjust_cost): Remove Atom case; fix core2/corei7/Haswell case.
3614 2013-10-03 Jan Hubicka <jh@suse.cz>
3616 * config/i386/i386.c (ix86_option_override_internal): Do not enable
3617 accumulate-outgoing-args when producing unwind info.
3619 2013-10-03 Wei Mi <wmi@google.com>
3621 * lra-constraints.c (insert_move_for_subreg): New function
3622 extracted from simplify_operand_subreg.
3623 (simplify_operand_subreg): Add reload for paradoxical subreg.
3625 2013-10-03 Rong Xu <xur@google.com>
3627 * ipa-inline-analysis.c (find_foldable_builtin_expect): Find
3628 the candidate of builtin_expect such that we should fix the
3629 size/time estimation.
3630 (estimate_function_body_sizes): Do the acutally size/time fix-up
3633 2013-10-03 Rong Xu <xur@google.com>
3635 * predict.c (tree_predict_by_opcode): Get the probability
3636 for builtin_expect from param builtin_expect_probability.
3637 * params.def (BUILTIN_EXPECT_PROBABILITY): New parameter.
3638 * predict.def (PRED_BUILTIN_EXPECT_RELAXED): Fix comments.
3639 * doc/invoke.texi: Add documentation for builtin-expect-probability.
3641 2013-10-03 Marc Glisse <marc.glisse@inria.fr>
3644 * common.opt (fcheck-new): Moved from c.opt. Make it 'Common'.
3645 * calls.c (alloca_call_p): Use get_callee_fndecl.
3646 * fold-const.c (tree_expr_nonzero_warnv_p): Handle operator new.
3647 * tree-vrp.c (gimple_stmt_nonzero_warnv_p, stmt_interesting_for_vrp):
3649 (vrp_visit_stmt): Remove duplicated code.
3651 2013-10-03 Michael Meissner <meissner@linux.vnet.ibm.com>
3653 * config/rs6000/rs6000-builtin.def (XSRDPIM): Use floatdf2,
3654 ceildf2, btruncdf2, instead of vsx_* name.
3656 * config/rs6000/vsx.md (vsx_add<mode>3): Change arithmetic
3657 iterators to only do V2DF and V4SF here. Move the DF code to
3658 rs6000.md where it is combined with SF mode. Replace <VSv> with
3659 just 'v' since only vector operations are handled with these insns
3660 after moving the DF support to rs6000.md.
3661 (vsx_sub<mode>3): Likewise.
3662 (vsx_mul<mode>3): Likewise.
3663 (vsx_div<mode>3): Likewise.
3664 (vsx_fre<mode>2): Likewise.
3665 (vsx_neg<mode>2): Likewise.
3666 (vsx_abs<mode>2): Likewise.
3667 (vsx_nabs<mode>2): Likewise.
3668 (vsx_smax<mode>3): Likewise.
3669 (vsx_smin<mode>3): Likewise.
3670 (vsx_sqrt<mode>2): Likewise.
3671 (vsx_rsqrte<mode>2): Likewise.
3672 (vsx_fms<mode>4): Likewise.
3673 (vsx_nfma<mode>4): Likewise.
3674 (vsx_copysign<mode>3): Likewise.
3675 (vsx_btrunc<mode>2): Likewise.
3676 (vsx_floor<mode>2): Likewise.
3677 (vsx_ceil<mode>2): Likewise.
3678 (vsx_smaxsf3): Delete scalar ops that were moved to rs6000.md.
3679 (vsx_sminsf3): Likewise.
3680 (vsx_fmadf4): Likewise.
3681 (vsx_fmsdf4): Likewise.
3682 (vsx_nfmadf4): Likewise.
3683 (vsx_nfmsdf4): Likewise.
3684 (vsx_cmpdf_internal1): Likewise.
3686 * config/rs6000/rs6000.h (TARGET_SF_SPE): Define macros to make it
3687 simpler to select whether a target has SPE or traditional floating
3688 point support in iterators.
3689 (TARGET_DF_SPE): Likewise.
3690 (TARGET_SF_FPR): Likewise.
3691 (TARGET_DF_FPR): Likewise.
3692 (TARGET_SF_INSN): Macros to say whether floating point support
3693 exists for a given operation for expanders.
3694 (TARGET_DF_INSN): Likewise.
3696 * config/rs6000/rs6000.c (Ftrad): New mode attributes to allow
3697 combining of SF/DF mode operations, using both traditional and VSX
3705 (abs<mode>2): Combine SF/DF modes using traditional floating point
3706 instructions. Add support for using the upper DF registers with
3707 VSX support, and SF registers with power8-vector support. Update
3708 expanders for operations supported by both the SPE and traditional
3709 floating point units.
3710 (abs<mode>2_fpr): Likewise.
3711 (nabs<mode>2): Likewise.
3712 (nabs<mode>2_fpr): Likewise.
3713 (neg<mode>2): Likewise.
3714 (neg<mode>2_fpr): Likewise.
3715 (add<mode>3): Likewise.
3716 (add<mode>3_fpr): Likewise.
3717 (sub<mode>3): Likewise.
3718 (sub<mode>3_fpr): Likewise.
3719 (mul<mode>3): Likewise.
3720 (mul<mode>3_fpr): Likewise.
3721 (div<mode>3): Likewise.
3722 (div<mode>3_fpr): Likewise.
3723 (sqrt<mode>3): Likewise.
3724 (sqrt<mode>3_fpr): Likewise.
3725 (fre<Fs>): Likewise.
3726 (rsqrt<mode>2): Likewise.
3727 (cmp<mode>_fpr): Likewise.
3728 (smax<mode>3): Likewise.
3729 (smin<mode>3): Likewise.
3730 (smax<mode>3_vsx): Likewise.
3731 (smin<mode>3_vsx): Likewise.
3732 (negsf2): Delete SF operations that are merged with DF.
3739 (fmasf4_fpr): Likewise.
3740 (fmssf4_fpr): Likewise.
3741 (nfmasf4_fpr): Likewise.
3742 (nfmssf4_fpr): Likewise.
3743 (sqrtsf2): Likewise.
3744 (rsqrtsf_internal1): Likewise.
3745 (smaxsf3): Likewise.
3746 (sminsf3): Likewise.
3747 (cmpsf_internal1): Likewise.
3748 (copysign<mode>3_fcpsgn): Add VSX/power8-vector support.
3749 (negdf2): Delete DF operations that are merged with SF.
3751 (nabsdf2): Likewise.
3757 (rsqrtdf_internal1): Likewise.
3758 (fmadf4_fpr): Likewise.
3759 (fmsdf4_fpr): Likewise.
3760 (nfmadf4_fpr): Likewise.
3761 (nfmsdf4_fpr): Likewise.
3762 (sqrtdf2): Likewise.
3763 (smaxdf3): Likewise.
3764 (smindf3): Likewise.
3765 (cmpdf_internal1): Likewise.
3766 (lrint<mode>di2): Use TARGET_<MODE>_FPR macro.
3767 (btrunc<mode>2): Delete separate expander, and combine with the
3768 insn and add VSX instruction support. Use TARGET_<MODE>_FPR.
3769 (btrunc<mode>2_fpr): Likewise.
3770 (ceil<mode>2): Likewise.
3771 (ceil<mode>2_fpr): Likewise.
3772 (floor<mode>2): Likewise.
3773 (floor<mode>2_fpr): Likewise.
3774 (fma<mode>4_fpr): Combine SF and DF fused multiply/add support.
3775 Add support for using the upper registers with VSX and
3776 power8-vector. Move insns to be closer to the define_expands. On
3777 VSX systems, prefer the traditional form of FMA over the VSX
3778 version, since the traditional form allows the target not to
3779 overlap with the inputs.
3780 (fms<mode>4_fpr): Likewise.
3781 (nfma<mode>4_fpr): Likewise.
3782 (nfms<mode>4_fpr): Likewise.
3784 2013-10-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3785 Richard Earnshaw <richard.earnshaw@arm.com>
3787 * config/arm/aarch-common-protos.h (struct alu_cost_table): New.
3788 (struct mult_cost_table): Likewise.
3789 (struct mem_cost_table): Likewise.
3790 (struct fp_cost_table): Likewise.
3791 (struct vector_cost_table): Likewise.
3792 (cpu_cost_table): Likewise.
3793 * config/arm/arm.opt (mold-rts-costs): New option.
3794 (mnew-generic-costs): Likewise.
3795 * config/arm/arm.c (generic_extra_costs): New table.
3796 (cortexa15_extra_costs): Likewise.
3797 (arm_slowmul_tune): Use NULL as new costs.
3798 (arm_fastmul_tune): Likewise.
3799 (arm_strongarm_tune): Likewise.
3800 (arm_xscale_tune): Likewise.
3801 (arm_9e_tune): Likewise.
3802 (arm_v6t2_tune): Likewise.
3803 (arm_cortex_a5_tune): Likewise.
3804 (arm_cortex_a9_tune): Likewise.
3805 (arm_v6m_tune): Likewise.
3806 (arm_fa726te_tune): Likewise.
3807 (arm_cortex_a15_tune): Use cortex15_extra_costs.
3808 (arm_cortex_tune): Use generict_extra_costs.
3809 (shifter_op_p): New function.
3810 (arm_unspec_cost): Likewise.
3811 (LIBCALL_COST): Define.
3812 (arm_new_rtx_costs): New function.
3813 (arm_rtx_costs): Use arm_new_rtx_costs when core-specific
3814 table is available. Use old costs otherwise unless mnew-generic-costs
3816 * config/arm/arm-protos.h (tune_params): Add insn_extra_cost field.
3817 (cpu_cost_table): Declare.
3819 2013-10-03 Marcus Shawcroft <marcus.shawcroft@arm.com>
3822 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>)
3823 (*subs_mul_imm_<mode>)
3824 (*add_<shift>_<mode>, *add_<shift>_si_uxtw,*add_mul_imm_<mode>)
3825 (*sub_<shift>_<mode>)
3826 (*sub_<shift>_si_uxtw,*sub_mul_imm_<mode>, *sub_mul_imm_si_uxtw):
3827 Remove k constraint.
3829 2013-10-03 Ian Bolton <ian.bolton@arm.com>
3831 * config/aarch64/aarch64.c (aarch64_secondary_reload): Remove legacy
3833 * config/aarch64/aarch64.md (reload_sp_immediate): Likewise.
3835 2013-10-02 Teresa Johnson <tejohnson@google.com>
3837 * predict.c (probably_never_executed): New function.
3838 (probably_never_executed_bb_p): Invoke probably_never_executed.
3839 (probably_never_executed_edge_p): Ditto.
3840 * bb-reorder.c (find_rarely_executed_basic_blocks_and_crossing_edges):
3841 Treat profile insanities conservatively.
3843 2013-10-02 John David Anglin <danglin@gcc.gnu.org>
3845 * config.gcc (hppa*64*-*-linux*): Don't add pa/t-linux to tmake_file.
3847 2013-10-02 Vladimir Makarov <vmakarov@redhat.com>
3849 * lra-constraints.c (process_alt_operand): Calculate scratch_p and
3850 use it. Use smaller increase for scratch. Don't increase reject
3851 for early clobber scratch.
3852 * lra-eliminations.c (eliminate_regs_in_insn): Remove all insns
3853 setting eliminated regs except setting fp from hfp.
3854 (lra_eliminate): Check lra_insn_recog_data on NULL.
3856 2013-10-02 Michael Meissner <meissner@linux.vnet.ibm.com>
3859 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Turn off
3860 setting -mvsx-timode by default until the underlying problem is fixed.
3861 (RS6000_CPU, power7 defaults): Likewise.
3863 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
3865 * config/x-linux (host-linux.o): Remove header dependencies.
3866 Use $(COMPILE) and $(POSTCOMPILE).
3867 * config/t-linux-android (linux-android.o): Ditto.
3869 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
3871 * Makefile.in (expmed.o-warn): Remove.
3873 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
3875 * graphite-scop-detection.c: Include tree-ssa-propagate,h.
3876 * graphite-sese-to-poly.c: Include tree-ssa-propagate.h.
3878 2013-10-02 Teresa Johnson <tejohnson@google.com>
3880 * dojump.c (do_jump_1): Divide probability between
3881 both conditions of a TRUTH_ANDIF_EXPR/TRUTH_ORIF_EXPR.
3883 2013-10-02 Tom Tromey <tromey@redhat.com>
3885 * Makefile.in (DRIVER_DEFINES): Use $(if), not $(and).
3887 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
3889 * tree-flow.h: Remove some prototypes.
3890 * tree-ssa-dce.c (mark_virtual_operand_for_renaming,
3891 mark_virtual_phi_result_for_renaming): Move to tree-into-ssa.c.
3892 * tree-into-ssa.c (mark_virtual_operand_for_renaming,
3893 mark_virtual_phi_result_for_renaming): Relocate here.
3894 * tree-into-ssa.h: Add prototypes.
3895 * tree-ssa-phiopt.c: (tree_ssa_phiopt_worker) Use
3896 single_pred_before_succ_order.
3897 (blocks_in_phiopt_order): Rename and move to cfganal.c.
3898 (nonfreeing_call_p) Move to gimple.c.
3899 * cfganal.c (single_pred_before_succ_order): Move and renamed from
3901 * basic-block.h (single_pred_before_succ_order): Add prototype.
3902 * gimple.c (nonfreeing_call_p): Relocate here.
3903 * gimple.h: Add prototype.
3904 * tree-ssa-ifcombine.c: Include tree-ssa-phiopt.h.
3905 * tree-ssa-dom.h: New file. Relocate prototypes here.
3906 * tree-ssa.h: Include tree-ssa-dom.h.
3908 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
3910 * config/i386/x-i386 (driver-i386.o): Remove header dependencies.
3911 Use $(COMPILE) and $(POSTCOMPILE).
3913 * config/alpha/x-alpha (driver-alpha.o): Ditto.
3915 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
3917 * tree-flow.h: Remove some prototypes.
3918 * gimple-fold.h: Add prototypes from gimple.h and tree-flow.h.
3919 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
3920 * tree-ssa-copy.c (may_propagate*, propagate_value, replace_exp,
3921 propagate_tree_value*): Move from here to...
3922 * tree-ssa-propagate.c (may_propagate*, propagate_value, replace_exp,
3923 propagate_tree_value*): Relocate here.
3924 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
3925 * gimple.h: Include gimple-fold.h, move prototypes into gimple-fold.h.
3926 * gimple-fold.c: Remove gimple-fold.h from include list.
3927 * tree-vrp.c: Remove gimple-fold.h from include list.
3928 * tree-ssa-sccvn.c: Remove gimple-fold.h from include list.
3929 * tree-ssa-ccp.c: Remove gimple-fold.h from include list.
3930 * tree-scalar-evolution.c: Add tree-ssa-propagate.h to include list.
3931 * tree-ssa-pre.c: Add tree-ssa-propagate.h to include list.
3932 * sese.c: Add tree-ssa-propagate.h to include list.
3934 2013-10-02 Richard Biener <rguenther@suse.de>
3936 * tree-loop-distribution.c: Include tree-vectorizer.h for
3938 (enum partition_kind): Remove PKIND_REDUCTION.
3939 (struct partition_s): Remove has_writes member, add reduction_p member.
3940 (partition_alloc): Adjust.
3941 (partition_builtin_p): Likewise.
3942 (partition_has_writes): Remove.
3943 (partition_reduction_p): New function.
3944 (partition_merge_into): Likewise.
3945 (generate_code_for_partition): Commonize builtin partition
3947 (rdg_cannot_recompute_vertex_p): Remove.
3948 (already_processed_vertex_p): Likewise.
3949 (rdg_flag_vertex): Do not set has_writes.
3950 (classify_partition): Adjust.
3951 (rdg_build_partitions): Do not set has_writes, treat all
3952 partitions as useful.
3953 (distribute_loop): Record number of library calls generated. Adjust.
3954 (tree_loop_distribution): Report number of loops and library
3955 calls generated as opt-info.
3957 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
3959 * tree-flow.h: Include new .h files. Move prototypes.
3960 * tree-cfgcleanup.h: New file. Add prototypes from tree-flow.h.
3961 * tree-dfa.h: New File. Add prototypes from tree-flow.h.
3962 (get_addr_base_and_unit_offset_1) Move from tree-flow-inline.h.
3963 * tree-pretty-print.h: Add prototypes from tree-flow.h.
3964 * tree-into-ssa.h: New File. Add prototypes from tree-flow.h.
3965 ({debug|dump}*): Move debugging prototypes out of tree-into-ssa.c.
3966 * tree-into-ssa.c ({debug|dump}*): Move prototypes to header file.
3967 * tree.h (get_ref_base_and_extent): Move prototype out.
3968 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Move to
3970 * gimple-low.h: New File. Add prototypes from tree-flow.h.
3971 * gimple-low.c (try_catch_may_fallthru, block_may_fallthru): Move to...
3972 * tree.c (try_catch_may_fallthru, block_may_fallthru): Here.
3973 * tree-scalar-evolution.c: Include tree.h.
3974 * sese.c: Include tree.h.
3975 * dumpfile.c: Move gimple-pretty-print.h include after tree.h.
3976 * dwarf2out.c: Include tree-dfa.h.
3977 * tree-chrec.c: Include tree.h.
3978 * tree-data-ref.c: Include tree.h.
3980 2013-10-02 Yufeng Zhang <yufeng.zhang@arm.com>
3982 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref):
3985 2013-10-02 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3987 * config/t-sol2 (sol2-c.o): Remove header dependencies.
3988 Use $(COMPILE) and $(POSTCOMPILE).
3989 (sol2-cxx.o): Likewise.
3990 (sol2-stubs.o): Likewise.
3992 * config/x-solaris (host-solaris.o): Likewise.
3994 * config/sparc/t-sparc (sparc.o): Remove.
3995 (sparc-c.o): Remove header dependencies.
3996 Use $(COMPILE) and $(POSTCOMPILE).
3997 * config/sparc/x-sparc: Likewise.
3999 2013-10-02 Joern Rennecke <joern.rennecke@embecosm.com>
4001 * config/arc/arc-opts.h: Add 2013 to Copyright years.
4002 * config/arc/arc700.md: Likewise.
4003 * config/arc/arc-modes.def: Likewise.
4004 * config/arc/arc-simd.h: Likewise.
4005 * config/arc/t-arc-uClibc: Likewise.
4006 * config/arc/t-arc-newlib: Likewise.
4008 2013-10-02 Renlin Li <renlin.li@arm.com>
4010 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
4012 (aarch64_expand_epilogue): Likewise.
4014 2013-10-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4015 Yufeng Zhang <yufeng.zhang@arm.com>
4017 * gimple-ssa-strength-reduction.c (legal_cast_p_1): Forward
4019 (backtrace_base_for_ref): Call get_unwidened with 'base_in' if
4020 'base_in' represent a conversion and legal_cast_p_1 holds; set
4021 'base_in' with the returned value from get_unwidened.
4023 2013-10-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4025 * config/arm/arm.c (arm_legitimize_reload_address): Explain why
4026 plus_constant is not used.
4028 2013-10-01 Wei Mi <wmi@google.com>
4030 * config/i386/x86-tune.def (DEF_TUNE): Remove m_CORE_ALL.
4031 * config/i386/i386.md: Add define_peephole2 to
4032 break partial reg stall for cvtss2sd/cvtsd2ss.
4034 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
4036 * config/arc/arc.c (pass_arc_ifcvt::clone):
4037 Update for ctxt_ -> m_ctxt change.
4039 2013-10-01 Jeff Law <law@redhat.com>
4041 * tree-ssa-threadupdate.c (struct redirection_data): Delete
4042 outgoing_edge and intermediate_edge fields. Instead store the path.
4043 (redirection_data::hash): Hash on the last edge's destination index.
4044 (redirection_data::equal): Check the entire thread path.
4045 (lookup_redirectio_data): Corresponding changes.
4046 (create_edge_and_update_destination_phis): Likewise.
4047 (thread_single_edge): Likewise.
4049 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
4050 Diego Novillo <dnovillo@google.com>
4052 * config/arc/simdext.md (UNSPEC_ARC_SIMD_VLD32WH): Delete.
4053 (UNSPEC_ARC_SIMD_VLD32WL): Likewise.
4054 (vld32wh_insn, vld32wl_insn): Delete commented-out old
4055 versions of these patterns.
4057 * doc/extend.texi (long_call/medium_call/short_call): Typo fix.
4058 (__builtin_arc_aligned): Likewise.
4060 * config/arc/arc.md: Expand adc_0 comment stating the intended
4061 purpose and why it isn't ready.
4062 Replace commented out call_value_via_label_mixed with a
4063 plain comment about bl_s.
4065 * config/arc/arc.c (stdio.h): Don't include directly.
4066 (arc_expand_epilogue): Remove [0]: Remove fp_restored_p.
4067 Remove if (1) condition.
4068 (arc_encode_section_info): Fix comment.
4070 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
4072 * config/arc/arc.c (arc_conditional_register_usage):
4073 Use ARC_FIRST_SIMD_VR_REG / ARC_LAST_SIMD_VR_REG.
4074 Also set reg_alloc_order for DMA config regs.
4076 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
4077 Jeremy Bennett <jeremy.bennett@embecosm.com>
4079 * doc/install.texi (--with-cpu): Mention ARC.
4080 (arc-*-elf32): New paragraph.
4081 (arc-linux-uclibc): Likewise.
4082 * doc/md.texi (Machine Constraints): Add ARC part.
4083 * doc/invoke.texi: (menu): Add ARC Options.
4084 (Machine Dependent Options) <ARC Options>: Add synopsis.
4085 (node ARC Options): Add.
4086 * doc/extend.texi (long_call / short_call attribute): Add ARC.
4087 (ARC Built-in Functions): New section defining
4088 generic ARC built-in functions.
4089 (ARC SIMD Built-in Functions): New section defining SIMD specific
4091 (Declaring Attributes of Functions): Extended
4092 description of short_call and long_call attributes for ARC and
4093 added index entries.
4095 2013-10-01 Saurabh Verma <saurabh.verma@codito.com>
4096 Ramana Radhakrishnan <ramana.radhakrishnan@codito.com>
4097 Joern Rennecke <joern.rennecke@embecosm.com>
4098 Muhammad Khurram Riaz <khurram.riaz@arc.com>
4099 Brendan Kehoe <brendan@zen.org>
4100 Michael Eager <eager@eagercon.com>
4101 Simon Cook <simon.cook@embecosm.com>
4102 Jeremy Bennett <jeremy.bennett@embecosm.com>
4104 * config/arc, common/config/arc: New directories.
4106 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
4107 Brendan Kehoe <brendan@zen.org>
4108 Simon Cook <simon.cook@embecosm.com>
4110 * config.gcc (arc*-*-elf*, arc*-*-linux-uclibc*): New configurations.
4112 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
4114 * tree-ssa-live.h (coalesce_ssa_name): Move Prototype to...
4115 * tree-ssa-coalesce.h: New. Move prototype to here.
4116 * tree-outof-ssa.h: Include tree-ssa-coalesce.h.
4117 * tree-ssa-coalesce.c: Include tree-outof-ssa.h.
4118 (gimple_can_coalesce_p): Move to...
4119 * gimple.c (gimple_can_coalesce_p): Here.
4121 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
4123 * tree-into-ssa.c (enum need_phi_state): Relocate from tree-flow.h.
4124 (dump_decl_set): Move to gimple.c.
4125 * gimple.h: Don't include tree-ssa-operands.h.
4126 (dump_decl_set): Add prototype.
4127 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
4128 Move to gimple-ssa.h.
4129 (phi_ssa_name_p, phi_nodes, phi_nodes_ptr, gimple_phi_arg_def,
4130 gimple_phi_arg_def_ptr, gimple_phi_arg_edge, gimple_phi_arg_location,
4131 gimple_phi_arg_location_from_edge, gimple_phi_arg_set_location,
4132 gimple_phi_arg_has_location): Relocate from tree-flow-inline.h
4133 * gimple.c (walk_stmt_load_store_ops): Use gimple_phi_arg_def rather
4135 (dump_decl_set): Relocate here.
4136 * gimple-ssa.h: New file.
4137 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
4138 Relocate from gimple.h.
4139 * tree-cfg.c (has_zero_uses_1, single_imm_use_1): Move to...
4140 * tree-ssa-operands.c (swap_ssa_operands): Rename from
4141 swap_tree_operands and remove non-ssa path.
4142 (has_zero_uses_1, single_imm_use_1): Relocate from tree-cfg.c.
4143 * tree-ssa-reassoc.c (linearize_expr_tree, repropagate_negates): Use
4145 * tree-vect-loop.c (destroy_loop_vec_info, vect_is_slp_reduction,
4146 vect_is_simple_reduction_1): Use swap_ssa_operands.
4147 * tree-flow.h: Move various prototypes to tree-phinodes.h.
4148 (enum need_phi_state): Move to tree-into-ssa.c.
4149 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
4150 BREAK_FROM_IMM_USE_STMT): Move to ssa-iterators.h.
4151 (swap_tree_operands): Rename and move prototype to tree-ssa-operands.h.
4152 * tree-flow-inline.h (delink_imm_use, link_imm_use_to_list,
4153 link_imm_use, set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
4154 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
4155 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
4156 num_imm_uses): Move to ssa-iterators.h.
4157 (get_use_from_ptr, get_def_from_ptr): Move to tree-ssa-operands.h
4158 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Move to
4160 (op_iter_done, op_iter_next_def, op_iter_next_tree,
4161 clear_and_done_ssa_iter, op_iter_init, op_iter_init_use,
4162 op_iter_init_def, op_iter_init_tree, single_ssa_tree_operand,
4163 single_ssa_use_operand, single_ssa_def_operand, zero_ssa_operands,
4164 num_ssa_operands, delink_stmt_imm_use, single_phi_def,
4165 op_iter_init_phiuse, op_iter_init_phidef, end_imm_use_stmt_p,
4166 end_imm_use_stmt_traverse, move_use_after_head, link_use_stmts_after,
4167 first_imm_use_stmt, next_imm_use_stmt, first_imm_use_on_stmt,
4168 end_imm_use_on_stmt_p, next_imm_use_on_stmt): Move to ssa-iterators.h.
4169 (gimple_phi_arg_def, gimple_phi_arg_def_ptr, gimple_phi_arg_edge,
4170 gimple_phi_arg_location, gimple_phi_arg_location_from_edge,
4171 gimple_phi_arg_set_location, gimple_phi_arg_has_location, phi_nodes,
4172 phi_nodes_ptr, phi_ssa_name_p): Move to gimple.h.
4173 (set_phi_nodes): Move to tree-phinodes.h.
4174 * tree-ssa-operands.h (enum ssa_op_iter_type,
4175 struct ssa_operand_iterator_d, SSA_OP*, FOR_EACH_SSA*, SINGLE_SSA*,
4176 ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS): Move to ssa-iterators.h.
4177 (dump_decl_set): Remove prototype.
4178 (get_use_from_ptr, get_def_from_ptr): Relocate from tree-flow.h.
4179 * tree-phinodes.h: New file. Move some prototypes from tree-flow.h.
4180 (set_phi_nodes): Relocate from tree-flow-inline.h.
4181 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Relocate from
4183 * tree-ssa.h: Add tree-phinodes.h, gimple-ssa.h, ssa-iterators.h to
4184 include list. Temporarily add gimple.h to include list.
4185 * ssa-iterators.h: New file.
4186 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
4187 BREAK_FROM_IMM_USE_STMT): Relocate from tree-flow.h.
4188 (enum ssa_op_iter_type, struct ssa_operand_iterator_d, SSA_OP*,
4189 FOR_EACH_SSA*, SINGLE_SSA*, ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS):
4190 Relocate from tree-ssa-operands.h.
4191 (delink_imm_use, link_imm_use_to_list, link_imm_use,
4192 set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
4193 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
4194 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
4195 num_imm_uses, get_use_from_ptr, get_def_from_ptr,
4196 phi_arg_index_from_use, op_iter_done, op_iter_next_def,
4197 op_iter_next_tree, clear_and_done_ssa_iter, op_iter_init,
4198 op_iter_init_use, op_iter_init_def, op_iter_init_tree,
4199 single_ssa_tree_operand, single_ssa_use_operand, single_ssa_def_operand,
4200 zero_ssa_operands, num_ssa_operands, delink_stmt_imm_use,
4201 single_phi_def, op_iter_init_phiuse, op_iter_init_phidef,
4202 end_imm_use_stmt_p, end_imm_use_stmt_traverse, move_use_after_head,
4203 link_use_stmts_after, first_imm_use_stmt, next_imm_use_stmt,
4204 first_imm_use_on_stmt, end_imm_use_on_stmt_p, next_imm_use_on_stmt):
4205 Relocate from tree-flow-inline.h.
4206 * tree-outof-ssa.h: Change _SSAEXPAND_H macro to GCC_TREE_OUTOF_SSA_H.
4208 2013-10-01 Vidya Praveen <vidyapraveen@arm.com>
4211 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l2<mode>_internal): Rename to ...
4212 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): ... this;
4213 Insert '\t' to output template.
4214 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): New.
4215 (aarch64_saddl2<mode>, aarch64_uaddl2<mode>): Modify to call
4216 gen_aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal() instead.
4217 (aarch64_ssubl2<mode>, aarch64_usubl2<mode>): Ditto.
4219 2013-10-01 Uros Bizjak <ubizjak@gmail.com>
4221 * doc/install.texi (Host/target specific installation notes for GCC):
4222 Put @anchor before @heading.
4223 * doc/gcc.texi (titlepage): Use @uref and http:// prefix for website.
4224 Use @email for email addresses.
4226 2013-10-01 Jeff Law <law@redhat.com>
4228 * tree-ssa-threadedge.c (thread_across_edge): Make path a pointer to
4229 a vec. Only delete the path if we create one without successfully
4230 registering a jump thread.
4231 * tree-ssa-threadupdate.h (register_jump_thread): Pass in path vector
4233 * tree-ssa-threadupdate.c (threaded_edges): Remove. No longer used
4234 (paths): New vector of jump threading paths.
4235 (THREAD_TARGET, THREAD_TARGET2): Remove accessor macros.
4236 (THREAD_PATH): New accessor macro for the entire thread path.
4237 (lookup_redirection_data): Get intermediate and final outgoing edge
4238 from the thread path.
4239 (create_edge_and_update_destination_phis): Copy the threading path.
4240 (ssa_fix_duplicate_block_edges): Get edges and block types from the
4241 jump threading path.
4242 (ssa_redirect_edges): Get edges and block types from the jump threading
4243 path. Free the path vector.
4244 (thread_block): Get edges from the jump threading path. Look at the
4245 entire path to see if we thread to a loop exit. If we cancel a jump
4246 thread request, then free the path vector.
4247 (thread_single_edge): Get edges and block types from the jump threading
4248 path. Free the path vector.
4249 (thread_through_loop_header): Get edges and block types from the jump
4250 threading path. Free the path vector.
4251 (mark_threaded_blocks): Iterate over the vector of paths and store
4252 the path on the appropriate edge. Get edges and block types from the
4253 jump threading path.
4254 (mark_threaded_blocks): Get edges and block types from the jump
4255 threading path. Free the path vector.
4256 (thread_through_all_blocks): Use the vector of paths rather than
4257 a vector of 3-edge sets.
4258 (register_jump_thread): Accept pointer to a path vector rather
4259 than the path vector itself. Store the path vector for later use.
4262 2013-10-01 Jakub Jelinek <jakub@redhat.com>
4263 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4266 * config/s390/s390.c (s390_split_branches): Modify check for table
4268 (s390_chunkify_start): Rearrange table jump insn check in order to
4269 deal with compare and branch insns correctly.
4271 2013-10-01 Kugan Vivekanandarajah <kuganv@linaro.org>
4275 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
4276 * config/arm/arm.md (arm_ashldi3_1bit): define_insn into
4277 define_insn_and_split.
4278 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
4279 (shiftsi3_compare): New pattern.
4281 * config/arm/unspecs.md (UNSPEC_RRX): New.
4283 2013-10-01 Alan Modra <amodra@gmail.com>
4285 * stmt.c (expand_asm_operands): Revert part of 2013-09-24 special
4286 casing inout operands.
4288 2013-10-01 Richard Biener <rguenther@suse.de>
4290 PR tree-optimization/58553
4291 * tree-loop-distribution.c (struct partition_s): Add niter member.
4292 (classify_partition): Populate niter member for the partition
4293 and properly identify whether the relevant store happens before
4294 or after the loop exit.
4295 (generate_memset_builtin): Use niter member from the partition.
4296 (generate_memcpy_builtin): Likewise.
4298 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
4300 * vec.h (vec_prefix, vec): Prefix member names with "m_".
4301 * vec.c (vec_prefix::calculate_allocation): Update accordingly.
4303 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
4305 * basic-block.h (edge_list): Prefix member names with "m_".
4306 * context.h (context): Likewise.
4307 * domwalk.h (dom_walker): Likewise.
4308 * gengtype-state.c (s_expr_writer, state_writer): Likewise.
4309 * graphite-sese-to-poly.c (sese_dom_walker): Likewise.
4310 * hash-table.h (hash_table): Likewise.
4311 * machmode.h (bit_field_mode_iterator): Likewise.
4312 * pass_manager.h (pass_list): Likewise.
4313 * tree-into-ssa.c (mark_def_dom_walker): Likewise.
4314 * tree-pass.h (pass_data): Likewise.
4315 * tree-ssa-dom.c (dom_opt_dom_walker): Likewise.
4316 * tree-ssa-phiopt.c (nontrapping_dom_walker): Likewise,
4317 * tree-ssa-uncprop.c (uncprop_dom_walker): Likewise.
4318 * asan.c (pass_data_asan): Update accordingly.
4319 * cfganal.c (control_dependences::find_control_dependence): Likewise.
4320 (control_dependences::control_dependences): Likewise.
4321 (control_dependences::~control_dependences): Likewise.
4322 (control_dependences::~control_dependences): Likewise.
4323 (control_dependences::get_edges_dependent_on): Likewise.
4324 * cgraphbuild.c (pass_data_rebuild_cgraph_edges::clone): Likewise.
4325 (pass_data_remove_cgraph_callee_edges::clone): Likewise.
4326 * context.c (gcc::context::context): Likewise.
4327 * cprop.c (pass_rtl_cprop::clone): Likewise.
4328 * domwalk.c (dom_walker::walk): Likewise.
4329 * ipa-inline-analysis.c (pass_inline_parameters::clone): Likewise.
4330 * ipa-pure-const.c (pass_local_pure_const::clone): Likewise.
4331 * mode-switching.c (pass_mode_switching::clone): Likewise.
4332 * passes.c (opt_pass::opt_pass): Likewise.
4333 (pass_manager::pass_manager): Likewise.
4334 * predict.c (pass_strip_predict_hints::clone): Likewise.
4335 * recog.c (pass_data pass_data_peephole2::clone): Likewise.
4336 (pass_split_all_insns::clone): Likewise.
4337 * stor-layout.c (bit_field_mode_iterator::bit_field_mode_iterator):
4339 (bit_field_mode_iterator::next_mode): Likewise.
4340 (bit_field_mode_iterator::prefer_smaller_modes): Likewise.
4341 * tree-cfg.c (pass_split_crit_edges::clone): Likewise.
4342 * tree-cfgcleanup.c (pass_merge_phi::clone): Likewise.
4343 * tree-complex.c (pass_lower_complex::clone): Likewise.
4344 * tree-eh.c (pass_cleanup_eh::clone): Likewise.
4345 * tree-object-size.c (pass_object_sizes::clone): Likewise.
4346 * tree-optimize.c (pass_fixup_cfg::clone): Likewise.
4347 * tree-ssa-ccp.c (pass_data_ccp::clone): Likewise.
4348 (pass_fold_builtins::clone): Likewise.
4349 * tree-ssa-copy.c (pass_data_copy_prop::clone): Likewise.
4350 * tree-ssa-copyrename.c (pass_rename_ssa_copies::clone): Likewise.
4351 * tree-ssa-dce.c (pass_dce::clone, pass_dce_loop::clone): Likewise.
4352 (pass_cd_dce::clone): Likewise.
4353 * tree-ssa-dom.c (pass_dominator::clone): Likewise.
4354 (pass_phi_only_cprop::clone): Likewise.
4355 * tree-ssa-dse.c (pass_dse::clone): Likewise.
4356 * tree-ssa-forwprop.c (pass_forwprop::clone): Likewise.
4357 * tree-ssa-loop.c (pass_lim::clone): Likewise.
4358 * tree-ssa-phiopt.c (pass_phiopt::clone): Likewise.
4359 * tree-ssa-pre.c (pass_fre::clone): Likewise.
4360 * tree-ssa-reassoc.c (pass_reassoc::clone): Likewise.
4361 * tree-ssa-uninit.c (pass_late_warn_uninitialized::clone): Likewise.
4362 * tree-tailcall.c (pass_tail_recursion::clone): Likewise.
4363 * tree-vect-generic.c (pass_lower_vector_ssa::clone): Likewise.
4364 * tree-vrp.c (pass_vrp::clone): Likewise.
4365 * tsan.c (pass_tsan::clone): Likewise.
4367 2013-09-30 Jakub Jelinek <jakub@redhat.com>
4370 * fold-const.c (tree_unary_nonnegative_warnv_p): Use
4371 INTEGRAL_TYPE_P (t) instead of TREE_CODE (t) == INTEGER_TYPE.
4374 * fold-const.c (fold_ternary_loc): For A < 0 : <sign bit of A> : 0
4375 optimization, punt if sign_bit_p looked through any zero extension.
4377 2013-09-30 Teresa Johnson <tejohnson@google.com>
4379 * tree-ssa-threadupdate.c (ssa_fix_duplicate_block_edges):
4380 Update redirected out edge count in joiner case.
4381 (ssa_redirect_edges): Common the joiner and non-joiner cases
4382 so that joiner case gets profile updates.
4384 2013-09-30 Richard Biener <rguenther@suse.de>
4386 PR tree-optimization/58554
4387 * tree-loop-distribution.c (classify_partition): Require
4388 unconditionally executed stores for memcpy and memset recognition.
4389 (tree_loop_distribution): Calculate dominance info.
4391 2013-09-30 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
4393 * config/aarch64/aarch64.h (MCOUNT_NAME): Define.
4394 (NO_PROFILE_COUNTERS): Likewise.
4395 (PROFILE_HOOK): Likewise.
4396 (FUNCTION_PROFILER): Likewise.
4397 * config/aarch64/aarch64.c (aarch64_function_profiler): Remove.
4399 2013-09-30 Iain Sandoe <iain@codesourcery.com>
4401 * config/rs6000/darwin.md (load_macho_picbase_si): Wrap machopic
4402 calls and defines in TARGET_MACHO conditional.
4403 (load_macho_picbase_di): Likewise.
4404 (reload_macho_picbase): Likewise.
4405 (reload_macho_picbase_si): Likewise.
4406 (reload_macho_picbase_di): Likewise.
4407 (nonlocal_goto_receiver): Likewise.
4409 2013-09-30 Nick Clifton <nickc@redhat.com>
4411 * config/msp430/msp430.c (msp430x_names): New array. Lists MCUs
4412 that use the MSP430X ISA.
4413 (msp430_option_override): Scan -mmcu command line option for any
4414 MCU name that supports the MSP430X ISA.
4415 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add matches for known
4416 -mmcu options which enable the MSP430X ISA.
4418 2013-09-30 Richard Biener <rguenther@suse.de>
4421 * tree-cfg.c (make_abnormal_goto_edges): Skip debug statements
4422 before looking for setjmp-like calls.
4424 2013-09-29 Iain Sandoe <iain@codesourcery.com>
4427 * config/darwin-protos.h (machopic_get_function_picbase): New.
4428 * config/darwin.c (machopic_get_function_picbase): New.
4429 * config/rs6000/darwin.md (load_macho_picbase_si): Update picbase
4430 label for a new func. (load_macho_picbase_di): Likewise.
4431 (reload_macho_picbase): New expand.
4432 (reload_macho_picbase_si): New insn.
4433 (reload_macho_picbase_di): New insn.
4434 (nonlocal_goto_receiver): New define and split.
4435 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_RELD_MPIC.
4436 (unspecv enum): Add UNSPECV_NLGR.
4438 2013-09-29 Iain Sandoe <iain@codesourcery.com>
4440 * config/rs6000/rs6000.c (rs6000_init_dwarf_reg_sizes_extra): Ensure
4441 that altivec registers are correctly sized on Darwin.
4443 2013-09-29 Iain Sandoe <iain@codesourcery.com>
4445 * config/t-darwin (darwin.o, darwin-c.o, darwin-f.o,
4446 darwin-driver.o): Use COMPILE and POSTCOMPILE.
4447 * config/x-darwin (host-darwin.o): Likewise.
4448 * config/i386/x-darwin (host-i386-darwin.o): Likewise.
4449 * config/rs6000/x-darwin (host-ppc-darwin.o): Likewise.
4450 * config/rs6000/x-darwin64 (host-ppc64-darwin.o): Likewise.
4452 2013-09-29 Uros Bizjak <ubizjak@gmail.com>
4454 * doc/invoke.texi: Fix usage of @tie{} command.
4456 2013-09-29 Eric Botcazou <ebotcazou@adacore.com>
4458 * config/sparc/sync.md: Add peephole for consecutive memory barriers.
4460 2013-09-28 Jan Hubicka <jh@suse.cz>
4462 * config/i386/x86-tune.def: Add documentation for each of the options;
4465 2013-09-28 Jan Hubicka <jh@suse.cz>
4467 * x86-tune.def (X86_TUNE_SSE_UNALIGNED_LOAD_OPTIMAL): Enable for
4469 (X86_TUNE_SSE_UNALIGNED_STORE_OPTIMAL): Likewise.
4470 (X86_TUNE_FOUR_JUMP_LIMIT): Drop for generic and buldozer.
4471 (X86_TUNE_PAD_RETURNS): Drop for buldozer chips.
4472 (X86_TUNE_AVOID_VECTOR_DECODE): Drop for generic.
4473 (X86_TUNE_REASSOC_FP_TO_PARALLEL): Enable for generic.
4475 2013-09-28 Richard Sandiford <rdsandiford@googlemail.com>
4477 * alloc-pool.c, asan.c, auto-inc-dec.c, basic-block.h, bb-reorder.c,
4478 bitmap.c, bitmap.h, bt-load.c, builtins.c, calls.c, cfgcleanup.c,
4479 cfgexpand.c, cfghooks.c, cfgloop.c, cfgloopmanip.c, cfgrtl.c, cgraph.c,
4480 cgraph.h, cgraphbuild.c, cgraphclones.c, cgraphunit.c, collect2.c,
4481 combine-stack-adj.c, combine.c, compare-elim.c, context.c, context.h,
4482 cprop.c, cse.c, cselib.c, dbxout.c, dce.c, defaults.h, df-core.c,
4483 df-problems.c, df-scan.c, df.h, diagnostic.c, double-int.c, dse.c,
4484 dumpfile.c, dwarf2asm.c, dwarf2cfi.c, dwarf2out.c, emit-rtl.c,
4485 errors.c, except.c, expmed.c, expr.c, file-find.c, final.c,
4486 fixed-value.c, fold-const.c, function.c, fwprop.c, gcc-ar.c, gcc.c,
4487 gcov-io.c, gcov-io.h, gcov.c, gcse.c, genattr-common.c, genattr.c,
4488 genattrtab.c, genautomata.c, genconfig.c, genemit.c, genextract.c,
4489 genflags.c, gengenrtl.c, gengtype-state.c, gengtype.c, genmodes.c,
4490 genopinit.c, genoutput.c, genpeep.c, genpreds.c, genrecog.c,
4491 gensupport.c, ggc-common.c, ggc-page.c, gimple-fold.c, gimple-low.c,
4492 gimple-pretty-print.c, gimple-ssa-strength-reduction.c, gimple.c,
4493 gimple.h, godump.c, graphite-clast-to-gimple.c,
4494 graphite-optimize-isl.c, graphite-poly.h, graphite-sese-to-poly.c,
4495 graphite.c, haifa-sched.c, hash-table.c, hash-table.h, hwint.c,
4496 hwint.h, ifcvt.c, incpath.c, init-regs.c, input.h, intl.c, intl.h,
4497 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
4498 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c,
4499 ipa-utils.c, ipa.c, ira-build.c, ira.c, jump.c, loop-doloop.c,
4500 loop-init.c, loop-invariant.c, loop-iv.c, lower-subreg.c, lto-cgraph.c,
4501 lto-streamer-in.c, lto-streamer-out.c, lto-wrapper.c, mcf.c,
4502 mode-switching.c, modulo-sched.c, omp-low.c, optabs.c, opts.c,
4503 pass_manager.h, passes.c, plugin.c, postreload-gcse.c, postreload.c,
4504 predict.c, prefix.c, pretty-print.c, print-rtl.c, print-tree.c,
4505 profile.c, read-md.c, real.c, real.h, recog.c, ree.c, reg-stack.c,
4506 regcprop.c, reginfo.c, regmove.c, regrename.c, regs.h, regstat.c,
4507 reload1.c, reorg.c, rtl.c, rtl.h, rtlanal.c, sbitmap.c, sched-rgn.c,
4508 sdbout.c, sel-sched-ir.c, sel-sched.c, sparseset.c, stack-ptr-mod.c,
4509 statistics.c, stmt.c, stor-layout.c, store-motion.c, streamer-hooks.h,
4510 system.h, target-hooks-macros.h, targhooks.c, targhooks.h, toplev.c,
4511 tracer.c, trans-mem.c, tree-browser.c, tree-call-cdce.c, tree-cfg.c,
4512 tree-cfgcleanup.c, tree-complex.c, tree-data-ref.c, tree-data-ref.h,
4513 tree-eh.c, tree-emutls.c, tree-flow.h, tree-if-conv.c, tree-into-ssa.c,
4514 tree-iterator.c, tree-loop-distribution.c, tree-mudflap.c,
4515 tree-nested.c, tree-nomudflap.c, tree-nrv.c, tree-object-size.c,
4516 tree-optimize.c, tree-pass.h, tree-pretty-print.c, tree-profile.c,
4517 tree-scalar-evolution.c, tree-sra.c, tree-ssa-ccp.c,
4518 tree-ssa-coalesce.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
4519 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
4520 tree-ssa-ifcombine.c, tree-ssa-live.c, tree-ssa-loop-ch.c,
4521 tree-ssa-loop-im.c, tree-ssa-loop-ivopts.c, tree-ssa-loop-prefetch.c,
4522 tree-ssa-loop.c, tree-ssa-math-opts.c, tree-ssa-operands.c,
4523 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
4524 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
4525 tree-ssa-structalias.c, tree-ssa-threadedge.c, tree-ssa-threadupdate.c,
4526 tree-ssa-uncprop.c, tree-ssa-uninit.c, tree-ssa.c, tree-ssanames.c,
4527 tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
4528 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
4529 tree-vect-stmts.c, tree-vectorizer.c, tree-vectorizer.h, tree-vrp.c,
4530 tree.c, tree.h, tsan.c, tsystem.h, value-prof.c, var-tracking.c,
4531 varasm.c, vec.h, vmsdbgout.c, vtable-verify.c, web.c: Add missing
4532 whitespace before "(".
4534 2013-09-28 Sandra Loosemore <sandra@codesourcery.com>
4536 * expr.h (extract_bit_field): Remove packedp parameter.
4537 * expmed.c (extract_fixed_bit_field): Remove packedp parameter
4538 from forward declaration.
4539 (store_split_bit_field): Remove packedp arg from calls to
4540 extract_fixed_bit_field.
4541 (extract_bit_field_1): Remove packedp parameter and packedp
4542 argument from recursive calls and calls to extract_fixed_bit_field.
4543 (extract_bit_field): Remove packedp parameter and corresponding
4544 arg to extract_bit_field_1.
4545 (extract_fixed_bit_field): Remove packedp parameter. Remove code
4547 (extract_split_bit_field): Remove packedp arg from call to
4548 extract_fixed_bit_field.
4549 * expr.c (emit_group_load_1): Adjust calls to extract_bit_field.
4550 (copy_blkmode_from_reg): Likewise.
4551 (copy_blkmode_to_reg): Likewise.
4552 (read_complex_part): Likewise.
4553 (store_field): Likewise.
4554 (expand_expr_real_1): Likewise.
4555 * calls.c (store_unaligned_arguments_into_pseudos): Adjust call
4556 to extract_bit_field.
4557 * config/tilegx/tilegx.c (tilegx_expand_unaligned_load): Adjust
4558 call to extract_bit_field.
4559 * config/tilepro/tilepro.c (tilepro_expand_unaligned_load): Adjust
4560 call to extract_bit_field.
4561 * doc/invoke.texi (Code Gen Options): Remove mention of warnings
4562 and special packedp behavior from -fstrict-volatile-bitfields
4565 2013-09-27 Jan-Benedict Glaw <jbglaw@lug-owl.de>
4567 * lra-eliminations.c (init_elim_table): Guard value_p.
4569 2013-09-27 Michael Meissner <meissner@linux.vnet.ibm.com>
4571 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Allow
4572 DFmode, DImode, and SFmode in the upper VSX registers based on the
4573 -mupper-regs-{df,sf} flags. Fix wu constraint to be ALTIVEC_REGS
4574 if -mpower8-vector. Combine -mvsx-timode handling with the rest
4575 of the VSX register handling.
4577 * config/rs6000/rs6000.md (f32_lv): Use %x0 for VSX regsters.
4579 (zero_extendsidi2_lfiwzx): Add support for loading into the
4580 Altivec registers with -mpower8-vector. Use wu/wv constraints to
4581 only do VSX memory options on Altivec registers.
4582 (extendsidi2_lfiwax): Likewise.
4583 (extendsfdf2_fpr): Likewise.
4584 (mov<mode>_hardfloat, SF/SD modes): Likewise.
4585 (mov<mode>_hardfloat32, DF/DD modes): Likewise.
4586 (mov<mode>_hardfloat64, DF/DD modes): Likewise.
4587 (movdi_internal64): Likewise.
4589 2013-09-27 Xinliang David Li <davidxl@google.com>
4591 * opts.c (finish_options): Adjust parameters
4592 according to vect cost model.
4593 (common_handle_option): Set dynamic vect cost
4595 targhooks.c (default_add_stmt_cost): Compute stmt cost
4597 * tree-vect-loop.c (vect_estimate_min_profitable_iters):
4598 Use helper function.
4599 * tree-vectorizer.h (unlimited_cost_model): New function.
4600 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use helper function.
4601 * tree-vect-data-refs.c (vect_peeling_hash_insert): Use helper
4603 (vect_enhance_data_refs_alignment): Ditto.
4604 * flag-types.h: New enum.
4605 * common/config/i386/i386-common.c (ix86_option_init_struct):
4606 No need to initialize vect_cost_model flag.
4607 * config/i386/i386.c (ix86_add_stmt_cost): Compute stmt cost
4610 2013-09-27 Diego Novillo <dnovillo@google.com>
4612 * gimple.h (enum ssa_mode): Remove.
4614 2013-09-27 Paulo Matos <pmatos@broadcom.com>
4616 * cfgloop.h (number_of_loops): Fix typo in check for null.
4618 2013-09-27 Jakub Jelinek <jakub@redhat.com>
4621 * tree-cfg.c (move_sese_region_to_fn): Also move loops that
4622 are children of outermost saved_cfun's loop, and set it up to
4623 be moved to dest_cfun's outermost loop. Fix up num_nodes adjustments
4624 if loop != loop0 and SESE region contains bbs that belong to loop0.
4626 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
4628 * rtlanal.c (must_be_base_p, must_be_index_p): Delete.
4629 (binary_scale_code_p, get_base_term, get_index_term): New functions.
4630 (set_address_segment, set_address_base, set_address_index)
4631 (set_address_disp): Accept the argument unconditionally.
4632 (baseness): Remove must_be_base_p and must_be_index_p checks.
4633 (decompose_normal_address): Classify as much as possible in the
4636 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
4638 * cse.c (count_reg_usage): Handle INT_LIST.
4639 * lra-eliminations.c (lra_eliminate_regs_1): Likewise.
4640 * reginfo.c (reg_scan_mark_refs): Likewise.
4641 * reload1.c (eliminate_regs_1): Likewise.
4643 2013-09-27 Iain Sandoe <iain@codesourcery.com>
4646 * rtlanal.c (lsb_bitfield_op_p): Make both parts of the comparison
4649 2013-09-27 Richard Biener <rguenther@suse.de>
4651 PR tree-optimization/58459
4652 * tree-ssa-forwprop.c (forward_propagate_addr_expr): Remove
4653 restriction not propagating into loops.
4655 2013-09-26 Florian Weimer <fw@deneb.enyo.de>
4657 * tree-ssa.h (walk_use_def_chains_fn, walk_use_def_chains): Delete.
4658 * tree-ssa.c (walk_use_def_chains_1, walk_use_def_chains): Delete.
4659 * doc/tree-ssa.texi (Walking use-def chains): Delete.
4661 2013-09-26 Richard Biener <rguenther@suse.de>
4663 * tree-into-ssa.c (rewrite_into_ssa): Make more SSA names to anonymous.
4665 2013-09-26 Richard Biener <rguenther@suse.de>
4667 * alias.h (component_uses_parent_alias_set): Rename to ...
4668 (component_uses_parent_alias_set_from): ... this.
4669 * alias.c (component_uses_parent_alias_set): Rename to ...
4670 (component_uses_parent_alias_set_from): ... this and return
4672 (reference_alias_ptr_type_1): Use the result from
4673 component_uses_parent_alias_set_from instead of stripping
4674 components one at a time.
4675 * emit-rtl.c (set_mem_attributes_minus_bitpos): Adjust.
4677 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
4679 * tree-ssa-live.h (find_replaceable_exprs, dump_replaceable_exprs):
4680 Move prototypes to...
4681 * tree-ssa-ter.h: New File. Move prototypes here.
4682 * tree-flow.h (stmt_is_replaceable_p): Remove prototype.
4683 * tree-outof-ssa.h: New. Rename ssaexpand.h, include tree-ssa-ter.h.
4684 * tree-outof-ssa.c (ssa_is_replaceable_p): New. Refactor common bits
4685 from is_replaceable_p.
4686 * tree-ssa-ter.c (is_replaceable_p, stmt_is_replaceable_p): Delete.
4687 (ter_is_replaceable_p): New. Use new refactored ssa_is_replaceable_p.
4688 (process_replaceable): Use ter_is_replaceable_p.
4689 (find_replaceable_in_bb): Use ter_is_replaceable_p.
4690 * expr.c (stmt_is_replaceable_p): Relocate from tree-ssa-ter.c. Use
4691 newly refactored ssa_is_replaceable_p.
4692 * cfgexpand.c: Include tree-outof-ssa.h.
4693 * ssaexpand.h: Delete.
4695 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
4697 * gimple.c (gimple_replace_lhs): Move to tree-ssa.c and rename.
4698 (struct count_ptr_d, count_ptr_derefs, count_uses_and_derefs): Move to
4700 (create_gimple_tmp): Delete.
4701 (get_expr_type, build_assign, build_type_cast): Move to...
4702 * gimple-builder.c: New File.
4703 (get_expr_type): Relocate from gimple.c.
4704 (build_assign, build_type_cast): Change to only create ssanames.
4705 * gimple.h: Move prototypes to...
4706 * gimple-builder.h: New File. Here.
4707 * tree-ssa.h: And here.
4708 * tree-ssa.c (struct count_ptr_d, count_ptr_derefs,
4709 count_uses_and_derefs): Relocate from gimple.c.
4710 (gimple_replace_ssa_lhs): Renamed gimple_replace_ssa from gimple.c
4711 * tree-ssa-reassoc.c (repropagate_negates): Use gimple_replace_ssa_lhs.
4712 * tree-ssa-math-opts (execute_cse_reciprocals): Use
4713 gimple_replace_ssa_lhs.
4714 * asan.c: Include gimple-builder.h.
4715 * Makefile.in: Add gimple-builder.o.
4717 2013-09-26 Richard Biener <rguenther@suse.de>
4719 * tree-ssa-live.c (var_map_base_init): Handle SSA names with
4720 DECL_IGNORED_P base VAR_DECLs like anonymous SSA names.
4721 (loe_visit_block): Use gcc_checking_assert.
4722 * tree-ssa-coalesce.c (create_outofssa_var_map): Use
4723 gimple_assign_ssa_name_copy_p.
4724 (gimple_can_coalesce_p): Adjust according to the var_map_base_init
4727 2013-09-26 David Edelsohn <dje.gcc@gmail.com>
4729 * config/rs6000/t-rs6000 (rs6000.o): Remove.
4730 (rs6000-c.o): Use COMPILE and POSTCOMPILE.
4732 2013-09-26 Richard Biener <rguenther@suse.de>
4734 PR tree-optimization/58539
4735 * tree-vect-loop.c (vect_create_epilog_for_reduction): Honor
4736 the fact that debug statements are not taking part in loop-closed
4739 2013-09-26 Nick Clifton <nickc@redhat.com>
4741 * config/msp430/msp430.c (msp430_expand_epilogue): Fix compile
4742 time warning message.
4743 (msp430_print_operand_raw): Delete unused letter parameter.
4744 (TARGET_PRINT_OPERAND_ADDRESS): Define.
4745 (msp430_print_operand_address): New function.
4746 (msp430_print_operand): Move address printing code from here to
4748 * config/msp430/msp430.md (movsipsi2): Add comment in generated
4750 (zero_extendpsisi2): Likewise.
4751 (extendpsisi2): New pattern.
4752 (andneghi3): New pattern.
4754 2013-09-26 Yvan Roux <yvan.roux@linaro.org>
4756 * config/aarch64/aarch64.opt (mlra): New option.
4757 * config/aarch64/aarch64.c (aarch64_lra_p): New function.
4758 (TARGET_LRA_P): Define.
4760 2013-09-26 Eric Botcazou <ebotcazou@adacore.com>
4762 * expr.c (expand_assignment): Remove obsolete comment.
4764 2013-09-25 Jeff Law <law@redhat.com>
4766 * tree-flow.h (thread_through_all_blocks): Prototype moved into
4767 tree-ssa-threadupdate.h.
4768 (register_jump_thread): Similarly.
4769 * tree-ssa-threadupdate.h: New header file.
4770 * tree-ssa-dom.c: Include tree-ssa-threadupdate.h.
4771 * tree-vrp.c: Likewise.
4772 * tree-ssa-threadedge.c: Include tree-ssa-threadupdate.h.
4773 (thread_around_empty_blocks): Change type of path vector argument to
4774 an edge,type pair from just an edge. Initialize both elements when
4775 appending to a jump threading path. Tweak references to elements
4777 (thread_across_edge): Similarly. Release memory for the elements
4779 * tree-ssa-threadupdate.c: Include tree-ssa-threadupdate.h.
4780 (dump_jump_thread_path): New function broken out from
4781 register_jump_thread.
4782 (register_jump_thread): Use dump_jump_thread_path. Change type of
4783 path vector entries. Search the path for NULL edges and dump
4784 the path if one is found. Tweak the conversion of path to 3-edge
4785 form to use the block copy type information embedded in the path.
4787 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
4789 * lra.c (update_inc_notes): Remove all REG_DEAD and REG_UNUSED notes.
4791 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
4792 Vladimir Makarov <vmakarov@redhat.com>
4794 * rtlanal.c (lsb_bitfield_op_p): New predicate for bitfield operations
4795 from the least significant bit.
4796 (strip_address_mutations): Add bitfield operations handling.
4797 (must_be_index_p): Add shifting and rotate operations handling.
4798 (set_address_base): Use must_be_base_p predicate.
4799 (set_address_index): Use must_be_index_p predicate.
4801 2013-09-25 Alexander Ivchenko <alexander.ivchenko@intel.com>
4802 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4803 Sergey Lega <sergey.s.lega@intel.com>
4804 Anna Tikhonova <anna.tikhonova@intel.com>
4805 Ilya Tocar <ilya.tocar@intel.com>
4806 Andrey Turetskiy <andrey.turetskiy@intel.com>
4807 Ilya Verbin <ilya.verbin@intel.com>
4808 Kirill Yukhin <kirill.yukhin@intel.com>
4809 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4811 * config/i386/i386.c (ix86_avx256_split_vector_move_misalign):
4813 (ix86_expand_vector_move_misalign): Support new unaligned load and
4814 stores and use new names.
4815 (CODE_FOR_sse2_storedqu): Rename to ...
4816 (CODE_FOR_sse2_storedquv16qi): ... this.
4817 (CODE_FOR_sse2_loaddqu): Rename to ...
4818 (CODE_FOR_sse2_loaddquv16qi): ... this.
4819 (CODE_FOR_avx_loaddqu256): Rename to ...
4820 (CODE_FOR_avx_loaddquv32qi): ... this.
4821 (CODE_FOR_avx_storedqu256): Rename to ...
4822 (CODE_FOR_avx_storedquv32qi): ... this.
4823 * config/i386/i386.md (fpint_logic): New.
4824 * config/i386/sse.md (VMOVE): Extend for AVX512.
4828 (VI_UNALIGNED_LOADSTORE): Ditto.
4829 (sse2_avx_avx512f): Ditto.
4830 (sse2_avx2): Extend for AVX512.
4831 (sse4_1_avx2): Ditto.
4832 (avx2_avx512f): New.
4833 (sse): Extend for AVX512.
4836 (avxsizesuffix): Ditto.
4837 (sseintvecmode): Ditto.
4839 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Ditto.
4840 (<sse>_storeu<ssemodesuffix><avxsizesuffix>): Ditto.
4841 (<sse2>_loaddqu<avxsizesuffix>): Extend for AVX512 and rename to ...
4842 (<sse2_avx_avx512f>_loaddqu<mode>): ... this.
4843 (<sse2>_storedqu<avxsizesuffix>): Extend for AVX512 and rename to ...
4844 (<sse2_avx_avx512f>_storedqu<mode): ... this.
4845 (<sse>_movnt<mode>): Replace constraint "x" with "v".
4846 (STORENT_MODE): Extend for AVX512.
4847 (*absneg<mode>2): Replace constraint "x" with "v".
4848 (*mul<mode>3): Ditto.
4849 (*ieee_smin<mode>3): Ditto.
4850 (*ieee_smax<mode>3): Ditto.
4851 (avx_cmp<mode>3): Replace VF with VF_128_256.
4852 (*<sse>_maskcmp<mode>3_comm): Ditto.
4853 (<sse>_maskcmp<mode>3): Ditto.
4854 (<sse>_andnot<mode>3): Extend for AVX512.
4855 (<code><mode>3, anylogic): Replace VF with VF_128_256.
4856 (<code><mode>3, fpint_logic): New.
4857 (*<code><mode>3): Extend for AVX512.
4858 (avx512flogicsuff): New.
4859 (avx512f_<logic><mode>): Ditto.
4860 (<sse>_movmsk<ssemodesuffix><avxsizesuffix>): Replace VF with
4862 (<sse4_1>_blend<ssemodesuffix><avxsizesuffix>): Ditto.
4863 (<sse4_1>_blendv<ssemodesuffix><avxsizesuffix>): Ditto.
4864 (<sse4_1>_dp<ssemodesuffix><avxsizesuffix>): Ditto.
4865 (avx_vtest<ssemodesuffix><avxsizesuffix>): Ditto.
4866 (<sse4_1>_round<ssemodesuffix><avxsizesuffix>): Ditto.
4867 (xop_vpermil2<mode>3): Ditto.
4868 (*avx_vpermilp<mode>): Extend for AVX512 and rename to ...
4869 (*<sse2_avx_avx512f>_vpermilp<mode>): ... this.
4870 (avx_vpermilvar<mode>3): Extend for AVX512 and rename to ...
4871 (<sse2_avx_avx512f>_vpermilvar<mode>3): ... this.
4873 2013-09-25 Tom Tromey <tromey@redhat.com>
4875 * Makefile.in (PARTITION_H, LTO_SYMTAB_H, COMMON_TARGET_DEF_H)
4876 (RTL_ERROR_H, TRANS_MEM_H, COVERAGE_H, DEMANGLE_H, ALIAS_H)
4877 (SCHED_INT_H, SEL_SCHED_IR_H, SEL_SCHED_DUMP_H, VALTRACK_H, DDG_H)
4878 (GGC_INTERNAL_H, DECNUM_H, BACKTRACE_H, MKDEPS_H, TREE_HASHER_H)
4879 (TREE_SSA_LIVE_H, SSAEXPAND_H, DWARF2OUT_H, SCEV_H, OMEGA_H)
4880 (TREE_DATA_REF_H, IRA_INT_H, LRA_INT_H, DBGCNT_H, DATA_STREAMER_H)
4881 (GIMPLE_STREAMER_H, TREE_STREAMER_H, STREAMER_HOOKS_H)
4882 (TREE_VECTORIZER_H, IPA_INLINE_H, GSTAB_H, LIBFUNCS_H)
4883 (GRAPHITE_HTAB_H): Remove.
4885 2013-09-25 Tom Tromey <tromey@redhat.com>
4887 * config/mcore/t-mcore (CROSS_FLOAT_H): Remove.
4889 2013-09-25 Tom Tromey <tromey@redhat.com>
4891 * config/t-glibc (glibc-c.o): Use COMPILE and POSTCOMPILE.
4893 2013-09-25 Tom Tromey <tromey@redhat.com>
4895 * config/i386/t-i386 (i386.o): Remove.
4896 (i386-c.o): Use COMPILE and POSTCOMPILE.
4898 2013-09-25 Tom Tromey <tromey@redhat.com>
4900 * Makefile.in ($(out_object_file)): Use COMPILE and POSTCOMPILE.
4902 2013-09-25 Tom Tromey <tromey@redhat.com>
4904 * Makefile.in (graph.o, sbitmap.o, sparseset.o, gcc-ar.o)
4905 (gcc-ranlib.o, gcc-nm.o, collect2.o, collect2-aix.o, tlink.o)
4906 (lto-wrapper.o, default-c.o, attribs.o, incpath.o, prefix.o)
4907 (gcc.o, options.o, options-save.o, version.o, gtype-desc.o)
4908 (trans-mem.o, ggc-common.o, ggc-page.o, ggc-none.o, stringpool.o)
4909 (convert.o, double-int.o, lto-compress.o, data-streamer-in.o)
4910 (data-streamer-out.o, data-streamer.o, gimple-streamer-in.o)
4911 (gimple-streamer-out.o, tree-streamer.o, tree-streamer-in.o)
4912 (tree-streamer-out.o, streamer-hooks.o, lto-cgraph.o)
4913 (lto-streamer-in.o, lto-streamer-out.o, lto-section-in.o)
4914 (lto-section-out.o, lto-opts.o, lto-streamer.o, langhooks.o)
4915 (test-dump.o, tree.o, tree-dump.o, tree-inline.o, print-tree.o)
4916 (stor-layout.o, asan.o, tsan.o, ubsan.o, tree-ssa-tail-merge.o)
4917 (tree-ssa-structalias.o, tree-ssa-uninit.o, tree-ssa.o)
4918 (tree-into-ssa.o, tree-ssa-ter.o, tree-ssa-coalesce.o)
4919 (tree-outof-ssa.o, tree-ssa-dse.o, tree-ssa-forwprop.o)
4920 (tree-ssa-phiprop.o, tree-ssa-ifcombine.o, tree-ssa-phiopt.o)
4921 (tree-nrv.o, tree-ssa-copy.o, tree-ssa-propagate.o)
4922 (tree-ssa-dom.o, tree-ssa-uncprop.o, tree-ssa-threadedge.o)
4923 (tree-ssa-threadupdate.o, tree-ssanames.o, tree-phinodes.o)
4924 (domwalk.o, tree-ssa-live.o, tree-ssa-copyrename.o)
4925 (tree-ssa-pre.o, tree-ssa-sccvn.o)
4926 (gimple-ssa-strength-reduction.o, tree-vrp.o, tree-cfg.o)
4927 (tree-cfgcleanup.o, tree-tailcall.o, tree-ssa-sink.o)
4928 (tree-nested.o, tree-if-conv.o, tree-iterator.o, tree-dfa.o)
4929 (tree-ssa-operands.o, tree-eh.o, tree-ssa-loop.o)
4930 (tree-ssa-loop-unswitch.o, tree-ssa-address.o)
4931 (tree-ssa-loop-niter.o, tree-ssa-loop-ivcanon.o)
4932 (tree-ssa-loop-ch.o, tree-ssa-loop-prefetch.o, tree-predcom.o)
4933 (tree-ssa-loop-ivopts.o, tree-affine.o, tree-ssa-loop-manip.o)
4934 (tree-ssa-loop-im.o, tree-ssa-math-opts.o, tree-ssa-alias.o)
4935 (tree-ssa-reassoc.o, tree-optimize.o, gimplify.o)
4936 (gimple-iterator.o, gimple-fold.o, gimple-low.o, omp-low.o)
4937 (tree-browser.o, omega.o, tree-chrec.o, tree-scalar-evolution.o)
4938 (tree-data-ref.o, sese.o, graphite.o, graphite-blocking.o)
4939 (graphite-clast-to-gimple.o, graphite-dependences.o)
4940 (graphite-interchange.o, graphite-poly.o)
4941 (graphite-scop-detection.o, graphite-sese-to-poly.o)
4942 (graphite-optimize-isl.o, tree-vect-loop.o)
4943 (tree-vect-loop-manip.o, tree-vect-patterns.o, tree-vect-slp.o)
4944 (tree-vect-stmts.o, tree-vect-data-refs.o, tree-vectorizer.o)
4945 (vtable-verify.o, tree-loop-distribution.o, tree-parloops.o)
4946 (tree-stdarg.o, tree-object-size.o, internal-fn.o, gimple.o)
4947 (gimple-pretty-print.o, tree-mudflap.o, tree-nomudflap.o)
4948 (tree-pretty-print.o, tree-diagnostic.o, fold-const.o)
4949 (diagnostic.o, diagnostic-color.o, opts.o, opts-global.o)
4950 (opts-common.o, targhooks.o, common/common-targhooks.o, input.o)
4951 (toplev.o, hwint.o, passes.o, plugin.o, main.o, host-default.o)
4952 (rtl-error.o, rtl.o, print-rtl.o, rtlanal.o, varasm.o, function.o)
4953 (statistics.o, stmt.o, except.o, expr.o, dojump.o, builtins.o)
4954 (calls.o, expmed.o, explow.o, optabs.o, dbxout.o, debug.o)
4955 (sdbout.o, dwarf2out.o, dwarf2cfi.o, dwarf2asm.o, vmsdbgout.o)
4956 (xcoffout.o, godump.o, emit-rtl.o, real.o, realmpfr.o, dfp.o)
4957 (fixed-value.o, jump.o, simplify-rtx.o, symtab.o, cgraph.o)
4958 (cgraphunit.o, cgraphclones.o, cgraphbuild.o, varpool.o, ipa.o)
4959 (ipa-profile.o, ipa-devirt.o, ipa-prop.o, ipa-ref.o, ipa-cp.o)
4960 (ipa-split.o, ipa-inline.o, ipa-inline-analysis.o)
4961 (ipa-inline-transform.o, ipa-utils.o, ipa-reference.o)
4962 (ipa-pure-const.o, coverage.o, cselib.o, cse.o, dce.o, dumpfile.o)
4963 (dse.o, fwprop.o, web.o, ree.o, cprop.o, gcse.o, store-motion.o)
4964 (resource.o, lcm.o, mode-switching.o, tree-ssa-dce.o)
4965 (tree-call-cdce.o, tree-ssa-ccp.o, tree-ssa-strlen.o, tree-sra.o)
4966 (tree-switch-conversion.o, tree-complex.o, tree-emutls.o)
4967 (tree-vect-generic.o, df-core.o, df-problems.o, df-scan.o)
4968 (regstat.o, valtrack.o, var-tracking.o, profile.o, mcf.o)
4969 (tree-profile.o, value-prof.o, loop-doloop.o, alloc-pool.o)
4970 (auto-inc-dec.o, cfg.o, cfghooks.o, cfgexpand.o, cfgrtl.o)
4971 (cfganal.o, cfgbuild.o, cfgcleanup.o, cfgloop.o, cfgloopanal.o)
4972 (graphds.o, loop-iv.o, loop-invariant.o, cfgloopmanip.o)
4973 (loop-init.o, loop-unswitch.o, loop-unroll.o, dominance.o)
4974 (et-forest.o, combine.o, reginfo.o, bitmap.o, vec.o, hash-table.o)
4975 (reload.o, reload1.o, rtlhooks.o, postreload.o, postreload-gcse.o)
4976 (caller-save.o, bt-load.o, reorg.o, alias.o, stack-ptr-mod.o)
4977 (init-regs.o, ira-build.o, ira-costs.o, ira-conflicts.o)
4978 (ira-color.o, ira-emit.o, ira-lives.o, ira.o, lra.o)
4979 (lra-assigns.o, lra-coalesce.o, lra-constraints.o)
4980 (lra-eliminations.o, lra-lives.o, lra-spills.o, regmove.o)
4981 (combine-stack-adj.o, compare-elim.o, ddg.o, modulo-sched.o)
4982 (haifa-sched.o, sched-deps.o, sched-rgn.o, sched-ebb.o)
4983 (sched-vis.o, sel-sched.o, sel-sched-dump.o, sel-sched-ir.o)
4984 (final.o, recog.o, reg-stack.o, sreal.o, predict.o, lists.o)
4985 (bb-reorder.o, tracer.o, timevar.o, regcprop.o, regrename.o)
4986 (ifcvt.o, params.o, pointer-set.o, hooks.o, pretty-print.o)
4987 (errors.o, dbgcnt.o, lower-subreg.o, target-globals.o)
4988 (hw-doloop.o, file-find.o, context.o, $(common_out_object_file))
4989 (insn-attrtab.o, insn-automata.o, insn-dfatab.o, insn-emit.o)
4990 (insn-enums.o, insn-extract.o, insn-latencytab.o, insn-modes.o)
4991 (insn-opinit.o, insn-output.o, insn-peep.o, insn-preds.o)
4992 (insn-recog.o, intl.o, cppbuiltin.o, cppdefault.o, gcov.o)
4993 (gcov-dump.o): Remove.
4994 (default-c.o): Use COMPILE and POSTCOMPILE.
4995 (CFLAGS-gcc.o): New variable.
4996 ($(common_out_object_file)): Use COMPILE and POSTCOMPILE.
4998 2013-09-25 Tom Tromey <tromey@redhat.com>
5000 * Makefile.in (c-family/cppspec.o, c-family/c-common.o)
5001 (c-family/c-cppbuiltin.o, c-family/c-dump.o, c-family/c-format.o)
5002 (c-family/c-gimplify.o, c-family/c-lex.o, c-family/c-omp.o)
5003 (c-family/c-opts.o, c-family/c-pch.o, c-family/c-ppoutput.o)
5004 (c-family/c-pragma.o, c-family/c-pretty-print.o)
5005 (c-family/c-semantics.o, c-family/c-ada-spec.o)
5006 (c-family/array-notation-common.o, c-family/stub-objc.o)
5007 (c-family/c-ubsan.o): Remove.
5009 2013-09-25 Tom Tromey <tromey@redhat.com>
5011 * Makefile.in (C_TREE_H): Reference c/c-tree.h.
5013 2013-09-25 Tom Tromey <tromey@redhat.com>
5015 * Makefile.in (DRIVER_DEFINES): Use $(and), not shell code,
5016 to add -DENABLE_SHARED_LIBGCC.
5017 (gcc.o): Don't use subshell.
5019 2013-09-25 Tom Tromey <tromey@redhat.com>
5021 * Makefile.in (OUTPUT_OPTION): Define as "-o $@".
5022 * configure.ac: Don't invoke AM_PROG_CC_C_O.
5023 (NO_MINUS_C_MINUS_O, OUTPUT_OPTION): Don't subst.
5024 * configure, config.in: Rebuild.
5026 2013-09-25 Tom Tromey <tromey@redhat.com>
5028 * Makefile.in (CCDEPMODE, DEPDIR, depcomp, COMPILE.base)
5029 (COMPILE, POSTCOMPILE): New variables.
5030 (.cc.o .c.o): Use COMPILE, POSTCOMPILE.
5031 (DEPFILES): New variable.
5032 Include ".Po" files.
5033 * configure.ac: Add checks for dependency checking.
5034 * configure, aclocal.m4: Regenerate.
5036 2013-09-25 Tom Tromey <tromey@redhat.com>
5038 * Makefile.in (ALL_HOST_BACKEND_OBJS): Add lto-wrapper.o.
5039 ($(ALL_HOST_OBJS)): Move order-only dependency to end of file.
5041 2013-09-25 Tom Tromey <tromey@redhat.com>
5043 * Makefile.in (generated_files): Add options.h,
5044 target-hooks-def.h, insn-opinit.h,
5045 common/common-target-hooks-def.h, pass-instances.def,
5046 c-family/c-target-hooks-def.h.
5048 2013-09-25 Jeff Law <law@redhat.com>
5050 * tree-ssa-threadedge.c (thread_across_edge): Use foo.last () rather
5051 than foo[foo.length () - 1] to access last member in a vec.
5052 * tree-ssa-threadupdate.c (register_jump_thread): Similarly.
5054 2013-09-25 Richard Biener <rguenther@suse.de>
5057 * tree.c (iterative_hash_expr): Remove MEM_REF special handling.
5059 2013-09-25 Jan Hubicka <jh@suse.cz>
5061 * cgraph.c (cgraph_resolve_speculation): Use semantical equivalency
5064 2013-09-25 Marek Polacek <polacek@redhat.com>
5067 * ubsan.c (ubsan_type_descriptor): Handle IDENTIFIER_NODEs
5068 when determining the type name.
5070 2013-09-24 Oleg Endo <olegendo@gcc.gnu.org>
5072 * config/sh/sh.md: Fix formatting.
5074 2013-09-24 Xinliang David Li <davidxl@google.com>
5076 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Check
5077 max peel iterations parameter.
5078 * param.def: New parameter.
5079 * doc/invoke.texi: Document New parameter.
5081 2013-09-24 Christophe Lyon <christophe.lyon@linaro.org>
5083 * gimple-pretty-print.c: Various whitespace tweaks.
5084 * tree-core.h: Likewise.
5085 * tree-pretty-print.c: Likewise.
5086 * tree-ssa-alias.c: Likewise.
5087 * tree-ssa-copy.c: Likewise.
5088 * tree-ssanames.c: Likewise.
5089 * tree-ssanames.h: Likewise.
5090 * tree-vrp.c: Likewise.
5092 2013-09-24 Alan Modra <amodra@gmail.com>
5096 * stmt.c (expand_asm_operands): Call expand_expr with EXPAND_MEMORY
5097 for output operands that disallow regs. Don't use EXPAND_WRITE on
5100 2013-09-24 Richard Biener <rguenther@suse.de>
5103 * tree.c (reference_alias_ptr_type): Move ...
5104 * alias.c (reference_alias_ptr_type): ... here and implement
5105 in terms of the new reference_alias_ptr_type_1.
5106 (ref_all_alias_ptr_type_p): New helper.
5107 (get_deref_alias_set_1): Drop flag_strict_aliasing here,
5108 use ref_all_alias_ptr_type_p.
5109 (get_deref_alias_set): Add flag_strict_aliasing check here.
5110 (reference_alias_ptr_type_1): New function, split out from ...
5111 (get_alias_set): ... here.
5112 (alias_ptr_types_compatible_p): New function.
5113 * alias.h (reference_alias_ptr_type): Declare.
5114 (alias_ptr_types_compatible_p): Likewise.
5115 * tree.h (reference_alias_ptr_type): Remove.
5116 * fold-const.c (operand_equal_p): Use alias_ptr_types_compatible_p
5117 to compare MEM_REF alias types.
5119 2013-09-24 Richard Biener <rguenther@suse.de>
5121 * tree-vrp.c (vrp_finalize): Check for SSA name presence.
5123 2013-09-23 Michael Meissner <meissner@linux.vnet.ibm.com>
5125 * config/rs6000/rs6000.c (rs6000_vector_reload): Delete, combine
5126 reload helper function arrays into a single array reg_addr.
5127 (reload_fpr_gpr): Likewise.
5128 (reload_gpr_vsx): Likewise.
5129 (reload_vsx_gpr): Likewise.
5130 (struct rs6000_reg_addr): Likewise.
5131 (reg_addr): Likewise.
5132 (rs6000_debug_reg_global): Change rs6000_vector_reload,
5133 reload_fpr_gpr, reload_gpr_vsx, reload_vsx_gpr uses to reg_addr.
5134 (rs6000_init_hard_regno_mode_ok): Likewise.
5135 (rs6000_secondary_reload_direct_move): Likewise.
5136 (rs6000_secondary_reload): Likewise.
5138 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add new
5139 constraints: wu, ww, and wy. Repurpose wv constraint added during
5140 power8 changes. Put wg constraint in alphabetical order.
5142 * config/rs6000/rs6000.opt (-mvsx-scalar-float): New debug switch
5143 for future work to add ISA 2.07 VSX single precision support.
5144 (-mvsx-scalar-double): Change default from -1 to 1, update
5145 documentation comment.
5146 (-mvsx-scalar-memory): Rename debug switch to -mupper-regs-df.
5147 (-mupper-regs-df): New debug switch to control whether DF values
5148 can go in the traditional Altivec registers.
5149 (-mupper-regs-sf): New debug switch to control whether SF values
5150 can go in the traditional Altivec registers.
5152 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print wu, ww,
5154 (rs6000_init_hard_regno_mode_ok): Use ssize_t instead of int for
5155 loop variables. Rename -mvsx-scalar-memory to -mupper-regs-df.
5156 Add new constraints, wu/ww/wy. Repurpose wv constraint.
5157 (rs6000_debug_legitimate_address_p): Print if we are running
5158 before, during, or after reload.
5159 (rs6000_secondary_reload): Add a comment.
5160 (rs6000_opt_masks): Add -mupper-regs-df, -mupper-regs-sf.
5162 * config/rs6000/constraints.md (wa constraint): Sort w<x>
5163 constraints. Update documentation string.
5164 (wd constraint): Likewise.
5165 (wf constraint): Likewise.
5166 (wg constraint): Likewise.
5167 (wn constraint): Likewise.
5168 (ws constraint): Likewise.
5169 (wt constraint): Likewise.
5170 (wx constraint): Likewise.
5171 (wz constraint): Likewise.
5172 (wu constraint): New constraint for ISA 2.07 SFmode scalar
5174 (ww constraint): Likewise.
5175 (wy constraint): Likewise.
5176 (wv constraint): Repurpose ISA 2.07 constraint that we did not use
5177 in the previous submissions.
5178 * doc/md.texi (PowerPC and IBM RS6000): Likewise.
5180 2013-09-23 Richard Sandiford <rdsandiford@googlemail.com>
5182 * doc/rtl.texi (REG_NOTES): Say that int_list can also be used.
5183 (REG_BR_PROB): Say that the probability is stored in an int_list.
5184 * reg-notes.def: Update commentary to mention INT_LIST.
5185 * rtl.def (EXPR_LIST, INSN_LIST): Capitalize comments.
5186 (INT_LIST): New rtx.
5187 * rtl.h (add_int_reg_note, add_shallow_copy_of_reg_note): Declare.
5188 * rtlanal.c (int_reg_note_p): New function.
5189 (alloc_reg_note): Assert that the note does not have an int argument.
5190 (add_int_reg_note, add_shallow_copy_of_reg_note): New functions.
5191 * combine.c (distribute_notes): Use add_shallow_copy_of_rtx.
5192 * cse.c (cse_process_notes_1): Expect REG_EQUAL to be an EXPR_LIST
5193 rather than an INSN_LIST. Handle INT_LIST.
5194 * ifcvt.c (cond_exec_process_insns): Take the probability as an int
5195 rather than an rtx. Use gen_rtx_INT_LIST to create a REG_BR_PROB note.
5196 (cond_exec_process_if_block): Use XINT to extract REG_BR_PROB values.
5197 Manipulate them as ints rather than rtxes.
5198 * reg-stack.c (subst_asm_stack_regs): Only handle EXPR_LIST notes.
5199 * regmove.c (copy_src_to_dest): Likewise.
5200 * sched-vis.c (print_insn_with_notes): Handle INT_LIST.
5202 * config/i386/winnt.c (i386_pe_seh_unwind_emit): Sink pat assignment
5203 into the cases that need it.
5204 * config/arm/arm.c (arm_unwind_emit): Likewise.
5206 * asan.c (asan_clear_shadow): Use add_int_reg_note for REG_BR_PROB.
5207 * emit-rtl.c (try_split, emit_copy_of_insn_after): Likewise.
5208 * loop-doloop.c (add_test, doloop_modify): Likewise.
5209 * loop-unswitch.c (compare_and_jump_seq): Likewise.
5210 * optabs.c (emit_cmp_and_jump_insn_1): Likewise.
5211 * predict.c (combine_predictions_for_insn): Likewise.
5212 * print-rtl.c (print_rtx): Handle INT_LIST.
5213 * config/aarch64/aarch64.c (aarch64_emit_unlikely_jump): Likewise.
5214 * config/alpha/alpha.c (emit_unlikely_jump): Likewise.
5215 * config/arm/arm.c (emit_unlikely_jump): Likewise.
5216 * config/i386/i386.c (ix86_expand_split_stack_prologue): Likewise.
5217 (ix86_split_fp_branch, predict_jump): Likewise.
5218 * config/rs6000/rs6000.c (emit_unlikely_jump): Likewise.
5219 * config/sh/sh.c (expand_cbranchsi4): Likewise.
5220 * config/spu/spu.c (ea_load_store_inline): Likewise.
5222 * cfgbuild.c (compute_outgoing_frequencies): Use XINT to access the
5223 value of a REG_BR_PROB note.
5224 * cfgrtl.c (force_nonfallthru_and_redirect): Likewise.
5225 (update_br_prob_note, rtl_verify_edges, purge_dead_edges): Likewise.
5226 * emit-rtl.c (try_split): Likewise.
5227 * predict.c (br_prob_note_reliable_p): Likewise.
5228 (invert_br_probabilities, combine_predictions_for_insn): Likewise.
5229 * reorg.c (mostly_true_jump): Likewise.
5230 * config/bfin/bfin.c (cbranch_predicted_taken_p): Likewise.
5231 * config/frv/frv.c (frv_print_operand_jump_hint): Likewise.
5232 * config/i386/i386.c (ix86_print_operand): Likewise.
5233 * config/ia64/ia64.c (ia64_print_operand): Likewise.
5234 * config/mmix/mmix.c (mmix_print_operand): Likewise.
5235 * config/rs6000/rs6000.c (output_cbranch): Likewise.
5236 * config/s390/s390.c (s390_expand_tbegin): Likewise.
5237 * config/sh/sh.c (sh_print_operand, sh_adjust_cost): Likewise.
5238 * config/sparc/sparc.c (output_cbranch): Likewise.
5239 * config/spu/spu.c (get_branch_target): Likewise.
5240 * config/tilegx/tilegx.c (cbranch_predicted_p): Likewise.
5241 * config/tilepro/tilepro.c (cbranch_predicted_p): Likewise.
5243 2013-09-23 Jan Hubicka <jh@suse.cz>
5245 * ipa-cp.c (ipa_get_indirect_edge_target_1): Add sanity check
5247 * ipa-utils.h (possible_polymorphic_call_target_p): New function.
5248 * ipa-devirt.c (possible_polymorphic_call_target_p): Be tolerant
5250 * gimple-fold.c: Include ipa-utils.h and gimple-pretty-print.h
5251 (gimple_fold_call): Dump inconsistent devirtualizations; add
5252 sanity check for type based devirtualizations.
5253 * ipa-prop.c: Include ipa-utils.h
5254 (ipa_intraprocedural_devirtualization): Add sanity check.
5255 (try_make_edge_direct_virtual_call): Likewise.
5257 2013-09-23 Eric Botcazou <ebotcazou@adacore.com>
5259 * tree-ssa-ccp.c (insert_clobber_before_stack_restore): Recurse on copy
5260 assignment statements.
5262 2013-09-23 Kugan Vivekanandarajah <kuganv@linaro.org>
5264 * gimple-pretty-print.c (dump_ssaname_info): New function.
5265 (dump_gimple_phi): Call it.
5266 (pp_gimple_stmt_1): Likewise.
5267 * tree-core.h (tree_ssa_name): New union ssa_name_info_type field.
5268 (range_info_def): Declare.
5269 * tree-pretty-print.c (pp_double_int): New function.
5270 (dump_generic_node): Call it.
5271 * tree-pretty-print.h (pp_double_int): Declare.
5272 * tree-ssa-alias.c (dump_alias_info): Check pointer type.
5273 * tree-ssanames.h (range_info_def): New structure.
5274 (value_range_type): Move definition here.
5275 (set_range_info, value_range_type, duplicate_ssa_name_range_info):
5277 * tree-ssanames.c (make_ssa_name_fn): Check pointer type at
5279 (set_range_info): New function.
5280 (get_range_info): Likewise.
5281 (duplicate_ssa_name_range_info): Likewise.
5282 (duplicate_ssa_name_fn): Check pointer type and call
5283 duplicate_ssa_name_range_info.
5284 * tree-ssa-copy.c (fini_copy_prop): Likewise.
5285 * tree-vrp.c (value_range_type): Remove definition, now in
5287 (vrp_finalize): Call set_range_info to update value range of SSA_NAMEs.
5288 * tree.h (SSA_NAME_PTR_INFO): Macro changed to access via union.
5289 (SSA_NAME_RANGE_INFO): New macro.
5291 2013-09-23 Richard Biener <rguenther@suse.de>
5293 PR tree-optimization/58464
5294 * tree-ssa-pre.c (phi_trans_lookup): Remove.
5295 (phi_trans_add): Change to add conditionally on being not present.
5296 (phi_translate_1): Remove recursion detection here.
5297 (phi_translate): Pre-seed the cache with NULL to catch
5298 recursion here in a more generic way.
5299 (bitmap_find_leader): Adjust comment.
5300 (get_representative_for): Dump value-numbers.
5301 (create_expression_by_pieces): Likewise.
5302 (insert_into_preds_of_block): Likewise.
5304 2013-09-23 Christian Bruel <christian.bruel@st.com>
5307 * config/sh/sh.md (movsf_ie): Allow fpul_operand.
5308 * config/sh/predicate.md (arith_reg_operand): Disallow FPUL_REG.
5310 2013-09-23 James Greenhalgh <james.greenhalgh@arm.com>
5313 2013-09-20 Renlin Li <renlin.li@arm.com>
5315 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
5317 (aarch64_expand_epilogue): Likewise.
5318 (aarch64_legitimize_reload_address): Likewise.
5320 2013-09-22 Eric Botcazou <ebotcazou@adacore.com>
5322 * gimplify.c (gimplify_asm_expr): Reset the TREE_CHAIN of clobbers to
5323 NULL_TREE before pushing them onto the vector. Likewise for labels.
5325 2013-09-21 Eric Botcazou <ebotcazou@adacore.com>
5327 * config/ia64/predicates.md (ia64_cbranch_operator): Accept unordered
5328 comparison operators when -fno-trapping-math is in effect.
5329 * config/ia64/ia64.c (ia64_expand_compare): Add support for unordered
5330 comparison operators in TFmode and assert that unsupported operators
5332 (ia64_print_operand): Likewise.
5334 2013-09-21 Jan Hubicka <jh@suse.cz>
5336 * x86-tune.def (partial_reg_stall): Disable for CoreI7 and newer.
5337 (sse_typeless_stores): Enable for core
5338 (sse_load0_by_pxor): Likewise.
5339 (four_jump_limit): Disable for core.
5340 (pad_returns): Likewise.
5341 (avoid_vector_decode): Likewise.
5342 (fuse_cmp_and_branch): Enable for cores.
5343 * i386.c (x86_accumulate_outgoing_args): Disable for cores.
5345 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
5348 * config/pa/pa.c (pa_option_override): Disable auto increment and
5349 decrement instructions until reload is completed.
5351 * config/pa/pa-linux.h (TARGET_OS_CPP_BUILTINS): Define
5352 __GCC_HAVE_SYNC_COMPARE_AND_SWAP_1, __GCC_HAVE_SYNC_COMPARE_AND_SWAP_2,
5353 and __GCC_HAVE_SYNC_COMPARE_AND_SWAP_4.
5355 2013-09-20 DJ Delorie <dj@redhat.com>
5356 Nick Clifton <nickc@redhat.com>
5358 * config/rl78/rl78.c: Various whitespace and comment tweaks.
5359 (need_to_save): Save bank 0 on interrupts.
5360 (characterize_address): Strip far address wrappers.
5361 (rl78_as_legitimate_address): Likewise.
5362 (transcode_memory_rtx): Likewise.
5363 (rl78_peep_movhi_p): Disable this peephole after devirt.
5364 (rl78_propogate_register_origins): Forget all origins when a
5366 * config/rl78/rl78-virt.md: Various whitespace tweaks.
5367 * config/rl78/rl78-real.md: Various whitespace tweaks. Additional
5369 * config/rl78/rl78.md (sel_rb): Disable for G10 just in case.
5370 * config/rl78/rl78-expand.md (movqi): Check for subregs of consts.
5371 * config/rl78/rl78.h (LINK_SPEC): Pass -gc-sections unless
5373 * config/rl78/constraints.md: Various whitespace and paren tweaks.
5375 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
5377 * config/pa/pa.md: In "scc" insn patterns, change output template to
5378 handle const0_rtx in reg_or_0_operand operands.
5380 2013-09-20 Martin Husemann <martin@NetBSD.org>
5383 * config/vax/vax.c (vax_output_int_move): Use D format specifier.
5384 * config/vax/vax.md (ashldi3, <unnamed>): Ditto.
5386 2013-09-20 Richard Biener <rguenther@suse.de>
5389 * tree-scalar-evolution.c (struct scev_info_str): Shrink by
5390 remembering SSA name version and block index.
5391 (new_scev_info_str): Adjust.
5392 (hash_scev_info): Likewise. Also hash the block index.
5393 (eq_scev_info): Adjust.
5394 (find_var_scev_info): Likewise.
5395 (struct instantiate_cache_entry): Remove.
5396 (struct instantiate_cache_type): Use a htab to map name, block
5398 (instantiate_cache_type::~instantiate_cache_type): Adjust.
5399 (get_instantiated_value_entry): Likewise.
5400 (hash_idx_scev_info, eq_idx_scev_info): New functions.
5401 (instantiate_scev_name): Adjust.
5403 2013-09-20 Jeff Law <law@redhat.com>
5405 * tree-ssa-dom.c (record_temporary_equivalences): Add comment.
5407 2013-09-20 Yufeng Zhang <yufeng.zhang@arm.com>
5409 * config/aarch64/aarch64-builtins.c (aarch64_simd_expand_args):
5410 Call aarch64_simd_expand_args to update op[argc].
5412 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
5414 * plugin.c (parse_plugin_arg_opt): Accept equal sign inside
5417 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
5419 * gengtype.c (file_rules): Added rule for *.cc files.
5420 (get_output_file_with_visibility): Give fatal message when no
5423 2013-09-20 Renlin Li <renlin.li@arm.com>
5425 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use plus_constant.
5426 (aarch64_expand_epilogue): Likewise.
5427 (aarch64_legitimize_reload_address): Likewise.
5429 2013-09-20 Bernd Edlinger <bernd.edlinger@hotmail.de>
5432 * expr.c (expand_assignment): Remove misalignp code path.
5434 2013-09-20 Marek Polacek <polacek@redhat.com>
5437 * ubsan.c (get_ubsan_type_info_for_type): Use TYPE_SIZE instead of
5438 TYPE_PRECISION. Add asserts.
5440 2013-09-20 Richard Biener <rguenther@suse.de>
5442 PR tree-optimization/58453
5443 * tree-loop-distribution.c (distribute_loop): Apply the cost
5444 model for -ftree-loop-distribute-patterns, too.
5446 2013-09-20 Richard Biener <rguenther@suse.de>
5449 * tree-chrec.h (build_polynomial_chrec): Use gcc_checking_assert,
5450 make type comparison less strict.
5452 2013-09-20 Alan Modra <amodra@gmail.com>
5454 * configure: Regenerate.
5455 * aclocal.m4: Regenerate.
5457 2013-09-20 Marek Polacek <polacek@redhat.com>
5460 * doc/extend.texi: Document that attribute used is meant to be used
5461 on variables with static storage duration.
5463 2013-09-19 Jakub Jelinek <jakub@redhat.com>
5465 PR tree-optimization/58472
5466 * tree-vect-stmts.c (vectorizable_store, vectorizable_load): For
5467 simd_lane_access set inv_p = false.
5468 * omp-low.c (lower_rec_input_clauses): Set TREE_NO_WARNING on
5469 the simduid magic VAR_DECL.
5471 2013-09-19 Jan Hubicka <jh@suse.cz>
5473 * i386.c (generic_memcpy, generic_memset): Fix 32bit template.
5475 2013-09-17 Jeff Law <law@redhat.com>
5477 * tree-ssa-dom.c (record_temporary_equivalences): New function
5478 split out of dom_opt_dom_walker::after_dom_children.
5479 (dom_opt_dom_walker::thread_across_edge): Move common code
5480 in here from dom_opt_dom_walker::after_dom_children.
5481 (dom_opt_dom_walker::after_dom_children): Corresponding simplifictions.
5483 2013-09-19 Jan Hubicka <jh@suse.cz>
5485 * i386.h (TARGET_GENERIC32, TARGET_GENERIC64): Remove.
5486 (TARGET_GENERIC): Use PROCESOR_GENERIC
5487 (enum processor_type): Unify generic32 and 64.
5488 * i386.md (cpu): Likewise.
5489 * x86-tune.def (use_leave): Enable for generic32.
5490 (avoid_vector_decode, slow_imul_imm32_mem, slow_imul_imm8): Likewise.
5491 * athlon.md: Change generic64 to generic in all occurences.
5492 * i386-c.c (ix86_target_macros_internal): Unify generic64 and 32.
5493 (ix86_target_macros_internal): Likewise.
5494 * driver-i386.c (host_detect_local_cpu): Likewise.
5495 * i386.c (generic64_memcpy, generic64_memset, generic64_cost): Rename
5497 (generic_memcpy, generic_memset, generic_cost): This one.
5498 (generic32_memcpy, generic32_memset, generic32_cost): Remove.
5499 (m_GENERIC32, m_GENERIC64): Remove.
5500 (m_GENERIC): Turn into one flag.
5501 (processor_target): Unify generic tunnings.
5502 (ix86_option_override_internal): Replace generic32/64 by generic.
5503 (ix86_issue_rate): Likewise.
5504 (ix86_adjust_cost): Likewise.
5506 2013-09-19 Jan Hubicka <jh@suse.cz>
5508 * cgraph.c (cgraph_create_edge_1): Avoid uninitialized read
5509 of speculative flag.
5511 2013-09-19 Jakub Jelinek <jakub@redhat.com>
5513 * omp-low.c (expand_omp_sections): Always pass len - 1 to
5514 GOMP_sections_start, even if !exit_reachable.
5516 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
5518 * lra-constraints.c (need_for_all_save_p): Use macro
5519 HARD_REGNO_CALL_PART_CLOBBERED.
5520 * lra-lives.c (check_pseudos_live_through_calls): Use the macro to
5521 set up pseudo conflict hard regs.
5523 2013-09-18 Michael Meissner <meissner@linux.vnet.ibm.com>
5526 * config/rs6000/paired.md (movmisalignv2sf): Fix to allow memory
5529 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
5531 PR rtl-optimization/58438
5532 * lra.c (lra): Clear lra_optional_reload_pseudos in upper loop.
5533 * lra-constraints.c (undo_optional_reloads): Keep optional reloads
5534 from previous subpasses.
5536 2013-09-18 Richard Earnshaw <rearnsha@arm.com>
5538 * arm.c (arm_get_frame_offsets): Validate architecture supports
5539 LDRD/STRD before accepting the tuning preference.
5540 (arm_expand_prologue): Likewise.
5541 (arm_expand_epilogue): Likewise.
5543 2013-09-18 Richard Biener <rguenther@suse.de>
5545 PR tree-optimization/58417
5546 * tree-chrec.c (chrec_fold_plus_1): Assert that we do not
5547 have chrecs with symbols defined in the loop as operands.
5548 (chrec_fold_multiply): Likewise.
5549 * tree-scalar-evolution.c (interpret_rhs_expr): Instantiate
5550 parameters before folding binary operations.
5551 (struct instantiate_cache_entry_hasher): Remove.
5552 (struct instantiate_cache_type): Use a pointer-map.
5553 (instantiate_cache_type::instantiate_cache_type): New function.
5554 (instantiate_cache_type::get): Likewise.
5555 (instantiate_cache_type::set): Likewise.
5556 (instantiate_cache_type::~instantiate_cache_type): Adjust.
5557 (get_instantiated_value_entry): Likewise.
5558 (global_cache): New global.
5559 (instantiate_scev_r, instantiate_scev_poly, instantiate_scev_binary,
5560 instantiate_array_ref, instantiate_scev_convert, instantiate_scev_3,
5561 instantiate_scev_2, instantiate_scev_1): Do not pass along cache.
5562 (instantiate_scev_name): Adjust.
5563 (instantiate_scev): Construct global instead of local cache.
5564 (resolve_mixers): Likewise.
5566 2013-09-18 Daniel Morris <danielm@ecoscentric.com>
5567 Paolo Carlini <paolo.carlini@oracle.com>
5570 * doc/implement-cxx.texi: Fix references to the C++ standards.
5572 2013-09-18 Jakub Jelinek <jakub@redhat.com>
5574 * omp-low.c (copy_var_decl): Copy DECL_ATTRIBUTES.
5575 * tree-vect-data-refs.c (vect_analyze_data_refs): For
5576 simd_lane_access drs, update also DR_ALIGNED_TO.
5578 2013-09-18 Marek Polacek <polacek@redhat.com>
5581 * doc/extend.texi: Document no_sanitize_undefined attribute.
5582 * builtins.c (fold_builtin_0): Don't sanitize function if it has the
5583 no_sanitize_undefined attribute.
5585 2013-09-18 Nick Clifton <nickc@redhat.com>
5587 * config/msp430/msp430.h (ASM_SPEC): Pass -md on to the assembler.
5588 (ASM_DECLARE_FUNCTION_NAME): Define.
5590 2013-09-17 Trevor Saunders <tsaunders@mozilla.com>
5592 * compare-elim.c (find_comparison_dom_walker): New class
5593 (find_comparisons_in_bb): Rename to
5594 find_comparison_dom_walker::before_dom_children
5595 (find_comparisons): Adjust
5596 * domwalk.c (walk_dominator_tree): Rename to dom_walker::walk, and
5598 (init_walk_dominator_tree, fini_walk_dominator_tree): Remove
5599 * domwalk.h (dom_walk_data): Convert it To a class dom_walker.
5600 (init_walk_dominator_tree): Remove declaration.
5601 (fini_walk_dominator_tree): Remove declaration.
5602 * fwprop.c (single_def_use_dom_walker): New class
5603 (single_def_use_enter_block): Convert to
5604 single_def_use_dom_walker::before_dom_children.
5605 (single_def_use_leave_block): Convert to
5606 single_def_use_dom_walker::after_dom_children.
5607 (build_single_def_use_links): Adjust.
5608 * gimple-ssa-strength-reduction.c (find_candidates_dom_walker): New
5610 (find_candidates_in_block): Convert to
5611 find_candidates_dom_walker::before_dom_children.
5612 (execute_strength_reduction): Adjust.
5613 * graphite-sese-to-poly.c (struct bsc, build_sese_conditions): Remove.
5614 (sese_dom_walker): New class.
5615 (sese_dom_walker::sese_dom_walker): New constructor.
5616 (sese_dom_walker::~sese_dom_walker): New destructor.
5617 (build_sese_conditions_before): Convert to
5618 sese_dom_walker::before_dom_children.
5619 (build_sese_conditions_after): Convert to
5620 sese_dom_walker::after_dom_children.
5621 (build_poly_scop): Adjust
5622 * tree-into-ssa.c (rewrite_dom_walker): New class
5623 (rewrite_enter_block): Convert to
5624 rewrite_dom_walker::before_dom_children.
5625 (rewrite_leave_block): Convert to
5626 rewrite_dom_walker::after_dom_children.
5627 (rewrite_update_dom_walker): New class.
5628 (rewrite_update_enter_block): Convert to
5629 rewrite_update_dom_walker::before_dom_children.
5630 (rewrite_update_leave_block): Convert to
5631 rewrite_update_dom_walker::after_dom_children.
5632 (rewrite_blocks, rewrite_into_ssa): Adjust.
5633 (mark_def_dom_walker): New class.
5634 (mark_def_dom_walker::mark_def_dom_walker): New constructor.
5635 (mark_def_dom_walker::~mark_def_dom_walker): New destructor.
5636 (mark_def_sites_blocks): Convert to
5637 mark_def_dom_walker::before_dom_children.
5638 (mark_def_site_blocks): Remove.
5639 * tree-ssa-dom.c (dom_opt_dom_walker): New class.
5640 (tree_ssa_dominator_optimize): Adjust.
5641 (dom_thread_across_edge): Convert to method
5642 dom_opt_dom_walker::thread_across_edge.
5643 (dom_opt_enter_block): Convert to member function
5644 dom_opt_dom_walker::before_dom_children.
5645 (dom_opt_leave_block): Convert to member function
5646 dom_opt_dom_walker::after_dom_children.
5647 * tree-ssa-dse.c (dse_dom_walker): New class.
5648 (dse_enter_block): Convert to member function
5649 dse_dom_walker::before_dom_children.
5650 (tree_ssa_dse): Adjust.
5651 * tree-ssa-loop-im.c (invariantness_dom_walker): New class.
5652 (determine_invariantness_stmt): Convert to method
5653 invariantness_dom_walker::before_dom_children.
5654 (determine_invariantness): Remove
5655 (move_computations_dom_walker): New class.
5656 (move_computations_stmt): Convert to method
5657 move_computations_dom_walker::before_dom_children.
5658 (move_computations, tree_ssa_lim): Adjust.
5659 * tree-ssa-phiopt.c (nontrapping_dom_walker): New class.
5660 (nt_init_block): Convert to method
5661 notrappping_dom_walker::before_dom_children.
5662 (nt_fini_block): Convert to method
5663 method nontrapping_dom_walker::after_dom_children.
5664 (get_non_trapping): Adjust.
5665 * tree-ssa-pre.c (eliminate_dom_walker): New class.
5666 (eliminate_bb): Convert to method
5667 eliminate_dom_walker::before_dom_children.
5668 (eliminate_leave_block): Convert to method
5669 eliminate_dom_walker::after_dom_children.
5670 (eliminate): Adjust.
5671 * tree-ssa-strlen.c (strlen_dom_walker): New class.
5672 (strlen_enter_block): Convert to method
5673 strlen_dom_walker::before_dom_children.
5674 (strlen_leave_block): Convert to method
5675 method strlen_dom_walker::after_dom_children.
5676 (tree_ssa_strlen): Adjust.
5677 * tree-ssa-uncprop.c (uncprop_dom_walker): New class.
5678 (tree_ssa_uncprop): Adjust.
5679 (uncprop_leave_block): Convert to method
5680 uncprop_dom_walker::after_dom_children.
5681 (uncprop_leave_block): Convert to method
5682 uncprop_dom_walker::before_dom_children.
5684 2013-09-18 Bin Cheng <bin.cheng@arm.com>
5686 * config/arm/arm.c (thumb1_reorg): Search for flag setting insn before
5687 branch in same basic block. Check both src and dest of the move insn.
5689 2013-09-17 Nick Clifton <nickc@redhat.com>
5691 * config/rl78/rl78-real.md (bf): New pattern.
5693 * config/rl78/rl78.c (rl78_print_operand_1): Handle %B.
5694 (rl78_print_operand): Do not put a # before a %B.
5695 * config/rl78/rl78.opt: Tweak doc strings.
5697 2013-09-17 DJ Delorie <dj@redhat.com>
5699 * config/rl78/constraints.md (Wcv): Allow up to $r31.
5700 * config/rl78/rl78.c (rl78_asm_file_start: Likewise.
5701 (rl78_option_override): Likewise, if -mallregs.
5702 (is_virtual_register): Likewise.
5703 * config/rl78/rl78.h (reg_class): Extend VREGS to $r31.
5704 (REGNO_OK_FOR_BASE_P): Likewise.
5705 * config/rl78/rl78.opt (-mallregs): New.
5707 2013-09-17 Nick Clifton <nickc@redhat.com>
5709 * config/rl78/rl78.c (need_to_save): Change return type to bool.
5710 For interrupt functions: save all call clobbered registers if the
5711 interrupt handler is not a leaf function.
5712 (rl78_expand_prologue): Always recompute the frame information.
5713 For interrupt functions: only select bank 0 if one of the bank 0
5714 registers is going to be psuhed.
5716 2013-09-17 DJ Delorie <dj@redhat.com>
5718 * config/rl78/constraints.md: For each W* constraint, rename to C*
5719 and create a W* constraint that checks for an optional ES: prefix
5721 * config/rl78/rl78.md (UNS_ES_ADDR): New.
5722 (es_addr): New. Used to wrap far addresses.
5723 * config/rl78/rl78-protos.h (rl78_es_addr): New.
5724 (rl78_es_base): New.
5725 * config/rl78/rl78.c (rl78_as_legitimate_address): Accept "unspec"
5726 wrapped far addresses.
5727 (rl78_print_operand_1): Unwrap far addresses before processing.
5728 (rl78_lo16): Wrap far addresses in unspecs.
5729 (rl78_es_addr): New.
5730 (rl78_es_base): New.
5731 (insn_ok_now): Check for not-yet-wrapped far addresses.
5732 (transcode_memory_rtx): Properly re-wrap far addresses.
5734 2013-09-17 Sebastian Huber <sebastian.huber@embedded-brains.de>
5736 * config/sparc/t-rtems: Add leon3 multilibs.
5738 2013-09-17 Cong Hou <congh@google.com>
5740 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Fix a bug
5741 when checking the dot production pattern. The type of rhs operand
5742 of multiply is now checked correctly.
5744 2013-09-17 Jeff Law <law@redhat.com>
5746 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
5747 edge implied equivalences into successor phis.
5748 * tree-ssa-threadupdate.c (phi_args_equal_on_edges): Moved into
5749 here from tree-ssa-threadedge.c.
5750 (mark_threaded_blocks): When threading through a joiner, if both
5751 successors of the joiner's clone reach the same block, verify the
5752 PHI arguments are equal. If not, cancel the jump threading request.
5753 * tree-ssa-threadedge.c (phi_args_equal_on_edges): Moved into
5754 tree-ssa-threadupdate.c
5755 (thread_across_edge): Don't check PHI argument equality when
5756 threading through joiner block here.
5758 2013-09-17 Andrew MacLeod <amacleod@redhat.com>
5760 * tree-flow.h (ssa_undefined_value_p): Remove prototype.
5761 * tree-ssa.c (ssa_undefined_value_p): Move pass independent parts here.
5762 (warn_uninit, warn_uninitialized_vars,
5763 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
5764 Move to tree-ssa-uninit.c.
5765 * tree-ssa-uninit.c (ssa_undefined_value_p): Move to tree-ssa.c.
5766 (has_undefined_value_p): New. Pass dependant parts of
5767 ssa_undefined_value_p.
5768 (uninit_undefined_value_p): Use has_undefined_value_p.
5769 (warn_uninit, warn_uninitialized_vars,
5770 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
5771 Move from tree-ssa.c.
5772 * tree-ssa.h: Adjust prototypes.
5774 2013-09-17 Jan Hubicka <jh@suse.cz>
5777 * cif-code.def (FUNCTION_NOT_OPTIMIZED): New CIF code.
5778 * ipa-inline.c (can_inline_edge_p): Do not downgrade
5779 FUNCTION_NOT_OPTIMIZED.
5780 * ipa-inline-analysis.c (compute_inline_parameters): Function
5781 not optimized is not inlinable unless it is alwaysinline.
5782 (inline_analyze_function): Force calls in not optimized
5783 function not inlinable.
5785 2013-09-17 Jan Hubicka <jh@suse.cz>
5788 * ipa-devirt.c (ipa_devirt): Be ready for symtab_nonoverwritable_alias
5790 * ipa.c (function_and_variable_visibility): Likewise.
5791 * ipa-profile.c (ipa_profile): Likewise.
5793 2013-09-17 Bernd Edlinger <bernd.edlinger@hotmail.de>
5796 * cgraph.c (cgraph_function_body_availability): Check for ifunc
5797 attribute, and don't inline the resolver in this case.
5799 2013-09-17 Teresa Johnson <tejohnson@google.com>
5801 * coverage.c (get_coverage_counts): Add missing newline.
5803 2013-09-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5805 PR tree-optimization/58088
5806 * fold-const.c (mask_with_trailing_zeros): New function.
5807 (fold_binary_loc): Make sure we don't recurse infinitely
5808 when the X in (X & C1) | C2 is a tree of the form (Y * K1) & K2.
5809 Use mask_with_trailing_zeros where appropriate.
5811 2013-09-17 Yuri Rumyantsev <ysrumyan@gmail.com>
5813 * config/i386/i386.c (distance_agu_use_in_bb) : Proper initialization
5814 of 'prev' var to get better distance estimation.
5816 2013-09-17 Eric Botcazou <ebotcazou@adacore.com>
5818 * tree-inline.h (struct copy_body_data): Add transform_parameter.
5819 * tree-inline.c (is_parameter_of): New predicate.
5820 (remap_gimple_op_r): Do not propagate TREE_THIS_NOTRAP on MEM_REF if
5821 a parameter has been remapped.
5822 (copy_tree_body_r): Likewise on INDIRECT_REF and MEM_REF.
5823 (optimize_inline_calls): Initialize transform_parameter.
5824 (copy_gimple_seq_and_replace_locals): Likewise.
5825 (tree_function_versioning): Likewise.
5826 (maybe_inline_call_in_expr): Likewise.
5828 2013-09-17 Nick Clifton <nickc@redhat.com>
5830 * config/msp430/msp430-protos.h: Add prototypes for new functions.
5831 * config/msp430/msp430.c (msp430_preserve_reg_p): Add support for
5833 (is_attr_func): New function.
5834 (msp430_is_interrupt_func): New function.
5835 (is_naked_func): New function.
5836 (is_reentrant_func): New function.
5837 (is_critical_func): New function.
5838 (msp430_start_function): Add annotations for function attributes.
5839 (msp430_attr): New function.
5840 (msp430_attribute_table): New.
