1 2013-11-08 Tom de Vries <tom@codesourcery.com>
3 * config/rs6000/t-xilinx: Remove duplicate contents.
5 2013-11-07 Andrew MacLeod <amacleod@redhat.com>
6 Joseph Myers <joseph@codesourcery.com>
8 * tree-core.h (enum cv_qualifier): Add TYPE_QUAL_ATOMIC.
9 (enum tree_index): Add TI_ATOMICQI_TYPE, TI_ATOMICHI_TYPE,
10 TI_ATOMICSI_TYPE, TI_ATOMICDI_TYPE and TI_ATOMICTI_TYPE.
11 (struct tree_base): Add atomic_flag field.
12 * tree.h (TYPE_ATOMIC): New accessor macro.
13 (TYPE_QUALS, TYPE_QUALS_NO_ADDR_SPACE): Add TYPE_QUAL_ATOMIC.
14 (TYPE_QUALS_NO_ADDR_SPACE_NO_ATOMIC): New macro.
15 (atomicQI_type_node, atomicHI_type_node, atomicSI_type_node)
16 (atomicDI_type_node, atomicTI_type_node): New macros for type
18 * tree.c (set_type_quals): Set TYPE_ATOMIC.
19 (find_atomic_core_type): New function.
20 (build_qualified_type): Adjust alignment for qualified types.
21 (build_atomic_base): New function
22 (build_common_tree_nodes): Build atomicQI_type_node,
23 atomicHI_type_node, atomicSI_type_node, atomicDI_type_node and
25 * print-tree.c (print_node): Print atomic qualifier.
26 * tree-pretty-print.c (dump_generic_node): Print atomic type
28 * target.def (atomic_assign_expand_fenv): New hook.
29 * doc/tm.texi.in (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): New @hook.
30 * doc/tm.texi: Regenerate.
31 * targhooks.c (default_atomic_assign_expand_fenv): New function.
32 * targhooks.h (default_atomic_assign_expand_fenv): Declare.
33 * sync-builtins.def (__atomic_feraiseexcept): New built-in
35 * config/i386/i386-builtin-types.def (VOID_FTYPE_PUSHORT): New
37 * config/i386/i386.c (enum ix86_builtins): Add
38 IX86_BUILTIN_FNSTENV, IX86_BUILTIN_FLDENV, IX86_BUILTIN_FNSTSW and
40 (bdesc_special_args): Add __builtin_ia32_fnstenv,
41 __builtin_ia32_fldenv, __builtin_ia32_fnstsw and
42 __builtin_ia32_fnclex.
43 (ix86_expand_builtin): Handle the new built-in functions.
44 (ix86_atomic_assign_expand_fenv): New function.
45 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): New macro.
46 * config/i386/i386.md (UNSPECV_FNSTENV, UNSPECV_FLDENV)
47 (UNSPECV_FNSTSW, UNSPECV_FNCLEX): New unspecs.
48 (fnstenv, fldenv, fnstsw, fnclex): New insns.
50 2013-11-07 Steve Ellcey <sellcey@mips.com>
52 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add fp64 directory.
53 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add -mfp64 flag.
54 (MULTILIB_DIRNAMES): Add fp64 directory.
55 (MULTILIB_EXCEPTIONS): Add new exclusions.
57 2013-11-07 Aldy Hernandez <aldyh@redhat.com>
59 * gimplify.c (gimple_regimplify_operands): Do not set
61 * graphite-sese-to-poly.c (remove_simple_copy_phi): Same.
62 (rewrite_close_phi_out_of_ssa): Same.
63 (rewrite_phi_out_of_ssa): Same.
64 (rewrite_degenerate_phi): Same.
65 (handle_scalar_deps_crossing_scop_limits): Same.
66 * tree-if-conv.c (predicate_scalar_phi): Same.
67 * tree-parloops.c (create_loads_for_reductions): Same.
68 (create_final_loads_for_reduction): Same.
69 (create_loads_and_stores_for_name): Same.
70 (transform_to_exit_first_loop): Same.
71 (create_parallel_loop): Same.
73 (move_computations_dom_walker::before_dom_children): Same.
74 * tree-ssa-loop-manip.c (rewrite_phi_with_iv): Same.
75 * tree-ssa-pre.c (eliminate_dom_walker::before_dom_children):
77 * tree-ssa-propagate.c (substitute_and_fold): Same.
78 * tree-vect-loop.c (vect_finalize_reduction): Same.
79 * tree-vect-stmts.c (vectorizable_call): Same.
81 2013-11-07 Mike Stump <mikestump@comcast.net>
83 * config/pdp11/pdp11.c: Include dbxout.h.
84 * config/picochip/picochip.c: Likewise.
86 2013-11-07 Cong Hou <congh@google.com>
88 PR tree-optimization/56764
89 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
90 Combine alias checks if it is possible to amortize the runtime
91 overhead. Return the number of alias checks after merging.
92 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
93 Use the function vect_create_cond_for_alias_checks () to check
94 the number of alias checks.
96 2013-11-07 Jeff Law <law@redhat.com>
98 * varpool.c (ctor_for_folding): Fix typo in comment.
100 2013-11-07 Joern Rennecke <joern.rennecke@embecosm.com>
102 * config/arc/arc.c (arc_ifcvt): Use commutativity, e.g.:
103 reg_a := reg_b + reg_a ==> reg_a := reg_a + reg_b
105 2013-11-07 Jeff Law <law@redhat.com>
107 * doc/invoke.texi (-fisolate-erroneous-paths): Document.
109 * gimple-ssa-isolate-paths.c (gate_isolate_erroneous_paths):
110 No longer check if we have __builtin_trap, assume it's
113 2013-11-07 Thomas Schwinge <thomas@codesourcery.com>
115 * doc/sourcebuild.texi (Top Level) <lto-plugin>: GNU ld can use
118 * config/arc/arc.h (LINK_COMMAND_SPEC): For
119 -ftree-parallelize-loops=*, link to libgomp and its dependencies.
120 * config/ia64/hpux.h (LIB_SPEC): Likewise.
121 * config/pa/pa-hpux11.h (LIB_SPEC): Likewise.
122 * config/pa/pa64-hpux.h (LIB_SPEC): Likewise.
123 * gcc.c (GOMP_SELF_SPECS): Update comment about libgomp's
126 2013-11-07 Jakub Jelinek <jakub@redhat.com>
128 * tree-ssa-loop-niter.c: Include tree-ssanames.h.
129 (determine_value_range): Add loop argument. Use get_range_info to
131 (bound_difference): Adjust caller.
133 2013-11-07 Richard Biener <rguenther@suse.de>
134 Jakub Jelinek <jakub@redhat.com>
136 * tree-vrp.c (find_assert_locations): Pre-seed live bitmaps for loop
137 latches from header PHI arguments from the latch edge.
139 2013-11-07 Paolo Carlini <paolo.carlini@oracle.com>
142 * varasm.c (output_constant): Handle NULLPTR_TYPE.
144 2013-11-07 H.J. Lu <hongjiu.lu@intel.com>
146 * config/i386/i386.c (ix86_expand_set_or_movmem): Don't set
147 misaligned_prologue_used when it has been set.
149 2013-11-07 Yury Gribov <y.gribov@samsung.com>
150 Jakub Jelinek <jakub@redhat.com>
153 * asan.c (get_mem_refs_of_builtin_call): Allow
154 integer literals as addresses in instrumented builtins.
156 2013-11-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
158 * config/aarch64/aarch64.c (aarch64_legitimize_reload_address):
159 Explain why plus_constant is not used.
161 2013-11-07 Richard Biener <rguenther@suse.de>
163 * tree-ssa-ccp.c (canonicalize_float_value): Rename to ...
164 (canonicalize_value): ... this. Also handle stripping of
166 (get_value, set_lattice_value, get_value_for_expr): Adjust.
167 * gimple-fold.c (canonicalize_constructor_val): Strip
169 * tree-ssa-threadedge.c (set_ssa_name_value): Likewise.
171 2013-11-07 Richard Biener <rguenther@suse.de>
173 * tree-dfa.c (get_ref_base_and_extent): Fix casting.
175 2013-11-07 H.J. Lu <hongjiu.lu@intel.com>
178 * config/i386/i386.md (push peepholer/splitter): Use Pmode
179 with stack_pointer_rtx.
181 2013-11-07 Bin Cheng <bin.cheng@arm.com>
183 * tree-ssa-loop-ivopts.c (get_shiftadd_cost): Check equality
184 using operand_equal_p.
186 2013-11-07 Bin Cheng <bin.cheng@arm.com>
188 * tree-ssa-loop-ivopts.c (alloc_iv): Lower address expressions.
189 * tree-affine.c (get_inner_reference_aff): Return base.
190 * tree-affine.h (get_inner_reference_aff): Change prototype.
192 2013-11-06 Tobias Burnus <burnus@net-b.de>
194 * doc/invoke.texi (Wdate-time): Fix typo.
196 2013-11-06 Oleg Endo <olegendo@gcc.gnu.org>
198 * config/sh/sh.md (addsf3, divsf3, divsf3_i, rsqrtsf2, cmpgtdf_t,
199 cmpeqdf_t, *ieee_ccmpeqdf_t, negdf2, sqrtdf2, absdf2): Use
200 fp_arith_reg_operand instead of arith_reg_operand.
202 2013-11-06 Oleg Endo <olegendo@gcc.gnu.org>
204 * config/sh/sh.md (adddi3): Remove empty constraints.
205 Remove can_create_pseudo_p and arith_reg_operand check.
206 (adddi3_compact, subdi3_compact, *negdi2): Remove constraints.
209 2013-11-06 Jeff Law <law@redhat.com>
210 Tom Tromey <tromey@redhat.com>
212 * gdbinit.in: Disable strict type checking.
214 2013-11-06 Vladimir Makarov <vmakarov@redhat.com>
216 * tree-pass.h (make_pass_live_range_shrinkage): New external.
217 * timevar.def (TV_LIVE_RANGE_SHRINKAGE): New.
218 * sched-rgn.c (gate_handle_live_range_shrinkage): New.
219 (rest_of_handle_live_range_shrinkage): Ditto
220 (class pass_live_range_shrinkage): Ditto.
221 (pass_data_live_range_shrinkage): Ditto.
222 (make_pass_live_range_shrinkage): Ditto.
223 * sched-int.h (initialize_live_range_shrinkage): New prototype.
224 (finish_live_range_shrinkage): Ditto.
225 * sched-deps.c (create_insn_reg_set): Make void return value.
226 * passes.def: Add pass_live_range_shrinkage.
227 * ira.c (update_equiv_regs): Don't move if
228 flag_live_range_shrinkage.
229 * haifa-sched.c (live_range_shrinkage_p): New.
230 (initialize_live_range_shrinkage, finish_live_range_shrinkage):
232 (rank_for_schedule): Add code for pressure relief through live
234 (schedule_insn): Print more debug info.
235 (sched_init): Setup SCHED_PRESSURE_WEIGHTED for pressure relief
236 through live range shrinkage.
237 * doc/invoke.texi (-flive-range-shrinkage): New.
238 * common.opt (flive-range-shrinkage): New.
240 2013-11-06 Uros Bizjak <ubizjak@gmail.com>
243 * config/i386/i386.c (ix86_avx_u128_mode_needed): Require
244 AVX_U128_DIRTY mode for call_insn RTXes that use AVX256 registers.
245 (ix86_avx_u128_mode_needed): Return AVX_U128_DIRTY mode for call_insn
246 RTXes that return in AVX256 register.
248 2013-11-06 Richard Biener <rguenther@suse.de>
250 PR tree-optimization/58653
251 * tree-predcom.c (ref_at_iteration): Rewrite to generate a MEM_REF.
252 (prepare_initializers_chain): Adjust.
254 2013-11-06 Andrew MacLeod <amacleod@redhat.com>
256 * gimple.h (block_in_transaction): Move to basic-block.h and rename.
257 (gimple_in_transaction): Use bb_in_transaction.
258 * basic-block.h (bb_in_transaction): Relocate here and rename.
259 * tree-ssa-loop-im.c (execute_sm): Use bb_in_transaction.
261 2013-11-06 Richard Biener <rguenther@suse.de>
263 * tree.c (drop_tree_overflow): New function.
264 * tree.h (drop_tree_overflow): Declare.
265 * gimplify.c (gimplify_expr): Drop TREE_OVERFLOW.
266 * tree-vrp.c (range_int_cst_singleton_p): Use
267 is_overflow_infinity instead of testing TREE_OVERFLOW.
268 (extract_range_from_assert): Likewise.
269 (zero_nonzero_bits_from_vr): Likewise.
270 (extract_range_basic): Likewise.
271 (register_new_assert_for): Use drop_tree_overflow.
272 (vrp_visit_phi_node): Likewise.
274 2013-11-06 Eric Botcazou <ebotcazou@adacore.com>
276 * config/i386/i386.c (ix86_expand_prologue): Optimize stack
277 checking for leaf functions without dynamic stack allocation.
278 * config/ia64/ia64.c (ia64_emit_probe_stack_range): Adjust.
279 (ia64_expand_prologue): Likewise.
280 * config/mips/mips.c (mips_expand_prologue): Likewise.
281 * config/rs6000/rs6000.c (rs6000_emit_prologue): Likewise.
282 * config/sparc/sparc.c (sparc_expand_prologue): Likewise.
283 (sparc_flat_expand_prologue): Likewise.
285 2013-11-06 James Greenhalgh <james.greenhalgh@arm.com>
287 * config/aarch64/arm_neon.h (__ST2_LANE_FUNC): Better model data size.
288 (__ST3_LANE_FUNC): Likewise.
289 (__ST4_LANE_FUNC): Likewise.
291 2013-11-06 Nick Clifton <nickc@redhat.com>
293 * config/msp430/msp430.h (TARGET_CPU_CPP_BUILTINS): Define the
294 name returned by msp430_mcu_name.
295 (LIB_SPEC): If a -T option has not been specified then set a
296 default, mcu-specific, linker script.
297 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more mcu names.
298 * config/msp430/msp430.c (msp430x_names): Likewise.
299 Alpha sort the names for ease of comparison.
300 (msp430_mcu_name): New function: Returns a string suitable for
301 use as a C preprocessor symbol based upon the name of the MCU
303 (msp430_option_override): Accept msp430x and msp430xv2 as generic
305 * config/msp430/msp430-protos.h (msp430_mcu_name): Prototype.
307 * gcc.c (do_spec_1): Do not insert a space after a %* substitution
308 unless it is the last part of a spec substring.
309 * doc/invoke.texi (Spec Files): Document space insertion
312 2013-11-06 Christian Bruel <christian.bruel@st.com>
314 * config/sh/sh-mem.cc (sh_expand_cmpnstr, sh_expand_cmpstr):
315 Factorize probabilities, Use adjust_address instead of
316 adjust_automodify_address when possible. Enable for optimize.
317 (sh_expand_strlen): New function.
318 * config/sh/sh-protos.h (sh_expand_strlen): Declare.
319 * config/sh/sh.md (strlensi): New pattern.
320 (UNSPEC_BUILTIN_STRLEN): Define.
322 2013-11-06 Jakub Jelinek <jakub@redhat.com>
325 * expr.c (get_bit_range): Handle *offset == NULL_TREE.
326 (expand_assignment): If *bitpos is negative, set *offset
327 and adjust *bitpos, so that it is not negative.
329 2013-11-06 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
331 * config/i386/bdver3.md : Added two additional decoder units
332 to support issue rate of 4 and remodeled vector unit.
333 * config/i386/i386.c (ix86_issue_rate): Issue rate for BD
334 architectures is set to 4.
335 * config/i386/i386.c (ia32_multipass_dfa_lookahead): DFA
336 lookahead is set to 4 for BD architectures.
338 2013-11-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
340 * config/rs6000/rs6000.c (rs6000_option_override_internal):
341 Remove restriction against use of VSX instructions when generating
342 code for little endian mode.
344 2013-11-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
346 * config/rs6000/altivec.md (mulv4si3): Ensure we generate vmulouh
347 for both big and little endian.
348 (mulv8hi3): Swap input operands for merge high and merge low
349 instructions for little endian.
351 2013-11-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
353 * config/rs6000/altivec.md (vec_widen_umult_even_v16qi): Change
354 define_insn to define_expand that uses even patterns for big
355 endian and odd patterns for little endian.
356 (vec_widen_smult_even_v16qi): Likewise.
357 (vec_widen_umult_even_v8hi): Likewise.
358 (vec_widen_smult_even_v8hi): Likewise.
359 (vec_widen_umult_odd_v16qi): Likewise.
360 (vec_widen_smult_odd_v16qi): Likewise.
361 (vec_widen_umult_odd_v8hi): Likewise.
362 (vec_widen_smult_odd_v8hi): Likewise.
363 (altivec_vmuleub): New define_insn.
364 (altivec_vmuloub): Likewise.
365 (altivec_vmulesb): Likewise.
366 (altivec_vmulosb): Likewise.
367 (altivec_vmuleuh): Likewise.
368 (altivec_vmulouh): Likewise.
369 (altivec_vmulesh): Likewise.
370 (altivec_vmulosh): Likewise.
372 2013-11-05 Mike Stump <mikestump@comcast.net>
374 * Makefile.in (mostlyclean): Remove c-family objects.
376 2013-11-05 Ian Lance Taylor <iant@google.com>
378 * config/i386/sync.md (atomic_compare_and_swap<dwi>_doubleword):
379 If possible, add .cfi directives to record change to bx.
380 * config/i386/i386.c (ix86_emit_cfi): New function.
381 * config/i386/i386-protos.h (ix86_emit_cfi): Declare.
383 2013-11-05 Steven Bosscher <steven@gcc.gnu.org>
385 * rtlanal.c (tablejump_p): Expect a JUMP_TABLE_DATA to always follow
386 immediately after a label for a tablejump pattern.
388 * config/arm/arm.c (is_jump_table): Remove.
389 (create_fix_barrier): Use tablejump_p instead.
390 (arm_reorg): Likewise.
391 (thumb1_output_casesi): Expect JUMP_TABLE_DATA to always be NEXT_INSN.
392 (thumb2_output_casesi): Likewise.
393 * config/aarch64/aarch64.c (aarch64_output_casesi): Likewise.
394 * config/sh/sh.md (casesi_worker_1, casesi_worker_2,
395 casesi_shift_media, casesi_load_media): Likewise.
396 * config/iq2000/iq2000.md: Likewise (in anonymous define_insn).
397 * config/microblaze/microblaze.md: Likewise.
399 2013-11-05 Tobias Burnus <burnus@net-b.de>
401 * doc/invoke.texi (-Wdate-time): Document.
403 2013-11-05 Richard Sandiford <rdsandiford@googlemail.com>
405 * double-int.c (lshift_double, rshift_double): Remove
406 SHIFT_COUNT_TRUNCATED handling.
408 2013-11-05 Jeff Law <law@redhat.com>
410 * Makefile.in (OBJS): Add gimple-ssa-isolate-paths.o
411 * common.opt (-fisolate-erroneous-paths): Add option and documentation.
412 * gimple-ssa-isolate-paths.c: New file.
413 * gimple.c (check_loadstore): New function.
414 (infer_nonnull_range): Moved into gimple.c from tree-vrp.c
415 Verify OP is in the argument list and the argument corresponding
416 to OP is a pointer type. Use operand_equal_p rather than
417 pointer equality when testing if OP is on the nonnull list.
418 Use check_loadstore rather than count_ptr_derefs. Handle
419 GIMPLE_RETURN statements.
420 * tree-vrp.c (infer_nonnull_range): Remove.
421 * gimple.h (infer_nonnull_range): Declare.
422 * opts.c (default_options_table): Add OPT_fisolate_erroneous_paths.
423 * passes.def: Add pass_isolate_erroneous_paths.
424 * timevar.def (TV_ISOLATE_ERRONEOUS_PATHS): New timevar.
425 * tree-pass.h (make_pass_isolate_erroneous_paths): Declare.
426 * tree-ssa.c (struct count_ptr_d): Remove.
427 (count_ptr_derefs, count_uses_and_derefs): Remove.
428 * tree-ssa.h (count_uses_and_derefs): Remove.
430 2013-11-05 Jakub Jelinek <jakub@redhat.com>
432 PR rtl-optimization/58997
433 * loop-iv.c (iv_subreg): For IV_UNKNOWN_EXTEND, expect
434 get_iv_value to be in iv->mode rather than iv->extend_mode.
435 (iv_extend): Likewise. Otherwise, if iv->extend != extend,
436 use lowpart_subreg on get_iv_value before calling simplify_gen_unary.
437 * loop-unswitch.c (may_unswitch_on): Make sure op[i] is in the right
440 2013-11-05 Andrew MacLeod <amacleod@redhat.com>
442 * gimple.h: Move some prototypes to gimple-expr.h and add to include
444 (extract_ops_from_tree, gimple_call_addr_fndecl, is_gimple_reg_type):
445 Move to gimple-expr.h.
446 * gimple-expr.h: New file. Relocate some prototypes from gimple.h.
447 (types_compatible_p, is_gimple_reg_type, is_gimple_variable,
448 is_gimple_id, virtual_operand_p, is_gimple_addressable,
449 is_gimple_constant, extract_ops_from_tree, gimple_call_addr_fndecl):
451 * gimple.c (extract_ops_from_tree_1, gimple_cond_get_ops_from_tree,
452 gimple_set_body, gimple_body, gimple_has_body_p, is_gimple_lvalue,
453 is_gimple_condexpr, is_gimple_addressable, is_gimple_constant,
454 is_gimple_address, is_gimple_invariant_address,
455 is_gimple_ip_invariant_address, is_gimple_min_invariant,
456 is_gimple_ip_invariant, is_gimple_variable, is_gimple_id,
457 virtual_operand_p, is_gimple_reg, is_gimple_val, is_gimple_asm_val,
458 is_gimple_min_lval, is_gimple_call_addr, is_gimple_mem_ref_addr,
459 gimple_decl_printable_name, useless_type_conversion_p,
460 types_compatible_p, gimple_can_coalesce_p, copy_var_decl): Move to
462 * gimple-expr.c: New File.
463 (useless_type_conversion_p, gimple_set_body, gimple_body,
464 gimple_has_body_p, gimple_decl_printable_name, copy_var_decl,
465 gimple_can_coalesce_p, extract_ops_from_tree_1,
466 gimple_cond_get_ops_from_tree, is_gimple_lvalue, is_gimple_condexpr,
467 is_gimple_address, is_gimple_invariant_address,
468 is_gimple_ip_invariant_address, is_gimple_min_invariant,
469 is_gimple_ip_invariant, is_gimple_reg, is_gimple_val,
470 is_gimple_asm_val, is_gimple_min_lval, is_gimple_call_addr,
471 is_gimple_mem_ref_addr): Relocate here.
472 * Makefile.in (OBJS): Add gimple-expr.o.
474 2013-11-05 David Malcolm <dmalcolm@redhat.com>
476 * gengtype-parse.c (struct_field_seq): Support empty structs.
478 2013-11-05 Uros Bizjak <ubizjak@gmail.com>
480 * config/i386/t-rtems (MULTILIB_MATCHES): Fix option typos.
482 2013-11-05 Uros Bizjak <ubizjak@gmail.com>
484 * config/i386/i386-c.c (ix86_target_macros): Define _SOFT_FLOAT
486 * config/i386/rtemself.h (TARGET_OS_CPP_BUILTINS): Do not define
488 (LONG_DOUBLE_TYPE_SIZE): New define.
489 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Ditto.
491 2013-11-05 Paolo Carlini <paolo.carlini@oracle.com>
494 * doc/extend.texi [visibility ("visibility_type")]: Add example
495 about visibility attribute on namespace declaration.
497 2013-11-05 Richard Biener <rguenther@suse.de>
500 * passes.def (all_passes): Start with pass_fixup_cfg again.
502 2013-11-05 Richard Biener <rguenther@suse.de>
504 PR tree-optimization/58955
505 * tree-loop-distribution.c (pg_add_dependence_edges): Fix
508 2013-11-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
510 * config/rs6000/vector.md (vec_pack_sfix_trunc_v2df): Adjust for
512 (vec_pack_ufix_trunc_v2df): Likewise.
514 2013-11-05 H.J. Lu <hongjiu.lu@intel.com>
517 * doc/md.texi (@code{movmem@var{m}}): Specify Pmode as mode of
518 pattern, instead of word_mode.
520 * expr.c (emit_block_move_via_movmem): Don't use mode wider than
522 (set_storage_via_setmem): Likewise.
524 2013-11-05 Andrew MacLeod <amacleod@redhat.com>
526 * tree-outof-ssa.c (queue_phi_copy_p): Combine phi_ssa_name_p from
527 gimple.h and the rest of the condition in eliminate_build.
528 (eliminate_build): Call new routine.
529 * gimple.h (phi_ssa_name_p): Delete.
531 2013-11-05 Trevor Saunders <tsaunders@mozilla.com>
533 * vec.c (vec_prefix::calculate_allocation): Don't try to handle the
534 case of no prefix and reserving zero slots, because when that's the
535 case we'll never get here.
536 * vec.h (va_heap::reserve): Don't try and handle
537 vec_prefix::calculate_allocation returning zero because that should
540 2013-11-05 Richard Biener <rguenther@suse.de>
543 * tree-dfa.c (get_ref_base_and_extent): Merge common code
544 in MEM_REF and TARGET_MEM_REF handling. Make sure to
545 process trailing array detection before diving into the
546 view-converted object (and possibly apply some extra offset).
548 2013-11-05 Joseph Myers <joseph@codesourcery.com>
550 * config/i386/i386.c (ix86_float_exceptions_rounding_supported_p):
552 (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P): Define.
554 2013-11-05 Marc Glisse <marc.glisse@inria.fr>
556 PR tree-optimization/58958
557 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Use
558 get_addr_base_and_unit_offset instead of get_ref_base_and_extent.
560 2013-11-05 Marc Glisse <marc.glisse@inria.fr>
562 * tree-ssa-alias.h (ranges_overlap_p): Handle negative offsets.
563 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Likewise.
565 2013-11-05 Jakub Jelinek <jakub@redhat.com>
567 PR tree-optimization/58984
568 * ipa-prop.c (ipa_load_from_parm_agg_1): Add SIZE_P argument,
569 set *SIZE_P if non-NULL on success.
570 (ipa_load_from_parm_agg, ipa_analyze_indirect_call_uses): Adjust
572 (ipcp_transform_function): Likewise. Punt if size of access
573 is different from TYPE_SIZE on v->value's type.
575 2013-11-05 Tobias Burnus <burnus@net-b.de>
577 * doc/invoke.texi (-fopenmp-simd): Document new option.
578 * gimplify.c (gimplify_body): Accept -fopenmp-simd.
579 * omp-low.c (execute_expand_omp, execute_lower_omp): Ditto.
580 * tree.c (attribute_value_equal): Ditto.
582 2013-11-04 Wei Mi <wmi@google.com>
584 * sched-rgn.c (add_branch_dependences): Keep insns in
585 a SCHED_GROUP at the end of BB to remain their location.
587 2013-11-04 Wei Mi <wmi@google.com>
589 * config/i386/i386.c (memory_address_length): Extract a part
590 of code to rip_relative_addr_p.
591 (rip_relative_addr_p): New Function.
592 (ix86_macro_fusion_p): Ditto.
593 (ix86_macro_fusion_pair_p): Ditto.
594 * config/i386/i386.h: Add new tune features about macro-fusion.
595 * config/i386/x86-tune.def (DEF_TUNE): Ditto.
596 * doc/tm.texi: Generated.
597 * doc/tm.texi.in: Ditto.
598 * haifa-sched.c (try_group_insn): New Function.
599 (group_insns_for_macro_fusion): Ditto.
600 (sched_init): Call group_insns_for_macro_fusion.
601 * target.def: Add two hooks: macro_fusion_p and
604 2013-11-04 Kostya Serebryany <kcc@google.com>
606 Update to match the changed asan API.
607 * asan.c (asan_function_start): New function.
608 (asan_emit_stack_protection): Update the string stored in the
609 stack red zone to match new API. Store the PC of the current
610 function in the red zone.
611 (asan_global_struct): Update the __asan_global definition to match
613 (asan_add_global): Ditto.
614 * asan.h (asan_function_start): New prototype.
615 * final.c (final_start_function): Call asan_function_start.
616 * sanitizer.def (BUILT_IN_ASAN_INIT): Rename __asan_init_v1
619 2013-11-04 Wei Mi <wmi@google.com>
621 * config/i386/i386-c.c (ix86_target_macros_internal): Separate
622 PROCESSOR_COREI7_AVX out from PROCESSOR_COREI7.
623 * config/i386/i386.c (ix86_option_override_internal): Ditto.
624 (ix86_issue_rate): Ditto.
625 (ix86_adjust_cost): Ditto.
626 (ia32_multipass_dfa_lookahead): Ditto.
627 (ix86_sched_init_global): Ditto.
628 (get_builtin_code_for_version): Ditto.
629 * config/i386/i386.h (enum target_cpu_default): Ditto.
630 (enum processor_type): Ditto.
631 * config/i386/x86-tune.def (DEF_TUNE): Ditto.
633 2013-11-04 Vladimir Makarov <vmakarov@redhat.com>
635 PR rtl-optimization/58967
636 * config/rs6000/rs6000.c (legitimate_lo_sum_address_p): Remove
637 !lra_in_progress for mode sizes bigger word.
639 2013-11-04 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
641 * config/rs6000/altivec.md (vec_widen_umult_hi_v16qi): Swap
642 arguments to merge instruction for little endian.
643 (vec_widen_umult_lo_v16qi): Likewise.
644 (vec_widen_smult_hi_v16qi): Likewise.
645 (vec_widen_smult_lo_v16qi): Likewise.
646 (vec_widen_umult_hi_v8hi): Likewise.
647 (vec_widen_umult_lo_v8hi): Likewise.
648 (vec_widen_smult_hi_v8hi): Likewise.
649 (vec_widen_smult_lo_v8hi): Likewise.
651 2013-11-04 Ian Lance Taylor <iant@google.com>
653 * builtins.def (ATTR_NOTHROWCALL_LEAF_LIST): Define.
654 * sync-builtins.def: Use ATTR_NOTHROWCALL_LEAF_LIST for all sync
655 builtins that take pointers.
656 * lto-opts.c (lto_write_options): Write -fnon-call-exceptions if set.
657 * lto-wrapper.c (merge_and_complain): Collect OPT_fnon_call_exceptions.
658 (run_gcc): Pass -fnon-call-exceptions.
660 2013-11-04 Jakub Jelinek <jakub@redhat.com>
662 * optabs.c (expand_vec_perm): Revert one incorrect line from
665 PR tree-optimization/58978
666 * tree-vrp.c (all_imm_uses_in_stmt_or_feed_cond): Don't modify
667 use_stmt by single_imm_use directly. Only call single_imm_use
670 2013-11-04 Vladimir Makarov <vmakarov@redhat.com>
672 PR rtl-optimization/58968
673 * lra-spills.c (return_regno_p): New function.
674 (lra_final_code_change): Use it.
676 2013-11-04 Joseph Myers <joseph@codesourcery.com>
678 * doc/cpp.texi (__GCC_IEC_559, __GCC_IEC_559_COMPLEX): Document macros.
679 * target.def (float_exceptions_rounding_supported_p): New hook.
680 * targhooks.c (default_float_exceptions_rounding_supported_p): New
682 * targhooks.h (default_float_exceptions_rounding_supported_p): Declare.
683 * doc/tm.texi.in (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P):
685 * doc/tm.texi: Regenerate.
686 * config.gcc (powerpc*-*-linux*): Set extra_objs.
687 * config/rs6000/rs6000-linux.c: New file.
688 * config/rs6000/rs6000-protos.h
689 (rs6000_linux_float_exceptions_rounding_supported_p): Declare.
690 * config/rs6000/linux.h
691 (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P): New macro.
692 * config/rs6000/linux64.h
693 (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P): Likewise.
694 * config/rs6000/t-linux (rs6000-linux.o): New rule.
695 * config/rs6000/t-linux64 (rs6000-linux.o): Likewise.
697 2013-11-04 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
699 * config/rs6000/vsx.md (*vsx_le_perm_store_<mode> for VSX_D):
700 Replace the define_insn_and_split with a define_insn and two
701 define_splits, with the split after reload re-permuting the source
702 register to its original value.
703 (*vsx_le_perm_store_<mode> for VSX_W): Likewise.
704 (*vsx_le_perm_store_v8hi): Likewise.
705 (*vsx_le_perm_store_v16qi): Likewise.
707 2013-11-04 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
709 * config/rs6000/vector.md (vec_pack_trunc_v2df): Adjust for
712 2013-11-04 Jakub Jelinek <jakub@redhat.com>
714 PR tree-optimization/58946
715 * tree-ssa-reassoc.c (maybe_optimize_range_tests): Update all
716 bbs with bbinfo[idx].op != NULL before all blocks with
717 bbinfo[idx].op == NULL.
719 2013-11-04 Richard Sandiford <rdsandiford@googlemail.com>
721 * config/avr/avr-log.c (avr_double_int_pop_digit): Delete.
722 (avr_dump_double_int_hex): Likewise.
723 (avr_log_vadump): Remove %D and %X handling.
724 * config/avr/avr.c (avr_double_int_push_digit): Delete.
725 (avr_map_op_t): Change map from double_int to unsigned int.
726 (avr_map_op): Update accordingly.
727 (avr_map, avr_map_metric, avr_has_nibble_0xf, avr_map_decompose)
728 (avr_move_bits, avr_out_insert_bits, avr_fold_builtin): Operate on
729 unsigned ints rather than double_ints.
731 2013-11-03 Marek Polacek <polacek@redhat.com>
733 Implement -fsanitize=vla-bound.
734 * opts.c (common_handle_option): Handle vla-bound.
735 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE): Define.
736 * flag-types.h (enum sanitize_code): Add SANITIZE_VLA.
737 * asan.c (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR.
739 2013-11-02 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
741 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Adjust for
744 2013-11-02 Uros Bizjak <ubizjak@gmail.com>
746 * config/i386/constraints.md (Ts, Tv): New address constrains.
747 * config/i386/i386.md (*lea<mode>, *<mode>_<bndcheck>): Use Ts
748 constraint for address_no_seg_operand.
749 * config/i386/sse.md (*avx512pf_gatherpf<mode>_mask)
750 (*avx512pf_gatherpf<mode>, *avx512pf_scatterpf<mode>_mask)
751 (*avx512pf_scatterpf<mode>, *avx2_gathersi<mode>)
752 (*avx2_gathersi<mode>_2, *avx2_gatherdi<mode>, *avx2_gatherdi<mode>_2)
753 (*avx2_gatherdi<mode>_3, *avx2_gatherdi<mode>_4)
754 (*avx512f_gathersi<mode>, *avx512f_gathersi<mode>_2)
755 (*avx512f_gatherdi<mode>, *avx512f_gatherdi<mode>_2)
756 (*avx512f_scattersi<mode> *avx512f_scatterdi<mode>): Use Tv
757 constraint for vsib_address_operand.
759 2013-11-02 Steven Bosscher <steven@gcc.gnu.org>
761 * gcse.c (pre_delete): Remove references to regmove from comments.
762 * recog.c: (validate_replace_rtx_1): Likewise.
763 * config/rl78/rl78.c: Likewise.
764 * config/v850/v850.h: Likewise, and remove unused ENABLE_REGMOVE_PASS.
765 * common/config/m32r/m32r-common.c: Don't manipulate OPT_fregmove.
766 * common/config/mmix/mmix-common.c: Likewise.
768 2013-11-01 Trevor Saunders <tsaunders@mozilla.com>
770 * function.c (reorder_blocks): Convert block_stack to a stack_vec.
771 * gimplify.c (gimplify_compound_lval): Likewise.
772 * graphite-clast-to-gimple.c (gloog): Likewise.
773 * graphite-dependences.c (loop_is_parallel_p): Likewise.
774 * graphite-scop-detection.c (scopdet_basic_block_info): Likewise.
775 (limit_scop); Likewise.
776 (build_scops): Likewise.
777 (dot_scop): Likewise.
778 * graphite-sese-to-poly.c (sese_dom_walker): Likewise.
779 (build_scop_drs): Likewise.
780 (insert_stmts): Likewise.
781 (insert_out_of_ssa_copy): Likewise.
782 (remove_phi): Likewise.
783 (rewrite_commutative_reductions_out_of_ssa_close_phi): Likewise.
784 * hw-doloop.c (discover_loop): Likewise.
785 * tree-call-cdce.c (shrink_wrap_one_built_in_call): Likewise.
786 * tree-dfa.c (dump_enumerated_decls): Likewise.
787 * tree-if-conv.c (if_convertable_loop_p): Likewise.
788 * tree-inline.c (tree_function_versioning): Likewise.
789 * tree-loop-distribution.c (build_rdg): Likewise.
790 (rdg_flag_vertex_and_dependent): Likewise.
791 (distribute_loop): Likewise.
792 * tree-parloops.c (loop_parallel_p): Likewise.
793 (eliminate_local_variables): Likewise.
794 (separate_decls_in_region): Likewise.
795 * tree-predcom.c (tree_predictive_commoning_loop): Likewise.
796 * tree-ssa-phiopt.c (cond_if_else_store_replacement): Likewise.
797 * tree-ssa-uncprop.c (uncprop_dom_walker): Likewise.
798 * tree-vect-loop.c (vect_analyze_scaler_cycles_1): Likewise.
799 * tree-vect-patterns.c (vect_pattern_recog): Likewise.
800 * tree-vect-stmts.c (vect_mark_stmts_to_be_vectorized): Likewise.
801 (vectorizable_condition): Likewise.
803 2013-11-01 Uros Bizjak <ubizjak@gmail.com>
805 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): Always define as 0/1.
806 * configure: Regenerate.
807 * config/i386/i386.md (*movdi_internal): Change
808 HAVE_AS_IX86_INTERUNIT_MOVQ to runtime check.
809 (*movdf_internal): Ditto.
810 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
811 * config/i386/sse.md (vec_concatv2di): Output interunit movq
812 for HAVE_AS_IX86_INTERUNIT_MOVQ targets.
814 2013-10-31 Robert Suchanek <Robert.Suchanek@imgtec.com>
816 * lra-spills.c (assign_spill_hard_regs): Remove statement terminator
819 2013-10-31 David Malcolm <dmalcolm@redhat.com>
821 Automated part of renaming of symtab_node_base to symtab_node.
823 Patch autogenerated by rename_symtab.py from
824 https://github.com/davidmalcolm/gcc-refactoring-scripts
825 revision 58bb219cc090b2f4516a9297d868c245495ee622
826 with ChangeLog entry fixed up by hand.
828 * cgraph.c (x_cgraph_nodes_queue): Rename symtab_node_base to
830 (cgraph_node_for_asm): Likewise.
831 * cgraph.h (symtab_node_base): Likewise.
832 (cgraph_node): Likewise.
833 (varpool_node): Likewise.
834 (is_a_helper <cgraph_node>::test): Likewise.
835 (is_a_helper <varpool_node>::test): Likewise.
836 (symtab_nodes): Likewise.
837 (symtab_register_node): Likewise.
838 (symtab_unregister_node): Likewise.
839 (symtab_remove_node): Likewise.
840 (symtab_get_node): Likewise.
841 (symtab_node_for_asm): Likewise.
842 (symtab_node_asm_name): Likewise.
843 (symtab_node_name): Likewise.
844 (symtab_insert_node_to_hashtable): Likewise.
845 (symtab_add_to_same_comdat_group): Likewise.
846 (symtab_dissolve_same_comdat_group_list): Likewise.
847 (dump_symtab_node): Likewise.
848 (debug_symtab_node): Likewise.
849 (dump_symtab_base): Likewise.
850 (verify_symtab_node): Likewise.
851 (verify_symtab_base): Likewise.
852 (symtab_used_from_object_file_p): Likewise.
853 (symtab_alias_ultimate_target): Likewise.
854 (symtab_resolve_alias): Likewise.
855 (fixup_same_cpp_alias_visibility): Likewise.
856 (symtab_for_node_and_aliases): Likewise.
857 (symtab_nonoverwritable_alias): Likewise.
858 (availability symtab_node_availability): Likewise.
859 (symtab_semantically_equivalent_p): Likewise.
860 (fixup_same_cpp_alias_visibility): Likewise.
861 (symtab_prevail_in_asm_name_hash): Likewise.
864 (varpool_first_variable): Likewise.
865 (varpool_next_variable): Likewise.
866 (varpool_first_static_initializer): Likewise.
867 (varpool_next_static_initializer): Likewise.
868 (varpool_first_defined_variable): Likewise.
869 (varpool_next_defined_variable): Likewise.
870 (cgraph_first_defined_function): Likewise.
871 (cgraph_next_defined_function): Likewise.
872 (cgraph_first_function): Likewise.
873 (cgraph_next_function): Likewise.
874 (cgraph_first_function_with_gimple_body): Likewise.
875 (cgraph_next_function_with_gimple_body): Likewise.
876 (symtab_alias_target): Likewise.
877 (symtab_real_symbol_p): Likewise.
878 (symtab_can_be_discarded): Likewise.
879 * cgraphbuild.c (mark_address): Likewise.
880 (mark_load): Likewise.
881 (mark_store): Likewise.
882 * cgraphunit.c (decide_is_symbol_needed): Likewise.
884 (enqueue_node): Likewise.
885 (referred_to_p): Likewise.
886 (cgraph_process_same_body_aliases): Likewise.
887 (analyze_functions): Likewise.
888 (handle_alias_pairs): Likewise.
889 (output_weakrefs): Likewise.
891 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
892 * ipa-inline-analysis.c (inline_write_summary): Likewise.
893 * ipa-prop.c (remove_described_reference): Likewise.
894 (try_decrement_rdesc_refcount): Likewise.
895 (ipa_edge_duplication_hook): Likewise.
896 * ipa-ref.c (ipa_record_reference): Likewise.
897 (ipa_maybe_record_reference): Likewise.
898 (ipa_clone_ref): Likewise.
899 (ipa_clone_references): Likewise.
900 (ipa_clone_referring): Likewise.
901 (ipa_find_reference): Likewise.
902 (ipa_remove_stmt_references): Likewise.
903 (ipa_clear_stmts_in_references): Likewise.
904 * ipa-ref.h (symtab_node_base): Likewise.
906 (ipa_record_reference): Likewise.
907 (ipa_maybe_record_reference): Likewise.
908 (ipa_clone_references): Likewise.
909 (ipa_clone_referring): Likewise.
910 (ipa_clone_ref): Likewise.
911 (ipa_find_reference): Likewise.
912 (ipa_remove_stmt_references): Likewise.
913 (ipa_clear_stmts_in_references): Likewise.
914 * ipa-reference.c (ipa_reference_write_optimization_summary):
916 * ipa.c (enqueue_node): Likewise.
917 (process_references): Likewise.
918 (walk_polymorphic_call_targets): Likewise.
919 (symtab_remove_unreachable_nodes): Likewise.
920 (address_taken_from_non_vtable_p): Likewise.
921 (comdat_can_be_unshared_p_1): Likewise.
922 (comdat_can_be_unshared_p): Likewise.
923 (can_replace_by_local_alias): Likewise.
924 (function_and_variable_visibility): Likewise.
925 * is-a.h: Likewise (within example in comment).
926 * lto-cgraph.c (input_cgraph_opt_summary): Likewise.
927 (lto_symtab_encoder_encode): Likewise.
928 (lto_symtab_encoder_delete_node): Likewise.
929 (lto_symtab_encoder_in_partition_p): Likewise.
930 (lto_set_symtab_encoder_in_partition): Likewise.
931 (output_refs): Likewise.
932 (compute_ltrans_boundary): Likewise.
933 (output_symtab): Likewise.
934 (input_node): Likewise.
935 (input_ref): Likewise.
936 (input_edge): Likewise.
937 (input_cgraph_1): Likewise.
938 (input_refs): Likewise.
939 (output_cgraph_opt_summary): Likewise.
940 (input_node_opt_summary): Likewise.
941 (input_cgraph_opt_section): Likewise.
942 * lto-section-in.c (lto_free_function_in_decl_state_for_node):
944 * lto-streamer-out.c (lto_output): Likewise.
945 (output_symbol_p): Likewise.
946 (produce_symtab): Likewise.
947 * lto-streamer.h (lto_encoder_entry): Likewise.
948 (lto_free_function_in_decl_state_for_node): Likewise.
949 (lto_symtab_encoder_encode): Likewise.
950 (lto_symtab_encoder_delete_node): Likewise.
951 (lto_symtab_encoder_in_partition_p): Likewise.
952 (lto_set_symtab_encoder_in_partition): Likewise.
953 (lto_symtab_encoder_lookup): Likewise.
954 (lsei_node): Likewise.
955 (lto_symtab_encoder_deref): Likewise.
956 * symtab.c (symtab_hash): Likewise.
957 (assembler_name_hash): Likewise.
958 (symtab_nodes): Likewise.
959 (hash_node): Likewise.
961 (hash_node_by_assembler_name): Likewise.
962 (eq_assembler_name): Likewise.
963 (insert_to_assembler_name_hash): Likewise.
964 (unlink_from_assembler_name_hash): Likewise.
965 (symtab_prevail_in_asm_name_hash): Likewise.
966 (symtab_register_node): Likewise.
967 (symtab_insert_node_to_hashtable): Likewise.
968 (symtab_unregister_node): Likewise.
969 (symtab_get_node): Likewise.
970 (symtab_remove_node): Likewise.
971 (symtab_initialize_asm_name_hash): Likewise.
972 (symtab_node_for_asm): Likewise.
973 (symtab_add_to_same_comdat_group): Likewise.
974 (symtab_dissolve_same_comdat_group_list): Likewise.
975 (symtab_node_asm_name): Likewise.
976 (symtab_node_name): Likewise.
977 (dump_symtab_base): Likewise.
978 (dump_symtab_node): Likewise.
979 (dump_symtab): Likewise.
980 (debug_symtab_node): Likewise.
981 (verify_symtab_base): Likewise.
982 (verify_symtab_node): Likewise.
983 (verify_symtab): Likewise.
984 (symtab_used_from_object_file_p): Likewise.
985 (symtab_node_availability): Likewise.
986 (symtab_alias_ultimate_target): Likewise.
987 (fixup_same_cpp_alias_visibility): Likewise.
988 (symtab_resolve_alias): Likewise.
989 (symtab_for_node_and_aliases): Likewise.
990 (symtab_for_node_and_aliases): Likewise.
991 (symtab_nonoverwritable_alias_1): Likewise.
992 (symtab_nonoverwritable_alias): Likewise.
993 (symtab_semantically_equivalent_p): Likewise.
994 * value-prof.c (init_node_map): Likewise.
995 * varasm.c (find_decl): Likewise.
996 * varpool.c (varpool_node_for_asm): Likewise.
997 (varpool_remove_unreferenced_decls): Likewise.
999 2013-10-31 David Malcolm <dmalcolm@redhat.com>
1001 Manual part of renaming of symtab_node_base to symtab_node.
1003 * ipa-ref.h (symtab_node): Remove typedef to pointer type, as it
1004 clashes with the preferred name for the base class.
1005 (const_symtab_node): Remove redundant typedef.
1007 2013-10-31 Jakub Jelinek <jakub@redhat.com>
1009 * optabs.c (expand_vec_perm): Avoid vector mode punning
1010 SUBREGs in SET_DEST.
1011 * expmed.c (store_bit_field_1): Likewise.
1012 * config/i386/sse.md (movdi_to_sse, vec_pack_sfix_trunc_v2df,
1013 vec_pack_sfix_v2df, vec_shl_<mode>, vec_shr_<mode>,
1014 vec_interleave_high<mode>, vec_interleave_low<mode>): Likewise.
1015 * config/i386/i386.c (ix86_expand_vector_move_misalign,
1016 ix86_expand_sse_movcc, ix86_expand_int_vcond, ix86_expand_vec_perm,
1017 ix86_expand_sse_unpack, ix86_expand_args_builtin,
1018 ix86_expand_vector_init_duplicate, ix86_expand_vector_set,
1019 emit_reduc_half, expand_vec_perm_blend, expand_vec_perm_pshufb,
1020 expand_vec_perm_interleave2, expand_vec_perm_pshufb2,
1021 expand_vec_perm_vpshufb2_vpermq,
1022 expand_vec_perm_vpshufb2_vpermq_even_odd, expand_vec_perm_even_odd_1,
1023 expand_vec_perm_broadcast_1, expand_vec_perm_vpshufb4_vpermq2,
1024 ix86_expand_sse2_mulv4si3, ix86_expand_pinsr): Likewise.
1025 (expand_vec_perm_palignr): Likewise. Modify a copy of *d rather
1028 2013-10-31 Uros Bizjak <ubizjak@gmail.com>
1030 * config/i386/i386.c (ix86_expand_sse2_abs): Rename function arguments.
1031 Use gcc_unreachable for unhandled modes. Do not check results of
1032 expand_simple_binop. If not expanded to target, move the result.
1034 2013-10-31 Chung-Ju Wu <jasonwucj@gmail.com>
1035 Shiva Chen <shiva0217@gmail.com>
1037 * config.gcc (nds32*-*-*): Add nds32 target.
1038 * config/nds32/nds32.c: New file.
1039 * config/nds32/nds32.h: New file.
1040 * config/nds32/nds32.md: New file.
1041 * config/nds32/constants.md: New file.
1042 * config/nds32/constraints.md: New file.
1043 * config/nds32/iterators.md: New file.
1044 * config/nds32/nds32-doubleword.md: New file.
1045 * config/nds32/nds32-intrinsic.md: New file.
1046 * config/nds32/nds32_intrinsic.h: New file.
1047 * config/nds32/nds32-modes.def: New file.
1048 * config/nds32/nds32-multiple.md: New file.
1049 * config/nds32/nds32.opt: New file.
1050 * config/nds32/nds32-opts.h: New file.
1051 * config/nds32/nds32-protos.h: New file.
1052 * config/nds32/nds32-peephole2.md: New file.
1053 * config/nds32/pipelines.md: New file.
1054 * config/nds32/predicates.md: New file.
1055 * config/nds32/t-mlibs: New file.
1056 * common/config/nds32: New directory and files.
1058 * doc/invoke.texi (NDS32 options): Document nds32 specific options.
1059 * doc/md.texi (NDS32 family): Document nds32 specific constraints.
1060 * doc/install.texi (Cross-Compiler-Specific Options): Document
1061 --with-nds32-lib for nds32 target.
1062 * doc/extend.texi (Function Attributes, Target Builtins): Document
1063 nds32 specific attributes.
1065 2013-10-31 Vladimir Makarov <vmakarov@redhat.com>
1067 * lra-constraints (process_alt_operands): Use the result
1068 elimination register for operand when matching constraints.
1070 2013-10-31 Jakub Jelinek <jakub@redhat.com>
1072 * tree-vrp.c (maybe_set_nonzero_bits): New function.
1073 (remove_range_assertions): Call it.
1075 * tree.c (tree_ctz): New function.
1076 * tree.h (tree_ctz): New prototype.
1077 * tree-ssanames.h (get_range_info, get_nonzero_bits): Change
1078 first argument from tree to const_tree.
1079 * tree-ssanames.c (get_range_info, get_nonzero_bits): Likewise.
1080 * tree-vectorizer.h (vect_generate_tmps_on_preheader): New prototype.
1081 * tree-vect-loop-manip.c (vect_generate_tmps_on_preheader): No longer
1083 * expr.c (highest_pow2_factor): Reimplemented using tree_ctz.
1084 * tree-vect-loop.c (vect_analyze_loop_operations,
1085 vect_transform_loop): Don't force scalar loop for bound just because
1086 number of iterations is unknown, only do it if it is not known to be
1087 a multiple of vectorization_factor.
1088 * builtins.c (get_object_alignment_2): Use tree_ctz on offset.
1090 * gimple-pretty-print.c (dump_ssaname_info): Print newline also
1091 in case of VR_VARYING. Print get_nonzero_bits if not all ones.
1092 * tree-ssanames.h (struct range_info_def): Add nonzero_bits field.
1093 (set_nonzero_bits, get_nonzero_bits): New prototypes.
1094 * tree-ssa-ccp.c (get_default_value): Use get_range_info to see if
1095 a default def isn't partially constant.
1096 (ccp_finalize): If after IPA, set_range_info if integral SSA_NAME
1097 is known to be partially zero.
1098 (evaluate_stmt): If we'd return otherwise VARYING, use get_range_info
1099 to see if a default def isn't partially constant.
1100 * tree-ssanames.c (set_range_info): Initialize nonzero_bits upon
1101 creation of a range, if VR_RANGE, try to improve nonzero_bits from
1103 (set_nonzero_bits, get_nonzero_bits): New functions.
1105 * tree-cfg.c (assert_unreachable_fallthru_edge_p): New function.
1106 * tree-cfg.h (assert_unreachable_fallthru_edge_p): New prototype.
1107 * tree-vrp.c (all_imm_uses_in_stmt_or_feed_cond): New function.
1108 (remove_range_assertions): If ASSERT_EXPR_VAR has no other immediate
1109 uses but in the condition and ASSERT_EXPR and the other successor of
1110 the predecessor bb is __builtin_unreachable (), set_range_info of the
1111 ASSERT_EXPR_VAR to the range info of the ASSERT_EXPR's lhs.
1113 2013-10-31 Martin Jambor <mjambor@suse.cz>
1115 PR rtl-optimization/58934
1117 2013-10-30 Martin Jambor <mjambor@suse.cz>
1118 PR rtl-optimization/10474
1119 * ira.c (find_moveable_pseudos): Do not calculate dominance info
1121 (interesting_dest_for_shprep): New function.
1122 (split_live_ranges_for_shrink_wrap): Likewise.
1123 (ira): Calculate dominance info and df analysis. Call
1124 split_live_ranges_for_shrink_wrap.
1126 2013-10-31 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
1127 Yury Gribov <y.gribov@samsung.com>
1130 * asan.c (asan_clear_shadow): Allocate a new vreg for temporary
1131 shadow pointer to avoid clobbering the main one.
1133 2013-10-31 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1135 * lower-subreg.c (resolve_simple_move): Copy REG_INC note.
1137 2013-10-30 Vladimir Makarov <vmakarov@redhat.com>
1140 * ira-color.c (update_costs_from_copies): Add new parameter. Use
1141 it for calling update_costs_from_allocno.
1142 (assign_hard_reg): Call restore_costs_from_copies only for
1143 !retry_p. Pass new argument to update_costs_from_copies.
1144 (color_pass): Pass new argument to update_costs_from_copies.
1145 (ira_mark_allocation_change): Ditto.
1147 2013-10-30 Sharad Singhai <singhai@google.com>
1150 * opts.c (common_handle_option): Remove deprecated option
1151 -ftree-vectorizer-verbose.
1152 * doc/invoke.texi (Debugging Options): Ditto.
1153 * opts-global.c (handle_common_deferred_options): Ditto.
1154 (dump_remap_tree_vectorizer_verbose): Delete.
1155 * common.opt: Set -ftree-vectorizer-verbose as an ignored option.
1157 2013-10-30 DJ Delorie <dj@redhat.com>
1159 * config/rx/rx.c (ADD_RX_BUILTIN0): New macro, used for builtins
1160 that take no arguments.
1162 2013-10-30 Joern Rennecke <joern.rennecke@embecosm.com>
1165 * reload1.c (update_eliminables_and_spill): New function, broken
1167 (reload): Use it. Check for frame size change after frame size
1168 alignment, and call update_eliminables_and_spill first if continue-ing.
1170 2013-10-30 Cong Hou <congh@google.com>
1173 * config/i386/i386-protos.h (ix86_expand_sse2_abs): New function.
1174 * config/i386/i386.c (ix86_expand_sse2_abs): New function.
1175 * config/i386/sse.md: Add SSE2 support to abs (8/16/32-bit-int).
1177 2013-10-18 Mikael Pettersson <mikpelinux@gmail.com>
1179 PR rtl-optimization/58369
1180 * reload1.c (compute_reload_subreg_offset): New function.
1181 (choose_reload_regs): Use it to pass endian-correct
1182 offset to subreg_regno_offset.
1184 2013-10-30 Tobias Burnus <burnus@net-b.de>
1187 * tree-cfg.c (replace_loop_annotate): Replace warning by
1190 2013-10-30 Jason Merrill <jason@redhat.com>
1192 * configure.ac (loose_warn): Add -Wno-format if
1193 --disable-build-format-warnings.
1195 2013-10-30 David Malcolm <dmalcolm@redhat.com>
1197 * cgraphunit.c (analyze_functions): Split symtab_node declarations
1198 onto multiple lines to make things easier for rename_symtab.py.
1200 * symtab.c (symtab_dissolve_same_comdat_group_list): Likewise.
1201 (symtab_semantically_equivalent_p): Likewise.
1203 2013-10-30 Vladimir Makarov <vmakarov@redhat.com>
1206 * lra.c (check_rtl): Remove address check before LRA work.
1208 2013-10-30 Marc Glisse <marc.glisse@inria.fr>
1210 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Look for a
1211 POINTER_PLUS_EXPR in the defining statement.
1213 2013-10-30 Vladimir Makarov <vmakarov@redhat.com>
1215 * regmove.c: Remove.
1216 * tree-pass.h (make_pass_regmove): Remove.
1217 * timevar.def (TV_REGMOVE): Remove.
1218 * passes.def (pass_regmove): Remove.
1219 * opts.c (default_options_table): Remove entry for regmove.
1220 * doc/passes.texi: Remove regmove pass description.
1221 * doc/invoke.texi (-foptimize-register-move, -fregmove): Remove
1223 (-fdump-rtl-regmove): Ditto.
1224 * common.opt (foptimize-register-move, fregmove): Ignore.
1225 * Makefile.in (OBJS): Remove regmove.o.
1226 * regmove.c: Remove.
1227 * ira-int.h (struct ira_allocno_pref, ira_pref_t): New structure
1229 (struct ira_allocno) New member allocno_prefs.
1230 (ALLOCNO_PREFS): New macro.
1231 (ira_prefs, ira_prefs_num): New external vars.
1232 (ira_setup_alts, ira_get_dup_out_num, ira_debug_pref): New prototypes.
1233 (ira_debug_prefs, ira_debug_allocno_prefs, ira_create_pref): Ditto.
1234 (ira_add_allocno_pref, ira_remove_pref, ira_remove_allocno_prefs):
1236 (ira_add_allocno_copy_to_list): Remove prototype.
1237 (ira_swap_allocno_copy_ends_if_necessary): Ditto.
1238 (ira_pref_iterator): New type.
1239 (ira_pref_iter_init, ira_pref_iter_cond): New functions.
1240 (FOR_EACH_PREF): New macro.
1241 * ira.c (commutative_constraint_p): Move from ira-conflicts.c.
1242 (ira_get_dup_out_num): Ditto. Rename from get_dup_num. Modify the
1244 (ira_setup_alts): New function.
1245 (decrease_live_ranges_number): New function.
1246 (ira): Call the above function.
1247 * ira-build.c (ira_prefs, ira_prefs_num): New global vars.
1248 (ira_create_allocno): Initialize allocno prefs.
1249 (pref_pool, pref_vec): New static vars.
1250 (initiate_prefs, find_allocno_pref, ira_create_pref): New
1252 (add_allocno_pref_to_list, ira_add_allocno_pref, print_pref): Ditto.
1253 (ira_debug_pref, print_prefs, ira_debug_prefs): Ditto.
1254 (print_allocno_prefs, ira_debug_allocno_prefs, finish_pref): Ditto.
1255 (ira_remove_pref, ira_remove_allocno_prefs, finish_prefs): Ditto.
1256 (ira_add_allocno_copy_to_list): Make static. Rename to
1257 add_allocno_copy_to_list.
1258 (ira_swap_allocno_copy_ends_if_necessary): Make static. Rename to
1259 swap_allocno_copy_ends_if_necessary.
1260 (remove_unnecessary_allocnos, remove_low_level_allocnos): Call
1261 ira_remove_allocno_prefs.
1262 (ira_flattening): Ditto.
1263 (ira_build): Call initiate_prefs, print_prefs.
1264 (ira_destroy): Call finish_prefs.
1265 * ira-color.c (struct update_cost_record): New.
1266 (struct allocno_color_data): Add new member update_cost_records.
1267 (update_cost_record_pool): New static var.
1268 (init_update_cost_records, get_update_cost_record): New functions.
1269 (free_update_cost_record_list, finish_update_cost_records): Ditto.
1270 (struct update_cost_queue_elem): Add member from.
1271 (initiate_cost_update): Call init_update_cost_records.
1272 (finish_cost_update): Call finish_update_cost_records.
1273 (queue_update_cost, get_next_update_cost): Add new param from.
1274 (Update_allocno_cost, update_costs_from_allocno): New functions.
1275 (update_costs_from_prefs): Ditto.
1276 (update_copy_costs): Rename to update_costs_from_copies.
1277 (restore_costs_from_copies): New function.
1278 (update_conflict_hard_regno_costs): Don't go back.
1279 (assign_hard_reg): Call restore_costs_from_copies. Add printing
1281 (pop_allocnos): Add priniting more debug info.
1282 (color_allocnos): Remove prefs for conflicting hard regs.
1283 Call update_costs_from_prefs.
1284 * ira-conflicts.c (commutative_constraint_p): Move to ira.c
1285 (get_dup_num): Rename, modify, and move to ira.c
1286 (process_regs_for_copy): Add prefs.
1287 (add_insn_allocno_copies): Put src as first arg of
1288 process_regs_for_copy. Remove dead code. Call ira_setup_alts.
1289 * ira-costs.c (record_reg_classes): Modify and move code into
1290 record_operands_costs.
1291 (find_costs_and_classes): Create prefs for the hard reg of small
1293 (process_bb_node_for_hard_reg_moves): Add prefs.
1295 2013-10-30 Richard Biener <rguenther@suse.de>
1298 * basic-block.h (pre_and_rev_post_order_compute_fn): New function.
1299 * cfganal.c (pre_and_rev_post_order_compute_fn): New function
1301 (pre_and_rev_post_order_compute): ... which now wraps it.
1302 * graph.c (draw_cfg_nodes_no_loops): Use
1303 pre_and_rev_post_order_compute_fn to avoid ICEing and dependence
1306 2013-10-30 Christian Bruel <christian.bruel@st.com>
1308 * config/sh/sh-mem.cc (sh_expand_cmpnstr): New function.
1309 (sh_expand_cmpstr): Handle known align and schedule improvements.
1310 * config/sh/sh-protos.h (sh_expand_cmpstrn): Declare.
1311 * config/sh/sh.md (cmpstrnsi): New pattern.
1313 2013-10-30 Martin Jambor <mjambor@suse.cz>
1315 PR rtl-optimization/10474
1316 * ira.c (find_moveable_pseudos): Do not calculate dominance info
1318 (interesting_dest_for_shprep): New function.
1319 (split_live_ranges_for_shrink_wrap): Likewise.
1320 (ira): Calculate dominance info and df analysis. Call
1321 split_live_ranges_for_shrink_wrap.
1323 2013-10-30 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1326 * config/arm/arm.c (arm_expand_epilogue_apcs_frame): Emit blockage.
1328 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
1330 * tree-core.h (tree_index): Add TI_POINTER_BOUNDS_TYPE.
1331 * tree.h (POINTER_BOUNDS_P): New.
1332 (BOUNDED_TYPE_P): New.
1334 (pointer_bounds_type_node): New.
1335 * tree.c (build_common_tree_nodes): Initialize
1336 pointer_bounds_type_node.
1337 * gimple.h (gimple_call_get_nobnd_arg_index): New.
1338 (gimple_call_num_nobnd_args): New.
1339 (gimple_call_nobnd_arg): New.
1340 (gimple_return_retbnd): New.
1341 (gimple_return_set_retbnd): New
1342 * gimple.c (gimple_build_return): Increase number of ops
1343 for return statement.
1344 (gimple_call_get_nobnd_arg_index): New.
1345 * gimple-pretty-print.c (dump_gimple_return): Print second op.
1347 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
1349 * ipa.c (cgraph_build_static_cdtor_1): Support contructors
1350 with "chkp ctor" and "bnd_legacy" attributes.
1351 * gimplify.c (gimplify_init_constructor): Avoid infinite
1352 loop during gimplification of bounds initializer.
1354 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
1356 * c-family/c-common.c (handle_bnd_variable_size_attribute): New.
1357 (handle_bnd_legacy): New.
1358 (c_common_attribute_table): Add bnd_variable_size and bnd_legacy.
1359 * doc/extend.texi: Document bnd_variable_size and bnd_legacy
1362 2013-10-29 Ilya Enkovich <ilya.enkovich@intel.com>
1364 * builtin-types.def (BT_FN_VOID_CONST_PTR): New.
1365 (BT_FN_PTR_CONST_PTR): New.
1366 (BT_FN_CONST_PTR_CONST_PTR): New.
1367 (BT_FN_PTR_CONST_PTR_SIZE): New.
1368 (BT_FN_PTR_CONST_PTR_CONST_PTR): New.
1369 (BT_FN_VOID_PTRPTR_CONST_PTR): New.
1370 (BT_FN_VOID_CONST_PTR_SIZE): New.
1371 (BT_FN_PTR_CONST_PTR_CONST_PTR_SIZE): New.
1372 * chkp-builtins.def: New.
1373 * builtins.def: include chkp-builtins.def.
1374 (DEF_CHKP_BUILTIN): New.
1375 * builtins.c (expand_builtin): Support BUILT_IN_CHKP_INIT_PTR_BOUNDS,
1376 BUILT_IN_CHKP_NULL_PTR_BOUNDS, BUILT_IN_CHKP_COPY_PTR_BOUNDS,
1377 BUILT_IN_CHKP_CHECK_PTR_LBOUNDS, BUILT_IN_CHKP_CHECK_PTR_UBOUNDS,
1378 BUILT_IN_CHKP_CHECK_PTR_BOUNDS, BUILT_IN_CHKP_SET_PTR_BOUNDS,
1379 BUILT_IN_CHKP_NARROW_PTR_BOUNDS, BUILT_IN_CHKP_STORE_PTR_BOUNDS,
1380 BUILT_IN_CHKP_GET_PTR_LBOUND, BUILT_IN_CHKP_GET_PTR_UBOUND,
1381 BUILT_IN_CHKP_BNDMK, BUILT_IN_CHKP_BNDSTX, BUILT_IN_CHKP_BNDCL,
1382 BUILT_IN_CHKP_BNDCU, BUILT_IN_CHKP_BNDLDX, BUILT_IN_CHKP_BNDRET,
1383 BUILT_IN_CHKP_INTERSECT, BUILT_IN_CHKP_ARG_BND, BUILT_IN_CHKP_NARROW,
1384 BUILT_IN_CHKP_EXTRACT_LOWER, BUILT_IN_CHKP_EXTRACT_UPPER.
1385 * common.opt (fcheck-pointer-bounds): New.
1386 * toplev.c (process_options): Check Pointer Bounds Checker is
1388 * doc/extend.texi: Document Pointer Bounds Checker built-in functions.
1390 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
1392 * target.def (builtin_chkp_function): New.
1393 (chkp_bound_type): New.
1394 (chkp_bound_mode): New.
1395 (fn_abi_va_list_bounds_size): New.
1396 (load_bounds_for_arg): New.
1397 (store_bounds_for_arg): New.
1398 * targhooks.h (default_load_bounds_for_arg): New.
1399 (default_store_bounds_for_arg): New.
1400 (default_fn_abi_va_list_bounds_size): New.
1401 (default_chkp_bound_type): New.
1402 (default_chkp_bound_mode): New.
1403 (default_builtin_chkp_function): New.
1404 * targhooks.c (default_load_bounds_for_arg): New.
1405 (default_store_bounds_for_arg): New.
1406 (default_fn_abi_va_list_bounds_size): New.
1407 (default_chkp_bound_type): New.
1408 (default_chkp_bound_mode); New.
1409 (default_builtin_chkp_function): New.
1410 * doc/tm.texi.in (TARGET_FN_ABI_VA_LIST_BOUNDS_SIZE): New.
1411 (TARGET_LOAD_BOUNDS_FOR_ARG): New.
1412 (TARGET_STORE_BOUNDS_FOR_ARG): New.
1413 (TARGET_BUILTIN_CHKP_FUNCTION): New.
1414 (TARGET_CHKP_BOUND_TYPE): New.
1415 (TARGET_CHKP_BOUND_MODE): New.
1416 * doc/tm.texi: Regenerated.
1417 * langhooks.h (lang_hooks): Add chkp_supported field.
1418 * langhooks-def.h (LANG_HOOKS_CHKP_SUPPORTED): New.
1419 (LANG_HOOKS_INITIALIZER); Add LANG_HOOKS_CHKP_SUPPORTED.
1421 2013-10-29 Andrew Pinski <apinski@cavium.com>
1423 * tree-ssa-ifcombine.c: Include rtl.h and tm_p.h.
1424 (ifcombine_ifandif): Handle cases where maybe_fold_and_comparisons
1425 fails, combining the branches anyways.
1426 (tree_ssa_ifcombine): Inverse the order of the basic block walk,
1427 increases the number of combinings.
1428 * gimple.h (gsi_start_nondebug_after_labels_bb): New function.
1430 2013-10-29 Mike Stump <mikestump@comcast.net>
1432 * machmode.def (PARTIAL_INT_MODE): Add precision and name.
1433 * genmodes.c (PARTIAL_INT_MODE): Add precision and name.
1434 (make_vector_mode): Increase namebuf to 16.
1435 (emit_insn_modes_h): When processing BImode, don't
1436 also match partial int modes.
1437 (emit_class_narrowest_mode): Likewise.
1439 * config/bfin/bfin-modes.def: Add precision to PDI.
1440 * config/m32c/m32c-modes.def: Add precision to PSI.
1441 * config/msp430/msp430-modes.def: Add precision to PSI.
1442 * config/rs6000/rs6000-modes.def: Add precision to PTI.
1443 * config/sh/sh-modes.def: Add precision to PSI and PDI.
1445 2013-10-29 Oleg Endo <olegendo@gcc.gnu.org>
1448 * config/sh/sh.md (*addc): Rename existing variations to ...
1449 (*addc_r_r_1, *addc_2r_1, *addc_r_1): ... these.
1450 (*addc_r_lsb, *addc_r_r_lsb, *addc_r_lsb_r, *addc_2r_lsb, *addc_r_msb,
1451 *addc_r_r_msb, *addc_2r_msb): New insn_and_split patterns.
1452 * config/sh/sh.c (addsubcosts): Handle some addc special cases.
1454 2013-10-29 Teresa Johnson <tejohnson@google.com>
1457 * tree-ssa-tail-merge.c (replace_block_by): Tolerate profile
1458 insanities when updating probabilities.
1460 2013-10-29 David Malcolm <dmalcolm@redhat.com>
1462 * gdbhooks.py (CGraphNodePrinter.to_string): Update gdb
1463 prettyprinter for cgraph_node to reflect the conversion of the
1464 symtable types to a C++ class hierarchy: it now *is* a
1465 symtab_node_base, rather than having one (named "symbol").
1467 2013-10-29 Balaji V. Iyer <balaji.v.iyer@intel.com>
1469 * builtins.c (is_builtin_name): Added a check for __cilkrts_detach and
1470 __cilkrts_pop_frame. If matched, then return true for built-in
1472 (expand_builtin): Added BUILT_IN_CILK_DETACH and
1473 BUILT_IN_CILK_POP_FRAME case.
1474 * langhooks-def.h (lhd_install_body_with_frame_cleanup): New prototype.
1475 (lhs_cilk_detect_spawn): Likewise.
1476 (LANG_HOOKS_DECLS): Added LANG_HOOKS_CILKPLUS.
1477 (LANG_HOOKS_CILKPLUS_DETECT_SPAWN_AND_UNWRAP): New #define.
1478 (LANG_HOOKS_CILKPLUS_FRAME_CLEANUP): Likewise.
1479 (LANG_HOOKS_CILKPLUS_GIMPLIFY_SPAWN): Likewise.
1480 (LANG_HOOKS_CILKPLUS): Likewise.
1481 * tree.h (CILK_SPAWN_FN): Likewise.
1482 * builtin.def (DEF_CILK_BUILTIN_STUB): Likewise.
1483 * Makefile.in (C_COMMON_OBJS): Added c-family/cilk.o.
1484 (OBJS): Added cilk-common.o.
1485 (BUILTINS_DEF): Added cilk-builtins.def.
1486 * langhooks.c (lhd_install_body_with_frame_cleanup): New function.
1487 (lhd_cilk_detect_spawn): Likewise.
1488 * langhooks.h (lang_hooks_for_cilkplus): New struct.
1489 (struct lang_hooks): Added new field called "cilkplus."
1490 * cilk-common.c: New file.
1492 * cilk-builtins.def: Likewise.
1493 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Added
1494 "__cilk" macro and set it to 200.
1495 * function.h (struct function::cilk_frame_decl): New field.
1496 (struct function::is_cilk_function): Likewise.
1497 (struct function::calls_cilk_spawn): Likewise.
1498 * gimplify.c (gimplify_call_expr): Added a check if the function call
1499 being gimplified is a spawn detach point. If so, then add pop_frame
1500 and detach function calls.
1501 (gimplify_expr): Added a CILK_SPAWN_STMT and CILK_SYNC_STMT case
1502 for gimplifying _Cilk_spawn and _Cilk_sync statements.
1503 (gimplify_return_expr): Added a check for _Cilk_spawn usage in
1504 function. If so, added a _Cilk_sync and gimplified it.
1505 (gimplify_modify_expr): Added a check for _Cilk_spawn in MODIFY and
1506 INIT_EXPRs. If so, then call gimplify_cilk_spawn.
1507 * ipa-inline-analysis (initialize_inline_failed): Prevent inlining of
1509 (can_inline_edge_p): Prevent inling of spawnee function.
1510 * ira.c (ira_setup_eliminable_regset): Force usage of frame pointer
1511 for functions that use Cilk keywords.
1512 * tree-inline.h (struct copy_body_data::remap_var_for_cilk): New field.
1513 * tree-pretty-print.c (dump_generic_node): Added CILK_SPAWN_STMT and
1514 CILK_SYNC_STMT cases.
1515 * tree.def (DEFTREECODE): Added CILK_SPAWN_STMT and CILK_SYNC_STMT
1517 * generic.texi (CILK_SPAWN_STMT): Added documentation for _Cilk_spawn.
1518 (CILK_SYNC_STMT): Added documentation for _Cilk_sync.
1519 * passes.texi (Cilk Keywords): New section that describes the compiler
1520 code changes for handling Cilk Keywords.
1522 2013-10-29 David Malcolm <dmalcolm@redhat.com>
1524 Patch autogenerated by refactor_symtab.py from
1525 https://github.com/davidmalcolm/gcc-refactoring-scripts
1526 revision 58bb219cc090b2f4516a9297d868c245495ee622
1528 * asan.c (asan_finish_file): Update for conversion of symtab types to
1529 a true class hierarchy.
1530 * cfgexpand.c (estimated_stack_frame_size): Likewise.
1531 * cgraph.c (cgraph_get_body): Likewise.
1532 (cgraph_get_create_real_symbol_node): Likewise.
1533 (verify_cgraph_node): Likewise.
1534 (verify_edge_corresponds_to_fndecl): Likewise.
1535 (verify_edge_count_and_frequency): Likewise.
1536 (cgraph_will_be_removed_from_program_if_no_direct_calls): Likewise.
1537 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
1538 (cgraph_can_remove_if_no_direct_calls_and_refs_p): Likewise.
1539 (cgraph_node_cannot_return): Likewise.
1540 (cgraph_set_pure_flag_1): Likewise.
1541 (cgraph_set_const_flag_1): Likewise.
1542 (cgraph_set_nothrow_flag_1): Likewise.
1543 (cgraph_make_node_local_1): Likewise.
1544 (cgraph_for_node_and_aliases): Likewise.
1545 (cgraph_for_node_thunks_and_aliases): Likewise.
1546 (cgraph_node_can_be_local_p): Likewise.
1547 (cgraph_node_cannot_be_local_p_1): Likewise.
1548 (cgraph_function_body_availability): Likewise.
1549 (dump_cgraph_node): Likewise.
1550 (cgraph_rtl_info): Likewise.
1551 (cgraph_mark_address_taken_node): Likewise.
1552 (cgraph_remove_node): Likewise.
1553 (cgraph_release_function_body): Likewise.
1554 (cgraph_update_edges_for_call_stmt_node): Likewise.
1555 (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
1556 (cgraph_make_edge_direct): Likewise.
1557 (cgraph_resolve_speculation): Likewise.
1558 (cgraph_speculative_call_info): Likewise.
1559 (cgraph_turn_edge_to_speculative): Likewise.
1560 (cgraph_create_edge_1): Likewise.
1561 (cgraph_set_call_stmt): Likewise.
1562 (cgraph_node_for_asm): Likewise.
1563 (cgraph_add_thunk): Likewise.
1564 (cgraph_same_body_alias): Likewise.
1565 (cgraph_create_function_alias): Likewise.
1566 (cgraph_create_node): Likewise.
1567 (cgraph_create_empty_node): Likewise.
1568 (record_function_versions): Likewise.
1569 (used_from_object_file_p): Likewise.
1570 * cgraph.h (symtab_can_be_discarded): Likewise.
1571 (symtab_real_symbol_p): Likewise.
1572 (cgraph_mark_force_output_node): Likewise.
1573 (cgraph_edge_recursive_p): Likewise.
1574 (symtab_alias_target): Likewise.
1575 (varpool_all_refs_explicit_p): Likewise.
1576 (varpool_can_remove_if_no_refs): Likewise.
1577 (cgraph_only_called_directly_or_aliased_p): Likewise.
1578 (cgraph_next_function_with_gimple_body): Likewise.
1579 (cgraph_first_function_with_gimple_body): Likewise.
1580 (cgraph_function_with_gimple_body_p): Likewise.
1581 (cgraph_next_function): Likewise.
1582 (cgraph_first_function): Likewise.
1583 (cgraph_next_defined_function): Likewise.
1584 (cgraph_first_defined_function): Likewise.
1585 (varpool_next_defined_variable): Likewise.
1586 (varpool_first_defined_variable): Likewise.
1587 (varpool_next_static_initializer): Likewise.
1588 (varpool_first_static_initializer): Likewise.
1589 (varpool_next_variable): Likewise.
1590 (varpool_first_variable): Likewise.
1591 (varpool_node_name): Likewise.
1592 (varpool): Likewise.
1594 (is_a_helper <varpool_node>::test): Likewise.
1595 (is_a_helper <cgraph_node>::test): Likewise.
1596 (varpool_variable_node): Likewise.
1597 (cgraph_function_or_thunk_node): Likewise.
1598 (varpool_alias_target): Likewise.
1599 (cgraph_alias_target): Likewise.
1600 (cgraph_node_name): Likewise.
1601 (varpool_node_asm_name): Likewise.
1602 (cgraph_node_asm_name): Likewise.
1603 * cgraphbuild.c (remove_cgraph_callee_edges): Likewise.
1604 (cgraph_rebuild_references): Likewise.
1605 (rebuild_cgraph_edges): Likewise.
1606 (record_eh_tables): Likewise.
1607 (build_cgraph_edges): Likewise.
1608 (mark_store): Likewise.
1609 (mark_load): Likewise.
1610 (mark_address): Likewise.
1611 (record_type_list): Likewise.
1612 (record_reference): Likewise.
1613 * cgraphclones.c (cgraph_materialize_all_clones): Likewise.
1614 (cgraph_materialize_clone): Likewise.
1615 (cgraph_function_versioning): Likewise.
1616 (cgraph_copy_node_for_versioning): Likewise.