5841 (msp430_function_section): New function.
5842 (TARGET_ASM_FUNCTION_SECTION): Define.
5843 (msp430_builtin): New enum.
5844 (msp430_init_builtins): New function.
5845 (msp430_builtin_devl): New function.
5846 (msp430_expand_builtin): New function.
5847 (TARGET_INIT_BUILTINS): Define.
5848 (TARGET_EXPAND_BUILTINS): Define.
5849 (TARGET_BUILTIN_DECL): Define.
5850 (msp430_expand_prologue): Add support for naked, interrupt,
5851 critical and reentrant functions.
5852 (msp430_expand_epilogue): Likewise.
5853 (msp430_print_operand): Handle 'O' character.
5854 * config/msp430/msp430.h (TARGET_CPU_CPP_BUILTINS): Define
5856 * config/msp430/msp430.md (unspec): Add UNS_DINT, UNS_EINT,
5857 UNS_PUSH_INTR, UNS_POP_INTR, UNS_BIC_SR, UNS_BIS_SR.
5858 (pushm): Use a 'n' rather than an 'i' constraint.
5859 (msp_return): Add generation of the interrupt return instruction.
5860 (disable_interrupts): New pattern.
5861 (enable_interrupts): New pattern.
5862 (push_intr_state): New pattern.
5863 (pop_intr_state): New pattern.
5864 (bic_SR): New pattern.
5865 (bis_SR): New pattern.
5866 * doc/extend.texi: Document MSP430 function attributes and builtin
5869 2013-09-17 Richard Biener <rguenther@suse.de>
5871 PR tree-optimization/58432
5872 * tree-loop-distribution.c (tree_loop_distribution): Also
5873 scan PHIs for outside loop uses and seed a partition from them.
5875 2013-09-17 Bin Cheng <bin.cheng@arm.com>
5877 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref): New.
5878 (restructure_reference): Call backtrace_base_for_ref.
5880 2013-09-17 Alan Modra <amodra@gmail.com>
5883 * config/rs6000/driver-rs6000.c (elf_platform): Revert 2013-06-11
5886 2013-09-16 DJ Delorie <dj@redhat.com>
5888 * config/rl78/rl78.c (rl78_asm_file_start): Specify alternate
5889 vregs location for RL78/G10.
5890 (rl78_expand_prologue): Avoid SEL on G10.
5891 (rl78_expand_epilogue): Likewise.
5892 (rl78_peep_movhi_p): Can't move a constant to memory in HImode.
5893 * config/rl78/rl78.h (TARGET_CPU_CPP_BUILTINS): Define
5894 __RL78_G10__ when appropriate.
5895 (ASM_SPEC): Pass -mg10 along to the assembler.
5896 * config/rl78/rl78.md (sel_rb): Disable for G10.
5897 * config/rl78/rl78.opt: Add -mg10 option.
5898 * config/rl78/t-rl78: Add -mg10 multilib.
5900 2013-09-16 Xinliang David Li <davidxl@google.com>
5902 * tree-if-conv.c (main_tree_if_conversion): Check new flag.
5903 * omp-low.c (omp_max_vf): Ditto.
5904 (expand_omp_simd): Ditto.
5905 * tree-vectorizer.c (vectorize_loops): Ditto.
5906 (gate_vect_slp): Ditto.
5907 (gate_increase_alignment): Ditto.
5908 * tree-ssa-pre.c (inhibit_phi_insertion): Ditto.
5909 * tree-ssa-loop.c (gate_tree_vectorize): Ditto.
5910 (gate_tree_vectorize): Name change.
5911 (tree_vectorize): Ditto.
5912 (pass_vectorize::gate): Call new function.
5913 (pass_vectorize::execute): Ditto.
5914 * opts.c: O3 default setting change.
5915 (finish_options): Check new flag.
5916 * doc/invoke.texi: Document new flags.
5917 * common.opt: New flags.
5919 2013-09-16 Andreas Schwab <schwab@linux-m68k.org>
5921 * doc/tm.texi.in (Cond Exec Macros): Remove node.
5922 (Condition Code): Don't reference it.
5923 * doc/tm.texi: Regenerate.
5925 2013-09-16 Vladimir Makarov <vmakarov@redhat.com>
5928 * lra-constraints.c (undo_optional_reloads): Consider all optional
5929 reload even if it did not get a hard reg.
5931 2013-09-16 Teresa Johnson <tejohnson@google.com>
5933 * dumpfile.c (dump_loc): Remove newline emission.
5934 * tree-vect-data-refs.c (vect_lanes_optab_supported_p): Add newline
5935 emission to dump_printf_loc calls where missing.
5936 (vect_mark_for_runtime_alias_test): Ditto.
5937 (vect_analyze_data_ref_dependence): Ditto.
5938 (vect_analyze_data_ref_dependences): Ditto.
5939 (vect_slp_analyze_data_ref_dependence): Ditto.
5940 (vect_slp_analyze_data_ref_dependences): Ditto.
5941 (vect_compute_data_ref_alignment): Ditto.
5942 (vect_update_misalignment_for_peel): Ditto.
5943 (vect_verify_datarefs_alignment): Ditto.
5944 (vector_alignment_reachable_p): Ditto.
5945 (vect_get_data_access_cost): Ditto.
5946 (vect_enhance_data_refs_alignment): Ditto.
5947 (vect_find_same_alignment_drs): Ditto.
5948 (vect_analyze_data_refs_alignment): Ditto.
5949 (vect_analyze_group_access): Ditto.
5950 (vect_analyze_data_ref_access): Ditto.
5951 (vect_analyze_data_ref_accesses): Ditto.
5952 (vect_prune_runtime_alias_test_list): Ditto.
5953 (vect_analyze_data_refs): Ditto.
5954 (vect_create_addr_base_for_vector_ref): Ditto.
5955 (vect_create_data_ref_ptr): Ditto.
5956 (vect_grouped_store_supported): Ditto.
5957 (vect_grouped_load_supported): Ditto.
5958 * value-prof.c (check_counter): Ditto.
5959 (check_ic_target): Ditto.
5960 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Ditto.
5961 (vect_recog_widen_mult_pattern): Ditto.
5962 (vect_recog_widen_sum_pattern): Ditto.
5963 (vect_recog_over_widening_pattern): Ditto.
5964 (vect_recog_widen_shift_pattern): Ditto.
5965 (vect_recog_rotate_pattern): Ditto.
5966 (vect_recog_vector_vector_shift_pattern): Ditto.
5967 (vect_recog_divmod_pattern): Ditto.
5968 (vect_recog_mixed_size_cond_pattern): Ditto.
5969 (vect_recog_bool_pattern): Ditto.
5970 (vect_pattern_recog_1): Ditto.
5971 (vect_pattern_recog): Ditto.
5972 * tree-vect-loop.c (vect_determine_vectorization_factor): Ditto.
5973 (vect_is_simple_iv_evolution): Ditto.
5974 (vect_analyze_scalar_cycles_1): Ditto.
5975 (vect_get_loop_niters): Ditto.
5976 (vect_analyze_loop_1): Ditto.
5977 (vect_analyze_loop_form): Ditto.
5978 (vect_analyze_loop_operations): Ditto.
5979 (vect_analyze_loop_2): Ditto.
5980 (vect_analyze_loop): Ditto.
5981 (report_vect_op): Ditto.
5982 (vect_is_slp_reduction): Ditto.
5983 (vect_is_simple_reduction_1): Ditto.
5984 (vect_get_known_peeling_cost): Ditto.
5985 (vect_estimate_min_profitable_iters): Ditto.
5986 (vect_model_reduction_cost): Ditto.
5987 (vect_model_induction_cost): Ditto.
5988 (get_initial_def_for_induction): Ditto.
5989 (vect_create_epilog_for_reduction): Ditto.
5990 (vectorizable_reduction): Ditto.
5991 (vectorizable_induction): Ditto.
5992 (vectorizable_live_operation): Ditto.
5993 (vect_loop_kill_debug_uses): Ditto.
5994 (vect_transform_loop): Ditto.
5995 * tree-vect-stmts.c (vect_mark_relevant): Ditto.
5996 (vect_stmt_relevant_p): Ditto.
5997 (process_use): Ditto.
5998 (vect_mark_stmts_to_be_vectorized): Ditto.
5999 (vect_model_simple_cost): Ditto.
6000 (vect_model_promotion_demotion_cost): Ditto.
6001 (vect_model_store_cost): Ditto.
6002 (vect_get_store_cost): Ditto.
6003 (vect_model_load_cost): Ditto.
6004 (vect_get_load_cost): Ditto.
6005 (vect_init_vector_1): Ditto.
6006 (vect_get_vec_def_for_operand): Ditto.
6007 (vect_finish_stmt_generation): Ditto.
6008 (vectorizable_call): Ditto.
6009 (vectorizable_conversion): Ditto.
6010 (vectorizable_assignment): Ditto.
6011 (vectorizable_shift): Ditto.
6012 (vectorizable_operation): Ditto.
6013 (vectorizable_store): Ditto.
6014 (vectorizable_load): Ditto.
6015 (vectorizable_condition): Ditto.
6016 (vect_analyze_stmt): Ditto.
6017 (vect_transform_stmt): Ditto.
6018 (vect_is_simple_use): Ditto.
6019 * tree-vect-loop-manip.c (slpeel_make_loop_iterate_ntimes): Ditto.
6020 (vect_can_advance_ivs_p): Ditto.
6021 (vect_update_ivs_after_vectorizer): Ditto.
6022 (vect_do_peeling_for_loop_bound): Ditto.
6023 (vect_gen_niters_for_prolog_loop): Ditto.
6024 (vect_update_inits_of_drs): Ditto.
6025 (vect_create_cond_for_alias_checks): Ditto.
6026 * tree-vect-slp.c (vect_get_and_check_slp_defs): Ditto.
6027 (vect_build_slp_tree_1): Ditto.
6028 (vect_supported_load_permutation_p): Ditto.
6029 (vect_analyze_slp_instance): Ditto.
6030 (vect_analyze_slp): Ditto.
6031 (vect_make_slp_decision): Ditto.
6032 (vect_detect_hybrid_slp): Ditto.
6033 (vect_bb_vectorization_profitable_p): Ditto.
6034 (vect_slp_analyze_bb_1): Ditto.
6035 (vect_update_slp_costs_according_to_vf): Ditto.
6036 (vect_get_mask_element): Ditto.
6037 (vect_transform_slp_perm_load): Ditto.
6038 (vect_schedule_slp_instance): Ditto.
6039 (vect_schedule_slp): Ditto.
6040 (vect_slp_transform_bb): Ditto.
6041 * profile.c (read_profile_edge_counts): Ditto.
6042 (compute_branch_probabilities): Ditto.
6043 * coverage.c (get_coverage_counts): Ditto.
6045 2013-09-16 Diego Novillo <dnovillo@google.com>
6047 * tree-core.h: Add missing comment lines from refactoring of tree.h.
6049 2013-09-16 Jan Hubicka <jh@suse.cz>
6051 * gimple-fold.c (can_refer_decl_in_current_unit_p): Do not accept
6052 abstract functions; for static functions check the presence of body.
6054 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
6056 * config/aarch64/aarch64-simd-builtins.def (fma): New.
6057 * config/aarch64/aarch64-simd.md
6058 (aarch64_mla_elt<mode>): New.
6059 (aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
6060 (aarch64_mls_elt<mode>): Likewise.
6061 (aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
6062 (aarch64_fma4_elt<mode>): Likewise.
6063 (aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
6064 (aarch64_fma4_elt_to_128v2df): Likewise.
6065 (aarch64_fma4_elt_to_64df): Likewise.
6066 (fnma<mode>4): Likewise.
6067 (aarch64_fnma4_elt<mode>): Likewise.
6068 (aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
6069 (aarch64_fnma4_elt_to_128v2df): Likewise.
6070 (aarch64_fnma4_elt_to_64df): Likewise.
6071 * config/aarch64/iterators.md (VDQSF): New.
6072 * config/aarch64/arm_neon.h
6073 (vfm<as><sdq>_lane<q>_f<32, 64>): Convert to C implementation.
6074 (vml<sa><q>_lane<q>_<fsu><16, 32, 64>): Likewise.
6076 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
6078 * config/aarch64/aarch64-simd.md (aarch64_mul3_elt<mode>): New.
6079 (aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
6080 (aarch64_mul3_elt_to_128df): Likewise.
6081 (aarch64_mul3_elt_to_64v2df): Likewise.
6082 * config/aarch64/iterators.md (VEL): Also handle DFmode.
6084 (VMUL_CHANGE_NLANES) Likewise.
6087 * config/aarch64/arm_neon.h
6088 (vmul<q>_lane<q>_<suf><16,32,64>): Convert to C implementation.
6090 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
6092 * config/aarch64/arm_neon.h
6093 (vcvtx_high_f32_f64): Fix parameters.
6095 2013-09-16 Jan-Benedict Glaw <jbglaw@lug-owl.de>
6096 Uros Bizjak <ubizjak@gmail.com>
6098 * config/alpha.c: Include tree-ssanames.h.
6100 2013-09-16 Richard Biener <rguenther@suse.de>
6102 * tree-loop-distribution.c (enum rdg_dep_type): Add control_dd.
6103 (dot_rdg_1): Handle control_dd.
6104 (create_edge_for_control_dependence): New function.
6105 (create_rdg_edges): Add control dependences if asked for.
6106 (build_rdg): Likewise.
6107 (generate_loops_for_partition): If there are not necessary
6108 control stmts remove all their dependencies.
6109 (collect_condition_stmts, rdg_flag_loop_exits): Remove.
6110 (distribute_loop): Pass on control dependences.
6111 (tree_loop_distribution): Compute control dependences and remove
6112 restriction on number of loop nodes.
6114 2013-09-16 Jakub Jelinek <jakub@redhat.com>
6116 * ipa-prop.c (ipa_compute_jump_functions_for_edge): Return early
6119 2013-09-16 Richard Sandiford <rdsandiford@googlemail.com>
6121 * cse.c (try_const_anchors): Punt on CC modes.
6123 2013-09-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
6125 * config/vax/constraints.md (T): Add missing CONSTANT_P check.
6127 2013-09-14 John David Anglin <danglin@gcc.gnu.org>
6130 * config/pa/pa.c (pa_expand_prologue): Change mode in gen_rtx_POST_INC
6133 2013-09-14 Iain Sandoe <iain@codesourcery.com>
6136 * config/darwin.c (darwin_objc2_section): Note if ObjC Metadata is
6138 (darwin_objc1_section): Likewise.
6139 (darwin_file_end): Emit Image Info section when required.
6141 2013-09-14 Jan Hubicka <jh@suse.cz>
6143 * tree-into-ssa.c (gate_into_ssa): New.
6144 (pass_data_build_ssa): Use it.
6145 * cgraph.h (expand_thunk): Update prototype.
6146 * cgraphunit.c (analyze_function): Expand thunks early.
6147 (expand_thunk): Fix DECL_CONTEXT of reust_decl;
6148 build proper cgraph; set in_ssa_p; clear bogus TREE_ASM_WRITTEN;
6149 set lowered flag; do not add new function.
6150 (assemble_thunks_and_aliases): Update.
6151 * tree-ssa.c (gate_init_datastructures): New gate.
6152 (pass_data_init_datastructures): Use it.
6154 2013-09-14 Iain Sandoe <iain@codesourcery.com>
6157 * config/i386/i386.c (ix86_function_arg_regno_p): Make Darwin use the
6158 xmm register set described in the psABI.
6160 2013-09-13 Evgeny Gavrin <e.gavrin@samsung.com>
6162 * dwarf2out.c (should_emit_struct_debug): Add check
6163 for type_decl variable is not NULL.
6165 2013-09-13 Jacek Caban <jacek@codeweavers.com>
6167 * config.gcc: Use new winnt-c.c target hooks
6168 * config/t-winnt: New file
6169 * config/winnt-c.c: New file
6170 * doc/tm.texi.in: Document new hook
6171 * doc/tm.texi: Regenerated
6173 2013-09-13 Jan Hubicka <jh@suse.cz>
6176 * ipa-inline.c (check_callers): New function.
6177 (check_caller_edge): Remove.
6178 (want_inline_function_to_all_callers_p): Also permit alises that are
6180 (inline_to_all_callers): Terminate the walk when devirtualization
6181 introduce new calls.
6183 2013-09-13 Jan Hubicka <jh@suse.cz>
6185 * ipa-inline-analysis.c (struct growth_data): Add node.
6186 (do_estimate_growth_1): Fix detection of recursion.
6188 2013-09-13 Jakub Jelinek <jakub@redhat.com>
6190 PR tree-optimization/58392
6191 * tree-cfg.c (move_sese_region_to_fn): Rename loop variable
6192 to avoid shadowing of outer loop variable. If
6193 saved_cfun->has_simduid_loops or saved_cfun->has_force_vect_loops,
6194 replace_by_duplicate_decl simduid of loops that have it set and
6195 set dest_cfun->has_simduid_loops and/or
6196 dest_cfun->has_force_vect_loops.
6197 * omp-low.c (build_outer_var_ref): Call maybe_lookup_decl_in_outer_ctx
6198 instead of maybe_lookup_decl.
6199 * tree-inline.c (copy_loops): Change blocks_to_copy argument to id.
6200 Use id->blocks_to_copy instead of blocks_to_copy. Adjust recursive
6201 call. Copy over force_vect and copy and remap simduid. Set
6202 cfun->has_simduid_loops and/or cfun->has_force_vect_loops.
6203 (copy_cfg_body): Remove blocks_to_copy argument. Use
6204 id->blocks_to_copy instead of blocks_to_copy. Adjust copy_loops
6205 caller. Don't set cfun->has_simduid_loops and/or
6206 cfun->has_force_vect_loops here.
6207 (copy_body): Remove blocks_to_copy argument. Adjust copy_cfg_body
6209 (expand_call_inline, tree_function_versioning): Adjust copy_body
6212 2013-09-13 Martin Jambor <mjambor@suse.cz>
6215 * ipa-prop.c (try_make_edge_direct_simple_call): Be less strict in
6216 the assert if the edge was a speculative one.
6218 2013-09-13 Richard Biener <rguenther@suse.de>
6220 * tree-data-ref.h (known_dependences_p): Move here ...
6221 * tree-loop-distribution.c (known_dependences_p): ... from here.
6222 (dump_rdg_component, debug_rdg_component): Remove.
6224 (generate_loops_for_partition): Use gimple_uid instead of
6225 relying on matching stmt visit order.
6226 (rdg_build_partitions): Take starting stmt vector.
6227 (ldist_gen): Merge into ...
6228 (distribute_loop): ... this function. Do not compute starting
6230 * tree-cfg.c (gimple_duplicate_bb): Copy UID for PHIs.
6232 2013-09-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6234 * config/arm/arm.md (arm_cmpsi_insn): Split rI alternative.
6235 Set type attribute correctly. Set predicable_short_it attribute.
6236 (cmpsi_shiftsi): Remove %? from output template.
6238 2013-09-13 Richard Biener <rguenther@suse.de>
6240 * tree-loop-distribution.c (struct rdg_component,
6241 rdg_defs_used_in_other_loops_p, free_rdg_components,
6242 rdg_build_components): Remove.
6243 (stmts_from_loop): Do not record virtual PHIs.
6244 (generate_loops_for_partition): Skip virtual PHIs.
6245 (build_rdg_partition_for_component): Rename to ...
6246 (build_rdg_partition_for_vertex): ... this and adjust.
6247 (rdg_build_partitions): Take a vector of starting vertices
6248 instead of components. Remove unnecessary leftover handling.
6249 (ldist_gen): Do not build components or record other stores.
6250 (distribute_loop): Do not distribute loops containing stmts
6253 2013-09-13 Christian Bruel <christian.bruel@st.com>
6256 * config/sh/sh.md (mov<mode>_reg_reg): Allow memory reloads.
6258 2013-09-13 Kai Tietz <ktietz@redhat.com>
6260 * config.gcc: Separate cases for mingw and cygwin targets,
6261 and add 64-bit cygwin target case.
6263 * config/i386/winnt-cxx.c (i386_pe_type_dllexport_p): Don't
6264 dll-export inline-functions.
6265 * config/i386/winnt.c (i386_pe_determine_dllexport_p): Likewise.
6268 2013-09-13 Jeff Law <law@redhat.com>
6272 2013-09-06 Jeff Law <law@redhat.com>
6274 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
6275 edge implied equivalences into successor phis.
6277 2013-09-12 DJ Delorie <dj@redhat.com>
6279 * config/rl78/rl78-virt.md: Change from | to \; for asm line
6282 2013-09-12 Brooks Moses <bmoses@google.com>
6285 * Makefile.in: Do not install driver binaries in $(target)/bin.
6287 2013-09-12 DJ Delorie <dj@redhat.com>
6289 * config/rl78/rl78.opt (mrelax): New.
6290 * config/rl78/rl78.h (ASM_SPEC): New, pass on -mrelax to gas.
6291 * config/rl78/rl78.h (LINK_SPEC): New, pass on -mrelax to ld.
6293 * config/rl78/rl78.c (rl78_expand_prologue): Use AX to copy
6295 (rl78_expand_epilogue): Likewise.
6297 2013-09-12 Vladimir Makarov <vmakarov@redhat.com>
6300 * lra-eliminations.c (remove_reg_equal_offset_note): New.
6301 (eliminate_regs_in_insn): Rewrite frame pointer to hard frame
6302 pointer elimination with using remove_reg_equal_offset_note.
6304 2013-09-12 DJ Delorie <dj@redhat.com>
6306 * config/msp430/: New port.
6307 * config.gcc (msp430): Added.
6308 * doc/invoke.texi: Document MSP430 options.
6309 * doc/install.texi: Document msp430-elf
6310 * doc/md.texi: Document msp430-elf
6311 * doc/contrib.texi: Document msp430-elf
6313 * cfgexpand.c (expand_debug_expr): Avoid sign-extending SImode to
6316 2013-09-12 Martin Jambor <mjambor@suse.cz>
6319 * ipa-prop.c (remove_described_reference): Give up if the edge in the
6320 reference descriptor is NULL.
6321 (ipa_edge_removal_hook): If owning a reference descriptor, set its
6324 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
6326 * tree-flow.h (FREE_SSANAMES): Move to tree-ssanames.c
6327 (SSANAMES, MODIFIED_NORETURN_CALLS, DEFAULT_DEFS, ptr_info_def,
6328 num_ssa_names, ssa_name): Move to tree-ssanames.h + prototypes.
6329 * tree-flow-inline.h (make_ssa_name, copy_ssa_name, duplicate_ssa_name,
6330 make_temp_ssa_name): move to tree-ssanames.h
6331 * tree-ssa-alias.h: Move prototype.
6332 * tree-ssa.h: Include tree-ssanames.h.
6333 * tree-ssanames.c (FREE_SSANAMES): Move to here.
6334 * tree-ssanames.h: New. Move items from tree-flow*.h
6335 * Makefile.in (tree-ssanames.h): Add to tree-ssanames.o and GTFILES.
6337 2013-09-12 Richard Biener <rguenther@suse.de>
6339 PR tree-optimization/58404
6340 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Also
6341 propagate non-invariant addresses into dereferences wrapped
6342 in component references.
6344 2013-09-12 Richard Biener <rguenther@suse.de>
6346 PR tree-optimization/58402
6347 * passes.def: Move pass_late_warn_uninitialized later.
6349 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
6351 * tree-ssa.h: New. Move content from tree-flow.h and
6353 * tree-flow.h (_edge_var_map, edge_var_map_vector): Move to tree-ssa.h.
6354 Move prototypes belonging to tree-ssa.c.
6355 * tree-flow-inline.h (redirect_edge_var_map_def,
6356 redirect_edge_var_map_result, redirect_edge_var_map_location): Move to
6358 * gimple.h: Adjust prototypes.
6359 * tree-ssa.c (useless_type_conversion_p, types_compatible_p): Move
6361 * gimple.c (useless_type_conversion_p, types_compatible_p): Here.
6362 * tree.h: Move prototype to tree-ssa.h.
6363 * gengtype.c (open_base_files): Replace tree-flow.h with tree-ssa.h.
6364 * Makefile.in: (TREE_SSA_H, TREE_FLOW_H): Adjust dependencies.
6365 * alias.c, asan.c, builtins.c, calls.c, cfgexpand.c, cfghooks.c,
6366 cfgloop.c, cfgloopmanip.c, cgraph.c, cgraphbuild.c, cgraphclones.c,
6367 cgraphunit.c, dse.c, except.c, expr.c, final.c, fold-const.c,
6368 ggc-page.c, gimple-fold.c, gimple-iterator.c, gimple-low.c,
6369 gimple-pretty-print.c, gimple-ssa-strength-reduction.c,
6370 gimple-streamer-in.c, gimple-streamer-out.c, gimple.c, gimplify.c,
6371 graphite-blocking.c, graphite-clast-to-gimple.c,
6372 graphite-dependences.c, graphite-interchange.c,
6373 graphite-optimize-isl.c, graphite-poly.c, graphite-scop-detection.c,
6374 graphite-sese-to-poly.c, graphite.c, ipa-cp.c, ipa-inline-analysis.c,
6375 ipa-inline-transform.c, ipa-inline.c, ipa-prop.c, ipa-pure-const.c,
6376 ipa-reference.c, ipa-split.c, ipa-utils.c,
6377 loop-init.c, lto-cgraph.c, lto-section-in.c, lto-section-out.c,
6378 lto-streamer-in.c, lto-streamer-out.c, lto-streamer.c, omp-low.c,
6379 passes.c, predict.c, print-tree.c, profile.c, sese.c, targhooks.c,
6380 tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
6381 tree-chrec.c, tree-complex.c, tree-data-ref.c, tree-dfa.c, tree-eh.c,
6382 tree-emutls.c, tree-if-conv.c, tree-inline.c, tree-into-ssa.c,
6383 tree-loop-distribution.c, tree-mudflap.c, tree-nested.c, tree-nrv.c,
6384 tree-object-size.c, tree-optimize.c, tree-outof-ssa.c, tree-parloops.c,
6385 tree-phinodes.c, tree-predcom.c, tree-pretty-print.c, tree-profile.c,
6386 tree-scalar-evolution.c, tree-sra.c, tree-ssa*.c, tree-stdarg.c,
6387 tree-streamer-in.c, tree-switch-conversion.c, tree-tailcall.c,
6388 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
6389 tree-vect-loop.c, tree-vect-patterns.c, tree-vect-slp.c,
6390 tree-vect-stmts.c, tree-vectorizer.c, tree-vrp.c, tsan.c,
6391 value-prof.c, var-tracking.c,
6392 varpool.c, vtable-verify.c: Replace tree-flow.h with tree-ssa.h
6394 2013-09-12 Richard Biener <rguenther@suse.de>
6396 PR tree-optimization/58396
6397 * tree-loop-distribution.c (create_rdg_edges): Free unused DDRs.
6398 (build_rdg): Take a loop-nest parameter, fix memleaks.
6399 (distribute_loop): Compute loop-nest here and pass it to build_rdg.
6401 2013-09-12 Yuri Rumyantsev <ysrumyan@gmail.com>
6403 * config/i386/x86-tune.def: Turn on X86_TUNE_AVOID_MEM_OPND_FOR_CMOVE
6406 2013-09-12 Cameron McInally <cameron.mcinally@nyu.edu>
6408 * doc/extend.texi: Fix errors in x86 FMA builtin naming.
6409 The FMA instruction names should have a 'v' prefix.
6411 2013-09-12 Richard Biener <rguenther@suse.de>
6413 * tree-loop-distribution.c (dot_rdg_1): Make graph prettier.
6414 (dot_rdg): Use popen instead of system in optional code.
6415 (remaining_stmts, upstream_mem_writes): Remove global bitmaps.
6416 (already_processed_vertex_p): Adjust.
6417 (has_anti_or_output_dependence, predecessor_has_mem_write,
6418 mark_nodes_having_upstream_mem_writes, has_upstream_mem_writes,
6419 rdg_flag_uses): Remove.
6420 (rdg_flag_vertex): Simplify.
6421 (rdg_flag_vertex_and_dependent): Rely on a correct RDG and
6423 (build_rdg_partition_for_component): Process the first vertex
6424 of a component only.
6425 (ldist_gen): Do not compute remaining_stmts or upstream_mem_writes.
6427 2013-09-12 Alan Modra <amodra@gmail.com>
6429 * config/rs6000/rs6000.c (toc_relative_expr_p): Use add_cint_operand.
6431 2013-09-11 DJ Delorie <dj@redhat.com>
6432 Nick Clifton <nickc@redhat.com>
6434 * config/rl78/predicates.md (rl78_cmp_operator_signed): New.
6435 (rl78_stack_based_mem): New.
6436 * config/rl78/constraints.md (Iv08): New.
6445 * config/rl78/rl78-expand.md (movqi): Reject more SUBREG operands.
6447 (movsi): Change from expand to insn-and-split.
6448 (ashrsi3): Clobber AX.
6452 * config/rl78/rl78.md (CC_REG): Fix.
6453 (addsi3): Allow memory and immediate operands.
6454 (addsi3_internal): Split into...
6455 (addsi3_internal_virt): ...new, and ...
6456 (addsi3_internal_real): ...new.
6458 (subsi3_internal_virt): New.
6459 (subsi3_internal_real): New.
6460 (mulsi3): Add memory operand.
6461 (mulsi3_rl78): Likewise.
6462 (mulsi3_g13): Likewise.
6463 * config/rl78/rl78-real.md (cbranchqi4_real_signed): New.
6464 (cbranchqi4_real): Add more constraint options.
6465 (cbranchhi4_real): Expand pattern.
6466 (cbranchhi4_real_signed): New.
6467 (cbranchhi4_real_inverted): New.
6468 (cbranchsi4_real_lt): New.
6469 (cbranchsi4_real_ge): New.
6470 (cbranchsi4_real_signed): New.
6471 (cbranchsi4_real): New.
6473 * config/rl78/rl78-virt.md (ashrsi3_virt): Add custom cases for
6475 (lshrsi3_virt): Likewise.
6476 (ashlsi3_virt): Likewise.
6477 (cbranchqi4_virt_signed): New.
6478 (cbranchhi4_virt_signed): New.
6479 (cbranchsi4_virt): New.
6480 * config/rl78/rl78.c: Whitespace fixes throughout.
6481 (move_elim_pass): New.
6482 (pass_data_rl78_move_elim): New.
6483 (pass_rl78_move_elim): New.
6484 (make_pass_rl78_move_elim): New.
6485 (rl78_devirt_info): Run devirt earlier.
6486 (rl78_move_elim_info): New.
6487 (rl78_asm_file_start): Register it.
6488 (rl78_split_movsi): New.
6489 (rl78_as_legitimate_address): Allow virtual base registers when
6491 (rl78_addr_space_convert): Remove spurious debug stuff.
6492 (rl78_print_operand_1): Add z,s,S,r,E modifiers.
6493 (rl78_print_operand): More cases for not printing '#'.
6494 (rl78_expand_compare): Remove most of the logic.
6495 (content_memory): New.
6496 (clear_content_memory): New.
6497 (get_content_index): New.
6498 (get_content_name): New.
6499 (display_content_memory): New.
6500 (update_content): New.
6501 (record_content): New.
6502 (already_contains): New.
6503 (insn_ok_now): Re-recog insns with virtual registers.
6504 (add_postponed_content_update): New.
6505 (process_postponed_content_update): New.
6506 (gen_and_emit_move): New.
6507 (transcode_memory_rtx): Record new location content.
6508 Use gen_and_emit_move.
6509 (force_into_acc): New.
6510 (move_to_acc): Use gen_and_emit_move.
6511 (move_from_acc): Likewise.
6512 (move_acc_to_reg): Likewise.
6513 (move_to_x): Likewise.
6514 (move_to_hl): Likewise.
6515 (move_to_de): Likewise.
6516 (rl78_alloc_physical_registers_op1): Record location content.
6517 (has_constraint): New.
6518 (rl78_alloc_physical_registers_op2): Record location content.
6520 (rl78_alloc_physical_registers_ro1): Likewise.
6521 (rl78_alloc_physical_registers_cmp): Likewise.
6522 (rl78_alloc_physical_registers_umul): Likewise.
6523 (rl78_alloc_address_registers_macax): New.
6524 (rl78_alloc_physical_registers): Initialize and set location
6525 content memory as needed.
6526 (rl78_reorg): Make sure split2 is called.
6527 (rl78_rtx_costs): New.
6529 2013-09-11 Richard Sandiford <rdsandiford@googlemail.com>
6531 * simplify-rtx.c (simplify_unary_operation_1): Use simplify_gen_binary
6532 for (not (neg ...)) and (neg (not ...)) cases.
6534 2013-09-11 Richard Biener <rguenther@suse.de>
6537 * passes.def: Split critical edges before late uninit warning passes.
6538 * tree-cfg.c (pass_split_crit_edges): Implement clone method.
6540 2013-09-11 Jakub Jelinek <jakub@redhat.com>
6542 PR tree-optimization/58385
6543 * fold-const.c (build_range_check): If both low and high are NULL,
6544 use omit_one_operand_loc to preserve exp side-effects.
6546 2013-09-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6548 * config/arm/arm.md (arm_shiftsi3): New alternative l/l/M.
6550 2013-09-11 Richard Biener <rguenther@suse.de>
6552 * tree-data-ref.c (dump_rdg_vertex, debug_rdg_vertex,
6553 dump_rdg_component, debug_rdg_component, dump_rdg, debug_rdg,
6554 dot_rdg_1, dot_rdg, rdg_vertex_for_stmt, create_rdg_edge_for_ddr,
6555 create_rdg_edges_for_scalar, create_rdg_edges, create_rdg_vertices,
6556 stmts_from_loop, known_dependences_p, build_empty_rdg,
6557 build_rdg, free_rdg, rdg_defs_used_in_other_loops_p): Move ...
6558 * tree-loop-distribution.c: ... here.
6559 * tree-data-ref.h (struct rdg_vertex, RDGV_STMT, RDGV_DATAREFS,
6560 RDGV_HAS_MEM_WRITE, RDGV_HAS_MEM_READS, RDG_STMT, RDG_DATAREFS,
6561 RDG_MEM_WRITE_STMT, RDG_MEM_READS_STMT, enum rdg_dep_type,
6562 struct rdg_edge, RDGE_TYPE, RDGE_LEVEL, RDGE_RELATION): Move ...
6563 * tree-loop-distribution.c: ... here.
6564 * tree-loop-distribution.c: Include gimple-pretty-print.h.
6565 (struct partition_s): Add loops member.
6566 (partition_alloc, partition_free, rdg_flag_uses, rdg_flag_vertex,
6567 rdg_flag_vertex_and_dependent, rdg_flag_loop_exits,
6568 build_rdg_partition_for_component, rdg_build_partitions): Adjust.
6570 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
6571 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
6572 Sergey Lega <sergey.s.lega@intel.com>
6573 Anna Tikhonova <anna.tikhonova@intel.com>
6574 Ilya Tocar <ilya.tocar@intel.com>
6575 Andrey Turetskiy <andrey.turetskiy@intel.com>
6576 Ilya Verbin <ilya.verbin@intel.com>
6577 Kirill Yukhin <kirill.yukhin@intel.com>
6578 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
6580 * config/i386/constraints.md (k): New.
6582 * config/i386/i386.c (const regclass_map): Add new mask registers.
6583 (dbx_register_map): Ditto.
6584 (dbx64_register_map): Ditto.
6585 (svr4_dbx_register_map): Ditto.
6586 (ix86_conditional_register_usage): Squash mask registers if AVX512F is
6588 (ix86_preferred_reload_class): Disable constants for mask registers.
6589 (ix86_secondary_reload): Do spill of mask register using 32-bit insn.
6590 (ix86_hard_regno_mode_ok): Support new mask registers.
6591 (x86_order_regs_for_local_alloc): Ditto.
6592 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Update.
6593 (FIXED_REGISTERS): Add new mask registers.
6594 (CALL_USED_REGISTERS): Ditto.
6595 (REG_ALLOC_ORDER): Ditto.
6596 (VALID_MASK_REG_MODE): New.
6597 (FIRST_MASK_REG): Ditto.
6598 (LAST_MASK_REG): Ditto.
6599 (reg_class): Add MASK_EVEX_REGS, MASK_REGS.
6600 (MAYBE_MASK_CLASS_P): New.
6601 (REG_CLASS_NAMES): Add MASK_EVEX_REGS, MASK_REGS.
6602 (REG_CLASS_CONTENTS): Ditto.
6603 (MASK_REGNO_P): New.
6604 (ANY_MASK_REG_P): Ditto.
6605 (HI_REGISTER_NAMES): Add new mask registers.
6606 * config/i386/i386.md (MASK0_REG, MASK1_REG, MASK2_REG, MASK3_REG,
6607 MASK4_REG, MASK5_REG, MASK6_REG, MASK7_REG): Constants for new
6609 (attribute "type"): Add mskmov, msklog.
6610 (attribute "length_immediate"): Support them.
6611 (attribute "memory"): Ditto.
6612 (attribute "prefix_0f"): Ditto.
6613 (*movhi_internal): Support new mask registers.
6614 (*movqi_internal): Ditto.
6615 (define_split): Split out clobber pattern is a logic
6616 insn on mask registers.
6617 (*k<logic><mode>): New.
6618 (*andhi_1): Extend to support mask regs.
6619 (*andqi_1): Extend to support mask regs.
6621 (define_split): Split and-not to and and not if operands
6623 (*<code><mode>_1): Separate HI mode to new pattern...
6624 (*<code>hi_1): This.
6625 (*<code>qi_1): Extend to support mask regs.
6627 (kortestzhi): Ditto.
6628 (kortestchi): Ditto.
6630 (*one_cmpl<mode>2_1): Remove HImode and handle it...
6631 (*one_cmplhi2_1): ...Here, now with mask registers support.
6632 (*one_cmplqi2_1): Support new mask registers.
6633 (HI/QImode arithmetics splitter): Don't split if mask registers
6635 (HI/QImode not splitter): Ditto.
6636 * config/i386/predicated.md (mask_reg_operand): New.
6637 (general_reg_operand): Ditto.
6639 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
6641 * doc/invoke.texi: Document fxsr, xsave and xsaveopt options.
6642 * doc/extend.texi: Document fxsr, xsave and xsaveopt builtins.
6644 2013-09-10 Jeff Law <law@redhat.com>
6646 PR tree-optimization/58380
6647 * tree-ssa-threadupdate.c (thread_block): Recognize another case
6648 of threading through a buried loop header.
6650 * tree-ssa-threadedge.c (thread_around_empty_blocks): Correct
6651 return value for single successor case.
6653 2013-09-10 Jan Hubicka <jh@suse.cz>
6655 * ipa-devirt.c (ipa_devirt): Enable with LTO.
6657 2013-09-10 Richard Earnshaw <rearnsha@arm.com>
6660 * arm/vfp.md (combine_vcvt_f32_<FCVTI32typename>): Fix pattern to
6661 support conditional execution.
6662 (combine_vcvt_f64_<FCVTI32typename>): Likewise.
6664 2013-09-10 Vladimir Makarov <vmakarov@redhat.com>
6666 * lra.c (lra): Clear lra_optional_reload_pseudos before every
6668 * lra-constraints.c (curr_insn_transform): Switch on optional reloads.
6669 Check destination too to check move insn.
6670 (undo_optional_reloads): Add check that the original peudo did not
6671 changed its allocation and the optional reload was inherited on last
6672 inheritance pass. Break loop after deciding to keep optional reload.
6673 (lra_undo_inheritance): Add check that inherited pseudo still in
6676 2013-09-10 James Greenhalgh <james.greenhalgh@arm.com>
6678 * config/aarch64/aarch64.md (generic_sched): New.
6679 * config/aarch64/aarch64-generic.md (load): Make conditional
6680 on generic_sched attribute.
6681 (nonload): Likewise.
6683 2013-09-10 Jan Hubicka <jh@suse.cz>
6685 * lto-cgraph.c: Include ipa-utils.h.
6686 (compute_ltrans_boundary): Also add possible targets into the boundary.
6688 2013-09-10 Jan Hubicka <jh@suse.cz>
6690 * gimple-fold.c (gimple_get_virt_method_for_binfo): Pass real
6691 VAR_DECL of vtable rather than full expression.
6693 2013-09-10 Jan Hubicka <jh@suse.cz>
6694 Paolo Carlini <paolo.carlini@oracle.com>
6696 * cgraphunit.c (analyze_functions): Save input_location, set it
6697 to UNKNOWN_LOCATION and restore it at the end.
6699 2013-09-10 Martin Jambor <mjambor@suse.cz>
6701 * ipa-cp.c (propagate_constants_topo): Do not ignore SCC
6702 represented by a thunk.
6704 2013-09-10 Jeff Law <law@redhat.com>
6706 PR tree-optimization/58343
6707 * tree-ssa-threadupdate.c (thread_block): Identify and disable
6708 jump threading requests through loop headers buried in the middle
6709 of a jump threading path.
6711 * tree-ssa-threadedge.c (thread_around_empty_blocks): Fix thinko
6712 in return value/type.
6714 2013-09-10 Jakub Jelinek <jakub@redhat.com>
6716 PR rtl-optimization/58365
6717 * cfgcleanup.c (merge_memattrs): Also clear MEM_READONLY_P
6718 resp. MEM_NOTRAP_P if they differ, or set MEM_VOLATILE_P if
6721 2013-09-10 Richard Biener <rguenther@suse.de>
6723 * tree-data-ref.h (build_rdg): Drop all parameters but loop.
6724 * tree-data-ref.c (create_rdg_vertices): Collect all data
6725 references, signal failure to the caller, use data-ref API.
6726 (build_rdg): Compute data references only once. Maintain lifetime
6727 of data references and data dependences from within RDG.
6728 (free_rdg): Free dependence relations.
6729 * tree-loop-distribution.c (rdg_flag_uses): Drop weird code
6730 inventing extra dependences.
6731 (distribute_loop): Update for RDG API changes.
6733 2013-09-10 Kai Tietz <ktietz@redhat.com>
6735 * doc/invoke.texi (fms-extensions): Document changed
6736 behavior for ms-abi targets.
6737 * config/i386/i386.c (ix86_option_override_internal):
6738 Set default value of option -fms-extension for ms-abi targets.
6740 2013-09-10 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
6742 * config/i386/i386.c (ix86_expand_movmem): Fix epilogue generation.
6744 2013-09-10 Alan Modra <amodra@gmail.com>
6747 * config/rs6000/rs6000.md (bswapdi2_64bit): Disable for volatile mems.
6749 2013-09-10 Alan Modra <amodra@gmail.com>
6751 * config/rs6000/predicates.md (add_cint_operand): New.
6752 (reg_or_add_cint_operand, small_toc_ref): Use add_cint_operand.
6753 * config/rs6000/rs6000.md (largetoc_high_plus): Restrict offset
6754 using add_cint_operand.
6755 (largetoc_high_plus_aix): Likewise.
6757 2013-09-09 Jakub Jelinek <jakub@redhat.com>
6759 PR tree-optimization/58364
6760 * tree-ssa-reassoc.c (init_range_entry): For BIT_NOT_EXPR on
6761 BOOLEAN_TYPE, only invert in_p and continue with arg0 if
6762 the current range can't be an unconditional true or false.
6764 2013-09-09 James Greenhalgh <james.greenhalgh@arm.com>
6766 * config/aarch64/arm_neon.h (vrsqrte_f64): Fix parameter type.
6768 2013-09-09 Uros Bizjak <ubizjak@gmail.com>
6770 * ipa-prop.c (ipa_modify_call_arguments): Initialize deref_align.
6772 2013-09-09 Paolo Carlini <paolo.carlini@oracle.com>
6775 * doc/invoke.texi (-Wdelete-incomplete): Document it.
6777 2013-09-09 Ian Bolton <ian.bolton@arm.com>
6779 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Return
6780 NO_REGS for immediate that can't be moved directly into FP_REGS.
6782 2013-09-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6784 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return CC_SWP for
6785 comparison with negated operand.
6786 * config/aarch64/aarch64.md (compare_neg<mode>): Match canonical
6789 2013-09-09 Richard Biener <rguenther@suse.de>
6792 * cfgloopmanip.c (fix_bb_placements): When fixing the placement
6793 of a subloop record all its block as affecting loop-closed SSA form.
6795 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
6797 * expmed.c (lshift_value): Take an unsigned HOST_WIDE_INT instead
6798 of an rtx/bitpos pair.
6799 (store_fixed_bit_field): Update accordingly.
6801 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
6803 * asan.c (asan_emit_stack_protection): Use gen_int_mode instead of
6805 * builtins.c (expand_errno_check): Likewise.
6806 * dwarf2cfi.c (init_return_column_size): Likewise.
6807 * except.c (sjlj_mark_call_sites): Likewise.
6808 * expr.c (move_by_pieces_1, store_by_pieces_2): Likewise.
6809 * lra-constraints.c (emit_inc): Likewise.
6810 * ree.c (combine_set_extension): Likewise.
6811 * regmove.c (fixup_match_2): Likewise.
6812 * reload1.c (inc_for_reload): Likewise.
6814 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
6816 * combine.c (simplify_set, expand_field_assignment, extract_left_shift)
6817 (force_to_mode, simplify_shift_const_1, simplify_comparison):
6818 Use gen_int_mode with the mode of the associated simplify_* call.
6819 * explow.c (probe_stack_range, anti_adjust_stack_and_probe): Likewise.
6820 * expmed.c (expand_shift_1): Likewise.
6821 * function.c (instantiate_virtual_regs_in_insn): Likewise.
6822 * loop-iv.c (iv_number_of_iterations): Likewise.
6823 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
6824 * simplify-rtx.c (simplify_binary_operation_1): Likewise.
6826 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
6828 * asan.c (asan_clear_shadow): Use gen_int_mode with the mode
6829 of the associated expand_* call.
6830 (asan_emit_stack_protection): Likewise.
6831 * builtins.c (round_trampoline_addr): Likewise.
6832 * explow.c (allocate_dynamic_stack_space, probe_stack_range): Likewise.
6833 * expmed.c (expand_smod_pow2, expand_sdiv_pow2, expand_divmod)
6834 (emit_store_flag): Likewise.
6835 * expr.c (emit_move_resolve_push, push_block, emit_single_push_insn_1)
6836 (emit_push_insn, optimize_bitfield_assignment_op, expand_expr_real_1):
6838 * function.c (instantiate_virtual_regs_in_insn): Likewise.
6839 * ifcvt.c (noce_try_store_flag_constants): Likewise.
6840 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
6841 * modulo-sched.c (generate_prolog_epilog): Likewise.
6842 * optabs.c (expand_binop, widen_leading, expand_doubleword_clz)
6843 (expand_ctz, expand_ffs, expand_unop): Likewise.
6845 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
6847 * alias.c (addr_side_effect_eval): Use gen_int_mode with the mode
6848 of the associated gen_rtx_* call.
6849 * caller-save.c (init_caller_save): Likewise.
6850 * combine.c (find_split_point, make_extraction): Likewise.
6851 (make_compound_operation): Likewise.
6852 * dwarf2out.c (mem_loc_descriptor): Likewise.
6853 * explow.c (plus_constant, probe_stack_range): Likewise.
6854 * expmed.c (expand_mult_const): Likewise.
6855 * expr.c (emit_single_push_insn_1, do_tablejump): Likewise.
6856 * reload1.c (init_reload): Likewise.
6857 * valtrack.c (cleanup_auto_inc_dec): Likewise.
6858 * var-tracking.c (adjust_mems): Likewise.
6859 * modulo-sched.c (sms_schedule): Likewise, but use gen_rtx_GT
6860 rather than gen_rtx_fmt_ee.
6862 2013-09-09 Jan Hubicka <jh@suse.cz>
6865 * value-prof.c (gimple_ic): Copy also abnormal edges.
6867 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
6869 * asan.c (asan_shadow_cst): Use gen_int_mode.
6871 2013-09-08 Jan Hubicka <jh@suse.cz>
6873 * ipa-profile.c: Add toplevel comment.
6874 (ipa_propagate_frequency_1): Be more conservative when profile is read.
6875 (contains_hot_call_p): New function.
6876 (ipa_propagate_frequency): Set frequencies based on counts when
6878 * predict.c (compute_function_frequency): Use PROFILE_READ gueard for
6879 profile; do not tamper with profile after inlining if it is read.
6881 2013-09-08 Jan Hubicka <jh@suse.cz>
6883 * ipa-prop.c (try_make_edge_direct_simple_call): Do not special case
6886 2013-09-08 Jan Hubicka <jh@suse.cz>
6888 * ipa.c (walk_polymorphic_call_targets): Fix redirection before IPA
6891 2013-09-08 Jeff Law <law@redhat.com>
6894 * tree-ssa-threadedge.c (thread_across_edge): Fix initialization
6897 2013-09-08 Andi Kleen <ak@linux.intel.com>
6899 * tree-inline.c (estimate_num_insns): Limit asm cost to 1000.
6901 2013-09-08 Jan Hubicka <jh@suse.cz>
6903 * ipa.c (walk_polymorphic_call_targets): Fix inliner summary update.
6905 2013-09-08 Richard Sandiford <rdsandiford@googlemail.com>
6907 * ira.c (update_equiv_regs): Only call set_paradoxical_subreg
6908 for non-debug insns.
6909 * lra.c (new_insn_reg): Take the containing insn as a parameter.
6910 Only modify lra_reg_info[].biggest_mode if it's non-debug insn.
6911 (collect_non_operand_hard_regs, add_regs_to_insn_regno_info): Update
6914 2013-09-08 Jan Hubicka <jh@suse.cz>
6916 * cgraphunit.c (walk_polymorphic_call_targets): Permit 0 possible
6917 targets and devirtualize to BUILT_IN_UNREACHABLE.
6918 * timevar.def (TV_IPA_UNREACHABLE): New timevar.
6919 * ipa.c (walk_polymorphic_call_targets): New function.
6920 (symtab_remove_unreachable_nodes): Use it; do not keep all virtual
6921 functions; use the new timevar.
6922 * ipa-devirt.c (maybe_record_node): Do not insert static nodes that
6923 was removed from the program.
6924 (record_binfo): If BINFO corresponds to an anonymous namespace, we may
6925 not consider it in the walk when its vtable is dead.
6926 (possible_polymorphic_call_targets_1): Pass anonymous flag to
6928 (devirt_variable_node_removal_hook): New function.
6929 (possible_polymorphic_call_targets): Also register
6930 devirt_variable_node_removal_hook.
6931 (ipa_devirt): Do not do non-speculative devirtualization.
6932 (gate_ipa_devirt): One execute if devirtualizing speculatively.
6934 2013-09-08 Jan Hubicka <jh@suse.cz>
6936 * cgraph.h (varpool_node_hook, varpool_node_hook_list,
6937 varpool_add_node_removal_hook, varpool_add_variable_insertion_hook,
6938 varpool_remove_variable_insertion_hook): Declare.
6939 * varpool.c (varpool_node_hook_list): New structure.
6940 (first_varpool_node_removal_hook,
6941 first_varpool_variable_insertion_hook): New variables.
6942 (varpool_add_node_removal_hook, varpool_remove_node_removal_hook,
6943 varpool_call_node_removal_hooks, varpool_add_variable_insertion_hook,
6944 varpool_remove_variable_insertion_hook,
6945 varpool_call_variable_insertion_hooks): New functions.
6946 (varpool_remove_node): Use it.
6948 2013-09-08 Paolo Carlini <paolo.carlini@oracle.com>
6951 * diagnostic.c (diagnostic_build_prefix): When s.file is
6952 "<built-in>" don't output line and column numbers.
6954 2013-09-06 Jan Hubicka <jh@suse.cz>
6956 * cgraphunit.c (expand_thunk): Get body before touching arguments.
6957 * lto-streamer-out.c: Stream thunks, too.
6958 * lto-streamer-in.c (input_function): Pop cfun here
6959 (lto_read_body): Instead of here.
6961 2013-09-06 Caroline Tice <cmtice@google.com>
6963 * doc/install.texi: Add documentation for the --enable-vtable-verify
6964 and the --disable-libvtv configure options.
6966 2013-09-06 Jeff Law <law@redhat.com>
6968 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
6969 edge implied equivalences into successor phis.
6971 2013-09-06 Joern Rennecke <joern.rennecke@embecosm.com>
6973 * resource.c (mark_referenced_resources): Handle COND_EXEC.
6975 2013-09-06 Claudiu Zissulescu <claziss@synopsys.com>
6977 * resource.c (mark_target_live_regs): Compute resources taking
6978 into account if a call is predicated or not.
6980 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
6982 * toplev.c (output_stack_usage): Be prepared for suffixes created by
6983 the compiler in the function names.
6985 2013-09-06 Jan Hubicka <jh@suse.cz>
6988 * ipa-inline.c (has_caller_p): New function.
6989 (want_inline_function_to_all_callers_p): Use it.
6990 (sum_callers, inline_to_all_callers): Break out from ...
6991 (ipa_inline): ... here.
6993 2013-09-06 Jan Hubicka <jh@suse.cz>
6995 * config/i386/i386.c (ix86_hard_regno_mode_ok): AVX modes are valid
6996 only when AVX is enabled.
6998 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
7000 * config/aarch64/aarch64.md
7001 (*movtf_aarch64): Use neon_<ls>dm_2 as type where v8type
7002 is fpsimd_<load/store>2.
7003 (load_pair<mode>): Likewise.
7004 (store_pair<mode>): Likewise.
7006 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
7008 * config/arm/types.md (type): Add "mrs" type.
7009 * config/aarch64/aarch64.md
7010 (aarch64_load_tp_hard): Make type "mrs".
7012 (load_tp_hard): Make type "mrs".
7013 * config/arm/cortex-a15.md: Update with new attributes.
7014 * config/arm/cortex-a5.md: Update with new attributes.
7015 * config/arm/cortex-a53.md: Update with new attributes.
7016 * config/arm/cortex-a7.md: Update with new attributes.
7017 * config/arm/cortex-a8.md: Update with new attributes.
7018 * config/arm/cortex-a9.md: Update with new attributes.
7019 * config/arm/cortex-m4.md: Update with new attributes.
7020 * config/arm/cortex-r4.md: Update with new attributes.
7021 * config/arm/fa526.md: Update with new attributes.
7022 * config/arm/fa606te.md: Update with new attributes.
7023 * config/arm/fa626te.md: Update with new attributes.
7024 * config/arm/fa726te.md: Update with new attributes.
7026 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
7028 * config/aarch64/aarch64.md
7029 (*movti_aarch64): Use "multiple" for type where v8type is "move2".
7030 (*movtf_aarch64): Likewise.
7032 (thumb1_movdi_insn): Use "multiple" for type where more than one
7033 instruction is used for a move.
7034 (*arm32_movhf): Likewise.
7035 (*thumb_movdf_insn): Likewise.
7037 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
7039 * config/arm/types.md (type): Rename fcpys to fmov.
7041 (*arm_movsi_vfp): Rename type fcpys as fmov.
7042 (*thumb2_movsi_vfp): Likewise
7043 (*movhf_vfp_neon): Likewise
7044 (*movhf_vfp): Likewise
7045 (*movsf_vfp): Likewise
7046 (*thumb2_movsf_vfp): Likewise
7047 (*movsfcc_vfp): Likewise
7048 (*thumb2_movsfcc_vfp): Likewise
7049 * config/aarch64/aarch64-simd.md
7050 (move_lo_quad_<mode>): Replace type mov_reg with fmovs.
7051 * config/aarch64/aarch64.md
7052 (*movsi_aarch64): Replace type mov_reg with fmovs.
7053 (*movdi_aarch64): Likewise
7054 (*movsf_aarch64): Likewise
7055 (*movdf_aarch64): Likewise
7057 (cortexa7_older_only): Rename TYPE_FCPYS to TYPE_FMOV.
7058 * config/arm/iwmmxt.md
7059 (*iwmmxt_movsi_insn): Rename type fcpys as fmov.
7060 * config/arm/arm1020e.md: Update with new attributes.
7061 * config/arm/cortex-a15-neon.md: Update with new attributes.
7062 * config/arm/cortex-a5.md: Update with new attributes.
7063 * config/arm/cortex-a53.md: Update with new attributes.
7064 * config/arm/cortex-a7.md: Update with new attributes.
7065 * config/arm/cortex-a8-neon.md: Update with new attributes.
7066 * config/arm/cortex-a9.md: Update with new attributes.
7067 * config/arm/cortex-m4-fpu.md: Update with new attributes.
7068 * config/arm/cortex-r4f.md: Update with new attributes.
7069 * config/arm/marvell-pj4.md: Update with new attributes.
7070 * config/arm/vfp11.md: Update with new attributes.
7072 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
7074 * config/aarch64/aarch64.md
7075 (*madd<mode>): Fix type attribute.
7076 (*maddsi_uxtw): Likewise.
7077 (*msub<mode>): Likewise.
7078 (*msubsi_uxtw): Likewise.
7079 (<su_optab>maddsidi4): Likewise.
7080 (<su_optab>msubsidi4): Likewise.
7082 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
7084 * config/arm/types.md: Split fdiv<sd> as fsqrt<sd>, fdiv<sd>.
7085 * config/arm/arm.md (core_cycles): Remove fdiv.
7086 * config/arm/vfp.md:
7087 (*sqrtsf2_vfp): Update for attribute changes.
7088 (*sqrtdf2_vfp): Likewise.
7089 * config/aarch64/aarch64.md:
7090 (sqrt<mode>2): Update for attribute changes.
7091 * config/arm/arm1020e.md: Update with new attributes.
7092 * config/arm/cortex-a15-neon.md: Update with new attributes.
7093 * config/arm/cortex-a5.md: Update with new attributes.
7094 * config/arm/cortex-a53.md: Update with new attributes.
7095 * config/arm/cortex-a7.md: Update with new attributes.
7096 * config/arm/cortex-a8-neon.md: Update with new attributes.
7097 * config/arm/cortex-a9.md: Update with new attributes.
7098 * config/arm/cortex-m4-fpu.md: Update with new attributes.
7099 * config/arm/cortex-r4f.md: Update with new attributes.
7100 * config/arm/marvell-pj4.md: Update with new attributes.
7101 * config/arm/vfp11.md: Update with new attributes.
7103 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
7105 * config/arm/types.md
7106 (type): Split f_cvt as f_cvt, f_cvtf2i, f_cvti2f.
7107 * config/aarch64/aarch64.md
7108 (l<fcvt_pattern><su_optab><GPF:mode><GPI:mode>2): Update with
7110 (fix_trunc<GPF:mode><GPI:mode>2): Likewise.
7111 (fixuns_trunc<GPF:mode><GPI:mode>2): Likewise.
7112 (float<GPI:mode><GPF:mode>2): Likewise.
7114 (*truncsisf2_vfp): Update with new attributes.
7115 (*truncsidf2_vfp): Likewise.
7116 (fixuns_truncsfsi2): Likewise.
7117 (fixuns_truncdfsi2): Likewise.
7118 (*floatsisf2_vfp): Likewise.
7119 (*floatsidf2_vfp): Likewise.
7120 (floatunssisf2): Likewise.
7121 (floatunssidf2): Likewise.
7122 (*combine_vcvt_f32_<FCVTI32typename>): Likewise.
7123 (*combine_vcvt_f64_<FCVTI32typename>): Likewise.
7124 * config/arm/arm1020e.md: Update with new attributes.
7125 * config/arm/cortex-a15-neon.md: Update with new attributes.
7126 * config/arm/cortex-a5.md: Update with new attributes.
7127 * config/arm/cortex-a53.md: Update with new attributes.
7128 * config/arm/cortex-a7.md: Update with new attributes.
7129 * config/arm/cortex-a8-neon.md: Update with new attributes.
7130 * config/arm/cortex-a9.md: Update with new attributes.
7131 * config/arm/cortex-m4-fpu.md: Update with new attributes.
7132 * config/arm/cortex-r4f.md: Update with new attributes.
7133 * config/arm/marvell-pj4.md: Update with new attributes.
7134 * config/arm/vfp11.md: Update with new attributes.
7136 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
7138 * config/aarch64/arm_neon.h
7139 (vqtbl<1,2,3,4><q>_s8): Fix control vector parameter type.
7140 (vqtbx<1,2,3,4><q>_s8): Likewise.
7142 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
7144 * config/arm/types.md: Add "no_insn", "multiple" and "untyped" types.
7145 * config/arm/arm-fixed.md: Add type attribute to all insn patterns.
7146 (add<mode>3): Add type attribute.
7147 (add<mode>3): Likewise.
7148 (usadd<mode>3): Likewise.
7149 (ssadd<mode>3): Likewise.
7150 (sub<mode>3): Likewise.
7151 (sub<mode>3): Likewise.
7152 (ussub<mode>3): Likewise.
7153 (sssub<mode>3): Likewise.
7154 (ssmulsa3): Likewise.
7155 (usmulusa3): Likewise.
7156 (arm_usatsihi): Likewise.
7158 (*movdi_vfp): Add types for all instructions.
7159 (*movdi_vfp_cortexa8): Likewise.
7160 (*movhf_vfp_neon): Likewise.
7161 (*movhf_vfp): Likewise.
7162 (*movdf_vfp): Likewise.
7163 (*thumb2_movdf_vfp): Likewise.
7164 (*thumb2_movdfcc_vfp): Likewise.
7165 * config/arm/arm.md: Add type attribute to all insn patterns.
7166 (*thumb1_adddi3): Add type attribute.
7167 (*arm_adddi3): Likewise.
7168 (*adddi_sesidi_di): Likewise.
7169 (*adddi_zesidi_di): Likewise.
7170 (*thumb1_addsi3): Likewise.
7171 (addsi3_compare0): Likewise.
7172 (*addsi3_compare0_scratch): Likewise.
7173 (*compare_negsi_si): Likewise.
7174 (cmpsi2_addneg): Likewise.
7175 (*addsi3_carryin_<optab>): Likewise.
7176 (*addsi3_carryin_alt2_<optab>): Likewise.
7177 (*addsi3_carryin_clobercc_<optab>): Likewise.
7178 (*subsi3_carryin): Likewise.
7179 (*subsi3_carryin_const): Likewise.
7180 (*subsi3_carryin_compare): Likewise.
7181 (*subsi3_carryin_compare_const): Likewise.
7182 (*arm_subdi3): Likewise.
7183 (*thumb_subdi3): Likewise.
7184 (*subdi_di_zesidi): Likewise.
7185 (*subdi_di_sesidi): Likewise.
7186 (*subdi_zesidi_di): Likewise.
7187 (*subdi_sesidi_di): Likewise.
7188 (*subdi_zesidi_ze): Likewise.
7189 (thumb1_subsi3_insn): Likewise.
7190 (*arm_subsi3_insn): Likewise.
7191 (*anddi3_insn): Likewise.
7192 (*anddi_zesidi_di): Likewise.
7193 (*anddi_sesdi_di): Likewise.
7194 (*ne_zeroextracts): Likewise.
7195 (*ne_zeroextracts): Likewise.
7196 (*ite_ne_zeroextr): Likewise.
7197 (*ite_ne_zeroextr): Likewise.
7198 (*anddi_notdi_di): Likewise.
7199 (*anddi_notzesidi): Likewise.
7200 (*anddi_notsesidi): Likewise.
7201 (andsi_notsi_si): Likewise.
7202 (thumb1_bicsi3): Likewise.
7203 (*iordi3_insn): Likewise.
7204 (*iordi_zesidi_di): Likewise.
7205 (*iordi_sesidi_di): Likewise.
7206 (*thumb1_iorsi3_insn): Likewise.
7207 (*xordi3_insn): Likewise.
7208 (*xordi_zesidi_di): Likewise.
7209 (*xordi_sesidi_di): Likewise.
7210 (*arm_xorsi3): Likewise.
7211 (*andsi_iorsi3_no): Likewise.
7212 (*smax_0): Likewise.
7213 (*smax_m1): Likewise.
7214 (*arm_smax_insn): Likewise.
7215 (*smin_0): Likewise.
7216 (*arm_smin_insn): Likewise.
7217 (*arm_umaxsi3): Likewise.
7218 (*arm_uminsi3): Likewise.
7219 (*minmax_arithsi): Likewise.
7220 (*minmax_arithsi_): Likewise.
7221 (*satsi_<SAT:code>): Likewise.
7222 (arm_ashldi3_1bit): Likewise.
7223 (arm_ashrdi3_1bit): Likewise.
7224 (arm_lshrdi3_1bit): Likewise.
7225 (*arm_negdi2): Likewise.
7226 (*thumb1_negdi2): Likewise.
7227 (*arm_negsi2): Likewise.
7228 (*thumb1_negsi2): Likewise.
7229 (*negdi_extendsid): Likewise.
7230 (*negdi_zero_extend): Likewise.
7231 (*arm_abssi2): Likewise.
7232 (*thumb1_abssi2): Likewise.
7233 (*arm_neg_abssi2): Likewise.
7234 (*thumb1_neg_abss): Likewise.
7235 (one_cmpldi2): Likewise.
7236 (extend<mode>di2): Likewise.
7237 (*compareqi_eq0): Likewise.
7238 (*arm_extendhisi2addsi): Likewise.
7239 (*arm_movdi): Likewise.
7240 (*thumb1_movdi_insn): Likewise.
7241 (*arm_movt): Likewise.
7242 (*thumb1_movsi_insn): Likewise.
7243 (pic_add_dot_plus_four): Likewise.
7244 (pic_add_dot_plus_eight): Likewise.
7245 (tls_load_dot_plus_eight): Likewise.
7246 (*thumb1_movhi_insn): Likewise.
7247 (*thumb1_movsf_insn): Likewise.
7248 (*movdf_soft_insn): Likewise.
7249 (*thumb_movdf_insn): Likewise.
7250 (cbranchsi4_insn): Likewise.
7251 (cbranchsi4_scratch): Likewise.
7252 (*negated_cbranchsi4): Likewise.
7253 (*tbit_cbranch): Likewise.
7254 (*tlobits_cbranch): Likewise.
7255 (*tstsi3_cbranch): Likewise.
7256 (*cbranchne_decr1): Likewise.
7257 (*addsi3_cbranch): Likewise.
7258 (*addsi3_cbranch_scratch): Likewise.
7259 (*arm_cmpdi_insn): Likewise.
7260 (*arm_cmpdi_unsig): Likewise.
7261 (*arm_cmpdi_zero): Likewise.
7262 (*thumb_cmpdi_zero): Likewise.
7263 (*deleted_compare): Likewise.
7264 (*mov_scc): Likewise.
7265 (*mov_negscc): Likewise.
7266 (*mov_notscc): Likewise.
7267 (*cstoresi_eq0_thumb1_insn): Likewise.
7268 (cstoresi_nltu_thumb1): Likewise.
7269 (cstoresi_ltu_thu): Likewise.
7270 (thumb1_addsi3_addgeu): Likewise.
7271 (*arm_jump): Likewise.
7272 (*thumb_jump): Likewise.
7273 (*check_arch2): Likewise.
7274 (arm_casesi_internal): Likewise.
7275 (thumb1_casesi_dispatch): Likewise.
7276 (*arm_indirect_jump): Likewise.
7277 (*thumb1_indirect_jump): Likewise.
7279 (*and_scc): Likewise.
7280 (*ior_scc): Likewise.
7281 (*compare_scc): Likewise.
7282 (*cond_move): Likewise.
7283 (*cond_arith): Likewise.
7284 (*cond_sub): Likewise.
7285 (*cmp_ite0): Likewise.
7286 (*cmp_ite1): Likewise.
7287 (*cmp_and): Likewise.
7288 (*cmp_ior): Likewise.
7289 (*ior_scc_scc): Likewise.
7290 (*ior_scc_scc_cmp): Likewise.
7291 (*and_scc_scc): Likewise.
7292 (*and_scc_scc_cmp): Likewise.
7293 (*and_scc_scc_nod): Likewise.
7294 (*negscc): Likewise.
7295 (movcond_addsi): Likewise.
7296 (movcond): Likewise.
7297 (*ifcompare_plus_move): Likewise.
7298 (*if_plus_move): Likewise.
7299 (*ifcompare_move_plus): Likewise.
7300 (*if_move_plus): Likewise.
7301 (*ifcompare_arith_arith): Likewise.
7302 (*if_arith_arith): Likewise.
7303 (*ifcompare_arith_move): Likewise.
7304 (*if_arith_move): Likewise.
7305 (*ifcompare_move_arith): Likewise.
7306 (*if_move_arith): Likewise.
7307 (*ifcompare_move_not): Likewise.
7308 (*if_move_not): Likewise.
7309 (*ifcompare_not_move): Likewise.
7310 (*if_not_move): Likewise.
7311 (*ifcompare_shift_move): Likewise.
7312 (*if_shift_move): Likewise.
7313 (*ifcompare_move_shift): Likewise.
7314 (*if_move_shift): Likewise.
7315 (*ifcompare_shift_shift): Likewise.
7316 (*ifcompare_not_arith): Likewise.
7317 (*ifcompare_arith_not): Likewise.
7318 (*if_arith_not): Likewise.
7319 (*ifcompare_neg_move): Likewise.
7320 (*if_neg_move): Likewise.
7321 (*ifcompare_move_neg): Likewise.
7322 (*if_move_neg): Likewise.
7323 (prologue_thumb1_interwork): Likewise.
7324 (*cond_move_not): Likewise.
7325 (*sign_extract_onebit): Likewise.
7326 (*not_signextract_onebit): Likewise.
7327 (stack_tie): Likewise.
7328 (align_4): Likewise.
7329 (align_8): Likewise.
7330 (consttable_end): Likewise.
7331 (consttable_1): Likewise.
7332 (consttable_2): Likewise.
7333 (consttable_4): Likewise.
7334 (consttable_8): Likewise.
7335 (consttable_16): Likewise.
7336 (*thumb1_tablejump): Likewise.
7337 (prefetch): Likewise.
7338 (force_register_use): Likewise.
7339 (thumb_eh_return): Likewise.
7340 (load_tp_hard): Likewise.
7341 (load_tp_soft): Likewise.
7342 (tlscall): Likewise.
7343 (*arm_movtas_ze): Likewise.
7344 (*arm_rev): Likewise.
7345 (*arm_revsh): Likewise.
7346 (*arm_rev16): Likewise.
7347 * config/arm/thumb2.md
7348 (*thumb2_smaxsi3): Likewise.
7349 (*thumb2_sminsi3): Likewise.
7350 (*thumb32_umaxsi3): Likewise.
7351 (*thumb2_uminsi3): Likewise.
7352 (*thumb2_negdi2): Likewise.
7353 (*thumb2_abssi2): Likewise.
7354 (*thumb2_neg_abss): Likewise.
7355 (*thumb2_movsi_insn): Likewise.
7356 (tls_load_dot_plus_four): Likewise.
7357 (*thumb2_movhi_insn): Likewise.
7358 (*thumb2_mov_scc): Likewise.
7359 (*thumb2_mov_negs): Likewise.
7360 (*thumb2_mov_negs): Likewise.
7361 (*thumb2_mov_nots): Likewise.
7362 (*thumb2_mov_nots): Likewise.
7363 (*thumb2_movsicc_): Likewise.
7364 (*thumb2_movsfcc_soft_insn): Likewise.
7365 (*thumb2_indirect_jump): Likewise.
7366 (*thumb2_and_scc): Likewise.
7367 (*thumb2_ior_scc): Likewise.
7368 (*thumb2_ior_scc_strict_it): Likewise.
7369 (*thumb2_cond_move): Likewise.
7370 (*thumb2_cond_arith): Likewise.
7371 (*thumb2_cond_ari): Likewise.
7372 (*thumb2_cond_sub): Likewise.
7373 (*thumb2_negscc): Likewise.
7374 (*thumb2_movcond): Likewise.
7375 (thumb2_casesi_internal): Likewise.
7376 (thumb2_casesi_internal_pic): Likewise.
7377 (*thumb2_alusi3_short): Likewise.
7378 (*thumb2_mov<mode>_shortim): Likewise.
7379 (*thumb2_addsi_short): Likewise.
7380 (*thumb2_subsi_short): Likewise.
7381 (thumb2_addsi3_compare0): Likewise.
7382 (*thumb2_cbz): Likewise.
7383 (*thumb2_cbnz): Likewise.
7384 (*thumb2_one_cmplsi2_short): Likewise.
7385 (*thumb2_negsi2_short): Likewise.
7386 (*orsi_notsi_si): Likewise.
7387 * config/arm/arm1020e.md: Update with new attributes.
7388 * config/arm/arm1026ejs.md: Update with new attributes.
7389 * config/arm/arm1136jfs.md: Update with new attributes.
7390 * config/arm/arm926ejs.md: Update with new attributes.
7391 * config/arm/cortex-a15.md: Update with new attributes.
7392 * config/arm/cortex-a5.md: Update with new attributes.
7393 * config/arm/cortex-a53.md: Update with new attributes.
7394 * config/arm/cortex-a7.md: Update with new attributes.
7395 * config/arm/cortex-a8.md: Update with new attributes.
7396 * config/arm/cortex-a9.md: Update with new attributes.
7397 * config/arm/cortex-m4.md: Update with new attributes.
7398 * config/arm/cortex-r4.md: Update with new attributes.
7399 * config/arm/fa526.md: Update with new attributes.
7400 * config/arm/fa606te.md: Update with new attributes.
7401 * config/arm/fa626te.md: Update with new attributes.
7402 * config/arm/fa726te.md: Update with new attributes.
7404 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
7406 * config/aarch64/aarch64-simd.md
7407 (aarch64_sqdml<SBINQOPS:as>l_n<mode>_internal): Use
7408 <vwx> iterator to ensure correct register choice.
7409 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
7410 (aarch64_sqdmull_n<mode>): Likewise.
7411 (aarch64_sqdmull2_n<mode>_internal): Likewise.
7412 * config/aarch64/arm_neon.h
7413 (vml<as><q>_lane<q>_<su>16): Use 'x' constraint for element vector.
7414 (vml<as><q>_n_<su>16): Likewise.
7415 (vml<as>l_high_lane<q>_<su>16): Likewise.
7416 (vml<as>l_high_n_<su>16): Likewise.
7417 (vml<as>l_lane<q>_<su>16): Likewise.
7418 (vml<as>l_n_<su>16): Likewise.
7419 (vmul<q>_lane<q>_<su>16): Likewise.
7420 (vmul<q>_n_<su>16): Likewise.
7421 (vmull_lane<q>_<su>16): Likewise.
7422 (vmull_n_<su>16): Likewise.
7423 (vmull_high_lane<q>_<su>16): Likewise.
7424 (vmull_high_n_<su>16): Likewise.
7425 (vqrdmulh<q>_n_s16): Likewise.
7427 2013-09-06 Tejas Belagod <tejas.belagod@arm.com>
7429 * config/aarch64/arm_neon.h: Fix all vdup<bhsd_lane<q> intrinsics to
7430 have the correct lane parameter.
7432 2013-09-06 Richard Biener <rguenther@suse.de>
7434 * cfganal.c (control_dependences::~control_dependences):
7435 Properly free all of the vector.
7437 2013-09-06 Kirill Yukhin <kirill.yukhin@intel.com>
7440 * config/i386/i386.c (ix86_conditional_register_usage):
7441 Proper initialize extended SSE registers.
7443 2013-09-06 Jan Hubicka <jh@suse.cz>
7445 PR tree-optimization/58311
7446 * ipa-devirt.c (gate_ipa_devirt): Only execute when optimizing.
7448 2013-09-06 Jan Hubicka <jh@suse.cz>
7450 * Makefile.in (tree-sra.o): Update dependencies.
7451 * tree-sra.c: Include ipa-utils.h
7452 (scan_function): Use recursive_call_p.
7453 (has_caller_p): New function.
7454 (cgraph_for_node_and_aliases): Count also callers of aliases.
7456 2013-09-06 Jan Hubicka <jh@suse.cz>
7459 * cgraph.h (symtab_semantically_equivalent_p): Declare.
7460 * tree-tailcall.c: Include ipa-utils.h.
7461 (find_tail_calls): Use it.
7462 * ipa-pure-const.c (check_call): Likewise.
7463 * ipa-utils.c (recursive_call_p): New function.
7464 * ipa-utils.h (recursive_call_p): Dclare.
7465 * symtab.c (symtab_nonoverwritable_alias): Fix formatting.
7466 (symtab_semantically_equivalent_p): New function.
7467 * Makefile.in (tree-tailcall.o): Update dependencies.
7469 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
7471 * ipa-split.c (split_function): Set DECL_NO_INLINE_WARNING_P on the
7474 2013-09-06 Richard Biener <rguenther@suse.de>
7476 * lto-streamer.h (lto_global_var_decls): Remove.
7477 * Makefile.in (OBJS): Remove lto-symtab.o.
7478 (lto-symtab.o): Remove.
7479 (GTFILES): Remove lto-symtab.c
7480 * lto-symtab.c: Move to lto/
7482 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7484 * config/s390/s390.md (UNSPEC_FPINT_FLOOR, UNSPEC_FPINT_BTRUNC)
7485 (UNSPEC_FPINT_ROUND, UNSPEC_FPINT_CEIL, UNSPEC_FPINT_NEARBYINT)
7486 (UNSPEC_FPINT_RINT): New constant definitions.
7487 (FPINT, fpint_name, fpint_roundingmode): New integer iterator
7488 definition with 2 attributes.
7489 ("<FPINT:fpint_name><BFP:mode>2", "rint<BFP:mode>2")
7490 ("<FPINT:fpint_name><DFP:mode>2", "rint<DFP:mode>2"): New pattern
7493 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7495 * config/s390/s390.md: Add "bcr_flush" value to mnemonic attribute.
7496 ("mem_thread_fence_1"): Use bcr 14,0 for z196 and later.
7497 Set the mnemonic attribute to "bcr_flush". Set the "z196prop"
7498 attribute to "z196_alone".
7499 * config/s390/2827.md: Add "bcr_flush" to "ooo_groupalone" and
7502 2013-09-06 Richard Biener <rguenther@suse.de>
7504 * basic-block.h (class control_dependences): New.
7505 * tree-ssa-dce.c (control_dependence_map): Remove.
7507 (EXECUTE_IF_CONTROL_DEPENDENT): Remove.
7508 (set_control_dependence_map_bit, clear_control_dependence_bitmap,
7509 find_pdom, find_control_dependence, find_all_control_dependences):
7511 (mark_control_dependent_edges_necessary,
7512 find_obviously_necessary_stmts, propagate_necessity, tree_dce_init,
7513 tree_dce_done, perform_tree_ssa_dce): Adjust.
7514 * cfganal.c (set_control_dependence_map_bit,
7515 clear_control_dependence_bitmap, find_pdom, find_control_dependence,
7516 find_all_control_dependences): Move from tree-ssa-dce.c and
7517 implement as methods of control_dependences class.
7518 (control_dependences::control_dependences): New.
7519 (control_dependences::~control_dependences): Likewise.
7520 (control_dependences::get_edges_dependent_on): Likewise.
7521 (control_dependences::get_edge): Likewise.
7523 2013-09-04 Jan Hubicka <jh@suse.cz>
7525 * tree.c (types_same_for_odr): Drop overactive check.
7526 * ipa-devirt.c (hash_type_name): Likewise.
7528 2013-09-04 Jan Hubicka <jh@suse.cz>
7530 * cgraphunit.c (walk_polymorphic_call_targets): Break out from ...
7531 (analyze_functions): ... here.
7533 2013-09-04 Jan Hubicka <jh@suse.cz>
7536 * cgraphunit.c (analyze_functions): Clear AUX fields
7537 after processing; initialize assembler name has.
7539 2013-09-05 Jeff Law <law@redhat.com>
7541 * tree-ssa-threadedge.c (thread_around_empty_blocks): Renamed
7542 from thread_around_empty_block. Record threading path into PATH.
7543 Recurse if threading through the initial block is successful.
7544 (thread_across_edge): Corresponding changes to slightly simplify.
7546 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
7548 * config/aarch64/aarch64.md
7549 (type): Remove frecpe, frecps, frecpx.
7550 (aarch64_frecp<FRECP:frecp_suffix><mode>): Move to aarch64-simd.md,
7551 fix to be a TARGET_SIMD instruction.
7552 (aarch64_frecps): Remove.
7553 * config/aarch64/aarch64-simd.md
7554 (aarch64_frecp<FRECP:frecp_suffix><mode>): New, moved from aarch64.md
7555 (aarch64_frecps<mode>): Handle all float/vector of float modes.
7557 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
7558 Sofiane Naci <sofiane.naci@arm.com>
7560 * config/arm/types.md (define_attr "type"): Expand "arlo_imm"
7561 into "adr", "alu_imm", "alus_imm", "logic_imm", "logics_imm".
7562 Expand "arlo_reg" into "adc_reg", "adc_imm", "adcs_reg", "adcs_imm",
7563 "alu_ext", "alu_reg", "alus_ext", "alus_reg", "bfm", "csel",
7564 "logic_reg", "logics_reg", "rev". Expand "arlo_shift" into
7565 "alu_shift_imm", "alus_shift_imm", "logic_shift_imm",
7566 "logics_shift_imm". Expand "arlo_shift_reg" into "alu_shift_reg",
7567 "alus_shift_reg", "logic_shift_reg", "logics_shift_reg". Expand "clz"
7568 into "clz, "rbit". Rename "shift" to "shift_imm".
7569 * config/arm/arm.md (define_attr "core_cycles"): Update for attribute
7570 changes. Update for attribute changes all occurrences of arlo_* and
7572 * config/arm/arm-fixed.md: Update for attribute changes
7573 all occurrences of arlo_* types.
7574 * config/arm/thumb2.md: Update for attribute changes all occurrences
7576 * config/arm/arm.c (xscale_sched_adjust_cost): (rtx insn, rtx
7577 (cortexa7_older_only): Likewise.
7578 (cortexa7_younger): Likewise.
7579 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
7580 (1020alu_shift_op): Likewise.
7581 (1020alu_shift_reg_op): Likewise.
7582 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
7583 (alu_shift_op): Likewise.
7584 (alu_shift_reg_op): Likewise.
7585 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
7586 (11_alu_shift_op): Likewise.
7587 (11_alu_shift_reg_op): Likewise.
7588 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
7589 (9_alu_shift_reg_op): Likewise.
7590 * config/arm/cortex-a15.md (cortex_a15_alu): Update for
7592 (cortex_a15_alu_shift): Likewise.
7593 (cortex_a15_alu_shift_reg): Likewise.
7594 * config/arm/cortex-a5.md (cortex_a5_alu): Update for
7596 (cortex_a5_alu_shift): Likewise.
7597 * config/arm/cortex-a53.md (cortex_a53_alu): Update for
7599 (cortex_a53_alu_shift): Likewise.
7600 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for
7602 (cortex_a7_alu_reg): Likewise.
7603 (cortex_a7_alu_shift): Likewise.
7604 * config/arm/cortex-a8.md (cortex_a8_alu): Update for
7606 (cortex_a8_alu_shift): Likewise.
7607 (cortex_a8_alu_shift_reg): Likewise.
7608 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
7609 (cortex_a9_dp_shift): Likewise.
7610 * config/arm/cortex-m4.md (cortex_m4_alu): Update for
7612 * config/arm/cortex-r4.md
7613 (cortex_r4_alu): Update for attribute changes.
7614 (cortex_r4_mov): Likewise.
7615 (cortex_r4_alu_shift_reg): Likewise.
7616 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
7617 (526_alu_shift_op): Likewise.
7618 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
7619 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
7620 (626te_alu_shift_op): Likewise.
7621 * config/arm/fa726te.md (726te_alu_op): Update for attribute changes.
7622 (726te_alu_shift_op): Likewise.
7623 (726te_alu_shift_reg_op): Likewise.
7624 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
7625 (mp626_alu_shift_op): Likewise.
7626 * config/arm/marvell-pj4.md (pj4_alu): Update for attribute changes.
7627 (pj4_alu_conds): Likewise.
7628 (pj4_shift): Likewise.
7629 (pj4_shift_conds): Likewise.
7630 (pj4_alu_shift): Likewise.
7631 (pj4_alu_shift_conds): Likewise.
7632 * config/aarch64/aarch64.md: Update for attribute change
7633 all occurrences of arlo_* and shift* types.
7635 2013-09-05 Mike Stump <mikestump@comcast.net>
7637 * tree.h: Move documentation for tree_function_decl to tree-core.h
7638 with the declaration.
7640 2013-09-05 Peter Bergner <bergner@vnet.ibm.com>
7643 * reginfo.c (choose_hard_reg_mode): Scan through all mode classes
7644 looking for widest mode.
7646 2013-09-05 Eric Botcazou <ebotcazou@adacore.com>
7648 * config.gcc (*-*-vxworks*): Do not override an existing extra_objs.
7650 2013-09-05 Richard Biener <rguenther@suse.de>
7652 PR tree-optimization/58137
7653 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size):
7654 Do not create vectors of pointers.
7655 * tree-vect-loop.c (get_initial_def_for_induction): Use proper
7656 types for the components of the vector initializer.
7657 * tree-cfg.c (verify_gimple_assign_binary): Remove special-casing
7658 allowing pointer vectors with PLUS_EXPR/MINUS_EXPR.
7660 2013-09-05 Martin Jambor <mjambor@suse.cz>
7662 * ipa-prop.c (remove_described_reference): Accept missing references,
7663 return false if that hppens, otherwise return true.
7664 (cgraph_node_for_jfunc): New function.
7665 (try_decrement_rdesc_refcount): Likewise.
7666 (try_make_edge_direct_simple_call): Use them.
7667 (ipa_edge_removal_hook): Remove references from rdescs.
7668 (ipa_edge_duplication_hook): Clone rdescs and their references
7669 when the new edge has the same caller as the old one.
7670 * cgraph.c (cgraph_resolve_speculation): Remove speculative
7671 reference before removing any edges.
7673 2013-09-05 Richard Earnshaw <rearnsha@arm.com>
7675 * arm.c (thumb2_emit_strd_push): Rewrite to use pre-decrement on
7677 * thumb2.md (thumb2_storewb_parisi): New pattern.
7679 2013-09-05 Yufeng Zhang <yufeng.zhang@arm.com>
7681 * config/aarch64/aarch64-option-extensions.def: Add
7682 AARCH64_OPT_EXTENSION of 'crc'.
7683 * config/aarch64/aarch64.h (AARCH64_FL_CRC): New define.
7684 (AARCH64_ISA_CRC): Ditto.
7685 * doc/invoke.texi (-march and -mcpu feature modifiers): Add
7686 description of the CRC extension.
7688 2013-09-05 Alexander Ivchenko <alexander.ivchenko@intel.com>
7690 * config/rs6000/linux64.h: Define OPTION_BIONIC and OPTION_UCLIBC.
7691 * config/rs6000/linux.h: Ditto.
7692 * alpha/linux.h: Ditto.
7693 * config/bfin/uclinux.h: Define TARGET_LIBC_HAS_FUNCTION as
7694 no_c99_libc_has_function.
7695 * config/c6x/uclinux-elf.h: Ditto.
7696 * config/lm32/uclinux-elf.h: Ditto.
7697 * config/m68k/uclinux.h: Ditto.
7698 * config/moxie/uclinux.h: Ditto.
7699 * config.gcc (bfin*-linux-uclibc*): Add t-linux-android to tmake_file.
7700 (crisv32-*-linux*, cris-*-linux*): Ditto.
7701 * config/bfin/bfin.c: Include "tm_p.h".
7703 2013-09-05 Richard Biener <rguenther@suse.de>
7705 * tree-vect-loop.c (vect_analyze_loop_operations): Properly
7706 check for a definition without a basic-block.
7708 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
7709 Sofiane Naci <sofiane.naci@arm.com>
7711 * config/aarch64/aarch64.md
7712 (*movti_aarch64): Rename r_2_f and f_2_r.
7713 (*movsf_aarch64): Likewise.
7714 (*movdf_aarch64): Likewise.
7715 (*movtf_aarch64): Likewise.
7716 (aarch64_movdi_<mode>low): Likewise.
7717 (aarch64_movdi_<mode>high): Likewise.
7718 (aarch64_mov<mode>high_di): Likewise.
7719 (aarch64_mov<mode>low_di): Likewise.
7720 (aarch64_movtilow_tilow): Likewise.
7721 * config/arm/arm.md (attribute "neon_type"): Delete. Move attribute
7722 values to config/arm/types.md
7723 (attribute "conds"): Update for attribute change.
7724 (anddi3_insn): Likewise.
7725 (iordi3_insn): Likewise.
7726 (xordi3_insn): Likewise.
7727 (one_cmpldi2): Likewise.
7728 * config/arm/types.md (type): Add Neon types.
7729 * config/arm/neon.md (neon_mov<mode>): Remove "neon_type" attribute,
7730 use "type" attribute.
7731 (movmisalign<mode>_neon_store): Likewise.
7732 (movmisalign<mode>_neon_load): Likewise.
7733 (vec_set<mode>_internal): Likewise.
7734 (vec_setv2di_internal): Likewise.
7735 (vec_extract<mode>): Likewise.
7736 (vec_extractv2di): Likewise.
7737 (add<mode>3_neon): Likewise.
7738 (adddi3_neon): Likewise.
7739 (sub<mode>3_neon): Likewise.
7740 (subdi3_neon): Likewise.
7741 (mul<mode>3_neon): Likewise.
7742 (mul<mode>3add<mode>_neon): Likewise.
7743 (mul<mode>3neg<mode>add<mode>_neon): Likewise.
7744 (fma<VCVTF:mode>4)): Likewise.
7745 (fma<VCVTF:mode>4_intrinsic): Likewise.
7746 (fmsub<VCVTF:mode>4)): Likewise.
7747 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
7748 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
7749 (ior<mode>3): Likewise.
7750 (and<mode>3): Likewise.
7751 (anddi3_neon): Likewise.
7752 (orn<mode>3_neon): Likewise.
7753 (orndi3_neon): Likewise.
7754 (bic<mode>3_neon): Likewise.
7755 (bicdi3_neon): Likewise.
7756 (xor<mode>3): Likewise.
7757 (one_cmpl<mode>2): Likewise.
7758 (abs<mode>2): Likewise.
7759 (neg<mode>2): Likewise.
7760 (umin<mode>3_neon): Likewise.
7761 (umax<mode>3_neon): Likewise.
7762 (smin<mode>3_neon): Likewise.
7763 (smax<mode>3_neon): Likewise.
7764 (vashl<mode>3): Likewise.
7765 (vashr<mode>3_imm): Likewise.
7766 (vlshr<mode>3_imm): Likewise.
7767 (ashl<mode>3_signed): Likewise.
7768 (ashl<mode>3_unsigned): Likewise.
7769 (neon_load_count): Likewise.
7770 (ashldi3_neon_noclobber): Likewise.
7771 (signed_shift_di3_neon): Likewise.
7772 (unsigned_shift_di3_neon): Likewise.
7773 (ashrdi3_neon_imm_noclobber): Likewise.
7774 (lshrdi3_neon_imm_noclobber): Likewise.
7775 (widen_ssum<mode>3): Likewise.
7776 (widen_usum<mode>3): Likewise.
7777 (quad_halves_<code>v4si): Likewise.
7778 (quad_halves_<code>v4sf): Likewise.
7779 (quad_halves_<code>v8hi): Likewise.
7780 (quad_halves_<code>v16qi): Likewise.
7781 (reduc_splus_v2di): Likewise.
7782 (neon_vpadd_internal<mode>): Likewise.
7783 (neon_vpsmin<mode>): Likewise.
7784 (neon_vpsmax<mode>): Likewise.
7785 (neon_vpumin<mode>): Likewise.
7786 (neon_vpumax<mode>): Likewise.
7787 (ss_add<mode>_neon): Likewise.
7788 (us_add<mode>_neon): Likewise.
7789 (ss_sub<mode>_neon): Likewise.
7790 (us_sub<mode>_neon): Likewise.
7791 (neon_vadd<mode>_unspec): Likewise.
7792 (neon_vaddl<mode>): Likewise.
7793 (neon_vaddw<mode>): Likewise.
7794 (neon_vhadd<mode>): Likewise.
7795 (neon_vqadd<mode>): Likewise.
7796 (neon_vaddhn<mode>): Likewise.
7797 (neon_vmul<mode>): Likewise.
7798 (neon_vmla<mode>): Likewise.
7799 (neon_vmlal<mode>): Likewise.
7800 (neon_vmls<mode>): Likewise.
7801 (neon_vmlsl<mode>): Likewise.
7802 (neon_vqdmulh<mode>): Likewise.
7803 (neon_vqdmlal<mode>): Likewise.
7804 (neon_vqdmlsl<mode>): Likewise.
7805 (neon_vmull<mode>): Likewise.
7806 (neon_vqdmull<mode>): Likewise.
7807 (neon_vsub<mode>_unspec): Likewise.
7808 (neon_vsubl<mode>): Likewise.
7809 (neon_vsubw<mode>): Likewise.
7810 (neon_vqsub<mode>): Likewise.
7811 (neon_vhsub<mode>): Likewise.
7812 (neon_vsubhn<mode>): Likewise.
7813 (neon_vceq<mode>): Likewise.
7814 (neon_vcge<mode>): Likewise.
7815 (neon_vcgeu<mode>): Likewise.
7816 (neon_vcgt<mode>): Likewise.
7817 (neon_vcgtu<mode>): Likewise.
7818 (neon_vcle<mode>): Likewise.
7819 (neon_vclt<mode>): Likewise.
7820 (neon_vcage<mode>): Likewise.
7821 (neon_vcagt<mode>): Likewise.
7822 (neon_vtst<mode>): Likewise.
7823 (neon_vabd<mode>): Likewise.
7824 (neon_vabdl<mode>): Likewise.
7825 (neon_vaba<mode>): Likewise.
7826 (neon_vabal<mode>): Likewise.
7827 (neon_vmax<mode>): Likewise.
7828 (neon_vmin<mode>): Likewise.
7829 (neon_vpaddl<mode>): Likewise.
7830 (neon_vpadal<mode>): Likewise.
7831 (neon_vpmax<mode>): Likewise.
7832 (neon_vpmin<mode>): Likewise.
7833 (neon_vrecps<mode>): Likewise.
7834 (neon_vrsqrts<mode>): Likewise.
7835 (neon_vqabs<mode>): Likewise.
7836 (neon_vqneg<mode>): Likewise.
7837 (neon_vcls<mode>): Likewise.
7838 (clz<mode>2): Likewise.
7839 (popcount<mode>2): Likewise.
7840 (neon_vrecpe): Likewise.
7841 (neon_vrsqrte): Likewise.
7842 (neon_vget_lane<mode>_sext_internal): Likewise.
7843 (neon_vget_lane<mode>_zext_internal): Likewise.
7844 (neon_vdup_n<mode>): Likewise.
7845 (neon_vdup_nv2di): Likewise.
7846 (neon_vdpu_lane<mode>_internal): Likewise.
7847 (neon_vswp<mode>): Likewise.
7848 (float<mode><V_cvtto>2): Likewise.
7849 (floatuns<mode><V_cvtto>2): Likewise.
7850 (fix_trunc<mode><V_cvtto>)2): Likewise
7851 (fixuns_trunc<mode><V_cvtto)2): Likewise.
7852 (neon_vcvt<mode>): Likewise.
7853 (neon_vcvtv4sfv4hf): Likewise.
7854 (neon_vcvtv4hfv4sf): Likewise.
7855 (neon_vcvt_n<mode>): Likewise.
7856 (neon_vmovn<mode>): Likewise.
7857 (neon_vqmovn<mode>): Likewise.
7858 (neon_vqmovun<mode>): Likewise.
7859 (neon_vmovl<mode>): Likewise.
7860 (neon_vmul_lane<mode>): Likewise.
7861 (neon_vmull_lane<mode>): Likewise.
7862 (neon_vqdmull_lane<mode>): Likewise.
7863 (neon_vqdmulh_lane<mode>): Likewise.
7864 (neon_vmla_lane<mode>): Likewise.
7865 (neon_vmlal_lane<mode>): Likewise.
7866 (neon_vqdmlal_lane<mode>): Likewise.
7867 (neon_vmls_lane<mode>): Likewise.
7868 (neon_vmlsl_lane<mode>): Likewise.
7869 (neon_vqdmlsl_lane<mode>): Likewise.
7870 (neon_vext<mode>): Likewise.
7871 (neon_vrev64<mode>): Likewise.
7872 (neon_vrev32<mode>): Likewise.
7873 (neon_vrev16<mode>): Likewise.
7874 (neon_vbsl<mode>_internal): Likewise.
7875 (neon_vshl<mode>): Likewise.
7876 (neon_vqshl<mode>): Likewise.
7877 (neon_vshr_n<mode>): Likewise.
7878 (neon_vshrn_n<mode>): Likewise.
7879 (neon_vqshrn_n<mode>): Likewise.
7880 (neon_vqshrun_n<mode>): Likewise.
7881 (neon_vshl_n<mode>): Likewise.
7882 (neon_vqshl_n<mode>): Likewise.
7883 (neon_vqshlu_n<mode>): Likewise.
7884 (neon_vshll_n<mode>): Likewise.
7885 (neon_vsra_n<mode>): Likewise.
7886 (neon_vsri_n<mode>): Likewise.
7887 (neon_vsli_n<mode>): Likewise.
7888 (neon_vtbl1v8qi): Likewise.
7889 (neon_vtbl2v8qi): Likewise.
7890 (neon_vtbl3v8qi): Likewise.
7891 (neon_vtbl4v8qi): Likewise.
7892 (neon_vtbx1v8qi): Likewise.
7893 (neon_vtbx2v8qi): Likewise.
7894 (neon_vtbx3v8qi): Likewise.
7895 (neon_vtbx4v8qi): Likewise.
7896 (neon_vtrn<mode>_internal): Likewise.
7897 (neon_vzip<mode>_internal): Likewise.
7898 (neon_vuzp<mode>_internal): Likewise.
7899 (neon_vld1<mode>): Likewise.
7900 (neon_vld1_lane<mode>): Likewise.
7901 (neon_vld1_dup<mode>): Likewise.
7902 (neon_vld1_dupv2di): Likewise.
7903 (neon_vst1<mode>): Likewise.
7904 (neon_vst1_lane<mode>): Likewise.
7905 (neon_vld2<mode>): Likewise.
7906 (neon_vld2_lane<mode>): Likewise.
7907 (neon_vld2_dup<mode>): Likewise.
7908 (neon_vst2<mode>): Likewise.
7909 (neon_vst2_lane<mode>): Likewise.
7910 (neon_vld3<mode>): Likewise.
7911 (neon_vld3qa<mode>): Likewise.
7912 (neon_vld3qb<mode>): Likewise.
7913 (neon_vld3_lane<mode>): Likewise.
7914 (neon_vld3_dup<mode>): Likewise.
7915 (neon_vst3<mode>): Likewise.
7916 (neon_vst3qa<mode>): Likewise.
7917 (neon_vst3qb<mode>): Likewise.
7918 (neon_vst3_lane<mode>): Likewise.
7919 (neon_vld4<mode>): Likewise.
7920 (neon_vld4qa<mode>): Likewise.
7921 (neon_vld4qb<mode>): Likewise.
7922 (neon_vld4_lane<mode>): Likewise.
7923 (neon_vld4_dup<mode>): Likewise.
7924 (neon_vst4<mode>): Likewise.
7925 (neon_vst4qa<mode>): Likewise.
7926 (neon_vst4qb<mode>): Likewise.
7927 (neon_vst4_lane<mode>): Likewise.
7928 (neon_vec_unpack<US>_lo_<mode>): Likewise.
7929 (neon_vec_unpack<US>_hi_<mode>): Likewise.
7930 (neon_vec_<US>mult_lo_<mode>): Likewise.
7931 (neon_vec_<US>mult_hi_<mode>): Likewise.
7932 (neon_vec_<US>shiftl_<mode>): Likewise.
7933 (neon_unpack<US>_<mode>): Likewise.
7934 (neon_vec_<US>mult_<mode>): Likewise.
7935 (vec_pack_trunc_<mode>): Likewise.
7936 (neon_vec_pack_trunk_<mode>): Likewise.
7937 (neon_vabd<mode>_2): Likewise.
7938 (neon_vabd<mode>_3): Likewise.
7939 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
7940 (thumb2_movsi_vfp): Likewise.
7941 (movdi_vfp): Likewise.
7942 (movdi_vfp_cortexa8): Likewise.
7943 (movhf_vfp_neon): Likewise.
7944 (movhf_vfp): Likewiwse.
7945 (movsf_vfp): Likewiwse.
7946 (thumb2_movsf_vfp): Likewiwse.
7947 (movdf_vfp): Likewise.
7948 (thumb2_movdf_vfp): Likewise.
7949 (movsfcc_vfp): Likewise.
7950 (thumb2_movsfcc_vfp): Likewise.
7951 (movdfcc_vfp): Likewise.
7952 (thumb2_movdfcc_vfp): Likewise.
7953 * config/arm/arm.c (cortexa7_older_only): Update for attribute change.
7954 * config/arm/arm1020e.md (v10_c2v): Update for attribute change.
7955 (v10_v2c): Likewise.
7956 * config/arm/cortex-a15-neon.md (cortex_a15_neon_int_1): Update for
7958 (cortex_a15_neon_int_2): Likewise.
7959 (cortex_a15_neon_int_3): Likewise.
7960 (cortex_a15_neon_int_4): Likewise.
7961 (cortex_a15_neon_int_5): Likewise.
7962 (cortex_a15_neon_vqneg_vqabs): Likewise.
7963 (cortex_a15_neon_vmov): Likewise.
7964 (cortex_a15_neon_vaba): Likewise.
7965 (cortex_a15_neon_vaba_qqq): Likewise.
7966 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7967 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
7968 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_\
7969 scalar_64_32_long_scalar): Likewise.
7970 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7971 (cortex_a15_neon_mla_qqq_8_16): Likewise.
7972 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
7973 lotype_qdd_64_32_long): Likewise.
7974 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
7975 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
7976 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
7977 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
7978 (cortex_a15_neon_shift_1): Likewise.
7979 (cortex_a15_neon_shift_2): Likewise.
7980 (cortex_a15_neon_shift_3): Likewise.
7981 (cortex_a15_neon_vshl_ddd): Likewise.
7982 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
7983 (cortex_a15_neon_vsra_vrsra): Likewise.
7984 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
7985 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
7986 (cortex_a15_neon_fp_vmul_ddd): Likewise.
7987 (cortex_a15_neon_fp_vmul_qqd): Likewise.
7988 (cortex_a15_neon_fp_vmla_ddd): Likewise.
7989 (cortex_a15_neon_fp_vmla_qqq): Likewise.
7990 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
7991 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
7992 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
7993 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
7994 (cortex_a15_neon_bp_simple): Likewise.
7995 (cortex_a15_neon_bp_2cycle): Likewise.
7996 (cortex_a15_neon_bp_3cycle): Likewise.
7997 (cortex_a15_neon_vld1_1_2_regs): Likewise.
7998 (cortex_a15_neon_vld1_3_4_regs): Likewise.
7999 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
8000 (cortex_a15_neon_vld2_4_regs): Likewise.
8001 (cortex_a15_neon_vld3_vld4): Likewise.
8002 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
8003 (cortex_a15_neon_vst1_3_4_regs): Likewise.
8004 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Likewise.
8005 (cortex_a15_neon_vst3_vst4): Likewise.
8006 (cortex_a15_neon_vld1_vld2_lane): Likewise.
8007 (cortex_a15_neon_vld3_vld4_lane" 10
8008 (cortex_a15_neon_vst1_vst2_lane): Likewise.
8009 (cortex_a15_neon_vst3_vst4_lane): Likewise.
8010 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
8011 (cortex_a15_neon_ldm_2): Likewise.0
8012 (cortex_a15_neon_stm_2): Likewise.
8013 (cortex_a15_neon_mcr): Likewise.
8014 (cortex_a15_neon_mcr_2_mcrr): Likewise.
8015 (cortex_a15_neon_mrc): Likewise.
8016 (cortex_a15_neon_mrrc): Likewise.
8017 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
8019 (cortex_a15_alu_shift): Likewise.
8020 (cortex_a15_alu_shift_reg): Likewise.
8021 (cortex_a15_mult32): Likewise.
8022 (cortex_a15_mult64): Likewise.
8023 (cortex_a15_block): Likewise.
8024 (cortex_a15_branch): Likewise.
8025 (cortex_a15_load1): Likewise.
8026 (cortex_a15_load3): Likewise.
8027 (cortex_a15_store1): Likewise.
8028 (cortex_a15_store3): Likewise.
8029 (cortex_a15_call): Likewise.
8030 * config/arm/cortex-a5.md (cortex_a5_r2f): Update for attribute change.
8031 (cortex_a5_f2r): Likewise.
8032 * config/arm/cortex-a53.md (cortex_a53_r2f): Update for attribute
8034 (cortex_a53_f2r): Likewise.
8035 * config/arm/cortex-a7.md
8036 (cortex_a7_branch): Update for attribute change.
8037 (cortex_a7_call): Likewise.
8038 (cortex_a7_alu_imm): Likewise.
8039 (cortex_a7_alu_reg): Likewise.
8040 (cortex_a7_alu_shift): Likewise.
8041 (cortex_a7_mul): Likewise.
8042 (cortex_a7_load1): Likewise.
8043 (cortex_a7_store1): Likewise.
8044 (cortex_a7_load2): Likewise.
8045 (cortex_a7_store2): Likewise.
8046 (cortex_a7_load3): Likewise.
8047 (cortex_a7_store3): Likewise.
8048 (cortex_a7_load4): Likewise.
8049 (cortex_a7_store4): Likewise.
8050 (cortex_a7_fpalu): Likewise.
8051 (cortex_a7_fconst): Likewise.
8052 (cortex_a7_fpmuls): Likewise.
8053 (cortex_a7_neon_mul): Likewise.
8054 (cortex_a7_fpmacs): Likewise.
8055 (cortex_a7_neon_mla: Likewise.
8056 (cortex_a7_fpmuld: Likewise.
8057 (cortex_a7_fpmacd: Likewise.
8058 (cortex_a7_fpfmad: Likewise.
8059 (cortex_a7_fdivs: Likewise.
8060 (cortex_a7_fdivd: Likewise.
8061 (cortex_a7_r2f: Likewise.
8062 (cortex_a7_f2r: Likewise.
8063 (cortex_a7_f_flags: Likewise.
8064 (cortex_a7_f_loads: Likewise.
8065 (cortex_a7_f_loadd: Likewise.
8066 (cortex_a7_f_stores: Likewise.
8067 (cortex_a7_f_stored: Likewise.
8068 (cortex_a7_neon): Likewise.
8069 * config/arm/cortex-a8-neon.md
8070 (cortex_a8_neon_mrc): Update for attribute change.
8071 (cortex_a8_neon_mrrc): Likewise.
8072 (cortex_a8_neon_int_1): Likewise.
8073 (cortex_a8_neon_int_2): Likewise.
8074 (cortex_a8_neon_int_3): Likewise.
8075 (cortex_a8_neon_int_4): Likewise.
8076 (cortex_a8_neon_int_5): Likewise.
8077 (cortex_a8_neon_vqneg_vqabs): Likewise.
8078 (cortex_a8_neon_vmov): Likewise.
8079 (cortex_a8_neon_vaba): Likewise.
8080 (cortex_a8_neon_vaba_qqq): Likewise.
8081 (cortex_a8_neon_vsma): Likewise.
8082 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
8083 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
8084 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
8085 long_scalar): Likewise.
8086 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
8087 (cortex_a8_neon_mla_qqq_8_16): Likewise.
8088 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
8089 long_scalar_qdd_64_32_long): Likewise.
8090 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
8091 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
8092 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
8093 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
8094 (cortex_a8_neon_shift_1): Likewise.
8095 (cortex_a8_neon_shift_2): Likewise.
8096 (cortex_a8_neon_shift_3): Likewise.
8097 (cortex_a8_neon_vshl_ddd): Likewise.
8098 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
8099 (cortex_a8_neon_vsra_vrsra): Likewise.
8100 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
8101 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
8102 (cortex_a8_neon_fp_vsum): Likewise.
8103 (cortex_a8_neon_fp_vmul_ddd): Likewise.
8104 (cortex_a8_neon_fp_vmul_qqd): Likewise.
8105 (cortex_a8_neon_fp_vmla_ddd): Likewise.
8106 (cortex_a8_neon_fp_vmla_qqq): Likewise.
8107 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
8108 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
8109 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
8110 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
8111 (cortex_a8_neon_bp_simple): Likewise.
8112 (cortex_a8_neon_bp_2cycle): Likewise.
8113 (cortex_a8_neon_bp_3cycle): Likewise.
8114 (cortex_a8_neon_ldr): Likewise.
8115 (cortex_a8_neon_str): Likewise.
8116 (cortex_a8_neon_vld1_1_2_regs): Likewise.
8117 (cortex_a8_neon_vld1_3_4_regs): Likewise.
8118 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
8119 (cortex_a8_neon_vld2_4_regs): Likewise.
8120 (cortex_a8_neon_vld3_vld4): Likewise.
8121 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
8122 (cortex_a8_neon_vst1_3_4_regs): Likewise.
8123 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
8124 (cortex_a8_neon_vst3_vst4): Likewise.
8125 (cortex_a8_neon_vld1_vld2_lane): Likewise.
8126 (cortex_a8_neon_vld3_vld4_lane): Likewise.
8127 (cortex_a8_neon_vst1_vst2_lane): Likewise.
8128 (cortex_a8_neon_vst3_vst4_lane): Likewise.
8129 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
8130 (cortex_a8_neon_mcr): Likewise.
8131 (cortex_a8_neon_mcr_2_mcrr): Likewise.
8132 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
8133 * config/arm/cortex-a9-neon.md (ca9_neon_mrc): Update for attribute
8135 (ca9_neon_mrrc): Likewise.
8136 (cortex_a9_neon_int_1): Likewise.
8137 (cortex_a9_neon_int_2): Likewise.
8138 (cortex_a9_neon_int_3): Likewise.
8139 (cortex_a9_neon_int_4): Likewise.
8140 (cortex_a9_neon_int_5): Likewise.
8141 (cortex_a9_neon_vqneg_vqabs): Likewise.
8142 (cortex_a9_neon_vmov): Likewise.
8143 (cortex_a9_neon_vaba): Likewise.
8144 (cortex_a9_neon_vaba_qqq): Likewise.
8145 (cortex_a9_neon_vsma): Likewise.
8146 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
8147 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
8148 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
8149 long_scalar): Likewise.
8150 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
8151 (cortex_a9_neon_mla_qqq_8_16): Likewise.
8152 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
8153 long_scalar_qdd_64_32_long): Likewise.
8154 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
8155 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
8156 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
8157 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
8158 (cortex_a9_neon_shift_1): Likewise.
8159 (cortex_a9_neon_shift_2): Likewise.
8160 (cortex_a9_neon_shift_3): Likewise.
8161 (cortex_a9_neon_vshl_ddd): Likewise.
8162 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
8163 (cortex_a9_neon_vsra_vrsra): Likewise.
8164 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
8165 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
8166 (cortex_a9_neon_fp_vsum): Likewise.
8167 (cortex_a9_neon_fp_vmul_ddd): Likewise.
8168 (cortex_a9_neon_fp_vmul_qqd): Likewise.
8169 (cortex_a9_neon_fp_vmla_ddd): Likewise.
8170 (cortex_a9_neon_fp_vmla_qqq): Likewise.
8171 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
8172 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
8173 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
8174 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
8175 (cortex_a9_neon_bp_simple): Likewise.
8176 (cortex_a9_neon_bp_2cycle): Likewise.
8177 (cortex_a9_neon_bp_3cycle): Likewise.
8178 (cortex_a9_neon_ldr): Likewise.
8179 (cortex_a9_neon_str): Likewise.
8180 (cortex_a9_neon_vld1_1_2_regs): Likewise.
8181 (cortex_a9_neon_vld1_3_4_regs): Likewise.
8182 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
8183 (cortex_a9_neon_vld2_4_regs): Likewise.
8184 (cortex_a9_neon_vld3_vld4): Likewise.
8185 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
8186 (cortex_a9_neon_vst1_3_4_regs): Likewise.
8187 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
8188 (cortex_a9_neon_vst3_vst4): Likewise.
8189 (cortex_a9_neon_vld1_vld2_lane): Likewise.
8190 (cortex_a9_neon_vld3_vld4_lane): Likewise.
8191 (cortex_a9_neon_vst1_vst2_lane): Likewise.
8192 (cortex_a9_neon_vst3_vst4_lane): Likewise.
8193 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
8194 (cortex_a9_neon_mcr): Likewise.
8195 (cortex_a9_neon_mcr_2_mcrr): Likewise.
8196 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
8197 (cortex_a9_fps): Likewise.
8198 * config/arm/cortex-m4-fpu.md (cortex_m4_vmov_2): Update for attribute
8200 (cortex_m4_fmuls): Likewise.
8201 * config/arm/cortex-r4f.md (cortex_r4_mcr): Update for attribute
8203 (cortex_r4_mrc): Likewise.
8204 * config/arm/iterators.md: Update comment referring to neon_type.
8205 * config/arm/iwmmxt.md (iwmmxt_arm_movdi): Update for attribute change.
8206 (iwmmxt_movsi_insn): Likewise.
8207 * config/arm/marvell-pj4.md (pj4_vfp_to_core): Update for
8209 (pj4_core_to_vfp): Likewise.
8210 * config/arm/neon-schedgen.ml (emit_insn_reservations): Update for
8212 * config/arm/vfp11.md (vfp_fload): Update for attribute change.
8213 (vfp_fstore): Likewise.
8214 * doc/md.texi: Change references to neon_type to refer to type.
8216 2013-09-04 Dodji Seketeli <dodji@redhat.com>
8218 * tree.h (DECL_BUILT_IN): Fix typo in comment.
8220 2013-09-04 David Edelsohn <dje.gcc@gmail.com>
8222 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Only emit
8225 2013-09-04 Easwaran Raman <eraman@google.com>
8228 * tree-ssa-reassoc.c (get_stmt_uid_with_default): New function,
8229 (build_and_add_sum): Use it.
8230 (appears_later_in_bb): Simplify code.
8232 2013-09-04 Teresa Johnson <tejohnson@google.com>
8234 * dumpfile.c (dump_finish): Don't close stderr/stdout.
8236 2013-09-04 James Greenhalgh <james.greenhalgh@arm.com>
8238 * config/aarch64/arm_neon.h (vaddvq_<su>64): Fix return types.
8240 2013-09-04 Jan Hubicka <jh@suse.cz>
8242 * Makefile.in (ipa-devirt.o): Add dependency on diagnostic.h
8243 * ipa-devirt.c: Include diganostic.h
8244 (odr_type_d): Add types and types_set.
8245 (hash_type_name): Work for types with vtables during LTO.
8246 (odr_hasher::remove): Fix comment; destroy types_set.
8247 (add_type_duplicate): New function,
8248 (get_odr_type): Use it.
8249 (dump_type_inheritance_graph): Dump type duplicates.
8250 * ipa.c (symtab_remove_unreachable_nodes): Build type inheritance
8252 * tree.c (types_same_for_odr): Give exact answers on types with
8255 2013-09-04 Dodji Seketeli <dodji@redhat.com>
8257 * tree.h (DECL_BUILT_IN, DECL_IS_BUILTIN): Add more comments
8258 explaining their differences.
8260 2013-09-04 Sandeep Kumar Singh<Sandeep.Singh2@kpitcummins.com>
8262 * config/rx/rx.h: Add option -mcpu for target variants RX100 and RX200.
8264 2013-09-03 Jeff Law <law@redhat.com>
8266 * tree-ssa-threadedge.c (thread_across_edge): Record entire path
8267 when not threading through a joiner block. Pass joiner/no joiner
8268 state to register_jump_thread.
8269 * tree-ssa-threadupdate.c (register_jump_thread): Get joiner/no joiner
8270 state from argument rather than implying on path length.
8271 Dump the entire jump thread path into debugging dump.
8272 * tree-flow.h (register_jump_thread): Update prototype.
8274 2013-08-29 Xinliang David Li <davidxl@google.com>
8276 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
8277 Remove a trivial gcc_assert.
8279 2013-08-29 Xinliang David Li <davidxl@google.com>
8281 * tree-vect-slp.c (destroy_bb_vec_info): Data ref cleanup.
8282 * tree-vect-loop.c (destroy_bb_vec_info): Ditto.
8283 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
8284 Delay base decl alignment adjustment.
8285 * tree-vectorizer.c (vect_destroy_datarefs): New function.
8286 * tree-vectorizer.h: New data structure.
8287 (set_dr_misalignment): New function.
8288 (dr_misalignment): Ditto.
8289 * tree-vect-stmts.c (vectorizable_store): Ensure alignment.
8290 (vectorizable_load): Ditto.
8291 (ensure_base_align): New function.
8292 (vectorize_loops): Add dbg_cnt support.
8293 (execute_vect_slp): Ditto.
8294 * dbgcnt.def: New debug counter.
8295 * Makefile: New dependency.
8297 2013-09-03 Meador Inge <meadori@codesourcery.com>
8301 2013-08-30 Meador Inge <meadori@codesourcery.com>
8303 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
8305 2013-09-03 David Edelsohn <dje.gcc@gmail.com>
8307 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Emit lglobl for
8308 function descriptor.
8310 2013-09-03 Richard Biener <rguenther@suse.de>
8312 * tree-affine.c (add_elt_to_tree): Fix association issue,
8313 avoid useless converts and make sure to always return a
8314 properly typed result.
8316 2013-09-03 Richard Biener <rguenther@suse.de>
8319 * fold-const.c (negate_expr_p): Fix division case.
8320 (negate_expr): Likewise.
8322 2013-09-03 Richard Biener <rguenther@suse.de>
8325 * tree-streamer-out.c: Include tm.h.
8326 * Makefile.in (tree-streamer-out.o): Depend on $(TM_H).
8328 2013-09-03 Jan Hubicka <jh@suse.cz>
8330 * tree-profile.c (tree_profiling): Cleanup CFG when done.
8332 2013-09-03 Alan Modra <amodra@gmail.com>
8334 * config.gcc (powerpc*-*-linux*): Add support for little-endian
8335 multilibs to big-endian target and vice versa.
8336 * config/rs6000/t-linux64: Use := assignment on all vars.
8337 (MULTILIB_EXTRA_OPTS): Remove fPIC.
8338 (MULTILIB_OSDIRNAMES): Specify using mapping from multilib_options.
8339 * config/rs6000/t-linux64le: New file.
8340 * config/rs6000/t-linux64bele: New file.
8341 * config/rs6000/t-linux64lebe: New file.
8343 2013-09-02 Jan Hubicka <jh@suse.cz>
8345 * ipa-inline-transform.c (inline_transform): Do not
8346 optimize_inline_calls when not optimizing.
8348 2013-09-02 Jan Hubicka <jh@suse.cz>
8350 * lto-symtab.c (lto_symtab_merge_symbols): Add comments; merge
8351 duplicated nodes for assembler names.
8352 * symtab.c (symtab_unregister_node): Do not attempt to unlink
8353 hard registers from assembler name hash.
8355 2013-09-02 Jan Hubicka <jh@suse.cz>
8357 * ipa-split.c (execute_split_functions): Split externally visible
8358 functions called once.
8360 2013-09-02 Martin Jambor <mjambor@suse.cz>
8363 * ipa-prop.c (ipa_edge_duplication_hook): Always put new rdesc to the
8364 linked list. When finding the correct duplicate, also consider also
8365 the caller in additon to its inlined_to node.
8367 2013-09-02 James Greenhalgh <james.greenhalgh@arm.com>
8369 * config/aarch64/aarch64-simd-builtins.def
8370 (dup_lane_scalar): Remove.
8371 * config/aarch64/aarch64-simd.md
8372 (aarch64_simd_dup): Add 'w->w' alternative.
8373 (aarch64_dup_lane<mode>): Allow for VALL.
8374 (aarch64_dup_lane_scalar<mode>): Remove.
8375 (aarch64_dup_lane_<vswap_width_name><mode>): New.
8376 (aarch64_get_lane_signed<mode>): Add w->w altenative.
8377 (aarch64_get_lane_unsigned<mode>): Likewise.
8378 (aarch64_get_lane<mode>): Likewise.
8379 * config/aarch64/aarch64.c (aarch64_evpc_dup): New.
8380 (aarch64_expand_vec_perm_const_1): Use aarch64_evpc_dup.
8381 * config/aarch64/iterators.md (VSWAP_WIDTH): New.
8382 (VCON): Change container of V2SF.
8383 (vswap_width_name): Likewise.
8384 * config/aarch64/arm_neon.h
8385 (__aarch64_vdup_lane_any): New.
8386 (__aarch64_vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
8387 (vdup<q>_n_<psuf><8,16,32,64>): Convert to C implementation.
8388 (vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
8390 2013-09-02 Eric Botcazou <ebotcazou@adacore.com>
8393 * expr.c (emit_move_complex): Do not move complex FP values as parts if
8394 the source or the destination is a single hard register.
8396 2013-09-02 Richard Biener <rguenther@suse.de>
8399 * tree-scalar-evolution.c (instantiate_scev_name): Allow
8402 2013-09-02 Richard Biener <rguenther@suse.de>
8404 * tree-affine.c (add_elt_to_tree): Avoid converting all pointer
8405 arithmetic to sizetype.
8407 2013-09-02 Bin Cheng <bin.cheng@arm.com>
8409 * tree-ssa-loop-ivopts.c (set_autoinc_for_original_candidates):
8410 Find auto-increment use both before and after candidate.
8412 2013-09-02 Marek Polacek <polacek@redhat.com>
8414 * Makefile.in (ubsan.o): Add $(TM_P_H) dependency.
8416 2013-09-01 Jan Hubicka <jh@suse.cz>
8418 * Makefile.in: Add ipa-profile.o
8419 (ipa.o, ipa-devrit.o, ipa-inline-analysis.o): Adjust dependencies.
8420 * cgraph.c (struct cgraph_propagate_frequency_data,
8421 cgraph_propagate_frequency_1, cgraph_propagate_frequency): Move to
8422 ipa-profile.c; replace cgraph_ by ipa_ prefix.
8423 * cgraph.h (cgraph_propagate_frequency): Remove.
8424 * ipa-inline-analysis.c: Include ipa-utils.h;
8425 drop duplicated cfgloop.h.
8426 (inline_update_callee_summaries): Update.
8427 * ipa-profile.c: New file.
8428 * ipa-utils.h (ipa_propagate_frequency): Declare.
8429 * ipa.c: Do not include pointer-set.h, hash-table.h, lto-streamer.h,
8430 data-streamer.h, value-prof.h.
8431 (symtab_remove_unreachable_nodes): Update profile.
8432 (struct histogram_entry, histogram, histogram_pool, histogram_hash,
8433 account_time_size, cmp_counts, dump_histogram,
8434 ipa_profile_generate_summary, ipa_profile_write_summary,
8435 ipa_profile_read_summary, ipa_profile, gate_ipa_profile,
8436 pass_data_ipa_profile, pass_ipa_profile, make_pass_ipa_profile):
8437 Move to ipa-profile.c.
8439 2013-09-01 John David Anglin <danglin@gcc.gnu.org>
8441 * config/pa/pa.md: Allow "const 0" operand 1 in "scc" insns.
8443 2013-09-01 Jan Hubicka <jh@suse.cz>
8445 * common.opt (fdevirtualize-speculatively): New function.
8446 * invoke.texi (fdevirtualize-speculatively): Document.
8447 * ipa-devirt.c: Include ipa-inline.h
8448 (likely_target_p): New function.
8449 (ipa_devirt): New function.
8450 (gate_ipa_devirt): New function.
8451 (pass_data_ipa_devirt): New static var.
8452 (pass_ipa_devirt): Likewise.
8453 (make_pass_ipa_devirt): New function.
8454 * opts.c (default_options): Add OPT_fdevirtualize_speculatively.
8455 (common_handle_option): Disable devirtualization when
8456 value range profiling is available.
8457 * passes.def (pass_ipa_devirt): Add.
8458 * timever.def (TV_IPA_DEVIRT): New timevar.
8459 * tree-pass.h (make_pass_ipa_devirt):
8461 2013-09-01 Iain Sandoe <iain@codesourcery.com>
8463 * config/darwin.h (LINK_COMMAND_SPEC_A): Revise sanitizer specs to
8464 include sanitize(undefined).
8466 2013-08-31 Diego Novillo <dnovillo@google.com>
8468 * Makefile.in (TREE_CORE_H): Define.
8470 (GTFILES): Add tree-core.h.
8471 * builtins.c (built_in_class_names): Use BUILT_IN_LAST to
8473 * tree-core.h: New file.
8474 Move all data structures, enum, typedefs, global
8475 declarations and constants from ...
8478 2013-08-31 Jan Hubicka <jh@suse.cz>
8480 * bulitins.c (expand_builtin): Do not early exit for gcov
8481 instrumented functions.
8483 2013-08-31 Marek Polacek <polacek@redhat.com>
8485 * ubsan.c: Include tm_p.h.
8487 2013-08-31 Jan Hubicka <jh@suse.cz>
8489 * gimple-streamer-in.c (input_gimple_stmt): Silence parameter unused
8492 * cgraph.c (cgraph_get_body): Update call of lto_input_function_body.
8493 * gimple-streamer-in.c (input_gimple_stmt): Move sanity check to ...
8494 * tree-cfg.c (verify_gimple_label): ... here.
8495 * ipa-utils.c: Include lto-streamer.h, ipa-inline.h
8496 (ipa_merge_profiles): New function.
8497 * lto-streamer-in.c (lto_read_body): Take node instead of fn_decl.
8498 (lto_input_function_body): Likewise.
8499 * ipa-utils.h (ipa_merge_profiles): Declare.
8500 * lto-streamer.h (lto_input_function_body): Update prototype.
8501 (emit_label_in_global_context_p): Remove.
8502 * lto-symtab.c: Include ipa-utils.h
8503 (lto_cgraph_replace_node): Use ipa_merge_profiles.
8505 2013-08-31 Jan Hubicka <jh@suse.cz>
8507 * cgraph.c (cgraph_speculative_call_info): Fix ref lookup
8509 2013-08-31 Jan Hubicka <jh@suse.cz>
8511 * basic-block.h (apply_scale): Make scale parmeter gcov_type.
8513 2013-08-31 Uros Bizjak <ubizjak@gmail.com>
8515 * config/alpha/alpha.c (alpha_emit_conditional_move): Update
8516 "cmp" RTX before signed_comparison_operator check to account
8519 2013-08-30 Jan Hubicka <jh@suse.cz>
8521 * ipa-prop.c (ipa_set_jf_known_type): Check that we add only records.
8522 (detect_type_change_1): Rename to ...
8523 (detect_type_change): ... this one; early return on non-polymorphic
8525 (detect_type_change_ssa): Add comp_type parameter; update
8526 use of detect_type_change.
8527 (compute_complex_assign_jump_func): Add param_type parameter;
8528 update use of detect_type_change_ssa.
8529 (compute_complex_ancestor_jump_func): Likewise.
8530 (ipa_get_callee_param_type): New function.
8531 (ipa_compute_jump_functions_for_edge): Compute parameter type;
8532 update calls to the jump function computation functions.
8534 2013-08-30 Teresa Johnson <tejohnson@google.com>
8535 Steven Bosscher <steven@gcc.gnu.org>
8537 * cfgrtl.c (fixup_new_cold_bb): New routine.
8538 (commit_edge_insertions): Invoke fixup_partitions.
8539 (find_partition_fixes): New routine.
8540 (fixup_partitions): Ditto.
8541 (verify_hot_cold_block_grouping): Update comments.
8542 (rtl_verify_edges): Invoke find_partition_fixes.
8543 (rtl_verify_bb_pointers): Update comments.
8544 (rtl_verify_bb_layout): Ditto.
8545 * basic-block.h (probably_never_executed_edge_p): Declare.
8546 (fixup_partitions): Ditto.
8547 * cfgcleanup.c (try_optimize_cfg): Invoke fixup_partitions.
8548 * bb-reorder.c (sanitize_hot_paths): New function.
8549 (find_rarely_executed_basic_blocks_and_crossing_edges): Invoke
8551 * predict.c (probably_never_executed_edge_p): New routine.
8552 * cfg.c (check_bb_profile): Add partition insanity warnings.
8554 2013-08-30 Meador Inge <meadori@codesourcery.com>
8556 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
8558 2013-08-30 Marek Polacek <polacek@redhat.com>
8560 * Makefile.in (ubsan.o): Add.
8561 (c-family/c-ubsan.o): Add.
8562 (builtins.o): Add ubsan.h dependency.
8563 * ubsan.h: New file.
8564 * ubsan.c: New file.
8565 * common.opt: Add -fsanitize=undefined option.
8566 (flag_sanitize): Add variable.
8567 (fsanitize=): Add option. Add Driver.
8568 (fsanitize=thread): Remove option.
8569 (fsanitize=address): Likewise.
8570 (static-libubsan): New option.
8571 * doc/invoke.texi: Document the new flag and -static-libubsan.
8572 * sanitizer.def (DEF_SANITIZER_BUILTIN): Define.
8573 (BUILT_IN_UBSAN_HANDLE_BUILTIN_UNREACHABLE): Define.
8574 * builtin-attrs.def (ATTR_COLD): Define.
8575 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
8576 * builtins.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW,
8577 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS): Define.
8578 * flag-types.h (sanitize_code): New enum.
8579 * opts.c (common_handle_option): Parse command line arguments
8580 of -fsanitize=. Add -fsanitize=unreachable option.
8581 * varasm.c (get_variable_section): Adjust.
8582 (assemble_noswitch_variable): Likewise.
8583 (assemble_variable): Likewise.
8584 (output_constant_def_contents): Likewise.
8585 (categorize_decl_for_section): Likewise.
8586 (place_block_symbol): Likewise.
8587 (output_object_block): Likewise.
8588 * builtins.def: Likewise.
8589 * toplev.c (compile_file): Likewise.
8590 (process_options): Likewise.
8591 * cppbuiltin.c: Likewise.
8592 * tsan.c (tsan_pass): Likewise.
8593 (tsan_gate): Likewise.
8594 (tsan_gate_O0): Likewise.
8595 * cfgexpand.c (partition_stack_vars): Likewise.
8596 (expand_stack_vars): Likewise.
8597 (defer_stack_allocation): Likewise.
8598 (expand_used_vars): Likewise.
8599 * cfgcleanup.c (old_insns_match_p): Likewise.
8600 * asan.c (asan_finish_file): Likewise.
8601 (asan_instrument): Likewise.
8602 (gate_asan): Likewise.
8603 (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR_PTR.
8604 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
8605 (asan_global_struct): Use pointer_sized_int_node instead
8606 calling build_nonstandard_integer_type.
8607 (initialize_sanitizer_builtins): Likewise.
8608 (asan_finish_file): Likewise.
8609 * gcc.c: Document %{%:function(args):X}.
8610 (static_spec_functions): Add sanitize.
8611 (handle_spec_function): Add retval_nonnull argument and if non-NULL,
8612 store funcval != NULL there.
8613 (do_spec_1): Adjust handle_spec_function caller.
8614 (handle_braces): Allow %:function(args) as condition.
8615 (sanitize_spec_function): New function.
8616 (ADD_STATIC_LIBUBSAN_LIBS): Define.
8617 (LIBUBSAN_SPEC): Likewise.
8618 (LIBUBSAN_EARLY_SPEC): Likewise.
8619 (SANITIZER_SPEC): Handle libubsan.
8620 (SANITIZER_EARLY_SPEC): Likewise.
8621 * config/darwin.h (LINK_COMMAND_SPEC_A): Use %:sanitize(address)
8622 instead of fsanitize=address.
8623 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Use %:sanitize(address)
8624 instead of fsanitize=address*.
8625 * builtins.c: Include ubsan.h.
8626 (fold_builtin_0): Instrument __builtin_unreachable.
8627 * config/rs6000/rs6000.h (FRAME_GROWS_DOWNWARD): Use flag_sanitize
8628 instead of flag_asan.
8629 * tree.h (enum tree_index): Add TI_POINTER_SIZED_TYPE.
8630 (pointer_sized_int_node): Define.
8631 * tree.c (build_common_tree_nodes): Initialize pointer_sized_int_node.
8633 2013-08-30 Mike Stump <mikestump@comcast.net>
8635 * doc/install.texi (Prerequisites): Note regression in Tcl 8.6
8638 2013-08-29 Jan Hubicka <jh@suse.cz>
8640 * cgraph.c (cgraph_function_body_availability): Handle weakref
8642 * passes.def: Remove pass_fixup_cfg.
8643 * ipa-inline.c (ipa_inline): When not optimizing, do not inline;
8644 track when we need to remove functions.
8645 (gate_ipa_inline): Execute inlining always; add comment why.
8646 (pass_data_ipa_inline): Remove TODO_remove_functions.
8647 * ipa-inline-analysis.c (inline_generate_summary): When not optimizing
8648 do not produce summaries.
8649 * symtab.c (change_decl_assembler_name): Handle renaming of weakrefs.
8650 (symtab_nonoverwritable_alias): Assert we are not called on weakref.
8651 * varpool.c (cgraph_variable_initializer_availability): Fix weakrefs,
8652 constant pool and vtable.
8654 2013-08-30 Tejas Belagod <tejas.belagod@arm.com>
8656 * config/aarch64/arm_neon.h (__AARCH64_UINT64_C, __AARCH64_INT64_C):
8657 New arm_neon.h's internal macros to specify 64-bit constants.
8658 Avoid using stdint.h's macros.
8660 2013-08-30 Joern Rennecke <joern.rennecke@embecosm.com>
8662 * recog.c (verify_changes): Verify that changes[i].old is non-zero
8663 before applying REG_P.
8665 2013-08-30 Jakub Jelinek <jakub@redhat.com>
8667 PR tree-optimization/58277
8668 * tree-ssa-strlen.c (strlen_enter_block): If do_invalidate gave up
8669 after seeing too many stmts with vdef in between dombb and current
8670 bb, invalidate everything.
8672 2013-08-30 Richard Biener <rguenther@suse.de>
8674 * fold-const.c (fold_single_bit_test): Fix overflow test.
8676 2013-08-30 Eric Botcazou <ebotcazou@adacore.com>
8678 * function.c (assign_parm_setup_reg): For a parameter passed by pointer
8679 and which can live in a register, always retrieve the value on entry.
8680 * var-tracking.c (add_stores): Treat the copy on entry for a parameter
8681 passed by invisible reference specially.
8682 (emit_notes_in_bb) <MO_VAL_USE>: Emit notes before the instruction.
8683 (vt_add_function_parameter): Correctly deal with a parameter passed by
8684 invisible reference.
8686 2013-08-30 Jan Hubicka <jh@suse.cz>
8688 * tree.c (set_call_expr_flags): Fix handling of TM_PURE.
8690 2013-08-30 Richard Biener <rguenther@suse.de>
8692 PR tree-optimization/58228
8693 * tree-vect-data-refs.c (vect_analyze_data_ref_access): Do not
8694 allow invariant loads in nested loop vectorization.
8696 2013-08-30 Richard Biener <rguenther@suse.de>
8698 PR tree-optimization/58223
8699 * tree-loop-distribution.c (has_anti_dependence): Rename to ...
8700 (has_anti_or_output_dependence): ... this and adjust to also
8701 look for output dependences.
8702 (mark_nodes_having_upstream_mem_writes): Adjust.
8703 (rdg_flag_uses): Likewise.
8705 2013-08-30 Richard Biener <rguenther@suse.de>
8707 PR tree-optimization/58010
8708 * tree-vect-loop.c (vect_create_epilog_for_reduction): Remove
8709 assert that we have a loop-closed PHI.
8711 2013-08-29 Jan Hubicka <jh@suse.cz>
8713 * lto-symtab.c (lto_cgraph_replace_node): Free decl_in_state.
8714 * cgraph.c (cgraph_release_function_body): Free decl_in_state.
8715 * lto-section-in.c (lto_free_function_in_decl_state): New function.
8716 (lto_free_function_in_decl_state_for_node): New function.
8718 2013-08-29 Xinliang David Li <davidxl@google.com>
8720 * loop-unroll.c (report_unroll_peel): Minor message change.
8721 * tree-vect-loop-manip.c (vect_do_peeling_for_alignment):
8722 Emit alignment peeling message with default -fopt-info.
8723 (vect_loop_versioning): Emit loop version info message.
8724 * tree-vectorizer.c (vectorize_loops): Minor message change.
8725 (execute_vect_slp): Ditto.
8727 2013-08-29 Eric Botcazou <ebotcazou@adacore.com>
8729 * cgraphclones.c (cgraph_create_virtual_clone): Compute the DECL_NAME
8730 of the clone from the DECL_NAME of the original function.
8732 2013-08-29 Oleg Endo <olegendo@gcc.gnu.org>
8734 * passes.c (register_pass): Add overload.
8735 * tree-pass.h (register_pass): Forward declare it. Add comment.
8737 2013-08-29 Jan Hubicka <jh@suse.cz>
8739 * lto-streamer-out.c (hash_tree): Stream DECL_FINAL_P,
8740 DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P and TYPE_FINAL_P.
8741 * lto-streamer-in.c (unpack_ts_decl_with_vis_value_fields): Stream
8742 DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
8743 (unpack_ts_type_common_value_fields): Stream TYPE_FINAL_P.
8744 * tree-streamer-out.c (pack_ts_decl_with_vis_value_fields):
8745 Add DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
8746 (pack_ts_type_common_value_fields): Add TYPE_FINAL_P.
8748 2013-08-29 Teresa Johnson <tejohnson@google.com>
8750 * dumpfile.c (dump_loc): Output column number.
8751 * dumpfile.h (OPTGROUP_OTHER): Add and enable under OPTGROUP_ALL.
8752 * doc/invoke.texi: Document optall -fopt-info flag.
8753 * profile.c (read_profile_edge_counts): Use new dump framework.
8754 (compute_branch_probabilities): Ditto.
8755 * passes.c (pass_manager::register_one_dump_file): Use OPTGROUP_OTHER
8756 when pass not in any opt group.
8757 * pass_manager.h (pass_manager::get_pass_profile): New method.
8758 * value-prof.c (check_counter): Use new dump framework.
8759 (check_ic_target): Ditto.
8760 * coverage.c (get_coverage_counts): Ditto.
8761 (coverage_init): Setup new dump framework.
8763 2013-08-29 Richard Biener <rguenther@suse.de>
8765 PR tree-optimization/58246
8766 * tree-ssa-dce.c (mark_aliased_reaching_defs_necessary_1): Properly
8767 handle the dominance check inside a basic-block.
8769 2013-08-29 Richard Biener <rguenther@suse.de>
8772 * tree-ssa-copy.c (may_propagate_copy): Allow propagating
8773 of default defs that appear in abnormal PHI nodes.
8775 2013-08-29 Richard Biener <rguenther@suse.de>
8777 PR tree-optimization/57685
8778 * tree-vrp.c (register_edge_assert_for_1): Recurse only for
8779 single-use operands to avoid exponential complexity.
8781 2013-08-28 Dehao Chen <dehao@google.com>
8783 * ipa-inline.c (edge_badness): Fix integer underflow.
8785 2013-08-28 Uros Bizjak <ubizjak@gmail.com>
8787 * gtm-builtins.def (_ITM_free): Declare leaf.
8789 2013-08-28 Jakub Jelinek <jakub@redhat.com>
8792 * config/i386/i386.md (*tls_global_dynamic_64_largepic): New insn.
8793 (*tls_local_dynamic_base_64_largepic): Likewise.
8794 (tls_global_dynamic_64_<mode>, tls_local_dynamic_base_64_<mode>):
8795 Remove predicate from call operand.
8796 * config/i386/i386.c (ix86_tls_get_addr): For -mcmodel=large -fpic
8797 return sum of pic_offset_table_rtx and UNSPEC_PLTOFF of the symbol.
8799 2013-08-28 Jeff Law <law@redhat.com>
8801 * tree-ssa-threadedge.c (thread_around_empty_block): Remove
8802 checks for the number of predecessors and successors allowed.
8803 * tree-ssa-threadupdate.c (mark_threaded_blocks): Ignore requests
8804 which require copying a joiner block if there is a request which
8805 is a subpath that requires no joiner block copying.
8807 2013-08-28 Jan Hubicka <jh@suse.cz>
8809 * lto-streamer-out.c (DFS_write_tree_body): Drop
8810 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX and BINFO_VPTR_INDEX.
8811 (hash_tree): Do not hash DECL_DEFER_OUTPUT, BINFO_INHERITANCE_CHAIN,
8812 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX, DECL_IN_TEXT_SECTION.
8813 * tree-streamer-in.c (unpack_ts_decl_common_value_fields):
8814 Do not read DECL_ERROR_ISSUED.
8815 (unpack_ts_decl_with_vis_value_fields): Do not read
8817 (lto_input_ts_binfo_tree_pointers): Do not read
8818 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX
8819 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Do not
8820 write DECL_ERROR_ISSUED..
8821 (pack_ts_decl_with_vis_value_fields): Do not write
8823 (write_ts_binfo_tree_pointers): Do not read BINFO_INHERITANCE_CHAIN,
8824 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX.
8825 * print-tree.c (print_node): Do not print DECL_ERROR_ISSUED.
8826 * tree.h (tree_decl_common): Update comment.
8827 (DECL_ERROR_ISSUED): Remove.
8829 2013-08-28 Jakub Jelinek <jakub@redhat.com>
8832 * omp-low.c (copy_var_decl): Copy over TREE_NO_WARNING flag.
8834 2013-08-28 Jan Hubicka <jh@suse.cz>
8836 * builtins.def (free): Declare leaf.
8838 2013-08-27 David Malcolm <dmalcolm@redhat.com>
8841 * configure.ac (gdbinit.in): Add import of gcc/gdbhooks.py.
8842 * configure: Regenerate.
8844 2013-08-27 Martin Jambor <mjambor@suse.cz>
8846 * ipa-prop.h (ipa_pass_through_data): New field type_preserved.
8847 (ipa_ancestor_jf_data): Likewise.
8848 (ipa_get_jf_pass_through_agg_preserved): Fix comment typo.
8849 (ipa_get_jf_pass_through_type_preserved): New function.
8850 (ipa_get_jf_ancestor_agg_preserved): Fix comment typo.
8851 (ipa_get_jf_ancestor_type_preserved): New function.
8852 * ipa-cp.c (ipa_get_jf_pass_through_result): Honor type_preserved flag.
8853 (ipa_get_jf_ancestor_result): Likewise.
8854 (propagate_vals_accross_pass_through): Use
8855 ipa_get_jf_pass_through_result to do all the value mappings.
8856 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Dump the
8857 type_preserved flag.
8858 (ipa_set_jf_cst_copy): New function.
8859 (ipa_set_jf_simple_pass_through): Set the type_preserved flag.
8860 (ipa_set_jf_arith_pass_through): Likewise.
8861 (ipa_set_ancestor_jf): Likewise.
8862 (compute_complex_assign_jump_func): Set type_preserved instead of
8864 (ipa_compute_jump_functions_for_edge): Likewise.
8865 (combine_known_type_and_ancestor_jfs): Honor type_preserved.
8866 (update_jump_functions_after_inlining): Update type_preserved.
8867 Explicitely create jump functions when combining one with pass_through.
8868 (ipa_write_jump_function): Stream the type_preserved flags.
8869 (ipa_read_jump_function): Likewise.
8871 2013-08-27 Jakub Jelinek <jakub@redhat.com>
8872 Aldy Hernandez <aldyh@redhat.com>
8874 * Makefile.in (omp-low.o): Depend on $(TARGET_H).
8875 * cfgloop.h (struct loop): Add safelen, force_vect, simduid.
8876 * function.h (struct function): Add has_force_vect_loops and
8878 * gimple-pretty-print.c (dump_gimple_omp_for): Handle GF_OMP_FOR_KIND*.
8879 * gimple.c (gimple_build_omp_critical): Add KIND argument and
8881 * gimple.def: Update CLAUSES comments.
8882 * gimple.h (enum gf_mask): Add GF_OMP_FOR_KIND_{FOR,SIMD}.
8883 (gimple_build_omp_for): Add argument to prototype.
8884 (gimple_omp_for_kind): New.
8885 (gimple_omp_for_set_kind): New.
8886 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_LINEAR to
8887 GOVD_DATA_SHARE_CLASS.
8888 (enum omp_region_type): Add ORT_SIMD.
8889 (gimple_add_tmp_var): Handle ORT_SIMD.
8890 (gimplify_var_or_parm_decl): Same.
8891 (is_gimple_stmt): Same.
8892 (omp_firstprivatize_variable): Same.
8893 (omp_add_variable): Only use splay_tree_insert if lookup failed.
8894 (omp_notice_variable): Handle ORT_SIMD.
8895 (omp_is_private): Add SIMD argument and handle it as well as ORT_SIMD.
8896 (omp_check_private): Handle ORT_SIMD.
8897 (gimplify_scan_omp_clauses): Handle OMP_CLAUSE_LINEAR and
8899 (gimplify_adjust_omp_clauses_1): Handle GOVD_LINEAR.
8900 Handle OMP_CLAUSE_LASTPRIVATE.
8901 (gimplify_adjust_omp_clauses): Handle OMP_CLAUSE_LINEAR and
8903 (gimplify_omp_for): Handle OMP_SIMD and OMP_CLAUSE_LINEAR.
8904 (gimplify_expr): Handle OMP_SIMD.
8905 * internal-fn.c (expand_GOMP_SIMD_LANE): New.
8906 (expand_GOMP_SIMD_VF): New.
8907 (expand_GOMP_SIMD_LAST_LANE): New.
8908 * internal-fn.def (GOMP_SIMD_LANE): New.
8909 (GOMP_SIMD_VF): New.
8910 (GOMP_SIMD_LAST_LANE): New.
8911 * omp-low.c: Include target.h.
8912 (extract_omp_for_data): Handle OMP_SIMD, OMP_CLAUSE_LINEAR,
8914 (check_omp_nesting_restrictions): Same.
8916 (lower_rec_simd_input_clauses): New.
8917 (lower_rec_input_clauses): Handle OMP_SIMD, GF_OMP_FOR_KIND_SIMD,
8919 (lower_lastprivate_clauses): Handle OMP_CLAUSE_LINEAR,
8920 GF_OMP_FOR_KIND_SIMD, OMP_SIMD.
8921 (expand_omp_build_assign): New.
8922 (expand_omp_for_init_counts): New.
8923 (expand_omp_for_init_vars): New.
8924 (extract_omp_for_update_vars): New.
8925 (expand_omp_for_generic): Use expand_omp_for_{init,update}_vars
8926 and rewrite accordingly.
8927 (expand_omp_simd): New.
8928 (expand_omp_for): Use expand_omp_simd.
8929 (lower_omp_for_lastprivate): Unshare vinit when appropriate.
8930 (lower_omp_for): Do not lower the body.
8931 * tree-data-ref (get_references_in_stmt): Allow IFN_GOMP_SIMD_LANE
8933 * tree-flow.h (find_omp_clause): Remove prototype.
8934 * tree-if-conv.c (main_tree_if_conversion): Run if doing if conversion,
8935 forcing vectorization of the loop, or if flag_tree_vectorize.
8936 (gate_tree_if_conversion): Similarly.
8937 * tree-inline.c (remap_gimple_stmt): Pass for kind argument to
8938 gimple_build_omp_for.
8939 (copy_cfg_body): set has_force_vect_loops and has_simduid_loops.
8940 * tree-parloops (create_parallel_loop): Pass kind argument to
8941 gimple_build_omp_for.
8942 * tree-pretty-print.c (dump_omp_clause): Add cases for
8943 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_LINEAR, OMP_CLAUSE_SAFELEN,
8944 OMP_CLAUSE__SIMDUID_.
8945 (dump_generic_node): Handle OMP_SIMD.
8946 * tree-ssa-ccp.c (likely_value): Handle IFN_GOMP_SIMD*.
8947 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Do not
8948 unroll OMP_SIMD loops here.
8949 * tree-ssa-loop.c (gate_tree_vectorize): Run if has_force_vect_loops.
8950 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Handle
8952 (vect_analyze_data_refs): Handle simd loops.
8953 * tree-vect-loop.c (vectorizable_live_operation): Handle
8955 * tree-vect-stmts.c (vectorizable_call): Handle IFN_GOMP_SIMD_LANE.
8956 (vectorizable_store): Handle STMT_VINFO_SIMD_LANE_ACCESS_P.
8957 (vectorizable_load): Same.
8958 * tree-vectorizer.c: Include hash-table.h and tree-ssa-propagate.h.
8959 (struct simduid_to_vf): New.
8960 (simduid_to_vf::hash): New.
8961 (simduid_to-vf::equal): New.
8962 (struct simd_array_to_simduid): New.
8963 (simd_array_to_simduid::hash): New.
8964 (simd_array_to_simduid::equal): New.
8965 (adjust_simduid_builtins): New.
8966 (struct note_simd_array_uses_struct): New.
8967 (note_simd_array_uses_cb): New.
8968 (note_simd_array_uses): New.
8969 (vectorize_loops): Handle simd hints and adjust simd builtins
8971 * tree-vectorizer.h (struct _stmt_vec_info): Add
8972 simd_lane_access_p field.
8973 (STMT_VINFO_SIMD_LANE_ACCESS_P): New macro.
8974 * tree.c (omp_clause_num_ops): Add entries for OMP_CLAUSE_LINEAR,
8975 OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_UNIFORM.
8976 (omp_clause_code_name): Same.
8977 (walk_tree_1): Handle OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN,
8978 OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_LINEAR.
8979 * tree.def (OMP_SIMD): New entry.
8980 * tree.h (enum omp_clause_code): Add entries for OMP_CLAUSE_LINEAR,
8981 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_.
8982 (OMP_CLAUSE_DECL): Adjust range for new clauses.
8983 (OMP_CLAUSE_LINEAR_NO_COPYIN): New.
8984 (OMP_CLAUSE_LINEAR_NO_COPYOUT): New.
8985 (OMP_CLAUSE_LINEAR_STEP): New.
8986 (OMP_CLAUSE_SAFELEN_EXPR): New.
8987 (OMP_CLAUSE__SIMDUID__DECL): New.
8988 (find_omp_clause): New prototype.
8990 2013-08-27 H.J. Lu <hongjiu.lu@intel.com>
8992 * config/i386/driver-i386.c (host_detect_local_cpu): Update
8993 Haswell processor detection.
8995 2013-08-27 Christian Widmer <shadow@umbrox.de>
8998 * config/i386/driver-i386.c (host_detect_local_cpu): Add detection
8999 of Ivy Bridge and Haswell processors. Assume core-avx2 for unknown
9000 AVX2 capable processors.
9002 2013-08-27 Tejas Belagod <tejas.belagod@arm.com>
9004 * config/aarch64/arm_neon.h: Replace all inline asm implementations
9005 of vget_low_* with implementations in terms of other intrinsics.
9007 2013-08-27 Marc Glisse <marc.glisse@inria.fr>
9010 * doc/extend.texi (__builtin_isinf_sign): Restrict the return
9011 values to -1, 0 and 1.
9013 2013-08-27 Vidya Praveen <vidyapraveen@arm.com>
9015 * config/aarch64/aarch64.md (unspec): Add UNSPEC_SISD_SSHL,
9016 UNSPEC_SISD_USHL, UNSPEC_USHL_2S, UNSPEC_SSHL_2S, UNSPEC_SISD_NEG.
9017 (<optab><mode>3_insn): Remove.
9018 (aarch64_ashl_sisd_or_int_<mode>3): New Pattern.
9019 (aarch64_lshr_sisd_or_int_<mode>3): Likewise.
9020 (aarch64_ashr_sisd_or_int_<mode>3): Likewise.
9021 (define_split for aarch64_lshr_sisd_or_int_di3): Likewise.
9022 (define_split for aarch64_lshr_sisd_or_int_si3): Likewise.
9023 (define_split for aarch64_ashr_sisd_or_int_di3): Likewise.
9024 (define_split for aarch64_ashr_sisd_or_int_si3): Likewise.
9025 (aarch64_sisd_ushl, aarch64_sisd_sshl): Likewise.
9026 (aarch64_ushl_2s, aarch64_sshl_2s, aarch64_sisd_neg_qi): Likewise.
9027 (ror<mode>3_insn): Likewise.
9028 * config/aarch64/predicates.md (aarch64_simd_register): New.
9030 2013-08-27 Richard Biener <rguenther@suse.de>
9032 PR tree-optimization/57521
9033 * tree-if-conv.c (if_convertible_bb_p): Verify that at least
9034 one edge is non-critical.
9035 (find_phi_replacement_condition): Make sure to use a non-critical
9036 edge. Cleanup and remove old bug workarounds.
9037 (bb_postdominates_preds): Remove.
9038 (if_convertible_loop_p_1): Do not compute post-dominators.
9039 (combine_blocks): Do not free post-dominators.
9040 (main_tree_if_conversion): Likewise.
9041 (pass_data_if_conversion): Add TODO_verify_ssa.
9043 2013-08-27 DJ Delorie <dj@redhat.com>
9045 * config/i386/djgpp.h (ASM_DECLARE_FUNCTION_NAME): New.
9047 2013-08-27 Yufeng Zhang <yufeng.zhang@arm.com>
9049 * function.c (assign_parm_find_data_types): Set passed_mode and
9050 nominal_mode to the TYPE_MODE of nominal_type for the built
9051 pointer type in case of the struct-pass-by-reference.
9053 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
9055 * config/avr/avr-stdint.h (INT16_TYPE): Change default to "int".
9056 (UINT16_TYPE): Change default to "unsigned int".
9058 * config/avr/avr.opt (mfract-convert-truncate): New option.
9059 * config/avr/avr.c (avr_out_fract): Unless TARGET_FRACT_CONV_TRUNC
9060 is set, round negative fractional integers according to n1169
9061 when converting to integer types.
9063 2013-08-26 Jan Hubicka <jh@suse.cz>
9065 * cgraph.c (cgraph_propagate_frequency): Do not assume that virtual
9066 methods can not be called indirectly when their address is not taken.
9068 2013-08-26 Jan Hubicka <jh@suse.cz>
9070 * gimple-fold.c (gimple_get_virt_method_for_binfo): Use
9073 2013-08-26 Jan Hubicka <jh@suse.cz>
9075 * ipa.c (comdat_can_be_unshared_p_1): C++ constructors and destructors
9078 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
9080 * reload.c (find_valid_class): Allow classes that do not include
9081 FIRST_PSEUDO_REGISTER - 1.
9083 2013-08-26 Jan Hubicka <jh@suse.cz>
9085 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix formatting;
9086 fix edge count/frequency when speculation failed; fix type check
9087 for the direct call.
9089 2013-08-26 Jan Hubicka <jh@suse.cz>
9091 * ipa-prop.c (ipa_print_node_params): Do not ICE during WPA.
9093 2013-08-26 Jan Hubicka <jh@suse.cz>
9095 * ipa-inline-transform.c (inline_transform): Be ready for basic block
9096 to be changed by edge redirection.
9098 2013-08-26 Jan Hubicka <jh@suse.cz>
9100 * cgraph.c (cgraph_speculative_call_info): Fix parameter order and
9101 formating; add sanity check.
9102 (cgraph_resolve_speculation): Add FIXME about scaling profiles.
9103 (cgraph_redirect_edge_call_stmt_to_callee): Fix ICE in debug dump.
9104 * ipa-inline.c (heap_edge_removal_hook): Reset node growth cache.
9105 (resolve_noninline_speculation): Update callee keys, too.
9107 2013-08-26 Jan Hubicka <jh@suse.cz>
9109 * tree.h (tree_decl_with_vis): Add cxx_constructor, cxx_destructor.
9110 (DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P): New macros.
9112 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
9114 * config/i386/i386.c (x86_64_elf_select_section): Put ATTRIBUTE_UNUSED
9117 2013-08-26 Uros Bizjak <ubizjak@gmail.com>
9119 * config/i386/i386.c (ix86_debug_options): Remove prototype.
9120 (x86_64_elf_select_section): Ditto.
9121 (ix86_handle_tm_regparm_attribute): Remove ATTRIBUTE_UNUSED on used
9123 (ix86_pass_by_reference): Ditto.
9124 (output_set_got): Ditto.
9125 (ix86_unary_operator_ok): Ditto.
9126 (ix86_expand_builtin): Ditto.
9128 2013-08-23 Jan Hubicka <jh@suse.cz>
9130 * cgraph.c (cgraph_turn_edge_to_speculative): Fix debug output.
9132 2013-08-23 Jan Hubicka <jh@suse.cz>
9134 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
9135 (tree_decl_with_vis): Add FINAL field.
9137 2013-08-23 Jeff Law <law@redhat.com>
9139 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
9140 the debugging dump when the expression is fully redundant.
9142 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
9144 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
9145 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
9146 * pretty-print.c (pp_formatted_text_data): Likewise.
9147 (pp_write_text_to_stream): Likewise.
9148 (pp_write_text_as_dot_label_to_stream): Likewise.
9149 (pp_append_r): Likewise.
9150 (pp_format): Likewise.
9151 (pp_flush): Likewise.
9152 (pp_clear_output_area): Likewise.
9153 (pp_append_text): Likewise.
9154 (pp_formatted_text): Likewise.
9155 (pp_remaining_character_count_for_line): Likewise.
9156 (pp_newline): Likewise.
9157 (pp_character): Likewise.
9158 (output_buffer::~output_buffer): Define.
9159 (pretty_printer::~pretty_printer): Destruct output buffer.
9160 * pretty-print.h (output_buffer::~output_buffer): Declare.
9161 (pretty_printer::~pretty_printer): Declare virtual.
9163 2013-08-24 Marc Glisse <marc.glisse@inria.fr>
9166 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
9167 HOST_WIDE_INT_M1U): New macros.
9168 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
9169 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
9170 unsigned -1 for lshift.
9171 * cse.c (cse_insn): Likewise.
9172 * double-int.c (rshift_double, lshift_double): Likewise.
9173 * builtins.c (fold_builtin_bitop): Likewise.
9174 * combine.c (force_to_mode): Likewise.
9175 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
9176 * simplify-rtx.c (simplify_const_unary_operation,
9177 simplify_const_binary_operation): Likewise.
9178 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
9179 check_va_list_escapes): Likewise.
9180 * rtlanal.c (nonzero_bits1): Likewise.
9181 * expmed.c (expand_smod_pow2): Likewise.
9182 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
9184 2013-08-23 Jan Hubicka <jh@suse.cz>
9186 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
9187 as having address taken.
9189 2013-08-23 Jan Hubicka <jh@suse.cz>
9191 * ipa-utils.h (method_class_type): Declare.
9192 * ipa-devirt.c (method_class_type): Export.
9194 * cgraphunit.c (analyze_functions): Do basic devirtualization;
9195 do not walk base classes of anonymous types.
9197 2013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
9199 PR rtl-optimization/58220
9201 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
9202 handle SEQUENCE insns properly.
9204 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
9206 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
9208 (pp_newline_and_indent): Likewise.
9209 (pp_separate_with): Likewise.
9210 * pretty-print.c (pp_newline_and_flush): Define.
9211 (pp_newline_and_indent): Likewise.
9212 (pp_separate_with): Likewise.
9214 2013-08-23 Jakub Jelinek <jakub@redhat.com>
9217 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
9218 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
9220 2013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
9222 * config/i386/predicates.md (ext_sse_reg_operand): New.
9223 * config/i386/i386.md (*movti_internal): Use
9224 predicate to determine if EVEX is needed.
9225 (*movsi_internal): Ditto.
9226 (*movdf_internal): Ditto.
9227 (*movsf_internal): Ditto.
9228 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
9230 2013-08-23 Jakub Jelinek <jakub@redhat.com>
9232 PR tree-optimization/58209
9233 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
9234 (find_tail_calls): Give up for pointer result types if m is non-NULL.
9235 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
9236 emit POINTER_PLUS_EXPR.
9237 (create_tailcall_accumulator): For pointer result type accumulate in
9240 2013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
9242 * configure.ac: Add backslashes missing from the last change.
9243 * configure: Regenerate.
9245 2013-08-22 Jan Hubicka <jh@susue.cz>
9247 * ipa.c (function_and_variable_visibility): First remember function
9248 was global and then make it local.
9250 2013-08-22 Julian Brown <julian@codesourcery.com>
9252 * configure.ac: Add aarch64 to list of arches which use "nop" in
9254 * configure: Regenerate.
9256 2013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9258 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
9259 gnu_libc_has_function.
9260 * config/s390/tpf.h: Likewise.
9262 2013-08-22 Jan Hubicka <jh@susue.cz>
9264 * timevar.c (validate_phases): Add cast.
9266 2013-08-22 Jan Hubicka <jh@susue.cz>
9268 * timevar.c (validate_phases): Use size_t for memory.
9269 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
9271 2013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
9273 * pretty-print.h (output_buffer::output_buffer): Declare.
9274 (pretty_printer::pretty_printer): Likewise.
9275 (pp_construct): Remove.
9276 * pretty-print.c (output_buffer::output_buffer): Define.
9277 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
9278 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
9279 (print_gimple_expr): Likewise.
9280 (print_gimple_seq): Likewise.
9281 (gimple_dump_bb): Likewise.
9282 * sched-vis.c (dump_value_slim): Likewise.
9283 (dump_insn_slim): Likewise.
9284 (dump_rtl_slim): Likewise.
9285 (str_pattern_slim): Likewise.
9286 * tree-mudflap.c (mf_varname_tree): Likewise.
9287 * graph.c (print_graph_cfg): Likewise.
9288 (start_graph_dump): Likewise.
9289 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
9291 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
9292 pretty printer initialization.
9293 * coretypes.h (diagnostic_context): Remove superflous type alias
9295 (pretty_printer): Likewise. Declare directly as a class.
9296 (pretty_print_info): Remove declaration as class.
9297 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
9298 and pp_clear_output_area.
9299 (asan_add_global): Likewise.
9301 2013-08-22 Jan Hubicka <jh@suse.cz>
9303 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
9304 * ipa-utils.h (update_type_inheritance_graph): Declare.
9305 (possible_polymorphic_call_target_p): Declare.
9306 (possible_polymorphic_call_target_p): New.
9307 * ipa-devirt.c: Update toplevel comments.
9308 (cached_polymorphic_call_targets): Move up.
9309 (odr_type_d): Move ID down.
9310 (polymorphic_type_binfo_p): Update comment.
9311 (odr_hasher::remove): Likewise;
9312 (get_odr_type): Set anonymous_namespace.
9313 (dump_odr_type): Dump it.
9314 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
9315 (maybe_record_node): Record node in cached_polymorphic_call_targets.
9316 (record_binfo): Add comment.
9317 (free_polymorphic_call_targets_hash): Do not ICE when cache is not
9319 (devirt_node_removal_hook): Do not iCE when cache is freed.
9320 (possible_polymorphic_call_target_p): New predicate.
9321 (update_type_inheritance_graph): New function.
9323 2013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
9324 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
9325 Sergey Lega <sergey.s.lega@intel.com>
9326 Anna Tikhonova <anna.tikhonova@intel.com>
9327 Ilya Tocar <ilya.tocar@intel.com>
9328 Andrey Turetskiy <andrey.turetskiy@intel.com>
9329 Ilya Verbin <ilya.verbin@intel.com>
9330 Kirill Yukhin <kirill.yukhin@intel.com>
9331 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
9333 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
9334 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
9335 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
9336 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
9337 (OPTION_MASK_ISA_AVX2_UNSET): Update.
9338 (OPTION_MASK_ISA_AVX512F_UNSET): New.
9339 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
9340 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
9341 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
9342 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
9343 OPT_mavx512pf, OPT_mavx512er cases.
9344 * config/i386/constraints.md (v): New constraint.
9345 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
9346 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
9347 (bit_AVX512CD): New.
9348 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
9349 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
9350 * config/i386/i386-c.c (ix86_target_macros_internal):
9351 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
9353 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
9354 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
9355 * config/i386/i386.c (regclass_map, dbx_register_map)
9356 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
9357 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
9358 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
9360 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
9361 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
9362 -mavx512pf options. Fix formatting.
9363 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
9364 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
9365 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
9366 -mavx512cd, -mavx512pf options.
9367 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
9368 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
9369 (ix86_preferred_output_reload_class): Replace SSE_REGS with
9371 (ix86_hard_regno_mode_ok): Support 512-bit registers.
9372 (ix86_set_reg_reg_cost): Ditto.
9373 (x86_order_regs_for_local_alloc): Ditto.
9374 (MAX_VECT_LEN): Extend to 64-byte.
9375 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
9376 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
9377 (TARGET_AVX512ER, TARGET_AVX512CD): New.
9378 (BIGGEST_ALIGNMENT): Extend to 512-bits.
9379 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
9380 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
9381 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
9382 (SSE_REG_MODE_P): Support new modes.
9383 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
9384 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
9385 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
9386 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
9387 (REG_CLASS_CONTENTS): Add new registers.
9388 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
9389 (EXT_REX_SSE_REGNO_P): New.
9390 (HI_REGISTER_NAMES): Add new registers.
9391 * config/i386/i386.md: Define constants for new registers.
9392 (mode): Add new 512-bit modes.
9393 (prefix): Support evex prefix.
9394 (isa): Support avx512f, noavx512f, fma_avx512f.
9395 (ssemodesuffix): Add new 512-bit modes.
9397 (*movxi_internal_avx512f): Ditto.
9398 (*movdi_internal): Replace constraint "x" with the new constraint "v".
9400 (*movsi_internal): Likewise.
9401 (*movdf_internal): Likewise.
9402 (*movsf_internal): Likewise.
9403 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
9404 (<code><mode>3): Likewise.
9405 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd):
9407 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
9408 with the new constraint "v".
9409 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
9411 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
9412 with the new constraint "v".
9413 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
9414 (<sse2>_storedqu<avxsizesuffix>): Likewise.
9415 (*<plusminus_insn><mode>3): Likewise.
9416 (<sse>_vm<plusminus_insn><mode>3): Likewise.
9417 (*mul<mode>3): Likewise.
9418 (<sse>_vmmul<mode>3): Likewise.
9419 (<sse>_div<mode>3): Likewise.
9420 (<sse>_vmdiv<mode>3): Likewise.
9421 (<sse>_sqrt<mode>2): Likewise.
9422 (<sse>_vmsqrt<mode>2): Likewise.
9423 (*<code><mode>3_finite): Likewise.
9424 (*<code><mode>3) <smaxmin>: Likewise.
9425 (<sse>_vm<code><mode>3): Likewise.
9426 (*<code><mode>3) <any_logic>: Likewise.
9427 (*fma_fmadd_<mode>): Likewise.
9428 (*fma_fmsub_<mode>): Likewise.
9429 (*fma_fnmadd_<mode>): Likewise.
9430 (*fma_fnmsub_<mode>): Likewise.
9431 (*fma_fmaddsub_<mode>): Likewise.
9432 (*fma_fmsubadd_<mode>): Likewise.
9433 (*fmai_fmadd_<mode>): Likewise.
9434 (*fmai_fmsub_<mode>): Likewise.
9435 (*fmai_fnmadd_<mode>): Likewise.
9436 (*fmai_fnmsub_<mode>): Likewise.
9437 (sse_cvtsi2ss): Likewise.
9438 (sse_cvtsi2ssq): Likewise.
9439 (sse_cvtss2si): Likewise.
9440 (sse_cvtss2si_2): Likewise.
9441 (sse_cvtss2siq): Likewise.
9442 (sse_cvtss2siq_2): Likewise.
9443 (sse_cvttss2si): Likewise.
9444 (sse_cvtss2siq_2): Likewise.
9445 (float<sseintvecmodelower><mode>2): Likewise.
9446 (sse2_cvtsd2si_2): Likewise.
9447 (sse2_cvtsd2siq_2): Likewise.
9448 (*<plusminus_insn><mode>3): Likewise.
9449 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
9450 (*<sse4_1_avx2>_mul<mode>3): Likewise.
9451 (ashr<mode>3): Likewise.
9452 (<shift_insn><mode>3): Likewise.
9453 (avx2_<code><mode>3): Likewise.
9454 (*avx2_<code><mode>3): Likewise.
9455 (*andnot<mode>3): Likewise.
9456 (*<code><mode>3) <any_logic>: Likewise.
9457 (abs<mode>2): Likewise.
9458 (avx2_permvar<mode>): Likewise.
9459 (avx2_perm<mode>_1): Likewise.
9460 (*avx_vpermilp<mode>): Likewise.
9461 (avx_vpermilvar<mode>3): Likewise.
9462 (avx2_ashrv<mode>): Likewise.
9463 (avx2_<shift_insn>v<mode>): Likewise.
9464 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
9466 * doc/rtl.texi: Document XImode.
9468 2013-08-21 Jeff Law <law@redhat.com>
9470 * tree-flow.h (register_jump_thread): Pass vector of edges
9471 instead of each important edge.
9472 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
9473 thread path into a vector and pass that to register_jump_thread.
9474 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
9475 passed in edge vector to the current 3-edge form.
9478 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
9480 * dce.c (fini_dce): Call df_analyze again just in case
9481 delete_unmarked_insns removed anything.
9483 2013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
9485 * reload.h (struct reg_equivs): Rename to ..
9486 (struct reg_equivs_s): .. this.
9488 2013-08-20 Martin Liska <marxin.liska@gmail.com>
9490 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
9492 2013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
9494 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest comment.
9496 2013-08-21 Jeff Law <law@redhat.com>
9498 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
9499 simplify assignments too. If the RHS collapses to a singleton
9500 range, then return the value for the range.
9502 2013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
9504 * config/i386/sse.md (V16): Rename to...
9506 (mov<mode>): Update iterator name.
9507 (*mov<mode>_internal): Ditto.
9508 (push<mode>1): Ditto.
9509 (movmisalign<mode>): Ditto.
9511 2013-08-20 Jan Hubicka <jh@suse.cz>
9514 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
9515 entry for direct edges.
9516 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
9518 2013-08-20 David Malcolm <dmalcolm@redhat.com>
9520 Revert my last two changes, r201865 and r201864:
9523 2013-08-20 David Malcolm <dmalcolm@redhat.com>
9525 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
9526 instances can own GC refs.
9528 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
9529 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
9530 (gcc::context::gt_pch_nx): Likewise.
9531 (gcc::context::gt_pch_nx): Likewise.
9532 * ggc.h (gt_ggc_mx <T>): New.
9533 (gt_pch_nx_with_op <T>): New.
9534 (gt_pch_nx <T>): New.
9535 * passes.c (opt_pass::gt_ggc_mx): New.
9536 (opt_pass::gt_pch_nx): New.
9537 (opt_pass::gt_pch_nx_with_op): New.
9538 (pass_manager::gt_ggc_mx): New.
9539 (pass_manager::gt_pch_nx): New.
9540 (pass_manager::gt_pch_nx_with_op): New.
9541 (pass_manager::operator new): Use
9542 ggc_internal_cleared_alloc_stat rather than xcalloc.
9543 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
9544 (pass_manager::gt_ggc_mx): New.
9545 (pass_manager::gt_pch_nx): New.
9546 (pass_manager::gt_pch_nx_with_op): New.
9547 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
9548 (opt_pass::operator new): New.
9549 (opt_pass::gt_ggc_mx): New.
9550 (opt_pass::gt_pch_nx): New.
9551 (opt_pass::gt_pch_nx_with_op): New.
9554 2013-08-20 David Malcolm <dmalcolm@redhat.com>
9556 * Makefile.in (GTFILES): Add context.h.
9557 * context.c (gcc::context::operator new): New.
9558 (gcc::context::gt_ggc_mx): New.
9559 (gcc::context::gt_pch_nx): New.
9560 (gcc::context::gt_pch_nx): New.
9561 * context.h (gcc::context): Add GTY((user)) marking.
9562 (gcc::context::operator new): New.
9563 (gcc::context::gt_ggc_mx): New.
9564 (gcc::context::gt_pch_nx): New.
9565 (gcc::context::gt_pch_nx): New.
9566 (g): Add GTY marking.
9567 (gt_ggc_mx (gcc::context *)): New.
9568 (gt_pch_nx (gcc::context *)): New.
9569 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
9570 void *cookie)): New.
9571 * gengtype.c (open_base_files) <ifiles>: Add context.h.
9573 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
9575 * dce.c (fini_dce): Call df_analyze again just in case
9576 delete_unmarked_insns removed anything.
9578 2013-08-20 Teresa Johnson <tejohnson@google.com>
9580 PR rtl-optimizations/57451
9581 * final.c (reemit_insn_block_notes): Prevent lexical blocks
9582 from crossing split section boundaries.
9584 2013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
9586 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
9587 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
9588 with MULTLIB_DEFAULTS.
9590 2013-08-20 Nick Clifton <nickc@redhat.com>
9592 * target.def (narrow_volatile_bitfield): Note that the default
9593 value is false, not !TARGET_STRICT_ALIGN.
9594 * doc/tm.texi: Regenerate.
9596 2013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
9598 Fix LIB_SPEC for systems without libpthread.
9600 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
9601 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
9603 * config/i386/linux-common.h: Likewise.
9604 * config/mips/linux-common.h: Likewise.
9606 2013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
9608 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
9611 2013-08-20 David Malcolm <dmalcolm@redhat.com>
9613 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
9614 instances can own GC refs.
9616 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
9617 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
9618 (gcc::context::gt_pch_nx): Likewise.
9619 (gcc::context::gt_pch_nx): Likewise.
9620 * ggc.h (gt_ggc_mx <T>): New.
9621 (gt_pch_nx_with_op <T>): New.
9622 (gt_pch_nx <T>): New.
9623 * passes.c (opt_pass::gt_ggc_mx): New.
9624 (opt_pass::gt_pch_nx): New.
9625 (opt_pass::gt_pch_nx_with_op): New.
9626 (pass_manager::gt_ggc_mx): New.
9627 (pass_manager::gt_pch_nx): New.
9628 (pass_manager::gt_pch_nx_with_op): New.
9629 (pass_manager::operator new): Use
9630 ggc_internal_cleared_alloc_stat rather than xcalloc.
9631 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
9632 (pass_manager::gt_ggc_mx): New.
9633 (pass_manager::gt_pch_nx): New.
9634 (pass_manager::gt_pch_nx_with_op): New.
9635 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
9636 (opt_pass::operator new): New.
9637 (opt_pass::gt_ggc_mx): New.
9638 (opt_pass::gt_pch_nx): New.
9639 (opt_pass::gt_pch_nx_with_op): New.
9641 2013-08-20 David Malcolm <dmalcolm@redhat.com>
9643 * Makefile.in (GTFILES): Add context.h.
9644 * context.c (gcc::context::operator new): New.
9645 (gcc::context::gt_ggc_mx): New.
9646 (gcc::context::gt_pch_nx): New.
9647 (gcc::context::gt_pch_nx): New.
9648 * context.h (gcc::context): Add GTY((user)) marking.
9649 (gcc::context::operator new): New.
9650 (gcc::context::gt_ggc_mx): New.
9651 (gcc::context::gt_pch_nx): New.
9652 (gcc::context::gt_pch_nx): New.
9653 (g): Add GTY marking.
9654 (gt_ggc_mx (gcc::context *)): New.
9655 (gt_pch_nx (gcc::context *)): New.
9656 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
9657 void *cookie)): New.
9658 * gengtype.c (open_base_files) <ifiles>: Add context.h.
9660 2013-08-20 Alan Modra <amodra@gmail.com>
9663 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
9664 (rs6000_emit_epilogue): Likewise.
9666 2013-08-19 Dehao Chen <dehao@google.com>
9668 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
9670 2013-08-19 Peter Bergner <bergner@vnet.ibm.com>
9671 Jakub Jelinek <jakub@redhat.com>
9673 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
9674 (BUILT_IN_FABSD64): Likewise.
9675 (BUILT_IN_FABSD128): Likewise.
9676 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
9677 (fold_builtin_1): Likewise.
9678 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping
9679 destination and source operands.
9680 (*abstd2_fpr): Likewise.
9681 (*nabstd2_fpr): Likewise.
9683 2013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
9685 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
9688 2013-08-19 Aldy Hernandez <aldyh@redhat.com>
9690 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
9693 2013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
9695 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
9696 * builtins.c (default_libc_has_function): New.
9697 (gnu_libc_has_function): Ditto.
9698 (no_c99_libc_has_function): Ditto.
9699 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
9700 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
9701 (fold_builtin_sincos): Likewise.
9702 (fold_builtin_cexp): Likewise.
9703 * builtins.def (DEF_C94_BUILTIN): Likewise.
9704 (DEF_C99_BUILTIN): Likewise.
9705 (DEF_C99_C90RES_BUILTIN): Likewise.
9706 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
9707 definitions to using this define.
9708 * config/darwin-protos.h (darwin_libc_has_function): New.
9709 * config/darwin.c (darwin_libc_has_function): Ditto.
9710 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
9711 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
9712 * config/darwin.h: Ditto.
9713 * config/elfos.h: Ditto.
9714 * config/freebsd.h: Ditto.
9715 * config/i386/cygming.h: Ditto.
9716 * config/i386/djgpp.h: Ditto.
9717 * config/i386/i386-interix.h: Ditto.
9718 * config/microblaze/microblaze.h: Ditto.
9719 * config/mmix/mmix.h: Ditto.
9720 * config/gnu-user.h: Ditto.
9721 * config/ia64/hpux.h: Ditto.
9722 * config/pa/pa-hpux.h: Ditto.
9723 * config/pdp11/pdp11.h: Ditto.
9724 * config/picochip/picochip.h: Ditto.
9725 * config/linux.h: Ditto.
9726 * config/netbsd.h: Ditto.
9727 * config/openbsd.h: Ditto.
9728 * config/rs6000/aix43.h: Ditto.
9729 * config/rs6000/aix51.h: Ditto.
9730 * config/rs6000/aix52.h: Ditto.
9731 * config/rs6000/aix53.h: Ditto.
9732 * config/rs6000/aix61.h: Ditto.
9733 * config/rs6000/darwin.h: Ditto.
9734 * config/rs6000/linux.h: Ditto.
9735 * config/rs6000/linux64.h: Ditto.
9736 * config/s390/tpf.h: Ditto.
9737 * config/sol2-10.h: Ditto.
9738 * config/sol2.h: Ditto.
9739 * config/vms/vms.h: Ditto.
9740 * config/vxworks.h: Ditto.
9741 * config/linux-android.c (linux_android_libc_has_function):
9742 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
9743 * config/linux-protos.h (linux_android_libc_has_function):
9745 * config/i386/i386.c (ix86_libc_has_function): New.
9746 * config/i386/i386-protos.h
9747 (ix86_libc_has_function): New declaration.
9748 * config/i386/i386.md
9749 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
9750 ("isinf<mode>2): Likewise.
9751 * convert.c (convert_to_integer): Using new target hook
9752 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
9753 TARGET_C99_FUNCTIONS.
9754 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
9755 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
9756 * coretypes.h (function_class): New enum for different
9757 classes of functions.
9758 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
9759 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
9760 (TARGET_HAS_SINCOS): Likewise.
9761 (TARGET_LIBC_HAS_FUNCTION): New.
9762 * doc/tm.texi: Regenerated.
9763 * targhooks.h (default_libc_has_function): New declaration.
9764 (no_c99_libc_has_function): Ditto.
9765 (gnu_libc_has_function): Ditto.
9766 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
9767 and TARGET_HAS_SINCOS.
9769 2013-08-18 Jan Hubicka <jh@suse.cz>
9771 * Makeifle-in (ipa-devirt.o): New.
9772 (GTFILES): Add ipa-utils.h and ipa-devirt.c
9773 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
9774 (analyze_functions): Look into possible targets of polymorphic call.
9775 * dumpfile.c (dump_files): Add type-inheritance dump.
9776 * dumpfile.h (TDI_inheritance): New.
9777 * ipa-devirt.c: New file.
9778 * ipa-utils.h (odr_type_d): Forward declare.
9779 (odr_type): New type.
9780 (build_type_inheritance_graph): Declare.
9781 (possible_polymorphic_call_targets): Declare and introduce inline
9782 variant when only edge is pased.
9783 (dump_possible_polymorphic_call_targets): Likewise.
9784 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
9785 * tree.c (type_in_anonymous_namespace_p): Break out from ...
9786 (types_same_for_odr): ... here.
9787 * tree.h (type_in_anonymous_namespace_p): Declare.
9789 2013-08-18 Jakub Jelinek <jakub@redhat.com>
9791 PR tree-optimization/58006
9792 * tree-parloops.c (take_address_of): Don't ICE if get_name
9794 (eliminate_local_variables_stmt): Remove clobber stmts.
9796 2013-08-18 Eric Botcazou <ebotcazou@adacore.com>
9798 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
9799 error message is issued for an alias to undefined symbol.
9801 2013-08-18 Jan Hubicka <jh@suse.cz>
9803 * cgraph.c (cgraph_create_indirect_edge): Discover
9804 polymorphic calls and record basic info into indirect_info.
9805 * gimple-fold.c (gimple_fold_call): When doing BINFO based
9806 devirtualization, ignore objc function calls.
9807 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
9808 call with no parm index info.
9809 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
9810 * tree.c (virtual_method_call_p): New function.
9811 * tree.h (virtual_method_call_p): Declare.
9813 2013-08-16 Jan Hubicka <jh@suse.cz>
9816 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
9818 2013-08-16 David Edelsohn <dje.gcc@gmail.com>
9820 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
9823 2013-08-16 David Malcolm <dmalcolm@redhat.com>
9825 * gengtype.c (type_for_name): Add special-case support for
9826 locating types within the "gcc::" namespace.
9827 (open_base_files): Emit a "using namespace gcc" directive.
9829 2013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
9832 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
9833 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
9835 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
9836 array instead of each individual operand as a separate argument.
9837 (emit_fusion_gpr_load): Likewise.
9838 (expand_fusion_gpr_load): Add new function declaration.
9840 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
9841 signature to have the operands passed as an array, instead of as
9842 separate arguments. Allow ZERO_EXTEND to be in the memory
9843 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
9844 depend on the register live/dead flags when peepholes are run.
9845 (expand_fusion_gpr_load): New function to be called from the
9846 peephole2 pass, to change the register that addis sets to be the
9848 (emit_fusion_gpr_load): Change the calling signature to have the
9849 operands passed as an array, instead of as separate arguments.
9850 Allow ZERO_EXTEND to be in the memory address, and also
9851 SIGN_EXTEND if -mpower8-fusion-sign.
9853 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
9855 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
9856 adjust the register addis loads up in the peephole2 pass. Do not
9857 depend on the register live/dead state when the peephole pass is done.
9859 2013-08-16 David Malcolm <dmalcolm@redhat.com>
9861 * gengtype.c (create_user_defined_type): Ensure that the kind
9862 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
9863 declaration is seen before the GTY((user)) marking.
9865 2013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
9868 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
9870 2013-08-16 Jan Hubicka <jh@suse.cz>
9872 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
9873 arugment expected_type.
9874 (gimple_fold_call): Use it.
9875 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
9876 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
9877 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
9878 (try_make_edge_direct_virtual_call): Likewise.
9879 * tree.c (obj_type_ref_class): New.
9880 * tree.h (obj_type_ref_class): Use it.
9882 2013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
9884 * sched-vis.c (rtl_slim_pp_initialized): Remove.
9885 (rtl_slim_pp): Likewise.
9886 (init_rtl_slim_pretty_print): Likewise.
9887 (dump_value_slim): Don't call it. Use local pretty printer.
9888 (dump_insn_slim): Likewise.
9889 (dump_rtl_slim): Likewise.
9890 (str_pattern_slim): Likewise.
9891 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
9894 2013-08-16 Jakub Jelinek <jakub@redhat.com>
9896 PR tree-optimization/58164
9897 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
9898 walk gimple_goto_dest of GIMPLE_GOTO.
9900 PR tree-optimization/58165
9901 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
9902 bi_call must be the last stmt in a bb, don't split_block, instead
9903 use fallthru edge from it and give up if there is none.
9904 Release conds vector when returning early.
9906 2013-08-14 Xinliang David Li <davidxl@google.com>
9908 * config/i386/i386.c (ix86_option_override_internal):
9909 Remove unused variable and field.
9911 2013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9914 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
9915 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
9916 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
9917 and Linux, correct BLKmode alignment when 128-bit alignment is
9918 required and compatibility flag is not set.
9919 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
9920 for zero-size arguments when compatibility flag is not set.
9922 2013-08-14 Jakub Jelinek <jakub@redhat.com>
9924 PR tree-optimization/58145
9925 * tree-sra.c (build_ref_for_offset): If prev_base has
9926 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
9928 2013-08-14 Xinliang David Li <davidxl@google.com>
9930 * config/i386/i386.c (ix86_option_override_internal):
9931 Fix uninitialized variable error.
9933 2013-08-14 Xinliang David Li <davidxl@google.com>
9935 * config/i386/i386.opt: Define two new options.
9936 * config/i386/x86-tune.def: Add arch selector field in macros.
9937 * config/i386/i386.h: Adjust macro definition.
9938 * config/i386/i386.c (ix86_option_override_internal):
9940 (parse_mtune_ctrl_str): New function.
9941 (set_ix86_tune_features): New function.
9942 (ix86_function_specific_restore): Call the new helper function.
9944 2013-08-14 Andrey Belevantsev <abel@ispras.ru>
9946 PR rtl-optimization/57662
9947 * sel-sched.c (code_motion_process_successors): When the current insn
9948 is removed after the recursive traversal, break from the loop.
9949 Add comments and debug printouts.
9951 2013-08-14 Jakub Jelinek <jakub@redhat.com>
9952 Alexandre Oliva <aoliva@redhat.com>
9955 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
9956 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
9957 there also UNSPEC_PLTOFF.
9959 2013-08-14 Marek Polacek <polacek@redhat.com>
9961 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
9962 NUM_CONDITIONS bit positions.
9964 2013-08-13 Cary Coutant <ccoutant@google.com>
9966 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
9967 (attr_checksum): Hash vector contents instead of pointer.
9968 (attr_checksum_ordered): Likewise.
9970 2013-08-13 Uros Bizjak <ubizjak@gmail.com>
9972 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
9973 when Pmode != word_mode. Add length_address attribute.
9974 (sse3_monitor_<mode>): Merge from sse3_monitor and
9975 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
9976 Pmode != word_mode. Update insn length attribute.
9977 * config/i386/i386.c (ix86_option_override_internal): Update
9978 ix86_gen_monitor selection for merged sse3_monitor insn.
9980 2013-08-13 Julian Brown <julian@codesourcery.com>
9982 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
9983 perform invalid legitimization on greater-than-word-size modes for
9986 2013-08-13 Vladimir Makarov <vmakarov@redhat.com>
9988 * ira.c (setup_class_translate_array): Use aclass instead of cl
9989 for classes not fully covered by allocno classes.
9991 2013-08-13 Jakub Jelinek <jakub@redhat.com>
9993 PR tree-optimization/57661
9994 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
9995 * tree-inline.c (tree_function_versioning): Initialize it.
9996 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
9997 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
9998 that is not being copied.
10001 * asan.c (instrument_strlen_call): Fix typo in comment.
10002 Use char * type even for the lhs of POINTER_PLUS_EXPR.
10004 2013-08-13 Steve Ellcey <sellcey@mips.com>
10006 * config/mips/mips.md (prefetch): Use lw instead of ld on
10007 loongson in 32bit mode.
10009 2013-08-13 Nick Clifton <nickc@redhat.com>
10011 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
10013 2013-08-13 Jan Hubicka <jh@suse.cz>
10015 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
10016 introduced edge; fix typo in sanity check.
10017 (cgraph_resolve_speculation): Export; improve diagnostic.
10018 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
10019 speculation at type mismatch.
10020 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
10021 (cgraph_resolve_speculation): Declare.
10022 (symtab_can_be_discarded): New function.
10023 * value-prof.c (gimple_ic_transform): Remove actual transform code.
10024 * ipa-inline-transform.c (speculation_removed): New global var.
10025 (clone_inlined_nodes): See if speculation can be removed.
10026 (inline_call): If speculations was removed, we growths may not match.
10027 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
10028 (speculation_useful_p): New function.
10029 (resolve_noninline_speculation): New function.
10030 (inline_small_functions): Resolve useless speculations.
10031 * ipa-inline.h (speculation_useful_p): Declare
10032 * ipa.c (can_replace_by_local_alias): Simplify.
10033 (ipa_profile): Produce speculative calls in non-lto, too;
10034 add simple cost model; produce local aliases.
10036 2013-08-13 David Malcolm <dmalcolm@redhat.com>
10038 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
10041 2013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
10043 * config/i386/i386.c (ix86_function_versions): Use error + inform.
10045 2013-08-12 Uros Bizjak <ubizjak@gmail.com>
10047 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
10048 iterator instead of X87MODEF.
10050 2013-08-12 Perez Read <netfirewall@gmail.com>
10053 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
10054 operand 0 for intel asm alternative.
10055 (*movabs<mode>_2): Ditto for operand 1.
10057 2013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
10059 * config/aarch64/arm_none.h
10060 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
10062 2013-08-12 Nick Clifton <nickc@redhat.com>
10064 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
10066 2013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
10068 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
10069 expand for QI/HImode operand to produce more effictive code for
10070 unsigned char(short) --> float(double) conversion.
10072 2013-08-12 Alexander Monakov <amonakov@ispras.ru>
10074 * doc/invoke.texi: Mention that -ftls-model does not force the final
10077 2013-08-12 Marek Polacek <polacek@redhat.com>
10078 Marc Glisse <marc.glisse@inria.fr>
10080 PR tree-optimization/57980
10081 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
10082 when creating -1 constant.
10084 2013-08-10 Jan Hubicka <jh@suse.cz>
10086 Workaround binutils PR14342.
10087 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
10088 (gimple_init_edge_profiler): Likewise.
10089 (gimple_gen_ic_func_profiler): Likewise.
10091 2013-08-09 Jan Hubicka <jh@suse.cz>
10093 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
10095 2013-08-09 Xinliang David Li <davidxl@google.com>
10097 * config/i386/stringop.def: New file.
10098 * config/i386/stringop.opt: New file.
10099 * config/i386/i386-opts.h: Include stringopt.def.
10100 * config/i386/i386.opt: Include stringopt.opt.
10101 * config/i386/i386.c (ix86_option_override_internal):
10102 Override default size based stringop inline strategies with options.
10103 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
10106 2013-08-09 Jan Hubicka <jh@suse.cz>
10108 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
10110 2013-08-09 Jan Hubicka <jh@suse.cz>
10112 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
10114 (dump_cgraph_node): Dump profile-id.
10115 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
10116 and common_target_probability.
10117 * lto-cgraph.c (lto_output_edge): Stream common targets.
10118 (lto_output_node): Stream profile ids.
10119 (input_node): Stream profile ids.
10120 (input_edge): Stream common targets.
10121 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
10122 * ipa.c: Include value-prof.h
10123 (ipa_profile_generate_summary): Turn indirect call statement histograms
10124 into common targets.
10125 (ipa_profile): Turn common targets into speculative edges.
10127 2013-08-09 Jan Hubicka <jh@suse.cz>
10129 * cgraph.h (cgraph_node): Add profile_id.
10130 * value-prof.c (cgraph_node_map): Turn into pointer_map.
10131 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
10132 (del_node_map): Update.
10133 (find_func_by_funcdef_no): Replace by ...
10134 (find_func_by_profile_id): ... this one.
10135 (gimple_ic_transform): Do not remove useful histograms when
10136 speculation is not done; dump info when indirect call removal
10138 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
10139 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
10140 (__gcov_indirect_call_profiler_v2): .. this one.
10141 * profile.h (init_node_map): Update.
10142 * coverage.c (coverage_compute_profile_id): New function.
10143 * coverage.h (coverage_compute_profile_id): Declare.
10144 * tree-profile.c (init_ic_make_global_vars): Make
10145 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
10146 (gimple_init_edge_profiler): Update prototype of
10147 __gcov_indirect_call_profiler.
10148 (gimple_gen_ic_func_profiler): Simplify.
10149 (tree_profiling): Use init_node_map
10151 2013-08-09 Jan Hubicka <jh@suse.cz>
10153 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
10154 non-speculative refs.
10155 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
10156 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
10157 (cgraph_set_call_stmt): Likewise.
10158 (cgraph_create_edge_1): Fix release checking compilatoin;
10159 clear lto_stmt_uid.
10160 (cgraph_free_edge): Free indirect info.
10161 (cgraph_turn_edge_to_speculative): New function.
10162 (cgraph_speculative_call_info): New function.
10163 (cgraph_make_edge_direct): Return direct edge; handle speculation.
10164 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
10165 (dump_cgraph_node): Dump speculation.
10166 (verify_edge_count_and_frequency): Accept speculative edges.
10167 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
10168 (verify_cgraph_node): Handle speculation.
10169 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
10170 (cgraph_set_call_stmt): Update prototype.
10171 (cgraph_make_edge_direct): Update prototype.
10172 (cgraph_speculative_call_info): Declare.
10173 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
10174 to change; update call of ipa_find_references.
10175 * ipa-ref.c (ipa_record_reference): Fix return value; clear
10176 lto_stmt_uid and speculative flags.
10177 (ipa_dump_references): Dump speculation.
10178 (ipa_clone_references): Clone speculative flag.
10179 (ipa_clone_referring): Likewise.
10180 (ipa_clone_ref): New function.
10181 (ipa_find_reference): Look into lto_stmt_uids
10182 (ipa_clear_stmts_in_references): Do not clear speculative calls.
10183 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
10184 (ipa_find_reference): Update declaration.
10185 (ipa_clone_ref): Declare.
10186 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
10187 stream speculative flag.
10188 (lto_output_ref): Stream statements uids and speculation.
10189 (input_ref): Likewise.
10190 (input_edge): Stream speuclation.
10191 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
10192 (cgraph_set_call_stmt_including_clones): Handle speculation.
10193 * ipa-inline.c (heap_edge_removal_hook): New function.
10194 (inline_small_functions): Register it.
10195 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
10196 also initialize refs.
10197 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
10199 (try_make_edge_direct_simple_call): Likewise.
10200 (try_make_edge_direct_simple_call): Likewise.
10201 (update_indirect_edges_after_inlining): Likewise.
10202 (remove_described_reference): Look proper lto_stmt_uid.
10203 (propagate_controlled_uses): Likewise.
10204 (propagate_controlled_uses): Liekwise.
10205 * tree-inline.c (copy_bb): Copy speculative edges.
10206 (redirect_all_calls): New function.
10207 (copy_cfg_body): Do redirection after loop info is updated.
10208 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
10210 2013-08-09 Jan Hubicka <jh@suse.cz>
10212 * lto-streamer-out.c (output_function): Renumber PHIs.
10213 * lto-streamer-in.c (input_function): Likewise.
10215 2013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
10217 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
10218 (get_lane_unsigned): Likewise.
10219 (dup_lane_scalar): Likewise.
10220 (get_lane): enable for VALL.
10221 * config/aarch64/aarch64-simd.md
10222 (aarch64_dup_lane_scalar<mode>): Remove.
10223 (aarch64_get_lane_signed<mode>): Likewise.
10224 (aarch64_get_lane_unsigned<mode>): Likewise.
10225 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
10226 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
10227 (aarch64_get_lane<mode>): Enable for all vector modes.
10228 (aarch64_get_lanedi): Remove misleading constraints.
10229 * config/aarch64/arm_neon.h
10230 (__aarch64_vget_lane_any): Define.
10231 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
10232 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
10233 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
10234 * config/aarch64/iterators.md (VDQQH): New.
10235 (VDQQHS): Likewise.
10236 (vwcore): Likewise.
10238 2013-08-09 Eric Botcazou <ebotcazou@adacore.com>
10240 * configure.ac: Add GAS check for LEON instructions on SPARC.
10241 * configure: Regenerate.
10242 * config.in: Likewise.
10243 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
10245 * config/sparc/sparc.opt (LEON, LEON3): New masks.
10246 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
10248 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
10249 (AS_LEON_FLAG): New macro.
10250 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
10251 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
10252 Deal with LEON and LEON3 for the memory model.
10253 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
10254 (atomic_compare_and_swap<mode>_1): Likewise.
10255 (*atomic_compare_and_swap<mode>_1): Likewise.
10257 2013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
10259 * config/arm/neon.md (vcond): Fix floating-point vector
10260 comparisons against 0.
10262 2013-08-08 Vladimir Makarov <vmakarov@redhat.com>
10264 * lra-constraints.c (emit_spill_move): Remove assert.
10265 (process_alt_operands): Add more debugging
10266 output. Increase reject for spilling into memory. Decrease
10267 reject for reloading scratch.
10268 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
10270 2013-08-08 Steve Ellcey <sellcey@mips.com>
10272 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
10273 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
10274 micromips incompatible. Add nan2008.
10275 (MULTILIB_DIRNAMES): Add nan2008.
10276 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
10277 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
10278 and micromips incompatible. Add nan2008.
10279 (MULTILIB_DIRNAMES): Add nan2008.
10280 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
10282 2013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
10284 PR rtl-optimization/58079
10285 * combine.c (combine_simplify_rtx): Avoid using SUBST if
10286 simplify_comparison has widened a comparison with an integer.
10288 2013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10290 * config/arm/neon.md (movmisalign<mode>): Disable when we
10291 don't allow unaligned accesses.
10292 (*movmisalign<mode>_neon_store): Likewise.
10293 (*movmisalign<mode>_neon_load): Likewise.
10294 (*movmisalign<mode>_neon_store): Likewise.
10295 (*movmisalign<mode>_neon_load): Likewise.
10297 2013-08-08 Jan Hubicka <jh@suse.cz>
10299 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
10300 (make_pass_rebuild_cgraph_edges): Also clear references.
10301 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
10302 * ipa-inline-transform.c (inline_transform): Remove all references
10304 * cgraphunit.c (expand_function): Remove all references after
10306 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
10307 (ipa_find_reference): Rewrite to iterator.
10308 (remove_stmt_references): Likewise.
10309 (ipa_clear_stmts_in_references): New function.
10310 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
10311 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
10313 * ipa-split.c (split_function): Remove references in split function.
10315 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
10318 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
10319 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
10321 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
10324 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
10325 suggested mode for the assignment isn't compatible with the
10326 registers required.
10328 2013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
10331 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
10333 2013-08-07 Xinliang David Li <davidxl@google.com>
10335 * config/i386/i386.opt: New option -mtune-ctrl=.
10336 * config/i386/x86-tune.def: New file.
10337 * config/i386/i386.h: include x86-tune.def.
10338 * config/i386/i386.c (ix86_option_override_internal):
10339 Parsing -mtune-ctrl= option and set tune features.
10341 2013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
10344 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
10345 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
10348 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
10350 * rtl.h (update_alignments): Declare.
10351 * final.c (grow_label_align): New function extracted from...
10352 (shorten_branches): ...here. Call it.
10353 (update_alignments): New function.
10354 * reorg.c (sibling_labels): New variable.
10355 (get_label_before): Add SIBLING parameter. If it is non-zero, push
10356 the new label along with it onto the sibling_labels vector.
10357 (fill_simple_delay_slots): Adjust call to get_label_before.
10358 (fill_slots_from_thread): Likewise.
10359 (relax_delay_slots): Likewise.
10360 (make_return_insns): Likewise.
10361 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
10363 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
10365 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
10366 document its semantics.
10367 (diagnostic_report_diagnostic): Adjust accordingly.
10369 2013-08-07 David Malcolm <dmalcolm@redhat.com>
10371 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
10372 (sparc_option_override): ...and port to new C++ pass API.
10373 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
10375 2013-08-07 Peter Bergner <bergner@vnet.ibm.com>
10377 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
10379 2013-08-06 Caroline Tice <cmtice@google.com>
10381 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
10382 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
10383 * tree-pass.h: Add pass_vtable_verify.
10384 * varasm.c (assemble_variable): Add code to properly set the comdat
10385 section and name for the .vtable_map_vars section.
10386 (assemble_vtyv_preinit_initializer): New function.
10387 (default_sectin_type_flags): Make sure .vtable_map_vars section has
10389 * output.h: Add function decl for assemble_vtv_preinit_initializer.
10390 * vtable-verify.c: New file.
10391 * vtable-verify.h: New file.
10392 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
10393 initialiation levels.
10394 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
10395 * passes.def: Insert pass_vtable_verify.
10396 * aclocal.m4: Reorder includes.
10397 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
10398 -fvtv-counts options.
10399 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
10400 as appropriate, if -fvtable-verify=... is used.
10401 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
10402 -fvtable-verify=... is used.
10403 * Makefile.in (OBJS): Add vtable-verify.o to list.
10404 (vtable-verify.o): Add new build rule.
10405 (GTFILES): Add vtable-verify.c to list.
10406 * common.opt (fvtable-verify=): New flag.
10407 (vtv_priority): Values for fvtable-verify= flag.
10408 (fvtv-counts): New flag.
10409 (fvtv-debug): New flag.
10410 * tree.h (save_vtable_map_decl): New extern function decl.
10412 2013-08-07 David Malcolm <dmalcolm@redhat.com>
10414 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
10415 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
10416 (pass_data_rl78_devirt): ...new pass_data instance and...
10417 (make_pass_rl78_devirt): ...new function.
10418 (rl78_asm_file_start): Port pass registration to new C++ API.
10420 2013-08-07 David Malcolm <dmalcolm@redhat.com>
10422 * coretypes.h (rtl_opt_pass): Add.
10423 (gcc::context): Add.
10424 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
10425 (epiphany_init): Port to new C++ pass API.
10426 (epiphany_optimize_mode_switching): Likewise.
10427 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
10428 (pass_manager::get_pass_mode_switching): New.
10429 (pass_manager::get_pass_peephole2): New.
10430 * mode-switching.c (pass_mode_switching): Add clone method.
10431 * recog.c (pass_peephole2): Add clone method.
10432 (pass_split_all_insns): Add clone method.
10434 2013-08-06 David Malcolm <dmalcolm@redhat.com>
10436 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
10437 (mips_option_override): ...here, porting to new C++ API for passes.
10439 2013-08-06 Jan Hubicka <jh@suse.cz>
10441 * cgraph.c (cgraph_get_body): New function based on lto.c
10443 * cgraph.h (cgraph_get_body): Declare.
10444 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
10446 * cgraphunit.c (expand_function): Get body prior expanding.
10447 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
10448 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
10450 * passes.c (do_per_function_toporder): Get body.
10451 * tree-inline.c (expand_call_inline): Get body prior inlining it.
10452 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
10454 2013-08-06 Martin Jambor <mjambor@suse.cz>
10457 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
10458 is not re-finalized. Rename second parameter to no_collect.
10460 2013-08-06 Martin Jambor <mjambor@suse.cz>
10462 PR middle-end/58041
10463 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
10464 MEM_REF has proper alignment information.
10466 2013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
10469 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
10471 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
10472 rtx (*) (rtx, ...) with insn_gen_fn.
10473 * genoutput.c (output_insn_data): Cast gen_? function pointers to
10474 insn_gen_fn::stored_funcptr. Add initializer braces.
10476 2013-08-05 David Malcolm <dmalcolm@redhat.com>
10478 Rewrite how instances of passes are cloned to remove assumptions
10479 about their sizes (thus allowing pass subclasses to have
10480 additional data fields, albeit non-GC-managed ones at this point).
10482 * passes.c (make_pass_instance): Now that passes have clone
10483 methods, rewrite this function to eliminate XNEW and memcpy
10484 calls that used hardcoded sizes. Since this function no longer
10485 creates pass instances, rename it to...
10486 (add_pass_instance): ...this. Document the old way that passes were
10487 numbered and flagged, and rework this function to continue using it.
10488 (next_pass_1): Add an initial_pass argument for use by
10490 (position_pass): When adding multiple instances of a pass, use
10491 the pass's clone method, rather than relying on the XNEW/memcpy
10492 within the former make_pass_instance (now add_pass_instance).
10493 (pass_manager::pass_manager): When invoking next_pass_1, also supply
10494 the initial instance of the current pass within the pass manager.
10496 2013-08-05 David Malcolm <dmalcolm@redhat.com>
10498 This is the automated part of the conversion of passes from C
10499 structs to C++ classes.
10501 Patch autogenerated by refactor_passes.py from
10502 https://github.com/davidmalcolm/gcc-refactoring-scripts
10503 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
10505 * asan.c (pass_asan): Convert from a global struct to a subclass of
10506 gimple_opt_pass along with...
10507 (pass_data_asan): ...new pass_data instance and...
10508 (make_pass_asan): ...new function.
10509 (pass_asan_O0): Convert from a global struct to a subclass of
10510 gimple_opt_pass along with...
10511 (pass_data_asan_O0): ...new pass_data instance and...
10512 (make_pass_asan_O0): ...new function.
10513 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
10514 subclass of rtl_opt_pass along with...
10515 (pass_data_inc_dec): ...new pass_data instance and...
10516 (make_pass_inc_dec): ...new function.
10517 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
10518 a subclass of rtl_opt_pass along with...
10519 (pass_data_reorder_blocks): ...new pass_data instance and...
10520 (make_pass_reorder_blocks): ...new function.
10521 (pass_duplicate_computed_gotos): Convert from a global struct to a
10522 subclass of rtl_opt_pass along with...
10523 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
10524 (make_pass_duplicate_computed_gotos): ...new function.
10525 (pass_partition_blocks): Convert from a global struct to a subclass of
10526 rtl_opt_pass along with...
10527 (pass_data_partition_blocks): ...new pass_data instance and...
10528 (make_pass_partition_blocks): ...new function.
10529 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
10530 struct to a subclass of rtl_opt_pass along with...
10531 (pass_data_branch_target_load_optimize1): ...new pass_data instance
10533 (make_pass_branch_target_load_optimize1): ...new function.
10534 (pass_branch_target_load_optimize2): Convert from a global struct to a
10535 subclass of rtl_opt_pass along with...
10536 (pass_data_branch_target_load_optimize2): ...new pass_data instance
10538 (make_pass_branch_target_load_optimize2): ...new function.
10539 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
10540 of rtl_opt_pass along with...
10541 (pass_data_jump): ...new pass_data instance and...
10542 (make_pass_jump): ...new function.
10543 (pass_jump2): Convert from a global struct to a subclass of
10544 rtl_opt_pass along with...
10545 (pass_data_jump2): ...new pass_data instance and...
10546 (make_pass_jump2): ...new function.
10547 * cfgexpand.c (pass_expand): Convert from a global struct to a
10548 subclass of rtl_opt_pass along with...
10549 (pass_data_expand): ...new pass_data instance and...
10550 (make_pass_expand): ...new function.
10551 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
10552 of rtl_opt_pass along with...
10553 (pass_data_free_cfg): ...new pass_data instance and...
10554 (make_pass_free_cfg): ...new function.
10555 (pass_into_cfg_layout_mode): Convert from a global struct to a
10556 subclass of rtl_opt_pass along with...
10557 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
10558 (make_pass_into_cfg_layout_mode): ...new function.
10559 (pass_outof_cfg_layout_mode): Convert from a global struct to a
10560 subclass of rtl_opt_pass along with...
10561 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
10562 (make_pass_outof_cfg_layout_mode): ...new function.
10563 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
10564 struct to a subclass of gimple_opt_pass along with...
10565 (pass_data_build_cgraph_edges): ...new pass_data instance and...
10566 (make_pass_build_cgraph_edges): ...new function.
10567 (pass_rebuild_cgraph_edges): Convert from a global struct to a
10568 subclass of gimple_opt_pass along with...
10569 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
10570 (make_pass_rebuild_cgraph_edges): ...new function.
10571 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
10572 subclass of gimple_opt_pass along with...
10573 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
10575 (make_pass_remove_cgraph_callee_edges): ...new function.
10576 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
10577 struct to a subclass of rtl_opt_pass along with...
10578 (pass_data_stack_adjustments): ...new pass_data instance and...
10579 (make_pass_stack_adjustments): ...new function.
10580 * combine.c (pass_combine): Convert from a global struct to a subclass
10581 of rtl_opt_pass along with...
10582 (pass_data_combine): ...new pass_data instance and...
10583 (make_pass_combine): ...new function.
10584 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
10585 global struct to a subclass of rtl_opt_pass along with...
10586 (pass_data_compare_elim_after_reload): ...new pass_data instance
10588 (make_pass_compare_elim_after_reload): ...new function.
10589 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
10590 of rtl_opt_pass along with...
10591 (pass_data_rtl_cprop): ...new pass_data instance and...
10592 (make_pass_rtl_cprop): ...new function.
10593 * cse.c (pass_cse): Convert from a global struct to a subclass of
10594 rtl_opt_pass along with...
10595 (pass_data_cse): ...new pass_data instance and...
10596 (make_pass_cse): ...new function.
10597 (pass_cse2): Convert from a global struct to a subclass of
10598 rtl_opt_pass along with...
10599 (pass_data_cse2): ...new pass_data instance and...
10600 (make_pass_cse2): ...new function.
10601 (pass_cse_after_global_opts): Convert from a global struct to a
10602 subclass of rtl_opt_pass along with...
10603 (pass_data_cse_after_global_opts): ...new pass_data instance and...
10604 (make_pass_cse_after_global_opts): ...new function.
10605 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
10606 of rtl_opt_pass along with...
10607 (pass_data_ud_rtl_dce): ...new pass_data instance and...
10608 (make_pass_ud_rtl_dce): ...new function.
10609 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
10610 rtl_opt_pass along with...
10611 (pass_data_fast_rtl_dce): ...new pass_data instance and...
10612 (make_pass_fast_rtl_dce): ...new function.
10613 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
10614 a subclass of rtl_opt_pass along with...
10615 (pass_data_df_initialize_opt): ...new pass_data instance and...
10616 (make_pass_df_initialize_opt): ...new function.
10617 (pass_df_initialize_no_opt): Convert from a global struct to a
10618 subclass of rtl_opt_pass along with...
10619 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
10620 (make_pass_df_initialize_no_opt): ...new function.
10621 (pass_df_finish): Convert from a global struct to a subclass of
10622 rtl_opt_pass along with...
10623 (pass_data_df_finish): ...new pass_data instance and...
10624 (make_pass_df_finish): ...new function.
10625 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
10626 rtl_opt_pass along with...
10627 (pass_data_rtl_dse1): ...new pass_data instance and...
10628 (make_pass_rtl_dse1): ...new function.
10629 (pass_rtl_dse2): Convert from a global struct to a subclass of
10630 rtl_opt_pass along with...
10631 (pass_data_rtl_dse2): ...new pass_data instance and...
10632 (make_pass_rtl_dse2): ...new function.
10633 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
10634 subclass of rtl_opt_pass along with...
10635 (pass_data_dwarf2_frame): ...new pass_data instance and...
10636 (make_pass_dwarf2_frame): ...new function.
10637 * except.c (pass_set_nothrow_function_flags): Convert from a global
10638 struct to a subclass of rtl_opt_pass along with...
10639 (pass_data_set_nothrow_function_flags): ...new pass_data instance
10641 (make_pass_set_nothrow_function_flags): ...new function.
10642 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
10643 subclass of rtl_opt_pass along with...
10644 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
10646 (make_pass_convert_to_eh_region_ranges): ...new function.
10647 * final.c (pass_compute_alignments): Convert from a global struct to a
10648 subclass of rtl_opt_pass along with...
10649 (pass_data_compute_alignments): ...new pass_data instance and...
10650 (make_pass_compute_alignments): ...new function.
10651 (pass_final): Convert from a global struct to a subclass of
10652 rtl_opt_pass along with...
10653 (pass_data_final): ...new pass_data instance and...
10654 (make_pass_final): ...new function.
10655 (pass_shorten_branches): Convert from a global struct to a subclass of
10656 rtl_opt_pass along with...
10657 (pass_data_shorten_branches): ...new pass_data instance and...
10658 (make_pass_shorten_branches): ...new function.
10659 (pass_clean_state): Convert from a global struct to a subclass of
10660 rtl_opt_pass along with...
10661 (pass_data_clean_state): ...new pass_data instance and...
10662 (make_pass_clean_state): ...new function.
10663 * function.c (pass_instantiate_virtual_regs): Convert from a global
10664 struct to a subclass of rtl_opt_pass along with...
10665 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
10666 (make_pass_instantiate_virtual_regs): ...new function.
10667 (pass_leaf_regs): Convert from a global struct to a subclass of
10668 rtl_opt_pass along with...
10669 (pass_data_leaf_regs): ...new pass_data instance and...
10670 (make_pass_leaf_regs): ...new function.
10671 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
10672 subclass of rtl_opt_pass along with...
10673 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
10675 (make_pass_thread_prologue_and_epilogue): ...new function.
10676 (pass_match_asm_constraints): Convert from a global struct to a
10677 subclass of rtl_opt_pass along with...
10678 (pass_data_match_asm_constraints): ...new pass_data instance and...
10679 (make_pass_match_asm_constraints): ...new function.
10680 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
10681 subclass of rtl_opt_pass along with...
10682 (pass_data_rtl_fwprop): ...new pass_data instance and...
10683 (make_pass_rtl_fwprop): ...new function.
10684 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
10685 rtl_opt_pass along with...
10686 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
10687 (make_pass_rtl_fwprop_addr): ...new function.
10688 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
10689 rtl_opt_pass along with...
10690 (pass_data_rtl_pre): ...new pass_data instance and...
10691 (make_pass_rtl_pre): ...new function.
10692 (pass_rtl_hoist): Convert from a global struct to a subclass of
10693 rtl_opt_pass along with...
10694 (pass_data_rtl_hoist): ...new pass_data instance and...
10695 (make_pass_rtl_hoist): ...new function.
10696 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
10697 subclass of gimple_opt_pass along with...
10698 (pass_data_lower_cf): ...new pass_data instance and...
10699 (make_pass_lower_cf): ...new function.
10700 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
10701 from a global struct to a subclass of gimple_opt_pass along with...
10702 (pass_data_strength_reduction): ...new pass_data instance and...
10703 (make_pass_strength_reduction): ...new function.
10704 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
10705 of rtl_opt_pass along with...
10706 (pass_data_rtl_ifcvt): ...new pass_data instance and...
10707 (make_pass_rtl_ifcvt): ...new function.
10708 (pass_if_after_combine): Convert from a global struct to a subclass of
10709 rtl_opt_pass along with...
10710 (pass_data_if_after_combine): ...new pass_data instance and...
10711 (make_pass_if_after_combine): ...new function.
10712 (pass_if_after_reload): Convert from a global struct to a subclass of
10713 rtl_opt_pass along with...
10714 (pass_data_if_after_reload): ...new pass_data instance and...
10715 (make_pass_if_after_reload): ...new function.
10716 * init-regs.c (pass_initialize_regs): Convert from a global struct to
10717 a subclass of rtl_opt_pass along with...
10718 (pass_data_initialize_regs): ...new pass_data instance and...
10719 (make_pass_initialize_regs): ...new function.
10720 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
10721 of ipa_opt_pass_d along with...
10722 (pass_data_ipa_cp): ...new pass_data instance and...
10723 (make_pass_ipa_cp): ...new function.
10724 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
10725 global struct to a subclass of gimple_opt_pass along with...
10726 (pass_data_inline_parameters): ...new pass_data instance and...
10727 (make_pass_inline_parameters): ...new function.
10728 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
10729 subclass of gimple_opt_pass along with...
10730 (pass_data_early_inline): ...new pass_data instance and...
10731 (make_pass_early_inline): ...new function.
10732 (pass_ipa_inline): Convert from a global struct to a subclass of
10733 ipa_opt_pass_d along with...
10734 (pass_data_ipa_inline): ...new pass_data instance and...
10735 (make_pass_ipa_inline): ...new function.
10736 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
10737 struct to a subclass of gimple_opt_pass along with...
10738 (pass_data_local_pure_const): ...new pass_data instance and...
10739 (make_pass_local_pure_const): ...new function.
10740 (pass_ipa_pure_const): Convert from a global struct to a subclass of
10741 ipa_opt_pass_d along with...
10742 (pass_data_ipa_pure_const): ...new pass_data instance and...
10743 (make_pass_ipa_pure_const): ...new function.
10744 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
10745 to a subclass of ipa_opt_pass_d along with...
10746 (pass_data_ipa_reference): ...new pass_data instance and...
10747 (make_pass_ipa_reference): ...new function.
10748 * ipa-split.c (pass_split_functions): Convert from a global struct to
10749 a subclass of gimple_opt_pass along with...
10750 (pass_data_split_functions): ...new pass_data instance and...
10751 (make_pass_split_functions): ...new function.
10752 (pass_feedback_split_functions): Convert from a global struct to a
10753 subclass of gimple_opt_pass along with...
10754 (pass_data_feedback_split_functions): ...new pass_data instance and...
10755 (make_pass_feedback_split_functions): ...new function.
10756 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
10757 global struct to a subclass of simple_ipa_opt_pass along with...
10758 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
10760 (make_pass_ipa_function_and_variable_visibility): ...new function.
10761 (pass_ipa_free_inline_summary): Convert from a global struct to a
10762 subclass of simple_ipa_opt_pass along with...
10763 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
10764 (make_pass_ipa_free_inline_summary): ...new function.
10765 (pass_ipa_whole_program_visibility): Convert from a global struct to a
10766 subclass of ipa_opt_pass_d along with...
10767 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
10769 (make_pass_ipa_whole_program_visibility): ...new function.
10770 (pass_ipa_profile): Convert from a global struct to a subclass of
10771 ipa_opt_pass_d along with...
10772 (pass_data_ipa_profile): ...new pass_data instance and...
10773 (make_pass_ipa_profile): ...new function.
10774 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
10775 ipa_opt_pass_d along with...
10776 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
10777 (make_pass_ipa_cdtor_merge): ...new function.
10778 * ira.c (pass_ira): Convert from a global struct to a subclass of
10779 rtl_opt_pass along with...
10780 (pass_data_ira): ...new pass_data instance and...
10781 (make_pass_ira): ...new function.
10782 (pass_reload): Convert from a global struct to a subclass of
10783 rtl_opt_pass along with...
10784 (pass_data_reload): ...new pass_data instance and...
10785 (make_pass_reload): ...new function.
10786 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
10787 subclass of rtl_opt_pass along with...
10788 (pass_data_cleanup_barriers): ...new pass_data instance and...
10789 (make_pass_cleanup_barriers): ...new function.
10790 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
10791 of rtl_opt_pass along with...
10792 (pass_data_loop2): ...new pass_data instance and...
10793 (make_pass_loop2): ...new function.
10794 (pass_rtl_loop_init): Convert from a global struct to a subclass of
10795 rtl_opt_pass along with...
10796 (pass_data_rtl_loop_init): ...new pass_data instance and...
10797 (make_pass_rtl_loop_init): ...new function.
10798 (pass_rtl_loop_done): Convert from a global struct to a subclass of
10799 rtl_opt_pass along with...
10800 (pass_data_rtl_loop_done): ...new pass_data instance and...
10801 (make_pass_rtl_loop_done): ...new function.
10802 (pass_rtl_move_loop_invariants): Convert from a global struct to a
10803 subclass of rtl_opt_pass along with...
10804 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
10805 (make_pass_rtl_move_loop_invariants): ...new function.
10806 (pass_rtl_unswitch): Convert from a global struct to a subclass of
10807 rtl_opt_pass along with...
10808 (pass_data_rtl_unswitch): ...new pass_data instance and...
10809 (make_pass_rtl_unswitch): ...new function.
10810 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
10811 subclass of rtl_opt_pass along with...
10812 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
10814 (make_pass_rtl_unroll_and_peel_loops): ...new function.
10815 (pass_rtl_doloop): Convert from a global struct to a subclass of
10816 rtl_opt_pass along with...
10817 (pass_data_rtl_doloop): ...new pass_data instance and...
10818 (make_pass_rtl_doloop): ...new function.
10819 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
10820 a subclass of rtl_opt_pass along with...
10821 (pass_data_lower_subreg): ...new pass_data instance and...
10822 (make_pass_lower_subreg): ...new function.
10823 (pass_lower_subreg2): Convert from a global struct to a subclass of
10824 rtl_opt_pass along with...
10825 (pass_data_lower_subreg2): ...new pass_data instance and...
10826 (make_pass_lower_subreg2): ...new function.
10827 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
10828 struct to a subclass of ipa_opt_pass_d along with...
10829 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
10830 (make_pass_ipa_lto_gimple_out): ...new function.
10831 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
10832 of ipa_opt_pass_d along with...
10833 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
10834 (make_pass_ipa_lto_finish_out): ...new function.
10835 * mode-switching.c (pass_mode_switching): Convert from a global struct
10836 to a subclass of rtl_opt_pass along with...
10837 (pass_data_mode_switching): ...new pass_data instance and...
10838 (make_pass_mode_switching): ...new function.
10839 * modulo-sched.c (pass_sms): Convert from a global struct to a
10840 subclass of rtl_opt_pass along with...
10841 (pass_data_sms): ...new pass_data instance and...
10842 (make_pass_sms): ...new function.
10843 * omp-low.c (pass_expand_omp): Convert from a global struct to a
10844 subclass of gimple_opt_pass along with...
10845 (pass_data_expand_omp): ...new pass_data instance and...
10846 (make_pass_expand_omp): ...new function.
10847 (pass_lower_omp): Convert from a global struct to a subclass of
10848 gimple_opt_pass along with...
10849 (pass_data_lower_omp): ...new pass_data instance and...
10850 (make_pass_lower_omp): ...new function.
10851 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
10852 of gimple_opt_pass along with...
10853 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
10854 (make_pass_diagnose_omp_blocks): ...new function.
10855 * passes.c (pass_early_local_passes): Convert from a global struct to
10856 a subclass of simple_ipa_opt_pass along with...
10857 (pass_data_early_local_passes): ...new pass_data instance and...
10858 (make_pass_early_local_passes): ...new function.
10859 (pass_all_early_optimizations): Convert from a global struct to a
10860 subclass of gimple_opt_pass along with...
10861 (pass_data_all_early_optimizations): ...new pass_data instance and...
10862 (make_pass_all_early_optimizations): ...new function.
10863 (pass_all_optimizations): Convert from a global struct to a subclass
10864 of gimple_opt_pass along with...
10865 (pass_data_all_optimizations): ...new pass_data instance and...
10866 (make_pass_all_optimizations): ...new function.
10867 (pass_all_optimizations_g): Convert from a global struct to a subclass
10868 of gimple_opt_pass along with...
10869 (pass_data_all_optimizations_g): ...new pass_data instance and...
10870 (make_pass_all_optimizations_g): ...new function.
10871 (pass_rest_of_compilation): Convert from a global struct to a subclass
10872 of rtl_opt_pass along with...
10873 (pass_data_rest_of_compilation): ...new pass_data instance and...
10874 (make_pass_rest_of_compilation): ...new function.
10875 (pass_postreload): Convert from a global struct to a subclass of
10876 rtl_opt_pass along with...
10877 (pass_data_postreload): ...new pass_data instance and...
10878 (make_pass_postreload): ...new function.
10879 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
10880 subclass of rtl_opt_pass along with...
10881 (pass_data_gcse2): ...new pass_data instance and...
10882 (make_pass_gcse2): ...new function.
10883 * postreload.c (pass_postreload_cse): Convert from a global struct to
10884 a subclass of rtl_opt_pass along with...
10885 (pass_data_postreload_cse): ...new pass_data instance and...
10886 (make_pass_postreload_cse): ...new function.
10887 * predict.c (pass_profile): Convert from a global struct to a subclass
10888 of gimple_opt_pass along with...
10889 (pass_data_profile): ...new pass_data instance and...
10890 (make_pass_profile): ...new function.
10891 (pass_strip_predict_hints): Convert from a global struct to a subclass
10892 of gimple_opt_pass along with...
10893 (pass_data_strip_predict_hints): ...new pass_data instance and...
10894 (make_pass_strip_predict_hints): ...new function.
10895 * recog.c (pass_peephole2): Convert from a global struct to a subclass
10896 of rtl_opt_pass along with...
10897 (pass_data_peephole2): ...new pass_data instance and...
10898 (make_pass_peephole2): ...new function.
10899 (pass_split_all_insns): Convert from a global struct to a subclass of
10900 rtl_opt_pass along with...
10901 (pass_data_split_all_insns): ...new pass_data instance and...
10902 (make_pass_split_all_insns): ...new function.
10903 (pass_split_after_reload): Convert from a global struct to a subclass
10904 of rtl_opt_pass along with...
10905 (pass_data_split_after_reload): ...new pass_data instance and...
10906 (make_pass_split_after_reload): ...new function.
10907 (pass_split_before_regstack): Convert from a global struct to a
10908 subclass of rtl_opt_pass along with...
10909 (pass_data_split_before_regstack): ...new pass_data instance and...
10910 (make_pass_split_before_regstack): ...new function.
10911 (pass_split_before_sched2): Convert from a global struct to a subclass
10912 of rtl_opt_pass along with...
10913 (pass_data_split_before_sched2): ...new pass_data instance and...
10914 (make_pass_split_before_sched2): ...new function.
10915 (pass_split_for_shorten_branches): Convert from a global struct to a
10916 subclass of rtl_opt_pass along with...
10917 (pass_data_split_for_shorten_branches): ...new pass_data instance
10919 (make_pass_split_for_shorten_branches): ...new function.
10920 * ree.c (pass_ree): Convert from a global struct to a subclass of
10921 rtl_opt_pass along with...
10922 (pass_data_ree): ...new pass_data instance and...
10923 (make_pass_ree): ...new function.
10924 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
10925 subclass of rtl_opt_pass along with...
10926 (pass_data_stack_regs): ...new pass_data instance and...
10927 (make_pass_stack_regs): ...new function.
10928 (pass_stack_regs_run): Convert from a global struct to a subclass of
10929 rtl_opt_pass along with...
10930 (pass_data_stack_regs_run): ...new pass_data instance and...
10931 (make_pass_stack_regs_run): ...new function.
10932 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
10933 subclass of rtl_opt_pass along with...
10934 (pass_data_cprop_hardreg): ...new pass_data instance and...
10935 (make_pass_cprop_hardreg): ...new function.
10936 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
10937 subclass of rtl_opt_pass along with...
10938 (pass_data_reginfo_init): ...new pass_data instance and...
10939 (make_pass_reginfo_init): ...new function.
10940 * regmove.c (pass_regmove): Convert from a global struct to a subclass
10941 of rtl_opt_pass along with...
10942 (pass_data_regmove): ...new pass_data instance and...
10943 (make_pass_regmove): ...new function.
10944 * regrename.c (pass_regrename): Convert from a global struct to a
10945 subclass of rtl_opt_pass along with...
10946 (pass_data_regrename): ...new pass_data instance and...
10947 (make_pass_regrename): ...new function.
10948 * reorg.c (pass_delay_slots): Convert from a global struct to a
10949 subclass of rtl_opt_pass along with...
10950 (pass_data_delay_slots): ...new pass_data instance and...
10951 (make_pass_delay_slots): ...new function.
10952 (pass_machine_reorg): Convert from a global struct to a subclass of
10953 rtl_opt_pass along with...
10954 (pass_data_machine_reorg): ...new pass_data instance and...
10955 (make_pass_machine_reorg): ...new function.
10956 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
10957 of rtl_opt_pass along with...
10958 (pass_data_sched): ...new pass_data instance and...
10959 (make_pass_sched): ...new function.
10960 (pass_sched2): Convert from a global struct to a subclass of
10961 rtl_opt_pass along with...
10962 (pass_data_sched2): ...new pass_data instance and...
10963 (make_pass_sched2): ...new function.
10964 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
10965 to a subclass of rtl_opt_pass along with...
10966 (pass_data_stack_ptr_mod): ...new pass_data instance and...
10967 (make_pass_stack_ptr_mod): ...new function.
10968 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
10969 to a subclass of rtl_opt_pass along with...
10970 (pass_data_rtl_store_motion): ...new pass_data instance and...
10971 (make_pass_rtl_store_motion): ...new function.
10972 * tracer.c (pass_tracer): Convert from a global struct to a subclass
10973 of gimple_opt_pass along with...
10974 (pass_data_tracer): ...new pass_data instance and...
10975 (make_pass_tracer): ...new function.
10976 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
10977 to a subclass of gimple_opt_pass along with...
10978 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
10979 (make_pass_diagnose_tm_blocks): ...new function.
10980 (pass_lower_tm): Convert from a global struct to a subclass of
10981 gimple_opt_pass along with...
10982 (pass_data_lower_tm): ...new pass_data instance and...
10983 (make_pass_lower_tm): ...new function.
10984 (pass_tm_init): Convert from a global struct to a subclass of
10985 gimple_opt_pass along with...
10986 (pass_data_tm_init): ...new pass_data instance and...
10987 (make_pass_tm_init): ...new function.
10988 (pass_tm_mark): Convert from a global struct to a subclass of
10989 gimple_opt_pass along with...
10990 (pass_data_tm_mark): ...new pass_data instance and...
10991 (make_pass_tm_mark): ...new function.