1617 (update_call_expr): Likewise.
1618 (cgraph_find_replacement_node): Likewise.
1619 (cgraph_create_virtual_clone): Likewise.
1620 (cgraph_clone_node): Likewise.
1621 * cgraphunit.c (compile): Likewise.
1622 (output_weakrefs): Likewise.
1623 (output_in_order): Likewise.
1624 (expand_function): Likewise.
1625 (assemble_thunks_and_aliases): Likewise.
1626 (expand_thunk): Likewise.
1627 (mark_functions_to_output): Likewise.
1628 (handle_alias_pairs): Likewise.
1629 (analyze_functions): Likewise.
1630 (walk_polymorphic_call_targets): Likewise.
1631 (varpool_finalize_decl): Likewise.
1632 (process_function_and_variable_attributes): Likewise.
1633 (cgraph_process_same_body_aliases): Likewise.
1634 (analyze_function): Likewise.
1635 (cgraph_add_new_function): Likewise.
1636 (cgraph_finalize_function): Likewise.
1637 (referred_to_p): Likewise.
1638 (cgraph_reset_node): Likewise.
1639 (cgraph_process_new_functions): Likewise.
1640 (enqueue_node): Likewise.
1641 (decide_is_symbol_needed): Likewise.
1642 * coverage.c (coverage_compute_profile_id): Likewise.
1643 * dbxout.c (dbxout_expand_expr): Likewise.
1644 * dwarf2out.c (premark_types_used_by_global_vars_helper): Likewise.
1645 (reference_to_unused): Likewise.
1646 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
1647 * gimplify.c (unvisit_body): Likewise.
1648 (unshare_body): Likewise.
1649 * ipa-cp.c (ipcp_generate_summary): Likewise.
1650 (ipcp_decision_stage): Likewise.
1651 (identify_dead_nodes): Likewise.
1652 (decide_whether_version_node): Likewise.
1653 (decide_about_value): Likewise.
1654 (perhaps_add_new_callers): Likewise.
1655 (create_specialized_node): Likewise.
1656 (update_profiling_info): Likewise.
1657 (ipcp_propagate_stage): Likewise.
1658 (estimate_local_effects): Likewise.
1659 (good_cloning_opportunity_p): Likewise.
1660 (devirtualization_time_bonus): Likewise.
1661 (propagate_constants_accross_call): Likewise.
1662 (initialize_node_lattices): Likewise.
1663 (ipcp_cloning_candidate_p): Likewise.
1664 (determine_versionability): Likewise.
1665 (print_all_lattices): Likewise.
1666 (print_lattice): Likewise.
1667 (ipcp_discover_new_direct_edges): Likewise.
1668 * ipa-devirt.c (ipa_devirt): Likewise.
1669 (likely_target_p): Likewise.
1670 (update_type_inheritance_graph): Likewise.
1671 (possible_polymorphic_call_target_p): Likewise.
1672 (dump_possible_polymorphic_call_targets): Likewise.
1673 (devirt_variable_node_removal_hook): Likewise.
1674 (record_binfo): Likewise.
1675 (maybe_record_node): Likewise.
1676 (build_type_inheritance_graph): Likewise.
1677 * ipa-inline-analysis.c (inline_write_summary): Likewise.
1678 (inline_generate_summary): Likewise.
1679 (inline_analyze_function): Likewise.
1680 (do_estimate_growth): Likewise.
1681 (simple_edge_hints): Likewise.
1682 (estimate_node_size_and_time): Likewise.
1683 (estimate_edge_devirt_benefit): Likewise.
1684 (compute_inline_parameters): Likewise.
1685 (estimate_function_body_sizes): Likewise.
1686 (compute_bb_predicates): Likewise.
1687 (initialize_inline_failed): Likewise.
1688 (dump_inline_summary): Likewise.
1689 (dump_inline_edge_summary): Likewise.
1690 * ipa-inline-transform.c (inline_transform): Likewise.
1691 (preserve_function_body_p): Likewise.
1692 (save_inline_function_body): Likewise.
1693 (inline_call): Likewise.
1694 (clone_inlined_nodes): Likewise.
1695 (can_remove_node_now_p): Likewise.
1696 (can_remove_node_now_p_1): Likewise.
1697 * ipa-inline.c (early_inliner): Likewise.
1698 (early_inline_small_functions): Likewise.
1699 (inline_always_inline_functions): Likewise.
1700 (ipa_inline): Likewise.
1701 (flatten_function): Likewise.
1702 (inline_small_functions): Likewise.
1703 (speculation_useful_p): Likewise.
1704 (recursive_inlining): Likewise.
1705 (update_caller_keys): Likewise.
1706 (reset_edge_caches): Likewise.
1707 (update_edge_key): Likewise.
1708 (edge_badness): Likewise.
1709 (relative_time_benefit): Likewise.
1710 (want_inline_self_recursive_call_p): Likewise.
1711 (want_inline_small_function_p): Likewise.
1712 (want_early_inline_function_p): Likewise.
1713 (num_calls): Likewise.
1714 (can_early_inline_edge_p): Likewise.
1715 (can_inline_edge_p): Likewise.
1716 (report_inline_failed_reason): Likewise.
1717 * ipa-profile.c (ipa_profile): Likewise.
1718 (ipa_propagate_frequency): Likewise.
1719 (ipa_propagate_frequency_1): Likewise.
1720 (ipa_profile_generate_summary): Likewise.
1721 * ipa-prop.c (ipcp_transform_function): Likewise.
1722 (read_replacements_section): Likewise.
1723 (ipa_prop_read_section): Likewise.
1724 (ipa_modify_call_arguments): Likewise.
1725 (ipa_print_node_params): Likewise.
1726 (propagate_controlled_uses): Likewise.
1727 (update_indirect_edges_after_inlining): Likewise.
1728 (remove_described_reference): Likewise.
1729 (ipa_make_edge_direct_to_target): Likewise.
1730 (ipa_analyze_node): Likewise.
1731 (ipa_analyze_params_uses): Likewise.
1732 (ipa_compute_jump_functions): Likewise.
1733 (ipa_get_callee_param_type): Likewise.
1734 (ipa_print_node_jump_functions): Likewise.
1735 (ipa_initialize_node_params): Likewise.
1736 (ipa_populate_param_decls): Likewise.
1737 (ipa_func_spec_opts_forbid_analysis_p): Likewise.
1738 (write_agg_replacement_chain): Likewise.
1739 (ipa_write_node_info): Likewise.
1740 (ipa_edge_duplication_hook): Likewise.
1741 (try_decrement_rdesc_refcount): Likewise.
1742 * ipa-pure-const.c (propagate_nothrow): Likewise.
1743 (propagate_pure_const): Likewise.
1744 (pure_const_read_summary): Likewise.
1745 (pure_const_write_summary): Likewise.
1746 (analyze_function): Likewise.
1747 * ipa-ref-inline.h (ipa_ref_referred_ref_list): Likewise.
1748 (ipa_ref_referring_ref_list): Likewise.
1749 * ipa-ref.c (ipa_clear_stmts_in_references): Likewise.
1750 (ipa_remove_stmt_references): Likewise.
1751 (ipa_find_reference): Likewise.
1752 (ipa_dump_referring): Likewise.
1753 (ipa_dump_references): Likewise.
1754 (ipa_record_reference): Likewise.
1755 * ipa-reference.c (ipa_reference_read_optimization_summary): Likewise.
1756 (ipa_reference_write_optimization_summary): Likewise.
1757 (write_node_summary_p): Likewise.
1758 (propagate): Likewise.
1759 (read_write_all_from_decl): Likewise.
1760 (generate_summary): Likewise.
1761 (analyze_function): Likewise.
1762 (propagate_bits): Likewise.
1763 (ipa_reference_get_not_written_global): Likewise.
1764 (ipa_reference_get_not_read_global): Likewise.
1765 * ipa-split.c (execute_split_functions): Likewise.
1766 (split_function): Likewise.
1767 * ipa-utils.c (ipa_merge_profiles): Likewise.
1768 (dump_cgraph_node_set): Likewise.
1769 (ipa_reverse_postorder): Likewise.
1770 (ipa_edge_within_scc): Likewise.
1771 (ipa_get_nodes_in_cycle): Likewise.
1772 (ipa_free_postorder_info): Likewise.
1773 (ipa_reduced_postorder): Likewise.
1774 (searchc): Likewise.
1775 (recursive_call_p): Likewise.
1776 * ipa.c (ipa_cdtor_merge): Likewise.
1777 (record_cdtor_fn): Likewise.
1778 (function_and_variable_visibility): Likewise.
1779 (varpool_externally_visible_p): Likewise.
1780 (cgraph_externally_visible_p): Likewise.
1781 (comdat_can_be_unshared_p): Likewise.
1782 (comdat_can_be_unshared_p_1): Likewise.
1783 (address_taken_from_non_vtable_p): Likewise.
1784 (ipa_discover_readonly_nonaddressable_vars): Likewise.
1785 (symtab_remove_unreachable_nodes): Likewise.
1786 (walk_polymorphic_call_targets): Likewise.
1787 (process_references): Likewise.
1788 (enqueue_node): Likewise.
1789 (has_addr_references_p): Likewise.
1790 (cgraph_non_local_node_p_1): Likewise.
1791 * is-a.h (varpool_analyze_node): Likewise.
1792 * lto-cgraph.c (input_symtab): Likewise.
1793 (merge_profile_summaries): Likewise.
1794 (input_cgraph_1): Likewise.
1795 (input_edge): Likewise.
1796 (input_varpool_node): Likewise.
1797 (input_node): Likewise.
1798 (input_overwrite_node): Likewise.
1799 (compute_ltrans_boundary): Likewise.
1800 (output_refs): Likewise.
1801 (lto_output_varpool_node): Likewise.
1802 (lto_output_node): Likewise.
1803 (reachable_from_other_partition_p): Likewise.
1804 (referenced_from_other_partition_p): Likewise.
1805 (lto_output_edge): Likewise.
1806 (output_node_opt_summary): Likewise.
1807 (add_node_to): Likewise.
1808 (reachable_from_this_partition_p): Likewise.
1809 (lto_set_symtab_encoder_in_partition): Likewise.
1810 (lto_symtab_encoder_in_partition_p): Likewise.
1811 (lto_set_symtab_encoder_encode_initializer): Likewise.
1812 (lto_symtab_encoder_encode_initializer_p): Likewise.
1813 (lto_set_symtab_encoder_encode_body): Likewise.
1814 (lto_symtab_encoder_encode_body_p): Likewise.
1815 * lto-section-in.c (lto_free_function_in_decl_state_for_node):
1817 * lto-streamer-in.c (lto_read_body): Likewise.
1818 (fixup_call_stmt_edges): Likewise.
1819 (fixup_call_stmt_edges_1): Likewise.
1820 * lto-streamer-out.c (produce_symtab): Likewise.
1821 (output_symbol_p): Likewise.
1822 (write_symbol): Likewise.
1823 (lto_output): Likewise.
1824 (copy_function): Likewise.
1825 (output_function): Likewise.
1826 * passes.c (function_called_by_processed_nodes_p): Likewise.
1827 (ipa_write_optimization_summaries): Likewise.
1828 (ipa_write_summaries): Likewise.
1829 (do_per_function_toporder): Likewise.
1830 (do_per_function): Likewise.
1831 (dump_passes): Likewise.
1832 * symtab.c (symtab_semantically_equivalent_p): Likewise.
1833 (symtab_nonoverwritable_alias): Likewise.
1834 (symtab_nonoverwritable_alias_1): Likewise.
1835 (symtab_for_node_and_aliases): Likewise.
1836 (symtab_resolve_alias): Likewise.
1837 (fixup_same_cpp_alias_visibility): Likewise.
1838 (symtab_alias_ultimate_target): Likewise.
1839 (symtab_used_from_object_file_p): Likewise.
1840 (verify_symtab_base): Likewise.
1841 (dump_symtab_base): Likewise.
1842 (symtab_node_name): Likewise.
1843 (symtab_node_asm_name): Likewise.
1844 (symtab_dissolve_same_comdat_group_list): Likewise.
1845 (symtab_add_to_same_comdat_group): Likewise.
1846 (symtab_unregister_node): Likewise.
1847 (symtab_insert_node_to_hashtable): Likewise.
1848 (symtab_register_node): Likewise.
1849 (unlink_from_assembler_name_hash): Likewise.
1850 (insert_to_assembler_name_hash): Likewise.
1851 (eq_assembler_name): Likewise.
1852 (hash_node_by_assembler_name): Likewise.
1853 (eq_node): Likewise.
1854 (hash_node): Likewise.
1855 * toplev.c (wrapup_global_declaration_2): Likewise.
1856 * trans-mem.c (ipa_tm_execute): Likewise.
1857 (ipa_tm_transform_clone): Likewise.
1858 (ipa_tm_transform_transaction): Likewise.
1859 (ipa_tm_transform_calls_redirect): Likewise.
1860 (ipa_tm_insert_gettmclone_call): Likewise.
1861 (ipa_tm_insert_irr_call): Likewise.
1862 (ipa_tm_create_version): Likewise.
1863 (ipa_tm_create_version_alias): Likewise.
1864 (ipa_tm_mark_forced_by_abi_node): Likewise.
1865 (ipa_tm_mark_force_output_node): Likewise.
1866 (ipa_tm_diagnose_tm_safe): Likewise.
1867 (ipa_tm_mayenterirr_function): Likewise.
1868 (ipa_tm_scan_irr_function): Likewise.
1869 (ipa_tm_note_irrevocable): Likewise.
1870 (ipa_tm_scan_calls_clone): Likewise.
1871 (get_cg_data): Likewise.
1872 * tree-eh.c (tree_could_trap_p): Likewise.
1873 * tree-emutls.c (ipa_lower_emutls): Likewise.
1874 (create_emultls_var): Likewise.
1875 (lower_emutls_function_body): Likewise.
1876 (gen_emutls_addr): Likewise.
1877 (emutls_decl): Likewise.
1878 (new_emutls_decl): Likewise.
1879 * tree-inline.c (tree_function_versioning): Likewise.
1880 (optimize_inline_calls): Likewise.
1881 (expand_call_inline): Likewise.
1882 (estimate_num_insns): Likewise.
1883 (copy_bb): Likewise.
1884 (delete_unreachable_blocks_update_callgraph): Likewise.
1885 * tree-nested.c (gimplify_all_functions): Likewise.
1886 (create_nesting_tree): Likewise.
1887 (check_for_nested_with_variably_modified): Likewise.
1888 * tree-pretty-print.c (dump_function_header): Likewise.
1889 * tree-profile.c (tree_profiling): Likewise.
1890 * tree-sra.c (ipa_sra_preliminary_function_checks): Likewise.
1891 (modify_function): Likewise.
1892 (convert_callers): Likewise.
1893 (convert_callers_for_node): Likewise.
1894 * tree-ssa-structalias.c (ipa_pta_execute): Likewise.
1895 (associate_varinfo_to_alias): Likewise.
1896 (create_variable_info_for): Likewise.
1897 (get_constraint_for_ssa_var): Likewise.
1898 * tree-vectorizer.c (increase_alignment): Likewise.
1899 * tree.c (find_decls_types_in_var): Likewise.
1900 (find_decls_types_in_node): Likewise.
1901 (free_lang_data_in_decl): Likewise.
1902 * value-prof.c (gimple_ic_transform): Likewise.
1903 (gimple_ic): Likewise.
1904 (check_ic_target): Likewise.
1905 (init_node_map): Likewise.
1906 * varasm.c (decl_binds_to_current_def_p): Likewise.
1907 (default_binds_local_p_1): Likewise.
1908 (dump_tm_clone_pairs): Likewise.
1909 (assemble_alias): Likewise.
1910 (find_decl): Likewise.
1911 (mark_decl_referenced): Likewise.
1912 * varpool.c (varpool_for_node_and_aliases): Likewise.
1913 (varpool_extra_name_alias): Likewise.
1914 (varpool_create_variable_alias): Likewise.
1915 (add_new_static_var): Likewise.
1916 (varpool_finalize_named_section_flags): Likewise.
1917 (varpool_remove_unreferenced_decls): Likewise.
1918 (enqueue_node): Likewise.
1919 (varpool_assemble_decl): Likewise.
1920 (assemble_aliases): Likewise.
1921 (varpool_analyze_node): Likewise.
1922 (cgraph_variable_initializer_availability): Likewise.
1923 (varpool_add_new_variable): Likewise.
1924 (ctor_for_folding): Likewise.
1925 (dump_varpool_node): Likewise.
1926 (varpool_remove_initializer): Likewise.
1927 (varpool_remove_node): Likewise.
1928 (varpool_node_for_decl): Likewise.
1929 (varpool_create_empty_node): Likewise.
1930 * config/i386/i386.c (ix86_generate_version_dispatcher_body): Likewise.
1931 (ix86_get_function_versions_dispatcher): Likewise.
1933 2013-10-29 David Malcolm <dmalcolm@redhat.com>
1935 * cgraph.h (symtab_node_base): Convert to a class;
1936 add GTY((desc ("%h.type"), tag ("SYMTAB_SYMBOL"))), and take
1937 chain_next/prev from symtab_node_def.
1938 (cgraph_node): Inherit from symtab_node; add GTY option
1939 tag ("SYMTAB_FUNCTION").
1940 (varpool_node): Inherit from symtab_node; add GTY option
1941 tag ("SYMTAB_VARIABLE").
1942 (symtab_node_def): Remove.
1943 (is_a_helper <cgraph_node>::test (symtab_node_def *)): Convert to...
1944 (is_a_helper <cgraph_node>::test (symtab_node_base *)): ...this.
1945 (is_a_helper <varpool_node>::test (symtab_node_def *)): Convert to...
1946 (is_a_helper <varpool_node>::test (symtab_node_base *)): ...this.
1948 * ipa-ref.h (symtab_node_def): Drop.
1949 (symtab_node): Change underlying type from symtab_node_def to
1951 (const_symtab_node): Likwise.
1953 * is-a.h: Update examples in comment.
1955 * symtab.c (symtab_hash): Change symtab_node_def to symtab_node_base.
1956 (assembler_name_hash): Likewise.
1958 2013-10-29 Martin Liska <marxin.liska@gmail.com>
1960 * doc/tree-ssa.texi (gimple_phi_result): Document.
1961 (gimple_phi_num_args, gimple_phi_arg): Likewise.
1962 (gimple_phi_arg_edge, gimple_phi_arg_def): Likewise.
1963 (PHI_RESULT, PHI_NUM_ARGS): Remove.
1964 (PHI_ARG_ELT, PHI_ARG_EDGE, PHI_ARG_DEF): Likewise.
1966 2013-10-29 Andrew MacLeod <amacleod@redhat.com>
1968 * expr.h: Revert change and include tree-core.h.
1969 * rtl.h: Revert change and don't include tree-core.h.
1971 2013-10-29 Andrew MacLeod <amacleod@redhat.com>
1973 * config/darwin.c: Include gimple.h.
1974 * config/i386/winnt.c: Likewise.
1976 2013-10-29 Marc Glisse <marc.glisse@inria.fr>
1978 PR tree-optimization/19831
1979 * tree-ssa-alias.c (stmt_kills_ref_p_1): Handle BUILT_IN_FREE.
1981 2013-10-29 Andrew MacLeod <amacleod@redhat.com>
1983 * tree-outof-ssa.h: Remove include files.
1984 * tree-outof-ssa.c: Add required include files from tree-outof-ssa.h.
1986 * tree-ssa-coalesce.c: Likewise.
1987 * cfgexpand.c: Likewise.
1988 * tree-ssa-ter.c: Likewise.
1989 * ipa-prop.h: Remove gimple.h and tree-core.h from include list.
1990 * lto-streamer.h: Likewise.
1991 * cgraphbuild.c: Add gimple.h to include list.
1992 * data-streamer-in.c: Likewise.
1993 * ipa-cp.c: Likewise.
1994 * tree-streamer.c: Likewise.
1995 * lto-compress.c: Likewise.
1996 * ipa-reference.c: Likewise.
1997 * data-streamer-out.c: Likewise.
1998 * lto-cgraph.c: Likewise.
1999 * cgraphclones.c: Likewise.
2000 * ipa-utils.c: Likewise.
2001 * data-streamer.c: Likewise.
2002 * ipa-split.c: Likewise.
2003 * lto-section-in.c: Likewise.
2004 * tree-streamer-out.c: Likewise.
2005 * ipa-prop.c: Likewise.
2006 * tree-streamer-in.c: Likewise.
2007 * symtab.c: Likewise.
2008 * opts-global.c: Likewise.
2009 * lto-opts.c: Likewise.
2010 * lto-section-out.c: Likewise.
2011 * lto-streamer.c: Likewise.
2012 * rtl.h: Add tree-core.h to include list.
2013 * expr.h: Remove tree-core.h from include list.
2014 * gimple.h: Likewise.
2015 * ipa-utils.h: Likewise.
2016 * streamer-hooks.h: Likewise.
2017 * streamer-hooks.c: Include input.h.
2019 2013-10-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2021 * config/arm/arm.c (cortexa7_extra_costs): New table.
2022 (arm_cortex_a7_tune): New.
2023 * config/arm/arm-cores.def: Use cortex_a7 tuning for cortex-a7.
2025 2013-10-29 Eric Botcazou <ebotcazou@adacore.com>
2027 * expr.c (expand_expr_real_1) <MEM_EXPR>: Eliminate small redundancy.
2029 2013-10-29 David Malcolm <dmalcolm@redhat.com>
2031 * doc/gty.texi ("Inheritance and GTY"): Make it clear that
2032 to use autogenerated markers for a class-hierarchy, every class
2033 must have a GTY marker.
2034 * gengtype.h (struct type): Add linked list of subclasses to
2035 the "s" member of the union.
2036 (add_subclass): New decl.
2037 * gengtype-state.c (read_state_struct_type): Set up subclass
2039 * gengtype.c (get_ultimate_base_class): New.
2040 (add_subclass): New.
2041 (new_structure): Set up subclass linked list.
2042 (set_gc_used_type): Propagate usage information to subclasses.
2043 (output_mangled_typename): Use get_ultimate_base_class.
2044 (walk_subclasses): Use the subclass linked list, avoiding an
2045 O(N^2) when writing out all types.
2046 (walk_type): Issue an error if the base class is missing a tag,
2047 rather than generating bogus C code. Add a gcc_unreachable
2048 default case, in case people omit tags from concrete subclasses,
2049 or get the values wrong.
2050 (write_func_for_structure): Issue an error for subclasses for
2051 which the base doesn't have a "desc", since otherwise the
2052 autogenerated routines for the base would silently fail to visit
2053 any subclass fields.
2054 (write_root): Use get_ultimate_base_class, tweaking constness of
2055 tp to match that function's signature.
2057 2013-10-29 David Malcolm <dmalcolm@redhat.com>
2059 * doc/gty.texi (GTY Options): Add note about inheritance to
2060 description of desc and tag.
2061 (Inheritance and GTY): New.
2063 2013-10-29 David Malcolm <dmalcolm@redhat.com>
2065 * gengtype-parse.c (opts_have): Drop "static" so that
2066 we can use this from gengtype.c.
2067 * gengtype.c (set_gc_used_type): Mark any base class as used;
2068 update field traversal to visit inherited fields.
2069 (output_mangled_typename): Convert references to classes within
2070 an inheritance hierarchy to reference the ultimate base class,
2071 since only it will have gt_ functions.
2072 (get_string_option): New.
2073 (walk_subclasses): New.
2074 (walk_type): Treat GTY structs that have a "desc" as being the
2075 root of an inheritance hierarchy. Generate a switch on it
2076 within the marking function which walks all subclasses, adding
2077 cases for them via walk_subclasses. For subclasses, visit all
2078 fields of the type (including inherited ones).
2079 (write_func_for_structure): Don't write fns for subclasses, only
2080 for the ultimate base class within an inheritance hierarchy.
2081 Subclasses-marking will be handled by the base class marking functions.
2082 (write_types): Likewise.
2083 (write_local_func_for_structure): Likewise.
2084 (USED_BY_TYPED_GC_P): Emit allocators for subclasses that have
2085 a "tag" option (and are thus concrete subclasses).
2086 (write_root): Use the marker function for the ultimate base class.
2087 * gengtype.h (FOR_ALL_INHERITED_FIELDS): New.
2088 (opts_have): Add declaration.
2090 2013-10-28 Vladimir Makarov <vmakarov@redhat.com>
2092 * lra-spills.c (lra_final_code_change): Remove useless move insns
2093 originated from moves of pseudos.
2095 2013-10-28 Jeff Law <law@redhat.com>
2097 * lower-subreg.c (resolve_simple_move): Fix comment typo.
2099 2013-10-28 Trevor Saunders <tsaunders@mozilla.com>
2101 * df-scan.c (df_collection_rec): Adjust.
2102 (copy_defs): New constant.
2103 (copy_uses): Likewise.
2104 (copy_eq_uses): Likewise.
2105 (copy_mw): Likewise.
2106 (copy_all): Likewise.
2107 (df_insn_rescan): Adjust.
2108 (df_notes_rescan): Likewise.
2109 (df_swap_refs): Likewise.
2110 (df_sort_and_compress_refs): Likewise.
2111 (df_sort_and_compress_mws): Likewise.
2112 (df_install_refs): Likewise.
2113 (df_install_mws): Likewise.
2114 (df_refs_add_to_chains): Add flags parameter controlling which vectors
2116 (df_bb_refs_record): Adjust.
2117 (df_record_entry_block_defs): Likewise.
2118 (df_record_exit_block_defs): Likewise.
2119 (df_refs_verify): Likewise.
2120 (df_mws_verify): Likewise.
2121 (df_insn_refs_verify): Likewise.
2122 (df_bb_verify): Likewise.
2123 * ipa-pure-const.c (finish_state): Remove.
2124 (propagate): Adjust.
2125 * tree-data-ref.c tree-ssa-alias.c tree-ssa-loop-ivcanon.c
2126 tree-ssa-threadedge.c tree-vect-loop-manip.c tree-vect-slp.c
2127 var-tracking.c: Adjust.
2128 * vec.c (stack_vecs): Remove.
2129 (register_stack_vec): Likewise.
2130 (stack_vec_register_index): Likewise.
2131 (unregister_stack_vec): Likewise.
2132 * vec.h (struct va_stack): Remove.
2133 (struct vec<T, A, vl_ptr>): Specialize as
2134 struct vec<T, va_heap, vl_ptr> instead since va_heap is the only
2135 allocation strategy compatable with the vl_ptr layout.
2136 (struct vec<T, va_gc, vl_ptr>): Remove because it now gets an empty
2137 specialization anyway.
2138 (class stack_vec): New class.
2139 (vec_stack_alloc): Remove.
2140 (vec<T, va_heap, vl_ptr>::using_auto_storage): New method.
2142 2013-10-28 Alexander Ivchenko <alexander.ivchenko@intel.com>
2143 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2144 Sergey Lega <sergey.s.lega@intel.com>
2145 Anna Tikhonova <anna.tikhonova@intel.com>
2146 Ilya Tocar <ilya.tocar@intel.com>
2147 Andrey Turetskiy <andrey.turetskiy@intel.com>
2148 Ilya Verbin <ilya.verbin@intel.com>
2149 Kirill Yukhin <kirill.yukhin@intel.com>
2150 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2152 * config/i386/i386.md (prefetch): Allow TARGET_AVX512PF.
2153 (*prefetch_avx512pf_<mode>): New.
2154 * config/i386/sse.md (avx512f_vmcmp<mode>3): Ditto.
2155 (avx512f_maskcmp<mode>3): Ditto.
2156 (vashrv16si3): Ditto.
2158 2013-10-28 Alexander Ivchenko <alexander.ivchenko@intel.com>
2159 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2160 Sergey Lega <sergey.s.lega@intel.com>
2161 Anna Tikhonova <anna.tikhonova@intel.com>
2162 Ilya Tocar <ilya.tocar@intel.com>
2163 Andrey Turetskiy <andrey.turetskiy@intel.com>
2164 Ilya Verbin <ilya.verbin@intel.com>
2165 Kirill Yukhin <kirill.yukhin@intel.com>
2166 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2168 * config/i386/i386.md (any_truncate): New.
2169 (trunsuffix): Ditto.
2170 * config/i386/predicates.md (const_8_to_9_operand): New.
2171 (const_10_to_11_operand): Ditto.
2172 (const_12_to_13_operand): Ditto.
2173 (const_14_to_15_operand): Ditto.
2174 (const_16_to_19_operand): Ditto.
2175 (const_20_to_23_operand): Ditto.
2176 (const_24_to_27_operand): Ditto.
2177 (const_28_to_31_operand): Ditto.
2178 * config/i386/sse.md (unspec): Add UNSPEC_UNSIGNED_FIX_NOTRUNC.
2179 (cvtusi2<ssescalarmodesuffix>32): New.
2180 (cvtusi2<ssescalarmodesuffix>64): Ditto.
2181 (ufloatv16siv16sf2): Ditto.
2182 (avx512f_fix_notruncv16sfv16si): Ditto.
2183 (avx512f_ufix_notruncv16sfv16si): Ditto.
2184 (avx512f_vcvtss2usi): Ditto.
2185 (avx512f_vcvtss2usiq): Ditto.
2186 (avx512f_vcvttss2usi): Ditto.
2187 (avx512f_vcvttss2usiq): Ditto.
2188 (avx512f_vcvtsd2usi): Ditto.
2189 (avx512f_vcvtsd2usiq): Ditto.
2190 (avx512f_vcvttsd2usi): Ditto.
2191 (avx512f_vcvttsd2usiq): Ditto.
2192 (ufloatv8siv8df): Ditto.
2193 (avx512f_cvtdq2pd512_2): Ditto.
2194 (avx512f_cvtpd2dq512): Ditto.
2195 (avx512f_ufix_notruncv8dfv8si): Ditto.
2196 (avx512f_cvtpd2ps512): Ditto.
2197 (vec_unpacks_lo_v16sf): Ditto.
2198 (vec_unpacks_hi_v16sf): Ditto.
2199 (vec_unpacks_float_hi_v16si): Ditto.
2200 (vec_unpacks_float_lo_v16si): Ditto.
2201 (avx512f_unpckhps512): Ditto.
2202 (avx512f_unpcklps512): Ditto.
2203 (avx512f_movshdup512): Ditto.
2204 (avx512f_movsldup512): Ditto.
2205 (vec_extract_lo_v32hi): Ditto.
2206 (vec_extract_hi_v32hi): Ditto.
2207 (vec_extract_lo_v64qi): Ditto.
2208 (vec_extract_hi_v64qi): Ditto.
2209 (avx512f_unpckhpd512): Ditto.
2210 (avx512f_movddup512): Ditto.
2211 (avx512f_unpcklpd512): Ditto.
2212 (*avx512f_unpcklpd512): Ditto.
2213 (avx512f_shufps512_1): Ditto.
2214 (avx512f_shufpd512_1): Ditto.
2215 (avx512f_interleave_highv8di): Ditto.
2216 (avx512f_interleave_lowv8di): Ditto.
2217 (PMOV_DST_MODE): Ditto.
2218 (pmov_src_mode): Ditto.
2219 (pmov_src_lower): Ditto.
2221 (*avx512f_<code><pmov_src_lower><mode>2): Ditto.
2222 (*avx512f_<code>v8div16qi2): Ditto.
2223 (*avx512f_<code>v8div16qi2_store): Ditto.
2224 (vec_widen_umult_even_v16si): Ditto.
2225 (*vec_widen_umult_even_v16si): Ditto.
2226 (vec_widen_smult_even_v16si): Ditto.
2227 (*vec_widen_smult_even_v16si): Ditto.
2228 (avx512f_interleave_highv16si): Ditto.
2229 (avx512f_interleave_lowv16si): Ditto.
2230 (avx512f_<code>v16qiv16si2): Ditto.
2231 (avx512f_<code>v16hiv16si2): Ditto.
2232 (avx512f_<code>v8qiv8di2): Ditto.
2233 (avx512f_<code>v8hiv8di2): Ditto.
2234 (avx512f_<code>v8siv8di2): Ditto.
2235 (avx512cd_maskb_vec_dupv8di): Ditto.
2236 (avx512cd_maskw_vec_dupv16si): Ditto.
2237 (avx512f_vcvtph2ps512): Ditto.
2238 (avx512f_vcvtps2ph512): Ditto.
2239 (VEC_EXTRACT_MODE): Extened with wider modes.
2240 (VEC_PERM_AVX2): Ditto.
2241 (VEC_PERM_CONST): Ditto.
2243 2013-10-28 Joern Rennecke <joern.rennecke@embecosm.com>
2245 * config/arc/arc.c (arc_ccfsm_post_advance):
2246 Add comment about TYPE_RETURN.
2248 2013-10-28 Bin Cheng <bin.cheng@arm.com>
2250 * tree-ssa-loop-ivopts.c (strip_offset_1): Change parameter type.
2251 Count DECL_FIELD_BIT_OFFSET for COMPONENT_REF.
2252 (strip_offset): Convert offset to unsigned number.
2254 2013-10-27 Tom de Vries <tom@codesourcery.com>
2256 * cfgexpand.c (gimple_expand_cfg): Remove test for parm_birth_insn.
2257 Don't commit insertions after NOTE_INSN_FUNCTION_BEG.
2259 2013-10-27 Oleg Endo <olegendo@gcc.gnu.org>
2261 * config/sh/sh.c (MSW, LSW): Move and rename macros to...
2262 * config/sh/sh.h (SH_REG_MSW_OFFSET, SH_REG_LSW_OFFSET): ... here.
2263 (TARGET_BIG_ENDIAN): New macro.
2264 * config/sh/sh.md: Use it instead of !TARGET_LITTLE_ENDIAN.
2265 Use SH_REG_MSW_OFFSET and SH_REG_LSW_OFFSET.
2266 * config/sh/sh.c: Likewise.
2267 * config/sh/sh.h: Likewise.
2269 2013-10-27 Hans-Peter Nilsson <hp@axis.com>
2271 * config/cris/cris.c (cris_emit_trap_for_misalignment): Replace the
2272 removed PRED_MUDFLAP with PRED_NORETURN. Correct file-path in comment.
2274 2013-10-26 Oleg Endo <olegendo@gcc.gnu.org>
2276 * config/sh/sh.md (movmemsi): Remove empty constraints and predicates.
2278 (cmpstr_t, cmpstrsi): Fix formatting.
2280 2013-10-26 Oleg Endo <olegendo@gcc.gnu.org>
2283 * config/sh/predicates.md (general_movdst_operand): Allow reg+reg
2284 addressing, do not use general_operand for memory operands.
2286 2013-10-26 Vladimir Makarov <vmakarov@redhat.com>
2289 2013-10-25 Vladimir Makarov <vmakarov@redhat.com>
2290 * lra-spills.c (lra_final_code_change): Remove useless move insns.
2292 2013-10-26 Jeff Law <law@redhat.com>
2294 * predict.c (PRED_MUDFLAP): Remove.
2295 * targhooks.c (build_va_arg_indirect_ref): Remove mudflap support.
2297 * Makefile.in (C_COMMON_OBJS): Remove tree-mudflap.
2298 (OBJS): Remove tree-nomudflap.o
2299 (GTFILES): Remove tree-mudflap.c
2300 * builtins.c (expand_builtin_alloc): Remove mudflap support.
2301 * gcc.c (MFWRAP_SPEC, MFLIB_SPEC): Likewise.
2302 (mfwrap_spec, mflib_spec): Likewise.
2303 (cpp_unique_options, cc1_options, static_specs): Likewise.
2304 * gimplify (gimplify_vla_decl, build_va_arg_indirect_ref): Likewise.
2305 * passes.def: Likewise.
2306 * toplev.c (compile_file, process_options): Likewise.
2307 * tree-inline.c (copy_tree_r): Likewise.
2308 * tree-pass.,h (make_pass_mudflap_1, make_pass_mudflap_2): Likewise.
2309 * varasm.c (make_decl_rtl, make_decl_rtl_for_debug): Likewise.
2310 (build_constant_desc, output_constant_def_contents): Likewise.
2311 (categorize_decl_for_section): Likewise.
2312 * tree-mudflap.c: Removed.
2313 * tree-mudflap.h: Removed.
2314 * tree-nomudflap.c: Removed.
2315 * bfin/uclinux.h (MFWRAP_SPEC): Remove.
2316 * moxie/uclinux.h (MFWRAP_SPEC): Likewise.
2317 * rs6000/aix.h (MFWRAP_SPEC, MFLIB_SPEC): Likewise.
2318 * config/sol2.h (MFLIB_SPEC): Likewise.
2319 * doc/install.texi: Remove mudflap references.
2320 * doc/passes.texi: Similarly.
2321 * doc/sourcebuild.texi: Similarly.
2322 * doc/invoke.texi: Remove mudlfap related options.
2324 2013-10-25 Vladimir Makarov <vmakarov@redhat.com>
2326 PR rtl-optimization/58759
2327 * lra-constraints.c (lra_constraints): Remove wrong condition to
2328 remove insn setting up an equivalent pseudo.
2330 2013-10-25 Vladimir Makarov <vmakarov@redhat.com>
2332 * config/rs6000/rs6000-protos.h
2333 (rs6000_secondary_memory_needed_mode): New prototype.
2334 * config/rs6000/rs6000.c: Include ira.h.
2335 (TARGET_LRA_P): Redefine.
2336 (rs6000_legitimate_offset_address_p): Call
2337 legitimate_constant_pool_address_p in strict mode for LRA.
2338 (rs6000_legitimate_address_p): Ditto.
2339 (legitimate_lo_sum_address_p): Add code for LRA. Use lra_in_progress.
2340 (rs6000_emit_move): Add LRA version of code to generate load/store
2342 (rs6000_secondary_memory_needed_mode): New.
2343 (rs6000_alloc_sdmode_stack_slot): Do nothing for LRA.
2344 (rs6000_secondary_reload_class): Return NO_REGS for LRA for
2345 constants, memory, and FP registers.
2346 (rs6000_lra_p): New.
2347 * config/rs6000/rs6000.h (SECONDARY_MEMORY_NEEDED_MODE): New macro.
2348 * config/rs6000/rs6000.opt (mlra): New option.
2349 * lra-spills.c (lra_final_code_change): Remove useless move insns.
2351 2013-10-25 Yufeng Zhang <yufeng.zhang@arm.com>
2353 * tree-ssa-math-opts.c (convert_plusminus_to_widen): Call
2354 has_single_use () and not do the conversion if has_single_use ()
2355 returns false for the multiplication result.
2357 2013-10-25 David Malcolm <dmalcolm@redhat.com>
2359 * tree.h (EXCEPTIONAL_CLASS_P): Rename parameter from "CODE"
2360 to "NODE", since this works on a "tree", not an
2362 (CONSTANT_CLASS_P): Likewise.
2365 (INDIRECT_REF_P): Likewise.
2366 (REFERENCE_CLASS_P): Likewise.
2367 (COMPARISON_CLASS_P): Likewise.
2368 (UNARY_CLASS_P): Likewise.
2369 (BINARY_CLASS_P): Likewise.
2370 (STATEMENT_CLASS_P): Likewise.
2371 (VL_EXP_CLASS_P): Likewise.
2372 (EXPRESSION_CLASS_P): Likewise.
2373 (IS_TYPE_OR_DECL_P): Likewise.
2375 2013-10-25 Marc Glisse <marc.glisse@inria.fr>
2377 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Look for an
2378 ADDR_EXPR in the defining statement.
2380 2013-10-25 Richard Biener <rguenther@suse.de>
2382 PR tree-optimization/58626
2383 * tree-loop-distribution.c (enum rdg_dep_type): Remove
2384 anti_dd, output_dd and input_dd.
2385 (struct rdg_edge): Remove level and relation members.
2386 (RDGE_LEVEL, RDGE_RELATION): Remove.
2387 (dot_rdg_1): Adjust.
2388 (create_rdg_edge_for_ddr): Remove.
2389 (create_rdg_edges_for_scalar): Adjust.
2390 (create_edge_for_control_dependence): Likewise.
2391 (create_rdg_edges): Split into ...
2392 (create_rdg_flow_edges): ... this
2393 (create_rdg_cd_edges): ... and this.
2395 (build_rdg): Likewise, do not compute data dependences or
2397 (pg_add_dependence_edges): New function.
2399 (distribute_loop): First apply all non-dependence based
2400 partition mergings. Then compute dependences between partitions
2401 and merge and order partitions according to them.
2403 2013-10-25 Eric Botcazou <ebotcazou@adacore.com>
2405 PR rtl-optimization/58831
2406 * alias.c (init_alias_analysis): At the beginning of each iteration,
2407 set the reg_seen[N] bit if static_reg_base_value[N] is non-null.
2409 2013-10-25 Eric Botcazou <ebotcazou@adacore.com>
2411 * recog.c (search_ofs): New static variable moved from...
2412 (peep2_find_free_register): ...here.
2413 (peephole2_optimize): Initialize it.
2415 2013-10-25 Tobias Burnus <burnus@net-b.de>
2417 * doc/invoke.texi (fopenmp): Change supported OpenMP version to 4.0.
2419 2013-10-25 Uros Bizjak <ubizjak@gmail.com>
2421 * config/i386/i386.h (TARGET_MPX): New define.
2422 (TARGET_MPX_P): Ditto.
2424 2013-10-24 Ilya Enkovich <ilya.enkovich@intel.com>
2426 * config/i386/constraints.md (B): New.
2429 * config/i386/i386-c.c (ix86_target_macros_internal): Add __MPX__.
2430 * config/i386/i386-modes.def (BND32): New.
2432 * config/i386/i386-protos.h (ix86_bnd_prefixed_insn_p): New.
2433 * config/i386/i386.c (isa_opts): Add mmpx.
2434 (regclass_map): Add bound registers.
2435 (dbx_register_map): Likewise.
2436 (dbx64_register_map): Likewise.
2437 (svr4_dbx_register_map): Likewise.
2439 (ix86_option_override_internal): Support MPX ISA.
2440 (ix86_conditional_register_usage): Support bound registers.
2441 (print_reg): Likewise.
2442 (ix86_code_end): Add MPX bnd prefix.
2443 (output_set_got): Likewise.
2444 (ix86_output_call_insn): Likewise.
2445 (ix86_print_operand): Add '!' (MPX bnd) print prefix support.
2446 (ix86_print_operand_punct_valid_p): Likewise.
2447 (ix86_print_operand_address): Support UNSPEC_BNDMK_ADDR and
2449 (ix86_class_likely_spilled_p): Add bound regs support.
2450 (ix86_hard_regno_mode_ok): Likewise.
2451 (x86_order_regs_for_local_alloc): Likewise.
2452 (ix86_bnd_prefixed_insn_p): New.
2453 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Fix to new value.
2454 (FIXED_REGISTERS): Add bound registers.
2455 (CALL_USED_REGISTERS): Likewise.
2456 (REG_ALLOC_ORDER): Likewise.
2457 (HARD_REGNO_NREGS): Likewise.
2459 (VALID_BND_REG_MODE): New.
2460 (FIRST_BND_REG): New.
2461 (LAST_BND_REG): New.
2462 (reg_class): Add BND_REGS.
2463 (REG_CLASS_NAMES): Likewise.
2464 (REG_CLASS_CONTENTS): Likewise.
2466 (ANY_BND_REG_P): New.
2468 (HI_REGISTER_NAMES): Add bound registers.
2469 * config/i386/i386.md (UNSPEC_BNDMK): New.
2470 (UNSPEC_BNDMK_ADDR): New.
2471 (UNSPEC_BNDSTX): New.
2472 (UNSPEC_BNDLDX): New.
2473 (UNSPEC_BNDLDX_ADDR): New.
2474 (UNSPEC_BNDCL): New.
2475 (UNSPEC_BNDCU): New.
2476 (UNSPEC_BNDCN): New.
2477 (UNSPEC_MPX_FENCE): New.
2480 (type): Add mpxmov, mpxmk, mpxchk, mpxld, mpxst.
2481 (length_immediate): Likewise.
2482 (prefix_0f): Likewise.
2484 (prefix_rep): Check for bnd prefix.
2485 (length_nobnd): New.
2486 (length): Use length_nobnd if specified.
2491 (*jcc_1): Add bnd prefix and rename length attr to length_nobnd.
2494 (simple_return_internal): Likewise.
2495 (simple_return_pop_internal): Likewise.
2496 (*indirect_jump): Add MPX bnd prefix.
2497 (*tablejump_1): Likewise.
2498 (simple_return_internal_long): Likewise.
2499 (simple_return_indirect_internal): Likewise.
2503 (*mov<mode>_internal_mpx): New.
2504 (<mode>_<bndcheck>): New.
2505 (*<mode>_<bndcheck>): New.
2510 * config/i386/predicates.md (lea_address_operand): Rename to...
2511 (address_no_seg_operand): ... this.
2512 (address_mpx_no_base_operand): New.
2513 (address_mpx_no_index_operand): New.
2514 (bnd_mem_operator): New.
2515 * config/i386/i386.opt (mmpx): New.
2516 * doc/invoke.texi: Add documentation for the flags -mmpx, -mno-mpx.
2517 * doc/rtl.texi Add documentation for BND32mode and BND64mode.
2519 2013-10-24 Ilya Enkovich <ilya.enkovich@intel.com>
2521 * mode-classes.def (MODE_POINTER_BOUNDS): New.
2522 * tree.def (POINTER_BOUNDS_TYPE): New.
2523 * genmodes.c (complete_mode): Support MODE_POINTER_BOUNDS.
2524 (POINTER_BOUNDS_MODE): New.
2525 (make_pointer_bounds_mode): New.
2526 * machmode.h (POINTER_BOUNDS_MODE_P): New.
2527 * stor-layout.c (int_mode_for_mode): Support MODE_POINTER_BOUNDS.
2528 (layout_type): Support POINTER_BOUNDS_TYPE.
2529 * tree-pretty-print.c (dump_generic_node): Support POINTER_BOUNDS_TYPE.
2530 * tree.c (build_int_cst_wide): Support POINTER_BOUNDS_TYPE.
2531 (type_contains_placeholder_1): Likewise.
2532 * tree.h (POINTER_BOUNDS_TYPE_P): New.
2533 * varasm.c (output_constant): Support POINTER_BOUNDS_TYPE.
2534 * doc/rtl.texi (MODE_POINTER_BOUNDS): New.
2536 2013-10-24 Igor Shevlyakov <igor.shevlyakov@gmail.com>
2538 * expr.c (expand_expr_real_1): Use mode of memory reference rather than
2539 mode of address computation when calling memory_address_addr_space.
2541 2013-08-24 Richard Henderson <rth@twiddle.net>
2544 * optabs.c (maybe_emit_atomic_exchange): Use create_input_operand
2545 instead of create_convert_operand_to.
2546 (maybe_emit_sync_lock_test_and_set): Likewise.
2547 (expand_atomic_compare_and_swap): Likewise.
2548 (maybe_emit_compare_and_swap_exchange_loop): Don't convert_modes.
2550 2013-08-24 Sriraman Tallam <tmsriram@google.com>
2552 * cgraph.c (cgraph_fnver_htab): Move GTY((...)) to be before htab_t.
2553 Change param_is to use the struct and not the pointer to the struct.
2555 2013-10-24 Andrew MacLeod <amacleod@redhat.com>
2557 * builtins.c (dummy_object, gimplify_va_arg_expr): Move to gimplify.c.
2558 * gimplify.c (build_va_arg_indirect_ref, std_gimplify_va_arg_expr):
2559 Move to targhooks.c.
2560 (dummy_object, gimplify_va_arg_expr): Relocate from builtins.c.
2561 * targhooks.c (build_va_arg_indirect_ref, std_gimplify_va_arg_expr):
2562 Relocate from gimplify.c.
2563 * targhooks.h: Add 2 prototypes.
2564 * tree.h: Delete 2 prototypes.
2566 2013-10-24 Igor Shevlyakov <igor.shevlyakov@gmail.com>
2568 * tree-ssa-loop-ivopts.c (multiplier_allowed_in_address_p ): Check both
2569 [reg+mult*reg] and [mult*reg] to determine if multiplier is allowed.
2571 2013-10-24 Cong Hou <congh@google.com>
2573 * convert.c (convert_to_real): Guard those unsafe math function
2574 convertions with flag_unsafe_math_optimizations. Handle sqrt()
2577 2013-10-24 Markus Trippelsdorf <markus@trippelsdorf.de>
2580 * cgraph.c (cgraph_create_edge_1): Add indirect_unknown_callee
2582 (cgraph_create_edge): Use the new argument.
2583 (cgraph_create_indirect_edge): Likewise.
2585 2013-10-24 Joern Rennecke <joern.rennecke@embecosm.com>
2587 * config/arc/arc.c (arc_ccfsm_post_advance): Also handle
2589 (arc_ifcvt) <case 1 and 2>: Check that arc_ccfsm_post_advance
2590 changes statep->state.
2592 2013-10-24 Tobias Burnus <burnus@net-b.de>
2595 * tree-cfg.c (replace_loop_annotate): New function.
2596 (execute_build_cfg): Call it.
2597 * gimplify.c (gimple_boolify, gimplify_expr): Handle ANNOTATE_EXPR.
2598 * internal-fn.c (expand_ANNOTATE): New function.
2599 * internal-fn.def (ANNOTATE): Define as new internal function.
2600 * tree-core.h (tree_node_kind): Add annot_expr_ivdep_kind.
2601 * tree-pretty-print.c (dump_generic_node): Handle ANNOTATE_EXPR.
2602 * tree.def (ANNOTATE_EXPR): New DEFTREECODE.
2603 * doc/extend.texi (Pragmas): Document #pragma ivdep.
2604 * doc/generic.texi (Expressions): Document ANNOTATE_EXPR.
2606 2013-10-17 Ian Bolton <ian.bolton@arm.com>
2607 Marcus Shawcroft <marcus.shawcroft@arm.com>
2609 * config/aarch64/aarch64.c (aarch64_preferred_reload_class):
2610 Special case reload SP+C into none GENERAL_REGS.
2612 2013-10-24 Michael Matz <matz@suse.de>
2614 * gengtype.c (is_file_equal): Check that files will be same length.
2616 2013-10-25 Christian Bruel <christian.bruel@st.com>
2618 * config.gcc (sh-*): Add sh-mem.o to extra_obj.
2619 * config/sh/t-sh (sh-mem.o): New rule.
2620 * config/sh/sh-mem.cc (expand_block_move): Moved here.
2621 (sh_expand_cmpstr): New function.
2622 * config/sh/sh.c (force_into, expand_block_move): Move to sh-mem.c.
2623 * config/sh/sh-protos.h (sh_expand_cmpstr): Declare.
2624 * config/sh/sh.md (cmpstrsi, cmpstr_t): New patterns.
2625 (rotlhi3_8): Rename.
2627 2013-10-24 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2629 * configure.ac (ZW_PROG_COMPILER_DEPENDENCIES): Use CXX instead of CC.
2630 * Makefile.in (CXXDEPMODE): Assign and change users.
2631 (CCDEPMODE): Delete.
2632 * configure: Regenerate.
2634 2013-10-23 David Malcolm <dmalcolm@redhat.com>
2636 * gengtype-parse.c (require_without_advance): New.
2637 (type): For GTY-marked types that are not GTY((user)), parse any
2638 base classes, requiring them to be single-inheritance, and not
2639 be templates. For non-GTY-marked types and GTY((user)),
2640 continue to skip over any C++ inheritance specification.
2641 * gengtype-state.c (state_writer::write_state_struct_type):
2642 Write base class of type (if any).
2643 (read_state_struct_type): Read base class of type (if any).
2644 * gengtype.c (new_structure): Add a "base_class" parameter.
2645 (create_optional_field_): Update for new parameter to new_structure.
2646 (adjust_field_rtx_def): Likewise.
2647 (adjust_field_tree_exp): Likewise.
2648 * gengtype.h (struct type): Add "base_class" field to the s
2650 (new_structure): Add "base" parameter.
2652 2013-10-23 Sriraman Tallam <tmsriram@google.com>
2655 * config/i386/i386.c (ix86_option_override_internal):
2656 Change TARGET_SSE2 to TARGET_SSE2_P (opts->...)
2657 (ix86_valid_target_attribute_tree):
2658 Change TARGET_64BIT to TARGET_64BIT_P (opts->...)
2659 Change TARGET_SSE to TARGET_SSE_P (opts->...)
2661 2013-10-23 Andrew MacLeod <amacleod@redhat.com>
2663 * tree-ssa-loop.h: Remove include files.
2664 * gengtype.c (open_base_files): Adjust include list for gtype-desc.c.
2665 * cfgloopmanip.c: Move required includes from tree-ssa-loop.h.
2666 * graphite-clast-to-gimple.c: Likewise.
2667 * graphite-scop-detection.c: Likewise.
2668 * graphite-sese-to-poly.c: Likewise.
2669 * ipa-inline-analysis.c: Likewise.
2670 * ipa-pure-const.c: Likewise.
2671 * loop-init.c: Likewise.
2672 * passes.c: Likewise.
2673 * predict.c: Likewise.
2674 * tree-cfg.c: Likewise.
2675 * tree-cfgcleanup.c: Likewise.
2676 * tree-chrec.c: Likewise.
2677 * tree-data-ref.c: Likewise.
2678 * tree-loop-distribution.c: Likewise.
2679 * tree-parloops.c: Likewise.
2680 * tree-predcom.c: Likewise.
2681 * tree-scalar-evolution.c: Likewise.
2682 * tree-ssa-address.c: Likewise.
2683 * tree-ssa.c: Likewise.
2684 * tree-ssa-dce.c: Likewise.
2685 * tree-ssa-loop.c: Likewise.
2686 * tree-ssa-loop-im.c: Likewise.
2687 * tree-ssa-loop-ivcanon.c: Likewise.
2688 * tree-ssa-loop-ivopts.c: Likewise.
2689 * tree-ssa-loop-manip.c: Likewise.
2690 * tree-ssa-loop-niter.c: Likewise.
2691 * tree-ssa-loop-prefetch.c: Likewise.
2692 * tree-ssa-loop-unswitch.c: Likewise.
2693 * tree-ssa-reassoc.c: Likewise.
2694 * tree-vect-data-refs.c: Likewise.
2695 * tree-vect-loop.c: Likewise.
2696 * tree-vect-loop-manip.c: Likewise.
2697 * tree-vectorizer.c: Likewise.
2698 * tree-vect-stmts.c: Likewise.
2699 * tree-vrp.c: Likewise.
2701 2013-10-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2703 * config/rs6000/altivec.md (mulv8hi3): Adjust for little endian.
2705 2013-10-23 Jakub Jelinek <jakub@redhat.com>
2707 PR tree-optimization/58775
2708 PR tree-optimization/58791
2709 * tree-ssa-reassoc.c (reassoc_stmt_dominates_stmt_p): New function.
2710 (insert_stmt_after): Rewritten, don't move the stmt, but really
2712 (get_stmt_uid_with_default): Remove.
2713 (build_and_add_sum): Use insert_stmt_after and
2714 reassoc_stmt_dominates_stmt_p. Fix up uid if bb contains only labels.
2715 (update_range_test): Set uid on stmts added by
2716 force_gimple_operand_gsi. Don't immediately modify statements
2717 in inter-bb optimization, just update oe->op values.
2718 (optimize_range_tests): Return bool whether any changed have been made.
2719 (update_ops): New function.
2720 (struct inter_bb_range_test_entry): New type.
2721 (maybe_optimize_range_tests): Perform statement changes here.
2722 (not_dominated_by, appears_later_in_bb, get_def_stmt,
2723 ensure_ops_are_available): Remove.
2724 (find_insert_point): Rewritten.
2725 (rewrite_expr_tree): Remove MOVED argument, add CHANGED argument,
2726 return LHS of the (new resp. old) stmt. Don't call
2727 ensure_ops_are_available, don't reuse SSA_NAMEs, recurse first
2728 instead of last, move new stmt at the right place.
2729 (linearize_expr, repropagate_negates): Don't reuse SSA_NAMEs.
2730 (negate_value): Likewise. Set uids.
2731 (break_up_subtract_bb): Initialize uids.
2732 (reassociate_bb): Adjust rewrite_expr_tree caller.
2733 (do_reassoc): Don't call renumber_gimple_stmt_uids.
2735 2013-10-23 David Edelsohn <dje.gcc@gmail.com>
2738 * config/rs6000/rs6000.md (mulsi3_internal1 and splitter): Add
2739 TARGET_32BIT final condition.
2740 (mulsi3_internal2 and splitter): Same.
2742 2013-10-23 Jeff Law <law@redhat.com>
2744 * tree-ssa-threadedge.c (thread_across_edge): Do not allow threading
2745 through joiner blocks with abnormal outgoing edges.
2747 * tree-ssa-threadupdate.c (thread_block_1): Renamed from thread_block.
2748 Add parameter JOINERS, to allow/disallow threading through joiner
2750 (thread_block): New. Call thread_block_1.
2751 (mark_threaded_blocks): Remove code to filter out certain cases
2752 of threading through joiner blocks.
2753 (thread_through_all_blocks): Document how we can have a dangling
2754 edge AUX field and clear it.
2756 2013-10-23 Ian Lance Taylor <iant@google.com>
2758 * doc/invoke.texi (Option Summary): Remove -fno-default-inline.
2759 (C++ Dialect Options): Likewise.
2760 (Optimize Options): Likewise.
2762 2013-10-23 Tom de Vries <tom@codesourcery.com>
2764 PR tree-optimization/58805
2765 * tree-ssa-tail-merge.c (stmt_local_def): Add gimple_vdef check.
2767 2013-10-23 Jakub Jelinek <jakub@redhat.com>
2769 * tree-vect-patterns.c (vect_recog_divmod_pattern): Optimize
2770 sequence based on get_range_info returned range.
2772 2013-10-23 Andrew MacLeod <amacleod@redhat.com>
2774 * tree-ssa.h: Remove all #include's
2775 * gengtype.c (open_base_files): Adjust include list for gtype-desc.c.
2776 * alias.c: Move required includes from tree-ssa.h.
2778 * builtins.c: Likewise.
2779 * calls.c: Likewise.
2780 * cfgexpand.c: Likewise.
2781 * cfghooks.c: Likewise.
2782 * cfgloop.c: Likewise.
2783 * cfgloopmanip.c: Likewise.
2784 * cgraph.c: Likewise.
2785 * cgraphbuild.c: Likewise.
2786 * cgraphclones.c: Likewise.
2787 * cgraphunit.c: Likewise.
2789 * except.c: Likewise.
2791 * final.c: Likewise.
2792 * fold-const.c: Likewise.
2793 * ggc-page.c: Likewise.
2794 * gimple-builder.c: Likewise.
2795 * gimple-fold.c: Likewise.
2796 * gimple-iterator.c: Likewise.
2797 * gimple-low.c: Likewise.
2798 * gimple-pretty-print.c: Likewise.
2799 * gimple-ssa-strength-reduction.c: Likewise.
2800 * gimple-streamer-in.c: Likewise.
2801 * gimple-streamer-out.c: Likewise.
2802 * gimplify.c: Likewise.
2803 * graphite-blocking.c: Likewise.
2804 * graphite-clast-to-gimple.c: Likewise.
2805 * graphite-dependences.c: Likewise.
2806 * graphite-interchange.c: Likewise.
2807 * graphite-optimize-isl.c: Likewise.
2808 * graphite-poly.c: Likewise.
2809 * graphite-scop-detection.c: Likewise.
2810 * graphite-sese-to-poly.c: Likewise.
2811 * graphite.c: Likewise.
2812 * ipa-cp.c: Likewise.
2813 * ipa-inline-analysis.c: Likewise.
2814 * ipa-inline-transform.c: Likewise.
2815 * ipa-inline.c: Likewise.
2816 * ipa-prop.c: Likewise.
2817 * ipa-pure-const.c: Likewise.
2818 * ipa-reference.c: Likewise.
2819 * ipa-split.c: Likewise.
2820 * ipa-utils.c: Likewise.
2821 * loop-init.c: Likewise.
2822 * lto-cgraph.c: Likewise.
2823 * lto-section-in.c: Likewise.
2824 * lto-section-out.c: Likewise.
2825 * lto-streamer-in.c: Likewise.
2826 * lto-streamer-out.c: Likewise.
2827 * lto-streamer.c: Likewise.
2828 * omp-low.c: Likewise.
2829 * passes.c: Likewise.
2830 * predict.c: Likewise.
2831 * print-tree.c: Likewise.
2832 * profile.c: Likewise.
2834 * targhooks.c: Likewise.
2835 * tracer.c: Likewise.
2836 * trans-mem.c: Likewise.
2837 * tree-call-cdce.c: Likewise.
2838 * tree-cfg.c: Likewise.
2839 * tree-cfgcleanup.c: Likewise.
2840 * tree-chrec.c: Likewise.
2841 * tree-complex.c: Likewise.
2842 * tree-data-ref.c: Likewise.
2843 * tree-dfa.c: Likewise.
2844 * tree-eh.c: Likewise.
2845 * tree-emutls.c: Likewise.
2846 * tree-if-conv.c: Likewise.
2847 * tree-inline.c: Likewise.
2848 * tree-into-ssa.c: Likewise.
2849 * tree-loop-distribution.c: Likewise.
2850 * tree-mudflap.c: Likewise.
2851 * tree-nested.c: Likewise.
2852 * tree-nrv.c: Likewise.
2853 * tree-object-size.c: Likewise.
2854 * tree-outof-ssa.c: Likewise.
2855 * tree-parloops.c: Likewise.
2856 * tree-phinodes.c: Likewise.
2857 * tree-predcom.c: Likewise.
2858 * tree-pretty-print.c: Likewise.
2859 * tree-profile.c: Likewise.
2860 * tree-scalar-evolution.c: Likewise.
2861 * tree-sra.c: Likewise.
2862 * tree-ssa-address.c: Likewise.
2863 * tree-ssa-alias.c: Likewise.
2864 * tree-ssa-ccp.c: Likewise.
2865 * tree-ssa-coalesce.c: Likewise.
2866 * tree-ssa-copy.c: Likewise.
2867 * tree-ssa-copyrename.c: Likewise.
2868 * tree-ssa-dce.c: Likewise.
2869 * tree-ssa-dom.c: Likewise.
2870 * tree-ssa-dse.c: Likewise.
2871 * tree-ssa-forwprop.c: Likewise.
2872 * tree-ssa-ifcombine.c: Likewise.
2873 * tree-ssa-live.c: Likewise.
2874 * tree-ssa-loop-ch.c: Likewise.
2875 * tree-ssa-loop-im.c: Likewise.
2876 * tree-ssa-loop-ivcanon.c: Likewise.
2877 * tree-ssa-loop-ivopts.c: Likewise.
2878 * tree-ssa-loop-manip.c: Likewise.
2879 * tree-ssa-loop-niter.c: Likewise.
2880 * tree-ssa-loop-prefetch.c: Likewise.
2881 * tree-ssa-loop-unswitch.c: Likewise.
2882 * tree-ssa-loop.c: Likewise.
2883 * tree-ssa-math-opts.c: Likewise.
2884 * tree-ssa-operands.c: Likewise.
2885 * tree-ssa-phiopt.c: Likewise.
2886 * tree-ssa-phiprop.c: Likewise.
2887 * tree-ssa-pre.c: Likewise.
2888 * tree-ssa-propagate.c: Likewise.
2889 * tree-ssa-reassoc.c: Likewise.
2890 * tree-ssa-sccvn.c: Likewise.
2891 * tree-ssa-sink.c: Likewise.
2892 * tree-ssa-strlen.c: Likewise.
2893 * tree-ssa-structalias.c: Likewise.
2894 * tree-ssa-tail-merge.c: Likewise.
2895 * tree-ssa-ter.c: Likewise.
2896 * tree-ssa-threadedge.c: Likewise.
2897 * tree-ssa-threadupdate.c: Likewise.
2898 * tree-ssa-uncprop.c: Likewise.
2899 * tree-ssa-uninit.c: Likewise.
2900 * tree-ssa.c: Likewise.
2901 * tree-ssanames.c: Likewise.
2902 * tree-stdarg.c: Likewise.
2903 * tree-streamer-in.c: Likewise.
2904 * tree-switch-conversion.c: Likewise.
2905 * tree-tailcall.c: Likewise.
2906 * tree-vect-data-refs.c: Likewise.
2907 * tree-vect-generic.c: Likewise.
2908 * tree-vect-loop-manip.c: Likewise.
2909 * tree-vect-loop.c: Likewise.
2910 * tree-vect-patterns.c: Likewise.
2911 * tree-vect-slp.c: Likewise.
2912 * tree-vect-stmts.c: Likewise.
2913 * tree-vectorizer.c: Likewise.
2914 * tree-vrp.c: Likewise.
2917 * value-prof.c: Likewise.
2918 * var-tracking.c: Likewise.
2919 * varpool.c: Likewise.
2920 * vtable-verify.c: Likewise.
2922 2013-10-23 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2924 * config/tilegx/tilegx.c: Include "tree.h".
2926 2013-10-23 Jakub Jelinek <jakub@redhat.com>
2928 * gimple-pretty-print.c (dump_ssaname_info): Always print "# " before
2929 the info, not after it.
2930 (gump_gimple_phi): Add COMMENT argument, if true, print "# " after
2931 dump_ssaname_info call.
2932 (pp_gimple_stmt_1): Adjust caller.
2933 (dump_phi_nodes): Likewise. Don't print "# " here.
2935 2013-10-22 Jan Hubicka <jh@suse.cz>
2937 * i386.h (TARGET_MISALIGNED_MOVE_STRING_PROLOGUES_EPILOGUES): New
2939 * x86-tune.def (TARGET_MISALIGNED_MOVE_STRING_PROLOGUES): Define it.
2940 * i386.c (expand_small_movmem_or_setmem): New function.
2941 (expand_set_or_movmem_prologue_epilogue_by_misaligned_moves): New
2943 (alg_usable_p): Add support for value ranges; cleanup.
2944 (ix86_expand_set_or_movmem): Add support for misaligned moves.
2946 2013-10-22 Sterling Augustine <saugustine@google.com>
2948 * doc/invoke.texi: Document -ggnu-pubnames.
2949 * common.opt: Add new option -ggnu-pubnames and modify -gpubnames
2951 * dwarf2out.c: Include gdb/gdb-index.h.
2952 (DEBUG_PUBNAMES_SECTION, DEBUG_PUBTYPES_SECTION): Handle
2953 debug_generate_pub_sections.
2954 (is_java, output_pubtables, output_pubname): New functions.
2955 (include_pubname_in_output): Handle debug_generate_pub_sections at
2957 (size_of_pubnames): Use new local space_for_flags based on
2958 debug_generate_pub_sections.
2959 (output_pubnames): Unify pubnames and pubtypes output logic.
2960 Genericize comments. Call output_pubname.
2961 (dwarf2out_finish): Move logic to output_pubnames and call it.
2963 2013-10-22 Uros Bizjak <ubizjak@gmail.com>
2966 * config/i386/i386.c (put_condition_code) <case GTU, case LEU>:
2967 Remove CCCmode handling.
2968 <case LTU>: Return 'c' suffix for CCCmode.
2969 <case GEU>: Return 'nc' suffix for CCCmode.
2970 (ix86_cc_mode) <case GTU, case LEU>: Do not generate overflow checks.
2971 * config/i386/i386.md (*sub<mode>3_cconly_overflow): Remove.
2972 (*sub<mode>3_cc_overflow): Ditto.
2973 (*subsi3_zext_cc_overflow): Ditto.
2975 2013-10-22 Steve Ellcey <sellcey@mips.com>
2977 * config/mips/mips.c (mips_rtx_costs): Fix cost estimate for nor
2978 (AND (NOT OP1) (NOT OP2)).
2980 2013-10-22 Bill Schmidt <wschmidt@vnet.ibm.com>
2982 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
2983 meaning of merge-high and merge-low masks for little endian; avoid
2984 use of vector-pack masks for little endian for mismatched modes.
2986 2013-10-22 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2988 * config/tilepro/tilepro.c: Include "tree.h".
2990 2013-10-22 Andreas Schwab <schwab@suse.de>
2992 * config/m68k/m68k.c (notice_update_cc): Handle register conflict
2995 2013-10-22 Paolo Carlini <paolo.carlini@oracle.com>
2997 * doc/contrib.texi ([Fran@,{c}ois Dumont], [Tim Shen],
2998 [Ed Smith-Rowland]): New entries.
2999 ([Stephen M. Webb]): Update.
3001 2013-10-22 Andrew MacLeod <amacleod@redhat.com>
3003 * tree-ssa-ter.h: Remove duplicate copy of file contents.
3005 2013-10-21 Marek Polacek <polacek@redhat.com>
3008 * fold-const.c (fold_range_test): Return 0 if the type is not
3011 2013-10-21 Richard Sandiford <rdsandiford@googlemail.com>
3013 * system.h: Move hwint.h include further down.
3014 * hwint.h (sext_hwi, zext_hwi): Define unconditionally. Add
3015 gcc_checking_asserts.
3016 * hwint.c (sext_hwi, zext_hwi): Delete ENABLE_CHECKING versions.
3018 2013-10-21 Bernd Edlinger <bernd.edlinger@hotmail.de>
3020 Fix volatile issues in optimize_bit_field_compare.
3021 * fold-const.c (optimize_bit_field_compare): Bail out if
3022 lvolatilep or rvolatilep.
3024 2013-10-21 Bernd Edlinger <bernd.edlinger@hotmail.de>
3026 Fix DECL_BIT_FIELD depencency on flag_strict_volatile_bitfields
3027 and get_inner_reference returning different pmode for non-volatile
3028 bit-field members dependent on flag_strict_volatile_bitfields.
3029 * stor-layout.c (layout_decl): Remove special handling of
3030 flag_strict_volatile_bitfields.
3031 * expr.c (get_inner_reference): Don't use DECL_BIT_FIELD
3032 if flag_strict_volatile_bitfields > 0 and TREE_THIS_VOLATILE.
3034 2013-10-21 Paulo Matos <pmatos@broadcom.com>
3036 * ipa-inline.c (edge_badness): Cap edge->count at max_count for badness
3039 2013-10-21 Jeff Law <law@redhat.com>
3041 * tree-ssa-threadedge.c (thread_through_normal_block): New
3042 argument VISITED. Remove VISISTED as a local variable. When we
3043 have a threadable jump, verify the destination of the jump has not
3045 (thread_across_edge): Allocate VISITED bitmap once at function
3046 scope and use it throughout. Make sure to set appropriate bits in
3047 VISITED for E (start of jump thread path).
3048 * tree-ssa-threadupdate.c (mark_threaded_blocks): Reject threading
3049 through a joiner if any edge on the path has a recorded jump thread.
3051 2013-10-21 Ian Lance Taylor <iant@google.com>
3053 * doc/invoke.texi (Optimize Options): For -fno-toplevel-reorder,
3054 don't imply that attributes can solve all problems.
3055 (Directory Options): Fix typo.
3057 2013-10-21 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3059 * config/arm/arm.c (cortexa9_extra_costs): Update mult costs for
3060 extend and extend_add.
3062 2013-10-21 Richard Biener <rguenther@suse.de>
3064 PR tree-optimization/58794
3065 * fold-const.c (operand_equal_p): Compare FIELD_DECL operand
3066 of COMPONENT_REFs with OEP_CONSTANT_ADDRESS_OF left in place.
3068 2013-10-21 Richard Biener <rguenther@suse.de>
3071 * fold-const.c (fold_binary_loc): Fold ((T) (X /[ex] C)) * C
3072 to (T) X for sign-changing conversions (or no conversion).
3074 2013-10-20 Uros Bizjak <ubizjak@gmail.com>
3076 * config/i386/i386.md (kxnor<mode>): Add FLAGS_REG clobber.
3078 2013-10-20 Jan Hubicka <jh@suse.cz>
3080 * config/i386/i386-tune.def: Add comment; organize into categories
3082 2013-10-21 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
3084 * config/i386/i386.c (expand_set_or_movmem_via_loop): Add issetmem
3085 argument. Update function comment.
3086 (expand_set_or_movmem_via_rep): New function combining
3087 expand_movmem_via_rep_mov and expand_setmem_via_rep_stos.
3088 (expand_movmem_via_rep_mov): Remove.
3089 expand_setmem_via_rep_stos): Remove.
3090 (expand_movmem_epilogue): Update calls correspondingly.
3091 (expand_setmem_epilogue_via_loop): Likewise.
3093 (expand_setmem_epilogue): Add VEC_VALUE argument, refactor.
3094 (expand_set_or_movmem_prologue): New function combining
3095 expand_movmem_prologue and expand_setmem_prologue.
3096 (expand_movmem_prologue): Remove.
3097 (expand_setmem_prologue): Remove.
3098 (expand_set_or_movmem_constant_prologue): New function combining
3099 expand_constant_movmem_prologue and expand_constant_setmem_prologue.
3100 (expand_constant_movmem_prologue): Remove.
3101 (expand_constant_setmem_prologue): Remove.
3102 (promote_duplicated_reg): Allow vector-const0 value.
3103 (ix86_expand_set_or_movmem): New function combining ix86_expand_movmem
3104 and ix86_expand_setmem.
3105 (ix86_expand_movmem): Call ix86_expand_set_or_movmem.
3106 (ix86_expand_setmem): Call ix86_expand_set_or_movmem.
3108 2013-10-21 Diego Novillo <dnovillo@google.com>
3110 * asan.c: Include tree.h
3111 * bb-reorder.c: Likewise.
3112 * cfgcleanup.c: Likewise.
3113 * cfgloopmanip.c: Likewise.
3114 * data-streamer-in.c: Likewise.
3115 * data-streamer-out.c: Likewise.
3116 * data-streamer.c: Likewise.
3117 * dwarf2cfi.c: Likewise.
3118 * graphite-blocking.c: Likewise.
3119 * graphite-clast-to-gimple.c: Likewise.
3120 * graphite-dependences.c: Likewise.
3121 * graphite-interchange.c: Likewise.
3122 * graphite-optimize-isl.c: Likewise.
3123 * graphite-poly.c: Likewise.
3124 * graphite-scop-detection.c: Likewise.
3125 * graphite-sese-to-poly.c: Likewise.
3126 * graphite.c: Likewise.
3127 * ipa-devirt.c: Likewise.
3128 * ipa-profile.c: Likewise.
3131 * loop-init.c: Likewise.
3132 * loop-unroll.c: Likewise.
3133 * lower-subreg.c: Likewise.
3134 * lto/lto-object.c: Likewise.
3135 * recog.c: Likewise.
3136 * reginfo.c: Likewise.
3137 * tree-loop-distribution.c: Likewise.
3138 * tree-parloops.c: Likewise.
3139 * tree-ssa-strlen.c: Likewise.
3140 * tree-streamer.c: Likewise.
3141 * value-prof.c: Likewise.
3142 * target-globals.c: Likewise.
3143 * expr.h: Include tree-core.h instead of tree.h.
3144 * gimple.h: Likewise.
3145 * ipa-prop.h: Likewise.
3146 * ipa-utils.h: Likewise.
3147 * lto-streamer.h: Likewise.
3148 * streamer-hooks.h: Likewise.
3149 * ipa-reference.h: Include cgraph.h instead of tree.h.
3150 * cgraph.h: Include basic-block.h instead of tree.h.
3151 * tree-streamer.h: Do not include tree.h.
3152 * genattrtab.c (write_header): Generate inclusion of
3154 * genautomata.c (main): Likewise.
3155 * genemit.c: Likewise.
3156 * genopinit.c: Likewise.
3157 * genoutput.c (output_prologue): Likewise.
3158 * genpeep.c: Likewise.
3160 2013-10-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3162 * config/rs6000/altivec.md (vec_unpacku_hi_v16qi): Adjust for
3164 (vec_unpacku_hi_v8hi): Likewise.
3165 (vec_unpacku_lo_v16qi): Likewise.
3166 (vec_unpacku_lo_v8hi): Likewise.
3168 2013-10-20 Jan Hubicka <jh@suse.cz>
3170 * config/i386/x86-tune.def (X86_TUNE_SLOW_IMUL_IMM32_MEM,
3171 X86_TUNE_SLOW_IMUL_IMM8): Keep enabled only for K8 and AMDFAM10.
3172 (X86_TUNE_USE_VECTOR_FP_CONVERTS): Disable for generic.
3174 2013-10-20 Richard Sandiford <rdsandiford@googlemail.com>
3176 * config/mips/mips.h (ISA_HAS_WSBH): Define.
3177 * config/mips/mips.md (UNSPEC_WSBH, UNSPEC_DSBH, UNSPEC_DSHD): New
3179 (bswaphi2, bswapsi2, bswapdi2, wsbh, dsbh, dshd): New patterns.
3181 2013-10-19 John David Anglin <danglin@gcc.gnu.org>
3184 * system.h: Undef m_slot.
3186 2013-10-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3188 * config/rs6000/rs6000.c (vspltis_constant): Make sure we check
3189 all elements for both endian flavors.
3191 2013-10-19 Uros Bizjak <ubizjak@gmail.com>
3194 * config/i386/i386.c (ix86_function_value_regno): Add DX_REG,
3195 ST1_REG and XMM1_REG for 32bit and 64bit targets. Also add DI_REG
3196 and SI_REG for 64bit SYSV ABI targets.
3198 2013-10-19 Uros Bizjak <ubizjak@gmail.com>
3200 * mode-switching.c (create_pre_exit): Rename maybe_builtin_apply
3201 to multi_reg_return. Clarify that we are skipping USEs of multiple
3202 return registers. Use bool type where appropriate.
3204 2013-10-18 Jan Hubicka <jh@suse.cz>
3206 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Disable accumulation
3208 * x86-tune.def (X86_TUNE_USE_LEAVE): Update comment.
3209 (X86_TUNE_PUSH_MEMORY): Likewise.
3210 (X86_TUNE_AVX256_UNALIGNED_LOAD_OPTIMAL,
3211 X86_TUNE_AVX256_UNALIGNED_STORE_OPTIMAL): New.
3212 (X86_TUNE_ACCUMULATE_OUTGOING_ARGS, X86_TUNE_ALWAYS_FANCY_MATH_387):
3214 * i386.c (x86_accumulate_outgoing_args, x86_arch_always_fancy_math_387,
3215 x86_avx256_split_unaligned_load, x86_avx256_split_unaligned_store):
3217 (ix86_option_override_internal): Update to use tune features instead
3220 2013-10-18 Cong Hou <congh@google.com>
3222 PR tree-optimization/58508
3223 * tree-vect-loop-manip.c (vect_loop_versioning): Hoist loop invariant
3224 statement that contains data refs with zero-step.
3226 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
3228 * tree-ssa.h: Don't include gimple-low.h, tree-ssa-address.h,
3229 sbitmap.h, tree-ssa-threadedge.h, tree-ssa-dom.h and tree-cfgcleanup.h.
3230 * gimple-low.c (gimple_check_call_arg,
3231 gimple_check_call_matching_types): Move to cgraph.c.
3232 * gimple-low.h: Remove prototype.
3233 * cgraph.c: (gimple_check_call_arg, gimple_check_call_matching_types):
3234 Relocate from gimple-low.c.
3235 * cgraph.h: Add prototype. Don't include basic-block.h.
3236 * gimplify.c: Add gimple-low to include list.
3237 * omp-low.c: Add gimple-low and tree-cfgcleanup.h to include list.
3238 * tree-eh.c: Add gimple-low to include list.
3239 * tree-nested.c: Likewise.
3240 * cfgexpand.c: Add tree-ssa-address.h to include list.
3242 * gimple-fold.c: Likewise.
3243 * gimple-ssa-strength-reduction.c: Likewise.
3244 * trans-mem.c: Likewise.