10992 (pass_tm_edges): Convert from a global struct to a subclass of
10993 gimple_opt_pass along with...
10994 (pass_data_tm_edges): ...new pass_data instance and...
10995 (make_pass_tm_edges): ...new function.
10996 (pass_tm_memopt): Convert from a global struct to a subclass of
10997 gimple_opt_pass along with...
10998 (pass_data_tm_memopt): ...new pass_data instance and...
10999 (make_pass_tm_memopt): ...new function.
11000 (pass_ipa_tm): Convert from a global struct to a subclass of
11001 simple_ipa_opt_pass along with...
11002 (pass_data_ipa_tm): ...new pass_data instance and...
11003 (make_pass_ipa_tm): ...new function.
11004 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
11005 subclass of gimple_opt_pass along with...
11006 (pass_data_call_cdce): ...new pass_data instance and...
11007 (make_pass_call_cdce): ...new function.
11008 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
11009 subclass of gimple_opt_pass along with...
11010 (pass_data_build_cfg): ...new pass_data instance and...
11011 (make_pass_build_cfg): ...new function.
11012 (pass_split_crit_edges): Convert from a global struct to a subclass of
11013 gimple_opt_pass along with...
11014 (pass_data_split_crit_edges): ...new pass_data instance and...
11015 (make_pass_split_crit_edges): ...new function.
11016 (pass_warn_function_return): Convert from a global struct to a
11017 subclass of gimple_opt_pass along with...
11018 (pass_data_warn_function_return): ...new pass_data instance and...
11019 (make_pass_warn_function_return): ...new function.
11020 (pass_warn_function_noreturn): Convert from a global struct to a
11021 subclass of gimple_opt_pass along with...
11022 (pass_data_warn_function_noreturn): ...new pass_data instance and...
11023 (make_pass_warn_function_noreturn): ...new function.
11024 (pass_warn_unused_result): Convert from a global struct to a subclass
11025 of gimple_opt_pass along with...
11026 (pass_data_warn_unused_result): ...new pass_data instance and...
11027 (make_pass_warn_unused_result): ...new function.
11028 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
11029 a subclass of gimple_opt_pass along with...
11030 (pass_data_merge_phi): ...new pass_data instance and...
11031 (make_pass_merge_phi): ...new function.
11032 * tree-complex.c (pass_lower_complex): Convert from a global struct to
11033 a subclass of gimple_opt_pass along with...
11034 (pass_data_lower_complex): ...new pass_data instance and...
11035 (make_pass_lower_complex): ...new function.
11036 (pass_lower_complex_O0): Convert from a global struct to a subclass of
11037 gimple_opt_pass along with...
11038 (pass_data_lower_complex_O0): ...new pass_data instance and...
11039 (make_pass_lower_complex_O0): ...new function.
11040 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
11041 subclass of gimple_opt_pass along with...
11042 (pass_data_lower_eh): ...new pass_data instance and...
11043 (make_pass_lower_eh): ...new function.
11044 (pass_refactor_eh): Convert from a global struct to a subclass of
11045 gimple_opt_pass along with...
11046 (pass_data_refactor_eh): ...new pass_data instance and...
11047 (make_pass_refactor_eh): ...new function.
11048 (pass_lower_resx): Convert from a global struct to a subclass of
11049 gimple_opt_pass along with...
11050 (pass_data_lower_resx): ...new pass_data instance and...
11051 (make_pass_lower_resx): ...new function.
11052 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
11053 of gimple_opt_pass along with...
11054 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
11055 (make_pass_lower_eh_dispatch): ...new function.
11056 (pass_cleanup_eh): Convert from a global struct to a subclass of
11057 gimple_opt_pass along with...
11058 (pass_data_cleanup_eh): ...new pass_data instance and...
11059 (make_pass_cleanup_eh): ...new function.
11060 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
11061 to a subclass of simple_ipa_opt_pass along with...
11062 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
11063 (make_pass_ipa_lower_emutls): ...new function.
11064 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
11065 a subclass of gimple_opt_pass along with...
11066 (pass_data_if_conversion): ...new pass_data instance and...
11067 (make_pass_if_conversion): ...new function.
11068 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
11069 subclass of gimple_opt_pass along with...
11070 (pass_data_build_ssa): ...new pass_data instance and...
11071 (make_pass_build_ssa): ...new function.
11072 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
11073 global struct to a subclass of gimple_opt_pass along with...
11074 (pass_data_loop_distribution): ...new pass_data instance and...
11075 (make_pass_loop_distribution): ...new function.
11076 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
11077 subclass of gimple_opt_pass along with...
11078 (pass_data_mudflap_1): ...new pass_data instance and...
11079 (make_pass_mudflap_1): ...new function.
11080 (pass_mudflap_2): Convert from a global struct to a subclass of
11081 gimple_opt_pass along with...
11082 (pass_data_mudflap_2): ...new pass_data instance and...
11083 (make_pass_mudflap_2): ...new function.
11084 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
11085 subclass of gimple_opt_pass along with...
11086 (pass_data_mudflap_1): ...new pass_data instance and...
11087 (make_pass_mudflap_1): ...new function.
11088 (pass_mudflap_2): Convert from a global struct to a subclass of
11089 gimple_opt_pass along with...
11090 (pass_data_mudflap_2): ...new pass_data instance and...
11091 (make_pass_mudflap_2): ...new function.
11092 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
11093 gimple_opt_pass along with...
11094 (pass_data_nrv): ...new pass_data instance and...
11095 (make_pass_nrv): ...new function.
11096 (pass_return_slot): Convert from a global struct to a subclass of
11097 gimple_opt_pass along with...
11098 (pass_data_return_slot): ...new pass_data instance and...
11099 (make_pass_return_slot): ...new function.
11100 * tree-object-size.c (pass_object_sizes): Convert from a global struct
11101 to a subclass of gimple_opt_pass along with...
11102 (pass_data_object_sizes): ...new pass_data instance and...
11103 (make_pass_object_sizes): ...new function.
11104 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
11105 global struct to a subclass of gimple_opt_pass along with...
11106 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
11108 (make_pass_cleanup_cfg_post_optimizing): ...new function.
11109 (pass_fixup_cfg): Convert from a global struct to a subclass of
11110 gimple_opt_pass along with...
11111 (pass_data_fixup_cfg): ...new pass_data instance and...
11112 (make_pass_fixup_cfg): ...new function.
11113 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
11114 (make_pass_mudflap_1): ...new function.
11115 (pass_mudflap_2): Replace declaration with that of...
11116 (make_pass_mudflap_2): ...new function.
11117 (pass_asan): Replace declaration with that of...
11118 (make_pass_asan): ...new function.
11119 (pass_asan_O0): Replace declaration with that of...
11120 (make_pass_asan_O0): ...new function.
11121 (pass_tsan): Replace declaration with that of...
11122 (make_pass_tsan): ...new function.
11123 (pass_tsan_O0): Replace declaration with that of...
11124 (make_pass_tsan_O0): ...new function.
11125 (pass_lower_cf): Replace declaration with that of...
11126 (make_pass_lower_cf): ...new function.
11127 (pass_refactor_eh): Replace declaration with that of...
11128 (make_pass_refactor_eh): ...new function.
11129 (pass_lower_eh): Replace declaration with that of...
11130 (make_pass_lower_eh): ...new function.
11131 (pass_lower_eh_dispatch): Replace declaration with that of...
11132 (make_pass_lower_eh_dispatch): ...new function.
11133 (pass_lower_resx): Replace declaration with that of...
11134 (make_pass_lower_resx): ...new function.
11135 (pass_build_cfg): Replace declaration with that of...
11136 (make_pass_build_cfg): ...new function.
11137 (pass_early_tree_profile): Replace declaration with that of...
11138 (make_pass_early_tree_profile): ...new function.
11139 (pass_cleanup_eh): Replace declaration with that of...
11140 (make_pass_cleanup_eh): ...new function.
11141 (pass_sra): Replace declaration with that of...
11142 (make_pass_sra): ...new function.
11143 (pass_sra_early): Replace declaration with that of...
11144 (make_pass_sra_early): ...new function.
11145 (pass_early_ipa_sra): Replace declaration with that of...
11146 (make_pass_early_ipa_sra): ...new function.
11147 (pass_tail_recursion): Replace declaration with that of...
11148 (make_pass_tail_recursion): ...new function.
11149 (pass_tail_calls): Replace declaration with that of...
11150 (make_pass_tail_calls): ...new function.
11151 (pass_tree_loop): Replace declaration with that of...
11152 (make_pass_tree_loop): ...new function.
11153 (pass_tree_loop_init): Replace declaration with that of...
11154 (make_pass_tree_loop_init): ...new function.
11155 (pass_lim): Replace declaration with that of...
11156 (make_pass_lim): ...new function.
11157 (pass_tree_unswitch): Replace declaration with that of...
11158 (make_pass_tree_unswitch): ...new function.
11159 (pass_predcom): Replace declaration with that of...
11160 (make_pass_predcom): ...new function.
11161 (pass_iv_canon): Replace declaration with that of...
11162 (make_pass_iv_canon): ...new function.
11163 (pass_scev_cprop): Replace declaration with that of...
11164 (make_pass_scev_cprop): ...new function.
11165 (pass_empty_loop): Replace declaration with that of...
11166 (make_pass_empty_loop): ...new function.
11167 (pass_record_bounds): Replace declaration with that of...
11168 (make_pass_record_bounds): ...new function.
11169 (pass_graphite): Replace declaration with that of...
11170 (make_pass_graphite): ...new function.
11171 (pass_graphite_transforms): Replace declaration with that of...
11172 (make_pass_graphite_transforms): ...new function.
11173 (pass_if_conversion): Replace declaration with that of...
11174 (make_pass_if_conversion): ...new function.
11175 (pass_loop_distribution): Replace declaration with that of...
11176 (make_pass_loop_distribution): ...new function.
11177 (pass_vectorize): Replace declaration with that of...
11178 (make_pass_vectorize): ...new function.
11179 (pass_slp_vectorize): Replace declaration with that of...
11180 (make_pass_slp_vectorize): ...new function.
11181 (pass_complete_unroll): Replace declaration with that of...
11182 (make_pass_complete_unroll): ...new function.
11183 (pass_complete_unrolli): Replace declaration with that of...
11184 (make_pass_complete_unrolli): ...new function.
11185 (pass_parallelize_loops): Replace declaration with that of...
11186 (make_pass_parallelize_loops): ...new function.
11187 (pass_loop_prefetch): Replace declaration with that of...
11188 (make_pass_loop_prefetch): ...new function.
11189 (pass_iv_optimize): Replace declaration with that of...
11190 (make_pass_iv_optimize): ...new function.
11191 (pass_tree_loop_done): Replace declaration with that of...
11192 (make_pass_tree_loop_done): ...new function.
11193 (pass_ch): Replace declaration with that of...
11194 (make_pass_ch): ...new function.
11195 (pass_ccp): Replace declaration with that of...
11196 (make_pass_ccp): ...new function.
11197 (pass_phi_only_cprop): Replace declaration with that of...
11198 (make_pass_phi_only_cprop): ...new function.
11199 (pass_build_ssa): Replace declaration with that of...
11200 (make_pass_build_ssa): ...new function.
11201 (pass_build_alias): Replace declaration with that of...
11202 (make_pass_build_alias): ...new function.
11203 (pass_build_ealias): Replace declaration with that of...
11204 (make_pass_build_ealias): ...new function.
11205 (pass_dominator): Replace declaration with that of...
11206 (make_pass_dominator): ...new function.
11207 (pass_dce): Replace declaration with that of...
11208 (make_pass_dce): ...new function.
11209 (pass_dce_loop): Replace declaration with that of...
11210 (make_pass_dce_loop): ...new function.
11211 (pass_cd_dce): Replace declaration with that of...
11212 (make_pass_cd_dce): ...new function.
11213 (pass_call_cdce): Replace declaration with that of...
11214 (make_pass_call_cdce): ...new function.
11215 (pass_merge_phi): Replace declaration with that of...
11216 (make_pass_merge_phi): ...new function.
11217 (pass_split_crit_edges): Replace declaration with that of...
11218 (make_pass_split_crit_edges): ...new function.
11219 (pass_pre): Replace declaration with that of...
11220 (make_pass_pre): ...new function.
11221 (pass_profile): Replace declaration with that of...
11222 (make_pass_profile): ...new function.
11223 (pass_strip_predict_hints): Replace declaration with that of...
11224 (make_pass_strip_predict_hints): ...new function.
11225 (pass_lower_complex_O0): Replace declaration with that of...
11226 (make_pass_lower_complex_O0): ...new function.
11227 (pass_lower_complex): Replace declaration with that of...
11228 (make_pass_lower_complex): ...new function.
11229 (pass_lower_vector): Replace declaration with that of...
11230 (make_pass_lower_vector): ...new function.
11231 (pass_lower_vector_ssa): Replace declaration with that of...
11232 (make_pass_lower_vector_ssa): ...new function.
11233 (pass_lower_omp): Replace declaration with that of...
11234 (make_pass_lower_omp): ...new function.
11235 (pass_diagnose_omp_blocks): Replace declaration with that of...
11236 (make_pass_diagnose_omp_blocks): ...new function.
11237 (pass_expand_omp): Replace declaration with that of...
11238 (make_pass_expand_omp): ...new function.
11239 (pass_expand_omp_ssa): Replace declaration with that of...
11240 (make_pass_expand_omp_ssa): ...new function.
11241 (pass_object_sizes): Replace declaration with that of...
11242 (make_pass_object_sizes): ...new function.
11243 (pass_strlen): Replace declaration with that of...
11244 (make_pass_strlen): ...new function.
11245 (pass_fold_builtins): Replace declaration with that of...
11246 (make_pass_fold_builtins): ...new function.
11247 (pass_stdarg): Replace declaration with that of...
11248 (make_pass_stdarg): ...new function.
11249 (pass_early_warn_uninitialized): Replace declaration with that of...
11250 (make_pass_early_warn_uninitialized): ...new function.
11251 (pass_late_warn_uninitialized): Replace declaration with that of...
11252 (make_pass_late_warn_uninitialized): ...new function.
11253 (pass_cse_reciprocals): Replace declaration with that of...
11254 (make_pass_cse_reciprocals): ...new function.
11255 (pass_cse_sincos): Replace declaration with that of...
11256 (make_pass_cse_sincos): ...new function.
11257 (pass_optimize_bswap): Replace declaration with that of...
11258 (make_pass_optimize_bswap): ...new function.
11259 (pass_optimize_widening_mul): Replace declaration with that of...
11260 (make_pass_optimize_widening_mul): ...new function.
11261 (pass_warn_function_return): Replace declaration with that of...
11262 (make_pass_warn_function_return): ...new function.
11263 (pass_warn_function_noreturn): Replace declaration with that of...
11264 (make_pass_warn_function_noreturn): ...new function.
11265 (pass_cselim): Replace declaration with that of...
11266 (make_pass_cselim): ...new function.
11267 (pass_phiopt): Replace declaration with that of...
11268 (make_pass_phiopt): ...new function.
11269 (pass_forwprop): Replace declaration with that of...
11270 (make_pass_forwprop): ...new function.
11271 (pass_phiprop): Replace declaration with that of...
11272 (make_pass_phiprop): ...new function.
11273 (pass_tree_ifcombine): Replace declaration with that of...
11274 (make_pass_tree_ifcombine): ...new function.
11275 (pass_dse): Replace declaration with that of...
11276 (make_pass_dse): ...new function.
11277 (pass_nrv): Replace declaration with that of...
11278 (make_pass_nrv): ...new function.
11279 (pass_rename_ssa_copies): Replace declaration with that of...
11280 (make_pass_rename_ssa_copies): ...new function.
11281 (pass_sink_code): Replace declaration with that of...
11282 (make_pass_sink_code): ...new function.
11283 (pass_fre): Replace declaration with that of...
11284 (make_pass_fre): ...new function.
11285 (pass_check_data_deps): Replace declaration with that of...
11286 (make_pass_check_data_deps): ...new function.
11287 (pass_copy_prop): Replace declaration with that of...
11288 (make_pass_copy_prop): ...new function.
11289 (pass_vrp): Replace declaration with that of...
11290 (make_pass_vrp): ...new function.
11291 (pass_uncprop): Replace declaration with that of...
11292 (make_pass_uncprop): ...new function.
11293 (pass_return_slot): Replace declaration with that of...
11294 (make_pass_return_slot): ...new function.
11295 (pass_reassoc): Replace declaration with that of...
11296 (make_pass_reassoc): ...new function.
11297 (pass_rebuild_cgraph_edges): Replace declaration with that of...
11298 (make_pass_rebuild_cgraph_edges): ...new function.
11299 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
11300 (make_pass_remove_cgraph_callee_edges): ...new function.
11301 (pass_build_cgraph_edges): Replace declaration with that of...
11302 (make_pass_build_cgraph_edges): ...new function.
11303 (pass_local_pure_const): Replace declaration with that of...
11304 (make_pass_local_pure_const): ...new function.
11305 (pass_tracer): Replace declaration with that of...
11306 (make_pass_tracer): ...new function.
11307 (pass_warn_unused_result): Replace declaration with that of...
11308 (make_pass_warn_unused_result): ...new function.
11309 (pass_diagnose_tm_blocks): Replace declaration with that of...
11310 (make_pass_diagnose_tm_blocks): ...new function.
11311 (pass_lower_tm): Replace declaration with that of...
11312 (make_pass_lower_tm): ...new function.
11313 (pass_tm_init): Replace declaration with that of...
11314 (make_pass_tm_init): ...new function.
11315 (pass_tm_mark): Replace declaration with that of...
11316 (make_pass_tm_mark): ...new function.
11317 (pass_tm_memopt): Replace declaration with that of...
11318 (make_pass_tm_memopt): ...new function.
11319 (pass_tm_edges): Replace declaration with that of...
11320 (make_pass_tm_edges): ...new function.
11321 (pass_split_functions): Replace declaration with that of...
11322 (make_pass_split_functions): ...new function.
11323 (pass_feedback_split_functions): Replace declaration with that of...
11324 (make_pass_feedback_split_functions): ...new function.
11325 (pass_strength_reduction): Replace declaration with that of...
11326 (make_pass_strength_reduction): ...new function.
11327 (pass_ipa_lower_emutls): Replace declaration with that of...
11328 (make_pass_ipa_lower_emutls): ...new function.
11329 (pass_ipa_function_and_variable_visibility): Replace declaration with
11331 (make_pass_ipa_function_and_variable_visibility): ...new function.
11332 (pass_ipa_tree_profile): Replace declaration with that of...
11333 (make_pass_ipa_tree_profile): ...new function.
11334 (pass_early_local_passes): Replace declaration with that of...
11335 (make_pass_early_local_passes): ...new function.
11336 (pass_ipa_whole_program_visibility): Replace declaration with that
11338 (make_pass_ipa_whole_program_visibility): ...new function.
11339 (pass_ipa_lto_gimple_out): Replace declaration with that of...
11340 (make_pass_ipa_lto_gimple_out): ...new function.
11341 (pass_ipa_increase_alignment): Replace declaration with that of...
11342 (make_pass_ipa_increase_alignment): ...new function.
11343 (pass_ipa_inline): Replace declaration with that of...
11344 (make_pass_ipa_inline): ...new function.
11345 (pass_ipa_free_lang_data): Replace declaration with that of...
11346 (make_pass_ipa_free_lang_data): ...new function.
11347 (pass_ipa_free_inline_summary): Replace declaration with that of...
11348 (make_pass_ipa_free_inline_summary): ...new function.
11349 (pass_ipa_cp): Replace declaration with that of...
11350 (make_pass_ipa_cp): ...new function.
11351 (pass_ipa_reference): Replace declaration with that of...
11352 (make_pass_ipa_reference): ...new function.
11353 (pass_ipa_pure_const): Replace declaration with that of...
11354 (make_pass_ipa_pure_const): ...new function.
11355 (pass_ipa_pta): Replace declaration with that of...
11356 (make_pass_ipa_pta): ...new function.
11357 (pass_ipa_lto_finish_out): Replace declaration with that of...
11358 (make_pass_ipa_lto_finish_out): ...new function.
11359 (pass_ipa_tm): Replace declaration with that of...
11360 (make_pass_ipa_tm): ...new function.
11361 (pass_ipa_profile): Replace declaration with that of...
11362 (make_pass_ipa_profile): ...new function.
11363 (pass_ipa_cdtor_merge): Replace declaration with that of...
11364 (make_pass_ipa_cdtor_merge): ...new function.
11365 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
11367 (make_pass_cleanup_cfg_post_optimizing): ...new function.
11368 (pass_init_datastructures): Replace declaration with that of...
11369 (make_pass_init_datastructures): ...new function.
11370 (pass_fixup_cfg): Replace declaration with that of...
11371 (make_pass_fixup_cfg): ...new function.
11372 (pass_expand): Replace declaration with that of...
11373 (make_pass_expand): ...new function.
11374 (pass_instantiate_virtual_regs): Replace declaration with that of...
11375 (make_pass_instantiate_virtual_regs): ...new function.
11376 (pass_rtl_fwprop): Replace declaration with that of...
11377 (make_pass_rtl_fwprop): ...new function.
11378 (pass_rtl_fwprop_addr): Replace declaration with that of...
11379 (make_pass_rtl_fwprop_addr): ...new function.
11380 (pass_jump): Replace declaration with that of...
11381 (make_pass_jump): ...new function.
11382 (pass_jump2): Replace declaration with that of...
11383 (make_pass_jump2): ...new function.
11384 (pass_lower_subreg): Replace declaration with that of...
11385 (make_pass_lower_subreg): ...new function.
11386 (pass_cse): Replace declaration with that of...
11387 (make_pass_cse): ...new function.
11388 (pass_fast_rtl_dce): Replace declaration with that of...
11389 (make_pass_fast_rtl_dce): ...new function.
11390 (pass_ud_rtl_dce): Replace declaration with that of...
11391 (make_pass_ud_rtl_dce): ...new function.
11392 (pass_rtl_dce): Replace declaration with that of...
11393 (make_pass_rtl_dce): ...new function.
11394 (pass_rtl_dse1): Replace declaration with that of...
11395 (make_pass_rtl_dse1): ...new function.
11396 (pass_rtl_dse2): Replace declaration with that of...
11397 (make_pass_rtl_dse2): ...new function.
11398 (pass_rtl_dse3): Replace declaration with that of...
11399 (make_pass_rtl_dse3): ...new function.
11400 (pass_rtl_cprop): Replace declaration with that of...
11401 (make_pass_rtl_cprop): ...new function.
11402 (pass_rtl_pre): Replace declaration with that of...
11403 (make_pass_rtl_pre): ...new function.
11404 (pass_rtl_hoist): Replace declaration with that of...
11405 (make_pass_rtl_hoist): ...new function.
11406 (pass_rtl_store_motion): Replace declaration with that of...
11407 (make_pass_rtl_store_motion): ...new function.
11408 (pass_cse_after_global_opts): Replace declaration with that of...
11409 (make_pass_cse_after_global_opts): ...new function.
11410 (pass_rtl_ifcvt): Replace declaration with that of...
11411 (make_pass_rtl_ifcvt): ...new function.
11412 (pass_into_cfg_layout_mode): Replace declaration with that of...
11413 (make_pass_into_cfg_layout_mode): ...new function.
11414 (pass_outof_cfg_layout_mode): Replace declaration with that of...
11415 (make_pass_outof_cfg_layout_mode): ...new function.
11416 (pass_loop2): Replace declaration with that of...
11417 (make_pass_loop2): ...new function.
11418 (pass_rtl_loop_init): Replace declaration with that of...
11419 (make_pass_rtl_loop_init): ...new function.
11420 (pass_rtl_move_loop_invariants): Replace declaration with that of...
11421 (make_pass_rtl_move_loop_invariants): ...new function.
11422 (pass_rtl_unswitch): Replace declaration with that of...
11423 (make_pass_rtl_unswitch): ...new function.
11424 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
11425 (make_pass_rtl_unroll_and_peel_loops): ...new function.
11426 (pass_rtl_doloop): Replace declaration with that of...
11427 (make_pass_rtl_doloop): ...new function.
11428 (pass_rtl_loop_done): Replace declaration with that of...
11429 (make_pass_rtl_loop_done): ...new function.
11430 (pass_web): Replace declaration with that of...
11431 (make_pass_web): ...new function.
11432 (pass_cse2): Replace declaration with that of...
11433 (make_pass_cse2): ...new function.
11434 (pass_df_initialize_opt): Replace declaration with that of...
11435 (make_pass_df_initialize_opt): ...new function.
11436 (pass_df_initialize_no_opt): Replace declaration with that of...
11437 (make_pass_df_initialize_no_opt): ...new function.
11438 (pass_reginfo_init): Replace declaration with that of...
11439 (make_pass_reginfo_init): ...new function.
11440 (pass_inc_dec): Replace declaration with that of...
11441 (make_pass_inc_dec): ...new function.
11442 (pass_stack_ptr_mod): Replace declaration with that of...
11443 (make_pass_stack_ptr_mod): ...new function.
11444 (pass_initialize_regs): Replace declaration with that of...
11445 (make_pass_initialize_regs): ...new function.
11446 (pass_combine): Replace declaration with that of...
11447 (make_pass_combine): ...new function.
11448 (pass_if_after_combine): Replace declaration with that of...
11449 (make_pass_if_after_combine): ...new function.
11450 (pass_ree): Replace declaration with that of...
11451 (make_pass_ree): ...new function.
11452 (pass_partition_blocks): Replace declaration with that of...
11453 (make_pass_partition_blocks): ...new function.
11454 (pass_match_asm_constraints): Replace declaration with that of...
11455 (make_pass_match_asm_constraints): ...new function.
11456 (pass_regmove): Replace declaration with that of...
11457 (make_pass_regmove): ...new function.
11458 (pass_split_all_insns): Replace declaration with that of...
11459 (make_pass_split_all_insns): ...new function.
11460 (pass_fast_rtl_byte_dce): Replace declaration with that of...
11461 (make_pass_fast_rtl_byte_dce): ...new function.
11462 (pass_lower_subreg2): Replace declaration with that of...
11463 (make_pass_lower_subreg2): ...new function.
11464 (pass_mode_switching): Replace declaration with that of...
11465 (make_pass_mode_switching): ...new function.
11466 (pass_sms): Replace declaration with that of...
11467 (make_pass_sms): ...new function.
11468 (pass_sched): Replace declaration with that of...
11469 (make_pass_sched): ...new function.
11470 (pass_ira): Replace declaration with that of...
11471 (make_pass_ira): ...new function.
11472 (pass_reload): Replace declaration with that of...
11473 (make_pass_reload): ...new function.
11474 (pass_clean_state): Replace declaration with that of...
11475 (make_pass_clean_state): ...new function.
11476 (pass_branch_prob): Replace declaration with that of...
11477 (make_pass_branch_prob): ...new function.
11478 (pass_value_profile_transformations): Replace declaration with that
11480 (make_pass_value_profile_transformations): ...new function.
11481 (pass_postreload_cse): Replace declaration with that of...
11482 (make_pass_postreload_cse): ...new function.
11483 (pass_gcse2): Replace declaration with that of...
11484 (make_pass_gcse2): ...new function.
11485 (pass_split_after_reload): Replace declaration with that of...
11486 (make_pass_split_after_reload): ...new function.
11487 (pass_branch_target_load_optimize1): Replace declaration with that
11489 (make_pass_branch_target_load_optimize1): ...new function.
11490 (pass_thread_prologue_and_epilogue): Replace declaration with that
11492 (make_pass_thread_prologue_and_epilogue): ...new function.
11493 (pass_stack_adjustments): Replace declaration with that of...
11494 (make_pass_stack_adjustments): ...new function.
11495 (pass_peephole2): Replace declaration with that of...
11496 (make_pass_peephole2): ...new function.
11497 (pass_if_after_reload): Replace declaration with that of...
11498 (make_pass_if_after_reload): ...new function.
11499 (pass_regrename): Replace declaration with that of...
11500 (make_pass_regrename): ...new function.
11501 (pass_cprop_hardreg): Replace declaration with that of...
11502 (make_pass_cprop_hardreg): ...new function.
11503 (pass_reorder_blocks): Replace declaration with that of...
11504 (make_pass_reorder_blocks): ...new function.
11505 (pass_branch_target_load_optimize2): Replace declaration with that
11507 (make_pass_branch_target_load_optimize2): ...new function.
11508 (pass_leaf_regs): Replace declaration with that of...
11509 (make_pass_leaf_regs): ...new function.
11510 (pass_split_before_sched2): Replace declaration with that of...
11511 (make_pass_split_before_sched2): ...new function.
11512 (pass_compare_elim_after_reload): Replace declaration with that of...
11513 (make_pass_compare_elim_after_reload): ...new function.
11514 (pass_sched2): Replace declaration with that of...
11515 (make_pass_sched2): ...new function.
11516 (pass_stack_regs): Replace declaration with that of...
11517 (make_pass_stack_regs): ...new function.
11518 (pass_stack_regs_run): Replace declaration with that of...
11519 (make_pass_stack_regs_run): ...new function.
11520 (pass_df_finish): Replace declaration with that of...
11521 (make_pass_df_finish): ...new function.
11522 (pass_compute_alignments): Replace declaration with that of...
11523 (make_pass_compute_alignments): ...new function.
11524 (pass_duplicate_computed_gotos): Replace declaration with that of...
11525 (make_pass_duplicate_computed_gotos): ...new function.
11526 (pass_variable_tracking): Replace declaration with that of...
11527 (make_pass_variable_tracking): ...new function.
11528 (pass_free_cfg): Replace declaration with that of...
11529 (make_pass_free_cfg): ...new function.
11530 (pass_machine_reorg): Replace declaration with that of...
11531 (make_pass_machine_reorg): ...new function.
11532 (pass_cleanup_barriers): Replace declaration with that of...
11533 (make_pass_cleanup_barriers): ...new function.
11534 (pass_delay_slots): Replace declaration with that of...
11535 (make_pass_delay_slots): ...new function.
11536 (pass_split_for_shorten_branches): Replace declaration with that of...
11537 (make_pass_split_for_shorten_branches): ...new function.
11538 (pass_split_before_regstack): Replace declaration with that of...
11539 (make_pass_split_before_regstack): ...new function.
11540 (pass_convert_to_eh_region_ranges): Replace declaration with that
11542 (make_pass_convert_to_eh_region_ranges): ...new function.
11543 (pass_shorten_branches): Replace declaration with that of...
11544 (make_pass_shorten_branches): ...new function.
11545 (pass_set_nothrow_function_flags): Replace declaration with that of...
11546 (make_pass_set_nothrow_function_flags): ...new function.
11547 (pass_dwarf2_frame): Replace declaration with that of...
11548 (make_pass_dwarf2_frame): ...new function.
11549 (pass_final): Replace declaration with that of...
11550 (make_pass_final): ...new function.
11551 (pass_rtl_seqabstr): Replace declaration with that of...
11552 (make_pass_rtl_seqabstr): ...new function.
11553 (pass_release_ssa_names): Replace declaration with that of...
11554 (make_pass_release_ssa_names): ...new function.
11555 (pass_early_inline): Replace declaration with that of...
11556 (make_pass_early_inline): ...new function.
11557 (pass_inline_parameters): Replace declaration with that of...
11558 (make_pass_inline_parameters): ...new function.
11559 (pass_update_address_taken): Replace declaration with that of...
11560 (make_pass_update_address_taken): ...new function.
11561 (pass_convert_switch): Replace declaration with that of...
11562 (make_pass_convert_switch): ...new function.
11563 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
11564 to a subclass of simple_ipa_opt_pass along with...
11565 (pass_data_ipa_tree_profile): ...new pass_data instance and...
11566 (make_pass_ipa_tree_profile): ...new function.
11567 * tree-sra.c (pass_sra_early): Convert from a global struct to a
11568 subclass of gimple_opt_pass along with...
11569 (pass_data_sra_early): ...new pass_data instance and...
11570 (make_pass_sra_early): ...new function.
11571 (pass_sra): Convert from a global struct to a subclass of
11572 gimple_opt_pass along with...
11573 (pass_data_sra): ...new pass_data instance and...
11574 (make_pass_sra): ...new function.
11575 (pass_early_ipa_sra): Convert from a global struct to a subclass of
11576 gimple_opt_pass along with...
11577 (pass_data_early_ipa_sra): ...new pass_data instance and...
11578 (make_pass_early_ipa_sra): ...new function.
11579 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
11580 subclass of gimple_opt_pass along with...
11581 (pass_data_ccp): ...new pass_data instance and...
11582 (make_pass_ccp): ...new function.
11583 (pass_fold_builtins): Convert from a global struct to a subclass of
11584 gimple_opt_pass along with...
11585 (pass_data_fold_builtins): ...new pass_data instance and...
11586 (make_pass_fold_builtins): ...new function.
11587 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
11588 subclass of gimple_opt_pass along with...
11589 (pass_data_copy_prop): ...new pass_data instance and...
11590 (make_pass_copy_prop): ...new function.
11591 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
11592 global struct to a subclass of gimple_opt_pass along with...
11593 (pass_data_rename_ssa_copies): ...new pass_data instance and...
11594 (make_pass_rename_ssa_copies): ...new function.
11595 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
11596 subclass of gimple_opt_pass along with...
11597 (pass_data_dce): ...new pass_data instance and...
11598 (make_pass_dce): ...new function.
11599 (pass_dce_loop): Convert from a global struct to a subclass of
11600 gimple_opt_pass along with...
11601 (pass_data_dce_loop): ...new pass_data instance and...
11602 (make_pass_dce_loop): ...new function.
11603 (pass_cd_dce): Convert from a global struct to a subclass of
11604 gimple_opt_pass along with...
11605 (pass_data_cd_dce): ...new pass_data instance and...
11606 (make_pass_cd_dce): ...new function.
11607 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
11608 subclass of gimple_opt_pass along with...
11609 (pass_data_dominator): ...new pass_data instance and...
11610 (make_pass_dominator): ...new function.
11611 (pass_phi_only_cprop): Convert from a global struct to a subclass of
11612 gimple_opt_pass along with...
11613 (pass_data_phi_only_cprop): ...new pass_data instance and...
11614 (make_pass_phi_only_cprop): ...new function.
11615 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
11616 subclass of gimple_opt_pass along with...
11617 (pass_data_dse): ...new pass_data instance and...
11618 (make_pass_dse): ...new function.
11619 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
11620 a subclass of gimple_opt_pass along with...
11621 (pass_data_forwprop): ...new pass_data instance and...
11622 (make_pass_forwprop): ...new function.
11623 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
11624 struct to a subclass of gimple_opt_pass along with...
11625 (pass_data_tree_ifcombine): ...new pass_data instance and...
11626 (make_pass_tree_ifcombine): ...new function.
11627 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
11628 subclass of gimple_opt_pass along with...
11629 (pass_data_ch): ...new pass_data instance and...
11630 (make_pass_ch): ...new function.
11631 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
11632 subclass of gimple_opt_pass along with...
11633 (pass_data_tree_loop): ...new pass_data instance and...
11634 (make_pass_tree_loop): ...new function.
11635 (pass_tree_loop_init): Convert from a global struct to a subclass of
11636 gimple_opt_pass along with...
11637 (pass_data_tree_loop_init): ...new pass_data instance and...
11638 (make_pass_tree_loop_init): ...new function.
11639 (pass_lim): Convert from a global struct to a subclass of
11640 gimple_opt_pass along with...
11641 (pass_data_lim): ...new pass_data instance and...
11642 (make_pass_lim): ...new function.
11643 (pass_tree_unswitch): Convert from a global struct to a subclass of
11644 gimple_opt_pass along with...
11645 (pass_data_tree_unswitch): ...new pass_data instance and...
11646 (make_pass_tree_unswitch): ...new function.
11647 (pass_predcom): Convert from a global struct to a subclass of
11648 gimple_opt_pass along with...
11649 (pass_data_predcom): ...new pass_data instance and...
11650 (make_pass_predcom): ...new function.
11651 (pass_vectorize): Convert from a global struct to a subclass of
11652 gimple_opt_pass along with...
11653 (pass_data_vectorize): ...new pass_data instance and...
11654 (make_pass_vectorize): ...new function.
11655 (pass_graphite): Convert from a global struct to a subclass of
11656 gimple_opt_pass along with...
11657 (pass_data_graphite): ...new pass_data instance and...
11658 (make_pass_graphite): ...new function.
11659 (pass_graphite_transforms): Convert from a global struct to a subclass
11660 of gimple_opt_pass along with...
11661 (pass_data_graphite_transforms): ...new pass_data instance and...
11662 (make_pass_graphite_transforms): ...new function.
11663 (pass_check_data_deps): Convert from a global struct to a subclass of
11664 gimple_opt_pass along with...
11665 (pass_data_check_data_deps): ...new pass_data instance and...
11666 (make_pass_check_data_deps): ...new function.
11667 (pass_iv_canon): Convert from a global struct to a subclass of
11668 gimple_opt_pass along with...
11669 (pass_data_iv_canon): ...new pass_data instance and...
11670 (make_pass_iv_canon): ...new function.
11671 (pass_scev_cprop): Convert from a global struct to a subclass of
11672 gimple_opt_pass along with...
11673 (pass_data_scev_cprop): ...new pass_data instance and...
11674 (make_pass_scev_cprop): ...new function.
11675 (pass_record_bounds): Convert from a global struct to a subclass of
11676 gimple_opt_pass along with...
11677 (pass_data_record_bounds): ...new pass_data instance and...
11678 (make_pass_record_bounds): ...new function.
11679 (pass_complete_unroll): Convert from a global struct to a subclass of
11680 gimple_opt_pass along with...
11681 (pass_data_complete_unroll): ...new pass_data instance and...
11682 (make_pass_complete_unroll): ...new function.
11683 (pass_complete_unrolli): Convert from a global struct to a subclass of
11684 gimple_opt_pass along with...
11685 (pass_data_complete_unrolli): ...new pass_data instance and...
11686 (make_pass_complete_unrolli): ...new function.
11687 (pass_parallelize_loops): Convert from a global struct to a subclass
11688 of gimple_opt_pass along with...
11689 (pass_data_parallelize_loops): ...new pass_data instance and...
11690 (make_pass_parallelize_loops): ...new function.
11691 (pass_loop_prefetch): Convert from a global struct to a subclass of
11692 gimple_opt_pass along with...
11693 (pass_data_loop_prefetch): ...new pass_data instance and...
11694 (make_pass_loop_prefetch): ...new function.
11695 (pass_iv_optimize): Convert from a global struct to a subclass of
11696 gimple_opt_pass along with...
11697 (pass_data_iv_optimize): ...new pass_data instance and...
11698 (make_pass_iv_optimize): ...new function.
11699 (pass_tree_loop_done): Convert from a global struct to a subclass of
11700 gimple_opt_pass along with...
11701 (pass_data_tree_loop_done): ...new pass_data instance and...
11702 (make_pass_tree_loop_done): ...new function.
11703 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
11704 struct to a subclass of gimple_opt_pass along with...
11705 (pass_data_cse_reciprocals): ...new pass_data instance and...
11706 (make_pass_cse_reciprocals): ...new function.
11707 (pass_cse_sincos): Convert from a global struct to a subclass of
11708 gimple_opt_pass along with...
11709 (pass_data_cse_sincos): ...new pass_data instance and...
11710 (make_pass_cse_sincos): ...new function.
11711 (pass_optimize_bswap): Convert from a global struct to a subclass of
11712 gimple_opt_pass along with...
11713 (pass_data_optimize_bswap): ...new pass_data instance and...
11714 (make_pass_optimize_bswap): ...new function.
11715 (pass_optimize_widening_mul): Convert from a global struct to a
11716 subclass of gimple_opt_pass along with...
11717 (pass_data_optimize_widening_mul): ...new pass_data instance and...
11718 (make_pass_optimize_widening_mul): ...new function.
11719 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
11720 subclass of gimple_opt_pass along with...
11721 (pass_data_phiopt): ...new pass_data instance and...
11722 (make_pass_phiopt): ...new function.
11723 (pass_cselim): Convert from a global struct to a subclass of
11724 gimple_opt_pass along with...
11725 (pass_data_cselim): ...new pass_data instance and...
11726 (make_pass_cselim): ...new function.
11727 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
11728 subclass of gimple_opt_pass along with...
11729 (pass_data_phiprop): ...new pass_data instance and...
11730 (make_pass_phiprop): ...new function.
11731 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
11732 subclass of gimple_opt_pass along with...
11733 (pass_data_pre): ...new pass_data instance and...
11734 (make_pass_pre): ...new function.
11735 (pass_fre): Convert from a global struct to a subclass of
11736 gimple_opt_pass along with...
11737 (pass_data_fre): ...new pass_data instance and...
11738 (make_pass_fre): ...new function.
11739 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
11740 subclass of gimple_opt_pass along with...
11741 (pass_data_reassoc): ...new pass_data instance and...
11742 (make_pass_reassoc): ...new function.
11743 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
11744 subclass of gimple_opt_pass along with...
11745 (pass_data_sink_code): ...new pass_data instance and...
11746 (make_pass_sink_code): ...new function.
11747 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
11748 subclass of gimple_opt_pass along with...
11749 (pass_data_strlen): ...new pass_data instance and...
11750 (make_pass_strlen): ...new function.
11751 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
11752 struct to a subclass of gimple_opt_pass along with...
11753 (pass_data_build_alias): ...new pass_data instance and...
11754 (make_pass_build_alias): ...new function.
11755 (pass_build_ealias): Convert from a global struct to a subclass of
11756 gimple_opt_pass along with...
11757 (pass_data_build_ealias): ...new pass_data instance and...
11758 (make_pass_build_ealias): ...new function.
11759 (pass_ipa_pta): Convert from a global struct to a subclass of
11760 simple_ipa_opt_pass along with...
11761 (pass_data_ipa_pta): ...new pass_data instance and...
11762 (make_pass_ipa_pta): ...new function.
11763 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
11764 subclass of gimple_opt_pass along with...
11765 (pass_data_uncprop): ...new pass_data instance and...
11766 (make_pass_uncprop): ...new function.
11767 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
11768 global struct to a subclass of gimple_opt_pass along with...
11769 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
11770 (make_pass_late_warn_uninitialized): ...new function.
11771 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
11772 to a subclass of gimple_opt_pass along with...
11773 (pass_data_init_datastructures): ...new pass_data instance and...
11774 (make_pass_init_datastructures): ...new function.
11775 (pass_early_warn_uninitialized): Convert from a global struct to a
11776 subclass of gimple_opt_pass along with...
11777 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
11778 (make_pass_early_warn_uninitialized): ...new function.
11779 (pass_update_address_taken): Convert from a global struct to a
11780 subclass of gimple_opt_pass along with...
11781 (pass_data_update_address_taken): ...new pass_data instance and...
11782 (make_pass_update_address_taken): ...new function.
11783 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
11784 struct to a subclass of gimple_opt_pass along with...
11785 (pass_data_release_ssa_names): ...new pass_data instance and...
11786 (make_pass_release_ssa_names): ...new function.
11787 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
11788 subclass of gimple_opt_pass along with...
11789 (pass_data_stdarg): ...new pass_data instance and...
11790 (make_pass_stdarg): ...new function.
11791 * tree-switch-conversion.c (pass_convert_switch): Convert from a
11792 global struct to a subclass of gimple_opt_pass along with...
11793 (pass_data_convert_switch): ...new pass_data instance and...
11794 (make_pass_convert_switch): ...new function.
11795 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
11796 to a subclass of gimple_opt_pass along with...
11797 (pass_data_tail_recursion): ...new pass_data instance and...
11798 (make_pass_tail_recursion): ...new function.
11799 (pass_tail_calls): Convert from a global struct to a subclass of
11800 gimple_opt_pass along with...
11801 (pass_data_tail_calls): ...new pass_data instance and...
11802 (make_pass_tail_calls): ...new function.
11803 * tree-vect-generic.c (pass_lower_vector): Convert from a global
11804 struct to a subclass of gimple_opt_pass along with...
11805 (pass_data_lower_vector): ...new pass_data instance and...
11806 (make_pass_lower_vector): ...new function.
11807 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
11808 gimple_opt_pass along with...
11809 (pass_data_lower_vector_ssa): ...new pass_data instance and...
11810 (make_pass_lower_vector_ssa): ...new function.
11811 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
11812 to a subclass of gimple_opt_pass along with...
11813 (pass_data_slp_vectorize): ...new pass_data instance and...
11814 (make_pass_slp_vectorize): ...new function.
11815 (pass_ipa_increase_alignment): Convert from a global struct to a
11816 subclass of simple_ipa_opt_pass along with...
11817 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
11818 (make_pass_ipa_increase_alignment): ...new function.
11819 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
11820 gimple_opt_pass along with...
11821 (pass_data_vrp): ...new pass_data instance and...
11822 (make_pass_vrp): ...new function.
11823 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
11824 subclass of simple_ipa_opt_pass along with...
11825 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
11826 (make_pass_ipa_free_lang_data): ...new function.
11827 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
11828 gimple_opt_pass along with...
11829 (pass_data_tsan): ...new pass_data instance and...
11830 (make_pass_tsan): ...new function.
11831 (pass_tsan_O0): Convert from a global struct to a subclass of
11832 gimple_opt_pass along with...
11833 (pass_data_tsan_O0): ...new pass_data instance and...
11834 (make_pass_tsan_O0): ...new function.
11835 * var-tracking.c (pass_variable_tracking): Convert from a global
11836 struct to a subclass of rtl_opt_pass along with...
11837 (pass_data_variable_tracking): ...new pass_data instance and...
11838 (make_pass_variable_tracking): ...new function.
11839 * web.c (pass_web): Convert from a global struct to a subclass of
11840 rtl_opt_pass along with...
11841 (pass_data_web): ...new pass_data instance and...
11842 (make_pass_web): ...new function.
11843 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
11844 declaration with that of...
11845 (make_pass_mode_switch_use): ...new function.
11846 (pass_resolve_sw_modes): Replace declaration with that of...
11847 (make_pass_resolve_sw_modes): ...new function.
11848 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
11849 from a global struct to a subclass of rtl_opt_pass along with...
11850 (pass_data_mode_switch_use): ...new pass_data instance and...
11851 (make_pass_mode_switch_use): ...new function.
11852 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
11853 from a global struct to a subclass of rtl_opt_pass along with...
11854 (pass_data_resolve_sw_modes): ...new pass_data instance and...
11855 (make_pass_resolve_sw_modes): ...new function.
11856 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
11857 struct to a subclass of rtl_opt_pass along with...
11858 (pass_data_insert_vzeroupper): ...new pass_data instance and...
11859 (make_pass_insert_vzeroupper): ...new function.
11860 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
11861 global struct to a subclass of rtl_opt_pass along with...
11862 (pass_data_work_around_errata): ...new pass_data instance and...
11863 (make_pass_work_around_errata): ...new function.
11864 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
11865 struct to a subclass of rtl_opt_pass along with...
11866 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
11867 (make_pass_mips_machine_reorg2): ...new function.
11869 2013-08-05 David Malcolm <dmalcolm@redhat.com>
11871 * passes.c (pass_manager::operator new): New.
11873 2013-08-05 David Malcolm <dmalcolm@redhat.com>
11875 Handwritten part of conversion of passes to C++ classes.
11877 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
11878 (toplev.o): Add dep on PASS_MANAGER_H.
11879 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
11880 of early local pases to reflect this moving from a global to a
11881 member of gcc::pass_manager.
11882 (cgraph_add_new_function): Likewise.
11883 * lto-cgraph.c (lto_output_node): Update for conversion of
11884 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
11885 * passes.c (opt_pass::clone): New.
11886 (opt_pass::gate): New.
11887 (opt_pass::execute): New.
11888 (opt_pass::opt_pass): New.
11889 (pass_manager::execute_early_local_passes): New.
11890 (pass_manager::execute_pass_mode_switching): new.
11891 (finish_optimization_passes): Convert to...
11892 (pass_manager::finish_optimization_passes): ...this.
11893 (finish_optimization_passes): Update for conversion of passes to
11895 (register_dump_files_1): Use has_gate since we cannot portably
11896 check a vtable entry against NULL.
11897 (dump_one_pass): Likewise.
11898 (ipa_write_summaries_2): Likewise.
11899 (ipa_write_optimization_summaries_1): Likewise.
11900 (ipa_read_summaries_1): Likewise.
11901 (ipa_read_optimization_summaries_1): Likewise.
11902 (execute_ipa_stmt_fixups): Likewise.
11903 (pass_manager::pass_manager): Rewrite pass-creation, invoking
11904 pass-creation functions rather than wiring up globals, and
11905 storing the results in fields of pass_manager generated using
11906 pass-instances.def.
11907 (pass_manager::dump_profile_report): Update for conversion of
11908 passes to C++ classes.
11909 (pass_manager::execute_ipa_summary_passes): Likewise.
11910 (execute_one_ipa_transform_pass): Likewise.
11911 (execute_one_pass): Use has_gate and has_execute since we cannot
11912 portably check a vtable entry against NULL.
11913 * pass_manager.h (pass_manager::finish_optimization_passes): New.
11914 (pass_manager): Use pass-instances.def to add fields for the
11915 various pass instances.
11916 * toplev.c (finalize): Update for move of
11917 finish_optimization_passes to a method of gcc::pass_manager.
11918 * toplev.h (finish_optimization_passes): Move to method of class
11920 * tree-pass.h (struct pass_data): New.
11921 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
11922 (opt_pass::gate): Convert to virtual function.
11923 (opt_pass::~opt_pass): New.
11924 (opt_pass::clone): New.
11925 (opt_pass::execute): Convert to virtual function.
11926 (opt_pass::opt_pass): New.
11927 (opt_pass::ctxt_): new.
11928 (gimple_opt_pass): Convert to subclass of opt_pass.
11929 (gimple_opt_pass::gimple_opt_pass): New.
11930 (rtl_opt_pass): Convert to subclass of opt_pass.
11931 (rtl_opt_pass::rtl_opt_pass): New.
11932 (ipa_opt_pass_d): Convert to subclass of opt_pass.
11933 (ipa_opt_pass_d::ipa_opt_pass_d): New.
11934 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
11935 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
11936 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
11937 invocation of pass_mode_switching to reflect this moving from a
11938 global to a member of gcc::pass_manager.
11939 (ix86_option_override): Rework how pass_insert_vzeroupper is
11940 added to the pass_manager to reflect autogenerated changes.
11941 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
11943 2013-08-05 Richard Earnshaw <rearnsha@arm.com>
11945 PR rtl-optimization/57708
11946 * recog.c (peep2_find_free_register): Validate all regs in a
11949 2013-08-05 Jan Hubicka <jh@suse.cz>
11952 * cgraph.c (verify_cgraph_node): Accept local flags from other
11954 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
11955 (function_and_variable_visibility): Likewise.
11956 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
11958 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
11960 * graph.c (init_graph_slim_pretty_print): Remove.
11961 (print_graph_cfg): Do not call it. Use local pretty printer.
11962 (start_graph_dump): Likewise.
11964 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
11966 * gimple-pretty-print.c (buffer): Remove.
11967 (initialized): Likewise.
11968 (maybe_init_pretty_print): Likewise.
11969 (print_gimple_stmt): Do not call it. Use non-static local
11970 pretty_printer variable.
11971 (print_gimple_expr): Likewise.
11972 (print_gimple_seq): Likewise.
11973 (gimple_dump_bb): Likewise.
11975 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
11977 * asan.c (asan_pp): Remove.
11978 (asan_pp_initialized): Likewise.
11979 (asan_pp_initialize): Likewise.
11980 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
11981 (asan_emit_stack_protection): Tidy. Use local pretty printer.
11982 (asan_add_global): Likewise.
11984 2013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
11986 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
11987 * diagnostic.h (diagnostic_flush_buffer): Adjust.
11988 * pretty-print.c (pp_formatted_text_data): Likewise.
11989 (pp_indent): Rename from pp_base_indent.
11990 (pp_format): Rename from pp_base_format.
11991 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
11992 (pp_format_verbatim): Rename from pp_base_format_verbatim.
11993 (pp_flush): Rename from pp_base_flush.
11994 (pp_set_line_maximum_length): Rename from
11995 pp_base_set_line_maximum_length.
11996 (pp_clear_output_area): Rename from pp_base_clear_output_area.
11997 (pp_set_prefix): Rename from pp_base_set_prefix.
11998 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
11999 (pp_emit_prefix): Rename from pp_base_emit_prefix.
12000 (pp_append_text): Rename from pp_base_append_text.
12001 (pp_formatted_text): Rename from pp_base_formatted_text.
12002 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
12003 (pp_remaining_character_count_for_line): Rename from
12004 pp_base_remaining_character_count_for_line.
12005 (pp_newline): Rename from pp_base_newline.
12006 (pp_character): Rename from pp_base_character.
12007 (pp_string): Rename from pp_base_string.
12008 (pp_maybe_space): Rename from pp_base_maybe_space.
12009 * asan.c (asan_pp_string): Adjust.
12010 (asan_emit_stack_protection): Likewise.
12011 (asan_add_global): Likewise.
12012 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
12013 * tree-mudflap.c (mf_varname_tree): Likewise.
12014 * tree-pretty-print.c (pp_tree_identifier): Rename from
12015 pp_base_tree_identifier.
12016 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
12017 Declare as function.
12019 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
12021 * pretty-print.h (pp_bar_bar): New.
12022 (pp_ampersand_ampersand): Likewise.
12023 (pp_less_equal): Likewise.
12024 (pp_greater_equal): Likewise.
12025 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
12026 printer functions instead of pp_string or operators and punctuators.
12027 (dump_gimple_call): Likewise.
12028 (dump_gimple_omp_for): Likewise.
12029 (dump_gimple_transaction): Likewise.
12030 (dump_gimple_phi): Likewise.
12031 (pp_gimple_stmt_1): Likewise.
12032 * sched-vis.c (print_insn): Likewise.
12033 * tree-mudflap.c (mf_varname_tree): Likewise.
12034 * tree-pretty-print.c (dump_block_node): Likewise.
12035 (dump_generic_node): Likewise.
12037 2013-08-02 Jan Hubicka <jh@suse.cz>
12039 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
12041 * lto-streamer-out.c (tree_is_indexable): Results decls and
12042 parm decls are not indexable.
12043 (DFS_write_tree_body): Do not follow args and results.
12044 (hash_tree): Likewise.
12045 (output_functions): Rearrange so struct function is needed
12046 only when real body is output; be able to also ouptut abstract
12047 functions; output DECL_ARGUMENTS and DECL_RESULT.
12048 (lto_output): When not in WPA, ale store abstract functions.
12049 (write_symbol): Do not care about RESULT_DECL.
12050 (output_symbol_p): Handle correctly sbtract decls.
12051 * lto-streamer-in.c (input_function): Rearrange so struct
12052 function can be NULL at entry; allow streaming of
12053 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
12054 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
12055 sanity check during LTO.
12056 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
12057 RESULT_DECl and DECL_ARGUMENTS.
12058 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
12061 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
12063 * pretty-print.h (pp_underscore): New.
12065 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
12066 printer functions instead of pp_character.
12067 (dump_binary_rhs): Likewise.
12068 (dump_ternary_rhs): Likewise.
12069 (dump_gimple_call_args): Likewise.
12070 (pp_points_to_solution): Likewise.
12071 (dump_gimple_call): Likewise.
12072 (dump_gimple_switch): Likewise.
12073 (dump_gimple_cond): Likewise.
12074 (dump_gimple_bind): Likewise.
12075 (dump_gimple_try): Likewise.
12076 (dump_gimple_omp_for): Likewise.
12077 (dump_gimple_omp_continue): Likewise.
12078 (dump_gimple_omp_single): Likewise.
12079 (dump_gimple_omp_sections): Likewise.
12080 (dump_gimple_omp_block): Likewise.
12081 (dump_gimple_omp_critical): Likewise.
12082 (dump_gimple_transaction): Likewise.
12083 (dump_gimple_asm): Likewise.
12084 (dump_gimple_phi): Likewise.
12085 (dump_gimple_omp_parallel): Likewise.
12086 (dump_gimple_omp_task): Likewise.
12087 (dump_gimple_omp_atomic_load): Likewise.
12088 (dump_gimple_omp_atomic_store): Likewise.
12089 (dump_gimple_mem_ops): Likewise.
12090 (pp_gimple_stmt_1): Likewise.
12091 (pp_cfg_jump): Likewise.
12092 (dump_implicit_edges): Likewise.
12093 (gimple_dump_bb_for_graph): Likewise.
12094 * graph.c (draw_cfg_node): Likewise.
12095 * langhooks.c (lhd_print_error_function): Likewise.
12096 * sched-vis.c (print_exp): Likewise.
12097 (print_value): Likewise.
12098 (print_pattern): Likewise.
12099 (print_insn): Likewise.
12100 (rtl_dump_bb_for_graph): Likewise.
12101 * tree-pretty-print.c (dump_function_declaration): Likewise.
12102 (dump_array_domain): Likewise.
12103 (dump_omp_clause): Likewise.
12104 (dump_location): Likewise.
12105 (dump_generic_node): Likewise.
12106 (print_struct_decl): Likewise.
12107 * diagnostic.c (diagnostic_show_locus): Use pp_space.
12109 2013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
12111 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
12112 candidate table when replacing a candidate statement.
12113 (replace_rhs_if_not_dup): Likewise.
12114 (replace_one_candidate): Likewise.
12116 2013-08-02 Jan Hubicka <jh@suse.cz>
12117 Martin Liska <marxin.liska@gmail.com>
12119 * cgraphunit.c (add_new_function): Fix logic when adding from
12121 (assemble_thunk): Rename to ...
12122 (expand_thunk); .. this one; export; get it working with
12123 general functions; make produced gimple valid.
12124 * cgraph.h (expand_thunk): Declare.
12126 2013-08-02 Jan Hubicka <jh@suse.cz>
12128 * ipa-cp.c (gather_context_independent_values): Use
12129 ipa_get_param_move_cost.
12130 (get_replacement_map): Remove PARAM; move parameter folding
12132 (create_specialized_node): Update.
12133 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
12134 assert that we have gimple body; update move_cost.
12135 (count_formal_params): Assert that we have gimple body.
12136 (ipa_dump_param): New function.
12137 (ipa_alloc_node_params): Break out from ...
12138 (ipa_initialize_node_params): ... here.
12139 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
12140 (ipa_write_node_info): Stream move costs.
12141 (ipa_read_node_info): Read move costs.
12142 (ipa_update_after_lto_read): Do not recompute node params.
12143 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
12144 (ipa_get_param): Check we are not in WPA.
12145 (ipa_get_param_move_cost): New.
12146 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
12147 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
12148 parm numbers to be present.
12150 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
12152 PR rtl-optimization/58048
12153 * lra-constraints.c (process_alt_operands): Don't check asm
12154 operand on register.
12156 2013-08-02 Eric Botcazou <ebotcazou@adacore.com>
12158 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
12159 the implied StoreLoad barrier for atomic operations if before.
12161 2013-08-02 Jan Hubicka <jh@suse.cz>
12162 Martin Liska <marxin.liska@gmail.com>
12164 * cgraph.c (cgraph_function_body_availability): Do not check
12166 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
12167 symtab_node_availability): Declare.
12168 * ipa.c (can_replace_by_local_alias): New.
12169 (function_and_variable_visibility): Use it.
12170 * symtab.c (symtab_for_node_and_aliases,
12171 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
12173 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
12175 PR rtl-optimization/57963
12176 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
12177 (lra_constraints): Use them.
12179 2013-08-02 Sofiane Naci <sofiane.naci@arm.com>
12181 * config/arm/types.md (define_attr "type"): Add "load_acq"
12183 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
12185 (cortex_a53_store1): Likewise.
12187 2013-08-01 Jan Hubicka <jh@suse.cz>
12189 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
12190 partitions are not needed.
12192 2013-08-01 Uros Bizjak <ubizjak@gmail.com>
12194 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
12195 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
12196 MAYBE_NON_Q_CLASS_P where appropriate.
12198 2013-08-01 Jan Hubicka <jh@suse.cz>
12200 * cgraph.h (release_function_body): Declare.
12201 * tree.c (free_lang_data_in_decl): Free, parameters and return values
12202 of unused delcarations.
12204 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12206 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
12207 RTL form when subtracting a constant.
12209 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12211 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
12212 Generate canonical plus rtx with negated immediate instead of minus
12214 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
12216 2013-08-01 Jan Hubicka <jh@suse.cz>
12218 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
12219 (cgraph_release_function_body): Likewise.
12220 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
12221 * cgraph.h (cgrpah_node): Rename abstract_and_needed
12222 to used_as_abstract_origin.
12223 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
12224 symbols used as abstract origins.
12225 * cgraphunit.c (analyze_functions): Update.
12226 * ipa.c (symtab_remove_unreachable_nodes): Recompute
12227 used_as_abstract_origin.
12228 * tree-inline.c (tree_function_versioning): Update
12229 used_as_abstract_origin; be ready for DECL_RESULT and
12230 DECL_ARGUMENTS to be NULL.
12232 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
12233 for abstract functions.
12234 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
12237 2013-08-01 Jan Hubicka <jh@suse.cz>
12239 * profile.c (compute_value_histograms): Fix thinko.
12241 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
12243 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
12244 aarch-common-protos.h to extra_headers.
12245 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
12246 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
12247 * config/aarch64/t-aarch64 (aarch-common.o): Define.
12249 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
12251 * config/aarch64/aarch64.md (define_attr "type"): Delete.
12252 Include "../arm/types.md". Define "type" attribute for all patterns.
12253 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
12256 2013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
12258 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
12259 to support power8 load fusion.
12260 (fusion_gpr_mem_load): Likewise.
12262 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
12264 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
12265 declarations for power8 load fusion.
12266 (emit_fusion_gpr_load): Likewise.
12268 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
12269 tuning for power8, turn on fusion mode by default. Turn on sign
12270 extending fusion mode if normal fusion mode is on, and we are at
12272 (fusion_gpr_load_p): New function, return true if we can fuse an
12273 addis instruction with a dependent load to a GPR.
12274 (emit_fusion_gpr_load): Emit the instructions for power8 load
12277 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
12278 (VSX load fusion peepholes): New peepholes to fuse together an
12279 addi instruction with a VSX load instruction.
12281 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
12282 peepholes to fuse an addis instruction with a load to a GPR base
12283 register. If we are supporting sign extending fusions, convert
12284 sign extending loads to zero extending loads and add an explicit
12287 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
12289 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
12290 aarch-common-protos.h to extra_headers.
12291 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
12292 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
12293 (arm_early_store_addr_dep): Likewise.
12294 (arm_no_early_alu_shift_dep): Likewise.
12295 (arm_no_early_alu_shift_value_dep): Likewise.
12296 (arm_no_early_mul_dep): Likewise.
12297 (arm_no_early_store_addr_dep): Likewise.
12298 (arm_mac_accumulator_is_mul_result): Likewise.
12299 (arm_mac_accumulator_is_result): Likewise.
12300 * config/arm/aarch-common.c: ... here. New file.
12301 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
12303 (arm_early_store_addr_dep): Likewise.
12304 (arm_no_early_alu_shift_dep): Likewise.
12305 (arm_no_early_alu_shift_value_dep): Likewise.
12306 (arm_no_early_mul_dep): Likewise.
12307 (arm_no_early_store_addr_dep): Likewise.
12308 (arm_mac_accumulator_is_mul_result): Likewise.
12309 (arm_mac_accumulator_is_result): Likewise.
12310 * config/arm/aarch-common-protos.h: ... here. New file.
12311 * config/arm/t-arm (aarch-common.o): Define.
12313 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
12315 * config/arm/arm.md: Include new file "types.md".
12316 (define_attr "type"): Move from here to ...
12317 (define_attr "mul32"): Likewise.
12318 (define_attr "mul64"): Likewise.
12319 * config/arm/types.md: ... here. New file.
12321 2013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
12323 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
12324 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
12326 2013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
12328 * gen-pass-instances.awk: Fix offset of substr().
12330 2013-07-31 David Malcolm <dmalcolm@redhat.com>
12332 * Makefile.in (pass-instances.def): New.
12333 (passes.o): Replace dependency on passes.def with one on
12336 * gen-pass-instances.awk: New.
12338 * passes.c (pass_manager::pass_manager): Use pass-instances.def
12339 rather than passes.def, updating local definition of NEXT_PASS
12340 macro to add an extra NUM parameter (currently unused).
12342 2013-07-30 David Malcolm <dmalcolm@redhat.com>
12344 * Makefile.in (PASS_MANAGER_H): New.
12345 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
12346 (passes.o): Likewise.
12347 (statistics.o): Likewise.
12348 (cgraphunit.o): Likewise.
12349 (context.o): Depend on PASS_MANAGER_H.
12351 * pass_manager.h: New.
12353 * cgraphunit.c (cgraph_add_new_function): Update for moves
12354 of globals to fields of pass_manager.
12355 (analyze_function): Likewise.
12356 (expand_function): Likewise.
12357 (ipa_passes): Likewise.
12358 (compile): Likewise.
12360 * context.c (context::context): New.
12361 * context.h (context::context): New.
12362 (context::get_passes): New.
12363 (context::passes_): New.
12365 * lto-cgraph.c (input_node): Update for moves of globals to
12366 fields of pass_manager.
12368 * passes.c (all_passes): Remove, in favor of a field of the
12369 same name within the new class pass_manager.
12370 (all_small_ipa_passes): Likewise.
12371 (all_lowering_passes): Likewise.
12372 (all_regular_ipa_passes): Likewise.
12373 (all_late_ipa_passes): Likewise.
12374 (all_lto_gen_passes): Likewise.
12375 (passes_by_id): Likewise.
12376 (passes_by_id_size): Likewise.
12377 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
12378 the new class pass_manager.
12379 (set_pass_for_id): Convert to...
12380 (pass_manager::set_pass_for_id): ...method.
12381 (get_pass_for_id): Convert to...
12382 (pass_manager::get_pass_for_id): ...method.
12383 (register_one_dump_file): Move body of implementation into...
12384 (pass_manager::register_one_dump_file): ...here.
12385 (register_dump_files_1): Convert to...
12386 (pass_manager::register_dump_files_1): ...method.
12387 (register_dump_files): Convert to...
12388 (pass_manager::register_dump_files): ...method.
12389 (create_pass_tab): Update for moves of globals to fields of
12391 (dump_passes): Move body of implementation into...
12392 (pass_manager::dump_passes): ...here.
12393 (register_pass): Move body of implementation into...
12394 (pass_manager::register_pass): ...here.
12395 (init_optimization_passes): Convert into...
12396 (pass_manager::pass_manager): ...constructor for new
12397 pass_manager class, and initialize the pass_lists array.
12398 (check_profile_consistency): Update for moves of globals to
12399 fields of pass_manager.
12400 (dump_profile_report): Move body of implementation into...
12401 (pass_manager::dump_profile_report): ...here.
12402 (ipa_write_summaries_1): Update for moves of pass lists from
12403 being globals to fields of pass_manager.
12404 (ipa_write_optimization_summaries): Likewise.
12405 (ipa_read_summaries): Likewise.
12406 (ipa_read_optimization_summaries): Likewise.
12407 (execute_all_ipa_stmt_fixups): Likewise.
12409 * statistics.c (statistics_fini): Update for moves of globals to
12410 fields of pass_manager.
12412 * toplev.c (general_init): Replace call to
12413 init_optimization_passes with construction of the pass_manager
12416 * tree-pass.h (all_passes): Remove, in favor of a field of the
12417 same name within the new class pass_manager.
12418 (all_small_ipa_passes): Likewise.
12419 (all_lowering_passes): Likewise.
12420 (all_regular_ipa_passes): Likewise.
12421 (all_lto_gen_passes): Likewise.
12422 (all_late_ipa_passes): Likewise.
12423 (passes_by_id): Likewise.
12424 (passes_by_id_size): Likewise.
12425 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
12426 the new class pass_manager.
12427 (get_pass_for_id): Remove.
12429 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
12431 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
12434 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
12436 * arm.md (mulhi3): New expand pattern.
12438 2013-07-30 Jan Hubicka <jh@suse.cz>
12439 Martin Liska <marxin.liska@gmail.com>
12441 * profile.c (compute_value_histograms): Do not ICE when
12442 there is mismatch only on some counters.
12444 2013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
12446 PR rtl-optimization/57637
12447 * function.c (move_insn_for_shrink_wrap): Also check the
12448 GEN set of the LIVE problem for the liveness analysis
12449 if it exists, otherwise give up.
12451 2013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
12453 PR tree-optimization/57993
12454 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
12455 replaced statement in the candidate table.
12456 (phi_add_costs): Return infinite cost when the hidden basis does
12457 not dominate all phis on which the candidate is dependent.
12458 (replace_one_candidate): Record replaced statement in the
12461 2013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
12463 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
12464 (ashlv2si3): New expander.
12465 (*ashlv2si3_i): New define_insn_and_split.
12466 * predicates.md (float_operation): Allow patterns with three
12467 basic sub-patterns.
12469 PR rtl-optimization/58021
12470 * mode-switching.c (create_pre_exit): Always split off preceding
12471 insns if we are not at the basic block head.
12473 2013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
12475 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
12476 (UCLIBC_DYNAMIC_LINKER): New macro.
12477 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
12479 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
12480 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
12482 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
12483 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
12484 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
12485 for SF and DF modes. Use ieee_quad_format for TF mode.
12486 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
12487 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
12488 (mips_option_override): Handle `-mnan=legacy'.
12489 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
12490 `-mabs=2008' and `-mnan=2008'.
12491 (OPTION_DEFAULT_SPECS): Add "nan" default.
12492 (ASM_SPEC): Handle `-mnan='.
12493 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
12494 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
12495 comment accordingly.
12496 (neg<mode>2): Likewise.
12497 * config/mips/mips.opt (mabs, mnan): New options.
12498 * doc/install.texi (Configuration): Document `--with-nan=' option.
12499 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
12501 (MIPS Options): Document them.
12502 * config.gcc <mips*-*-*>: Handle `--with-nan='.
12503 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
12504 * configure: Regenerate.
12505 * config.in: Regenerate.
12507 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
12509 * config/i386/i386.md (float post-reload splitters): Do not check
12510 for subregs of SSE registers.
12512 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
12513 H.J. Lu <hongjiu.lu@intel.com>
12517 * config/i386/i386.md (post-reload splitter
12518 to avoid partial SSE reg dependency stalls): New pattern.
12520 2013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
12522 * config/s390/s390.md ("movcc"): Swap load and store instructions.
12524 2013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
12526 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
12527 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
12529 2013-07-26 Cary Coutant <ccoutant@google.com>
12531 * dwarf2out.c (die_checksum_ordered): Don't include template
12532 instantiations in signature.
12533 (is_template_parameter): New function.
12534 (is_template_instantiation): New function.
12535 (generate_skeleton_bottom_up): Don't include template instantiations
12537 (generate_skeleton): Likewise.
12538 (break_out_comdat_types): Move recursive call to break out nested
12540 (prune_unused_types_mark_generic_parms_dies): Call
12541 is_template_parameter.
12543 2013-07-26 Ian Bolton <ian.bolton@arm.com>
12545 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
12546 uses vector registers.
12547 * config/aarch64/iterators.md: Add attributes rtn and vas.
12549 2013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12550 Richard Earnshaw <richard.earnshaw@arm.com>
12552 * combine.c (simplify_comparison): Re-canonicalize operands
12554 * config/arm/arm.md (movcond_addsi): New splitter.
12556 2013-07-25 Sterling Augustine <saugustine@google.com>
12558 * dwarf2out.c (size_of_pubnames): Move code to...
12559 (include_pubname_in_output): ...here. New.
12560 (want_pubnames): Rearrange.
12561 (output_pubnames): Call include_pubname_in_output. Move assertion.
12563 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
12565 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
12567 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
12570 * doc/extend.texi: Remove obsolete builtins. Fix
12571 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
12573 2013-07-25 Jan Hubicka <jh@suse.cz>
12575 * cgraph.c (release_function_body): Break out from ...
12576 (cgraph_release_function_body): ... this one; also release DECL_RESULT
12577 and DECL_ARGUMENTS.
12578 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
12579 old_tree in the map.
12580 (create_specialized_node): Update.
12581 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
12583 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
12584 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
12585 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
12586 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
12589 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12591 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
12592 addsi3_carryin_alt2_<optab>): Correct output template.
12594 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12596 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
12597 Adjust for arm_restrict_it.
12598 Remove trailing whitespace.
12600 2013-07-25  Mark Kettenis  <kettenis@openbsd.org>
12602 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
12603 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
12605 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
12607 2013-07-25 Vladimir Makarov <vmakarov@redhat.com>
12609 PR rtl-optimization/57960
12610 * lra-constraints.c (process_alt_operands): Use the right mode
12611 when checking strict_low.
12613 2013-07-25 Jan Hubicka <jh@suse.cz>
12615 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
12616 * cgraph.c (cgraph_remove_node): Do not release function body
12617 when in cgraph streaming.
12618 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
12619 in other partitions are not considered reachable; fix handling of
12622 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
12624 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
12626 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
12631 * config/arm/arm.md ("*sibcall_insn): Replace use of
12632 Ss with US. Adjust output for v5 and v4t.
12633 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
12635 * config/arm/constraints.md ("Ss"): Rename to US.
12637 2013-07-25 Terry Guo <terry.guo@arm.com>
12639 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
12640 shift_add/shift_sub0/shift_sub1 RTXs.
12642 2013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
12643 Anton Blanchard <anton@au1.ibm.com>
12645 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
12646 (altivec_vpks<VI_char>ss): Likewise.
12647 (altivec_vpks<VI_char>us): Likewise.
12648 (altivec_vpku<VI_char>us): Likewise.
12649 (altivec_vpku<VI_char>um): Likewise.
12651 2013-07-24 David Malcolm <dmalcolm@redhat.com>
12653 Introduce context class.
12655 * Makefile.in (CONTEXT_H): New.
12656 (OBJS): Add context.o.
12657 (toplev.o): Add CONTEXT_H to dependencies.
12660 * toplev.c (general_init): Create the singleton gcc::context instance.
12666 2013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
12668 PR rtl-optimization/57968
12669 * mode-switching.c (create_pre_exit): Allow instructions that
12670 don't set a return register to need a non-exit mode.
12672 2013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
12673 Anton Blanchard <anton@au1.ibm.com>
12675 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
12676 operands to vperm for little endian.
12677 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
12678 of lvsl to create the control mask for a vperm for little endian.
12680 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12681 Anton Blanchard <anton@au1.ibm.com>
12683 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
12684 two operands for little-endian.
12686 2013-07-23 Steve Ellcey <sellcey@mips.com>
12688 * config/mips/mips.c (mips_case_values_threshold): New.
12689 (TARGET_CASE_VALUES_THRESHOLD): Define.
12691 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12692 Anton Blanchard <anton@au1.ibm.com>
12694 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
12695 selection of field for vector splat in little endian mode.
12697 2013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
12699 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
12700 expanders to rs6000.md.
12701 (ior<mode>3): Likewise.
12702 (and<mode>3): Likewise.
12703 (one_cmpl<mode>2): Likewise.
12704 (nor<mode>3): Likewise.
12705 (andc<mode>3): Likewise.
12706 (eqv<mode>3): Likewise.
12707 (nand<mode>3): Likewise.
12708 (orc<mode>3): Likewise.
12710 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
12713 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
12714 to split multi-word logical operations.
12715 (rs6000_split_logical_di): Likewise.
12716 (rs6000_split_logical): Likewise.
12718 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
12719 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
12720 and allow TImode operations in 32-bit.
12721 (vsx_and<mode>3_64bit): Likewise.
12722 (vsx_ior<mode>3_32bit): Likewise.
12723 (vsx_ior<mode>3_64bit): Likewise.
12724 (vsx_xor<mode>3_32bit): Likewise.
12725 (vsx_xor<mode>3_64bit): Likewise.
12726 (vsx_one_cmpl<mode>2_32bit): Likewise.
12727 (vsx_one_cmpl<mode>2_64bit): Likewise.
12728 (vsx_nor<mode>3_32bit): Likewise.
12729 (vsx_nor<mode>3_64bit): Likewise.
12730 (vsx_andc<mode>3_32bit): Likewise.
12731 (vsx_andc<mode>3_64bit): Likewise.
12732 (vsx_eqv<mode>3_32bit): Likewise.
12733 (vsx_eqv<mode>3_64bit): Likewise.
12734 (vsx_nand<mode>3_32bit): Likewise.
12735 (vsx_nand<mode>3_64bit): Likewise.
12736 (vsx_orc<mode>3_32bit): Likewise.
12737 (vsx_orc<mode>3_64bit): Likewise.
12739 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
12740 logical types in GPRs.
12742 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
12743 logical insns to rs6000.md, and allow TImode operations in
12745 (altivec_ior<mode>3): Likewise.
12746 (altivec_xor<mode>3): Likewise.
12747 (altivec_one_cmpl<mode>2): Likewise.
12748 (altivec_nor<mode>3): Likewise.
12749 (altivec_andc<mode>3): Likewise.
12751 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
12752 attributes for moving the 128-bit logical operations into
12754 (BOOL_REGS_OUTPUT): Likewise.
12755 (BOOL_REGS_OP1): Likewise.
12756 (BOOL_REGS_OP2): Likewise.
12757 (BOOL_REGS_UNARY): Likewise.
12758 (BOOL_REGS_AND_CR0): Likewise.
12759 (one_cmpl<mode>2): Add support for DI logical operations on
12760 32-bit, splitting the operations to 32-bit.
12761 (anddi3): Likewise.
12762 (iordi3): Likewise.
12763 (xordi3): Likewise.
12764 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
12765 changes to combine the 32/64-bit code, allow logical operations on
12766 TI mode in 32-bit, and to use similar match_operator patterns like
12767 scalar mode uses. Combine the Altivec and VSX code for logical
12768 operations, and move it here.
12769 (ior<mode>3, 128-bit types): Likewise.
12770 (xor<mode>3, 128-bit types): Likewise.
12771 (one_cmpl<mode>3, 128-bit types): Likewise.
12772 (nor<mode>3, 128-bit types): Likewise.
12773 (andc<mode>3, 128-bit types): Likewise.
12774 (eqv<mode>3, 128-bit types): Likewise.
12775 (nand<mode>3, 128-bit types): Likewise.
12776 (orc<mode>3, 128-bit types): Likewise.
12777 (and<mode>3_internal): Likewise.
12778 (bool<mode>3_internal): Likewise.
12779 (boolc<mode>3_internal1): Likewise.
12780 (boolc<mode>3_internal2): Likewise.
12781 (boolcc<mode>3_internal1): Likewise.
12782 (boolcc<mode>3_internal2): Likewise.
12783 (eqv<mode>3_internal1): Likewise.
12784 (eqv<mode>3_internal2): Likewise.
12785 (one_cmpl1<mode>3_internal): Likewise.
12787 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
12789 * config/microblaze/microblaze.c (microblaze_expand_prologue):
12790 Rename flag_stack_usage to flag_stack_usage_info.
12792 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
12794 * config/microblaze/sync.md: New file.
12795 * config/microblaze/microblaze.md: Include sync.md
12796 * config/microblaze/microblaze.c: Add print_operand 'y'.
12797 * config/microblaze/constraints.md: Add memory_contraint
12798 'Q' which is a single register.
12800 2013-07-23 Eric Botcazou <ebotcazou@adacore.com>
12802 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
12804 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
12806 * reload.c (find_reloads): Exit loop once we find this operand
12807 cannot be reloaded somehow for this alternative.
12809 * reload.c (find_reloads): Exit loop once we find a hard register.
12811 * rtlanal.c (computed_jump_p): Exit loop once we find label
12814 * i386.c (ix86_pad_returns): Exit loop after setting replace.
12816 * cfgloopmanip.c (remove_path): Exit loop after setting
12819 * gensupport.c (subst_dup): Avoid loop if code is not
12820 MATCH_DUP nor MATCH_OP_DUP.
12822 2013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
12824 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
12826 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
12828 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
12829 true for SP_REGNUM if mode == ptr_mode.
12830 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
12831 with value R0_REGNUM + 31.
12833 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
12835 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
12836 pad pointer-typed argument downward.
12838 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
12840 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
12841 and __ILP32__ when the ILP32 model is in use.
12843 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
12845 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
12846 (aarch64_load_symref_appropriately): In the case of
12847 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
12848 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
12849 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
12850 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
12851 if 'mode' doesn't equal to 'ptr_mode'.
12852 (aarch64_output_mi_thunk): Add an assertion on the alignment of
12853 'vcall_offset'; change to call aarch64_emit_move differently depending
12854 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
12855 to calculate the upper bound of 'vcall_offset'.
12856 (aarch64_cannot_force_const_mem): Change to also return true if
12858 (aarch64_legitimize_reload_address): In the case of large
12859 displacements, add new local variable 'xmode' and an assertion
12860 based on it; change to use 'xmode' to generate the new rtx and
12862 (aarch64_asm_trampoline_template): Change to generate the template
12863 differently depending on TARGET_ILP32 or not; change to use
12864 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
12865 (aarch64_trampoline_size): Removed.
12866 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
12867 and replace immediate literals with it. Change to use 'ptr_mode'
12868 instead of 'DImode' and call convert_memory_address if the mode
12869 of 'fnaddr' doesn't equal to 'ptr_mode'.
12870 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
12872 (aarch64_elf_asm_destructor): Likewise.
12873 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
12874 on TARGET_ILP32 instead of aarch64_trampoline_size.
12875 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
12876 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
12877 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
12878 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
12879 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
12880 (storewb_pair<GPI:mode>_<P:mode>): ... this.
12881 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
12882 depending on the value of 'mode'.
12883 (add_losym_<mode>): New.
12884 (ldr_got_small_<mode>): New, based on ldr_got_small.
12885 (ldr_got_small): Remove.
12886 (ldr_got_small_sidi): New.
12887 * config/aarch64/iterators.md (P): New.
12888 (PTR): Change to 'ptr_mode' in the condition.
12890 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
12892 * config.gcc (aarch64*-*-*): Support --with-abi.
12893 (aarch64*-*-elf): Support --with-multilib-list.
12894 (aarch64*-*-linux*): Likewise.
12895 (supported_defaults): Add abi to aarch64*-*-*.
12896 * configure.ac: Mention AArch64 for --with-multilib-list.
12897 * configure: Re-generated.
12898 * config/aarch64/biarchilp32.h: New file.
12899 * config/aarch64/biarchlp64.h: New file.
12900 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
12902 (MULTILIB_DEFAULTS): Ditto.
12903 (DRIVER_SELF_SPECS): Ditto.
12904 (ASM_SPEC): Update to also substitute -mabi.
12905 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
12906 file whose name depends on -mabi= and -mbig-endian.
12907 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
12909 (POINTER_SIZE): New define.
12910 (POINTERS_EXTEND_UNSIGNED): Ditto.
12911 (enum aarch64_abi_type): New enumeration tag.
12912 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
12913 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
12914 (TARGET_ILP32): New define.
12915 * config/aarch64/aarch64.opt (mabi): New.
12916 (aarch64_abi): New.
12917 (ilp32, lp64): New values for -mabi.
12918 * config/aarch64/t-aarch64 (comma): New define.
12919 (MULTILIB_OPTIONS): Ditto.
12920 (MULTILIB_DIRNAMES): Ditto.
12921 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
12922 * doc/invoke.texi: Document -mabi for AArch64.
12924 2013-07-23 Georg-Johann Lay <avr@gjlay.de>
12926 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
12928 2013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
12929 Anton Blanchard <anton@au1.ibm.com>
12931 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
12932 endianness when selecting field to splat.
12934 2013-07-22 Eric Christopher <echristo@gmail.com>
12936 * dwarf2out.c (die_odr_checksum): New function to use
12937 CHECKSUM_ macros and ULEB128 for DIE tag.
12938 (generate_type_signature): Use.
12940 2013-07-22 Eric Botcazou <ebotcazou@adacore.com>
12942 * config.gcc (sparc*-*-*): Accept leon3 processor.
12943 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
12944 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
12945 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
12946 * config/sparc/sparc.opt (enum processor_type): Add leon3.
12947 (mfix-ut699): Adjust comment.
12948 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
12949 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
12950 (CPP_CPU_SPEC): Likewise.
12951 (ASM_CPU_SPEC): Likewise.
12952 * config/sparc/sparc.c (leon3_cost): New constant.
12953 (sparc_option_override): Add leon3 support.
12954 (mem_ref): New function.
12955 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
12956 (sparc_do_work_around_errata): Look into the instruction in the delay
12957 slot and adjust accordingly. Add fix for the data cache nullify issues
12958 of the UT699. Change insertion position for the NOP.
12959 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
12960 (leon3_load): New reservation.
12961 (leon_store): Bump latency to 2.
12962 (grfpu): New automaton.
12963 (grfpu_alu): New unit.
12964 (grfpu_ds): Likewise.
12965 (leon_fp_alu): Adjust.
12966 (leon_fp_mult): Delete.
12967 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
12968 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
12969 * config/sparc/sparc.md (cpu): Add leon3.
12970 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
12971 (swapsi): Likewise.
12972 (atomic_test_and_set): Likewise.
12973 (ldstub): Likewise.
12975 2013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
12977 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
12978 default for R5900 targets.
12979 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
12980 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
12981 * config/mips/mips.c (mips_option_override): Report an error for
12982 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
12983 for -march=r5900 -mhard-float.
12985 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
12987 * df-problems.c (can_move_insns_across): Exit loop once we
12988 find a non-fixed, non-global register.
12990 * ipa-pure-const.c (propagate_nothrow): Exit loop after
12993 * omega.c (omega_eliminate_red): Break after setting red_found.
12994 (omega_problem_has_red_equations): Similarly after setting found.
12995 (omega_query_variable): Similarly after setting coupled.
12997 2013-07-22 Marek Polacek <polacek@redhat.com>
12999 * gimplify.c: Don't include gimple.h twice.
13001 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13003 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
13004 instead of TARGET_THUMB1.
13005 (Pz): New constraint.
13006 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
13008 (compare_negsi_si): Likewise.
13009 (compare_addsi2_op0): Likewise.
13010 (compare_addsi2_op1): Likewise.
13011 (addsi3_carryin_<optab>): Likewise.
13012 (addsi3_carryin_alt2_<optab>): Likewise.
13013 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
13014 for arm_restrict_it.
13015 (subsi3_carryin): Likewise.
13016 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
13017 (minmax_arithsi): Disable for arm_restrict_it.
13018 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
13019 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
13020 (satsi_<SAT:code>_shift): Likewise.
13021 (arm_shiftsi3): Add alternative for 16-bit encoding.
13022 (arm32_movhf): Disable for arm_restrict_it.
13023 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
13024 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
13026 2013-07-22 Sofiane Naci <sofiane.naci@arm.com>
13028 * config/arm/arm.md (attribute "insn"): Delete.
13029 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
13030 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
13031 (not_shiftsi): Update for attribute change.
13032 (not_shiftsi_compare0): Likewise.
13033 (not_shiftsi_compare0_scratch): Likewise.
13034 (arm_one_cmplsi2): Likewise.
13035 (thumb1_one_cmplsi2): Likewise.
13036 (notsi_compare0): Likewise.
13037 (notsi_compare0_scratch): Likewise.
13038 (thumb1_movdi_insn): Likewise.
13039 (arm_movsi_insn): Likewise.
13040 (movhi_insn_arch4): Likewise.
13041 (movhi_bytes): Likewise.
13042 (arm_movqi_insn): Likewise.
13043 (thumb1_movqi_insn): Likewise.
13044 (arm32_movhf): Likewise.
13045 (thumb1_movhf): Likewise.
13046 (arm_movsf_soft_insn): Likewise.
13047 (thumb1_movsf_insn): Likewise.
13048 (thumb_movdf_insn): Likewise.
13049 (movsicc_insn): Likewise.
13050 (movsfcc_soft_insn): Likewise.
13051 (and_scc): Likewise.
13052 (cond_move): Likewise.
13053 (if_move_not): Likewise.
13054 (if_not_move): Likewise.
13055 (if_shift_move): Likewise.
13056 (if_move_shift): Likewise.
13057 (if_shift_shift): Likewise.
13058 (if_not_arith): Likewise.
13059 (if_arith_not): Likewise.
13060 (cond_move_not): Likewise.
13061 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
13062 (neon_mov<mode>): Likewise.
13063 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
13064 (thumb2_movsi_vfp): Likewise.
13065 (movsf_vfp): Likewise.
13066 (thumb2_movsf_vfp): Likewise.
13067 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
13069 (cortexa7_older_only): Likewise.
13070 (cortexa7_younger): Likewise.
13071 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
13072 (1020alu_shift_op): Likewise.
13073 (1020alu_shift_reg_op): Likewise.
13074 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
13075 (alu_shift_op): Likewise.
13076 (alu_shift_reg_op): Likewise.
13077 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
13078 (11_alu_shift_op): Likewise.
13079 (11_alu_shift_reg_op): Likewise.
13080 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
13081 (9_alu_shift_reg_op): Likewise.
13082 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
13084 (cortex_a15_alu_shift): Likewise.
13085 (cortex_a15_alu_shift_reg): Likewise.
13086 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
13087 (cortex_a5_alu_shift): Likewise.
13088 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
13090 (cortex_a53_alu_shift): Likewise.
13091 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
13093 (cortex_a7_alu_reg): Likewise.
13094 (cortex_a7_alu_shift): Likewise.
13095 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
13096 (cortex_a8_alu_shift): Likewise.
13097 (cortex_a8_alu_shift_reg): Likewise.
13098 (cortex_a8_mov): Likewise.
13099 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
13100 (cortex_a9_dp_shift): Likewise.
13101 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
13102 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
13103 (cortex_r4_mov): Likewise.
13104 (cortex_r4_alu_shift): Likewise.
13105 (cortex_r4_alu_shift_reg): Likewise.
13106 * config/arm/fa526.md (526_alu_op): Update for attribute change.
13107 (526_alu_shift_op): Likewise.
13108 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
13109 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
13110 (626te_alu_shift_op): Likewise.
13111 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
13112 (726te_alu_op): Likewise.
13113 (726te_alu_shift_op): Likewise.
13114 (726te_alu_shift_reg_op): Likewise.
13115 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
13116 (mp626_alu_shift_op): Likewise.
13117 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
13118 (pj4_alu_e1_conds): Likewise.
13119 (pj4_alu): Likewise.
13120 (pj4_alu_conds): Likewise.
13121 (pj4_shift): Likewise.
13122 (pj4_shift_conds): Likewise.
13123 (pj4_alu_shift): Likewise.
13124 (pj4_alu_shift_conds): Likewise.
13126 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13128 * config/arm/predicates.md (shiftable_operator_strict_it):
13130 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
13131 Disable cond_exec version for arm_restrict_it.
13132 (thumb2_smaxsi3): Convert to generate cond_exec.
13133 (thumb2_sminsi3): Likewise.
13134 (thumb32_umaxsi3): Likewise.
13135 (thumb2_uminsi3): Likewise.
13136 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
13137 (thumb2_neg_abssi2): Likewise.
13138 (thumb2_mov_scc): Add alternative for 16-bit encoding.
13139 (thumb2_movsicc_insn): Adjust alternatives.
13140 (thumb2_mov_negscc): Disable for arm_restrict_it.
13141 (thumb2_mov_negscc_strict_it): New pattern.
13142 (thumb2_mov_notscc_strict_it): New pattern.
13143 (thumb2_mov_notscc): Disable for arm_restrict_it.
13144 (thumb2_ior_scc): Likewise.
13145 (thumb2_ior_scc_strict_it): New pattern.
13146 (thumb2_cond_move): Adjust for arm_restrict_it.
13147 (thumb2_cond_arith): Disable for arm_restrict_it.
13148 (thumb2_cond_arith_strict_it): New pattern.
13149 (thumb2_cond_sub): Adjust for arm_restrict_it.
13150 (thumb2_movcond): Likewise.
13151 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
13152 (thumb2_zero_extendhisi2_v6): Likewise.
13153 (thumb2_zero_extendqisi2_v6): Likewise.
13154 (orsi_notsi_si): Likewise.
13155 (orsi_not_shiftsi_si): Likewise.
13157 2013-07-22 Georg-Johann Lay <avr@gjlay.de>
13159 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
13160 instruction sequence is 1 byte shorter.
13162 2013-07-22 Uros Bizjak <ubizjak@gmail.com>
13164 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
13165 it is not needed after split.
13167 2013-07-20 Iain Sandoe <iain@codesourcery.com>
13170 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
13171 second label for nonlocal goto receivers. Don't output pic base labels
13172 unless we're producing PIC; mark that action unreachable().
13173 (ix86_save_reg): If the function contains a nonlocal label, save the
13175 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
13176 * config/darwin.c (emitted_pic_label_num): New GTY.
13177 (update_pic_label_number_if_needed): New.
13178 (machopic_output_function_base_name): Adjust for nonlocal receiver
13180 (machopic_should_output_picbase_label): New.
13181 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
13182 (nonlocal_goto_receiver): New insn and split.
13184 2013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
13186 * config/aarch64/aarch64-builtins.c
13187 (aarch64_fold_builtin): Fold abs in all modes.
13188 * config/aarch64/aarch64-simd-builtins.def
13189 (abs): Enable for all modes.
13190 * config/aarch64/arm_neon.h
13191 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
13192 (vabs_f64): Add missing intrinsic.
13194 2013-07-19 Ian Bolton <ian.bolton@arm.com>
13196 * config/aarch64/arm_neon.h (vabs_s64): New function
13198 2013-07-19 Georg-Johann Lay <avr@gjlay.de>
13201 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
13202 * config/avr/avr.md (adjust_len): Add `round'.
13203 * config/avr/avr-protos.h (avr_out_round): New prototype.
13204 (avr_out_plus): Add `out_label' argument.
13205 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
13206 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
13207 Handle the case where `insn' is just a pattern.
13208 (avr_out_bitop): Handle the case where `insn' is just a pattern.
13209 (avr_out_round): New function.
13210 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
13212 2013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
13214 * config/microblaze/microblaze.c (microblaze_expand_prologue):
13215 Add check for flag_stack_usage to handle -fstack-usage support
13217 2013-07-18 Pat Haugen <pthaugen@us.ibm.com>
13219 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
13220 interaction for new Power8 flags and VSX.
13222 2013-07-18 Sriraman Tallam <tmsriram@google.com>
13224 PR middle-end/57698
13225 * tree-inline.c (expand_call_inline): Emit errors during
13226 early_inlining only if optimization is not turned on.
13228 2013-07-18 David Malcolm <dmalcolm@redhat.com>
13232 * passes.c (init_optimization_passes): Move the construction of
13233 the pass hierarchy into a new passes.def file.
13235 * Makefile.in (passes.o): Add dependency on passes.def.
13237 2013-07-18 David Malcolm <dmalcolm@redhat.com>
13239 * passes.c (init_optimization_passes): Introduce macros for
13240 constructing the tree of passes (INSERT_PASSES_AFTER,
13241 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
13242 TERMINATE_PASS_LIST).
13244 2013-07-18 Vladimir Makarov <vmakarov@redhat.com>
13245 Wei Mi <wmi@google.com>
13247 PR rtl-optimization/57878
13248 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
13250 (reload_pseudo_compare_func): Check nregs first for reload
13253 2013-07-18 David Malcolm <dmalcolm@redhat.com>
13255 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
13257 2013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
13259 * read-rtl.c (validate_const_int): Once an invalid character is
13260 seen, quit the loop.
13262 * gengtype.c (write_roots): Similarly once we find the "deletable"
13263 or "if_marked" option.
13265 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
13267 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
13268 "xtab" and "sat". Move value "clz" from here to ...
13269 (attriubte "type"): ... here.
13270 (satsi_<SAT:code>): Delete "insn" attribute.
13271 (satsi_<SAT:code>_shift): Likewise.
13272 (arm_zero_extendqisi2addsi): Likewise.
13273 (arm_extendqisi2addsi): Likewise.
13274 (clzsi2): Update for attribute changes.
13275 (rbitsi2): Likewise.
13276 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
13278 (arm_usatsihi): Likewise.
13279 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
13281 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
13283 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
13284 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
13285 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
13286 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
13287 in alphabetical order.
13288 (attribute "core_cycles"): Update for attribute changes.
13289 (arm_addsi3): Likewise.
13290 (addsi3_compare0): Likewise.
13291 (addsi3_compare0_scratch): Likewise.
13292 (addsi3_compare_op1): Likewise.
13293 (addsi3_compare_op2): Likewise.
13294 (compare_addsi2_op0): Likewise.
13295 (compare_addsi2_op1): Likewise.
13296 (addsi3_carryin_shift_<optab>): Likewise.
13297 (subsi3_carryin_shift): Likewise.
13298 (rsbsi3_carryin_shift): Likewise.
13299 (arm_subsi3_insn): Likewise.
13300 (subsi3_compare0): Likewise.
13301 (subsi3_compare): Likewise.
13302 (arm_andsi3_insn): Likewise.
13303 (thumb1_andsi3_insn): Likewise.
13304 (andsi3_compare0): Likewise.
13305 (andsi3_compare0_scratch): Likewise.
13306 (zeroextractsi_compare0_scratch
13307 (andsi_not_shiftsi_si): Likewise.
13308 (iorsi3_insn): Likewise.
13309 (iorsi3_compare0): Likewise.
13310 (iorsi3_compare0_scratch): Likewise.
13311 (arm_xorsi3): Likewise.
13312 (thumb1_xorsi3_insn): Likewise.
13313 (xorsi3_compare0): Likewise.
13314 (xorsi3_compare0_scratch): Likewise.
13315 (satsi_<SAT:code>_shift): Likewise.
13317 (arm_shiftsi3): Likewise.
13318 (shiftsi3_compare0): Likewise.
13319 (not_shiftsi): Likewise.
13320 (not_shiftsi_compare0): Likewise.
13321 (not_shiftsi_compare0_scratch): Likewise.
13322 (arm_one_cmplsi2): Likewise.
13323 (thumb_one_complsi2): Likewise.
13324 (notsi_compare0): Likewise.
13325 (notsi_compare0_scratch): Likewise.
13326 (thumb1_zero_extendhisi2): Likewise.
13327 (arm_zero_extendhisi2): Likewise.
13328 (arm_zero_extendhisi2_v6): Likewise.
13329 (arm_zero_extendhisi2addsi): Likewise.
13330 (thumb1_zero_extendqisi2): Likewise.
13331 (thumb1_zero_extendqisi2_v6): Likewise.
13332 (arm_zero_extendqisi2): Likewise.
13333 (arm_zero_extendqisi2_v6): Likewise.
13334 (arm_zero_extendqisi2addsi): Likewise.
13335 (thumb1_extendhisi2): Likewise.
13336 (arm_extendhisi2): Likewise.
13337 (arm_extendhisi2_v6): Likewise.
13338 (arm_extendqisi): Likewise.
13339 (arm_extendqisi_v6): Likewise.
13340 (arm_extendqisi2addsi): Likewise.
13341 (thumb1_extendqisi2): Likewise.
13342 (thumb1_movdi_insn): Likewise.
13343 (arm_movsi_insn): Likewise.
13344 (movsi_compare0): Likewise.
13345 (movhi_insn_arch4): Likewise.
13346 (movhi_bytes): Likewise.
13347 (arm_movqi_insn): Likewise.
13348 (thumb1_movqi_insn): Likewise.
13349 (arm32_movhf): Likewise.
13350 (thumb1_movhf): Likewise.
13351 (arm_movsf_soft_insn): Likewise.
13352 (thumb1_movsf_insn): Likewise.
13353 (movdf_soft_insn): Likewise.
13354 (thumb_movdf_insn): Likewise.
13355 (arm_cmpsi_insn): Likewise.
13356 (cmpsi_shiftsi): Likewise.
13357 (cmpsi_shiftsi_swp): Likewise.
13358 (arm_cmpsi_negshiftsi_si): Likewise.
13359 (movsicc_insn): Likewise.
13360 (movsfcc_soft_insn): Likewise.
13361 (arith_shiftsi): Likewise.
13362 (arith_shiftsi_compare0
13363 (arith_shiftsi_compare0_scratch
13364 (sub_shiftsi): Likewise.
13365 (sub_shiftsi_compare0
13366 (sub_shiftsi_compare0_scratch
13367 (and_scc): Likewise.
13368 (cond_move): Likewise.
13369 (if_plus_move): Likewise.
13370 (if_move_plus): Likewise.
13371 (if_move_not): Likewise.
13372 (if_not_move): Likewise.
13373 (if_shift_move): Likewise.
13374 (if_move_shift): Likewise.
13375 (if_shift_shift): Likewise.
13376 (if_not_arith): Likewise.
13377 (if_arith_not): Likewise.
13378 (cond_move_not): Likewise.
13379 (thumb1_ashlsi3): Set type attribute.
13380 (thumb1_ashrsi3): Likewise.
13381 (thumb1_lshrsi3): Likewise.
13382 (thumb1_rotrsi3): Likewise.
13383 (shiftsi3_compare0_scratch): Likewise.
13384 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
13385 (neon_mov<mode>): Likewise.
13386 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
13388 (thumb2_movsi_insn): Likewise.
13389 (thumb2_cmpsi_neg_shiftsi): Likewise.
13390 (thumb2_extendqisi_v6): Likewise.
13391 (thumb2_zero_extendhisi2_v6): Likewise.
13392 (thumb2_zero_extendqisi2_v6): Likewise.
13393 (thumb2_shiftsi3_short): Likewise.
13394 (thumb2_addsi3_compare0_scratch): Likewise.
13395 (orsi_not_shiftsi_si): Likewise.
13396 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
13397 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
13399 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
13400 (1020alu_shift_op): Likewise.
13401 (1020alu_shift_reg_op): Likewise.
13402 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
13403 (alu_shift_op): Likewise.
13404 (alu_shift_reg_op): Likewise.
13405 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
13406 (11_alu_shift_op): Likewise.
13407 (11_alu_shift_reg_op): Likewise.
13408 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
13409 (9_alu_shift_reg_op): Likewise.
13410 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
13412 (cortex_a15_alu_shift): Likewise.
13413 (cortex_a15_alu_shift_reg): Likewise.
13414 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
13416 (cortex_a5_alu_shift): Likewise.
13417 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
13419 (cortex_a53_alu_shift): Likewise.
13420 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
13422 (cortex_a7_alu_reg): Likewise.
13423 (cortex_a7_alu_shift): Likewise.
13424 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
13426 (cortex_a8_alu_shift): Likewise.
13427 (cortex_a8_alu_shift_reg): Likewise.
13428 (cortex_a8_mov): Likewise.
13429 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
13430 (cortex_a9_dp_shift): Likewise.
13431 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
13433 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
13435 (cortex_r4_mov): Likewise.
13436 (cortex_r4_alu_shift): Likewise.
13437 (cortex_r4_alu_shift_reg): Likewise.
13438 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
13439 (526_alu_shift_op): Likewise.
13440 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
13441 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
13442 (626te_alu_shift_op): Likewise.
13443 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
13444 (726te_alu_op): Likewise.
13445 (726te_alu_shift_op): Likewise.
13446 (726te_alu_shift_reg_op): Likewise.
13447 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
13448 (mp626_alu_shift_op): Likewise.
13449 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
13450 (pj4_alu_e1_conds): Likewise.
13451 (pj4_alu): Likewise.
13452 (pj4_alu_conds): Likewise.
13453 (pj4_shift): Likewise.
13454 (pj4_shift_conds): Likewise.
13455 (pj4_alu_shift): Likewise.
13456 (pj4_alu_shift_conds): Likewise.
13457 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
13459 (cortexa7_older_only): Likewise.
13460 (cortexa7_younger): Likewise.
13462 2013-07-18 David Malcolm <dmalcolm@redhat.com>
13464 * ipa-pure-const.c (generate_summary): Rename to...
13465 (pure_const_generate_summary): ... this.
13467 2013-07-17 Iain Sandoe <iain@codesourcery.com>
13469 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
13471 2013-07-17 Yvan Roux <yvan.roux@linaro.org>
13474 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
13477 2013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13479 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
13480 enabled without -march=zEC12.
13481 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
13484 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
13486 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
13487 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
13488 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
13489 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
13490 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
13491 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
13492 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
13494 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
13495 and ISA_HAS_NMADD3_NMSUB3.
13496 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
13497 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
13498 (nmsub4<mode>, nmsub3<mode>): Likewise.
13499 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
13501 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
13503 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
13504 TARGET_MIPS5400 checking.
13506 2013-07-16 Jakub Jelinek <jakub@redhat.com>
13507 Peter Bergner <bergner@vnet.ibm.com>
13509 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
13510 registers in the comment.
13511 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
13512 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
13513 rather than FIRST_PSEUDO_REGISTERS.
13515 2013-07-16 Peter Bergner <bergner@vnet.ibm.com>
13517 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
13518 enable extra ISA flags with TARGET_HTM.
13520 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
13522 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
13525 2013-07-15 Cong Hou <congh@google.com>
13527 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
13528 in compare function for sorting.
13530 2013-07-15 Peter Bergner <bergner@vnet.ibm.com>
13532 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
13533 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
13534 * config/rs6000/rs6000.opt: Add -mhtm option.
13535 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
13536 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
13537 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
13538 __HTM__ if the HTM instructions are available.
13539 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
13540 htm_spr_reg_operand): New define_predicates.
13541 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
13542 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
13544 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
13545 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
13546 HTM builtin functions.
13547 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
13548 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
13549 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
13550 (rs6000_builtin_mask_calculate): Likewise.
13551 (rs6000_option_override_internal): Likewise.
13552 (bdesc_htm): Add new HTM builtin support.
13553 (htm_spr_num): New function.
13554 (htm_spr_regno): Likewise.
13555 (rs6000_htm_spr_icode): Likewise.
13556 (htm_expand_builtin): Likewise.
13557 (htm_init_builtins): Likewise.
13558 (rs6000_expand_builtin): Add support for HTM builtin functions.
13559 (rs6000_init_builtins): Likewise.
13560 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
13562 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
13563 (TARGET_HTM, MASK_HTM): Define macros.
13564 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
13565 (FIXED_REGISTERS): Likewise.
13566 (CALL_USED_REGISTERS): Likewise.
13567 (CALL_REALLY_USED_REGISTERS): Likewise.
13568 (REG_ALLOC_ORDER): Likewise.
13569 (enum reg_class): Likewise.
13570 (REG_CLASS_NAMES): Likewise.
13571 (REG_CLASS_CONTENTS): Likewise.
13572 (REGISTER_NAMES): Likewise.
13573 (ADDITIONAL_REGISTER_NAMES): Likewise.
13574 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
13575 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
13576 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
13577 * config/rs6000/htm.md: New file.
13578 * config/rs6000/htmintrin.h: New file.
13579 * config/rs6000/htmxlintrin.h: New file.
13581 2013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
13583 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
13584 Define SYMBOL_TINY_GOT, update comment.
13585 * config/aarch64/aarch64.c
13586 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
13587 (aarch64_expand_mov_immediate): Likewise.
13588 (aarch64_print_operand): Likewise.
13589 (aarch64_classify_symbol): Likewise.
13590 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
13591 (ldr_got_tiny): Define.
13593 2013-07-13 Tobias Grosser <tobias@grosser.es>
13595 PR tree-optimization/54094
13596 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
13597 scheduling dimension for the parallelism check from the polyhedral
13598 information in the AST.
13599 * graphite-dependences.c (carries_deps): Do not assume the schedule is
13602 2013-07-13 Jason Merrill <jason@redhat.com>
13604 * print-tree.c (debug_vec_tree): Use debug_raw.
13605 (debug_raw (vec<tree, va_gc> &)): New.
13606 (debug_raw (vec<tree, va_gc> *)): New.
13607 * tree.h: Declare them.
13609 2013-07-13 Bin Cheng <bin.cheng@arm.com>
13611 * ifcvt.c (ifcvt_after_combine): New static variable.
13612 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
13614 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
13615 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
13616 rest_of_handle_if_after_reload): Pass new argument for if_convert.
13618 2013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
13620 * config/mips/mips.c (mips_expand_call): Remove empty statement.
13622 2013-07-12 Michael Matz <matz@suse.de>
13624 PR middle-end/55771
13625 * convert.c (convert_to_real): Reject non-float inner types.
13627 2013-07-12 Tejas Belagod <tejas.belagod@arm.com>
13629 * config/aarch64/aarch64-protos.h
13630 (aarch64_simd_immediate_valid_for_move): Remove.
13631 * config/aarch64/aarch64.c (simd_immediate_info): New member.
13632 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
13634 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
13636 2013-07-11 Steve Ellcey <sellcey@mips.com>
13638 * config/mips/mips.c (mips_conditional_register_usage): Do not
13639 use t[0-7] registers in MIPS16 mode when optimizing for size.
13641 2013-07-11 Sriraman Tallam <tmsriram@google.com>
13643 * config/i386/i386.c (dispatch_function_versions): Fix array
13644 indexing of function_version_info to match actual_versions.
13646 2013-07-11 Teresa Johnson <tejohnson@google.com>
13648 * vec.h (struct va_gc): Move release out-of-line.
13649 (va_gc::release): Call ggc_free on released vec.
13651 2013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
13653 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
13654 Require GOT register as additional operand in UNSPEC.
13655 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
13656 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
13657 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
13658 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
13659 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
13660 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
13661 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
13663 2013-07-11 Georg-Johann Lay <avr@gjlay.de>
13666 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
13667 name seen by assembler/linker rather if available.
13669 2013-07-11 Andreas Schwab <schwab@suse.de>
13671 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
13673 2013-07-10 Vladimir Makarov <vmakarov@redhat.com>
13675 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
13677 2013-07-10 Joseph Myers <joseph@codesourcery.com>
13679 * doc/tm.texi.in: Move hook documentation to ....
13680 * target.def: ... here.
13682 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
13683 text on @hook line.
13684 * doc/tm.texi: Regenerate.
13686 2013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
13689 * doc/invoke.texi: Document -Wconditionally-supported.
13691 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
13694 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
13697 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
13700 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
13701 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
13702 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
13703 Remove duplicate devices.
13704 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
13705 * config/avr/t-multilib: Regenerate.
13706 * config/avr/avr-tables.opt: Regenerate.
13707 * doc/avr-mmcu.texi: Regenerate.
13709 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
13712 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
13714 2013-07-10 Graham Stott <graham.stott@btinternet.com>
13716 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
13717 the cost of MULT when optimizing for size.
13719 2013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
13721 * config/cr16/cr16-protos.h: Don't include target.h.
13723 2013-07-09 Joseph Myers <joseph@codesourcery.com>
13725 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
13726 adjust register size for TDmode and TFmode for VSX registers.
13728 2013-07-08 Kai Tietz <ktietz@redhat.com>
13731 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
13732 hook_bool_const_tree_true.
13734 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13736 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
13737 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
13738 * config/s390/s390.md: Define FPR*_REGNUM constants.
13739 Fix FPR2_REGNUM constant (18 -> 17).
13740 ("*trunc<BFP:mode><DFP_ALL:mode>2")
13741 ("*trunc<DFP_ALL:mode><BFP:mode>2")
13742 ("trunc<BFP:mode><DFP_ALL:mode>2")
13743 ("trunc<DFP_ALL:mode><BFP:mode>2")
13744 ("*extend<BFP:mode><DFP_ALL:mode>2")
13745 ("*extend<DFP_ALL:mode><BFP:mode>2")
13746 ("extend<BFP:mode><DFP_ALL:mode>2")
13747 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
13750 2013-07-08 Graham Stott <graham.stott@btinternet.com>
13752 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
13754 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13756 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
13757 and cfun_fpr_bit_p to cfun_fpr_save_p.
13758 (s390_frame_area, s390_register_info, s390_frame_info)
13759 (s390_emit_prologue, s390_emit_epilogue)
13760 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
13762 * config/s390/s390.h: Define *_REGNUM macros for floating point
13765 2013-07-08 Eric Botcazou <ebotcazou@adacore.com>
13767 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
13769 2013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
13771 PR rtl-optimization/57786
13772 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
13773 and break out of the loop when it is set to false.
13775 2013-07-08 Jakub Jelinek <jakub@redhat.com>
13778 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
13779 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
13780 (const_int 63)) 0)).
13781 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
13782 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
13783 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
13785 PR rtl-optimization/57829
13786 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
13787 mask bits outside of mode are just sign-extension from mode to HWI.
13789 2013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
13791 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
13792 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
13793 adjust_address instead of change_address to keep info about alignment.
13794 (emit_strmov): Remove.
13795 (emit_memmov): New function.
13796 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
13797 (expand_movmem_epilogue): Likewise and return updated rtx for
13799 (expand_constant_movmem_prologue): Likewise and return updated rtx for
13800 destination and source.
13801 (decide_alignment): Refactor, handle vector_loop.
13802 (ix86_expand_movmem): Likewise.
13803 (ix86_expand_setmem): Likewise.
13804 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
13806 2013-07-07 Uros Bizjak <ubizjak@gmail.com>
13808 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
13809 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
13811 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
13813 * config/i386/sse.md (sse_movlhps): Change alternative 3
13814 of operand 2 to "m".
13816 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
13819 * config/i386/sse.md (iptr): New mode attribute.
13820 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
13821 (<sse>_vm<plusminus_insn><mode>3): Ditto.
13822 (<sse>_vmmul<mode>3): Ditto.
13823 (<sse>_vmdiv<mode>3): Ditto.
13824 (sse_vmrcpv4sf2): Ditto.
13825 (<sse>_vmsqrt<mode>2): Ditto.
13826 (sse_vmrsqrtv4sf2): Ditto.
13827 (<sse>_vm<code><mode>3): Ditto.
13828 (avx_vmcmp<mode>3): Ditto.
13829 (<sse>_vmmaskcmp<mode>3): Ditto.
13830 (<sse>_comi): Ditto.
13831 (<sse>_ucomi): Ditto.
13832 (*xop_vmfrcz_<mode>): Ditto.
13833 (*fmai_fmadd_<mode>): Ditto.
13834 (*fmai_fmsub_<mode>): Ditto.
13835 (*fmai_fnmadd_<mode>): Ditto.
13836 (*fmai_fnmsub_<mode>): Ditto.
13837 (*fma4i_vmfmadd_<mode>): Ditto.
13838 (*fma4i_vmfmsub_<mode>): Ditto.
13839 (*fma4i_vmfnmadd_<mode>): Ditto.
13840 (*fma4i_vmfnmsub_<mode>): Ditto.
13841 (*xop_vmfrcz_<mode>): Ditto.