3245 * tree-mudflap.c: Likewise.
3246 * tree-ssa-loop-ivopts.c: Likewise.
3247 * tree-ssa-dom.c: Include tree-ssa-threadedge.h and tree-ssa-dom.h.
3248 (degenerate_phi_result): Move to tree-phinodes.c.
3249 * tree-ssa-loop-ch.c: Include tree-ssa-threadedge.h.
3250 * tree-ssa-threadedge.c: Likewise.
3251 * tree-vrp.c: Likewise.
3252 * tree-phinodes.c (degenerate_phi_result): Relocate here.
3253 * tree-ssa-dom.h (degenerate_phi_result): Remove Prototype.
3254 * tree-phinodes.h (degenerate_phi_result): Add prototype.
3255 * tree-ssa-copy.c: Include tree-ssa-dom.h.
3256 * tree-ssa-forwprop.c: Likewise.
3257 * tree-cfgcleanup.c (execute_cleanup_cfg_post_optimizing,
3258 pass_data_cleanup_cfg_post_optimizing,
3259 make_pass_cleanup_cfg_post_optimizing): Relocate from tree-optimize.c.
3260 * tree-optimize.c: Delete File.
3261 * graphite.c: Include tree-cfgcleanup.h.
3262 * passes.c: Likewise.
3263 * tree-cfg.c: Likewise.
3264 * tree-profile.c: Likewise.
3265 * tree-ssa-dse.c: Likewise.
3266 * tree-ssa-loop-ivcanon.c: Likewise.
3267 * tree-switch-conversion.c: Don't include tree-ssa-operands.h.
3268 * tree-outof-ssa.c: Include sbitmap.h.
3269 * tree-ssa-live.c: Likewise.
3270 * tree-ssa-propagate.c: Likewise.
3271 * tree-ssa-structalias.c: Likewise.
3272 * tree-stdarg.c: Likewise.
3273 * Makefile.in (OBJS): Delete tree-optimize.o.
3274 * basic-block.h (gcov_type, gcov_type_unsigned): Move to coretypes.h.
3275 * coretypes.h (gcov_type, gcov_type_unsigned): Relocate here.
3276 * varasm.c: Include basic-block.h.
3277 * cfgloop.h: Include function.h instead of basic-block.h
3278 (bb_loop_depth): Move to cfgloop.c.
3279 * cfgloop.c (bb_loop_depth): Relocate from cfgloop.h.
3281 2013-10-18 Teresa Johnson <tejohnson@google.com>
3283 * predict.c (probably_never_executed): Compare frequency-based
3284 count to number of training runs.
3285 * params.def (UNLIKELY_BB_COUNT_FRACTION): New parameter.
3287 2013-10-18 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3289 * config/arm/arm.c (cortexa9_extra_costs): New table.
3290 (arm_cortex_a9_tune): Use cortexa9_extra_costs.
3292 2013-10-18 Jeff Law <law@redhat.com>
3294 * tree-ssa-threadupdate.c: Do not include "tm.h" or "tm_p.h".
3296 * tree-ssa-threadupdate.c: Include "dbgcnt.h".
3297 (register_jump_thread): Add "registered_jump_thread" debug
3299 * dbgcnt.def (registered_jump_thread): New debug counter.
3301 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
3303 * config/rs6000/rs6000.c: Include cgraph.h.
3305 2013-10-18 Teresa Johnson <tejohnson@google.com>
3307 * tree-ssa-tail-merge.c (replace_block_by): Update edge
3308 weights during merging.
3310 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
3312 * tree-cfg.h: Rename from tree-flow.h. Remove #includes.
3313 * tree-ssa.h: Relocate required #includes from tree-cfg.h.
3314 * tree-ssa-operands.h: Remove prototype.
3315 * tree-ssa-operands.c (virtual_operand_p): Move to gimple.c.
3316 * gimple.c (virtual_operand_p): Relocate from gimple.c.
3317 * gimple.h: Add prototype.
3318 * gimple-ssa.h: Include tree-ssa-operands.h.
3319 * tree-dump.c: Add tree-cfg.h to include list.
3320 * tree-ssa-alias.c: Add ipa-reference.h to include list.
3321 * config/alpha/alpha.c: Include gimple-ssa.h instead of tree-flow.h.
3322 * config/i386/i386.c: Don't include tree-flow.h.
3323 * config/rs6000/rs6000.c: Likewise.
3325 2013-10-18 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3327 * config/frv/frv.c (frv_init_cumulative_args): Fix wrong cast.
3329 2013-10-18 Richard Biener <rguenther@suse.de>
3331 * stor-layout.c (layout_type): Do not change TYPE_PRECISION
3332 or TYPE_UNSIGNED of integral types.
3333 (set_min_and_max_values_for_integral_type): Leave TYPE_MIN/MAX_VALUE
3334 NULL_TREE for zero-precision integral types.
3336 2013-10-18 James Greenhalgh <james.greenhalgh@arm.com>
3338 * config/aarch64/arm_neon.h
3339 (vcvt<ds>_n_<fsu><32,64>_<fsu><32,64>): Correct argument types.
3341 2013-10-17 Sriraman Tallam <tmsriram@google.com>
3344 * opth-gen.awk: Define target_flags_explicit.
3346 2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
3348 * config/rs6000/rs6000.c (enum rs6000_reload_reg_type): Add new
3349 fields to the reg_addr array that describes the valid addressing
3350 mode for any register, general purpose registers, floating point
3351 registers, and Altivec registers.
3352 (FIRST_RELOAD_REG_CLASS): Likewise.
3353 (LAST_RELOAD_REG_CLASS): Likewise.
3354 (struct reload_reg_map_type): Likewise.
3355 (reload_reg_map_type): Likewise.
3356 (RELOAD_REG_VALID): Likewise.
3357 (RELOAD_REG_MULTIPLE): Likewise.
3358 (RELOAD_REG_INDEXED): Likewise.
3359 (RELOAD_REG_OFFSET): Likewise.
3360 (RELOAD_REG_PRE_INCDEC): Likewise.
3361 (RELOAD_REG_PRE_MODIFY): Likewise.
3362 (reg_addr): Likewise.
3363 (mode_supports_pre_incdec_p): New helper functions to say whether
3364 a given mode supports PRE_INC, PRE_DEC, and PRE_MODIFY.
3365 (mode_supports_pre_modify_p): Likewise.
3366 (rs6000_debug_vector_unit): Rearrange the -mdebug=reg output to
3367 print the valid address mode bits for each mode.
3368 (rs6000_debug_print_mode): Likewise.
3369 (rs6000_debug_reg_global): Likewise.
3370 (rs6000_setup_reg_addr_masks): New function to set up the address
3371 mask bits for each type.
3372 (rs6000_init_hard_regno_mode_ok): Use memset to clear arrays.
3373 Call rs6000_setup_reg_addr_masks to set up the address mask bits.
3374 (rs6000_legitimate_address_p): Use mode_supports_pre_incdec_p and
3375 mode_supports_pre_modify_p to determine if PRE_INC, PRE_DEC, and
3376 PRE_MODIFY are supported.
3377 (rs6000_output_move_128bit): Change to use {src,dest}_vmx_p for altivec
3378 registers, instead of {src,dest}_av_p.
3379 (rs6000_print_options_internal): Tweak the debug output slightly.
3381 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
3383 * config/i386/sse.md (*vec_widen_smult_even_v8si): Remove
3386 2013-10-17 Andrew MacLeod <amacleod@redhat.com>
3388 * tree-flow.h (struct omp_region): Move to omp-low.c.
3389 Remove omp_ prototypes and variables.
3390 * gimple.h (omp_reduction_init): Move prototype to omp-low.h.
3391 (copy_var_decl): Relocate prototype from tree-flow.h.
3392 * gimple.c (copy_var_decl): Relocate from omp-low.c.
3393 * tree.h: Move prototype to omp-low.h.
3394 * omp-low.h: New File. Relocate prototypes here.
3395 * omp-low.c (struct omp_region): Make local here.
3396 (root_omp_region): Make static.
3397 (copy_var_decl) Move to gimple.c.
3398 (new_omp_region): Make static.
3399 (make_gimple_omp_edges): New. Refactored from tree-cfg.c make_edges.
3400 * tree-cfg.c: Include omp-low.h.
3401 (make_edges): Factor out OMP specific bits to make_gimple_omp_edges.
3402 * gimplify.c: Include omp-low.h.
3403 * tree-parloops.c: Likewise.
3405 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
3407 * config/i386/i386.c (ix86_fixup_binary_operands): When both source
3408 operands are in memory, prefer to force non-matched operand 1 to
3411 2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
3414 * config/rs6000/rs6000.c (rs6000_legitimate_address_p): Only
3415 restrict TImode addresses to single indirect registers if both
3416 -mquad-memory and -mvsx-timode are used.
3417 (rs6000_output_move_128bit): Use quad_load_store_p to determine if
3418 we should emit load/store quad. Remove using %y for quad memory
3421 * config/rs6000/rs6000.md (mov<mode>_ppc64, TI/PTImode): Add
3422 constraints to allow load/store quad on machines where TImode is
3423 not allowed in VSX registers. Use 'n' instead of 'F' constraint
3424 for TImode to load integer constants.
3426 2013-10-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3428 * config/aarch64/aarch64.c (aarch64_print_operand): Handle 'c'.
3430 2013-10-17 Marcus Shawcroft <marcus.shawcroft@arm.com>
3432 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Adjust
3433 handling of STACK_REG.
3435 2013-10-17 Richard Biener <rguenther@suse.de>
3437 PR tree-optimization/58143
3438 * tree-ssa-loop-im.c (arith_code_with_undefined_signed_overflow):
3440 (rewrite_to_defined_overflow): Likewise.
3441 (move_computations_dom_walker::before_dom): Rewrite stmts
3442 with undefined signed overflow that are not always executed
3443 into unsigned arithmetic.
3445 2013-10-16 Michael Meissner <meissner@linux.vnet.ibm.com>
3448 * config/rs6000/rs6000.opt (rs6000_isa_flags_explicit): Move the
3449 explicit isa flag to be an options variable, instead of using
3450 global_options_set. Remove define from rs6000.h.
3451 * config/rs6000/rs6000.h (rs6000_isa_flags_explicit): Likewise.
3453 * config/rs6000/rs6000.c (rs6000_option_override_internal):
3454 Initialize rs6000_isa_flags_explicit.
3455 (rs6000_function_specific_save): Add gcc_options* parameter, so
3456 that the powerpc builds after the 2013-10-15 changes.
3457 (rs6000_function_specific_restore): Likewise.
3459 2013-10-16 DJ Delorie <dj@redhat.com>
3461 * config/rl78/rl78.c (rl78_alloc_address_registers_macax): Verify
3462 op is a REG before checking REGNO.
3463 (rl78_alloc_physical_registers): Verify pattern is a SET before
3466 2013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3468 * config/rs6000/vector.md (vec_unpacks_hi_v4sf): Correct for
3470 (vec_unpacks_lo_v4sf): Likewise.
3471 (vec_unpacks_float_hi_v4si): Likewise.
3472 (vec_unpacks_float_lo_v4si): Likewise.
3473 (vec_unpacku_float_hi_v4si): Likewise.
3474 (vec_unpacku_float_lo_v4si): Likewise.
3476 2013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3478 * config/rs6000/vsx.md (vsx_concat_<mode>): Adjust output for LE.
3479 (vsx_concat_v2sf): Likewise.
3481 2013-10-16 James Greenhalgh <james.greenhalgh@arm.com>
3483 * config/aarch64/aarch64.md
3484 (*mov<mode>_aarch64): Fix output template for DUP (element) Scalar.
3486 2013-10-16 Andrew MacLeod <amacleod@redhat.com>
3488 PR tree-optimization/58697
3489 * cfgloop.c (get_estimated_loop_iterations_int): Rename from
3490 estimated_loop_iterations_int.
3491 (max_stmt_executions_int): Call get_max_loop_iterations_int.
3492 (get_max_loop_iterations_int): New. HWINT version of
3493 get_max_loop_iterations.
3494 * cfgloop.h: Add prototypes.
3495 * loop-iv.c (find_simple_exit): call get_estimated_loop_iterations_int.
3496 * loop-unroll.c (decide_peel_once_rolling): Call
3497 get_estimated_loop_iterations_int.
3498 * tree-ssa-loop-niter.c (estimated_loop_iterations_int): Add back.
3499 * tree-ssa-loop-niter.h: Tweak prototypes.
3501 2013-10-16 David Malcolm <dmalcolm@redhat.com>
3503 * gengtype-parse.c (struct_field_seq): Ignore access-control
3504 keywords ("public:" etc).
3506 2013-10-16 Marcus Shawcroft <marcus.shawcroft@arm.com>
3508 * config/aarch64/aarch64.c (aarch64_regno_regclass): Classify
3509 FRAME_POINTER_REGNUM and ARG_POINTER_REGNUM as POINTER_REGS.
3511 2013-10-16 Yvan Roux <yvan.roux@linaro.org>
3513 * config/arm/arm.opt (mlra): New option.
3514 * config/arm/arm.c (arm_lra_p): New function.
3515 (TARGET_LRA_P): Define.
3517 2013-10-16 Paulo Matos <pmatos@broadcom.com>
3519 * tree-core.h (tree_code_name): Remove.
3520 * tree.h (get_tree_code_name): New prototype.
3521 * tree.c (tree_code_name): Make static.
3522 (get_tree_code_name): New function.
3523 (dump_tree_statistics, tree_check_failed, tree_not_check_failed,
3524 tree_class_check_failed, tree_range_check_failed,
3525 tree_not_class_check_failed, omp_clause_check_failed,
3526 tree_contains_struct_check_failed, tree_operand_check_failed): Use new
3527 wrapper get_tree_code_name instead of calling tree_code_name directly.
3528 * tree-vrp.c (dump_asserts_for): Likewise.
3529 * tree-dump.c (dequeue_and_dump): Likewise.
3530 * tree-pretty-print.c (do_niy, dump_generic_node): Likewise.
3531 * tree-pretty-print.h (pp_unsupported_tree): Likewise.
3532 * lto-streamer-out.c (lto_write_tree, DFS_write_tree): Likewise.
3533 * tree-ssa-dom.c (print_expr_hash_elt): Likewise.
3534 * gimple-pretty-print.c (dump_unary_rhs, dump_binary_rhs,
3535 dump_ternary_rhs, dump_gimple_assign, dump_gimple_cond,
3536 dump_gimple_omp_for): Likewise.
3537 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Likewise.
3538 * tree-ssa-pre.c (print_pre_expr): Likewise.
3539 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Likewise.
3540 * print-tree.c (print_node_brief, print_node): Likewise.
3541 * gimple.c (gimple_check_failed): Likewise.
3542 * lto-streamer.c (lto_tag_name, print_lto_report): Likewise.
3543 * config/frv/frv.c (frv_init_cumulative_args): Likewise.
3544 * config/mep/mep.c (mep_validate_vliw): Likewise.
3545 * config/iq2000/iq2000.c (init_cumulative_args): Likewise.
3546 * config/rs6000/rs6000.c (init_cumulative_args): Likewise.
3548 2013-10-16 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
3550 * config/i386/i386.c (ix86_option_override_internal): Enable FMA4
3553 2013-10-16 Hans-Peter Nilsson <hp@axis.com>
3555 * config/cris/t-elfmulti (MULTILIB_OPTIONS, MULTILIB_DIRNAMES)
3556 (MULTILIB_MATCHES): Add multilib for -march=v8.
3558 2013-10-15 Sriraman Tallam <tmsriram@google.com>
3561 * optc-save-gen.awk: Add extra parameter to the save and restore
3563 * opth-gen.awk: Generate new TARGET_* macros to accept a parameter.
3564 * tree.c (build_optimization_node): New parameter. Add extra parameter
3565 to call to cl_optimization_save.
3566 (build_target_option_node): New parameter. Add extra parameter
3567 to call to cl_target_option_save.
3568 * tree.h (build_optimization_node): New parameter.
3569 (build_target_option_node): New parameter.
3570 * c-family/c-common.c (handle_optimize_attribute): Fix calls to
3571 build_optimization_node and build_target_option_node.
3572 * c-family/c-pragma.c (handle_pragma_optimize): Ditto.
3573 (handle_pragma_push_options): Ditto.
3574 * toplev.c (process_options): Ditto.
3575 * opts.c (init_options_struct): Check for opts_set non-null.
3576 * target.def (target_option.save): New parameter.
3577 (target_option.restore): New parameter.
3578 * tm.texi: Generate.
3579 * config/i386/i386-c.c (ix86_target_macros_internal): Ditto.
3580 (ix86_pragma_target_parse): Ditto.
3581 * config/i386/i386-protos.h (ix86_valid_target_attribute_tree): New
3583 * config/rs6000/rs6000.c (rs6000_option_override_internal): Fix calls
3584 to build_optimization_node and build_target_option_node.
3585 (rs6000_valid_attribute_p): Ditto.
3586 (rs6000_pragma_target_parse): Ditto.
3587 * config/i386/i386.opt (x_ix86_target_flags_explicit): New TargetSave
3589 * config/i386/i386.h:
3590 TARGET_64BIT_P: New Macro
3591 TARGET_MMX_P: New Macro.
3592 TARGET_3DNOW_P: New Macro.
3593 TARGET_3DNOW_A_P: New Macro.
3594 TARGET_SSE_P: New Macro.
3595 TARGET_SSE2_P: New Macro.
3596 TARGET_SSE3_P: New Macro.
3597 TARGET_SSSE3_P: New Macro.
3598 TARGET_SSE4_1_P: New Macro.
3599 TARGET_SSE4_2_P: New Macro.
3600 TARGET_AVX_P: New Macro.
3601 TARGET_AVX2_P: New Macro.
3602 TARGET_AVX512F_P: New Macro.
3603 TARGET_AVX512PF_P: New Macro.
3604 TARGET_AVX512ER_P: New Macro.
3605 TARGET_AVX512CD_P: New Macro.
3606 TARGET_FMA_P: New Macro.
3607 TARGET_SSE4A_P: New Macro.
3608 TARGET_FMA4_P: New Macro.
3609 TARGET_XOP_P: New Macro.
3610 TARGET_LWP_P: New Macro.
3611 TARGET_ABM_P: New Macro.
3612 TARGET_BMI_P: New Macro.
3613 TARGET_BMI2_P: New Macro.
3614 TARGET_LZCNT_P: New Macro.
3615 TARGET_TBM_P: New Macro.
3616 TARGET_POPCNT_P: New Macro.
3617 TARGET_SAHF_P: New Macro.
3618 TARGET_MOVBE_P: New Macro.
3619 TARGET_CRC32_P: New Macro.
3620 TARGET_AES_P: New Macro.
3621 TARGET_PCLMUL_P: New Macro.
3622 TARGET_CMPXCHG16B_P: New Macro.
3623 TARGET_FSGSBASE_P: New Macro.
3624 TARGET_RDRND_P: New Macro.
3625 TARGET_F16C_P: New Macro.
3626 TARGET_RTM_P: New Macro.
3627 TARGET_HLE_P: New Macro.
3628 TARGET_RDSEED_P: New Macro.
3629 TARGET_PRFCHW_P: New Macro.
3630 TARGET_ADX_P: New Macro.
3631 TARGET_FXSR_P: New Macro.
3632 TARGET_XSAVE_P: New Macro.
3633 TARGET_XSAVEOPT_P: New Macro.
3634 TARGET_LP64_P: New Macro.
3635 TARGET_X32_P: New Macro.
3636 TARGET_FPMATH_DEFAULT_P: New Macro.
3637 TARGET_FLOAT_RETURNS_IN_80387_P: New Macro.
3638 * config/i386/i386.c (ix86_option_override_internal): New parameters.
3640 Change ix86_tune_string to access opts->x_ix86_tune_string.
3641 Change ix86_isa_flags to access opts->x_ix86_isa_flags.
3642 Change ix86_arch_string to access opts->x_ix86_arch_string.
3643 Change ix86_stringop_alg to access opts->x_ix86_stringop_alg.
3644 Change ix86_pmode to access opts->x_ix86_pmode.
3645 Change ix86_abi to access opts->x_ix86_abi.
3646 Change ix86_cmodel to access opts->x_ix86_cmodel.
3647 Change ix86_asm_dialect to access opts->x_ix86_asm_dialect.
3648 Change ix86_isa_flags_explicit to access
3649 opts->x_ix86_isa_flags_explicit.
3650 Change ix86_dump_tunes to access opts->x_ix86_dump_tunes.
3651 Change ix86_regparm to access opts->x_ix86_regparm.
3652 Change ix86_branch_cost to access opts->x_ix86_branch_cost.
3653 Change ix86_preferred_stack_boundary_arg to access
3654 opts->x_ix86_preferred_stack_boundary_arg.
3655 Change ix86_force_align_arg_pointer to access
3656 opts->x_ix86_force_align_arg_pointer.
3657 Change ix86_incoming_stack_boundar_arg to access
3658 opts->x_ix86_incoming_stack_boundar_arg.
3659 Change ix86_fpmath to access opts->x_ix86_fpmath.
3660 Change ix86_veclibabi_type to access opts->x_ix86_veclibabi_type.
3661 Change ix86_recip_name to access opts->x_ix86_recip_name.
3662 Change ix86_stack_protector_guard to access
3663 opts->x_ix86_stack_protector_guard.
3664 Change ix86_tune_memcpy_strategy to access
3665 opts->x_ix86_tune_memcpy_strategy.
3666 Change ix86_tune_memset_strategy to access
3667 opts->x_ix86_tune_memset_strategy.
3668 Change global_options to access opts.
3669 Change global_options_set to access opts_set.
3670 Change TARGET_64BIT to TARGET_64BIT_P (opts->...).
3671 Change TARGET_MMX to TARGET_MMX_P (opts->...).
3672 Change TARGET_3DNOW to TARGET_3DNOW_P (opts->...).
3673 Change TARGET_3DNOW_A to TARGET_3DNOW_A_P (opts->...).
3674 Change TARGET_SSE to TARGET_SSE_P (opts->...).
3675 Change TARGET_SSE2 to TARGET_SSE2_P (opts->...).
3676 Change TARGET_SSE3 to TARGET_SSE3_P (opts->...).
3677 Change TARGET_SSSE3 to TARGET_SSSE3_P (opts->...).
3678 Change TARGET_SSE4_1 to TARGET_SSE4_1_P (opts->...).
3679 Change TARGET_SSE4_2 to TARGET_SSE4_2_P (opts->...).
3680 Change TARGET_AVX to TARGET_AVX_P (opts->...).
3681 Change TARGET_AVX2 to TARGET_AVX2_P (opts->...).
3682 Change TARGET_AVX512F to TARGET_AVX512F_P (opts->...).
3683 Change TARGET_AVX512PF to TARGET_AVX512PF_P (opts->...).
3684 Change TARGET_AVX512ER to TARGET_AVX512ER_P (opts->...).
3685 Change TARGET_AVX512CD to TARGET_AVX512CD_P (opts->...).
3686 Change TARGET_FMA to TARGET_FMA_P (opts->...).
3687 Change TARGET_SSE4A to TARGET_SSE4A_P (opts->...).
3688 Change TARGET_FMA4 to TARGET_FMA4_P (opts->...).
3689 Change TARGET_XOP to TARGET_XOP_P (opts->...).
3690 Change TARGET_LWP to TARGET_LWP_P (opts->...).
3691 Change TARGET_ABM to TARGET_ABM_P (opts->...).
3692 Change TARGET_BMI to TARGET_BMI_P (opts->...).
3693 Change TARGET_BMI2 to TARGET_BMI2_P (opts->...).
3694 Change TARGET_LZCNT to TARGET_LZCNT_P (opts->...).
3695 Change TARGET_TBM to TARGET_TBM_P (opts->...).
3696 Change TARGET_POPCNT to TARGET_POPCNT_P (opts->...).
3697 Change TARGET_SAHF to TARGET_SAHF_P (opts->...).
3698 Change TARGET_MOVBE to TARGET_MOVBE_P (opts->...).
3699 Change TARGET_CRC32 to TARGET_CRC32_P (opts->...).
3700 Change TARGET_AES to TARGET_AES_P (opts->...).
3701 Change TARGET_PCLMUL to TARGET_PCLMUL_P (opts->...).
3702 Change TARGET_CMPXCHG16B to TARGET_CMPXCHG16B_P (opts->...).
3703 Change TARGET_FSGSBASE to TARGET_FSGSBASE_P (opts->...).
3704 Change TARGET_RDRND to TARGET_RDRND_P (opts->...).
3705 Change TARGET_F16C to TARGET_F16C_P (opts->...).
3706 Change TARGET_RTM to TARGET_RTM_P (opts->...).
3707 Change TARGET_HLE to TARGET_HLE_P (opts->...).
3708 Change TARGET_RDSEED to TARGET_RDSEED_P (opts->...).
3709 Change TARGET_PRFCHW to TARGET_PRFCHW_P (opts->...).
3710 Change TARGET_ADX to TARGET_ADX_P (opts->...).
3711 Change TARGET_FXSR to TARGET_FXSR_P (opts->...).
3712 Change TARGET_XSAVE to TARGET_XSAVE_P (opts->...).
3713 Change TARGET_XSAVEOPT to TARGET_XSAVEOPT_P (opts->...).
3714 Change TARGET_LP64 to TARGET_LP64_P (opts->...).
3715 Change TARGET_X32 to TARGET_X32_P (opts->...).
3716 Change TARGET_FPMATH_DEFAULT to TARGET_FPMATH_DEFAULT_P (opts->...).
3717 Change TARGET_FLOAT_RETURNS_IN_80387 to
3718 TARGET_FLOAT_RETURNS_IN_80387_P (opts->...).
3719 (ix86_function_specific_save): New parameter. Use opts-> fields
3720 to replace global fields.
3721 (ix86_function_specific_restore): Ditto.
3722 (ix86_valid_target_attribute_inner_p): New parameters.
3724 Fix call to ix86_handle_option and set_option.
3725 (ix86_valid_target_attribute_tree): New parameters.
3726 Change global_options to access opts.
3727 Change global_options_set to access opts_set.
3728 Fix call to ix86_valid_target_attribute_inner_p.
3729 Change ix86_tune_string to access opts->x_ix86_tune_string.
3730 Change ix86_arch_string to access opts->x_ix86_arch_string.
3731 Change ix86_fpmath to access opts->x_ix86_fpmath
3732 Fix call to ix86_option_override_internal.
3733 Fix call to ix86_add_new_builtins.
3734 Fix calls to build_optimization_node and build_target_option_node.
3735 (ix86_valid_target_attribute_p): Remove access to global_options.
3736 Use new gcc_options structure func_options.
3737 Fix call to ix86_valid_target_attribute_tree.
3738 Fix call to build_optimization_node.
3739 (get_builtin_code_for_version): Fix call to
3740 ix86_valid_target_attribute_tree.
3742 2013-10-15 David Malcolm <dmalcolm@redhat.com>
3744 * Makefile.in (PICFLAG): New.
3745 (enable_host_shared): New.
3746 (INTERNAL_CFLAGS): Use PICFLAG.
3747 (LIBIBERTY): Use pic build of libiberty.a if configured with
3748 --enable-host-shared.
3749 * configure.ac: Add --enable-host-shared, setting up new
3751 * configure: Regenerate.
3752 * doc/install.texi (--enable-shared): Add note contrasting it with ...
3753 (--enable-host-shared): New option.
3755 2013-10-15 Richard Biener <rguenther@suse.de>
3757 * tree-tailcall.c (find_tail_calls): Don't use tail-call recursion
3758 for built-in functions.
3760 2013-10-15 Zhenqiang Chen <zhenqiang.chen@arm.com>
3762 * tree-ssa-reassoc.c: Include rtl.h and tm_p.h.
3763 (optimize_range_tests_1): New function,
3764 extracted from optimize_range_tests.
3765 (optimize_range_tests_xor): Similarly.
3766 (optimize_range_tests_diff): New function.
3767 (optimize_range_tests): Use optimize_range_tests_1.
3769 2013-10-15 Cong Hou <congh@google.com>
3771 * tree-vect-loop.c (vect_is_simple_reduction_1): Relax the
3772 requirement of the reduction pattern so that one operand of the
3773 reduction operation can come from outside of the loop.
3775 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3777 * config/arm/neon-schedgen.ml: Remove.
3778 * config/arm/cortex-a9-neon.md: Remove comment regarding
3781 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3783 * config/arm/types: Remove old neon types.
3785 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3787 * config/arm/cortex-a7.md
3788 (cortex_a7_neon_type): New.
3789 (cortex_a7_neon_mul): Update for new types.
3790 (cortex_a7_neon_mla): Likewise.
3791 (cortex_a7_neon): Likewise.
3793 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3795 * config/arm/cortex-a15-neon.md
3796 (cortex_a15_neon_type): New,
3798 (cortex_a15_neon_int_1): Remove.
3799 (cortex_a15_neon_int_2): Likewise.
3800 (cortex_a15_neon_int_3): Likewise.
3801 (cortex_a15_neon_int_4): Likewise.
3802 (cortex_a15_neon_int_5): Likewise.
3803 (cortex_a15_neon_vqneg_vqabs): Likewise.
3804 (cortex_a15_neon_vmov): Likewise.
3805 (cortex_a15_neon_vaba): Likewise.
3806 (cortex_a15_neon_vaba_qqq): Likewise.
3807 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3808 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
3809 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
3811 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3812 (cortex_a15_neon_mla_qqq_8_16): Likewise.
3813 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar): Likewise.
3814 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
3815 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
3816 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
3817 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
3818 (cortex_a15_neon_shift_1): Likewise.
3819 (cortex_a15_neon_shift_2): Likewise.
3820 (cortex_a15_neon_shift_3): Likewise.
3821 (cortex_a15_neon_vshl_ddd): Likewise.
3822 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
3823 (cortex_a15_neon_vsra_vrsra): Likewise.
3824 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
3825 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
3826 (cortex_a15_neon_bp_3cycle): Likewise.
3827 (cortex_a15_neon_ldm_2): Likewise.
3828 (cortex_a15_neon_stm_2): Likewise.
3829 (cortex_a15_neon_mcr): Likewise.
3830 (cortex_a15_neon_mrc): Likewise.
3831 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
3832 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
3833 (cortex_a15_neon_fp_vmul_ddd): Likewise.
3834 (cortex_a15_neon_fp_vmul_qqd): Likewise.
3835 (cortex_a15_neon_fp_vmla_ddd): Likewise.
3836 (cortex_a15_neon_fp_vmla_qqq): Likewise.
3837 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
3838 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
3839 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
3840 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
3841 (cortex_a15_neon_bp_simple): Likewise.
3842 (cortex_a15_neon_bp_2cycle): Likewise.
3843 (cortex_a15_neon_bp_3cycle): Likewise.
3844 (cortex_a15_neon_vld1_1_2_regs): Likewise.
3845 (cortex_a15_neon_vld1_3_4_regs): Likewise.
3846 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
3847 (cortex_a15_neon_vld2_4_regs): Likewise.
3848 (cortex_a15_neon_vld3_vld4): Likewise.
3849 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
3850 (cortex_a15_neon_vst1_3_4_regs): Likewise.
3851 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Rename to...
3852 (cortex_a15_neon_vst2_4_regs_vst3): ...This, update for new attributes.
3853 (cortex_a15_neon_vst3_vst4): Rename to...
3854 (cortex_a15_neon_vst4): This, update for new attributes.
3855 (cortex_a15_neon_vld1_vld2_lane): Update for new attributes.
3856 (cortex_a15_neon_vld3_vld4_lane): Likewise.
3857 (cortex_a15_neon_vst1_vst2_lane): Likewise.
3858 (cortex_a15_neon_vst3_vst4_lane): Likewise.
3859 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
3860 (cortex_a15_neon_ldm_2): Likewise.
3861 (cortex_a15_neon_stm_2): Likewise.
3862 (cortex_a15_neon_mcr): Likewise.
3863 (cortex_a15_neon_mcr_2_mcrr): Likewise.
3864 (cortex_a15_neon_mrc): Likewise.
3865 (cortex_a15_neon_mrrc): Likewise.
3867 (cortex_a15_neon_abd): New.
3868 (cortex_a15_neon_abd_q): Likewise.
3869 (cortex_a15_neon_aba): Likewise.
3870 (cortex_a15_neon_aba_q): Likewise.
3871 (cortex_a15_neon_acc): Likewise.
3872 (cortex_a15_neon_acc_q): Likewise.
3873 (cortex_a15_neon_arith_basic): Likewise.
3874 (cortex_a15_neon_arith_complex): Likewise.
3875 (cortex_a15_neon_multiply): Likewise.
3876 (cortex_a15_neon_multiply_q): Likewise.
3877 (cortex_a15_neon_mla): Likewise.
3878 (cortex_a15_neon_mla_q): Likewise.
3879 (cortex_a15_neon_sat_mla_long): Likewise.
3880 (cortex_a15_neon_shift_acc): Likewise.
3881 (cortex_a15_neon_shift_imm_basic): Likewise.
3882 (cortex_a15_neon_shift_imm_complex): Likewise.
3883 (cortex_a15_neon_shift_reg_basic): Likewise.
3884 (cortex_a15_neon_shift_reg_basic_q): Likewise.
3885 (cortex_a15_neon_shift_reg_complex): Likewise.
3886 (cortex_a15_neon_shift_reg_complex_q): Likewise.
3887 (cortex_a15_neon_fp_negabs): Likewise
3888 (cortex_a15_neon_fp_arith): Likewise
3889 (cortex_a15_neon_fp_arith_q): Likewise
3890 (cortex_a15_neon_fp_cvt_int): Likewise
3891 (cortex_a15_neon_fp_cvt_int_q): Likewise
3892 (cortex_a15_neon_fp_cvt_16): Likewise
3893 (cortex_a15_neon_fp_mul): Likewise
3894 (cortex_a15_neon_fp_mul_q): Likewise
3895 (cortex_a15_neon_fp_mla): Likewise
3896 (cortex_a15_neon_fp_mla_q): Likewise
3897 (cortex_a15_neon_fp_recps_rsqrte): Likewise.
3898 (cortex_a15_neon_fp_recps_rsqrte_q): Likewise.
3899 (cortex_a15_neon_bitops): Likewise.
3900 (cortex_a15_neon_bitops_q): Likewise.
3901 (cortex_a15_neon_from_gp): Likewise.
3902 (cortex_a15_neon_from_gp_q): Likewise.
3903 (cortex_a15_neon_tbl3_tbl4): Likewise.
3904 (cortex_a15_neon_zip_q): Likewise.
3905 (cortex_a15_neon_to_gp): Likewise.
3906 (cortex_a15_neon_load_a): Likewise.
3907 (cortex_a15_neon_load_b): Likewise.
3908 (cortex_a15_neon_load_c): Likewise.
3909 (cortex_a15_neon_load_d): Likewise.
3910 (cortex_a15_neon_load_e): Likewise.
3911 (cortex_a15_neon_load_f): Likewise.
3912 (cortex_a15_neon_store_a): Likewise.
3913 (cortex_a15_neon_store_b): Likewise.
3914 (cortex_a15_neon_store_c): Likewise.
3915 (cortex_a15_neon_store_d): Likewise.
3916 (cortex_a15_neon_store_e): Likewise.
3917 (cortex_a15_neon_store_f): Likewise.
3918 (cortex_a15_neon_store_g): Likewise.
3919 (cortex_a15_neon_store_h): Likewise.
3920 (cortex_a15_vfp_to_from_gp): Likewise.
3922 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3924 * config/arm/cortex-a9-neon.md (cortex_a9_neon_type): New.
3926 (cortex_a9_neon_vshl_ddd): Remove.
3927 (cortex_a9_neon_vst3_vst4): Likewise.
3928 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
3930 (cortex_a9_neon_bit_ops_q): New.
3932 (cortex_a9_neon_int_1): Use cortex_a8_neon_type.
3933 (cortex_a9_neon_int_2): Likewise.
3934 (cortex_a9_neon_int_3): Likewise.
3935 (cortex_a9_neon_int_4): Likewise.
3936 (cortex_a9_neon_int_5): Likewise.
3937 (cortex_a9_neon_vqneg_vqabs): Likewise.
3938 (cortex_a9_neon_vmov): Likewise.
3939 (cortex_a9_neon_vaba): Likewise.
3940 (cortex_a9_neon_vaba_qqq): Likewise.
3941 (cortex_a9_neon_shift_1): Likewise.
3942 (cortex_a9_neon_shift_2): Likewise.
3943 (cortex_a9_neon_shift_3): Likewise.
3944 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
3945 (cortex_a9_neon_vsra_vrsra): Likewise.
3946 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3947 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
3948 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
3950 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3951 (cortex_a9_neon_mla_qqq_8_16): Likewise.
3952 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
3954 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
3955 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
3956 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
3957 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
3958 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
3959 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
3960 (cortex_a9_neon_fp_vsum): Likewise.
3961 (cortex_a9_neon_fp_vmul_ddd): Likewise.
3962 (cortex_a9_neon_fp_vmul_qqd): Likewise.
3963 (cortex_a9_neon_fp_vmla_ddd): Likewise.
3964 (cortex_a9_neon_fp_vmla_qqq): Likewise.
3965 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
3966 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
3967 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
3968 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
3969 (cortex_a9_neon_bp_simple): Likewise.
3970 (cortex_a9_neon_bp_2cycle): Likewise.
3971 (cortex_a9_neon_bp_3cycle): Likewise.
3972 (cortex_a9_neon_ldr): Likewise.
3973 (cortex_a9_neon_str): Likewise.
3974 (cortex_a9_neon_vld1_1_2_regs): Likewise.
3975 (cortex_a9_neon_vld1_3_4_regs): Likewise.
3976 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
3977 (cortex_a9_neon_vld2_4_regs): Likewise.
3978 (cortex_a9_neon_vld3_vld4): Likewise.
3979 (cortex_a9_neon_vld1_vld2_lane): Likewise.
3980 (cortex_a9_neon_vld3_vld4_lane): Likewise.
3981 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
3982 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
3983 (cortex_a9_neon_vst1_3_4_regs): Likewise.
3984 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
3985 (cortex_a9_neon_vst1_vst2_lane): Likewise.
3986 (cortex_a9_neon_vst3_vst4_lane): Likewise.
3987 (cortex_a9_neon_mcr): Likewise.
3988 (cortex_a9_neon_mcr_2_mcrr): Likewise.
3989 (cortex_a9_neon_mrc): Likewise.
3990 (cortex_a9_neon_mrrc): Likewise.
3992 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3994 * config/arm/cortex-a8-neon.md (cortex_a8_neon_type): New.
3996 (cortex_a8_neon_vshl_ddd): Remove.
3997 (cortex_a8_neon_vst3_vst4): Likewise.
3998 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
4000 (cortex_a8_neon_bit_ops_q): New.
4002 (cortex_a8_neon_int_1): Use cortex_a8_neon_type.
4003 (cortex_a8_neon_int_2): Likewise..
4004 (cortex_a8_neon_int_3): Likewise.
4005 (cortex_a8_neon_int_5): Likewise.
4006 (cortex_a8_neon_vqneg_vqabs): Likewise.
4007 (cortex_a8_neon_int_4): Likewise.
4008 (cortex_a8_neon_vaba): Likewise.
4009 (cortex_a8_neon_vaba_qqq): Likewise.
4010 (cortex_a8_neon_shift_1): Likewise.
4011 (cortex_a8_neon_shift_2): Likewise.
4012 (cortex_a8_neon_shift_3): Likewise.
4013 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
4014 (cortex_a8_neon_vsra_vrsra): Likewise.
4015 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
4016 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
4017 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
4019 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
4020 (cortex_a8_neon_mla_qqq_8_16): Likewise.
4021 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
4023 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
4024 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
4025 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
4026 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
4027 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
4028 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
4029 (cortex_a8_neon_fp_vsum): Likewise.
4030 (cortex_a8_neon_fp_vmul_ddd): Likewise.
4031 (cortex_a8_neon_fp_vmul_qqd): Likewise.
4032 (cortex_a8_neon_fp_vmla_ddd): Likewise.
4033 (cortex_a8_neon_fp_vmla_qqq): Likewise.
4034 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
4035 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
4036 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
4037 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
4038 (cortex_a8_neon_bp_simple): Likewise.
4039 (cortex_a8_neon_bp_2cycle): Likewise.
4040 (cortex_a8_neon_bp_3cycle): Likewise.
4041 (cortex_a8_neon_ldr): Likewise.
4042 (cortex_a8_neon_str): Likewise.
4043 (cortex_a8_neon_vld1_1_2_regs): Likewise.
4044 (cortex_a8_neon_vld1_3_4_regs): Likewise.
4045 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
4046 (cortex_a8_neon_vld2_4_regs): Likewise.
4047 (cortex_a8_neon_vld3_vld4): Likewise.
4048 (cortex_a8_neon_vld1_vld2_lane): Likewise.
4049 (cortex_a8_neon_vld3_vld4_lane): Likewise.
4050 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
4051 (cortex_a8_neon_vst1_3_4_regs): Likewise.
4052 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
4053 (cortex_a8_neon_vst1_vst2_lane): Likewise.
4054 (cortex_a8_neon_vst3_vst4_lane): Likewise.
4055 (cortex_a8_neon_mcr): Likewise.
4056 (cortex_a8_neon_mcr_2_mcrr): Likewise.
4057 (cortex_a8_neon_mrc): Likewise.
4058 (cortex_a8_neon_mrrc): Likewise.
4060 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
4062 * config/aarch64/iterators.md (Vetype): Add SF and DF modes.
4064 * config/aarch64/aarch64-simd.md (neon_type): Remove.
4065 (aarch64_simd_dup<mode>): Add "type" attribute.
4066 (aarch64_dup_lane<mode>): Likewise.
4067 (aarch64_dup_lane_<vswap_width_name><mode>): Likewise.
4068 (*aarch64_simd_mov<mode>): Likewise.
4069 (aarch64_simd_mov_from_<mode>low): Likewise.
4070 (aarch64_simd_mov_from_<mode>high): Likewise.
4071 (orn<mode>3): Likewise.
4072 (bic<mode>3): Likewise.
4073 (add<mode>3): Likewise.
4074 (sub<mode>3): Likewise.
4075 (mul<mode>3): Likewise.
4076 (*aarch64_mul3_elt<mode>): Likewise.
4077 (*aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
4078 (*aarch64_mul3_elt_to_128df): Likewise.
4079 (*aarch64_mul3_elt_to_64v2df): Likewise.
4080 (neg<mode>2): Likewise.
4081 (abs<mode>2): Likewise.
4082 (abd<mode>_3): Likewise.
4083 (aba<mode>_3): Likewise.
4084 (fabd<mode>_3): Likewise.
4085 (*fabd_scalar<mode>3): Likewise.
4086 (and<mode>3): Likewise.
4087 (ior<mode>3): Likewise.
4088 (xor<mode>3): Likewise.
4089 (one_cmpl<mode>2): Likewise.
4090 (aarch64_simd_vec_set<mode>): Likewise.
4091 (aarch64_simd_lshr<mode>): Likewise.
4092 (aarch64_simd_ashr<mode>): Likewise.
4093 (aarch64_simd_imm_shl<mode>): Likewise.
4094 (aarch64_simd_reg_sshl<mode): Likewise.
4095 (aarch64_simd_reg_shl<mode>_unsigned): Likewise.
4096 (aarch64_simd_reg_shl<mode>_signed): Likewise.
4097 (aarch64_simd_vec_setv2di): Likewise.
4098 (aarch64_simd_vec_set<mode>): Likewise.
4099 (aarch64_mla<mode>): Likewise.
4100 (*aarch64_mla_elt<mode>): Likewise.
4101 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
4102 (aarch64_mls<mode>): Likewise.
4103 (*aarch64_mls_elt<mode>): Likewise.
4104 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
4105 (<su><maxmin><mode>3): Likewise.
4106 (move_lo_quad_<mode>): Likewise.
4107 (aarch64_simd_move_hi_quad_<mode>): Likewise.
4108 (aarch64_simd_vec_pack_trunc_<mode>): Likewise.
4109 (vec_pack_trunc_<mode>): Likewise.
4110 (aarch64_simd_vec_unpack<su>_lo_<mode>): Likewise.
4111 (aarch64_simd_vec_unpack<su>_hi_<mode>): Likewise.
4112 (*aarch64_<su>mlal_lo<mode>): Likewise.
4113 (*aarch64_<su>mlal_hi<mode>): Likewise.
4114 (*aarch64_<su>mlsl_lo<mode>): Likewise.
4115 (*aarch64_<su>mlsl_hi<mode>): Likewise.
4116 (*aarch64_<su>mlal<mode>): Likewise.
4117 (*aarch64_<su>mlsl<mode>): Likewise.
4118 (aarch64_simd_vec_<su>mult_lo_<mode>): Likewise.
4119 (aarch64_simd_vec_<su>mult_hi_<mode>): Likewise.
4120 (add<mode>3): Likewise.
4121 (sub<mode>3): Likewise.
4122 (mul<mode>3): Likewise.
4123 (div<mode>3): Likewise.
4124 (neg<mode>2): Likewise.
4125 (abs<mode>2): Likewise.
4126 (fma<mode>4): Likewise.
4127 (*aarch64_fma4_elt<mode>): Likewise.
4128 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
4129 (*aarch64_fma4_elt_to_128df): Likewise.
4130 (*aarch64_fma4_elt_to_64v2df): Likewise.
4131 (fnma<mode>4): Likewise.
4132 (*aarch64_fnma4_elt<mode>): Likewise.
4133 (*aarch64_fnma4_elt_<vswap_width_name><mode>
4134 (*aarch64_fnma4_elt_to_128df): Likewise.
4135 (*aarch64_fnma4_elt_to_64v2df): Likewise.
4136 (<frint_pattern><mode>2): Likewise.
4137 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): Likewise.
4138 (<optab><fcvt_target><VDQF:VDQF:mode>2): Likewise.
4139 (vec_unpacks_lo_v4sf): Likewise.
4140 (aarch64_float_extend_lo_v2df): Likewise.
4141 (vec_unpacks_hi_v4sf): Likewise.
4142 (aarch64_float_truncate_lo_v2sf): Likewise.
4143 (aarch64_float_truncate_hi_v4sf): Likewise.
4144 (aarch64_vmls<mode>): Likewise.
4145 (<su><maxmin><mode>3): Likewise.
4146 (<maxmin_uns><mode>3): Likewise.
4147 (reduc_<sur>plus_<mode>): Likewise.
4148 (reduc_<sur>plus_v2di): Likewise.
4149 (reduc_<sur>plus_v2si): Likewise.
4150 (reduc_<sur>plus_<mode>): Likewise.
4151 (aarch64_addpv4sf): Likewise.
4152 (clz<mode>2): Likewise.
4153 (reduc_<maxmin_uns>_<mode>): Likewise.
4154 (reduc_<maxmin_uns>_v2di): Likewise.
4155 (reduc_<maxmin_uns>_v2si): Likewise.
4156 (reduc_<maxmin_uns>_<mode>): Likewise.
4157 (reduc_<maxmin_uns>_v4sf): Likewise.
4158 (aarch64_simd_bsl<mode>_internal): Likewise.
4159 (*aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): Likewise.
4160 (*aarch64_get_lane_zero_extendsi<mode>): Likewise.
4161 (aarch64_get_lane<mode>): Likewise.
4162 (*aarch64_combinez<mode>): Likewise.
4163 (aarch64_combine<mode>): Likewise.
4164 (aarch64_simd_combine<mode>): Likewise.
4165 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): Likewise.
4166 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): Likewise.
4167 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>): Likewise.
4168 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w<mode>): Likewise.
4169 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w2<mode>_internal): Likewise.
4170 (aarch64_<sur>h<addsub><mode>): Likewise.
4171 (aarch64_<sur><addsub>hn<mode>): Likewise.
4172 (aarch64_<sur><addsub>hn2<mode>): Likewise.
4173 (aarch64_pmul<mode>): Likewise.
4174 (aarch64_<su_optab><optab><mode>): Likewise.
4175 (aarch64_<sur>qadd<mode>): Likewise.
4176 (aarch64_sqmovun<mode>): Likewise.
4177 (aarch64_<sur>qmovn<mode>): Likewise.
4178 (aarch64_s<optab><mode>): Likewise.
4179 (aarch64_sq<r>dmulh<mode>): Likewise.
4180 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
4181 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
4182 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
4183 (aarch64_sqdml<SBINQOPS:as>l<mode>): Likewise.
4184 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
4185 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
4186 (aarch64_sqdml<SBINQOPS:as>l_n<mode>): Likewise.
4187 (aarch64_sqdml<SBINQOPS:as>l2<mode>_internal): Likewise.
4188 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
4189 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
4190 (aarch64_sqdmull<mode>): Likewise.
4191 (aarch64_sqdmull_lane<mode>_internal): Likewise.
4192 (aarch64_sqdmull_n<mode>): Likewise.
4193 (aarch64_sqdmull2<mode>_internal): Likewise.
4194 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
4195 (aarch64_sqdmull2_n<mode>_internal): Likewise.
4196 (aarch64_<sur>shl<mode>): Likewise.
4197 (aarch64_<sur>q<r>shl<mode>
4198 (aarch64_<sur>shll_n<mode>): Likewise.
4199 (aarch64_<sur>shll2_n<mode>): Likewise.
4200 (aarch64_<sur>shr_n<mode>): Likewise.
4201 (aarch64_<sur>sra_n<mode>): Likewise.
4202 (aarch64_<sur>s<lr>i_n<mode>): Likewise.
4203 (aarch64_<sur>qshl<u>_n<mode>): Likewise.
4204 (aarch64_<sur>q<r>shr<u>n_n<mode>): Likewise.
4205 (aarch64_cm<optab><mode>): Likewise.
4206 (aarch64_cm<optab>di): Likewise.
4207 (aarch64_cm<optab><mode>): Likewise.
4208 (aarch64_cm<optab>di): Likewise.
4209 (aarch64_cmtst<mode>): Likewise.
4210 (aarch64_cmtstdi): Likewise.
4211 (aarch64_cm<optab><mode>): Likewise.
4212 (*aarch64_fac<optab><mode>): Likewise.
4213 (aarch64_addp<mode>): Likewise.
4214 (aarch64_addpdi): Likewise.
4215 (sqrt<mode>2): Likewise.
4216 (vec_load_lanesoi<mode>): Likewise.
4217 (vec_store_lanesoi<mode>): Likewise.
4218 (vec_load_lanesci<mode>): Likewise.
4219 (vec_store_lanesci<mode>): Likewise.
4220 (vec_load_lanesxi<mode>): Likewise.
4221 (vec_store_lanesxi<mode>): Likewise.
4222 (*aarch64_mov<mode>): Likewise.
4223 (aarch64_ld2<mode>_dreg): Likewise.
4224 (aarch64_ld2<mode>_dreg): Likewise.
4225 (aarch64_ld3<mode>_dreg): Likewise.
4226 (aarch64_ld3<mode>_dreg): Likewise.
4227 (aarch64_ld4<mode>_dreg): Likewise.
4228 (aarch64_ld4<mode>_dreg): Likewise.
4229 (aarch64_tbl1<mode>): Likewise.
4230 (aarch64_tbl2v16qi): Likewise.
4231 (aarch64_combinev16qi): Likewise.
4232 (aarch64_<PERMUTE:perm_insn><PERMUTE:perm_hilo><mode>): Likewise.
4233 (aarch64_st2<mode>_dreg): Likewise.
4234 (aarch64_st2<mode>_dreg): Likewise.
4235 (aarch64_st3<mode>_dreg): Likewise.
4236 (aarch64_st3<mode>_dreg): Likewise.
4237 (aarch64_st4<mode>_dreg): Likewise.
4238 (aarch64_st4<mode>_dreg): Likewise.
4239 (*aarch64_simd_ld1r<mode>): Likewise.
4240 (aarch64_frecpe<mode>): Likewise.
4241 (aarch64_frecp<FRECP:frecp_suffix><mode>): Likewise.
4242 (aarch64_frecps<mode>): Likewise.
4244 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
4246 * config/arm/iterators.md (V_elem_ch): New.
4248 (VQH_type): Likewise.
4249 * config/arm/arm.md (is_neon_type): New.
4250 (conds): Use is_neon_type.
4251 (anddi3_insn): Update type attribute.
4252 (xordi3_insn): Likewise.
4253 (one_cmpldi2): Likewise.
4254 * config/arm/vfp.md (movhf_vfp_neon): Update type attribute.
4255 * config/arm/neon.md (neon_mov): Update type attribute.
4256 (*movmisalign<mode>_neon_store): Likewise.
4257 (*movmisalign<mode>_neon_load): Likewise.
4258 (vec_set<mode>_internal): Likewise.
4259 (vec_set<mode>_internal): Likewise.
4260 (vec_setv2di_internal): Likewise.
4261 (vec_extract<mode>): Likewise.
4262 (vec_extract<mode>): Likewise.
4263 (vec_extractv2di): Likewise.
4264 (*add<mode>3_neon): Likewise.
4265 (adddi3_neon): Likewise.
4266 (*sub<mode>3_neon): Likewise.
4267 (subdi3_neon): Likewise.
4268 (fma<VCVTF:mode>4): Likewise.
4269 (fma<VCVTF:mode>4_intrinsic): Likewise.
4270 (*fmsub<VCVTF:mode>4): Likewise.
4271 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
4272 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
4273 (ior<mode>3): Likewise.
4274 (and<mode>3): Likewise.
4275 (orn<mode>3_neon): Likewise.
4276 (orndi3_neon): Likewise.
4277 (bic<mode>3_neon): Likewise.
4278 (bicdi3_neon): Likewise.
4279 (xor<mode>3): Likewise.
4280 (one_cmpl<mode>2): Likewise.
4281 (abs<mode>2): Likewise.
4282 (neg<mode>2): Likewise.
4283 (negdi2_neon): Likewise.
4284 (*umin<mode>3_neon): Likewise.
4285 (*umax<mode>3_neon): Likewise.
4286 (*smin<mode>3_neon): Likewise.
4287 (*smax<mode>3_neon): Likewise.
4288 (vashl<mode>3): Likewise.
4289 (vashr<mode>3_imm): Likewise.
4290 (vlshr<mode>3_imm): Likewise.
4291 (ashl<mode>3_signed): Likewise.
4292 (ashl<mode>3_unsigned): Likewise.
4293 (neon_load_count): Likewise.
4294 (ashldi3_neon_noclobber): Likewise.
4295 (ashldi3_neon): Likewise.
4296 (signed_shift_di3_neon): Likewise.
4297 (unsigned_shift_di3_neon): Likewise.
4298 (ashrdi3_neon_imm_noclobber): Likewise.
4299 (lshrdi3_neon_imm_noclobber): Likewise.
4300 (<shift>di3_neon): Likewise.
4301 (widen_ssum<mode>3): Likewise.
4302 (widen_usum<mode>3): Likewise.
4303 (quad_halves_<code>v4si): Likewise.
4304 (quad_halves_<code>v4sf): Likewise.
4305 (quad_halves_<code>v8hi): Likewise.
4306 (quad_halves_<code>v16qi): Likewise.
4307 (reduc_splus_v2di): Likewise.
4308 (neon_vpadd_internal<mode>): Likewise.
4309 (neon_vpsmin<mode>): Likewise.
4310 (neon_vpsmax<mode>): Likewise.
4311 (neon_vpumin<mode>): Likewise.
4312 (neon_vpumax<mode>): Likewise.
4313 (*ss_add<mode>_neon): Likewise.
4314 (*us_add<mode>_neon): Likewise.
4315 (*ss_sub<mode>_neon): Likewise.
4316 (*us_sub<mode>_neon): Likewise.
4317 (neon_vadd<mode>_unspec): Likewise.
4318 (neon_vaddl<mode>): Likewise.
4319 (neon_vaddw<mode>): Likewise.
4320 (neon_vhadd<mode>): Likewise.
4321 (neon_vqadd<mode>): Likewise.
4322 (neon_vaddhn<mode>): Likewise.
4323 (neon_vmul<mode>): Likewise.
4324 (neon_vfms<VCVTF:mode>): Likewise.
4325 (neon_vmlal<mode>): Likewise.
4326 (neon_vmls<mode>): Likewise.
4327 (neon_vmlsl<mode>): Likewise.
4328 (neon_vqdmulh<mode>): Likewise.
4329 (neon_vqdmlal<mode>): Likewise.
4330 (neon_vqdmlsl<mode>): Likewise.
4331 (neon_vmull<mode>): Likewise.
4332 (neon_vqdmull<mode>): Likewise.
4333 (neon_vsub<mode>_unspec): Likewise.
4334 (neon_vsubl<mode>): Likewise.
4335 (neon_vsubw<mode>): Likewise.
4336 (neon_vqsub<mode>): Likewise.
4337 (neon_vhsub<mode>): Likewise.
4338 (neon_vsubhn<mode>): Likewise.
4339 (neon_vceq<mode>): Likewise.
4340 (neon_vcge<mode>): Likewise.
4341 (neon_vcgeu<mode>): Likewise.
4342 (neon_vcgt<mode>): Likewise.
4343 (neon_vcgtu<mode>): Likewise.
4344 (neon_vcle<mode>): Likewise.
4345 (neon_vclt<mode>): Likewise.
4346 (neon_vcage<mode>): Likewise.
4347 (neon_vcagt<mode>): Likewise.
4348 (neon_vtst<mode>): Likewise.
4349 (neon_vabd<mode>): Likewise.
4350 (neon_vabdl<mode>): Likewise.
4351 (neon_vaba<mode>): Likewise.
4352 (neon_vabal<mode>): Likewise.
4353 (neon_vmax<mode>): Likewise.
4354 (neon_vmin<mode>): Likewise.
4355 (neon_vpaddl<mode>): Likewise.
4356 (neon_vpadal<mode>): Likewise.
4357 (neon_vpmax<mode>): Likewise.
4358 (neon_vpmin<mode>): Likewise.
4359 (neon_vrecps<mode>): Likewise.
4360 (neon_vrsqrts<mode>): Likewise.
4361 (neon_vqabs<mode>): Likewise.
4362 (neon_vqneg<mode>): Likewise.
4363 (neon_vcls<mode>): Likewise.
4364 (clz<mode>2): Likewise.
4365 (popcount<mode>2): Likewise.
4366 (neon_vrecpe<mode>): Likewise.
4367 (neon_vrsqrte<mode>): Likewise.
4368 (neon_vget_lane<mode>_sext_internal): Likewise.
4369 (neon_vget_lane<mode>_zext_internal): Likewise.
4370 (neon_vdup_n<mode>): Likewise.
4371 (neon_vdup_n<mode>): Likewise.
4372 (neon_vdup_nv2di): Likewise.
4373 (neon_vdup_lane<mode>_interal): Likewise.
4374 (*neon_vswp<mode>): Likewise.
4375 (neon_vcombine<mode>): Likewise.
4376 (float<mode><V_cvtto>2): Likewise.
4377 (floatuns<mode><V_cvtto>2): Likewise.
4378 (fix_trunc<mode><V_cvtto>2): Likewise.
4379 (fixuns_trunc<mode><V_cvtto>2
4380 (neon_vcvt<mode>): Likewise.
4381 (neon_vcvt<mode>): Likewise.
4382 (neon_vcvtv4sfv4hf): Likewise.
4383 (neon_vcvtv4hfv4sf): Likewise.
4384 (neon_vcvt_n<mode>): Likewise.
4385 (neon_vcvt_n<mode>): Likewise.
4386 (neon_vmovn<mode>): Likewise.
4387 (neon_vqmovn<mode>): Likewise.
4388 (neon_vqmovun<mode>): Likewise.
4389 (neon_vmovl<mode>): Likewise.
4390 (neon_vmul_lane<mode>): Likewise.
4391 (neon_vmul_lane<mode>): Likewise.
4392 (neon_vmull_lane<mode>): Likewise.
4393 (neon_vqdmull_lane<mode>): Likewise.
4394 (neon_vqdmulh_lane<mode>): Likewise.
4395 (neon_vqdmulh_lane<mode>): Likewise.
4396 (neon_vmla_lane<mode>): Likewise.
4397 (neon_vmla_lane<mode>): Likewise.
4398 (neon_vmlal_lane<mode>): Likewise.
4399 (neon_vqdmlal_lane<mode>): Likewise.
4400 (neon_vmls_lane<mode>): Likewise.
4401 (neon_vmls_lane<mode>): Likewise.
4402 (neon_vmlsl_lane<mode>): Likewise.
4403 (neon_vqdmlsl_lane<mode>): Likewise.
4404 (neon_vext<mode>): Likewise.
4405 (neon_vrev64<mode>): Likewise.
4406 (neon_vrev32<mode>): Likewise.
4407 (neon_vrev16<mode>): Likewise.
4408 (neon_vbsl<mode>_internal): Likewise.
4409 (neon_vshl<mode>): Likewise.
4410 (neon_vqshl<mode>): Likewise.
4411 (neon_vshr_n<mode>): Likewise.
4412 (neon_vshrn_n<mode>): Likewise.
4413 (neon_vqshrn_n<mode>): Likewise.
4414 (neon_vqshrun_n<mode>): Likewise.
4415 (neon_vshl_n<mode>): Likewise.
4416 (neon_vqshl_n<mode>): Likewise.
4417 (neon_vqshlu_n<mode>): Likewise.
4418 (neon_vshll_n<mode>): Likewise.
4419 (neon_vsra_n<mode>): Likewise.
4420 (neon_vsri_n<mode>): Likewise.
4421 (neon_vsli_n<mode>): Likewise.
4422 (neon_vtbl1v8qi): Likewise.
4423 (neon_vtbl2v8qi): Likewise.
4424 (neon_vtbl3v8qi): Likewise.
4425 (neon_vtbl4v8qi): Likewise.
4426 (neon_vtbl1v16qi): Likewise.
4427 (neon_vtbl2v16qi): Likewise.
4428 (neon_vcombinev16qi): Likewise.
4429 (neon_vtbx1v8qi): Likewise.
4430 (neon_vtbx2v8qi): Likewise.
4431 (neon_vtbx3v8qi): Likewise.
4432 (neon_vtbx4v8qi): Likewise.
4433 (*neon_vtrn<mode>_insn): Likewise.
4434 (*neon_vzip<mode>_insn): Likewise.
4435 (*neon_vuzp<mode>_insn): Likewise.
4436 (neon_vld1<mode>): Likewise.
4437 (neon_vld1_lane<mode>): Likewise.
4438 (neon_vld1_lane<mode>): Likewise.
4439 (neon_vld1_dup<mode>): Likewise.
4440 (neon_vld1_dup<mode>): Likewise.
4441 (neon_vld1_dupv2di): Likewise.
4442 (neon_vst1<mode>): Likewise.
4443 (neon_vst1_lane<mode>): Likewise.
4444 (neon_vst1_lane<mode>): Likewise.
4445 (neon_vld2<mode>): Likewise.
4446 (neon_vld2<mode>): Likewise.
4447 (neon_vld2_lane<mode>): Likewise.
4448 (neon_vld2_lane<mode>): Likewise.
4449 (neon_vld2_dup<mode>): Likewise.
4450 (neon_vst2<mode>): Likewise.
4451 (neon_vst2<mode>): Likewise.
4452 (neon_vst2_lane<mode>): Likewise.
4453 (neon_vst2_lane<mode>): Likewise.
4454 (neon_vld3<mode>): Likewise.
4455 (neon_vld3qa<mode>): Likewise.
4456 (neon_vld3qb<mode>): Likewise.
4457 (neon_vld3_lane<mode>): Likewise.
4458 (neon_vld3_lane<mode>): Likewise.
4459 (neon_vld3_dup<mode>): Likewise.
4460 (neon_vst3<mode>): Likewise.
4461 (neon_vst3qa<mode>): Likewise.
4462 (neon_vst3qb<mode>): Likewise.
4463 (neon_vst3_lane<mode>): Likewise.
4464 (neon_vst3_lane<mode>): Likewise.
4465 (neon_vld4<mode>): Likewise.
4466 (neon_vld4qa<mode>): Likewise.
4467 (neon_vld4qb<mode>): Likewise.
4468 (neon_vld4_lane<mode>): Likewise.
4469 (neon_vld4_lane<mode>): Likewise.
4470 (neon_vld4_dup<mode>): Likewise.
4471 (neon_vst4<mode>): Likewise.
4472 (neon_vst4qa<mode>): Likewise.
4473 (neon_vst4qb<mode>): Likewise.
4474 (neon_vst4_lane<mode>): Likewise.
4475 (neon_vst4_lane<mode>): Likewise.
4476 (neon_vec_unpack<US>_lo_<mode>): Likewise.
4477 (neon_vec_unpack<US>_hi_<mode>): Likewise.
4478 (neon_vec_<US>mult_lo_<mode>): Likewise.
4479 (neon_vec_<US>mult_hi_<mode>): Likewise.
4480 (neon_vec_<US>shiftl_<mode>): Likewise.
4481 (neon_unpack<US>_<mode>): Likewise.
4482 (neon_vec_<US>mult_<mode>): Likewise.
4483 (vec_pack_trunc_<mode>): Likewise.
4484 (neon_vec_pack_trunc_<mode>): Likewise.
4485 (neon_vabd<mode>_2): Likewise.
4486 (neon_vabd<mode>_3): Likewise.
4488 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
4490 * config/aarch64/aarch64.md (movtf_aarch64): Update type attribute.
4491 (load_pair): Update type attribute.
4492 (store_pair): Update type attribute.
4493 * config/aarch64/iterators.md (q): New.
4495 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
4497 * config/arm/types.md: Add new types for Neon insns.
4499 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
4500 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4501 Sergey Lega <sergey.s.lega@intel.com>
4502 Anna Tikhonova <anna.tikhonova@intel.com>
4503 Ilya Tocar <ilya.tocar@intel.com>
4504 Andrey Turetskiy <andrey.turetskiy@intel.com>
4505 Ilya Verbin <ilya.verbin@intel.com>
4506 Kirill Yukhin <kirill.yukhin@intel.com>
4507 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4509 * config/i386/sse.md (unspec): Add UNSPEC_RCP14, UNSPEC_RSQRT14,
4510 UNSPEC_FIXUPIMM, UNSPEC_SCALEF, UNSPEC_GETEXP, UNSPEC_GETMANT,
4511 UNSPEC_EXP2, UNSPEC_RCP28, UNSPEC_RSQRT28.
4513 (srcp14<mode>): Ditto.
4514 (rsqrt14<mode>): Ditto.
4515 (rsqrt14<mode>): Ditto.
4516 (avx512f_vmscalef<mode>): Ditto.
4517 (avx512f_scalef<mode>): Ditto.
4518 (avx512f_getexp<mode>): Ditto.
4519 (avx512f_sgetexp<mode>): Ditto.
4520 (avx512f_fixupimm<mode>): Ditto.
4521 (avx512f_sfixupimm<mode>): Ditto.
4522 (avx512f_rndscale<mode>): Ditto.
4523 (*avx512er_exp2<mode>): Ditto.
4524 (*avx512er_rcp28<mode>): Ditto.
4525 (avx512er_rsqrt28<mode>): Ditto.
4526 (avx512f_getmant<mode>): Ditto.
4527 (avx512f_getmant<mode>): Ditto.
4528 (avx512f_rndscale<mode>): Fix formatting.
4530 2013-10-15 Martin Jambor <mjambor@suse.cz>
4532 * ipa-utils.h (ipa_edge_within_scc): Declare.
4533 * ipa-cp.c (edge_within_scc): Moved...
4534 * ipa-utils.c (ipa_edge_within_scc): ...here. Updated all callers.
4536 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
4537 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4538 Sergey Lega <sergey.s.lega@intel.com>
4539 Anna Tikhonova <anna.tikhonova@intel.com>
4540 Ilya Tocar <ilya.tocar@intel.com>
4541 Andrey Turetskiy <andrey.turetskiy@intel.com>
4542 Ilya Verbin <ilya.verbin@intel.com>
4543 Kirill Yukhin <kirill.yukhin@intel.com>
4544 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4546 * config/i386/predicates.md (const_8_to_15_operand): New.
4547 (const_16_to_31_operand): Ditto.
4548 * config/i386/sse.md (V8FI): New.
4550 (reduc_splus_v8df): Ditto.
4551 (reduc_splus_v16sf): Ditto.
4552 (avx512f_vextract<shuffletype>32x4_1): Ditto.
4553 (vec_extract_hi_<mode>): Ditto.
4554 (avx512f_vinsert<shuffletype>32x4_1): Ditto.
4555 (vec_set_lo_<mode>): Ditto.
4556 (vec_set_hi_<mode>): Ditto.
4557 (avx512f_shuf_<shuffletype>64x2_1): Ditto.
4558 (avx512f_shuf_<shuffletype>32x4_1): Ditto.
4559 (avx512f_pshufd_1): Ditto.
4560 (avx512f_broadcast<mode>): Ditto.
4561 (avx512f_broadcast<mode>): Ditto.
4562 (define_split): Split vec_extract_lo into move.
4563 (ssequartermode): Ditto.
4564 (ssedoublemode): Extened with wider modes.
4565 (vec_extract_lo_<mode>): Ditto.
4567 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
4568 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4569 Sergey Lega <sergey.s.lega@intel.com>
4570 Anna Tikhonova <anna.tikhonova@intel.com>
4571 Ilya Tocar <ilya.tocar@intel.com>
4572 Andrey Turetskiy <andrey.turetskiy@intel.com>
4573 Ilya Verbin <ilya.verbin@intel.com>
4574 Kirill Yukhin <kirill.yukhin@intel.com>
4575 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4577 * config/i386/predicates.md (register_or_constm1_operand): New.
4578 * config/i386/sse.md (unspec): Add UNSPEC_UNSIGNED_PCMP, UNSPEC_TESTM,
4579 UNSPEC_TESTNM, UNSPEC_VTERNLOG, UNSPEC_ALIGN, UNSPEC_CONFLICT,
4580 UNSPEC_MASKED_EQ, UNSPEC_MASKED_GT, UNSPEC_GATHER_PREFETCH,
4581 UNSPEC_SCATTER_PREFETCH
4583 (avx512f_ucmp<mode>3): Ditto.
4584 (avx512f_vternlog<mode>): Ditto.
4585 (avx512f_align<mode>): Ditto.
4586 (<shift_insn><mode>3): Ditto.
4587 (avx512f_<rotate>v<mode>): Ditto.
4588 (avx512f_<rotate><mode>): Ditto.
4589 (avx512f_eq<mode>3): Ditto.
4590 (avx512f_eq<mode>3_1): Ditto.
4591 (avx512f_gt<mode>3): Ditto.
4592 (avx512f_testm<mode>3): Ditto.
4593 (avx512f_testnm<mode>3): Ditto.
4594 (avx512pf_gatherpf<mode>): Ditto.
4595 (*avx512pf_gatherpf<mode>_mask): Ditto.
4596 (*avx512pf_gatherpf<mode>): Ditto.
4597 (avx512pf_scatterpf<mode>): Ditto.
4598 (*avx512pf_scatterpf<mode>_mask): Ditto.
4599 (*avx512pf_scatterpf<mode>): Ditto.
4600 (avx512f_vec_dup_gpr<mode>): Ditto.
4601 (clz<mode>2): Ditto.
4602 (conflict<mode>): Ditto.
4603 (REDUC_SMINMAX_MODE): Extened with wider modes.
4604 (reduc_<code>_<mode>): Ditto.
4605 (vlshr<mode>3): Ditto.
4606 (vashl<mode>3): Ditto.
4608 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
4609 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4610 Sergey Lega <sergey.s.lega@intel.com>
4611 Anna Tikhonova <anna.tikhonova@intel.com>
4612 Ilya Tocar <ilya.tocar@intel.com>
4613 Andrey Turetskiy <andrey.turetskiy@intel.com>
4614 Ilya Verbin <ilya.verbin@intel.com>
4615 Kirill Yukhin <kirill.yukhin@intel.com>
4616 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4618 * config/i386/sse.md (unspec): Added UNSPEC_VPERMI2, UNSPEC_VPERMT2,
4621 (avx512fmaskmode): Ditto.
4622 (bcstscalarsuff): Ditto.
4623 (avx512f_blendm<mode>): Ditto.
4624 (cmp_imm_predicate): Ditto.
4625 (avx512f_cmp<mode>3): Ditto.
4626 (avx512f_vec_dup<mode>): Ditto.
4627 (avx512f_vec_dup_mem<mode>): Ditto.
4628 (avx512f_vpermi2var<mode>3): Ditto.
4629 (avx512f_vpermt2var<mode>3): Ditto.
4630 (vec_init<mode>): Ditto.
4631 (avx512f_gathersi<mode>): Ditto.
4632 (*avx512f_gathersi<mode>): Ditto.
4633 (*avx512f_gathersi<mode>_2): Ditto.
4634 (avx512f_gatherdi<mode>): Ditto.
4635 (*avx512f_gatherdi<mode>): Ditto.
4636 (*avx512f_gatherdi<mode>_2): Ditto.
4637 (avx512f_scattersi<mode>): Ditto.
4638 (*avx512f_scattersi<mode>): Ditto.
4639 (avx512f_scatterdi<mode>): Ditto.
4640 (*avx512f_scatterdi<mode>): Ditto.
4641 (sseintprefix): Extened with wider modes.
4642 (VEC_GATHER_IDXSI): Ditto.
4643 (VEC_GATHER_IDXDI): Ditto.
4644 (VEC_GATHER_SRCDI): Ditto.
4646 2013-10-15 Matthew Gretton-Dann <matthew.gretton-dann@arm.com>
4647 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4649 * config/arm/t-aprofile: New file.
4650 * config.gcc: Handle --with-multilib-list option.
4652 2013-10-15 Bernd Schmidt <bernds@codesourcery.com>
4654 * reload1.c (reloads_unique_chain_p): Ensure that r1 is
4657 2013-10-15 Richard Biener <rguenther@suse.de>
4659 * tree-loop-distribution.c (build_empty_rdg): Inline into
4661 (rdg_flag_vertex): Inline into single user.
4662 (rdg_flag_vertex_and_dependent): Likewise.
4663 (build_rdg_partition_for_vertex): Remove processed bitmap.
4664 (rdg_build_partitions): Simplify.
4666 2013-10-15 Richard Biener <rguenther@suse.de>
4668 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1):
4669 Restructure forwarding through conversions and copies to
4670 avoid performing copy-propagation the wrong way. Adjust
4671 recursion invocations.
4672 (forward_propagate_addr_expr): Add argument stating if we
4673 are recursing from a single-use.
4674 (ssa_forward_propagate_and_combine): Adjust.
4676 2013-10-14 David Malcolm <dmalcolm@redhat.com>
4678 * dumpfile.h (gcc::dump_manager): New class, to hold state
4679 relating to dumpfile management.
4680 (get_dump_file_name): Remove in favor of method of dump_manager.
4681 (dump_initialized_p): Likewise.
4682 (dump_start): Likewise.
4683 (dump_finish): Likewise.
4684 (dump_switch_p): Likewise.
4685 (dump_register): Likewise.
4686 (get_dump_file_info): Likewise.
4687 * context.c (gcc::context::context): Construct the dump_manager
4689 * context.h (gcc::context::get_dumps): New.
4690 (gcc::context::m_dumps): New.
4691 * coverage.c (coverage_init): Port to dump_manager API.
4692 * dumpfile.c (extra_dump_files): Convert to field of gcc::dump_manager.
4693 (extra_dump_files_in_use): Likewise.
4694 (extra_dump_files_alloced): Likewise.
4695 (gcc::dump_manager::dump_manager): New.
4696 (dump_register): Convert to...
4697 (gcc::dump_manager::dump_register): ...method, replacing
4698 function-static next_dump with m_next_dump field.
4699 (get_dump_file_info): Convert to...
4700 (gcc::dump_manager::get_dump_file_info): ...method.
4701 (get_dump_file_name): Convert to...
4702 (gcc::dump_manager::get_dump_file_name): ...method.
4703 (dump_start): Convert to...
4704 (gcc::dump_manager::dump_start): ...method.
4705 (dump_finish): Convert to...
4706 (gcc::dump_manager::dump_finish): ...method.
4707 (dump_begin): Replace body with...
4708 (gcc::dump_manager::dump_begin): ...new method.
4709 (dump_phase_enabled_p): Convert to...
4710 (gcc::dump_manager::dump_phase_enabled_p): ...method.
4711 (dump_phase_enabled_p): Convert to...
4712 (gcc::dump_manager::dump_phase_enabled_p): ...method.
4713 (dump_initialized_p): Convert to...
4714 (gcc::dump_manager::dump_initialized_p): ...method.
4715 (dump_flag_name): Replace body with...
4716 (gcc::dump_manager::dump_flag_name): ...new method.
4717 (dump_enable_all): Convert to...
4718 (gcc::dump_manager::dump_enable_all): ...new method.
4719 (opt_info_enable_passes): Convert to...
4720 (gcc::dump_manager::opt_info_enable_passes): ...new method.
4721 (dump_switch_p_1): Convert to...
4722 (gcc::dump_manager::dump_switch_p_1): ...new method.
4723 (dump_switch_p): Convert to...
4724 (gcc::dump_manager::dump_switch_p): ...new method.
4725 (opt_info_switch_p): Port to dump_manager API.
4726 (enable_rtl_dump_file): Likewise.
4727 * opts-global.c (handle_common_deferred_options): Port to new
4729 * passes.c (pass_manager::finish_optimization_passes): Likewise.
4730 (pass_manager::register_one_dump_file): Likewise.
4731 (pass_manager::register_pass): Likewise.
4732 (pass_init_dump_file): Likewise.
4733 (pass_fini_dump_file): Likewise.
4734 * statistics.c (statistics_early_init): Likewise.
4736 2013-10-14 Richard Biener <rguenther@suse.de>
4738 * gimple.c (gimple_canonical_types, canonical_type_hash_cache,
4739 iterative_hash_canonical_type, gimple_canonical_type_hash,
4740 gimple_canonical_types_compatible_p, gimple_canonical_type_eq,
4741 gimple_register_canonical_type, print_gimple_types_stats,
4742 free_gimple_type_tables): Move to lto/lto.c
4743 (gt-gimple.h): Do not include.
4744 * gimple.h (gimple_register_canonical_type,
4745 print_gimple_types_stats, free_gimple_type_tables): Remove.
4746 * Makefile.in (GTFILES): Remove gimple.c.
4748 2013-10-14 Travis Snoozy <quandary@remstate.com>
4751 * config/msp430/msp430.c (msp430_option_override): Correct thinko
4752 scanning for msp430x targets.
4754 2013-10-14 Eric Botcazou <ebotcazou@adacore.com>
4757 * config/sparc/sparc-protos.h (widen_mem_for_ldd_peep): Declare.
4758 (registers_ok_for_ldd_peep): Move around.
4759 * config/sparc/sparc.c (widen_mem_for_ldd_peep): New.
4760 * config/sparc/sparc.md (widening peepholes): Use it.
4762 2013-10-14 Richard Biener <rguenther@suse.de>
4766 * gimple.c (iterative_hash_canonical_type): Make sure to
4767 record the hash into the correct hashtable slot.
4769 2013-10-13 Eric Botcazou <ebotcazou@adacore.com>
4771 PR rtl-optimization/58662
4772 * combine.c (try_combine): Take into account death nodes on I2 when
4773 splitting a PARALLEL of two independent SETs. Fix dump message.
4775 2013-10-12 Oleg Endo <olegendo@gcc.gnu.org>
4778 * config/sh/sh_treg_combine.cc: New SH specific RTL pass.
4779 * config.gcc (SH extra_objs): Add sh_ifcvt.o.
4780 * config/sh/t-sh (sh_treg_combine.o): New entry.
4781 * config/sh/sh.c (sh_fixed_condition_code_regs): New function that
4782 implements the target hook TARGET_FIXED_CONDITION_CODE_REGS.
4783 (register_sh_passes): New function. Register sh_treg_combine pass.
4784 (sh_option_override): Invoke it.
4785 (sh_canonicalize_comparison): Handle op0_preserve_value.
4786 * sh.md (*cbranch_t"): Do not try to optimize missed test and branch
4787 opportunities. Canonicalize branch condition.
4788 (nott): Allow only if pseudos can be created for non-SH2A.
4790 2013-10-12 H.J. Lu <hongjiu.lu@intel.com>
4793 * config/i386/i386.c (ix86_copy_addr_to_reg): New function.
4794 (ix86_expand_movmem): Replace copy_addr_to_reg with
4795 ix86_copy_addr_to_reg.
4796 (ix86_expand_setmem): Likewise.
4798 2013-10-12 Alexander Monakov <amonakov@ispras.ru>
4800 * config/i386/i386.c (ix86_expand_sse_compare_and_jump): Use mode
4801 provided by ix86_fp_compare_mode instead of CCFPUmode.
4803 2013-10-12 James Greenhalgh <james.greenhalgh@arm.com>
4805 * config/aarch64/arm_neon.h
4806 (vtbx<1,3>_<psu>8): Fix register constriants.
4808 2013-10-11 Jeff Law <law@redhat.com>
4810 PR tree-optimization/58640
4811 * tree-ssa-threadupdate.c (mark_threaded_blocks): Truncate jump
4812 threading paths that cross over two loop entry points.
4814 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4816 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): Generalize to
4817 handle vector float as well.
4818 (*vsx_le_perm_load_v4si): Likewise.
4819 (*vsx_le_perm_store_v2di): Likewise.
4820 (*vsx_le_perm_store_v4si): Likewise.
4822 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4824 * config/rs6000/vector.md (vec_realign_load<mode>): Generate vperm
4825 directly to circumvent subtract from splat{31} workaround.
4826 * config/rs6000/rs6000-protos.h (altivec_expand_vec_perm_le): New
4828 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): New.
4829 * config/rs6000/altivec.md (define_c_enum "unspec"): Add
4830 UNSPEC_VPERM_X and UNSPEC_VPERM_UNS_X.
4831 (altivec_vperm_<mode>): Convert to define_insn_and_split to
4832 separate big and little endian logic.
4833 (*altivec_vperm_<mode>_internal): New define_insn.
4834 (altivec_vperm_<mode>_uns): Convert to define_insn_and_split to
4835 separate big and little endian logic.
4836 (*altivec_vperm_<mode>_uns_internal): New define_insn.
4837 (vec_permv16qi): Add little endian logic.
4839 2013-10-11 Marc Glisse <marc.glisse@inria.fr>
4841 * doc/extend.texi (returns_nonnull): Remove arguments.
4843 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4844 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4845 Sergey Lega <sergey.s.lega@intel.com>
4846 Anna Tikhonova <anna.tikhonova@intel.com>
4847 Ilya Tocar <ilya.tocar@intel.com>
4848 Andrey Turetskiy <andrey.turetskiy@intel.com>
4849 Ilya Verbin <ilya.verbin@intel.com>
4850 Kirill Yukhin <kirill.yukhin@intel.com>
4851 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4853 * config/i386/sse.md (VI48F_256_512): New.
4854 (avx2_permvar<mode>): Change to ...
4855 (<avx2_avx512f>_permvar<mode>): This.
4857 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4858 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4859 Sergey Lega <sergey.s.lega@intel.com>
4860 Anna Tikhonova <anna.tikhonova@intel.com>
4861 Ilya Tocar <ilya.tocar@intel.com>
4862 Andrey Turetskiy <andrey.turetskiy@intel.com>
4863 Ilya Verbin <ilya.verbin@intel.com>
4864 Kirill Yukhin <kirill.yukhin@intel.com>
4865 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4867 * config/i386/i386.c (bdesc_args): Change corresponding pattern for
4868 __builtin_ia32_cvtps2dq, __builtin_ia32_cvtps2dq256.
4869 * config/i386/sse.md (VI4_AVX): New.
4870 (sf2simodelower): Ditto.
4871 (sse2_cvtps2dq): Change to ...
4872 (<sse2_avx_avx512f>_fix_notrunc<sf2simodelower><mode>): This.
4874 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4875 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4876 Sergey Lega <sergey.s.lega@intel.com>
4877 Anna Tikhonova <anna.tikhonova@intel.com>
4878 Ilya Tocar <ilya.tocar@intel.com>
4879 Andrey Turetskiy <andrey.turetskiy@intel.com>
4880 Ilya Verbin <ilya.verbin@intel.com>
4881 Kirill Yukhin <kirill.yukhin@intel.com>
4882 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4884 * config/i386/sse.md (V_512): New.
4886 (vcond<V_512:mode><VF_512:mode>): Ditto.
4887 (vcond<V_512:mode><VI_512:mode>): Ditto.
4888 (vcondu<V_512:mode><VI_512:mode>): Ditto.
4890 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4891 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4892 Sergey Lega <sergey.s.lega@intel.com>
4893 Anna Tikhonova <anna.tikhonova@intel.com>
4894 Ilya Tocar <ilya.tocar@intel.com>
4895 Andrey Turetskiy <andrey.turetskiy@intel.com>
4896 Ilya Verbin <ilya.verbin@intel.com>
4897 Kirill Yukhin <kirill.yukhin@intel.com>
4898 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4900 * config/i386/i386.c (ix86_rtx_costs): Enable fma for TARGET_AVX512F.
4901 * config/i386/sse.md (FMAMODEM): Changed modes and conditions.
4903 (fma<mode>4): Removed condition.
4904 (fms<mode>4): Ditto.
4905 (fnma<mode>4): Ditto.
4906 (fnms<mode>4): Ditto.
4907 (fma4i_fmadd_<mode>): Ditto.
4908 (*fma_fmadd_<mode>): Ditto.
4909 (*fma_fmsub_<mode>): Ditto.
4910 (*fma_fnmadd_<mode>): Ditto.
4911 (*fma_fnmsub_<mode>): Ditto.
4912 (fmaddsub_<mode>): Allow for TARGET_AVX512F.
4913 (*fma_fmaddsub_<mode>): Ditto.
4914 (*fma_fmsubadd_<mode>): Ditto.
4915 (*fmai_fmadd_<mode>): Ditto.
4916 (*fmai_fmsub_<mode>): Ditto.
4917 (*fmai_fnmadd_<mode>): Ditto.
4918 (*fmai_fnmsub_<mode>): Ditto.
4920 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4921 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4922 Sergey Lega <sergey.s.lega@intel.com>
4923 Anna Tikhonova <anna.tikhonova@intel.com>
4924 Ilya Tocar <ilya.tocar@intel.com>
4925 Andrey Turetskiy <andrey.turetskiy@intel.com>
4926 Ilya Verbin <ilya.verbin@intel.com>
4927 Kirill Yukhin <kirill.yukhin@intel.com>
4928 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4930 * config/i386/sse.md (VI248_AVX2_8_AVX512F): New.
4931 (VI124_256): Changed to ...
4932 (VI124_256_48_512): This.
4933 (ssepackmode): Extended with wider modes.
4934 (<code><mode>3): Changed iterator.
4935 (*avx2_<code><mode>3): Ditto.
4936 (vec_pack_trunc_<mode>): Ditto.
4938 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4939 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4940 Sergey Lega <sergey.s.lega@intel.com>
4941 Anna Tikhonova <anna.tikhonova@intel.com>
4942 Ilya Tocar <ilya.tocar@intel.com>
4943 Andrey Turetskiy <andrey.turetskiy@intel.com>
4944 Ilya Verbin <ilya.verbin@intel.com>
4945 Kirill Yukhin <kirill.yukhin@intel.com>
4946 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4948 * config/i386/sse.md (VI124_AVX2_48_AVX512F): New.
4949 (VI8F_256_512): Ditto.
4950 (abs<mode>2): Changed iterator.
4951 (avx2_perm<mode>): Changed to ...
4952 (<avx2_avx512f>_perm<mode>): This.
4953 (avx2_perm<mode>_1): Changed to ...
4954 (<avx2_avx512f>_perm<mode>_1): This.
4956 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4957 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4958 Sergey Lega <sergey.s.lega@intel.com>
4959 Anna Tikhonova <anna.tikhonova@intel.com>
4960 Ilya Tocar <ilya.tocar@intel.com>
4961 Andrey Turetskiy <andrey.turetskiy@intel.com>
4962 Ilya Verbin <ilya.verbin@intel.com>
4963 Kirill Yukhin <kirill.yukhin@intel.com>
4964 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4966 * config/i386/sse.md (VI48_AVX512F): New.
4967 (VI48_AVX2): Changed to ...
4968 (VI48_AVX2_48_AVX512F): This.
4969 (avx2_ashrv<mode>): Changed to ...
4970 (<avx2_avx512f>_ashrv<mode>): This.
4971 (avx2_<shift_insn>v<mode>): Changed to ...
4972 (<avx2_avx512f>_<shift_insn>v<mode>): This.
4974 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4975 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4976 Sergey Lega <sergey.s.lega@intel.com>
4977 Anna Tikhonova <anna.tikhonova@intel.com>
4978 Ilya Tocar <ilya.tocar@intel.com>
4979 Andrey Turetskiy <andrey.turetskiy@intel.com>
4980 Ilya Verbin <ilya.verbin@intel.com>
4981 Kirill Yukhin <kirill.yukhin@intel.com>
4982 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4984 * config/i386/sse.md (VI4_AVX512F): New.
4985 (VI8_AVX2_AVX512F): Ditto.
4986 (mul<mode>3): Extended with wider modes.
4987 (*<sse4_1_avx2>_mul<mode>3): Ditto.
4988 (mul<mode>3): Ditto.
4989 (vec_widen_<s>mult_odd_<mode>): Ditto.
4991 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4992 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4993 Sergey Lega <sergey.s.lega@intel.com>
4994 Anna Tikhonova <anna.tikhonova@intel.com>
4995 Ilya Tocar <ilya.tocar@intel.com>
4996 Andrey Turetskiy <andrey.turetskiy@intel.com>
4997 Ilya Verbin <ilya.verbin@intel.com>
4998 Kirill Yukhin <kirill.yukhin@intel.com>
4999 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5001 * config/i386/sse.md (VI2_AVX512F): New.
5002 (VI124_AVX512F): Ditto.
5003 (sseunpackmode): Extended with wider modes.
5004 (sseunpackfltmode): Ditto.
5005 (vec_unpacks_float_hi_<mode>): Ditto.
5006 (vec_unpacks_float_lo_<mode>): Ditto.
5007 (vec_unpacku_float_hi_<mode>): Ditto.
5008 (vec_unpacku_float_lo_<mode>): Ditto.
5009 (vec_unpacks_lo_<mode>): Ditto.
5010 (vec_unpacks_hi_<mode>): Ditto.
5011 (vec_unpacku_lo_<mode>): Ditto.
5012 (vec_unpacku_hi_<mode>): Ditto.
5014 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5015 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5016 Sergey Lega <sergey.s.lega@intel.com>
5017 Anna Tikhonova <anna.tikhonova@intel.com>
5018 Ilya Tocar <ilya.tocar@intel.com>
5019 Andrey Turetskiy <andrey.turetskiy@intel.com>
5020 Ilya Verbin <ilya.verbin@intel.com>
5021 Kirill Yukhin <kirill.yukhin@intel.com>
5022 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5024 * config/i386/i386.md (multdiv): New.
5025 (multdiv_mnemonic): Ditto.
5026 * config/i386/sse.md (<sse>_vmmul<mode>3): Changed to...
5027 (<sse>_vm<multdiv_mnemonic><mode>3): This.
5028 (<sse>_vmdiv<mode>3): Removed.
5030 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5031 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5032 Sergey Lega <sergey.s.lega@intel.com>
5033 Anna Tikhonova <anna.tikhonova@intel.com>
5034 Ilya Tocar <ilya.tocar@intel.com>
5035 Andrey Turetskiy <andrey.turetskiy@intel.com>
5036 Ilya Verbin <ilya.verbin@intel.com>
5037 Kirill Yukhin <kirill.yukhin@intel.com>
5038 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5040 * config/i386/sse.md (V): Extended with wider modes.
5042 (ssehalfvecmode): Ditto.
5044 (ssepackfltmode): Ditto.
5045 (avx_vec_concat<mode>): Ditto.
5046 (V_256_512): New iterator.
5047 (VF2_512_256): Ditto.
5048 (si2dfmode): New attribute.
5049 (si2dfmodelower): Ditto.
5051 (concat_tg_mode): Ditto.
5052 (floatv4siv4df2): Changed to ...
5053 (float<si2dfmodelower><mode>2): This.
5054 (avx_cvtps2pd256): Changed to ...
5055 (<sse2_avx_avx512f>_cvtps2pd<avxsizesuffix>): This.
5056 (vec_pack_trunc_v4df): Changed to ...
5057 (vec_pack_trunc_<mode>): This.
5058 (avx_vpermil<mode>): Changed to ...
5059 (<sse2_avx_avx512f>_vpermil<mode>): This.
5060 (<fixsuffix>fix_truncv8dfv8si2): New.
5061 (vec_pack_sfix_trunc_v8df): Ditto.
5062 (avx512f_rndscale<mode>): Ditto.
5063 (avx512f_roundpd512): Ditto.
5064 (vec_pack_ufix_trunc_<mode>): Updated iterator.
5066 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5067 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5068 Sergey Lega <sergey.s.lega@intel.com>
5069 Anna Tikhonova <anna.tikhonova@intel.com>
5070 Ilya Tocar <ilya.tocar@intel.com>
5071 Andrey Turetskiy <andrey.turetskiy@intel.com>
5072 Ilya Verbin <ilya.verbin@intel.com>
5073 Kirill Yukhin <kirill.yukhin@intel.com>
5074 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5076 * config/i386/i386.md (any_fix): New iterator.
5077 (fixsuffix): New attribute.
5078 * config/i386/sse.md (VF1): Extened with wider modes.
5082 (sseintvecmodelower): Ditto.
5083 (ssescalarmode): Ditto.
5084 (ssescalarnum): Ditto.
5086 (ssexmmmode): Ditto.
5087 (<fixsuffix>fix_truncv16sfv16si2): Ditto.
5088 (<sse>_rcp<mode>2): Change iterator.
5089 (rsqrt<mode>2): Ditto.
5090 (<sse>_rsqrt<mode>2): Ditto.
5091 (avx2_vec_dup<mode>): Ditto.
5092 (<sse4_1>_round<ssemodesuffix>_sfix<avxsizesuffix>): Ditto.
5093 (round<mode>2_sfix): Ditto.
5094 (avx2_pbroadcast<mode>): Ditto.
5095 (*andnot<mode>3): Handle XI mode.
5096 (*<code><mode>3): Ditto.
5097 (AVXTOSSEMODE): Removed.
5098 (avx_vpermil<mode>): Changed to ...
5099 (<sse2_avx_avx512f>_vpermil<mode>): This.
5101 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5102 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5103 Sergey Lega <sergey.s.lega@intel.com>
5104 Anna Tikhonova <anna.tikhonova@intel.com>
5105 Ilya Tocar <ilya.tocar@intel.com>
5106 Andrey Turetskiy <andrey.turetskiy@intel.com>
5107 Ilya Verbin <ilya.verbin@intel.com>
5108 Kirill Yukhin <kirill.yukhin@intel.com>
5109 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5111 * config/i386/sse.md (<sse2>_movnt<mode>): Update constraint to "v".
5112 (<sse>_comi): Ditto.
5113 (<sse>_ucomi): Ditto.
5114 (sse_cvtss2siq_2): Ditto.
5115 (sse2_cvtsd2si): Ditto.
5116 (sse2_cvtsd2siq): Ditto.
5117 (sse2_cvttsd2si): Ditto.
5118 (sse2_cvttsd2siq): Ditto.
5119 (<shift_insn><mode>3): Ditto.
5120 (sse2_cvtsi2sdq): Update constraint and prefix.
5121 (sse_cvtsi2ss): Update prefix.
5122 (sse_cvtsi2ssq): Ditto.
5124 2013-10-11 Jakub Jelinek <jakub@redhat.com>
5126 * tree-vrp.c (infer_nonnull_range): Use is_gimple_call,
5127 ignore internal calls.
5129 2013-10-11 Richard Biener <rguenther@suse.de>
5131 * tree-pretty-print.c (dump_generic_node): Allow to dump both (D)
5132 and (ab) for SSA_NAMEs. Mark INTEGER_CSTs with (OVF) if
5133 TREE_OVERFLOW is set.
5135 2013-10-11 Thomas Schwinge <thomas@codesourcery.com>
5137 * tree.h (OMP_CLAUSE_CODE): Remove duplicate definition.
5139 * gimple.c: GIMPLE statements have subcodes, not sub-codes.
5140 * gimple.h: Likewise.
5142 * doc/generic.texi (OpenMP): OMP_CLAUSE_* are subcodes, not sub-codes.
5144 * doc/generic.texi (Adding new DECL node types): Explain *_CHECK
5147 * doc/gimple.texi (is_gimple_omp): Move into the correct section.
5149 * acinclude.m4 (gcc_GAS_FLAGS): Add more gcc_cv_as_flags overrides.
5150 * configure: Regenerate.
5152 2013-10-11 Jakub Jelinek <jakub@redhat.com>
5154 * tree-pretty-print.c (dump_omp_clause): Handle OMP_CLAUSE__LOOPTEMP_
5155 and new OpenMP 4.0 clauses, handle UDR OMP_CLAUSE_REDUCTION,
5156 formatting fixes, use pp_colon instead of pp_character (..., ':'),
5157 similarly pp_right_paren.
5158 (dump_generic_node): Handle OMP_DISTRIBUTE, OMP_TEAMS,
5159 OMP_TARGET_DATA, OMP_TARGET, OMP_TARGET_UPDATE, OMP_TASKGROUP,
5160 allow OMP_FOR_INIT to be NULL, handle OMP_ATOMIC_SEQ_CST.
5161 * tree.c (omp_clause_num_ops, omp_clause_code_name): Add OpenMP 4.0
5163 (omp_declare_simd_clauses_equal,
5164 omp_remove_redundant_declare_simd_attrs): New functions.
5165 (attribute_value_equal): Use omp_declare_simd_clauses_equal.
5166 (walk_tree_1): Handle new OpenMP 4.0 clauses.
5167 * tree.h (OMP_LOOP_CHECK): Define.
5168 (OMP_FOR_BODY, OMP_FOR_CLAUSES, OMP_FOR_INIT, OMP_FOR_COND,
5169 OMP_FOR_INCR, OMP_FOR_PRE_BODY): Use it.
5170 (OMP_TASKGROUP_BODY, OMP_TEAMS_BODY, OMP_TEAMS_CLAUSES,
5171 OMP_TARGET_DATA_BODY, OMP_TARGET_DATA_CLAUSES, OMP_TARGET_BODY,
5172 OMP_TARGET_CLAUSES, OMP_TARGET_UPDATE_CLAUSES, OMP_CLAUSE_SIZE,
5173 OMP_ATOMIC_SEQ_CST, OMP_CLAUSE_DEPEND_KIND, OMP_CLAUSE_MAP_KIND,
5174 OMP_CLAUSE_MAP_ZERO_BIAS_ARRAY_SECTION, OMP_CLAUSE_PROC_BIND_KIND,
5175 OMP_CLAUSE_REDUCTION_OMP_ORIG_REF, OMP_CLAUSE_ALIGNED_ALIGNMENT,
5176 OMP_CLAUSE_NUM_TEAMS_EXPR, OMP_CLAUSE_THREAD_LIMIT_EXPR,
5177 OMP_CLAUSE_DEVICE_ID, OMP_CLAUSE_DIST_SCHEDULE_CHUNK_EXPR,
5178 OMP_CLAUSE_SIMDLEN_EXPR): Define.
5179 (OMP_CLAUSE_DECL): Change range up to OMP_CLAUSE__LOOPTEMP_.
5180 (omp_remove_redundant_declare_simd_attrs): New prototype.
5181 * gimple.def (GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET,
5182 GIMPLE_OMP_TEAMS): New codes.
5183 (GIMPLE_OMP_RETURN): Use GSS_OMP_ATOMIC_STORE instead of GSS_BASE.
5184 * omp-low.c (struct omp_context): Add cancel_label and cancellable
5186 (target_nesting_level): New variable.
5187 (extract_omp_for_data): Handle GF_OMP_FOR_KIND_DISTRIBUTE and
5188 OMP_CLAUSE_DIST_SCHEDULE. Don't fallback to library implementation
5189 for collapse > 1 static schedule unless ordered.
5190 (get_ws_args_for): Add par_stmt argument. Handle combined loops.
5191 (determine_parallel_type): Adjust get_ws_args_for caller.
5192 (install_var_field): Handle mask & 4 for double indirection.
5193 (scan_sharing_clauses): Ignore shared clause on teams construct.
5194 Handle OMP_CLAUSE__LOOPTEMP_ and new OpenMP 4.0 clauses.
5195 (create_omp_child_function): If inside target or declare target
5196 constructs, set "omp declare target" attribute on the child function.
5197 (find_combined_for): New function.
5198 (scan_omp_parallel): Handle combined loops.
5199 (scan_omp_target, scan_omp_teams): New functions.
5200 (check_omp_nesting_restrictions): Check new OpenMP 4.0 nesting
5201 restrictions and set ctx->cancellable for cancellable constructs.
5202 (scan_omp_1_stmt): Call check_omp_nesting_restrictions also on
5203 selected builtin calls. Handle GIMPLE_OMP_TASKGROUP,
5204 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS.
5205 (build_omp_barrier): Add lhs argument, return gimple rather than tree.
5206 (omp_clause_aligned_alignment): New function.
5207 (lower_rec_simd_input_clauses): Only call SET_DECL_VALUE_EXPR on decls.
5208 (lower_rec_input_clauses): Add FD argument. Ignore shared clauses
5209 on teams constructs. Handle user defined reductions and new
5211 (lower_reduction_clauses): Don't set placeholder to address of ref
5212 if it has already the right type.
5213 (lower_send_clauses): Handle OMP_CLAUSE__LOOPTEMP_.
5214 (expand_parallel_call): Use the new non-_start suffixed builtins,
5215 handle OMP_CLAUSE_PROC_BIND, don't call the outlined function
5216 and GOMP_parallel_end after the call.
5217 (expand_task_call): Handle OMP_CLAUSE_DEPEND.
5218 (expand_omp_for_init_counts): Handle combined loops.
5219 (expand_omp_for_init_vars): Add inner_stmt argument, handle combined
5221 (expand_omp_for_generic): Likewise. Use GOMP_loop_end_cancel at the
5222 end of cancellable loops.
5223 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
5224 Likewise. Handle collapse > 1 loops.
5225 (expand_omp_simd): Handle combined loops.
5226 (expand_omp_for): Add inner_stmt argument, adjust callers of
5227 expand_omp_for* functions, use expand_omp_for_static*chunk even
5228 for collapse > 1 unless ordered.
5229 (expand_omp_sections): Use GOMP_sections_end_cancel at the end
5230 of cancellable sections.
5231 (expand_omp_single): Remove need_barrier variable, just rely on
5232 gimple_omp_return_nowait_p. Adjust build_omp_barrier caller.
5233 (expand_omp_synch): Allow GIMPLE_OMP_TASKGROUP and GIMPLE_OMP_TEAMS.
5234 (expand_omp_atomic_load, expand_omp_atomic_store,
5235 expand_omp_atomic_fetch_op): Handle gimple_omp_atomic_seq_cst_p.
5236 (expand_omp_target): New function.
5237 (expand_omp): Handle combined loops. Handle GIMPLE_OMP_TASKGROUP,
5238 GIMPLE_OMP_TEAMS, GIMPLE_OMP_TARGET.
5239 (build_omp_regions_1): Immediately close region for
5240 GF_OMP_TARGET_KIND_UPDATE.
5241 (maybe_add_implicit_barrier_cancel): New function.
5242 (lower_omp_sections): Adjust lower_rec_input_clauses caller. Handle
5244 (lower_omp_single): Likewise. Add clobber after the barrier.
5245 (lower_omp_taskgroup): New function.
5246 (lower_omp_for): Handle combined loops. Adjust
5247 lower_rec_input_clauses caller. Handle cancellation.
5248 (lower_depend_clauses): New function.
5249 (lower_omp_taskreg): Lower depend clauses. Adjust
5250 lower_rec_input_clauses caller. Add clobber after the call. Handle
5252 (lower_omp_target, lower_omp_teams): New functions.
5253 (lower_omp_1): Handle cancellation. Handle GIMPLE_OMP_TASKGROUP,
5254 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GOMP_barrier, GOMP_cancel
5255 and GOMP_cancellation_point calls.
5256 (lower_omp): Fold stmts inside of target region.
5257 (diagnose_sb_1, diagnose_sb_2): Handle GIMPLE_OMP_TASKGROUP,
5258 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
5259 * builtin-types.def (DEF_FUNCTION_TYPE_8): Document.
5260 (BT_FN_VOID_OMPFN_PTR_UINT,
5261 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG,
5262 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG,
5263 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT): Remove.
5264 (BT_FN_VOID_OMPFN_PTR_UINT_UINT_UINT,
5265 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_UINT,
5266 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG_UINT,
5267 BT_FN_BOOL_INT, BT_FN_BOOL_INT_BOOL, BT_FN_VOID_UINT_UINT,
5268 BT_FN_VOID_INT_PTR_SIZE_PTR_PTR_PTR,
5269 BT_FN_VOID_INT_OMPFN_PTR_SIZE_PTR_PTR_PTR,
5270 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR): New.
5271 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1,
5272 call_may_clobber_ref_p_1): Handle BUILT_IN_GOMP_BARRIER_CANCEL,
5273 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_LOOP_END_CANCEL,
5274 BUILT_IN_GOMP_SECTIONS_END_CANCEL. Don't handle
5275 BUILT_IN_GOMP_PARALLEL_END.
5276 * gimple-low.c (lower_stmt): Handle GIMPLE_OMP_TASKGROUP,
5277 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
5278 * gimple-pretty-print.c (dump_gimple_omp_for): Handle
5279 GF_OMP_FOR_KIND_DISTRIBUTE.
5280 (dump_gimple_omp_target, dump_gimple_omp_teams): New functions.
5281 (dump_gimple_omp_block): Handle GIMPLE_OMP_TASKGROUP.
5282 (dump_gimple_omp_return): Print lhs if it has any.
5283 (dump_gimple_omp_atomic_load, dump_gimple_omp_atomic_store): Handle
5284 gimple_omp_atomic_seq_cst_p.
5285 (pp_gimple_stmt_1): Handle GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET
5286 and GIMPLE_OMP_TEAMS.
5287 * langhooks.c (lhd_omp_mappable_type): New function.
5288 * tree-vectorizer.c (struct simd_array_to_simduid): Fix up comment.
5289 * langhooks.h (struct lang_hooks_for_types): Add omp_mappable_type
5291 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_MAP,
5292 GOVD_ALIGNED and GOVD_MAP_TO_ONLY.
5293 (enum omp_region_type): Add ORT_TEAMS, ORT_TARGET_DATA and ORT_TARGET.
5294 (struct gimplify_omp_ctx): Add combined_loop field.
5295 (gimplify_call_expr, gimplify_modify_expr): Don't call fold_stmt
5296 on stmts inside of target region.
5297 (is_gimple_stmt): Return true for OMP_DISTRIBUTE and OMP_TASKGROUP.
5298 (omp_firstprivatize_variable): Handle GOVD_MAP, GOVD_ALIGNED,
5299 ORT_TARGET and ORT_TARGET_DATA.
5300 (omp_add_variable): Avoid checks on readding var for GOVD_ALIGNED.
5302 (omp_notice_threadprivate_variable): Complain about threadprivate
5303 variables in target region.
5304 (omp_notice_variable): Complain about vars with non-mappable type
5305 in target region. Handle ORT_TEAMS, ORT_TARGET and ORT_TARGET_DATA.
5306 (omp_check_private): Ignore ORT_TARGET* regions.
5307 (gimplify_scan_omp_clauses, gimplify_adjust_omp_clauses_1,
5308 gimplify_adjust_omp_clauses): Handle new OpenMP 4.0 clauses.
5309 (find_combined_omp_for): New function.
5310 (gimplify_omp_for): Handle gimplification of combined loops.
5311 (gimplify_omp_workshare): Gimplify also OMP_TARGET, OMP_TARGET_DATA,
5313 (gimplify_omp_target_update): New function.
5314 (gimplify_omp_atomic): Handle OMP_ATOMIC_SEQ_CST.
5315 (gimplify_expr): Handle OMP_DISTRIBUTE, OMP_TARGET, OMP_TARGET_DATA,
5316 OMP_TARGET_UPDATE, OMP_TEAMS, OMP_TASKGROUP.
5317 (gimplify_body): If fndecl has "omp declare target" attribute, add
5318 implicit ORT_TARGET context around it.
5319 * tree.def (OMP_DISTRIBUTE, OMP_TEAMS, OMP_TARGET_DATA, OMP_TARGET,
5320 OMP_TASKGROUP, OMP_TARGET_UPDATE): New tree codes.
5321 * tree-nested.c (convert_nonlocal_reference_stmt,
5322 convert_local_reference_stmt, convert_gimple_call): Handle
5323 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
5324 * omp-builtins.def (BUILT_IN_GOMP_TASK): Use
5325 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR
5326 instead of BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT.
5327 (BUILT_IN_GOMP_TARGET, BUILT_IN_GOMP_TARGET_DATA,
5328 BUILT_IN_GOMP_TARGET_END_DATA, BUILT_IN_GOMP_TARGET_UPDATE,
5329 BUILT_IN_GOMP_TEAMS, BUILT_IN_BARRIER_CANCEL,
5330 BUILT_IN_GOMP_LOOP_END_CANCEL,
5331 BUILT_IN_GOMP_SECTIONS_END_CANCEL, BUILT_IN_OMP_GET_TEAM_NUM,
5332 BUILT_IN_OMP_GET_NUM_TEAMS, BUILT_IN_GOMP_TASKGROUP_START,
5333 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_PARALLEL_LOOP_STATIC,
5334 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC,
5335 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED,
5336 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME, BUILT_IN_GOMP_PARALLEL,
5337 BUILT_IN_GOMP_PARALLEL_SECTIONS, BUILT_IN_GOMP_CANCEL,
5338 BUILT_IN_GOMP_CANCELLATION_POINT): New built-ins.
5339 (BUILT_IN_GOMP_PARALLEL_LOOP_STATIC_START,
5340 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC_START,
5341 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED_START,
5342 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME_START,
5343 BUILT_IN_GOMP_PARALLEL_START, BUILT_IN_GOMP_PARALLEL_END,
5344 BUILT_IN_GOMP_PARALLEL_SECTIONS_START): Remove.
5345 * tree-inline.c (remap_gimple_stmt, estimate_num_insns):
5346 Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
5347 * gimple.c (gimple_build_omp_taskgroup, gimple_build_omp_target,
5348 gimple_build_omp_teams): New functions.
5349 (walk_gimple_op): Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and
5350 GIMPLE_OMP_TASKGROUP. Walk optional lhs on GIMPLE_OMP_RETURN.
5351 (walk_gimple_stmt, gimple_copy): Handle GIMPLE_OMP_TARGET,
5352 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
5353 * gimple.h (enum gf_mask): GF_OMP_FOR_KIND_DISTRIBUTE,
5354 GF_OMP_FOR_COMBINED, GF_OMP_FOR_COMBINED_INTO,
5355 GF_OMP_TARGET_KIND_MASK, GF_OMP_TARGET_KIND_REGION,
5356 GF_OMP_TARGET_KIND_DATA, GF_OMP_TARGET_KIND_UPDATE,
5357 GF_OMP_ATOMIC_SEQ_CST): New.
5358 (gimple_build_omp_taskgroup, gimple_build_omp_target,
5359 gimple_build_omp_teams): New prototypes.
5360 (gimple_has_substatements): Handle GIMPLE_OMP_TARGET,
5361 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
5362 (gimple_omp_subcode): Use GIMPLE_OMP_TEAMS instead of
5363 GIMPLE_OMP_SINGLE as end of range.
5364 (gimple_omp_return_set_lhs, gimple_omp_return_lhs,
5365 gimple_omp_return_lhs_ptr, gimple_omp_atomic_seq_cst_p,
5366 gimple_omp_atomic_set_seq_cst, gimple_omp_for_combined_p,
5367 gimple_omp_for_set_combined_p, gimple_omp_for_combined_into_p,
5368 gimple_omp_for_set_combined_into_p, gimple_omp_target_clauses,
5369 gimple_omp_target_clauses_ptr, gimple_omp_target_set_clauses,
5370 gimple_omp_target_kind, gimple_omp_target_set_kind,
5371 gimple_omp_target_child_fn, gimple_omp_target_child_fn_ptr,
5372 gimple_omp_target_set_child_fn, gimple_omp_target_data_arg,
5373 gimple_omp_target_data_arg_ptr, gimple_omp_target_set_data_arg,
5374 gimple_omp_teams_clauses, gimple_omp_teams_clauses_ptr,
5375 gimple_omp_teams_set_clauses): New inlines.
5376 (CASE_GIMPLE_OMP): Add GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS
5377 and GIMPLE_OMP_TASKGROUP.
5378 * tree-core.h (enum omp_clause_code): Add new OpenMP 4.0 clause codes.
5379 (enum omp_clause_depend_kind, enum omp_clause_map_kind,
5380 enum omp_clause_proc_bind_kind): New.
5381 (union omp_clause_subcode): Add depend_kind, map_kind and
5382 proc_bind_kind fields.
5383 * tree-cfg.c (make_edges): Handle GIMPLE_OMP_TARGET,
5384 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
5385 * langhooks-def.h (lhd_omp_mappable_type): New prototype.
5386 (LANG_HOOKS_OMP_MAPPABLE_TYPE): Define.
5387 (LANG_HOOKS_FOR_TYPES_INITIALIZER): Add it.
5389 2013-10-10 Teresa Johnson <tejohnson@google.com>
5391 * predict.c (tree_estimate_probability): Add new parameter
5392 for estimate_bb_frequencies.
5393 (estimate_bb_frequencies): Add new parameter to force estimation.
5394 (rebuild_frequencies): When max frequency in function is small,
5395 recompute counts from frequencies.
5396 * predict.h (estimate_bb_frequencies): New parameter.
5398 2013-10-10 David Malcolm <dmalcolm@redhat.com>
5400 * ipa-inline.c (ipa_inline): Fix leak of "order" when
5401 optimizations are disabled.
5403 2013-10-10 David Malcolm <dmalcolm@redhat.com>
5405 * coverage.c (coverage_finish): Fix leak of da_file_name.
5407 2013-10-10 Jan Hubicka <jh@suse.cz>
5409 * config/i386/x86-tune.def: Enable X86_TUNE_SSE_TYPELESS_STORES
5410 for generic, enable X86_TUNE_SSE_LOAD0_BY_PXOR for Bulldozer,
5413 2013-10-10 Jakub Jelinek <jakub@redhat.com>
5416 * stmt.c (expand_asm_operands): Add FALLTHRU_BB argument,
5417 if any labels are in FALLTHRU_BB, use a special label emitted
5418 immediately after the asm goto insn rather than label_rtx
5420 (expand_asm_stmt): Adjust caller.
5421 * cfgrtl.c (commit_one_edge_insertion): Force splitting of
5422 edge if the last insn in predecessor is a jump with single successor,
5423 but it isn't simplejump_p.
5425 2013-10-10 Richard Biener <rguenther@suse.de>
5427 PR tree-optimization/58656
5428 * tree-ssa-pre.c (phi_translate): Do not cache failed translations.
5430 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
5432 * gimplify.c: Include expr.h and tm_p.h for targets with special
5433 va-arg padding requirements.
5435 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
5437 * tree-flow.h: Move some prototypes to gimple.h.
5438 (gimple_fold_indirect_ref): Move prototype to gimple-fold.h.
5439 * gimple.h: Relocate some prototypes from tree-flow.h
5440 * builtins.c (std_gimplify_va_arg_expr, build_va_arg_indirect_ref):
5442 * gimplify.c (gimple_fold_indirect_ref): Move to gimple-fold.c.
5443 (build_va_arg_indirect_ref): Relocate and make static.
5444 (std_gimplify_va_arg_expr): Relocate here.
5445 * gimple-fold.c (gimple_fold_indirect_ref): Relocate here.
5446 * gimple-fold.h (gimple_fold_indirect_ref): Add prototype.
5448 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5450 * doc/md.texi: Document the mnemonic attribute.
5452 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5455 * gensupport.c (gen_mnemonic_attr): Handle (set (attr x) y) and
5456 (set_attr_alternative x ...) when searching for user defined
5459 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
5461 * config/aplha/alpha.c: Add gimple-ssa.h to include list.
5463 2013-10-09 Easwaran Raman <eraman@google.com>
5465 * params.def (PARAM_MIN_SIZE_FOR_STACK_SHARING): New param...
5466 * cfgexpand.c (defer_stack_allocation): ...use here
5467 * doc/invoke.texi: Add documentation for min-size-for-stack-sharing.
5469 2013-10-09 Zhenqiang Chen <zhenqiang.chen@arm.com>
5471 * tree-ssa-phiopts.c (rhs_is_fed_for_value_replacement): New function.
5472 (operand_equal_for_value_replacement): New function, extracted from
5473 value_replacement and enhanced to catch more cases.
5474 (value_replacement): Use operand_equal_for_value_replacement.
5476 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
5478 * loop-doloop.c (doloop_modify, doloop_optimize): Use
5479 get_max_loop_iterations.
5481 2013-10-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5483 * config/arm/aarch-common.c (arm_early_load_addr_dep):
5484 Place comment above function.
5486 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
5488 * tree-flow.h: Remove all remaining prototypes, enums and structs that
5489 are not related to tree-cfg.c.
5490 * tree-ssa-address.h: New file. Relocate prototypes.
5491 * tree-ssa-address.c: (struct mem_address): Relocate from tree-flow.h.
5492 (addr_for_mem_ref): New. Combine call to get_address_description and
5493 return addr_for_mem_ref.
5494 * expr.c (expand_expr_real_1): Use new addr_for_mem_ref routine.
5495 * tree-ssa-live.h: Adjust prototypes.
5496 * passes.c: Include tree-ssa-live.h.
5497 * gimple-pretty-print.h (gimple_dump_bb): Add prototype.
5498 * graphite.c (graphite_transform_loops): Make static.
5499 (graphite_transforms, gate_graphite_transforms, pass_data_graphite,
5500 make_pass_graphite, pass_data_graphite_transforms,
5501 make_pass_graphite_transforms): Relocate here from tree-ssa-loop.c.
5502 * ipa-pure-const.c (warn_function_noreturn): Make static.
5503 (execute_warn_function_noreturn, gate_warn_function_noreturn,
5504 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
5505 Relocate from tree-cfg.c
5506 * tree-cfg.c (tree_node_can_be_shared, gimple_empty_block_p): Make
5508 (execute_warn_function_noreturn, gate_warn_function_noreturn,
5509 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
5510 Move to ipa-pure-const.c.
5511 (execute_fixup_cfg, class pass_fixup_cfg, make_pass_fixup_cfg):
5512 Relocate from tree-optimize.c.
5513 * tree-optimize.c (execute_fixup_cfg, class pass_fixup_cfg,
5514 make_pass_fixup_cfg): Move to tree-cfg.c.
5515 * tree-chrec.h: (enum ev_direction): Relocate here from tree-flow.h.
5516 Relocate some prototypes.
5517 * tree-data-ref.h (tree_check_data_deps) Add prototype.
5518 * tree-dump.c (dump_function_to_file): Remove prototype.
5519 Add tree-flow.h to the include file.
5520 * tree-dump.h: Remove prototype.
5521 * tree-parloops.h: New File. Add prototypes.
5522 * tree-parloops.c (gate_tree_parallelize_loops, tree_parallelize_loops,
5523 pass_data_parallelize_loops, make_pass_parallelize_loops): Relocate
5524 from tree-ssa-loop.c.
5525 * tree-predcom.c (run_tree_predictive_commoning,
5526 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom):
5527 Relocate here from tree-ssa-loop.c.
5528 * tree-ssa-dom.c (tree_ssa_dominator_optimize) Don't call
5529 ssa_name_values.release ().
5530 * tree-ssa-threadedge.h: New File. Relocate prototypes here.
5531 (ssa_name_values): Relocate from tree-flow.h.
5532 * tree-ssa.h: Include tree-ssa-threadedge.h and tree-ssa-address.h.
5533 * tree-ssa-loop.c (run_tree_predictive_commoning,
5534 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom,
5535 graphite_transforms, gate_graphite_transforms, pass_data_graphite,
5536 make_pass_graphite, pass_data_graphite_transforms,
5537 make_pass_graphite_transforms, gate_tree_parallelize_loops,
5538 tree_parallelize_loops, pass_data_parallelize_loops,
5539 make_pass_parallelize_loops): Move to other files.
5540 * tree-vectorizer.h (lpeel_tree_duplicate_loop_to_edge_cfg): Prototype
5542 * tree.h: Remove prototypes from tree-address.c.
5544 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
5546 * tree-flow.h (tm_restart_node, gimple_df): Move to gimple-ssa.h.
5547 (struct int_tree_map): Move to tree-hasher.h
5548 (SCALE, LABEL, PERCENT): Move to gimple.h
5549 * tree-flow-inline.h: Delete. Move functions to other files.
5550 (unmodifiable_var_p, ref_contains_array_ref): Unused, so delete.
5551 * gimple-ssa.h (tm_restart_node, gimple_df): Relocate from tree-flow.h.
5552 (gimple_in_ssa_p, gimple_vop): Relocate from tree-flow-inline.h
5553 * gimple.h (imple_stmt_max_uid, set_gimple_stmt_max_uid,
5554 inc_gimple_stmt_max_uid, get_lineno): Relocate from tree-flow-inline.h.
5555 (SCALE, LABEL, PERCENT): Relocate from tree-flow.h
5556 * tree-hasher.h: Don't include tree-flow.h.
5557 (struct int_tree_map): Relocate from tree-flow.h.
5558 * tree-sra.c (contains_view_convert_expr_p): Relocate from
5559 tree-flow-inline.h and make static.
5560 * tree-ssa-alias.h (ranges_overlap_p): Relocate from
5562 * tree-ssa-operands.c (gimple_ssa_operands): Relocate from
5563 tree-flow-inline.h and make static.
5564 * tree.h (is_global_var, may_be_aliased): Relocate from
5566 * Makefile.in (GTFILES): Remove tree-flow.h and add gimple-ssa.h.
5567 * value-prof.c: No longer include tree-flow-inline.h.
5568 * tree-switch-conversion.c: No longer include tree-flow-inline.h.
5570 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
5572 * tree-flow.h: Move some protoypes. Include new tree-ssa-loop.h.
5573 (struct affine_iv, struct tree_niter_desc): Move to tree-ssa-loop.h.
5574 (enum move_pos): Move to tree-ssa-loop-im.h
5575 * cfgloop.h: Move some prototypes.
5576 (gcov_type_to_double_int): relocate from tree-ssa-loop.niter.c.
5577 * tree-flow-inline.h (loop_containing_stmt): Move to tree-ssa-loop.h.
5578 * tree-ssa-loop.h: New File. Include other tree-ssa-loop-*.h files.
5579 (struct affine_iv, struct tree_niter_desc): Relocate from tree-flow.h.
5580 (loop_containing_stmt): Relocate from tree-flow-inline.h.
5581 * tree-ssa-loop-ch.c: (do_while_loop_p): Make static.
5582 * tree-ssa-loop-im.c (for_each_index): Move to tree-ssa-loop.c.
5583 (enum move_pos): Relocate here.
5584 (lsm_tmp_name_add, gen_lsm_tmp_name, get_lsm_tmp_name): Move to
5586 (execute_sm_if_changed_flag_set): Change get_lsm_tmp_name call.
5587 (tree_ssa_loop_im, gate_tree_ssa_loop_im, pass_data_lim,
5588 make_pass_lim): Relocate here from tree-ssa-loop.c.
5589 * tree-ssa-loop-ivcanon.c (tree_num_loop_insns): Move to
5591 (loop_edge_to_cancel, unloop_loops): Make static.
5592 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
5593 make_pass_iv_canon): Relocate from tree-ssa-loop.c.
5594 (tree_complete_unroll, gate_tree_complete_unroll,
5595 pass_data_complete_unroll, make_pass_complete_unroll): Relocate here.
5596 (tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
5597 pass_data_complete_unrolli, make_pass_complete_unrolli): Relocate here.
5598 * tree-ssa-loop-ivopts.c: Remove local prototypes.
5599 (stmt_invariant_in_loop_p): Remove unused function.
5600 * tree-ssa-loop-ivopts.h: New file. Add prototypes.
5601 * tree-ssa-loop-manip.h: New file. Add prototypes.
5602 * tree-ssa-loop-niter.c (record_niter_bound): Move to cfgloop.c.
5603 (gcov_type_to_double_int): Move to cfgloop.h.
5604 (double_int_cmp, bound_index,
5605 estimate_numbers_of_iterations_loop): Make static.
5606 (estimated_loop_iterations): Factor out get_estimated_loop_iterations.
5607 (max_loop_iterations): Factor out get_max_loop_iterations.
5608 (estimated_loop_iterations_int, max_stmt_executions_int): Move to
5610 * tree-ssa-loop-niter.h: New file. Add prototypes.
5611 * tree-ssa-loop-prefetch.c (tree_ssa_loop_prefetch,
5612 gate_tree_ssa_loop_prefetch, pass_data_loop_prefetch,
5613 make_pass_loop_prefetch): Relocate from tree-ssa-loop.c.
5614 * tree-ssa-loop-unswitch.c (tree_ssa_loop_unswitch,
5615 gate_tree_ssa_loop_unswitch, pass_data_tree_unswitch,
5616 make_pass_tree_unswitch): Relocate from tree-ssa-loop.c.
5617 * tree-ssa-loop.c (tree_ssa_loop_im, gate_tree_ssa_loop_im,
5618 pass_data_lim, make_pass_lim): Move to tree-ssa-loop-im.c.
5619 (tree_ssa_loop_unswitch, gate_tree_ssa_loop_unswitch,
5620 pass_data_tree_unswitch, make_pass_tree_unswitch): Move.
5621 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
5622 make_pass_iv_canon, tree_complete_unroll, gate_tree_complete_unroll,
5623 pass_data_complete_unroll, make_pass_complete_unroll,
5624 tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
5625 pass_data_complete_unrolli, make_pass_complete_unrolli): Move to
5626 tree-ssa-loop-ivcanon.c.
5627 (tree_ssa_loop_prefetch, gate_tree_ssa_loop_prefetch,
5628 pass_data_loop_prefetch, make_pass_loop_prefetch): Move to
5629 tree-ssa-loop-prefetch.c.
5630 (for_each_index, lsm_tmp_name_add, gen_lsm_tmp_name): Relocate from
5632 (get_lsm_tmp_name): Relocate and add suffix parameter.
5633 (tree_num_loop_insns): Relocate from tree-ssa-ivcanon.c.
5634 * tree-scalar-evolution.h (simple_iv): Don't use affive_iv typedef.
5635 * cfgloop.c (record_niter_bound, estimated_loop_iterations_int,
5636 max_stmt_executions_int): Move from tree-ssa-loop-niter.c.
5637 (get_estimated_loop_iterations): Factor out accessor from
5638 estimated_loop_iterations in tree-ssa-loop-niter.c.
5639 (get_max_loop_iterations): Factor out accessor from
5640 _max_loop_iterations in tree-ssa-niter.c.
5641 * loop-unroll.c (decide_unroll_constant_iterations,
5642 decide_unroll_runtime_iterations, decide_peel_simple,
5643 decide_unroll_stupid): Use new get_* accessors.
5645 2013-10-09 Marc Glisse <marc.glisse@inria.fr>
5647 PR tree-optimization/20318
5648 * doc/extend.texi (returns_nonnull): New function attribute.
5649 * fold-const.c (tree_expr_nonzero_warnv_p): Look for returns_nonnull
5651 * tree-vrp.c (gimple_stmt_nonzero_warnv_p): Likewise.
5652 (stmt_interesting_for_vrp): Accept all GIMPLE_CALL.
5654 2013-10-09 Eric Botcazou <ebotcazou@adacore.com>
5657 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): Return
5658 false if both components are bitfields.
5660 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
5662 * config/aarch64/arm_neon.h (vclz_s8, vclz_s16, vclz_s32)
5663 (vclzq_s8, vclzq_s16, vclzq_s32, vclz_u8, vclz_u16, vclz_u32)
5664 (vclzq_u8, vclzq_u16, vclzq_u32): Replace ASM with C.
5665 * config/aarch64/aarch64.h
5666 (CLZ_DEFINED_VALUE_AT_ZERO): Macro fixed for clz.
5667 * config/aarch64/aarch64-simd-builtins.def
5668 (VAR1 (UNOP, clz, 0, v4si)): Replaced with iterator.
5670 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
5672 * config/aarch64/arm_neon.h (vadd_f64, vsub_f64): Implementation added.
5674 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
5676 * config/aarch64/arm_neon.h (vdiv_f64): Added.
5678 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
5680 * config/aarch64/arm_neon.h (vneg_f32): Asm replaced with C.
5681 (vneg_f64): New intrinsic.
5682 (vneg_s8): Asm replaced with C.
5683 (vneg_s16): Likewise.
5684 (vneg_s32): Likewise.
5685 (vneg_s64): New intrinsic.
5686 (vnegq_f32): Asm replaced with C.
5687 (vnegq_f64): Likewise.
5688 (vnegq_s8): Likewise.
5689 (vnegq_s16): Likewise.
5690 (vnegq_s32): Likewise.
5691 (vnegq_s64): Likewise.
5693 2013-10-09 Renlin Li <Renlin.Li@arm.com>
5695 * config/arm/arm.c (arm_output_mi_thunk): Use plus_constant.
5697 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5699 * config/s390/s390.c (s390_register_info_stdarg_fpr): Remove
5700 packed stack special handling.
5701 (s390_frame_info, s390_emit_prologue, s390_emit_epilogue): Switch
5702 back to fixed stack slots for FPRs saved due to stdarg.
5704 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5706 * config/s390/s390.c (s390_frame_info): Restructure function.
5708 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5710 * config/s390/s390.c (struct s390_frame_layout): New field
5712 (cfun_save_arg_fprs_p, cfun_gpr_save_slot): New macros.
5713 (s390_reg_clobbered_rtx, s390_regs_ever_clobbered): Change type of
5714 regs_ever_clobbered to char*.
5715 (s390_regs_ever_clobbered): Check crtl->saves_all_registers instead
5716 of cfun->has_nonlocal_label. Ignore frame related restore INSNs.
5717 (s390_register_info): Enable FPR save slots. Move/Copy some
5718 functionality into ...
5719 (s390_register_info_gprtofpr, s390_register_info_stdarg_fpr)
5720 (s390_register_info_stdarg_gpr, s390_optimize_register_info): New
5722 (s390_frame_info): Do gpr slot allocation here now. stdarg does
5723 not imply a stack frame.
5724 (s390_init_frame_layout): Remove variable clobbered_regs.
5725 (s390_update_register_info): Remove function.
5726 (s390_hard_regno_rename_ok): Call-saved regs without a save slot
5727 cannot be used for register renaming.
5728 (s390_hard_regno_scratch_ok): New function.
5729 (TARGET_HARD_REGNO_SCRATCH_OK): Define target hook.
5730 (s390_initial_elimination_offset): Change offset calculation of
5731 the return address pointer.
5732 (save_gprs): Deal with only r6 being saved from the call-saved regs.
5733 (restore_gprs): Set frame related flag.
5734 (s390_save_gprs_to_fprs, s390_restore_gprs_from_fprs): New functions.
5735 (s390_emit_prologue): Call s390_register_info instead of
5736 s390_update_frame_layout. Call s390_save_gprs_to_fprs.
5737 (s390_emit_epilogue): Call s390_restore_gprs_from_fprs.
5738 (s390_optimize_prologue): Call s390_optimize_register_info.
5739 Try to remove also FPR slot save/restore INSNs. Remove frame
5740 related flags from restore INSNs.
5742 2013-10-08 DJ Delorie <dj@redhat.com>
5744 * config/rl78/rl78-expand.md (movqi): use operands[] not operandN.
5747 * config/rl78/rl78.c (rl78_print_operand_1): Change %c to %C to
5748 avoid conflict with the MI use of %c.
5749 * config/rl78/rl78-real.md: change %c to %C throughout.
5750 * config/rl78/rl78-virt.md: Likewise.
5752 2013-10-08 Jan Hubicka <jh@suse.cz>
5754 * config/i386/i386.c (ix86_option_override_internal): Switch
5755 to SSE math for -ffast-math when target ISA supports SSE2.
5757 2013-10-08 Andrew MacLeod <amacleod@redhat.com>
5759 * tree-flow.h: Remove some prototypes.
5760 * tree.h: Remove some protypes, add a couple.
5761 * tree.c (using_eh_for_cleanups_flag, using_eh_for_cleanups,
5762 using_eh_for_cleanups_p): Add interface routines for front ends.
5763 * tree-eh.h: New file. Add protoptyes.
5764 * tree-eh.c (using_eh_for_cleanups_p, using_eh_for_cleanups): Delete.
5765 (add_stmt_to_eh_lp_fn): Make static.
5766 (lower_try_finally): Use new using_eh_for_cleanups_p.
5767 * emit-rtl.c: Include tree-eh.h.
5768 * gimple.h: Include tree-eh.h.
5770 2013-10-08 Marc Glisse <marc.glisse@inria.fr>
5772 PR tree-optimization/58480
5773 * tree-vrp.c (infer_nonnull_range): New function.
5774 (infer_value_range): Call infer_nonnull_range.
5776 2013-10-08 Dehao Chen <dehao@google.com>
5778 PR tree-optimization/58619
5779 * tree-inline.c (copy_phis_for_bb): Combine location data
5782 2013-10-08 Zhenqiang Chen <zhenqiang.chen@linaro.org>
5785 * config/arm/arm.c (arm_emit_ldrd_pop): Attach
5786 RTX_FRAME_RELATED_P on INSN.
5788 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5790 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const_le): New.
5791 (altivec_expand_vec_perm_const): Call it.
5793 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5795 * config/rs6000/vector.md (mov<mode>): Emit permuted move
5796 sequences for LE VSX loads and stores at expand time.
5797 * config/rs6000/rs6000-protos.h (rs6000_emit_le_vsx_move): New
5799 * config/rs6000/rs6000.c (rs6000_const_vec): New.
5800 (rs6000_gen_le_vsx_permute): New.
5801 (rs6000_gen_le_vsx_load): New.
5802 (rs6000_gen_le_vsx_store): New.
5803 (rs6000_gen_le_vsx_move): New.
5804 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): New.
5805 (*vsx_le_perm_load_v4si): New.
5806 (*vsx_le_perm_load_v8hi): New.
5807 (*vsx_le_perm_load_v16qi): New.
5808 (*vsx_le_perm_store_v2di): New.
5809 (*vsx_le_perm_store_v4si): New.
5810 (*vsx_le_perm_store_v8hi): New.
5811 (*vsx_le_perm_store_v16qi): New.
5812 (*vsx_xxpermdi2_le_<mode>): New.
5813 (*vsx_xxpermdi4_le_<mode>): New.
5814 (*vsx_xxpermdi8_le_V8HI): New.
5815 (*vsx_xxpermdi16_le_V16QI): New.
5816 (*vsx_lxvd2x2_le_<mode>): New.
5817 (*vsx_lxvd2x4_le_<mode>): New.
5818 (*vsx_lxvd2x8_le_V8HI): New.
5819 (*vsx_lxvd2x16_le_V16QI): New.
5820 (*vsx_stxvd2x2_le_<mode>): New.
5821 (*vsx_stxvd2x4_le_<mode>): New.
5822 (*vsx_stxvd2x8_le_V8HI): New.
5823 (*vsx_stxvd2x16_le_V16QI): New.
5825 2013-10-07 Renlin Li <Renlin.Li@arm.com>
5827 * config/arm/arm-cores.def (cortex-a53): Use cortex tuning.
5829 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5831 * config/s390/s390.c (s390_register_info): Make the call-saved FPR
5832 loop to work also for 31bit ABI.
5833 Save the stack pointer for frame_size > 0.
5835 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5837 * config/s390/s390.md ("tbegin", "tbegin_nofloat", "tbegin_retry")
5838 ("tbegin_retry_nofloat", "tend", "tabort", "tx_assist"): Remove
5839 constraint letters from expanders.
5840 ("tbegin_retry", "tbegin_retry_nofloat"): Change predicate of the
5841 retry count to general_operand.
5842 ("tabort"): Give operand 0 a mode.
5843 ("tabort_1"): Add mode and constraint letter for operand 0.
5844 * doc/extend.texi: Fix protoype of __builtin_non_tx_store.
5846 2013-10-04 Jeff Law <law@redhat.com>
5848 * tree-ssa-threadedge.c: Fix some trailing whitespace problems.
5850 * tree-ssa-threadedge.c (thread_through_normal_block): Broken
5852 (thread_across_edge): Here. Call it.
5854 2013-10-04 Cary Coutant <ccoutant@google.com>
5856 * dwarf2out.c (dw_sra_loc_expr): Release addr_table entries when
5857 discarding a location list expression (or a piece of one).
5859 2013-10-03 Jan Hubicka <jh@suse.cz>
5861 * config/i386/i386.c (ix86_issue_rate): Pentium4, Nocona has issue
5862 rate of 2. Core2, Corei7 and Haswell has issue rate of 4.
5863 (ix86_adjust_cost): Remove Atom case; fix core2/corei7/Haswell case.
5865 2013-10-03 Jan Hubicka <jh@suse.cz>
5867 * config/i386/i386.c (ix86_option_override_internal): Do not enable
5868 accumulate-outgoing-args when producing unwind info.
5870 2013-10-03 Wei Mi <wmi@google.com>
5872 * lra-constraints.c (insert_move_for_subreg): New function
5873 extracted from simplify_operand_subreg.
5874 (simplify_operand_subreg): Add reload for paradoxical subreg.
5876 2013-10-03 Rong Xu <xur@google.com>
5878 * ipa-inline-analysis.c (find_foldable_builtin_expect): Find
5879 the candidate of builtin_expect such that we should fix the
5880 size/time estimation.
5881 (estimate_function_body_sizes): Do the acutally size/time fix-up
5884 2013-10-03 Rong Xu <xur@google.com>
5886 * predict.c (tree_predict_by_opcode): Get the probability
5887 for builtin_expect from param builtin_expect_probability.
5888 * params.def (BUILTIN_EXPECT_PROBABILITY): New parameter.
5889 * predict.def (PRED_BUILTIN_EXPECT_RELAXED): Fix comments.
5890 * doc/invoke.texi: Add documentation for builtin-expect-probability.
5892 2013-10-03 Marc Glisse <marc.glisse@inria.fr>
5895 * common.opt (fcheck-new): Moved from c.opt. Make it 'Common'.
5896 * calls.c (alloca_call_p): Use get_callee_fndecl.
5897 * fold-const.c (tree_expr_nonzero_warnv_p): Handle operator new.
5898 * tree-vrp.c (gimple_stmt_nonzero_warnv_p, stmt_interesting_for_vrp):
5900 (vrp_visit_stmt): Remove duplicated code.
5902 2013-10-03 Michael Meissner <meissner@linux.vnet.ibm.com>
5904 * config/rs6000/rs6000-builtin.def (XSRDPIM): Use floatdf2,
5905 ceildf2, btruncdf2, instead of vsx_* name.
5907 * config/rs6000/vsx.md (vsx_add<mode>3): Change arithmetic
5908 iterators to only do V2DF and V4SF here. Move the DF code to
5909 rs6000.md where it is combined with SF mode. Replace <VSv> with
5910 just 'v' since only vector operations are handled with these insns
5911 after moving the DF support to rs6000.md.
5912 (vsx_sub<mode>3): Likewise.
5913 (vsx_mul<mode>3): Likewise.
5914 (vsx_div<mode>3): Likewise.
5915 (vsx_fre<mode>2): Likewise.
5916 (vsx_neg<mode>2): Likewise.
5917 (vsx_abs<mode>2): Likewise.
5918 (vsx_nabs<mode>2): Likewise.
5919 (vsx_smax<mode>3): Likewise.
5920 (vsx_smin<mode>3): Likewise.
5921 (vsx_sqrt<mode>2): Likewise.
5922 (vsx_rsqrte<mode>2): Likewise.
5923 (vsx_fms<mode>4): Likewise.
5924 (vsx_nfma<mode>4): Likewise.
5925 (vsx_copysign<mode>3): Likewise.
5926 (vsx_btrunc<mode>2): Likewise.
5927 (vsx_floor<mode>2): Likewise.
5928 (vsx_ceil<mode>2): Likewise.
5929 (vsx_smaxsf3): Delete scalar ops that were moved to rs6000.md.
5930 (vsx_sminsf3): Likewise.
5931 (vsx_fmadf4): Likewise.
5932 (vsx_fmsdf4): Likewise.
5933 (vsx_nfmadf4): Likewise.
5934 (vsx_nfmsdf4): Likewise.
5935 (vsx_cmpdf_internal1): Likewise.
5937 * config/rs6000/rs6000.h (TARGET_SF_SPE): Define macros to make it
5938 simpler to select whether a target has SPE or traditional floating
5939 point support in iterators.
5940 (TARGET_DF_SPE): Likewise.
5941 (TARGET_SF_FPR): Likewise.
5942 (TARGET_DF_FPR): Likewise.
5943 (TARGET_SF_INSN): Macros to say whether floating point support
5944 exists for a given operation for expanders.
5945 (TARGET_DF_INSN): Likewise.
5947 * config/rs6000/rs6000.c (Ftrad): New mode attributes to allow
5948 combining of SF/DF mode operations, using both traditional and VSX
5956 (abs<mode>2): Combine SF/DF modes using traditional floating point
5957 instructions. Add support for using the upper DF registers with
5958 VSX support, and SF registers with power8-vector support. Update
5959 expanders for operations supported by both the SPE and traditional
5960 floating point units.
5961 (abs<mode>2_fpr): Likewise.
5962 (nabs<mode>2): Likewise.
5963 (nabs<mode>2_fpr): Likewise.
5964 (neg<mode>2): Likewise.
5965 (neg<mode>2_fpr): Likewise.
5966 (add<mode>3): Likewise.
5967 (add<mode>3_fpr): Likewise.
5968 (sub<mode>3): Likewise.
5969 (sub<mode>3_fpr): Likewise.
5970 (mul<mode>3): Likewise.
5971 (mul<mode>3_fpr): Likewise.
5972 (div<mode>3): Likewise.
5973 (div<mode>3_fpr): Likewise.
5974 (sqrt<mode>3): Likewise.
5975 (sqrt<mode>3_fpr): Likewise.
5976 (fre<Fs>): Likewise.
5977 (rsqrt<mode>2): Likewise.
5978 (cmp<mode>_fpr): Likewise.
5979 (smax<mode>3): Likewise.
5980 (smin<mode>3): Likewise.
5981 (smax<mode>3_vsx): Likewise.
5982 (smin<mode>3_vsx): Likewise.
5983 (negsf2): Delete SF operations that are merged with DF.
5990 (fmasf4_fpr): Likewise.
5991 (fmssf4_fpr): Likewise.
5992 (nfmasf4_fpr): Likewise.
5993 (nfmssf4_fpr): Likewise.
5994 (sqrtsf2): Likewise.
5995 (rsqrtsf_internal1): Likewise.
5996 (smaxsf3): Likewise.
5997 (sminsf3): Likewise.
5998 (cmpsf_internal1): Likewise.
5999 (copysign<mode>3_fcpsgn): Add VSX/power8-vector support.
6000 (negdf2): Delete DF operations that are merged with SF.
6002 (nabsdf2): Likewise.
6008 (rsqrtdf_internal1): Likewise.
6009 (fmadf4_fpr): Likewise.
6010 (fmsdf4_fpr): Likewise.
6011 (nfmadf4_fpr): Likewise.
6012 (nfmsdf4_fpr): Likewise.
6013 (sqrtdf2): Likewise.
6014 (smaxdf3): Likewise.
6015 (smindf3): Likewise.
6016 (cmpdf_internal1): Likewise.
6017 (lrint<mode>di2): Use TARGET_<MODE>_FPR macro.
6018 (btrunc<mode>2): Delete separate expander, and combine with the
6019 insn and add VSX instruction support. Use TARGET_<MODE>_FPR.
6020 (btrunc<mode>2_fpr): Likewise.
6021 (ceil<mode>2): Likewise.
6022 (ceil<mode>2_fpr): Likewise.
6023 (floor<mode>2): Likewise.
6024 (floor<mode>2_fpr): Likewise.
6025 (fma<mode>4_fpr): Combine SF and DF fused multiply/add support.
6026 Add support for using the upper registers with VSX and
6027 power8-vector. Move insns to be closer to the define_expands. On
6028 VSX systems, prefer the traditional form of FMA over the VSX
6029 version, since the traditional form allows the target not to
6030 overlap with the inputs.
6031 (fms<mode>4_fpr): Likewise.
6032 (nfma<mode>4_fpr): Likewise.
6033 (nfms<mode>4_fpr): Likewise.
6035 2013-10-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6036 Richard Earnshaw <richard.earnshaw@arm.com>
6038 * config/arm/aarch-common-protos.h (struct alu_cost_table): New.
6039 (struct mult_cost_table): Likewise.
6040 (struct mem_cost_table): Likewise.
6041 (struct fp_cost_table): Likewise.
6042 (struct vector_cost_table): Likewise.
6043 (cpu_cost_table): Likewise.
6044 * config/arm/arm.opt (mold-rts-costs): New option.
6045 (mnew-generic-costs): Likewise.
6046 * config/arm/arm.c (generic_extra_costs): New table.
6047 (cortexa15_extra_costs): Likewise.
6048 (arm_slowmul_tune): Use NULL as new costs.
6049 (arm_fastmul_tune): Likewise.
6050 (arm_strongarm_tune): Likewise.
6051 (arm_xscale_tune): Likewise.
6052 (arm_9e_tune): Likewise.
6053 (arm_v6t2_tune): Likewise.
6054 (arm_cortex_a5_tune): Likewise.
6055 (arm_cortex_a9_tune): Likewise.
6056 (arm_v6m_tune): Likewise.
6057 (arm_fa726te_tune): Likewise.
6058 (arm_cortex_a15_tune): Use cortex15_extra_costs.
6059 (arm_cortex_tune): Use generict_extra_costs.
6060 (shifter_op_p): New function.
6061 (arm_unspec_cost): Likewise.
6062 (LIBCALL_COST): Define.
6063 (arm_new_rtx_costs): New function.
6064 (arm_rtx_costs): Use arm_new_rtx_costs when core-specific
6065 table is available. Use old costs otherwise unless mnew-generic-costs
6067 * config/arm/arm-protos.h (tune_params): Add insn_extra_cost field.
6068 (cpu_cost_table): Declare.
6070 2013-10-03 Marcus Shawcroft <marcus.shawcroft@arm.com>
6073 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>)
6074 (*subs_mul_imm_<mode>)
6075 (*add_<shift>_<mode>, *add_<shift>_si_uxtw,*add_mul_imm_<mode>)
6076 (*sub_<shift>_<mode>)
6077 (*sub_<shift>_si_uxtw,*sub_mul_imm_<mode>, *sub_mul_imm_si_uxtw):
6078 Remove k constraint.
6080 2013-10-03 Ian Bolton <ian.bolton@arm.com>
6082 * config/aarch64/aarch64.c (aarch64_secondary_reload): Remove legacy
6084 * config/aarch64/aarch64.md (reload_sp_immediate): Likewise.
6086 2013-10-02 Teresa Johnson <tejohnson@google.com>
6088 * predict.c (probably_never_executed): New function.
6089 (probably_never_executed_bb_p): Invoke probably_never_executed.
6090 (probably_never_executed_edge_p): Ditto.
6091 * bb-reorder.c (find_rarely_executed_basic_blocks_and_crossing_edges):
6092 Treat profile insanities conservatively.
6094 2013-10-02 John David Anglin <danglin@gcc.gnu.org>
6096 * config.gcc (hppa*64*-*-linux*): Don't add pa/t-linux to tmake_file.
6098 2013-10-02 Vladimir Makarov <vmakarov@redhat.com>
6100 * lra-constraints.c (process_alt_operand): Calculate scratch_p and
6101 use it. Use smaller increase for scratch. Don't increase reject
6102 for early clobber scratch.
6103 * lra-eliminations.c (eliminate_regs_in_insn): Remove all insns
6104 setting eliminated regs except setting fp from hfp.
6105 (lra_eliminate): Check lra_insn_recog_data on NULL.
6107 2013-10-02 Michael Meissner <meissner@linux.vnet.ibm.com>
6110 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Turn off
6111 setting -mvsx-timode by default until the underlying problem is fixed.
6112 (RS6000_CPU, power7 defaults): Likewise.
6114 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
6116 * config/x-linux (host-linux.o): Remove header dependencies.
6117 Use $(COMPILE) and $(POSTCOMPILE).
6118 * config/t-linux-android (linux-android.o): Ditto.
6120 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
6122 * Makefile.in (expmed.o-warn): Remove.
6124 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
6126 * graphite-scop-detection.c: Include tree-ssa-propagate,h.
6127 * graphite-sese-to-poly.c: Include tree-ssa-propagate.h.
6129 2013-10-02 Teresa Johnson <tejohnson@google.com>
6131 * dojump.c (do_jump_1): Divide probability between
6132 both conditions of a TRUTH_ANDIF_EXPR/TRUTH_ORIF_EXPR.
6134 2013-10-02 Tom Tromey <tromey@redhat.com>
6136 * Makefile.in (DRIVER_DEFINES): Use $(if), not $(and).
6138 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
6140 * tree-flow.h: Remove some prototypes.
6141 * tree-ssa-dce.c (mark_virtual_operand_for_renaming,
6142 mark_virtual_phi_result_for_renaming): Move to tree-into-ssa.c.
6143 * tree-into-ssa.c (mark_virtual_operand_for_renaming,
6144 mark_virtual_phi_result_for_renaming): Relocate here.
6145 * tree-into-ssa.h: Add prototypes.
6146 * tree-ssa-phiopt.c: (tree_ssa_phiopt_worker) Use
6147 single_pred_before_succ_order.
6148 (blocks_in_phiopt_order): Rename and move to cfganal.c.
6149 (nonfreeing_call_p) Move to gimple.c.
6150 * cfganal.c (single_pred_before_succ_order): Move and renamed from
6152 * basic-block.h (single_pred_before_succ_order): Add prototype.
6153 * gimple.c (nonfreeing_call_p): Relocate here.
6154 * gimple.h: Add prototype.
6155 * tree-ssa-ifcombine.c: Include tree-ssa-phiopt.h.
6156 * tree-ssa-dom.h: New file. Relocate prototypes here.
6157 * tree-ssa.h: Include tree-ssa-dom.h.
6159 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
6161 * config/i386/x-i386 (driver-i386.o): Remove header dependencies.
6162 Use $(COMPILE) and $(POSTCOMPILE).
6164 * config/alpha/x-alpha (driver-alpha.o): Ditto.
6166 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
6168 * tree-flow.h: Remove some prototypes.
6169 * gimple-fold.h: Add prototypes from gimple.h and tree-flow.h.
6170 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
6171 * tree-ssa-copy.c (may_propagate*, propagate_value, replace_exp,
6172 propagate_tree_value*): Move from here to...
6173 * tree-ssa-propagate.c (may_propagate*, propagate_value, replace_exp,
6174 propagate_tree_value*): Relocate here.
6175 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
6176 * gimple.h: Include gimple-fold.h, move prototypes into gimple-fold.h.
6177 * gimple-fold.c: Remove gimple-fold.h from include list.
6178 * tree-vrp.c: Remove gimple-fold.h from include list.
6179 * tree-ssa-sccvn.c: Remove gimple-fold.h from include list.
6180 * tree-ssa-ccp.c: Remove gimple-fold.h from include list.
6181 * tree-scalar-evolution.c: Add tree-ssa-propagate.h to include list.
6182 * tree-ssa-pre.c: Add tree-ssa-propagate.h to include list.
6183 * sese.c: Add tree-ssa-propagate.h to include list.
6185 2013-10-02 Richard Biener <rguenther@suse.de>
6187 * tree-loop-distribution.c: Include tree-vectorizer.h for
6189 (enum partition_kind): Remove PKIND_REDUCTION.
6190 (struct partition_s): Remove has_writes member, add reduction_p member.
6191 (partition_alloc): Adjust.
6192 (partition_builtin_p): Likewise.
6193 (partition_has_writes): Remove.
6194 (partition_reduction_p): New function.
6195 (partition_merge_into): Likewise.
6196 (generate_code_for_partition): Commonize builtin partition
6198 (rdg_cannot_recompute_vertex_p): Remove.
6199 (already_processed_vertex_p): Likewise.
6200 (rdg_flag_vertex): Do not set has_writes.
6201 (classify_partition): Adjust.
6202 (rdg_build_partitions): Do not set has_writes, treat all
6203 partitions as useful.
6204 (distribute_loop): Record number of library calls generated. Adjust.
6205 (tree_loop_distribution): Report number of loops and library
6206 calls generated as opt-info.
6208 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
6210 * tree-flow.h: Include new .h files. Move prototypes.
6211 * tree-cfgcleanup.h: New file. Add prototypes from tree-flow.h.
6212 * tree-dfa.h: New File. Add prototypes from tree-flow.h.
6213 (get_addr_base_and_unit_offset_1) Move from tree-flow-inline.h.
6214 * tree-pretty-print.h: Add prototypes from tree-flow.h.
6215 * tree-into-ssa.h: New File. Add prototypes from tree-flow.h.
6216 ({debug|dump}*): Move debugging prototypes out of tree-into-ssa.c.
6217 * tree-into-ssa.c ({debug|dump}*): Move prototypes to header file.
6218 * tree.h (get_ref_base_and_extent): Move prototype out.
6219 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Move to
6221 * gimple-low.h: New File. Add prototypes from tree-flow.h.
6222 * gimple-low.c (try_catch_may_fallthru, block_may_fallthru): Move to...
6223 * tree.c (try_catch_may_fallthru, block_may_fallthru): Here.
6224 * tree-scalar-evolution.c: Include tree.h.
6225 * sese.c: Include tree.h.
6226 * dumpfile.c: Move gimple-pretty-print.h include after tree.h.
6227 * dwarf2out.c: Include tree-dfa.h.
6228 * tree-chrec.c: Include tree.h.
6229 * tree-data-ref.c: Include tree.h.
6231 2013-10-02 Yufeng Zhang <yufeng.zhang@arm.com>
6233 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref):
6236 2013-10-02 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
6238 * config/t-sol2 (sol2-c.o): Remove header dependencies.
6239 Use $(COMPILE) and $(POSTCOMPILE).
6240 (sol2-cxx.o): Likewise.
6241 (sol2-stubs.o): Likewise.
6243 * config/x-solaris (host-solaris.o): Likewise.
6245 * config/sparc/t-sparc (sparc.o): Remove.
6246 (sparc-c.o): Remove header dependencies.
6247 Use $(COMPILE) and $(POSTCOMPILE).
6248 * config/sparc/x-sparc: Likewise.
6250 2013-10-02 Joern Rennecke <joern.rennecke@embecosm.com>
6252 * config/arc/arc-opts.h: Add 2013 to Copyright years.
6253 * config/arc/arc700.md: Likewise.
6254 * config/arc/arc-modes.def: Likewise.
6255 * config/arc/arc-simd.h: Likewise.
6256 * config/arc/t-arc-uClibc: Likewise.
6257 * config/arc/t-arc-newlib: Likewise.
6259 2013-10-02 Renlin Li <renlin.li@arm.com>
6261 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
6263 (aarch64_expand_epilogue): Likewise.
6265 2013-10-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6266 Yufeng Zhang <yufeng.zhang@arm.com>
6268 * gimple-ssa-strength-reduction.c (legal_cast_p_1): Forward
6270 (backtrace_base_for_ref): Call get_unwidened with 'base_in' if
6271 'base_in' represent a conversion and legal_cast_p_1 holds; set
6272 'base_in' with the returned value from get_unwidened.
6274 2013-10-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6276 * config/arm/arm.c (arm_legitimize_reload_address): Explain why
6277 plus_constant is not used.
6279 2013-10-01 Wei Mi <wmi@google.com>
6281 * config/i386/x86-tune.def (DEF_TUNE): Remove m_CORE_ALL.
6282 * config/i386/i386.md: Add define_peephole2 to
6283 break partial reg stall for cvtss2sd/cvtsd2ss.
6285 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
6287 * config/arc/arc.c (pass_arc_ifcvt::clone):
6288 Update for ctxt_ -> m_ctxt change.
6290 2013-10-01 Jeff Law <law@redhat.com>
6292 * tree-ssa-threadupdate.c (struct redirection_data): Delete
6293 outgoing_edge and intermediate_edge fields. Instead store the path.
6294 (redirection_data::hash): Hash on the last edge's destination index.
6295 (redirection_data::equal): Check the entire thread path.
6296 (lookup_redirectio_data): Corresponding changes.
6297 (create_edge_and_update_destination_phis): Likewise.
6298 (thread_single_edge): Likewise.
6300 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
6301 Diego Novillo <dnovillo@google.com>
6303 * config/arc/simdext.md (UNSPEC_ARC_SIMD_VLD32WH): Delete.
6304 (UNSPEC_ARC_SIMD_VLD32WL): Likewise.
6305 (vld32wh_insn, vld32wl_insn): Delete commented-out old
6306 versions of these patterns.
6308 * doc/extend.texi (long_call/medium_call/short_call): Typo fix.
6309 (__builtin_arc_aligned): Likewise.
6311 * config/arc/arc.md: Expand adc_0 comment stating the intended
6312 purpose and why it isn't ready.
6313 Replace commented out call_value_via_label_mixed with a
6314 plain comment about bl_s.
6316 * config/arc/arc.c (stdio.h): Don't include directly.
6317 (arc_expand_epilogue): Remove [0]: Remove fp_restored_p.
6318 Remove if (1) condition.
6319 (arc_encode_section_info): Fix comment.
6321 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
6323 * config/arc/arc.c (arc_conditional_register_usage):
6324 Use ARC_FIRST_SIMD_VR_REG / ARC_LAST_SIMD_VR_REG.
6325 Also set reg_alloc_order for DMA config regs.
6327 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
6328 Jeremy Bennett <jeremy.bennett@embecosm.com>
6330 * doc/install.texi (--with-cpu): Mention ARC.
6331 (arc-*-elf32): New paragraph.
6332 (arc-linux-uclibc): Likewise.
6333 * doc/md.texi (Machine Constraints): Add ARC part.
6334 * doc/invoke.texi: (menu): Add ARC Options.
6335 (Machine Dependent Options) <ARC Options>: Add synopsis.
6336 (node ARC Options): Add.
6337 * doc/extend.texi (long_call / short_call attribute): Add ARC.
6338 (ARC Built-in Functions): New section defining
6339 generic ARC built-in functions.
6340 (ARC SIMD Built-in Functions): New section defining SIMD specific
6342 (Declaring Attributes of Functions): Extended
6343 description of short_call and long_call attributes for ARC and
6344 added index entries.