13842 (sse_cvtps2pi): Ditto.
13843 (sse_cvttps2pi): Ditto.
13844 (sse_cvtss2si): Ditto.
13845 (sse_cvtss2si_2): Ditto.
13846 (sse_cvtss2siq_2): Ditto.
13847 (sse_cvttss2si): Ditto.
13848 (sse_cvttss2siq): Ditto.
13849 (sse_cvtsd2si): Ditto.
13850 (sse_cvtsd2si_2): Ditto.
13851 (sse_cvtsd2siq_2): Ditto.
13852 (sse_cvttsd2si): Ditto.
13853 (sse_cvttsd2siq): Ditto.
13854 (sse_cvtsd2ss): Ditto.
13855 (sse_cvtss2sd): Ditto.
13856 (avx2_pbroadcast<mode>): Ditto.
13857 (avx2_pbroadcast<mode>_1): Ditto.
13858 (*avx_vperm_broadcast_v4sf): Ditto.
13860 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
13861 (sse_movlhps): Ditto.
13862 (sse_storehps): Ditto.
13863 (sse_loadhps): Ditto.
13864 (sse_storelps): Ditto.
13865 (sse_loadlps): Ditto.
13866 (*vec_concatv4sf): Ditto.
13867 (*vec_interleave_highv2df): Ditto.
13868 (*vec_interleave_lowv2df): Ditto.
13869 (*vec_extractv2df_1_sse): Ditto.
13870 (*vec_extractv2df_0_sse): Ditto.
13871 (sse2_storelpd): Ditto.
13872 (sse2_loadlpd): Ditto.
13873 (sse2_movsd): Ditto.
13874 (*vec_concatv4si): Ditto.
13875 (vec_concatv2di): Ditto.
13877 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
13878 for Intel asm dialect.
13879 (mmx_punpcklwd): Ditto.
13880 (mmx_punpckldq): Ditto.
13882 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
13883 for intel assembler dialect.
13885 2013-07-06 Jakub Jelinek <jakub@redhat.com>
13888 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
13889 for BUILT_IN_C{LZ,LRSB}*.
13890 * tree.h (CASE_INT_FN): Add FN##IMAX case.
13891 * tree-vrp.c (extract_range_basic): Handle
13892 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
13893 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
13894 fall thru to code calling set_value*.
13895 * builtins.c (expand_builtin): Remove *IMAX cases.
13896 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
13897 if width is bigger than 2*HWI.
13899 2013-07-05 Vladimir Makarov <vmakarov@redhat.com>
13901 PR rtl-optimization/55342
13902 * lra-int.h (lra_subreg_reload_pseudos): New.
13903 * lra.c: Add undoing optional reloads to the block diagram.
13904 (lra_subreg_reload_pseudos): New.
13905 (lra_optional_reload_pseudos): Change comments.
13906 (lra): Init and clear lra_subreg_reload_pseudos. Clear
13907 lra_optional_reload_pseudos after undo transformations.
13908 * lra-assigns.c (pseudo_prefix_title): New.
13909 (lra_setup_reg_renumber): Use it.
13910 (spill_for): Ditto. Check subreg reload pseudos too.
13911 (assign_by_spills): Consider subreg reload pseudos too.
13912 * lra-constraints.c (simplify_operand_subreg): Use
13913 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
13914 (curr_insn_transform): Recognize and do optional reloads.
13915 (undo_optional_reloads): New.
13916 (lra_undo_inheritance): Call undo_optional_reloads.
13918 2013-07-05 Thomas Quinot <quinot@adacore.com>
13920 * tree-complex.c (expand_complex_operations_1): Fix typo.
13922 2013-07-04 Tejas Belagod <tejas.belagod@arm.com>
13924 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
13925 (tune_params): New member 'const vec_costs'.
13926 * config/aarch64/aarch64.c (generic_vector_cost): New.
13927 (generic_tunings): New member 'generic_vector_cost'.
13928 (aarch64_builtin_vectorization_cost): New.
13929 (aarch64_add_stmt_cost): New.
13930 (TARGET_VECTORIZE_ADD_STMT_COST): New.
13931 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
13933 2013-07-03 Jakub Jelinek <jakub@redhat.com>
13936 * config/i386/predicates.md (vsib_address_operand): Disallow
13937 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
13939 2013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
13941 PR middle-end/55030
13942 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
13943 expand_builtin_setjmp_receiver.
13944 (expand_label): Adjust, call expand_builtin_setjmp_receiver
13945 with NULL for the label parameter.
13946 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
13947 the frame-pointer. Adjust comments.
13948 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
13949 only if LABEL is non-NULL.
13951 2013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
13953 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
13954 (ARM_ABI_AAPCS64): Ditto.
13956 (ARM_DEFAULT_ABI): Ditto.
13958 2013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
13960 * config/aarch64/aarch64-builtins.c
13961 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
13962 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
13964 * config/aarch64/aarch64-simd.md
13965 (aarch64_ld1<VALL:mode>): New.
13966 (aarch64_st1<VALL:mode>): Likewise.
13967 * config/aarch64/arm_neon.h
13968 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
13970 2013-07-02 Sriraman Tallam <tmsriram@google.com>
13972 * config/i386/i386.c (gate_insert_vzeroupper): Check if
13974 (ix86_option_override_internal):Turn on all -mavx target flags by
13975 default as they are dependent on AVX anyway.
13977 2013-07-02 Cary Coutant <ccoutant@google.com>
13979 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
13980 deterministic hash.
13981 (loc_checksum_ordered): Likewise.
13982 (hash_loc_operands): Remove inline keyword.
13984 2013-07-02 Jakub Jelinek <jakub@redhat.com>
13986 PR tree-optimization/57741
13987 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
13988 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
13989 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
13990 Allow REAL_CST step_exprs if flag_associative_math.
13991 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
13993 2013-07-02 Ian Bolton <ian.bolton@arm.com>
13995 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
13997 2013-07-02 Ian Bolton <ian.bolton@arm.com>
13999 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
14001 2013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14003 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
14005 (iorsi3_insn): Likewise.
14006 (arm_xorsi3): Likewise.
14008 2013-07-01 Sofiane Naci <sofiane.naci@arm.com>
14010 * arm.md (attribute "wtype"): Delete. Move attribute values from here
14012 (attribute "type"): ... here, and prefix with "wmmx_".
14013 (attribute "core_cycles"): Update for attribute changes.
14014 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
14015 (tbcstv4hi): Likewise.
14016 (tbcstv2si): Likewise.
14017 (iwmmxt_iordi3): Likewise.
14018 (iwmmxt_xordi3): Likewise.
14019 (iwmmxt_anddi3): Likewise.
14020 (iwmmxt_nanddi3): Likewise.
14021 (iwmmxt_arm_movdi): Likewise.
14022 (iwmmxt_movsi_insn): Likewise.
14023 (mov<mode>_internal): Likewise.
14024 (and<mode>3_iwmmxt): Likewise.
14025 (ior<mode>3_iwmmxt): Likewise.
14026 (xor<mode>3_iwmmxt): Likewise.
14027 (add<mode>3_iwmmxt): Likewise.
14028 (ssaddv8qi3): Likewise.
14029 (ssaddv4hi3): Likewise.
14030 (ssaddv2si3): Likewise.
14031 (usaddv8qi3): Likewise.
14032 (usaddv4hi3): Likewise.
14033 (usaddv2si3): Likewise.
14034 (sub<mode>3_iwmmxt): Likewise.
14035 (sssubv8qi3): Likewise.
14036 (sssubv4hi3): Likewise.
14037 (sssubv2si3): Likewise.
14038 (ussubv8qi3): Likewise.
14039 (ussubv4hi3): Likewise.
14040 (ussubv2si3): Likewise.
14041 (mulv4hi3_iwmmxt): Likewise.
14042 (smulv4hi3_highpart): Likewise.
14043 (umulv4hi3_highpart): Likewise.
14044 (iwmmxt_wmacs): Likewise.
14045 (iwmmxt_wmacsz): Likewise.
14046 (iwmmxt_wmacu): Likewise.
14047 (iwmmxt_wmacuz): Likewise.
14048 (iwmmxt_clrdi): Likewise.
14049 (iwmmxt_clrv8qi): Likewise.
14050 (iwmmxt_clr4hi): Likewise.
14051 (iwmmxt_clr2si): Likewise.
14052 (iwmmxt_uavgrndv8qi3): Likewise.
14053 (iwmmxt_uavgrndv4hi3): Likewise.
14054 (iwmmxt_uavgv8qi3): Likewise.
14055 (iwmmxt_uavgv4hi3): Likewise.
14056 (iwmmxt_tinsrb): Likewise.
14057 (iwmmxt_tinsrh): Likewise.
14058 (iwmmxt_tinsrw): Likewise.
14059 (iwmmxt_textrmub): Likewise.
14060 (iwmmxt_textrmsb): Likewise.
14061 (iwmmxt_textrmuh): Likewise.
14062 (iwmmxt_textrmsh): Likewise.
14063 (iwmmxt_textrmw): Likewise.
14064 (iwmxxt_wshufh): Likewise.
14065 (eqv8qi3): Likewise.
14066 (eqv4hi3): Likewise.
14067 (eqv2si3): Likewise.
14068 (gtuv8qi3): Likewise.
14069 (gtuv4hi3): Likewise.
14070 (gtuv2si3): Likewise.
14071 (gtv8qi3): Likewise.
14072 (gtv4hi3): Likewise.
14073 (gtv2si3): Likewise.
14074 (smax<mode>3_iwmmxt): Likewise.
14075 (umax<mode>3_iwmmxt): Likewise.
14076 (smin<mode>3_iwmmxt): Likewise.
14077 (umin<mode>3_iwmmxt): Likewise.
14078 (iwmmxt_wpackhss): Likewise.
14079 (iwmmxt_wpackwss): Likewise.
14080 (iwmmxt_wpackdss): Likewise.
14081 (iwmmxt_wpackhus): Likewise.
14082 (iwmmxt_wpackwus): Likewise.
14083 (iwmmxt_wpackdus): Likewise.
14084 (iwmmxt_wunpckihb): Likewise.
14085 (iwmmxt_wunpckihh): Likewise.
14086 (iwmmxt_wunpckihw): Likewise.
14087 (iwmmxt_wunpckilb): Likewise.
14088 (iwmmxt_wunpckilh): Likewise.
14089 (iwmmxt_wunpckilw): Likewise.
14090 (iwmmxt_wunpckehub): Likewise.
14091 (iwmmxt_wunpckehuh): Likewise.
14092 (iwmmxt_wunpckehuw): Likewise.
14093 (iwmmxt_wunpckehsb): Likewise.
14094 (iwmmxt_wunpckehsh): Likewise.
14095 (iwmmxt_wunpckehsw): Likewise.
14096 (iwmmxt_wunpckelub): Likewise.
14097 (iwmmxt_wunpckeluh): Likewise.
14098 (iwmmxt_wunpckeluw): Likewise.
14099 (iwmmxt_wunpckelsb): Likewise.
14100 (iwmmxt_wunpckelsh): Likewise.
14101 (iwmmxt_wunpckelsw): Likewise.
14102 (ror<mode>3): Likewise.
14103 (ashr<mode>3_iwmmxt): Likewise.
14104 (lshr<mode>3_iwmmxt): Likewise.
14105 (ashl<mode>3_iwmmxt): Likewise.
14106 (ror<mode>3_di): Likewise.
14107 (ashr<mode>3_di): Likewise.
14108 (lshr<mode>3_di): Likewise.
14109 (ashl<mode>3_di): Likewise.
14110 (iwmmxt_wmadds): Likewise.
14111 (iwmmxt_wmaddu): Likewise.
14112 (iwmmxt_tmia): Likewise.
14113 (iwmmxt_tmiaph): Likewise.
14114 (iwmmxt_tmiabb): Likewise.
14115 (iwmmxt_tmiatb): Likewise.
14116 (iwmmxt_tmiabt): Likewise.
14117 (iwmmxt_tmiatt): Likewise.
14118 (iwmmxt_tmovmskb): Likewise.
14119 (iwmmxt_tmovmskh): Likewise.
14120 (iwmmxt_tmovmskw): Likewise.
14121 (iwmmxt_waccb): Likewise.
14122 (iwmmxt_wacch): Likewise.
14123 (iwmmxt_waccw): Likewise.
14124 (iwmmxt_waligni): Likewise.
14125 (iwmmxt_walignr): Likewise.
14126 (iwmmxt_walignr0): Likewise.
14127 (iwmmxt_walignr1): Likewise.
14128 (iwmmxt_walignr2): Likewise.
14129 (iwmmxt_walignr3): Likewise.
14130 (iwmmxt_wsadb): Likewise.
14131 (iwmmxt_wsadh): Likewise.
14132 (iwmmxt_wsadbz): Likewise.
14133 (iwmmxt_wsadhz): Likewise.
14134 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
14135 (iwmmxt_wabsdiffb): Likewise.
14136 (iwmmxt_wabsdiffh): Likewise.
14137 (iwmmxt_wabsdiffw): Likewise.
14138 (iwmmxt_waddsubhx): Likewise
14139 (iwmmxt_wsubaddhx): Likewise.
14140 (addc<mode>3): Likewise.
14141 (iwmmxt_avg4): Likewise.
14142 (iwmmxt_avg4r): Likewise.
14143 (iwmmxt_wmaddsx): Likewise.
14144 (iwmmxt_wmaddux): Likewise.
14145 (iwmmxt_wmaddsn): Likewise.
14146 (iwmmxt_wmaddun): Likewise.
14147 (iwmmxt_wmulwsm): Likewise.
14148 (iwmmxt_wmulwum): Likewise.
14149 (iwmmxt_wmulsmr): Likewise.
14150 (iwmmxt_wmulumr): Likewise.
14151 (iwmmxt_wmulwsmr): Likewise.
14152 (iwmmxt_wmulwumr): Likewise.
14153 (iwmmxt_wmulwl): Likewise.
14154 (iwmmxt_wqmulm): Likewise.
14155 (iwmmxt_wqmulwm): Likewise.
14156 (iwmmxt_wqmulmr): Likewise.
14157 (iwmmxt_wqmulwmr): Likewise.
14158 (iwmmxt_waddbhusm): Likewise.
14159 (iwmmxt_waddbhusl): Likewise.
14160 (iwmmxt_wqmiabb): Likewise.
14161 (iwmmxt_wqmiabt): Likewise.
14162 (iwmmxt_wqmiatb): Likewise.
14163 (iwmmxt_wqmiatt): Likewise.
14164 (iwmmxt_wqmiabbn): Likewise.
14165 (iwmmxt_wqmiabtn): Likewise.
14166 (iwmmxt_wqmiatbn): Likewise.
14167 (iwmmxt_wqmiattn): Likewise.
14168 (iwmmxt_wmiabb): Likewise.
14169 (iwmmxt_wmiabt): Likewise.
14170 (iwmmxt_wmiatb): Likewise.
14171 (iwmmxt_wmiatt): Likewise.
14172 (iwmmxt_wmiabbn): Likewise.
14173 (iwmmxt_wmiabtn): Likewise.
14174 (iwmmxt_wmiatbn): Likewise.
14175 (iwmmxt_wmiattn): Likewise.
14176 (iwmmxt_wmiawbb): Likewise.
14177 (iwmmxt_wmiawbt): Likewise.
14178 (iwmmxt_wmiawtb): Likewise.
14179 (iwmmxt_wmiawtt): Likewise.
14180 (iwmmxt_wmiawbbn): Likewise.
14181 (iwmmxt_wmiawbtn): Likewise.
14182 (iwmmxt_wmiawtbn): Likewise.
14183 (iwmmxt_wmiawttn): Likewise.
14184 (iwmmxt_wmerge): Likewise.
14185 (iwmmxt_tandc<mode>3): Likewise.
14186 (iwmmxt_torc<mode>3): Likewise.
14187 (iwmmxt_torvsc<mode>3): Likewise.
14188 (iwmmxt_textrc<mode>3): Likewise.
14189 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
14190 (wmmxt_pack): Likewise.
14191 (wmmxt_mult_c1): Likewise.
14192 (wmmxt_mult_c2): Likewise.
14193 (wmmxt_alu_c1): Likewise.
14194 (wmmxt_alu_c2): Likewise.
14195 (wmmxt_alu_c3): Likewise.
14196 (wmmxt_transfer_c1): Likewise.
14197 (wmmxt_transfer_c2): Likewise.
14198 (wmmxt_transfer_c3): Likewise.
14199 (marvell_f_iwmmxt_wstr): Likewise.
14200 (marvell_f_iwmmxt_wldr): Likewise.
14202 2013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
14204 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
14206 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
14209 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
14210 * lra-constraints.c (need_for_split_p): Check call used hard regs
14211 living through calls.
14213 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
14214 call used regs for call insn.
14216 2013-06-28 Jakub Jelinek <jakub@redhat.com>
14219 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
14220 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
14222 2013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
14224 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
14226 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
14228 * lra-constraints.c (need_for_split_p): Check call used hard regs
14229 living through calls.
14231 2013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
14234 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
14235 to tie with any other modes. Eliminate Altivec vector mode tests,
14236 since these are a subset of ALTIVEC or VSX vector modes. Simplify
14237 code, to return 0 if testing MODE2 for a condition, if we've
14238 already tested MODE1 for the same condition.
14240 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
14242 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
14245 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
14247 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
14248 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
14250 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
14252 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
14254 (aarch64_symbolic_constant_p): Remove.
14255 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
14256 static. Fix line length and white space.
14257 (aarch64_symbolic_constant_p): Remove.
14258 * config/aarch64/predicates.md (aarch64_valid_symref):
14259 Use aarch64_classify_symbol_expression.
14261 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14263 * config/arm/constraints.md (Ts): New constraint.
14264 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
14266 (compare_scc): Use "Ts" constraint for operand 0.
14267 (ior_scc_scc): Likewise.
14268 (and_scc_scc): Likewise.
14269 (and_scc_scc_nodom): Likewise.
14270 (ior_scc_scc_cmp): Likewise for operand 7.
14271 (and_scc_scc_cmp): Likewise.
14272 * config/arm/thumb2.md (thumb2_movsi_insn):
14273 Add alternatives for 16-bit encodings.
14274 (thumb2_movhi_insn): Likewise.
14275 (thumb2_movsicc_insn): Likewise.
14276 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
14277 (thumb2_negscc): Use "Ts" constraint.
14278 Move mvn instruction outside cond_exec block.
14279 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
14280 for 16-bit encodings.
14282 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14284 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
14286 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
14287 (mulsi3subsi): Likewise.
14288 (mulsidi3adddi): Likewise.
14289 (mulsidi3_v6): Likewise.
14290 (umulsidi3_v6): Likewise.
14291 (umulsidi3adddi_v6): Likewise.
14292 (smulsi3_highpart_v6): Likewise.
14293 (umulsi3_highpart_v6): Likewise.
14294 (mulhisi3tb): Likewise.
14295 (mulhisi3bt): Likewise.
14296 (mulhisi3tt): Likewise.
14297 (maddhisi4): Likewise.
14298 (maddhisi4tb): Likewise.
14299 (maddhisi4tt): Likewise.
14300 (maddhidi4): Likewise.
14301 (maddhidi4tb): Likewise.
14302 (maddhidi4tt): Likewise.
14303 (zeroextractsi_compare0_scratch): Likewise.
14304 (insv_zero): Likewise.
14305 (insv_t2): Likewise.
14306 (anddi_notzesidi_di): Likewise.
14307 (anddi_notsesidi_di): Likewise.
14308 (andsi_notsi_si): Likewise.
14309 (iordi_zesidi_di): Likewise.
14310 (xordi_zesidi_di): Likewise.
14311 (andsi_iorsi3_notsi): Likewise.
14312 (smax_0): Likewise.
14313 (smax_m1): Likewise.
14314 (smin_0): Likewise.
14315 (not_shiftsi): Likewise.
14316 (unaligned_loadsi): Likewise.
14317 (unaligned_loadhis): Likewise.
14318 (unaligned_loadhiu): Likewise.
14319 (unaligned_storesi): Likewise.
14320 (unaligned_storehi): Likewise.
14321 (extv_reg): Likewise.
14322 (extzv_t2): Likewise.
14323 (divsi3): Likewise.
14324 (udivsi3): Likewise.
14325 (arm_zero_extendhisi2addsi): Likewise.
14326 (arm_zero_extendqisi2addsi): Likewise.
14327 (compareqi_eq0): Likewise.
14328 (arm_extendhisi2_v6): Likewise.
14329 (arm_extendqisi2addsi): Likewise.
14330 (arm_movt): Likewise.
14331 (thumb2_ldrd): Likewise.
14332 (thumb2_ldrd_base): Likewise.
14333 (thumb2_ldrd_base_neg): Likewise.
14334 (thumb2_strd): Likewise.
14335 (thumb2_strd_base): Likewise.
14336 (thumb2_strd_base_neg): Likewise.
14337 (arm_negsi2): Add alternative for 16-bit encoding.
14338 (arm_one_cmplsi2): Likewise.
14340 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14342 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
14343 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
14344 (movdfcc): Likewise.
14345 * config/arm/vfp.md (*thumb2_movsf_vfp):
14346 Disable predication for arm_restrict_it.
14347 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
14348 (*thumb2_movdfcc_vfp): Likewise.
14349 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
14350 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
14351 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
14352 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
14353 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
14354 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
14355 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
14356 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
14357 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
14358 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
14359 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
14360 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
14361 Disable predication for arm_restrict_it.
14363 2013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
14365 * config/i386/bmiintrin.h (_bextr_u32): New.
14366 (_bextr_u64): Ditto.
14368 2013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
14370 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
14371 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
14372 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
14373 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
14374 * config/mips/n32-elf.h: ...this new file.
14376 2013-06-27 Marc Glisse <marc.glisse@inria.fr>
14379 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
14380 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
14382 2013-06-27 Catherine Moore <clm@codesourcery.com>
14384 * config/mips/mips-tables.opt: Regenerate.
14385 * config/mips/mips-cpus.def: Add m14ke and m14kec.
14386 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
14387 * doc/invoke.texi: Add -m14kc.
14389 2013-06-27 Jakub Jelinek <jakub@redhat.com>
14392 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
14393 constraints of operand 1 and 2.
14396 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
14397 to match RTL canonicalization. Swap predicates and
14398 constraints of operand 1 and 2.
14400 2013-06-27 Vladimir Makarov <vmakarov@redhat.com>
14402 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
14403 Process OP_INOUT regs for splitting too.
14405 2013-06-27 Jakub Jelinek <jakub@redhat.com>
14407 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
14408 decl before the loop, initialize to NULL.
14409 (vectorizable_load): Initialize ptr_incr to NULL.
14411 2013-06-27 Martin Jambor <mjambor@suse.cz>
14414 * ipa-ref.h (ipa_maybe_record_reference): Declare.
14415 * ipa-ref.c (ipa_maybe_record_reference): New function.
14416 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
14417 * ipa-cp.c (create_specialized_node): Record potential references from
14419 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
14421 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
14423 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
14424 parameter 'mode' of type 'enum machine_mode mode'; change to pass
14425 'mode' to force_reg.
14426 (aarch64_add_offset): Update calls to aarch64_force_temporary.
14427 (aarch64_expand_mov_immediate): Likewise.
14429 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
14431 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
14432 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
14434 2013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14436 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
14437 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
14438 (struct machine_function): Add tbegin_p.
14439 (s390_canonicalize_comparison): Fold CC mode compares to
14440 conditional jump if possible.
14441 (s390_emit_jump): Return the emitted jump.
14442 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
14443 Handle CCRAWmode compares.
14444 (s390_option_override): Default to -mhtm if available.
14445 (s390_reg_clobbered_rtx): Handle floating point regs as well.
14446 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
14447 FPRs instead of df_regs_ever_live_p.
14448 (s390_optimize_nonescaping_tx): New function.
14449 (s390_init_frame_layout): Extend clobbered_regs array to cover
14451 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
14452 (s390_expand_tbegin): New function.
14453 (enum s390_builtin): New enum definition.
14454 (code_for_builtin): New array definition.
14455 (s390_init_builtins): New function.
14456 (s390_expand_builtin): New function.
14457 (TARGET_INIT_BUILTINS): Define.
14458 (TARGET_EXPAND_BUILTIN): Define.
14459 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
14460 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
14461 (s390_alc_comparison): Likewise.
14462 * config/s390/s390-modes.def: Add CCRAWmode.
14463 * config/s390/s390.h (processor_flags): Add PF_TX.
14464 (TARGET_CPU_HTM): Define macro.
14465 (TARGET_HTM): Define macro.
14466 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
14467 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
14468 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
14469 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
14470 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
14471 (TBEGIN_MASK, TBEGINC_MASK): New constants.
14472 ("*cc_to_int"): Move up.
14473 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
14474 constants other than 0.
14475 ("*ccraw_to_int"): New insn and splitter definition.
14476 ("tbegin", "tbegin_nofloat", "tbegin_retry")
14477 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
14478 ("tx_assist"): New expander.
14479 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
14480 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
14481 * config/s390/s390.opt: Add -mhtm option.
14482 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
14483 * config/s390/htmxlintrin.h: New file.
14484 * config/s390/htmintrin.h: New file.
14485 * config/s390/s390intrin.h: New file.
14486 * doc/extend.texi: Document htm builtins.
14487 * config.gcc: Add the new header files to extra_headers.
14489 2013-06-26 Thomas Schwinge <thomas@codesourcery.com>
14491 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
14492 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
14494 2013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
14495 Pat Haugen <pthaugen@us.ibm.com>
14496 Peter Bergner <bergner@vnet.ibm.com>
14498 * config/rs6000/power8.md: New.
14499 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
14500 setting for power8 entry.
14501 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
14502 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
14503 test for Power4/Power5 only.
14504 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
14506 (force_new_group): Adjust comment.
14507 * config/rs6000/rs6000.md: Include power8.md.
14509 2013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
14511 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
14512 * config/arm/arm-protos.h (arm_max_conditional_execute): New
14514 (tune_params): Update comment.
14515 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
14516 (arm_max_conditional_execute): New function.
14517 (thumb2_final_prescan_insn): Use max_insn_skipped and
14518 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
14520 2013-06-25 Jakub Jelinek <jakub@redhat.com>
14522 PR tree-optimization/57705
14523 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
14524 SSA_NAME step, provided that it is not defined inside the loop.
14525 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
14526 (get_initial_def_for_induction): Handle SSA_NAME IV step.
14528 2013-06-25 Martin Jambor <mjambor@suse.cz>
14530 PR middle-end/57670
14531 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
14532 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
14534 (ipa_note_param_call): Initialize member_ptr flag.
14535 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
14536 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
14537 (ipa_write_indirect_edge_info): Stream member_ptr flag.
14538 (ipa_read_indirect_edge_info): Likewise.
14540 2013-06-25 Richard Biener <rguenther@suse.de>
14542 PR middle-end/56977
14543 * passes.c (init_optimization_passes): Move pass_fold_builtins
14544 and pass_dce earlier with -Og.
14546 2013-06-25 Eric Botcazou <ebotcazou@adacore.com>
14548 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
14549 <BIT_FIELD_REF>: Remove trailing TAB.
14550 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
14553 2013-06-24 Martin Jambor <mjambor@suse.cz>
14555 PR tree-optimization/57358
14556 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
14557 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
14558 (ipa_analyze_params_uses): Generate pessimistic info when true.
14560 2013-06-24 Martin Jambor <mjambor@suse.cz>
14562 PR tree-optimization/57539
14563 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
14564 global.inlined_to of the new node to it. All callers changed.
14565 * ipa-inline-transform.c (clone_inlined_nodes): New variable
14566 inlining_into, pass it to cgraph_clone_node.
14567 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
14568 ipa_free_edge_args_substructures.
14569 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
14570 rdesc linked list. Do not assert rdesc edges have inlined caller.
14571 Assert we have found an rdesc in the rdesc list.
14573 2013-06-24 Richard Biener <rguenther@suse.de>
14575 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
14576 (pointer_set_lookup): Declare.
14577 (class pointer_map): New template class implementing a
14578 generic pointer to T map.
14579 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
14580 pointer_map<T>::contains, pointer_map<T>::insert,
14581 pointer_map<T>::traverse): New functions.
14582 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
14583 (pointer_set_lookup): New function.
14584 (pointer_set_contains): Use pointer_set_lookup.
14585 (pointer_set_insert): Likewise.
14586 (insert_aux): Remove.
14587 (struct pointer_map_t): Embed a pointer_set_t.
14588 (pointer_map_create): Adjust.
14589 (pointer_map_destroy): Likewise.
14590 (pointer_map_contains): Likewise.
14591 (pointer_map_insert): Likewise.
14592 (pointer_map_traverse): Likewise.
14593 * tree-streamer.h (struct streamer_tree_cache_d): Use a
14594 pointer_map<unsigned> instead of a pointer_map_t.
14595 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
14596 (streamer_tree_cache_lookup): Likewise.
14597 (streamer_tree_cache_create): Likewise.
14598 (streamer_tree_cache_delete): Likewise.
14599 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
14600 pointer_map<unsigned> instead of a pointer_map_t.
14601 (lto_init_tree_ref_encoder): Adjust.
14602 (lto_destroy_tree_ref_encoder): Likewise.
14603 * lto-section-out.c (lto_output_decl_index): Likewise.
14604 (lto_record_function_out_decl_state): Likewise.
14605 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
14607 2013-06-24 Richard Biener <rguenther@suse.de>
14609 PR tree-optimization/57488
14610 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
14612 2013-06-24 Alan Modra <amodra@gmail.com>
14614 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
14615 (gen_easy_altivec_constant): Likewise.
14616 * config/rs6000/predicates.md (easy_vector_constant_add_self,
14617 easy_vector_constant_msb): Likewise.
14619 2013-06-23 Jakub Jelinek <jakub@redhat.com>
14622 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
14623 add missing return true.
14625 2013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
14628 * config/sh/predicates.md (general_extend_operand): Invoke
14629 general_movsrc_operand for memory operands.
14630 (general_movsrc_operand): Allow reg+reg addressing, do not use
14631 general_operand for memory operands.
14633 2013-06-23 Sriraman Tallam <tmsriram@google.com>
14635 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
14636 when current target options does not apply.
14637 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
14638 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
14639 * config/i386/bmiintrin.h: Pass appropriate target
14640 attributes to header.
14641 * config/i386/mmintrin.h: Ditto.
14642 * config/i386/nmmintrin.h: Ditto.
14643 * config/i386/avx2intrin.h: Ditto.
14644 * config/i386/fxsrintrin.h: Ditto.
14645 * config/i386/tbmintrin.h: Ditto.
14646 * config/i386/xsaveintrin.h: Ditto.
14647 * config/i386/f16cintrin.h: Ditto.
14648 * config/i386/xtestintrin.h: Ditto.
14649 * config/i386/xsaveoptintrin.h: Ditto.
14650 * config/i386/bmi2intrin.h: Ditto.
14651 * config/i386/lzcntintrin.h: Ditto.
14652 * config/i386/smmintrin.h: Ditto.
14653 * config/i386/wmmintrin.h: Ditto.
14654 * config/i386/x86intrin.h: Remove all header include guards.
14655 * config/i386/prfchwintrin.h: Ditto.
14656 * config/i386/pmmintrin.h: Ditto.
14657 * config/i386/tmmintrin.h: Ditto.
14658 * config/i386/xmmintrin.h: Ditto.
14659 * config/i386/popcntintrin.h: Ditto.
14660 * config/i386/rdseedintrin.h: Ditto.
14661 * config/i386/ammintrin.h: Ditto.
14662 * config/i386/emmintrin.h: Ditto.
14663 * config/i386/immintrin.h: Remove all header include guards.
14664 * config/i386/fma4intrin.h: Ditto.
14665 * config/i386/lwpintrin.h: Ditto.
14666 * config/i386/xopintrin.h: Ditto.
14667 * config/i386/ia32intrin.h: Ditto.
14668 * config/i386/avxintrin.h: Ditto.
14669 * config/i386/rtmintrin.h: Ditto.
14670 * config/i386/fmaintrin.h: Ditto.
14671 * config/i386/mm3dnow.h: Ditto.
14673 2013-06-22 Sriraman Tallam <tmsriram@google.com>
14675 * common/config/i386/i386-common.c: Handle LZCNT.
14677 2013-06-22 Andi Kleen <ak@linux.intel.com>
14679 * doc/extend.texi: Use __atomic_store_n instead of
14680 __atomic_store in HLE example.
14682 2013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
14684 * config/sh/sh.c: Remove <cstdlib> workaround.
14686 2013-06-21 Andi Kleen <ak@linux.intel.com>
14688 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
14690 2013-06-21 Andi Kleen <ak@linux.intel.com>
14692 * doc/extend.texi: Document that __atomic_clear and
14693 __atomic_test_and_set should only be used with bool.
14695 2013-06-20 Jan Hubicka <jh@suse.cz>
14697 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
14698 types_same_for_odr.
14699 * tree.c (decls_same_for_odr): New function.
14700 (same_for_edr): New function.
14701 (types_same_for_odr): New function.
14702 (get_binfo_at_offset): Use it.
14703 * tree.h (types_same_for_odr): Declare.
14705 2013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
14706 Jason Merrill <jason@redhat.com>
14708 * system.h: Include <cstdlib> as well as <stdlib.h>.
14710 2013-06-20 Uros Bizjak <ubizjak@gmail.com>
14713 * config/i386/i386.c (construct_container): Report error if
14714 long double is used with disabled x87 float returns.
14716 2013-06-20 Jan Hubicka <jh@suse.cz>
14718 * lto-cgraph.c (input_symtab): Do not set cgraph state.
14720 2013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
14722 PR rtl-optimization/57425
14723 PR rtl-optimization/57569
14724 * alias.c (write_dependence_p): Remove parameters mem_mode and
14725 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
14726 Changed all callers.
14727 (canon_anti_dependence): Get comments and semantics in sync.
14728 Add parameter mem_canonicalized. Changed all callers.
14729 * rtl.h (canon_anti_dependence): Update prototype.
14731 2013-06-20 Richard Biener <rguenther@suse.de>
14733 * data-streamer-in.c (streamer_read_uhwi): Optimize single
14734 byte case, inline streamer_read_uchar and defer section
14737 2013-06-20 Richard Biener <rguenther@suse.de>
14739 PR tree-optimization/57584
14740 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
14741 SSA names into the expanded expression that take part in
14742 abnormal coalescing.
14744 2013-06-19 Sharad Singhai <singhai@google.com>
14746 * gcov.c (print_usage): Handle new option.
14747 (process_args): Ditto.
14748 (get_gcov_intermediate_filename): New function.
14749 (output_intermediate_file): New function.
14750 (output_gcov_file): New function
14751 (generate_results): Handle new option.
14752 (release_function): Relase demangled name.
14753 (read_graph_file): Handle demangled name.
14754 (output_lines): Ditto.
14755 * doc/gcov.texi: Document gcov intermediate format.
14757 2013-06-19 Vladimir Makarov <vmakarov@redhat.com>
14760 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
14761 (lra_emit_add): Use the functions. Add comment about Y as an
14764 2013-06-19 David Edelsohn <dje.gcc@gmail.com>
14767 * collect2.c (collect_atexit): New.
14768 (collect_exit): Delete.
14769 (main): Register collect_atexit with atexit.
14770 (collect_wait): Change collect_exit to exit.
14772 * collect2.h (collect_exit): Delete.
14773 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
14775 2013-06-19 Wei Mi <wmi@google.com>
14777 PR rtl-optimization/57518
14778 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
14779 if regno is used in paradoxical subreg.
14780 (update_equiv_regs): Check pdx_subregs[regno] before
14781 set a reg to be equivalent with a mem.
14783 2013-06-19 Matthias Klose <doko@ubuntu.com>
14786 * file-find.h (find_a_file): Add a mode parameter.
14787 * file-find.c (find_a_file): Likewise.
14788 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
14789 with X_OK for the executables.
14790 * collect2.c (main): Call find_a_file with X_OK.
14792 2013-06-19 Steve Ellcey <sellcey@mips.com>
14795 * config/mips/mips.md (casesi_internal_mips16_<mode>):
14796 Use NEXT_INSN instead of next_real_insn.
14798 2013-06-19 Jan Hubicka <jh@suse.cz>
14800 * cgraph.h (const_value_known_p): Replace by ...
14801 (ctor_for_folding): .. this one.
14802 * cgraphunit.c (process_function_and_variable_attributes): Use it.
14803 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
14804 * expr.c (expand_expr_real_1): Likewise.
14805 (string_constant): Likewise.
14806 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
14807 * ipa.c (process_references): Likewise.
14808 (symtab_remove_unreachable_nodes): Likewise.
14809 * ipa-inline-analysis.c (param_change_prob): Likewise.
14810 * gimple-fold.c (canonicalize_constructor_val): Likewise.
14811 (get_base_constructor): Likwise.
14812 * varpool.c (varpool_remove_node): Likewise.
14813 (varpool_remove_initializer): LIkewise.
14814 (dump_varpool_node): LIkwise.
14815 (const_value_known_p): Rewrite to ...
14816 (ctor_for_folding): ... this one.
14818 2013-06-19 Jakub Jelinek <jakub@redhat.com>
14821 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
14822 PERSONALITY in $PATH derived prefixes.
14824 2013-06-19 Jeff Law <law@redhat.com>
14826 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
14829 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
14830 (simplify_bitwise_binary): Use it to simpify certain binary ops on
14833 2013-06-19 Sofiane Naci <sofiane.naci@arm.com>
14835 * config/arm/vfp.md: Move VFP instruction classification documentation
14837 * config/arm/arm.md: ... here. Update instruction classification
14840 2013-06-19 Richard Earnshaw <rearnsha@arm.com>
14842 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
14843 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
14844 pattern. Use more efficient sequences on ARMv5 and Thumb2.
14846 2013-06-19 Steven Bosscher <steven@gcc.gnu.org>
14849 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
14850 with NEXT_INSN. Use tablejump_p to check for jump table data
14853 2013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
14856 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
14857 that now in C++ the value is correct per the C++ standards.
14859 2013-06-19 Richard Biener <rguenther@suse.de>
14861 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
14862 for global context.
14864 2013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14867 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14870 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
14871 with next_active_insn.
14873 2013-06-18 Sriraman Tallam <tmsriram@google.com>
14875 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
14876 functions are inlined during failures to flag an error.
14877 * tree-inline.c (expand_call_inline): Allow the error to be flagged
14878 in early inline pass.
14880 2013-06-18 H.J. Lu <hongjiu.lu@intel.com>
14882 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
14885 2013-06-18 Julian Brown <julian@codesourcery.com>
14887 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
14888 Permit virtual register pre-reload if !strict.
14889 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
14891 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
14893 * config/arm/neon.md (movmisalign<mode>): Use
14894 neon_perm_struct_or_reg_operand instead of
14895 neon_struct_or_register_operand.
14896 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
14897 neon_permissive_struct_operand instead of neon_struct_operand.
14898 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
14899 neon_vector_mem_operand.
14900 * config/arm/predicates.md (neon_struct_operand): Adjust call to
14901 neon_vector_mem_operand.
14902 (neon_permissive_struct_operand): New.
14903 (neon_struct_or_register_operand): Rename to...
14904 (neon_perm_struct_or_reg_operand): This. Adjust call to
14905 neon_vector_mem_operand.
14907 2013-06-18 Richard Biener <rguenther@suse.de>
14909 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
14910 * lto-streamer.h: Include pointer-set.h.
14911 (struct lto_decl_slot): Remove.
14912 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
14913 Remove next_index entry.
14914 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
14915 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
14916 (lto_init_tree_ref_encoder): Adjust.
14917 (lto_destroy_tree_ref_encoder): Likewise.
14918 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
14919 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
14920 (lto_output_decl_index): Adjust.
14921 (lto_new_out_decl_state): Likewise.
14922 (lto_record_function_out_decl_state): Likewise.
14923 * lto-streamer-out.c (copy_function): Likewise.
14925 2013-06-18 Richard Biener <rguenther@suse.de>
14927 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
14928 * cgraphunit.c: Include cfgloop.h.
14929 (init_lowered_empty_function): Initialize the loop tree.
14930 (assemble_thunk): Insert new BBs into loops.
14932 2013-06-18 Richard Biener <rguenther@suse.de>
14934 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
14935 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
14936 the map from cache entry to cache index optional.
14937 (streamer_tree_cache_replace_tree): Adjust accordingly.
14938 (streamer_tree_cache_append): Likewise.
14939 (streamer_tree_cache_delete): Likewise.
14940 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
14941 streamer cache map from cache entry to cache index.
14942 * lto-streamer-out.c (create_output_block): Adjust.
14944 2013-06-18 Sofiane Naci <sofiane.naci@arm.com>
14946 * config/arm/arm.md (attribute "insn"): Move multiplication and
14947 division attributes to...
14948 (attribute "type"): ... here. Remove mult.
14949 (attribute "mul32"): New attribute.
14950 (attribute "mul64"): Add umaal.
14951 (*arm_mulsi3): Update attributes.
14952 (*arm_mulsi3_v6): Likewise.
14953 (*thumb_mulsi3): Likewise.
14954 (*thumb_mulsi3_v6): Likewise.
14955 (*mulsi3_compare0): Likewise.
14956 (*mulsi3_compare0_v6): Likewise.
14957 (*mulsi_compare0_scratch): Likewise.
14958 (*mulsi_compare0_scratch_v6): Likewise.
14959 (*mulsi3addsi): Likewise.
14960 (*mulsi3addsi_v6): Likewise.
14961 (*mulsi3addsi_compare0): Likewise.
14962 (*mulsi3addsi_compare0_v6): Likewise.
14963 (*mulsi3addsi_compare0_scratch): Likewise.
14964 (*mulsi3addsi_compare0_scratch_v6): Likewise.
14965 (*mulsi3subsi): Likewise.
14966 (*mulsidi3adddi): Likewise.
14967 (*mulsi3addsi_v6): Likewise.
14968 (*mulsidi3adddi_v6): Likewise.
14969 (*mulsidi3_nov6): Likewise.
14970 (*mulsidi3_v6): Likewise.
14971 (*umulsidi3_nov6): Likewise.
14972 (*umulsidi3_v6): Likewise.
14973 (*umulsidi3adddi): Likewise.
14974 (*umulsidi3adddi_v6): Likewise.
14975 (*smulsi3_highpart_nov6): Likewise.
14976 (*smulsi3_highpart_v6): Likewise.
14977 (*umulsi3_highpart_nov6): Likewise.
14978 (*umulsi3_highpart_v6): Likewise.
14979 (mulhisi3): Likewise.
14980 (*mulhisi3tb): Likewise.
14981 (*mulhisi3bt): Likewise.
14982 (*mulhisi3tt): Likewise.
14983 (maddhisi4): Likewise.
14984 (*maddhisi4tb): Likewise.
14985 (*maddhisi4tt): Likewise.
14986 (maddhidi4): Likewise.
14987 (*maddhidi4tb): Likewise.
14988 (*maddhidi4tt): Likewise.
14989 (divsi3): Likewise.
14990 (udivsi3): Likewise.
14991 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
14992 (thumb2_mulsi_short_compare0): Likewise.
14993 (thumb2_mulsi_short_compare0_scratch): Likewise.
14994 * config/arm/arm1020e.md (1020mult1): Update attribute change.
14995 (1020mult2): Likewise.
14996 (1020mult3): Likewise.
14997 (1020mult4): Likewise.
14998 (1020mult5): Likewise.
14999 (1020mult6): Likewise.
15000 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
15002 (cortex_a15_mult64): Likewise.
15003 (cortex_a15_sdiv): Likewise.
15004 (cortex_a15_udiv): Likewise.
15005 * config/arm/arm1026ejs.md (mult1): Update attribute change.
15011 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
15012 (pj4_ir_div): Likewise.
15013 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
15014 (11_mult2): Likewise.
15015 (11_mult3): Likewise.
15016 (11_mult4): Likewise.
15017 (11_mult5): Likewise.
15018 (11_mult6): Likewise.
15019 (11_mult7): Likewise.
15020 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
15021 (cortex_a8_mla): Likewise.
15022 (cortex_a8_mull): Likewise.
15023 (cortex_a8_smulwy): Likewise.
15024 (cortex_a8_smlald): Likewise.
15025 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
15026 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
15027 (cortex_r4_mul_3): Likewise.
15028 (cortex_r4_mla_4): Likewise.
15029 (cortex_r4_mla_3): Likewise.
15030 (cortex_r4_smlald): Likewise.
15031 (cortex_r4_mull): Likewise.
15032 (cortex_r4_sdiv): Likewise.
15033 (cortex_r4_udiv): Likewise.
15034 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
15035 (cortex_a7_idiv): Likewise.
15036 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
15037 (9_mult2): Likewise.
15038 (9_mult3): Likewise.
15039 (9_mult4): Likewise.
15040 (9_mult5): Likewise.
15041 (9_mult6): Likewise.
15042 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
15043 (cortex_a53_sdiv): Likewise.
15044 (cortex_a53_udiv): Likewise.
15045 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
15046 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
15047 (mp626_mult2): Likewise.
15048 (mp626_mult3): Likewise.
15049 (mp626_mult4): Likewise.
15050 * config/arm/fa526.md (526_mult1): Update attribute change.
15051 (526_mult2): Likewise.
15052 * config/arm/arm-generic.md (mult): Update attribute change.
15053 (mult_ldsched_strongarm): Likewise.
15054 (mult_ldsched): Likewise.
15055 (multi_cycle): Likewise.
15056 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
15057 * config/arm/fa606te.md (606te_mult1): Update attribute change.
15058 (606te_mult2): Likewise.
15059 (606te_mult3): Likewise.
15060 (606te_mult4): Likewise.
15061 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
15062 (cortex_a9_mac16): Likewise.
15063 (cortex_a9_multiply): Likewise.
15064 (cortex_a9_mac): Likewise.
15065 (cortex_a9_multiply_long): Likewise.
15066 * config/arm/fa626te.md (626te_mult1): Update attribute change.
15067 (626te_mult2): Likewise.
15068 (626te_mult3): Likewise.
15069 (626te_mult4): Likewise.
15071 2013-06-18 Richard Biener <rguenther@suse.de>
15074 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
15076 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
15079 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
15080 with next_active_insn.
15082 2013-06-18 Alan Modra <amodra@gmail.com>
15084 * config/rs6000/rs6000.h (enum data_align): New.
15085 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
15086 (DATA_ABI_ALIGNMENT): Define.
15087 (CONSTANT_ALIGNMENT): Correct comment.
15088 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
15089 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
15091 2013-06-17 David Malcolm <dmalcolm@redhat.com>
15093 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
15094 ATTRIBUTE_UNUSED marking.
15096 2013-06-17 Sofiane Naci <sofiane.naci@arm.com>
15098 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
15099 alternative and update.
15100 (aarch64_dup_lanedi): Delete.
15101 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
15102 * config/aarch64/aarch64-simd-builtins.def: Update.
15104 2013-06-17 Richard Biener <rguenther@suse.de>
15106 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
15107 (lto_input_scc): Declare.
15108 (lto_input_tree_1): Likewise.
15109 (struct lto_stats_d): Add num_tree_bodies_output and
15110 num_pickle_refs_output.
15111 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
15112 (lto_read_tree_1): Split out from ...
15113 (lto_read_tree): ... this.
15114 (lto_input_scc): New function.
15115 (lto_input_tree_1): Split out from ...
15116 (lto_input_tree): ... this. Handle LTO_tree_scc.
15117 (lto_data_in_create): Create the streamer cache without hashes.
15118 * lto-streamer-out.c (create_output_block): Create the streamer
15119 cache with hashes when not doing WPA.
15120 (lto_write_tree_1): Split out from ...
15121 (lto_write_tree): ... this.
15122 (get_symbol_initial_value): New function.
15123 (lto_output_tree_1): Split out from ...
15124 (lto_output_tree): ... this. Write trees as series of SCCs
15125 using a DFS walk via DFS_write_tree.
15126 (struct sccs, struct scc_entry): New types.
15127 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
15128 (DFS_write_tree_body): New function.
15129 (DFS_write_tree): Likewise.
15130 (hash_tree): Likewise.
15131 (scc_entry_compare): Likewise.
15132 (hash_scc): Likewise.
15133 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
15134 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
15135 TREE_CHAIN as regular reference.
15136 (streamer_read_integer_cst): Remove.
15137 (streamer_get_pickled_tree): Adjust.
15138 * tree-streamer-out.c (streamer_write_chain): Disable streaming
15139 of DECL_EXTERNALs in BLOCK_VARS for now.
15140 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
15142 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
15143 Add hash value argument and record that if hashes are recorded
15145 (streamer_tree_cache_insert_1): Adjust.
15146 (streamer_tree_cache_insert): Likewise.
15147 (streamer_tree_cache_insert_at): Rename to ...
15148 (streamer_tree_cache_replace_tree): ... this and adjust.
15149 (streamer_tree_cache_append): Adjust.
15150 (record_common_node): Likewise.
15151 (streamer_tree_cache_create): Add argument whether to
15152 record hash values together with trees.
15153 (streamer_tree_cache_delete): Adjust.
15154 * tree-streamer.h (struct streamer_tree_cache_d): Add
15156 (streamer_read_integer_cst): Remove.
15157 (streamer_tree_cache_insert): Adjust.
15158 (streamer_tree_cache_append): Likewise.
15159 (streamer_tree_cache_insert_at): Rename to ...
15160 (streamer_tree_cache_replace_tree): ... this and adjust.
15161 (streamer_tree_cache_create): Add argument whether to record hashes.
15162 (streamer_tree_cache_get): Rename to ...
15163 (streamer_tree_cache_get_tree): ... this.
15164 (streamer_tree_cache_get_hash): New function.
15165 * tree.c (cache_integer_cst): New function.
15166 * tree.h (cache_integer_cst): Declare.
15167 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
15168 * lto-symtab.c (lto_varpool_replace_node): Only release
15169 DECL_INITIAL of non-prevailing decls.
15170 * varpool.c (varpool_remove_initializer): Do not release
15171 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
15173 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
15175 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
15176 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
15177 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
15178 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
15179 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
15180 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
15181 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
15182 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
15183 instead of TARGET_64BIT.
15184 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
15185 Require ISA_HAS_<D>DIV.
15187 2013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
15189 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
15190 (mips*-*-linux*): Move default with_llsc setting to where other
15192 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
15194 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
15195 Likewise. Remove default with_tune setting. Move default float
15196 setting to its own block. Handle with_llsc in the same block as above.
15198 2013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
15200 PR rtl-optimization/57425
15201 PR rtl-optimization/57569
15202 * alias.c (write_dependence_p): Add new parameters mem_mode,
15203 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
15204 Changed all callers.
15205 (canon_anti_dependence): New function.
15206 * cse.c (check_dependence): Use canon_anti_dependence.
15207 * cselib.c (cselib_invalidate_mem): Likewise.
15208 * rtl.h (canon_anti_dependence): Declare.
15210 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
15212 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
15213 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
15214 ".set mips3" for 64-bit targets.
15216 2013-06-15 Dehao Chen <dehao@google.com>
15218 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
15219 * gimple-low.c (gimple_check_call_matching_types): Likewise.
15220 (gimple_check_call_args): Likewise.
15221 * value-prof.c (check_ic_target): Likewise.
15222 * ipa-inline.c (early_inliner): Likewise.
15223 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
15224 * cgraph.c (cgraph_create_edge_1): Likewise.
15225 (cgraph_make_edge_direct): Likewise.
15227 2013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
15230 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
15231 rs6000_output_move_128bit to handle emitting quad memory
15232 operations. Set attribute length to 8 bytes.
15234 2013-06-14 Vidya Praveen <vidyapraveen@arm.com>
15236 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
15238 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
15239 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
15240 (aarch64_<su>mlsl<mode>): Likewise.
15242 2013-06-14 Mike Stump <mikestump@comcast.net>
15244 * Makefile.in (TARGET_H): Add insn-codes.h.
15246 2013-06-14 Alan Modra <amodra@gmail.com>
15248 PR middle-end/57134
15249 PR middle-end/57586
15250 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
15251 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
15252 bitfield expansion when EXPAND_MEMORY.
15253 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
15255 2013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
15257 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
15258 test for clearing quad memory on 32-bit later.
15260 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
15262 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
15263 (fold_negate_expr): Likewise.
15264 (fold_real_zero_addition_p): Handle vectors.
15265 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
15267 2013-06-14 Alan Modra <amodra@gmail.com>
15269 * varasm.c (force_const_mem): Revert 2013-06-07 change.
15271 2013-06-13 Jan Hubicka <jh@suse.cz>
15273 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
15274 Local comdats are not externally visible.
15275 * symtab.c (dump_symtab_base): Dump externally visible.
15276 (verify_symtab_base): Verify back links in the symtab hash.
15278 2013-06-13 Bin Cheng <bin.cheng@arm.com>
15280 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
15281 CONVERT_EXPR as equal nodes.
15283 2013-06-13 Bin Cheng <bin.cheng@arm.com>
15285 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
15287 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
15289 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
15290 Generalize to complex and vector.
15291 * tree.c (build_all_ones_cst): New function.
15292 * tree.h (build_all_ones_cst): Declare it.
15294 2013-06-13 Alan Modra <amodra@gmail.com>
15296 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
15297 * config/rs6000/rs6000.md (signbittf2): New insn.
15298 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
15299 (abstf2_internal, cmptf_internal2): Likewise.
15300 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
15302 2013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
15303 Pat Haugen <pthaugen@us.ibm.com>
15304 Peter Bergner <bergner@vnet.ibm.com>
15306 * config/rs6000/rs6000.c (emit_load_locked): Add support for
15307 power8 byte, half-word, and quad-word atomic instructions.
15308 (emit_store_conditional): Likewise.
15309 (rs6000_expand_atomic_compare_and_swap): Likewise.
15310 (rs6000_expand_atomic_op): Likewise.
15312 * config/rs6000/sync.md (larx): Add new modes for power8.
15314 (AINT): New mode iterator to include TImode as well as normal
15315 integer modes on power8.
15316 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
15317 that VSX registers are not considered. Use AINT mode iterator
15318 instead of INT1 to allow inclusion of quad word atomic operations
15320 (load_locked<mode>): Likewise.
15321 (store_conditional<mode>): Likewise.
15322 (atomic_compare_and_swap<mode>): Likewise.
15323 (atomic_exchange<mode>): Likewise.
15324 (atomic_nand<mode>): Likewise.
15325 (atomic_fetch_<fetchop_name><mode>): Likewise.
15326 (atomic_nand_fetch<mode>): Likewise.
15327 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
15329 (ATOMIC): On power8, add QImode, HImode modes.
15330 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
15331 modes that promote to SImode.
15332 (load_lockedti): Convert TImode arguments to PTImode, so that we
15333 get a guaranteed even/odd register pair.
15334 (load_lockedpti): Likewise.
15335 (store_conditionalti): Likewise.
15336 (store_conditionalpti): Likewise.
15338 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
15339 atomic load/store instructions.
15342 2013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
15344 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
15346 (insn_count): New attribute, with most cases extracted from...
15347 (length): ...here. Redefine most cases in terms of insn_count.
15348 (single_insn): Delete.
15349 (can_delay): Use insn_count to check for single instructions.
15350 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
15351 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
15352 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
15353 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
15354 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
15355 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
15356 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
15357 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
15358 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
15359 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
15360 rather than "length".
15361 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
15362 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
15363 Use "insn_count" rather than "length".
15364 * config/mips/mips-dsp.md
15365 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
15366 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
15369 2013-06-12 Marc Glisse <marc.glisse@inria.fr>
15371 PR tree-optimization/57361
15372 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
15374 2013-06-12 Sofiane Naci <sofiane.naci@arm.com>
15376 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
15378 (aarch64_simd_combine<mode>): New instruction expansion.
15379 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
15380 function prototype.
15381 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
15382 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
15384 2013-06-12 Jan Hubicka <jh@suse.cz>
15386 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
15387 decl has when in streaming stage.
15388 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
15389 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
15391 2013-06-12 Roland Stigge <stigge@antcom.de>
15394 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
15396 2013-06-12 Jakub Jelinek <jakub@redhat.com>
15398 PR tree-optimization/57537
15399 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
15400 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
15402 2013-06-12 Richard Biener <rguenther@suse.de>
15404 * data-streamer.h (streamer_write_char_stream): CSE
15405 obs->current_pointer.
15406 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
15407 streamer_write_char_stream manually and optimize the resulting loop.
15408 (streamer_write_hwi_stream): Likewise.
15410 2013-06-12 Jan Hubicka <jh@suse.cz>
15412 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
15413 * cgraph.h (varpool_create_empty_node): Declare.
15414 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
15416 * symtab.c (symtab_unregister_node): Be lax about missin entries
15418 (symtab_get_node): Update comment.
15419 * varpool.c (varpool_create_empty_node): Break out from ...
15420 (varpool_node_for_decl): ... here.
15421 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
15423 2013-06-12 Eric Botcazou <ebotcazou@adacore.com>
15425 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
15426 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
15427 part. Use straight-line flow at the end.
15428 <COMPONENT_REF>: Remove superfluous else.
15429 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
15431 2013-06-12 Jakub Jelinek <jakub@redhat.com>
15434 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
15435 target hook even for !TREE_PUBLIC decls. If no resolution info
15436 is available, return false for common and external decls.
15438 2013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
15440 * config/rl78/constraints.md (U): New constraint.
15441 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
15444 2013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
15447 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
15448 to allow returning address to AT_PLATFORM name.
15450 2013-06-11 Jan Hubicka <jh@suse.cz>
15452 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
15453 * cgraph.h (symtab_node_base): Add weakref flag.
15454 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
15455 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
15456 (output_weakrefs): Use weakref flag.
15457 * fold-const.c (simple_operand_p): Handle WEAK.
15458 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
15459 * ipa.c (varpool_externally_visible_p): Drop weakref.
15460 (function_and_variable_visibility): Update comment; fix weakref
15461 sanity checks; do not clear DECL_WEAK on them.
15462 * lto-cgraph.c (lto_output_node): update.
15463 (lto_output_varpool_node): Update.
15464 (input_overwrite_node): Update.
15465 (input_node): Update.
15466 (input_varpool_node): Update.
15467 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
15468 (lto_symtab_merge_symbols): Add sanity check.
15469 (lto_symtab_prevailing_decl): Do not special case weakrefs.
15470 * passes.c (rest_of_decl_compilation): Set static flag, too.
15471 * symtab.c (dump_symtab_base): Dump weakref.
15472 (verify_symtab_base): Sanity check weakrefs.
15473 (symtab_make_decl_local): Remove duplicated code.
15474 (symtab_alias_ultimate_target): Simplify.
15475 * varpool.c (varpool_create_variable_alias): Set weakref flag.
15477 2013-06-11 Tom de Vries <tom@codesourcery.com>
15479 * genautomata.c (gen_regexp_sequence): Handle els_num == -1. Handle
15480 sequence_vect == NULL.
15482 2013-06-11 DJ Delorie <dj@redhat.com>
15484 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
15485 (rl78_unwind_word_mode): New.
15487 2013-06-11 David Malcolm <dmalcolm@redhat.com>
15489 * final.c (debug_prefix_maps): Make static.
15491 2013-06-11 David Malcolm <dmalcolm@redhat.com>
15493 * function.c (initial_trampoline): Remove stray copy.
15495 2013-06-11 Sofiane Naci <sofiane.naci@arm.com>
15497 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
15499 2013-06-11 Martin Jambor <mjambor@suse.cz>
15501 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
15502 within bounds at the beginning of the function.
15504 2013-06-11 Alan Modra <amodra@gmail.com>
15506 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
15508 (get_named_section): Don't NULL !DECL_P decl.
15510 2013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
15512 * doc/invoke.texi (core-avx2): Document.
15514 (atom): Updated with MOVBE.
15516 2013-06-11 Richard Biener <rguenther@suse.de>
15518 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
15520 2013-06-11 Anton Blanchard <anton@samba.org>
15522 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
15523 correct shift value in little-endian mode.
15525 2013-06-11 Jakub Jelinek <jakub@redhat.com>
15528 * varasm.c (get_variable_align): Move #endif to the right place.
15530 2013-06-10 Cary Coutant <ccoutant@google.com>
15532 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
15533 for hash so that hash table traversal order is deterministic.
15535 2013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
15536 Pat Haugen <pthaugen@us.ibm.com>
15537 Peter Bergner <bergner@vnet.ibm.com>
15539 * config/rs6000/vector.md (GPR move splitter): Do not split moves
15540 of vectors in GPRS if they are direct moves or quad word load or
15543 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
15545 (direct_move_p): Likewise.
15546 (quad_load_store_p): Likewise.
15548 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
15549 classes into bins based on the physical register type.
15550 (reg_class_to_reg_type): Likewise.
15551 (IS_STD_REG_TYPE): Likewise.
15552 (IS_FP_VECT_REG_TYPE): Likewise.
15553 (reload_fpr_gpr): Arrays to determine what insn to use if we can
15554 use direct move instructions.
15555 (reload_gpr_vsx): Likewise.
15556 (reload_vsx_gpr): Likewise.
15557 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
15558 information that is a simplification of register classes. Also
15559 precalculate direct move reload helpers.
15560 (direct_move_p): New function to return true if the operation can
15561 be done as a direct move instruciton.
15562 (quad_load_store_p): New function to return true if the operation
15563 is a quad memory operation.
15564 (rs6000_legitimize_address): If quad memory, only allow register
15565 indirect for TImode addresses.
15566 (rs6000_legitimate_address_p): Likewise.
15567 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
15568 (rs6000_reload_register_type): Likewise.
15569 (register_to_reg_type): Return register type.
15570 (rs6000_secondary_reload_simple_move): New helper function for
15571 secondary reload and secondary memory needed to identify anything
15572 that is a simple move, and does not need reloading.
15573 (rs6000_secondary_reload_direct_move): New helper function for
15574 secondary reload to identify cases that can be done with several
15575 instructions via the direct move instructions.
15576 (rs6000_secondary_reload_move): New helper function for secondary
15577 reload to identify moves between register types that can be done.
15578 (rs6000_secondary_reload): Add support for quad memory operations
15579 and for direct move.
15580 (rs6000_secondary_memory_needed): Likewise.
15581 (rs6000_debug_secondary_memory_needed): Change argument names.
15582 (rs6000_output_move_128bit): New function to return the move to
15583 use for 128-bit moves, including knowing about the various
15584 limitations of quad memory operations.
15586 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
15587 memory operations. call rs6000_output_move_128bit for the actual
15588 instruciton(s) to generate.
15589 (vsx_movti_64bit): Likewise.
15591 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
15592 (UNSPEC_P8V_MTVSRWZ): Likewise.
15593 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
15594 (UNSPEC_P8V_MTVSRD): Likewise.
15595 (UNSPEC_P8V_XXPERMDI): Likewise.
15596 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
15597 (UNSPEC_FUSION_GPR): Likewise.
15598 (FMOVE128_GPR): New iterator for direct move.
15599 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
15600 (f32_sv): Likewise.
15601 (f32_dm): Likewise.
15602 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
15603 loads and direct move instructions.
15604 (zero_extendsidi2_lfiwzx): Likewise.
15605 (extendsidi2_lfiwax): Likewise.
15606 (extendsidi2_nocell): Likewise.
15607 (floatsi<mode>2_lfiwax): Likewise.
15608 (lfiwax): Likewise.
15609 (floatunssi<mode>2_lfiwzx): Likewise.
15610 (lfiwzx): Likewise.
15611 (fix_trunc<mode>_stfiwx): Likewise.
15612 (fixuns_trunc<mode>_stfiwx): Likewise.
15613 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
15614 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
15615 (parity<mode>2_cmpb): Set length/type attr.
15616 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
15617 for 'mr.' to fast_compare.
15618 (bpermd_<mode>): Change type attr to popcnt.
15619 (p8_fmrgow_<mode>): New insns for power8 direct move support.
15620 (p8_mtvsrwz_1): Likewise.
15621 (p8_mtvsrwz_2): Likewise.
15622 (reload_fpr_from_gpr<mode>): Likewise.
15623 (p8_mtvsrd_1): Likewise.
15624 (p8_mtvsrd_2): Likewise.
15625 (p8_xxpermdi_<mode>): Likewise.
15626 (reload_vsx_from_gpr<mode>): Likewise.
15627 (reload_vsx_from_gprsf): Likewise.
15628 (p8_mfvsrd_3_<mode>): LIkewise.
15629 (reload_gpr_from_vsx<mode>): Likewise.
15630 (reload_gpr_from_vsxsf): Likewise.
15631 (p8_mfvsrd_4_disf): Likewise.
15632 (multi-word GPR splits): Do not split direct moves or quad memory
15635 2013-06-10 David Malcolm <dmalcolm@redhat.com>
15637 * tree-into-ssa.c (interesting_blocks): Make static.
15639 2013-06-10 Jakub Jelinek <jakub@redhat.com>
15642 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
15643 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
15644 Use DATA_ABI_ALIGNMENT for that case instead if defined.
15645 (get_variable_align): New function.
15646 (get_variable_section, emit_bss, emit_common,
15647 assemble_variable_contents, place_block_symbol): Use
15648 get_variable_align instead of DECL_ALIGN.
15649 (assemble_noswitch_variable): Add align argument, use it
15650 instead of DECL_ALIGN.
15651 (assemble_variable): Adjust caller. Use get_variable_align
15652 instead of DECL_ALIGN.
15653 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
15655 (DATA_ABI_ALIGNMENT): Define.
15656 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
15657 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
15658 opt is false, only return the psABI mandated alignment increase.
15659 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
15660 (DATA_ABI_ALIGNMENT): ... this.
15661 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
15662 (DATA_ABI_ALIGNMENT): ... this.
15663 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
15664 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
15665 (DATA_ABI_ALIGNMENT): ... this.
15666 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
15667 * doc/tm.texi: Regenerated.
15669 2013-06-10 Uros Bizjak <ubizjak@gmail.com>
15671 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
15672 cmp_code to construct REG_EQUAL note.
15674 2013-06-09 Jakub Jelinek <jakub@redhat.com>
15677 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
15678 that operands[2] doesn't overlap with operands[0].
15680 2013-06-09 David Edelsohn <dje.gcc@gmail.com>
15681 Jan Hubicka <jh@suse.cz>
15683 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
15684 hack to mark symbols as used.
15686 2013-06-08 Vladimir Makarov <vmakarov@redhat.com>
15688 PR rtl-optimization/57559
15689 * lra-constraints.c (process_alt_operands): Don't discourage
15690 memory with known offset for offsetable memory constraint.
15691 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
15693 2013-06-08 Eric Botcazou <ebotcazou@adacore.com>
15695 * varasm.c (struct oc_local_state): Reorder fields.
15696 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
15697 and adjust accordingly.
15698 (output_constructor): Reorder initialization code and adjust call to
15699 output_constructor_bitfield.
15701 2013-06-07 Jan Hubicka <jh@suse.cz>
15703 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
15705 2013-06-07 David Malcolm <dmalcolm@redhat.com>
15707 * tree-object-size.c (unknown): Make const.
15709 2013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
15711 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
15712 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
15713 for last alternative in the cpu_facility attribute.
15715 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15718 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
15719 (xordi3): Change operand 2 constraint to arm_xordi_operand.
15720 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
15721 * config/arm/constraints.md (Dg): New constraint.
15722 * config/arm/neon.md (xordi3_neon): Remove.
15723 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
15724 * config/arm/predicates.md (arm_xordi_operand): New predicate.
15726 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15728 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
15729 Clean up alternatives.
15731 2013-06-07 Alan Modra <amodra@gmail.com>
15733 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
15736 2013-06-07 Alan Modra <amodra@gmail.com>
15738 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
15740 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15742 * config/arm/constraints.md (Df): New constraint.
15743 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
15744 Correct length attribute for last two alternatives.
15746 2013-06-07 Alan Modra <amodra@gmail.com>
15748 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
15749 override user -mfp-in-toc.
15750 (offsettable_ok_by_alignment): Consider just the current access
15751 rather than the whole object, unless BLKmode. Handle
15752 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
15753 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
15754 for -mcmodel=medium.
15755 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
15756 override user -mfp-in-toc or -msum-in-toc. Default to
15757 -mno-fp-in-toc for -mcmodel=medium.
15759 2013-06-06 DJ Delorie <dj@redhat.com>
15761 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
15762 TARGET_VALID_POINTER_MODE.
15764 2013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
15765 Pat Haugen <pthaugen@us.ibm.com>
15766 Peter Bergner <bergner@vnet.ibm.com>
15768 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
15769 Document new power8 builtins.
15771 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
15772 condition code register, to allow 128-bit logical operations to be
15773 done in the VSX or GPR registers.
15774 (nor<mode>3): Use the canonical form for nor.
15775 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
15776 vclz*, and vpopcnt* vector instructions.
15777 (nand<mode>3): Likewise.
15778 (orc<mode>3): Likewise.
15779 (clz<mode>2): LIkewise.
15780 (popcount<mode>2): Likewise.
15782 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
15783 that only the GPRs are recognized.
15785 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
15786 support for new power8 builtins.
15788 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
15790 (xscvdpspn): Likewise.
15796 (vpopcnt): Likewise.
15797 (vpopcntb): Likewise.
15798 (vpopcnth): Likewise.
15799 (vpopcntw): Likewise.
15800 (vpopcntd): Likewise.
15802 (vmrgew): Likewise.
15803 (vmrgow): Likewise.
15805 (eqv_v16qi3): Likewise.
15806 (eqv_v8hi3): Likewise.
15807 (eqv_v4si3): Likewise.
15808 (eqv_v2di3): Likewise.
15809 (eqv_v4sf3): Likewise.
15810 (eqv_v2df3): Likewise.
15812 (nand_v16qi3): Likewise.
15813 (nand_v8hi3): Likewise.
15814 (nand_v4si3): Likewise.
15815 (nand_v2di3): Likewise.
15816 (nand_v4sf3): Likewise.
15817 (nand_v2df3): Likewise.
15819 (orc_v16qi3): Likewise.
15820 (orc_v8hi3): Likewise.
15821 (orc_v4si3): Likewise.
15822 (orc_v2di3): Likewise.
15823 (orc_v4sf3): Likewise.
15824 (orc_v2df3): Likewise.
15826 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
15827 allow power8 quad mode in 64-bit.
15828 (rs6000_builtin_vectorized_function): Add support to vectorize
15829 ISA 2.07 count leading zeros, population count builtins.
15830 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
15831 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
15832 (builtin_function_type): Add vgbbd builtin function which takes an
15834 (altivec_expand_vec_perm_const): Add support for new power8 merge
15837 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
15838 that does not include TImdoe for use with 32-bit.
15839 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
15841 (UNSPEC_VSX_CVDPSPN): Likewise.
15842 (vsx_xscvdpspn): Likewise.
15843 (vsx_xscvspdpn): Likewise.
15844 (vsx_xscvdpspn_scalar): Likewise.
15845 (vsx_xscvspdpn_directmove): Likewise.
15846 (vsx_and<mode>3): Split logical operations into 32-bit and
15847 64-bit. Add support to do logical operations on TImode as well as
15848 VSX vector types. Allow logical operations to be done in either
15849 VSX registers or in general purpose registers in 64-bit mode. Add
15850 splitters if GPRs were used. For AND, add clobber of CCmode to
15851 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
15853 (vsx_and<mode>3_32bit): Likewise.
15854 (vsx_and<mode>3_64bit): Likewise.
15855 (vsx_ior<mode>3): Likewise.
15856 (vsx_ior<mode>3_32bit): Likewise.
15857 (vsx_ior<mode>3_64bit): Likewise.
15858 (vsx_xor<mode>3): Likewise.
15859 (vsx_xor<mode>3_32bit): Likewise.
15860 (vsx_xor<mode>3_64bit): Likewise.
15861 (vsx_one_cmpl<mode>2): Likewise.
15862 (vsx_one_cmpl<mode>2_32bit): Likewise.
15863 (vsx_one_cmpl<mode>2_64bit): Likewise.
15864 (vsx_nor<mode>3): Likewise.
15865 (vsx_nor<mode>3_32bit): Likewise.
15866 (vsx_nor<mode>3_64bit): Likewise.
15867 (vsx_andc<mode>3): Likewise.
15868 (vsx_andc<mode>3_32bit): Likewise.
15869 (vsx_andc<mode>3_64bit): Likewise.
15870 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
15871 and xxlorc instructions.
15872 (vsx_eqv<mode>3_64bit): Likewise.
15873 (vsx_nand<mode>3_32bit): Likewise.
15874 (vsx_nand<mode>3_64bit): Likewise.
15875 (vsx_orc<mode>3_32bit): Likewise.
15876 (vsx_orc<mode>3_64bit): Likewise.
15878 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
15880 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
15882 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
15883 (p8_vmrgow): Likewise.
15884 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
15885 GPRs to be split under VSX.
15886 (p8v_clz<mode>2): Add power8 count leading zero support.
15887 (p8v_popcount<mode>2): Add power8 population count support.
15888 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
15891 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
15894 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
15896 (vec_nand): Likewise.
15897 (vec_vclz): Likewise.
15898 (vec_vclzb): Likewise.
15899 (vec_vclzd): Likewise.
15900 (vec_vclzh): Likewise.
15901 (vec_vclzw): Likewise.
15902 (vec_vgbbd): Likewise.
15903 (vec_vmrgew): Likewise.
15904 (vec_vmrgow): Likewise.
15905 (vec_vpopcnt): Likewise.
15906 (vec_vpopcntb): Likewise.
15907 (vec_vpopcntd): Likewise.
15908 (vec_vpopcnth): Likewise.
15909 (vec_vpopcntw): Likewise.
15911 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
15913 PR rtl-optimization/57468
15914 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
15917 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
15919 PR rtl-optimization/57459
15920 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
15921 type when setting live regs.
15923 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
15925 * config/s390/s390.opt (mlra): New option.
15926 * config/s390/s390.c (s390_decompose_address): Check displacement
15927 for all registers for LRA.
15928 (s390_secondary_reload): Don't used secondary reloads for LRA.
15929 (s390_lra_p): New function.
15930 (TARGET_LRA_P): Define.
15931 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
15932 of attribute cpu_facility to zarch for the last alternative.
15933 (*cmpmem_short): Ditto.
15935 2013-06-06 Eric Botcazou <ebotcazou@adacore.com>
15937 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
15938 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
15939 (arm_expand_prologue): Likewise.
15941 2013-06-06 Teresa Johnson <tejohnson@google.com>
15944 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
15945 as this is now done by redirect_edge_and_branch_force.
15946 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
15947 barriers, and fix interaction with splitting.
15948 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
15949 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
15950 reflect changes made in the routine.
15951 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
15952 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
15953 since this is called in cfglayout mode, and replace partition fixup
15954 with assert as that is now done by force_nonfallthru_and_redirect.
15955 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
15956 already be marked with region crossing note.
15957 (insert_section_boundary_note): Make non-static, gate on flag
15958 has_bb_partition, rewrite to also check for multiple partitions.
15959 (rest_of_handle_reorder_blocks): Remove call to
15960 insert_section_boundary_note, now done later during free_cfg.
15961 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
15962 * bb-reorder.h (insert_section_boundary_note): Declare.
15963 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
15964 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
15965 invoke insert_section_boundary_note.
15966 (try_redirect_by_replacing_jump): Remove unnecessary
15967 check for region crossing note.
15968 (fixup_partition_crossing): New function.
15969 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
15970 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
15971 in non-cfglayout mode.
15972 (force_nonfallthru_and_redirect): Fixup partition boundaries,
15973 remove old code that tried to do this. Emit barrier correctly
15974 when we are in cfglayout mode.
15975 (last_bb_in_partition): New function.
15976 (rtl_split_edge): Correctly fixup partition boundaries.
15977 (commit_one_edge_insertion): Remove old code that tried to
15978 fixup region crossing edge since this is now handled in
15979 split_block, and set up insertion point correctly since
15980 block may now end in a jump.
15981 (verify_hot_cold_block_grouping): Guard against checking when not in
15982 linearized RTL mode.
15983 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
15985 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
15986 rtl_verify_flow_info, so not called in cfglayout mode.
15987 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
15988 (fixup_reorder_chain): Remove old code that attempted to fixup region
15989 crossing note as this is now handled in force_nonfallthru_and_redirect.
15990 (duplicate_insn_chain): Don't duplicate switch section notes.
15991 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
15993 * basic-block.h (emit_barrier_after_bb): Declare.
15995 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15997 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
15998 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
15999 arm_usatsihi): Adjust alternatives for arm_restrict_it.
16001 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16003 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
16005 * config/arm/ldmstm.md: Regenerate.
16007 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16009 * config/arm/sync.md (atomic_loaddi_1):
16010 Disable predication for arm_restrict_it.
16011 (arm_load_exclusive<mode>): Likewise.
16012 (arm_load_exclusivesi): Likewise.
16013 (arm_load_exclusivedi): Likewise.
16014 (arm_load_acquire_exclusive<mode>): Likewise.
16015 (arm_load_acquire_exclusivesi): Likewise.
16016 (arm_load_acquire_exclusivedi): Likewise.
16017 (arm_store_exclusive<mode>): Likewise.
16018 (arm_store_exclusive<mode>): Likewise.
16019 (arm_store_release_exclusivedi): Likewise.
16020 (arm_store_release_exclusive<mode>): Likewise.
16022 2013-06-06 Richard Biener <rguenther@suse.de>
16024 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
16026 (lto_tag_is_tree_code_p): Adjust.
16027 (lto_tag_is_gimple_code_p): Likewise.
16028 (lto_gimple_code_to_tag): Likewise.
16029 (lto_tag_to_gimple_code): Likewise.
16030 (lto_tree_code_to_tag): Likewise.
16031 (lto_tag_to_tree_code): Likewise.
16032 * data-streamer.h (streamer_write_hwi_in_range): Use
16033 uhwi streaming to stream the normalized range.
16034 (streamer_read_hwi_in_range): Likewise.
16036 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16038 * config/arm/arm.md (enabled_for_depr_it): New attribute.
16039 (predicable_short_it): Likewise.
16040 (predicated): Likewise.
16041 (enabled): Handle above.
16042 (define_cond_exec): Set predicated attribute to yes.
16044 2013-06-05 Mike Stump <mikestump@comcast.net>
16046 * gdbinit.in (__FUNCTION__): Add.
16048 2013-06-05 Uros Bizjak <ubizjak@gmail.com>
16050 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
16051 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
16053 2013-06-05 Jan Hubicka <jh@suse.cz>
16055 * varasm.c (mark_decl_referenced): Revert the removal until targets
16058 2013-06-05 David Edelsohn <dje.gcc@gmail.com>
16060 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
16061 instead of mark_decl_referenced.
16063 2013-06-05 Jan Hubicka <jh@suse.cz>
16065 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
16066 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
16067 and symtab_used_from_object_file_p.
16068 (cgraph_make_node_local_1): Clear forced_by_abi.
16069 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
16070 * cgraph.h (symtab_node_base): Add forced_by_abi.
16071 (decide_is_variable_needed): Remove.
16072 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
16073 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
16074 (decide_is_symbol_needed): ... this one; handle symbols in general;
16075 always analyze virtuals; honnor forced_by_abi.
16076 (cgraph_finalize_function): Update.
16077 (varpool_finalize_decl): Update.
16078 (symbol_defined_and_needed): Remove.
16079 (analyze_functions): Update.
16080 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
16081 output_refs, input_overwrite_node): Handle forced_by_abi.
16082 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
16083 (address_taken_from_non_vtable_p): ... this one.
16084 (comdat_can_be_unshared_p_1): New function.
16085 (cgraph_comdat_can_be_unshared_p): Rename to ...
16086 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
16087 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
16088 (function_and_variable_visibility): Clear forced_by_abi as needed.
16089 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
16090 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
16091 * symtab.c (dump_symtab_base): Dump forced_by_abi.
16092 * varpool.c (decide_is_variable_needed): Remove.
16094 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16096 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
16097 (arm_option_override): Override arm_restrict_it where appropriate.
16098 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
16099 * config/arm/arm.opt (mrestrict-it): New command-line option.
16100 * doc/invoke.texi: Document -mrestrict-it.
16102 2013-06-05 David Malcolm <dmalcolm@redhat.com>
16104 * tsan.c (tsan_atomic_table): Make const.
16106 2013-06-05 Richard Biener <rguenther@suse.de>
16108 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
16109 index associated with the tree we are supposed to replace.
16110 * tree-streamer-out.c (pack_ts_base_value_fields): Output
16111 TREE_ASM_WRITTEN as zero for everything but SSA names.
16113 2013-06-05 David Malcolm <dmalcolm@redhat.com>
16115 * tree-ssa-structalias.c (call_stmt_vars): Make static.
16117 2013-06-04 Jan Hubicka <jh@suse.cz>
16119 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
16120 (input_node, input_varpool_node): Handle correctly external same
16122 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
16123 nodes at ltrans stage.
16125 2013-06-04 Jan Hubicka <jh@suse.cz>
16127 * ipa-inline.c (update_caller_keys): Fix availability test.
16128 (update_callee_keys): Likewise.
16129 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
16130 to follow ELF standard.
16132 2013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
16134 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
16135 (mips64r5900el-*-elf*): New configurations.
16136 * config/mips/mips-cpus.def (r5900): New processor.
16137 * config/mips/mips-tables.opt: Regenerate.
16138 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
16139 (mips_issue_rate): Handle PROCESSOR_R5900.
16140 (mips_reorg_process_insns): Force reorder mode for the R5900.
16141 * config/mips/mips.h (TARGET_MIPS5900): Define.
16142 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
16144 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
16146 * config/mips/mips.md (processor): Add r5900.
16147 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
16149 2013-06-04 Ian Bolton <ian.bolton@arm.com>
16151 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
16152 into function to generate MOVI instruction.
16153 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
16154 (aarch64_preferred_simd_mode): Turn into wrapper.
16155 (aarch64_output_scalar_simd_mov_immediate): New function.
16156 * config/aarch64/aarch64-protos.h: Add prototype for above.
16158 2013-06-04 Ian Bolton <ian.bolton@arm.com>
16160 * config/aarch64/aarch64.c (simd_immediate_info): Remove
16161 element_char member.
16162 (sizetochar): Return signed char.
16163 (aarch64_simd_valid_immediate): Remove elchar and other
16164 unnecessary variables.
16165 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
16166 Calculate element_char as required.
16167 * config/aarch64/aarch64-protos.h: Update and move prototype
16168 for aarch64_output_simd_mov_immediate.
16169 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
16172 2013-06-04 Ian Bolton <ian.bolton@arm.com>
16174 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
16175 information completed by aarch64_simd_valid_immediate.
16176 (aarch64_legitimate_constant_p): Update arguments.
16177 (aarch64_simd_valid_immediate): Work with struct rather than many
16179 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
16180 (aarch64_simd_make_constant): Update arguments.
16181 (aarch64_output_simd_mov_immediate): Work with struct rather than
16182 many pointers. Output immediate directly rather than as operand.
16183 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
16185 * config/aarch64/constraints.md (Dn): Update arguments.
16187 2013-06-04 Ian Bolton <ian.bolton@arm.com>
16189 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
16191 (aarch64_simd_immediate_valid_for_move): Remove.
16192 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
16193 (aarch64_simd_make_constant): Update call.
16194 (aarch64_output_simd_mov_immediate): Update call.
16195 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
16197 * config/aarch64/constraints.md (Dn): Update call.
16199 2013-06-04 Ian Bolton <ian.bolton@arm.com>
16201 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
16202 return type to bool for prototype.
16203 (aarch64_legitimate_constant_p): Check for true instead of not -1.
16204 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
16205 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
16207 2013-06-04 Catherine Moore <clm@codesourcery.com>
16209 * config/mips/mips.opt (meva): New.
16210 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
16211 (ASM_SPEC): Handle -meva.
16212 * doc/invoke.texi (meva): Document.
16214 2013-06-04 Alan Modra <amodra@gmail.com>
16216 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
16219 2013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16221 * rtl.def: Add extra fourth optional field to define_cond_exec.
16222 * gensupport.c (process_one_cond_exec): Process attributes from
16224 * doc/md.texi: Document fourth field in define_cond_exec.
16226 2013-06-04 Eric Botcazou <ebotcazou@adacore.com>
16228 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
16229 out the processing order as in store_bit_field_1.
16231 2013-06-04 Jan Hubicka <jh@suse.cz>
16233 PR middle-end/57500
16234 * cgraphunit.c (cgraph_process_same_body_aliases): Create
16235 non-VAR_DECL node if it does not exist yet.
16237 2013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
16239 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
16240 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
16241 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
16242 target_cpu_default setting.
16244 2013-06-03 Teresa Johnson <tejohnson@google.com>
16246 * dumpfile.c (opt_info_switch_p): Change -fopt-info
16247 default to -fopt-info=optimized instead of all.
16248 * doc/invoke.texi: Ditto.
16249 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
16250 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
16251 (execute_vect_slp): Emit BB vectorization success under
16252 MSG_OPTIMIZED_LOCATIONS.
16253 * tree-vect-slp.c (vect_slp_transform_bb): Change
16254 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
16255 * tree-vect-loop.c (vect_transform_loop): Ditto.
16257 2013-06-03 Jason Merrill <jason@redhat.com>
16260 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
16261 Use TARGET_EXPR for C++.
16263 2013-06-03 Jakub Jelinek <jakub@redhat.com>
16265 PR rtl-optimization/57268
16266 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
16267 if DEBUG_INSN_P (insn).
16270 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
16272 PR rtl-optimization/57268
16273 * sched-deps.c (sched_analyze_2): Flush dependence lists if
16274 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
16276 2013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
16278 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
16279 (ix86_avoid_lea_for_addr): Likewise.
16280 (exact_dependency_1): Likewise.
16281 (ix86_adjust_cost): Likewise.
16282 (swap_top_of_ready_list): Fix formatting and !reload_completed check
16284 (do_reorder_for_imul): Fix typo, formatting and
16285 !reload_completed check removed.
16286 (ix86_sched_reorder): Fix typo and formatting.
16287 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
16290 2013-06-03 Sofiane Naci <sofiane.naci@arm.com>
16292 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
16294 2013-06-03 Eric Botcazou <ebotcazou@adacore.com>
16296 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
16297 <STRING_CST>: Likewise.
16298 <VECTOR_CST>: Likewise.
16300 2013-06-01 Janus Weil <janus@gcc.gnu.org>
16301 Mikael Morin <mikael@gcc.gnu.org>
16303 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
16304 * config.in: Regenerated.
16305 * configure: Regenerated.
16307 2013-06-01 Jan Hubicka <jh@suse.cz>
16309 PR middle-end/57366
16310 * cgraphunit.c (compile): When weakref is not supported,
16311 set up transparent aliases before final output pass.
16312 * varasm.c (assemble_alias): Do not try to do it here.
16314 2013-06-01 Jan Hubicka <jh@suse.cz>
16316 PR middle-end/57467
16317 * passes.c (for_per_function): Skip unanalyzed functions.
16319 2013-06-01 Jan Hubicka <jh@suse.cz>
16321 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
16322 (lto_symtab_merge_symbols_1): ... this one.
16323 (lto_symtab_merge_cgraph_nodes): Rename to ...
16324 (lto_symtab_merge_symbols): ... this one; simplify.
16325 * cgraph.c (same_body_aliases_done): Rename to ...
16326 (cpp_implicit_aliases_done): ... this one.
16327 (cgraph_create_function_alias): Update.
16328 (cgraph_same_body_alias): Update.
16329 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
16330 (verify_edge_corresponds_to_fndecl): Simplify.
16331 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
16332 (cgraph_node): Remove same_body_alias.
16333 (varpool_node): Remove alias_of and extra_name_alias.
16334 (same_body_aliases_done): Rename to ..
16335 (cpp_implicit_aliases_done): ... this one.
16336 (symtab_alias_ultimate_target): Add default parameter.
16337 (symtab_resolve_alias): New function.
16338 (fixup_same_cpp_alias_visibility): Declare.
16339 (cgraph_function_node): Add default parameter.
16340 (cgraph_node_asm_name): Likewise.
16341 (cgraph_function_or_thunk_node): Add default parameter; do
16342 not ICE when it is NULL.
16343 (varpool_variable_node): Likewise.
16344 * tree-emutls.c (create_emultls_var): Update.
16345 (ipa_lower_emutls): Update.
16346 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
16347 (cgraph_reset_node): Reset alias info.
16348 (cgraph_finalize_function): Update.
16349 (fixup_same_cpp_alias_visibility): Move to symtab.c.
16350 (analyze_function): Simplify.
16351 (cgraph_process_same_body_aliases): Simplify.
16352 (analyze_functions): Fixup same body aliases.
16353 (handle_alias_pairs): Simplify.
16354 (assemble_thunk): Update.
16355 (assemble_thunks_and_aliases): Update.
16356 (output_weakrefs): Rewrite.
16357 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
16358 (lto_output_varpool_node): Likewise.
16359 (compute_ltrans_boundary): Remve assert.
16360 (get_alias_symbol): New functoin.
16361 (input_node): Rewrite alias handling.
16362 (input_varpool_node): Likewise.
16363 * ipa-pure-const.c (propagate_pure_const): Fix formating.
16364 * ipa.c (process_references): Handle weakrefs correctly.
16365 (symtab_remove_unreachable_nodes): Likewise.
16366 * trans-mem.c (get_cg_data): Update.
16367 (ipa_tm_create_version_alias): Update.
16368 (ipa_tm_execute): Update.
16369 * symtab.c (dump_symtab_base): Dump aliases.
16370 (verify_symtab_base): Verify aliases.
16371 (symtab_node_availability): New function.
16372 (symtab_alias_ultimate_target): Simplify.
16373 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
16374 handle all the fixup cases.
16375 (symtab_resolve_alias): New function.
16376 * passes.c (ipa_write_summaries): Handle weakrefs.
16377 * varpool.c (varpool_analyze_node): Simplify.
16378 (assemble_aliases): Update.
16379 (varpool_create_variable_alias): Simplify.
16380 (varpool_extra_name_alias): Simplify.
16381 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
16382 (lto_symtab_merge_symbols): ... this one.
16384 2013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
16387 PR rtl-optimization/57268
16388 * sched-deps.c (sched_analyze_2): Flush dependence lists if
16389 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
16391 2013-06-01 Tobias Burnus <burnus@net-b.de>
16393 Partially reverted:
16394 2013-05-31 Tobias Burnus <burnus@net-b.de>
16396 PR middle-end/57073
16397 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
16400 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
16402 PR rtl-optimization/57268
16403 * sched-deps.c (sched_analyze_2): Flush dependence lists if
16404 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
16406 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
16408 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
16409 unordered comparison operators when -fno-trapping-math is in effect
16411 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
16412 and implement unordered comparison operators properly on the e500.
16414 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
16416 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
16417 for constant scalar integers.
16418 (simplify_relational_operation_1): Likewise.
16420 2013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
16422 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
16423 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
16426 2013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
16427 Igor Zamyatin <igor.zamyatin@intel.com>
16429 Silvermont (SLM) architecture performance tuning.
16430 * config/i386/i386.h (enum ix86_tune_indices): Add
16431 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
16432 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
16434 * config/i386/i386.c (initial_ix86_tune_features)
16435 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
16436 (ix86_lea_outperforms): Handle Silvermont tuning.
16437 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
16439 (ix86_use_lea_for_mov): Likewise.
16440 (ix86_avoid_lea_for_addr): Likewise.
16441 (ix86_lea_for_add_ok): Likewise.
16442 (exact_dependency_1): New function.
16443 (exact_store_load_dependency): Likewise.
16444 (ix86_adjust_cost): Handle Silvermont tuning.
16445 (do_reoder_for_imul): Likewise.
16446 (swap_top_of_ready_list): New function.
16447 (ix86_sched_reorder): Changed to handle Silvermont tuning.
16449 * config/i386/i386.md (peepholes that split memory operand in fp
16452 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
16454 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
16455 Remove un-necessary braces.
16457 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
16459 * config/aarch64/aarch64.c (aarch64_classify_symbol):
16460 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
16462 2013-05-31 Tobias Burnus <burnus@net-b.de>
16464 PR middle-end/57073
16465 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
16467 2013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16470 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
16471 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
16472 * config/arm/neon.md (iordi3_neon): Remove.
16473 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
16474 * config/arm/predicates.md (imm_for_neon_logic_operand):
16475 Move to earlier in the file.
16476 (neon_logic_op2): Likewise.
16477 (arm_iordi_operand_neon): New predicate.
16479 2013-05-31 Richard Biener <rguenther@suse.de>
16481 PR tree-optimization/57478
16482 PR tree-optimization/57453
16483 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
16486 2013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
16488 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
16489 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
16491 2013-05-30 Tobias Burnus <burnus@net-b.de>
16492 Thomas Koenig <tkoenig@gcc.gnu.org>
16494 PR middle-end/57073
16495 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
16496 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
16498 2013-05-30 Steven Bosscher <steven@gcc.gnu.org>
16500 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
16502 2013-05-30 Vladimir Makarov <vmakarov@redhat.com>
16504 * target.def (register_usage_leveling_p): New hook.
16505 * targhooks.c (default_register_usage_leveling_p): New.
16506 * targhooks.h (default_register_usage_leveling_p): New prototype.
16507 * lra-assigns.c (register_usage_leveling_p): Use the hook.
16508 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
16509 * doc/tm.texi: Update.
16510 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
16512 2013-05-30 Ian Bolton <ian.bolton@arm.com>
16514 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
16515 (*insv_reg<mode>): New define_insn.
16517 2013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
16519 PR rtl-optimization/57439
16520 * postreload.c (move2add_valid_value_p): Check that we have
16521 a zero subreg_regno_offset when accessing the register in
16522 the requested mode.
16524 2013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
16525 Igor Zamyatin <igor.zamyatin@intel.com>
16527 Silvermont (SLM) architecture pipeline model, tuning and
16529 * config.gcc: Add slm config options and target.
16531 * config/i386/slm.md: New.
16533 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
16535 * config/i386/i386-c.c (ix86_target_macros_internal): New case
16537 (ix86_target_macros_internal): Likewise.
16539 * config/i386/i386.c (slm_cost): New cost.
16540 (m_SLM): New macro flag.
16541 (initial_ix86_tune_features): Set m_SLM.
16542 (x86_accumulate_outgoing_args): Likewise.
16543 (x86_arch_always_fancy_math_387): Likewise.
16544 (processor_target_table): Add slm cost.
16545 (cpu_names): Add slm cpu name.
16546 (x86_option_override_internal): Set SLM ISA.
16547 (ix86_issue_rate): New case PROCESSOR_SLM.
16548 (ia32_multipass_dfa_lookahead): Likewise.
16549 (fold_builtin_cpu): Add slm.
16551 * config/i386/i386.h (TARGET_SLM): New target macro.
16552 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
16553 (processor_type): Add PROCESSOR_SLM.
16555 * config/i386/i386.md (cpu): Add new value "slm".
16556 (slm.md): Include slm.md.
16558 2013-05-30 Bernd Schmidt <bernds@codesourcery.com>
16559 Zhenqiang Chen <zhenqiang.chen@linaro.org>
16561 * config/arm/arm-protos.h: Add and update function protos.
16562 * config/arm/arm.c (use_simple_return_p): New added.
16563 (thumb2_expand_return): Check simple_return flag.
16564 * config/arm/arm.md: Add simple_return and conditional simple_return.
16565 * config/arm/iterators.md: Add iterator for return and simple_return.
16567 2013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
16569 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
16570 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
16571 (arm_emit_vfp_multi_reg_pop): Likewise.
16572 (thumb2_emit_ldrd_pop): Likewise.
16573 (arm_expand_epilogue): Add misc REG_CFA notes.
16574 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
16576 2013-05-29 Lawrence Crowl <crowl@google.com>
16578 * config/arm/t-arm: Update for below.
16580 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
16581 Change type to hash_table. Update dependent calls and types.
16583 * config/i386/t-cygming: Update for below.
16585 * config/i386/t-interix: Update for below.
16587 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
16588 Change type to hash_table. Update dependent calls and types.
16589 (i386_find_on_wrapper_list::wrappers): Likewise.
16591 * config/ia64/t-ia64: Update for below.
16593 * config/ia64/ia64.c (bundle_state_table):
16594 Change type to hash_table. Update dependent calls and types.
16596 * config/mips/mips.c (mips_reorg_process_insns::htab):
16597 Change type to hash_table. Update dependent calls and types.
16599 * config/sol2.c (solaris_comdat_htab):
16600 Change type to hash_table. Update dependent calls and types.
16602 * config/t-sol2: Update for above.
16604 2013-05-29 Teresa Johnson <tejohnson@google.com>
16606 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
16607 functions are not yet marked as defined.
16609 2013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
16610 Pat Haugen <pthaugen@us.ibm.com>
16611 Peter Bergner <bergner@vnet.ibm.com>
16613 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
16617 (vrotl<mode>3): Likewise.
16618 (vashl<mode>3): Likewise.
16619 (vlshr<mode>3): Likewise.
16620 (vashr<mode>3): Likewise.
16622 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
16623 support for power8 V2DI builtins.
16625 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
16626 power8 V2DI builtins.
16627 (vupkhsw): Likewise.
16628 (vupklsw): Likewise.
16629 (vaddudm): Likewise.
16630 (vminsd): Likewise.
16631 (vmaxsd): Likewise.
16632 (vminud): Likewise.
16633 (vmaxud): Likewise.
16634 (vpkudum): Likewise.
16635 (vpksdss): Likewise.
16636 (vpkudus): Likewise.
16637 (vpksdus): Likewise.
16642 (vsubudm): Likewise.
16643 (vcmpequd): Likewise.
16644 (vcmpgtsd): Likewise.
16645 (vcmpgtud): Likewise.
16646 (vcmpequd_p): Likewise.
16647 (vcmpgtsd_p): Likewise.
16648 (vcmpgtud_p): Likewise.
16649 (vupkhsw): Likewise.
16650 (vupklsw): Likewise.
16651 (vaddudm): Likewise.
16652 (vmaxsd): Likewise.
16653 (vmaxud): Likewise.
16654 (vminsd): Likewise.
16655 (vminud): Likewise.
16656 (vpksdss): Likewise.
16657 (vpksdus): Likewise.
16658 (vpkudum): Likewise.
16659 (vpkudus): Likewise.
16664 (vsubudm): Likewise.
16666 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
16667 support for power8 V2DI instructions.
16669 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
16670 power8 V2DI instructions. Combine pack and unpack insns to use an
16671 iterator for each mode. Check whether a particular mode supports
16672 Altivec instructions instead of just checking TARGET_ALTIVEC.
16673 (UNSPEC_VPKUWUM): Likewise.
16674 (UNSPEC_VPKSHSS): Likewise.
16675 (UNSPEC_VPKSWSS): Likewise.
16676 (UNSPEC_VPKUHUS): Likewise.
16677 (UNSPEC_VPKSHUS): Likewise.
16678 (UNSPEC_VPKUWUS): Likewise.
16679 (UNSPEC_VPKSWUS): Likewise.
16680 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
16681 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
16682 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
16683 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
16684 (UNSPEC_VUPKHSB): Likewise.
16685 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
16686 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
16687 (UNSPEC_VUPKHSH): Likewise.
16688 (UNSPEC_VUPKLSB): Likewise.
16689 (UNSPEC_VUPKLSH): Likewise.
16691 (VI_char): Likewise.
16692 (VI_scalar): Likewise.
16693 (VI_unit): Likewise.
16695 (VP_small): Likewise.
16696 (VP_small_lc): Likewise.
16697 (VU_char): Likewise.
16698 (add<mode>3): Likewise.
16699 (altivec_vaddcuw): Likewise.
16700 (altivec_vaddu<VI_char>s): Likewise.
16701 (altivec_vadds<VI_char>s): Likewise.
16702 (sub<mode>3): Likewise.
16703 (altivec_vsubcuw): Likewise.
16704 (altivec_vsubu<VI_char>s): Likewise.
16705 (altivec_vsubs<VI_char>s): Likewise.
16706 (altivec_vavgs<VI_char>): Likewise.
16707 (altivec_vcmpbfp): Likewise.
16708 (altivec_eq<mode>): Likewise.
16709 (altivec_gt<mode>): Likewise.
16710 (altivec_gtu<mode>): Likewise.
16711 (umax<mode>3): Likewise.
16712 (smax<mode>3): Likewise.
16713 (umin<mode>3): Likewise.
16714 (smin<mode>3): Likewise.
16715 (altivec_vpkuhum): Likewise.
16716 (altivec_vpkuwum): Likewise.
16717 (altivec_vpkshss): Likewise.
16718 (altivec_vpkswss): Likewise.
16719 (altivec_vpkuhus): Likewise.
16720 (altivec_vpkshus): Likewise.
16721 (altivec_vpkuwus): Likewise.
16722 (altivec_vpkswus): Likewise.
16723 (altivec_vpks<VI_char>ss): Likewise.
16724 (altivec_vpks<VI_char>us): Likewise.
16725 (altivec_vpku<VI_char>us): Likewise.
16726 (altivec_vpku<VI_char>um): Likewise.
16727 (altivec_vrl<VI_char>): Likewise.
16728 (altivec_vsl<VI_char>): Likewise.
16729 (altivec_vsr<VI_char>): Likewise.
16730 (altivec_vsra<VI_char>): Likewise.
16731 (altivec_vsldoi_<mode>): Likewise.
16732 (altivec_vupkhsb): Likewise.
16733 (altivec_vupkhs<VU_char>): Likewise.
16734 (altivec_vupkls<VU_char>): Likewise.
16735 (altivec_vupkhsh): Likewise.
16736 (altivec_vupklsb): Likewise.
16737 (altivec_vupklsh): Likewise.
16738 (altivec_vcmpequ<VI_char>_p): Likewise.
16739 (altivec_vcmpgts<VI_char>_p): Likewise.
16740 (altivec_vcmpgtu<VI_char>_p): Likewise.
16741 (abs<mode>2): Likewise.
16742 (vec_unpacks_hi_v16qi): Likewise.
16743 (vec_unpacks_hi_v8hi): Likewise.
16744 (vec_unpacks_lo_v16qi): Likewise.
16745 (vec_unpacks_hi_<VP_small_lc>): Likewise.
16746 (vec_unpacks_lo_v8hi): Likewise.
16747 (vec_unpacks_lo_<VP_small_lc>): Likewise.
16748 (vec_pack_trunc_v8h): Likewise.
16749 (vec_pack_trunc_v4si): Likewise.
16750 (vec_pack_trunc_<mode>): Likewise.
16752 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
16754 (vec_vmaxsd): Likewise.
16755 (vec_vmaxud): Likewise.
16756 (vec_vminsd): Likewise.
16757 (vec_vminud): Likewise.
16758 (vec_vpksdss): Likewise.
16759 (vec_vpksdus): Likewise.
16760 (vec_vpkudum): Likewise.
16761 (vec_vpkudus): Likewise.
16762 (vec_vrld): Likewise.
16763 (vec_vsld): Likewise.
16764 (vec_vsrad): Likewise.
16765 (vec_vsrd): Likewise.
16766 (vec_vsubudm): Likewise.
16767 (vec_vupkhsw): Likewise.
16768 (vec_vupklsw): Likewise.
16770 2013-05-29 Jan Hubicka <jh@suse.cz>
16772 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
16773 flags; reorder rest of fields in more consistent way.
16774 (varpool_node): Remove analyzed, finalized and alias.
16775 (cgraph_ndoe): Likewise.
16776 (symtab_alias_ultimate_target): New function.
16777 (cgraph_function_node): Move offline.
16778 (cgraph_reset_node): Declare.
16779 (cgraph_comdat_can_be_unshared_p): Remove.
16780 (varpool_remove_initializer): Declare.
16781 (varpool_first_defined_variable, varpool_next_defined_variable
16782 cgraph_first_defined_function, cgraph_next_defined_function): Update.
16783 (cgraph_function_with_gimple_body_p): Update.
16784 (varpool_all_refs_explicit_p): Update.
16785 (symtab_alias_target): New function.
16786 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
16787 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
16788 (cgraph_function_or_thunk_node): Simplify using
16789 symtab_alias_ultimate_target.
16790 (varpool_variable_node): Likewise.
16791 * cgraph.c (cgraph_create_function_alias): Update.
16792 (cgraph_add_thunk): Update.
16793 (cgraph_remove_node): Update.
16794 (dump_cgraph_node): Do not dump removed flags.
16795 (cgraph_function_body_availability): Update.
16796 (cgraph_propagate_frequency): Update.
16797 (verify_cgraph_node): Check sanity of local flag.
16798 (cgraph_function_node): Move here from cgraph.h; revamp for
16799 cgraph_function_or_thunk_node.
16800 * lto-symtab.c (lto_varpool_replace_node): Update.
16801 (lto_symtab_resolve_can_prevail_p): Update.
16802 (lto_symtab_merge_cgraph_nodes): Update.
16803 * ipa-cp.c (determine_versionability, initialize_node_lattices,
16804 propagate_constants_accross_call, devirtualization_time_bonus,
16805 ipcp_propagate_stage): Update.
16806 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
16807 * ipa-inline-transform.c (clone_inlined_nodes,
16808 preserve_function_body_p): Update.
16809 * ipa-reference.c (propagate): Update.
16810 (write_node_summary_p): Update.
16811 * toplev.c (wrapup_global_declaration_2): Update.
16812 * cgraphunit.c (cgraph_analyze_function): Rename to ...
16813 (analyze_function) ... this one.
16814 (cgraph_process_new_functions): Update.
16815 (cgraph_reset_node): Export.
16816 (cgraph_finalize_function): Update.
16817 (cgraph_add_new_function): Update.
16818 (process_function_and_variable_attributes): Update.
16819 (varpool_finalize_decl): Update.
16820 (symbol_finalized): Remove.
16821 (symbol_finalized_and_needed): Rename to ...
16822 (symbol_defined_and_needed): ... update.
16823 (cgraph_analyze_functions): Update.
16824 (handle_alias_pairs): Update.
16825 (mark_functions_to_output): Update.
16826 (assemble_thunk): Update.
16827 (output_in_order): Update.
16828 (output_weakrefs): Update.
16829 (finalize_compilation_unit): Update.
16830 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
16831 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
16832 input_node, input_varpool_node): Update.
16833 * dbxout.c (dbxout_expand_expr): Update.
16834 * cgraphclones.c (cgraph_clone_node): Update.
16835 (cgraph_copy_node_for_versioning): Update.
16836 (cgraph_materialize_clone): Update.
16837 (cgraph_materialize_all_clones): Update.
16838 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
16839 propagate_pure_const, propagate_nothrow): Update.
16840 * lto-streamer-out.c (lto_output, write_symbol): Update.
16841 * ipa-utils.c (ipa_reverse_postorder): Update.
16842 * ipa-inline.c (can_inline_edge_p): Update.
16843 (update_caller_keys, ipa_inline): Update.
16844 * dwarf2out.c (reference_to_unused,
16845 premark_types_used_by_global_vars_helper): Update.
16846 * tree-eh.c (tree_could_trap_p): Update.
16847 * ipa-split.c (consider_split, execute_split_functions): Update.
16848 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
16849 has_addr_references_p): Update; move ahead in file for better
16851 (process_references): Simplify.
16852 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
16853 bodies are removed.
16854 (cgraph_comdat_can_be_unshared_p): Make static.
16855 (cgraph_externally_visible_p): Update.
16856 (varpool_externally_visible_p): Update.
16857 (function_and_variable_visibility): Update.
16858 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
16859 ipa_tm_mark_force_output_node): Update.
16860 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
16861 estimate_edge_devirt_benefit, inline_generate_summary,
16862 inline_write_summary): Update.
16863 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
16864 * ipa-prop.c (ipa_compute_jump_functions): Update.
16865 (ipa_print_node_params, ipa_prop_read_section,
16866 ipa_update_after_lto_read, read_replacements_section): Update.
16867 * varasm.c (mark_decl_referenced): Update.
16868 (assemble_alias, dump_tm_clone_pairs): Update.
16869 * tree-inline.c (copy_bb): Update.
16870 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
16872 * symtab.c (dump_symtab_base): Print new flags.
16873 (verify_symtab_base): Verify new flags.
16874 (symtab_alias_ultimate_target): New function.
16875 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
16876 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
16878 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
16880 * i386.c (ix86_get_function_versions_dispatcher,
16881 ix86_generate_version_dispatcher_body): Update.
16882 (fold_builtin_cpu): Use varpool_add_new_variable.
16883 * varpool.c (varpool_remove_initializer): Break out from ...
16884 (varpool_remove_node): ... this one.
16885 (dump_varpool_node, varpool_node_for_asm,
16886 cgraph_variable_initializer_availability, varpool_analyze_node,
16887 varpool_assemble_decl, varpool_remove_unreferenced_decls,
16888 varpool_finalize_named_section_flags, varpool_create_variable_alias):
16891 2013-05-29 Jan Hubicka <jh@suse.cz>
16893 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
16895 2013-05-29 Easwaran Raman <eraman@google.com>
16897 PR tree-optimization/57442
16898 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
16899 when control exits the main loop.
16901 2013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
16903 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
16905 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
16906 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
16907 * rx/t-rx: Add rx100 under multi library matches option for nofpu
16910 2013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
16912 PR tree-optimization/57441
16913 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
16914 Don't limit size of incr_vec to number of candidates.
16916 2013-05-29 Steve Ellcey <sellcey@imgtec.com>
16918 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
16919 and mips16 directories.
16920 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
16921 (MULTILIB_DIRNAMES): Ditto.
16922 (MULTILIB_EXCEPTIONS): Add new exceptions.
16923 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
16924 (MULTILIB_DIRNAMES): Ditto.
16925 (MULTILIB_EXCEPTIONS): Add new exceptions.
16927 2012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
16928 Marcus Shawcroft <marcus.shawcroft@arm.com>
16930 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
16931 SYMBOL_TINY_ABSOLUTE.
16932 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
16933 SYMBOL_TINY_ABSOLUTE.
16934 (aarch64_expand_mov_immediate): Likewise.
16935 (aarch64_classify_symbol): Likewise.
16936 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
16937 Permit SYMBOL_TINY_ABSOLUTE.
16938 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
16940 2013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
16941 Marcus Shawcroft <marcus.shawcroft@arm.com>
16943 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
16944 Refactor if/switch. Replace gcc_assert with if.
16946 2013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
16948 * config/i386/i386.c (initial_ix86_tune_features): Enable
16949 FP Reassociation for AMD bdver1 and bdver2.
16951 2013-05-29 Martin Jambor <mjambor@suse.cz>
16953 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
16954 and IMAGPART_EXPR do not occur within other handled_components.
16956 2013-05-29 Richard Biener <rguenther@suse.de>
16958 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
16959 access on whether the use is in the BB we currently try to
16961 (vect_bb_vectorization_profitable_p): Pass the BB we currently
16962 vectorize to vect_bb_slp_scalar_cost.
16964 2013-05-29 Richard Biener <rguenther@suse.de>
16966 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
16967 computing scalar cost offsetted by stmts that are kept live
16969 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
16970 for computation of scalar cost.
16972 2013-05-28 Steve Ellcey <sellcey@mips.com>
16974 * config/mips/mips-cpus.def (mips32r2): Change processor type.
16976 2013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
16978 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
16979 array notation built-in reduction functions.
16980 * doc/passes.texi (Passes): Added documentation about changes done
16982 * doc/invoke.texi (C Dialect Options): Added documentation about
16983 the -fcilkplus flag.
16984 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
16985 (BUILTINS_DEF): Depend on cilkplus.def.
16986 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
16987 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
16988 * cilkplus.def: New file.
16990 2013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
16992 PR rtl-optimization/57439
16993 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
16995 2013-05-28 Easwaran Raman <eraman@google.com>
16997 PR tree-optimization/57337
16998 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
16999 (find_insert_point): Correctly identify the insertion point
17000 when two statements with the same UID is compared.
17002 2013-05-28 Richard Biener <rguenther@suse.de>
17004 PR tree-optimization/56787
17005 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
17006 from the list of data references.
17007 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
17009 (vect_analyze_loop_operations): Likewise.
17010 (vect_transform_loop): Remove clobbers.
17012 2013-05-28 Martin Jambor <mjambor@suse.cz>
17014 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
17015 and REALPART_EXPRs have scalar type.
17017 2013-05-28 Richard Biener <rguenther@suse.de>
17019 PR tree-optimization/57411
17020 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
17022 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
17023 virtual operand propagation.
17025 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
17027 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
17028 destination register for bmasksi_vis.
17029 (vector_init_bshuffle): Likewise.
17030 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
17032 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
17034 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
17035 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
17036 mode if the instruction isn't available in the original mode.
17037 * config/sparc/sparc.opt (mfix-ut699): New option.
17038 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
17039 (divdf3): Turn into expander.
17040 (divdf3_nofix): New insn.
17041 (divdf3_fix): Likewise.
17042 (divsf3): Disable if -mfix-ut699.
17043 (sqrtdf2): Turn into expander.
17044 (sqrtdf2_nofix): New insn.
17045 (sqrtdf2_fix): Likewise.
17046 (sqrtsf2): Disable if -mfix-ut699.
17048 2013-05-27 Richard Biener <rguenther@suse.de>
17050 PR middle-end/57412
17051 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
17052 block for the new loop.
17054 2013-05-27 Richard Biener <rguenther@suse.de>
17056 PR tree-optimization/57343
17057 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
17058 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
17059 (number_of_iterations_cond): Do not build the folded tree.
17061 2013-05-27 Richard Biener <rguenther@suse.de>
17064 PR middle-end/57381
17065 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
17066 OEP_CONSTANT_ADDRESS_OF retained.
17068 PR tree-optimization/57417
17069 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
17070 for unchanged base.
17071 (set_ssa_val_to): Compare addresses using
17072 get_addr_base_and_unit_offset.
17074 2013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
17076 PR rtl-optimization/56833
17077 * postreload.c (move2add_record_mode): New function.
17078 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
17079 (move2add_use_add2_insn): Use move2add_record_sym_value.
17080 (move2add_use_add3_insn): Likewise.
17081 (reload_cse_move2add): Use move2add_valid_value_p and
17082 move2add_record_mode. Invalidate call-clobbered and REG_INC
17083 affected regs by setting reg_mode to VOIDmode.
17084 (move2add_note_store): Don't pretend the inside of a SUBREG is
17085 the actual destination. Invalidate single/leading registers by
17086 setting reg_mode to VOIDmode.
17087 Use move2add_record_sym_value, move2add_valid_value_p and
17088 move2add_record_mode.