6346 2013-10-01 Saurabh Verma <saurabh.verma@codito.com>
6347 Ramana Radhakrishnan <ramana.radhakrishnan@codito.com>
6348 Joern Rennecke <joern.rennecke@embecosm.com>
6349 Muhammad Khurram Riaz <khurram.riaz@arc.com>
6350 Brendan Kehoe <brendan@zen.org>
6351 Michael Eager <eager@eagercon.com>
6352 Simon Cook <simon.cook@embecosm.com>
6353 Jeremy Bennett <jeremy.bennett@embecosm.com>
6355 * config/arc, common/config/arc: New directories.
6357 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
6358 Brendan Kehoe <brendan@zen.org>
6359 Simon Cook <simon.cook@embecosm.com>
6361 * config.gcc (arc*-*-elf*, arc*-*-linux-uclibc*): New configurations.
6363 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
6365 * tree-ssa-live.h (coalesce_ssa_name): Move Prototype to...
6366 * tree-ssa-coalesce.h: New. Move prototype to here.
6367 * tree-outof-ssa.h: Include tree-ssa-coalesce.h.
6368 * tree-ssa-coalesce.c: Include tree-outof-ssa.h.
6369 (gimple_can_coalesce_p): Move to...
6370 * gimple.c (gimple_can_coalesce_p): Here.
6372 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
6374 * tree-into-ssa.c (enum need_phi_state): Relocate from tree-flow.h.
6375 (dump_decl_set): Move to gimple.c.
6376 * gimple.h: Don't include tree-ssa-operands.h.
6377 (dump_decl_set): Add prototype.
6378 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
6379 Move to gimple-ssa.h.
6380 (phi_ssa_name_p, phi_nodes, phi_nodes_ptr, gimple_phi_arg_def,
6381 gimple_phi_arg_def_ptr, gimple_phi_arg_edge, gimple_phi_arg_location,
6382 gimple_phi_arg_location_from_edge, gimple_phi_arg_set_location,
6383 gimple_phi_arg_has_location): Relocate from tree-flow-inline.h
6384 * gimple.c (walk_stmt_load_store_ops): Use gimple_phi_arg_def rather
6386 (dump_decl_set): Relocate here.
6387 * gimple-ssa.h: New file.
6388 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
6389 Relocate from gimple.h.
6390 * tree-cfg.c (has_zero_uses_1, single_imm_use_1): Move to...
6391 * tree-ssa-operands.c (swap_ssa_operands): Rename from
6392 swap_tree_operands and remove non-ssa path.
6393 (has_zero_uses_1, single_imm_use_1): Relocate from tree-cfg.c.
6394 * tree-ssa-reassoc.c (linearize_expr_tree, repropagate_negates): Use
6396 * tree-vect-loop.c (destroy_loop_vec_info, vect_is_slp_reduction,
6397 vect_is_simple_reduction_1): Use swap_ssa_operands.
6398 * tree-flow.h: Move various prototypes to tree-phinodes.h.
6399 (enum need_phi_state): Move to tree-into-ssa.c.
6400 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
6401 BREAK_FROM_IMM_USE_STMT): Move to ssa-iterators.h.
6402 (swap_tree_operands): Rename and move prototype to tree-ssa-operands.h.
6403 * tree-flow-inline.h (delink_imm_use, link_imm_use_to_list,
6404 link_imm_use, set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
6405 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
6406 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
6407 num_imm_uses): Move to ssa-iterators.h.
6408 (get_use_from_ptr, get_def_from_ptr): Move to tree-ssa-operands.h
6409 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Move to
6411 (op_iter_done, op_iter_next_def, op_iter_next_tree,
6412 clear_and_done_ssa_iter, op_iter_init, op_iter_init_use,
6413 op_iter_init_def, op_iter_init_tree, single_ssa_tree_operand,
6414 single_ssa_use_operand, single_ssa_def_operand, zero_ssa_operands,
6415 num_ssa_operands, delink_stmt_imm_use, single_phi_def,
6416 op_iter_init_phiuse, op_iter_init_phidef, end_imm_use_stmt_p,
6417 end_imm_use_stmt_traverse, move_use_after_head, link_use_stmts_after,
6418 first_imm_use_stmt, next_imm_use_stmt, first_imm_use_on_stmt,
6419 end_imm_use_on_stmt_p, next_imm_use_on_stmt): Move to ssa-iterators.h.
6420 (gimple_phi_arg_def, gimple_phi_arg_def_ptr, gimple_phi_arg_edge,
6421 gimple_phi_arg_location, gimple_phi_arg_location_from_edge,
6422 gimple_phi_arg_set_location, gimple_phi_arg_has_location, phi_nodes,
6423 phi_nodes_ptr, phi_ssa_name_p): Move to gimple.h.
6424 (set_phi_nodes): Move to tree-phinodes.h.
6425 * tree-ssa-operands.h (enum ssa_op_iter_type,
6426 struct ssa_operand_iterator_d, SSA_OP*, FOR_EACH_SSA*, SINGLE_SSA*,
6427 ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS): Move to ssa-iterators.h.
6428 (dump_decl_set): Remove prototype.
6429 (get_use_from_ptr, get_def_from_ptr): Relocate from tree-flow.h.
6430 * tree-phinodes.h: New file. Move some prototypes from tree-flow.h.
6431 (set_phi_nodes): Relocate from tree-flow-inline.h.
6432 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Relocate from
6434 * tree-ssa.h: Add tree-phinodes.h, gimple-ssa.h, ssa-iterators.h to
6435 include list. Temporarily add gimple.h to include list.
6436 * ssa-iterators.h: New file.
6437 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
6438 BREAK_FROM_IMM_USE_STMT): Relocate from tree-flow.h.
6439 (enum ssa_op_iter_type, struct ssa_operand_iterator_d, SSA_OP*,
6440 FOR_EACH_SSA*, SINGLE_SSA*, ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS):
6441 Relocate from tree-ssa-operands.h.
6442 (delink_imm_use, link_imm_use_to_list, link_imm_use,
6443 set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
6444 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
6445 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
6446 num_imm_uses, get_use_from_ptr, get_def_from_ptr,
6447 phi_arg_index_from_use, op_iter_done, op_iter_next_def,
6448 op_iter_next_tree, clear_and_done_ssa_iter, op_iter_init,
6449 op_iter_init_use, op_iter_init_def, op_iter_init_tree,
6450 single_ssa_tree_operand, single_ssa_use_operand, single_ssa_def_operand,
6451 zero_ssa_operands, num_ssa_operands, delink_stmt_imm_use,
6452 single_phi_def, op_iter_init_phiuse, op_iter_init_phidef,
6453 end_imm_use_stmt_p, end_imm_use_stmt_traverse, move_use_after_head,
6454 link_use_stmts_after, first_imm_use_stmt, next_imm_use_stmt,
6455 first_imm_use_on_stmt, end_imm_use_on_stmt_p, next_imm_use_on_stmt):
6456 Relocate from tree-flow-inline.h.
6457 * tree-outof-ssa.h: Change _SSAEXPAND_H macro to GCC_TREE_OUTOF_SSA_H.
6459 2013-10-01 Vidya Praveen <vidyapraveen@arm.com>
6462 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l2<mode>_internal): Rename to ...
6463 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): ... this;
6464 Insert '\t' to output template.
6465 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): New.
6466 (aarch64_saddl2<mode>, aarch64_uaddl2<mode>): Modify to call
6467 gen_aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal() instead.
6468 (aarch64_ssubl2<mode>, aarch64_usubl2<mode>): Ditto.
6470 2013-10-01 Uros Bizjak <ubizjak@gmail.com>
6472 * doc/install.texi (Host/target specific installation notes for GCC):
6473 Put @anchor before @heading.
6474 * doc/gcc.texi (titlepage): Use @uref and http:// prefix for website.
6475 Use @email for email addresses.
6477 2013-10-01 Jeff Law <law@redhat.com>
6479 * tree-ssa-threadedge.c (thread_across_edge): Make path a pointer to
6480 a vec. Only delete the path if we create one without successfully
6481 registering a jump thread.
6482 * tree-ssa-threadupdate.h (register_jump_thread): Pass in path vector
6484 * tree-ssa-threadupdate.c (threaded_edges): Remove. No longer used
6485 (paths): New vector of jump threading paths.
6486 (THREAD_TARGET, THREAD_TARGET2): Remove accessor macros.
6487 (THREAD_PATH): New accessor macro for the entire thread path.
6488 (lookup_redirection_data): Get intermediate and final outgoing edge
6489 from the thread path.
6490 (create_edge_and_update_destination_phis): Copy the threading path.
6491 (ssa_fix_duplicate_block_edges): Get edges and block types from the
6492 jump threading path.
6493 (ssa_redirect_edges): Get edges and block types from the jump threading
6494 path. Free the path vector.
6495 (thread_block): Get edges from the jump threading path. Look at the
6496 entire path to see if we thread to a loop exit. If we cancel a jump
6497 thread request, then free the path vector.
6498 (thread_single_edge): Get edges and block types from the jump threading
6499 path. Free the path vector.
6500 (thread_through_loop_header): Get edges and block types from the jump
6501 threading path. Free the path vector.
6502 (mark_threaded_blocks): Iterate over the vector of paths and store
6503 the path on the appropriate edge. Get edges and block types from the
6504 jump threading path.
6505 (mark_threaded_blocks): Get edges and block types from the jump
6506 threading path. Free the path vector.
6507 (thread_through_all_blocks): Use the vector of paths rather than
6508 a vector of 3-edge sets.
6509 (register_jump_thread): Accept pointer to a path vector rather
6510 than the path vector itself. Store the path vector for later use.
6513 2013-10-01 Jakub Jelinek <jakub@redhat.com>
6514 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6517 * config/s390/s390.c (s390_split_branches): Modify check for table
6519 (s390_chunkify_start): Rearrange table jump insn check in order to
6520 deal with compare and branch insns correctly.
6522 2013-10-01 Kugan Vivekanandarajah <kuganv@linaro.org>
6526 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
6527 * config/arm/arm.md (arm_ashldi3_1bit): define_insn into
6528 define_insn_and_split.
6529 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
6530 (shiftsi3_compare): New pattern.
6532 * config/arm/unspecs.md (UNSPEC_RRX): New.
6534 2013-10-01 Alan Modra <amodra@gmail.com>
6536 * stmt.c (expand_asm_operands): Revert part of 2013-09-24 special
6537 casing inout operands.
6539 2013-10-01 Richard Biener <rguenther@suse.de>
6541 PR tree-optimization/58553
6542 * tree-loop-distribution.c (struct partition_s): Add niter member.
6543 (classify_partition): Populate niter member for the partition
6544 and properly identify whether the relevant store happens before
6545 or after the loop exit.
6546 (generate_memset_builtin): Use niter member from the partition.
6547 (generate_memcpy_builtin): Likewise.
6549 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
6551 * vec.h (vec_prefix, vec): Prefix member names with "m_".
6552 * vec.c (vec_prefix::calculate_allocation): Update accordingly.
6554 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
6556 * basic-block.h (edge_list): Prefix member names with "m_".
6557 * context.h (context): Likewise.
6558 * domwalk.h (dom_walker): Likewise.
6559 * gengtype-state.c (s_expr_writer, state_writer): Likewise.
6560 * graphite-sese-to-poly.c (sese_dom_walker): Likewise.
6561 * hash-table.h (hash_table): Likewise.
6562 * machmode.h (bit_field_mode_iterator): Likewise.
6563 * pass_manager.h (pass_list): Likewise.
6564 * tree-into-ssa.c (mark_def_dom_walker): Likewise.
6565 * tree-pass.h (pass_data): Likewise.
6566 * tree-ssa-dom.c (dom_opt_dom_walker): Likewise.
6567 * tree-ssa-phiopt.c (nontrapping_dom_walker): Likewise,
6568 * tree-ssa-uncprop.c (uncprop_dom_walker): Likewise.
6569 * asan.c (pass_data_asan): Update accordingly.
6570 * cfganal.c (control_dependences::find_control_dependence): Likewise.
6571 (control_dependences::control_dependences): Likewise.
6572 (control_dependences::~control_dependences): Likewise.
6573 (control_dependences::~control_dependences): Likewise.
6574 (control_dependences::get_edges_dependent_on): Likewise.
6575 * cgraphbuild.c (pass_data_rebuild_cgraph_edges::clone): Likewise.
6576 (pass_data_remove_cgraph_callee_edges::clone): Likewise.
6577 * context.c (gcc::context::context): Likewise.
6578 * cprop.c (pass_rtl_cprop::clone): Likewise.
6579 * domwalk.c (dom_walker::walk): Likewise.
6580 * ipa-inline-analysis.c (pass_inline_parameters::clone): Likewise.
6581 * ipa-pure-const.c (pass_local_pure_const::clone): Likewise.
6582 * mode-switching.c (pass_mode_switching::clone): Likewise.
6583 * passes.c (opt_pass::opt_pass): Likewise.
6584 (pass_manager::pass_manager): Likewise.
6585 * predict.c (pass_strip_predict_hints::clone): Likewise.
6586 * recog.c (pass_data pass_data_peephole2::clone): Likewise.
6587 (pass_split_all_insns::clone): Likewise.
6588 * stor-layout.c (bit_field_mode_iterator::bit_field_mode_iterator):
6590 (bit_field_mode_iterator::next_mode): Likewise.
6591 (bit_field_mode_iterator::prefer_smaller_modes): Likewise.
6592 * tree-cfg.c (pass_split_crit_edges::clone): Likewise.
6593 * tree-cfgcleanup.c (pass_merge_phi::clone): Likewise.
6594 * tree-complex.c (pass_lower_complex::clone): Likewise.
6595 * tree-eh.c (pass_cleanup_eh::clone): Likewise.
6596 * tree-object-size.c (pass_object_sizes::clone): Likewise.
6597 * tree-optimize.c (pass_fixup_cfg::clone): Likewise.
6598 * tree-ssa-ccp.c (pass_data_ccp::clone): Likewise.
6599 (pass_fold_builtins::clone): Likewise.
6600 * tree-ssa-copy.c (pass_data_copy_prop::clone): Likewise.
6601 * tree-ssa-copyrename.c (pass_rename_ssa_copies::clone): Likewise.
6602 * tree-ssa-dce.c (pass_dce::clone, pass_dce_loop::clone): Likewise.
6603 (pass_cd_dce::clone): Likewise.
6604 * tree-ssa-dom.c (pass_dominator::clone): Likewise.
6605 (pass_phi_only_cprop::clone): Likewise.
6606 * tree-ssa-dse.c (pass_dse::clone): Likewise.
6607 * tree-ssa-forwprop.c (pass_forwprop::clone): Likewise.
6608 * tree-ssa-loop.c (pass_lim::clone): Likewise.
6609 * tree-ssa-phiopt.c (pass_phiopt::clone): Likewise.
6610 * tree-ssa-pre.c (pass_fre::clone): Likewise.
6611 * tree-ssa-reassoc.c (pass_reassoc::clone): Likewise.
6612 * tree-ssa-uninit.c (pass_late_warn_uninitialized::clone): Likewise.
6613 * tree-tailcall.c (pass_tail_recursion::clone): Likewise.
6614 * tree-vect-generic.c (pass_lower_vector_ssa::clone): Likewise.
6615 * tree-vrp.c (pass_vrp::clone): Likewise.
6616 * tsan.c (pass_tsan::clone): Likewise.
6618 2013-09-30 Jakub Jelinek <jakub@redhat.com>
6621 * fold-const.c (tree_unary_nonnegative_warnv_p): Use
6622 INTEGRAL_TYPE_P (t) instead of TREE_CODE (t) == INTEGER_TYPE.
6625 * fold-const.c (fold_ternary_loc): For A < 0 : <sign bit of A> : 0
6626 optimization, punt if sign_bit_p looked through any zero extension.
6628 2013-09-30 Teresa Johnson <tejohnson@google.com>
6630 * tree-ssa-threadupdate.c (ssa_fix_duplicate_block_edges):
6631 Update redirected out edge count in joiner case.
6632 (ssa_redirect_edges): Common the joiner and non-joiner cases
6633 so that joiner case gets profile updates.
6635 2013-09-30 Richard Biener <rguenther@suse.de>
6637 PR tree-optimization/58554
6638 * tree-loop-distribution.c (classify_partition): Require
6639 unconditionally executed stores for memcpy and memset recognition.
6640 (tree_loop_distribution): Calculate dominance info.
6642 2013-09-30 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
6644 * config/aarch64/aarch64.h (MCOUNT_NAME): Define.
6645 (NO_PROFILE_COUNTERS): Likewise.
6646 (PROFILE_HOOK): Likewise.
6647 (FUNCTION_PROFILER): Likewise.
6648 * config/aarch64/aarch64.c (aarch64_function_profiler): Remove.
6650 2013-09-30 Iain Sandoe <iain@codesourcery.com>
6652 * config/rs6000/darwin.md (load_macho_picbase_si): Wrap machopic
6653 calls and defines in TARGET_MACHO conditional.
6654 (load_macho_picbase_di): Likewise.
6655 (reload_macho_picbase): Likewise.
6656 (reload_macho_picbase_si): Likewise.
6657 (reload_macho_picbase_di): Likewise.
6658 (nonlocal_goto_receiver): Likewise.
6660 2013-09-30 Nick Clifton <nickc@redhat.com>
6662 * config/msp430/msp430.c (msp430x_names): New array. Lists MCUs
6663 that use the MSP430X ISA.
6664 (msp430_option_override): Scan -mmcu command line option for any
6665 MCU name that supports the MSP430X ISA.
6666 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add matches for known
6667 -mmcu options which enable the MSP430X ISA.
6669 2013-09-30 Richard Biener <rguenther@suse.de>
6672 * tree-cfg.c (make_abnormal_goto_edges): Skip debug statements
6673 before looking for setjmp-like calls.
6675 2013-09-29 Iain Sandoe <iain@codesourcery.com>
6678 * config/darwin-protos.h (machopic_get_function_picbase): New.
6679 * config/darwin.c (machopic_get_function_picbase): New.
6680 * config/rs6000/darwin.md (load_macho_picbase_si): Update picbase
6681 label for a new func. (load_macho_picbase_di): Likewise.
6682 (reload_macho_picbase): New expand.
6683 (reload_macho_picbase_si): New insn.
6684 (reload_macho_picbase_di): New insn.
6685 (nonlocal_goto_receiver): New define and split.
6686 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_RELD_MPIC.
6687 (unspecv enum): Add UNSPECV_NLGR.
6689 2013-09-29 Iain Sandoe <iain@codesourcery.com>
6691 * config/rs6000/rs6000.c (rs6000_init_dwarf_reg_sizes_extra): Ensure
6692 that altivec registers are correctly sized on Darwin.
6694 2013-09-29 Iain Sandoe <iain@codesourcery.com>
6696 * config/t-darwin (darwin.o, darwin-c.o, darwin-f.o,
6697 darwin-driver.o): Use COMPILE and POSTCOMPILE.
6698 * config/x-darwin (host-darwin.o): Likewise.
6699 * config/i386/x-darwin (host-i386-darwin.o): Likewise.
6700 * config/rs6000/x-darwin (host-ppc-darwin.o): Likewise.
6701 * config/rs6000/x-darwin64 (host-ppc64-darwin.o): Likewise.
6703 2013-09-29 Uros Bizjak <ubizjak@gmail.com>
6705 * doc/invoke.texi: Fix usage of @tie{} command.
6707 2013-09-29 Eric Botcazou <ebotcazou@adacore.com>
6709 * config/sparc/sync.md: Add peephole for consecutive memory barriers.
6711 2013-09-28 Jan Hubicka <jh@suse.cz>
6713 * config/i386/x86-tune.def: Add documentation for each of the options;
6716 2013-09-28 Jan Hubicka <jh@suse.cz>
6718 * x86-tune.def (X86_TUNE_SSE_UNALIGNED_LOAD_OPTIMAL): Enable for
6720 (X86_TUNE_SSE_UNALIGNED_STORE_OPTIMAL): Likewise.
6721 (X86_TUNE_FOUR_JUMP_LIMIT): Drop for generic and buldozer.
6722 (X86_TUNE_PAD_RETURNS): Drop for buldozer chips.
6723 (X86_TUNE_AVOID_VECTOR_DECODE): Drop for generic.
6724 (X86_TUNE_REASSOC_FP_TO_PARALLEL): Enable for generic.
6726 2013-09-28 Richard Sandiford <rdsandiford@googlemail.com>
6728 * alloc-pool.c, asan.c, auto-inc-dec.c, basic-block.h, bb-reorder.c,
6729 bitmap.c, bitmap.h, bt-load.c, builtins.c, calls.c, cfgcleanup.c,
6730 cfgexpand.c, cfghooks.c, cfgloop.c, cfgloopmanip.c, cfgrtl.c, cgraph.c,
6731 cgraph.h, cgraphbuild.c, cgraphclones.c, cgraphunit.c, collect2.c,
6732 combine-stack-adj.c, combine.c, compare-elim.c, context.c, context.h,
6733 cprop.c, cse.c, cselib.c, dbxout.c, dce.c, defaults.h, df-core.c,
6734 df-problems.c, df-scan.c, df.h, diagnostic.c, double-int.c, dse.c,
6735 dumpfile.c, dwarf2asm.c, dwarf2cfi.c, dwarf2out.c, emit-rtl.c,
6736 errors.c, except.c, expmed.c, expr.c, file-find.c, final.c,
6737 fixed-value.c, fold-const.c, function.c, fwprop.c, gcc-ar.c, gcc.c,
6738 gcov-io.c, gcov-io.h, gcov.c, gcse.c, genattr-common.c, genattr.c,
6739 genattrtab.c, genautomata.c, genconfig.c, genemit.c, genextract.c,
6740 genflags.c, gengenrtl.c, gengtype-state.c, gengtype.c, genmodes.c,
6741 genopinit.c, genoutput.c, genpeep.c, genpreds.c, genrecog.c,
6742 gensupport.c, ggc-common.c, ggc-page.c, gimple-fold.c, gimple-low.c,
6743 gimple-pretty-print.c, gimple-ssa-strength-reduction.c, gimple.c,
6744 gimple.h, godump.c, graphite-clast-to-gimple.c,
6745 graphite-optimize-isl.c, graphite-poly.h, graphite-sese-to-poly.c,
6746 graphite.c, haifa-sched.c, hash-table.c, hash-table.h, hwint.c,
6747 hwint.h, ifcvt.c, incpath.c, init-regs.c, input.h, intl.c, intl.h,
6748 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
6749 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c,
6750 ipa-utils.c, ipa.c, ira-build.c, ira.c, jump.c, loop-doloop.c,
6751 loop-init.c, loop-invariant.c, loop-iv.c, lower-subreg.c, lto-cgraph.c,
6752 lto-streamer-in.c, lto-streamer-out.c, lto-wrapper.c, mcf.c,
6753 mode-switching.c, modulo-sched.c, omp-low.c, optabs.c, opts.c,
6754 pass_manager.h, passes.c, plugin.c, postreload-gcse.c, postreload.c,
6755 predict.c, prefix.c, pretty-print.c, print-rtl.c, print-tree.c,
6756 profile.c, read-md.c, real.c, real.h, recog.c, ree.c, reg-stack.c,
6757 regcprop.c, reginfo.c, regmove.c, regrename.c, regs.h, regstat.c,
6758 reload1.c, reorg.c, rtl.c, rtl.h, rtlanal.c, sbitmap.c, sched-rgn.c,
6759 sdbout.c, sel-sched-ir.c, sel-sched.c, sparseset.c, stack-ptr-mod.c,
6760 statistics.c, stmt.c, stor-layout.c, store-motion.c, streamer-hooks.h,
6761 system.h, target-hooks-macros.h, targhooks.c, targhooks.h, toplev.c,
6762 tracer.c, trans-mem.c, tree-browser.c, tree-call-cdce.c, tree-cfg.c,
6763 tree-cfgcleanup.c, tree-complex.c, tree-data-ref.c, tree-data-ref.h,
6764 tree-eh.c, tree-emutls.c, tree-flow.h, tree-if-conv.c, tree-into-ssa.c,
6765 tree-iterator.c, tree-loop-distribution.c, tree-mudflap.c,
6766 tree-nested.c, tree-nomudflap.c, tree-nrv.c, tree-object-size.c,
6767 tree-optimize.c, tree-pass.h, tree-pretty-print.c, tree-profile.c,
6768 tree-scalar-evolution.c, tree-sra.c, tree-ssa-ccp.c,
6769 tree-ssa-coalesce.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
6770 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
6771 tree-ssa-ifcombine.c, tree-ssa-live.c, tree-ssa-loop-ch.c,
6772 tree-ssa-loop-im.c, tree-ssa-loop-ivopts.c, tree-ssa-loop-prefetch.c,
6773 tree-ssa-loop.c, tree-ssa-math-opts.c, tree-ssa-operands.c,
6774 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
6775 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
6776 tree-ssa-structalias.c, tree-ssa-threadedge.c, tree-ssa-threadupdate.c,
6777 tree-ssa-uncprop.c, tree-ssa-uninit.c, tree-ssa.c, tree-ssanames.c,
6778 tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
6779 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
6780 tree-vect-stmts.c, tree-vectorizer.c, tree-vectorizer.h, tree-vrp.c,
6781 tree.c, tree.h, tsan.c, tsystem.h, value-prof.c, var-tracking.c,
6782 varasm.c, vec.h, vmsdbgout.c, vtable-verify.c, web.c: Add missing
6783 whitespace before "(".
6785 2013-09-28 Sandra Loosemore <sandra@codesourcery.com>
6787 * expr.h (extract_bit_field): Remove packedp parameter.
6788 * expmed.c (extract_fixed_bit_field): Remove packedp parameter
6789 from forward declaration.
6790 (store_split_bit_field): Remove packedp arg from calls to
6791 extract_fixed_bit_field.
6792 (extract_bit_field_1): Remove packedp parameter and packedp
6793 argument from recursive calls and calls to extract_fixed_bit_field.
6794 (extract_bit_field): Remove packedp parameter and corresponding
6795 arg to extract_bit_field_1.
6796 (extract_fixed_bit_field): Remove packedp parameter. Remove code
6798 (extract_split_bit_field): Remove packedp arg from call to
6799 extract_fixed_bit_field.
6800 * expr.c (emit_group_load_1): Adjust calls to extract_bit_field.
6801 (copy_blkmode_from_reg): Likewise.
6802 (copy_blkmode_to_reg): Likewise.
6803 (read_complex_part): Likewise.
6804 (store_field): Likewise.
6805 (expand_expr_real_1): Likewise.
6806 * calls.c (store_unaligned_arguments_into_pseudos): Adjust call
6807 to extract_bit_field.
6808 * config/tilegx/tilegx.c (tilegx_expand_unaligned_load): Adjust
6809 call to extract_bit_field.
6810 * config/tilepro/tilepro.c (tilepro_expand_unaligned_load): Adjust
6811 call to extract_bit_field.
6812 * doc/invoke.texi (Code Gen Options): Remove mention of warnings
6813 and special packedp behavior from -fstrict-volatile-bitfields
6816 2013-09-27 Jan-Benedict Glaw <jbglaw@lug-owl.de>
6818 * lra-eliminations.c (init_elim_table): Guard value_p.
6820 2013-09-27 Michael Meissner <meissner@linux.vnet.ibm.com>
6822 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Allow
6823 DFmode, DImode, and SFmode in the upper VSX registers based on the
6824 -mupper-regs-{df,sf} flags. Fix wu constraint to be ALTIVEC_REGS
6825 if -mpower8-vector. Combine -mvsx-timode handling with the rest
6826 of the VSX register handling.
6828 * config/rs6000/rs6000.md (f32_lv): Use %x0 for VSX regsters.
6830 (zero_extendsidi2_lfiwzx): Add support for loading into the
6831 Altivec registers with -mpower8-vector. Use wu/wv constraints to
6832 only do VSX memory options on Altivec registers.
6833 (extendsidi2_lfiwax): Likewise.
6834 (extendsfdf2_fpr): Likewise.
6835 (mov<mode>_hardfloat, SF/SD modes): Likewise.
6836 (mov<mode>_hardfloat32, DF/DD modes): Likewise.
6837 (mov<mode>_hardfloat64, DF/DD modes): Likewise.
6838 (movdi_internal64): Likewise.
6840 2013-09-27 Xinliang David Li <davidxl@google.com>
6842 * opts.c (finish_options): Adjust parameters
6843 according to vect cost model.
6844 (common_handle_option): Set dynamic vect cost
6846 targhooks.c (default_add_stmt_cost): Compute stmt cost
6848 * tree-vect-loop.c (vect_estimate_min_profitable_iters):
6849 Use helper function.
6850 * tree-vectorizer.h (unlimited_cost_model): New function.
6851 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use helper function.
6852 * tree-vect-data-refs.c (vect_peeling_hash_insert): Use helper
6854 (vect_enhance_data_refs_alignment): Ditto.
6855 * flag-types.h: New enum.
6856 * common/config/i386/i386-common.c (ix86_option_init_struct):
6857 No need to initialize vect_cost_model flag.
6858 * config/i386/i386.c (ix86_add_stmt_cost): Compute stmt cost
6861 2013-09-27 Diego Novillo <dnovillo@google.com>
6863 * gimple.h (enum ssa_mode): Remove.
6865 2013-09-27 Paulo Matos <pmatos@broadcom.com>
6867 * cfgloop.h (number_of_loops): Fix typo in check for null.
6869 2013-09-27 Jakub Jelinek <jakub@redhat.com>
6872 * tree-cfg.c (move_sese_region_to_fn): Also move loops that
6873 are children of outermost saved_cfun's loop, and set it up to
6874 be moved to dest_cfun's outermost loop. Fix up num_nodes adjustments
6875 if loop != loop0 and SESE region contains bbs that belong to loop0.
6877 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
6879 * rtlanal.c (must_be_base_p, must_be_index_p): Delete.
6880 (binary_scale_code_p, get_base_term, get_index_term): New functions.
6881 (set_address_segment, set_address_base, set_address_index)
6882 (set_address_disp): Accept the argument unconditionally.
6883 (baseness): Remove must_be_base_p and must_be_index_p checks.
6884 (decompose_normal_address): Classify as much as possible in the
6887 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
6889 * cse.c (count_reg_usage): Handle INT_LIST.
6890 * lra-eliminations.c (lra_eliminate_regs_1): Likewise.
6891 * reginfo.c (reg_scan_mark_refs): Likewise.
6892 * reload1.c (eliminate_regs_1): Likewise.
6894 2013-09-27 Iain Sandoe <iain@codesourcery.com>
6897 * rtlanal.c (lsb_bitfield_op_p): Make both parts of the comparison
6900 2013-09-27 Richard Biener <rguenther@suse.de>
6902 PR tree-optimization/58459
6903 * tree-ssa-forwprop.c (forward_propagate_addr_expr): Remove
6904 restriction not propagating into loops.
6906 2013-09-26 Florian Weimer <fw@deneb.enyo.de>
6908 * tree-ssa.h (walk_use_def_chains_fn, walk_use_def_chains): Delete.
6909 * tree-ssa.c (walk_use_def_chains_1, walk_use_def_chains): Delete.
6910 * doc/tree-ssa.texi (Walking use-def chains): Delete.
6912 2013-09-26 Richard Biener <rguenther@suse.de>
6914 * tree-into-ssa.c (rewrite_into_ssa): Make more SSA names to anonymous.
6916 2013-09-26 Richard Biener <rguenther@suse.de>
6918 * alias.h (component_uses_parent_alias_set): Rename to ...
6919 (component_uses_parent_alias_set_from): ... this.
6920 * alias.c (component_uses_parent_alias_set): Rename to ...
6921 (component_uses_parent_alias_set_from): ... this and return
6923 (reference_alias_ptr_type_1): Use the result from
6924 component_uses_parent_alias_set_from instead of stripping
6925 components one at a time.
6926 * emit-rtl.c (set_mem_attributes_minus_bitpos): Adjust.
6928 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
6930 * tree-ssa-live.h (find_replaceable_exprs, dump_replaceable_exprs):
6931 Move prototypes to...
6932 * tree-ssa-ter.h: New File. Move prototypes here.
6933 * tree-flow.h (stmt_is_replaceable_p): Remove prototype.
6934 * tree-outof-ssa.h: New. Rename ssaexpand.h, include tree-ssa-ter.h.
6935 * tree-outof-ssa.c (ssa_is_replaceable_p): New. Refactor common bits
6936 from is_replaceable_p.
6937 * tree-ssa-ter.c (is_replaceable_p, stmt_is_replaceable_p): Delete.
6938 (ter_is_replaceable_p): New. Use new refactored ssa_is_replaceable_p.
6939 (process_replaceable): Use ter_is_replaceable_p.
6940 (find_replaceable_in_bb): Use ter_is_replaceable_p.
6941 * expr.c (stmt_is_replaceable_p): Relocate from tree-ssa-ter.c. Use
6942 newly refactored ssa_is_replaceable_p.
6943 * cfgexpand.c: Include tree-outof-ssa.h.
6944 * ssaexpand.h: Delete.
6946 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
6948 * gimple.c (gimple_replace_lhs): Move to tree-ssa.c and rename.
6949 (struct count_ptr_d, count_ptr_derefs, count_uses_and_derefs): Move to
6951 (create_gimple_tmp): Delete.
6952 (get_expr_type, build_assign, build_type_cast): Move to...
6953 * gimple-builder.c: New File.
6954 (get_expr_type): Relocate from gimple.c.
6955 (build_assign, build_type_cast): Change to only create ssanames.
6956 * gimple.h: Move prototypes to...
6957 * gimple-builder.h: New File. Here.
6958 * tree-ssa.h: And here.
6959 * tree-ssa.c (struct count_ptr_d, count_ptr_derefs,
6960 count_uses_and_derefs): Relocate from gimple.c.
6961 (gimple_replace_ssa_lhs): Renamed gimple_replace_ssa from gimple.c
6962 * tree-ssa-reassoc.c (repropagate_negates): Use gimple_replace_ssa_lhs.
6963 * tree-ssa-math-opts (execute_cse_reciprocals): Use
6964 gimple_replace_ssa_lhs.
6965 * asan.c: Include gimple-builder.h.
6966 * Makefile.in: Add gimple-builder.o.
6968 2013-09-26 Richard Biener <rguenther@suse.de>
6970 * tree-ssa-live.c (var_map_base_init): Handle SSA names with
6971 DECL_IGNORED_P base VAR_DECLs like anonymous SSA names.
6972 (loe_visit_block): Use gcc_checking_assert.
6973 * tree-ssa-coalesce.c (create_outofssa_var_map): Use
6974 gimple_assign_ssa_name_copy_p.
6975 (gimple_can_coalesce_p): Adjust according to the var_map_base_init
6978 2013-09-26 David Edelsohn <dje.gcc@gmail.com>
6980 * config/rs6000/t-rs6000 (rs6000.o): Remove.
6981 (rs6000-c.o): Use COMPILE and POSTCOMPILE.
6983 2013-09-26 Richard Biener <rguenther@suse.de>
6985 PR tree-optimization/58539
6986 * tree-vect-loop.c (vect_create_epilog_for_reduction): Honor
6987 the fact that debug statements are not taking part in loop-closed
6990 2013-09-26 Nick Clifton <nickc@redhat.com>
6992 * config/msp430/msp430.c (msp430_expand_epilogue): Fix compile
6993 time warning message.
6994 (msp430_print_operand_raw): Delete unused letter parameter.
6995 (TARGET_PRINT_OPERAND_ADDRESS): Define.
6996 (msp430_print_operand_address): New function.
6997 (msp430_print_operand): Move address printing code from here to
6999 * config/msp430/msp430.md (movsipsi2): Add comment in generated
7001 (zero_extendpsisi2): Likewise.
7002 (extendpsisi2): New pattern.
7003 (andneghi3): New pattern.
7005 2013-09-26 Yvan Roux <yvan.roux@linaro.org>
7007 * config/aarch64/aarch64.opt (mlra): New option.
7008 * config/aarch64/aarch64.c (aarch64_lra_p): New function.
7009 (TARGET_LRA_P): Define.
7011 2013-09-26 Eric Botcazou <ebotcazou@adacore.com>
7013 * expr.c (expand_assignment): Remove obsolete comment.
7015 2013-09-25 Jeff Law <law@redhat.com>
7017 * tree-flow.h (thread_through_all_blocks): Prototype moved into
7018 tree-ssa-threadupdate.h.
7019 (register_jump_thread): Similarly.
7020 * tree-ssa-threadupdate.h: New header file.
7021 * tree-ssa-dom.c: Include tree-ssa-threadupdate.h.
7022 * tree-vrp.c: Likewise.
7023 * tree-ssa-threadedge.c: Include tree-ssa-threadupdate.h.
7024 (thread_around_empty_blocks): Change type of path vector argument to
7025 an edge,type pair from just an edge. Initialize both elements when
7026 appending to a jump threading path. Tweak references to elements
7028 (thread_across_edge): Similarly. Release memory for the elements
7030 * tree-ssa-threadupdate.c: Include tree-ssa-threadupdate.h.
7031 (dump_jump_thread_path): New function broken out from
7032 register_jump_thread.
7033 (register_jump_thread): Use dump_jump_thread_path. Change type of
7034 path vector entries. Search the path for NULL edges and dump
7035 the path if one is found. Tweak the conversion of path to 3-edge
7036 form to use the block copy type information embedded in the path.
7038 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
7040 * lra.c (update_inc_notes): Remove all REG_DEAD and REG_UNUSED notes.
7042 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
7043 Vladimir Makarov <vmakarov@redhat.com>
7045 * rtlanal.c (lsb_bitfield_op_p): New predicate for bitfield operations
7046 from the least significant bit.
7047 (strip_address_mutations): Add bitfield operations handling.
7048 (must_be_index_p): Add shifting and rotate operations handling.
7049 (set_address_base): Use must_be_base_p predicate.
7050 (set_address_index): Use must_be_index_p predicate.
7052 2013-09-25 Alexander Ivchenko <alexander.ivchenko@intel.com>
7053 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
7054 Sergey Lega <sergey.s.lega@intel.com>
7055 Anna Tikhonova <anna.tikhonova@intel.com>
7056 Ilya Tocar <ilya.tocar@intel.com>
7057 Andrey Turetskiy <andrey.turetskiy@intel.com>
7058 Ilya Verbin <ilya.verbin@intel.com>
7059 Kirill Yukhin <kirill.yukhin@intel.com>
7060 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
7062 * config/i386/i386.c (ix86_avx256_split_vector_move_misalign):
7064 (ix86_expand_vector_move_misalign): Support new unaligned load and
7065 stores and use new names.
7066 (CODE_FOR_sse2_storedqu): Rename to ...
7067 (CODE_FOR_sse2_storedquv16qi): ... this.
7068 (CODE_FOR_sse2_loaddqu): Rename to ...
7069 (CODE_FOR_sse2_loaddquv16qi): ... this.
7070 (CODE_FOR_avx_loaddqu256): Rename to ...
7071 (CODE_FOR_avx_loaddquv32qi): ... this.
7072 (CODE_FOR_avx_storedqu256): Rename to ...
7073 (CODE_FOR_avx_storedquv32qi): ... this.
7074 * config/i386/i386.md (fpint_logic): New.
7075 * config/i386/sse.md (VMOVE): Extend for AVX512.
7079 (VI_UNALIGNED_LOADSTORE): Ditto.
7080 (sse2_avx_avx512f): Ditto.
7081 (sse2_avx2): Extend for AVX512.
7082 (sse4_1_avx2): Ditto.
7083 (avx2_avx512f): New.
7084 (sse): Extend for AVX512.
7087 (avxsizesuffix): Ditto.
7088 (sseintvecmode): Ditto.
7090 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Ditto.
7091 (<sse>_storeu<ssemodesuffix><avxsizesuffix>): Ditto.
7092 (<sse2>_loaddqu<avxsizesuffix>): Extend for AVX512 and rename to ...
7093 (<sse2_avx_avx512f>_loaddqu<mode>): ... this.
7094 (<sse2>_storedqu<avxsizesuffix>): Extend for AVX512 and rename to ...
7095 (<sse2_avx_avx512f>_storedqu<mode): ... this.
7096 (<sse>_movnt<mode>): Replace constraint "x" with "v".
7097 (STORENT_MODE): Extend for AVX512.
7098 (*absneg<mode>2): Replace constraint "x" with "v".
7099 (*mul<mode>3): Ditto.
7100 (*ieee_smin<mode>3): Ditto.
7101 (*ieee_smax<mode>3): Ditto.
7102 (avx_cmp<mode>3): Replace VF with VF_128_256.
7103 (*<sse>_maskcmp<mode>3_comm): Ditto.
7104 (<sse>_maskcmp<mode>3): Ditto.
7105 (<sse>_andnot<mode>3): Extend for AVX512.
7106 (<code><mode>3, anylogic): Replace VF with VF_128_256.
7107 (<code><mode>3, fpint_logic): New.
7108 (*<code><mode>3): Extend for AVX512.
7109 (avx512flogicsuff): New.
7110 (avx512f_<logic><mode>): Ditto.
7111 (<sse>_movmsk<ssemodesuffix><avxsizesuffix>): Replace VF with
7113 (<sse4_1>_blend<ssemodesuffix><avxsizesuffix>): Ditto.
7114 (<sse4_1>_blendv<ssemodesuffix><avxsizesuffix>): Ditto.
7115 (<sse4_1>_dp<ssemodesuffix><avxsizesuffix>): Ditto.
7116 (avx_vtest<ssemodesuffix><avxsizesuffix>): Ditto.
7117 (<sse4_1>_round<ssemodesuffix><avxsizesuffix>): Ditto.
7118 (xop_vpermil2<mode>3): Ditto.
7119 (*avx_vpermilp<mode>): Extend for AVX512 and rename to ...
7120 (*<sse2_avx_avx512f>_vpermilp<mode>): ... this.
7121 (avx_vpermilvar<mode>3): Extend for AVX512 and rename to ...
7122 (<sse2_avx_avx512f>_vpermilvar<mode>3): ... this.
7124 2013-09-25 Tom Tromey <tromey@redhat.com>
7126 * Makefile.in (PARTITION_H, LTO_SYMTAB_H, COMMON_TARGET_DEF_H)
7127 (RTL_ERROR_H, TRANS_MEM_H, COVERAGE_H, DEMANGLE_H, ALIAS_H)
7128 (SCHED_INT_H, SEL_SCHED_IR_H, SEL_SCHED_DUMP_H, VALTRACK_H, DDG_H)
7129 (GGC_INTERNAL_H, DECNUM_H, BACKTRACE_H, MKDEPS_H, TREE_HASHER_H)
7130 (TREE_SSA_LIVE_H, SSAEXPAND_H, DWARF2OUT_H, SCEV_H, OMEGA_H)
7131 (TREE_DATA_REF_H, IRA_INT_H, LRA_INT_H, DBGCNT_H, DATA_STREAMER_H)
7132 (GIMPLE_STREAMER_H, TREE_STREAMER_H, STREAMER_HOOKS_H)
7133 (TREE_VECTORIZER_H, IPA_INLINE_H, GSTAB_H, LIBFUNCS_H)
7134 (GRAPHITE_HTAB_H): Remove.
7136 2013-09-25 Tom Tromey <tromey@redhat.com>
7138 * config/mcore/t-mcore (CROSS_FLOAT_H): Remove.
7140 2013-09-25 Tom Tromey <tromey@redhat.com>
7142 * config/t-glibc (glibc-c.o): Use COMPILE and POSTCOMPILE.
7144 2013-09-25 Tom Tromey <tromey@redhat.com>
7146 * config/i386/t-i386 (i386.o): Remove.
7147 (i386-c.o): Use COMPILE and POSTCOMPILE.
7149 2013-09-25 Tom Tromey <tromey@redhat.com>
7151 * Makefile.in ($(out_object_file)): Use COMPILE and POSTCOMPILE.
7153 2013-09-25 Tom Tromey <tromey@redhat.com>
7155 * Makefile.in (graph.o, sbitmap.o, sparseset.o, gcc-ar.o)
7156 (gcc-ranlib.o, gcc-nm.o, collect2.o, collect2-aix.o, tlink.o)
7157 (lto-wrapper.o, default-c.o, attribs.o, incpath.o, prefix.o)
7158 (gcc.o, options.o, options-save.o, version.o, gtype-desc.o)
7159 (trans-mem.o, ggc-common.o, ggc-page.o, ggc-none.o, stringpool.o)
7160 (convert.o, double-int.o, lto-compress.o, data-streamer-in.o)
7161 (data-streamer-out.o, data-streamer.o, gimple-streamer-in.o)
7162 (gimple-streamer-out.o, tree-streamer.o, tree-streamer-in.o)
7163 (tree-streamer-out.o, streamer-hooks.o, lto-cgraph.o)
7164 (lto-streamer-in.o, lto-streamer-out.o, lto-section-in.o)
7165 (lto-section-out.o, lto-opts.o, lto-streamer.o, langhooks.o)
7166 (test-dump.o, tree.o, tree-dump.o, tree-inline.o, print-tree.o)
7167 (stor-layout.o, asan.o, tsan.o, ubsan.o, tree-ssa-tail-merge.o)
7168 (tree-ssa-structalias.o, tree-ssa-uninit.o, tree-ssa.o)
7169 (tree-into-ssa.o, tree-ssa-ter.o, tree-ssa-coalesce.o)
7170 (tree-outof-ssa.o, tree-ssa-dse.o, tree-ssa-forwprop.o)
7171 (tree-ssa-phiprop.o, tree-ssa-ifcombine.o, tree-ssa-phiopt.o)
7172 (tree-nrv.o, tree-ssa-copy.o, tree-ssa-propagate.o)
7173 (tree-ssa-dom.o, tree-ssa-uncprop.o, tree-ssa-threadedge.o)
7174 (tree-ssa-threadupdate.o, tree-ssanames.o, tree-phinodes.o)
7175 (domwalk.o, tree-ssa-live.o, tree-ssa-copyrename.o)
7176 (tree-ssa-pre.o, tree-ssa-sccvn.o)
7177 (gimple-ssa-strength-reduction.o, tree-vrp.o, tree-cfg.o)
7178 (tree-cfgcleanup.o, tree-tailcall.o, tree-ssa-sink.o)
7179 (tree-nested.o, tree-if-conv.o, tree-iterator.o, tree-dfa.o)
7180 (tree-ssa-operands.o, tree-eh.o, tree-ssa-loop.o)
7181 (tree-ssa-loop-unswitch.o, tree-ssa-address.o)
7182 (tree-ssa-loop-niter.o, tree-ssa-loop-ivcanon.o)
7183 (tree-ssa-loop-ch.o, tree-ssa-loop-prefetch.o, tree-predcom.o)
7184 (tree-ssa-loop-ivopts.o, tree-affine.o, tree-ssa-loop-manip.o)
7185 (tree-ssa-loop-im.o, tree-ssa-math-opts.o, tree-ssa-alias.o)
7186 (tree-ssa-reassoc.o, tree-optimize.o, gimplify.o)
7187 (gimple-iterator.o, gimple-fold.o, gimple-low.o, omp-low.o)
7188 (tree-browser.o, omega.o, tree-chrec.o, tree-scalar-evolution.o)
7189 (tree-data-ref.o, sese.o, graphite.o, graphite-blocking.o)
7190 (graphite-clast-to-gimple.o, graphite-dependences.o)
7191 (graphite-interchange.o, graphite-poly.o)
7192 (graphite-scop-detection.o, graphite-sese-to-poly.o)
7193 (graphite-optimize-isl.o, tree-vect-loop.o)
7194 (tree-vect-loop-manip.o, tree-vect-patterns.o, tree-vect-slp.o)
7195 (tree-vect-stmts.o, tree-vect-data-refs.o, tree-vectorizer.o)
7196 (vtable-verify.o, tree-loop-distribution.o, tree-parloops.o)
7197 (tree-stdarg.o, tree-object-size.o, internal-fn.o, gimple.o)
7198 (gimple-pretty-print.o, tree-mudflap.o, tree-nomudflap.o)
7199 (tree-pretty-print.o, tree-diagnostic.o, fold-const.o)
7200 (diagnostic.o, diagnostic-color.o, opts.o, opts-global.o)
7201 (opts-common.o, targhooks.o, common/common-targhooks.o, input.o)
7202 (toplev.o, hwint.o, passes.o, plugin.o, main.o, host-default.o)
7203 (rtl-error.o, rtl.o, print-rtl.o, rtlanal.o, varasm.o, function.o)
7204 (statistics.o, stmt.o, except.o, expr.o, dojump.o, builtins.o)
7205 (calls.o, expmed.o, explow.o, optabs.o, dbxout.o, debug.o)
7206 (sdbout.o, dwarf2out.o, dwarf2cfi.o, dwarf2asm.o, vmsdbgout.o)
7207 (xcoffout.o, godump.o, emit-rtl.o, real.o, realmpfr.o, dfp.o)
7208 (fixed-value.o, jump.o, simplify-rtx.o, symtab.o, cgraph.o)
7209 (cgraphunit.o, cgraphclones.o, cgraphbuild.o, varpool.o, ipa.o)
7210 (ipa-profile.o, ipa-devirt.o, ipa-prop.o, ipa-ref.o, ipa-cp.o)
7211 (ipa-split.o, ipa-inline.o, ipa-inline-analysis.o)
7212 (ipa-inline-transform.o, ipa-utils.o, ipa-reference.o)
7213 (ipa-pure-const.o, coverage.o, cselib.o, cse.o, dce.o, dumpfile.o)
7214 (dse.o, fwprop.o, web.o, ree.o, cprop.o, gcse.o, store-motion.o)
7215 (resource.o, lcm.o, mode-switching.o, tree-ssa-dce.o)
7216 (tree-call-cdce.o, tree-ssa-ccp.o, tree-ssa-strlen.o, tree-sra.o)
7217 (tree-switch-conversion.o, tree-complex.o, tree-emutls.o)
7218 (tree-vect-generic.o, df-core.o, df-problems.o, df-scan.o)
7219 (regstat.o, valtrack.o, var-tracking.o, profile.o, mcf.o)
7220 (tree-profile.o, value-prof.o, loop-doloop.o, alloc-pool.o)
7221 (auto-inc-dec.o, cfg.o, cfghooks.o, cfgexpand.o, cfgrtl.o)
7222 (cfganal.o, cfgbuild.o, cfgcleanup.o, cfgloop.o, cfgloopanal.o)
7223 (graphds.o, loop-iv.o, loop-invariant.o, cfgloopmanip.o)
7224 (loop-init.o, loop-unswitch.o, loop-unroll.o, dominance.o)
7225 (et-forest.o, combine.o, reginfo.o, bitmap.o, vec.o, hash-table.o)
7226 (reload.o, reload1.o, rtlhooks.o, postreload.o, postreload-gcse.o)
7227 (caller-save.o, bt-load.o, reorg.o, alias.o, stack-ptr-mod.o)
7228 (init-regs.o, ira-build.o, ira-costs.o, ira-conflicts.o)
7229 (ira-color.o, ira-emit.o, ira-lives.o, ira.o, lra.o)
7230 (lra-assigns.o, lra-coalesce.o, lra-constraints.o)
7231 (lra-eliminations.o, lra-lives.o, lra-spills.o, regmove.o)
7232 (combine-stack-adj.o, compare-elim.o, ddg.o, modulo-sched.o)
7233 (haifa-sched.o, sched-deps.o, sched-rgn.o, sched-ebb.o)
7234 (sched-vis.o, sel-sched.o, sel-sched-dump.o, sel-sched-ir.o)
7235 (final.o, recog.o, reg-stack.o, sreal.o, predict.o, lists.o)
7236 (bb-reorder.o, tracer.o, timevar.o, regcprop.o, regrename.o)
7237 (ifcvt.o, params.o, pointer-set.o, hooks.o, pretty-print.o)
7238 (errors.o, dbgcnt.o, lower-subreg.o, target-globals.o)
7239 (hw-doloop.o, file-find.o, context.o, $(common_out_object_file))
7240 (insn-attrtab.o, insn-automata.o, insn-dfatab.o, insn-emit.o)
7241 (insn-enums.o, insn-extract.o, insn-latencytab.o, insn-modes.o)
7242 (insn-opinit.o, insn-output.o, insn-peep.o, insn-preds.o)
7243 (insn-recog.o, intl.o, cppbuiltin.o, cppdefault.o, gcov.o)
7244 (gcov-dump.o): Remove.
7245 (default-c.o): Use COMPILE and POSTCOMPILE.
7246 (CFLAGS-gcc.o): New variable.
7247 ($(common_out_object_file)): Use COMPILE and POSTCOMPILE.
7249 2013-09-25 Tom Tromey <tromey@redhat.com>
7251 * Makefile.in (c-family/cppspec.o, c-family/c-common.o)
7252 (c-family/c-cppbuiltin.o, c-family/c-dump.o, c-family/c-format.o)
7253 (c-family/c-gimplify.o, c-family/c-lex.o, c-family/c-omp.o)
7254 (c-family/c-opts.o, c-family/c-pch.o, c-family/c-ppoutput.o)
7255 (c-family/c-pragma.o, c-family/c-pretty-print.o)
7256 (c-family/c-semantics.o, c-family/c-ada-spec.o)
7257 (c-family/array-notation-common.o, c-family/stub-objc.o)
7258 (c-family/c-ubsan.o): Remove.
7260 2013-09-25 Tom Tromey <tromey@redhat.com>
7262 * Makefile.in (C_TREE_H): Reference c/c-tree.h.
7264 2013-09-25 Tom Tromey <tromey@redhat.com>
7266 * Makefile.in (DRIVER_DEFINES): Use $(and), not shell code,
7267 to add -DENABLE_SHARED_LIBGCC.
7268 (gcc.o): Don't use subshell.
7270 2013-09-25 Tom Tromey <tromey@redhat.com>
7272 * Makefile.in (OUTPUT_OPTION): Define as "-o $@".
7273 * configure.ac: Don't invoke AM_PROG_CC_C_O.
7274 (NO_MINUS_C_MINUS_O, OUTPUT_OPTION): Don't subst.
7275 * configure, config.in: Rebuild.
7277 2013-09-25 Tom Tromey <tromey@redhat.com>
7279 * Makefile.in (CCDEPMODE, DEPDIR, depcomp, COMPILE.base)
7280 (COMPILE, POSTCOMPILE): New variables.
7281 (.cc.o .c.o): Use COMPILE, POSTCOMPILE.
7282 (DEPFILES): New variable.
7283 Include ".Po" files.
7284 * configure.ac: Add checks for dependency checking.
7285 * configure, aclocal.m4: Regenerate.
7287 2013-09-25 Tom Tromey <tromey@redhat.com>
7289 * Makefile.in (ALL_HOST_BACKEND_OBJS): Add lto-wrapper.o.
7290 ($(ALL_HOST_OBJS)): Move order-only dependency to end of file.
7292 2013-09-25 Tom Tromey <tromey@redhat.com>
7294 * Makefile.in (generated_files): Add options.h,
7295 target-hooks-def.h, insn-opinit.h,
7296 common/common-target-hooks-def.h, pass-instances.def,
7297 c-family/c-target-hooks-def.h.
7299 2013-09-25 Jeff Law <law@redhat.com>
7301 * tree-ssa-threadedge.c (thread_across_edge): Use foo.last () rather
7302 than foo[foo.length () - 1] to access last member in a vec.
7303 * tree-ssa-threadupdate.c (register_jump_thread): Similarly.
7305 2013-09-25 Richard Biener <rguenther@suse.de>
7308 * tree.c (iterative_hash_expr): Remove MEM_REF special handling.
7310 2013-09-25 Jan Hubicka <jh@suse.cz>
7312 * cgraph.c (cgraph_resolve_speculation): Use semantical equivalency
7315 2013-09-25 Marek Polacek <polacek@redhat.com>
7318 * ubsan.c (ubsan_type_descriptor): Handle IDENTIFIER_NODEs
7319 when determining the type name.
7321 2013-09-24 Oleg Endo <olegendo@gcc.gnu.org>
7323 * config/sh/sh.md: Fix formatting.
7325 2013-09-24 Xinliang David Li <davidxl@google.com>
7327 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Check
7328 max peel iterations parameter.
7329 * param.def: New parameter.
7330 * doc/invoke.texi: Document New parameter.
7332 2013-09-24 Christophe Lyon <christophe.lyon@linaro.org>
7334 * gimple-pretty-print.c: Various whitespace tweaks.
7335 * tree-core.h: Likewise.
7336 * tree-pretty-print.c: Likewise.
7337 * tree-ssa-alias.c: Likewise.
7338 * tree-ssa-copy.c: Likewise.
7339 * tree-ssanames.c: Likewise.
7340 * tree-ssanames.h: Likewise.
7341 * tree-vrp.c: Likewise.
7343 2013-09-24 Alan Modra <amodra@gmail.com>
7347 * stmt.c (expand_asm_operands): Call expand_expr with EXPAND_MEMORY
7348 for output operands that disallow regs. Don't use EXPAND_WRITE on
7351 2013-09-24 Richard Biener <rguenther@suse.de>
7354 * tree.c (reference_alias_ptr_type): Move ...
7355 * alias.c (reference_alias_ptr_type): ... here and implement
7356 in terms of the new reference_alias_ptr_type_1.
7357 (ref_all_alias_ptr_type_p): New helper.
7358 (get_deref_alias_set_1): Drop flag_strict_aliasing here,
7359 use ref_all_alias_ptr_type_p.
7360 (get_deref_alias_set): Add flag_strict_aliasing check here.
7361 (reference_alias_ptr_type_1): New function, split out from ...
7362 (get_alias_set): ... here.
7363 (alias_ptr_types_compatible_p): New function.
7364 * alias.h (reference_alias_ptr_type): Declare.
7365 (alias_ptr_types_compatible_p): Likewise.
7366 * tree.h (reference_alias_ptr_type): Remove.
7367 * fold-const.c (operand_equal_p): Use alias_ptr_types_compatible_p
7368 to compare MEM_REF alias types.
7370 2013-09-24 Richard Biener <rguenther@suse.de>
7372 * tree-vrp.c (vrp_finalize): Check for SSA name presence.
7374 2013-09-23 Michael Meissner <meissner@linux.vnet.ibm.com>
7376 * config/rs6000/rs6000.c (rs6000_vector_reload): Delete, combine
7377 reload helper function arrays into a single array reg_addr.
7378 (reload_fpr_gpr): Likewise.
7379 (reload_gpr_vsx): Likewise.
7380 (reload_vsx_gpr): Likewise.
7381 (struct rs6000_reg_addr): Likewise.
7382 (reg_addr): Likewise.
7383 (rs6000_debug_reg_global): Change rs6000_vector_reload,
7384 reload_fpr_gpr, reload_gpr_vsx, reload_vsx_gpr uses to reg_addr.
7385 (rs6000_init_hard_regno_mode_ok): Likewise.
7386 (rs6000_secondary_reload_direct_move): Likewise.
7387 (rs6000_secondary_reload): Likewise.
7389 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add new
7390 constraints: wu, ww, and wy. Repurpose wv constraint added during
7391 power8 changes. Put wg constraint in alphabetical order.
7393 * config/rs6000/rs6000.opt (-mvsx-scalar-float): New debug switch
7394 for future work to add ISA 2.07 VSX single precision support.
7395 (-mvsx-scalar-double): Change default from -1 to 1, update
7396 documentation comment.
7397 (-mvsx-scalar-memory): Rename debug switch to -mupper-regs-df.
7398 (-mupper-regs-df): New debug switch to control whether DF values
7399 can go in the traditional Altivec registers.
7400 (-mupper-regs-sf): New debug switch to control whether SF values
7401 can go in the traditional Altivec registers.
7403 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print wu, ww,
7405 (rs6000_init_hard_regno_mode_ok): Use ssize_t instead of int for
7406 loop variables. Rename -mvsx-scalar-memory to -mupper-regs-df.
7407 Add new constraints, wu/ww/wy. Repurpose wv constraint.
7408 (rs6000_debug_legitimate_address_p): Print if we are running
7409 before, during, or after reload.
7410 (rs6000_secondary_reload): Add a comment.
7411 (rs6000_opt_masks): Add -mupper-regs-df, -mupper-regs-sf.
7413 * config/rs6000/constraints.md (wa constraint): Sort w<x>
7414 constraints. Update documentation string.
7415 (wd constraint): Likewise.
7416 (wf constraint): Likewise.
7417 (wg constraint): Likewise.
7418 (wn constraint): Likewise.
7419 (ws constraint): Likewise.
7420 (wt constraint): Likewise.
7421 (wx constraint): Likewise.
7422 (wz constraint): Likewise.
7423 (wu constraint): New constraint for ISA 2.07 SFmode scalar
7425 (ww constraint): Likewise.
7426 (wy constraint): Likewise.
7427 (wv constraint): Repurpose ISA 2.07 constraint that we did not use
7428 in the previous submissions.
7429 * doc/md.texi (PowerPC and IBM RS6000): Likewise.
7431 2013-09-23 Richard Sandiford <rdsandiford@googlemail.com>
7433 * doc/rtl.texi (REG_NOTES): Say that int_list can also be used.
7434 (REG_BR_PROB): Say that the probability is stored in an int_list.
7435 * reg-notes.def: Update commentary to mention INT_LIST.
7436 * rtl.def (EXPR_LIST, INSN_LIST): Capitalize comments.
7437 (INT_LIST): New rtx.
7438 * rtl.h (add_int_reg_note, add_shallow_copy_of_reg_note): Declare.
7439 * rtlanal.c (int_reg_note_p): New function.
7440 (alloc_reg_note): Assert that the note does not have an int argument.
7441 (add_int_reg_note, add_shallow_copy_of_reg_note): New functions.
7442 * combine.c (distribute_notes): Use add_shallow_copy_of_rtx.
7443 * cse.c (cse_process_notes_1): Expect REG_EQUAL to be an EXPR_LIST
7444 rather than an INSN_LIST. Handle INT_LIST.
7445 * ifcvt.c (cond_exec_process_insns): Take the probability as an int
7446 rather than an rtx. Use gen_rtx_INT_LIST to create a REG_BR_PROB note.
7447 (cond_exec_process_if_block): Use XINT to extract REG_BR_PROB values.
7448 Manipulate them as ints rather than rtxes.
7449 * reg-stack.c (subst_asm_stack_regs): Only handle EXPR_LIST notes.
7450 * regmove.c (copy_src_to_dest): Likewise.
7451 * sched-vis.c (print_insn_with_notes): Handle INT_LIST.
7453 * config/i386/winnt.c (i386_pe_seh_unwind_emit): Sink pat assignment
7454 into the cases that need it.
7455 * config/arm/arm.c (arm_unwind_emit): Likewise.
7457 * asan.c (asan_clear_shadow): Use add_int_reg_note for REG_BR_PROB.
7458 * emit-rtl.c (try_split, emit_copy_of_insn_after): Likewise.
7459 * loop-doloop.c (add_test, doloop_modify): Likewise.
7460 * loop-unswitch.c (compare_and_jump_seq): Likewise.
7461 * optabs.c (emit_cmp_and_jump_insn_1): Likewise.
7462 * predict.c (combine_predictions_for_insn): Likewise.
7463 * print-rtl.c (print_rtx): Handle INT_LIST.
7464 * config/aarch64/aarch64.c (aarch64_emit_unlikely_jump): Likewise.
7465 * config/alpha/alpha.c (emit_unlikely_jump): Likewise.
7466 * config/arm/arm.c (emit_unlikely_jump): Likewise.
7467 * config/i386/i386.c (ix86_expand_split_stack_prologue): Likewise.
7468 (ix86_split_fp_branch, predict_jump): Likewise.
7469 * config/rs6000/rs6000.c (emit_unlikely_jump): Likewise.
7470 * config/sh/sh.c (expand_cbranchsi4): Likewise.
7471 * config/spu/spu.c (ea_load_store_inline): Likewise.
7473 * cfgbuild.c (compute_outgoing_frequencies): Use XINT to access the
7474 value of a REG_BR_PROB note.
7475 * cfgrtl.c (force_nonfallthru_and_redirect): Likewise.
7476 (update_br_prob_note, rtl_verify_edges, purge_dead_edges): Likewise.
7477 * emit-rtl.c (try_split): Likewise.
7478 * predict.c (br_prob_note_reliable_p): Likewise.
7479 (invert_br_probabilities, combine_predictions_for_insn): Likewise.
7480 * reorg.c (mostly_true_jump): Likewise.
7481 * config/bfin/bfin.c (cbranch_predicted_taken_p): Likewise.
7482 * config/frv/frv.c (frv_print_operand_jump_hint): Likewise.
7483 * config/i386/i386.c (ix86_print_operand): Likewise.
7484 * config/ia64/ia64.c (ia64_print_operand): Likewise.
7485 * config/mmix/mmix.c (mmix_print_operand): Likewise.
7486 * config/rs6000/rs6000.c (output_cbranch): Likewise.
7487 * config/s390/s390.c (s390_expand_tbegin): Likewise.
7488 * config/sh/sh.c (sh_print_operand, sh_adjust_cost): Likewise.
7489 * config/sparc/sparc.c (output_cbranch): Likewise.
7490 * config/spu/spu.c (get_branch_target): Likewise.
7491 * config/tilegx/tilegx.c (cbranch_predicted_p): Likewise.
7492 * config/tilepro/tilepro.c (cbranch_predicted_p): Likewise.
7494 2013-09-23 Jan Hubicka <jh@suse.cz>
7496 * ipa-cp.c (ipa_get_indirect_edge_target_1): Add sanity check
7498 * ipa-utils.h (possible_polymorphic_call_target_p): New function.
7499 * ipa-devirt.c (possible_polymorphic_call_target_p): Be tolerant
7501 * gimple-fold.c: Include ipa-utils.h and gimple-pretty-print.h
7502 (gimple_fold_call): Dump inconsistent devirtualizations; add
7503 sanity check for type based devirtualizations.
7504 * ipa-prop.c: Include ipa-utils.h
7505 (ipa_intraprocedural_devirtualization): Add sanity check.
7506 (try_make_edge_direct_virtual_call): Likewise.
7508 2013-09-23 Eric Botcazou <ebotcazou@adacore.com>
7510 * tree-ssa-ccp.c (insert_clobber_before_stack_restore): Recurse on copy
7511 assignment statements.
7513 2013-09-23 Kugan Vivekanandarajah <kuganv@linaro.org>
7515 * gimple-pretty-print.c (dump_ssaname_info): New function.
7516 (dump_gimple_phi): Call it.
7517 (pp_gimple_stmt_1): Likewise.
7518 * tree-core.h (tree_ssa_name): New union ssa_name_info_type field.
7519 (range_info_def): Declare.
7520 * tree-pretty-print.c (pp_double_int): New function.
7521 (dump_generic_node): Call it.
7522 * tree-pretty-print.h (pp_double_int): Declare.
7523 * tree-ssa-alias.c (dump_alias_info): Check pointer type.
7524 * tree-ssanames.h (range_info_def): New structure.
7525 (value_range_type): Move definition here.
7526 (set_range_info, value_range_type, duplicate_ssa_name_range_info):
7528 * tree-ssanames.c (make_ssa_name_fn): Check pointer type at
7530 (set_range_info): New function.
7531 (get_range_info): Likewise.
7532 (duplicate_ssa_name_range_info): Likewise.
7533 (duplicate_ssa_name_fn): Check pointer type and call
7534 duplicate_ssa_name_range_info.
7535 * tree-ssa-copy.c (fini_copy_prop): Likewise.
7536 * tree-vrp.c (value_range_type): Remove definition, now in
7538 (vrp_finalize): Call set_range_info to update value range of SSA_NAMEs.
7539 * tree.h (SSA_NAME_PTR_INFO): Macro changed to access via union.
7540 (SSA_NAME_RANGE_INFO): New macro.
7542 2013-09-23 Richard Biener <rguenther@suse.de>
7544 PR tree-optimization/58464
7545 * tree-ssa-pre.c (phi_trans_lookup): Remove.
7546 (phi_trans_add): Change to add conditionally on being not present.
7547 (phi_translate_1): Remove recursion detection here.
7548 (phi_translate): Pre-seed the cache with NULL to catch
7549 recursion here in a more generic way.
7550 (bitmap_find_leader): Adjust comment.
7551 (get_representative_for): Dump value-numbers.
7552 (create_expression_by_pieces): Likewise.
7553 (insert_into_preds_of_block): Likewise.
7555 2013-09-23 Christian Bruel <christian.bruel@st.com>
7558 * config/sh/sh.md (movsf_ie): Allow fpul_operand.
7559 * config/sh/predicate.md (arith_reg_operand): Disallow FPUL_REG.
7561 2013-09-23 James Greenhalgh <james.greenhalgh@arm.com>
7564 2013-09-20 Renlin Li <renlin.li@arm.com>
7566 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
7568 (aarch64_expand_epilogue): Likewise.
7569 (aarch64_legitimize_reload_address): Likewise.
7571 2013-09-22 Eric Botcazou <ebotcazou@adacore.com>
7573 * gimplify.c (gimplify_asm_expr): Reset the TREE_CHAIN of clobbers to
7574 NULL_TREE before pushing them onto the vector. Likewise for labels.
7576 2013-09-21 Eric Botcazou <ebotcazou@adacore.com>
7578 * config/ia64/predicates.md (ia64_cbranch_operator): Accept unordered
7579 comparison operators when -fno-trapping-math is in effect.
7580 * config/ia64/ia64.c (ia64_expand_compare): Add support for unordered
7581 comparison operators in TFmode and assert that unsupported operators
7583 (ia64_print_operand): Likewise.
7585 2013-09-21 Jan Hubicka <jh@suse.cz>
7587 * x86-tune.def (partial_reg_stall): Disable for CoreI7 and newer.
7588 (sse_typeless_stores): Enable for core
7589 (sse_load0_by_pxor): Likewise.
7590 (four_jump_limit): Disable for core.
7591 (pad_returns): Likewise.
7592 (avoid_vector_decode): Likewise.
7593 (fuse_cmp_and_branch): Enable for cores.
7594 * i386.c (x86_accumulate_outgoing_args): Disable for cores.
7596 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
7599 * config/pa/pa.c (pa_option_override): Disable auto increment and
7600 decrement instructions until reload is completed.
7602 * config/pa/pa-linux.h (TARGET_OS_CPP_BUILTINS): Define
7603 __GCC_HAVE_SYNC_COMPARE_AND_SWAP_1, __GCC_HAVE_SYNC_COMPARE_AND_SWAP_2,
7604 and __GCC_HAVE_SYNC_COMPARE_AND_SWAP_4.
7606 2013-09-20 DJ Delorie <dj@redhat.com>
7607 Nick Clifton <nickc@redhat.com>
7609 * config/rl78/rl78.c: Various whitespace and comment tweaks.
7610 (need_to_save): Save bank 0 on interrupts.
7611 (characterize_address): Strip far address wrappers.
7612 (rl78_as_legitimate_address): Likewise.
7613 (transcode_memory_rtx): Likewise.
7614 (rl78_peep_movhi_p): Disable this peephole after devirt.
7615 (rl78_propogate_register_origins): Forget all origins when a
7617 * config/rl78/rl78-virt.md: Various whitespace tweaks.
7618 * config/rl78/rl78-real.md: Various whitespace tweaks. Additional
7620 * config/rl78/rl78.md (sel_rb): Disable for G10 just in case.
7621 * config/rl78/rl78-expand.md (movqi): Check for subregs of consts.
7622 * config/rl78/rl78.h (LINK_SPEC): Pass -gc-sections unless
7624 * config/rl78/constraints.md: Various whitespace and paren tweaks.
7626 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
7628 * config/pa/pa.md: In "scc" insn patterns, change output template to
7629 handle const0_rtx in reg_or_0_operand operands.
7631 2013-09-20 Martin Husemann <martin@NetBSD.org>
7634 * config/vax/vax.c (vax_output_int_move): Use D format specifier.
7635 * config/vax/vax.md (ashldi3, <unnamed>): Ditto.
7637 2013-09-20 Richard Biener <rguenther@suse.de>
7640 * tree-scalar-evolution.c (struct scev_info_str): Shrink by
7641 remembering SSA name version and block index.
7642 (new_scev_info_str): Adjust.
7643 (hash_scev_info): Likewise. Also hash the block index.
7644 (eq_scev_info): Adjust.
7645 (find_var_scev_info): Likewise.
7646 (struct instantiate_cache_entry): Remove.
7647 (struct instantiate_cache_type): Use a htab to map name, block
7649 (instantiate_cache_type::~instantiate_cache_type): Adjust.
7650 (get_instantiated_value_entry): Likewise.
7651 (hash_idx_scev_info, eq_idx_scev_info): New functions.
7652 (instantiate_scev_name): Adjust.
7654 2013-09-20 Jeff Law <law@redhat.com>
7656 * tree-ssa-dom.c (record_temporary_equivalences): Add comment.
7658 2013-09-20 Yufeng Zhang <yufeng.zhang@arm.com>
7660 * config/aarch64/aarch64-builtins.c (aarch64_simd_expand_args):
7661 Call aarch64_simd_expand_args to update op[argc].
7663 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
7665 * plugin.c (parse_plugin_arg_opt): Accept equal sign inside
7668 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
7670 * gengtype.c (file_rules): Added rule for *.cc files.
7671 (get_output_file_with_visibility): Give fatal message when no
7674 2013-09-20 Renlin Li <renlin.li@arm.com>
7676 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use plus_constant.
7677 (aarch64_expand_epilogue): Likewise.
7678 (aarch64_legitimize_reload_address): Likewise.
7680 2013-09-20 Bernd Edlinger <bernd.edlinger@hotmail.de>
7683 * expr.c (expand_assignment): Remove misalignp code path.
7685 2013-09-20 Marek Polacek <polacek@redhat.com>
7688 * ubsan.c (get_ubsan_type_info_for_type): Use TYPE_SIZE instead of
7689 TYPE_PRECISION. Add asserts.
7691 2013-09-20 Richard Biener <rguenther@suse.de>
7693 PR tree-optimization/58453
7694 * tree-loop-distribution.c (distribute_loop): Apply the cost
7695 model for -ftree-loop-distribute-patterns, too.
7697 2013-09-20 Richard Biener <rguenther@suse.de>
7700 * tree-chrec.h (build_polynomial_chrec): Use gcc_checking_assert,
7701 make type comparison less strict.
7703 2013-09-20 Alan Modra <amodra@gmail.com>
7705 * configure: Regenerate.
7706 * aclocal.m4: Regenerate.
7708 2013-09-20 Marek Polacek <polacek@redhat.com>
7711 * doc/extend.texi: Document that attribute used is meant to be used
7712 on variables with static storage duration.
7714 2013-09-19 Jakub Jelinek <jakub@redhat.com>
7716 PR tree-optimization/58472
7717 * tree-vect-stmts.c (vectorizable_store, vectorizable_load): For
7718 simd_lane_access set inv_p = false.
7719 * omp-low.c (lower_rec_input_clauses): Set TREE_NO_WARNING on
7720 the simduid magic VAR_DECL.
7722 2013-09-19 Jan Hubicka <jh@suse.cz>
7724 * i386.c (generic_memcpy, generic_memset): Fix 32bit template.
7726 2013-09-17 Jeff Law <law@redhat.com>
7728 * tree-ssa-dom.c (record_temporary_equivalences): New function
7729 split out of dom_opt_dom_walker::after_dom_children.
7730 (dom_opt_dom_walker::thread_across_edge): Move common code
7731 in here from dom_opt_dom_walker::after_dom_children.
7732 (dom_opt_dom_walker::after_dom_children): Corresponding simplifictions.
7734 2013-09-19 Jan Hubicka <jh@suse.cz>
7736 * i386.h (TARGET_GENERIC32, TARGET_GENERIC64): Remove.
7737 (TARGET_GENERIC): Use PROCESOR_GENERIC
7738 (enum processor_type): Unify generic32 and 64.
7739 * i386.md (cpu): Likewise.
7740 * x86-tune.def (use_leave): Enable for generic32.
7741 (avoid_vector_decode, slow_imul_imm32_mem, slow_imul_imm8): Likewise.
7742 * athlon.md: Change generic64 to generic in all occurences.
7743 * i386-c.c (ix86_target_macros_internal): Unify generic64 and 32.
7744 (ix86_target_macros_internal): Likewise.
7745 * driver-i386.c (host_detect_local_cpu): Likewise.
7746 * i386.c (generic64_memcpy, generic64_memset, generic64_cost): Rename
7748 (generic_memcpy, generic_memset, generic_cost): This one.
7749 (generic32_memcpy, generic32_memset, generic32_cost): Remove.
7750 (m_GENERIC32, m_GENERIC64): Remove.
7751 (m_GENERIC): Turn into one flag.
7752 (processor_target): Unify generic tunnings.
7753 (ix86_option_override_internal): Replace generic32/64 by generic.
7754 (ix86_issue_rate): Likewise.
7755 (ix86_adjust_cost): Likewise.
7757 2013-09-19 Jan Hubicka <jh@suse.cz>
7759 * cgraph.c (cgraph_create_edge_1): Avoid uninitialized read
7760 of speculative flag.
7762 2013-09-19 Jakub Jelinek <jakub@redhat.com>
7764 * omp-low.c (expand_omp_sections): Always pass len - 1 to
7765 GOMP_sections_start, even if !exit_reachable.
7767 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
7769 * lra-constraints.c (need_for_all_save_p): Use macro
7770 HARD_REGNO_CALL_PART_CLOBBERED.
7771 * lra-lives.c (check_pseudos_live_through_calls): Use the macro to
7772 set up pseudo conflict hard regs.
7774 2013-09-18 Michael Meissner <meissner@linux.vnet.ibm.com>
7777 * config/rs6000/paired.md (movmisalignv2sf): Fix to allow memory
7780 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
7782 PR rtl-optimization/58438
7783 * lra.c (lra): Clear lra_optional_reload_pseudos in upper loop.
7784 * lra-constraints.c (undo_optional_reloads): Keep optional reloads
7785 from previous subpasses.
7787 2013-09-18 Richard Earnshaw <rearnsha@arm.com>
7789 * arm.c (arm_get_frame_offsets): Validate architecture supports
7790 LDRD/STRD before accepting the tuning preference.
7791 (arm_expand_prologue): Likewise.
7792 (arm_expand_epilogue): Likewise.
7794 2013-09-18 Richard Biener <rguenther@suse.de>
7796 PR tree-optimization/58417
7797 * tree-chrec.c (chrec_fold_plus_1): Assert that we do not
7798 have chrecs with symbols defined in the loop as operands.
7799 (chrec_fold_multiply): Likewise.
7800 * tree-scalar-evolution.c (interpret_rhs_expr): Instantiate
7801 parameters before folding binary operations.
7802 (struct instantiate_cache_entry_hasher): Remove.
7803 (struct instantiate_cache_type): Use a pointer-map.
7804 (instantiate_cache_type::instantiate_cache_type): New function.
7805 (instantiate_cache_type::get): Likewise.
7806 (instantiate_cache_type::set): Likewise.
7807 (instantiate_cache_type::~instantiate_cache_type): Adjust.
7808 (get_instantiated_value_entry): Likewise.
7809 (global_cache): New global.
7810 (instantiate_scev_r, instantiate_scev_poly, instantiate_scev_binary,
7811 instantiate_array_ref, instantiate_scev_convert, instantiate_scev_3,
7812 instantiate_scev_2, instantiate_scev_1): Do not pass along cache.
7813 (instantiate_scev_name): Adjust.
7814 (instantiate_scev): Construct global instead of local cache.
7815 (resolve_mixers): Likewise.
7817 2013-09-18 Daniel Morris <danielm@ecoscentric.com>
7818 Paolo Carlini <paolo.carlini@oracle.com>
7821 * doc/implement-cxx.texi: Fix references to the C++ standards.
7823 2013-09-18 Jakub Jelinek <jakub@redhat.com>
7825 * omp-low.c (copy_var_decl): Copy DECL_ATTRIBUTES.
7826 * tree-vect-data-refs.c (vect_analyze_data_refs): For
7827 simd_lane_access drs, update also DR_ALIGNED_TO.
7829 2013-09-18 Marek Polacek <polacek@redhat.com>
7832 * doc/extend.texi: Document no_sanitize_undefined attribute.
7833 * builtins.c (fold_builtin_0): Don't sanitize function if it has the
7834 no_sanitize_undefined attribute.
7836 2013-09-18 Nick Clifton <nickc@redhat.com>
7838 * config/msp430/msp430.h (ASM_SPEC): Pass -md on to the assembler.
7839 (ASM_DECLARE_FUNCTION_NAME): Define.
7841 2013-09-17 Trevor Saunders <tsaunders@mozilla.com>
7843 * compare-elim.c (find_comparison_dom_walker): New class
7844 (find_comparisons_in_bb): Rename to
7845 find_comparison_dom_walker::before_dom_children
7846 (find_comparisons): Adjust
7847 * domwalk.c (walk_dominator_tree): Rename to dom_walker::walk, and
7849 (init_walk_dominator_tree, fini_walk_dominator_tree): Remove
7850 * domwalk.h (dom_walk_data): Convert it To a class dom_walker.
7851 (init_walk_dominator_tree): Remove declaration.
7852 (fini_walk_dominator_tree): Remove declaration.
7853 * fwprop.c (single_def_use_dom_walker): New class
7854 (single_def_use_enter_block): Convert to
7855 single_def_use_dom_walker::before_dom_children.
7856 (single_def_use_leave_block): Convert to
7857 single_def_use_dom_walker::after_dom_children.
7858 (build_single_def_use_links): Adjust.
7859 * gimple-ssa-strength-reduction.c (find_candidates_dom_walker): New
7861 (find_candidates_in_block): Convert to
7862 find_candidates_dom_walker::before_dom_children.
7863 (execute_strength_reduction): Adjust.
7864 * graphite-sese-to-poly.c (struct bsc, build_sese_conditions): Remove.
7865 (sese_dom_walker): New class.
7866 (sese_dom_walker::sese_dom_walker): New constructor.
7867 (sese_dom_walker::~sese_dom_walker): New destructor.
7868 (build_sese_conditions_before): Convert to
7869 sese_dom_walker::before_dom_children.
7870 (build_sese_conditions_after): Convert to
7871 sese_dom_walker::after_dom_children.
7872 (build_poly_scop): Adjust
7873 * tree-into-ssa.c (rewrite_dom_walker): New class
7874 (rewrite_enter_block): Convert to
7875 rewrite_dom_walker::before_dom_children.
7876 (rewrite_leave_block): Convert to
7877 rewrite_dom_walker::after_dom_children.
7878 (rewrite_update_dom_walker): New class.
7879 (rewrite_update_enter_block): Convert to
7880 rewrite_update_dom_walker::before_dom_children.
7881 (rewrite_update_leave_block): Convert to
7882 rewrite_update_dom_walker::after_dom_children.
7883 (rewrite_blocks, rewrite_into_ssa): Adjust.
7884 (mark_def_dom_walker): New class.
7885 (mark_def_dom_walker::mark_def_dom_walker): New constructor.
7886 (mark_def_dom_walker::~mark_def_dom_walker): New destructor.
7887 (mark_def_sites_blocks): Convert to
7888 mark_def_dom_walker::before_dom_children.
7889 (mark_def_site_blocks): Remove.
7890 * tree-ssa-dom.c (dom_opt_dom_walker): New class.
7891 (tree_ssa_dominator_optimize): Adjust.
7892 (dom_thread_across_edge): Convert to method
7893 dom_opt_dom_walker::thread_across_edge.
7894 (dom_opt_enter_block): Convert to member function
7895 dom_opt_dom_walker::before_dom_children.
7896 (dom_opt_leave_block): Convert to member function
7897 dom_opt_dom_walker::after_dom_children.
7898 * tree-ssa-dse.c (dse_dom_walker): New class.
7899 (dse_enter_block): Convert to member function
7900 dse_dom_walker::before_dom_children.
7901 (tree_ssa_dse): Adjust.
7902 * tree-ssa-loop-im.c (invariantness_dom_walker): New class.
7903 (determine_invariantness_stmt): Convert to method
7904 invariantness_dom_walker::before_dom_children.
7905 (determine_invariantness): Remove
7906 (move_computations_dom_walker): New class.
7907 (move_computations_stmt): Convert to method
7908 move_computations_dom_walker::before_dom_children.
7909 (move_computations, tree_ssa_lim): Adjust.
7910 * tree-ssa-phiopt.c (nontrapping_dom_walker): New class.
7911 (nt_init_block): Convert to method
7912 notrappping_dom_walker::before_dom_children.
7913 (nt_fini_block): Convert to method
7914 method nontrapping_dom_walker::after_dom_children.
7915 (get_non_trapping): Adjust.
7916 * tree-ssa-pre.c (eliminate_dom_walker): New class.
7917 (eliminate_bb): Convert to method
7918 eliminate_dom_walker::before_dom_children.
7919 (eliminate_leave_block): Convert to method
7920 eliminate_dom_walker::after_dom_children.
7921 (eliminate): Adjust.
7922 * tree-ssa-strlen.c (strlen_dom_walker): New class.
7923 (strlen_enter_block): Convert to method
7924 strlen_dom_walker::before_dom_children.
7925 (strlen_leave_block): Convert to method
7926 method strlen_dom_walker::after_dom_children.
7927 (tree_ssa_strlen): Adjust.
7928 * tree-ssa-uncprop.c (uncprop_dom_walker): New class.
7929 (tree_ssa_uncprop): Adjust.
7930 (uncprop_leave_block): Convert to method
7931 uncprop_dom_walker::after_dom_children.
7932 (uncprop_leave_block): Convert to method
7933 uncprop_dom_walker::before_dom_children.
7935 2013-09-18 Bin Cheng <bin.cheng@arm.com>
7937 * config/arm/arm.c (thumb1_reorg): Search for flag setting insn before
7938 branch in same basic block. Check both src and dest of the move insn.
7940 2013-09-17 Nick Clifton <nickc@redhat.com>
7942 * config/rl78/rl78-real.md (bf): New pattern.
7944 * config/rl78/rl78.c (rl78_print_operand_1): Handle %B.
7945 (rl78_print_operand): Do not put a # before a %B.
7946 * config/rl78/rl78.opt: Tweak doc strings.
7948 2013-09-17 DJ Delorie <dj@redhat.com>
7950 * config/rl78/constraints.md (Wcv): Allow up to $r31.
7951 * config/rl78/rl78.c (rl78_asm_file_start: Likewise.
7952 (rl78_option_override): Likewise, if -mallregs.
7953 (is_virtual_register): Likewise.
7954 * config/rl78/rl78.h (reg_class): Extend VREGS to $r31.
7955 (REGNO_OK_FOR_BASE_P): Likewise.
7956 * config/rl78/rl78.opt (-mallregs): New.
7958 2013-09-17 Nick Clifton <nickc@redhat.com>
7960 * config/rl78/rl78.c (need_to_save): Change return type to bool.
7961 For interrupt functions: save all call clobbered registers if the
7962 interrupt handler is not a leaf function.
7963 (rl78_expand_prologue): Always recompute the frame information.
7964 For interrupt functions: only select bank 0 if one of the bank 0
7965 registers is going to be psuhed.
7967 2013-09-17 DJ Delorie <dj@redhat.com>
7969 * config/rl78/constraints.md: For each W* constraint, rename to C*
7970 and create a W* constraint that checks for an optional ES: prefix
7972 * config/rl78/rl78.md (UNS_ES_ADDR): New.
7973 (es_addr): New. Used to wrap far addresses.
7974 * config/rl78/rl78-protos.h (rl78_es_addr): New.
7975 (rl78_es_base): New.
7976 * config/rl78/rl78.c (rl78_as_legitimate_address): Accept "unspec"
7977 wrapped far addresses.
7978 (rl78_print_operand_1): Unwrap far addresses before processing.
7979 (rl78_lo16): Wrap far addresses in unspecs.
7980 (rl78_es_addr): New.
7981 (rl78_es_base): New.
7982 (insn_ok_now): Check for not-yet-wrapped far addresses.
7983 (transcode_memory_rtx): Properly re-wrap far addresses.
7985 2013-09-17 Sebastian Huber <sebastian.huber@embedded-brains.de>
7987 * config/sparc/t-rtems: Add leon3 multilibs.
7989 2013-09-17 Cong Hou <congh@google.com>
7991 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Fix a bug
7992 when checking the dot production pattern. The type of rhs operand
7993 of multiply is now checked correctly.
7995 2013-09-17 Jeff Law <law@redhat.com>
7997 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
7998 edge implied equivalences into successor phis.
7999 * tree-ssa-threadupdate.c (phi_args_equal_on_edges): Moved into
8000 here from tree-ssa-threadedge.c.
8001 (mark_threaded_blocks): When threading through a joiner, if both
8002 successors of the joiner's clone reach the same block, verify the
8003 PHI arguments are equal. If not, cancel the jump threading request.
8004 * tree-ssa-threadedge.c (phi_args_equal_on_edges): Moved into
8005 tree-ssa-threadupdate.c
8006 (thread_across_edge): Don't check PHI argument equality when
8007 threading through joiner block here.
8009 2013-09-17 Andrew MacLeod <amacleod@redhat.com>
8011 * tree-flow.h (ssa_undefined_value_p): Remove prototype.
8012 * tree-ssa.c (ssa_undefined_value_p): Move pass independent parts here.
8013 (warn_uninit, warn_uninitialized_vars,
8014 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
8015 Move to tree-ssa-uninit.c.
8016 * tree-ssa-uninit.c (ssa_undefined_value_p): Move to tree-ssa.c.
8017 (has_undefined_value_p): New. Pass dependant parts of
8018 ssa_undefined_value_p.
8019 (uninit_undefined_value_p): Use has_undefined_value_p.
8020 (warn_uninit, warn_uninitialized_vars,
8021 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
8022 Move from tree-ssa.c.
8023 * tree-ssa.h: Adjust prototypes.
8025 2013-09-17 Jan Hubicka <jh@suse.cz>
8028 * cif-code.def (FUNCTION_NOT_OPTIMIZED): New CIF code.
8029 * ipa-inline.c (can_inline_edge_p): Do not downgrade
8030 FUNCTION_NOT_OPTIMIZED.
8031 * ipa-inline-analysis.c (compute_inline_parameters): Function
8032 not optimized is not inlinable unless it is alwaysinline.
8033 (inline_analyze_function): Force calls in not optimized
8034 function not inlinable.
8036 2013-09-17 Jan Hubicka <jh@suse.cz>
8039 * ipa-devirt.c (ipa_devirt): Be ready for symtab_nonoverwritable_alias
8041 * ipa.c (function_and_variable_visibility): Likewise.
8042 * ipa-profile.c (ipa_profile): Likewise.
8044 2013-09-17 Bernd Edlinger <bernd.edlinger@hotmail.de>
8047 * cgraph.c (cgraph_function_body_availability): Check for ifunc
8048 attribute, and don't inline the resolver in this case.
8050 2013-09-17 Teresa Johnson <tejohnson@google.com>
8052 * coverage.c (get_coverage_counts): Add missing newline.
8054 2013-09-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8056 PR tree-optimization/58088
8057 * fold-const.c (mask_with_trailing_zeros): New function.
8058 (fold_binary_loc): Make sure we don't recurse infinitely
8059 when the X in (X & C1) | C2 is a tree of the form (Y * K1) & K2.
8060 Use mask_with_trailing_zeros where appropriate.
8062 2013-09-17 Yuri Rumyantsev <ysrumyan@gmail.com>
8064 * config/i386/i386.c (distance_agu_use_in_bb) : Proper initialization
8065 of 'prev' var to get better distance estimation.
8067 2013-09-17 Eric Botcazou <ebotcazou@adacore.com>
8069 * tree-inline.h (struct copy_body_data): Add transform_parameter.
8070 * tree-inline.c (is_parameter_of): New predicate.
8071 (remap_gimple_op_r): Do not propagate TREE_THIS_NOTRAP on MEM_REF if
8072 a parameter has been remapped.
8073 (copy_tree_body_r): Likewise on INDIRECT_REF and MEM_REF.
8074 (optimize_inline_calls): Initialize transform_parameter.
8075 (copy_gimple_seq_and_replace_locals): Likewise.
8076 (tree_function_versioning): Likewise.
8077 (maybe_inline_call_in_expr): Likewise.
8079 2013-09-17 Nick Clifton <nickc@redhat.com>
8081 * config/msp430/msp430-protos.h: Add prototypes for new functions.
8082 * config/msp430/msp430.c (msp430_preserve_reg_p): Add support for
8084 (is_attr_func): New function.
8085 (msp430_is_interrupt_func): New function.
8086 (is_naked_func): New function.
8087 (is_reentrant_func): New function.
8088 (is_critical_func): New function.
8089 (msp430_start_function): Add annotations for function attributes.
8090 (msp430_attr): New function.
8091 (msp430_attribute_table): New.
8092 (msp430_function_section): New function.
8093 (TARGET_ASM_FUNCTION_SECTION): Define.
8094 (msp430_builtin): New enum.
8095 (msp430_init_builtins): New function.
8096 (msp430_builtin_devl): New function.
8097 (msp430_expand_builtin): New function.
8098 (TARGET_INIT_BUILTINS): Define.
8099 (TARGET_EXPAND_BUILTINS): Define.
8100 (TARGET_BUILTIN_DECL): Define.
8101 (msp430_expand_prologue): Add support for naked, interrupt,
8102 critical and reentrant functions.
8103 (msp430_expand_epilogue): Likewise.
8104 (msp430_print_operand): Handle 'O' character.
8105 * config/msp430/msp430.h (TARGET_CPU_CPP_BUILTINS): Define
8107 * config/msp430/msp430.md (unspec): Add UNS_DINT, UNS_EINT,
8108 UNS_PUSH_INTR, UNS_POP_INTR, UNS_BIC_SR, UNS_BIS_SR.
8109 (pushm): Use a 'n' rather than an 'i' constraint.
8110 (msp_return): Add generation of the interrupt return instruction.
8111 (disable_interrupts): New pattern.
8112 (enable_interrupts): New pattern.
8113 (push_intr_state): New pattern.
8114 (pop_intr_state): New pattern.
8115 (bic_SR): New pattern.
8116 (bis_SR): New pattern.
8117 * doc/extend.texi: Document MSP430 function attributes and builtin
8120 2013-09-17 Richard Biener <rguenther@suse.de>
8122 PR tree-optimization/58432
8123 * tree-loop-distribution.c (tree_loop_distribution): Also
8124 scan PHIs for outside loop uses and seed a partition from them.
8126 2013-09-17 Bin Cheng <bin.cheng@arm.com>
8128 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref): New.
8129 (restructure_reference): Call backtrace_base_for_ref.
8131 2013-09-17 Alan Modra <amodra@gmail.com>
8134 * config/rs6000/driver-rs6000.c (elf_platform): Revert 2013-06-11
8137 2013-09-16 DJ Delorie <dj@redhat.com>
8139 * config/rl78/rl78.c (rl78_asm_file_start): Specify alternate
8140 vregs location for RL78/G10.
8141 (rl78_expand_prologue): Avoid SEL on G10.
8142 (rl78_expand_epilogue): Likewise.
8143 (rl78_peep_movhi_p): Can't move a constant to memory in HImode.
8144 * config/rl78/rl78.h (TARGET_CPU_CPP_BUILTINS): Define
8145 __RL78_G10__ when appropriate.
8146 (ASM_SPEC): Pass -mg10 along to the assembler.
8147 * config/rl78/rl78.md (sel_rb): Disable for G10.
8148 * config/rl78/rl78.opt: Add -mg10 option.
8149 * config/rl78/t-rl78: Add -mg10 multilib.
8151 2013-09-16 Xinliang David Li <davidxl@google.com>
8153 * tree-if-conv.c (main_tree_if_conversion): Check new flag.
8154 * omp-low.c (omp_max_vf): Ditto.
8155 (expand_omp_simd): Ditto.
8156 * tree-vectorizer.c (vectorize_loops): Ditto.
8157 (gate_vect_slp): Ditto.
8158 (gate_increase_alignment): Ditto.
8159 * tree-ssa-pre.c (inhibit_phi_insertion): Ditto.
8160 * tree-ssa-loop.c (gate_tree_vectorize): Ditto.
8161 (gate_tree_vectorize): Name change.
8162 (tree_vectorize): Ditto.
8163 (pass_vectorize::gate): Call new function.
8164 (pass_vectorize::execute): Ditto.
8165 * opts.c: O3 default setting change.
8166 (finish_options): Check new flag.
8167 * doc/invoke.texi: Document new flags.
8168 * common.opt: New flags.
8170 2013-09-16 Andreas Schwab <schwab@linux-m68k.org>
8172 * doc/tm.texi.in (Cond Exec Macros): Remove node.
8173 (Condition Code): Don't reference it.
8174 * doc/tm.texi: Regenerate.
8176 2013-09-16 Vladimir Makarov <vmakarov@redhat.com>
8179 * lra-constraints.c (undo_optional_reloads): Consider all optional
8180 reload even if it did not get a hard reg.
8182 2013-09-16 Teresa Johnson <tejohnson@google.com>
8184 * dumpfile.c (dump_loc): Remove newline emission.
8185 * tree-vect-data-refs.c (vect_lanes_optab_supported_p): Add newline
8186 emission to dump_printf_loc calls where missing.
8187 (vect_mark_for_runtime_alias_test): Ditto.
8188 (vect_analyze_data_ref_dependence): Ditto.
8189 (vect_analyze_data_ref_dependences): Ditto.
8190 (vect_slp_analyze_data_ref_dependence): Ditto.
8191 (vect_slp_analyze_data_ref_dependences): Ditto.
8192 (vect_compute_data_ref_alignment): Ditto.
8193 (vect_update_misalignment_for_peel): Ditto.
8194 (vect_verify_datarefs_alignment): Ditto.
8195 (vector_alignment_reachable_p): Ditto.
8196 (vect_get_data_access_cost): Ditto.
8197 (vect_enhance_data_refs_alignment): Ditto.
8198 (vect_find_same_alignment_drs): Ditto.
8199 (vect_analyze_data_refs_alignment): Ditto.
8200 (vect_analyze_group_access): Ditto.
8201 (vect_analyze_data_ref_access): Ditto.
8202 (vect_analyze_data_ref_accesses): Ditto.
8203 (vect_prune_runtime_alias_test_list): Ditto.
8204 (vect_analyze_data_refs): Ditto.
8205 (vect_create_addr_base_for_vector_ref): Ditto.
8206 (vect_create_data_ref_ptr): Ditto.
8207 (vect_grouped_store_supported): Ditto.
8208 (vect_grouped_load_supported): Ditto.
8209 * value-prof.c (check_counter): Ditto.
8210 (check_ic_target): Ditto.
8211 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Ditto.
8212 (vect_recog_widen_mult_pattern): Ditto.
8213 (vect_recog_widen_sum_pattern): Ditto.
8214 (vect_recog_over_widening_pattern): Ditto.
8215 (vect_recog_widen_shift_pattern): Ditto.
8216 (vect_recog_rotate_pattern): Ditto.
8217 (vect_recog_vector_vector_shift_pattern): Ditto.
8218 (vect_recog_divmod_pattern): Ditto.
8219 (vect_recog_mixed_size_cond_pattern): Ditto.
8220 (vect_recog_bool_pattern): Ditto.
8221 (vect_pattern_recog_1): Ditto.
8222 (vect_pattern_recog): Ditto.
8223 * tree-vect-loop.c (vect_determine_vectorization_factor): Ditto.
8224 (vect_is_simple_iv_evolution): Ditto.
8225 (vect_analyze_scalar_cycles_1): Ditto.
8226 (vect_get_loop_niters): Ditto.
8227 (vect_analyze_loop_1): Ditto.
8228 (vect_analyze_loop_form): Ditto.
8229 (vect_analyze_loop_operations): Ditto.
8230 (vect_analyze_loop_2): Ditto.
8231 (vect_analyze_loop): Ditto.
8232 (report_vect_op): Ditto.
8233 (vect_is_slp_reduction): Ditto.
8234 (vect_is_simple_reduction_1): Ditto.
8235 (vect_get_known_peeling_cost): Ditto.
8236 (vect_estimate_min_profitable_iters): Ditto.
8237 (vect_model_reduction_cost): Ditto.
8238 (vect_model_induction_cost): Ditto.
8239 (get_initial_def_for_induction): Ditto.
8240 (vect_create_epilog_for_reduction): Ditto.
8241 (vectorizable_reduction): Ditto.
8242 (vectorizable_induction): Ditto.
8243 (vectorizable_live_operation): Ditto.
8244 (vect_loop_kill_debug_uses): Ditto.
8245 (vect_transform_loop): Ditto.
8246 * tree-vect-stmts.c (vect_mark_relevant): Ditto.
8247 (vect_stmt_relevant_p): Ditto.
8248 (process_use): Ditto.
8249 (vect_mark_stmts_to_be_vectorized): Ditto.
8250 (vect_model_simple_cost): Ditto.
8251 (vect_model_promotion_demotion_cost): Ditto.
8252 (vect_model_store_cost): Ditto.
8253 (vect_get_store_cost): Ditto.
8254 (vect_model_load_cost): Ditto.
8255 (vect_get_load_cost): Ditto.
8256 (vect_init_vector_1): Ditto.
8257 (vect_get_vec_def_for_operand): Ditto.
8258 (vect_finish_stmt_generation): Ditto.
8259 (vectorizable_call): Ditto.
8260 (vectorizable_conversion): Ditto.
8261 (vectorizable_assignment): Ditto.
8262 (vectorizable_shift): Ditto.
8263 (vectorizable_operation): Ditto.
8264 (vectorizable_store): Ditto.
8265 (vectorizable_load): Ditto.
8266 (vectorizable_condition): Ditto.
8267 (vect_analyze_stmt): Ditto.
8268 (vect_transform_stmt): Ditto.
8269 (vect_is_simple_use): Ditto.
8270 * tree-vect-loop-manip.c (slpeel_make_loop_iterate_ntimes): Ditto.
8271 (vect_can_advance_ivs_p): Ditto.
8272 (vect_update_ivs_after_vectorizer): Ditto.
8273 (vect_do_peeling_for_loop_bound): Ditto.
8274 (vect_gen_niters_for_prolog_loop): Ditto.
8275 (vect_update_inits_of_drs): Ditto.
8276 (vect_create_cond_for_alias_checks): Ditto.
8277 * tree-vect-slp.c (vect_get_and_check_slp_defs): Ditto.
8278 (vect_build_slp_tree_1): Ditto.
8279 (vect_supported_load_permutation_p): Ditto.
8280 (vect_analyze_slp_instance): Ditto.
8281 (vect_analyze_slp): Ditto.
8282 (vect_make_slp_decision): Ditto.
8283 (vect_detect_hybrid_slp): Ditto.
8284 (vect_bb_vectorization_profitable_p): Ditto.
8285 (vect_slp_analyze_bb_1): Ditto.
8286 (vect_update_slp_costs_according_to_vf): Ditto.
8287 (vect_get_mask_element): Ditto.
8288 (vect_transform_slp_perm_load): Ditto.
8289 (vect_schedule_slp_instance): Ditto.
8290 (vect_schedule_slp): Ditto.
8291 (vect_slp_transform_bb): Ditto.
8292 * profile.c (read_profile_edge_counts): Ditto.
8293 (compute_branch_probabilities): Ditto.
8294 * coverage.c (get_coverage_counts): Ditto.
8296 2013-09-16 Diego Novillo <dnovillo@google.com>
8298 * tree-core.h: Add missing comment lines from refactoring of tree.h.
8300 2013-09-16 Jan Hubicka <jh@suse.cz>
8302 * gimple-fold.c (can_refer_decl_in_current_unit_p): Do not accept
8303 abstract functions; for static functions check the presence of body.
8305 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
8307 * config/aarch64/aarch64-simd-builtins.def (fma): New.
8308 * config/aarch64/aarch64-simd.md
8309 (aarch64_mla_elt<mode>): New.
8310 (aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
8311 (aarch64_mls_elt<mode>): Likewise.
8312 (aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
8313 (aarch64_fma4_elt<mode>): Likewise.
8314 (aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
8315 (aarch64_fma4_elt_to_128v2df): Likewise.
8316 (aarch64_fma4_elt_to_64df): Likewise.
8317 (fnma<mode>4): Likewise.
8318 (aarch64_fnma4_elt<mode>): Likewise.
8319 (aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
8320 (aarch64_fnma4_elt_to_128v2df): Likewise.
8321 (aarch64_fnma4_elt_to_64df): Likewise.
8322 * config/aarch64/iterators.md (VDQSF): New.
8323 * config/aarch64/arm_neon.h
8324 (vfm<as><sdq>_lane<q>_f<32, 64>): Convert to C implementation.
8325 (vml<sa><q>_lane<q>_<fsu><16, 32, 64>): Likewise.
8327 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
8329 * config/aarch64/aarch64-simd.md (aarch64_mul3_elt<mode>): New.
8330 (aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
8331 (aarch64_mul3_elt_to_128df): Likewise.
8332 (aarch64_mul3_elt_to_64v2df): Likewise.
8333 * config/aarch64/iterators.md (VEL): Also handle DFmode.
8335 (VMUL_CHANGE_NLANES) Likewise.
8338 * config/aarch64/arm_neon.h
8339 (vmul<q>_lane<q>_<suf><16,32,64>): Convert to C implementation.
8341 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
8343 * config/aarch64/arm_neon.h
8344 (vcvtx_high_f32_f64): Fix parameters.
8346 2013-09-16 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8347 Uros Bizjak <ubizjak@gmail.com>
8349 * config/alpha.c: Include tree-ssanames.h.
8351 2013-09-16 Richard Biener <rguenther@suse.de>
8353 * tree-loop-distribution.c (enum rdg_dep_type): Add control_dd.
8354 (dot_rdg_1): Handle control_dd.
8355 (create_edge_for_control_dependence): New function.
8356 (create_rdg_edges): Add control dependences if asked for.
8357 (build_rdg): Likewise.
8358 (generate_loops_for_partition): If there are not necessary
8359 control stmts remove all their dependencies.
8360 (collect_condition_stmts, rdg_flag_loop_exits): Remove.
8361 (distribute_loop): Pass on control dependences.
8362 (tree_loop_distribution): Compute control dependences and remove
8363 restriction on number of loop nodes.
8365 2013-09-16 Jakub Jelinek <jakub@redhat.com>
8367 * ipa-prop.c (ipa_compute_jump_functions_for_edge): Return early
8370 2013-09-16 Richard Sandiford <rdsandiford@googlemail.com>
8372 * cse.c (try_const_anchors): Punt on CC modes.
8374 2013-09-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8376 * config/vax/constraints.md (T): Add missing CONSTANT_P check.
8378 2013-09-14 John David Anglin <danglin@gcc.gnu.org>
8381 * config/pa/pa.c (pa_expand_prologue): Change mode in gen_rtx_POST_INC
8384 2013-09-14 Iain Sandoe <iain@codesourcery.com>
8387 * config/darwin.c (darwin_objc2_section): Note if ObjC Metadata is
8389 (darwin_objc1_section): Likewise.
8390 (darwin_file_end): Emit Image Info section when required.
8392 2013-09-14 Jan Hubicka <jh@suse.cz>
8394 * tree-into-ssa.c (gate_into_ssa): New.
8395 (pass_data_build_ssa): Use it.
8396 * cgraph.h (expand_thunk): Update prototype.
8397 * cgraphunit.c (analyze_function): Expand thunks early.
8398 (expand_thunk): Fix DECL_CONTEXT of reust_decl;
8399 build proper cgraph; set in_ssa_p; clear bogus TREE_ASM_WRITTEN;
8400 set lowered flag; do not add new function.
8401 (assemble_thunks_and_aliases): Update.
8402 * tree-ssa.c (gate_init_datastructures): New gate.
8403 (pass_data_init_datastructures): Use it.
8405 2013-09-14 Iain Sandoe <iain@codesourcery.com>
8408 * config/i386/i386.c (ix86_function_arg_regno_p): Make Darwin use the
8409 xmm register set described in the psABI.
8411 2013-09-13 Evgeny Gavrin <e.gavrin@samsung.com>
8413 * dwarf2out.c (should_emit_struct_debug): Add check
8414 for type_decl variable is not NULL.
8416 2013-09-13 Jacek Caban <jacek@codeweavers.com>
8418 * config.gcc: Use new winnt-c.c target hooks
8419 * config/t-winnt: New file
8420 * config/winnt-c.c: New file
8421 * doc/tm.texi.in: Document new hook
8422 * doc/tm.texi: Regenerated
8424 2013-09-13 Jan Hubicka <jh@suse.cz>
8427 * ipa-inline.c (check_callers): New function.
8428 (check_caller_edge): Remove.
8429 (want_inline_function_to_all_callers_p): Also permit alises that are
8431 (inline_to_all_callers): Terminate the walk when devirtualization
8432 introduce new calls.
8434 2013-09-13 Jan Hubicka <jh@suse.cz>
8436 * ipa-inline-analysis.c (struct growth_data): Add node.
8437 (do_estimate_growth_1): Fix detection of recursion.
8439 2013-09-13 Jakub Jelinek <jakub@redhat.com>
8441 PR tree-optimization/58392
8442 * tree-cfg.c (move_sese_region_to_fn): Rename loop variable
8443 to avoid shadowing of outer loop variable. If
8444 saved_cfun->has_simduid_loops or saved_cfun->has_force_vect_loops,
8445 replace_by_duplicate_decl simduid of loops that have it set and
8446 set dest_cfun->has_simduid_loops and/or
8447 dest_cfun->has_force_vect_loops.
8448 * omp-low.c (build_outer_var_ref): Call maybe_lookup_decl_in_outer_ctx
8449 instead of maybe_lookup_decl.
8450 * tree-inline.c (copy_loops): Change blocks_to_copy argument to id.
8451 Use id->blocks_to_copy instead of blocks_to_copy. Adjust recursive
8452 call. Copy over force_vect and copy and remap simduid. Set
8453 cfun->has_simduid_loops and/or cfun->has_force_vect_loops.
8454 (copy_cfg_body): Remove blocks_to_copy argument. Use
8455 id->blocks_to_copy instead of blocks_to_copy. Adjust copy_loops
8456 caller. Don't set cfun->has_simduid_loops and/or
8457 cfun->has_force_vect_loops here.
8458 (copy_body): Remove blocks_to_copy argument. Adjust copy_cfg_body
8460 (expand_call_inline, tree_function_versioning): Adjust copy_body
8463 2013-09-13 Martin Jambor <mjambor@suse.cz>
8466 * ipa-prop.c (try_make_edge_direct_simple_call): Be less strict in
8467 the assert if the edge was a speculative one.
8469 2013-09-13 Richard Biener <rguenther@suse.de>
8471 * tree-data-ref.h (known_dependences_p): Move here ...
8472 * tree-loop-distribution.c (known_dependences_p): ... from here.
8473 (dump_rdg_component, debug_rdg_component): Remove.
8475 (generate_loops_for_partition): Use gimple_uid instead of
8476 relying on matching stmt visit order.
8477 (rdg_build_partitions): Take starting stmt vector.
8478 (ldist_gen): Merge into ...
8479 (distribute_loop): ... this function. Do not compute starting
8481 * tree-cfg.c (gimple_duplicate_bb): Copy UID for PHIs.
8483 2013-09-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8485 * config/arm/arm.md (arm_cmpsi_insn): Split rI alternative.
8486 Set type attribute correctly. Set predicable_short_it attribute.
8487 (cmpsi_shiftsi): Remove %? from output template.
8489 2013-09-13 Richard Biener <rguenther@suse.de>
8491 * tree-loop-distribution.c (struct rdg_component,
8492 rdg_defs_used_in_other_loops_p, free_rdg_components,
8493 rdg_build_components): Remove.
8494 (stmts_from_loop): Do not record virtual PHIs.
8495 (generate_loops_for_partition): Skip virtual PHIs.
8496 (build_rdg_partition_for_component): Rename to ...
8497 (build_rdg_partition_for_vertex): ... this and adjust.
8498 (rdg_build_partitions): Take a vector of starting vertices
8499 instead of components. Remove unnecessary leftover handling.
8500 (ldist_gen): Do not build components or record other stores.
8501 (distribute_loop): Do not distribute loops containing stmts
8504 2013-09-13 Christian Bruel <christian.bruel@st.com>
8507 * config/sh/sh.md (mov<mode>_reg_reg): Allow memory reloads.
8509 2013-09-13 Kai Tietz <ktietz@redhat.com>
8511 * config.gcc: Separate cases for mingw and cygwin targets,
8512 and add 64-bit cygwin target case.
8514 * config/i386/winnt-cxx.c (i386_pe_type_dllexport_p): Don't
8515 dll-export inline-functions.
8516 * config/i386/winnt.c (i386_pe_determine_dllexport_p): Likewise.
8518 2013-09-13 Jeff Law <law@redhat.com>
8522 2013-09-06 Jeff Law <law@redhat.com>
8524 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
8525 edge implied equivalences into successor phis.
8527 2013-09-12 DJ Delorie <dj@redhat.com>
8529 * config/rl78/rl78-virt.md: Change from | to \; for asm line
8532 2013-09-12 Brooks Moses <bmoses@google.com>
8535 * Makefile.in: Do not install driver binaries in $(target)/bin.
8537 2013-09-12 DJ Delorie <dj@redhat.com>
8539 * config/rl78/rl78.opt (mrelax): New.
8540 * config/rl78/rl78.h (ASM_SPEC): New, pass on -mrelax to gas.
8541 * config/rl78/rl78.h (LINK_SPEC): New, pass on -mrelax to ld.
8543 * config/rl78/rl78.c (rl78_expand_prologue): Use AX to copy
8545 (rl78_expand_epilogue): Likewise.
8547 2013-09-12 Vladimir Makarov <vmakarov@redhat.com>
8550 * lra-eliminations.c (remove_reg_equal_offset_note): New.
8551 (eliminate_regs_in_insn): Rewrite frame pointer to hard frame
8552 pointer elimination with using remove_reg_equal_offset_note.
8554 2013-09-12 DJ Delorie <dj@redhat.com>
8556 * config/msp430/: New port.
8557 * config.gcc (msp430): Added.
8558 * doc/invoke.texi: Document MSP430 options.
8559 * doc/install.texi: Document msp430-elf
8560 * doc/md.texi: Document msp430-elf
8561 * doc/contrib.texi: Document msp430-elf
8563 * cfgexpand.c (expand_debug_expr): Avoid sign-extending SImode to
8566 2013-09-12 Martin Jambor <mjambor@suse.cz>
8569 * ipa-prop.c (remove_described_reference): Give up if the edge in the
8570 reference descriptor is NULL.
8571 (ipa_edge_removal_hook): If owning a reference descriptor, set its
8574 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
8576 * tree-flow.h (FREE_SSANAMES): Move to tree-ssanames.c
8577 (SSANAMES, MODIFIED_NORETURN_CALLS, DEFAULT_DEFS, ptr_info_def,
8578 num_ssa_names, ssa_name): Move to tree-ssanames.h + prototypes.
8579 * tree-flow-inline.h (make_ssa_name, copy_ssa_name, duplicate_ssa_name,
8580 make_temp_ssa_name): move to tree-ssanames.h
8581 * tree-ssa-alias.h: Move prototype.
8582 * tree-ssa.h: Include tree-ssanames.h.
8583 * tree-ssanames.c (FREE_SSANAMES): Move to here.
8584 * tree-ssanames.h: New. Move items from tree-flow*.h
8585 * Makefile.in (tree-ssanames.h): Add to tree-ssanames.o and GTFILES.
8587 2013-09-12 Richard Biener <rguenther@suse.de>
8589 PR tree-optimization/58404
8590 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Also
8591 propagate non-invariant addresses into dereferences wrapped
8592 in component references.
8594 2013-09-12 Richard Biener <rguenther@suse.de>
8596 PR tree-optimization/58402
8597 * passes.def: Move pass_late_warn_uninitialized later.
8599 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
8601 * tree-ssa.h: New. Move content from tree-flow.h and
8603 * tree-flow.h (_edge_var_map, edge_var_map_vector): Move to tree-ssa.h.
8604 Move prototypes belonging to tree-ssa.c.
8605 * tree-flow-inline.h (redirect_edge_var_map_def,
8606 redirect_edge_var_map_result, redirect_edge_var_map_location): Move to
8608 * gimple.h: Adjust prototypes.
8609 * tree-ssa.c (useless_type_conversion_p, types_compatible_p): Move
8611 * gimple.c (useless_type_conversion_p, types_compatible_p): Here.
8612 * tree.h: Move prototype to tree-ssa.h.
8613 * gengtype.c (open_base_files): Replace tree-flow.h with tree-ssa.h.
8614 * Makefile.in: (TREE_SSA_H, TREE_FLOW_H): Adjust dependencies.
8615 * alias.c, asan.c, builtins.c, calls.c, cfgexpand.c, cfghooks.c,
8616 cfgloop.c, cfgloopmanip.c, cgraph.c, cgraphbuild.c, cgraphclones.c,
8617 cgraphunit.c, dse.c, except.c, expr.c, final.c, fold-const.c,
8618 ggc-page.c, gimple-fold.c, gimple-iterator.c, gimple-low.c,
8619 gimple-pretty-print.c, gimple-ssa-strength-reduction.c,
8620 gimple-streamer-in.c, gimple-streamer-out.c, gimple.c, gimplify.c,
8621 graphite-blocking.c, graphite-clast-to-gimple.c,
8622 graphite-dependences.c, graphite-interchange.c,
8623 graphite-optimize-isl.c, graphite-poly.c, graphite-scop-detection.c,
8624 graphite-sese-to-poly.c, graphite.c, ipa-cp.c, ipa-inline-analysis.c,
8625 ipa-inline-transform.c, ipa-inline.c, ipa-prop.c, ipa-pure-const.c,
8626 ipa-reference.c, ipa-split.c, ipa-utils.c,
8627 loop-init.c, lto-cgraph.c, lto-section-in.c, lto-section-out.c,
8628 lto-streamer-in.c, lto-streamer-out.c, lto-streamer.c, omp-low.c,
8629 passes.c, predict.c, print-tree.c, profile.c, sese.c, targhooks.c,
8630 tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
8631 tree-chrec.c, tree-complex.c, tree-data-ref.c, tree-dfa.c, tree-eh.c,
8632 tree-emutls.c, tree-if-conv.c, tree-inline.c, tree-into-ssa.c,
8633 tree-loop-distribution.c, tree-mudflap.c, tree-nested.c, tree-nrv.c,
8634 tree-object-size.c, tree-optimize.c, tree-outof-ssa.c, tree-parloops.c,
8635 tree-phinodes.c, tree-predcom.c, tree-pretty-print.c, tree-profile.c,
8636 tree-scalar-evolution.c, tree-sra.c, tree-ssa*.c, tree-stdarg.c,
8637 tree-streamer-in.c, tree-switch-conversion.c, tree-tailcall.c,
8638 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
8639 tree-vect-loop.c, tree-vect-patterns.c, tree-vect-slp.c,
8640 tree-vect-stmts.c, tree-vectorizer.c, tree-vrp.c, tsan.c,
8641 value-prof.c, var-tracking.c,
8642 varpool.c, vtable-verify.c: Replace tree-flow.h with tree-ssa.h
8644 2013-09-12 Richard Biener <rguenther@suse.de>
8646 PR tree-optimization/58396
8647 * tree-loop-distribution.c (create_rdg_edges): Free unused DDRs.
8648 (build_rdg): Take a loop-nest parameter, fix memleaks.
8649 (distribute_loop): Compute loop-nest here and pass it to build_rdg.
8651 2013-09-12 Yuri Rumyantsev <ysrumyan@gmail.com>
8653 * config/i386/x86-tune.def: Turn on X86_TUNE_AVOID_MEM_OPND_FOR_CMOVE
8656 2013-09-12 Cameron McInally <cameron.mcinally@nyu.edu>
8658 * doc/extend.texi: Fix errors in x86 FMA builtin naming.
8659 The FMA instruction names should have a 'v' prefix.
8661 2013-09-12 Richard Biener <rguenther@suse.de>
8663 * tree-loop-distribution.c (dot_rdg_1): Make graph prettier.
8664 (dot_rdg): Use popen instead of system in optional code.
8665 (remaining_stmts, upstream_mem_writes): Remove global bitmaps.
8666 (already_processed_vertex_p): Adjust.
8667 (has_anti_or_output_dependence, predecessor_has_mem_write,
8668 mark_nodes_having_upstream_mem_writes, has_upstream_mem_writes,
8669 rdg_flag_uses): Remove.
8670 (rdg_flag_vertex): Simplify.
8671 (rdg_flag_vertex_and_dependent): Rely on a correct RDG and
8673 (build_rdg_partition_for_component): Process the first vertex
8674 of a component only.
8675 (ldist_gen): Do not compute remaining_stmts or upstream_mem_writes.
8677 2013-09-12 Alan Modra <amodra@gmail.com>
8679 * config/rs6000/rs6000.c (toc_relative_expr_p): Use add_cint_operand.
8681 2013-09-11 DJ Delorie <dj@redhat.com>
8682 Nick Clifton <nickc@redhat.com>
8684 * config/rl78/predicates.md (rl78_cmp_operator_signed): New.
8685 (rl78_stack_based_mem): New.
8686 * config/rl78/constraints.md (Iv08): New.
8695 * config/rl78/rl78-expand.md (movqi): Reject more SUBREG operands.
8697 (movsi): Change from expand to insn-and-split.
8698 (ashrsi3): Clobber AX.
8702 * config/rl78/rl78.md (CC_REG): Fix.
8703 (addsi3): Allow memory and immediate operands.
8704 (addsi3_internal): Split into...
8705 (addsi3_internal_virt): ...new, and ...
8706 (addsi3_internal_real): ...new.
8708 (subsi3_internal_virt): New.
8709 (subsi3_internal_real): New.
8710 (mulsi3): Add memory operand.
8711 (mulsi3_rl78): Likewise.
8712 (mulsi3_g13): Likewise.
8713 * config/rl78/rl78-real.md (cbranchqi4_real_signed): New.
8714 (cbranchqi4_real): Add more constraint options.
8715 (cbranchhi4_real): Expand pattern.
8716 (cbranchhi4_real_signed): New.
8717 (cbranchhi4_real_inverted): New.
8718 (cbranchsi4_real_lt): New.
8719 (cbranchsi4_real_ge): New.
8720 (cbranchsi4_real_signed): New.
8721 (cbranchsi4_real): New.
8723 * config/rl78/rl78-virt.md (ashrsi3_virt): Add custom cases for
8725 (lshrsi3_virt): Likewise.
8726 (ashlsi3_virt): Likewise.
8727 (cbranchqi4_virt_signed): New.
8728 (cbranchhi4_virt_signed): New.
8729 (cbranchsi4_virt): New.
8730 * config/rl78/rl78.c: Whitespace fixes throughout.
8731 (move_elim_pass): New.
8732 (pass_data_rl78_move_elim): New.
8733 (pass_rl78_move_elim): New.
8734 (make_pass_rl78_move_elim): New.
8735 (rl78_devirt_info): Run devirt earlier.
8736 (rl78_move_elim_info): New.
8737 (rl78_asm_file_start): Register it.
8738 (rl78_split_movsi): New.
8739 (rl78_as_legitimate_address): Allow virtual base registers when
8741 (rl78_addr_space_convert): Remove spurious debug stuff.
8742 (rl78_print_operand_1): Add z,s,S,r,E modifiers.
8743 (rl78_print_operand): More cases for not printing '#'.
8744 (rl78_expand_compare): Remove most of the logic.
8745 (content_memory): New.
8746 (clear_content_memory): New.
8747 (get_content_index): New.
8748 (get_content_name): New.
8749 (display_content_memory): New.
8750 (update_content): New.
8751 (record_content): New.
8752 (already_contains): New.
8753 (insn_ok_now): Re-recog insns with virtual registers.
8754 (add_postponed_content_update): New.
8755 (process_postponed_content_update): New.
8756 (gen_and_emit_move): New.
8757 (transcode_memory_rtx): Record new location content.
8758 Use gen_and_emit_move.
8759 (force_into_acc): New.
8760 (move_to_acc): Use gen_and_emit_move.
8761 (move_from_acc): Likewise.
8762 (move_acc_to_reg): Likewise.
8763 (move_to_x): Likewise.
8764 (move_to_hl): Likewise.
8765 (move_to_de): Likewise.
8766 (rl78_alloc_physical_registers_op1): Record location content.
8767 (has_constraint): New.
8768 (rl78_alloc_physical_registers_op2): Record location content.
8770 (rl78_alloc_physical_registers_ro1): Likewise.
8771 (rl78_alloc_physical_registers_cmp): Likewise.
8772 (rl78_alloc_physical_registers_umul): Likewise.
8773 (rl78_alloc_address_registers_macax): New.
8774 (rl78_alloc_physical_registers): Initialize and set location
8775 content memory as needed.
8776 (rl78_reorg): Make sure split2 is called.
8777 (rl78_rtx_costs): New.
8779 2013-09-11 Richard Sandiford <rdsandiford@googlemail.com>
8781 * simplify-rtx.c (simplify_unary_operation_1): Use simplify_gen_binary
8782 for (not (neg ...)) and (neg (not ...)) cases.
8784 2013-09-11 Richard Biener <rguenther@suse.de>
8787 * passes.def: Split critical edges before late uninit warning passes.
8788 * tree-cfg.c (pass_split_crit_edges): Implement clone method.
8790 2013-09-11 Jakub Jelinek <jakub@redhat.com>
8792 PR tree-optimization/58385
8793 * fold-const.c (build_range_check): If both low and high are NULL,
8794 use omit_one_operand_loc to preserve exp side-effects.
8796 2013-09-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8798 * config/arm/arm.md (arm_shiftsi3): New alternative l/l/M.
8800 2013-09-11 Richard Biener <rguenther@suse.de>
8802 * tree-data-ref.c (dump_rdg_vertex, debug_rdg_vertex,
8803 dump_rdg_component, debug_rdg_component, dump_rdg, debug_rdg,
8804 dot_rdg_1, dot_rdg, rdg_vertex_for_stmt, create_rdg_edge_for_ddr,
8805 create_rdg_edges_for_scalar, create_rdg_edges, create_rdg_vertices,
8806 stmts_from_loop, known_dependences_p, build_empty_rdg,
8807 build_rdg, free_rdg, rdg_defs_used_in_other_loops_p): Move ...
8808 * tree-loop-distribution.c: ... here.
8809 * tree-data-ref.h (struct rdg_vertex, RDGV_STMT, RDGV_DATAREFS,
8810 RDGV_HAS_MEM_WRITE, RDGV_HAS_MEM_READS, RDG_STMT, RDG_DATAREFS,
8811 RDG_MEM_WRITE_STMT, RDG_MEM_READS_STMT, enum rdg_dep_type,
8812 struct rdg_edge, RDGE_TYPE, RDGE_LEVEL, RDGE_RELATION): Move ...
8813 * tree-loop-distribution.c: ... here.
8814 * tree-loop-distribution.c: Include gimple-pretty-print.h.
8815 (struct partition_s): Add loops member.
8816 (partition_alloc, partition_free, rdg_flag_uses, rdg_flag_vertex,
8817 rdg_flag_vertex_and_dependent, rdg_flag_loop_exits,
8818 build_rdg_partition_for_component, rdg_build_partitions): Adjust.
8820 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
8821 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
8822 Sergey Lega <sergey.s.lega@intel.com>
8823 Anna Tikhonova <anna.tikhonova@intel.com>
8824 Ilya Tocar <ilya.tocar@intel.com>
8825 Andrey Turetskiy <andrey.turetskiy@intel.com>
8826 Ilya Verbin <ilya.verbin@intel.com>
8827 Kirill Yukhin <kirill.yukhin@intel.com>
8828 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
8830 * config/i386/constraints.md (k): New.
8832 * config/i386/i386.c (const regclass_map): Add new mask registers.
8833 (dbx_register_map): Ditto.
8834 (dbx64_register_map): Ditto.
8835 (svr4_dbx_register_map): Ditto.
8836 (ix86_conditional_register_usage): Squash mask registers if AVX512F is
8838 (ix86_preferred_reload_class): Disable constants for mask registers.
8839 (ix86_secondary_reload): Do spill of mask register using 32-bit insn.
8840 (ix86_hard_regno_mode_ok): Support new mask registers.
8841 (x86_order_regs_for_local_alloc): Ditto.
8842 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Update.
8843 (FIXED_REGISTERS): Add new mask registers.
8844 (CALL_USED_REGISTERS): Ditto.
8845 (REG_ALLOC_ORDER): Ditto.
8846 (VALID_MASK_REG_MODE): New.
8847 (FIRST_MASK_REG): Ditto.
8848 (LAST_MASK_REG): Ditto.
8849 (reg_class): Add MASK_EVEX_REGS, MASK_REGS.
8850 (MAYBE_MASK_CLASS_P): New.
8851 (REG_CLASS_NAMES): Add MASK_EVEX_REGS, MASK_REGS.
8852 (REG_CLASS_CONTENTS): Ditto.
8853 (MASK_REGNO_P): New.
8854 (ANY_MASK_REG_P): Ditto.
8855 (HI_REGISTER_NAMES): Add new mask registers.
8856 * config/i386/i386.md (MASK0_REG, MASK1_REG, MASK2_REG, MASK3_REG,
8857 MASK4_REG, MASK5_REG, MASK6_REG, MASK7_REG): Constants for new
8859 (attribute "type"): Add mskmov, msklog.
8860 (attribute "length_immediate"): Support them.
8861 (attribute "memory"): Ditto.
8862 (attribute "prefix_0f"): Ditto.
8863 (*movhi_internal): Support new mask registers.
8864 (*movqi_internal): Ditto.
8865 (define_split): Split out clobber pattern is a logic
8866 insn on mask registers.
8867 (*k<logic><mode>): New.
8868 (*andhi_1): Extend to support mask regs.
8869 (*andqi_1): Extend to support mask regs.
8871 (define_split): Split and-not to and and not if operands
8873 (*<code><mode>_1): Separate HI mode to new pattern...
8874 (*<code>hi_1): This.
8875 (*<code>qi_1): Extend to support mask regs.
8877 (kortestzhi): Ditto.
8878 (kortestchi): Ditto.
8880 (*one_cmpl<mode>2_1): Remove HImode and handle it...
8881 (*one_cmplhi2_1): ...Here, now with mask registers support.
8882 (*one_cmplqi2_1): Support new mask registers.
8883 (HI/QImode arithmetics splitter): Don't split if mask registers
8885 (HI/QImode not splitter): Ditto.
8886 * config/i386/predicated.md (mask_reg_operand): New.
8887 (general_reg_operand): Ditto.
8889 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
8891 * doc/invoke.texi: Document fxsr, xsave and xsaveopt options.
8892 * doc/extend.texi: Document fxsr, xsave and xsaveopt builtins.
8894 2013-09-10 Jeff Law <law@redhat.com>
8896 PR tree-optimization/58380
8897 * tree-ssa-threadupdate.c (thread_block): Recognize another case
8898 of threading through a buried loop header.
8900 * tree-ssa-threadedge.c (thread_around_empty_blocks): Correct
8901 return value for single successor case.
8903 2013-09-10 Jan Hubicka <jh@suse.cz>
8905 * ipa-devirt.c (ipa_devirt): Enable with LTO.
8907 2013-09-10 Richard Earnshaw <rearnsha@arm.com>
8910 * arm/vfp.md (combine_vcvt_f32_<FCVTI32typename>): Fix pattern to
8911 support conditional execution.
8912 (combine_vcvt_f64_<FCVTI32typename>): Likewise.
8914 2013-09-10 Vladimir Makarov <vmakarov@redhat.com>
8916 * lra.c (lra): Clear lra_optional_reload_pseudos before every
8918 * lra-constraints.c (curr_insn_transform): Switch on optional reloads.
8919 Check destination too to check move insn.
8920 (undo_optional_reloads): Add check that the original peudo did not
8921 changed its allocation and the optional reload was inherited on last
8922 inheritance pass. Break loop after deciding to keep optional reload.
8923 (lra_undo_inheritance): Add check that inherited pseudo still in
8926 2013-09-10 James Greenhalgh <james.greenhalgh@arm.com>
8928 * config/aarch64/aarch64.md (generic_sched): New.
8929 * config/aarch64/aarch64-generic.md (load): Make conditional
8930 on generic_sched attribute.
8931 (nonload): Likewise.
8933 2013-09-10 Jan Hubicka <jh@suse.cz>
8935 * lto-cgraph.c: Include ipa-utils.h.
8936 (compute_ltrans_boundary): Also add possible targets into the boundary.
8938 2013-09-10 Jan Hubicka <jh@suse.cz>
8940 * gimple-fold.c (gimple_get_virt_method_for_binfo): Pass real
8941 VAR_DECL of vtable rather than full expression.
8943 2013-09-10 Jan Hubicka <jh@suse.cz>
8944 Paolo Carlini <paolo.carlini@oracle.com>
8946 * cgraphunit.c (analyze_functions): Save input_location, set it
8947 to UNKNOWN_LOCATION and restore it at the end.
8949 2013-09-10 Martin Jambor <mjambor@suse.cz>
8951 * ipa-cp.c (propagate_constants_topo): Do not ignore SCC
8952 represented by a thunk.
8954 2013-09-10 Jeff Law <law@redhat.com>
8956 PR tree-optimization/58343
8957 * tree-ssa-threadupdate.c (thread_block): Identify and disable
8958 jump threading requests through loop headers buried in the middle
8959 of a jump threading path.
8961 * tree-ssa-threadedge.c (thread_around_empty_blocks): Fix thinko
8962 in return value/type.
8964 2013-09-10 Jakub Jelinek <jakub@redhat.com>
8966 PR rtl-optimization/58365
8967 * cfgcleanup.c (merge_memattrs): Also clear MEM_READONLY_P
8968 resp. MEM_NOTRAP_P if they differ, or set MEM_VOLATILE_P if
8971 2013-09-10 Richard Biener <rguenther@suse.de>
8973 * tree-data-ref.h (build_rdg): Drop all parameters but loop.
8974 * tree-data-ref.c (create_rdg_vertices): Collect all data
8975 references, signal failure to the caller, use data-ref API.
8976 (build_rdg): Compute data references only once. Maintain lifetime
8977 of data references and data dependences from within RDG.
8978 (free_rdg): Free dependence relations.
8979 * tree-loop-distribution.c (rdg_flag_uses): Drop weird code
8980 inventing extra dependences.
8981 (distribute_loop): Update for RDG API changes.
8983 2013-09-10 Kai Tietz <ktietz@redhat.com>
8985 * doc/invoke.texi (fms-extensions): Document changed
8986 behavior for ms-abi targets.
8987 * config/i386/i386.c (ix86_option_override_internal):
8988 Set default value of option -fms-extension for ms-abi targets.
8990 2013-09-10 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
8992 * config/i386/i386.c (ix86_expand_movmem): Fix epilogue generation.
8994 2013-09-10 Alan Modra <amodra@gmail.com>
8997 * config/rs6000/rs6000.md (bswapdi2_64bit): Disable for volatile mems.
8999 2013-09-10 Alan Modra <amodra@gmail.com>
9001 * config/rs6000/predicates.md (add_cint_operand): New.
9002 (reg_or_add_cint_operand, small_toc_ref): Use add_cint_operand.
9003 * config/rs6000/rs6000.md (largetoc_high_plus): Restrict offset
9004 using add_cint_operand.
9005 (largetoc_high_plus_aix): Likewise.
9007 2013-09-09 Jakub Jelinek <jakub@redhat.com>
9009 PR tree-optimization/58364
9010 * tree-ssa-reassoc.c (init_range_entry): For BIT_NOT_EXPR on
9011 BOOLEAN_TYPE, only invert in_p and continue with arg0 if
9012 the current range can't be an unconditional true or false.
9014 2013-09-09 James Greenhalgh <james.greenhalgh@arm.com>
9016 * config/aarch64/arm_neon.h (vrsqrte_f64): Fix parameter type.
9018 2013-09-09 Uros Bizjak <ubizjak@gmail.com>
9020 * ipa-prop.c (ipa_modify_call_arguments): Initialize deref_align.
9022 2013-09-09 Paolo Carlini <paolo.carlini@oracle.com>
9025 * doc/invoke.texi (-Wdelete-incomplete): Document it.
9027 2013-09-09 Ian Bolton <ian.bolton@arm.com>
9029 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Return
9030 NO_REGS for immediate that can't be moved directly into FP_REGS.
9032 2013-09-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9034 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return CC_SWP for
9035 comparison with negated operand.
9036 * config/aarch64/aarch64.md (compare_neg<mode>): Match canonical
9039 2013-09-09 Richard Biener <rguenther@suse.de>
9042 * cfgloopmanip.c (fix_bb_placements): When fixing the placement
9043 of a subloop record all its block as affecting loop-closed SSA form.
9045 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
9047 * expmed.c (lshift_value): Take an unsigned HOST_WIDE_INT instead
9048 of an rtx/bitpos pair.
9049 (store_fixed_bit_field): Update accordingly.
9051 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
9053 * asan.c (asan_emit_stack_protection): Use gen_int_mode instead of
9055 * builtins.c (expand_errno_check): Likewise.
9056 * dwarf2cfi.c (init_return_column_size): Likewise.
9057 * except.c (sjlj_mark_call_sites): Likewise.
9058 * expr.c (move_by_pieces_1, store_by_pieces_2): Likewise.
9059 * lra-constraints.c (emit_inc): Likewise.
9060 * ree.c (combine_set_extension): Likewise.
9061 * regmove.c (fixup_match_2): Likewise.
9062 * reload1.c (inc_for_reload): Likewise.
9064 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
9066 * combine.c (simplify_set, expand_field_assignment, extract_left_shift)
9067 (force_to_mode, simplify_shift_const_1, simplify_comparison):
9068 Use gen_int_mode with the mode of the associated simplify_* call.
9069 * explow.c (probe_stack_range, anti_adjust_stack_and_probe): Likewise.
9070 * expmed.c (expand_shift_1): Likewise.
9071 * function.c (instantiate_virtual_regs_in_insn): Likewise.
9072 * loop-iv.c (iv_number_of_iterations): Likewise.
9073 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
9074 * simplify-rtx.c (simplify_binary_operation_1): Likewise.
9076 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
9078 * asan.c (asan_clear_shadow): Use gen_int_mode with the mode
9079 of the associated expand_* call.
9080 (asan_emit_stack_protection): Likewise.
9081 * builtins.c (round_trampoline_addr): Likewise.
9082 * explow.c (allocate_dynamic_stack_space, probe_stack_range): Likewise.
9083 * expmed.c (expand_smod_pow2, expand_sdiv_pow2, expand_divmod)
9084 (emit_store_flag): Likewise.
9085 * expr.c (emit_move_resolve_push, push_block, emit_single_push_insn_1)
9086 (emit_push_insn, optimize_bitfield_assignment_op, expand_expr_real_1):
9088 * function.c (instantiate_virtual_regs_in_insn): Likewise.
9089 * ifcvt.c (noce_try_store_flag_constants): Likewise.
9090 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
9091 * modulo-sched.c (generate_prolog_epilog): Likewise.
9092 * optabs.c (expand_binop, widen_leading, expand_doubleword_clz)
9093 (expand_ctz, expand_ffs, expand_unop): Likewise.
9095 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
9097 * alias.c (addr_side_effect_eval): Use gen_int_mode with the mode
9098 of the associated gen_rtx_* call.
9099 * caller-save.c (init_caller_save): Likewise.
9100 * combine.c (find_split_point, make_extraction): Likewise.
9101 (make_compound_operation): Likewise.
9102 * dwarf2out.c (mem_loc_descriptor): Likewise.
9103 * explow.c (plus_constant, probe_stack_range): Likewise.
9104 * expmed.c (expand_mult_const): Likewise.
9105 * expr.c (emit_single_push_insn_1, do_tablejump): Likewise.
9106 * reload1.c (init_reload): Likewise.
9107 * valtrack.c (cleanup_auto_inc_dec): Likewise.
9108 * var-tracking.c (adjust_mems): Likewise.
9109 * modulo-sched.c (sms_schedule): Likewise, but use gen_rtx_GT
9110 rather than gen_rtx_fmt_ee.
9112 2013-09-09 Jan Hubicka <jh@suse.cz>
9115 * value-prof.c (gimple_ic): Copy also abnormal edges.
9117 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
9119 * asan.c (asan_shadow_cst): Use gen_int_mode.
9121 2013-09-08 Jan Hubicka <jh@suse.cz>
9123 * ipa-profile.c: Add toplevel comment.
9124 (ipa_propagate_frequency_1): Be more conservative when profile is read.
9125 (contains_hot_call_p): New function.
9126 (ipa_propagate_frequency): Set frequencies based on counts when
9128 * predict.c (compute_function_frequency): Use PROFILE_READ gueard for
9129 profile; do not tamper with profile after inlining if it is read.
9131 2013-09-08 Jan Hubicka <jh@suse.cz>
9133 * ipa-prop.c (try_make_edge_direct_simple_call): Do not special case
9136 2013-09-08 Jan Hubicka <jh@suse.cz>
9138 * ipa.c (walk_polymorphic_call_targets): Fix redirection before IPA
9141 2013-09-08 Jeff Law <law@redhat.com>
9144 * tree-ssa-threadedge.c (thread_across_edge): Fix initialization
9147 2013-09-08 Andi Kleen <ak@linux.intel.com>
9149 * tree-inline.c (estimate_num_insns): Limit asm cost to 1000.
9151 2013-09-08 Jan Hubicka <jh@suse.cz>
9153 * ipa.c (walk_polymorphic_call_targets): Fix inliner summary update.
9155 2013-09-08 Richard Sandiford <rdsandiford@googlemail.com>
9157 * ira.c (update_equiv_regs): Only call set_paradoxical_subreg
9158 for non-debug insns.
9159 * lra.c (new_insn_reg): Take the containing insn as a parameter.
9160 Only modify lra_reg_info[].biggest_mode if it's non-debug insn.
9161 (collect_non_operand_hard_regs, add_regs_to_insn_regno_info): Update
9164 2013-09-08 Jan Hubicka <jh@suse.cz>
9166 * cgraphunit.c (walk_polymorphic_call_targets): Permit 0 possible
9167 targets and devirtualize to BUILT_IN_UNREACHABLE.
9168 * timevar.def (TV_IPA_UNREACHABLE): New timevar.
9169 * ipa.c (walk_polymorphic_call_targets): New function.
9170 (symtab_remove_unreachable_nodes): Use it; do not keep all virtual
9171 functions; use the new timevar.
9172 * ipa-devirt.c (maybe_record_node): Do not insert static nodes that
9173 was removed from the program.
9174 (record_binfo): If BINFO corresponds to an anonymous namespace, we may
9175 not consider it in the walk when its vtable is dead.
9176 (possible_polymorphic_call_targets_1): Pass anonymous flag to
9178 (devirt_variable_node_removal_hook): New function.
9179 (possible_polymorphic_call_targets): Also register
9180 devirt_variable_node_removal_hook.
9181 (ipa_devirt): Do not do non-speculative devirtualization.
9182 (gate_ipa_devirt): One execute if devirtualizing speculatively.
9184 2013-09-08 Jan Hubicka <jh@suse.cz>
9186 * cgraph.h (varpool_node_hook, varpool_node_hook_list,
9187 varpool_add_node_removal_hook, varpool_add_variable_insertion_hook,
9188 varpool_remove_variable_insertion_hook): Declare.
9189 * varpool.c (varpool_node_hook_list): New structure.
9190 (first_varpool_node_removal_hook,
9191 first_varpool_variable_insertion_hook): New variables.
9192 (varpool_add_node_removal_hook, varpool_remove_node_removal_hook,
9193 varpool_call_node_removal_hooks, varpool_add_variable_insertion_hook,
9194 varpool_remove_variable_insertion_hook,
9195 varpool_call_variable_insertion_hooks): New functions.
9196 (varpool_remove_node): Use it.
9198 2013-09-08 Paolo Carlini <paolo.carlini@oracle.com>
9201 * diagnostic.c (diagnostic_build_prefix): When s.file is
9202 "<built-in>" don't output line and column numbers.
9204 2013-09-06 Jan Hubicka <jh@suse.cz>
9206 * cgraphunit.c (expand_thunk): Get body before touching arguments.
9207 * lto-streamer-out.c: Stream thunks, too.
9208 * lto-streamer-in.c (input_function): Pop cfun here
9209 (lto_read_body): Instead of here.
9211 2013-09-06 Caroline Tice <cmtice@google.com>
9213 * doc/install.texi: Add documentation for the --enable-vtable-verify
9214 and the --disable-libvtv configure options.
9216 2013-09-06 Jeff Law <law@redhat.com>
9218 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
9219 edge implied equivalences into successor phis.
9221 2013-09-06 Joern Rennecke <joern.rennecke@embecosm.com>
9223 * resource.c (mark_referenced_resources): Handle COND_EXEC.
9225 2013-09-06 Claudiu Zissulescu <claziss@synopsys.com>
9227 * resource.c (mark_target_live_regs): Compute resources taking
9228 into account if a call is predicated or not.
9230 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
9232 * toplev.c (output_stack_usage): Be prepared for suffixes created by
9233 the compiler in the function names.
9235 2013-09-06 Jan Hubicka <jh@suse.cz>
9238 * ipa-inline.c (has_caller_p): New function.
9239 (want_inline_function_to_all_callers_p): Use it.
9240 (sum_callers, inline_to_all_callers): Break out from ...
9241 (ipa_inline): ... here.
9243 2013-09-06 Jan Hubicka <jh@suse.cz>
9245 * config/i386/i386.c (ix86_hard_regno_mode_ok): AVX modes are valid
9246 only when AVX is enabled.
9248 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9250 * config/aarch64/aarch64.md
9251 (*movtf_aarch64): Use neon_<ls>dm_2 as type where v8type
9252 is fpsimd_<load/store>2.
9253 (load_pair<mode>): Likewise.
9254 (store_pair<mode>): Likewise.
9256 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9258 * config/arm/types.md (type): Add "mrs" type.
9259 * config/aarch64/aarch64.md
9260 (aarch64_load_tp_hard): Make type "mrs".
9262 (load_tp_hard): Make type "mrs".
9263 * config/arm/cortex-a15.md: Update with new attributes.
9264 * config/arm/cortex-a5.md: Update with new attributes.
9265 * config/arm/cortex-a53.md: Update with new attributes.
9266 * config/arm/cortex-a7.md: Update with new attributes.
9267 * config/arm/cortex-a8.md: Update with new attributes.
9268 * config/arm/cortex-a9.md: Update with new attributes.
9269 * config/arm/cortex-m4.md: Update with new attributes.
9270 * config/arm/cortex-r4.md: Update with new attributes.
9271 * config/arm/fa526.md: Update with new attributes.
9272 * config/arm/fa606te.md: Update with new attributes.
9273 * config/arm/fa626te.md: Update with new attributes.
9274 * config/arm/fa726te.md: Update with new attributes.
9276 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9278 * config/aarch64/aarch64.md
9279 (*movti_aarch64): Use "multiple" for type where v8type is "move2".
9280 (*movtf_aarch64): Likewise.
9282 (thumb1_movdi_insn): Use "multiple" for type where more than one
9283 instruction is used for a move.
9284 (*arm32_movhf): Likewise.
9285 (*thumb_movdf_insn): Likewise.
9287 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9289 * config/arm/types.md (type): Rename fcpys to fmov.
9291 (*arm_movsi_vfp): Rename type fcpys as fmov.
9292 (*thumb2_movsi_vfp): Likewise
9293 (*movhf_vfp_neon): Likewise
9294 (*movhf_vfp): Likewise
9295 (*movsf_vfp): Likewise
9296 (*thumb2_movsf_vfp): Likewise
9297 (*movsfcc_vfp): Likewise
9298 (*thumb2_movsfcc_vfp): Likewise
9299 * config/aarch64/aarch64-simd.md
9300 (move_lo_quad_<mode>): Replace type mov_reg with fmovs.
9301 * config/aarch64/aarch64.md
9302 (*movsi_aarch64): Replace type mov_reg with fmovs.
9303 (*movdi_aarch64): Likewise
9304 (*movsf_aarch64): Likewise
9305 (*movdf_aarch64): Likewise
9307 (cortexa7_older_only): Rename TYPE_FCPYS to TYPE_FMOV.
9308 * config/arm/iwmmxt.md
9309 (*iwmmxt_movsi_insn): Rename type fcpys as fmov.
9310 * config/arm/arm1020e.md: Update with new attributes.
9311 * config/arm/cortex-a15-neon.md: Update with new attributes.
9312 * config/arm/cortex-a5.md: Update with new attributes.
9313 * config/arm/cortex-a53.md: Update with new attributes.
9314 * config/arm/cortex-a7.md: Update with new attributes.
9315 * config/arm/cortex-a8-neon.md: Update with new attributes.
9316 * config/arm/cortex-a9.md: Update with new attributes.
9317 * config/arm/cortex-m4-fpu.md: Update with new attributes.
9318 * config/arm/cortex-r4f.md: Update with new attributes.
9319 * config/arm/marvell-pj4.md: Update with new attributes.
9320 * config/arm/vfp11.md: Update with new attributes.
9322 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9324 * config/aarch64/aarch64.md
9325 (*madd<mode>): Fix type attribute.
9326 (*maddsi_uxtw): Likewise.
9327 (*msub<mode>): Likewise.
9328 (*msubsi_uxtw): Likewise.
9329 (<su_optab>maddsidi4): Likewise.
9330 (<su_optab>msubsidi4): Likewise.
9332 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9334 * config/arm/types.md: Split fdiv<sd> as fsqrt<sd>, fdiv<sd>.
9335 * config/arm/arm.md (core_cycles): Remove fdiv.
9336 * config/arm/vfp.md:
9337 (*sqrtsf2_vfp): Update for attribute changes.
9338 (*sqrtdf2_vfp): Likewise.
9339 * config/aarch64/aarch64.md:
9340 (sqrt<mode>2): Update for attribute changes.
9341 * config/arm/arm1020e.md: Update with new attributes.
9342 * config/arm/cortex-a15-neon.md: Update with new attributes.
9343 * config/arm/cortex-a5.md: Update with new attributes.
9344 * config/arm/cortex-a53.md: Update with new attributes.
9345 * config/arm/cortex-a7.md: Update with new attributes.
9346 * config/arm/cortex-a8-neon.md: Update with new attributes.
9347 * config/arm/cortex-a9.md: Update with new attributes.
9348 * config/arm/cortex-m4-fpu.md: Update with new attributes.
9349 * config/arm/cortex-r4f.md: Update with new attributes.
9350 * config/arm/marvell-pj4.md: Update with new attributes.
9351 * config/arm/vfp11.md: Update with new attributes.
9353 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9355 * config/arm/types.md
9356 (type): Split f_cvt as f_cvt, f_cvtf2i, f_cvti2f.
9357 * config/aarch64/aarch64.md
9358 (l<fcvt_pattern><su_optab><GPF:mode><GPI:mode>2): Update with
9360 (fix_trunc<GPF:mode><GPI:mode>2): Likewise.
9361 (fixuns_trunc<GPF:mode><GPI:mode>2): Likewise.
9362 (float<GPI:mode><GPF:mode>2): Likewise.
9364 (*truncsisf2_vfp): Update with new attributes.
9365 (*truncsidf2_vfp): Likewise.
9366 (fixuns_truncsfsi2): Likewise.
9367 (fixuns_truncdfsi2): Likewise.
9368 (*floatsisf2_vfp): Likewise.
9369 (*floatsidf2_vfp): Likewise.
9370 (floatunssisf2): Likewise.
9371 (floatunssidf2): Likewise.
9372 (*combine_vcvt_f32_<FCVTI32typename>): Likewise.
9373 (*combine_vcvt_f64_<FCVTI32typename>): Likewise.
9374 * config/arm/arm1020e.md: Update with new attributes.
9375 * config/arm/cortex-a15-neon.md: Update with new attributes.
9376 * config/arm/cortex-a5.md: Update with new attributes.
9377 * config/arm/cortex-a53.md: Update with new attributes.
9378 * config/arm/cortex-a7.md: Update with new attributes.
9379 * config/arm/cortex-a8-neon.md: Update with new attributes.
9380 * config/arm/cortex-a9.md: Update with new attributes.
9381 * config/arm/cortex-m4-fpu.md: Update with new attributes.
9382 * config/arm/cortex-r4f.md: Update with new attributes.
9383 * config/arm/marvell-pj4.md: Update with new attributes.
9384 * config/arm/vfp11.md: Update with new attributes.
9386 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9388 * config/aarch64/arm_neon.h
9389 (vqtbl<1,2,3,4><q>_s8): Fix control vector parameter type.
9390 (vqtbx<1,2,3,4><q>_s8): Likewise.
9392 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9394 * config/arm/types.md: Add "no_insn", "multiple" and "untyped" types.
9395 * config/arm/arm-fixed.md: Add type attribute to all insn patterns.
9396 (add<mode>3): Add type attribute.
9397 (add<mode>3): Likewise.
9398 (usadd<mode>3): Likewise.
9399 (ssadd<mode>3): Likewise.
9400 (sub<mode>3): Likewise.
9401 (sub<mode>3): Likewise.
9402 (ussub<mode>3): Likewise.
9403 (sssub<mode>3): Likewise.
9404 (ssmulsa3): Likewise.
9405 (usmulusa3): Likewise.
9406 (arm_usatsihi): Likewise.
9408 (*movdi_vfp): Add types for all instructions.
9409 (*movdi_vfp_cortexa8): Likewise.
9410 (*movhf_vfp_neon): Likewise.
9411 (*movhf_vfp): Likewise.
9412 (*movdf_vfp): Likewise.
9413 (*thumb2_movdf_vfp): Likewise.
9414 (*thumb2_movdfcc_vfp): Likewise.
9415 * config/arm/arm.md: Add type attribute to all insn patterns.
9416 (*thumb1_adddi3): Add type attribute.
9417 (*arm_adddi3): Likewise.
9418 (*adddi_sesidi_di): Likewise.
9419 (*adddi_zesidi_di): Likewise.
9420 (*thumb1_addsi3): Likewise.
9421 (addsi3_compare0): Likewise.
9422 (*addsi3_compare0_scratch): Likewise.
9423 (*compare_negsi_si): Likewise.
9424 (cmpsi2_addneg): Likewise.
9425 (*addsi3_carryin_<optab>): Likewise.
9426 (*addsi3_carryin_alt2_<optab>): Likewise.
9427 (*addsi3_carryin_clobercc_<optab>): Likewise.
9428 (*subsi3_carryin): Likewise.
9429 (*subsi3_carryin_const): Likewise.
9430 (*subsi3_carryin_compare): Likewise.
9431 (*subsi3_carryin_compare_const): Likewise.
9432 (*arm_subdi3): Likewise.
9433 (*thumb_subdi3): Likewise.
9434 (*subdi_di_zesidi): Likewise.
9435 (*subdi_di_sesidi): Likewise.
9436 (*subdi_zesidi_di): Likewise.
9437 (*subdi_sesidi_di): Likewise.
9438 (*subdi_zesidi_ze): Likewise.
9439 (thumb1_subsi3_insn): Likewise.
9440 (*arm_subsi3_insn): Likewise.
9441 (*anddi3_insn): Likewise.
9442 (*anddi_zesidi_di): Likewise.
9443 (*anddi_sesdi_di): Likewise.
9444 (*ne_zeroextracts): Likewise.
9445 (*ne_zeroextracts): Likewise.
9446 (*ite_ne_zeroextr): Likewise.
9447 (*ite_ne_zeroextr): Likewise.
9448 (*anddi_notdi_di): Likewise.
9449 (*anddi_notzesidi): Likewise.
9450 (*anddi_notsesidi): Likewise.
9451 (andsi_notsi_si): Likewise.
9452 (thumb1_bicsi3): Likewise.
9453 (*iordi3_insn): Likewise.
9454 (*iordi_zesidi_di): Likewise.
9455 (*iordi_sesidi_di): Likewise.
9456 (*thumb1_iorsi3_insn): Likewise.
9457 (*xordi3_insn): Likewise.
9458 (*xordi_zesidi_di): Likewise.
9459 (*xordi_sesidi_di): Likewise.
9460 (*arm_xorsi3): Likewise.
9461 (*andsi_iorsi3_no): Likewise.
9462 (*smax_0): Likewise.
9463 (*smax_m1): Likewise.
9464 (*arm_smax_insn): Likewise.
9465 (*smin_0): Likewise.
9466 (*arm_smin_insn): Likewise.
9467 (*arm_umaxsi3): Likewise.
9468 (*arm_uminsi3): Likewise.
9469 (*minmax_arithsi): Likewise.
9470 (*minmax_arithsi_): Likewise.
9471 (*satsi_<SAT:code>): Likewise.
9472 (arm_ashldi3_1bit): Likewise.
9473 (arm_ashrdi3_1bit): Likewise.
9474 (arm_lshrdi3_1bit): Likewise.
9475 (*arm_negdi2): Likewise.
9476 (*thumb1_negdi2): Likewise.
9477 (*arm_negsi2): Likewise.
9478 (*thumb1_negsi2): Likewise.
9479 (*negdi_extendsid): Likewise.
9480 (*negdi_zero_extend): Likewise.
9481 (*arm_abssi2): Likewise.
9482 (*thumb1_abssi2): Likewise.
9483 (*arm_neg_abssi2): Likewise.
9484 (*thumb1_neg_abss): Likewise.
9485 (one_cmpldi2): Likewise.
9486 (extend<mode>di2): Likewise.
9487 (*compareqi_eq0): Likewise.
9488 (*arm_extendhisi2addsi): Likewise.
9489 (*arm_movdi): Likewise.
9490 (*thumb1_movdi_insn): Likewise.
9491 (*arm_movt): Likewise.
9492 (*thumb1_movsi_insn): Likewise.
9493 (pic_add_dot_plus_four): Likewise.
9494 (pic_add_dot_plus_eight): Likewise.
9495 (tls_load_dot_plus_eight): Likewise.
9496 (*thumb1_movhi_insn): Likewise.
9497 (*thumb1_movsf_insn): Likewise.
9498 (*movdf_soft_insn): Likewise.
9499 (*thumb_movdf_insn): Likewise.
9500 (cbranchsi4_insn): Likewise.
9501 (cbranchsi4_scratch): Likewise.
9502 (*negated_cbranchsi4): Likewise.
9503 (*tbit_cbranch): Likewise.
9504 (*tlobits_cbranch): Likewise.
9505 (*tstsi3_cbranch): Likewise.
9506 (*cbranchne_decr1): Likewise.
9507 (*addsi3_cbranch): Likewise.
9508 (*addsi3_cbranch_scratch): Likewise.
9509 (*arm_cmpdi_insn): Likewise.
9510 (*arm_cmpdi_unsig): Likewise.
9511 (*arm_cmpdi_zero): Likewise.
9512 (*thumb_cmpdi_zero): Likewise.
9513 (*deleted_compare): Likewise.
9514 (*mov_scc): Likewise.
9515 (*mov_negscc): Likewise.
9516 (*mov_notscc): Likewise.