17090 2013-05-27 Richard Biener <rguenther@suse.de>
17092 PR tree-optimization/57396
17093 * tree-affine.c (double_int_constant_multiple_p): Properly
17094 return false for val == 0 and div != 0.
17096 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
17098 * config/mips/mips.h: Use #elif in preprocessor conditions.
17100 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
17103 * config/mips/constraints.md (kl): New constraint.
17104 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
17105 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
17106 constraint for operand 0. Split after CSE for MIPS16. Emit a move
17107 from LO for MIPS16.
17108 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
17110 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
17113 * config/mips/mips.c (mips_can_inline_p): New function.
17114 (TARGET_CAN_INLINE_P): Define.
17116 2013-05-25 Steven Bosscher <steven@gcc.gnu.org>
17118 * sched-int.h (ds_t, dw_t): Make unsigned int.
17119 Fix documentation that describes how all the ds_t bits are used.
17120 Reserve the last bit for delayed-branch scheduling.
17121 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
17122 (BITS_PER_DEP_WEAK): Fix definition and documentation.
17123 (gen_dep_weak_1): Remove prototype.
17124 * sched-deps.c (get_dep_weak_1): Make static.
17125 * target.def (speculate_insn, needs_block_p, gen_spec_check,
17126 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
17127 * doc/tm.texi: Regenerate.
17128 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
17130 2013-05-24 Steven Bosscher <steven@gcc.gnu.org>
17133 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
17135 2013-05-24 Nathan Sidwell <nathan@codesourcery.com>
17136 Sandra Loosemore <sandra@codesourcery.com>
17138 * config.gcc (powerpc-*): Allow native for with-cpu.
17140 2013-05-24 Jeff Law <law@redhat.com>
17142 PR tree-optimization/57124
17143 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
17144 conversion feeding a condition if the range has an overflow
17145 if -fstrict-overflow. Add warnings for when we do make the
17148 2013-05-24 Dehao Chen <dehao@google.com>
17150 * tree-cfg.c (locus_discrim_map): Fix the typo.
17151 (locus_discrim_hasher): Likewise.
17152 (locus_discrim_hasher::hash): Likewise.
17153 (locus_discrim_hasher::equal): Likewise.
17155 2013-05-24 Martin Jambor <mjambor@suse.cz>
17157 PR tree-optimization/57294
17158 * cgraph.h (ipa_record_stmt_references): Declare.
17159 * cgraphbuild.c (ipa_record_stmt_references): New function.
17160 (build_cgraph_edges): Use ipa_record_stmt_references.
17161 (rebuild_cgraph_edges): Likewise.
17162 (cgraph_rebuild_references): Likewise.
17163 * ipa-prop.c (ipa_modify_call_arguments): Discard references
17164 associated with the old statement and build references from the
17165 newly built statements.
17166 * ipa-ref.c (ipa_remove_stmt_references): New function.
17167 * ipa-ref.h (ipa_remove_stmt_references): Declare.
17169 2013-05-24 Vladimir Makarov <vmakarov@redhat.com>
17171 * lra-constraints.c (emit_spill_move): Use smaller mode for
17173 (check_and_process_move): Consider mem-reg moves for secondary
17175 (curr_insn_transform): Don't lose insns emitted before for
17176 secondary memory moves.
17177 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
17178 reg set up in the current insn.
17180 2013-05-24 Dehao Chen <dehao@google.com>
17182 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
17184 (locus_descrim_hasher::equal): Likewise.
17185 (build_gimple_cfg): New discriminator assignment algorithm.
17186 (make_edges): Likewise.
17187 (next_discriminator_for_locus): Likewise.
17188 (same_line_p): Likewise.
17189 (assign_discriminators): Likewise.
17190 (make_cond_expr_edges): Likewise.
17191 (make_gimple_switch_edges): Likewise.
17192 (make_goto_expr_edges): Likewise.
17193 (make_gimple_asm_edges): Likewise.
17195 2013-05-24 Ian Bolton <ian.bolton@arm.com>
17197 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
17198 X format specifier to only display bottom 16 bits.
17199 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
17200 immediate to match for operand 2, since it will be masked.
17202 2013-05-24 Richard Biener <rguenther@suse.de>
17204 PR tree-optimization/57287
17205 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
17206 all SSA names that occur in abnormal PHIs.
17208 2013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
17211 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
17212 that index is not negative.
17214 2013-05-24 Eric Botcazou <ebotcazou@adacore.com>
17216 PR rtl-optimization/55177
17217 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
17218 (simplify_byte_swapping_operation): New.
17219 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
17220 (simplify_relational_operation_1): Deal with BSWAP.
17222 2013-05-23 Richard Henderson <rth@redhat.com>
17225 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
17226 (ix86_reorg): Call it.
17228 2013-05-23 Uros Bizjak <ubizjak@gmail.com>
17231 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
17232 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
17233 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
17235 2013-05-23 Christian Bruel <christian.bruel@st.com>
17238 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
17240 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
17241 Marcus Shawcroft <marcus.shawcroft@arm.com>
17243 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
17244 * config/aarch64/constraints.md (Usa): Remove.
17245 * doc/md.texi (AArch64 Usa): Remove.
17247 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
17248 Marcus Shawcroft <marcus.shawcroft@arm.com>
17250 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
17251 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
17252 * config/aarch64/predicates.md (aarch64_const_address): Remove.
17253 (aarch64_mov_operand): Use aarch64_mov_operand_p.
17255 2013-05-23 Vidya Praveen <vidyapraveen@arm.com>
17257 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
17258 instruction (AdvSIMD).
17259 * config/aarch64/aarch64-builtins.c
17260 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
17261 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
17263 2013-05-23 Martin Jambor <mjambor@suse.cz>
17265 PR middle-end/57347
17266 * tree.h (contains_bitfld_component_ref_p): Declare.
17267 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
17268 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
17270 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
17271 not access a bit-field. Assert all final offsets are byte-aligned.
17273 2013-05-23 Richard Biener <rguenther@suse.de>
17275 PR tree-optimization/57380
17276 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
17277 least one invariant or re-used load.
17278 * passes.c (init_optimization_passes): Move pass_phiprop before
17281 2013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
17283 * config/aarch64/aarch64-simd.md
17284 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
17286 2013-05-23 Richard Biener <rguenther@suse.de>
17288 PR middle-end/57381
17289 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
17290 OEP_CONSTANT_ADDRESS_OF retained.
17292 2013-05-23 Jakub Jelinek <jakub@redhat.com>
17294 PR middle-end/57344
17295 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
17296 don't lower unit. Handle unit not being always BITS_PER_WORD.
17298 2013-05-23 Richard Biener <rguenther@suse.de>
17300 PR rtl-optimization/57341
17301 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
17302 instead of true_dependence.
17304 2013-05-22 David Malcolm <dmalcolm@redhat.com>
17306 * bb-reorder.c (branch_threshold): Make const.
17307 (exec_threshold): Ditto.
17309 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
17310 Pat Haugen <pthaugen@us.ibm.com>
17311 Peter Bergner <bergner@vnet.ibm.com>
17313 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
17314 documentation for the power8 crypto builtins.
17316 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
17318 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
17319 macros for defining power8 builtin functions.
17320 (BU_P8V_AV_2): Likewise.
17321 (BU_P8V_AV_P): Likewise.
17322 (BU_P8V_VSX_1): Likewise.
17323 (BU_P8V_OVERLOAD_1): Likewise.
17324 (BU_P8V_OVERLOAD_2): Likewise.
17325 (BU_CRYPTO_1): Likewise.
17326 (BU_CRYPTO_2): Likewise.
17327 (BU_CRYPTO_3): Likewise.
17328 (BU_CRYPTO_OVERLOAD_1): Likewise.
17329 (BU_CRYPTO_OVERLOAD_2): Likewise.
17330 (XSCVSPDP): Fix typo, point to the correct instruction.
17331 (VCIPHER): Add power8 crypto builtins.
17332 (VCIPHERLAST): Likewise.
17333 (VNCIPHER): Likewise.
17334 (VNCIPHERLAST): Likewise.
17335 (VPMSUMB): Likewise.
17336 (VPMSUMH): Likewise.
17337 (VPMSUMW): Likewise.
17338 (VPERMXOR_V2DI): Likewise.
17339 (VPERMXOR_V4SI: Likewise.
17340 (VPERMXOR_V8HI: Likewise.
17341 (VPERMXOR_V16QI: Likewise.
17342 (VSHASIGMAW): Likewise.
17343 (VSHASIGMAD): Likewise.
17344 (VPMSUM): Likewise.
17345 (VPERMXOR): Likewise.
17346 (VSHASIGMA): Likewise.
17348 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
17349 __CRYPTO__ if the crypto instructions are available.
17350 (altivec_overloaded_builtins): Add support for overloaded power8
17353 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
17354 support for power8 crypto builtins.
17355 (builtin_function_type): Likewise.
17356 (altivec_init_builtins): Add support for builtins that take vector
17357 long long (V2DI) arguments.
17359 * config/rs6000/crypto.md: New file, define power8 crypto
17362 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
17363 Pat Haugen <pthaugen@us.ibm.com>
17364 Peter Bergner <bergner@vnet.ibm.com>
17366 * doc/invoke.texi (Option Summary): Add power8 options.
17367 (RS/6000 and PowerPC Options): Likewise.
17369 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
17370 constraints.md instead of rs6000.h. Reorder w* constraints. Add
17371 wm, wn, wr documentation.
17373 * config/rs6000/constraints.md (wm): New constraint for VSX
17374 registers if direct move instructions are enabled.
17375 (wn): New constraint for no registers.
17376 (wq): New constraint for quad word even GPR registers.
17377 (wr): New constraint if 64-bit instructions are enabled.
17378 (wv): New constraint if power8 vector instructions are enabled.
17379 (wQ): New constraint for quad word memory locations.
17381 * config/rs6000/predicates.md (const_0_to_15_operand): New
17382 constraint for 0..15 for crypto instructions.
17383 (gpc_reg_operand): If VSX allow registers in VSX registers as well
17384 as GPR and floating point registers.
17385 (int_reg_operand): New predicate to match only GPR registers.
17386 (base_reg_operand): New predicate to match base registers.
17387 (quad_int_reg_operand): New predicate to match even GPR registers
17388 for quad memory operations.
17389 (vsx_reg_or_cint_operand): New predicate to allow vector logical
17390 operations in both GPR and VSX registers.
17391 (quad_memory_operand): New predicate for quad memory operations.
17392 (reg_or_indexed_operand): New predicate for direct move support.
17394 * config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
17395 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
17396 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
17397 (POWERPC_MASKS): Add power8 options.
17398 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
17401 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
17402 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
17404 * config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
17405 (-mpower8-fusion): New power8 options.
17406 (-mpower8-fusion-sign): Likewise.
17407 (-mpower8-vector): Likewise.
17408 (-mcrypto): Likewise.
17409 (-mdirect-move): Likewise.
17410 (-mquad-memory): Likewise.
17412 * config/rs6000/rs6000.c (power8_cost): Initial definition for power8.
17413 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
17415 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
17416 (rs6000_debug_vector_unit): Add p8_vector.
17417 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
17418 definitions. Also print fusion state.
17419 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
17420 (rs6000_builtin_mask_calculate): Add power8 builtin support.
17421 (rs6000_option_override_internal): Add support for power8.
17422 (rs6000_common_init_builtins): Add debugging for skipped builtins
17423 if -mdebug=builtin.
17424 (rs6000_adjust_cost): Add power8 support.
17425 (rs6000_issue_rate): Likewise.
17426 (insn_must_be_first_in_group): Likewise.
17427 (insn_must_be_last_in_group): Likewise.
17428 (force_new_group): Likewise.
17429 (rs6000_register_move_cost): Likewise.
17430 (rs6000_opt_masks): Likewise.
17432 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
17433 power8 capable assembler, default to power7 options.
17434 (TARGET_DIRECT_MOVE): Likewise.
17435 (TARGET_CRYPTO): Likewise.
17436 (TARGET_P8_VECTOR): Likewise.
17437 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
17438 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
17439 (VECTOR_MEM_P8_VECTOR_P): Likewise.
17440 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
17441 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
17442 (TARGET_XSCVDPSPN): Likewise.
17443 (TARGET_XSCVSPDPN): Likewsie.
17444 (TARGET_SYNC_HI_QI): Likewise.
17445 (TARGET_SYNC_TI): Likewise.
17446 (MASK_CRYPTO): Likewise.
17447 (MASK_DIRECT_MOVE): Likewise.
17448 (MASK_P8_FUSION): Likewise.
17449 (MASK_P8_VECTOR): Likewise.
17450 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
17451 temporary used by some of the direct move instructions to get two FP
17452 temporary registers does not force creation of a stack frame.
17453 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
17454 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
17455 that any VSX registers are tieable, even if they are also an
17456 Altivec vector mode.
17457 (r6000_reg_class_enum): Add wm, wr, wv constraints.
17458 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
17459 (RS6000_BTM_CRYPTO): Likewise.
17460 (RS6000_BTM_COMMON): Likewise.
17462 * config/rs6000/rs6000.md (cpu attribute): Add power8.
17463 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
17464 (enum rs6000_vector): Add power8 vector support.
17466 2013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
17470 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
17471 (any_sibcall_could_use_r3): this and handle indirect calls.
17472 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
17474 2013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17476 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
17478 2013-05-22 Richard Biener <rguenther@suse.de>
17480 PR middle-end/57349
17481 * profile.c (branch_prob): Do not split blocks that are
17482 abnormally receiving from ECF_RETURNS_TWICE functions.
17484 2013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
17486 * recog.c (offsettable_address_addr_space_p): Fix calculation of
17487 address mode. Move pointer mode initialization to the same place.
17489 2013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
17491 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
17492 while it has any effect.
17494 2013-05-21 Easwaran Raman <eraman@google.com>
17496 PR tree-optimization/57322
17497 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
17498 UID of the statement added to the BB to be 1.
17500 2013-05-21 Jakub Jelinek <jakub@redhat.com>
17502 PR tree-optimization/57331
17503 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
17504 of conversion from pointer type to integral type with integer.
17506 2013-05-21 Martin Jambor <mjambor@suse.cz>
17509 * ipa-prop.c (ipa_read_node_info): Process param_used and
17510 controlled_uses in the same order as when writing.
17512 2013-05-21 Magnus Granberg <baldrick@free.fr>
17515 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
17517 2013-05-21 Richard Biener <rguenther@suse.de>
17519 PR tree-optimization/57318
17520 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
17521 estimate stmts with side-effects as likely eliminated.
17523 2013-05-21 Richard Biener <rguenther@suse.de>
17525 PR tree-optimization/57330
17526 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
17527 preserve the call stmts fntype.
17529 2013-05-21 Richard Biener <rguenther@suse.de>
17531 PR tree-optimization/57303
17532 * tree-ssa-sink.c (statement_sink_location): Improve killing
17533 stmt detection and properly handle self-assignments.
17535 2013-05-21 Christian Bruel <christian.bruel@st.com>
17537 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
17538 spanning registers. LEAF_REG_REMAP is supported only for contiguous
17539 registers. Set register size out of the PARALLEL loop.
17541 2013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
17544 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
17545 (*fmasf4, *fmasf4_media): New insns.
17547 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
17549 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
17550 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
17551 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
17552 (mips_idiv_insns): Update the comments to say that the returned
17553 instruction counts are in units of BASE_INSN_LENGTH.
17554 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
17555 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
17556 using 2 rather than 4 as the length of indirect MIPS16 and
17557 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
17558 length of a NOP. Don't divide MIPS16 lengths by 2.
17559 (mips16_split_long_branches): Assume a branch is long if the
17560 length is greater than 4 rather than 8.
17561 * config/mips/mips.md (length): Give MIPS16 lengths directly,
17562 rather than multiplying them by 2. Multiply instruction counts
17563 by BASE_INSN_LENGTH rather than 4.
17564 (*jump_mips16, tls_get_tp_mips16_<mode>)
17565 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
17567 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
17569 * config/mips/mips.md (extended_mips16): Remove branch case.
17570 (length): Remove duplicated extended_mips16 test.
17572 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
17574 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
17576 2013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
17578 * recog.h (Recog_data): Rename to...
17579 (recog_data_d): ...this.
17580 (recog_data): Update accordingly.
17581 * recog.c (recog_data): Likewise.
17582 * reload.c (save_recog_data): Likewise.
17583 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
17584 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
17586 2013-05-17 Julian Brown <julian@codesourcery.com>
17588 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
17589 found in a REG_EQUAL note, invalidate it.
17591 2013-05-17 Easwaran Raman <eraman@google.com>
17593 * tree-ssa-reassoc.c (find_insert_point): New function.
17594 (insert_stmt_after): Likewise.
17595 (get_def_stmt): Likewise.
17596 (ensure_ops_are_available): Likewise.
17597 (not_dominated_by): Likewise.
17598 (rewrite_expr_tree): Do not move statements beyond what is
17599 necessary. Remove call to swap_ops_for_binary_stmt...
17600 (reassociate_bb): ... and move it here.
17601 (build_and_add_sum): Assign UIDs for new statements.
17602 (linearize_expr): Likewise.
17603 (do_reassoc): Renumber gimple statement UIDs.
17605 2013-05-17 Jan Hubicka <jh@suse.cz>
17607 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
17609 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
17610 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
17612 (output_weakrefs): Update.
17614 2013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
17615 Martin Jambor <mjambor@suse.cz>
17617 PR middle-end/57276
17618 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
17619 value that corresponds to the given aggval is found in values vector.
17621 2013-05-17 Uros Bizjak <ubizjak@gmail.com>
17623 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
17624 sse, sse2, sse3, ssse3 and sse4a flags to options.
17626 2013-05-17 David Malcolm <dmalcolm@redhat.com>
17628 * gengtype-state.c: (s_expr_writer): New class, to handle
17629 prettifying of output layout of s-expressions.
17630 (state_writer): New class, to write out gtype.state.
17631 (state_written_type_count): Move this variable into member data of
17633 (s_expr_writer::s_expr_writer): New code: constructor for new class
17634 (state_writer::state_writer(): ditto
17635 (s_expr_writer::write_new_line): New function
17636 (s_expr_writer::write_any_indent): ditto
17637 (s_expr_writer::begin_s_expr): ditto
17638 (s_expr_writer::end_s_expr): ditto
17639 (write_state_fileloc): convert to method of state_writer...
17640 (state_writer:: write_state_fileloc): ...and use methods of
17641 s_expr_writer to write indentation into the gtype.state output file
17642 to visually represent the hierarchical structure of the list
17644 (write_state_fields): ditto, renaming to...
17645 (state_writer::write_state_fields)
17646 (write_state_a_string): ditto, renaming to...
17647 (state_writer::write_state_a_string)
17648 (write_state_string_option): ditto, renaming to...
17649 (state_writer::write_state_string_option)
17650 (write_state_type_option): ditto, renaming to...
17651 (state_writer::write_state_type_option)
17652 (write_state_nested_option): ditto, renaming to...
17653 (state_writer::write_state_nested_option)
17654 (write_state_option): ditto, renaming to...
17655 (state_writer::write_state_option)
17656 (write_state_options): ditto, renaming to...
17657 (state_writer::write_state_options)
17658 (write_state_lang_bitmap): ditto, renaming to...
17659 (state_writer::write_state_lang_bitmap)
17660 (write_state_version): ditto, renaming to...
17661 (state_writer::write_state_version)
17662 (write_state_scalar_type): ditto, renaming to...
17663 (state_writer::write_state_scalar_type)
17664 (write_state_string_type): ditto, renaming to...
17665 (state_writer::write_state_string_type)
17666 (write_state_undefined_type): ditto, renaming to...
17667 (state_writer::write_state_undefined_type)
17668 (write_state_struct_union_type): ditto, renaming to...
17669 (state_writer::write_state_struct_union_type)
17670 (write_state_struct_type): ditto, renaming to...
17671 (state_writer::write_state_struct_type)
17672 (write_state_user_struct_type): ditto, renaming to...
17673 (state_writer::write_state_user_struct_type)
17674 (write_state_lang_struct_type): ditto, renaming to...
17675 (state_writer::write_state_lang_struct_type)
17676 (write_state_param_struct_type): ditto, renaming to...
17677 (state_writer::write_state_param_struct_type)
17678 (write_state_pointer_type): ditto, renaming to...
17679 (state_writer::write_state_pointer_type)
17680 (write_state_array_type): ditto, renaming to...
17681 (state_writer::write_state_array_type)
17682 (write_state_gc_used): ditto, renaming to...
17683 (state_writer::write_state_gc_used)
17684 (write_state_common_type_content): ditto, renaming to...
17685 (state_writer::write_state_common_type_content)
17686 (write_state_type): ditto, renaming to...
17687 (state_writer::write_state_type)
17688 (write_state_pair_list): ditto, renaming to...
17689 (state_writer::write_state_pair_list)
17690 (write_state_pair): ditto, renaming to...
17691 (state_writer::write_state_pair)
17692 (write_state_typedefs): ditto, renaming to...
17693 (state_writer::write_state_typedefs)
17694 (write_state_structures): ditto, renaming to...
17695 (state_writer::write_state_structures)
17696 (write_state_param_structs): ditto, renaming to...
17697 (state_writer::write_state_param_structs)
17698 (write_state_variables): ditto, renaming to...
17699 (state_writer::write_state_variables)
17700 (write_state_srcdir): ditto, renaming to...
17701 (state_writer::write_state_srcdir)
17702 (write_state_files_list): ditto, renaming to...
17703 (state_writer::write_state_files_list)
17704 (write_state_languages): ditto, renaming to...
17705 (state_writer::write_state_languages)
17706 (write_state): create a state_writer instance and use it when
17707 writing out the state file
17709 2013-05-17 Mike Stump <mikestump@comcast.net>
17711 PR rtl-optimization/57304
17712 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
17713 accessing DF_REF_REAL_LOC.
17715 2013-05-17 Jakub Jelinek <jakub@redhat.com>
17717 PR rtl-optimization/57281
17718 PR rtl-optimization/57300
17719 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
17720 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
17721 what the other splitter did if the registers are dead.
17723 2013-05-17 Richard Biener <rguenther@suse.de>
17725 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
17728 2013-05-17 Jakub Jelinek <jakub@redhat.com>
17730 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
17733 2013-05-17 Marek Polacek <polacek@redhat.com>
17735 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
17736 length when doing non-zero store of storing '\0' to '\0'.
17738 2013-05-17 Jakub Jelinek <jakub@redhat.com>
17740 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
17741 vect_external_def oprnd1 with loop_vinfo, try to emit
17742 optional cast, negation and and stmts on the loop preheader
17743 edge instead of into the pattern def seq.
17745 PR tree-optimization/57051
17746 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
17747 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
17749 2013-05-16 Nick Clifton <nickc@redhat.com>
17751 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
17752 (rl78_is_naked_func): New function.
17753 (rl78_expand_prologue): Skip prologue generation for naked functions.
17754 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
17755 * doc/extend.texi (naked): Add RL78 to the list of processors
17756 that supports this attribute.
17758 2013-05-16 Jeff Law <law@redhat.com>
17760 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
17762 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
17764 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
17765 cache parameters using detect_caches_amd also for CYRIX,
17766 NSC and TM2 signatures.
17768 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
17769 Dzianis Kahanovich <mahatma@eu.by>
17773 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
17774 VIA/Centaur processors and determine their cache parameters
17775 using detect_caches_amd.
17777 2013-05-16 Teresa Johnson <tejohnson@google.com>
17779 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
17780 (rtl_verify_edges): New function.
17781 (rtl_verify_bb_insns): Ditto.
17782 (rtl_verify_bb_pointers): Ditto.
17783 (rtl_verify_bb_insn_chain): Ditto.
17784 (rtl_verify_fallthru): Ditto.
17785 (rtl_verify_bb_layout): Ditto.
17786 (rtl_verify_flow_info_1): Outline checks into new functions.
17787 (rtl_verify_flow_info): Ditto.
17789 2013-05-16 Steve Ellcey <sellcey@imgtec.com>
17791 * cfghooks.c (copy_bbs): Add update_dominance argument.
17792 * cfghooks.h (copy_bbs): Update prototype.
17793 * tree-cfg.c (gimple_duplicate_sese_region):
17794 Add update_dominance argument.
17795 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
17796 * tree-ssa-loop-ch.c (copy_loop_headers): Update
17797 gimple_duplicate_sese_region call.
17798 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
17799 Update copy_bbs call.
17800 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
17801 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
17803 2013-05-16 Jakub Jelinek <jakub@redhat.com>
17805 * tree-vectorizer.h (NUM_PATTERNS): Increment.
17806 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
17807 vect_recog_rotate_pattern.
17808 (vect_recog_rotate_pattern): New function.
17810 2013-05-16 Jason Merrill <jason@redhat.com>
17812 * Makefile.in (LLINKER): New variable.
17813 (mostlyclean): Remove link mutex.
17814 * configure.ac: Handle --enable-link-mutex.
17815 * lock-and-run.sh: New script.
17817 2013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
17820 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
17823 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
17825 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
17827 2013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
17829 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
17830 * config/arm/arm.c (next_consecutive_mem): New function.
17831 (gen_movmem_ldrd_strd): Likewise.
17832 * config/arm/arm.md (movmemqi): Update condition and code.
17833 (unaligned_loaddi, unaligned_storedi): New patterns.
17835 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
17837 * config.gcc: Obsolete *-*-solaris2.9*.
17838 * doc/install.texi (Specific, *-*-solaris2*): Document it.
17840 2013-05-16 Richard Biener <rguenther@suse.de>
17842 * passes.c (init_optimization_passes): Move pass_parallelize_loops
17843 earlier, after GRAPHITE transforms and IV canonicalization.
17845 2013-05-16 Jakub Jelinek <jakub@redhat.com>
17847 * omp-low.c (extract_omp_for_data): For collapsed loops,
17848 if at least one of the loops is known at compile time to
17849 iterate zero times, set count to 0.
17850 (expand_omp_regimplify_p): New function.
17851 (expand_omp_for_generic): For collapsed loops, if at least
17852 one of the loops isn't known to iterate at least once,
17853 add runtime check with setting count to 0.
17854 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
17855 For unsigned types if it isn't known at compile time that
17856 the loop will iterate at least once, add runtime check to bypass
17857 the whole loop if initial condition isn't true.
17859 2013-05-16 Nathan Sidwell <nathan@codesourcery.com>
17861 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
17863 2013-05-16 Marc Glisse <marc.glisse@inria.fr>
17865 PR middle-end/57286
17866 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
17867 transformations to avoid an infinite loop.
17869 2013-05-16 Marek Polacek <polacek@redhat.com>
17871 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
17873 2013-05-15 Leif Ekblad <leif@rdos.net>
17875 * config/i386/i386.c (ix86_decompose_address): Use
17876 DEFAULT_TLS_SEG_REG to access TLS segment register.
17877 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
17878 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
17879 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
17881 2013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
17884 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
17885 sibling calls to functions that would normally be lazily bound,
17886 unless $gp is call-clobbered.
17888 2013-05-15 Uros Bizjak <ubizjak@gmail.com>
17890 * config/i386/i386.c (ix86_option_override_internal): Update
17891 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
17892 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
17893 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
17894 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
17896 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
17898 2013-05-15 Andreas Schwab <schwab@suse.de>
17900 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
17901 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
17904 2013-05-15 Teresa Johnson <tejohnson@google.com>
17906 * loop-unroll.c (report_unroll_peel): Check decision before
17907 emitting unroll/peel message.
17909 2013-05-15 Teresa Johnson <tejohnson@google.com>
17911 * function.h (has_bb_partition): New rtl_data flag.
17912 (bb_reorder_complete): Ditto.
17913 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
17914 instead of flag_reorder_blocks_and_partition.
17915 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
17916 with some enhancements.
17917 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
17918 * bb-reorder.c (connect_traces): Check for has_bb_partition
17919 instead of flag_reorder_blocks_and_partition.
17920 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
17921 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
17922 verify_hot_cold_block_grouping.
17923 (partition_hot_cold_basic_blocks): Set has_bb_partition.
17925 2013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
17928 * config/arm/predicates.md (call_insn_operand): New predicate.
17929 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
17930 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
17931 if insn is not a tail call.
17932 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
17934 * config/arm/arm.h (enum reg_class): New caller save register class.
17935 (REG_CLASS_NAMES): Likewise.
17936 (REG_CLASS_CONTENTS): Likewise.
17937 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
17940 2013-05-15 Richard Biener <rguenther@suse.de>
17942 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
17943 of MSG_OPTIMIZED_LOCATIONS.
17944 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
17945 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
17947 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
17948 of MSG_OPTIMIZED_LOCATIONS.
17949 (execute_vect_slp): Likewise.
17950 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
17951 (vect_create_cond_for_alias_checks): Likewise.
17952 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
17953 (vect_recog_widen_mult_pattern): Likewise.
17954 (vect_recog_widen_sum_pattern): Likewise.
17955 (vect_recog_over_widening_pattern): Likewise.
17956 (vect_recog_widen_shift_pattern): Likewise.
17957 (vect_recog_vector_vector_shift_pattern): Likewise.
17958 (vect_recog_divmod_pattern): Likewise.
17959 (vect_recog_mixed_size_cond_pattern): Likewise.
17960 (vect_recog_bool_pattern): Likewise.
17961 (vect_pattern_recog_1): Likewise.
17963 2013-05-15 Martin Jambor <mjambor@suse.cz>
17965 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
17966 non-functions to builtin_unreachable.
17967 * ipa-inline-transform.c (inline_call): Do not assert estimates were
17968 correct when new direct edges were discovered.
17970 2013-05-15 Martin Jambor <mjambor@suse.cz>
17972 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
17973 header, print symbol order instead of node uid, print more information
17974 about indirect edge targets.
17975 (ipa_make_edge_direct_to_target): Print symbol order instead of node
17977 (ipa_make_edge_direct_to_target): Likewise.
17978 (remove_described_reference): Likewise.
17979 (propagate_controlled_uses): Likewise.
17980 (ipa_print_node_params): Also print symbol order.
17981 (ipcp_transform_function): Print symbol order instead of node uids.
17982 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
17983 (cgraph_get_create_real_symbol_node): Likewise.
17984 * ipa-cp.c (print_lattice): Likewise.
17985 (print_all_lattices): Likewise.
17986 (determine_versionability): Likewise.
17987 (initialize_node_lattices): Likewise.
17988 (estimate_local_effects): Likewise.
17989 (update_profiling_info): Likewise.
17990 (create_specialized_node): Likewise.
17991 (perhaps_add_new_callers): Likewise.
17992 (decide_about_value): Likewise.
17993 (decide_whether_version_node): Likewise.
17994 (identify_dead_nodes): Likewise.
17995 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
17996 (dump_inline_summary): Likewise.
17997 (estimate_node_size_and_time): Likewise.
17998 (inline_analyze_function): Likewise.
17999 * ipa-inline.c (report_inline_failed_reason): Likewise.
18000 (want_early_inline_function_p): Likewise.
18001 (edge_badness): Likewise.
18002 (update_edge_key): Likewise.
18003 (inline_small_functions): Likewise. Add dumping of order to two other
18005 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
18006 instead of node uids.
18007 (propagate_pure_const): Likewise.
18008 (propagate_pure_const): Likewise.
18009 * ipa-utils.c (dump_cgraph_node_set): Likewise.
18010 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
18011 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
18013 * tree-pretty-print.c (dump_function_header): Likewise.
18014 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
18015 Print symbol order instead of node uids.
18017 2013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
18019 * config/s390/s390.c (s390_register_move_cost): Don't impose the
18020 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
18022 2013-05-15 Richard Biener <rguenther@suse.de>
18024 PR tree-optimization/57275
18025 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
18026 return value for fail to do runtime alias checks for gather loads.
18028 2013-05-15 Jan Hubicka <jh@suse.cz>
18032 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
18033 weakrefs are not external.
18034 (lto_symtab_merge_decls): Fix thinko when dealing with
18035 non-lto_symtab decls.
18036 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
18037 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
18038 * varpool.c (dump_varpool_node): Dump more flags.
18040 2013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
18042 * config/i386/i386.c (processor_alias_table): Add instruction
18043 FSGSBASE for AMD bdver3 architecture.
18045 2013-05-14 Jakub Jelinek <jakub@redhat.com>
18047 * tree.c (warn_deprecated_use): Print file:line using locus color.
18048 * diagnostic.c (diagnostic_report_current_module): Print file:line
18049 and file:line:column using locus color.
18051 2013-05-14 Mike Stump <mikestump@comcast.net>
18053 * gdbinit.in: Add __null.
18055 2013-05-14 Mike Stump <mikestump@comcast.net>
18057 * recog.h: Rename struct recog_data to Recog_data.
18058 * recog.c: Likewise.
18059 * reload.c (can_reload_into): Likewise.
18060 * config/picochip/picochip.c: Likewise.
18062 2013-05-14 Mike Stump <mikestump@comcast.net>
18064 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
18066 2013-05-14 Steven Bosscher <steven@gcc.gnu.org>
18068 * resource.h (struct resources): Remove unch_memory member.
18069 (CLEAR_RESOURCE): Don't clear unch_memory.
18070 * resource.c (mark_referenced_resources): Don't set it.
18071 (mark_set_resources): Likewise.
18072 (mark_target_live_regs): Don't clear it.
18073 (init_resource_info): Likewise.
18074 * reorg.c (resource_conflicts_p): Don't compare it.
18075 (redundant_insn): Don't set it.
18077 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
18079 * emit-rtl.c (next_label): Remove unused function.
18080 (skip_consecutive_labels, link_cc0_insns): Move to ...
18081 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
18082 only place where these functions are used, and make them static.
18084 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
18086 * fold-const.c (fold_negate_expr): Handle vectors.
18087 (fold_truth_not_expr): Make it static.
18088 (fold_invert_truthvalue): New static function.
18089 (invert_truthvalue_loc): Handle vectors. Do not call
18090 fold_truth_not_expr directly.
18091 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
18092 <TRUTH_NOT_EXPR>: Do not cast to boolean.
18093 (fold_comparison): Handle vector constants.
18094 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
18095 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
18096 * tree.h (fold_truth_not_expr): Remove declaration.
18098 2013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
18100 * config/aarch64/aarch64-simd.md
18101 (aarch64_vcond_internal<mode>): Rename to...
18102 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
18103 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
18104 float modes. Clarify all iterator modes.
18105 (vcond<mode><mode>): Use new name for vcond expanders.
18106 (vcond<v_cmp_result><mode>): Likewise.
18107 (vcondu<mode><mode>: Likewise.
18108 * config/aarch64/iterators.md (VDQF_COND): New.
18110 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
18113 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
18114 variable for the shift amount. Check that we shift by non-negative
18117 2013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
18120 * config/arm/arm.h (EPILOGUE_USES): Only return true
18121 for LR_REGNUM after epilogue_completed.
18123 2013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
18125 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
18126 is error_mark_node.
18128 2013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
18131 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
18132 and Solaris 11+/x86 with gld.
18133 * configure: Regenerate.
18135 2013-05-14 Jakub Jelinek <jakub@redhat.com>
18137 * expmed.c (expand_shift_1): Canonicalize rotates by
18138 constant bitsize / 2 to bitsize - 1.
18139 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
18140 case ROTATERT>: Likewise.
18143 2013-05-10 Jakub Jelinek <jakub@redhat.com>
18145 * config/i386/i386.md (rotateinv): New code attr.
18146 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
18147 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
18148 roll $31, %eax, etc.
18150 2013-05-14 Richard Biener <rguenther@suse.de>
18152 PR middle-end/57235
18153 * tree-eh.c (sink_clobbers): Give up for successors with
18154 multiple predecessors and no virtual uses.
18156 2013-05-14 Eric Botcazou <ebotcazou@adacore.com>
18158 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
18159 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
18161 2013-05-14 Jakub Jelinek <jakub@redhat.com>
18163 PR middle-end/57251
18164 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
18165 the case when both op0 and op1 have VOIDmode.
18167 2013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
18169 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
18170 in multiply-accumulate mode.
18172 2013-05-13 Guozhi Wei <carrot@google.com>
18174 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
18176 2013-05-13 Kai Tietz <ktietz@redhat.com>
18179 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
18180 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
18181 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
18182 * config/i386/i386.c (ix86_option_override_internal): Likewise.
18183 (ix86_expand_prologue): Likewise.
18184 (ix86_expand_split_stack_prologue): Likewise.
18185 (legitimate_pic_address_disp_p): Likewise.
18186 (legitimize_pic_address): Likewise.
18187 (legitimize_tls_address): Likewise.
18188 (legitimize_pe_coff_symbol): Likewise.
18189 (output_pic_addr_const): Likewise.
18190 (construct_plt_address): Likewise.
18191 (ix86_expand_call): Likewise.
18192 (x86_output_mi_thunk): Likewise.
18193 (x86_function_profiler): Likewise.
18195 2013-05-13 Sofiane Naci <sofiane.naci@arm.com>
18197 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
18198 similar switch cases.
18199 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
18200 (aarch64_simd_mov_to_<mode>low): Delete.
18201 (aarch64_simd_mov_to_<mode>high): Delete.
18202 (move_lo_quad_<mode>): Add w<-r alternative.
18203 (aarch64_simd_move_hi_quad_<mode>): Likewise.
18204 (aarch64_simd_mov_from_*): Update type attribute.
18205 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
18208 2013-05-13 Jan Hubicka <jh@suse.cz>
18210 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
18211 * config/i386/i386.c (ix86_compute_frame_layout,
18212 ix86_expand_epilogue, emit_i387_cw_initialization,
18213 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
18214 ix86_local_alignment): Fix use of size/speed predicates.
18216 2013-05-13 Jakub Jelinek <jakub@redhat.com>
18218 PR tree-optimization/45216
18219 PR tree-optimization/57157
18220 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
18221 the (-Y) & (B - 1) variant if OP is |.
18222 * expmed.c (expand_shift_1): For rotations by const0_rtx just
18223 return shifted. Use (-op1) & (prec - 1) as other_amount
18224 instead of prec - op1.
18226 2013-05-13 Martin Jambor <mjambor@suse.cz>
18228 PR middle-end/42371
18229 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
18230 (ipa_constant_data): New type.
18231 (ipa_jump_func): Use ipa_constant_data to hold information about
18232 constant jump functions.
18233 (ipa_get_jf_constant): Adjust to jump function type changes.
18234 (ipa_get_jf_constant_rdesc): New function.
18235 (ipa_param_descriptor): New field controlled_uses.
18236 (ipa_get_controlled_uses): New function.
18237 (ipa_set_controlled_uses): Likewise.
18238 * ipa-ref.h (ipa_find_reference): Declare.
18239 * ipa-prop.c (ipa_cst_ref_desc): New type.
18240 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
18242 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
18243 New parameter cs. Adjust all callers.
18244 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
18245 (remove_described_reference): New function.
18246 (jfunc_rdesc_usable): Likewise.
18247 (try_make_edge_direct_simple_call): Decrement controlled use count,
18248 attempt to remove reference if it hits zero.
18249 (combine_controlled_uses_counters): New function.
18250 (propagate_controlled_uses): Likewise.
18251 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
18252 (ipa_edge_duplication_hook): Duplicate reference descriptions.
18253 (ipa_print_node_params): Print described use counter.
18254 (ipa_write_jump_function): Adjust to jump function type changes.
18255 (ipa_read_jump_function): New parameter CS, pass it to
18256 ipa_set_jf_constant. Adjust caller.
18257 (ipa_write_node_info): Stream controlled use count
18258 (ipa_read_node_info): Likewise.
18259 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
18261 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
18262 count. Remove cloning-added reference if it reaches zero.
18263 * ipa-ref.c (ipa_find_reference): New function.
18265 2013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
18267 * config/i386/i386.c (processor_target_table): Modified default
18268 alignment values for AMD BD and BT architectures.
18270 2013-05-13 Marc Glisse <marc.glisse@inria.fr>
18272 * tree-vect-generic.c (uniform_vector_p): Move ...
18273 * tree.c (uniform_vector_p): ... here.
18274 * tree.h (uniform_vector_p): Declare it.
18275 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
18278 2013-05-13 Jakub Jelinek <jakub@redhat.com>
18280 PR tree-optimization/57230
18281 * tree-ssa-strlen.c (handle_char_store): Record length for
18282 array store from STRING_CST.
18284 PR tree-optimization/57230
18285 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
18288 2013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
18290 * config/epiphany/epiphany.c (epiphany_init): Check size of
18291 NUM_MODES_FOR_MODE_SWITCHING.
18292 (epiphany_expand_prologue):
18293 Remove CONFIG_REGNUM initial value handling code.
18294 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
18295 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
18296 (emit_set_fp_mode, epiphany_mode_after): Likewise.
18297 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
18298 Don't return 1 for FP_MODE_NONE.
18299 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
18300 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
18301 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
18302 * config/epiphany/epiphany.md (save_config): New pattern.
18304 2013-05-12 Uros Bizjak <ubizjak@gmail.com>
18306 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
18308 2013-05-10 Uros Bizjak <ubizjak@gmail.com>
18310 * config/i386/i386.md (memory): Handle sseishft1.
18311 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
18312 (*vec_extractv2di_1): Ditto.
18314 2013-05-10 Vladimir Makarov <vmakarov@redhat.com>
18316 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
18319 2013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
18321 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
18322 Add mthumb/march=armv7-a multilib.
18323 Add mthumb/march=armv7-r multilib.
18324 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
18326 2013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
18328 * config/v850/t-rtems: Add more multilibs.
18330 2013-05-10 Richard Biener <rguenther@suse.de>
18332 PR tree-optimization/57214
18333 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
18334 not propagate from SSA names that occur in abnormal PHI nodes.
18336 2013-05-10 Marc Glisse <marc.glisse@inria.fr>
18338 * stor-layout.c (element_precision): New function.
18339 * machmode.h (element_precision): Declare it.
18340 * tree.c (build_minus_one_cst): New function.
18341 (element_precision): Likewise.
18342 * tree.h (build_minus_one_cst): Declare new function.
18343 (element_precision): Likewise.
18344 * fold-const.c (operand_equal_p): Use element_precision.
18345 (fold_binary_loc): Handle vector types.
18346 * convert.c (convert_to_integer): Use element_precision.
18347 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
18350 2013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
18352 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
18353 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
18354 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
18355 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
18356 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
18357 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
18358 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
18359 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
18360 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
18361 (Uuw8): New constraints.
18362 (Usb4): Move into alphabetical order.
18363 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
18364 (sd8_operand, ub8_operand, uw8_operand): New predicates.
18365 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
18366 previously unnamed patterns.
18367 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
18368 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
18369 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
18370 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
18371 of set_attr_alternative/if_then_else. Use extended_mips16 instead
18372 of specific lengths.
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 PR tree-optimization/45216
18382 PR tree-optimization/57157
18383 * tree-ssa-forwprop.c (simplify_rotate): New function.
18384 (ssa_forward_propagate_and_combine): Call it.
18386 2013-05-10 Richard Biener <rguenther@suse.de>
18388 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
18389 disable peeling when we version for aliasing.
18390 (vector_alignment_reachable_p): Honor explicit user alignment.
18391 (vect_supportable_dr_alignment): Likewise.
18392 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
18393 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
18394 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
18395 then peeling to arrange for the cost-model check to come first.
18397 2013-05-10 Alan Modra <amodra@gmail.com>
18399 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
18400 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
18401 * configure: Regenerate.
18403 2013-05-10 Alan Modra <amodra@gmail.com>
18406 * varasm.c (default_elf_select_section): Move !DECL_P check..
18407 (get_named_section): ..to here before calling get_section_name.
18409 (default_section_type_flags): Add DECL_P check.
18410 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
18411 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
18413 2013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
18415 * config/epiphany/epiphany.c (epiphany_expand_prologue):
18416 When using gen_stack_adjust_str with a register offset, add a
18417 REG_FRAME_RELATED_EXPR note.
18419 2013-05-09 Uros Bizjak <ubizjak@gmail.com>
18421 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
18422 (*vec_extractv4si_zext_mem): Ditto.
18423 (*vec_extractv2di): Add 0->x and x->x alternatives.
18424 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
18425 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
18427 2013-05-09 Jason Merrill <jason@redhat.com>
18429 N3639 C++1y VLA support
18430 * gimplify.c (gimplify_vla_decl): Don't touch an existing
18433 * tree.c (build_constructor_va): New.
18434 * tree.h: Declare it.
18436 2013-05-09 Martin Jambor <mjambor@suse.cz>
18439 * gimple-fold.c (canonicalize_constructor_val): Call
18440 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
18442 2013-05-09 Jan Hubicka <jh@suse.cz>
18443 Richard Biener <rguenther@suse.de>
18446 * symtab.c (symtab_make_decl_local): Do not add private names.
18448 2013-05-09 Jan Hubicka <jh@suse.cz>
18451 * symtab.c (insert_to_assembler_name_hash): Handle clones.
18452 (unlink_from_assembler_name_hash): Likewise.
18453 (symtab_prevail_in_asm_name_hash, symtab_register_node,
18454 symtab_unregister_node, symtab_initialize_asm_name_hash,
18455 change_decl_assembler_name): Update.
18457 2013-05-09 Sofiane Naci <sofiane.naci@arm.com>
18459 * config/aarch64/aarch64.md: New movtf split.
18460 (*movtf_aarch64): Update.
18461 (aarch64_movdi_tilow): Handle TF modes and rename to
18462 aarch64_movdi_<mode>low.
18463 (aarch64_movdi_tihigh): Handle TF modes and rename to
18464 aarch64_movdi_<mode>high
18465 (aarch64_movtihigh_di): Handle TF modes and rename to
18466 aarch64_mov<mode>high_di
18467 (aarch64_movtilow_di): Handle TF modes and rename to
18468 aarch64_mov<mode>low_di
18469 (aarch64_movtilow_tilow): Remove spurious whitespace.
18470 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
18472 (aarch64_print_operand): Update.
18474 2013-05-09 Alan Modra <amodra@gmail.com>
18476 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
18478 * configure: Regenerate.
18480 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
18482 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
18483 splitter preparation statements.
18484 * config/i386/sse.md (*vec_extract* splitters): Ditto.
18485 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
18488 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18490 * gimple-ssa-strength-reduction.c (count_candidates): Change
18491 return value to int.
18492 (analyze_candidates_and_replace): Change type of length to int.
18494 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
18496 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
18497 (*vec_extract<mode>): Use VI12_128 mode iterator.
18498 (*vec_extract<mode>_mem): Ditto.
18499 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
18502 2013-05-08 Diego Novillo <dnovillo@google.com>
18507 2012-08-17 Diego Novillo <dnovillo@google.com>
18510 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
18511 * config.in: Regenerate.
18512 * configure: Regenerate.
18513 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
18515 2013-05-08 Jan Hubicka <jh@suse.cz>
18518 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
18519 * cgraph.h (symtab_node_base): Add unique_name.
18520 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
18521 input_overwrite_node, input_varpool_node): Stream unique_name.
18522 * cgraphclones.c (cgraph_create_virtual_clone,
18523 cgraph_function_versioning): Set unique_name.
18524 * ipa.c (function_and_variable_visibility): Set unique_name.
18526 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18528 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
18529 (alloc_cand_and_find_basis): Restrict conditional candidate
18530 processing to CAND_MULTs.
18532 2013-05-08 Jan Hubicka <jh@suse.cz>
18535 lto-symtab.c (lto_symtab_symbol_p): New function.
18536 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
18537 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
18538 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
18539 Skip static symbols.
18541 2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
18543 PR tree-optimization/57200
18544 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
18545 Only call inform if the preceding warning_at returns true.
18547 2013-05-07 Han Shen <shenhan@google.com>
18549 * cfgexpand.c (record_or_union_type_has_array_p): New function.
18550 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
18551 * common.opt (fstack-protector-strong): New option.
18552 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
18553 * doc/invoke.texi (Optimization Options): Document
18554 "-fstack-protector-strong".
18555 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
18557 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
18559 * config/mips/mips.c (mips_machine_reorg2): Return 0.
18561 2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
18563 * ira.c (update_equiv_regs): Add insn having equiv memory even if
18564 it is not lhs of the insn.
18565 (setup_reg_equiv): Remove insn having equiv memory which it is not
18567 * lra-constraints.c (process_address): Try to improve generation
18568 code for address base + disp.
18569 (lra_constraints): Make correct the code for checking insn setting
18570 up backward equivalence. Remove insn only if it is in the init
18572 * lra-eliminations.c (update_reg_eliminate): Change return value.
18573 (lra_eliminate): Use the result.
18575 2013-05-07 Uros Bizjak <ubizjak@gmail.com>
18577 * config/i386/sse.md (ssescalarnummask): New mode attribute.
18578 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
18579 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
18580 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
18581 register target operands.
18582 (*vec_extractv8hi_sse2): New pattern.
18583 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
18584 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
18585 (*vec_extract<mode>_mem): New insn and split pattern.
18587 2013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
18589 * config/arm/arm.c (arm_asan_shadow_offset): New function.
18590 (TARGET_ASAN_SHADOW_OFFSET): Define.
18591 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
18592 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
18594 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18596 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
18597 (incr_vec_index): Return -1 if increment not found.
18598 (create_add_on_incoming_edge): Assert if increment not found.
18599 (record_increment): Limit number of increments recorded.
18600 (all_phi_incrs_profitable): Return false if an increment not found.
18601 (replace_profitable_candidates): Don't process increments that were
18603 (analyze_candidates_and_replace): Limit size of incr_vec.
18605 2013-05-07 Richard Biener <rguenther@suse.de>
18607 * calls.c (special_function_p): setjmp-like functions are leaf.
18608 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
18609 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
18611 2013-05-07 Sofiane Naci <sofiane.naci@arm.com>
18613 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
18614 (aarch64_simd_mov<mode>): New expander.
18615 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
18616 (aarch64_simd_mov_to_<mode>high): Likewise.
18617 (aarch64_simd_mov_from_<mode>low): Likewise.
18618 (aarch64_simd_mov_from_<mode>high): Likewise.
18619 (aarch64_dup_lane<mode>): Update.
18620 (aarch64_dup_lanedi): New instruction pattern.
18621 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
18622 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
18624 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18626 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
18627 (replace_mult_candidate): Remove unnecessary argument; remove
18628 unnecessary parameter from call to introduce_cast_before_cand.
18629 (replace_unconditional_candidate): Remove unnecessary parameter
18630 from call to replace_mult_candidate.
18631 (replace_conditional_candidate): Likewise.
18632 (insert_initializers): Use make_temp_ssa_name.
18633 (introduce_cast_before_cand): Remove unnecessary argument; use
18634 make_temp_ssa_name.
18635 (replace_one_candidate): Remove unnecessary argument; remove
18636 unnecessary parameter from calls to introduce_cast_before_cand.
18637 (replace_profitable_candidates): Remove unnecessary parameters
18638 from calls to replace_one_candidate.
18640 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18642 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
18643 phi def as possibly hiding a basis for a CAND_ADD whose operands
18644 have been commuted in the analysis.
18645 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
18647 2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
18649 * config/aarch64/aarch64.md
18650 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
18651 shift value between 0-4.
18653 2013-05-07 Richard Biener <rguenther@suse.de>
18655 * double-int.h (rshift): New overload.
18656 * double-int.c (rshift): New function.
18657 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
18658 (create_reference_ops_from_ref): Remove.
18659 (vn_reference_insert): Use shared ops for constructing the
18660 reference and copy it.
18662 2013-05-07 Richard Biener <rguenther@suse.de>
18664 PR middle-end/57190
18665 * tree-eh.c (sink_clobbers): Properly propagate
18666 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
18668 2013-05-07 Jakub Jelinek <jakub@redhat.com>
18670 PR tree-optimization/57149
18671 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
18672 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
18673 collect_phi_def_edges, execute_late_warn_uninitialized): Use
18674 uninit_undefined_value_p instead of ssa_undefined_value_p.
18677 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
18678 for modifier == EXPAND_INITIALIZER.
18680 2013-05-07 Anton Blanchard <anton@samba.org>
18682 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
18683 for powerpc64 little endian.
18684 * configure: Regenerate.
18686 2013-05-06 Graham Stott <grahams@btinternet.com>
18688 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
18689 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
18690 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
18691 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
18693 2013-05-06 Graham Stott <grahams@btinternet.com>
18695 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
18696 codes which allow non-lvalues.
18698 2013-05-06 Marc Glisse <marc.glisse@inria.fr>
18700 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
18701 components are all 1s.
18702 (integer_minus_onep): New function.
18703 * tree.h (integer_minus_onep): Declare it.
18704 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
18705 integer_minus_onep instead of integer_all_onesp.
18707 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
18710 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
18711 variations of these patterns.
18713 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
18715 * config/i386/i386.md (isa): Add x64_sse4 member.
18716 (enabled): Handle x64_sse4.
18717 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
18718 instruction for 64bit SSE4_1 targets. Update insn attributes.
18719 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
18720 instruction for SSE4_1 targets. Update insn attributes.
18721 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
18722 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
18723 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
18725 (*vec_extractv4si): Rename from *sse4_1_pextrd.
18726 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
18727 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
18729 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
18732 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
18734 2013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
18736 * final.c (do_assembler_dialects): Don't handle curly braces and
18737 vertical bar escaped by % as dialect delimiters.
18738 (output_asm_insn): Print curly braces and vertical bar if escaped
18739 by % and ASSEMBLER_DIALECT defined.
18740 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
18741 * doc/tm.texi: Regenerated.
18743 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
18745 * config/mips/mips.c: Include tree-pass.h.
18746 (mips_reorg): Split in pre- and post-dbr_schedule parts.
18747 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
18748 (pass_mips_machine_reorg2): New machine specific pass.
18749 (insert_pass_mips_machine_reorg2): New pass plugin definition.
18750 (mips_option_override): Register the new pass.
18751 * rtl.h (cleanup_barriers): Remove prototype.
18752 (dbr_schedule): Likewise.
18753 * jump.c (cleanup_barriers): Make static.
18754 * reorg.c (dbr_schedule): Likewise.
18756 2013-05-06 Richard Biener <rguenther@suse.de>
18758 PR tree-optimization/57185
18759 * tree-parloops.c (add_field_for_reduction): Handle anonymous
18760 SSA names properly.
18762 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
18765 * config/i386/i386.c (add_parameter_dependencies): Add dependence
18766 between "first_arg" and "insn", not "last" and "insn".
18768 2013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18770 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
18771 (find_candidates_in_block): Re-enable slsr_process_phi.
18772 (create_phi_basis): Fix double counting of candidate adjustment.
18774 2013-05-06 Richard Biener <rguenther@suse.de>
18776 PR middle-end/57147
18777 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
18778 the edge is also fallthru, preserve it and just clear the
18780 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
18781 also complex, preserve that and just clear the fallthru flag.
18782 * tree-inline.c (update_ssa_across_abnormal_edges): Also
18783 update virtual operands.
18785 2013-05-06 Alan Modra <amodra@gmail.com>
18787 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
18788 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
18789 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
18790 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
18791 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
18792 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
18794 2013-05-06 Alan Modra <amodra@gmail.com>
18796 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
18797 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
18798 (DEFAULT_ASM_ENDIAN): Define.
18799 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
18800 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
18801 Update -K PIC clause from sysv4.h.
18802 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
18803 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
18805 2013-05-06 Alan Modra <amodra@gmail.com>
18807 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
18808 twice for little-endian.
18809 (ashrdi3_no_power, ashrdi3): Support little-endian.
18811 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
18814 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
18815 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
18817 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
18818 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
18819 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
18822 2013-05-05 Steven Bosscher <steven@gcc.gnu.org>
18823 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
18825 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
18826 * config/pa/pa.opt: Make mbig-switch a no-op.
18827 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
18828 (CASE_VECTOR_MODE): Always return SImode.
18829 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
18830 for the !TARGET_BIG_SWITCH case.
18831 * config/pa/pa-linux.h: Likewise.
18832 * config/pa/pa-openbsd.h: Likewise.
18833 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
18834 * config/pa/pa.md (short_jump): Remove define_insn.
18835 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
18836 (casesi0): Remove define_insn.
18837 (type): Remove btable_branch.
18838 (pa_combine_type): Likewise.
18839 (in_nullified_branch_delay): Likewise.
18840 (in_call_delay): Likewise.
18841 (define_delay): Likewise.
18842 (define_insn_reservation "Z3"): Likewise.
18843 (define_insn_reservation "Z4"): Likewise.
18844 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
18845 (pa_adjust_insn_length): Remove adjustment for btable branches.
18846 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
18849 2013-05-05 Uros Bizjak <ubizjak@gmail.com>
18851 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
18852 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
18853 Add m->r,x alternatives.
18854 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
18855 splitters using SWI48x mode iterator.
18856 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
18857 TARGET_64BIT. Add m->x alternative.
18858 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
18859 Add o->x alternative. Enable for TARGET_SSE.
18860 (sse_storeq): Remove expander.
18861 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
18862 with memory input operand.
18863 (*vec_extractv2di_1 splitter): New.
18864 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
18865 * config/i386/i386.md (ssevecmodelower): New mode attribute.
18867 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
18869 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
18870 (INT_LOWPART): Delete.
18871 (extract_MB): Adjust.
18872 (extract_ME): Adjust.
18873 (print_operand): Adjust.
18875 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
18877 * config/rs6000/predicates.md (reg_or_add_cint_operand,
18878 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
18879 (reg_or_logical_cint_operand, easy_fp_constant,
18880 logical_const_operand): Delete "CONST_DOUBLE" case.
18881 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
18882 "HOST_BITS_PER_WIDE_INT == 64" test.
18883 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
18884 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
18885 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
18886 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
18888 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
18889 CONST_DOUBLE DImode/VOIDmode case.
18890 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
18891 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
18892 CONST_DOUBLE VOIDmode case.
18893 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
18894 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
18895 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
18896 Delete CONST_DOUBLE case.
18897 (splitters for mov FMOVE64 const_double): Delete
18898 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
18899 "HOST_BITS_PER_WIDE_INT >= 64" test.
18900 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
18902 (mov DI const_double): Delete.
18904 2013-05-04 Jakub Jelinek <jakub@redhat.com>
18906 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
18907 on op shows all bits zero in mode of a lowpart subreg, return zero.
18909 2013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
18912 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
18913 to save TFmode registers and DImode to save TImode registers for
18914 caller save operations.
18915 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
18916 mark being partially clobbered since they only use the first
18919 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
18920 and TDmode only use the upper 64-bits of each VSX register.
18922 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18924 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
18925 (find_candidates_in_block): Disable slsr_process_phi.
18927 2013-05-03 Guozhi Wei <carrot@google.com>
18929 * coverage.c (coverage_obj_init): Move the construction of gcov
18931 (build_init_ctor): ... here.
18933 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18935 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
18936 (slsr_cand_d): Redefine def_phi.
18937 (stride_status, phi_adjust_status, count_phis_status): New enums.
18938 (find_phi_def): New.
18939 (find_basis_for_base_expr): New.
18940 (find_basis_for_candidate): Handle hidden bases.
18941 (alloc_cand_and_find_basis): Handle phi candidates.
18942 (slsr_process_phi): New.
18943 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
18944 (create_mul_imm_cand): Likewise.
18945 (create_add_ssa_cand): Exclude phi base candidates.
18946 (create_add_imm_cand): Likewise.
18947 (slsr_process_cast): Likewise.
18948 (slsr_process_copy): Likewise.
18949 (find_candidates_in_block): Handle phi candidates.
18950 (dump_candidate): Likewise.
18951 (unconditional_cands): Delete.
18952 (unconditional_cands_with_known_stride_p): Delete.
18953 (phi_dependent_cand_p): New.
18954 (cand_increment): Handle phi-dependent candidates.
18955 (replace_dependent): Delete.
18956 (replace_mult_candidate): New.
18957 (replace_unconditional_candidate): New.
18958 (incr_vec_index): Move to avoid forward reference.
18959 (create_add_on_incoming_edge): New.
18960 (create_phi_basis): New.
18961 (replace_dependents): Delete.
18962 (replace_conditional_candidate): New.
18963 (phi_add_costs): New.
18964 (replace_uncond_cands_and_profitable_phis): New.
18965 (record_increment): Handle phi adjustments.
18966 (record_phi_increments): New.
18967 (record_increments): Handle phi adjustments.
18968 (phi_incr_cost): New.
18969 (lowest_cost_path): Handle phis.
18970 (total_savings): Likewise.
18971 (analyze_increments): Likewise.
18972 (ncd_with_phi): New.
18973 (ncd_of_cand_and_phis): New.
18974 (nearest_common_dominator_for_cands): Handle phi increments.
18975 (all_phi_incrs_profitable): New.
18976 (replace_profitable_candidates): Handle phi-dependent candidates.
18977 (analyze_candidates_and_replace): Likewise.
18979 2013-05-03 Teresa Johnson <tejohnson@google.com>
18982 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
18983 do not exceed REG_BR_PROB_BASE.
18985 2013-05-03 Jeff Law <law@redhat.com>
18987 PR tree-optimization/57144
18988 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
18989 operand of the condition will bit into the new type when eliminating
18990 a cast feeding a condition.
18992 2013-05-03 Jakub Jelinek <jakub@redhat.com>
18994 PR rtl-optimization/57130
18995 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
18996 of COMPARE as in_code to the recursive call if needed.
18998 2013-05-03 Uros Bizjak <ubizjak@gmail.com>
19000 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
19001 (enabled): Handle new members.
19002 * config/i386/sse.md (*vec_concatv2si): Merge from
19003 *vec_concatv2si_sse2 and vec_concatv2si_sse.
19004 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
19006 2013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
19008 PR tree-optimization/57027
19009 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
19010 for fnms opportunity, check we got the prerequisite kind
19011 of tree / gimple before using accessor functions.
19013 2013-05-03 Richard Biener <rguenther@suse.de>
19015 * double-int.h (lshift): New overload without precision
19016 and arith argument.
19017 (operator *=, operator +=, operator -=): Move ...
19018 * double-int.c (operator *=, operator +=, operator -=): ... here
19019 and implement more efficiently.
19020 (mul_double_with_sign): Remove.
19021 (lshift_double): Adjust to take unsinged shift argument, push
19022 dispatching code to callers.
19023 (mul_double_wide_with_sign): Add early out for callers that
19024 are not interested in high parts or overflow.
19025 (lshift): New function.
19026 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
19027 dispatch code here.
19028 (lrotate, rrotate): Use logical shifts.
19029 * expr.c (get_inner_reference): Use lshift.
19030 * fixed-value.c (do_fixed_divide): Likewise.
19031 * tree-dfa.c (get_ref_base_and_extent): Likewise.
19032 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
19033 (indirect_refs_may_alias_p): Likewise.
19034 (stmt_kills_ref_p_1): Likewise.
19036 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
19038 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
19040 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
19042 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
19043 scalar form of FABD instruction.
19045 2013-05-02 Vladimir Makarov <vmakarov@redhat.com>
19047 * lra-constraints.c (process_alt_operands): Add checking alt
19048 number to choose the best alternative.
19050 2013-05-02 Richard Biener <rguenther@suse.de>
19052 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
19053 bitmap and its handling.
19054 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
19056 2013-05-02 Richard Biener <rguenther@suse.de>
19058 PR middle-end/57140
19059 * tree-inline.c (copy_loops): Properly handle removed loops.
19060 (copy_cfg_body): Mark destination loops for fixup if source
19061 loops needed fixup.
19063 2013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
19066 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
19067 generating simple_return for naked functions.
19069 2013-05-02 Martin Jambor <mjambor@suse.cz>
19071 PR middle-end/56988
19072 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
19073 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
19075 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
19076 ipa_agg_replacement_value structures.
19077 (known_aggs_to_agg_replacement_list): Likewise.
19078 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
19079 (read_agg_replacement_chain): Likewise.
19080 (ipcp_transform_function): Also check that by_ref flags match.
19082 2013-05-02 Richard Biener <rguenther@suse.de>
19084 * graphds.h (struct graph): Add obstack member.
19085 * graphds.c (new_graph): Initialize obstack and allocate
19087 (add_edge): Allocate edge from the obstack.
19088 (free_graph): Free the obstack instead of all edges and vertices.
19090 2013-05-02 Teresa Johnson <tejohnson@google.com>
19092 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
19094 * cfg.c (update_bb_profile_for_threading): Ditto.
19095 * tree-inline.c (copy_bb): Ditto.
19096 (copy_edges_for_bb): Ditto.
19097 (initialize_cfun): Ditto.
19098 (copy_cfg_body): Ditto.
19099 (expand_call_inline): Ditto.
19100 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
19101 (estimate_node_size_and_time): Ditto.
19102 (inline_merge_summary): Ditto.
19103 * cgraphclones.c (cgraph_clone_edge): Ditto.
19104 (cgraph_clone_node): Ditto.
19105 * sched-rgn.c (compute_dom_prob_ps): Ditto.
19106 (compute_trg_info): Ditto.
19108 2013-05-02 Ian Bolton <ian.bolton@arm.com>
19110 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
19111 S reg when fp attribute set.
19112 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
19114 2013-05-02 Ian Bolton <ian.bolton@arm.com>
19116 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
19118 (*and_one_cmplsi3_compare0_uxtw): Likewise.
19119 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
19120 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
19122 2013-05-02 Richard Biener <rguenther@suse.de>
19124 * tree-scalar-evolution.c (scev_info_hasher): Remove.
19125 (struct instantiate_cache_entry): New type.
19126 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
19127 (struct instantiate_cache_type): New type.
19128 (set_instantiated_value, get_instantiated_value): Remove.
19129 (get_instantiated_value_entry): New function.
19130 (instantiate_scev_name): Use the new cache and adjust.
19131 (instantiate_scev_poly): Adjust.
19132 (instantiate_scev_binary): Likewise.
19133 (instantiate_array_ref): Likewise.
19134 (instantiate_scev_convert): Likewise.
19135 (instantiate_scev_not): Likewise.
19136 (instantiate_scev_3): Likewise.
19137 (instantiate_scev_2): Likewise.
19138 (instantiate_scev_r): Likewise.
19139 (instantiate_scev): Likewise.
19140 (resolve_mixers): Likewise.
19142 2013-05-01 Vladimir Makarov <vmakarov@redhat.com>
19145 * lra-constraints.c (best_small_class_operands_num): Remove.
19146 (process_alt_operands): Remove small_class_operands_num. Take
19147 small classes operands into losers and only if the operand is not
19148 matched. Modify debugging output.
19149 (curr_insn_transform): Remove best_small_class_operands_num.
19152 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
19154 * config/aarch64/aarch64-builtins.c
19155 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
19156 * config/aarch64/aarch64-simd-builtins.def
19157 (reduc_splus_): Add new modes.
19158 (reduc_uplus_): New.
19159 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
19160 (reduc_uplus_v4sf): Likewise.
19161 (reduc_splus_v4sf): Likewise.
19162 (aarch64_addv<mode>): Likewise.
19163 (reduc_uplus_<mode>): Likewise.
19164 (reduc_splus_<mode>): Likewise.
19165 (aarch64_addvv2di): Likewise.
19166 (reduc_uplus_v2di): Likewise.
19167 (reduc_splus_v2di): Likewise.
19168 (aarch64_addvv2si): Likewise.
19169 (reduc_uplus_v2si): Likewise.
19170 (reduc_splus_v2si): Likewise.
19171 (reduc_<sur>plus_<mode>): New.
19172 (reduc_<sur>plus_v2di): Likewise.
19173 (reduc_<sur>plus_v2si): Likewise.
19174 (reduc_<sur>plus_v4sf): Likewise.
19175 (aarch64_addpv4sf): Likewise.
19176 * config/aarch64/arm_neon.h
19177 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
19178 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
19179 add UNSPEC_SADDV, UNSPEC_UADDV.
19181 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
19183 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
19185 * config/aarch64/arm_neon.h
19186 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
19188 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
19190 * config/aarch64/aarch64-builtins
19191 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
19193 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
19195 * config/aarch64/aarch64-simd-builtins.def
19196 (reduc_smax_): New.
19197 (reduc_smin_): Likewise.
19198 (reduc_umax_): Likewise.
19199 (reduc_umin_): Likewise.
19200 (reduc_smax_nan_): Likewise.
19201 (reduc_smin_nan_): Likewise.
19204 (smax): Update for V2SF, V4SF and V2DF modes.
19207 (smin_nan): Likewise.
19208 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
19209 (<su><maxmin><mode>3): ...This, refactor.
19210 (s<maxmin><mode>3): New.
19211 (<maxmin_uns><mode>3): Likewise.
19212 (reduc_<maxmin_uns>_<mode>): Refactor.
19213 (reduc_<maxmin_uns>_v4sf): Likewise.
19214 (reduc_<maxmin_uns>_v2si): Likewise.
19215 (aarch64_<fmaxmin><mode>: Remove.
19216 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
19218 (vmin<q>_f<32,64>): Likewise.
19219 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
19221 (su): Add mappings for smax, smin, umax, umin.
19223 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
19224 (FMAXMIN): Rename as...
19225 (FMAXMIN_UNS): ...This.
19227 (fmaxminv): Likewise.
19228 (fmaxmin): Likewise.
19230 (maxmin_uns_op): Likewise.
19232 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
19234 * config/aarch64/arm_neon.h
19235 (vac<ge, gt><sd>_f<32, 64>): Rename to...
19236 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
19237 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
19239 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
19241 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
19242 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
19244 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
19246 * config/aarch64/aarch64-simd.md
19247 (vcond<mode>_internal): Handle special cases for constant masks.
19248 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
19249 (vcondu<mode><mode>): Likewise.
19250 (vcond<v_cmp_result><mode>): New.
19252 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
19254 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
19255 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
19256 * config/aarch64/aarch64-simd-builtins.def
19257 (cmeq): Update to BUILTIN_VALLDI.
19262 * config/aarch64/arm_neon.h
19263 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
19264 to builtins or C as appropriate.
19266 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
19268 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
19270 (cmhi): Rename to...
19272 * config/aarch64/aarch64-simd.md
19273 (simd_mode): Add SF.
19274 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
19275 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
19276 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
19277 (cstore<mode>_neg): ...This.
19278 * config/aarch64/iterators.md
19280 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
19281 (COMPARISONS): New.
19282 (UCOMPARISONS): Likewise.
19283 (optab): Add missing comparisons.
19289 (VCMP_S): Likewise.
19290 (VCMP_U): Likewise.
19291 (V_cmp_result): Add DF, SF modes.
19292 (v_cmp_result): Likewise.
19294 (vmtype): Likewise.
19295 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
19297 2013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
19299 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
19300 define_insn to define_insn_and_split.
19301 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
19302 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
19303 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
19304 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
19305 (thumb2_negscc): Likewise.
19307 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
19309 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
19311 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
19313 * config/arm/thumb2.md: Remove trailing whitespaces.
19315 2013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
19317 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
19318 Use gen_int_mode rather than GEN_INT.
19320 2013-04-30 H.J. Lu <hongjiu.lu@intel.com>
19322 * value-prof.c (stream_in_histogram_value): Remove the strayed
19325 2013-04-30 Richard Biener <rguenther@suse.de>
19327 PR middle-end/57122
19328 * cfghooks.c (split_edge): Properly check for the loop latch edge.
19330 2013-04-30 Richard Biener <rguenther@suse.de>
19332 PR middle-end/57107
19333 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
19335 2013-04-30 Andrey Belevantsev <abel@ispras.ru>
19337 PR rtl-optimization/56957
19338 PR rtl-optimization/57105
19339 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
19340 variable. Use just INSN_UID for determining whether an insn
19341 should be only disconnected from the insn stream.
19342 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
19344 2013-04-30 Jakub Jelinek <jakub@redhat.com>
19346 PR tree-optimization/57104
19347 * tsan.c (instrument_expr): Don't instrument accesses to
19348 DECL_HARD_REGISTER VAR_DECLs.
19350 2013-04-30 Richard Biener <rguenther@suse.de>
19352 * function.h (loops_for_fn): New inline function.
19353 (set_loops_for_fn): Likewise.
19354 * cfgloop.h (place_new_loop): Add struct function parameter.
19355 (get_loop): Likewise.
19356 (get_loops): Likewise.
19357 (number_of_loops): Likewise.
19358 (fel_next): Adjust.
19359 (fel_init): Likewise.
19360 * cfg.c (get_loop_copy): Adjust.
19361 * cfgloop.c (flow_loops_dump): Likewise.
19362 (record_loop_exits): Likewise.
19363 (verify_loop_structure): Likewise.
19364 * cfgloopanal.c (mark_irreducible_loops): Likewise.
19365 (estimate_reg_pressure_cost): Likewise.
19366 (mark_loop_exit_edges): Likewise.
19367 * cfgloopmanip.c (place_new_loop): Likewise.
19368 (add_loop): Likewise.
19369 (duplicate_loop): Likewise.
19370 * graph.c (draw_cfg_nodes): Likewise.
19371 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
19372 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
19373 (extract_affine_chrec): Likewise.
19374 (build_scop_iteration_domain): Likewise.
19375 * graphite.c (graphite_initialize): Likewise.
19376 * ira-build.c (create_loop_tree_nodes): Likewise.
19377 (more_one_region_p): Likewise.
19378 (rebuild_regno_allocno_maps): Likewise.
19379 (mark_loops_for_removal): Likewise.
19380 (mark_all_loops_for_removal): Likewise.
19381 (remove_unnecessary_regions): Likewise.
19382 (ira_build): Likewise.
19383 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
19384 * loop-init.c (fix_loop_structure): Likewise.
19385 (gate_rtl_move_loop_invariants): Likewise.
19386 (gate_rtl_unswitch): Likewise.
19387 (gate_rtl_unroll_and_peel_loops): Likewise.
19388 (rtl_doloop): Likewise.
19389 * lto-streamer-in.c (input_cfg): Likewise.
19390 * lto-streamer-out.c (output_cfg): Likewise.
19391 * modulo-sched.c (sms_schedule): Likewise.
19392 * predict.c (tree_estimate_probability): Likewise.
19393 (tree_estimate_probability_driver): Likewise.
19394 (estimate_loops): Likewise.
19395 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
19396 (move_sese_region_to_fn): Likewise.
19397 (debug_loop_num): Likewise.
19398 * tree-chrec.c (chrec_evaluate): Likewise.
19399 (hide_evolution_in_other_loops_than_loop): Likewise.
19400 (chrec_component_in_loop_num): Likewise.
19401 (reset_evolution_in_loop): Likewise.
19402 (evolution_function_is_invariant_rec_p): Likewise.
19403 * tree-if-conv.c (main_tree_if_conversion): Likewise.
19404 * tree-inline.c (copy_loops): Likewise.
19405 (copy_cfg_body): Likewise.
19406 (tree_function_versioning): Likewise.
19407 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
19408 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
19410 (add_to_evolution_1): Likewise.
19411 (scev_const_prop): Likewise.
19412 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
19413 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
19414 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
19415 (tree_ssa_lim_initialize): Likewise.
19416 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
19417 (verify_loop_closed_ssa): Likewise.
19418 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
19419 (tree_ssa_loop_im): Likewise.
19420 (tree_ssa_loop_unswitch): Likewise.
19421 (tree_vectorize): Likewise.
19422 (check_data_deps): Likewise.
19423 (tree_ssa_loop_ivcanon): Likewise.
19424 (tree_ssa_loop_bounds): Likewise.
19425 (tree_complete_unroll): Likewise.
19426 (tree_complete_unroll_inner): Likewise.
19427 (tree_parallelize_loops): Likewise.
19428 (tree_ssa_loop_prefetch): Likewise.
19429 (tree_ssa_loop_ivopts): Likewise.
19430 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
19431 * tree-vectorizer.c (vectorize_loops): Likewise.
19433 2013-04-29 Mike Frysinger <vapier@gentoo.org>
19435 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
19436 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
19437 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
19438 with EABI_LINK_SPEC.
19440 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
19443 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
19446 2013-04-29 Vladimir Makarov <vmakarov@redhat.com>
19449 * lra-constraints.c (process_alt_operands): Discourage a bit more
19450 using memory for pseudos. Print cost dump for alternatives.
19451 Modify cost values for conflicts with early clobbers.
19452 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
19454 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
19457 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
19459 2013-04-29 Ian Bolton <ian.bolton@arm.com>
19461 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
19462 from/to S register.
19463 (movdi_aarch64): Support LDR/STR from/to D register.
19465 2013-04-29 Ian Bolton <ian.bolton@arm.com>
19467 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
19468 or higher by default.
19470 2013-04-29 Richard Biener <rguenther@suse.de>
19472 PR middle-end/57075
19473 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
19474 even if not adding abnormal edges for calls that can make
19477 2013-04-29 Richard Biener <rguenther@suse.de>
19479 PR middle-end/57103
19480 * tree-cfg.c (move_stmt_op): Fix condition under which to update
19482 (move_stmt_r): Remove redundant checking.
19484 2013-04-29 Teresa Johnson <tejohnson@google.com>
19487 * basic-block.h (apply_scale): New function.
19488 (apply_probability): Use apply_scale.
19489 * gimple-streamer-in.c (input_bb): Ditto.
19490 * lto-streamer-in.c (input_cfg): Ditto.
19491 * lto-cgraph.c (merge_profile_summaries): Ditto.
19492 * tree-optimize.c (execute_fixup_cfg): Ditto.
19493 * tree-inline.c (copy_bb): Update comment to use apply_scale.
19494 (copy_edges_for_bb): Ditto.
19495 (copy_cfg_body): Ditto.
19497 2013-04-29 Tom de Vries <tom@codesourcery.com>
19499 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
19500 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
19501 (tail_merge_optimize): Handle current_loops == NULL.
19503 2013-04-26 Jeff Law <law@redhat.com>
19505 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
19506 (simplify_cond_using_ranges): Generalize code to simplify
19507 COND_EXPRs where one argument is a constant and the other
19508 is an SSA_NAME created by an integral type conversion.
19510 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19512 * config/arm/arm.md (store_minmaxsi): Use only when
19513 optimize_insn_for_size_p.
19515 2013-04-29 Christian Bruel <christian.bruel@st.com>
19518 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
19520 2013-04-29 Richard Biener <rguenther@suse.de>
19522 PR middle-end/57089
19523 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
19524 loop tree make sure to schedule a fixup for the child as well.
19525 (expand_omp_for_generic): Properly add loops.
19526 (expand_omp_for_static_nochunk): Likewise.
19527 (expand_omp_for_static_chunk): Likewise.
19528 (expand_omp_for): For the degenerate case fixup loops.
19529 (expand_omp_sections): Fix default bb placement in loops.
19530 (expand_omp_atomic_pipeline): Properly add loops.
19532 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19534 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
19536 2013-04-29 Tom de Vries <tom@codesourcery.com>
19538 * tree-ssa-tail-merge.c: Update header comment.
19540 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
19542 * config/aarch64/arm_neon.h
19543 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
19544 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
19545 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
19546 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
19547 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
19548 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
19549 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
19550 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
19552 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
19554 * config/aarch64/aarch64-simd.md
19555 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
19556 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
19557 fix_trunc, fixuns_trunc.
19558 (ftrunc<VDQF:mode>2): New.
19559 * config/aarch64/iterators.md (optab): Add fix, fixuns.
19560 (fix_trunc_optab): New.
19562 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
19564 * config/aarch64/aarch64-builtins.c
19565 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
19566 iceilf, lround, iroundf.
19568 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
19571 * config/i386/i386.h (enum ix86_tune_indices)
19572 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
19573 New, split from X86_TUNE_INTER_UNIT_MOVES.
19574 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
19575 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
19576 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
19577 (TARGET_INTER_UNIT_MOVES): Remove.
19578 * config/i386/i386.c (initial_ix86_tune_features): Update.
19579 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
19580 (ix86_expand_convert_uns_didf_sse): Use
19581 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
19582 (ix86_expand_vector_init_one_nonzero): Ditto.
19583 (ix86_expand_vector_init_interleave): Ditto.
19584 (inline_secondary_memory_needed): Return true for moves from SSE class
19585 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
19586 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
19587 * config/i386/constraints.md (Yi, Ym): Depend on
19588 TARGET_INTER_UNIT_MOVES_TO_VEC.
19589 (Yj, Yn): New constraints.
19590 * config/i386/i386.md (*movdi_internal): Change constraints of
19591 operand 1 from Yi to Yj and from Ym to Yn.
19592 (*movsi_internal): Ditto.
19593 (*movdf_internal): Ditto.
19594 (*movsf_internal): Ditto.
19595 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
19596 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
19597 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
19598 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
19599 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
19600 * config/i386/sse.md (movdi_to_sse): Ditto.
19601 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
19602 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
19603 TARGET_INTER_UNIT_MOVES.
19604 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
19605 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
19606 instead of TARGET_INTER_UNIT_MOVES.
19607 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
19608 operand 1 from Yi to Yj and from Ym to Yn.
19610 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
19612 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
19613 (float_truncate_hi_): Likewise.
19614 (float_extend_lo_): Likewise.
19615 (float_truncate_lo_): Likewise.
19616 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
19617 (aarch64_float_extend_lo_v2df): Likewise.
19618 (vec_unpacks_hi_v4sf): Likewise.
19619 (aarch64_float_truncate_lo_v2sf): Likewise.
19620 (aarch64_float_truncate_hi_v4sf): Likewise.
19621 (vec_pack_trunc_v2df): Likewise.
19622 (vec_pack_trunc_df): Likewise.
19624 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
19626 * config/aarch64/aarch64-builtins.c
19627 (aarch64_fold_builtin): Fold float conversions.
19628 * config/aarch64/aarch64-simd-builtins.def
19629 (floatv2si, floatv4si, floatv2di): New.
19630 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
19631 * config/aarch64/aarch64-simd.md
19632 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
19633 * config/aarch64/iterators.md (FLOATUORS): New.
19634 (optab): Add float, floatuns.
19635 (su_optab): Likewise.
19637 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
19639 * config/aarch64/aarch64-builtins.c
19640 (aarch64_builtin_vectorized_function): Use new names for
19642 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
19643 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
19644 (fcvtzu): Split as...
19645 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
19646 (fcvtas): Split as...
19647 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
19648 (fcvtau): Split as...
19649 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
19650 (fcvtps): Split as...
19651 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
19652 (fcvtpu): Split as...
19653 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
19654 (fcvtms): Split as...
19655 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
19656 (fcvtmu): Split as...
19657 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
19658 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
19659 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
19660 (lfrintnusf, lfrintnudf): Likewise.
19661 * config/aarch64/aarch64-simd.md
19662 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
19664 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
19665 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
19666 (fcvt_pattern): Likewise.
19668 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
19670 * config/aarch64/aarch64-simd.md
19671 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
19672 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
19674 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
19676 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
19677 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
19678 (vrnd<a,m,n,p>_f32): Implement using builtins.
19679 (vrnd<i,x><q>_f<32, 64>): New.
19681 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
19683 * config/aarch64/aarch64-builtins.c
19684 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
19685 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
19686 (frintz): Rename to...
19688 (frintp): Rename to...
19690 (frintm): Rename to...
19692 (frinti): Rename to...
19693 (nearbyint): ...this.
19694 (frintx): Rename to...
19696 (frinta): Rename to...
19698 * config/aarch64/aarch64-simd.md
19699 (aarch64_frint<frint_suffix><mode>): Delete.
19700 (<frint_pattern><mode>2): Convert to insn.
19701 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
19702 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
19703 (frint_pattern): Likewise.
19704 (frint_suffix): Likewise.
19706 2013-04-29 Richard Biener <rguenther@suse.de>
19708 PR tree-optimization/57081
19709 * loop-init.c: Include tree-flow.h.
19710 (loop_optimizer_finalize): Free number of iteration estimates.
19711 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
19713 2013-04-29 Jakub Jelinek <jakub@redhat.com>
19715 PR tree-optimization/57083
19716 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
19717 non-singleton shift count range, zero extend low_bound for uns case.
19719 * config/i386/predicates.md (general_vector_operand): New predicate.
19720 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
19721 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
19722 if they aren't nonimmediate operands. If their original values
19723 satisfy const_vector_equal_evenodd_p, don't shift them.
19724 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
19725 predicates. For the SSE4.1 case force operands[{1,2}] into registers
19726 if not nonimmediate_operand.
19727 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
19728 instead of register_operand.
19729 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
19731 2013-04-28 Eric Botcazou <ebotcazou@adacore.com>
19733 * stor-layout.c (finalize_size_functions): Allocate a structure and
19734 reset cfun before dumping the functions.
19736 2013-04-27 Jakub Jelinek <jakub@redhat.com>
19738 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
19741 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
19742 use xop_pmacsdqh if uns_p.
19743 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
19744 the immediate rotate count.
19746 2013-04-26 Vladimir Makarov <vmakarov@redhat.com>
19748 * rtl.h (struct rtx_def): Add comment for field jump.
19749 (LRA_SUBREG_P): New macro.
19750 * recog.c (register_operand): Check LRA_SUBREG_P.
19751 * lra.c (lra): Add note at the end of RTL code. Align non-empty
19753 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
19754 (lra_final_code_change): Skip subreg change for operators.
19755 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
19756 if there are no operand changes.
19757 * lra-constraints.c (curr_insn_set): New.
19758 (match_reload): Set LRA_SUBREG_P.
19759 (emit_spill_move): Ditto.
19760 (check_and_process_move): Use curr_insn_set. Process only single
19761 set insns. Don't initialize sec_mem_p and change_p.
19762 (simplify_operand_subreg): Use LRA_SUBREG_P.
19763 (reg_in_class_p): New function.
19764 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
19765 of #ifdef. Add code to remove cycling.
19766 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
19767 non-null disp. Reload inner instead of disp when base and index
19768 are null. Try to put lo_sum into register.
19769 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
19770 (check_and_process_move): Move code for move cost check to
19771 simple_move_p. Remove equiv_substitution.
19772 (simple_move_p): New function.
19773 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
19774 curr_insn_set. Call check_and_process_move only for single set
19775 insns. Use the new function. Move call of check_and_process_move
19776 after operand equiv substitution and address process.
19778 2013-04-26 Jakub Jelinek <jakub@redhat.com>
19781 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
19782 with nonlocal goto receivers or returns twice calls, ignore
19783 unininitialized values from abnormal edges to nl goto receiver
19784 or returns twice call.
19786 2013-04-26 Jakub Jelinek <jakub@redhat.com>
19788 PR tree-optimization/57051
19789 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
19790 and VEC_RSHIFT_EXPR if shift count is a multiple of element
19793 2013-04-26 Richard Biener <rguenther@suse.de>
19795 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
19796 (expand_omp_taskreg): Likewise. Mark loops for fixup.
19797 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
19798 (fixup_loop_arrays_after_move): New function.
19799 (move_sese_region_to_fn): Properly outline the loop tree parts
19800 of the SESE region.
19802 2013-04-26 Uros Bizjak <ubizjak@gmail.com>
19804 * config/i386/i386.md (type, unit): Fix long lines.
19806 2013-04-26 Richard Biener <rguenther@suse.de>
19808 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
19809 (lto-streamer-out.o): Likewise.
19810 * cfgloop.c (init_loops_structure): Export, add struct function
19811 argument and adjust.
19812 (flow_loops_find): Adjust.
19813 * cfgloop.h (enum loop_estimation): Add EST_LAST.
19814 (init_loops_structure): Declare.
19815 * lto-streamer-in.c: Include cfgloop.h.
19816 (input_cfg): Input the loop tree.
19817 * lto-streamer-out.c: Include cfgloop.h.
19818 (output_cfg): Output the loop tree.
19819 (output_struct_function_base): Do not drop PROP_loops.
19821 2013-03-26 Richard Biener <rguenther@suse.de>
19823 * tree-cfg.c (execute_build_cfg): Build the loop tree.
19824 (pass_build_cfg): Provide PROP_loops.
19825 (move_sese_region_to_fn): Remove loops that are outlined into fn
19827 * tree-inline.c: Include cfgloop.h.
19828 (initialize_cfun): Do not drop PROP_loops.
19829 (copy_loops): New function.
19830 (copy_cfg_body): Copy loop structure.
19831 (tree_function_versioning): Initialize destination loop tree.
19832 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
19833 (pass_parallelize_loops): Do IL verification.
19834 * loop-init.c (loop_optimizer_init): Fixup loops if required.
19835 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
19836 the CFG make sure we fixup loops as well.
19837 * tree-ssa-tail-merge.c: Include cfgloop.h.
19838 (replace_block_by): When merging loop latches mark loops for fixup.
19839 * lto-streamer-out.c (output_struct_function_base): Drop
19840 PROP_loops for now.
19841 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
19842 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
19843 * ipa-split.c: Include cfgloop.h.
19844 (split_function): Add the new return block to the loop tree root.
19845 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
19846 whether we have removed the forwarder block.
19847 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
19848 * cfgloop.h (place_new_loop): Declare.
19849 * cfgloopmanip.c (place_new_loop): Export.
19850 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
19851 (tree-switch-conversion.o): Likewise.
19852 (tree-complex.o): Likewise.
19853 (tree-inline.o): Likewise.
19854 (tree-ssa-tailmerge.o): Likewise.
19855 (ipa-split.o): Likewise.
19856 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
19857 (tree-ssa-copy.o): Likewise.
19858 * tree-switch-conversion.c: Include cfgloop.h
19859 (process_switch): If we emit a bit-test cascade, schedule loops
19861 * tree-complex.c: Include cfgloop.h.
19862 (expand_complex_div_wide): Properly add new basic-blocks to loops.
19863 * asan.c: Include cfgloop.h.
19864 (create_cond_insert_point): Properly add new basic-blocks to
19865 loops, schedule loop fixup.
19866 * cfgloop.c (verify_loop_structure): Check that looks are not
19868 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
19870 (expand_omp_for_generic): Likewise.
19871 (expand_omp_sections): Likewise.
19872 (expand_omp_atomic_pipeline): Schedule loops for fixup.
19873 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
19874 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
19875 is initialized, not when loops are present.
19876 * tree-parloops.c (parallelize_loops): Remove checking here.
19877 * passes.c (init_optimization_passes): Schedule a copy-propagation
19878 pass before complete unrolling of inner loops.
19880 2013-04-26 Jakub Jelinek <jakub@redhat.com>
19882 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
19883 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
19884 (colorize_init): Add argument to _WIN32 version.
19885 * toplev.c: Include diagnostic-color.h.
19886 (process_options): Default to -fdiagnostics-color=auto if
19887 GCC_COLORS env var is in the environment.
19888 * common.opt (fdiagnostics-color=): Add Var and Init.
19889 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
19890 env var is in the environment, the default is auto rather than never.
19892 * diagnostic.h (file_name_as_prefix): Add context argument.
19893 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
19894 the string as locus.
19895 * langhooks.c (lhd_print_error_function): Adjust caller.
19897 2013-04-25 Lawrence Crowl <crowl@google.com>
19899 * var-tracking.c (shared_hash_def::htab):
19900 Change type to hash_table. Update dependent calls and types.
19902 2013-04-25 Lawrence Crowl <crowl@google.com>
19904 * Makefile.in: Update as needed below.
19906 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
19907 Move declaration to after the type's method definitons.
19909 * attribs.c (htab_t scoped_attributes::attribute_hash):
19910 Change type to hash_table. Update dependent calls and types.
19912 * bitmap.c (htab_t bitmap_desc_hash):
19913 Change type to hash_table. Update dependent calls and types.
19915 * cselib.c (htab_t cselib_hash_table):
19916 Change type to hash_table. Update dependent calls and types.
19918 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
19919 (hash_string_slot_node): Move implementation into lto-streamer.h
19920 struct string_slot_hasher.
19921 (eq_string_slot_node): Likewise.
19923 * data-streamer-out.c: Update output_block::string_hash_table
19924 dependent calls and types.
19926 * dwarf2cfi.c (htab_t trace_index):
19927 Change type to hash_table. Update dependent calls and types.
19929 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
19930 Change type to hash_table. Update dependent calls and types.
19931 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
19932 (htab_t optimize_external_refs::map): Likewise.
19933 (htab_t output_comp_unit::extern_map): Likewise.
19934 (htab_t output_comdat_type_unit::extern_map): Likewise.
19935 (htab_t output_macinfo::macinfo_htab): Likewise.
19936 (htab_t optimize_location_lists::htab): Likewise.
19937 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
19939 * except.c (htab_t ehspec_hash_type):
19940 Change type to hash_table. Update dependent calls and types.
19941 (assign_filter_values::ttypes): Likewise.
19942 (assign_filter_values::ehspec): Likewise.
19943 (sjlj_assign_call_site_values::ar_hash): Likewise.
19944 (convert_to_eh_region_ranges::ar_hash): Likewise.
19946 * gcse.c (htab_t pre_ldst_table):
19947 Change type to hash_table. Update dependent calls and types.
19949 * ggc-common.c (htab_t saving_htab):
19950 Change type to hash_table. Update dependent calls and types.
19951 (htab_t loc_hash): Likewise.
19952 (htab_t ptr_hash): Likewise.
19953 (call_count): Rename ggc_call_count.
19954 (call_alloc): Rename ggc_call_alloc.
19955 (loc_descriptor): Rename make_loc_descriptor.
19956 (add_statistics): Rename ggc_add_statistics.
19958 * ggc-common.c (saving_htab):
19959 Change type to hash_table. Update dependent calls and types.
19961 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
19962 (push_gimplify_context): Likewise.
19963 (pop_gimplify_context): Likewise.
19964 (struct gimple_temp_hash_elt): Added.
19965 (struct gimplify_hasher): Likewise.
19966 (struct gimplify_ctx.temp_htab):
19967 Change type to hash_table. Update dependent calls and types.
19969 * gimple-fold.c: Include gimplify-ctx.h.
19971 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
19972 Change type to hash_table. Update dependent calls and types.
19973 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
19974 avoid potential global name collision.
19976 * gimplify.c: Include gimplify-ctx.h.
19977 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
19978 (htab_t gimplify_ctx::temp_htab):
19979 Update dependent calls and types for new type hash_table.
19980 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
19981 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
19983 * gimplify-ctx.h: New.
19984 (struct gimple_temp_hash_elt): Move from gimplify.c.
19985 (class gimplify_hasher): New.
19986 (struct gimplify_ctx): Move from gimple.h.
19987 (htab_t gimplify_ctx::temp_htab):
19988 Change type to hash_table. Update dependent calls and types.
19990 * graphite-clast-to-gimple.c: Include graphite-htab.h.
19991 (htab_t ivs_params::newivs_index):
19992 Change type to hash_table. Update dependent calls and types.
19993 (htab_t ivs_params::params_index): Likewise.
19994 (htab_t print_generated_program::params_index): Likewise.
19995 (htab_t gloog::newivs_index): Likewise.
19996 (htab_t gloog::params_index): Likewise.
19998 * graphite.c: Include graphite-htab.h.
19999 4htab_t graphite_transform_loops::bb_pbb_mapping):
20000 Change type to hash_table. Update dependent calls and types.
20002 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
20003 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
20004 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
20006 * graphite-dependences.c: Include graphite-htab.h.
20007 (loop_is_parallel_p): Change hash table type of parameter.
20009 * graphite-htab.h: New.
20010 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
20011 (extern find_pbb_via_hash): Move from graphite-poly.h.
20012 (extern loop_is_parallel_p): Move from graphite-poly.h.
20013 (extern get_loop_body_pbbs): Move from graphite-poly.h.
20015 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
20016 (extern loop_is_parallel_p): Move to graphite-htab.h.
20017 (extern get_loop_body_pbbs): Move to graphite-htab.h.
20019 * haifa-sched.c (htab_t delay_htab):
20020 Change type to hash_table. Update dependent calls and types.
20021 (htab_t delay_htab_i2): Likewise.
20023 * ira-color.c (htab_t allocno_hard_regs_htab):
20024 Change type to hash_table. Update dependent calls and types.
20026 * ira-costs.c (htab_t cost_classes_htab):
20027 Change type to hash_table. Update dependent calls and types.
20029 * loop-invariant.c (htab_t merge_identical_invariants::eq):
20030 Change type to hash_table. Update dependent calls and types.
20032 * loop-iv.c (htab_t bivs):
20033 Change type to hash_table. Update dependent calls and types.
20035 * loop-unroll.c (htab_t opt_info::insns_to_split):
20036 Change type to hash_table. Update dependent calls and types.
20037 (htab_t opt_info::insns_with_var_to_expand): Likewise.
20039 * lto-streamer.h (struct string_slot): Move from data-streamer.h
20040 (struct string_slot_hasher): New.
20041 (htab_t output_block::string_hash_table):
20042 Change type to hash_table. Update dependent calls and types.
20044 * lto-streamer-in.c (freeing_string_slot_hasher): New.
20045 (htab_t file_name_hash_table):
20046 Change type to hash_table. Update dependent calls and types.
20048 * lto-streamer-out.c: Update output_block::string_hash_table dependent
20051 * lto-streamer.c (htab_t tree_htab):
20052 Change type to hash_table. Update dependent calls and types.
20054 * omp-low.c: Include gimplify-ctx.h.
20056 * passes.c (htab_t name_to_pass_map):
20057 Change type to hash_table. Update dependent calls and types.
20058 (pass_traverse): Rename to passes_pass_traverse.
20060 * plugin.c (htab_t event_tab):
20061 Change type to hash_table. Update dependent calls and types.
20063 * postreload-gcse.c (htab_t expr_table):
20064 Change type to hash_table. Update dependent calls and types.
20065 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
20067 * sese.c (debug_rename_map_1): Make extern.
20068 (htab_t copy_bb_and_scalar_dependences::rename_map):
20069 Change type to hash_table. Update dependent calls and types.
20071 * sese.h (extern debug_rename_map): Move to .c file.
20073 * store-motion.c (htab_t store_motion_mems_table):
20074 Change type to hash_table. Update dependent calls and types.
20076 * trans-mem.c (htab_t tm_new_mem_hash):
20077 Change type to hash_table. Update dependent calls and types.
20079 * tree-browser.c (htab_t TB_up_ht):
20080 Change type to hash_table. Update dependent calls and types.
20082 * tree-cfg.c (htab_t discriminator_per_locus):
20083 Change type to hash_table. Update dependent calls and types.
20085 * tree-complex.c: Include tree-hasher.h
20086 (htab_t complex_variable_components):
20087 Change type to hash_table. Update dependent calls and types.
20089 * tree-eh.c (htab_t finally_tree):
20090 Change type to hash_table. Update dependent calls and types.
20092 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
20093 struct int_tree_hasher.
20094 (extern int_tree_map_eq): Likewise.
20095 (uid_decl_map_hash): Removed.
20096 (extern decl_tree_map_eq): Likewise.
20098 * tree-hasher.h: New.
20099 (struct int_tree_hasher): New.
20100 (typedef int_tree_htab_type): New.
20102 * tree-inline.c: Include gimplify-ctx.h.
20104 * tree-mudflap.c: Include gimplify-ctx.h.
20106 * tree-parloops.c: Include tree-hasher.h.
20107 (htab_t eliminate_local_variables_stmt::decl_address):
20108 Change type to hash_table. Update dependent calls and types.
20109 (htab_t separate_decls_in_region::decl_copies): Likewise.
20111 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
20112 Change type to hash_table. Update dependent calls and types.
20114 * tree-sra.c (candidates):
20115 Change type to hash_table. Update dependent calls and types.
20117 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
20119 (int_tree_map_hash): Likewise.
20121 * tree-ssa-dom.c (htab_t avail_exprs):
20122 Change type to hash_table. Update dependent calls and types.
20124 * tree-ssa-live.c (var_map_base_init::tree_to_index):
20125 Change type to hash_table. Update dependent calls and types.
20127 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
20128 Change type to hash_table. Update dependent calls and types.
20130 * tree-ssa-phiopt.c (seen_ssa_names):
20131 Change type to hash_table. Update dependent calls and types.
20133 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
20134 Change type to hash_table. Update dependent calls and types.
20136 * tree-ssa-uncprop.c (equiv):
20137 Change type to hash_table. Update dependent calls and types.
20139 2013-04-25 Jakub Jelinek <jakub@redhat.com>
20141 PR rtl-optimization/57003
20142 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
20143 call note_stores with kill_clobbered_value callback again after
20144 killing regs_invalidated_by_call.
20146 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
20148 * config/aarch64/aarch64-simd.md
20149 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
20150 (aarch64_simd_bsl<mode>): Likewise.
20151 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
20153 2013-04-25 Marek Polacek <polacek@redhat.com>
20155 PR tree-optimization/57066
20156 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
20158 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
20160 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
20162 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
20164 * config/aarch64/aarch64-builtins.c
20165 (aarch64_fold_builtin): New.
20166 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
20167 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
20168 * config/aarch64/aarch64-simd-builtins.def (abs): New.
20169 * config/aarch64/arm_neon.h
20170 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
20172 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
20173 Tejas Belagod <tejas.belagod@arm.com>
20175 * config/aarch64/aarch64-builtins.c
20176 (aarch64_gimple_fold_builtin): New.
20177 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
20178 * config/aarch64/aarch64-simd-builtins.def (addv): New.
20179 * config/aarch64/aarch64-simd.md (addpv4sf): New.
20180 (addvv4sf): Update.
20181 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
20183 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
20185 * config/aarch64/aarch64.md
20186 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
20188 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
20190 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
20191 (*ngcsi_uxtw): New pattern.
20193 2013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20194 Julian Brown <julian@codesourcery.com>
20196 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
20197 (TB_DREG): Add T_V4HF.
20198 (v4hf_UP): New macro.
20199 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
20200 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
20201 Handle initialisation of V4HF. Adjust initialisation of reinterpret
20203 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
20204 (arm_vector_mode_supported_p): Handle V4HF.
20205 (arm_mangle_map): Handle V4HFmode.
20206 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
20207 * config/arm/arm_neon_builtins.def: Add entries for
20208 vcvtv4hfv4sf, vcvtv4sfv4hf.
20209 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
20210 (neon_vcvtv4hfv4sf): Likewise.
20211 * config/arm/neon-gen.ml: Handle half-precision floating point
20213 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
20214 * config/arm/arm_neon.h: Regenerate.
20215 * config/arm/neon.ml (type elts): Add F16.
20216 (type vectype): Add T_float16x4, T_floatHF.
20217 (type vecmode): Add V4HF.
20218 (type features): Add Requires_FP_bit feature.
20219 (elt_width): Handle F16.
20220 (elt_class): Likewise.
20221 (elt_of_class_width): Likewise.
20222 (mode_of_elt): Refactor.
20223 (type_for_elt): Handle F16, fix error messages.
20224 (vectype_size): Handle T_float16x4.
20225 (vcvt_sh): New function.
20226 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
20227 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
20228 (string_of_mode): Handle V4HF.
20229 * doc/arm-neon-intrinsics.texi: Regenerate.
20231 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
20233 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
20234 format specifier in 'X' case.
20236 2013-04-25 Alan Modra <amodra@gmail.com>
20239 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
20240 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
20241 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
20242 Repeat for many other rotate/shift and mask patterns using subregs.
20243 Name lshiftrt insns.
20244 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
20245 on WORDS_BIG_ENDIAN.
20247 2013-04-25 Alan Modra <amodra@gmail.com>
20249 * config.gcc: Support little-endian powerpc-linux targets.
20250 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
20251 (LINK_OS_LINUX_SPEC): Define.
20252 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
20253 Preserve MASK_LITTLE_ENDIAN.
20254 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
20255 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
20256 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
20257 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
20258 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
20259 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
20260 Correct fp word order for little-endian. Don't shift toc entries
20261 smaller than a word for little-endian.
20262 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
20263 (bswapdi2 splits): Correct low-part subreg for little-endian.
20264 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
20265 low/high where such is correct only for be.
20266 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
20267 little-endian for -mcall-aixdesc.
20269 2013-04-25 Alan Modra <amodra@gmail.com>
20271 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
20272 replace_equiv_address_nv.
20274 2013-04-25 Alan Modra <amodra@gmail.com>
20276 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
20278 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
20281 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
20282 * rtl.h (struct rtx_def): ...
20284 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
20286 PR rtl-optimizations/57046
20287 * lra-constraints (split_reg): Set up lra_risky_transformations_p
20288 for multi-reg splits.
20290 2013-04-24 H.J. Lu <hongjiu.lu@intel.com>
20292 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
20294 2013-04-24 Sterling Augustine <saugustine@google.com>
20296 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
20297 (comp_dir_string, debug_str_dwo_section): New.
20298 (DEBUG_STR_DWO_SECTION): Rename to ...
20299 (DEBUG_DWO_STR_SECTION): ... this.
20300 (DEBUG_NORM_STR_SECTION): Delete.
20301 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
20302 (DEBUG_STR_DWO_SECTION_FLAGS): New.
20303 (find_AT_string): Move most logic to ...
20304 (find_AT_string_in_table): ... here. New.
20305 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
20306 add_skeleton_AT_string. Delete logic.
20307 (output_skeleton_debug_sections): Remove call to
20308 add_top_level_skeleton_die_attrs.
20309 (add_comp_dir_attribute): Move logic to comp_dir_string.
20310 (dwarf2out_init): Initialize debug_str_dwo_section.
20311 (output_indirect_string): Call find_string_form.
20312 (output_indirect_strings): Rewrite.
20313 (prune_unused_types): Empty skeleton_debug_str_hash.
20314 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
20315 (dwarf2out_finish): Call output_indirect_strings.
20317 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
20319 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
20321 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
20323 * rtl.h (struct rtx_def): Add comment for field jump.
20324 (LRA_SUBREG_P): New macro.
20325 * recog.c (register_operand): Check LRA_SUBREG_P.
20326 * lra.c (lra): Add note at the end of RTL code. Align non-empty
20328 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
20329 (lra_final_code_change): Skip subreg change for operators.
20330 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
20331 if there are no operand changes.
20332 * lra-constraints.c (curr_insn_set): New.
20333 (match_reload): Set LRA_SUBREG_P.
20334 (emit_spill_move): Ditto.
20335 (check_and_process_move): Use curr_insn_set. Process only single
20336 set insns. Don't initialize sec_mem_p and change_p.
20337 (simplify_operand_subreg): Use LRA_SUBREG_P.
20338 (reg_in_class_p): New function.
20339 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
20340 of #ifdef. Add code to remove cycling.
20341 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
20342 non-null disp. Reload inner instead of disp when base and index
20343 are null. Try to put lo_sum into register.
20344 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
20345 (check_and_process_move): Move code for move cost check to
20346 simple_move_p. Remove equiv_substitution.
20347 (simple_move_p): New function.
20348 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
20349 curr_insn_set. Call check_and_process_move only for single set
20350 insns. Use the new function. Move call of check_and_process_move
20351 after operand equiv substitution and address process.
20353 2013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
20355 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
20356 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
20357 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
20359 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
20361 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
20363 2013-04-24 Marek Polacek <polacek@redhat.com>
20365 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
20366 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
20367 (select_loops_exit_conditions): Likewise.
20368 (number_of_iterations_for_all_loops): Likewise.
20369 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
20370 (scev_analysis): Likewise.
20372 2013-04-02 Catherine Moore <clm@codesourcery.com>
20373 Chao-ying Fu <fu@mips.com>
20375 * config/mips/micromips.md (jraddiusp): New pattern.
20376 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
20377 instruction if possible.
20379 2013-04-24 Alan Modra <amodra@gmail.com>
20381 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
20383 2013-04-24 Julian Brown <julian@codesourcery.com>
20384 Chung-Lin Tang <cltang@codesourcery.com>
20386 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
20387 dependency behavior in enumeration type DIE generation. Add TODO note
20388 to comments about future DW_FORM_sdata/udata re-work of related code.
20390 2013-04-23 Lawrence Crowl <crowl@google.com>
20392 * Makefile.in: Update as needed below.
20394 * hash-table.h (class hash_table):
20395 Correct many methods with parameter types compare_type to the correct
20396 value_type. (Correct code was unlikely to notice the change.)
20397 (hash_table::elements_with_deleted) New.
20398 (class hashtable::iterator): New.
20399 (hashtable::begin()): New.
20400 (hashtable::end()): New.
20401 (FOR_EACH_HASH_TABLE_ELEMENT): New.
20403 * statistics.c (statistics_hashes):
20404 Change type to hash_table. Update dependent calls and types.
20406 * tree-into-ssa.c (var_infos):
20407 Change type to hash_table. Update dependent calls and types.
20409 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
20410 Change type to hash_table. Update dependent calls and types.
20412 * tree-ssa-loop-im.c (struct mem_ref.refs):
20413 Change type to hash_table. Update dependent calls and types.
20415 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
20416 Change type to hash_table. Update dependent calls and types.
20418 * tree-ssa-sccvn.c (vn_tables_s::nary):
20419 Change type to hash_table. Update dependent calls and types.
20420 (vn_tables_s::phis): Likewise.
20421 (vn_tables_s::references): Likewise.
20423 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
20424 (vn_reference_eq): Update parameter and return types.
20426 * tree-ssa-structalias.c (pointer_equiv_class_table):
20427 Change type to hash_table. Update dependent calls and types.
20428 (location_equiv_class_table): Likewise.
20430 * tree-vect-data-refs.c: Consequential changes for making
20431 peeling a hash_table.
20433 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
20434 (destroy_loop_vec_info): Dependent hash_table update.
20436 * tree-vectorizer.h (peeling_htab):
20437 Change type to hash_table. Update dependent calls and types.
20439 2013-04-23 Shiva Chen <shiva0217@gmail.com>
20441 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
20442 to check the register content is equal or not.
20443 * lra-constraints.c (match_reload): Use lra_assign_reg_val
20444 to assign register content record.
20445 * lra-eliminations.c (update_reg_eliminate): Use
20446 lra_update_reg_val_offset to update register content offset.
20447 * lra-int.h (struct lra_reg): Add offset member.
20448 (lra_reg_val_equal_p): New static inline function.
20449 (lra_update_reg_val_offset): New static inline function.
20450 (lra_assign_reg_val): New static inline function.
20451 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
20452 to assign register content record.
20453 (initialize_lra_reg_info_element): Initial offset to zero.
20455 2013-04-23 Catherine Moore <clm@codesourcery.com>
20457 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
20458 operands. Record compression.
20460 2013-04-23 Xinliang David Li <davidxl@google.com>
20462 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
20464 2013-04-23 Richard Biener <rguenther@suse.de>
20466 PR middle-end/57036
20467 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
20468 parameter, only add abnormal goto edges from the copied body
20469 if the call could perform abnormal gotos.
20470 (copy_cfg_body): Adjust.
20472 2013-04-23 Sofiane Naci <sofiane.naci@arm.com>
20474 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
20476 2013-04-23 Andreas Schwab <schwab@linux-m68k.org>
20478 * coretypes.h (gimple_stmt_iterator): Add struct to make
20481 2013-04-23 Richard Biener <rguenther@suse.de>
20483 PR tree-optimization/57026
20484 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
20485 from SSA names occuring in abnormal PHI nodes.
20487 2013-04-22 Andi Kleen <ak@linux.intel.com>
20489 * lto/lto.c (print_lto_report_1): Fix LTO report names.
20491 2013-04-22 Andi Kleen <ak@linux.intel.com>
20493 * lto/lto.c (print_lto_report_1): Declare early.
20494 (read_cgraph_and_symbols): Call print_lto_report_1 early.
20496 2013-04-22 Andi Kleen <ak@linux.intel.com>
20498 * common.opt (-flto-report-wpa): Add.
20499 * doc/invoke.texi (-flto-report-wpa): Add.
20500 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
20503 2013-04-22 Xinliang David Li <davidxl@google.com>
20505 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
20506 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
20507 * Makefile.in: New dependency
20509 David Daney <ddaney.cavm@gmail.com>
20511 * configure.ac (gcc_cv_as_micromips_support): Use the
20512 --fatal-warnings option.
20513 * configure: Regenerate.
20515 2013-04-22 Marek Polacek <polacek@redhat.com>
20518 * tsan.c (instrument_expr): Don't instrument expression
20519 in case its size is zero.
20521 2013-04-22 Uros Bizjak <ubizjak@gmail.com>
20525 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
20527 * config/alpha/alpha.c (TARGET_LRA_P): New define.
20529 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
20531 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
20532 (gimple_stmt_iterator): New typedef.
20533 * gimple.h (gimple_stmt_iterator): Rename to...
20534 (gimple_stmt_iterator_d): ... This.
20535 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
20536 trees be valid for GIMPLE and GENERIC.
20537 (TARGET_GIMPLE_FOLD_BUILTIN): New.
20538 * gimple-fold.c (gimple_fold_call): Call target hook
20539 gimple_fold_builtin.
20540 * hooks.c (hook_bool_gsiptr_false): New.
20541 * hooks.h (hook_bool_gsiptr_false): New.
20542 * target.def (fold_stmt): New.
20543 * doc/tm.texi: Regenerate.
20545 2013-04-22 Vladimir Makarov <vmakarov@redhat.com>
20548 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
20549 a set sp if no stack realignment.
20551 2013-04-22 Nick Clifton <nickc@redhat.com>
20553 * config.gcc (tilegx-linux): Extend extra_objs rather than
20555 (tilepro-linux): Likewise.
20557 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
20559 * config/aarch64/aarch64-builtins.c
20561 (CF0, CF1, CF2, CF3, CF4, CF10): New.
20562 (VAR<1-12>): Add MAP parameter.
20563 (BUILTIN_*): Likewise.
20564 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
20565 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
20566 (aarch64_ushl_n<mode>): Likewise.
20567 (aarch64_sshr_n<mode>): Likewise.
20568 (aarch64_ushr_n<mode>): Likewise.
20569 (aarch64_<maxmin><mode>): Likewise.
20570 (aarch64_sqrt<mode>): Likewise.
20571 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
20572 (vshr<q>_n_*): Likewise.
20574 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
20576 * config/aarch64/aarch64-builtins.c
20577 (aarch64_simd_builtin_type_mode): Handle SF types.
20579 (BUILTIN_GPF): Define.
20580 (aarch64_init_simd_builtins): Handle SF types.
20581 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
20582 (frecps): Likewise.
20583 (frecpx): Likewise.
20584 * config/aarch64/aarch64-simd.md
20585 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
20586 (aarch64_frecpe<mode>): New.
20587 (aarch64_frecps<mode>): Likewise.
20588 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
20589 (v8type): Add frecp<esx>.
20590 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
20591 (aarch64_frecps<mode>): Likewise.
20592 * config/aarch64/iterators.md (FRECP): New.
20593 (frecp_suffix): Likewise.
20594 * config/aarch64/arm_neon.h
20595 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
20597 2013-04-22 Christian Bruel <christian.bruel@st.com>
20600 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
20601 (REG_CLASS_NAMES): Idem.
20602 (REG_CLASS_CONTENTS): Idem.