9517 (*cstoresi_eq0_thumb1_insn): Likewise.
9518 (cstoresi_nltu_thumb1): Likewise.
9519 (cstoresi_ltu_thu): Likewise.
9520 (thumb1_addsi3_addgeu): Likewise.
9521 (*arm_jump): Likewise.
9522 (*thumb_jump): Likewise.
9523 (*check_arch2): Likewise.
9524 (arm_casesi_internal): Likewise.
9525 (thumb1_casesi_dispatch): Likewise.
9526 (*arm_indirect_jump): Likewise.
9527 (*thumb1_indirect_jump): Likewise.
9529 (*and_scc): Likewise.
9530 (*ior_scc): Likewise.
9531 (*compare_scc): Likewise.
9532 (*cond_move): Likewise.
9533 (*cond_arith): Likewise.
9534 (*cond_sub): Likewise.
9535 (*cmp_ite0): Likewise.
9536 (*cmp_ite1): Likewise.
9537 (*cmp_and): Likewise.
9538 (*cmp_ior): Likewise.
9539 (*ior_scc_scc): Likewise.
9540 (*ior_scc_scc_cmp): Likewise.
9541 (*and_scc_scc): Likewise.
9542 (*and_scc_scc_cmp): Likewise.
9543 (*and_scc_scc_nod): Likewise.
9544 (*negscc): Likewise.
9545 (movcond_addsi): Likewise.
9546 (movcond): Likewise.
9547 (*ifcompare_plus_move): Likewise.
9548 (*if_plus_move): Likewise.
9549 (*ifcompare_move_plus): Likewise.
9550 (*if_move_plus): Likewise.
9551 (*ifcompare_arith_arith): Likewise.
9552 (*if_arith_arith): Likewise.
9553 (*ifcompare_arith_move): Likewise.
9554 (*if_arith_move): Likewise.
9555 (*ifcompare_move_arith): Likewise.
9556 (*if_move_arith): Likewise.
9557 (*ifcompare_move_not): Likewise.
9558 (*if_move_not): Likewise.
9559 (*ifcompare_not_move): Likewise.
9560 (*if_not_move): Likewise.
9561 (*ifcompare_shift_move): Likewise.
9562 (*if_shift_move): Likewise.
9563 (*ifcompare_move_shift): Likewise.
9564 (*if_move_shift): Likewise.
9565 (*ifcompare_shift_shift): Likewise.
9566 (*ifcompare_not_arith): Likewise.
9567 (*ifcompare_arith_not): Likewise.
9568 (*if_arith_not): Likewise.
9569 (*ifcompare_neg_move): Likewise.
9570 (*if_neg_move): Likewise.
9571 (*ifcompare_move_neg): Likewise.
9572 (*if_move_neg): Likewise.
9573 (prologue_thumb1_interwork): Likewise.
9574 (*cond_move_not): Likewise.
9575 (*sign_extract_onebit): Likewise.
9576 (*not_signextract_onebit): Likewise.
9577 (stack_tie): Likewise.
9578 (align_4): Likewise.
9579 (align_8): Likewise.
9580 (consttable_end): Likewise.
9581 (consttable_1): Likewise.
9582 (consttable_2): Likewise.
9583 (consttable_4): Likewise.
9584 (consttable_8): Likewise.
9585 (consttable_16): Likewise.
9586 (*thumb1_tablejump): Likewise.
9587 (prefetch): Likewise.
9588 (force_register_use): Likewise.
9589 (thumb_eh_return): Likewise.
9590 (load_tp_hard): Likewise.
9591 (load_tp_soft): Likewise.
9592 (tlscall): Likewise.
9593 (*arm_movtas_ze): Likewise.
9594 (*arm_rev): Likewise.
9595 (*arm_revsh): Likewise.
9596 (*arm_rev16): Likewise.
9597 * config/arm/thumb2.md
9598 (*thumb2_smaxsi3): Likewise.
9599 (*thumb2_sminsi3): Likewise.
9600 (*thumb32_umaxsi3): Likewise.
9601 (*thumb2_uminsi3): Likewise.
9602 (*thumb2_negdi2): Likewise.
9603 (*thumb2_abssi2): Likewise.
9604 (*thumb2_neg_abss): Likewise.
9605 (*thumb2_movsi_insn): Likewise.
9606 (tls_load_dot_plus_four): Likewise.
9607 (*thumb2_movhi_insn): Likewise.
9608 (*thumb2_mov_scc): Likewise.
9609 (*thumb2_mov_negs): Likewise.
9610 (*thumb2_mov_negs): Likewise.
9611 (*thumb2_mov_nots): Likewise.
9612 (*thumb2_mov_nots): Likewise.
9613 (*thumb2_movsicc_): Likewise.
9614 (*thumb2_movsfcc_soft_insn): Likewise.
9615 (*thumb2_indirect_jump): Likewise.
9616 (*thumb2_and_scc): Likewise.
9617 (*thumb2_ior_scc): Likewise.
9618 (*thumb2_ior_scc_strict_it): Likewise.
9619 (*thumb2_cond_move): Likewise.
9620 (*thumb2_cond_arith): Likewise.
9621 (*thumb2_cond_ari): Likewise.
9622 (*thumb2_cond_sub): Likewise.
9623 (*thumb2_negscc): Likewise.
9624 (*thumb2_movcond): Likewise.
9625 (thumb2_casesi_internal): Likewise.
9626 (thumb2_casesi_internal_pic): Likewise.
9627 (*thumb2_alusi3_short): Likewise.
9628 (*thumb2_mov<mode>_shortim): Likewise.
9629 (*thumb2_addsi_short): Likewise.
9630 (*thumb2_subsi_short): Likewise.
9631 (thumb2_addsi3_compare0): Likewise.
9632 (*thumb2_cbz): Likewise.
9633 (*thumb2_cbnz): Likewise.
9634 (*thumb2_one_cmplsi2_short): Likewise.
9635 (*thumb2_negsi2_short): Likewise.
9636 (*orsi_notsi_si): Likewise.
9637 * config/arm/arm1020e.md: Update with new attributes.
9638 * config/arm/arm1026ejs.md: Update with new attributes.
9639 * config/arm/arm1136jfs.md: Update with new attributes.
9640 * config/arm/arm926ejs.md: Update with new attributes.
9641 * config/arm/cortex-a15.md: Update with new attributes.
9642 * config/arm/cortex-a5.md: Update with new attributes.
9643 * config/arm/cortex-a53.md: Update with new attributes.
9644 * config/arm/cortex-a7.md: Update with new attributes.
9645 * config/arm/cortex-a8.md: Update with new attributes.
9646 * config/arm/cortex-a9.md: Update with new attributes.
9647 * config/arm/cortex-m4.md: Update with new attributes.
9648 * config/arm/cortex-r4.md: Update with new attributes.
9649 * config/arm/fa526.md: Update with new attributes.
9650 * config/arm/fa606te.md: Update with new attributes.
9651 * config/arm/fa626te.md: Update with new attributes.
9652 * config/arm/fa726te.md: Update with new attributes.
9654 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9656 * config/aarch64/aarch64-simd.md
9657 (aarch64_sqdml<SBINQOPS:as>l_n<mode>_internal): Use
9658 <vwx> iterator to ensure correct register choice.
9659 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
9660 (aarch64_sqdmull_n<mode>): Likewise.
9661 (aarch64_sqdmull2_n<mode>_internal): Likewise.
9662 * config/aarch64/arm_neon.h
9663 (vml<as><q>_lane<q>_<su>16): Use 'x' constraint for element vector.
9664 (vml<as><q>_n_<su>16): Likewise.
9665 (vml<as>l_high_lane<q>_<su>16): Likewise.
9666 (vml<as>l_high_n_<su>16): Likewise.
9667 (vml<as>l_lane<q>_<su>16): Likewise.
9668 (vml<as>l_n_<su>16): Likewise.
9669 (vmul<q>_lane<q>_<su>16): Likewise.
9670 (vmul<q>_n_<su>16): Likewise.
9671 (vmull_lane<q>_<su>16): Likewise.
9672 (vmull_n_<su>16): Likewise.
9673 (vmull_high_lane<q>_<su>16): Likewise.
9674 (vmull_high_n_<su>16): Likewise.
9675 (vqrdmulh<q>_n_s16): Likewise.
9677 2013-09-06 Tejas Belagod <tejas.belagod@arm.com>
9679 * config/aarch64/arm_neon.h: Fix all vdup<bhsd_lane<q> intrinsics to
9680 have the correct lane parameter.
9682 2013-09-06 Richard Biener <rguenther@suse.de>
9684 * cfganal.c (control_dependences::~control_dependences):
9685 Properly free all of the vector.
9687 2013-09-06 Kirill Yukhin <kirill.yukhin@intel.com>
9690 * config/i386/i386.c (ix86_conditional_register_usage):
9691 Proper initialize extended SSE registers.
9693 2013-09-06 Jan Hubicka <jh@suse.cz>
9695 PR tree-optimization/58311
9696 * ipa-devirt.c (gate_ipa_devirt): Only execute when optimizing.
9698 2013-09-06 Jan Hubicka <jh@suse.cz>
9700 * Makefile.in (tree-sra.o): Update dependencies.
9701 * tree-sra.c: Include ipa-utils.h
9702 (scan_function): Use recursive_call_p.
9703 (has_caller_p): New function.
9704 (cgraph_for_node_and_aliases): Count also callers of aliases.
9706 2013-09-06 Jan Hubicka <jh@suse.cz>
9709 * cgraph.h (symtab_semantically_equivalent_p): Declare.
9710 * tree-tailcall.c: Include ipa-utils.h.
9711 (find_tail_calls): Use it.
9712 * ipa-pure-const.c (check_call): Likewise.
9713 * ipa-utils.c (recursive_call_p): New function.
9714 * ipa-utils.h (recursive_call_p): Dclare.
9715 * symtab.c (symtab_nonoverwritable_alias): Fix formatting.
9716 (symtab_semantically_equivalent_p): New function.
9717 * Makefile.in (tree-tailcall.o): Update dependencies.
9719 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
9721 * ipa-split.c (split_function): Set DECL_NO_INLINE_WARNING_P on the
9724 2013-09-06 Richard Biener <rguenther@suse.de>
9726 * lto-streamer.h (lto_global_var_decls): Remove.
9727 * Makefile.in (OBJS): Remove lto-symtab.o.
9728 (lto-symtab.o): Remove.
9729 (GTFILES): Remove lto-symtab.c
9730 * lto-symtab.c: Move to lto/
9732 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9734 * config/s390/s390.md (UNSPEC_FPINT_FLOOR, UNSPEC_FPINT_BTRUNC)
9735 (UNSPEC_FPINT_ROUND, UNSPEC_FPINT_CEIL, UNSPEC_FPINT_NEARBYINT)
9736 (UNSPEC_FPINT_RINT): New constant definitions.
9737 (FPINT, fpint_name, fpint_roundingmode): New integer iterator
9738 definition with 2 attributes.
9739 ("<FPINT:fpint_name><BFP:mode>2", "rint<BFP:mode>2")
9740 ("<FPINT:fpint_name><DFP:mode>2", "rint<DFP:mode>2"): New pattern
9743 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9745 * config/s390/s390.md: Add "bcr_flush" value to mnemonic attribute.
9746 ("mem_thread_fence_1"): Use bcr 14,0 for z196 and later.
9747 Set the mnemonic attribute to "bcr_flush". Set the "z196prop"
9748 attribute to "z196_alone".
9749 * config/s390/2827.md: Add "bcr_flush" to "ooo_groupalone" and
9752 2013-09-06 Richard Biener <rguenther@suse.de>
9754 * basic-block.h (class control_dependences): New.
9755 * tree-ssa-dce.c (control_dependence_map): Remove.
9757 (EXECUTE_IF_CONTROL_DEPENDENT): Remove.
9758 (set_control_dependence_map_bit, clear_control_dependence_bitmap,
9759 find_pdom, find_control_dependence, find_all_control_dependences):
9761 (mark_control_dependent_edges_necessary,
9762 find_obviously_necessary_stmts, propagate_necessity, tree_dce_init,
9763 tree_dce_done, perform_tree_ssa_dce): Adjust.
9764 * cfganal.c (set_control_dependence_map_bit,
9765 clear_control_dependence_bitmap, find_pdom, find_control_dependence,
9766 find_all_control_dependences): Move from tree-ssa-dce.c and
9767 implement as methods of control_dependences class.
9768 (control_dependences::control_dependences): New.
9769 (control_dependences::~control_dependences): Likewise.
9770 (control_dependences::get_edges_dependent_on): Likewise.
9771 (control_dependences::get_edge): Likewise.
9773 2013-09-04 Jan Hubicka <jh@suse.cz>
9775 * tree.c (types_same_for_odr): Drop overactive check.
9776 * ipa-devirt.c (hash_type_name): Likewise.
9778 2013-09-04 Jan Hubicka <jh@suse.cz>
9780 * cgraphunit.c (walk_polymorphic_call_targets): Break out from ...
9781 (analyze_functions): ... here.
9783 2013-09-04 Jan Hubicka <jh@suse.cz>
9786 * cgraphunit.c (analyze_functions): Clear AUX fields
9787 after processing; initialize assembler name has.
9789 2013-09-05 Jeff Law <law@redhat.com>
9791 * tree-ssa-threadedge.c (thread_around_empty_blocks): Renamed
9792 from thread_around_empty_block. Record threading path into PATH.
9793 Recurse if threading through the initial block is successful.
9794 (thread_across_edge): Corresponding changes to slightly simplify.
9796 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
9798 * config/aarch64/aarch64.md
9799 (type): Remove frecpe, frecps, frecpx.
9800 (aarch64_frecp<FRECP:frecp_suffix><mode>): Move to aarch64-simd.md,
9801 fix to be a TARGET_SIMD instruction.
9802 (aarch64_frecps): Remove.
9803 * config/aarch64/aarch64-simd.md
9804 (aarch64_frecp<FRECP:frecp_suffix><mode>): New, moved from aarch64.md
9805 (aarch64_frecps<mode>): Handle all float/vector of float modes.
9807 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
9808 Sofiane Naci <sofiane.naci@arm.com>
9810 * config/arm/types.md (define_attr "type"): Expand "arlo_imm"
9811 into "adr", "alu_imm", "alus_imm", "logic_imm", "logics_imm".
9812 Expand "arlo_reg" into "adc_reg", "adc_imm", "adcs_reg", "adcs_imm",
9813 "alu_ext", "alu_reg", "alus_ext", "alus_reg", "bfm", "csel",
9814 "logic_reg", "logics_reg", "rev". Expand "arlo_shift" into
9815 "alu_shift_imm", "alus_shift_imm", "logic_shift_imm",
9816 "logics_shift_imm". Expand "arlo_shift_reg" into "alu_shift_reg",
9817 "alus_shift_reg", "logic_shift_reg", "logics_shift_reg". Expand "clz"
9818 into "clz, "rbit". Rename "shift" to "shift_imm".
9819 * config/arm/arm.md (define_attr "core_cycles"): Update for attribute
9820 changes. Update for attribute changes all occurrences of arlo_* and
9822 * config/arm/arm-fixed.md: Update for attribute changes
9823 all occurrences of arlo_* types.
9824 * config/arm/thumb2.md: Update for attribute changes all occurrences
9826 * config/arm/arm.c (xscale_sched_adjust_cost): (rtx insn, rtx
9827 (cortexa7_older_only): Likewise.
9828 (cortexa7_younger): Likewise.
9829 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
9830 (1020alu_shift_op): Likewise.
9831 (1020alu_shift_reg_op): Likewise.
9832 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
9833 (alu_shift_op): Likewise.
9834 (alu_shift_reg_op): Likewise.
9835 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
9836 (11_alu_shift_op): Likewise.
9837 (11_alu_shift_reg_op): Likewise.
9838 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
9839 (9_alu_shift_reg_op): Likewise.
9840 * config/arm/cortex-a15.md (cortex_a15_alu): Update for
9842 (cortex_a15_alu_shift): Likewise.
9843 (cortex_a15_alu_shift_reg): Likewise.
9844 * config/arm/cortex-a5.md (cortex_a5_alu): Update for
9846 (cortex_a5_alu_shift): Likewise.
9847 * config/arm/cortex-a53.md (cortex_a53_alu): Update for
9849 (cortex_a53_alu_shift): Likewise.
9850 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for
9852 (cortex_a7_alu_reg): Likewise.
9853 (cortex_a7_alu_shift): Likewise.
9854 * config/arm/cortex-a8.md (cortex_a8_alu): Update for
9856 (cortex_a8_alu_shift): Likewise.
9857 (cortex_a8_alu_shift_reg): Likewise.
9858 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
9859 (cortex_a9_dp_shift): Likewise.
9860 * config/arm/cortex-m4.md (cortex_m4_alu): Update for
9862 * config/arm/cortex-r4.md
9863 (cortex_r4_alu): Update for attribute changes.
9864 (cortex_r4_mov): Likewise.
9865 (cortex_r4_alu_shift_reg): Likewise.
9866 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
9867 (526_alu_shift_op): Likewise.
9868 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
9869 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
9870 (626te_alu_shift_op): Likewise.
9871 * config/arm/fa726te.md (726te_alu_op): Update for attribute changes.
9872 (726te_alu_shift_op): Likewise.
9873 (726te_alu_shift_reg_op): Likewise.
9874 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
9875 (mp626_alu_shift_op): Likewise.
9876 * config/arm/marvell-pj4.md (pj4_alu): Update for attribute changes.
9877 (pj4_alu_conds): Likewise.
9878 (pj4_shift): Likewise.
9879 (pj4_shift_conds): Likewise.
9880 (pj4_alu_shift): Likewise.
9881 (pj4_alu_shift_conds): Likewise.
9882 * config/aarch64/aarch64.md: Update for attribute change
9883 all occurrences of arlo_* and shift* types.
9885 2013-09-05 Mike Stump <mikestump@comcast.net>
9887 * tree.h: Move documentation for tree_function_decl to tree-core.h
9888 with the declaration.
9890 2013-09-05 Peter Bergner <bergner@vnet.ibm.com>
9893 * reginfo.c (choose_hard_reg_mode): Scan through all mode classes
9894 looking for widest mode.
9896 2013-09-05 Eric Botcazou <ebotcazou@adacore.com>
9898 * config.gcc (*-*-vxworks*): Do not override an existing extra_objs.
9900 2013-09-05 Richard Biener <rguenther@suse.de>
9902 PR tree-optimization/58137
9903 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size):
9904 Do not create vectors of pointers.
9905 * tree-vect-loop.c (get_initial_def_for_induction): Use proper
9906 types for the components of the vector initializer.
9907 * tree-cfg.c (verify_gimple_assign_binary): Remove special-casing
9908 allowing pointer vectors with PLUS_EXPR/MINUS_EXPR.
9910 2013-09-05 Martin Jambor <mjambor@suse.cz>
9912 * ipa-prop.c (remove_described_reference): Accept missing references,
9913 return false if that hppens, otherwise return true.
9914 (cgraph_node_for_jfunc): New function.
9915 (try_decrement_rdesc_refcount): Likewise.
9916 (try_make_edge_direct_simple_call): Use them.
9917 (ipa_edge_removal_hook): Remove references from rdescs.
9918 (ipa_edge_duplication_hook): Clone rdescs and their references
9919 when the new edge has the same caller as the old one.
9920 * cgraph.c (cgraph_resolve_speculation): Remove speculative
9921 reference before removing any edges.
9923 2013-09-05 Richard Earnshaw <rearnsha@arm.com>
9925 * arm.c (thumb2_emit_strd_push): Rewrite to use pre-decrement on
9927 * thumb2.md (thumb2_storewb_parisi): New pattern.
9929 2013-09-05 Yufeng Zhang <yufeng.zhang@arm.com>
9931 * config/aarch64/aarch64-option-extensions.def: Add
9932 AARCH64_OPT_EXTENSION of 'crc'.
9933 * config/aarch64/aarch64.h (AARCH64_FL_CRC): New define.
9934 (AARCH64_ISA_CRC): Ditto.
9935 * doc/invoke.texi (-march and -mcpu feature modifiers): Add
9936 description of the CRC extension.
9938 2013-09-05 Alexander Ivchenko <alexander.ivchenko@intel.com>
9940 * config/rs6000/linux64.h: Define OPTION_BIONIC and OPTION_UCLIBC.
9941 * config/rs6000/linux.h: Ditto.
9942 * alpha/linux.h: Ditto.
9943 * config/bfin/uclinux.h: Define TARGET_LIBC_HAS_FUNCTION as
9944 no_c99_libc_has_function.
9945 * config/c6x/uclinux-elf.h: Ditto.
9946 * config/lm32/uclinux-elf.h: Ditto.
9947 * config/m68k/uclinux.h: Ditto.
9948 * config/moxie/uclinux.h: Ditto.
9949 * config.gcc (bfin*-linux-uclibc*): Add t-linux-android to tmake_file.
9950 (crisv32-*-linux*, cris-*-linux*): Ditto.
9951 * config/bfin/bfin.c: Include "tm_p.h".
9953 2013-09-05 Richard Biener <rguenther@suse.de>
9955 * tree-vect-loop.c (vect_analyze_loop_operations): Properly
9956 check for a definition without a basic-block.
9958 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
9959 Sofiane Naci <sofiane.naci@arm.com>
9961 * config/aarch64/aarch64.md
9962 (*movti_aarch64): Rename r_2_f and f_2_r.
9963 (*movsf_aarch64): Likewise.
9964 (*movdf_aarch64): Likewise.
9965 (*movtf_aarch64): Likewise.
9966 (aarch64_movdi_<mode>low): Likewise.
9967 (aarch64_movdi_<mode>high): Likewise.
9968 (aarch64_mov<mode>high_di): Likewise.
9969 (aarch64_mov<mode>low_di): Likewise.
9970 (aarch64_movtilow_tilow): Likewise.
9971 * config/arm/arm.md (attribute "neon_type"): Delete. Move attribute
9972 values to config/arm/types.md
9973 (attribute "conds"): Update for attribute change.
9974 (anddi3_insn): Likewise.
9975 (iordi3_insn): Likewise.
9976 (xordi3_insn): Likewise.
9977 (one_cmpldi2): Likewise.
9978 * config/arm/types.md (type): Add Neon types.
9979 * config/arm/neon.md (neon_mov<mode>): Remove "neon_type" attribute,
9980 use "type" attribute.
9981 (movmisalign<mode>_neon_store): Likewise.
9982 (movmisalign<mode>_neon_load): Likewise.
9983 (vec_set<mode>_internal): Likewise.
9984 (vec_setv2di_internal): Likewise.
9985 (vec_extract<mode>): Likewise.
9986 (vec_extractv2di): Likewise.
9987 (add<mode>3_neon): Likewise.
9988 (adddi3_neon): Likewise.
9989 (sub<mode>3_neon): Likewise.
9990 (subdi3_neon): Likewise.
9991 (mul<mode>3_neon): Likewise.
9992 (mul<mode>3add<mode>_neon): Likewise.
9993 (mul<mode>3neg<mode>add<mode>_neon): Likewise.
9994 (fma<VCVTF:mode>4)): Likewise.
9995 (fma<VCVTF:mode>4_intrinsic): Likewise.
9996 (fmsub<VCVTF:mode>4)): Likewise.
9997 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
9998 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
9999 (ior<mode>3): Likewise.
10000 (and<mode>3): Likewise.
10001 (anddi3_neon): Likewise.
10002 (orn<mode>3_neon): Likewise.
10003 (orndi3_neon): Likewise.
10004 (bic<mode>3_neon): Likewise.
10005 (bicdi3_neon): Likewise.
10006 (xor<mode>3): Likewise.
10007 (one_cmpl<mode>2): Likewise.
10008 (abs<mode>2): Likewise.
10009 (neg<mode>2): Likewise.
10010 (umin<mode>3_neon): Likewise.
10011 (umax<mode>3_neon): Likewise.
10012 (smin<mode>3_neon): Likewise.
10013 (smax<mode>3_neon): Likewise.
10014 (vashl<mode>3): Likewise.
10015 (vashr<mode>3_imm): Likewise.
10016 (vlshr<mode>3_imm): Likewise.
10017 (ashl<mode>3_signed): Likewise.
10018 (ashl<mode>3_unsigned): Likewise.
10019 (neon_load_count): Likewise.
10020 (ashldi3_neon_noclobber): Likewise.
10021 (signed_shift_di3_neon): Likewise.
10022 (unsigned_shift_di3_neon): Likewise.
10023 (ashrdi3_neon_imm_noclobber): Likewise.
10024 (lshrdi3_neon_imm_noclobber): Likewise.
10025 (widen_ssum<mode>3): Likewise.
10026 (widen_usum<mode>3): Likewise.
10027 (quad_halves_<code>v4si): Likewise.
10028 (quad_halves_<code>v4sf): Likewise.
10029 (quad_halves_<code>v8hi): Likewise.
10030 (quad_halves_<code>v16qi): Likewise.
10031 (reduc_splus_v2di): Likewise.
10032 (neon_vpadd_internal<mode>): Likewise.
10033 (neon_vpsmin<mode>): Likewise.
10034 (neon_vpsmax<mode>): Likewise.
10035 (neon_vpumin<mode>): Likewise.
10036 (neon_vpumax<mode>): Likewise.
10037 (ss_add<mode>_neon): Likewise.
10038 (us_add<mode>_neon): Likewise.
10039 (ss_sub<mode>_neon): Likewise.
10040 (us_sub<mode>_neon): Likewise.
10041 (neon_vadd<mode>_unspec): Likewise.
10042 (neon_vaddl<mode>): Likewise.
10043 (neon_vaddw<mode>): Likewise.
10044 (neon_vhadd<mode>): Likewise.
10045 (neon_vqadd<mode>): Likewise.
10046 (neon_vaddhn<mode>): Likewise.
10047 (neon_vmul<mode>): Likewise.
10048 (neon_vmla<mode>): Likewise.
10049 (neon_vmlal<mode>): Likewise.
10050 (neon_vmls<mode>): Likewise.
10051 (neon_vmlsl<mode>): Likewise.
10052 (neon_vqdmulh<mode>): Likewise.
10053 (neon_vqdmlal<mode>): Likewise.
10054 (neon_vqdmlsl<mode>): Likewise.
10055 (neon_vmull<mode>): Likewise.
10056 (neon_vqdmull<mode>): Likewise.
10057 (neon_vsub<mode>_unspec): Likewise.
10058 (neon_vsubl<mode>): Likewise.
10059 (neon_vsubw<mode>): Likewise.
10060 (neon_vqsub<mode>): Likewise.
10061 (neon_vhsub<mode>): Likewise.
10062 (neon_vsubhn<mode>): Likewise.
10063 (neon_vceq<mode>): Likewise.
10064 (neon_vcge<mode>): Likewise.
10065 (neon_vcgeu<mode>): Likewise.
10066 (neon_vcgt<mode>): Likewise.
10067 (neon_vcgtu<mode>): Likewise.
10068 (neon_vcle<mode>): Likewise.
10069 (neon_vclt<mode>): Likewise.
10070 (neon_vcage<mode>): Likewise.
10071 (neon_vcagt<mode>): Likewise.
10072 (neon_vtst<mode>): Likewise.
10073 (neon_vabd<mode>): Likewise.
10074 (neon_vabdl<mode>): Likewise.
10075 (neon_vaba<mode>): Likewise.
10076 (neon_vabal<mode>): Likewise.
10077 (neon_vmax<mode>): Likewise.
10078 (neon_vmin<mode>): Likewise.
10079 (neon_vpaddl<mode>): Likewise.
10080 (neon_vpadal<mode>): Likewise.
10081 (neon_vpmax<mode>): Likewise.
10082 (neon_vpmin<mode>): Likewise.
10083 (neon_vrecps<mode>): Likewise.
10084 (neon_vrsqrts<mode>): Likewise.
10085 (neon_vqabs<mode>): Likewise.
10086 (neon_vqneg<mode>): Likewise.
10087 (neon_vcls<mode>): Likewise.
10088 (clz<mode>2): Likewise.
10089 (popcount<mode>2): Likewise.
10090 (neon_vrecpe): Likewise.
10091 (neon_vrsqrte): Likewise.
10092 (neon_vget_lane<mode>_sext_internal): Likewise.
10093 (neon_vget_lane<mode>_zext_internal): Likewise.
10094 (neon_vdup_n<mode>): Likewise.
10095 (neon_vdup_nv2di): Likewise.
10096 (neon_vdpu_lane<mode>_internal): Likewise.
10097 (neon_vswp<mode>): Likewise.
10098 (float<mode><V_cvtto>2): Likewise.
10099 (floatuns<mode><V_cvtto>2): Likewise.
10100 (fix_trunc<mode><V_cvtto>)2): Likewise
10101 (fixuns_trunc<mode><V_cvtto)2): Likewise.
10102 (neon_vcvt<mode>): Likewise.
10103 (neon_vcvtv4sfv4hf): Likewise.
10104 (neon_vcvtv4hfv4sf): Likewise.
10105 (neon_vcvt_n<mode>): Likewise.
10106 (neon_vmovn<mode>): Likewise.
10107 (neon_vqmovn<mode>): Likewise.
10108 (neon_vqmovun<mode>): Likewise.
10109 (neon_vmovl<mode>): Likewise.
10110 (neon_vmul_lane<mode>): Likewise.
10111 (neon_vmull_lane<mode>): Likewise.
10112 (neon_vqdmull_lane<mode>): Likewise.
10113 (neon_vqdmulh_lane<mode>): Likewise.
10114 (neon_vmla_lane<mode>): Likewise.
10115 (neon_vmlal_lane<mode>): Likewise.
10116 (neon_vqdmlal_lane<mode>): Likewise.
10117 (neon_vmls_lane<mode>): Likewise.
10118 (neon_vmlsl_lane<mode>): Likewise.
10119 (neon_vqdmlsl_lane<mode>): Likewise.
10120 (neon_vext<mode>): Likewise.
10121 (neon_vrev64<mode>): Likewise.
10122 (neon_vrev32<mode>): Likewise.
10123 (neon_vrev16<mode>): Likewise.
10124 (neon_vbsl<mode>_internal): Likewise.
10125 (neon_vshl<mode>): Likewise.
10126 (neon_vqshl<mode>): Likewise.
10127 (neon_vshr_n<mode>): Likewise.
10128 (neon_vshrn_n<mode>): Likewise.
10129 (neon_vqshrn_n<mode>): Likewise.
10130 (neon_vqshrun_n<mode>): Likewise.
10131 (neon_vshl_n<mode>): Likewise.
10132 (neon_vqshl_n<mode>): Likewise.
10133 (neon_vqshlu_n<mode>): Likewise.
10134 (neon_vshll_n<mode>): Likewise.
10135 (neon_vsra_n<mode>): Likewise.
10136 (neon_vsri_n<mode>): Likewise.
10137 (neon_vsli_n<mode>): Likewise.
10138 (neon_vtbl1v8qi): Likewise.
10139 (neon_vtbl2v8qi): Likewise.
10140 (neon_vtbl3v8qi): Likewise.
10141 (neon_vtbl4v8qi): Likewise.
10142 (neon_vtbx1v8qi): Likewise.
10143 (neon_vtbx2v8qi): Likewise.
10144 (neon_vtbx3v8qi): Likewise.
10145 (neon_vtbx4v8qi): Likewise.
10146 (neon_vtrn<mode>_internal): Likewise.
10147 (neon_vzip<mode>_internal): Likewise.
10148 (neon_vuzp<mode>_internal): Likewise.
10149 (neon_vld1<mode>): Likewise.
10150 (neon_vld1_lane<mode>): Likewise.
10151 (neon_vld1_dup<mode>): Likewise.
10152 (neon_vld1_dupv2di): Likewise.
10153 (neon_vst1<mode>): Likewise.
10154 (neon_vst1_lane<mode>): Likewise.
10155 (neon_vld2<mode>): Likewise.
10156 (neon_vld2_lane<mode>): Likewise.
10157 (neon_vld2_dup<mode>): Likewise.
10158 (neon_vst2<mode>): Likewise.
10159 (neon_vst2_lane<mode>): Likewise.
10160 (neon_vld3<mode>): Likewise.
10161 (neon_vld3qa<mode>): Likewise.
10162 (neon_vld3qb<mode>): Likewise.
10163 (neon_vld3_lane<mode>): Likewise.
10164 (neon_vld3_dup<mode>): Likewise.
10165 (neon_vst3<mode>): Likewise.
10166 (neon_vst3qa<mode>): Likewise.
10167 (neon_vst3qb<mode>): Likewise.
10168 (neon_vst3_lane<mode>): Likewise.
10169 (neon_vld4<mode>): Likewise.
10170 (neon_vld4qa<mode>): Likewise.
10171 (neon_vld4qb<mode>): Likewise.
10172 (neon_vld4_lane<mode>): Likewise.
10173 (neon_vld4_dup<mode>): Likewise.
10174 (neon_vst4<mode>): Likewise.
10175 (neon_vst4qa<mode>): Likewise.
10176 (neon_vst4qb<mode>): Likewise.
10177 (neon_vst4_lane<mode>): Likewise.
10178 (neon_vec_unpack<US>_lo_<mode>): Likewise.
10179 (neon_vec_unpack<US>_hi_<mode>): Likewise.
10180 (neon_vec_<US>mult_lo_<mode>): Likewise.
10181 (neon_vec_<US>mult_hi_<mode>): Likewise.
10182 (neon_vec_<US>shiftl_<mode>): Likewise.
10183 (neon_unpack<US>_<mode>): Likewise.
10184 (neon_vec_<US>mult_<mode>): Likewise.
10185 (vec_pack_trunc_<mode>): Likewise.
10186 (neon_vec_pack_trunk_<mode>): Likewise.
10187 (neon_vabd<mode>_2): Likewise.
10188 (neon_vabd<mode>_3): Likewise.
10189 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
10190 (thumb2_movsi_vfp): Likewise.
10191 (movdi_vfp): Likewise.
10192 (movdi_vfp_cortexa8): Likewise.
10193 (movhf_vfp_neon): Likewise.
10194 (movhf_vfp): Likewiwse.
10195 (movsf_vfp): Likewiwse.
10196 (thumb2_movsf_vfp): Likewiwse.
10197 (movdf_vfp): Likewise.
10198 (thumb2_movdf_vfp): Likewise.
10199 (movsfcc_vfp): Likewise.
10200 (thumb2_movsfcc_vfp): Likewise.
10201 (movdfcc_vfp): Likewise.
10202 (thumb2_movdfcc_vfp): Likewise.
10203 * config/arm/arm.c (cortexa7_older_only): Update for attribute change.
10204 * config/arm/arm1020e.md (v10_c2v): Update for attribute change.
10205 (v10_v2c): Likewise.
10206 * config/arm/cortex-a15-neon.md (cortex_a15_neon_int_1): Update for
10208 (cortex_a15_neon_int_2): Likewise.
10209 (cortex_a15_neon_int_3): Likewise.
10210 (cortex_a15_neon_int_4): Likewise.
10211 (cortex_a15_neon_int_5): Likewise.
10212 (cortex_a15_neon_vqneg_vqabs): Likewise.
10213 (cortex_a15_neon_vmov): Likewise.
10214 (cortex_a15_neon_vaba): Likewise.
10215 (cortex_a15_neon_vaba_qqq): Likewise.
10216 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
10217 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
10218 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_\
10219 scalar_64_32_long_scalar): Likewise.
10220 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
10221 (cortex_a15_neon_mla_qqq_8_16): Likewise.
10222 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
10223 lotype_qdd_64_32_long): Likewise.
10224 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
10225 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
10226 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
10227 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
10228 (cortex_a15_neon_shift_1): Likewise.
10229 (cortex_a15_neon_shift_2): Likewise.
10230 (cortex_a15_neon_shift_3): Likewise.
10231 (cortex_a15_neon_vshl_ddd): Likewise.
10232 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
10233 (cortex_a15_neon_vsra_vrsra): Likewise.
10234 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
10235 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
10236 (cortex_a15_neon_fp_vmul_ddd): Likewise.
10237 (cortex_a15_neon_fp_vmul_qqd): Likewise.
10238 (cortex_a15_neon_fp_vmla_ddd): Likewise.
10239 (cortex_a15_neon_fp_vmla_qqq): Likewise.
10240 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
10241 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
10242 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
10243 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
10244 (cortex_a15_neon_bp_simple): Likewise.
10245 (cortex_a15_neon_bp_2cycle): Likewise.
10246 (cortex_a15_neon_bp_3cycle): Likewise.
10247 (cortex_a15_neon_vld1_1_2_regs): Likewise.
10248 (cortex_a15_neon_vld1_3_4_regs): Likewise.
10249 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
10250 (cortex_a15_neon_vld2_4_regs): Likewise.
10251 (cortex_a15_neon_vld3_vld4): Likewise.
10252 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
10253 (cortex_a15_neon_vst1_3_4_regs): Likewise.
10254 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Likewise.
10255 (cortex_a15_neon_vst3_vst4): Likewise.
10256 (cortex_a15_neon_vld1_vld2_lane): Likewise.
10257 (cortex_a15_neon_vld3_vld4_lane" 10
10258 (cortex_a15_neon_vst1_vst2_lane): Likewise.
10259 (cortex_a15_neon_vst3_vst4_lane): Likewise.
10260 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
10261 (cortex_a15_neon_ldm_2): Likewise.0
10262 (cortex_a15_neon_stm_2): Likewise.
10263 (cortex_a15_neon_mcr): Likewise.
10264 (cortex_a15_neon_mcr_2_mcrr): Likewise.
10265 (cortex_a15_neon_mrc): Likewise.
10266 (cortex_a15_neon_mrrc): Likewise.
10267 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
10269 (cortex_a15_alu_shift): Likewise.
10270 (cortex_a15_alu_shift_reg): Likewise.
10271 (cortex_a15_mult32): Likewise.
10272 (cortex_a15_mult64): Likewise.
10273 (cortex_a15_block): Likewise.
10274 (cortex_a15_branch): Likewise.
10275 (cortex_a15_load1): Likewise.
10276 (cortex_a15_load3): Likewise.
10277 (cortex_a15_store1): Likewise.
10278 (cortex_a15_store3): Likewise.
10279 (cortex_a15_call): Likewise.
10280 * config/arm/cortex-a5.md (cortex_a5_r2f): Update for attribute change.
10281 (cortex_a5_f2r): Likewise.
10282 * config/arm/cortex-a53.md (cortex_a53_r2f): Update for attribute
10284 (cortex_a53_f2r): Likewise.
10285 * config/arm/cortex-a7.md
10286 (cortex_a7_branch): Update for attribute change.
10287 (cortex_a7_call): Likewise.
10288 (cortex_a7_alu_imm): Likewise.
10289 (cortex_a7_alu_reg): Likewise.
10290 (cortex_a7_alu_shift): Likewise.
10291 (cortex_a7_mul): Likewise.
10292 (cortex_a7_load1): Likewise.
10293 (cortex_a7_store1): Likewise.
10294 (cortex_a7_load2): Likewise.
10295 (cortex_a7_store2): Likewise.
10296 (cortex_a7_load3): Likewise.
10297 (cortex_a7_store3): Likewise.
10298 (cortex_a7_load4): Likewise.
10299 (cortex_a7_store4): Likewise.
10300 (cortex_a7_fpalu): Likewise.
10301 (cortex_a7_fconst): Likewise.
10302 (cortex_a7_fpmuls): Likewise.
10303 (cortex_a7_neon_mul): Likewise.
10304 (cortex_a7_fpmacs): Likewise.
10305 (cortex_a7_neon_mla: Likewise.
10306 (cortex_a7_fpmuld: Likewise.
10307 (cortex_a7_fpmacd: Likewise.
10308 (cortex_a7_fpfmad: Likewise.
10309 (cortex_a7_fdivs: Likewise.
10310 (cortex_a7_fdivd: Likewise.
10311 (cortex_a7_r2f: Likewise.
10312 (cortex_a7_f2r: Likewise.
10313 (cortex_a7_f_flags: Likewise.
10314 (cortex_a7_f_loads: Likewise.
10315 (cortex_a7_f_loadd: Likewise.
10316 (cortex_a7_f_stores: Likewise.
10317 (cortex_a7_f_stored: Likewise.
10318 (cortex_a7_neon): Likewise.
10319 * config/arm/cortex-a8-neon.md
10320 (cortex_a8_neon_mrc): Update for attribute change.
10321 (cortex_a8_neon_mrrc): Likewise.
10322 (cortex_a8_neon_int_1): Likewise.
10323 (cortex_a8_neon_int_2): Likewise.
10324 (cortex_a8_neon_int_3): Likewise.
10325 (cortex_a8_neon_int_4): Likewise.
10326 (cortex_a8_neon_int_5): Likewise.
10327 (cortex_a8_neon_vqneg_vqabs): Likewise.
10328 (cortex_a8_neon_vmov): Likewise.
10329 (cortex_a8_neon_vaba): Likewise.
10330 (cortex_a8_neon_vaba_qqq): Likewise.
10331 (cortex_a8_neon_vsma): Likewise.
10332 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
10333 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
10334 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
10335 long_scalar): Likewise.
10336 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
10337 (cortex_a8_neon_mla_qqq_8_16): Likewise.
10338 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
10339 long_scalar_qdd_64_32_long): Likewise.
10340 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
10341 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
10342 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
10343 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
10344 (cortex_a8_neon_shift_1): Likewise.
10345 (cortex_a8_neon_shift_2): Likewise.
10346 (cortex_a8_neon_shift_3): Likewise.
10347 (cortex_a8_neon_vshl_ddd): Likewise.
10348 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
10349 (cortex_a8_neon_vsra_vrsra): Likewise.
10350 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
10351 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
10352 (cortex_a8_neon_fp_vsum): Likewise.
10353 (cortex_a8_neon_fp_vmul_ddd): Likewise.
10354 (cortex_a8_neon_fp_vmul_qqd): Likewise.
10355 (cortex_a8_neon_fp_vmla_ddd): Likewise.
10356 (cortex_a8_neon_fp_vmla_qqq): Likewise.
10357 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
10358 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
10359 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
10360 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
10361 (cortex_a8_neon_bp_simple): Likewise.
10362 (cortex_a8_neon_bp_2cycle): Likewise.
10363 (cortex_a8_neon_bp_3cycle): Likewise.
10364 (cortex_a8_neon_ldr): Likewise.
10365 (cortex_a8_neon_str): Likewise.
10366 (cortex_a8_neon_vld1_1_2_regs): Likewise.
10367 (cortex_a8_neon_vld1_3_4_regs): Likewise.
10368 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
10369 (cortex_a8_neon_vld2_4_regs): Likewise.
10370 (cortex_a8_neon_vld3_vld4): Likewise.
10371 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
10372 (cortex_a8_neon_vst1_3_4_regs): Likewise.
10373 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
10374 (cortex_a8_neon_vst3_vst4): Likewise.
10375 (cortex_a8_neon_vld1_vld2_lane): Likewise.
10376 (cortex_a8_neon_vld3_vld4_lane): Likewise.
10377 (cortex_a8_neon_vst1_vst2_lane): Likewise.
10378 (cortex_a8_neon_vst3_vst4_lane): Likewise.
10379 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
10380 (cortex_a8_neon_mcr): Likewise.
10381 (cortex_a8_neon_mcr_2_mcrr): Likewise.
10382 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
10383 * config/arm/cortex-a9-neon.md (ca9_neon_mrc): Update for attribute
10385 (ca9_neon_mrrc): Likewise.
10386 (cortex_a9_neon_int_1): Likewise.
10387 (cortex_a9_neon_int_2): Likewise.
10388 (cortex_a9_neon_int_3): Likewise.
10389 (cortex_a9_neon_int_4): Likewise.
10390 (cortex_a9_neon_int_5): Likewise.
10391 (cortex_a9_neon_vqneg_vqabs): Likewise.
10392 (cortex_a9_neon_vmov): Likewise.
10393 (cortex_a9_neon_vaba): Likewise.
10394 (cortex_a9_neon_vaba_qqq): Likewise.
10395 (cortex_a9_neon_vsma): Likewise.
10396 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
10397 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
10398 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
10399 long_scalar): Likewise.
10400 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
10401 (cortex_a9_neon_mla_qqq_8_16): Likewise.
10402 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
10403 long_scalar_qdd_64_32_long): Likewise.
10404 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
10405 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
10406 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
10407 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
10408 (cortex_a9_neon_shift_1): Likewise.
10409 (cortex_a9_neon_shift_2): Likewise.
10410 (cortex_a9_neon_shift_3): Likewise.
10411 (cortex_a9_neon_vshl_ddd): Likewise.
10412 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
10413 (cortex_a9_neon_vsra_vrsra): Likewise.
10414 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
10415 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
10416 (cortex_a9_neon_fp_vsum): Likewise.
10417 (cortex_a9_neon_fp_vmul_ddd): Likewise.
10418 (cortex_a9_neon_fp_vmul_qqd): Likewise.
10419 (cortex_a9_neon_fp_vmla_ddd): Likewise.
10420 (cortex_a9_neon_fp_vmla_qqq): Likewise.
10421 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
10422 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
10423 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
10424 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
10425 (cortex_a9_neon_bp_simple): Likewise.
10426 (cortex_a9_neon_bp_2cycle): Likewise.
10427 (cortex_a9_neon_bp_3cycle): Likewise.
10428 (cortex_a9_neon_ldr): Likewise.
10429 (cortex_a9_neon_str): Likewise.
10430 (cortex_a9_neon_vld1_1_2_regs): Likewise.
10431 (cortex_a9_neon_vld1_3_4_regs): Likewise.
10432 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
10433 (cortex_a9_neon_vld2_4_regs): Likewise.
10434 (cortex_a9_neon_vld3_vld4): Likewise.
10435 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
10436 (cortex_a9_neon_vst1_3_4_regs): Likewise.
10437 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
10438 (cortex_a9_neon_vst3_vst4): Likewise.
10439 (cortex_a9_neon_vld1_vld2_lane): Likewise.
10440 (cortex_a9_neon_vld3_vld4_lane): Likewise.
10441 (cortex_a9_neon_vst1_vst2_lane): Likewise.
10442 (cortex_a9_neon_vst3_vst4_lane): Likewise.
10443 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
10444 (cortex_a9_neon_mcr): Likewise.
10445 (cortex_a9_neon_mcr_2_mcrr): Likewise.
10446 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
10447 (cortex_a9_fps): Likewise.
10448 * config/arm/cortex-m4-fpu.md (cortex_m4_vmov_2): Update for attribute
10450 (cortex_m4_fmuls): Likewise.
10451 * config/arm/cortex-r4f.md (cortex_r4_mcr): Update for attribute
10453 (cortex_r4_mrc): Likewise.
10454 * config/arm/iterators.md: Update comment referring to neon_type.
10455 * config/arm/iwmmxt.md (iwmmxt_arm_movdi): Update for attribute change.
10456 (iwmmxt_movsi_insn): Likewise.
10457 * config/arm/marvell-pj4.md (pj4_vfp_to_core): Update for
10459 (pj4_core_to_vfp): Likewise.
10460 * config/arm/neon-schedgen.ml (emit_insn_reservations): Update for
10462 * config/arm/vfp11.md (vfp_fload): Update for attribute change.
10463 (vfp_fstore): Likewise.
10464 * doc/md.texi: Change references to neon_type to refer to type.
10466 2013-09-04 Dodji Seketeli <dodji@redhat.com>
10468 * tree.h (DECL_BUILT_IN): Fix typo in comment.
10470 2013-09-04 David Edelsohn <dje.gcc@gmail.com>
10472 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Only emit
10473 lglobl if not weak.
10475 2013-09-04 Easwaran Raman <eraman@google.com>
10477 PR middle-end/57370
10478 * tree-ssa-reassoc.c (get_stmt_uid_with_default): New function,
10479 (build_and_add_sum): Use it.
10480 (appears_later_in_bb): Simplify code.
10482 2013-09-04 Teresa Johnson <tejohnson@google.com>
10484 * dumpfile.c (dump_finish): Don't close stderr/stdout.
10486 2013-09-04 James Greenhalgh <james.greenhalgh@arm.com>
10488 * config/aarch64/arm_neon.h (vaddvq_<su>64): Fix return types.
10490 2013-09-04 Jan Hubicka <jh@suse.cz>
10492 * Makefile.in (ipa-devirt.o): Add dependency on diagnostic.h
10493 * ipa-devirt.c: Include diganostic.h
10494 (odr_type_d): Add types and types_set.
10495 (hash_type_name): Work for types with vtables during LTO.
10496 (odr_hasher::remove): Fix comment; destroy types_set.
10497 (add_type_duplicate): New function,
10498 (get_odr_type): Use it.
10499 (dump_type_inheritance_graph): Dump type duplicates.
10500 * ipa.c (symtab_remove_unreachable_nodes): Build type inheritance
10502 * tree.c (types_same_for_odr): Give exact answers on types with
10505 2013-09-04 Dodji Seketeli <dodji@redhat.com>
10507 * tree.h (DECL_BUILT_IN, DECL_IS_BUILTIN): Add more comments
10508 explaining their differences.
10510 2013-09-04 Sandeep Kumar Singh<Sandeep.Singh2@kpitcummins.com>
10512 * config/rx/rx.h: Add option -mcpu for target variants RX100 and RX200.
10514 2013-09-03 Jeff Law <law@redhat.com>
10516 * tree-ssa-threadedge.c (thread_across_edge): Record entire path
10517 when not threading through a joiner block. Pass joiner/no joiner
10518 state to register_jump_thread.
10519 * tree-ssa-threadupdate.c (register_jump_thread): Get joiner/no joiner
10520 state from argument rather than implying on path length.
10521 Dump the entire jump thread path into debugging dump.
10522 * tree-flow.h (register_jump_thread): Update prototype.
10524 2013-08-29 Xinliang David Li <davidxl@google.com>
10526 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
10527 Remove a trivial gcc_assert.
10529 2013-08-29 Xinliang David Li <davidxl@google.com>
10531 * tree-vect-slp.c (destroy_bb_vec_info): Data ref cleanup.
10532 * tree-vect-loop.c (destroy_bb_vec_info): Ditto.
10533 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
10534 Delay base decl alignment adjustment.
10535 * tree-vectorizer.c (vect_destroy_datarefs): New function.
10536 * tree-vectorizer.h: New data structure.
10537 (set_dr_misalignment): New function.
10538 (dr_misalignment): Ditto.
10539 * tree-vect-stmts.c (vectorizable_store): Ensure alignment.
10540 (vectorizable_load): Ditto.
10541 (ensure_base_align): New function.
10542 (vectorize_loops): Add dbg_cnt support.
10543 (execute_vect_slp): Ditto.
10544 * dbgcnt.def: New debug counter.
10545 * Makefile: New dependency.
10547 2013-09-03 Meador Inge <meadori@codesourcery.com>
10551 2013-08-30 Meador Inge <meadori@codesourcery.com>
10553 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
10555 2013-09-03 David Edelsohn <dje.gcc@gmail.com>
10557 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Emit lglobl for
10558 function descriptor.
10560 2013-09-03 Richard Biener <rguenther@suse.de>
10562 * tree-affine.c (add_elt_to_tree): Fix association issue,
10563 avoid useless converts and make sure to always return a
10564 properly typed result.
10566 2013-09-03 Richard Biener <rguenther@suse.de>
10568 PR middle-end/57656
10569 * fold-const.c (negate_expr_p): Fix division case.
10570 (negate_expr): Likewise.
10572 2013-09-03 Richard Biener <rguenther@suse.de>
10575 * tree-streamer-out.c: Include tm.h.
10576 * Makefile.in (tree-streamer-out.o): Depend on $(TM_H).
10578 2013-09-03 Jan Hubicka <jh@suse.cz>
10580 * tree-profile.c (tree_profiling): Cleanup CFG when done.
10582 2013-09-03 Alan Modra <amodra@gmail.com>
10584 * config.gcc (powerpc*-*-linux*): Add support for little-endian
10585 multilibs to big-endian target and vice versa.
10586 * config/rs6000/t-linux64: Use := assignment on all vars.
10587 (MULTILIB_EXTRA_OPTS): Remove fPIC.
10588 (MULTILIB_OSDIRNAMES): Specify using mapping from multilib_options.
10589 * config/rs6000/t-linux64le: New file.
10590 * config/rs6000/t-linux64bele: New file.
10591 * config/rs6000/t-linux64lebe: New file.
10593 2013-09-02 Jan Hubicka <jh@suse.cz>
10595 * ipa-inline-transform.c (inline_transform): Do not
10596 optimize_inline_calls when not optimizing.
10598 2013-09-02 Jan Hubicka <jh@suse.cz>
10600 * lto-symtab.c (lto_symtab_merge_symbols): Add comments; merge
10601 duplicated nodes for assembler names.
10602 * symtab.c (symtab_unregister_node): Do not attempt to unlink
10603 hard registers from assembler name hash.
10605 2013-09-02 Jan Hubicka <jh@suse.cz>
10607 * ipa-split.c (execute_split_functions): Split externally visible
10608 functions called once.
10610 2013-09-02 Martin Jambor <mjambor@suse.cz>
10613 * ipa-prop.c (ipa_edge_duplication_hook): Always put new rdesc to the
10614 linked list. When finding the correct duplicate, also consider also
10615 the caller in additon to its inlined_to node.
10617 2013-09-02 James Greenhalgh <james.greenhalgh@arm.com>
10619 * config/aarch64/aarch64-simd-builtins.def
10620 (dup_lane_scalar): Remove.
10621 * config/aarch64/aarch64-simd.md
10622 (aarch64_simd_dup): Add 'w->w' alternative.
10623 (aarch64_dup_lane<mode>): Allow for VALL.
10624 (aarch64_dup_lane_scalar<mode>): Remove.
10625 (aarch64_dup_lane_<vswap_width_name><mode>): New.
10626 (aarch64_get_lane_signed<mode>): Add w->w altenative.
10627 (aarch64_get_lane_unsigned<mode>): Likewise.
10628 (aarch64_get_lane<mode>): Likewise.
10629 * config/aarch64/aarch64.c (aarch64_evpc_dup): New.
10630 (aarch64_expand_vec_perm_const_1): Use aarch64_evpc_dup.
10631 * config/aarch64/iterators.md (VSWAP_WIDTH): New.
10632 (VCON): Change container of V2SF.
10633 (vswap_width_name): Likewise.
10634 * config/aarch64/arm_neon.h
10635 (__aarch64_vdup_lane_any): New.
10636 (__aarch64_vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
10637 (vdup<q>_n_<psuf><8,16,32,64>): Convert to C implementation.
10638 (vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
10640 2013-09-02 Eric Botcazou <ebotcazou@adacore.com>
10642 PR middle-end/56382
10643 * expr.c (emit_move_complex): Do not move complex FP values as parts if
10644 the source or the destination is a single hard register.
10646 2013-09-02 Richard Biener <rguenther@suse.de>
10648 PR middle-end/57511
10649 * tree-scalar-evolution.c (instantiate_scev_name): Allow
10652 2013-09-02 Richard Biener <rguenther@suse.de>
10654 * tree-affine.c (add_elt_to_tree): Avoid converting all pointer
10655 arithmetic to sizetype.
10657 2013-09-02 Bin Cheng <bin.cheng@arm.com>
10659 * tree-ssa-loop-ivopts.c (set_autoinc_for_original_candidates):
10660 Find auto-increment use both before and after candidate.
10662 2013-09-02 Marek Polacek <polacek@redhat.com>
10664 * Makefile.in (ubsan.o): Add $(TM_P_H) dependency.
10666 2013-09-01 Jan Hubicka <jh@suse.cz>
10668 * Makefile.in: Add ipa-profile.o
10669 (ipa.o, ipa-devrit.o, ipa-inline-analysis.o): Adjust dependencies.
10670 * cgraph.c (struct cgraph_propagate_frequency_data,
10671 cgraph_propagate_frequency_1, cgraph_propagate_frequency): Move to
10672 ipa-profile.c; replace cgraph_ by ipa_ prefix.
10673 * cgraph.h (cgraph_propagate_frequency): Remove.
10674 * ipa-inline-analysis.c: Include ipa-utils.h;
10675 drop duplicated cfgloop.h.
10676 (inline_update_callee_summaries): Update.
10677 * ipa-profile.c: New file.
10678 * ipa-utils.h (ipa_propagate_frequency): Declare.
10679 * ipa.c: Do not include pointer-set.h, hash-table.h, lto-streamer.h,
10680 data-streamer.h, value-prof.h.
10681 (symtab_remove_unreachable_nodes): Update profile.
10682 (struct histogram_entry, histogram, histogram_pool, histogram_hash,
10683 account_time_size, cmp_counts, dump_histogram,
10684 ipa_profile_generate_summary, ipa_profile_write_summary,
10685 ipa_profile_read_summary, ipa_profile, gate_ipa_profile,
10686 pass_data_ipa_profile, pass_ipa_profile, make_pass_ipa_profile):
10687 Move to ipa-profile.c.
10689 2013-09-01 John David Anglin <danglin@gcc.gnu.org>
10691 * config/pa/pa.md: Allow "const 0" operand 1 in "scc" insns.
10693 2013-09-01 Jan Hubicka <jh@suse.cz>
10695 * common.opt (fdevirtualize-speculatively): New function.
10696 * invoke.texi (fdevirtualize-speculatively): Document.
10697 * ipa-devirt.c: Include ipa-inline.h
10698 (likely_target_p): New function.
10699 (ipa_devirt): New function.
10700 (gate_ipa_devirt): New function.
10701 (pass_data_ipa_devirt): New static var.
10702 (pass_ipa_devirt): Likewise.
10703 (make_pass_ipa_devirt): New function.
10704 * opts.c (default_options): Add OPT_fdevirtualize_speculatively.
10705 (common_handle_option): Disable devirtualization when
10706 value range profiling is available.
10707 * passes.def (pass_ipa_devirt): Add.
10708 * timever.def (TV_IPA_DEVIRT): New timevar.
10709 * tree-pass.h (make_pass_ipa_devirt):
10711 2013-09-01 Iain Sandoe <iain@codesourcery.com>
10713 * config/darwin.h (LINK_COMMAND_SPEC_A): Revise sanitizer specs to
10714 include sanitize(undefined).
10716 2013-08-31 Diego Novillo <dnovillo@google.com>
10718 * Makefile.in (TREE_CORE_H): Define.
10720 (GTFILES): Add tree-core.h.
10721 * builtins.c (built_in_class_names): Use BUILT_IN_LAST to
10723 * tree-core.h: New file.
10724 Move all data structures, enum, typedefs, global
10725 declarations and constants from ...
10726 * tree.h: ... here.
10728 2013-08-31 Jan Hubicka <jh@suse.cz>
10730 * bulitins.c (expand_builtin): Do not early exit for gcov
10731 instrumented functions.
10733 2013-08-31 Marek Polacek <polacek@redhat.com>
10735 * ubsan.c: Include tm_p.h.
10737 2013-08-31 Jan Hubicka <jh@suse.cz>
10739 * gimple-streamer-in.c (input_gimple_stmt): Silence parameter unused
10742 * cgraph.c (cgraph_get_body): Update call of lto_input_function_body.
10743 * gimple-streamer-in.c (input_gimple_stmt): Move sanity check to ...
10744 * tree-cfg.c (verify_gimple_label): ... here.
10745 * ipa-utils.c: Include lto-streamer.h, ipa-inline.h
10746 (ipa_merge_profiles): New function.
10747 * lto-streamer-in.c (lto_read_body): Take node instead of fn_decl.
10748 (lto_input_function_body): Likewise.
10749 * ipa-utils.h (ipa_merge_profiles): Declare.
10750 * lto-streamer.h (lto_input_function_body): Update prototype.
10751 (emit_label_in_global_context_p): Remove.
10752 * lto-symtab.c: Include ipa-utils.h
10753 (lto_cgraph_replace_node): Use ipa_merge_profiles.
10755 2013-08-31 Jan Hubicka <jh@suse.cz>
10757 * cgraph.c (cgraph_speculative_call_info): Fix ref lookup
10759 2013-08-31 Jan Hubicka <jh@suse.cz>
10761 * basic-block.h (apply_scale): Make scale parmeter gcov_type.
10763 2013-08-31 Uros Bizjak <ubizjak@gmail.com>
10765 * config/alpha/alpha.c (alpha_emit_conditional_move): Update
10766 "cmp" RTX before signed_comparison_operator check to account
10767 for "code" changes.
10769 2013-08-30 Jan Hubicka <jh@suse.cz>
10771 * ipa-prop.c (ipa_set_jf_known_type): Check that we add only records.
10772 (detect_type_change_1): Rename to ...
10773 (detect_type_change): ... this one; early return on non-polymorphic
10775 (detect_type_change_ssa): Add comp_type parameter; update
10776 use of detect_type_change.
10777 (compute_complex_assign_jump_func): Add param_type parameter;
10778 update use of detect_type_change_ssa.
10779 (compute_complex_ancestor_jump_func): Likewise.
10780 (ipa_get_callee_param_type): New function.
10781 (ipa_compute_jump_functions_for_edge): Compute parameter type;
10782 update calls to the jump function computation functions.
10784 2013-08-30 Teresa Johnson <tejohnson@google.com>
10785 Steven Bosscher <steven@gcc.gnu.org>
10787 * cfgrtl.c (fixup_new_cold_bb): New routine.
10788 (commit_edge_insertions): Invoke fixup_partitions.
10789 (find_partition_fixes): New routine.
10790 (fixup_partitions): Ditto.
10791 (verify_hot_cold_block_grouping): Update comments.
10792 (rtl_verify_edges): Invoke find_partition_fixes.
10793 (rtl_verify_bb_pointers): Update comments.
10794 (rtl_verify_bb_layout): Ditto.
10795 * basic-block.h (probably_never_executed_edge_p): Declare.
10796 (fixup_partitions): Ditto.
10797 * cfgcleanup.c (try_optimize_cfg): Invoke fixup_partitions.
10798 * bb-reorder.c (sanitize_hot_paths): New function.
10799 (find_rarely_executed_basic_blocks_and_crossing_edges): Invoke
10800 sanitize_hot_paths.
10801 * predict.c (probably_never_executed_edge_p): New routine.
10802 * cfg.c (check_bb_profile): Add partition insanity warnings.
10804 2013-08-30 Meador Inge <meadori@codesourcery.com>
10806 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
10808 2013-08-30 Marek Polacek <polacek@redhat.com>
10810 * Makefile.in (ubsan.o): Add.
10811 (c-family/c-ubsan.o): Add.
10812 (builtins.o): Add ubsan.h dependency.
10813 * ubsan.h: New file.
10814 * ubsan.c: New file.
10815 * common.opt: Add -fsanitize=undefined option.
10816 (flag_sanitize): Add variable.
10817 (fsanitize=): Add option. Add Driver.
10818 (fsanitize=thread): Remove option.
10819 (fsanitize=address): Likewise.
10820 (static-libubsan): New option.
10821 * doc/invoke.texi: Document the new flag and -static-libubsan.
10822 * sanitizer.def (DEF_SANITIZER_BUILTIN): Define.
10823 (BUILT_IN_UBSAN_HANDLE_BUILTIN_UNREACHABLE): Define.
10824 * builtin-attrs.def (ATTR_COLD): Define.
10825 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
10826 * builtins.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW,
10827 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS): Define.
10828 * flag-types.h (sanitize_code): New enum.
10829 * opts.c (common_handle_option): Parse command line arguments
10830 of -fsanitize=. Add -fsanitize=unreachable option.
10831 * varasm.c (get_variable_section): Adjust.
10832 (assemble_noswitch_variable): Likewise.
10833 (assemble_variable): Likewise.
10834 (output_constant_def_contents): Likewise.
10835 (categorize_decl_for_section): Likewise.
10836 (place_block_symbol): Likewise.
10837 (output_object_block): Likewise.
10838 * builtins.def: Likewise.
10839 * toplev.c (compile_file): Likewise.
10840 (process_options): Likewise.
10841 * cppbuiltin.c: Likewise.
10842 * tsan.c (tsan_pass): Likewise.
10843 (tsan_gate): Likewise.
10844 (tsan_gate_O0): Likewise.
10845 * cfgexpand.c (partition_stack_vars): Likewise.
10846 (expand_stack_vars): Likewise.
10847 (defer_stack_allocation): Likewise.
10848 (expand_used_vars): Likewise.
10849 * cfgcleanup.c (old_insns_match_p): Likewise.
10850 * asan.c (asan_finish_file): Likewise.
10851 (asan_instrument): Likewise.
10852 (gate_asan): Likewise.
10853 (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR_PTR.
10854 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
10855 (asan_global_struct): Use pointer_sized_int_node instead
10856 calling build_nonstandard_integer_type.
10857 (initialize_sanitizer_builtins): Likewise.
10858 (asan_finish_file): Likewise.
10859 * gcc.c: Document %{%:function(args):X}.
10860 (static_spec_functions): Add sanitize.
10861 (handle_spec_function): Add retval_nonnull argument and if non-NULL,
10862 store funcval != NULL there.
10863 (do_spec_1): Adjust handle_spec_function caller.
10864 (handle_braces): Allow %:function(args) as condition.
10865 (sanitize_spec_function): New function.
10866 (ADD_STATIC_LIBUBSAN_LIBS): Define.
10867 (LIBUBSAN_SPEC): Likewise.
10868 (LIBUBSAN_EARLY_SPEC): Likewise.
10869 (SANITIZER_SPEC): Handle libubsan.
10870 (SANITIZER_EARLY_SPEC): Likewise.
10871 * config/darwin.h (LINK_COMMAND_SPEC_A): Use %:sanitize(address)
10872 instead of fsanitize=address.
10873 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Use %:sanitize(address)
10874 instead of fsanitize=address*.
10875 * builtins.c: Include ubsan.h.
10876 (fold_builtin_0): Instrument __builtin_unreachable.
10877 * config/rs6000/rs6000.h (FRAME_GROWS_DOWNWARD): Use flag_sanitize
10878 instead of flag_asan.
10879 * tree.h (enum tree_index): Add TI_POINTER_SIZED_TYPE.
10880 (pointer_sized_int_node): Define.
10881 * tree.c (build_common_tree_nodes): Initialize pointer_sized_int_node.
10883 2013-08-30 Mike Stump <mikestump@comcast.net>
10885 * doc/install.texi (Prerequisites): Note regression in Tcl 8.6
10888 2013-08-29 Jan Hubicka <jh@suse.cz>
10890 * cgraph.c (cgraph_function_body_availability): Handle weakref
10892 * passes.def: Remove pass_fixup_cfg.
10893 * ipa-inline.c (ipa_inline): When not optimizing, do not inline;
10894 track when we need to remove functions.
10895 (gate_ipa_inline): Execute inlining always; add comment why.
10896 (pass_data_ipa_inline): Remove TODO_remove_functions.
10897 * ipa-inline-analysis.c (inline_generate_summary): When not optimizing
10898 do not produce summaries.
10899 * symtab.c (change_decl_assembler_name): Handle renaming of weakrefs.
10900 (symtab_nonoverwritable_alias): Assert we are not called on weakref.
10901 * varpool.c (cgraph_variable_initializer_availability): Fix weakrefs,
10902 constant pool and vtable.
10904 2013-08-30 Tejas Belagod <tejas.belagod@arm.com>
10906 * config/aarch64/arm_neon.h (__AARCH64_UINT64_C, __AARCH64_INT64_C):
10907 New arm_neon.h's internal macros to specify 64-bit constants.
10908 Avoid using stdint.h's macros.
10910 2013-08-30 Joern Rennecke <joern.rennecke@embecosm.com>
10912 * recog.c (verify_changes): Verify that changes[i].old is non-zero
10913 before applying REG_P.
10915 2013-08-30 Jakub Jelinek <jakub@redhat.com>
10917 PR tree-optimization/58277
10918 * tree-ssa-strlen.c (strlen_enter_block): If do_invalidate gave up
10919 after seeing too many stmts with vdef in between dombb and current
10920 bb, invalidate everything.
10922 2013-08-30 Richard Biener <rguenther@suse.de>
10924 * fold-const.c (fold_single_bit_test): Fix overflow test.
10926 2013-08-30 Eric Botcazou <ebotcazou@adacore.com>
10928 * function.c (assign_parm_setup_reg): For a parameter passed by pointer
10929 and which can live in a register, always retrieve the value on entry.
10930 * var-tracking.c (add_stores): Treat the copy on entry for a parameter
10931 passed by invisible reference specially.
10932 (emit_notes_in_bb) <MO_VAL_USE>: Emit notes before the instruction.
10933 (vt_add_function_parameter): Correctly deal with a parameter passed by
10934 invisible reference.
10936 2013-08-30 Jan Hubicka <jh@suse.cz>
10938 * tree.c (set_call_expr_flags): Fix handling of TM_PURE.
10940 2013-08-30 Richard Biener <rguenther@suse.de>
10942 PR tree-optimization/58228
10943 * tree-vect-data-refs.c (vect_analyze_data_ref_access): Do not
10944 allow invariant loads in nested loop vectorization.
10946 2013-08-30 Richard Biener <rguenther@suse.de>
10948 PR tree-optimization/58223
10949 * tree-loop-distribution.c (has_anti_dependence): Rename to ...
10950 (has_anti_or_output_dependence): ... this and adjust to also
10951 look for output dependences.
10952 (mark_nodes_having_upstream_mem_writes): Adjust.
10953 (rdg_flag_uses): Likewise.
10955 2013-08-30 Richard Biener <rguenther@suse.de>
10957 PR tree-optimization/58010
10958 * tree-vect-loop.c (vect_create_epilog_for_reduction): Remove
10959 assert that we have a loop-closed PHI.
10961 2013-08-29 Jan Hubicka <jh@suse.cz>
10963 * lto-symtab.c (lto_cgraph_replace_node): Free decl_in_state.
10964 * cgraph.c (cgraph_release_function_body): Free decl_in_state.
10965 * lto-section-in.c (lto_free_function_in_decl_state): New function.
10966 (lto_free_function_in_decl_state_for_node): New function.
10968 2013-08-29 Xinliang David Li <davidxl@google.com>
10970 * loop-unroll.c (report_unroll_peel): Minor message change.
10971 * tree-vect-loop-manip.c (vect_do_peeling_for_alignment):
10972 Emit alignment peeling message with default -fopt-info.
10973 (vect_loop_versioning): Emit loop version info message.
10974 * tree-vectorizer.c (vectorize_loops): Minor message change.
10975 (execute_vect_slp): Ditto.
10977 2013-08-29 Eric Botcazou <ebotcazou@adacore.com>
10979 * cgraphclones.c (cgraph_create_virtual_clone): Compute the DECL_NAME
10980 of the clone from the DECL_NAME of the original function.
10982 2013-08-29 Oleg Endo <olegendo@gcc.gnu.org>
10984 * passes.c (register_pass): Add overload.
10985 * tree-pass.h (register_pass): Forward declare it. Add comment.
10987 2013-08-29 Jan Hubicka <jh@suse.cz>
10989 * lto-streamer-out.c (hash_tree): Stream DECL_FINAL_P,
10990 DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P and TYPE_FINAL_P.
10991 * lto-streamer-in.c (unpack_ts_decl_with_vis_value_fields): Stream
10992 DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
10993 (unpack_ts_type_common_value_fields): Stream TYPE_FINAL_P.
10994 * tree-streamer-out.c (pack_ts_decl_with_vis_value_fields):
10995 Add DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
10996 (pack_ts_type_common_value_fields): Add TYPE_FINAL_P.
10998 2013-08-29 Teresa Johnson <tejohnson@google.com>
11000 * dumpfile.c (dump_loc): Output column number.
11001 * dumpfile.h (OPTGROUP_OTHER): Add and enable under OPTGROUP_ALL.
11002 * doc/invoke.texi: Document optall -fopt-info flag.
11003 * profile.c (read_profile_edge_counts): Use new dump framework.
11004 (compute_branch_probabilities): Ditto.
11005 * passes.c (pass_manager::register_one_dump_file): Use OPTGROUP_OTHER
11006 when pass not in any opt group.
11007 * pass_manager.h (pass_manager::get_pass_profile): New method.
11008 * value-prof.c (check_counter): Use new dump framework.
11009 (check_ic_target): Ditto.
11010 * coverage.c (get_coverage_counts): Ditto.
11011 (coverage_init): Setup new dump framework.
11013 2013-08-29 Richard Biener <rguenther@suse.de>
11015 PR tree-optimization/58246
11016 * tree-ssa-dce.c (mark_aliased_reaching_defs_necessary_1): Properly
11017 handle the dominance check inside a basic-block.
11019 2013-08-29 Richard Biener <rguenther@suse.de>
11021 PR middle-end/57287
11022 * tree-ssa-copy.c (may_propagate_copy): Allow propagating
11023 of default defs that appear in abnormal PHI nodes.
11025 2013-08-29 Richard Biener <rguenther@suse.de>
11027 PR tree-optimization/57685
11028 * tree-vrp.c (register_edge_assert_for_1): Recurse only for
11029 single-use operands to avoid exponential complexity.
11031 2013-08-28 Dehao Chen <dehao@google.com>
11033 * ipa-inline.c (edge_badness): Fix integer underflow.
11035 2013-08-28 Uros Bizjak <ubizjak@gmail.com>
11037 * gtm-builtins.def (_ITM_free): Declare leaf.
11039 2013-08-28 Jakub Jelinek <jakub@redhat.com>
11042 * config/i386/i386.md (*tls_global_dynamic_64_largepic): New insn.
11043 (*tls_local_dynamic_base_64_largepic): Likewise.
11044 (tls_global_dynamic_64_<mode>, tls_local_dynamic_base_64_<mode>):
11045 Remove predicate from call operand.
11046 * config/i386/i386.c (ix86_tls_get_addr): For -mcmodel=large -fpic
11047 return sum of pic_offset_table_rtx and UNSPEC_PLTOFF of the symbol.
11049 2013-08-28 Jeff Law <law@redhat.com>
11051 * tree-ssa-threadedge.c (thread_around_empty_block): Remove
11052 checks for the number of predecessors and successors allowed.
11053 * tree-ssa-threadupdate.c (mark_threaded_blocks): Ignore requests
11054 which require copying a joiner block if there is a request which
11055 is a subpath that requires no joiner block copying.
11057 2013-08-28 Jan Hubicka <jh@suse.cz>
11059 * lto-streamer-out.c (DFS_write_tree_body): Drop
11060 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX and BINFO_VPTR_INDEX.
11061 (hash_tree): Do not hash DECL_DEFER_OUTPUT, BINFO_INHERITANCE_CHAIN,
11062 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX, DECL_IN_TEXT_SECTION.
11063 * tree-streamer-in.c (unpack_ts_decl_common_value_fields):
11064 Do not read DECL_ERROR_ISSUED.
11065 (unpack_ts_decl_with_vis_value_fields): Do not read
11067 (lto_input_ts_binfo_tree_pointers): Do not read
11068 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX
11069 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Do not
11070 write DECL_ERROR_ISSUED..
11071 (pack_ts_decl_with_vis_value_fields): Do not write
11073 (write_ts_binfo_tree_pointers): Do not read BINFO_INHERITANCE_CHAIN,
11074 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX.
11075 * print-tree.c (print_node): Do not print DECL_ERROR_ISSUED.
11076 * tree.h (tree_decl_common): Update comment.
11077 (DECL_ERROR_ISSUED): Remove.
11079 2013-08-28 Jakub Jelinek <jakub@redhat.com>
11081 PR middle-end/58257
11082 * omp-low.c (copy_var_decl): Copy over TREE_NO_WARNING flag.
11084 2013-08-28 Jan Hubicka <jh@suse.cz>
11086 * builtins.def (free): Declare leaf.
11088 2013-08-27 David Malcolm <dmalcolm@redhat.com>
11090 * gdbhooks.py: New.
11091 * configure.ac (gdbinit.in): Add import of gcc/gdbhooks.py.
11092 * configure: Regenerate.
11094 2013-08-27 Martin Jambor <mjambor@suse.cz>
11096 * ipa-prop.h (ipa_pass_through_data): New field type_preserved.
11097 (ipa_ancestor_jf_data): Likewise.
11098 (ipa_get_jf_pass_through_agg_preserved): Fix comment typo.
11099 (ipa_get_jf_pass_through_type_preserved): New function.
11100 (ipa_get_jf_ancestor_agg_preserved): Fix comment typo.
11101 (ipa_get_jf_ancestor_type_preserved): New function.
11102 * ipa-cp.c (ipa_get_jf_pass_through_result): Honor type_preserved flag.
11103 (ipa_get_jf_ancestor_result): Likewise.
11104 (propagate_vals_accross_pass_through): Use
11105 ipa_get_jf_pass_through_result to do all the value mappings.
11106 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Dump the
11107 type_preserved flag.
11108 (ipa_set_jf_cst_copy): New function.
11109 (ipa_set_jf_simple_pass_through): Set the type_preserved flag.
11110 (ipa_set_jf_arith_pass_through): Likewise.
11111 (ipa_set_ancestor_jf): Likewise.
11112 (compute_complex_assign_jump_func): Set type_preserved instead of
11114 (ipa_compute_jump_functions_for_edge): Likewise.
11115 (combine_known_type_and_ancestor_jfs): Honor type_preserved.
11116 (update_jump_functions_after_inlining): Update type_preserved.
11117 Explicitely create jump functions when combining one with pass_through.
11118 (ipa_write_jump_function): Stream the type_preserved flags.
11119 (ipa_read_jump_function): Likewise.
11121 2013-08-27 Jakub Jelinek <jakub@redhat.com>
11122 Aldy Hernandez <aldyh@redhat.com>
11124 * Makefile.in (omp-low.o): Depend on $(TARGET_H).
11125 * cfgloop.h (struct loop): Add safelen, force_vect, simduid.
11126 * function.h (struct function): Add has_force_vect_loops and
11128 * gimple-pretty-print.c (dump_gimple_omp_for): Handle GF_OMP_FOR_KIND*.
11129 * gimple.c (gimple_build_omp_critical): Add KIND argument and
11131 * gimple.def: Update CLAUSES comments.
11132 * gimple.h (enum gf_mask): Add GF_OMP_FOR_KIND_{FOR,SIMD}.
11133 (gimple_build_omp_for): Add argument to prototype.
11134 (gimple_omp_for_kind): New.
11135 (gimple_omp_for_set_kind): New.
11136 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_LINEAR to
11137 GOVD_DATA_SHARE_CLASS.
11138 (enum omp_region_type): Add ORT_SIMD.
11139 (gimple_add_tmp_var): Handle ORT_SIMD.
11140 (gimplify_var_or_parm_decl): Same.
11141 (is_gimple_stmt): Same.
11142 (omp_firstprivatize_variable): Same.
11143 (omp_add_variable): Only use splay_tree_insert if lookup failed.
11144 (omp_notice_variable): Handle ORT_SIMD.
11145 (omp_is_private): Add SIMD argument and handle it as well as ORT_SIMD.
11146 (omp_check_private): Handle ORT_SIMD.
11147 (gimplify_scan_omp_clauses): Handle OMP_CLAUSE_LINEAR and
11148 OMP_CLAUSE_SAFELEN.
11149 (gimplify_adjust_omp_clauses_1): Handle GOVD_LINEAR.
11150 Handle OMP_CLAUSE_LASTPRIVATE.
11151 (gimplify_adjust_omp_clauses): Handle OMP_CLAUSE_LINEAR and
11152 OMP_CLAUSE_SAFELEN.
11153 (gimplify_omp_for): Handle OMP_SIMD and OMP_CLAUSE_LINEAR.
11154 (gimplify_expr): Handle OMP_SIMD.
11155 * internal-fn.c (expand_GOMP_SIMD_LANE): New.