20603 (REGCLASS_HAS_FP_REG): Idem.
20604 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
20605 (sh_conditional_register_usage): Idem.
20607 2013-04-21 Jeff Law <law@redhat.com>
20609 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
20610 (ssa_forward_propagate_and_combine): Use it.
20612 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
20614 * lra.c: Update the flow chart diagram.
20616 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
20618 PR rtl-optimization/56847
20619 * lra-constraints.c (process_alt_operands): Discourage alternative
20620 with non-matche doffsettable memory constraint fro memory with
20623 2013-04-19 Richard Biener <rguenther@suse.de>
20625 PR tree-optimization/56982
20626 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
20628 * gimplify.c (gimplify_call_expr): Notice special calls.
20629 (gimplify_modify_expr): Likewise.
20630 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
20631 abnormal control flow receivers.
20632 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
20633 in the same way as cfun->has_nonlocal_labels.
20634 (gimple_purge_dead_abnormal_call_edges): Likewise.
20635 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
20636 receivers start a basic-block.
20638 2013-04-19 Richard Biener <rguenther@suse.de>
20640 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
20642 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
20643 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
20644 (SLP_TREE_LOAD_PERMUTATION): Add.
20645 (vect_transform_slp_perm_load): Adjust prototype.
20646 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
20647 (vect_free_slp_instance): Likewise.
20648 (vect_create_new_slp_node): Likewise.
20649 (vect_supported_slp_permutation_p): Remove.
20650 (vect_slp_rearrange_stmts): Adjust.
20651 (vect_supported_load_permutation_p): Likewise. Inline
20652 vect_supported_slp_permutation_p here.
20653 (vect_analyze_slp_instance): Compute load permutations per
20654 slp node instead of per instance.
20655 (vect_get_slp_defs): Adjust.
20656 (vect_transform_slp_perm_load): Likewise.
20657 (vect_schedule_slp_instance): Remove redundant code.
20658 (vect_schedule_slp): Remove hack for PR56270, add it ...
20659 * tree-vect-stmts.c (vectorizable_load): ... here, do not
20660 CSE loads for SLP. Adjust.
20662 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
20664 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
20665 spelling in two comments.
20667 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
20670 * config/arm/arm.c (load_multiple_sequence): Require SP
20671 as base register for loads if SP is in the register list.
20673 2013-04-19 Martin Jambor <mjambor@suse.cz>
20675 PR tree-optimization/56718
20676 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
20677 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
20678 and made public. Adjusted all callers.
20679 (ipa_intraprocedural_devirtualization): New function.
20680 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
20681 (ipa_intraprocedural_devirtualization): Likewise.
20682 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
20684 2013-04-19 Richard Biener <rguenther@suse.de>
20686 PR tree-optimization/57000
20687 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
20689 2013-04-19 Terry Guo <terry.guo@arm.com>
20691 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
20693 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
20694 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
20695 (cortex_m4_fmacs): Use new reservations.
20696 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
20698 2013-04-18 Vladimir Makarov <vmakarov@redhat.com>
20700 PR rtl-optimization/56999
20701 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
20703 (lra_coalesce): Remove split_origin_bitmap and related code.
20704 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
20705 ranges if necessary.
20707 2013-04-18 Uros Bizjak <ubizjak@gmail.com>
20709 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
20711 (ix86_expand_call): Remove clobbered_registers array and use
20712 x86_64_ms_sysv_extra_clobbered_registers instead.
20713 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
20715 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
20717 * config/i386/i386.md (*call_rex64_ms_sysv): Use
20718 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
20719 (*call_value_rex64_ms_sysv): Ditto.
20721 2013-04-18 Cary Coutant <ccoutant@google.com>
20723 * dwarf2out.c (output_pubnames): Check die_perennial_p of
20724 parent instead of die_mark.
20726 2013-04-18 Diego Novillo <dnovillo@google.com>
20728 * gimple.c (create_gimple_tmp): New.
20729 (get_expr_type): New.
20730 (build_assign): New.
20731 (build_type_cast): New.
20732 * gimple.h (enum ssa_mode): Define.
20733 (gimple_seq_set_location): New.
20734 * asan.c (build_check_stmt): Change some gimple_build_* calls
20735 to use build_assign and build_type_cast.
20737 2013-04-18 Richard Biener <rguenther@suse.de>
20739 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
20740 handle negative step. Remove redundant checks.
20741 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
20742 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
20743 for negative step and grouped loads fail to vectorize.
20745 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
20747 * emit-rtl.c (reset_insn_used_flags): New function.
20748 (reset_all_used_flags): Use it.
20749 (verify_insn_sharing): New function.
20750 (verify_rtl_sharing): Fix verification for SEQUENCEs.
20752 2013-04-18 Jakub Jelinek <jakub@redhat.com>
20754 PR tree-optimization/56984
20755 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
20756 and (x >> M) >= N don't register any assertion if N << M is the
20759 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
20761 * lower-subreg.c (resolve_simple_move): If called self-recursive,
20762 do not delete_insn insns that have not yet been emitted, only
20763 unlink them with remove_insn.
20764 * df-scan.c (df_insn_delete): Revert r197492.
20766 2013-04-17 Steven Bosscher <steven@gcc.gnu.org>
20768 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
20769 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
20771 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
20773 * config/arm/arm.md (movsicc_insn): Convert define_insn into
20774 define_insn_and_split.
20775 (and_scc,ior_scc,negscc): Likewise.
20776 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
20778 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
20780 * config/arm/arm.c (use_return_insn): Return 0 for targets that
20781 can benefit from using a sequence of LDRD instructions in epilogue
20782 instead of a single LDM instruction.
20784 2013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
20787 * doc/extend.texi: Fix typo.
20789 2013-04-17 Richard Biener <rguenther@suse.de>
20791 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
20792 (vect_build_slp_tree): ... here.
20793 (vect_build_slp_tree_1): Compute which stmts of the SLP group
20794 match. Remove special-casing of mismatched complex loads.
20795 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
20796 re-try the match with swapped commutative operands.
20797 (vect_supported_load_permutation_p): Remove special-casing of
20798 mismatched complex loads.
20799 (vect_analyze_slp_instance): Adjust.
20801 2013-04-17 Richard Biener <rguenther@suse.de>
20803 PR rtl-optimization/56921
20804 * cfgloop.h (struct loop): Add simple_loop_desc member.
20805 (struct niter_desc): Mark with GTY(()).
20806 (simple_loop_desc): Do not use aux field but simple_loop_desc.
20807 * loop-iv.c (get_simple_loop_desc): Likewise.
20808 (free_simple_loop_desc): Likewise.
20811 2013-04-16 Richard Biener <rguenther@suse.de>
20813 PR rtl-optimization/56921
20814 * loop-init.c (pass_rtl_move_loop_invariants): Add
20815 TODO_do_not_ggc_collect to todo_flags_finish.
20816 (pass_rtl_unswitch): Same.
20817 (pass_rtl_unroll_and_peel_loops): Same.
20818 (pass_rtl_doloop): Same.
20820 2013-04-17 Eric Botcazou <ebotcazou@adacore.com>
20822 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
20823 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
20824 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
20826 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
20827 * tree-streamer.c (record_common_node): Adjust reference in comment.
20829 2013-04-17 Terry Guo <terry.guo@arm.com>
20831 * config/arm/cortex-m4.md: Add a new bypass.
20833 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
20835 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
20837 (*subs_<optab><mode>_multp2): New pattern.
20838 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
20839 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
20841 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
20843 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
20844 (*subs_mul_imm_<mode>): New pattern.
20846 2013-04-16 David Edelsohn <dje.gcc@gmail.com>
20849 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
20850 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
20851 (vsx_movti_32bit): Change j->wa to O->wa.
20853 2013-04-16 Richard Biener <rguenther@suse.de>
20855 PR rtl-optimization/56921
20856 * loop-init.c (pass_rtl_move_loop_invariants): Add
20857 TODO_do_not_ggc_collect to todo_flags_finish.
20858 (pass_rtl_unswitch): Same.
20859 (pass_rtl_unroll_and_peel_loops): Same.
20860 (pass_rtl_doloop): Same.
20862 2013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
20864 * config/arm/arm.c (emit_multi_reg_push): New declaration
20865 for an existing function.
20866 (arm_emit_strd_push): New function.
20867 (arm_expand_prologue): Used here.
20868 (arm_emit_ldrd_pop): New function.
20869 (arm_expand_epilogue): Used here.
20870 (arm_get_frame_offsets): Update condition.
20871 (arm_emit_multi_reg_pop): Add a special case for load of a single
20872 register with writeback.
20874 2013-04-16 Uros Bizjak <ubizjak@gmail.com>
20876 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
20879 2013-04-16 Richard Biener <rguenther@suse.de>
20881 PR tree-optimization/56756
20882 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
20883 (first_mem_ref_loc): New.
20884 (execute_sm): Place the load temporarily before a previous
20885 access instead of in the latch edge to ensure its SSA dependencies
20886 are defined at points dominating the load.
20888 2013-04-16 Steven Bosscher <steven@gcc.gnu.org>
20890 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
20891 correct fix by moving header and footer insn to the footer of
20892 the merged basic block. Clear BB_END of the merged-away block.
20894 PR middle-end/43631
20895 * emit-rtl.c (make_note_raw): New function.
20896 (link_insn_into_chain): New static inline function.
20897 (add_insn): Use it.
20898 (add_insn_before, add_insn_after): Factor insn chain linking code...
20899 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
20900 using link_insn_into_chain.
20901 (note_outside_basic_block_p): New helper function for emit_note_after
20902 and emit_note_before.
20903 (emit_note_after): Use nobb variant of add_insn_after if the note
20904 should not be contained in a basic block.
20905 (emit_note_before): Use nobb variant of add_insn_before if the note
20906 should not be contained in a basic block.
20907 (emit_note_copy): Use make_note_raw.
20908 (emit_note): Likewise.
20909 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
20910 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
20911 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
20912 the moved barrier the tail of the basic block it follows.
20913 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
20915 2013-04-15 Jakub Jelinek <jakub@redhat.com>
20917 PR tree-optimization/56962
20918 * gimple-ssa-strength-reduction.c (record_increment): Only set
20919 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
20920 either rhs1 or rhs2 is equal to c->base_expr.
20922 2013-04-15 Richard Biener <rguenther@suse.de>
20924 PR tree-optimization/56933
20925 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
20927 (GROUP_READ_WRITE_DEPENDENCE): Remove.
20928 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
20929 * tree-vect-data-refs.c (vect_analyze_group_access): Move
20930 dependence check ...
20931 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
20933 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
20934 GROUP_READ_WRITE_DEPENDENCE.
20936 2013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
20938 * emit-rtl.c (reset_all_used_flags): New function.
20939 (verify_rtl_sharing): Call reset_all_used_flags before and after
20940 performing the checks.
20942 2013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20944 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
20945 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
20946 * config/arm/constraints.md (De): New constraint.
20947 * config/arm/neon.md (anddi3_neon): Delete.
20948 (neon_vand<mode>): Expand to standard anddi3 pattern.
20949 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
20950 Move earlier in the file.
20951 (neon_inv_logic_op2): Likewise.
20952 (arm_anddi_operand_neon): New predicate.
20954 2013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
20956 * configure.ac (gcc_cv_ld_as_needed): Set
20957 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
20958 Use -z ignore, -z record on *-*-solaris2*.
20959 (HAVE_LD_AS_NEEDED): Update comment.
20960 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
20961 * configure: Regenerate.
20962 * config.in: Regenerate.
20963 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
20964 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
20965 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
20966 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
20967 equivalents. Fix markup.
20968 * doc/tm.texi: Regenerate.
20970 2013-04-15 Andrew Hsieh <andrewhsieh.google.com>
20972 * config/i386/i386.opt: New option mstack-protector-guard=.
20973 * config/i386/i386-opts.h: Add enum stack_protector_guard.
20974 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
20975 TARGET_SSP_TLS_GUARD.
20976 * config/i386/i386.c (ix86_option_override_internal): Set
20977 ix86_stack_protector_guard.
20978 * config/i386/i386.md (stack_protect_set): Enable for
20979 TARGET_SSP_TLS_GUARD only.
20980 (stack_protect_set_<mode>): Ditto.
20981 (stack_protect_test): Ditto.
20982 (stack_protect_test_<mode>): Ditto.
20983 * doc/invoke.texi (i386 Option): Document.
20985 2013-04-15 Eric Botcazou <ebotcazou@adacore.com>
20988 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
20989 (S_MODES): Set H_MODE bit.
20990 (SF_MODES): Set only S_MODE and SF_MODE bits.
20991 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
20992 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
20993 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
20994 <MODE_FLOAT>: Likewise.
20996 2013-04-15 Joey Ye <joey.ye@arm.com>
20998 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
21000 2013-04-15 Joey Ye <joey.ye@arm.com>
21002 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
21004 (thumb_far_jump_used_p): Count instruction size and set
21007 2013-04-14 Eric Botcazou <ebotcazou@adacore.com>
21009 * reorg.c (fill_simple_delay_slots): Reindent block of code.
21010 * resource.c (mark_target_live_regs): Reformat conditional block.
21012 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
21014 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
21015 notes, they are emitted only just before final.
21016 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
21018 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
21020 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
21021 * cfgrtl.c (delete_insn): Call it here instead.
21022 * lra-spills.c (lra_final_code_change): Use delete_insn.
21023 * haifa-sched.c (sched_remove_insn): Likewise.
21024 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
21025 returning to the nop pool.
21026 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
21027 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
21029 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
21031 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
21032 * doc/tm.texi: Regenerated.
21034 2013-04-12 Uros Bizjak <ubizjak@gmail.com>
21036 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
21039 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
21041 * df-core.c (df_find_def): Compare register numbers.
21042 (df_find_use): Likewise.
21044 2013-04-12 Vladimir Makarov <vmakarov@redhat.com>
21047 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
21048 lra_in_progress for return.
21050 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
21052 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
21053 define_insn into define_insn_and_split and emit movsicc patterns.
21055 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
21057 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
21059 2013-04-12 Richard Biener <rguenther@suse.de>
21061 * tree-pass.h (TODO_do_not_ggc_collect): New.
21062 * passes.c (execute_one_ipa_transform_pass): Honor
21063 TODO_do_not_ggc_collect.
21064 (execute_one_pass): Likewise.
21067 2013-04-10 Richard Biener <rguenther@suse.de>
21069 * passes.c (init_optimization_passes): Remove reload pass.
21070 * ira.c (do_reload): Merge into ...
21072 (rest_of_handle_reload): Remove.
21073 (pass_reload): Likewise.
21074 * config/i386/i386.c (ix86_option_override): Refer to ira instead
21075 of reload for vzeroupper pass placement.
21077 2013-04-12 Jakub Jelinek <jakub@redhat.com>
21079 PR tree-optimization/56918
21080 PR tree-optimization/56920
21081 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
21082 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
21083 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
21084 use wide_mul_with_sign method.
21086 2013-04-12 Richard Biener <rguenther@suse.de>
21088 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
21089 not be considered a gimple constant.
21091 2013-04-12 Marc Glisse <marc.glisse@inria.fr>
21093 * fold-const.c (const_binop): Handle vector shifts by a scalar.
21094 (fold_binary_loc): Call const_binop also for mixed vector-scalar
21097 2013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
21098 Jakub Jelinek <jakub@redhat.com>
21100 * opts.c: Include diagnostic-color.h.
21101 (common_handle_option): Handle OPT_fdiagnostics_color_.
21102 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
21103 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
21104 (diagnostic-color.o): New.
21105 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
21106 (diagnostic_color_rule): New enum.
21107 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
21108 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
21109 the location string.
21110 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
21111 either NULL, or color kind.
21112 * diagnostic-color.c: New file.
21113 * diagnostic-color.h: New file.
21114 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
21116 * doc/invoke.texi (-fdiagnostics-color): Document.
21117 * pretty-print.h (pp_show_color): Define.
21118 (struct pretty_print_info): Add show_color field.
21119 * diagnostic.c: Include diagnostic-color.h.
21120 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
21121 macros. Colorize error:, warning: etc. strings and also the location
21123 (diagnostic_show_locus): Colorize the caret line.
21124 * pretty-print.c: Include diagnostic-color.h.
21125 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
21126 inside of %< %> quotes or quoted through q format modifier.
21128 2013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
21130 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
21132 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
21134 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
21135 code in CC_NZ mode.
21136 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
21139 2013-04-11 Marek Polacek <polacek@redhat.com>
21141 PR tree-optimization/48184
21142 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
21144 2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
21146 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
21147 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
21148 (skip_simple_arithmetic): Tidy up.
21149 * tree.h (skip_simple_constant_arithmetic): Declare.
21151 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
21153 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
21155 2013-04-11 Richard Biener <rguenther@suse.de>
21157 * tree-vect-loop.c (get_initial_def_for_induction): Properly
21158 generate vector constants.
21160 2013-04-11 Richard Biener <rguenther@suse.de>
21162 PR tree-optimization/56878
21163 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
21164 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
21166 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
21167 Prefer to align the DR with the most invariant base address.
21169 2013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
21171 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
21174 2013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
21176 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
21177 floating-point vector comparisons against 0.
21179 2013-04-11 Jakub Jelinek <jakub@redhat.com>
21181 PR tree-optimization/56899
21182 * fold-const.c (extract_muldiv_1): Apply distributive law
21183 only if TYPE_OVERFLOW_WRAPS (ctype).
21185 2013-04-11 Bin Cheng <bin.cheng@arm.com>
21188 * ira-costs.c (scan_one_insn): Check whether the source rtx of
21189 loading has side effect.
21191 2013-04-10 Steven Bosscher <steven@gcc.gnu.org>
21193 * config/sparc/sparc.c: Include tree-pass.h.
21194 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
21195 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
21196 head of file. Change return type. Split off gate function.
21197 (sparc_gate_work_around_errata): New function.
21198 (pass_work_around_errata): New pass definition.
21199 (insert_pass_work_around_errata) New pass insert definition to
21200 insert pass_work_around_errata just after delayed-branch scheduling.
21201 (sparc_option_override): Insert the pass.
21202 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
21204 2013-04-10 David S. Miller <davem@davemloft.net>
21206 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
21207 or -mcpu=hypersparc.
21209 * target.def (cstore_mode): New hook.
21210 * target.h: Include insn-codes.h
21211 * targhooks.c: Likewise.
21212 (default_cstore_mode): New function.
21213 * targhooks.h: Declare it.
21214 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
21215 * doc/tm.texi: Rebuild.
21216 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
21217 target hook, rather than inspecting the insn_data.
21218 * config/sparc/sparc.c (sparc_cstore_mode): New function.
21219 (TARGET_CSTORE_MODE): Redefine.
21220 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
21222 * config/sparc/predicates.md (cstore_result_operand): New special
21224 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
21225 Use it for operand 0.
21226 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
21227 (*snesi_special): Likewise.
21228 (*snesi_zero): Likewise.
21229 (*seqsi_zero): Likewise.
21230 (*sltu_insn): Likewise.
21231 (*sgeu_insn): Likewise.
21232 (*seqdi_special): Make operand 0 and comparison operation be of
21234 (*snedi_special): Likewise.
21235 (*snedi_special_vis3): Likewise.
21236 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
21237 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
21238 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
21239 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
21240 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
21241 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
21242 (*sltu_extend_sp64): Likewise.
21243 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
21244 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
21245 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
21246 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
21247 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
21249 2013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
21251 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
21252 (aarch64_start_file): Use the new function.
21254 2013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
21255 Jason Merrill <jason@redhat.com>
21257 * common.opt: Add -gdwarf.
21258 * opts.c (common_handle_option): Handle it.
21259 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
21261 2013-04-10 Richard Biener <rguenther@suse.de>
21263 * passes.c (execute_todo): Do not call ggc_collect conditional here.
21264 (execute_one_ipa_transform_pass): But unconditionally here.
21265 (execute_one_pass): And here.
21266 (init_optimization_passes): Remove reload pass.
21267 * tree-pass.h (TODO_ggc_collect): Remove.
21268 (pass_reload): Likewise.
21269 * ira.c (do_reload): Merge into ...
21271 (rest_of_handle_reload): Remove.
21272 (pass_reload): Likewise.
21273 * config/i386/i386.c (ix86_option_override): Refer to ira instead
21274 of reload for vzeroupper pass placement.
21275 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
21276 and todo_flags_finish of all passes.
21278 2013-04-10 Richard Biener <rguenther@suse.de>
21280 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
21281 first_const_oprnd field, rename first_def_type to first_op_type.
21282 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
21283 (vect_get_and_check_slp_defs): Always use the type of the
21284 operand. Allow mixed vect_external_def, vect_constant_def types.
21285 (vect_get_constant_vectors): Handle mixed vect_external_def,
21286 vect_constant_def types.
21288 2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
21290 PR tree-optimization/55524
21291 * tree-ssa-math-opts.c
21292 (convert_mult_to_fma): Don't use an fms construct
21293 when we don't have an fms operation, but fnma, and it looks
21294 likely that we'll be able to use the latter.
21296 2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
21298 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
21300 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
21301 inline fail caused by overwritable functions.
21303 2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
21305 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
21306 unnecessary bits in the constant power of two case.
21308 2013-04-10 Richard Biener <rguenther@suse.de>
21310 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
21311 broken code swapping operands.
21312 (vect_build_slp_tree): Do not compute load permutations here.
21313 (vect_analyze_slp_instance): Compute load permutations here,
21314 after building the SLP tree.
21316 2013-04-09 Christian Bruel <christian.bruel@st.com>
21318 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
21319 of next/prev_real_insn.
21321 2013-04-09 Jan Hubicka <jh@suse.cz>
21323 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
21324 Drop aliased parameter.
21325 (function_and_variable_visibility): Do not handle alias pairs.
21326 * cgraph.c (varpool_externally_visible_p): Update prototype.
21327 * varpool.c (varpool_add_new_variable): Update.
21329 2013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21331 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
21333 2013-04-09 Steven Bosscher <steven@gcc.gnu.org>
21335 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
21337 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
21339 2013-04-09 Marek Polacek <polacek@redhat.com>
21341 PR tree-optimization/48762
21342 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
21344 2013-04-09 Richard Biener <rguenther@suse.de>
21346 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
21348 (vect_build_slp_tree): Likewise.
21349 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
21350 calculating the cost of a SLP instance.
21351 (vect_analyze_slp_instance): Use it from here, after building
21354 2013-04-09 Jakub Jelinek <jakub@redhat.com>
21356 PR middle-end/56883
21357 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
21358 expand_omp_for_static_chunk): Use simple_p = true in
21359 force_gimple_operand_gsi calls when assigning to addressable decls.
21361 2013-04-09 Jeff Law <law@redhat.com>
21363 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
21364 when the boolean was created by converting a wider object which
21365 had a boolean range.
21367 2013-04-09 Richard Biener <rguenther@suse.de>
21369 * tree-vectorizer.h (slp_void_p): Remove.
21370 (slp_tree): Typedef before _slp_tree declaration.
21371 (struct _slp_tree): Use a vector of slp_tree as children.
21372 (vect_get_place_in_interleaving_chain): Remove.
21373 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
21375 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
21377 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
21378 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
21379 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
21380 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
21381 Use slp_node instead of slp_void_p and adjust.
21383 2013-04-09 Richard Biener <rguenther@suse.de>
21385 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
21386 work that is not necessary.
21388 2013-04-09 Jakub Jelinek <jakub@redhat.com>
21390 PR tree-optimization/56854
21391 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
21392 forward into clobber stmts if it would change MEM_REF lhs into
21395 2013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
21397 * tree.c (type_hash_lookup, type_hash_add): Make static.
21398 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
21400 2013-04-09 Richard Biener <rguenther@suse.de>
21402 * tree.h (unsave_expr_now): Remove.
21403 * tree-inline.c (mark_local_for_remap_r): Remove.
21404 (unsave_expr_1): Likewise.
21405 (unsave_r): Likewise.
21406 (unsave_expr_now): Likewise.
21407 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
21408 (propagate_tree_value): Likewise.
21410 2013-04-08 Steven Bosscher <steven@gcc.gnu.org>
21412 * doc/rtl.texi (sequence): Rewrite documentation to match the
21413 current use of SEQUENCE rtl objects.
21414 * rtl.def (SEQUENCE): Likewise.
21416 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
21417 Update documentation.
21418 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
21419 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
21421 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
21423 2013-04-08 Teresa Johnson <tejohnson@google.com>
21425 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
21426 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
21428 (estimate_edge_size_and_time): Add comment to suggest using rounding
21430 (estimate_node_size_and_time): Ditto.
21431 (remap_edge_change_prob): Use helper rounding divide methods.
21432 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
21433 (gimple_mod_pow2_value_transform): Ditto.
21434 (gimple_mod_subtract_transform): Ditto.
21435 (gimple_ic_transform): Ditto.
21436 (gimple_stringops_transform): Ditto.
21437 * stmt.c (conditional_probability): Ditto.
21438 (emit_case_dispatch_table): Ditto.
21439 * lto-cgraph.c (merge_profile_summaries): Ditto.
21440 * tree-optimize.c (execute_fixup_cfg): Ditto.
21441 * cfgcleanup.c (try_forward_edges): Ditto.
21442 * cfgloopmanip.c (scale_loop_profile): Ditto.
21444 (duplicate_loop_to_header_edge): Ditto.
21445 (lv_adjust_loop_entry_edge): Ditto.
21446 * tree-vect-loop.c (vect_transform_loop): Ditto.
21447 * profile.c (compute_branch_probabilities): Ditto.
21448 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
21449 * lto-streamer-in.c (input_cfg): Ditto.
21450 * gimple-streamer-in.c (input_bb): Ditto.
21451 * ipa-cp.c (update_profiling_info): Ditto.
21452 (update_specialized_profile): Ditto.
21453 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
21454 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
21456 * sched-rgn.c (compute_dom_prob_ps): Ditto.
21457 (compute_trg_info): Ditto.
21458 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
21459 (purge_dead_edges): Ditto.
21460 * loop-unswitch.c (unswitch_loop): Ditto.
21461 * cgraphclones.c (cgraph_clone_edge): Ditto.
21462 (cgraph_clone_node): Ditto.
21463 * tree-inline.c (copy_bb): Ditto.
21464 (copy_edges_for_bb): Ditto.
21465 (initialize_cfun): Ditto.
21466 (copy_cfg_body): Ditto.
21467 (expand_call_inline): Ditto.
21469 2013-04-08 Kai Tietz <ktietz@redhat.com>
21471 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
21472 TARGET_CYGWIN64 by TARGET_64BIT.
21474 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
21476 * config/epiphany/epiphany.md (GPR_1): New constant.
21477 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
21478 * config/epiphany/epiphany.c (gen_compare_reg):
21479 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
21480 is already in place.
21481 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
21482 Don't require being called during rtl expansion; If y operlaps r0,
21484 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
21485 (epiphany_expand_epilogue): Likewise.
21487 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
21488 Don't use CC_FPmode for ORDERED / UNORDERED.
21489 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
21491 * config/epiphany/constraints.md (CnL): New constraint.
21492 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
21493 * config/epiphany/predicates.md (add_operand): Allow 1024.
21495 * config/epiphany/epiphany.md (logical_op): New code iterator.
21496 (op_mnc): New code attribute.
21497 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
21498 (mov_f+1, mov_f+2): New peephole2 patterns.
21500 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
21501 (cstoresi4): Also allow re-use of zero result when doing a NE
21502 comparison to a non-zero operand.
21503 Use (clobber (scratch)) for first insn if the gpr output is not needed.
21505 * config/epiphany/epiphany.md (<insn_opname>v2si3):
21506 Use gen_addsi3_i / gen_subsi3_i.
21508 2013-04-08 Jakub Jelinek <jakub@redhat.com>
21512 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
21513 contain anything but clobbers, at most one __builtin_stack_restore,
21514 optionally debug stmts and final resx, and if it has at least one
21515 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
21516 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
21517 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
21518 which isn't defaut definition, remove them.
21519 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
21520 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
21521 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
21522 with MEM_REF LHS with SSA_NAME address.
21524 2013-04-08 Jeff Law <law@redhat.com>
21526 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
21528 2013-04-08 Richard Biener <rguenther@suse.de>
21530 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
21532 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
21533 determined vector type.
21534 (vect_analyze_data_refs): Likewise.
21535 (vect_get_new_vect_var): Adjust.
21536 (vect_create_destination_var): Preserve SSA name versions.
21537 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
21538 not dump anything here.
21540 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
21542 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
21543 Add member lr_slot_known.
21544 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
21546 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
21547 Remove code that sets lr_slot_offset according to what a previous
21548 version of epiphany_emit_save_restore used to do.
21549 (epiphany_emit_save_restore): When doing an lr save or restore,
21550 set/verify lr_slot_known and lr_slot_offset.
21552 2013-04-08 Xinyu Qi <xyqi@marvell.com>
21555 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
21558 2013-04-08 Richard Biener <rguenther@suse.de>
21560 * alias.c (find_base_term): Fix thinko in previous change.
21562 2013-04-08 Jakub Jelinek <jakub@redhat.com>
21564 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
21565 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
21566 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
21567 if possible to compute val.
21568 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
21569 For QImode integers don't require anything about precision. Use
21570 const_with_all_bytes_same to find out if the constant doesn't have
21571 repeated bytes in it.
21573 2013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
21575 * config/s390/s390.c (s390_expand_insv): Only accept insertions
21578 2013-04-08 Marek Polacek <polacek@redhat.com>
21580 PR rtl-optimization/48182
21581 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
21584 2013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
21587 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
21588 nuses, make sure we have a label.
21590 2013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
21593 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
21594 (rs6000_emit_swdiv_low_precision): Remove.
21595 (rs6000_emit_swdiv): Rewrite to handle between one and four
21596 iterations of Newton-Raphson generally; modify required number of
21597 iterations for some cases.
21598 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
21600 2013-04-05 Steven Bosscher <steven@gcc.gnu.org>
21602 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
21603 set-but-unused variable.
21605 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
21606 basic blocks of released function bodies garbage-collectable.
21608 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
21609 (struct rtl_opt_pass): Add TODO_df_finish.
21611 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
21613 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
21615 * config/arm/constraints.md (q): New constraint.
21616 * config/arm/ldrdstrd.md: New file.
21617 * config/arm/arm.md (ldrdstrd.md) New include.
21618 (arm_movdi): Use "q" instead of "r" constraint
21619 for double-word memory access.
21620 (movdf_soft_insn): Likewise.
21621 * config/arm/vfp.md (movdi_vfp): Likewise.
21622 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
21623 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
21624 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
21625 (mem_ok_for_ldrd_strd): Likewise.
21626 (output_move_double): Update assertion.
21628 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
21630 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
21632 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
21634 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
21635 define_insn_and_split.
21636 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
21638 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
21640 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
21641 define_insn_and_split.
21642 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
21643 (shiftsi3_compare): New pattern.
21644 (rrx): New pattern.
21645 * config/arm/unspecs.md (UNSPEC_RRX): New.
21647 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
21649 * config/arm/arm.md (negdi_extendsidi): New pattern.
21650 (negdi_zero_extendsidi): Likewise.
21652 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
21654 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
21655 define_insn_and_split.
21656 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
21657 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
21659 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
21661 * config/arm/arm.md (arm_subdi3): Convert define_insn into
21662 define_insn_and_split.
21663 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
21664 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
21666 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
21668 * config/arm/arm.md (subsi3_carryin): New pattern.
21669 (subsi3_carryin_const): Likewise.
21670 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
21671 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
21673 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
21675 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
21677 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
21679 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
21680 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
21682 2013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21684 * config/arm/arm.c (arm_expand_builtin): Change fcode
21685 type to unsigned int.
21687 2013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
21689 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
21691 2013-04-04 Ian Lance Taylor <iant@google.com>
21693 * doc/standards.texi (Standards): The Go frontend supports the Go 1
21696 2013-04-04 Steven Bosscher <steven@gcc.gnu.org>
21698 PR middle-end/56729
21699 * df-scan.c (df_insn_delete): Disable failing assert.
21701 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21703 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
21704 New function prototype.
21705 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
21706 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
21707 (arm_builtin_vectorized_function): New function.
21709 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21711 * config/arm/arm_neon_builtins.def: New file.
21712 * config/arm/arm.c (neon_builtin_data): Move contents to
21713 arm_neon_builtins.def.
21714 (enum arm_builtins): Include neon builtin definitions.
21715 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
21716 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
21718 2013-04-04 Marek Polacek <polacek@redhat.com>
21720 PR tree-optimization/48186
21721 * predict.c (maybe_hot_frequency_p): Return false if
21722 HOT_BB_FREQUENCY_FRACTION is 0.
21723 (cgraph_maybe_hot_edge_p): Likewise.
21725 2013-04-04 Richard Biener <rguenther@suse.de>
21727 PR tree-optimization/56826
21728 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
21731 2013-04-04 Richard Biener <rguenther@suse.de>
21733 PR tree-optimization/56213
21734 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
21735 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
21737 2013-04-04 Richard Biener <rguenther@suse.de>
21739 PR tree-optimization/56837
21740 * tree-loop-distribution.c (classify_partition): For non-zero
21741 values require that the value has the same precision as its
21742 mode to be useful as memset value.
21744 2013-04-03 Nick Clifton <nickc@redhat.com>
21746 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
21747 (fmssf4): Use fmsf.s on E3V5 architectures.
21748 (fnmasf4): Use fnmaf.s on E3V5 architectures.
21749 (fnmssf4): Use fnmsf.s on E3V5 architectures.
21751 2013-04-03 Jeff Law <law@redhat.com>
21753 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
21754 (lra-eliminations.o): Likewise.
21756 2013-04-03 Teresa Johnson <tejohnson@google.com>
21758 * gcov-io.c (compute_working_sets): Moved most of body of old
21759 compute_working_sets here from profile.c.
21760 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
21761 (gcov_working_set_t): Moved typedef here from basic-block.h
21762 (compute_working_set): Declare.
21763 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
21764 (get_working_sets): Renamed from compute_working_set,
21765 replace most of body with call to new compute_working_sets.
21766 (get_exec_counts): Replace call to compute_working_sets
21767 to get_working_sets.
21768 * profile.h (get_working_sets): Renamed from compute_working_set.
21769 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
21770 to get_working_sets.
21771 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
21772 * gcov-dump.c (dump_working_sets): New function.
21774 2013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
21776 * hwint.c (sext_hwi, zext_hwi): New functions.
21777 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
21778 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
21779 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
21780 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
21781 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
21782 (sext_hwi, zext_hwi): New functions.
21784 2013-04-03 Jeff Law <law@redhat.com>
21786 PR tree-optimization/56799
21787 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
21788 back test for widening conversion erroneously dropped in prior change.
21790 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21793 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
21794 instead of next_real_insn.
21796 2013-04-03 Marek Polacek <polacek@redhat.com>
21799 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
21801 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21804 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
21806 (thumb1_output_casesi): Likewise.
21807 (thumb2_output_casesi): Likewise.
21809 2013-04-03 Richard Biener <rguenther@suse.de>
21811 PR tree-optimization/56817
21812 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
21814 (tree_unroll_loops_completely_1): ... new function to manually
21815 walk the loop tree, properly defering outer loops of unrolled
21816 loops to later iterations.
21818 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
21820 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
21821 (vectorizable_load): Likewise.
21822 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
21823 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
21825 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
21827 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
21830 2013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
21832 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
21834 2013-04-03 Bin Cheng <bin.cheng@arm.com>
21836 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
21838 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
21840 PR tree-optimization/56790
21841 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
21844 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
21846 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
21848 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
21849 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
21852 2013-04-03 Jakub Jelinek <jakub@redhat.com>
21855 * tree.h (force_folding_builtin_constant_p): New decl.
21856 * builtins.c (force_folding_builtin_constant_p): New variable.
21857 (fold_builtin_constant_p): Fold immediately also if
21858 force_folding_builtin_constant_p.
21860 2013-04-03 Richard Biener <rguenther@suse.de>
21862 PR tree-optimization/56812
21863 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
21864 DRs of the same interleaving chain are independent.
21866 2013-04-02 Jason Merrill <jason@redhat.com>
21868 * gdbinit.in (pbb): Use debug fn.
21870 2013-04-02 Lawrence Crowl <crowl@google.com>
21872 * sese.h (struct ivtype_map_elt_s): Remove unused.
21873 (extern debug_ivtype_map): Remove unused.
21874 (extern eq_ivtype_map_elts): Remove unused.
21875 * sese.c (debug_ivtype_map): Removed unused.
21876 (debug_ivtype_map_1): Removed unused.
21877 (debug_ivtype_elt): Remove unused.
21878 (eq_ivtype_map_elts): Remove unused.
21881 2013-04-02 Kai Tietz <ktietz@redhat.com>
21884 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
21885 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
21886 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
21888 (legitimize_pe_coff_symbol): Likewise.
21889 (is_imported_p): New helper-function.
21890 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
21891 for Windows x64 targets.
21892 (ix86_expand_prologue): Optimize for pe-coff targets.
21893 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
21894 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
21895 medium/large code-model.
21896 (legitimize_pic_address): Likewise.
21897 (legitimize_tls_address): Likewise.
21898 (ix86_expand_call): Likewise.
21899 (x86_output_mi_thunk): Likewise.
21900 (get_dllimport_decl): Add new beimport argument.
21901 (construct_plt_address): Don't assert for x64 pe-coff targets.
21902 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
21904 (SYMBOL_FLAG_STUBVAR): New macro.
21905 (SYMBOL_REF_STUBVAR_P): Likewise.
21906 * config/i386/winnt.c (stub_list): New structure.
21907 (stub_head): New local variable.
21908 (i386_pe_record_stub): New function.
21909 (i386_pe_file_end): Emit refptr-stubs.
21911 2013-04-02 Jakub Jelinek <jakub@redhat.com>
21913 PR rtl-optimization/56745
21914 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
21915 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
21918 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
21919 and both of them are MEM_REFs, just compare first argument for
21920 equality and attempt to deal even with differing offsets.
21923 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
21924 of gimple_clobber_p to be MEM_REF.
21925 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
21926 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
21927 after gimplification.
21928 * asan.c (get_mem_ref_of_assignment): Don't instrument
21929 gimple_clobber_p stmts.
21930 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
21931 gimple_clobber_p stmt if they have MEM_REF lhs and
21932 are dead because of another gimple_clobber_p stmt.
21933 * tree-ssa-live.c (clear_unused_block_pointer): Treat
21934 gimple_clobber_p stmts like debug stmts.
21935 (remove_unused_locals): Remove clobbers with MEM_REF lhs
21936 that refer to unused VAR_DECLs or uninitialized values.
21937 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
21938 gimple_clobber_p stmts if they refer to removed parameters.
21939 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
21942 2013-04-02 Uros Bizjak <ubizjak@gmail.com>
21944 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
21945 using SWI48 mode attribute.
21947 2013-04-02 Wei Mi <wmi@google.com>
21949 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
21950 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
21951 *<rotate_insn><mode>3_mask in i386.md.
21953 2013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
21955 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
21957 2013-04-02 Richard Biener <rguenther@suse.de>
21959 PR tree-optimization/56778
21960 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
21961 Runtime alias tests are not supported for gather loads.
21962 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
21963 stmts referenced from SSA operands before updating SSA form.
21965 2013-04-02 Ian Caulfield <ian.caulfield@arm.com>
21966 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
21968 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
21969 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
21970 * config/arm/cortex-a53.md: New file.
21971 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
21972 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
21973 * config/arm/arm.c (arm_issue_rate): Likewise.
21974 * config/arm/arm-tune.md: Regenerate
21975 * config/arm/arm-tables.opt: Regenerate.
21976 * config/arm/arm-cores.def: Add cortex-a53.
21978 2013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
21980 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
21983 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
21985 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
21986 scalar load/store operations using B/H registers.
21987 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
21989 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
21991 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
21993 * config/aarch64/aarch64.c
21994 (aarch64_simd_scalar_immediate_valid_for_move): New.
21995 * config/aarch64/aarch64-protos.h
21996 (aarch64_simd_scalar_immediate_valid_for_move): New.
21997 * config/aarch64/constraints.md (Dh, Dq): New.
21998 * config/aarch64/iterators.md (hq): New.
22000 2013-04-02 Eric Botcazou <ebotcazou@adacore.com>
22002 * reorg.c (get_branch_condition): Deal with conditional returns.
22003 (fill_simple_delay_slots): Remove dead code dealing with jumps.
22005 2013-04-01 Wei Mi <wmi@google.com>
22007 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
22008 Truncate operand 2 using %b asm operand modifier.
22009 (*<shift_insn><mode>3_mask): Ditto.
22010 (*<rotate_insn><mode>3_mask): Ditto.
22012 2013-04-01 Steven Bosscher <steven@gcc.gnu.org>
22014 PR middle-end/56798
22015 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
22017 2013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
22019 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
22021 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
22023 2013-03-30 Lawrence Crowl <crowl@google.com>
22025 * dse.c (clear_alias_sets): Remove never set.
22026 (disqualified_clear_alias_sets): Remove never set.
22027 (clear_alias_mode_pool): Remove never set.
22028 (dse_step0): Remove condition that is never true.
22029 (canon_address): Remove condition that is never true.
22030 (dse_step7): Remove condition that is never true.
22031 (rest_of_handle_dse): Remove condition that is never true.
22032 (rest_of_handle_dse::did_global): Remove never read from above.
22033 (dse_step2_spill): Remove never called from above.
22034 (dse_step5_spill): Remove never called from above.
22036 2013-03-30 Steven Bosscher <steven@gcc.gnu.org>
22038 * doc/md.texi (Standard Names) <casesi>: Update documentation for
22039 JUMP_TABLE_DATA changes.
22040 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
22041 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
22042 (Insns) <jump_table_data>: New entry.
22043 * doc/tm.texi: Regenerate.
22045 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
22047 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
22048 for table jump at the end of a basic block using tablejump_p.
22049 * targhooks.c (default_invalid_within_doloop): Likewise.
22050 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
22051 target hook implementation that is identical to the default hook.
22052 (rs6000_invalid_within_doloop): Remove.
22054 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
22055 unused variable from tablejump_p call.
22057 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
22058 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
22059 (INSN_DELETED_P): Likewise.
22060 (emit_jump_table_data): New prototype.
22061 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
22062 after 4th as unused.
22063 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
22064 * sched-vis.c (print_insn): Likewise.
22065 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
22066 insn for compatibility with back ends that use next_active_insn to
22067 identify jump table data.
22068 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
22069 (remove_insn): Likewise.
22070 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
22072 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
22073 (emit_jump_table_data): New function.
22075 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
22076 basic block, a JUMP_TABLE_DATA never is.
22077 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
22078 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
22079 off from code handling real insns.
22080 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
22081 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
22082 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
22083 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
22084 is not a NONDEBUG_INSN_P.
22085 * ira-costs.c (scan_one_insn): Likewise.
22086 * jump.c (mark_all_labels): Likewise.
22087 (mark_jump_label_1): Likewise.
22088 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
22089 * lra.c (get_insn_freq): Expect all insns reaching here to be in
22091 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
22092 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
22093 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
22094 JUMP_TABLE_DATA_P insns.
22095 (calculate_elim_costs_all_insns): Likewise.
22096 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
22097 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
22098 (delete_output_reload): Code style fixups.
22099 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
22100 insn flags on this non-insn.
22101 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
22102 as scheduling barriers, for pre-change compatibility.
22103 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
22104 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
22106 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
22107 redundant JUMP_TABLE_DATA_P test.
22108 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
22109 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
22110 (frv_for_each_packet): Likewise.
22111 * config/i386/i386.c (min_insn_size): Likewise.
22112 (ix86_avoid_jump_mispredicts): Likewise.
22113 * config/m32r/m32r.c (m32r_is_insn): Likewise.
22114 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
22115 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
22116 (mips16_insn_length): Robustify.
22117 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
22118 (mips16_split_long_branches): Likewise.
22119 * config/pa/pa.c (pa_combine_instructions): Likewise.
22120 * config/rs6000/rs6000.c (get_next_active_insn): Treat
22121 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
22122 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
22123 as contributing to pool range lengths.
22124 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
22125 Remove redundant JUMP_TABLE_DATA_P test.
22126 (sh_loop_align): Likewise.
22127 (split_branches): Likewise.
22128 (sh_insn_length_adjustment): Likewise.
22129 * config/spu/spu.c (get_branch_target): Likewise.
22131 2013-03-29 Jan Hubicka <jh@suse.cz>
22133 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
22134 gcov streaming; stream hot bb threshold to ltrans.
22135 * predict.c (get_hot_bb_threshold): Break out from ....
22136 (maybe_hot_count_p): ... here.
22137 (set_hot_bb_threshold): New function.
22138 * lto-section-in.c (lto_section_name): Add profile.
22139 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
22140 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
22141 and data-streamer.h
22142 (histogram_entry): New structure.
22143 (histogram, histogram_pool): New global vars.
22144 (histogram_hash): New structure.
22145 (histogram_hash::hash): New method.
22146 (histogram_hash::equal): Likewise.
22147 (account_time_size): New function.
22148 (cmp_counts): New function.
22149 (dump_histogram): New function.
22150 (ipa_profile_generate_summary): New function.
22151 (ipa_profile_write_summary): New function.
22152 (ipa_profile_read_summary): New function.
22153 (ipa_profile): Decide on threshold.
22154 (pass_ipa_profile): Add ipa_profile_write_summary and
22155 ipa_profile_read_summary.
22156 * Makefile.in (ipa.o): Update dependencies.
22157 * lto-streamer.h (LTO_section_ipa_profile): New section.
22159 2013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
22161 * tree.h (VAR_P): New.
22163 2013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
22166 * doc/invoke.texi ([-fwhole-program]): Fix typo.
22168 2013-03-29 Steven Bosscher <steven@gcc.gnu.org>
22170 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
22171 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
22172 (control_flow_insn_p): Likewise.
22173 * cfgrtl.c (duplicate_insn_chain): Likewise.
22174 * final.c (get_attr_length_1): Likewise.
22175 (shorten_branches): Likewise.
22176 (final_scan_insn): Likewise.
22177 * function.c (instantiate_virtual_regs): Likewise.
22178 * gcse.c (insert_insn_end_basic_block): Likewise.
22179 * ira-costs.c (scan_one_insn): Likewise.
22180 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
22181 * lra.c (check_rtl): Likewise.
22182 * reload1.c (elimination_costs_in_insn): Likewise.
22183 * reorg.c (follow_jumps): Likewise.
22185 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
22186 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
22187 (thumb_far_jump_used_p): Likewise.
22188 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
22189 (workaround_speculation): Likewise.
22190 (add_sched_insns_for_speculation): Likewise.
22191 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
22192 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
22193 (frv_for_each_packet): Likewise.
22194 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
22195 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
22196 (final_emit_insn_group_barriers): Likewise.
22197 * config/m32r/m32r.c (m32r_is_insn): Likewise.
22198 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
22199 (mips16_insn_length): Likewise.
22200 * config/pa/pa.c (pa_reorg): Likewise.
22201 (pa_combine_instructions): Likewise.
22202 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
22203 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
22204 (sh_reorg): Likewise.
22205 (split_branches): Likewise.
22206 * config/spu/spu.c (get_branch_target): Likewise.
22208 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
22211 2013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
22213 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
22214 Fix declaration name.
22216 2013-03-28 Lawrence Crowl <crowl@google.com>
22218 * graphds.h (struct graph.indicies): Remove unused.
22219 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
22220 (SCOP_ORIGINAL_PDDRS): Remove unused.
22221 * sese.h (extern insert_loop_close_phis): Removed unused.
22222 (extern insert_guard_phis): Removed unused.
22223 (extern ivtype_map_elt_info): Removed unused.
22224 (new_ivtype_map_elt): Removed unused.
22225 * sese.c (ivtype_map_elt_info): Removed unused.
22227 2013-03-28 Lawrence Crowl <crowl@google.com>
22229 * Makefile.in: Add several missing include dependences.
22231 (test-dump.o): New. This object is not added to any executable,
22232 but is present for ad-hoc testing.
22234 (debug (const bitmap_head_def &)): New.
22235 (debug (const bitmap_head_def *)): New.
22237 (extern debug (const bitmap_head_def &)): New.
22238 (extern debug (const bitmap_head_def *)): New.
22240 (debug (edge_def &)): New.
22241 (debug (edge_def *)): New.
22243 (debug (basic_block_def &)): New.
22244 (debug (basic_block_def *)): New.
22246 (dump_node (const_tree, int, FILE *)): Correct source file.
22248 (debug (die_struct &)): New.
22249 (debug (die_struct *)): New.
22251 (extern debug (die_struct &)): New.
22252 (extern debug (die_struct *)): New.
22253 * gimple-pretty-print.c
22254 (debug (gimple_statement_d &)): New.
22255 (debug (gimple_statement_d *)): New.
22256 * gimple-pretty-print.h
22257 (extern debug (gimple_statement_d &)): New.
22258 (extern debug (gimple_statement_d *)): New.
22260 (debug (ira_allocno_copy &)): New.
22261 (debug (ira_allocno_copy *)): New.
22262 (debug (ira_allocno &)): New.
22263 (debug (ira_allocno *)): New.
22265 (extern debug (ira_allocno_copy &)): New.
22266 (extern debug (ira_allocno_copy *)): New.
22267 (extern debug (ira_allocno &)): New.
22268 (extern debug (ira_allocno *)): New.
22270 (debug (live_range &)): New.
22271 (debug (live_range *)): New.
22273 (debug (lra_live_range &)): New.
22274 (debug (lra_live_range *)): New.
22276 (debug (lra_live_range &)): New.
22277 (debug (lra_live_range *)): New.
22279 (debug (omega_pb_d &)): New.
22280 (debug (omega_pb_d *)): New.
22282 (extern debug (omega_pb_d &)): New.
22283 (extern debug (omega_pb_d *)): New.
22285 (debug (const rtx_def &)): New.
22286 (debug (const rtx_def *)): New.
22288 (debug_tree (tree): Move within file.
22289 (debug_raw (const tree_node &)): New.
22290 (debug_raw (const tree_node *)): New.
22291 (dump_tree_via_hooks (const tree_node *, int)): New.
22292 (debug (const tree_node &)): New.
22293 (debug (const tree_node *)): New.
22294 (debug_verbose (const tree_node &)): New.
22295 (debug_verbose (const tree_node *)): New.
22296 (debug_head (const tree_node &)): New.
22297 (debug_head (const tree_node *)): New.
22298 (debug_body (const tree_node &)): New.
22299 (debug_body (const tree_node *)): New.
22300 (debug_vec_tree (tree): Move and reimplement in terms of dump.
22301 (debug (vec<tree, va_gc> &)): New.
22302 (debug (vec<tree, va_gc> *)): New.
22304 (extern debug (const rtx_def &)): New.
22305 (extern debug (const rtx_def *)): New.
22307 (debug_raw (simple_bitmap_def &)): New.
22308 (debug_raw (simple_bitmap_def *)): New.
22309 (debug (simple_bitmap_def &)): New.
22310 (debug (simple_bitmap_def *)): New.
22312 (extern debug (simple_bitmap_def &)): New.
22313 (extern debug (simple_bitmap_def *)): New.
22314 (extern debug_raw (simple_bitmap_def &)): New.
22315 (extern debug_raw (simple_bitmap_def *)): New.
22317 (debug (vinsn_def &)): New.
22318 (debug (vinsn_def *)): New.
22319 (debug_verbose (vinsn_def &)): New.
22320 (debug_verbose (vinsn_def *)): New.
22321 (debug (expr_def &)): New.
22322 (debug (expr_def *)): New.
22323 (debug_verbose (expr_def &)): New.
22324 (debug_verbose (expr_def *)): New.
22325 (debug (vec<rtx> &)): New.
22326 (debug (vec<rtx> *)): New.
22328 (extern debug (vinsn_def &)): New.
22329 (extern debug (vinsn_def *)): New.
22330 (extern debug_verbose (vinsn_def &)): New.
22331 (extern debug_verbose (vinsn_def *)): New.
22332 (extern debug (expr_def &)): New.
22333 (extern debug (expr_def *)): New.
22334 (extern debug_verbose (expr_def &)): New.
22335 (extern debug_verbose (expr_def *)): New.
22336 (extern debug (vec<rtx> &)): New.
22337 (extern debug (vec<rtx> *)): New.
22339 (_list_iter_cond_expr): Make inline instead of static.
22341 (debug (sreal &)): New.
22342 (debug (sreal *)): New.
22344 (extern debug (sreal &)): New.
22345 (extern debug (sreal *)): New.
22347 (extern debug_raw (const tree_node &)): New.
22348 (extern debug_raw (const tree_node *)): New.
22349 (extern debug (const tree_node &)): New.
22350 (extern debug (const tree_node *)): New.
22351 (extern debug_verbose (const tree_node &)): New.
22352 (extern debug_verbose (const tree_node *)): New.
22353 (extern debug_head (const tree_node &)): New.
22354 (extern debug_head (const tree_node *)): New.
22355 (extern debug_body (const tree_node &)): New.
22356 (extern debug_body (const tree_node *)): New.
22357 (extern debug (vec<tree, va_gc> &)): New.
22358 (extern debug (vec<tree, va_gc> *)): New.
22360 (debug (struct loop &)): New.
22361 (debug (struct loop *)): New.
22362 (debug_verbose (struct loop &)): New.
22363 (debug_verbose (struct loop *)): New.
22364 * tree-dump.c: Add header dependence.
22366 (extern debug (struct loop &)): New.
22367 (extern debug (struct loop *)): New.
22368 (extern debug_verbose (struct loop &)): New.
22369 (extern debug_verbose (struct loop *)): New.
22371 (debug (data_reference &)): New.
22372 (debug (data_reference *)): New.
22373 (debug (vec<data_reference_p> &)): New.
22374 (debug (vec<data_reference_p> *)): New.
22375 (debug (vec<ddr_p> &)): New.
22376 (debug (vec<ddr_p> *)): New.
22378 (extern debug (data_reference &)): New.
22379 (extern debug (data_reference *)): New.
22380 (extern debug (vec<data_reference_p> &)): New.
22381 (extern debug (vec<data_reference_p> *)): New.
22382 (extern debug (vec<ddr_p> &)): New.
22383 (extern debug (vec<ddr_p> *)): New.
22385 (debug (pt_solution &)): New.
22386 (debug (pt_solution *)): New.
22388 (extern debug (pt_solution &)): New.
22389 (extern debug (pt_solution *)): New.
22391 (debug (_var_map &)): New.
22392 (debug (_var_map *)): New.
22393 (debug (tree_live_info_d &)): New.
22394 (debug (tree_live_info_d *)): New.
22396 (extern debug (_var_map &)): New.
22397 (extern debug (_var_map *)): New.
22398 (extern debug (tree_live_info_d &)): New.
22399 (extern debug (tree_live_info_d *)): New.
22401 2013-03-28 Jan Hubicka <jh@suse.cz>
22403 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
22405 2013-03-28 Ian Bolton <ian.bolton@arm.com>
22407 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
22408 record only when desired or required.
22410 2013-03-28 Uros Bizjak <ubizjak@gmail.com>
22412 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
22413 *vec_extractv2di_1_rex64. Use x64 isa attribute.
22415 2013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22417 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
22418 (*andsi3_compare0_uxtw): New pattern.
22419 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
22420 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
22422 2013-03-28 Jan Hubicka <jh@suse.cz>
22424 * data-streamer-in.c (streamer_read_gcov_count): New function.
22425 * gimple-streamer-out.c: Include value-prof.h.
22426 (output_gimple_stmt): Output histogram.
22427 (output_bb): Use streamer_write_gcov_count.
22428 * value-prof.c: Include data-streamer.h
22429 (dump_histogram_value): Add HIST_TYPE_MAX.
22430 (stream_out_histogram_value): New function.
22431 (stream_in_histogram_value): New function.
22432 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
22433 (stream_out_histogram_value, stream_in_histogram_value): Declare.
22434 * data-streamer-out.c (streamer_write_gcov_count): New function.
22435 (streamer_write_gcov_count_stream): New function.
22436 * lto-cgraph.c (lto_output_edge): Update counter streaming.
22437 (lto_output_node): Likewise.
22438 (input_node, input_edge): Likewise.
22439 * lto-streamer-out.c (output_cfg): Update streaming.
22440 * lto-streamer-in.c (input_cfg): Likewise.
22441 * data-streamer.h (streamer_write_gcov_count,
22442 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
22443 * gimple-streamer-in.c: Include value-prof.h
22444 (input_gimple_stmt): Input histograms.
22445 (input_bb): Update profile streaming.
22447 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
22449 * genmodes.c (emit_max_int): New function.
22450 (emit_insn_modes_h): Added call to emit_max_function.
22451 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
22453 * machmode.def: Fixed comment.
22455 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
22457 * combine.c (try_combine): Removed useless assert.
22458 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
22460 2013-03-28 Marek Polacek <polacek@redhat.com>
22461 Richard Biener <rguenther@suse.de>
22463 PR tree-optimization/56695
22464 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
22465 build signed result of a vector comparison.
22466 * tree-cfg.c (verify_gimple_comparison): Check that a result
22467 of a vector comparison has signed type.
22469 2013-03-28 Richard Biener <rguenther@suse.de>
22471 PR tree-optimization/37021
22472 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
22473 do not restrict gaps between groups.
22474 * tree-vect-stmts.c (vectorizable_load): Properly account for
22475 a gap between groups.
22477 2013-03-28 Eric Botcazou <ebotcazou@adacore.com>
22479 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
22480 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
22483 2013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
22485 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
22486 * doc/extend.texi (Named Address Spaces): Ditto.
22487 (Variable Attributes): Ditto.
22489 2013-03-27 Kai Tietz <ktietz@redhat.com>
22491 * config.build: Add support for cygwin x64 target.
22492 * config.gcc: Likewise.
22493 * config.host: Likewise.
22494 * configure.ac: Likewise
22495 * configure: Regenerated.
22497 2013-03-27 Kai Tietz <ktietz@redhat.com>
22499 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
22500 * config/i386/t-cygwin-w64: New file.
22501 * config/i386/cygwin-w64.h: New file.
22502 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
22503 and add support for x64-cygwin target.
22504 (CPP_SPEC): Likewise.
22505 (CXX_WRAP_SPEC_LIST): Undefine before define.
22506 (LIBGCJ_SONAME): Use 15 as version.
22508 2013-03-27 Richard Biener <rguenther@suse.de>
22510 PR tree-optimization/56716
22511 * tree-ssa-structalias.c (perform_var_substitution): Adjust
22512 dumping for ref nodes.
22514 2013-03-27 Martin Jambor <mjambor@suse.cz>
22516 PR tree-optimization/55334
22517 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
22518 restricted pointers to arrays.
22520 2013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
22522 * Makefile.in (.SUFFIXES): Add .cc.
22523 (.c.o): Apply same recipe for implicit rule .cc.o.
22525 2013-03-27 Richard Biener <rguenther@suse.de>
22527 PR tree-optimization/37021
22528 * tree-vect-data-refs.c (vect_check_strided_load): Allow
22529 REALPART/IMAGPART_EXPRs around the supported refs.
22530 * tree-ssa-structalias.c (find_func_aliases): Assume that
22531 floating-point values are not used to transfer pointers.
22533 2013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
22535 * target.def (TARGET_HAS_IFUNC_P): New target hook.
22536 * doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
22537 * doc/tm.texi: Regenerate.
22538 * targhooks.h (default_has_ifunc_p): New.
22539 * targhooks.c (default_has_ifunc_p): Ditto.
22540 * config/linux-protos.h: New file.
22541 * config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of this
22542 hook for linux which disables support of indirect functions in android.
22543 * config/linux-android.c: New file.
22544 * config/t-linux-android.c: Ditto.
22545 * config.gcc: Added new object file linux-android.o.
22546 * config/i386/i386.c (ix86_get_function_versions_dispatcher):
22547 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
22548 * varasm.c (do_assemble_alias): Likewise.
22549 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
22550 doesn't support indirect functions.
22551 * configure: Regenerate.
22553 2013-03-27 Bin Cheng <bin.cheng@arm.com>
22556 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
22557 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
22560 2013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
22562 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
22564 2013-03-27 Terry Guo <terry.guo@arm.com>
22566 * config/arm/arm-cores.def: Added core cortex-r7.
22567 * config/arm/arm-tune.md: Regenerated.
22568 * config/arm/arm-tables.opt: Regenerated.
22569 * doc/invoke.texi: Added entry for core cortex-r7.
22571 2013-03-27 Walter Lee <walt@tilera.com>
22573 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
22574 double-decrement of next_scratch_regno.
22576 2013-03-27 Walter Lee <walt@tilera.com>
22578 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
22580 (insn_v1mulus): Ditto.
22581 (insn_v2muls): Ditto.
22583 2013-03-27 Walter Lee <walt@tilera.com>
22585 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
22586 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
22588 2013-03-27 Walter Lee <walt@tilera.com>
22590 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
22591 (*sibcall_value): Ditto.
22593 2013-03-27 Walter Lee <walt@tilera.com>
22595 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
22596 (insn_mnz_v8qi): ... this ...
22597 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
22599 (insn_v<n>mnz): Replaced by ...
22600 (insn_v1mnz): ... this ...
22601 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
22603 (insn_mz_<mode>): Replaced by ...
22604 (insn_mz_v8qi): ... this ...
22605 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
22607 (insn_v<n>mz): Replaced by ...
22608 (insn_v1mz): ... this ...
22609 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
22612 2013-03-26 Eric Botcazou <ebotcazou@adacore.com>
22614 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
22616 2013-03-26 Roland McGrath <mcgrathr@google.com>
22618 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
22619 than fprintf with a non-constant, non-format string.
22621 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
22623 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
22624 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
22625 operand 0 predicate.
22626 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
22627 attribute. Use general_x64nomem_operand as operand 1 predicate.
22628 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
22629 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
22630 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
22631 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
22632 (mov<mode>_insv_1): Remove expander. Merge insn with
22633 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
22634 Use general_x64nomem_operand as operand 1 predicate.
22635 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
22636 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
22637 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
22638 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
22639 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
22640 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
22641 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
22642 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
22643 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
22644 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
22645 (general_x64nomem_operand): Ditto.
22647 2013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
22649 * config/rtems.opt: Add -pthread option.
22651 2013-03-26 Richard Biener <rguenther@suse.de>
22653 * alias.c (find_base_term): Avoid redundant and not used recursion.
22654 (base_alias_check): Get the initial base term from the caller.
22655 (true_dependence_1): Compute and pass base terms to base_alias_check.
22656 (write_dependence_p): Likewise.
22657 (may_alias_p): Likewise.
22659 2013-03-26 Sofiane Naci <sofiane.naci@arm.com>
22661 * config/aarch64/aarch64.c (aarch64_classify_address): Support
22662 PC-relative load in SI modes and above only.
22664 2013-03-26 Xinyu Qi <xyqi@marvell.com>
22666 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
22667 * config/arm/iwmmxt.md (WCGR0): Update.
22668 (WCGR1, WCGR2, WCGR3): Likewise.
22670 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
22672 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
22673 Use x64 and nox64 isa attributes.
22675 2013-03-26 Richard Biener <rguenther@suse.de>
22677 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
22678 alignment computations and rely on get_object_alignment_1
22679 for the !TYPE_P case.
22680 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
22682 2013-03-26 Walter Lee <walt@tilera.com>
22684 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
22685 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
22687 2013-03-25 Jeff Law <law@redhat.com>
22689 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
22690 check for INTEGRAL_TYPE_P that was missing due to checking in
22691 wrong version of prior patch.
22693 2013-03-25 Walter Lee <walt@tilera.com>
22695 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
22696 TILEGX_INSN_SHUFFLEBYTES1.
22697 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
22699 (tilegx_builtins): Ditto.
22700 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
22702 2013-03-25 Walter Lee <walt@tilera.com>
22704 * config/tilegx/tilegx.md (floatsisf2): New pattern.
22705 (floatunssisf2): New pattern.
22706 (floatsidf2): New pattern.
22707 (floatunssidf2): New pattern.
22709 2013-03-25 Walter Lee <walt@tilera.com>
22711 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
22712 tests for constraint J, K, N, P.
22714 2013-03-25 Walter Lee <walt@tilera.com>
22716 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
22717 Use indirect/pcrel encoding.
22718 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
22721 2013-03-25 Steve Ellcey <sellcey@mips.com>
22723 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
22724 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
22725 * config/mips/mips.c (mips_option_override): Set IMADD default.
22726 * config/mips/mips.h (PTF_AVOID_IMADD): New.
22727 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
22728 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
22729 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
22730 * doc/invoke.texi (-mimadd/-mno-imadd): New.
22732 2013-03-25 Jeff Law <law@redhat.com>
22734 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
22735 slightly to avoid creating and folding useless trees. Simplify
22736 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
22738 2013-03-25 Uros Bizjak <ubizjak@gmail.com>
22740 * config/i386/i386.md (*zero_extendsidi2): Merge with
22741 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
22742 * config/i386/predicates.md (x86_64_zext_operand): Rename from
22743 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
22744 targets. Clarify comment.
22746 2013-03-25 Martin Jambor <mjambor@suse.cz>
22748 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
22749 pass-through jump functions differently.
22750 (ipa_read_jump_function): Likewise. Also use setter functions to set
22753 2013-03-25 Martin Jambor <mjambor@suse.cz>
22755 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
22756 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
22758 (ipa_get_indirect_edge_target): New function.
22759 (devirtualization_time_bonus): New parameter known_aggs, pass it to
22760 ipa_get_indirect_edge_target. Update all callers.
22761 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
22762 ipa_get_indirect_edge_target_1 instead of calling
22763 ipa_get_indirect_edge_target.
22764 (create_specialized_node): Pass aggvlas to
22765 ipcp_discover_new_direct_edges.
22767 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22769 * config/arm/arm.md (f_sels, f_seld): New types.
22770 (*cmov<mode>): New pattern.
22771 * config/arm/predicates.md (arm_vsel_comparison_operator): New
22774 2013-03-25 Kai Tietz <ktietz@redhat.com>
22776 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
22777 POSIX-printf for mingw-hosted builds.
22779 2013-03-25 Richard Biener <rguenther@suse.de>
22781 PR middle-end/56694
22782 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
22783 must-not-throw stmt location.
22785 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22787 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
22788 Emit load-acquire versions when acq is true.
22789 (arm_emit_store_exclusive): Add rel parameter.
22790 Emit store-release versions when rel is true.
22791 (arm_split_compare_and_swap): Use acquire-release instructions
22793 of barriers when appropriate.
22794 (arm_split_atomic_op): Likewise.
22795 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
22796 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
22797 (VUNSPEC_SLX): Likewise.
22798 (VUNSPEC_LDA): Likewise.
22799 (VUNSPEC_STL): Likewise.
22800 * config/arm/sync.md (atomic_load<mode>): New pattern.
22801 (atomic_store<mode>): Likewise.
22802 (arm_load_acquire_exclusive<mode>): Likewise.
22803 (arm_load_acquire_exclusivesi): Likewise.
22804 (arm_load_acquire_exclusivedi): Likewise.
22805 (arm_store_release_exclusive<mode>): Likewise.
22807 2013-03-25 Catherine Moore <clm@codesourcery.com>
22809 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
22810 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
22811 * config/mip/predicates.md (lwsp_swsp_operand,
22812 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
22813 sb16_operand, db4_operand, db7_operand, ib3_operand,
22814 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
22815 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
22816 andi16_operand): New predicates.
22817 * config/mips/mips.md (compression): New attribute.
22818 (enabled): New attribute.
22819 (length): Consider compression in computing length.
22820 (shift_compression): New code attribute.
22821 (*add<mode>3): New operands. Record compression.
22822 (sub<mode>3): Likewise.
22823 (one_cmpl<mode>2): Likewise.
22824 (*and<mode>3): Likewise.
22825 (*ior<mode>3): Likewise.
22826 (unnamed pattern for xor): Likewise.
22827 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
22828 (*<optab><mode>3): Likewise.
22829 (*mov<mode>_internal: Likewise.
22830 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
22831 (mips_unsigned_immediate_p): New.
22832 (umips_lwsp_swsp_address_p): New.
22833 (m16_based_address_p): New.
22834 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
22835 (mips_unsigned_immediate_p): New prototype.
22836 (lwsp_swsp_address_p): New prototype.
22837 (m16_based_address_p): New prototype.
22838 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
22839 (mips_signed_immediate_p): New function.
22840 (m16_based_address_p): New function.
22841 (lwsp_swsp_address_p): New function.
22842 (mips_print_operand_punctuation): Recognize short delay slot insns
22843 for microMIPS.add<mode>3"
22845 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22848 * config/arm/iterators.md (v_cmp_result): New mode attribute.
22849 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
22851 2013-03-25 Richard Biener <rguenther@suse.de>
22853 PR tree-optimization/56689
22854 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
22857 2013-03-25 Richard Biener <rguenther@suse.de>
22859 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
22861 (memory_references): Likewise.
22862 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
22863 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
22864 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
22865 (gather_mem_refs_in_loops): Fold into ...
22866 (analyze_memory_references): ... this. Move initialization
22867 to tree_ssa_lim_initialize.
22868 (fill_always_executed_in): Rename to ...
22869 (fill_always_executed_in_1): ... this.
22870 (fill_always_executed_in): Move contains_call computation to
22871 this new function from ...
22872 (tree_ssa_lim_initialize): ... here.
22873 (tree_ssa_lim): Call fill_always_executed_in.
22875 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
22877 * postreload.c (reload_combine): Fix code detecting returns.
22879 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
22881 * function.c (emit_use_return_register_into_block): On cc0 targets,
22882 do not emit the sequence between cc0 setter and user.
22884 2013-03-25 Kai Tietz <ktietz@redhat.com>
22886 * config/i386/predicates.md (local_symbolic_operand): Interpret
22887 dll-imported symbols as none-local.
22889 2013-03-25 Richard Biener <rguenther@suse.de>
22891 * tree-ssa-loop-im.c (struct depend): Remove.
22892 (struct lim_aux_data): Make depends a vec of gimples.
22893 (free_lim_aux_data): Adjust.
22894 (add_dependency): Likewise.
22895 (set_level): Likewise.
22897 2013-03-25 Richard Biener <rguenther@suse.de>
22899 PR middle-end/56434
22900 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
22901 the pointer returned by calls with ECF_MALLOC set.
22903 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
22905 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
22907 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
22909 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
22910 using MMXMODE mode iterator.
22911 (*move<mode>_internal): Merge with *movv2sf_internal and
22912 *movv2sf_internal_rex64 using MMXMODE mode iterator.
22914 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
22916 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
22917 (record_last_mem_set_info): Likewise.
22919 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
22920 of XNEWVEC followed by memset.
22921 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
22923 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
22925 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
22926 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
22927 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
22928 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
22929 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
22930 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
22931 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
22932 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
22933 BARRIER_P instead of GET_CODE.
22935 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
22937 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
22938 inaccuracy in the probing code.
22940 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
22941 (ctrapdi4): Likewise.
22943 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
22945 * calls.c (expand_call): Add missing guard to code handling return
22946 of non-BLKmode structures in MSB.
22947 * function.c (expand_function_end): Likewise.
22949 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
22951 * combine.c (try_combine): Adjust comment. Do not add the set of
22952 insn #0 if the destination indirectly is set or dies in insn #2.
22953 Tidy up code to distribute a new note.
22955 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
22957 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
22958 also for alternatives 16 and 17.
22960 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
22962 * config/i386/sse.md (*mov<mode>_internal): Merge with
22963 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
22964 Emit insn template depending on type attribute. Use
22965 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
22966 movd instead of movq mnemonic for interunit moves. Rewrite mode
22967 attribute calculation. Remove unit attribute calculation.
22968 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
22969 Set prefix_data16 attribute for DImode ssemov types.
22970 Use Ym instead of y for SSE-MMX conversion alternatives.
22971 Reorder operand constraints.
22973 2013-03-22 Steven Bosscher <steven@gcc.gnu.org>
22975 * df.h (df_insn_delete): Adjust prototype.
22976 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
22977 and let it decide whether mark the basic block dirty.
22978 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
22979 * df-scan.c (df_insn_info_delete): New helper function, split
22980 off from df_insn_delete.
22981 (df_scan_free_bb_info): Use it.
22982 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
22984 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
22985 that the insn is actually an insn and it has a non-NULL basic block.
22986 Do not mark basic block dirty if only deleting a DEBUG_INSN.
22988 2013-03-22 Richard Biener <rguenther@suse.de>
22990 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
22992 (mem_ref_alloc): Do not allocate them.
22993 (refs_independent_p): Do not query or maintain a cache.
22995 2013-03-22 Richard Biener <rguenther@suse.de>
22997 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
22998 (gather_mem_refs_in_loops): Do not compute it.
22999 (analyze_memory_references): Do not allocate it.
23000 (tree_ssa_lim_finalize): Do not free it.
23001 (for_all_locs_in_loop): Do not query all_refs_in_loop.
23003 2013-03-22 Richard Biener <rguenther@suse.de>
23005 * is-a.h (as_a): Use gcc_checking_assert.
23007 2013-03-22 Ian Bolton <ian.bolton@arm.com>
23009 * config/aarch64/aarch64.c (aarch64_print_operand): New
23010 format specifier for printing a constant in hex.
23011 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
23012 format specifier for printing second operand.
23014 2013-03-22 Richard Biener <rguenther@suse.de>
23016 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
23018 (gather_mem_refs_in_loops): Perform store accumulation here.
23019 (create_vop_ref_mapping_loop): Remove.
23020 (create_vop_ref_mapping): Likewise.
23021 (analyze_memory_references): Initialize refs_stored_in_loop.
23022 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
23023 (record_indep_loop): Remove.
23024 (record_dep_loop): New function.
23025 (ref_indep_loop_p_1): Adjust to only walk over references
23026 in the loop, not its subloops.
23027 (ref_indep_loop_p): Rename to ...
23028 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
23029 maintaining a more fine-grained cache.
23030 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
23031 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
23033 2013-03-22 Richard Biener <rguenther@suse.de>
23035 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
23036 (struct mem_ref): Make accesses_in_loop a vec of a vec of
23037 aggregate mem_ref_loc.
23038 (free_mem_ref_locs): Inline into ...
23039 (memref_free): ... this and adjust.
23040 (mem_ref_alloc): Adjust.
23041 (mem_ref_locs_alloc): Remove.
23042 (record_mem_ref_loc): Adjust.
23043 (get_all_locs_in_loop): Rewrite into ...
23044 (for_all_locs_in_loop): ... this iterator.
23045 (rewrite_mem_ref_loc): New functor.
23046 (rewrite_mem_refs): Use for_all_locs_in_loop.
23047 (sm_set_flag_if_changed): New functor.
23048 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
23049 (ref_always_accessed): New functor.
23050 (ref_always_accessed_p): Use for_all_locs_in_loop.
23052 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
23054 * tree-pass.h (PROP_gimple_lvec): New.
23055 * passes.c (dump_properties): Handle PROP_gimple_lvec.
23056 (init_optimization_passes): Move pass_lower_vector.
23057 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
23059 (pass_lower_vector): Provide PROP_gimple_lvec.
23060 (pass_lower_vector_ssa): Likewise.
23061 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
23063 2013-03-21 Mark Wielaard <mjw@redhat.com>
23065 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
23067 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
23069 * config/i386/i386.md (*movdi_internal): Disparage slightly
23070 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
23071 conversion alternatives.
23073 2013-03-21 Jakub Jelinek <jakub@redhat.com>
23075 PR middle-end/48087
23076 * diagnostic.def (DK_WERROR): New kind.
23077 * diagnostic.h (werrorcount): Define.
23078 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
23079 promoted to DK_ERROR, increment DK_WERROR counter instead of
23081 * toplev.c (toplev_main): Call print_ignored_options even if
23082 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
23083 even if just werrorcount is non-zero.
23086 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
23088 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
23089 (string_cst_pool_decl): New function.
23090 (optimize_one_addr_into_implicit_ptr): New function.
23091 (resolve_addr_in_expr): Optimize DWARF location expression
23092 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
23093 which doesn't live in memory, but has DW_AT_location or
23094 DW_AT_const_value, or refers to a string literal, into
23095 DW_OP_GNU_implicit_pointer.
23096 (optimize_location_into_implicit_ptr): New function.
23097 (resolve_addr): If removing DW_AT_location of a variable because
23098 it was DW_OP_addr of address of the variable, but the variable doesn't
23099 live in memory, try to emit const value attribute for the initializer.
23101 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
23103 * tree.h (VECTOR_TYPE_P): New macro.
23104 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
23105 TYPE_MODE): Use it.
23106 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
23107 VEC_COND_EXPR cannot be lvalues.
23108 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
23110 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
23112 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
23113 Restrict the transformation to equal modes.
23115 2013-03-21 Richard Biener <rguenther@suse.de>
23117 PR tree-optimization/39326
23118 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
23119 (MEM_ANALYZABLE): Adjust.
23120 (record_mem_ref_loc): Move bitmap ops ...
23121 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
23122 unanalyzable refs, do not record locations for it.
23123 (analyze_memory_references): Allocate ref zero as shared
23125 (refs_independent_p): Do not test for unanalyzed mems here.
23126 (ref_indep_loop_p_1): Special-case disambiguation against
23127 the unanalyzed ref.
23128 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
23130 2013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
23132 * config/arm/arm-protos.h (tune_params): Add
23133 prefer_neon_for_64bits field.
23134 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
23135 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
23136 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
23137 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
23138 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
23139 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
23140 (arm_option_override): Handle -mneon-for-64bits new option.
23141 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
23142 (prefer_neon_for_64bits): Declare new variable.
23143 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
23144 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
23145 (arch_enabled): Handle new arch types. Remove support for onlya8
23147 (one_cmpldi2): Use new arch names.
23148 (zero_extend<mode>di2, extend<mode>di2): Ditto.
23149 * config/arm/arm.opt (mneon-for-64bits): Add option.
23150 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
23151 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
23152 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
23154 * doc/invoke.texi (-mneon-for-64bits): Document.
23156 2013-03-21 Richard Biener <rguenther@suse.de>
23158 PR tree-optimization/39326
23159 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
23160 (sort_bbs_in_loop_postorder_cmp): New function.
23161 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
23163 2013-03-21 Richard Biener <rguenther@suse.de>
23165 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
23166 (vect_insert_into_interleaving_chain): Likewise.
23167 (vect_drs_dependent_in_basic_block): Inline ...
23168 (vect_slp_analyze_data_ref_dependence): ... here. New function,
23170 (vect_analyze_data_ref_dependence): ... here. Simplify.
23171 (vect_check_interleaving): Simplify.
23172 (vect_analyze_data_ref_dependences): Likewise. Split out ...
23173 (vect_slp_analyze_data_ref_dependences): ... this new function.
23174 (dr_group_sort_cmp): New function.
23175 (vect_analyze_data_ref_accesses): Compute data-reference groups
23176 here instead of in vect_analyze_data_ref_dependence. Use
23177 a more efficient algorithm.
23178 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
23179 vect_slp_analyze_data_ref_dependences. Call
23180 vect_analyze_data_ref_accesses earlier.
23181 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
23182 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
23183 (vect_slp_analyze_data_ref_dependences): New prototype.
23185 2013-03-21 Richard Biener <rguenther@suse.de>
23187 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
23188 ref is stored in the loop.
23189 (find_refs_for_sm): Walk only over all stores.
23190 (store_motion_loop): Allocate from lim_bitmap_obstack.
23191 (store_motion): Likewise.
23193 2013-03-21 Richard Biener <rguenther@suse.de>
23195 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
23196 Update virtual SSA form.
23198 2013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
23200 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
23201 * configure: Regenerate.
23202 * config.in: Regenerate.
23203 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
23204 if !HAVE_LD_EH_FRAME_CIEV3.
23206 2013-03-21 Richard Biener <rguenther@suse.de>
23208 * tree-cfg.c (verify_expr_no_block): New function.
23209 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
23210 nor DECL_VALUE_EXPR have locations with associated blocks.
23211 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
23212 (clear_unused_block_pointer): Remove code dealing with
23213 blocks in DECL_DEBUG_EXPR locations.
23215 2013-03-21 Richard Biener <rguenther@suse.de>
23217 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
23218 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
23219 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
23220 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
23221 instead of DECL_DEBUG_EXPR_IS_FROM.
23222 * gimplify.c (gimplify_modify_expr): Likewise.
23223 * tree-cfg.c (verify_expr_location_1): Likewise.
23224 * tree-complex.c (create_one_component_var): Likewise.
23225 * tree-sra.c (create_access_replacement): Likewise.
23226 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
23227 (clear_unused_block_pointer): Likewise.
23228 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
23229 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
23230 * var-tracking.c (var_debug_decl): Likewise.
23231 (track_expr_p): Likewise.
23232 * tree-inline.c (add_local_variables): Likewise. Set
23233 DECL_HAS_DEBUG_EXPR_P after copying it.
23234 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
23235 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
23237 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
23240 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
23241 * configure: Regenerate.
23242 * config.in: Regenerate.
23243 * config/i386/i386.md (*movdf_internal): Use
23244 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
23245 movd instead of movq mnemonic for interunit moves.
23246 (*movdi_internal): Ditto.
23248 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
23250 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
23251 (abd<mode>_3): New pattern.
23252 (aba<mode>_3): New pattern.
23253 (fabd<mode>_3): New pattern.
23255 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
23257 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
23258 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
23259 occurrence of REGISTER_PREFIX as its empty string.
23261 2013-03-20 Jeff Law <law@redhat.com>
23263 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
23264 addititional equivalences for equality comparisons between an SSA_NAME
23265 and a constant where the SSA_NAME was set from a widening conversion.
23267 2013-03-20 Walter Lee <walt@tilera.com>
23269 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
23271 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
23273 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
23274 depending on type attribute.
23275 (*movti_internal): Ditto.
23276 (*movtf_internal): Ditto.
23277 (*movxf_internal): Ditto.
23278 (*movdf_internal): Ditto.
23279 (*movsf_internal): Ditto.
23281 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
23283 * config/i386/i386.md (*movti_internal): Set prefix attribute to
23284 maybe_vex for sselog1 and ssemov types.
23285 (*movdi_internal): Reorder operand constraints.
23286 (*movsi_internal): Ditto. Set prefix attribute to
23287 maybe_vex for sselog1 and ssemov types.
23288 (*movtf_internal): Set prefix attribute to maybe_vex
23289 for sselog1 and ssemov types.
23290 (*movdf_internal): Ditto. Set prefix_data16 attribute for
23291 DImode ssemov types. Reorder operand constraints.
23292 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
23293 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
23294 attribute for SImode ssemov types. Reorder operand constraints.
23296 2013-03-20 Martin Jambor <mjambor@suse.cz>
23298 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
23299 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
23301 2013-03-20 Pat Haugen <pthaugen@us.ibm.com>
23303 * config/rs6000/predicates.md (indexed_address, update_address_mem
23304 update_indexed_address_mem): New predicates.
23305 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
23306 attribute for load/store instructions.
23307 * config/rs6000/dfp.md (movsd_store): Likewise.
23308 (movsd_load): Likewise.
23309 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
23310 (unnamed HI->DI extend define_insn): Likewise.
23311 (unnamed SI->DI extend define_insn): Likewise.
23312 (unnamed QI->SI extend define_insn): Likewise.
23313 (unnamed QI->HI extend define_insn): Likewise.
23314 (unnamed HI->SI extend define_insn): Likewise.
23315 (unnamed HI->SI extend define_insn): Likewise.
23316 (extendsfdf2_fpr): Likewise.
23317 (movsi_internal1): Likewise.
23318 (movsi_internal1_single): Likewise.
23319 (movhi_internal): Likewise.
23320 (movqi_internal): Likewise.
23321 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
23322 attribute for load/store instructions.
23323 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
23325 (mov<mode>_softfloat): Likewise.
23326 (mov<mode>_hardfloat32): Likewise.
23327 (mov<mode>_hardfloat64): Likewise.
23328 (mov<mode>_softfloat64): Likewise.
23329 (movdi_internal32): Likewise.
23330 (movdi_internal64): Likewise.
23331 (probe_stack_<mode>): Likewise.
23333 2013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
23335 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
23336 floating point, and decimal floating point to reload iterator.
23338 * config/rs6000/constraints.md (wl constraint): New constraints to
23339 return FLOAT_REGS if certain options are used to reduce the number
23340 of separate patterns that exist in the file.
23341 (wx constraint): Likewise.
23342 (wz constraint): Likewise.
23344 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
23345 -mdebug=reg, print wg, wl, wx, and wz constraints.
23346 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
23347 Initialize the reload functions for 64-bit binary/decimal floating
23349 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
23350 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
23351 create the buffer on the stack to overcome not having a 32-bit
23353 (rs6000_emit_move): Likewise.
23354 (rs6000_secondary_memory_needed_rtx): Likewise.
23355 (rs6000_alloc_sdmode_stack_slot): Likewise.
23356 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
23357 via xxlxor, just like DFmode 0.0.
23359 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
23360 define as 1 if we are running on a power7 or newer.
23361 (enum r6000_reg_class_enum): Add new constraints.
23363 * config/rs6000/dfp.md (movsd): Delete, combine with binary
23364 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
23365 with other moves by using conditional constraits (wg). Use LFIWZX
23366 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
23367 (movsd splitter): Likewise.
23368 (movsd_hardfloat): Likewise.
23369 (movsd_softfloat): Likewise.
23371 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
23372 binary and decimal floating point moves.
23373 (fmove_ok): New attributes to combine binary and decimal floating
23374 point moves, and to combine power6x (mfpgpr) moves along normal
23376 (real_value_to_target): Likewise.
23377 (f32_lr): Likewise.
23378 (f32_lm): Likewise.
23379 (f32_li): Likewise.
23380 (f32_sr): Likewise.
23381 (f32_sm): Likewise.
23382 (f32_si): Likewise.
23383 (movsf): Combine binary and decimal floating point moves. Combine
23384 power6x (mfpgpr) moves with other moves by using conditional
23385 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
23386 (mov<mode> for SFmode/SDmode); Likewise.
23387 (SFmode/SDmode splitters): Likewise.
23388 (movsf_hardfloat): Likewise.
23389 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
23390 (movsf_softfloat): Likewise.
23391 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
23393 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
23394 wx and wz constraints.
23396 * config/rs6000/constraints.md (wg constraint): New constraint to
23397 return FLOAT_REGS if -mmfpgpr (power6x) was used.
23399 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
23402 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
23403 -mdebug=reg, print wg, wl, wx, and wz constraints.
23404 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
23405 Initialize the reload functions for 64-bit binary/decimal floating
23407 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
23408 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
23409 create the buffer on the stack to overcome not having a 32-bit
23411 (rs6000_emit_move): Likewise.
23412 (rs6000_secondary_memory_needed_rtx): Likewise.
23413 (rs6000_alloc_sdmode_stack_slot): Likewise.
23414 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
23415 via xxlxor, just like DFmode 0.0.
23417 * config/rs6000/dfp.md (movdd): Delete, combine with binary
23418 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
23419 with other moves by using conditional constraits (wg). Use LFIWZX
23420 and STFIWX for loading SDmode on power7.
23421 (movdd splitters): Likewise.
23422 (movdd_hardfloat32): Likewise.
23423 (movdd_softfloat32): Likewise.
23424 (movdd_hardfloat64_mfpgpr): Likewise.
23425 (movdd_hardfloat64): Likewise.
23426 (movdd_softfloat64): Likewise.
23428 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
23429 64-bit binary and decimal floating point moves.
23430 (FMOVE64X): Likewise.
23431 (movdf): Combine 64-bit binary and decimal floating point moves.
23432 Combine power6x (mfpgpr) moves with other moves by using
23433 conditional constraits (wg).
23434 (mov<mode> for DFmode/DDmode): Likewise.
23435 (DFmode/DDmode splitters): Likewise.
23436 (movdf_hardfloat32): Likewise.
23437 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
23438 (movdf_softfloat32): Likewise.
23439 (movdf_hardfloat64_mfpgpr): Likewise.
23440 (movdf_hardfloat64): Likewise.
23441 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
23442 (movdf_softfloat64): Likewise.
23443 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
23444 (reload_<mode>_load): Move to later in the file so they aren't in
23445 the middle of the floating point move insns.
23446 (reload_<mode>_store): Likewise.
23448 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
23451 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
23452 constraint if -mdebug=reg.
23453 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
23454 Enable using dd reload support if needed.
23456 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
23457 binary and decimal floating point moves in rs6000.md.
23458 (movtd_internal): Likewise.
23460 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
23461 decimal floating point moves.
23463 (movtf_internal): Likewise.
23464 (mov<mode>_internal, TDmode/TFmode): Likewise.
23465 (movtf_softfloat): Likewise.
23466 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
23468 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
23469 movdi_internal64, using wg constraint for move direct operations.
23470 (movdi_internal64): Likewise.
23472 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
23473 MODES_TIEABLE_P for selected modes. Print the numerical value of
23474 the various virtual registers. Use GPR/FPR first/last values,
23475 instead of hard coding the register numbers. Print which modes
23476 have reload functions registered.
23477 (rs6000_option_override_internal): If -mdebug=reg, trace the options
23478 settings before/after setting cpu, target and subtarget settings.
23479 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
23480 and for secondary reload failures in rs6000_secondary_reload_inner.
23481 (rs6000_secondary_reload_fail): Likewise.
23482 (rs6000_secondary_reload_inner): Likewise.
23484 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
23485 macros for first/last GPR and FPR registers.
23486 (LAST_GPR_REGNO): Likewise.
23487 (FIRST_FPR_REGNO): Likewise.
23488 (LAST_FPR_REGNO): Likewise.
23490 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
23491 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
23492 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
23493 (vcond<mode><mode>): Likewise.
23494 (vcondu<mode><mode>): Likewise.
23495 (vector_gtu<mode>): Likewise.
23496 (vector_gte<mode>): Likewise.
23497 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
23498 to prevent the compiler from converting DImode operations to TImode.
23499 (ior<mode>3): Likewise.
23500 (and<mode>3): Likewise.
23501 (one_cmpl<mode>2): Likewise.
23502 (nor<mode>3): Likewise.
23503 (andc<mode>3): Likewise.
23505 * config/rs6000/constraints.md (wt constraint): New constraint
23506 that returns VSX_REGS if TImode is allowed in VSX registers.
23508 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
23509 constant under VSX.
23511 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
23512 similar to TImode, but it is restricted to being in the GPRs.
23514 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
23515 TImode to occupy a single VSX register.
23517 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
23518 -mvsx-timode for power7/power8.
23519 (power7 cpu): Likewise.
23520 (power8 cpu): Likewise.
23522 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
23523 sure that TFmode/TDmode take up two registers if they are ever
23524 allowed in the upper VSX registers.
23525 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
23527 (rs6000_init_hard_regno_mode_ok): Likewise.
23528 (rs6000_debug_reg_global): Add debugging for PTImode and wt
23529 constraint. Print if LRA is turned on.
23530 (rs6000_option_override_internal): Give an error if -mvsx-timode
23531 and VSX is not enabled.
23532 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
23533 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
23534 to reg+offset addressing. Use PTImode when checking offset
23535 addresses for validity.
23536 (reg_offset_addressing_ok_p): Likewise.
23537 (rs6000_legitimate_offset_address_p): Likewise.
23538 (rs6000_legitimize_address): Likewise.
23539 (rs6000_legitimize_reload_address): Likewise.
23540 (rs6000_legitimate_address_p): Likewise.
23541 (rs6000_eliminate_indexed_memrefs): Likewise.
23542 (rs6000_emit_move): Likewise.
23543 (rs6000_secondary_reload): Likewise.
23544 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
23545 reloads to fpr registers to continue to use reg+offset addressing,
23546 but 64-bit reloads to altivec registers need reg+reg addressing.
23547 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
23548 it. Treat LO_SUM like a PLUS operation.
23549 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
23550 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
23551 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
23552 registers to share a register with a smaller sized type, since VSX
23553 puts scalars in the upper 64-bits.
23554 (print_operand): Add support for PTImode.
23555 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
23556 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
23557 registers, but don't have arithmetic support.
23558 (rs6000_memory_move_cost): Add test for VSX.
23559 (rs6000_opt_masks): Add -mvsx-timode.
23561 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
23564 (VSr): Use wt constraint for TImode.
23565 (VSv): Drop TImode support.
23566 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
23567 (vsx_movti_64bit): Likewise.
23568 (vsx_movti_32bit): Likewise.
23569 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
23570 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
23571 one '?' on the appropriate output constraint. Do not allow TImode
23572 logical operations on 32-bit systems.
23573 (vsx_ior<mode>3): Likewise.
23574 (vsx_xor<mode>3): Likewise.
23575 (vsx_one_cmpl<mode>2): Likewise.
23576 (vsx_nor<mode>3): Likewise.
23577 (vsx_andc<mode>3): Likewise.
23578 (vsx_concat_<mode>): Likewise.
23579 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
23581 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
23582 OPTION_MASK_VSX_TIMODE.
23583 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
23584 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
23586 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
23587 (TI2 iterator): New iterator for TImode, PTImode.
23588 (wd mode attribute): Add values for vector types.
23589 (movti_string): Replace TI move operations with operations for TImode
23590 and PTImode. Add support for TImode being allowed in VSX registers.
23591 (mov<mode>_string, TImode/PTImode): Likewise.
23592 (movti_ppc64): Likewise.
23593 (mov<mode>_ppc64, TImode/PTImode): Likewise.
23594 (TI mode splitters): Likewise.
23596 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
23599 2013-03-20 Marc Glisse <marc.glisse@inria.fr>
23601 PR tree-optimization/56355
23602 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
23603 Also handle integers with undefined overflow.
23605 2013-03-20 Catherine Moore <clm@codesourcery.com>
23606 Maciej W. Rozycki <macro@codesourcery.com>
23607 Tom de Vries <tom@codesourcery.com>
23608 Nathan Sidwell <nathan@codesourcery.com>
23609 Iain Sandoe <iain@codesourcery.com>
23610 Nathan Froyd <froydnj@codesourcery.com>
23611 Chao-ying Fu <fu@mips.com>
23613 * doc/extend.texi: (micromips, nomicromips, nocompression):
23614 Document new function attributes.
23615 * doc/invoke.texi (minterlink-compressed, mmicromips,
23616 m14k, m14ke, m14kec): Document new options.
23617 (minterlink-mips16): Update documentation.
23618 * doc/md.texi (ZC, ZD): Document new constraints.
23619 * configure.ac (gcc_cv_as_micromips): Check if linker
23620 supports the .set micromips directive.
23621 * configure: Regenerate.
23622 * config.in: Regenerate.
23623 * config/mips/mips-tables.opt: Regenerate.
23624 * config/mips/micromips.md: New file.
23625 * constraints.md (ZC, ZD): New constraints.
23626 * config/mips/predicates.md (movep_src_register): New predicate.
23627 (movep_src_operand): New predicate.
23628 (non_volatile_mem_operand): New predicate.
23629 * config/mips/mips.md (multimem): New type.
23630 (length): Differentiate between 17-bit and 18-bit branch offsets.
23631 (MOVEP1, MOVEP2): New mode iterator.
23632 (mov_<load>l): Use ZC constraint.
23633 (mov_<load>r): Likewise.
23634 (mov_<store>l): Likewise.
23635 (mov_<store>r): Likewise.
23636 (*branch_equality<mode>_inverted): Add microMIPS support.
23637 (*branch_equality<mode>): Likewise.
23638 (*jump_absolute): Likewise.
23639 (indirect_jump_<mode>): Likewise.
23640 (tablejump_<mode>): Likewise.
23641 (<optab>_internal): Likewise.
23642 (sibcall_internal): Likewise.
23643 (sibcall_value_internal): Likewise.
23644 (prefetch): Use constraint ZD.
23645 * config/mips/mips.opt (minterlink-compressed): New option.
23646 (minterlink-mips16): Now an alias for minterlink-compressed.
23647 (mmicromips): New option.
23648 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
23649 (compare_and_swap_12): Likewise.
23650 (sync_add<mode>): Likewise.
23651 (sync_<optab>_12): Likewise.
23652 (sync_old_<optab>_12): Likewise.
23653 (sync_new_<optab>_12): Likewise.
23654 (sync_nand_12): Likewise.
23655 (sync_old_nand_12): Likewise.
23656 (sync_new_nand_12): Likewise.
23657 (sync_sub<mode>): Likewise.
23658 (sync_old_add<mode>): Likewise.
23659 (sync_old_sub<mode>): Likewise.
23660 (sync_new_add<mode>): Likewise.
23661 (sync_new_sub<mode>): Likewise.
23662 (sync_<optab><mode>): Likewise.
23663 (sync_old_<optab><mode>): Likewise.
23664 (sync_new_<optab><mode>): Likewise.
23665 (sync_nand<mode>): Likewise.
23666 (sync_old_nand<mode>): Likewise.
23667 (sync_new_nand<mode>): Likewise.
23668 (sync_lock_test_and_set<mode>): Likewise.
23669 (test_and_set_12): Likewise.
23670 (atomic_compare_and_swap<mode>): Likewise.
23671 (atomic_exchange<mode>_llsc): Likewise.
23672 (atomic_fetch_add<mode>_llsc): Likewise.
23673 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
23674 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
23675 (umips_save_restore_pattern_p): Likewise.
23676 (umips_load_store_pair_p): Likewise.
23677 (umips_output_load_store_pair): Likewise.
23678 (umips_movep_target_p): Likewise.
23679 (umips_12bit_offset_address_p): Likewise.
23680 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
23681 (mips_base_mips16): Rename this...
23682 (mips_base_compression_flags): ...to this. Update all uses.
23683 (mips_attribute_table): Add micromips, nomicromips and nocompression.
23684 (mips_mips16_decl_p): Delete.
23685 (mips_nomips16_decl_p): Delete.
23686 (mips_get_compress_on_flags): New function.
23687 (mips_get_compress_off_flags): New function.
23688 (mips_get_compress_mode): New function.
23689 (mips_get_compress_on_name): New function.
23690 (mips_get_compress_off_name): New function.
23691 (mips_insert_attributes): Support multiple compression types.
23692 (mips_merge_decl_attributes): Likewise.
23693 (umips_12bit_offset_address_p): New function.
23694 (mips_start_function_definition): Emit .set micromips directive.
23695 (mips_call_may_need_jalx_p): New function.
23696 (mips_function_ok_for_sibcall): Add microMIPS support.
23697 (mips_print_operand_punctuation): Support short delay slots and
23699 (umips_swm_mask, umips_swm_encoding): New.
23700 (umips_build_save_restore): New function.
23701 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
23702 (was_mips16_p): Remove.
23703 (old_compression_mode): New.
23704 (mips_set_compression_mode): New function.
23705 (mips_set_current_function): Add microMIPS support.
23706 (mips_option_override): Likewise.
23707 (umips_save_restore_pattern_p): New function.
23708 (umips_output_save_restore): New function.
23709 (umips_load_store_pair_p_1): New function.
23710 (umips_load_store_pair_p): New function.
23711 (umips_output_load_store_pair_1): New function.
23712 (umips_output_load_store_pair): New function.
23713 (umips_movep_target_p) New function.
23714 (mips_prepare_pch_save): Add microMIPS support.
23715 * config/mips/mips.h (TARGET_COMPRESSION): New.
23716 (TARGET_CPU_CPP_BUILTINS): Update macro
23717 to use new compression flags and to support microMIPS.
23718 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
23719 (MIPS_ARCH_FLOAT_SPEC): Likewise.
23720 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
23721 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
23722 (ASM_SPEC): Support mmicromips and mno-micromips.
23723 (M16STORE_REG_P): New macro.
23724 (MIPS_CALL): Support TARGET_MICROMIPS.
23725 (MICROMIPS_J): New macro.
23726 (mips_base_mips16): Rename this...
23727 (mips_base_compression_flags): ...to this.
23728 (UMIPS_12BIT_OFFSET_P): New macro.
23729 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
23730 (MULTILIB_DIRNAMES): Likewise.
23731 2013-03-20 Richard Biener <rguenther@suse.de>
23733 PR tree-optimization/56661
23734 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
23735 the result does not have to be distinct.
23737 2013-03-20 Richard Biener <rguenther@suse.de>
23739 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
23742 2013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
23743 Steven Bosscher <steven@gcc.gnu.org>
23745 PR rtl-optimization/56605
23746 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
23748 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
23751 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
23752 that require movd instead of movq.
23754 2013-03-20 Richard Biener <rguenther@suse.de>
23756 * tree-ssa-structalias.c (struct variable_info): Add pointer
23757 to the first field of an aggregate with sub-vars. Make
23758 this and the pointer to the next subfield its ID.
23759 (vi_next): New function.
23760 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
23761 storedanything_id, integer_id): Increment by one.
23762 (new_var_info, get_call_vi, lookup_call_clobber_vi,
23763 get_call_clobber_vi): Adjust.
23764 (solution_set_expand): Simplify and speedup.
23765 (solution_set_add): Inline into ...
23766 (set_union_with_increment): ... this. Adjust accordingly.
23767 (do_sd_constraint): Likewise.
23768 (do_ds_constraint): Likewise.
23769 (do_complex_constraint): Simplify.
23770 (build_pred_graph): Adjust.
23771 (solve_graph): Likewise. Simplify and speedup.
23772 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
23773 get_constraint_for_component_ref, get_constraint_for_1,
23774 first_vi_for_offset, first_or_preceding_vi_for_offset,
23775 create_function_info_for, create_variable_info_for_1,
23776 create_variable_info_for, intra_create_variable_infos): Adjust.
23777 (init_base_vars): Push NULL for ID zero.
23778 (compute_points_to_sets): Adjust.
23780 2013-03-20 Richard Biener <rguenther@suse.de>
23782 * cfgloop.c (verify_loop_structure): Streamline and avoid
23783 ICEing on corrupt loop tree.
23784 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
23787 2013-03-20 Richard Biener <rguenther@suse.de>
23789 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
23790 check whether an SSA update is needed.
23792 2013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
23794 * config/mips/constraints.md (T): Rename to...
23798 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
23799 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
23801 2013-03-19 Ian Bolton <ian.bolton@arm.com>
23803 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
23804 (*subsi3_carryin_uxtw): Likewise.
23806 2013-03-19 Ian Bolton <ian.bolton@arm.com>
23808 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
23809 (*rorsi3_insn_uxtw): Likewise.
23811 2013-03-19 Ian Bolton <ian.bolton@arm.com>
23813 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
23814 (*extrsi5_insn_uxtw): Likewise.
23816 2013-03-19 Richard Biener <rguenther@suse.de>
23818 PR tree-optimization/56273
23819 * passes.c (init_optimization_passes): Move second VRP after DOM.
23821 2013-03-19 Uros Bizjak <ubizjak@gmail.com>
23823 * config/i386/i386.md (*movti_internal): Merge from
23824 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
23825 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
23826 nox64 isa attributes.
23828 2013-03-18 Richard Biener <rguenther@suse.de>
23830 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
23832 (merge_node_constraints): Likewise.
23833 (build_succ_graph): Likewise.
23834 (valid_graph_edge): Inline into single caller.
23835 (unify_nodes): Likewise. Use bitmap_set_bit return value
23837 (scc_visit): Fix formatting and variable use.
23838 (do_sd_constraint): Use gcc_checking_assert.
23839 (do_ds_constraint): Likewise.
23840 (do_complex_constraint): Likewise.
23841 (condense_visit): Likewise. Cleanup.
23842 (dump_pred_graph): New function.
23843 (perform_var_substitution): Dump the pred-graph before
23844 variable substitution.
23845 (find_equivalent_node): Use gcc_checking_assert.
23846 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
23848 2013-03-18 Richard Biener <rguenther@suse.de>
23850 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
23851 Remove cond_expr_stmt_list argument and do not gimplify the
23853 (vect_loop_versioning): Adjust.
23854 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
23855 Cleanup to use less temporaries.
23856 (vect_create_data_ref_ptr): Cleanup.
23858 2013-03-18 Jakub Jelinek <jakub@redhat.com>
23860 PR tree-optimization/56635
23861 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
23862 require types_compatible_p types.
23864 2013-03-18 Nick Clifton <nickc@redhat.com>
23866 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
23867 spurious backslash.
23869 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
23870 Add missing line to comment describing function.
23872 2013-03-18 Richard Biener <rguenther@suse.de>
23874 PR tree-optimization/56210
23875 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
23876 Handle string / character search functions.
23877 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
23879 2013-03-18 Richard Biener <rguenther@suse.de>
23881 PR middle-end/56483
23882 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
23883 and implement properly.
23884 * gimple.h (gimple_cond_single_var_p): Remove.
23886 2013-03-18 Richard Biener <rguenther@suse.de>
23888 * tree-data-ref.h (find_data_references_in_loop): Declare.
23889 * tree-data-ref.c (get_references_in_stmt): Use a stack
23890 vector pre-allocated in the callers.
23891 (find_data_references_in_stmt): Adjust.
23892 (graphite_find_data_references_in_stmt): Likewise.
23893 (create_rdg_vertices): Likewise.
23894 (find_data_references_in_loop): Export.
23895 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
23896 Compute dependences here...
23897 (vect_analyze_data_refs): ...not here. When we encounter
23898 a non-vectorizable data reference in basic-block vectorization
23899 truncate the data reference vector. Do not bother to
23900 fixup data-dependence information for gather loads.
23901 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
23902 of data references, as reported.
23904 2013-03-18 Richard Biener <rguenther@suse.de>
23906 PR tree-optimization/3713
23907 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
23908 has_constants and expr.
23909 (stmt_has_constants): Properly valueize SSA names when deciding
23910 whether the stmt has constants.
23912 2013-03-18 Richard Biener <rguenther@suse.de>
23914 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
23915 whole function when there is nothing to do.
23916 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
23917 * tree-vectorizer.c (vectorize_loops): Update virtual and
23918 loop-closed SSA once.
23919 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
23921 2013-03-18 Richard Biener <rguenther@suse.de>
23923 PR middle-end/56113
23924 * domwalk.c (bb_postorder): New global static.
23925 (cmp_bb_postorder): New function.
23926 (walk_dominator_tree): Replace scheme imposing an order for
23927 visiting dominator sons by one sorting them at the time they
23928 are pushed on the stack.
23930 2013-03-18 Richard Biener <rguenther@suse.de>
23932 PR tree-optimization/39326
23933 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
23934 (struct mem_ref): Replace mem member with ao_ref typed member.
23935 (MEM_ANALYZABLE): Adjust.
23936 (memref_eq): Likewise.
23937 (mem_ref_alloc): Likewise.
23938 (gather_mem_refs_stmt): Likewise.
23939 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
23940 (execute_sm_if_changed_flag_set): Adjust.
23941 (execute_sm): Likewise.
23942 (ref_always_accessed_p): Likewise.
23943 (refs_independent_p): Likewise.
23944 (can_sm_ref_p): Likewise.
23946 2013-03-18 Jakub Jelinek <jakub@redhat.com>
23949 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
23950 return 1 even for !unsignedp.
23952 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
23954 * config/i386/i386.md (isa): Add x64 and nox64.
23955 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
23956 (*pushtf): Enable *roF alternative for x64 isa only.
23957 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
23958 mode attribute of integer alternatives to DImode for TARGET_64BIT.
23959 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
23960 (*movtf_internal): Merge from *movtf_internal_rex64 and
23961 *movtf_internal_sse. Use x64 and nox64 isa attributes.
23962 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
23963 nox64 isa attributes.
23964 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
23965 nox64 isa attributes.
23966 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
23968 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
23970 * config/alpha/alpha.c (TARGET_LRA_P): New define.
23972 2013-03-17 Jakub Jelinek <jakub@redhat.com>
23975 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
23976 class names. Remove trailing comma after "ALL_REGS".
23978 2013-03-16 Jan Hubicka <jh@suse.cz>
23980 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
23981 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
23982 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
23983 of cgraph_get_create_node.
23984 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
23986 2013-03-16 Jason Merrill <jason@redhat.com>
23989 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
23990 with DW_AT_default_value.
23992 2013-03-16 Jakub Jelinek <jakub@redhat.com>
23994 * BASE-VER: Set to 4.9.0.
23996 2013-03-14 Andi Kleen <ak@linux.intel.com>
23999 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
24000 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
24001 Document _x* TSX intrinsics.
24003 2013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
24004 David Holsgrove <david.holsgrove@xilinx.com>
24006 * configure.ac: Add MicroBlaze TLS support detection.
24007 * configure: Regenerate.
24008 * config/microblaze/microblaze-protos.h
24009 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
24010 symbol_mentioned_p, label_mentioned_p): Add prototypes.
24011 * config/microblaze/microblaze.c (microblaze_address_type): Add
24012 ADDRESS_TLS and tls_reloc address types.
24013 (microblaze_address_info): Add tls_reloc.
24014 (TARGET_HAVE_TLS): Define.
24015 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
24016 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
24017 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
24018 load_tls_operand, microblaze_call_tls_get_addr,
24019 microblaze_legitimize_tls_address): New functions.
24020 (microblaze_classify_unspec): Handle UNSPEC_TLS.
24021 (get_base_reg): Use microblaze_tls_symbol_p.
24022 (microblaze_classify_address): Handle TLS.
24023 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
24024 label_mentioned_p and microblaze_tls_referenced_p.
24025 (microblaze_legitimize_address): Handle TLS.
24026 (microblaze_address_insns): Handle ADDRESS_TLS.
24027 (pic_address_needs_scratch): Handle TLS.
24028 (print_operand_address): Handle TLS.
24029 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
24030 (microblaze_expand_move): Handle TLS.
24031 (microblaze_legitimate_constant_p): Check
24032 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
24033 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
24034 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
24035 (PIC_OFFSET_TABLE_REGNUM): Set.
24036 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
24037 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
24038 (addsi3, movsi_internal2, movdf_internal): Update constraints
24039 * config/microblaze/predicates.md (arith_plus_operand): Define
24040 (move_operand): Redefine as move_src_operand,
24041 check microblaze_tls_referenced_p.
24043 2013-03-14 Ian Bolton <ian.bolton@arm.com>
24045 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
24046 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
24048 2013-03-14 Ian Bolton <ian.bolton@arm.com>
24050 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
24053 2013-03-14 Jakub Jelinek <jakub@redhat.com>
24055 PR tree-optimization/53265
24056 * common.opt (Waggressive-loop-optimizations): New option.
24057 * tree-ssa-loop-niter.c: Include tree-pass.h.
24058 (do_warn_aggressive_loop_optimizations): New function.
24059 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
24060 if number_of_latch_executions returned constant.
24061 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
24062 early. If number_of_latch_executions returned constant, set
24063 nb_iterations_upper_bound back to it.
24064 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
24066 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
24067 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
24069 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
24070 (MULTILIB_OSDIRNAMES): Set.
24071 * genmultilib: If defaultosdirname doesn't start with :: , set
24072 defaultosdirname2 instead, clear it and emit two . multilib_raw
24073 entries instead of just one.
24075 2013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
24077 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
24078 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
24079 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
24080 (SUBTARGET_OVERRIDE_OPTIONS): New.
24082 2013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
24085 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
24086 (musermode): Convert to Var(TARGET_USERMODE).
24087 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
24088 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
24089 * config/sh/sh.c (sh_option_override): Use
24090 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
24091 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
24093 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
24095 (udivsi3_i4_single, divsi3_i4_single): Use
24096 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
24098 2013-03-13 Dave Korn <dave.korn.cygwin@gmail.com>
24100 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
24103 2013-03-13 Richard Biener <rguenther@suse.de>
24105 PR tree-optimization/56608
24106 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
24107 calls when vectorizing basic-blocks.
24109 2013-03-13 Jakub Jelinek <jakub@redhat.com>
24112 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
24115 2013-03-12 Jakub Jelinek <jakub@redhat.com>
24117 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
24119 2013-03-11 Jan Hubicka <jh@suse.cz>
24122 * lto-streamer-out.c (output_symbol_p): Skip references from
24123 constructors of external variables.
24125 2013-03-11 Jan Hubicka <jh@suse.cz>
24127 PR middle-end/56571
24128 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
24130 * emit-rtl.c (verify_rtx_sharing): Likewise.
24131 (copy_insn_1): Likewise.
24132 * rtl.c (copy_rtx): Likewise.
24134 2013-03-11 Georg-Johann Lay <avr@gjlay.de>
24137 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
24138 output_operand_lossage message.
24140 2013-03-11 Richard Earnshaw <rearnsha@arm.com>
24143 * arm.c (shift_op): Validate RTL pattern on the fly.
24144 (arm_print_operand, case 'S'): Don't use shift_operator to validate
24147 2013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
24150 * config/pa/pa.md (call_value): Check for calls to powf and direct to
24151 new call patterns that clobber %fr12.
24152 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
24153 split and postreload patterns.
24154 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
24155 registers %fr12 and %fr12R as call used.
24157 2013-03-09 Steven Bosscher <steven@gcc.gnu.org>
24159 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
24160 (canon_address, record_store, replace_read, check_mem_read_rtx,
24161 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
24162 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
24163 rest_of_handle_dse): Likewise.
24165 2013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
24167 PR middle-end/56524
24168 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
24170 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
24171 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
24172 (save_optabs_if_changed): Replace with...
24173 (init_tree_optimization_optabs): ...this.
24174 * optabs.c (save_optabs_if_changed): Rename to...
24175 (init_tree_optimization_optabs): ...this. Take the optimization node
24176 as argument. Do nothing if the base optabs are already correct.
24177 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
24178 to recompute optabs.
24179 * function.h (function): Remove optabs field.
24180 * function.c (invoke_set_current_function_hook): Call
24181 init_tree_optimization_optabs. Use the result to initialize
24184 2013-02-27 Aldy Hernandez <aldyh@redhat.com>
24186 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
24187 if GTMA_HAS_NO_INSTRUMENTATION.
24188 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
24189 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
24190 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
24191 * gimple-pretty-print.c (dump_gimple_transaction): Handle
24192 GTMA_HAS_NO_INSTRUMENTATION.
24194 2013-03-08 Jakub Jelinek <jakub@redhat.com>
24196 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
24199 2013-03-08 Marek Polacek <polacek@redhat.com>
24200 Jakub Jelinek <jakub@redhat.com>
24202 PR tree-optimization/56478
24203 * predict.c (is_comparison_with_loop_invariant_p): Change the
24204 type of loop_step to tree.
24205 (predict_loops): Adjust.
24206 (predict_iv_comparison): Perform the computations on double_ints.
24208 2013-03-08 Richard Biener <rguenther@suse.de>
24210 PR tree-optimization/56570
24211 * tree-cfg.c (verify_expr_location_1): Verify locations for
24213 * tree-sra.c (create_access_replacement): Strip locations
24214 from DECL_DEBUG_EXPRs.
24216 2013-03-08 Richard Biener <rguenther@suse.de>
24218 * tree-inline.c (expand_call_inline): Do not associate
24219 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
24220 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
24222 2013-03-08 Richard Biener <rguenther@suse.de>
24224 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
24225 or block changes with -Og. Fix for location / block encoding
24226 changes and PHI arguments with locations.
24228 2013-03-07 Steven Bosscher <steven@gcc.gnu.org>
24230 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
24232 (struct output_info): Likewise.
24233 (register_overhead): Remove bad gcc_assert.
24234 (bitmap_find_bit): If there is only a single bitmap element, do not
24235 count a miss as a search.
24236 (print_statistics): Update for counter type changes.
24237 (dump_bitmap_statistics): Likewise. Print headers such that they
24238 are properly lined up with the printed counters.
24240 2013-03-07 Jakub Jelinek <jakub@redhat.com>
24242 PR tree-optimization/56559
24243 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
24244 check that it has only a single use.
24246 2013-03-07 Richard Biener <rguenther@suse.de>
24248 * doc/invoke.texi (fwhole-program): Discourage use in combination
24251 2013-03-06 Jakub Jelinek <jakub@redhat.com>
24253 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
24255 PR tree-optimization/56539
24256 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
24257 instead of GSI_CONTINUE_LINKING as last argument to
24258 force_gimple_operand_gsi. Adjust function comment.
24260 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
24263 PR middle-end/56548
24264 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
24265 promoted mode, convert the result back to the original mode.
24267 2013-03-06 Richard Biener <rguenther@suse.de>
24269 PR middle-end/56294
24270 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
24271 (insert_updated_phi_nodes_compare_uids): New function.
24272 (update_ssa): Sort symbols_to_rename after UID before
24273 traversing it to insert PHI nodes.
24275 2013-03-06 Richard Biener <rguenther@suse.de>
24277 PR middle-end/50494
24278 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
24279 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
24282 2013-02-13 Richard Biener <rguenther@suse.de>
24285 * varasm.c (output_constant_def_1): Get the decl representing
24286 the constant as argument.
24287 (output_constant_def): Wrap output_constant_def_1.
24288 (make_decl_rtl): Use output_constant_def_1 with the decl
24289 representing the constant.
24290 (build_constant_desc): Optionally re-use a decl already
24291 representing the constant.
24292 (tree_output_constant_def): Adjust.
24294 2013-03-06 Joey Ye <joey.ye@arm.com>
24297 * gcc.c (convert_white_space): New function.
24298 (main): Handles white space in function name.
24300 2013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
24303 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
24304 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
24305 to SH_DIV_CALL_TABLE for TARGET_SH2.
24306 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
24308 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
24309 call-table options.
24311 2013-03-05 Sterling Augustine <saugustine@google.com>
24312 Cary Coutant <ccoutant@google.com>
24315 * dwarf2out.c (resolve_addr): Don't call
24316 remove_loc_list_addr_table_entries a second time for the same
24319 2013-03-05 Jakub Jelinek <jakub@redhat.com>
24322 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
24323 (avoid_complex_debug_insns): New function.
24324 (expand_debug_locations): Call it.
24326 PR rtl-optimization/56484
24327 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
24328 lifetimes of hard registers on small register class machines.
24330 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
24332 * config/microblaze/microblaze-protos.h: Rename
24333 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
24334 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
24336 (microblaze_fast_interrupt_function_p): New function.
24337 (microblaze_is_interrupt_handler): Rename to
24338 microblaze_is_interrupt_variant and add fast_interrupt check.
24339 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
24340 (save_restore_insns): Likewise.
24341 (compute_frame_size): Likewise.
24342 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
24343 (microblaze_globalize_label): Likewise.
24344 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
24345 * config/microblaze/microblaze.md: Use wrapper
24346 microblaze_is_interrupt_variant.
24348 2013-03-05 Kai Tietz <ktietz@redhat.com>
24350 * sdbout.c (sdbout_one_type): Switch to current function's section
24351 supporting cold/hot.
24353 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
24355 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
24358 2013-03-05 Jakub Jelinek <jakub@redhat.com>
24360 PR middle-end/56461
24361 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
24362 if VALGRIND_GET_VBITS is defined, temporarily make object
24363 memory all defined, and restore previous valgrind addressability
24364 and definability afterwards. Free this_object at the end.
24366 PR middle-end/56461
24367 * lra.c (lra): Call lra_clear_live_ranges if live_p,
24368 right before calling lra_create_live_ranges, also call it
24369 when clearing live_p. Only call lra_clear_live_ranges
24370 at the end if live_p.
24372 PR middle-end/56461
24373 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
24375 2013-03-05 Richard Biener <rguenther@suse.de>
24377 PR tree-optimization/56521
24378 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
24381 2013-03-05 Steven Bosscher <steven@gcc.gnu.org>
24384 * except.h (remove_unreachable_eh_regions): New prototype.
24385 * except.c (remove_eh_handler_splicer): New function, split out
24386 of remove_eh_handler.
24387 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
24388 warning about running it on many EH regions one at a time.
24389 (remove_unreachable_eh_regions_worker): New function, walk the
24390 EH tree in depth-first order and remove non-marked regions.
24391 (remove_unreachable_eh_regions): New function.
24392 * tree-eh.c (mark_reachable_handlers): New function, split out
24393 from remove_unreachable_handlers.
24394 (remove_unreachable_handlers): Use mark_reachable_handlers and
24395 remove_unreachable_eh_regions.
24396 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
24397 and remove_unreachable_eh_regions.
24399 2013-03-05 Richard Biener <rguenther@suse.de>
24401 PR middle-end/56525
24402 * loop-init.c (fix_loop_structure): Remove loops in two stages,
24403 not freeing them until the end.
24405 2013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
24407 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
24409 2013-03-05 Richard Biener <rguenther@suse.de>
24411 PR tree-optimization/56270
24412 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
24413 of loads after scheduling an SLP instance.
24415 2013-03-05 Jakub Jelinek <jakub@redhat.com>
24417 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
24419 (check_gcc_parallelize): Run guality.exp as a separate job from
24420 vect.exp with unsorted.exp and $(dg_target_exps) separately from
24421 struct-layout-1.exp with stackalign.exp.
24423 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
24425 PR middle-end/56461
24426 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
24427 load_index sbitmap even if some bit in it isn't set.
24429 PR middle-end/56461
24430 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
24431 (discover_iteration_bound_by_body_walk): Change queues to
24432 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
24433 spelling in comment. Call safe_push on queues[bound_index] directly.
24434 Release queues[queue_index] in every iteration unconditionally.
24435 Release bounds vector.
24437 PR middle-end/56461
24438 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
24439 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
24440 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
24443 2013-03-05 Richard Biener <rguenther@suse.de>
24446 * tree-inline.c (remap_blocks_to_null): New function.
24447 (expand_call_inline): When expanding a call stmt without
24448 an associated block inline remap all callee blocks to NULL.
24450 2013-03-05 Jakub Jelinek <jakub@redhat.com>
24452 PR rtl-optimization/56494
24453 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
24454 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
24455 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
24457 PR middle-end/56461
24458 * sel-sched-ir.c (free_sched_pools): Release
24459 succs_info_pool.stack[succs_info_pool.max_top] vectors too
24460 if succs_info_pool.max_top isn't -1.
24463 * opts.c (opts_obstack, opts_concat): Moved to...
24464 * opts-common.c (opts_obstack, opts_concat): ... here.
24466 2013-03-04 Jakub Jelinek <jakub@redhat.com>
24468 PR middle-end/56461
24469 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
24471 2013-03-04 Martin Jambor <mjambor@suse.cz>
24473 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
24474 all appropriate places.
24476 2013-01-04 Eric Botcazou <ebotcazou@adacore.com>
24478 PR tree-optimization/56424
24479 * ipa-split.c (split_function): Do not set the RSO flag if result is
24480 not by reference and its type is a register type.
24482 2013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
24484 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
24485 (microblaze_legitimate_pic_operand): Likewise
24486 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
24487 new function microblaze_legitimate_pic_operand
24488 * config/microblaze/microblaze-protos.h
24489 (microblaze_legitimate_pic_operand): Declare.
24491 2013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24493 * config/microblaze/predicates.md (call_insn_simple_operand):
24494 New predicate for supported rtx code types.
24495 * config/microblaze/microblaze.md (call_internal1): Use
24496 call_insn_simple_operand predicate.
24498 2013-03-04 Jakub Jelinek <jakub@redhat.com>
24500 PR middle-end/56461
24501 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
24502 partitions.ordered_remove.
24504 PR middle-end/56461
24505 * tree-vect-stmts.c (vectorizable_conversion): Don't call
24506 vec_oprnds0.create (1) for modifier == NONE.
24508 PR middle-end/56461
24509 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
24510 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
24511 vec_oprnds1 right before pushing anything to it for
24514 PR middle-end/56461
24515 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
24516 set nbbs to 0 instead of having separate code path.
24517 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
24518 instead of false as last argument if returning NULL.
24520 2013-03-03 Sandra Loosemore <sandra@codesourcery.com>
24522 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
24523 the attribute is now called "target" instead of "option".
24524 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
24525 * doc/tm.texi.in (Target Attributes): Likewise document the correct
24526 attribute/pragma name for TARGET_OPTION_VALID_P and
24527 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
24528 * doc/tm.texi: Regenerated.
24530 2013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
24532 * config/microblaze/microblaze.c:
24533 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
24534 * config/microblaze/microblaze.h: Add -mxl-reorder to
24536 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
24537 instructions emitted if TARGET_REORDER.
24538 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
24539 or 0 for -m/-mno case, but initialises as 2 to detect default use case
24542 2013-03-01 Xinliang David Li <davidxl@google.com>
24544 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
24547 2013-03-01 Jakub Jelinek <jakub@redhat.com>
24549 PR middle-end/56461
24550 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
24551 vector even when returning true. Fix up function comment formatting.
24553 PR middle-end/56461
24554 * ira-build.c (ira_loop_nodes_count): New variable.
24555 (create_loop_tree_nodes): Initialize it.
24556 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
24558 PR middle-end/56461
24559 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
24560 method on dr_chain and result_chain.
24561 * tree-vect-stmts.c (vectorizable_store): Only call
24562 result_chain.create if j == 0.
24564 PR middle-end/56461
24565 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
24566 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
24567 before overwriting it.
24569 2013-03-01 Tobias Burnus <burnus@net-b.de>
24571 * doc/extended.texi (C Extensions): Change order in @menu
24573 (Other MIPS Built-in Functions): Move last MIPS entry before
24574 "picoChip Built-in Functions".
24575 (SH Built-in Functions): Move after RX Built-in Functions.
24576 * doc/gcc.texi (Introduction): Change order in @menu
24578 * doc/md.texi (Constraints): Ditto.
24579 * gty.texi (Type Information): Ditto.
24580 (User-provided marking routines for template types): Make
24582 * doc/invoke.texi (AArch64 Options): Move before
24583 "Adapteva Epiphany Options".
24585 2013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
24586 Jakub Jelinek <jakub@redhat.com>
24589 * asan.c (gate_asan): Lookup no_sanitize_address instead of
24590 no_address_safety_analysis attribute.
24591 * doc/extend.texi (no_address_safety_attribute): Rename to
24592 no_sanitize_address attribute, mention no_address_safety_analysis
24593 attribute as deprecated alias.
24595 2013-02-28 Jakub Jelinek <jakub@redhat.com>
24597 PR middle-end/56461
24598 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
24599 type to vec<vec<tree> > *.
24600 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
24601 to be vec<tree> instead of vec<tree> *, set vec_defs
24602 to vNULL and call vec_defs.create (number_of_vects), adjust other
24604 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
24605 vectorizable_condition): Adjust vect_get_slp_defs callers.
24607 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
24609 * config/aarch64/aarch64.c
24610 (aarch64_float_const_representable): Remove unused variable.
24612 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
24614 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
24616 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
24618 * config/aarch64/aarch64-builtins.c
24619 (aarch64_init_simd_builtins): Make static.
24621 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
24623 * config/aarch64/aarch64.c
24624 (aarch64_simd_make_constant): Make static.
24626 2013-02-28 Martin Jambor <mjambor@suse.cz>
24628 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
24629 with no initialization to the RHS of debug statements.
24631 2013-02-28 Martin Jambor <mjambor@suse.cz>
24633 PR tree-optimization/56294
24634 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
24636 (get_access_replacement): Do not call create_access_replacement.
24637 Assert a replacement exists.
24638 (get_repl_default_def_ssa_name): Create the replacement declaration
24641 2013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
24643 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
24644 final_end_function.
24646 2013-02-28 Marek Polacek <polacek@redhat.com>
24648 PR rtl-optimization/56466
24649 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
24650 if we're changing a loop.
24651 (peel_loops_completely): Likewise.
24653 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
24656 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
24658 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
24661 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
24662 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
24663 INTX_FTYPE_FX, FX_FTYPE_INTX.
24664 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
24666 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
24668 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
24669 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
24670 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
24671 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
24672 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
24673 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
24674 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
24675 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
24676 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
24677 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
24678 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
24679 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
24680 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
24681 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
24682 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
24683 (avrxmega6): Increase max flash segments from 5 to 6.
24684 * config/avr/t-multilib: Regenerate.
24685 * config/avr/avr-tables.opt: Regenerate.
24686 * doc/avr-mmcu.texi: Regenerate.
24688 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
24690 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
24691 (avr_device_to_arch): Rename to avr_device_to_ld.
24692 (avr_device_to_as): New prototype.
24693 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
24694 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
24695 * config/avr/driver-avr.c (avr_device_to_as): New.
24696 (avr_device_to_arch): Rename to avr_device_to_ld.
24698 2013-02-27 Jakub Jelinek <jakub@redhat.com>
24700 PR middle-end/56461
24701 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
24702 method on dr_chain and result_chain.
24704 PR middle-end/56461
24705 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
24706 pointer_set_destroy on not_executed_last_iteration.
24708 PR middle-end/56461
24709 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
24711 PR middle-end/56461
24712 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
24713 FOR_EACH_DEFINED_FUNCTION when freeing state.
24715 PR middle-end/56461
24716 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
24718 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
24721 PR middle-end/56461
24722 * ipa-cp.c (decide_whether_version_node): Call vec_free on
24723 known_aggs[i].items and release known_aggs vector.
24725 PR middle-end/56461
24726 * ipa-reference.c (propagate): Free node_info even for alias nodes.
24728 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24730 * config/microblaze/microblaze.c (microblaze_emit_compare):
24731 Use xor for EQ/NE comparisions.
24732 * config/microblaze/microblaze.md (cstoresf4): Add constraints
24733 (cbranchsf4): Adjust operator to comparison_operator.
24735 2013-02-27 Jakub Jelinek <jakub@redhat.com>
24737 PR middle-end/56461
24738 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
24740 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
24741 vec_safe_push, always update *slot.
24742 (redirect_edge_var_map_clear): Use vec_free.
24743 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
24744 (free_var_map_entry): Use vec_free.
24745 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
24746 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
24748 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
24750 PR middle-end/45472
24751 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
24752 when the may_trap_p bit of the exprs being merged differs.
24753 Reorder tests for speculativeness in the logical and operator.
24755 2013-02-27 Jakub Jelinek <jakub@redhat.com>
24757 * incpath.c (add_standard_paths): Use reconcat instead of concat
24758 where appropriate and avoid leaking memory.
24760 * opts.h: Include obstack.h.
24761 (opts_concat): New prototype.
24762 (opts_obstack): New declaration.
24763 * opts.c (opts_concat): New function.
24764 (opts_obstack): New variable.
24765 (init_options_struct): Call gcc_init_obstack on opts_obstack.
24766 (finish_options): Use opts_concat instead of concat
24767 and XOBNEWVEC instead of XNEWVEC.
24768 * opts-common.c (generate_canonical_option, decode_cmdline_option,
24769 generate_option): Likewise.
24770 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
24771 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
24774 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
24775 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
24777 2013-02-26 Jakub Jelinek <jakub@redhat.com>
24779 PR middle-end/56461
24780 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
24782 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
24784 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
24785 (arm_block_move_unaligned_straight): Likewise.
24786 (arm_adjust_block_mem): Likewise.
24788 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
24791 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
24792 temp, cond and label.
24793 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
24796 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
24797 * config/c6x/c6x.h (dbx_register_map): Update declaration.
24800 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
24801 of prologue/epilogue functions.
24804 * config/tilegx/tilegx.c (tilegx_expand_prologue):
24805 Remove unused variable cfa_offset.
24806 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
24809 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
24810 type promotion to unsigned.
24813 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
24814 for HOST_WIDE_INT of 32 bit / same size as int.
24815 (arm_block_move_unaligned_straight): Likewise.
24816 (arm_adjust_block_mem): Likewise.
24819 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
24822 2013-02-26 Marek Polacek <polacek@redhat.com>
24824 PR tree-optimization/56426
24825 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
24827 2013-02-26 Richard Biener <rguenther@suse.de>
24830 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
24831 unused variable loops.
24833 2013-02-26 Jakub Jelinek <jakub@redhat.com>
24835 PR tree-optimization/56448
24836 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
24837 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
24838 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
24839 later operands of the references, or even first operand for
24840 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
24842 PR tree-optimization/56443
24843 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
24844 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
24845 to type_for_mode langhook.
24847 2013-02-25 Matt Turner <mattst88@gmail.com>
24849 * doc/invoke.texi: Document r4700.
24851 2013-02-25 Richard Biener <rguenther@suse.de>
24853 PR tree-optimization/56175
24854 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
24856 (simplify_bitwise_binary): ... here. Also guard the conversion
24857 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
24859 2013-02-25 Catherine Moore <clm@codesourcery.com>
24862 2013-02-24 Catherine Moore <clm@codesourcery.com>
24863 Maciej W. Rozycki <macro@codesourcery.com>
24864 Tom de Vries <tom@codesourcery.com>
24865 Nathan Sidwell <nathan@codesourcery.com>
24866 Iain Sandoe <iain@codesourcery.com>
24867 Nathan Froyd <froydnj@codesourcery.com>
24868 Chao-ying Fu <fu@mips.com>
24870 * doc/extend.texi: (micromips, nomicromips, nocompression):
24871 Document new function attributes.
24872 * doc/invoke.texi (minterlink-compressed, mmicromips,
24873 m14k, m14ke, m14kec): Document new options.
24874 (minterlink-mips16): Update documentation.
24875 * doc/md.texi (ZC, ZD): Document new constraints.
24876 * configure.ac (gcc_cv_as_micromips): Check if linker
24877 supports the .set micromips directive.
24878 * configure: Regenerate.
24879 * config.in: Regenerate.
24880 * config/mips/mips-tables.opt: Regenerate.
24881 * config/mips/micromips.md: New file.
24882 * constraints.md (ZC, AD): New constraints.
24883 * config/mips/predicates.md (movep_src_register): New predicate.
24884 (movep_src_operand): New predicate.
24885 (non_volatile_mem_operand): New predicate.
24886 * config/mips/mips.md (multimem): New type.
24887 (length): Differentiate between 17-bit and 18-bit branch offsets.
24888 (MOVEP1, MOVEP2): New mode iterator.
24889 (mov_<load>l): Use ZC constraint.
24890 (mov_<load>r): Likewise.
24891 (mov_<store>l): Likewise.
24892 (mov_<store>r): Likewise.
24893 (*branch_equality<mode>_inverted): Add microMIPS support.
24894 (*branch_equality<mode>): Likewise.
24895 (*jump_absolute): Likewise.
24896 (indirect_jump_<mode>): Likewise.
24897 (tablejump_<mode>): Likewise.
24898 (<optab>_internal): Likewise.
24899 (sibcall_internal): Likewise.
24900 (sibcall_value_internal): Likewise.
24901 (prefetch): Use constraint ZD.
24902 * config/mips/mips.opt (minterlink-compressed): New option.
24903 (minterlink-mips16): Now an alias for minterlink-compressed.
24904 (mmicromips): New option.
24905 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
24906 (compare_and_swap_12): Likewise.
24907 (sync_add<mode>): Likewise.
24908 (sync_<optab>_12): Likewise.
24909 (sync_old_<optab>_12): Likewise.
24910 (sync_new_<optab>_12): Likewise.
24911 (sync_nand_12): Likewise.
24912 (sync_old_nand_12): Likewise.
24913 (sync_new_nand_12): Likewise.
24914 (sync_sub<mode>): Likewise.
24915 (sync_old_add<mode>): Likewise.
24916 (sync_old_sub<mode>): Likewise.
24917 (sync_new_add<mode>): Likewise.
24918 (sync_new_sub<mode>): Likewise.
24919 (sync_<optab><mode>): Likewise.
24920 (sync_old_<optab><mode>): Likewise.
24921 (sync_new_<optab><mode>): Likewise.
24922 (sync_nand<mode>): Likewise.
24923 (sync_old_nand<mode>): Likewise.
24924 (sync_new_nand<mode>): Likewise.
24925 (sync_lock_test_and_set<mode>): Likewise.
24926 (test_and_set_12): Likewise.
24927 (atomic_compare_and_swap<mode>): Likewise.
24928 (atomic_exchange<mode>_llsc): Likewise.
24929 (atomic_fetch_add<mode>_llsc): Likewise.
24930 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
24931 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
24932 (umips_save_restore_pattern_p): Likewise.
24933 (umips_load_store_pair_p): Likewise.
24934 (umips_output_load_store_pair): Likewise.
24935 (umips_movep_target_p): Likewise.
24936 (umips_12bit_offset_address_p): Likewise.
24937 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
24938 (mips_base_mips16): Rename this...
24939 (mips_base_compression_flags): ...to this. Update all uses.
24940 (mips_attribute_table): Add micromips, nomicromips and nocompression.
24941 (mips_mips16_decl_p): Delete.
24942 (mips_nomips16_decl_p): Delete.
24943 (mips_get_compress_on_flags): New function.
24944 (mips_get_compress_off_flags): New function.
24945 (mips_get_compress_mode): New function.
24946 (mips_get_compress_on_name): New function.
24947 (mips_get_compress_off_name): New function.
24948 (mips_insert_attributes): Support multiple compression types.
24949 (mips_merge_decl_attributes): Likewise.
24950 (umips_12bit_offset_address_p): New function.
24951 (mips_start_function_definition): Emit .set micromips directive.
24952 (mips_call_may_need_jalx_p): New function.
24953 (mips_function_ok_for_sibcall): Add microMIPS support.
24954 (mips_print_operand_punctuation): Support short delay slots and
24956 (umips_swm_mask, umips_swm_encoding): New.
24957 (umips_build_save_restore): New function.
24958 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
24959 (was_mips16_p): Remove.
24960 (old_compression_mode): New.
24961 (mips_set_compression_mode): New function.
24962 (mips_set_current_function): Add microMIPS support.
24963 (mips_option_override): Likewise.
24964 (umips_save_restore_pattern_p): New function.
24965 (umips_output_save_restore): New function.
24966 (umips_load_store_pair_p_1): New function.
24967 (umips_load_store_pair_p): New function.
24968 (umips_output_load_store_pair_1): New function.
24969 (umips_output_load_store_pair): New function.
24970 (umips_movep_target_p) New function.
24971 (mips_prepare_pch_save): Add microMIPS support.
24972 * config/mips/mips.h (TARGET_COMPRESSION): New.
24973 (TARGET_CPU_CPP_BUILTINS): Update macro
24974 to use new compression flags and to support microMIPS.
24975 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
24976 (MIPS_ARCH_FLOAT_SPEC): Likewise.
24977 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
24978 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
24979 (ASM_SPEC): Support mmicromips and mno-micromips.
24980 (M16STORE_REG_P): New macro.
24981 (MIPS_CALL): Support TARGET_MICROMIPS.
24982 (MICROMIPS_J): New macro.
24983 (mips_base_mips16): Rename this...
24984 (mips_base_compression_flags): ...to this.
24985 (UMIPS_12BIT_OFFSET_P): New macro.
24986 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
24987 (MULTILIB_DIRNAMES): Likewise.
24989 2013-02-25 Tom de Vries <tom@codesourcery.com>
24991 PR rtl-optimization/56131
24992 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
24993 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
24994 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
24996 2013-02-25 Tobias Burnus <burnus@net-b.de>
24998 * doc/invoke.texi (-fsanitize=): Move from optimization
24999 to debugging options.
25001 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
25003 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
25005 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
25006 Alexander Monakov <amonakov@ispras.ru>
25008 PR middle-end/56077
25009 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
25010 flush pending lists also on non-jumps. Adjust comment.
25012 2013-02-24 Catherine Moore <clm@codesourcery.com>
25013 Maciej W. Rozycki <macro@codesourcery.com>
25014 Tom de Vries <tom@codesourcery.com>
25015 Nathan Sidwell <nathan@codesourcery.com>
25016 Iain Sandoe <iain@codesourcery.com>
25017 Nathan Froyd <froydnj@codesourcery.com>
25018 Chao-ying Fu <fu@mips.com>
25020 * doc/extend.texi: (micromips, nomicromips, nocompression):
25021 Document new function attributes.
25022 * doc/invoke.texi (minterlink-compressed, mmicromips,
25023 m14k, m14ke, m14kec): Document new options.
25024 (minterlink-mips16): Update documentation.
25025 * doc/md.texi (ZC, ZD): Document new constraints.
25026 * configure.ac (gcc_cv_as_micromips): Check if linker
25027 supports the .set micromips directive.
25028 * configure: Regenerate.
25029 * config.in: Regenerate.
25030 * config/mips/mips-tables.opt: Regenerate.
25031 * config/mips/micromips.md: New file.
25032 * constraints.md (ZC, AD): New constraints.
25033 * config/mips/predicates.md (movep_src_register): New predicate.
25034 (movep_src_operand): New predicate.
25035 (non_volatile_mem_operand): New predicate.
25036 * config/mips/mips.md (multimem): New type.
25037 (length): Differentiate between 17-bit and 18-bit branch offsets.
25038 (MOVEP1, MOVEP2): New mode iterator.
25039 (mov_<load>l): Use ZC constraint.
25040 (mov_<load>r): Likewise.
25041 (mov_<store>l): Likewise.
25042 (mov_<store>r): Likewise.
25043 (*branch_equality<mode>_inverted): Add microMIPS support.
25044 (*branch_equality<mode>): Likewise.
25045 (*jump_absolute): Likewise.
25046 (indirect_jump_<mode>): Likewise.
25047 (tablejump_<mode>): Likewise.
25048 (<optab>_internal): Likewise.
25049 (sibcall_internal): Likewise.
25050 (sibcall_value_internal): Likewise.
25051 (prefetch): Use constraint ZD.
25052 * config/mips/mips.opt (minterlink-compressed): New option.
25053 (minterlink-mips16): Now an alias for minterlink-compressed.
25054 (mmicromips): New option.
25055 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
25056 (compare_and_swap_12): Likewise.
25057 (sync_add<mode>): Likewise.
25058 (sync_<optab>_12): Likewise.
25059 (sync_old_<optab>_12): Likewise.
25060 (sync_new_<optab>_12): Likewise.
25061 (sync_nand_12): Likewise.
25062 (sync_old_nand_12): Likewise.
25063 (sync_new_nand_12): Likewise.
25064 (sync_sub<mode>): Likewise.
25065 (sync_old_add<mode>): Likewise.
25066 (sync_old_sub<mode>): Likewise.
25067 (sync_new_add<mode>): Likewise.
25068 (sync_new_sub<mode>): Likewise.
25069 (sync_<optab><mode>): Likewise.
25070 (sync_old_<optab><mode>): Likewise.
25071 (sync_new_<optab><mode>): Likewise.
25072 (sync_nand<mode>): Likewise.
25073 (sync_old_nand<mode>): Likewise.
25074 (sync_new_nand<mode>): Likewise.
25075 (sync_lock_test_and_set<mode>): Likewise.
25076 (test_and_set_12): Likewise.
25077 (atomic_compare_and_swap<mode>): Likewise.
25078 (atomic_exchange<mode>_llsc): Likewise.
25079 (atomic_fetch_add<mode>_llsc): Likewise.
25080 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
25081 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
25082 (umips_save_restore_pattern_p): Likewise.
25083 (umips_load_store_pair_p): Likewise.
25084 (umips_output_load_store_pair): Likewise.
25085 (umips_movep_target_p): Likewise.
25086 (umips_12bit_offset_address_p): Likewise.
25087 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
25088 (mips_base_mips16): Rename this...
25089 (mips_base_compression_flags): ...to this. Update all uses.
25090 (mips_attribute_table): Add micromips, nomicromips and nocompression.
25091 (mips_mips16_decl_p): Delete.
25092 (mips_nomips16_decl_p): Delete.
25093 (mips_get_compress_on_flags): New function.
25094 (mips_get_compress_off_flags): New function.
25095 (mips_get_compress_mode): New function.
25096 (mips_get_compress_on_name): New function.
25097 (mips_get_compress_off_name): New function.
25098 (mips_insert_attributes): Support multiple compression types.
25099 (mips_merge_decl_attributes): Likewise.
25100 (umips_12bit_offset_address_p): New function.
25101 (mips_start_function_definition): Emit .set micromips directive.
25102 (mips_call_may_need_jalx_p): New function.
25103 (mips_function_ok_for_sibcall): Add microMIPS support.
25104 (mips_print_operand_punctuation): Support short delay slots and
25106 (umips_swm_mask, umips_swm_encoding): New.
25107 (umips_build_save_restore): New function.
25108 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
25109 (was_mips16_p): Remove.
25110 (old_compression_mode): New.
25111 (mips_set_compression_mode): New function.
25112 (mips_set_current_function): Add microMIPS support.
25113 (mips_option_override): Likewise.
25114 (umips_save_restore_pattern_p): New function.
25115 (umips_output_save_restore): New function.
25116 (umips_load_store_pair_p_1): New function.
25117 (umips_load_store_pair_p): New function.
25118 (umips_output_load_store_pair_1): New function.
25119 (umips_output_load_store_pair): New function.
25120 (umips_movep_target_p) New function.
25121 (mips_prepare_pch_save): Add microMIPS support.
25122 * config/mips/mips.h (TARGET_COMPRESSION): New.
25123 (TARGET_CPU_CPP_BUILTINS): Update macro
25124 to use new compression flags and to support microMIPS.
25125 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
25126 (MIPS_ARCH_FLOAT_SPEC): Likewise.
25127 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
25128 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
25129 (ASM_SPEC): Support mmicromips and mno-micromips.
25130 (M16STORE_REG_P): New macro.
25131 (MIPS_CALL): Support TARGET_MICROMIPS.
25132 (MICROMIPS_J): New macro.
25133 (mips_base_mips16): Rename this...
25134 (mips_base_compression_flags): ...to this.
25135 (UMIPS_12BIT_OFFSET_P): New macro.
25136 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
25137 (MULTILIB_DIRNAMES): Likewise.
25139 2013-02-24 Jakub Jelinek <jakub@redhat.com>
25142 * target-globals.c (save_target_globals): For init_reg_sets and
25143 target_reinit remporarily set this_fn_optabs to this_target_optabs.
25145 2013-02-22 James Greenhalgh <james.greenhalgh@arm.com>
25147 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
25148 * config/aarch64/t-aarch64
25149 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
25151 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
25153 PR inline-asm/56148
25154 * lra-constraints.c (process_alt_operands): Reload operand
25155 conflicting with earlier clobber only if no more other conflicting
25158 2013-02-22 Jakub Jelinek <jakub@redhat.com>
25161 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
25162 if not linking a shared library.
25164 2013-02-22 Seth LaForge <sethml@google.com>
25166 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
25168 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
25170 * config/arm/arm.md (split for extendsidi): Update condition.
25171 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
25172 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
25173 (qhs_zextenddi_cstr): Likewise.
25175 2013-02-21 Jakub Jelinek <jakub@redhat.com>
25177 PR middle-end/56420
25178 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
25179 avoid signed wrapping.
25180 (expand_mult): Handle properly multiplication by
25181 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
25182 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
25183 in the compiler if coeff is HOST_WIDE_INT_MIN.
25184 (expand_divmod): Don't make ext_op1 static, change it's type to
25185 uhwi. Avoid undefined behavior in -INTVAL (op1).
25187 PR rtl-optimization/50339
25188 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
25190 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
25191 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
25192 into splitting_ashiftrt field.
25193 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
25195 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
25198 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
25200 PR middle-end/56108
25201 * trans-mem.c (execute_tm_mark): Do not expand transactions that
25202 are sure to go irrevocable.
25204 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
25206 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
25207 scalars are valid operands.
25209 2013-02-21 Martin Jambor <mjambor@suse.cz>
25211 PR tree-optimization/56310
25212 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
25213 only matching indices and non-negative final offsets.
25214 (intersect_aggregates_with_edge): Pass src_idx to
25215 agg_replacements_to_vector. Pass src_idx insstead of index to
25216 intersect_with_agg_replacements.
25218 2013-02-21 Martin Jambor <mjambor@suse.cz>
25220 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
25221 instead of hard-wired defaults.
25223 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
25225 * doc/invoke.texi (MIPS Options): Update documentation of the
25226 floating-point multiply-accumulate instruction restrictions.
25228 2013-02-21 Kostya Serebryany <kcc@google.com>
25230 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
25231 asan_shadow_offset on x86_64 linux.
25233 2013-02-21 Richard Biener <rguenther@suse.de>
25235 PR tree-optimization/56415
25237 2013-02-11 Richard Biener <rguenther@suse.de>
25239 PR tree-optimization/56273
25240 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
25243 2013-02-21 Jakub Jelinek <jakub@redhat.com>
25246 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
25249 PR inline-asm/56405
25250 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
25251 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
25253 2013-02-20 Jan Hubicka <jh@suse.cz>
25255 PR tree-optimization/56265
25256 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
25257 when target is referenced for first time.
25259 2013-02-20 Richard Biener <rguenther@suse.de>
25261 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
25262 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
25263 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
25264 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
25265 not return anything.
25266 (rename_ssa_copies): Do not remove unused locals.
25267 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
25268 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
25269 * passes.c (execute_function_todo): Do not schedule unused locals
25270 removal if cleanup_tree_cfg did something.
25271 * tree-ssa-live.c (remove_unused_locals): Dump statistics
25272 about the number of removed locals.
25274 2013-02-20 Richard Biener <rguenther@suse.de>
25276 PR tree-optimization/56398
25277 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
25279 2013-02-20 Martin Jambor <mjambor@suse.cz>
25281 PR tree-optimization/55334
25282 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
25283 restricted pointers to arrays.
25285 2013-02-20 Richard Biener <rguenther@suse.de>
25286 Jakub Jelinek <jakub@redhat.com>
25288 PR tree-optimization/56396
25289 * tree-ssa-ccp.c (n_const_val): New static variable.
25290 (get_value): Return NULL for SSA names we don't have a lattice
25292 (ccp_initialize): Initialize n_const_val.
25293 * tree-ssa-copy.c (n_copy_of): New static variable.
25294 (init_copy_prop): Initialize n_copy_of.
25295 (get_value): Return NULL_TREE for SSA names we don't have a
25298 2013-02-20 Martin Jambor <mjambor@suse.cz>
25300 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
25302 2013-02-20 Richard Biener <rguenther@suse.de>
25304 * genpreds.c (write_lookup_constraint): Do not compare first
25305 letter of the constraint again.
25307 2013-02-20 Richard Biener <rguenther@suse.de>
25309 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
25311 (get_use_iv_cost): Terminate hashtable walk when coming across
25314 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
25316 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
25317 reassociation for avx2 targets.
25319 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
25321 * config/microblaze/microblaze.c: microblaze_has_clz = 0
25322 Add version check for v8.10.a to enable microblaze_has_clz
25323 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
25324 version and TARGET_PATTERN_COMPARE check
25325 * config/microblaze/microblaze.md: New clzsi2 instruction
25327 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
25329 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
25330 function before branching.
25332 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
25334 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
25336 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
25338 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
25340 PR middle-end/55889
25341 * sel-sched.c: Include ira.h.
25342 (implicit_clobber_conflict_p): New function.
25343 (moveup_expr): Use it.
25344 * Makefile.in (sel-sched.o): Depend on ira.h.
25346 2013-02-19 Richard Biener <rguenther@suse.de>
25348 PR tree-optimization/56384
25349 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
25350 (vn_hash_type): Split out from ...
25351 (vn_hash_constant_with_type): ... here.
25352 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
25353 (vn_phi_eq): Compare types from vn_phi_s structure.
25354 (vn_phi_lookup): Populate vn_phi_s type.
25355 (vn_phi_insert): Likewise.
25357 2013-02-19 Jakub Jelinek <jakub@redhat.com>
25359 PR tree-optimization/56350
25360 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
25361 if haven't found reduction or nested cycle operand, rather than
25362 asserting we must find it.
25364 PR tree-optimization/56381
25365 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
25368 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
25369 Jakub Jelinek <jakub@redhat.com>
25372 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
25373 (swap_optab_enable): Same.
25374 (init_all_optabs): Use argument instead of global.
25375 * tree.h (struct tree_optimization_option): New field target_optabs.
25376 * expr.h (init_all_optabs): Add argument to prototype.
25377 (TREE_OPTIMIZATION_OPTABS): New.
25378 (save_optabs_if_changed): Protoize.
25379 * optabs.h: Declare this_fn_optabs.
25380 * optabs.c (save_optabs_if_changed): New.
25381 Declare this_fn_optabs.
25382 (init_optabs): Add argument to init_all_optabs() call.
25383 * function.c (invoke_set_current_function_hook): Handle per
25385 * function.h (struct function): New field optabs.
25386 * config/mips/mips.c (mips_set_mips16_mode): Handle when
25387 optimization_current_node has changed.
25388 * target-globals.h (save_target_globals_default_opts): Protoize.
25389 * target-globals.c (save_target_globals_default_opts): New.
25391 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
25394 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
25395 registers %fr12 and %fr12R as call used.
25398 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
25399 and HImode, require all displacements to be an integer multiple of
25401 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
25402 only allow QImode and HImode when reload is in progress and strict is
25403 true. Likewise for symbolic addresses. Use base14_operand to check
25404 displacements in REG+BASE addresses.
25406 2013-02-18 Richard Biener <rguenther@suse.de>
25408 PR tree-optimization/56366
25409 * tree-vect-loop.c (get_initial_def_for_induction): Properly
25410 handle sign-conversion of outer-loop initial induction value.
25412 2013-02-18 Richard Biener <rguenther@suse.de>
25414 PR middle-end/56349
25415 * cfghooks.c (merge_blocks): If we merge a latch into another
25416 block adjust references to it.
25417 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
25418 (verify_loop_structure): Verify that a recorded latch is in fact
25421 2013-02-18 Richard Biener <rguenther@suse.de>
25423 PR tree-optimization/56321
25424 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
25425 order SSA name release and virtual operand unlinking.
25427 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
25429 * config/microblaze/microblaze.md (save_stack_block): Define.
25430 (restore_stack_block): Likewise.
25432 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
25434 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
25435 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
25436 * config/microblaze/microblaze.c (microblaze_option_override):
25437 Bail out early for PIC modes when target does not support PIC.
25439 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
25441 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
25442 Replace with a microblaze version.
25443 (microblaze_trampoline_init): Adapt for microblaze.
25444 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
25447 2013-02-16 Jakub Jelinek <jakub@redhat.com>
25448 Dodji Seketeli <dodji@redhat.com>
25451 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
25452 (instrument_mem_region_access): Do not forget to always put
25453 instrumentation of the of 'base' and 'base + len' in a "if (len !=
25454 0) statement, even for cases where either 'base' or 'base + len'
25455 are not instrumented -- because they have been previously
25456 instrumented. Simplify the logic by putting all the statements
25457 instrument 'base + len' inside a sequence, and then insert that
25458 sequence right before the current insertion point. Then, to
25459 instrument 'base + len', just get an iterator on that statement.
25460 And do not forget to update the pointer to iterator the function
25461 received as argument.
25463 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
25465 PR rtl-optimization/56348
25466 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
25468 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
25470 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
25471 (clean_graph_dump_file): Pass base to start_graph_dump.
25473 2013-02-14 Richard Henderson <rth@redhat.com>
25476 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
25478 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
25480 * collect2-aix.h: Define F_LOADONLY.
25482 2013-02-14 Richard Biener <rguenther@suse.de>
25485 * varasm.c (output_constant_def_1): Get the decl representing
25486 the constant as argument.
25487 (output_constant_def): Wrap output_constant_def_1.
25488 (make_decl_rtl): Use output_constant_def_1 with the decl
25489 representing the constant.
25490 (build_constant_desc): Optionally re-use a decl already
25491 representing the constant.
25492 (tree_output_constant_def): Adjust.
25494 2013-02-14 Dodji Seketeli <dodji@redhat.com>
25497 * asan.c (instrument_builtin_call): Really put the length of the
25498 second source argument into src1_len.
25500 2013-02-13 Jakub Jelinek <jakub@redhat.com>
25502 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
25503 argument. If it is false, don't create edge from then_bb to
25505 (insert_if_then_before_iter): Pass true to it.
25506 (build_check_stmt): Pass false to it.
25507 (transform_statements): Flush hash table only on extended basic
25508 block boundaries, rather than at the beginning of every bb.
25509 Don't flush hash table on nonfreeing_call_p calls.
25510 * tree-flow.h (nonfreeing_call_p): New prototype.
25511 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
25513 2013-02-13 David S. Miller <davem@davemloft.net>
25515 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
25517 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
25520 * ira.c (max_regno_before_ira): Move from ...
25522 (fix_reg_equiv_init): Use max_regno_before_ira instead of
25525 2013-02-13 Jakub Jelinek <jakub@redhat.com>
25527 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
25529 2013-02-13 Richard Biener <rguenther@suse.de>
25532 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
25533 globals in MEM_REFs.
25535 2013-02-13 Richard Biener <rguenther@suse.de>
25537 * loop-init.c (loop_optimizer_init): Clear loop state when
25538 re-initializing preserved loops.
25539 * loop-unswitch.c (unswitch_single_loop): Return whether
25540 we unswitched the loop. Do not verify loop state here.
25541 (unswitch_loops): When we unswitched a loop discover new loops.
25543 2013-02-13 Kostya Serebryany <kcc@google.com>
25545 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
25547 * sanitizer.def: Rename __asan_init to __asan_init_v1.
25549 2013-02-12 Dodji Seketeli <dodji@redhat.com>
25551 Avoid instrumenting duplicated memory access in the same basic block
25552 * Makefile.in (asan.o): Add new dependency on hash-table.h
25553 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
25554 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
25555 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
25556 (free_mem_ref_resources, has_mem_ref_been_instrumented)
25557 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
25558 (get_mem_ref_of_assignment): New functions.
25559 (get_mem_refs_of_builtin_call): Extract from
25560 instrument_builtin_call and tweak a little bit to make it fit with
25562 (instrument_builtin_call): Use the new
25563 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
25564 of is_gimple_builtin_call.
25565 (instrument_derefs, instrument_mem_region_access): Insert the
25566 instrumented memory reference into the hash table.
25567 (maybe_instrument_assignment): Renamed instrument_assignment into
25568 this, and change it to advance the iterator when instrumentation
25569 actually happened and return true in that case. This makes it
25570 homogeneous with maybe_instrument_assignment, and thus give a
25571 chance to callers to be more 'regular'.
25572 (transform_statements): Clear the memory reference hash table
25573 whenever we enter a new BB, when we cross a function call, or when
25574 we are done transforming statements. Use
25575 maybe_instrument_assignment instead of instrumentation. No more
25576 need to special case maybe_instrument_assignment and advance the
25577 iterator after calling it; it's now handled just like
25578 maybe_instrument_call. Update comment.
25580 2013-02-13 Richard Biener <rguenther@suse.de>
25582 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
25583 Fix loop discovery code.
25585 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
25587 PR inline-asm/56148
25588 * lra-constraints.c (process_alt_operands): Match early clobber
25589 operand with itself. Check conflicts with earlyclobber only if
25590 the operand is not reloaded. Prefer to reload conflicting operand
25591 if earlyclobber and matching operands are the same.
25593 2013-02-12 Richard Biener <rguenther@suse.de>
25596 * lto-streamer-out.c (write_symbol): Do not output symbols
25597 for hard register variables.
25599 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
25602 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
25603 (umulsidi3_insn, mulsidi3_insn): New insns.
25605 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
25607 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
25608 (struct tune_params): Add vec_costs field.
25609 * config/arm/arm.c (arm_builtin_vectorization_cost)
25610 (arm_add_stmt_cost): New functions.
25611 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
25612 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
25613 (arm_default_vec_cost): New struct of type cpu_vec_costs.
25614 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
25615 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
25616 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
25617 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
25619 2013-02-12 Richard Biener <rguenther@suse.de>
25622 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
25623 decls again if possible.
25625 2013-02-12 Richard Biener <rguenther@suse.de>
25627 PR middle-end/56288
25628 * tree-ssa.c (verify_ssa_name): Fix check, move
25629 SSA_NAME_IN_FREE_LIST check up.
25631 2013-02-12 Jakub Jelinek <jakub@redhat.com>
25632 Steven Bosscher <steven@gcc.gnu.org>
25634 PR rtl-optimization/56151
25635 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
25636 equal to op0 or op1, and last_insn pattern is CODE operation
25637 with MEM dest and one of the operands matches that MEM.
25639 2013-02-11 Sriraman Tallam <tmsriram@google.com>
25641 * doc/extend.texi: Document Function Multiversioning and "default"
25642 parameter string to target attribute.
25643 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
25644 target attribute parameter is "default".
25645 (ix86_compare_version_priority): Remove checks for target attribute.
25646 (ix86_mangle_function_version_assembler_name): Change error to sorry.
25647 Remove check for target attribute equal to NULL. Add assert.
25648 (ix86_generate_version_dispatcher_body): Change error to sorry.
25650 2013-02-11 Iain Sandoe <iain@codesourcery.com>
25651 Jack Howarth <howarth@bromo.med.uc.edu>
25652 Patrick Marlier <patrick.marlier@gmail.com>
25655 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
25656 define ENDFILE_SPEC as TM_DESTRUCTOR.
25657 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
25659 2013-02-11 Alexander Potapenko <glider@google.com>
25660 Jack Howarth <howarth@bromo.med.uc.edu>
25661 Jakub Jelinek <jakub@redhat.com>
25664 * config/darwin.c (cdtor_record): Rename ctor_record.
25665 (sort_cdtor_records): Rename sort_ctor_records.
25666 (finalize_dtors): New routine to sort destructors by
25667 priority before use in assemble_integer.
25668 (machopic_asm_out_destructor): Use finalize_dtors if needed.
25670 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
25672 PR rtl-optimization/56275
25673 * simplify-rtx.c (avoid_constant_pool_reference): Check that
25674 offset is non-negative and less than cmode size before
25675 calling simplify_subreg.
25677 2013-02-11 Richard Biener <rguenther@suse.de>
25679 PR tree-optimization/56264
25680 * cfgloop.h (fix_loop_structure): Adjust prototype.
25681 * loop-init.c (fix_loop_structure): Return the number of
25682 newly discovered loops.
25683 * tree-cfgcleanup.c (repair_loop_structures): When new loops
25684 are discovered, do a full loop-closed SSA rewrite.
25686 2013-02-11 Richard Biener <rguenther@suse.de>
25688 PR tree-optimization/56273
25689 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
25691 (check_array_ref): Fix missing newline in dumps.
25692 (search_for_addr_array): Likewise.
25694 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
25696 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
25698 2013-02-09 Jakub Jelinek <jakub@redhat.com>
25701 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
25703 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
25705 PR rtl-optimization/56246
25706 * lra-constraints.c (simplify_operand_subreg): Try to reuse
25708 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
25709 constraints are satisfied.
25711 2013-02-08 Jeff Law <law@redhat.com>
25714 * emit-rtl.c (reg_is_parm_p): New function.
25715 * regs.h (reg_is_parm_p): New prototype.
25716 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
25717 callee-clobbered registers.
25719 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
25722 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
25723 If there is no implicit builtin declaration, just return NULL.
25725 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
25727 * config/i386/sse.md (FMAMODEM): New mode iterator.
25728 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
25729 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
25731 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
25733 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
25734 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
25735 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
25737 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
25739 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
25740 (microblaze*-*-elf): Likewise.
25741 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
25743 * config/microblaze/microblaze-c.c: Add builtin defines for
25744 _LITTLE_ENDIAN and _BIG_ENDIAN.
25745 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
25746 add to TARGET_DEFAULT flags.
25747 Expand ASM_SPEC and LINK_SPEC.
25748 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
25749 * config/microblaze/microblaze.md: Update extendsidi2 and
25750 movdi_internal instructions to use low-order / high-order reg
25752 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
25753 options and inversemask / mask of LITTLE_ENDIAN.
25754 * config/microblaze/t-microblaze: Expand multilib options to
25755 include mlittle-endian (le) and update exceptions patterns.
25757 2013-02-08 Jakub Jelinek <jakub@redhat.com>
25759 PR rtl-optimization/56195
25760 * lra-constraints.c (get_reload_reg): Don't reuse regs
25761 if they have smaller mode than requested, if they have
25762 wider mode than requested, try to return a SUBREG.
25764 PR tree-optimization/56250
25765 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
25766 if type is unsigned and code isn't MULT_EXPR.
25768 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
25770 PR tree-optimization/56064
25771 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
25772 bits according to mode.
25773 * fixed-value.h (fixed_from_double_int)
25774 (const_fixed_from_double_int): Adjust comments.
25776 2013-02-08 Richard Biener <rguenther@suse.de>
25779 * lto-streamer.h (struct data_in): Remove current_file, current_line
25780 and current_col members.
25781 * lto-streamer-out.c (lto_output_location): Stream changed bits
25782 en-block for efficiency.
25783 * lto-streamer-in.c (clear_line_info): Remove.
25784 (lto_input_location): Cache current file, line and column
25785 globally via local statics. Read changed bits en-block.
25786 (input_function): Do not call clear_line_info.
25787 (lto_read_body): Likewise.
25788 (lto_input_toplevel_asms): Likewise.
25790 2013-02-08 Michael Matz <matz@suse.de>
25792 PR tree-optimization/52448
25793 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
25794 (nt_call_phase): New static.
25795 (add_or_mark_expr): Only mark accesses with newer phase than any
25797 (nonfreeing_call_p): New.
25798 (nt_init_block): Update nt_call_phase, mark blocks as visited.
25799 (nt_fini_block): Keep blocks marked as visited.
25800 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
25802 2013-02-08 Richard Biener <rguenther@suse.de>
25804 * ira.c (ira): Free broken dominator information.
25806 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
25808 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
25810 2013-02-08 Marek Polacek <polacek@redhat.com>
25812 * cfgloop.c (verify_loop_structure): Add more checking of headers.
25814 2013-02-08 Richard Biener <rguenther@suse.de>
25816 PR middle-end/56181
25817 * cfgloop.h (flow_loops_find): Adjust.
25818 (bb_loop_header_p): Declare.
25819 * cfgloop.c (bb_loop_header_p): New function split out from ...
25820 (flow_loops_find): ... here. Adjust function signature,
25821 support incremental loop structure update.
25822 (verify_loop_structure): Cleanup. Verify a loop is a loop.
25823 * cfgloopmanip.c (fix_loop_structure): Move ...
25824 * loop-init.c (fix_loop_structure): ... here.
25825 (apply_loop_flags): Split out from ...
25826 (loop_optimizer_init): ... here.
25827 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
25828 in incremental mode, only remove dead loops here.
25830 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
25833 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
25834 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
25835 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
25836 (*round<mode>3.libgcc): New insns for fixed-modes.
25837 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
25838 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
25839 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
25840 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
25841 implementations. Define to __builtin_avr_absFX,
25842 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
25843 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
25844 __builtin_avr_countlsFX, respectively.
25845 * config/avr/avr-c.c (target.h): Include it.
25846 (enum avr_builtin_id): New enum.
25847 (avr_resolve_overloaded_builtin): New static function.
25848 (avr_register_target_pragmas): Use it to set
25849 targetm.resolve_overloaded_builtin.
25850 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
25851 tree nodes used by DEF_BUILTIN.
25852 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
25853 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
25854 <AVR_BUILTIN_xxBITS>: Same.
25856 2013-02-08 Richard Biener <rguenther@suse.de>
25858 * cfgloop.c (verify_loop_structure): Properly handle
25859 a loop exiting to another loop header.
25860 * ira-int.h (ira_loops): Remove.
25861 * ira.c (ira_loops): Remove.
25862 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
25863 (do_reload): Use loop_optimizer_finalize.
25864 * ira-build.c (create_loop_tree_nodes): Use get_loops and
25865 number_of_loops to access the loop tree.
25866 (more_one_region_p): Likewise.
25867 (finish_loop_tree_nodes): Likewise.
25868 (rebuild_regno_allocno_maps): Likewise.
25869 (mark_loops_for_removal): Likewise.
25870 (mark_all_loops_for_removal): Likewise.
25871 (remove_unnecessary_regions): Likewise.
25872 (ira_build): Likewise.
25873 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
25875 2013-02-08 Richard Biener <rguenther@suse.de>
25877 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
25878 * ipa-pure-const.c (analyze_function): Avoid calling
25879 mark_irreducible_loops twice.
25880 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
25882 2013-02-07 David S. Miller <davem@davemloft.net>
25884 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
25886 * var-tracking.c (vt_add_function_parameter): Test the presence of
25887 HAVE_window_save properly and do not remap argument registers when
25888 we have a leaf function.
25890 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
25893 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
25895 * config/i386/i386.c (ix86_print_operand): Ditto.
25897 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
25899 * lra-constraints.c (process_alt_operands): Fix recently added comment.
25901 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
25903 PR rtl-optimization/56225
25904 * lra-constraints.c (process_alt_operands): Check that reload hard
25905 reg can hold value for strict_low_part.
25907 2013-02-07 Jakub Jelinek <jakub@redhat.com>
25910 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
25911 dwarf2out_end_function.
25912 (in_first_function_p, maybe_at_text_label_p,
25913 first_loclabel_num_not_at_text_label): New variables.
25914 (dwarf2out_var_location): In the first function find out
25915 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
25916 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
25919 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
25921 PR rtl-optimization/56178
25922 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
25923 SUBREG of a register. Tidy up related block of code.
25924 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
25925 note if the source is a register or a SUBREG of a register.
25927 2013-02-07 Jakub Jelinek <jakub@redhat.com>
25930 * config/rs6000/rs6000.md (ptrm): New mode attr.
25931 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
25932 call_value_indirect_aix<pttrsize>,
25933 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
25936 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
25938 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
25939 if -bnortl. Convert to strcmp and strncmp.
25941 2013-02-07 Alan Modra <amodra@gmail.com>
25944 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
25945 addresses won't wrap when offsetting.
25946 (rs6000_secondary_reload): Provide secondary reloads needed for
25947 wrapping LO_SUM addresses.
25949 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
25951 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
25952 MACH, just __MACH__.
25954 2013-02-06 Richard Biener <rguenther@suse.de>
25956 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
25957 instead of calling fix_loop_structure.
25959 2013-02-06 Jakub Jelinek <jakub@redhat.com>
25961 PR middle-end/56217
25962 * omp-low.c (use_pointer_for_field): Return false if
25963 lower_send_shared_vars doesn't generate any copy-out code.
25965 2013-02-06 Tom de Vries <tom@codesourcery.com>
25967 PR rtl-optimization/56131
25968 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
25969 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
25970 of the label is NULL. Add comment.
25972 2013-02-05 Jakub Jelinek <jakub@redhat.com>
25974 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
25977 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
25978 (STATIC_LIBTSAN_LIBS): Likewise.
25979 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
25980 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
25981 is defined, don't add anything else beyond that.
25982 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
25983 (LINK_COMMAND_SPEC): Use them.
25985 PR tree-optimization/56205
25986 * tree-stdarg.c (check_all_va_list_escapes): Return true if
25987 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
25988 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
25990 2013-02-05 Richard Biener <rguenther@suse.de>
25992 PR tree-optimization/53342
25993 PR tree-optimization/53185
25994 * tree-vectorizer.h (vect_check_strided_load): Remove.
25995 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
25996 not disallow peeling for vectorized strided loads.
25997 (vect_check_strided_load): Make static and simplify.
25998 (vect_analyze_data_refs): Adjust.
25999 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
26000 correctly when vectorizing strided loads.
26002 2013-02-05 Richard Biener <rguenther@suse.de>
26004 * doc/install.texi: Refer to ISL, not PPL.
26006 2013-02-05 Jan Hubicka <jh@suse.cz>
26008 PR tree-optimization/55789
26009 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
26011 2013-02-05 Jan Hubicka <jh@suse.cz>
26013 PR tree-optimization/55789
26014 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
26015 the dead call anyway.
26017 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
26020 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
26022 2013-02-04 Alexander Potapenko <glider@google.com>
26023 Jack Howarth <howarth@bromo.med.uc.edu>
26024 Jakub Jelinek <jakub@redhat.com>
26027 * config/darwin.c (sort_ctor_records): Stabilized qsort
26028 on constructor priority by using original position.
26029 (finalize_ctors): New routine to sort constructors by
26030 priority before use in assemble_integer.
26031 (machopic_asm_out_constructor): Use finalize_ctors if needed.
26033 2013-02-04 Jakub Jelinek <jakub@redhat.com>
26036 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
26037 about visibility on artificial decls.
26038 * config/sol2.c (solaris_assemble_visibility): Likewise.
26040 2013-02-04 Kai Tietz <ktietz@redhat.com>
26043 * config/i386/i386.c (function_value_ms_64): Add additional valtype
26044 argument and improve checking of return-argument types for 16-byte
26046 (ix86_function_value_1): Add additional valtype argument on call
26047 of function_value_64.
26048 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
26049 handling infunction_value_64 function.
26051 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
26053 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
26055 2013-02-04 Richard Biener <rguenther@suse.de>
26057 PR tree-optimization/56188
26058 * tree-ssa-structalias.c (label_visit): Consider case with
26059 initially non-empty points-to set.
26060 (perform_var_substitution): Dump node mapping and clean up.
26062 2013-02-04 Richard Guenther <rguenther@suse.de>
26065 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
26066 node prevail as last resort.
26067 (lto_symtab_merge_decls): Remove guard on LTRANS here.
26068 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
26070 2013-02-04 Richard Biener <rguenther@suse.de>
26072 PR tree-optimization/56113
26073 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
26075 (equiv_class_lookup_or_add): ... this.
26076 (label_visit): Adjust and fix error in previous patch.
26077 (perform_var_substitution): Adjust.
26079 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
26081 * config/sh/divtab.c: Fix formatting and comments throughout the file.
26082 * config/sh/sh4-300.md: Likewise.
26083 * config/sh/sh4a.md: Likewise.
26084 * config/sh/constraints.md: Likewise.
26085 * config/sh/sh.md: Likewise.
26086 * config/sh/netbsd-elf.h: Likewise.
26087 * config/sh/predicates.md: Likewise.
26088 * config/sh/sh-protos.h: Likewise.
26089 * config/sh/ushmedia.h: Likewise.
26090 * config/sh/linux.h: Likewise.
26091 * config/sh/sh.c: Likewise.
26092 * config/sh/superh.h: Likewise.
26093 * config/sh/elf.h: Likewise.
26094 * config/sh/sh4.md: Likewise.
26095 * config/sh/sh.h: Likewise.
26097 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26099 * config/pa/constraints.md: Adjust unused letters. Change "T"
26100 constraint to match_test floating_point_store_memory_operand().
26101 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
26102 (base14_operand): New.
26103 (floating_point_store_memory_operand): New.
26104 (integer_store_memory_operand): Revise to use base14_operand and
26105 reg_plus_base_memory_operand.
26106 (move_dest_operand): Allow symbolic_memory_operands.
26107 (symbolic_memory_operand): Check for LO_SOM.
26108 (symbolic_operand): Change default case to break.
26109 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
26110 CONST_DOUBLE values to be reloaded by putting them into memory when
26111 the destination is a floating point register.
26112 (movdf): Remove code to handle CONST_DOUBLE.
26114 (reload_indf_r1): New.
26115 (reload_insf_r1): New.
26116 Consistently use "Q" and "T" constraints with integer and floating
26117 point move instructions, respectively.
26118 (movdi): Remove FAIL.
26119 Change predicate for source operand unamed DImode move from
26120 general_operand to move_src_operand.
26121 (umulsidi3): Change predicate for destination operand to
26123 Likewise for similar unamed patterns.
26124 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
26125 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
26126 (hppa_legitimize_address): Simplify mask calculation.
26127 (pa_emit_move_sequence): Revised handling of secondary reloads from
26128 REG+D addresses for floating point loads and stores. Directly handle
26129 loading CONST0_RTX (mode) to a floating point register.
26130 (pa_secondary_reload): Handle reloading DF and SFmode constant values
26131 to floating point registers. Don't restrict secondary reloads to
26132 floating point registers to integer modes. Revise some comments and
26134 (TARGET_LEGITIMATE_ADDRESS_P): Define.
26135 (pa_legitimate_address_p): New.
26136 (pa_legitimize_reload_address): New.
26137 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
26138 (STRICT_REG_OK_FOR_BASE_P): New.
26139 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
26140 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
26142 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
26143 Andrew Dixie <andrewd@gentrack.com>
26145 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
26148 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
26150 * expmed.c (extract_bit_field_1): Pass the full width of the
26151 structure to get_best_reg_extraction_insn.
26153 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
26156 * configure.ac (use_cxa_atexit): Add AIX.
26157 * configure: Regenerate.
26159 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
26161 2013-02-01 Jakub Jelinek <jakub@redhat.com>
26164 * final.c (need_profile_function): New variable.
26165 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
26166 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
26167 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
26168 notes, targetm.asm_out.function_prologue doesn't emit anything,
26169 HAVE_prologue and profiler should be emitted before prologue,
26170 set need_profile_function instead of emitting it.
26171 (final_scan_insn): If need_profile_function, emit
26172 profile_function on the first NOTE_INSN_BASIC_BLOCK or
26173 NOTE_INSN_FUNCTION_BEG note.
26175 2013-02-01 Richard Henderson <rth@redhat.com>
26177 * config/rs6000/rs6000.md (smulditi3): New.
26180 * config/alpha/alpha.md (umulditi3): New.
26182 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
26184 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
26185 (ASM_OUTPUT_ALIGNED_LOCAL): New.
26187 2013-02-01 Richard Biener <rguenther@suse.de>
26189 PR tree-optimization/56113
26190 * tree-ssa-structalias.c (label_visit): Reduce work for
26191 single-predecessor nodes.
26193 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
26195 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
26196 range isn't testing for zero.
26198 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
26200 PR middle-end/56113
26201 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
26203 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
26204 Nick Clifton <nickc@redhat.com>
26206 * config/v850/constraints.md (Q): Define as a memory constraint.
26207 * config/v850/predicates.md (label_ref_operand): New predicate.
26208 (e3v5_shift_operand): New predicate.
26209 (ior_operator): New predicate.
26210 * config/v850/t-v850: Add e3v5 multilib.
26211 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
26212 (v850_gen_movdi): Prototype.
26213 * config/v850/v850.c: Add support for e3v5 architecture.
26214 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
26216 (construct_save_jarl): Add e3v5 long JARL support.
26217 (v850_adjust_insn_length): New function. Adjust length of call
26218 insns when using e3v5 instructions.
26219 (v850_gen_movdi): New function: Generate instructions to move a
26221 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
26222 (CPP_SPEC): Define __v850e3v5__ as appropriate.
26223 (TARGET_USE_FPU): Enable for e3v5.
26224 (CONST_OK_FOR_W): New macro.
26225 (ADJUST_INSN_LENGTH): Define.
26226 * config/v850/v850.md (UNSPEC_LOOP): Define.
26227 (attr cpu): Add v850e3v5.
26228 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
26229 (movdi): New pattern.
26230 (movdi_internal): New pattern.
26231 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
26232 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
26233 (cstoresf4): Likewise.
26234 (cstoredf4): Likewise.
26235 (insv): New pattern.
26236 (rotlso3_a): New pattern.
26237 (rotlsi3_b): New pattern
26238 (rotlsi3_v850e3v5): New pattern.
26239 (doloop_begin): New pattern.
26240 (fix_loop_counter): New pattern.
26241 (doloop_end): New pattern.
26242 (branch_normal): Add e3v5 long branch support.
26243 (branch_invert): Likewise.
26244 (branch_z_normal): Likewise.
26245 (branch_z_invert): Likewise.
26246 (branch_nz_normal): Likewise.
26247 (branch_nz_invert): Likewise.
26248 (call_internal_short): Add e3v5 register-indirect JARL support.
26249 (call_internal_long): Likewise.
26250 (call_value_internal_short): Likewise.
26251 (call_value_internal_long): Likewise.
26252 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
26253 (mloop): New option.
26254 * config.gcc: Add support for configuring v840e3v5 target.
26255 * doc/invoke.texi: Document new v850 specific command line options.
26257 2013-01-31 Paul Koning <ni1d@arrl.net>
26261 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
26262 children if parent is a class.
26263 (prune_unused_types_prune): Don't add DW_AT_declaration.
26265 2013-01-31 Richard Biener <rguenther@suse.de>
26267 PR tree-optimization/56157
26268 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
26269 match up operand with SLP child.
26271 2013-01-31 Jason Merrill <jason@redhat.com>
26274 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
26275 parameters the first time.
26276 (gen_scheduled_generic_parms_dies): Check completeness here.
26278 2013-01-31 Richard Biener <rguenther@suse.de>
26280 PR middle-end/53073
26281 * common.opt (faggressive-loop-optimizations): New flag,
26282 enabled by default.
26283 * doc/invoke.texi (faggressive-loop-optimizations): Document.
26284 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
26285 infer_loop_bounds_from_undefined by it.
26287 2013-01-31 Richard Biener <rguenther@suse.de>
26289 PR tree-optimization/56150
26290 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
26291 visit virtual operands.
26292 (find_uses_to_rename_bb): Likewise.
26294 2013-01-31 Richard Biener <rguenther@suse.de>
26296 PR tree-optimization/56150
26297 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
26298 mixed store non-store stmts.
26300 2013-01-30 Jakub Jelinek <jakub@redhat.com>
26303 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
26304 LIBASAN_EARLY_SPEC is defined.
26305 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
26306 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
26308 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
26311 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
26312 invalid args instead of ICEing on it.
26313 (ix86_valid_target_attribute_tree): Return error_mark_node if
26314 ix86_valid_target_attribute_inner_p failed.
26315 (ix86_valid_target_attribute_p): Return false only if
26316 ix86_valid_target_attribute_tree returned error_mark_node. Allow
26317 target("default") attribute.
26318 (sorted_attr_string): Change argument from const char * to tree,
26319 merge in all target attribute arguments rather than just one.
26320 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
26321 instead of free. Avoid using strcat.
26322 (ix86_mangle_function_version_assembler_name): Mangle
26323 target("default") as if no target attribute is present. Adjust
26324 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
26325 instead of xmalloc and XDELETEVEC instead of free.
26326 (ix86_function_versions): Don't return true if one of the decls
26327 doesn't have target attribute. If they don't and one of the decls
26328 is DECL_FUNCTION_VERSIONED, report an error. Adjust
26329 sorted_attr_string caller. Use XDELETEVEC instead of free.
26330 (ix86_supports_function_versions): Remove.
26331 (make_name): Fix up formatting.
26332 (make_dispatcher_decl): Remove resolver_name and its initialization.
26333 Avoid leaking memory.
26334 (is_function_default_version): Return true if there is
26335 target("default") attribute rather than no target attribute at all.
26336 (make_resolver_func): Avoid leaking memory.
26337 (ix86_generate_version_dispatcher_body): Likewise.
26338 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
26339 * target.def (supports_function_versions): Remove.
26340 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
26341 * doc/tm.texi: Regenerated.
26343 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
26345 PR rtl-optimization/56144
26346 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
26347 for values with side effects.
26349 2013-01-30 Richard Biener <rguenther@suse.de>
26351 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
26352 (sparseset_pop): Likewise.
26353 * cfganal.c (compute_idf): Likewise. Increase work-stack size
26354 to be able to use quick_push in the worker loop.
26356 2013-01-30 Marek Polacek <polacek@redhat.com>
26358 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
26360 2013-01-30 Richard Biener <rguenther@suse.de>
26363 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
26365 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
26367 PR tree-optimization/56064
26368 * fixed-value.c (fixed_from_double_int): New function.
26369 * fixed-value.h (fixed_from_double_int): New prototype.
26370 (const_fixed_from_double_int): New static inline function.
26371 * fold-const.c (native_interpret_fixed): New static function.
26372 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
26373 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
26374 (native_encode_fixed): New static function.
26375 (native_encode_expr) <FIXED_CST>: Use it.
26376 (native_interpret_int): Move double_int worker code to...
26377 * double-int.c (double_int::from_buffer): ...this new static method.
26378 * double-int.h (double_int::from_buffer): Prototype it.
26380 2013-01-30 Richard Biener <rguenther@suse.de>
26382 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
26383 New pointer-map and obstack.
26384 (init_alias_vars): Allocate pointer-map and obstack.
26385 (delete_points_to_sets): Free them.
26386 (find_what_var_points_to): Cache result.
26387 (find_what_p_points_to): Adjust for changed interface of
26388 find_what_var_points_to.
26389 (compute_points_to_sets): Likewise.
26390 (ipa_pta_execute): Likewise.
26392 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
26394 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
26395 * configure: Regenerate.
26396 * config.in: Regenerate.
26397 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
26398 #nobits/#progbits if supported.
26400 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
26403 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
26404 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
26405 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
26407 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
26409 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
26410 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
26412 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
26414 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
26415 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
26417 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
26419 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
26421 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
26422 * config/arm/cortex-a7.md: New bypasses using
26423 arm_mac_accumulator_is_result.
26425 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
26427 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
26428 (cortex_a7_neon_mla): Likewise.
26429 (cortex_a7_fpfmad): New reservation.
26430 (cortex_a7_fpmacs): Use ffmas and update required units.
26431 (cortex_a7_fpmuld): Update required units and latency.
26432 (cortex_a7_fpmacd): Likewise.
26433 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
26434 (cortex_a7_neon). Likewise.
26435 (bypass) Update participating units.
26437 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
26439 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
26440 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
26442 * config/arm/vfp11.md (vfp_farith): Use ffmas.
26443 (vfp_fmul): Use ffmad.
26444 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
26445 (cortex_r4_fmacd): Use ffmad.
26446 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
26447 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
26448 (cortex_a9_fmacd): Use ffmad.
26449 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
26450 (cortex_a8_vfp_macd): Use ffmad.
26451 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
26452 (cortex_a5_fpmacd): Use ffmad.
26453 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
26454 (cortex_a15_vfp_macd): Use ffmad.
26455 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
26457 2013-01-29 Jason Merrill <jason@redhat.com>
26460 * varasm.c (default_assemble_visibility): Don't warn about
26461 visibility on artificial decls.
26463 2013-01-29 Richard Biener <rguenther@suse.de>
26465 PR tree-optimization/56113
26466 * tree-ssa-structalias.c (equiv_class_lookup): Also return
26468 (label_visit): Free duplicate bitmaps and record the leader instead.
26469 (perform_var_substitution): Adjust.
26471 2013-01-29 Richard Biener <rguenther@suse.de>
26473 PR tree-optimization/55270
26474 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
26475 the CFG, schedule loops for fixup.
26477 2013-01-29 Nick Clifton <nickc@redhat.com>
26479 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
26482 2013-01-28 Leif Ekblad <leif@rdos.net>
26484 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
26485 * config/i386/i386.h (TARGET_RDOS): New macro.
26486 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
26487 * config/i386/i386.c (ix86_option_override_internal): For 64bit
26488 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
26489 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
26490 DEFAULT_LARGE_SECTION_THRESHOLD.
26491 * config/i386/i386.md (R14_REG, R15_REG): New constants.
26492 * config/i386/rdos.h: New file.
26493 * config/i386/rdos64.h: New file.
26495 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
26498 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
26501 2013-01-28 Jakub Jelinek <jakub@redhat.com>
26503 PR rtl-optimization/56117
26504 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
26505 call cselib_lookup_from_insn on the MEM before calling
26506 add_insn_mem_dependence.
26508 2013-01-28 Richard Biener <rguenther@suse.de>
26510 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
26511 to a stmt that didn't have one.
26512 (copy_phis_for_bb): Likewise for PHI arguments.
26513 (copy_debug_stmt): Likewise for debug stmts.
26515 2013-01-28 Richard Biener <rguenther@suse.de>
26517 PR tree-optimization/56034
26518 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
26519 (partition_builtin_p): Adjust.
26520 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
26521 it is the last partition.
26522 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
26523 up the vertex for the definition.
26524 (classify_partition): Classify whether a partition is a
26525 PKIND_REDUCTION, thus has uses outside of the loop.
26526 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
26527 Merge all PKIND_REDUCTION partitions into the last partition.
26528 (tree_loop_distribution): Seed partitions from reductions as well.
26530 2013-01-28 Jakub Jelinek <jakub@redhat.com>
26532 PR tree-optimization/56125
26533 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
26534 pow(x,c) into sqrt(x) * powi(x, n/2) or
26535 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
26536 optimizing for size.
26537 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
26538 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
26541 PR tree-optimization/56094
26542 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
26543 to UNKNOWN_LOCATION while gimplifying expr.
26545 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
26548 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
26549 operand 0 in movabs insn template for -masm=intel asm alternative.
26550 (*movabs<mode>_2): Ditto for operand 1.
26552 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
26555 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
26558 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
26560 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
26563 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
26565 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
26566 Undef to avoid warning.
26568 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
26570 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
26571 * configure: Regenerate.
26573 2013-01-25 Jakub Jelinek <jakub@redhat.com>
26575 PR tree-optimization/56098
26576 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
26577 for stmts with volatile ops.
26578 (cond_store_replacement): Don't optimize if assign has volatile ops.
26579 (cond_if_else_store_replacement_1): Don't optimize if either
26580 then_assign or else_assign have volatile ops.
26581 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
26584 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
26586 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
26588 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
26590 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
26591 missing ':' in asm example.
26593 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
26595 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
26596 entries into lane and laneq entries.
26597 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
26598 Remove AdvSIMD scalar modes.
26599 (aarch64_sq<r>dmulh_laneq<mode>): New.
26600 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
26602 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
26603 builtin implementations to relfect changes in RTL in aarch64-simd.md.
26604 * config/aarch64/iterators.md (VCOND): New.
26607 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
26610 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
26611 Add NULL LIBNAME argument to existing definitions.
26612 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
26613 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
26614 * config/avr/avr.c (DEF_BUILTIN): Same.
26615 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
26616 (avr_expand_builtin): Expand to a vanilla call if a libgcc
26617 implementation is available (DECL_ASSEMBLER_NAME is set).
26618 (avr_fold_absfx): New static function.
26619 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
26620 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
26621 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
26622 AVR_BUILTIN_ABSLLK.
26623 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
26624 (abshk, absk, abslk, absllk): Provide as static inline functions.
26626 2013-01-25 Marek Polacek <polacek@redhat.com>
26628 PR tree-optimization/56035
26629 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
26631 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
26633 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
26634 (*movtf_internal_rex64): Add (!o,C) alternative
26635 (*movxf_internal_rex64): Ditto.
26636 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
26638 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
26640 * doc/invoke.texi: fix typo.
26641 * doc/objc.texi: fix typo.
26643 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
26645 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
26646 for the first two alternatives.
26648 2013-01-24 Diego Novillo <dnovillo@google.com>
26650 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
26651 (ggc-zone.o): Remove.
26652 * configure.ac: Remove option --with-gc.
26653 * configure: Re-generate.
26654 * doc/install.texi: Remove documentation for --with-gc.
26655 * gengtype.c (write_enum_defn): Remove. Update all users.
26656 (write_Types_process_field): Remove generation of gt_e_* argument.
26657 (output_type_enum): Remove. Update all users.
26658 (write_enum_defn): Remove. Update all users.
26659 (enum alloc_zone): Remove. Update all users.
26660 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
26661 * ggc-common.c (ggc_splay_alloc): Remove first argument.
26662 Update all callers.
26663 (struct ptr_data): Remove field TYPE. Update all users.
26664 (gt_pch_note_object): Remove argument TYPE. Update all users.
26665 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
26667 * ggc-none.c (ggc_alloc_typed_stat): Remove.
26668 (struct alloc_zone): Remove.
26669 (ggc_internal_alloc_zone_stat): Remove.
26670 (ggc_internal_cleared_alloc_zone_stat): Remove.
26671 * ggc-page.c (ggc_alloc_typed_stat): Remove.
26672 (ggc_pch_count_object): Remove last argument. Update all users.
26673 (ggc_pch_alloc_object): Remove last argument. Update all users.
26674 (struct alloc_zone): Remove.
26675 * ggc-zone.c: Remove.
26676 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
26677 (struct alloc_zone): Remove.
26678 (ggc_alloc_typed_stat): Remove.
26679 (ggc_alloc_typed): Remove.
26680 (ggc_splay_alloc): Remove first argument.
26681 (rtl_zone): Remove. Update all users.
26682 (tree_zone): Remove. Update all users.
26683 (tree_id_zone): Remove. Update all users.
26684 (ggc_internal_zone_alloc_stat): Remove. Update all users.
26685 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
26686 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
26687 * tree-ssanames.c: Remove references to zone allocator in comments.
26689 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
26691 * config/avr/avr.c (avr_out_fract): Make register numbers that
26692 might be outside of source operand signed.
26694 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
26696 * config/i386/constraints.md (Yf): New constraint.
26697 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
26698 of f constraint to conditionaly disable x87 register preferences.
26699 (*movdf_internal): Ditto.
26700 (*movsf_internal): Ditto.
26702 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
26704 PR inline-asm/55934
26705 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
26706 that have operands with impossible constraints.
26707 Add a FIXME for a speed-up opportunity.
26708 * lra-constraints.c (process_alt_operands): Verify that a class
26709 selected from constraints on asms is valid for the operand mode.
26710 (curr_insn_transform): Remove incorrect comment.
26712 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
26714 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
26715 TOC operand is a valid symbol ref in the constant pool.
26717 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
26719 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
26721 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
26724 * config/avr/stdfix.h: New file.
26725 * t-avr (stdfix-gcc.h): New rule to build it.
26726 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
26728 2013-01-23 Kostya Serebryany <kcc@google.com>
26730 * config/darwin.h: remove dependency on
26731 CoreFoundation (asan on Mac OS).
26733 2013-01-23 Jakub Jelinek <jakub@redhat.com>
26736 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
26737 instead of cmpdi_operand for first comparison operand.
26738 Don't assert that comparison operands aren't both constants.
26740 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
26742 * doc/install.texi (Downloading the Source): Update references to
26743 downloading separate components.
26745 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
26747 * doc/extend.texi (__int128): Improve grammar.
26749 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
26752 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
26753 alternative to (o,r).
26754 (*movdi_internal_rex64): Remove (!o,n) alternative.
26755 (DImode immediate->memory splitter): Remove.
26756 (DImode immediate->memory peephole2): Remove.
26757 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
26758 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
26759 alternative to (!o,*r).
26760 (*movtf_internal_sse): New pattern.
26761 (*movxf_internal_rex64): New pattern.
26762 (*movxf_internal): Disable for TARGET_64BIT.
26763 (*movdf_internal_rex64): Remove (!o,F) alternative.
26765 2013-01-22 Jakub Jelinek <jakub@redhat.com>
26767 PR middle-end/56074
26768 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
26769 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
26770 * tree-vect-loop-manip.c (find_loop_location): Also ignore
26771 stmt locations where LOCATION_LOCUS of the stmt location is
26772 UNKNOWN_LOCATION or BUILTINS_LOCATION.
26775 * config/i386/i386.md (UNSPEC_STOS): New.
26776 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
26777 *strsetqi_1): Add UNSPEC_STOS.
26779 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
26782 * doc/invoke.texi: Remove left over -Wsynth example.
26784 2013-01-21 Jakub Jelinek <jakub@redhat.com>
26786 PR tree-optimization/56051
26787 * fold-const.c (fold_binary_loc): Don't fold
26788 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
26789 a narrowing conversion, or widening conversion from signed
26792 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
26794 PR rtl-optimization/56023
26795 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
26796 dependent on debug instruction.
26798 2013-01-21 Martin Jambor <mjambor@suse.cz>
26800 PR middle-end/56022
26801 * function.c (allocate_struct_function): Call
26802 invoke_set_current_function_hook earlier.
26804 2013-01-21 Jakub Jelinek <jakub@redhat.com>
26806 * reload1.c (init_reload): Only initialize reload_obstack
26807 during the first call.
26809 2013-01-21 Marek Polacek <polacek@redhat.com>
26811 * cfgloop.c (verify_loop_structure): Fix up grammar.
26813 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
26815 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
26816 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
26818 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
26821 * config/arm/marvell-pj4.md: Update copyright year.
26822 Fix up use of alu to alu_reg and simple_alu_imm.
26824 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
26826 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
26828 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
26831 * lra-constraints.c (curr_insn_transform): Don't reuse original
26832 insn for secondary memory move when memory mode should be different.
26834 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26836 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
26837 atomic_storedi_1): New patterns.
26839 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
26841 btver2 pipeline descriptions.
26842 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
26844 * config/i386/i386.md (btver2_decode): New type attributes.
26845 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
26847 * config/i386/btver2.md: New file describing btver2 pipelines.
26849 2013-01-19 Andrew Pinski <apinski@cavium.com>
26851 PR tree-optimization/52631
26852 * tree-ssa-sccvn (visit_use): Before looking up the original
26853 statement, try looking up the simplified expression.
26855 2013-01-19 Anthony Green <green@moxielogic.com>
26857 * config/moxie/moxie.c (moxie_expand_prologue): Set
26858 current_function_static_stack_size.
26860 2013-01-18 Jakub Jelinek <jakub@redhat.com>
26862 PR tree-optimization/56029
26863 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
26864 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
26866 2013-01-18 Sharad Singhai <singhai@google.com>
26868 PR tree-optimization/55995
26869 * dumpfile.c (dump_loc): Print location only if available.
26870 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
26872 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
26875 * lra-constraints.c (curr_insn_transform): Reuse original insn for
26876 secondary memory move.
26877 (inherit_reload_reg): Use rclass instead of cl for
26878 check_secondary_memory_needed_p.
26880 2013-01-18 Jakub Jelinek <jakub@redhat.com>
26882 PR middle-end/56015
26883 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
26884 the case where writing real complex part of target modifies op1.
26886 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
26888 * config/aarch64/aarch64-simd.md
26889 (aarch64_vcond_internal<mode>): Handle unordered cases.
26890 * config/aarch64/iterators.md (v_cmp_result): New.
26892 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
26893 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
26895 * config/arm/marvell-pj4.md: New file.
26896 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
26897 * config/arm/arm.md (generic_sched): Add marvell_pj4.
26898 (generic_vfp): Likewise.
26899 * config/arm/arm-cores.def: Add marvell-pj4.
26900 * config/arm/arm-tune.md: Regenerate.
26901 * config/arm/arm-tables.opt: Regenerate.
26902 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
26903 * doc/invoke.texi: Document marvell-pj4.
26905 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
26907 * config/aarch64/arm_neon.h: Map scalar types to standard types.
26909 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
26914 * var-tracking.c (negative_power_of_two_p): New.
26915 (global_get_addr_cache, local_get_addr_cache): New.
26916 (get_addr_from_global_cache, get_addr_from_local_cache): New.
26917 (vt_canonicalize_addr): Rewrite using the above. Adjust the
26919 (vt_stack_offset_p): Remove.
26920 (vt_canon_true_dep): Always canonicalize loc's address.
26921 (clobber_overlapping_mems): Make sure we have a MEM.
26922 (local_get_addr_clear_given_value): New.
26923 (val_reset): Clear local cached entries.
26924 (compute_bb_dataflow): Create and release the local cache.
26925 Disable duplicate MEMs clobbering.
26926 (emit_notes_in_bb): Clobber MEMs likewise.
26927 (vt_emit_notes): Create and release the local cache.
26928 (vt_initialize, vt_finalize): Create and release the global
26929 cache, respectively.
26930 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
26932 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
26934 PR libmudflap/53359
26935 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
26936 not found in the symtab.
26938 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
26941 PR rtl-optimization/55547
26942 PR rtl-optimization/53827
26945 * alias.c (offset_overlap_p): New, factored out of...
26946 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
26947 the conservative special case for symbolic constants. Don't
26948 adjust zero sizes on alignment.
26950 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
26952 PR rtl-optimization/52573
26953 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
26954 REG_UNUSED for the same register.
26956 2013-01-17 Richard Biener <rguenther@suse.de>
26957 Marek Polacek <polacek@redhat.com>
26959 PR rtl-optimization/55833
26960 * loop-unswitch.c (unswitch_loops): Move loop verification...
26961 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
26962 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
26963 Set it to true when we're removing a loop from hierarchy tree in
26964 an irreducible region.
26965 (fix_bb_placements): Adjust caller.
26966 (fix_loop_placements): Likewise.
26968 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
26970 * config/avr/builtins.def (DEF_BUILTIN): Factor out
26971 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
26972 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
26973 Remove ID. Adjust comments.
26974 * config/avr/avr-c.c (avr_builtin_name): Remove.
26975 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
26976 * config/avr/avr.c (avr_tolower): New static function.
26977 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
26978 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
26979 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
26982 2013-01-17 Jan Hubicka <jh@suse.cz>
26984 PR tree-optimization/55273
26985 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
26987 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
26990 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
26991 store through atomic_store<mode>_1.
26992 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
26994 2013-01-17 Martin Jambor <mjambor@suse.cz>
26996 PR tree-optimizations/55264
26997 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
26998 for virtual methods.
26999 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
27000 virtual methods before inlining is over.
27001 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
27003 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
27006 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
27008 PR rtl-optimization/56005
27009 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
27010 pending reads for prefetch.
27012 2013-01-16 Ian Bolton <ian.bolton@arm.com>
27014 * config/aarch64/aarch64.md
27015 (*cstoresi_neg_uxtw): New pattern.
27016 (*cmovsi_insn_uxtw): New pattern.
27017 (*<optab>si3_uxtw): New pattern.
27018 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
27019 (*<optab>si3_insn_uxtw): New pattern.
27020 (*bswapsi2_uxtw): New pattern.
27022 2013-01-16 Richard Biener <rguenther@suse.de>
27024 * tree-inline.c (tree_function_versioning): Remove set but
27025 never used variable.
27027 2013-01-16 Richard Biener <rguenther@suse.de>
27029 PR tree-optimization/55964
27030 * tree-flow.h (rename_variables_in_loop): Remove.
27031 (rename_variables_in_bb): Likewise.
27032 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
27033 (copy_loop_before): Adjust and delete update-ssa status.
27034 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
27035 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
27036 (rename_variables_in_loop): Remove.
27037 (slpeel_update_phis_for_duplicate_loop): Likewise.
27038 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
27039 use available cfg machinery instead of duplicating it.
27040 Update PHI nodes and perform poor-mans SSA update here.
27041 (slpeel_tree_peel_loop_to_edge): Adjust.
27043 2013-01-16 Richard Biener <rguenther@suse.de>
27045 PR tree-optimization/54767
27046 PR tree-optimization/53465
27047 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
27048 (vrp_visit_phi_node): For PHI arguments coming via backedges
27049 drop all symbolical range information.
27050 (execute_vrp): Compute backedges.
27052 2013-01-16 Richard Biener <rguenther@suse.de>
27054 * doc/install.texi: Update CLooG and ISL requirements to
27057 2013-01-16 Christian Bruel <christian.bruel@st.com>
27060 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
27061 (broken_move): Handle UNSPECV_SP_SWITCH_B.
27062 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
27064 2013-01-16 DJ Delorie <dj@redhat.com>
27066 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
27067 (UNSPECV_SP_SWITCH_E): New.
27068 (sp_switch_1): Change to an unspec.
27069 (sp_switch_2): Change to an unspec. Don't use post-inc when we
27072 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
27074 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
27075 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
27076 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
27077 (expand_mem_thread_fence): Ditto.
27078 (expand_mem_signal_fence): Ditto.
27079 (expand_atomic_load): Ditto.
27080 (expand_atomic_store): Ditto.
27082 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
27084 PR rtl-optimization/55547
27085 PR rtl-optimization/53827
27088 * alias.c (memrefs_conflict_p): Set sizes to negative after
27091 2013-01-15 Jakub Jelinek <jakub@redhat.com>
27094 * function.c (thread_prologue_and_epilogue_insns): Always
27095 add crtl->drap_reg to set_up_by_prologue.set, even if
27096 stack_realign_drap is false.
27098 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
27100 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
27101 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
27102 *call): Fix indention.
27104 2013-01-15 Tom de Vries <tom@codesourcery.com>
27107 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
27110 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
27112 PR rtl-optimization/55153
27113 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
27115 2013-01-15 Martin Jambor <mjambor@suse.cz>
27117 PR tree-optimization/55920
27118 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
27119 accesses as grp_to_be_debug_replaced.
27121 2013-01-15 Jakub Jelinek <jakub@redhat.com>
27123 PR tree-optimization/55920
27124 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
27125 there is non-useless type conversion needed from debug rhs to lhs,
27126 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
27128 2013-01-15 Joseph Myers <joseph@codesourcery.com>
27129 Mikael Pettersson <mikpe@it.uu.se>
27132 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
27134 (ASM_OUTPUT_CASE_LABEL): Remove.
27135 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
27136 * final.c (shorten_branches): Update alignment of labels before
27137 jump tables if CASE_VECTOR_SHORTEN_MODE.
27139 2013-01-15 Richard Biener <rguenther@suse.de>
27142 * system.h: Do not include gmp.h for building host tools.
27144 2013-01-15 Richard Biener <rguenther@suse.de>
27146 PR middle-end/55882
27147 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
27148 account for bitpos when computing alignment.
27150 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
27152 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
27153 (ix86_target_macros_internal): Likewise.
27155 * config/i386/i386.c (m_CORE2I7): Removed.
27156 (m_CORE_HASWELL): New macro.
27157 (m_CORE_ALL): Likewise.
27158 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
27159 (initial_ix86_arch_features): Likewise.
27160 (processor_target_table): Initializations for Core avx2.
27161 (cpu_names): New names "core-avx2".
27162 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
27163 PROCESSOR_CORE_HASWELL.
27164 (ix86_issue_rate): New case.
27165 (ia32_multipass_dfa_lookahead): Likewise.
27166 (ix86_sched_init_global): Likewise.
27168 * config/i386/i386.h (TARGET_HASWELL): New macro.
27169 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
27170 (processor_type): New PROCESSOR_HASWELL.
27172 2013-01-15 Jakub Jelinek <jakub@redhat.com>
27174 PR tree-optimization/55955
27175 * tree-vect-loop.c (vectorizable_reduction): Give up early on
27176 *SHIFT_EXPR and *ROTATE_EXPR codes.
27178 PR tree-optimization/48766
27179 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
27180 -ftrapv disable -fwrapv.
27182 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
27185 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
27186 etc. to 1 and not to __flash.
27187 Use LL suffix for __INT24_MAX__ with -mint8.
27188 Use ULL suffix for __UINT24_MAX__ with -mint8.
27190 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
27192 * config/avr/avr-arch.h
27193 (struct base_arch_s): Use typedef avr_arch_t instead.
27194 (struct arch_info_s): Use typedef avr_arch_info_t instead.
27195 (struct mcu_type_s): Use typedef avr_mcu_t instead.
27196 * config/avr/avr.c: Same.
27197 * config/avr/avr-devices.c: Same.
27198 * config/avr/driver-avr.c: Same.
27199 * config/avr/gen-avr-mmcu-texi.c: Same.
27200 * config/avr/avr-mcus.def: Adjust comment.
27202 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
27204 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
27205 * config/aarch64/iterators.md (VALLDI): New.
27207 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
27208 Andi Kleen <ak@linux.intel.com>
27211 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
27212 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
27215 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
27217 * config/avr/avr-stdint.h: Remove trailing blanks.
27218 * config/avr/avr-log.h: Same.
27219 * config/avr/avr-arch.h: Same.
27220 * config/avr/avr-devices.c: Same.
27221 * config/avr/avr-dimode.md: Same.
27222 * config/avr/predicates.md: Same.
27223 * config/avr/avr-c.c: Same. And fix typo.
27225 * config/avr/avr-protos.h: Same. And:
27226 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
27227 (init_cumulative_args): Rename to avr_init_cumulative_args.
27228 (expand_prologue): Rename to avr_expand_prologue.
27229 (expand_epilogue): Rename to avr_expand_epilogue.
27230 (adjust_insn_length): Rename to avr_adjust_insn_length.
27231 (notice_update_cc): Rename to avr_notice_update_cc.
27232 (final_prescan_insn): Rename to avr_final_prescan_insn.
27233 * config/avr/avr.c: Same.
27234 * config/avr/avr.h: Same.
27235 * config/avr/avr.md: Remove trailing blanks.
27236 (prologue): Use avr_expand_prologue.
27237 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
27239 2013-01-14 Richard Biener <rguenther@suse.de>
27241 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
27242 verify_location, collect_subblocks): New functions.
27243 (verify_gimple_in_cfg): Verify that locations only reference
27244 BLOCKs in the functions BLOCK tree.
27246 2013-01-14 Richard Biener <rguenther@suse.de>
27248 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
27250 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
27252 (insert_out_of_ssa_copy_on_edge): Likewise.
27253 (rewrite_close_phi_out_of_ssa): Likewise.
27254 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
27256 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
27257 propagated constants.
27258 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
27261 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
27263 * config/avr/avr-modes.def: Add GPL copyright notice.
27265 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
27267 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
27268 MEMMODEL_MASK to determine memory model.
27269 (atomic_store<mode>): Ditto from operands[2].
27270 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
27272 2013-01-13 Jakub Jelinek <jakub@redhat.com>
27275 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
27276 (fold_gimple_assign): Don't call unshare_expr here.
27277 (fold_ctor_reference): Call unshare_expr.
27279 2013-01-13 Terry Guo <terry.guo@arm.com>
27281 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
27282 * doc/fragments.texi: Document MULTILIB_REUSE.
27283 * gcc.c (multilib_reuse): New internal spec.
27284 (set_multilib_dir): Also search multilib from multilib_reuse.
27285 * genmultilib (tmpmultilib3): Refactor code.
27286 (tmpmultilib4): Ditto.
27287 (multilib_reuse): New multilib argument.
27289 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
27291 * Makefile.in: Update copyright.
27293 2013-01-12 Tom de Vries <tom@codesourcery.com>
27295 PR middle-end/55890
27296 * calls.c (expand_call): Check if arg_nr is valid.
27298 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
27300 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
27301 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
27302 documentation. Add missing '__' in front of
27303 __builtin_ia32_packssdw256.
27305 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
27308 * config/s390/s390.c (s390_preferred_reload_class): Do not return
27309 NO_REGS for larl operands.
27310 (s390_reload_larl_operand): Use s390_load_address instead of
27313 2013-01-11 Richard Biener <rguenther@suse.de>
27315 * tree-cfg.c (verify_node_sharing_1): Split out from ...
27316 (verify_node_sharing): ... here.
27317 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
27319 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
27321 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
27322 Substitute TREECHECKING.
27323 * configure: Regenerate.
27324 * Makefile.in (TREECHECKING): New.
27326 2013-01-11 Richard Guenther <rguenther@suse.de>
27328 PR tree-optimization/44061
27329 * tree-vrp.c (extract_range_basic): Compute zero as
27330 value-range for __builtin_constant_p of function parameters.
27332 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
27334 Update copyright years.
27336 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
27338 PR rtl-optimization/55672
27339 * lra-eliminations.c (mark_not_eliminable): Permit addition with
27340 const to be eliminable.
27342 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
27344 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
27345 * configure: Regenerate.
27347 2013-01-10 Richard Biener <rguenther@suse.de>
27349 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
27351 2013-01-10 Richard Biener <rguenther@suse.de>
27354 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
27355 locations for virtual PHI arguments.
27356 (rewrite_update_phi_arguments): Likewise.
27358 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
27360 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
27363 2013-01-10 Jakub Jelinek <jakub@redhat.com>
27365 PR tree-optimization/55921
27366 * tree-complex.c (expand_complex_asm): New function.
27367 (expand_complex_operations_1): Call it for GIMPLE_ASM.
27369 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
27372 * config/s390/s390.c (s390_symref_operand_p)
27373 (s390_loadrelative_operand_p): Merge the two functions.
27374 (s390_check_qrst_address, print_operand_address): Add parameters
27375 to s390_loadrelative_operand_p invokation.
27376 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
27377 (s390_reload_larl_operand, s390_secondary_reload): Use
27378 s390_loadrelative_operand_p instead of s390_symref_operand_p.
27379 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
27381 2013-01-09 Mike Stump <mikestump@comcast.net>
27383 * dse.c (record_store): Remove unnecessary assert.
27385 2013-01-09 Jan Hubicka <jh@suse.cz>
27387 PR tree-optimization/55569
27388 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
27389 * cfgloop.h (scale_loop_profile): Likewise.
27391 2013-01-09 Jan Hubicka <jh@suse.cz>
27394 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
27396 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
27398 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
27400 PR middle-end/55114
27401 * expr.h (maybe_emit_group_store): Declare.
27402 * expr.c (maybe_emit_group_store): New function.
27403 * builtins.c (expand_builtin_int_roundingfn): Call it.
27404 (expand_builtin_int_roundingfn_2): Likewise.
27406 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
27408 PR rtl-optimization/55829
27409 * lra-constraints.c (match_reload): Add code for absent output.
27410 (curr_insn_transform): Add code for reloads of matched inputs
27413 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
27415 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
27416 attribute of movddup insn to DF.
27417 (*vec_interleave_lowv2df): Ditto.
27418 (vec_dupv2df): Ditto.
27420 2013-01-09 Jan Hubicka <jh@suse.cz>
27422 PR tree-optimiation/55875
27423 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
27424 EVERY_ITERATION parameter.
27425 (number_of_iterations_exit): Check if exit is executed every iteration.
27426 (idx_infer_loop_bounds): Similarly here.
27427 (n_of_executions_at_most): Simplify
27428 to only test for cases where statement is dominated by the
27429 particular bound; handle correctly the "postdominance" test.
27430 (scev_probably_wraps_p): Use max loop iterations info
27431 as a global bound first.
27433 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
27434 Nick Clifton <nickc@redhat.com>
27436 * config/v850/v850.md (cbranchsf4): New pattern.
27437 (cstoresf4): New pattern.
27438 (cbranchdf4): New pattern.
27439 (cstoredf4): New pattern.
27440 (movsicc): Disallow floating point comparisons.
27441 (cmpsf_le_insn): Fix order of operators.
27442 (cmpsf_lt_insn): Likewise.
27443 (cmpsf_eq_insn): Likewise.
27444 (cmpdf_le_insn): Likewise.
27445 (cmpdf_lt_insn): Likewise.
27446 (cmpdf_eq_insn): Likewise.
27447 (cmpsf_ge_insn): Use LE comparison.
27448 (cmpdf_ge_insn): Likewise.
27449 (cmpsf_gt_insn): Use LT comparison.
27450 (cmpdf_gt_insn): Likewise.
27451 (cmpsf_ne_insn): Delete pattern.
27452 (cmpdf_ne_insn): Delete pattern.
27453 * config/v850/v850.c (v850_gen_float_compare): Use
27454 gen_cmpdf_eq_insn for NE comparison.
27455 (v850_float_z_comparison_operator)
27456 (v850_float_nz_comparison_operator): Move from here ...
27457 * config/v850/predicates.md: ... to here. Move GT and GE
27458 comparisons into v850_float_z_comparison_operator.
27459 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
27461 (v850_float_nz_comparison_operator): Likewise.
27463 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
27465 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
27466 with calls to gen_insvsi/gen_insvdi.
27468 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
27470 * config/i386/i386.c (initial_ix86_tune_features): Set up
27471 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
27473 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
27474 Jakub Jelinek <jakub@redhat.com>
27476 PR tree-optimization/48189
27477 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
27478 If nitercst is 0, don't predict the exit edge.
27480 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
27482 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
27483 in asm_fprintf with reg_names.
27484 (aarch64_print_operand_address): Likewise.
27485 (aarch64_return_addr): Likewise.
27486 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
27488 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
27490 * config/pa/pa.h (VAL_U6_BITS_P): Define.
27491 (INT_U6_BITS): Likewise.
27492 * config/pa/predicates.md (uint6_operand): New predicate.
27493 (shift5_operand, shift6_operand): Likewise.
27494 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
27496 (lshrdi3): Use shift6_operand.
27497 (shrpsi4, shrpdi4): New insn patterns.
27498 (extzv): Delete expander.
27499 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
27500 predicates in unamed zero extract patterns. Tighten common constraint.
27501 (extv): Delete expander.
27502 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
27503 predicates in unamed sign extract patterns. Tighten common constraint.
27504 (insv): Delete expander.
27505 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
27506 predicates in unamed insert patterns. Tighten common constraint.
27507 Change uint32_operand predicate to uint6_operand predicate in unamed
27508 DImode pattern to insert constant values of type 1...1xxxx.
27510 2013-01-04 Jan Hubicka <jh@suse.cz>
27512 PR tree-optimization/55823
27513 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
27516 2013-01-08 Jakub Jelinek <jakub@redhat.com>
27517 Uros Bizjak <ubizjak@gmail.com>
27519 PR rtl-optimization/55845
27520 * df-problems.c (can_move_insns_across): Stop scanning at
27521 volatile_insn_p source instruction or give up if
27522 across_from .. across_to range contains any volatile_insn_p
27525 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
27527 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
27528 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
27530 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
27531 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
27533 2013-01-08 Jakub Jelinek <jakub@redhat.com>
27536 * asan.c (asan_clear_shadow): New function.
27537 (asan_emit_stack_protection): Use it.
27539 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
27541 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
27542 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
27543 with tab instead of space.
27545 2013-01-08 Nick Clifton <nickc@redhat.com>
27547 * config/rl78/rl78.c (rl78_expand_prologue): Always select
27548 register bank 0 at the start of an interrupt handler.
27549 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
27552 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
27554 * config/aarch64/aarch64-simd.md
27555 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
27556 (aarch64_simd_bsl): Likewise.
27557 (aarch64_vcond_internal<mode>): Likewise.
27558 (vcond<mode><mode>): Likewise.
27559 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
27560 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
27562 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
27564 * config/aarch64/aarch64-builtins.c
27565 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
27567 2013-01-08 Martin Jambor <mjambor@suse.cz>
27570 * tree-sra.c (analyze_access_subtree): Return true also after
27571 potentially creating a debug-only replacement.
27573 2013-01-08 Jakub Jelinek <jakub@redhat.com>
27575 PR middle-end/55890
27576 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
27578 PR tree-optimization/54120
27579 * tree-vrp.c (range_fits_type_p): Don't allow
27580 src_precision < precision from signed vr to unsigned_p
27581 if vr->min or vr->max is negative.
27582 (simplify_float_conversion_using_ranges): Test can_float_p
27583 against CODE_FOR_nothing.
27585 2013-01-08 Jakub Jelinek <jakub@redhat.com>
27586 Richard Biener <rguenther@suse.de>
27588 PR middle-end/55851
27589 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
27590 types instead of just INTEGER_TYPE types.
27592 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
27594 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
27597 2013-01-07 Steve Ellcey <sellcey@mips.com>
27600 * config/mips/mips.opt: Change mad to mmad to match documentation.
27602 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
27605 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
27606 .progmemx.data now.
27608 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
27611 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
27612 (avr_addrspace_t): Add .section_name field.
27613 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
27615 (avr_addrspace): Same. Initialize .section_name. Remove last
27616 NULL entry. Put __memx into .progmemx.data.
27617 (progmem_section_prefix): Remove.
27618 (avr_asm_init_sections): No need to initialize progmem_section.
27619 (avr_asm_named_section): Use avr_addrspace[].section_name to get
27620 section name prefix.
27621 (avr_asm_select_section): Ditto. And use get_unnamed_section to
27622 retrieve the progmem section.
27623 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
27624 boundary to run over avr_addrspace[].
27625 (avr_register_target_pragmas): Ditto.
27627 2013-01-06 Jakub Jelinek <jakub@redhat.com>
27629 * varasm.c (output_constant_def_contents): For asan_protect_global
27630 protected strings, adjust DECL_ALIGN if needed, before testing for
27632 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
27633 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
27635 (output_object_block): For asan protected decls, emit asan padding
27636 after their contents.
27637 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
27638 (asan_finish_file): Test it here instead.
27640 2013-01-07 Nick Clifton <nickc@redhat.com>
27641 Matthias Klose <doko@debian.org>
27642 Doug Kwan <dougkwan@google.com>
27643 H.J. Lu <hongjiu.lu@intel.com>
27646 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
27648 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
27650 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
27652 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
27654 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
27656 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
27659 * doc/install.texi (Cross-Compiler-Specific Options): Document
27662 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
27664 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
27665 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
27666 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
27667 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
27668 vqmovun_high_s64): Fix source operand number and update copyright.
27670 2013-01-07 Richard Biener <rguenther@suse.de>
27672 PR middle-end/55890
27673 * gimple.h (gimple_call_builtin_p): New overload.
27674 * gimple.c (validate_call): New function.
27675 (gimple_call_builtin_p): Likewise.
27676 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
27677 Use gimple_call_builtin_p.
27678 (find_func_clobbers): Likewise.
27679 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
27680 (strlen_optimize_stmt): Likewise.
27682 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
27684 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
27685 (vld1q_dup_*): Likewise.
27686 (vld1_*): Likewise.
27687 (vld1q_*): Likewise.
27688 (vld1_lane_*): Likewise.
27689 (vld1q_lane_*): Likewise.
27691 2013-01-07 Richard Biener <rguenther@suse.de>
27693 * lto-streamer.h (LTO_minor_version): Bump to 2.
27695 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
27697 * config/aarch64/aarch64-protos.h
27698 (aarch64_const_double_zero_rtx_p): Rename to...
27699 (aarch64_float_const_zero_rtx_p): ...this.
27700 (aarch64_float_const_representable_p): New.
27701 (aarch64_output_simd_mov_immediate): Likewise.
27702 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
27703 move immediate case.
27704 * config/aarch64/aarch64.c
27705 (aarch64_const_double_zero_rtx_p): Rename to...
27706 (aarch64_float_const_zero_rtx_p): ...this.
27707 (aarch64_print_operand): Allow printing of new constants.
27708 (aarch64_valid_floating_const): New.
27709 (aarch64_legitimate_constant_p): Check for valid floating-point
27711 (aarch64_simd_valid_immediate): Likewise.
27712 (aarch64_vect_float_const_representable_p): New.
27713 (aarch64_float_const_representable_p): Likewise.
27714 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
27715 (aarch64_output_simd_mov_immediate): New.
27716 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
27717 (*movdf_aarch64): Likewise.
27718 * config/aarch64/constraints.md (Ufc): New.
27719 (Y): call aarch64_float_const_zero_rtx.
27720 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
27722 2013-01-07 Richard Biener <rguenther@suse.de>
27724 PR tree-optimization/55888
27725 PR tree-optimization/55862
27726 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
27727 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
27728 not if it is contained therein.
27730 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
27732 * config/avr/t-avr: Typo.
27734 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
27737 * config/avr/t-avr: Don't automatically rebuild
27738 $(srcdir)/config/avr/t-multilib
27739 $(srcdir)/config/avr/avr-tables.opt
27740 $(srcdir)/doc/avr-mmcu.texi
27741 (avr-mcus): New phony target to build them on request.
27742 (s-avr-mlib, s-avr-mmcu-texi): Remove.
27743 * avr/avr-mcus.def: Adjust comments.
27745 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
27747 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
27749 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
27751 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
27753 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
27755 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
27757 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
27759 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
27760 to generate profiling.
27761 * config/rs6000/aix64.h (LIB_SPEC): Same.
27763 2013-01-04 Andrew Pinski <apinski@cavium.com>
27765 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
27767 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
27769 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
27771 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
27773 (ix86_expand_move): Ditto.
27774 (ix86_zero_extend_to_Pmode): Ditto.
27775 (ix86_expand_call): Ditto.
27776 (ix86_expand_special_args_builtin): Ditto.
27777 (ix86_expand_builtin): Ditto.
27779 2013-01-04 Richard Biener <rguenther@suse.de>
27781 PR tree-optimization/55862
27782 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
27783 translating them through PHI nodes.
27785 2013-01-04 Martin Jambor <mjambor@suse.cz>
27787 PR tree-optimization/55755
27788 * tree-sra.c (sra_modify_assign): Do not check that an access has no
27789 children when trying to avoid producing a VIEW_CONVERT_EXPR.
27791 2013-01-04 Marek Polacek <polacek@redhat.com>
27793 PR middle-end/55859
27794 * opts.c (default_options_optimization): Clarify error message.
27796 2013-01-04 Richard Biener <rguenther@suse.de>
27798 PR middle-end/55863
27799 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
27802 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
27805 * config/pa/pa.md (movsi): Revert previous change.
27806 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
27809 2013-01-03 Richard Henderson <rth@redhat.com>
27811 * config/i386/i386.c (ix86_expand_move): Always assign to op1
27812 after eliminating TLS symbols.
27814 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
27817 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
27818 * graphite-poly.c (debug_gmp_value): Likewise.
27820 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
27823 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
27824 selected code model, define __code_mode_small__, __code_model_medium__,
27825 __code_model_large__, __code_model_32__ or __code_model_kernel__.
27826 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
27827 xchg temporary register with %k. Declare temporary register as
27829 [__x86_64__]: For medium and large code models, preserve %rbx register.
27831 2013-01-03 Richard Biener <rguenther@suse.de>
27833 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
27834 (dump_subscript): Adjust.
27835 (finalize_ddr_dependent): Do not dump redundant info.
27836 (analyze_siv_subscript): Adjust.
27837 (subscript_dependence_tester): Likewise.
27838 (compute_affine_dependence): Likewise.
27840 2013-01-03 Richard Biener <rguenther@suse.de>
27843 2013-01-03 Richard Biener <rguenther@suse.de>
27845 PR tree-optimization/55857
27846 * tree-vect-stmts.c (vectorizable_load): Do not setup
27847 re-alignment for invariant loads.
27849 2013-01-02 Richard Biener <rguenther@suse.de>
27851 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
27852 invariant load do not generate a vector load from the scalar location.
27854 2013-01-03 Richard Biener <rguenther@suse.de>
27856 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
27857 for not vectorizing.
27858 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
27859 not build INDIRECT_REFs, call get_name once only.
27860 (vect_create_data_ref_ptr): Likewise. Dump base object kind
27861 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
27863 2013-01-03 Richard Biener <rguenther@suse.de>
27865 PR tree-optimization/55857
27866 * tree-vect-stmts.c (vectorizable_load): Do not setup
27867 re-alignment for invariant loads.
27869 2013-01-03 Richard Biener <rguenther@suse.de>
27872 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
27873 prefer a built-in decl.
27875 2013-01-03 Jakub Jelinek <jakub@redhat.com>
27877 * gcc.c (process_command): Update copyright notice dates.
27878 * gcov.c (print_version): Likewise.
27879 * gcov-dump.c (print_version): Likewise.
27881 PR rtl-optimization/55838
27882 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
27883 iv0.step, iv1.step and step.
27885 2013-01-03 Jakub Jelinek <jakub@redhat.com>
27886 Marc Glisse <marc.glisse@inria.fr>
27888 PR tree-optimization/55832
27889 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
27890 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
27891 integer_{one,zero}_node.
27893 2013-01-03 Jakub Jelinek <jakub@redhat.com>
27896 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
27897 * var-tracking.c (reverse_op): Don't add reverse ops to
27898 VALUEs that have already
27899 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
27901 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
27903 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
27905 2013-01-02 Teresa Johnson <tejohnson@google.com>
27907 * dumpfile.c (dump_loc): Print filename with location.
27908 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
27909 new location_t parameter to emit complete unroll message with
27910 new dump framework.
27911 (canonicalize_loop_induction_variables): Compute loops location
27912 and pass to try_unroll_loop_completely.
27913 * loop-unroll.c (report_unroll_peel): New function.
27914 (peel_loops_completely): Use new dump format with location
27915 for main dumpfile message, and invoke report_unroll_peel on success.
27916 (decide_unrolling_and_peeling): Ditto.
27917 (decide_peel_once_rolling): Remove old dumpfile message subsumed
27918 by report_unroll_peel.
27919 (decide_peel_completely): Ditto.
27920 (decide_unroll_constant_iterations): Ditto.
27921 (decide_unroll_runtime_iterations): Ditto.
27922 (decide_peel_simple): Ditto.
27923 (decide_unroll_stupid): Ditto.
27924 * cfgloop.c (get_loop_location): New function.
27925 * cfgloop.h (get_loop_location): Declare.
27927 2013-01-02 Sriraman Tallam <tmsriram@google.com>
27929 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
27932 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
27934 PR middle-end/55198
27935 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
27936 BLKmode objects when EXPAND_MEMORY is specified.
27938 2013-01-02 Sriraman Tallam <tmsriram@google.com>
27940 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
27942 (fold_builtin_cpu): Do not share cpu model decls across statements.
27944 2013-01-02 Jason Merrill <jason@redhat.com>
27947 * tree.c (build_array_type_1): Revert earlier change.
27949 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
27951 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
27953 * config/aarch64/aarch64-tune.md: Re-generate.
27955 2013-01-02 Richard Biener <rguenther@suse.de>
27957 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
27958 invariant load do not generate a vector load from the scalar location.
27960 2013-01-02 Richard Biener <rguenther@suse.de>
27963 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
27964 * configure: Regenerate.
27966 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
27968 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
27969 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
27970 (expand_builtin_int_roundingfn_2): Keep the original target around
27971 for the fallback case.
27973 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
27975 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
27976 to be clear for sign changes.
27978 2013-01-01 Jan Hubicka <jh@suse.cz>
27980 * ipa-inline-analysis.c: Fix formatting.
27982 2013-01-01 Jakub Jelinek <jakub@redhat.com>
27984 PR tree-optimization/55831
27985 * tree-vect-loop.c (get_initial_def_for_induction): Use
27986 gsi_after_labels instead of gsi_start_bb.
27988 Copyright (C) 2013 Free Software Foundation, Inc.
27990 Copying and distribution of this file, with or without modification,
27991 are permitted in any medium without royalty provided the copyright
27992 notice and this notice are preserved.