11156 (expand_GOMP_SIMD_VF): New.
11157 (expand_GOMP_SIMD_LAST_LANE): New.
11158 * internal-fn.def (GOMP_SIMD_LANE): New.
11159 (GOMP_SIMD_VF): New.
11160 (GOMP_SIMD_LAST_LANE): New.
11161 * omp-low.c: Include target.h.
11162 (extract_omp_for_data): Handle OMP_SIMD, OMP_CLAUSE_LINEAR,
11163 OMP_CLAUSE_SAFELEN.
11164 (check_omp_nesting_restrictions): Same.
11166 (lower_rec_simd_input_clauses): New.
11167 (lower_rec_input_clauses): Handle OMP_SIMD, GF_OMP_FOR_KIND_SIMD,
11169 (lower_lastprivate_clauses): Handle OMP_CLAUSE_LINEAR,
11170 GF_OMP_FOR_KIND_SIMD, OMP_SIMD.
11171 (expand_omp_build_assign): New.
11172 (expand_omp_for_init_counts): New.
11173 (expand_omp_for_init_vars): New.
11174 (extract_omp_for_update_vars): New.
11175 (expand_omp_for_generic): Use expand_omp_for_{init,update}_vars
11176 and rewrite accordingly.
11177 (expand_omp_simd): New.
11178 (expand_omp_for): Use expand_omp_simd.
11179 (lower_omp_for_lastprivate): Unshare vinit when appropriate.
11180 (lower_omp_for): Do not lower the body.
11181 * tree-data-ref (get_references_in_stmt): Allow IFN_GOMP_SIMD_LANE
11182 in their own loops.
11183 * tree-flow.h (find_omp_clause): Remove prototype.
11184 * tree-if-conv.c (main_tree_if_conversion): Run if doing if conversion,
11185 forcing vectorization of the loop, or if flag_tree_vectorize.
11186 (gate_tree_if_conversion): Similarly.
11187 * tree-inline.c (remap_gimple_stmt): Pass for kind argument to
11188 gimple_build_omp_for.
11189 (copy_cfg_body): set has_force_vect_loops and has_simduid_loops.
11190 * tree-parloops (create_parallel_loop): Pass kind argument to
11191 gimple_build_omp_for.
11192 * tree-pretty-print.c (dump_omp_clause): Add cases for
11193 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_LINEAR, OMP_CLAUSE_SAFELEN,
11194 OMP_CLAUSE__SIMDUID_.
11195 (dump_generic_node): Handle OMP_SIMD.
11196 * tree-ssa-ccp.c (likely_value): Handle IFN_GOMP_SIMD*.
11197 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Do not
11198 unroll OMP_SIMD loops here.
11199 * tree-ssa-loop.c (gate_tree_vectorize): Run if has_force_vect_loops.
11200 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Handle
11202 (vect_analyze_data_refs): Handle simd loops.
11203 * tree-vect-loop.c (vectorizable_live_operation): Handle
11205 * tree-vect-stmts.c (vectorizable_call): Handle IFN_GOMP_SIMD_LANE.
11206 (vectorizable_store): Handle STMT_VINFO_SIMD_LANE_ACCESS_P.
11207 (vectorizable_load): Same.
11208 * tree-vectorizer.c: Include hash-table.h and tree-ssa-propagate.h.
11209 (struct simduid_to_vf): New.
11210 (simduid_to_vf::hash): New.
11211 (simduid_to-vf::equal): New.
11212 (struct simd_array_to_simduid): New.
11213 (simd_array_to_simduid::hash): New.
11214 (simd_array_to_simduid::equal): New.
11215 (adjust_simduid_builtins): New.
11216 (struct note_simd_array_uses_struct): New.
11217 (note_simd_array_uses_cb): New.
11218 (note_simd_array_uses): New.
11219 (vectorize_loops): Handle simd hints and adjust simd builtins
11221 * tree-vectorizer.h (struct _stmt_vec_info): Add
11222 simd_lane_access_p field.
11223 (STMT_VINFO_SIMD_LANE_ACCESS_P): New macro.
11224 * tree.c (omp_clause_num_ops): Add entries for OMP_CLAUSE_LINEAR,
11225 OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_UNIFORM.
11226 (omp_clause_code_name): Same.
11227 (walk_tree_1): Handle OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN,
11228 OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_LINEAR.
11229 * tree.def (OMP_SIMD): New entry.
11230 * tree.h (enum omp_clause_code): Add entries for OMP_CLAUSE_LINEAR,
11231 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_.
11232 (OMP_CLAUSE_DECL): Adjust range for new clauses.
11233 (OMP_CLAUSE_LINEAR_NO_COPYIN): New.
11234 (OMP_CLAUSE_LINEAR_NO_COPYOUT): New.
11235 (OMP_CLAUSE_LINEAR_STEP): New.
11236 (OMP_CLAUSE_SAFELEN_EXPR): New.
11237 (OMP_CLAUSE__SIMDUID__DECL): New.
11238 (find_omp_clause): New prototype.
11240 2013-08-27 H.J. Lu <hongjiu.lu@intel.com>
11242 * config/i386/driver-i386.c (host_detect_local_cpu): Update
11243 Haswell processor detection.
11245 2013-08-27 Christian Widmer <shadow@umbrox.de>
11248 * config/i386/driver-i386.c (host_detect_local_cpu): Add detection
11249 of Ivy Bridge and Haswell processors. Assume core-avx2 for unknown
11250 AVX2 capable processors.
11252 2013-08-27 Tejas Belagod <tejas.belagod@arm.com>
11254 * config/aarch64/arm_neon.h: Replace all inline asm implementations
11255 of vget_low_* with implementations in terms of other intrinsics.
11257 2013-08-27 Marc Glisse <marc.glisse@inria.fr>
11259 PR middle-end/57219
11260 * doc/extend.texi (__builtin_isinf_sign): Restrict the return
11261 values to -1, 0 and 1.
11263 2013-08-27 Vidya Praveen <vidyapraveen@arm.com>
11265 * config/aarch64/aarch64.md (unspec): Add UNSPEC_SISD_SSHL,
11266 UNSPEC_SISD_USHL, UNSPEC_USHL_2S, UNSPEC_SSHL_2S, UNSPEC_SISD_NEG.
11267 (<optab><mode>3_insn): Remove.
11268 (aarch64_ashl_sisd_or_int_<mode>3): New Pattern.
11269 (aarch64_lshr_sisd_or_int_<mode>3): Likewise.
11270 (aarch64_ashr_sisd_or_int_<mode>3): Likewise.
11271 (define_split for aarch64_lshr_sisd_or_int_di3): Likewise.
11272 (define_split for aarch64_lshr_sisd_or_int_si3): Likewise.
11273 (define_split for aarch64_ashr_sisd_or_int_di3): Likewise.
11274 (define_split for aarch64_ashr_sisd_or_int_si3): Likewise.
11275 (aarch64_sisd_ushl, aarch64_sisd_sshl): Likewise.
11276 (aarch64_ushl_2s, aarch64_sshl_2s, aarch64_sisd_neg_qi): Likewise.
11277 (ror<mode>3_insn): Likewise.
11278 * config/aarch64/predicates.md (aarch64_simd_register): New.
11280 2013-08-27 Richard Biener <rguenther@suse.de>
11282 PR tree-optimization/57521
11283 * tree-if-conv.c (if_convertible_bb_p): Verify that at least
11284 one edge is non-critical.
11285 (find_phi_replacement_condition): Make sure to use a non-critical
11286 edge. Cleanup and remove old bug workarounds.
11287 (bb_postdominates_preds): Remove.
11288 (if_convertible_loop_p_1): Do not compute post-dominators.
11289 (combine_blocks): Do not free post-dominators.
11290 (main_tree_if_conversion): Likewise.
11291 (pass_data_if_conversion): Add TODO_verify_ssa.
11293 2013-08-27 DJ Delorie <dj@redhat.com>
11295 * config/i386/djgpp.h (ASM_DECLARE_FUNCTION_NAME): New.
11297 2013-08-27 Yufeng Zhang <yufeng.zhang@arm.com>
11299 * function.c (assign_parm_find_data_types): Set passed_mode and
11300 nominal_mode to the TYPE_MODE of nominal_type for the built
11301 pointer type in case of the struct-pass-by-reference.
11303 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
11305 * config/avr/avr-stdint.h (INT16_TYPE): Change default to "int".
11306 (UINT16_TYPE): Change default to "unsigned int".
11308 * config/avr/avr.opt (mfract-convert-truncate): New option.
11309 * config/avr/avr.c (avr_out_fract): Unless TARGET_FRACT_CONV_TRUNC
11310 is set, round negative fractional integers according to n1169
11311 when converting to integer types.
11313 2013-08-26 Jan Hubicka <jh@suse.cz>
11315 * cgraph.c (cgraph_propagate_frequency): Do not assume that virtual
11316 methods can not be called indirectly when their address is not taken.
11318 2013-08-26 Jan Hubicka <jh@suse.cz>
11320 * gimple-fold.c (gimple_get_virt_method_for_binfo): Use
11323 2013-08-26 Jan Hubicka <jh@suse.cz>
11325 * ipa.c (comdat_can_be_unshared_p_1): C++ constructors and destructors
11328 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
11330 * reload.c (find_valid_class): Allow classes that do not include
11331 FIRST_PSEUDO_REGISTER - 1.
11333 2013-08-26 Jan Hubicka <jh@suse.cz>
11335 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix formatting;
11336 fix edge count/frequency when speculation failed; fix type check
11337 for the direct call.
11339 2013-08-26 Jan Hubicka <jh@suse.cz>
11341 * ipa-prop.c (ipa_print_node_params): Do not ICE during WPA.
11343 2013-08-26 Jan Hubicka <jh@suse.cz>
11345 * ipa-inline-transform.c (inline_transform): Be ready for basic block
11346 to be changed by edge redirection.
11348 2013-08-26 Jan Hubicka <jh@suse.cz>
11350 * cgraph.c (cgraph_speculative_call_info): Fix parameter order and
11351 formating; add sanity check.
11352 (cgraph_resolve_speculation): Add FIXME about scaling profiles.
11353 (cgraph_redirect_edge_call_stmt_to_callee): Fix ICE in debug dump.
11354 * ipa-inline.c (heap_edge_removal_hook): Reset node growth cache.
11355 (resolve_noninline_speculation): Update callee keys, too.
11357 2013-08-26 Jan Hubicka <jh@suse.cz>
11359 * tree.h (tree_decl_with_vis): Add cxx_constructor, cxx_destructor.
11360 (DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P): New macros.
11362 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
11364 * config/i386/i386.c (x86_64_elf_select_section): Put ATTRIBUTE_UNUSED
11367 2013-08-26 Uros Bizjak <ubizjak@gmail.com>
11369 * config/i386/i386.c (ix86_debug_options): Remove prototype.
11370 (x86_64_elf_select_section): Ditto.
11371 (ix86_handle_tm_regparm_attribute): Remove ATTRIBUTE_UNUSED on used
11373 (ix86_pass_by_reference): Ditto.
11374 (output_set_got): Ditto.
11375 (ix86_unary_operator_ok): Ditto.
11376 (ix86_expand_builtin): Ditto.
11378 2013-08-23 Jan Hubicka <jh@suse.cz>
11380 * cgraph.c (cgraph_turn_edge_to_speculative): Fix debug output.
11382 2013-08-23 Jan Hubicka <jh@suse.cz>
11384 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
11385 (tree_decl_with_vis): Add FINAL field.
11387 2013-08-23 Jeff Law <law@redhat.com>
11389 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
11390 the debugging dump when the expression is fully redundant.
11392 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
11394 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
11395 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
11396 * pretty-print.c (pp_formatted_text_data): Likewise.
11397 (pp_write_text_to_stream): Likewise.
11398 (pp_write_text_as_dot_label_to_stream): Likewise.
11399 (pp_append_r): Likewise.
11400 (pp_format): Likewise.
11401 (pp_flush): Likewise.
11402 (pp_clear_output_area): Likewise.
11403 (pp_append_text): Likewise.
11404 (pp_formatted_text): Likewise.
11405 (pp_remaining_character_count_for_line): Likewise.
11406 (pp_newline): Likewise.
11407 (pp_character): Likewise.
11408 (output_buffer::~output_buffer): Define.
11409 (pretty_printer::~pretty_printer): Destruct output buffer.
11410 * pretty-print.h (output_buffer::~output_buffer): Declare.
11411 (pretty_printer::~pretty_printer): Declare virtual.
11413 2013-08-24 Marc Glisse <marc.glisse@inria.fr>
11416 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
11417 HOST_WIDE_INT_M1U): New macros.
11418 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
11419 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
11420 unsigned -1 for lshift.
11421 * cse.c (cse_insn): Likewise.
11422 * double-int.c (rshift_double, lshift_double): Likewise.
11423 * builtins.c (fold_builtin_bitop): Likewise.
11424 * combine.c (force_to_mode): Likewise.
11425 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
11426 * simplify-rtx.c (simplify_const_unary_operation,
11427 simplify_const_binary_operation): Likewise.
11428 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
11429 check_va_list_escapes): Likewise.
11430 * rtlanal.c (nonzero_bits1): Likewise.
11431 * expmed.c (expand_smod_pow2): Likewise.
11432 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
11434 2013-08-23 Jan Hubicka <jh@suse.cz>
11436 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
11437 as having address taken.
11439 2013-08-23 Jan Hubicka <jh@suse.cz>
11441 * ipa-utils.h (method_class_type): Declare.
11442 * ipa-devirt.c (method_class_type): Export.
11444 * cgraphunit.c (analyze_functions): Do basic devirtualization;
11445 do not walk base classes of anonymous types.
11447 2013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
11449 PR rtl-optimization/58220
11450 PR regression/58221
11451 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
11452 handle SEQUENCE insns properly.
11454 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
11456 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
11458 (pp_newline_and_indent): Likewise.
11459 (pp_separate_with): Likewise.
11460 * pretty-print.c (pp_newline_and_flush): Define.
11461 (pp_newline_and_indent): Likewise.
11462 (pp_separate_with): Likewise.
11464 2013-08-23 Jakub Jelinek <jakub@redhat.com>
11467 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
11468 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
11470 2013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
11472 * config/i386/predicates.md (ext_sse_reg_operand): New.
11473 * config/i386/i386.md (*movti_internal): Use
11474 predicate to determine if EVEX is needed.
11475 (*movsi_internal): Ditto.
11476 (*movdf_internal): Ditto.
11477 (*movsf_internal): Ditto.
11478 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
11480 2013-08-23 Jakub Jelinek <jakub@redhat.com>
11482 PR tree-optimization/58209
11483 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
11484 (find_tail_calls): Give up for pointer result types if m is non-NULL.
11485 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
11486 emit POINTER_PLUS_EXPR.
11487 (create_tailcall_accumulator): For pointer result type accumulate in
11490 2013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
11492 * configure.ac: Add backslashes missing from the last change.
11493 * configure: Regenerate.
11495 2013-08-22 Jan Hubicka <jh@susue.cz>
11497 * ipa.c (function_and_variable_visibility): First remember function
11498 was global and then make it local.
11500 2013-08-22 Julian Brown <julian@codesourcery.com>
11502 * configure.ac: Add aarch64 to list of arches which use "nop" in
11504 * configure: Regenerate.
11506 2013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11508 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
11509 gnu_libc_has_function.
11510 * config/s390/tpf.h: Likewise.
11512 2013-08-22 Jan Hubicka <jh@susue.cz>
11514 * timevar.c (validate_phases): Add cast.
11516 2013-08-22 Jan Hubicka <jh@susue.cz>
11518 * timevar.c (validate_phases): Use size_t for memory.
11519 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
11521 2013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
11523 * pretty-print.h (output_buffer::output_buffer): Declare.
11524 (pretty_printer::pretty_printer): Likewise.
11525 (pp_construct): Remove.
11526 * pretty-print.c (output_buffer::output_buffer): Define.
11527 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
11528 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
11529 (print_gimple_expr): Likewise.
11530 (print_gimple_seq): Likewise.
11531 (gimple_dump_bb): Likewise.
11532 * sched-vis.c (dump_value_slim): Likewise.
11533 (dump_insn_slim): Likewise.
11534 (dump_rtl_slim): Likewise.
11535 (str_pattern_slim): Likewise.
11536 * tree-mudflap.c (mf_varname_tree): Likewise.
11537 * graph.c (print_graph_cfg): Likewise.
11538 (start_graph_dump): Likewise.
11539 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
11541 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
11542 pretty printer initialization.
11543 * coretypes.h (diagnostic_context): Remove superflous type alias
11545 (pretty_printer): Likewise. Declare directly as a class.
11546 (pretty_print_info): Remove declaration as class.
11547 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
11548 and pp_clear_output_area.
11549 (asan_add_global): Likewise.
11551 2013-08-22 Jan Hubicka <jh@suse.cz>
11553 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
11554 * ipa-utils.h (update_type_inheritance_graph): Declare.
11555 (possible_polymorphic_call_target_p): Declare.
11556 (possible_polymorphic_call_target_p): New.
11557 * ipa-devirt.c: Update toplevel comments.
11558 (cached_polymorphic_call_targets): Move up.
11559 (odr_type_d): Move ID down.
11560 (polymorphic_type_binfo_p): Update comment.
11561 (odr_hasher::remove): Likewise;
11562 (get_odr_type): Set anonymous_namespace.
11563 (dump_odr_type): Dump it.
11564 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
11565 (maybe_record_node): Record node in cached_polymorphic_call_targets.
11566 (record_binfo): Add comment.
11567 (free_polymorphic_call_targets_hash): Do not ICE when cache is not
11569 (devirt_node_removal_hook): Do not iCE when cache is freed.
11570 (possible_polymorphic_call_target_p): New predicate.
11571 (update_type_inheritance_graph): New function.
11573 2013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
11574 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
11575 Sergey Lega <sergey.s.lega@intel.com>
11576 Anna Tikhonova <anna.tikhonova@intel.com>
11577 Ilya Tocar <ilya.tocar@intel.com>
11578 Andrey Turetskiy <andrey.turetskiy@intel.com>
11579 Ilya Verbin <ilya.verbin@intel.com>
11580 Kirill Yukhin <kirill.yukhin@intel.com>
11581 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
11583 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
11584 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
11585 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
11586 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
11587 (OPTION_MASK_ISA_AVX2_UNSET): Update.
11588 (OPTION_MASK_ISA_AVX512F_UNSET): New.
11589 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
11590 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
11591 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
11592 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
11593 OPT_mavx512pf, OPT_mavx512er cases.
11594 * config/i386/constraints.md (v): New constraint.
11595 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
11596 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
11597 (bit_AVX512CD): New.
11598 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
11599 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
11600 * config/i386/i386-c.c (ix86_target_macros_internal):
11601 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
11603 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
11604 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
11605 * config/i386/i386.c (regclass_map, dbx_register_map)
11606 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
11607 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
11608 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
11609 -mavx512pf options.
11610 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
11611 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
11612 -mavx512pf options. Fix formatting.
11613 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
11614 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
11615 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
11616 -mavx512cd, -mavx512pf options.
11617 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
11618 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
11619 (ix86_preferred_output_reload_class): Replace SSE_REGS with
11621 (ix86_hard_regno_mode_ok): Support 512-bit registers.
11622 (ix86_set_reg_reg_cost): Ditto.
11623 (x86_order_regs_for_local_alloc): Ditto.
11624 (MAX_VECT_LEN): Extend to 64-byte.
11625 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
11626 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
11627 (TARGET_AVX512ER, TARGET_AVX512CD): New.
11628 (BIGGEST_ALIGNMENT): Extend to 512-bits.
11629 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
11630 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
11631 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
11632 (SSE_REG_MODE_P): Support new modes.
11633 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
11634 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
11635 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
11636 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
11637 (REG_CLASS_CONTENTS): Add new registers.
11638 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
11639 (EXT_REX_SSE_REGNO_P): New.
11640 (HI_REGISTER_NAMES): Add new registers.
11641 * config/i386/i386.md: Define constants for new registers.
11642 (mode): Add new 512-bit modes.
11643 (prefix): Support evex prefix.
11644 (isa): Support avx512f, noavx512f, fma_avx512f.
11645 (ssemodesuffix): Add new 512-bit modes.
11647 (*movxi_internal_avx512f): Ditto.
11648 (*movdi_internal): Replace constraint "x" with the new constraint "v".
11650 (*movsi_internal): Likewise.
11651 (*movdf_internal): Likewise.
11652 (*movsf_internal): Likewise.
11653 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
11654 (<code><mode>3): Likewise.
11655 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd):
11657 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
11658 with the new constraint "v".
11659 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
11661 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
11662 with the new constraint "v".
11663 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
11664 (<sse2>_storedqu<avxsizesuffix>): Likewise.
11665 (*<plusminus_insn><mode>3): Likewise.
11666 (<sse>_vm<plusminus_insn><mode>3): Likewise.
11667 (*mul<mode>3): Likewise.
11668 (<sse>_vmmul<mode>3): Likewise.
11669 (<sse>_div<mode>3): Likewise.
11670 (<sse>_vmdiv<mode>3): Likewise.
11671 (<sse>_sqrt<mode>2): Likewise.
11672 (<sse>_vmsqrt<mode>2): Likewise.
11673 (*<code><mode>3_finite): Likewise.
11674 (*<code><mode>3) <smaxmin>: Likewise.
11675 (<sse>_vm<code><mode>3): Likewise.
11676 (*<code><mode>3) <any_logic>: Likewise.
11677 (*fma_fmadd_<mode>): Likewise.
11678 (*fma_fmsub_<mode>): Likewise.
11679 (*fma_fnmadd_<mode>): Likewise.
11680 (*fma_fnmsub_<mode>): Likewise.
11681 (*fma_fmaddsub_<mode>): Likewise.
11682 (*fma_fmsubadd_<mode>): Likewise.
11683 (*fmai_fmadd_<mode>): Likewise.
11684 (*fmai_fmsub_<mode>): Likewise.
11685 (*fmai_fnmadd_<mode>): Likewise.
11686 (*fmai_fnmsub_<mode>): Likewise.
11687 (sse_cvtsi2ss): Likewise.
11688 (sse_cvtsi2ssq): Likewise.
11689 (sse_cvtss2si): Likewise.
11690 (sse_cvtss2si_2): Likewise.
11691 (sse_cvtss2siq): Likewise.
11692 (sse_cvtss2siq_2): Likewise.
11693 (sse_cvttss2si): Likewise.
11694 (sse_cvtss2siq_2): Likewise.
11695 (float<sseintvecmodelower><mode>2): Likewise.
11696 (sse2_cvtsd2si_2): Likewise.
11697 (sse2_cvtsd2siq_2): Likewise.
11698 (*<plusminus_insn><mode>3): Likewise.
11699 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
11700 (*<sse4_1_avx2>_mul<mode>3): Likewise.
11701 (ashr<mode>3): Likewise.
11702 (<shift_insn><mode>3): Likewise.
11703 (avx2_<code><mode>3): Likewise.
11704 (*avx2_<code><mode>3): Likewise.
11705 (*andnot<mode>3): Likewise.
11706 (*<code><mode>3) <any_logic>: Likewise.
11707 (abs<mode>2): Likewise.
11708 (avx2_permvar<mode>): Likewise.
11709 (avx2_perm<mode>_1): Likewise.
11710 (*avx_vpermilp<mode>): Likewise.
11711 (avx_vpermilvar<mode>3): Likewise.
11712 (avx2_ashrv<mode>): Likewise.
11713 (avx2_<shift_insn>v<mode>): Likewise.
11714 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
11716 * doc/rtl.texi: Document XImode.
11718 2013-08-21 Jeff Law <law@redhat.com>
11720 * tree-flow.h (register_jump_thread): Pass vector of edges
11721 instead of each important edge.
11722 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
11723 thread path into a vector and pass that to register_jump_thread.
11724 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
11725 passed in edge vector to the current 3-edge form.
11728 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
11730 * dce.c (fini_dce): Call df_analyze again just in case
11731 delete_unmarked_insns removed anything.
11733 2013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
11735 * reload.h (struct reg_equivs): Rename to ..
11736 (struct reg_equivs_s): .. this.
11738 2013-08-20 Martin Liska <marxin.liska@gmail.com>
11740 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
11742 2013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
11744 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest comment.
11746 2013-08-21 Jeff Law <law@redhat.com>
11748 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
11749 simplify assignments too. If the RHS collapses to a singleton
11750 range, then return the value for the range.
11752 2013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
11754 * config/i386/sse.md (V16): Rename to...
11756 (mov<mode>): Update iterator name.
11757 (*mov<mode>_internal): Ditto.
11758 (push<mode>1): Ditto.
11759 (movmisalign<mode>): Ditto.
11761 2013-08-20 Jan Hubicka <jh@suse.cz>
11764 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
11765 entry for direct edges.
11766 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
11768 2013-08-20 David Malcolm <dmalcolm@redhat.com>
11770 Revert my last two changes, r201865 and r201864:
11773 2013-08-20 David Malcolm <dmalcolm@redhat.com>
11775 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
11776 instances can own GC refs.
11778 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
11779 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
11780 (gcc::context::gt_pch_nx): Likewise.
11781 (gcc::context::gt_pch_nx): Likewise.
11782 * ggc.h (gt_ggc_mx <T>): New.
11783 (gt_pch_nx_with_op <T>): New.
11784 (gt_pch_nx <T>): New.
11785 * passes.c (opt_pass::gt_ggc_mx): New.
11786 (opt_pass::gt_pch_nx): New.
11787 (opt_pass::gt_pch_nx_with_op): New.
11788 (pass_manager::gt_ggc_mx): New.
11789 (pass_manager::gt_pch_nx): New.
11790 (pass_manager::gt_pch_nx_with_op): New.
11791 (pass_manager::operator new): Use
11792 ggc_internal_cleared_alloc_stat rather than xcalloc.
11793 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
11794 (pass_manager::gt_ggc_mx): New.
11795 (pass_manager::gt_pch_nx): New.
11796 (pass_manager::gt_pch_nx_with_op): New.
11797 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
11798 (opt_pass::operator new): New.
11799 (opt_pass::gt_ggc_mx): New.
11800 (opt_pass::gt_pch_nx): New.
11801 (opt_pass::gt_pch_nx_with_op): New.
11804 2013-08-20 David Malcolm <dmalcolm@redhat.com>
11806 * Makefile.in (GTFILES): Add context.h.
11807 * context.c (gcc::context::operator new): New.
11808 (gcc::context::gt_ggc_mx): New.
11809 (gcc::context::gt_pch_nx): New.
11810 (gcc::context::gt_pch_nx): New.
11811 * context.h (gcc::context): Add GTY((user)) marking.
11812 (gcc::context::operator new): New.
11813 (gcc::context::gt_ggc_mx): New.
11814 (gcc::context::gt_pch_nx): New.
11815 (gcc::context::gt_pch_nx): New.
11816 (g): Add GTY marking.
11817 (gt_ggc_mx (gcc::context *)): New.
11818 (gt_pch_nx (gcc::context *)): New.
11819 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
11820 void *cookie)): New.
11821 * gengtype.c (open_base_files) <ifiles>: Add context.h.
11823 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
11825 * dce.c (fini_dce): Call df_analyze again just in case
11826 delete_unmarked_insns removed anything.
11828 2013-08-20 Teresa Johnson <tejohnson@google.com>
11830 PR rtl-optimizations/57451
11831 * final.c (reemit_insn_block_notes): Prevent lexical blocks
11832 from crossing split section boundaries.
11834 2013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
11836 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
11837 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
11838 with MULTLIB_DEFAULTS.
11840 2013-08-20 Nick Clifton <nickc@redhat.com>
11842 * target.def (narrow_volatile_bitfield): Note that the default
11843 value is false, not !TARGET_STRICT_ALIGN.
11844 * doc/tm.texi: Regenerate.
11846 2013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
11848 Fix LIB_SPEC for systems without libpthread.
11850 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
11851 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
11853 * config/i386/linux-common.h: Likewise.
11854 * config/mips/linux-common.h: Likewise.
11856 2013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
11858 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
11861 2013-08-20 David Malcolm <dmalcolm@redhat.com>
11863 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
11864 instances can own GC refs.
11866 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
11867 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
11868 (gcc::context::gt_pch_nx): Likewise.
11869 (gcc::context::gt_pch_nx): Likewise.
11870 * ggc.h (gt_ggc_mx <T>): New.
11871 (gt_pch_nx_with_op <T>): New.
11872 (gt_pch_nx <T>): New.
11873 * passes.c (opt_pass::gt_ggc_mx): New.
11874 (opt_pass::gt_pch_nx): New.
11875 (opt_pass::gt_pch_nx_with_op): New.
11876 (pass_manager::gt_ggc_mx): New.
11877 (pass_manager::gt_pch_nx): New.
11878 (pass_manager::gt_pch_nx_with_op): New.
11879 (pass_manager::operator new): Use
11880 ggc_internal_cleared_alloc_stat rather than xcalloc.
11881 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
11882 (pass_manager::gt_ggc_mx): New.
11883 (pass_manager::gt_pch_nx): New.
11884 (pass_manager::gt_pch_nx_with_op): New.
11885 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
11886 (opt_pass::operator new): New.
11887 (opt_pass::gt_ggc_mx): New.
11888 (opt_pass::gt_pch_nx): New.
11889 (opt_pass::gt_pch_nx_with_op): New.
11891 2013-08-20 David Malcolm <dmalcolm@redhat.com>
11893 * Makefile.in (GTFILES): Add context.h.
11894 * context.c (gcc::context::operator new): New.
11895 (gcc::context::gt_ggc_mx): New.
11896 (gcc::context::gt_pch_nx): New.
11897 (gcc::context::gt_pch_nx): New.
11898 * context.h (gcc::context): Add GTY((user)) marking.
11899 (gcc::context::operator new): New.
11900 (gcc::context::gt_ggc_mx): New.
11901 (gcc::context::gt_pch_nx): New.
11902 (gcc::context::gt_pch_nx): New.
11903 (g): Add GTY marking.
11904 (gt_ggc_mx (gcc::context *)): New.
11905 (gt_pch_nx (gcc::context *)): New.
11906 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
11907 void *cookie)): New.
11908 * gengtype.c (open_base_files) <ifiles>: Add context.h.
11910 2013-08-20 Alan Modra <amodra@gmail.com>
11913 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
11914 (rs6000_emit_epilogue): Likewise.
11916 2013-08-19 Dehao Chen <dehao@google.com>
11918 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
11920 2013-08-19 Peter Bergner <bergner@vnet.ibm.com>
11921 Jakub Jelinek <jakub@redhat.com>
11923 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
11924 (BUILT_IN_FABSD64): Likewise.
11925 (BUILT_IN_FABSD128): Likewise.
11926 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
11927 (fold_builtin_1): Likewise.
11928 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping
11929 destination and source operands.
11930 (*abstd2_fpr): Likewise.
11931 (*nabstd2_fpr): Likewise.
11933 2013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
11935 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
11938 2013-08-19 Aldy Hernandez <aldyh@redhat.com>
11940 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
11943 2013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
11945 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
11946 * builtins.c (default_libc_has_function): New.
11947 (gnu_libc_has_function): Ditto.
11948 (no_c99_libc_has_function): Ditto.
11949 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
11950 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
11951 (fold_builtin_sincos): Likewise.
11952 (fold_builtin_cexp): Likewise.
11953 * builtins.def (DEF_C94_BUILTIN): Likewise.
11954 (DEF_C99_BUILTIN): Likewise.
11955 (DEF_C99_C90RES_BUILTIN): Likewise.
11956 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
11957 definitions to using this define.
11958 * config/darwin-protos.h (darwin_libc_has_function): New.
11959 * config/darwin.c (darwin_libc_has_function): Ditto.
11960 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
11961 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
11962 * config/darwin.h: Ditto.
11963 * config/elfos.h: Ditto.
11964 * config/freebsd.h: Ditto.
11965 * config/i386/cygming.h: Ditto.
11966 * config/i386/djgpp.h: Ditto.
11967 * config/i386/i386-interix.h: Ditto.
11968 * config/microblaze/microblaze.h: Ditto.
11969 * config/mmix/mmix.h: Ditto.
11970 * config/gnu-user.h: Ditto.
11971 * config/ia64/hpux.h: Ditto.
11972 * config/pa/pa-hpux.h: Ditto.
11973 * config/pdp11/pdp11.h: Ditto.
11974 * config/picochip/picochip.h: Ditto.
11975 * config/linux.h: Ditto.
11976 * config/netbsd.h: Ditto.
11977 * config/openbsd.h: Ditto.
11978 * config/rs6000/aix43.h: Ditto.
11979 * config/rs6000/aix51.h: Ditto.
11980 * config/rs6000/aix52.h: Ditto.
11981 * config/rs6000/aix53.h: Ditto.
11982 * config/rs6000/aix61.h: Ditto.
11983 * config/rs6000/darwin.h: Ditto.
11984 * config/rs6000/linux.h: Ditto.
11985 * config/rs6000/linux64.h: Ditto.
11986 * config/s390/tpf.h: Ditto.
11987 * config/sol2-10.h: Ditto.
11988 * config/sol2.h: Ditto.
11989 * config/vms/vms.h: Ditto.
11990 * config/vxworks.h: Ditto.
11991 * config/linux-android.c (linux_android_libc_has_function):
11992 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
11993 * config/linux-protos.h (linux_android_libc_has_function):
11995 * config/i386/i386.c (ix86_libc_has_function): New.
11996 * config/i386/i386-protos.h
11997 (ix86_libc_has_function): New declaration.
11998 * config/i386/i386.md
11999 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
12000 ("isinf<mode>2): Likewise.
12001 * convert.c (convert_to_integer): Using new target hook
12002 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
12003 TARGET_C99_FUNCTIONS.
12004 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
12005 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
12006 * coretypes.h (function_class): New enum for different
12007 classes of functions.
12008 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
12009 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
12010 (TARGET_HAS_SINCOS): Likewise.
12011 (TARGET_LIBC_HAS_FUNCTION): New.
12012 * doc/tm.texi: Regenerated.
12013 * targhooks.h (default_libc_has_function): New declaration.
12014 (no_c99_libc_has_function): Ditto.
12015 (gnu_libc_has_function): Ditto.
12016 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
12017 and TARGET_HAS_SINCOS.
12019 2013-08-18 Jan Hubicka <jh@suse.cz>
12021 * Makeifle-in (ipa-devirt.o): New.
12022 (GTFILES): Add ipa-utils.h and ipa-devirt.c
12023 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
12024 (analyze_functions): Look into possible targets of polymorphic call.
12025 * dumpfile.c (dump_files): Add type-inheritance dump.
12026 * dumpfile.h (TDI_inheritance): New.
12027 * ipa-devirt.c: New file.
12028 * ipa-utils.h (odr_type_d): Forward declare.
12029 (odr_type): New type.
12030 (build_type_inheritance_graph): Declare.
12031 (possible_polymorphic_call_targets): Declare and introduce inline
12032 variant when only edge is pased.
12033 (dump_possible_polymorphic_call_targets): Likewise.
12034 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
12035 * tree.c (type_in_anonymous_namespace_p): Break out from ...
12036 (types_same_for_odr): ... here.
12037 * tree.h (type_in_anonymous_namespace_p): Declare.
12039 2013-08-18 Jakub Jelinek <jakub@redhat.com>
12041 PR tree-optimization/58006
12042 * tree-parloops.c (take_address_of): Don't ICE if get_name
12044 (eliminate_local_variables_stmt): Remove clobber stmts.
12046 2013-08-18 Eric Botcazou <ebotcazou@adacore.com>
12048 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
12049 error message is issued for an alias to undefined symbol.
12051 2013-08-18 Jan Hubicka <jh@suse.cz>
12053 * cgraph.c (cgraph_create_indirect_edge): Discover
12054 polymorphic calls and record basic info into indirect_info.
12055 * gimple-fold.c (gimple_fold_call): When doing BINFO based
12056 devirtualization, ignore objc function calls.
12057 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
12058 call with no parm index info.
12059 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
12060 * tree.c (virtual_method_call_p): New function.
12061 * tree.h (virtual_method_call_p): Declare.
12063 2013-08-16 Jan Hubicka <jh@suse.cz>
12065 PR middle-end/58179
12066 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
12068 2013-08-16 David Edelsohn <dje.gcc@gmail.com>
12070 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
12073 2013-08-16 David Malcolm <dmalcolm@redhat.com>
12075 * gengtype.c (type_for_name): Add special-case support for
12076 locating types within the "gcc::" namespace.
12077 (open_base_files): Emit a "using namespace gcc" directive.
12079 2013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
12082 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
12083 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
12085 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
12086 array instead of each individual operand as a separate argument.
12087 (emit_fusion_gpr_load): Likewise.
12088 (expand_fusion_gpr_load): Add new function declaration.
12090 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
12091 signature to have the operands passed as an array, instead of as
12092 separate arguments. Allow ZERO_EXTEND to be in the memory
12093 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
12094 depend on the register live/dead flags when peepholes are run.
12095 (expand_fusion_gpr_load): New function to be called from the
12096 peephole2 pass, to change the register that addis sets to be the
12098 (emit_fusion_gpr_load): Change the calling signature to have the
12099 operands passed as an array, instead of as separate arguments.
12100 Allow ZERO_EXTEND to be in the memory address, and also
12101 SIGN_EXTEND if -mpower8-fusion-sign.
12103 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
12104 unspec enumeration.
12105 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
12106 adjust the register addis loads up in the peephole2 pass. Do not
12107 depend on the register live/dead state when the peephole pass is done.
12109 2013-08-16 David Malcolm <dmalcolm@redhat.com>
12111 * gengtype.c (create_user_defined_type): Ensure that the kind
12112 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
12113 declaration is seen before the GTY((user)) marking.
12115 2013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
12118 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
12120 2013-08-16 Jan Hubicka <jh@suse.cz>
12122 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
12123 arugment expected_type.
12124 (gimple_fold_call): Use it.
12125 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
12126 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
12127 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
12128 (try_make_edge_direct_virtual_call): Likewise.
12129 * tree.c (obj_type_ref_class): New.
12130 * tree.h (obj_type_ref_class): Use it.
12132 2013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
12134 * sched-vis.c (rtl_slim_pp_initialized): Remove.
12135 (rtl_slim_pp): Likewise.
12136 (init_rtl_slim_pretty_print): Likewise.
12137 (dump_value_slim): Don't call it. Use local pretty printer.
12138 (dump_insn_slim): Likewise.
12139 (dump_rtl_slim): Likewise.
12140 (str_pattern_slim): Likewise.
12141 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
12144 2013-08-16 Jakub Jelinek <jakub@redhat.com>
12146 PR tree-optimization/58164
12147 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
12148 walk gimple_goto_dest of GIMPLE_GOTO.
12150 PR tree-optimization/58165
12151 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
12152 bi_call must be the last stmt in a bb, don't split_block, instead
12153 use fallthru edge from it and give up if there is none.
12154 Release conds vector when returning early.
12156 2013-08-14 Xinliang David Li <davidxl@google.com>
12158 * config/i386/i386.c (ix86_option_override_internal):
12159 Remove unused variable and field.
12161 2013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12164 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
12165 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
12166 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
12167 and Linux, correct BLKmode alignment when 128-bit alignment is
12168 required and compatibility flag is not set.
12169 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
12170 for zero-size arguments when compatibility flag is not set.
12172 2013-08-14 Jakub Jelinek <jakub@redhat.com>
12174 PR tree-optimization/58145
12175 * tree-sra.c (build_ref_for_offset): If prev_base has
12176 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
12178 2013-08-14 Xinliang David Li <davidxl@google.com>
12180 * config/i386/i386.c (ix86_option_override_internal):
12181 Fix uninitialized variable error.
12183 2013-08-14 Xinliang David Li <davidxl@google.com>
12185 * config/i386/i386.opt: Define two new options.
12186 * config/i386/x86-tune.def: Add arch selector field in macros.
12187 * config/i386/i386.h: Adjust macro definition.
12188 * config/i386/i386.c (ix86_option_override_internal):
12190 (parse_mtune_ctrl_str): New function.
12191 (set_ix86_tune_features): New function.
12192 (ix86_function_specific_restore): Call the new helper function.
12194 2013-08-14 Andrey Belevantsev <abel@ispras.ru>
12196 PR rtl-optimization/57662
12197 * sel-sched.c (code_motion_process_successors): When the current insn
12198 is removed after the recursive traversal, break from the loop.
12199 Add comments and debug printouts.
12201 2013-08-14 Jakub Jelinek <jakub@redhat.com>
12202 Alexandre Oliva <aoliva@redhat.com>
12205 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
12206 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
12207 there also UNSPEC_PLTOFF.
12209 2013-08-14 Marek Polacek <polacek@redhat.com>
12211 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
12212 NUM_CONDITIONS bit positions.
12214 2013-08-13 Cary Coutant <ccoutant@google.com>
12216 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
12217 (attr_checksum): Hash vector contents instead of pointer.
12218 (attr_checksum_ordered): Likewise.
12220 2013-08-13 Uros Bizjak <ubizjak@gmail.com>
12222 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
12223 when Pmode != word_mode. Add length_address attribute.
12224 (sse3_monitor_<mode>): Merge from sse3_monitor and
12225 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
12226 Pmode != word_mode. Update insn length attribute.
12227 * config/i386/i386.c (ix86_option_override_internal): Update
12228 ix86_gen_monitor selection for merged sse3_monitor insn.
12230 2013-08-13 Julian Brown <julian@codesourcery.com>
12232 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
12233 perform invalid legitimization on greater-than-word-size modes for
12234 TARGET_E500_DOUBLE.
12236 2013-08-13 Vladimir Makarov <vmakarov@redhat.com>
12238 * ira.c (setup_class_translate_array): Use aclass instead of cl
12239 for classes not fully covered by allocno classes.
12241 2013-08-13 Jakub Jelinek <jakub@redhat.com>
12243 PR tree-optimization/57661
12244 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
12245 * tree-inline.c (tree_function_versioning): Initialize it.
12246 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
12247 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
12248 that is not being copied.
12251 * asan.c (instrument_strlen_call): Fix typo in comment.
12252 Use char * type even for the lhs of POINTER_PLUS_EXPR.
12254 2013-08-13 Steve Ellcey <sellcey@mips.com>
12256 * config/mips/mips.md (prefetch): Use lw instead of ld on
12257 loongson in 32bit mode.
12259 2013-08-13 Nick Clifton <nickc@redhat.com>
12261 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
12263 2013-08-13 Jan Hubicka <jh@suse.cz>
12265 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
12266 introduced edge; fix typo in sanity check.
12267 (cgraph_resolve_speculation): Export; improve diagnostic.
12268 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
12269 speculation at type mismatch.
12270 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
12271 (cgraph_resolve_speculation): Declare.
12272 (symtab_can_be_discarded): New function.
12273 * value-prof.c (gimple_ic_transform): Remove actual transform code.
12274 * ipa-inline-transform.c (speculation_removed): New global var.
12275 (clone_inlined_nodes): See if speculation can be removed.
12276 (inline_call): If speculations was removed, we growths may not match.
12277 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
12278 (speculation_useful_p): New function.
12279 (resolve_noninline_speculation): New function.
12280 (inline_small_functions): Resolve useless speculations.
12281 * ipa-inline.h (speculation_useful_p): Declare
12282 * ipa.c (can_replace_by_local_alias): Simplify.
12283 (ipa_profile): Produce speculative calls in non-lto, too;
12284 add simple cost model; produce local aliases.
12286 2013-08-13 David Malcolm <dmalcolm@redhat.com>
12288 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
12291 2013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
12293 * config/i386/i386.c (ix86_function_versions): Use error + inform.
12295 2013-08-12 Uros Bizjak <ubizjak@gmail.com>
12297 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
12298 iterator instead of X87MODEF.
12300 2013-08-12 Perez Read <netfirewall@gmail.com>
12303 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
12304 operand 0 for intel asm alternative.
12305 (*movabs<mode>_2): Ditto for operand 1.
12307 2013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
12309 * config/aarch64/arm_none.h
12310 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
12312 2013-08-12 Nick Clifton <nickc@redhat.com>
12314 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
12316 2013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
12318 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
12319 expand for QI/HImode operand to produce more effictive code for
12320 unsigned char(short) --> float(double) conversion.
12322 2013-08-12 Alexander Monakov <amonakov@ispras.ru>
12324 * doc/invoke.texi: Mention that -ftls-model does not force the final
12327 2013-08-12 Marek Polacek <polacek@redhat.com>
12328 Marc Glisse <marc.glisse@inria.fr>
12330 PR tree-optimization/57980
12331 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
12332 when creating -1 constant.
12334 2013-08-10 Jan Hubicka <jh@suse.cz>
12336 Workaround binutils PR14342.
12337 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
12338 (gimple_init_edge_profiler): Likewise.
12339 (gimple_gen_ic_func_profiler): Likewise.
12341 2013-08-09 Jan Hubicka <jh@suse.cz>
12343 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
12345 2013-08-09 Xinliang David Li <davidxl@google.com>
12347 * config/i386/stringop.def: New file.
12348 * config/i386/stringop.opt: New file.
12349 * config/i386/i386-opts.h: Include stringopt.def.
12350 * config/i386/i386.opt: Include stringopt.opt.
12351 * config/i386/i386.c (ix86_option_override_internal):
12352 Override default size based stringop inline strategies with options.
12353 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
12356 2013-08-09 Jan Hubicka <jh@suse.cz>
12358 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
12360 2013-08-09 Jan Hubicka <jh@suse.cz>
12362 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
12364 (dump_cgraph_node): Dump profile-id.
12365 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
12366 and common_target_probability.
12367 * lto-cgraph.c (lto_output_edge): Stream common targets.
12368 (lto_output_node): Stream profile ids.
12369 (input_node): Stream profile ids.
12370 (input_edge): Stream common targets.
12371 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
12372 * ipa.c: Include value-prof.h
12373 (ipa_profile_generate_summary): Turn indirect call statement histograms
12374 into common targets.
12375 (ipa_profile): Turn common targets into speculative edges.
12377 2013-08-09 Jan Hubicka <jh@suse.cz>
12379 * cgraph.h (cgraph_node): Add profile_id.
12380 * value-prof.c (cgraph_node_map): Turn into pointer_map.
12381 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
12382 (del_node_map): Update.
12383 (find_func_by_funcdef_no): Replace by ...
12384 (find_func_by_profile_id): ... this one.
12385 (gimple_ic_transform): Do not remove useful histograms when
12386 speculation is not done; dump info when indirect call removal
12388 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
12389 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
12390 (__gcov_indirect_call_profiler_v2): .. this one.
12391 * profile.h (init_node_map): Update.
12392 * coverage.c (coverage_compute_profile_id): New function.
12393 * coverage.h (coverage_compute_profile_id): Declare.
12394 * tree-profile.c (init_ic_make_global_vars): Make
12395 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
12396 (gimple_init_edge_profiler): Update prototype of
12397 __gcov_indirect_call_profiler.
12398 (gimple_gen_ic_func_profiler): Simplify.
12399 (tree_profiling): Use init_node_map
12401 2013-08-09 Jan Hubicka <jh@suse.cz>
12403 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
12404 non-speculative refs.
12405 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
12406 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
12407 (cgraph_set_call_stmt): Likewise.
12408 (cgraph_create_edge_1): Fix release checking compilatoin;
12409 clear lto_stmt_uid.
12410 (cgraph_free_edge): Free indirect info.
12411 (cgraph_turn_edge_to_speculative): New function.
12412 (cgraph_speculative_call_info): New function.
12413 (cgraph_make_edge_direct): Return direct edge; handle speculation.
12414 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
12415 (dump_cgraph_node): Dump speculation.
12416 (verify_edge_count_and_frequency): Accept speculative edges.
12417 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
12418 (verify_cgraph_node): Handle speculation.
12419 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
12420 (cgraph_set_call_stmt): Update prototype.
12421 (cgraph_make_edge_direct): Update prototype.
12422 (cgraph_speculative_call_info): Declare.
12423 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
12424 to change; update call of ipa_find_references.
12425 * ipa-ref.c (ipa_record_reference): Fix return value; clear
12426 lto_stmt_uid and speculative flags.
12427 (ipa_dump_references): Dump speculation.
12428 (ipa_clone_references): Clone speculative flag.
12429 (ipa_clone_referring): Likewise.
12430 (ipa_clone_ref): New function.
12431 (ipa_find_reference): Look into lto_stmt_uids
12432 (ipa_clear_stmts_in_references): Do not clear speculative calls.
12433 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
12434 (ipa_find_reference): Update declaration.
12435 (ipa_clone_ref): Declare.
12436 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
12437 stream speculative flag.
12438 (lto_output_ref): Stream statements uids and speculation.
12439 (input_ref): Likewise.
12440 (input_edge): Stream speuclation.
12441 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
12442 (cgraph_set_call_stmt_including_clones): Handle speculation.
12443 * ipa-inline.c (heap_edge_removal_hook): New function.
12444 (inline_small_functions): Register it.
12445 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
12446 also initialize refs.
12447 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
12449 (try_make_edge_direct_simple_call): Likewise.
12450 (try_make_edge_direct_simple_call): Likewise.
12451 (update_indirect_edges_after_inlining): Likewise.
12452 (remove_described_reference): Look proper lto_stmt_uid.
12453 (propagate_controlled_uses): Likewise.
12454 (propagate_controlled_uses): Liekwise.
12455 * tree-inline.c (copy_bb): Copy speculative edges.
12456 (redirect_all_calls): New function.
12457 (copy_cfg_body): Do redirection after loop info is updated.
12458 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
12460 2013-08-09 Jan Hubicka <jh@suse.cz>
12462 * lto-streamer-out.c (output_function): Renumber PHIs.
12463 * lto-streamer-in.c (input_function): Likewise.
12465 2013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
12467 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
12468 (get_lane_unsigned): Likewise.
12469 (dup_lane_scalar): Likewise.
12470 (get_lane): enable for VALL.
12471 * config/aarch64/aarch64-simd.md
12472 (aarch64_dup_lane_scalar<mode>): Remove.
12473 (aarch64_get_lane_signed<mode>): Likewise.
12474 (aarch64_get_lane_unsigned<mode>): Likewise.
12475 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
12476 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
12477 (aarch64_get_lane<mode>): Enable for all vector modes.
12478 (aarch64_get_lanedi): Remove misleading constraints.
12479 * config/aarch64/arm_neon.h
12480 (__aarch64_vget_lane_any): Define.
12481 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
12482 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
12483 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
12484 * config/aarch64/iterators.md (VDQQH): New.
12485 (VDQQHS): Likewise.
12486 (vwcore): Likewise.
12488 2013-08-09 Eric Botcazou <ebotcazou@adacore.com>
12490 * configure.ac: Add GAS check for LEON instructions on SPARC.
12491 * configure: Regenerate.
12492 * config.in: Likewise.
12493 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
12495 * config/sparc/sparc.opt (LEON, LEON3): New masks.
12496 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
12498 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
12499 (AS_LEON_FLAG): New macro.
12500 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
12501 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
12502 Deal with LEON and LEON3 for the memory model.
12503 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
12504 (atomic_compare_and_swap<mode>_1): Likewise.
12505 (*atomic_compare_and_swap<mode>_1): Likewise.
12507 2013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
12509 * config/arm/neon.md (vcond): Fix floating-point vector
12510 comparisons against 0.
12512 2013-08-08 Vladimir Makarov <vmakarov@redhat.com>
12514 * lra-constraints.c (emit_spill_move): Remove assert.
12515 (process_alt_operands): Add more debugging
12516 output. Increase reject for spilling into memory. Decrease
12517 reject for reloading scratch.
12518 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
12520 2013-08-08 Steve Ellcey <sellcey@mips.com>
12522 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
12523 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
12524 micromips incompatible. Add nan2008.
12525 (MULTILIB_DIRNAMES): Add nan2008.
12526 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
12527 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
12528 and micromips incompatible. Add nan2008.
12529 (MULTILIB_DIRNAMES): Add nan2008.
12530 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
12532 2013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
12534 PR rtl-optimization/58079
12535 * combine.c (combine_simplify_rtx): Avoid using SUBST if
12536 simplify_comparison has widened a comparison with an integer.
12538 2013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12540 * config/arm/neon.md (movmisalign<mode>): Disable when we
12541 don't allow unaligned accesses.
12542 (*movmisalign<mode>_neon_store): Likewise.
12543 (*movmisalign<mode>_neon_load): Likewise.
12544 (*movmisalign<mode>_neon_store): Likewise.
12545 (*movmisalign<mode>_neon_load): Likewise.
12547 2013-08-08 Jan Hubicka <jh@suse.cz>
12549 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
12550 (make_pass_rebuild_cgraph_edges): Also clear references.
12551 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
12552 * ipa-inline-transform.c (inline_transform): Remove all references
12554 * cgraphunit.c (expand_function): Remove all references after
12556 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
12557 (ipa_find_reference): Rewrite to iterator.
12558 (remove_stmt_references): Likewise.
12559 (ipa_clear_stmts_in_references): New function.
12560 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
12561 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
12563 * ipa-split.c (split_function): Remove references in split function.
12565 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
12568 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
12569 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
12571 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
12574 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
12575 suggested mode for the assignment isn't compatible with the
12576 registers required.
12578 2013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
12581 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
12583 2013-08-07 Xinliang David Li <davidxl@google.com>
12585 * config/i386/i386.opt: New option -mtune-ctrl=.
12586 * config/i386/x86-tune.def: New file.
12587 * config/i386/i386.h: include x86-tune.def.
12588 * config/i386/i386.c (ix86_option_override_internal):
12589 Parsing -mtune-ctrl= option and set tune features.
12591 2013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
12594 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
12595 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
12598 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
12600 * rtl.h (update_alignments): Declare.
12601 * final.c (grow_label_align): New function extracted from...
12602 (shorten_branches): ...here. Call it.
12603 (update_alignments): New function.
12604 * reorg.c (sibling_labels): New variable.
12605 (get_label_before): Add SIBLING parameter. If it is non-zero, push
12606 the new label along with it onto the sibling_labels vector.
12607 (fill_simple_delay_slots): Adjust call to get_label_before.
12608 (fill_slots_from_thread): Likewise.
12609 (relax_delay_slots): Likewise.
12610 (make_return_insns): Likewise.
12611 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
12613 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
12615 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
12616 document its semantics.
12617 (diagnostic_report_diagnostic): Adjust accordingly.
12619 2013-08-07 David Malcolm <dmalcolm@redhat.com>
12621 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
12622 (sparc_option_override): ...and port to new C++ pass API.
12623 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
12625 2013-08-07 Peter Bergner <bergner@vnet.ibm.com>
12627 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
12629 2013-08-06 Caroline Tice <cmtice@google.com>
12631 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
12632 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
12633 * tree-pass.h: Add pass_vtable_verify.
12634 * varasm.c (assemble_variable): Add code to properly set the comdat
12635 section and name for the .vtable_map_vars section.
12636 (assemble_vtyv_preinit_initializer): New function.
12637 (default_sectin_type_flags): Make sure .vtable_map_vars section has
12639 * output.h: Add function decl for assemble_vtv_preinit_initializer.
12640 * vtable-verify.c: New file.
12641 * vtable-verify.h: New file.
12642 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
12643 initialiation levels.
12644 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
12645 * passes.def: Insert pass_vtable_verify.
12646 * aclocal.m4: Reorder includes.
12647 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
12648 -fvtv-counts options.
12649 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
12650 as appropriate, if -fvtable-verify=... is used.
12651 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
12652 -fvtable-verify=... is used.
12653 * Makefile.in (OBJS): Add vtable-verify.o to list.
12654 (vtable-verify.o): Add new build rule.
12655 (GTFILES): Add vtable-verify.c to list.
12656 * common.opt (fvtable-verify=): New flag.
12657 (vtv_priority): Values for fvtable-verify= flag.
12658 (fvtv-counts): New flag.
12659 (fvtv-debug): New flag.
12660 * tree.h (save_vtable_map_decl): New extern function decl.
12662 2013-08-07 David Malcolm <dmalcolm@redhat.com>
12664 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
12665 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
12666 (pass_data_rl78_devirt): ...new pass_data instance and...
12667 (make_pass_rl78_devirt): ...new function.
12668 (rl78_asm_file_start): Port pass registration to new C++ API.
12670 2013-08-07 David Malcolm <dmalcolm@redhat.com>
12672 * coretypes.h (rtl_opt_pass): Add.
12673 (gcc::context): Add.
12674 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
12675 (epiphany_init): Port to new C++ pass API.
12676 (epiphany_optimize_mode_switching): Likewise.
12677 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
12678 (pass_manager::get_pass_mode_switching): New.
12679 (pass_manager::get_pass_peephole2): New.
12680 * mode-switching.c (pass_mode_switching): Add clone method.
12681 * recog.c (pass_peephole2): Add clone method.
12682 (pass_split_all_insns): Add clone method.
12684 2013-08-06 David Malcolm <dmalcolm@redhat.com>
12686 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
12687 (mips_option_override): ...here, porting to new C++ API for passes.
12689 2013-08-06 Jan Hubicka <jh@suse.cz>
12691 * cgraph.c (cgraph_get_body): New function based on lto.c
12693 * cgraph.h (cgraph_get_body): Declare.
12694 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
12696 * cgraphunit.c (expand_function): Get body prior expanding.
12697 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
12698 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
12700 * passes.c (do_per_function_toporder): Get body.
12701 * tree-inline.c (expand_call_inline): Get body prior inlining it.
12702 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
12704 2013-08-06 Martin Jambor <mjambor@suse.cz>
12707 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
12708 is not re-finalized. Rename second parameter to no_collect.
12710 2013-08-06 Martin Jambor <mjambor@suse.cz>
12712 PR middle-end/58041
12713 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
12714 MEM_REF has proper alignment information.
12716 2013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
12719 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
12721 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
12722 rtx (*) (rtx, ...) with insn_gen_fn.
12723 * genoutput.c (output_insn_data): Cast gen_? function pointers to
12724 insn_gen_fn::stored_funcptr. Add initializer braces.
12726 2013-08-05 David Malcolm <dmalcolm@redhat.com>
12728 Rewrite how instances of passes are cloned to remove assumptions
12729 about their sizes (thus allowing pass subclasses to have
12730 additional data fields, albeit non-GC-managed ones at this point).
12732 * passes.c (make_pass_instance): Now that passes have clone
12733 methods, rewrite this function to eliminate XNEW and memcpy
12734 calls that used hardcoded sizes. Since this function no longer
12735 creates pass instances, rename it to...
12736 (add_pass_instance): ...this. Document the old way that passes were
12737 numbered and flagged, and rework this function to continue using it.
12738 (next_pass_1): Add an initial_pass argument for use by
12740 (position_pass): When adding multiple instances of a pass, use
12741 the pass's clone method, rather than relying on the XNEW/memcpy
12742 within the former make_pass_instance (now add_pass_instance).
12743 (pass_manager::pass_manager): When invoking next_pass_1, also supply
12744 the initial instance of the current pass within the pass manager.
12746 2013-08-05 David Malcolm <dmalcolm@redhat.com>
12748 This is the automated part of the conversion of passes from C
12749 structs to C++ classes.
12751 Patch autogenerated by refactor_passes.py from
12752 https://github.com/davidmalcolm/gcc-refactoring-scripts
12753 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
12755 * asan.c (pass_asan): Convert from a global struct to a subclass of
12756 gimple_opt_pass along with...
12757 (pass_data_asan): ...new pass_data instance and...
12758 (make_pass_asan): ...new function.
12759 (pass_asan_O0): Convert from a global struct to a subclass of
12760 gimple_opt_pass along with...
12761 (pass_data_asan_O0): ...new pass_data instance and...
12762 (make_pass_asan_O0): ...new function.
12763 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
12764 subclass of rtl_opt_pass along with...
12765 (pass_data_inc_dec): ...new pass_data instance and...
12766 (make_pass_inc_dec): ...new function.
12767 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
12768 a subclass of rtl_opt_pass along with...
12769 (pass_data_reorder_blocks): ...new pass_data instance and...
12770 (make_pass_reorder_blocks): ...new function.
12771 (pass_duplicate_computed_gotos): Convert from a global struct to a
12772 subclass of rtl_opt_pass along with...
12773 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
12774 (make_pass_duplicate_computed_gotos): ...new function.
12775 (pass_partition_blocks): Convert from a global struct to a subclass of
12776 rtl_opt_pass along with...
12777 (pass_data_partition_blocks): ...new pass_data instance and...
12778 (make_pass_partition_blocks): ...new function.
12779 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
12780 struct to a subclass of rtl_opt_pass along with...
12781 (pass_data_branch_target_load_optimize1): ...new pass_data instance
12783 (make_pass_branch_target_load_optimize1): ...new function.
12784 (pass_branch_target_load_optimize2): Convert from a global struct to a
12785 subclass of rtl_opt_pass along with...
12786 (pass_data_branch_target_load_optimize2): ...new pass_data instance
12788 (make_pass_branch_target_load_optimize2): ...new function.
12789 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
12790 of rtl_opt_pass along with...
12791 (pass_data_jump): ...new pass_data instance and...
12792 (make_pass_jump): ...new function.
12793 (pass_jump2): Convert from a global struct to a subclass of
12794 rtl_opt_pass along with...
12795 (pass_data_jump2): ...new pass_data instance and...
12796 (make_pass_jump2): ...new function.
12797 * cfgexpand.c (pass_expand): Convert from a global struct to a
12798 subclass of rtl_opt_pass along with...
12799 (pass_data_expand): ...new pass_data instance and...
12800 (make_pass_expand): ...new function.
12801 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
12802 of rtl_opt_pass along with...
12803 (pass_data_free_cfg): ...new pass_data instance and...
12804 (make_pass_free_cfg): ...new function.
12805 (pass_into_cfg_layout_mode): Convert from a global struct to a
12806 subclass of rtl_opt_pass along with...
12807 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
12808 (make_pass_into_cfg_layout_mode): ...new function.
12809 (pass_outof_cfg_layout_mode): Convert from a global struct to a
12810 subclass of rtl_opt_pass along with...
12811 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
12812 (make_pass_outof_cfg_layout_mode): ...new function.
12813 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
12814 struct to a subclass of gimple_opt_pass along with...
12815 (pass_data_build_cgraph_edges): ...new pass_data instance and...
12816 (make_pass_build_cgraph_edges): ...new function.
12817 (pass_rebuild_cgraph_edges): Convert from a global struct to a
12818 subclass of gimple_opt_pass along with...
12819 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
12820 (make_pass_rebuild_cgraph_edges): ...new function.
12821 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
12822 subclass of gimple_opt_pass along with...
12823 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
12825 (make_pass_remove_cgraph_callee_edges): ...new function.
12826 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
12827 struct to a subclass of rtl_opt_pass along with...
12828 (pass_data_stack_adjustments): ...new pass_data instance and...
12829 (make_pass_stack_adjustments): ...new function.
12830 * combine.c (pass_combine): Convert from a global struct to a subclass
12831 of rtl_opt_pass along with...
12832 (pass_data_combine): ...new pass_data instance and...
12833 (make_pass_combine): ...new function.
12834 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
12835 global struct to a subclass of rtl_opt_pass along with...
12836 (pass_data_compare_elim_after_reload): ...new pass_data instance
12838 (make_pass_compare_elim_after_reload): ...new function.
12839 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
12840 of rtl_opt_pass along with...
12841 (pass_data_rtl_cprop): ...new pass_data instance and...
12842 (make_pass_rtl_cprop): ...new function.
12843 * cse.c (pass_cse): Convert from a global struct to a subclass of
12844 rtl_opt_pass along with...
12845 (pass_data_cse): ...new pass_data instance and...
12846 (make_pass_cse): ...new function.
12847 (pass_cse2): Convert from a global struct to a subclass of
12848 rtl_opt_pass along with...
12849 (pass_data_cse2): ...new pass_data instance and...
12850 (make_pass_cse2): ...new function.
12851 (pass_cse_after_global_opts): Convert from a global struct to a
12852 subclass of rtl_opt_pass along with...
12853 (pass_data_cse_after_global_opts): ...new pass_data instance and...
12854 (make_pass_cse_after_global_opts): ...new function.
12855 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
12856 of rtl_opt_pass along with...
12857 (pass_data_ud_rtl_dce): ...new pass_data instance and...
12858 (make_pass_ud_rtl_dce): ...new function.
12859 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
12860 rtl_opt_pass along with...
12861 (pass_data_fast_rtl_dce): ...new pass_data instance and...
12862 (make_pass_fast_rtl_dce): ...new function.
12863 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
12864 a subclass of rtl_opt_pass along with...
12865 (pass_data_df_initialize_opt): ...new pass_data instance and...
12866 (make_pass_df_initialize_opt): ...new function.
12867 (pass_df_initialize_no_opt): Convert from a global struct to a
12868 subclass of rtl_opt_pass along with...
12869 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
12870 (make_pass_df_initialize_no_opt): ...new function.
12871 (pass_df_finish): Convert from a global struct to a subclass of
12872 rtl_opt_pass along with...
12873 (pass_data_df_finish): ...new pass_data instance and...
12874 (make_pass_df_finish): ...new function.
12875 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
12876 rtl_opt_pass along with...
12877 (pass_data_rtl_dse1): ...new pass_data instance and...
12878 (make_pass_rtl_dse1): ...new function.
12879 (pass_rtl_dse2): Convert from a global struct to a subclass of
12880 rtl_opt_pass along with...
12881 (pass_data_rtl_dse2): ...new pass_data instance and...
12882 (make_pass_rtl_dse2): ...new function.
12883 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
12884 subclass of rtl_opt_pass along with...
12885 (pass_data_dwarf2_frame): ...new pass_data instance and...
12886 (make_pass_dwarf2_frame): ...new function.
12887 * except.c (pass_set_nothrow_function_flags): Convert from a global
12888 struct to a subclass of rtl_opt_pass along with...
12889 (pass_data_set_nothrow_function_flags): ...new pass_data instance
12891 (make_pass_set_nothrow_function_flags): ...new function.
12892 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
12893 subclass of rtl_opt_pass along with...
12894 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
12896 (make_pass_convert_to_eh_region_ranges): ...new function.
12897 * final.c (pass_compute_alignments): Convert from a global struct to a
12898 subclass of rtl_opt_pass along with...
12899 (pass_data_compute_alignments): ...new pass_data instance and...
12900 (make_pass_compute_alignments): ...new function.
12901 (pass_final): Convert from a global struct to a subclass of
12902 rtl_opt_pass along with...
12903 (pass_data_final): ...new pass_data instance and...
12904 (make_pass_final): ...new function.
12905 (pass_shorten_branches): Convert from a global struct to a subclass of
12906 rtl_opt_pass along with...
12907 (pass_data_shorten_branches): ...new pass_data instance and...
12908 (make_pass_shorten_branches): ...new function.
12909 (pass_clean_state): Convert from a global struct to a subclass of
12910 rtl_opt_pass along with...
12911 (pass_data_clean_state): ...new pass_data instance and...
12912 (make_pass_clean_state): ...new function.
12913 * function.c (pass_instantiate_virtual_regs): Convert from a global
12914 struct to a subclass of rtl_opt_pass along with...
12915 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
12916 (make_pass_instantiate_virtual_regs): ...new function.
12917 (pass_leaf_regs): Convert from a global struct to a subclass of
12918 rtl_opt_pass along with...
12919 (pass_data_leaf_regs): ...new pass_data instance and...
12920 (make_pass_leaf_regs): ...new function.
12921 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
12922 subclass of rtl_opt_pass along with...
12923 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
12925 (make_pass_thread_prologue_and_epilogue): ...new function.
12926 (pass_match_asm_constraints): Convert from a global struct to a
12927 subclass of rtl_opt_pass along with...
12928 (pass_data_match_asm_constraints): ...new pass_data instance and...
12929 (make_pass_match_asm_constraints): ...new function.
12930 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
12931 subclass of rtl_opt_pass along with...
12932 (pass_data_rtl_fwprop): ...new pass_data instance and...
12933 (make_pass_rtl_fwprop): ...new function.
12934 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
12935 rtl_opt_pass along with...
12936 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
12937 (make_pass_rtl_fwprop_addr): ...new function.
12938 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
12939 rtl_opt_pass along with...
12940 (pass_data_rtl_pre): ...new pass_data instance and...
12941 (make_pass_rtl_pre): ...new function.
12942 (pass_rtl_hoist): Convert from a global struct to a subclass of
12943 rtl_opt_pass along with...
12944 (pass_data_rtl_hoist): ...new pass_data instance and...
12945 (make_pass_rtl_hoist): ...new function.
12946 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
12947 subclass of gimple_opt_pass along with...
12948 (pass_data_lower_cf): ...new pass_data instance and...
12949 (make_pass_lower_cf): ...new function.
12950 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
12951 from a global struct to a subclass of gimple_opt_pass along with...
12952 (pass_data_strength_reduction): ...new pass_data instance and...
12953 (make_pass_strength_reduction): ...new function.
12954 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
12955 of rtl_opt_pass along with...
12956 (pass_data_rtl_ifcvt): ...new pass_data instance and...
12957 (make_pass_rtl_ifcvt): ...new function.
12958 (pass_if_after_combine): Convert from a global struct to a subclass of
12959 rtl_opt_pass along with...
12960 (pass_data_if_after_combine): ...new pass_data instance and...
12961 (make_pass_if_after_combine): ...new function.
12962 (pass_if_after_reload): Convert from a global struct to a subclass of
12963 rtl_opt_pass along with...
12964 (pass_data_if_after_reload): ...new pass_data instance and...
12965 (make_pass_if_after_reload): ...new function.
12966 * init-regs.c (pass_initialize_regs): Convert from a global struct to
12967 a subclass of rtl_opt_pass along with...
12968 (pass_data_initialize_regs): ...new pass_data instance and...
12969 (make_pass_initialize_regs): ...new function.
12970 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
12971 of ipa_opt_pass_d along with...
12972 (pass_data_ipa_cp): ...new pass_data instance and...
12973 (make_pass_ipa_cp): ...new function.
12974 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
12975 global struct to a subclass of gimple_opt_pass along with...
12976 (pass_data_inline_parameters): ...new pass_data instance and...
12977 (make_pass_inline_parameters): ...new function.
12978 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
12979 subclass of gimple_opt_pass along with...
12980 (pass_data_early_inline): ...new pass_data instance and...
12981 (make_pass_early_inline): ...new function.
12982 (pass_ipa_inline): Convert from a global struct to a subclass of
12983 ipa_opt_pass_d along with...
12984 (pass_data_ipa_inline): ...new pass_data instance and...
12985 (make_pass_ipa_inline): ...new function.
12986 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
12987 struct to a subclass of gimple_opt_pass along with...
12988 (pass_data_local_pure_const): ...new pass_data instance and...
12989 (make_pass_local_pure_const): ...new function.
12990 (pass_ipa_pure_const): Convert from a global struct to a subclass of
12991 ipa_opt_pass_d along with...
12992 (pass_data_ipa_pure_const): ...new pass_data instance and...
12993 (make_pass_ipa_pure_const): ...new function.
12994 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
12995 to a subclass of ipa_opt_pass_d along with...
12996 (pass_data_ipa_reference): ...new pass_data instance and...
12997 (make_pass_ipa_reference): ...new function.
12998 * ipa-split.c (pass_split_functions): Convert from a global struct to
12999 a subclass of gimple_opt_pass along with...
13000 (pass_data_split_functions): ...new pass_data instance and...
13001 (make_pass_split_functions): ...new function.
13002 (pass_feedback_split_functions): Convert from a global struct to a
13003 subclass of gimple_opt_pass along with...
13004 (pass_data_feedback_split_functions): ...new pass_data instance and...
13005 (make_pass_feedback_split_functions): ...new function.
13006 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
13007 global struct to a subclass of simple_ipa_opt_pass along with...
13008 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
13010 (make_pass_ipa_function_and_variable_visibility): ...new function.
13011 (pass_ipa_free_inline_summary): Convert from a global struct to a
13012 subclass of simple_ipa_opt_pass along with...
13013 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
13014 (make_pass_ipa_free_inline_summary): ...new function.
13015 (pass_ipa_whole_program_visibility): Convert from a global struct to a
13016 subclass of ipa_opt_pass_d along with...
13017 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
13019 (make_pass_ipa_whole_program_visibility): ...new function.
13020 (pass_ipa_profile): Convert from a global struct to a subclass of
13021 ipa_opt_pass_d along with...
13022 (pass_data_ipa_profile): ...new pass_data instance and...
13023 (make_pass_ipa_profile): ...new function.
13024 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
13025 ipa_opt_pass_d along with...
13026 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
13027 (make_pass_ipa_cdtor_merge): ...new function.
13028 * ira.c (pass_ira): Convert from a global struct to a subclass of
13029 rtl_opt_pass along with...
13030 (pass_data_ira): ...new pass_data instance and...
13031 (make_pass_ira): ...new function.
13032 (pass_reload): Convert from a global struct to a subclass of
13033 rtl_opt_pass along with...
13034 (pass_data_reload): ...new pass_data instance and...
13035 (make_pass_reload): ...new function.
13036 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
13037 subclass of rtl_opt_pass along with...
13038 (pass_data_cleanup_barriers): ...new pass_data instance and...
13039 (make_pass_cleanup_barriers): ...new function.
13040 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
13041 of rtl_opt_pass along with...
13042 (pass_data_loop2): ...new pass_data instance and...
13043 (make_pass_loop2): ...new function.
13044 (pass_rtl_loop_init): Convert from a global struct to a subclass of
13045 rtl_opt_pass along with...
13046 (pass_data_rtl_loop_init): ...new pass_data instance and...
13047 (make_pass_rtl_loop_init): ...new function.
13048 (pass_rtl_loop_done): Convert from a global struct to a subclass of
13049 rtl_opt_pass along with...
13050 (pass_data_rtl_loop_done): ...new pass_data instance and...
13051 (make_pass_rtl_loop_done): ...new function.
13052 (pass_rtl_move_loop_invariants): Convert from a global struct to a
13053 subclass of rtl_opt_pass along with...
13054 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
13055 (make_pass_rtl_move_loop_invariants): ...new function.
13056 (pass_rtl_unswitch): Convert from a global struct to a subclass of
13057 rtl_opt_pass along with...
13058 (pass_data_rtl_unswitch): ...new pass_data instance and...
13059 (make_pass_rtl_unswitch): ...new function.
13060 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
13061 subclass of rtl_opt_pass along with...
13062 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
13064 (make_pass_rtl_unroll_and_peel_loops): ...new function.
13065 (pass_rtl_doloop): Convert from a global struct to a subclass of
13066 rtl_opt_pass along with...
13067 (pass_data_rtl_doloop): ...new pass_data instance and...
13068 (make_pass_rtl_doloop): ...new function.
13069 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
13070 a subclass of rtl_opt_pass along with...
13071 (pass_data_lower_subreg): ...new pass_data instance and...
13072 (make_pass_lower_subreg): ...new function.
13073 (pass_lower_subreg2): Convert from a global struct to a subclass of
13074 rtl_opt_pass along with...
13075 (pass_data_lower_subreg2): ...new pass_data instance and...
13076 (make_pass_lower_subreg2): ...new function.
13077 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
13078 struct to a subclass of ipa_opt_pass_d along with...
13079 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
13080 (make_pass_ipa_lto_gimple_out): ...new function.
13081 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
13082 of ipa_opt_pass_d along with...
13083 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
13084 (make_pass_ipa_lto_finish_out): ...new function.
13085 * mode-switching.c (pass_mode_switching): Convert from a global struct
13086 to a subclass of rtl_opt_pass along with...
13087 (pass_data_mode_switching): ...new pass_data instance and...
13088 (make_pass_mode_switching): ...new function.
13089 * modulo-sched.c (pass_sms): Convert from a global struct to a
13090 subclass of rtl_opt_pass along with...
13091 (pass_data_sms): ...new pass_data instance and...
13092 (make_pass_sms): ...new function.
13093 * omp-low.c (pass_expand_omp): Convert from a global struct to a
13094 subclass of gimple_opt_pass along with...
13095 (pass_data_expand_omp): ...new pass_data instance and...
13096 (make_pass_expand_omp): ...new function.
13097 (pass_lower_omp): Convert from a global struct to a subclass of
13098 gimple_opt_pass along with...
13099 (pass_data_lower_omp): ...new pass_data instance and...
13100 (make_pass_lower_omp): ...new function.
13101 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
13102 of gimple_opt_pass along with...
13103 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
13104 (make_pass_diagnose_omp_blocks): ...new function.
13105 * passes.c (pass_early_local_passes): Convert from a global struct to
13106 a subclass of simple_ipa_opt_pass along with...
13107 (pass_data_early_local_passes): ...new pass_data instance and...
13108 (make_pass_early_local_passes): ...new function.
13109 (pass_all_early_optimizations): Convert from a global struct to a
13110 subclass of gimple_opt_pass along with...
13111 (pass_data_all_early_optimizations): ...new pass_data instance and...
13112 (make_pass_all_early_optimizations): ...new function.
13113 (pass_all_optimizations): Convert from a global struct to a subclass
13114 of gimple_opt_pass along with...
13115 (pass_data_all_optimizations): ...new pass_data instance and...
13116 (make_pass_all_optimizations): ...new function.
13117 (pass_all_optimizations_g): Convert from a global struct to a subclass
13118 of gimple_opt_pass along with...
13119 (pass_data_all_optimizations_g): ...new pass_data instance and...
13120 (make_pass_all_optimizations_g): ...new function.
13121 (pass_rest_of_compilation): Convert from a global struct to a subclass
13122 of rtl_opt_pass along with...
13123 (pass_data_rest_of_compilation): ...new pass_data instance and...
13124 (make_pass_rest_of_compilation): ...new function.
13125 (pass_postreload): Convert from a global struct to a subclass of
13126 rtl_opt_pass along with...
13127 (pass_data_postreload): ...new pass_data instance and...
13128 (make_pass_postreload): ...new function.
13129 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
13130 subclass of rtl_opt_pass along with...
13131 (pass_data_gcse2): ...new pass_data instance and...
13132 (make_pass_gcse2): ...new function.
13133 * postreload.c (pass_postreload_cse): Convert from a global struct to
13134 a subclass of rtl_opt_pass along with...
13135 (pass_data_postreload_cse): ...new pass_data instance and...
13136 (make_pass_postreload_cse): ...new function.
13137 * predict.c (pass_profile): Convert from a global struct to a subclass
13138 of gimple_opt_pass along with...
13139 (pass_data_profile): ...new pass_data instance and...
13140 (make_pass_profile): ...new function.
13141 (pass_strip_predict_hints): Convert from a global struct to a subclass
13142 of gimple_opt_pass along with...
13143 (pass_data_strip_predict_hints): ...new pass_data instance and...
13144 (make_pass_strip_predict_hints): ...new function.
13145 * recog.c (pass_peephole2): Convert from a global struct to a subclass
13146 of rtl_opt_pass along with...
13147 (pass_data_peephole2): ...new pass_data instance and...
13148 (make_pass_peephole2): ...new function.
13149 (pass_split_all_insns): Convert from a global struct to a subclass of
13150 rtl_opt_pass along with...
13151 (pass_data_split_all_insns): ...new pass_data instance and...
13152 (make_pass_split_all_insns): ...new function.
13153 (pass_split_after_reload): Convert from a global struct to a subclass
13154 of rtl_opt_pass along with...
13155 (pass_data_split_after_reload): ...new pass_data instance and...
13156 (make_pass_split_after_reload): ...new function.
13157 (pass_split_before_regstack): Convert from a global struct to a
13158 subclass of rtl_opt_pass along with...
13159 (pass_data_split_before_regstack): ...new pass_data instance and...
13160 (make_pass_split_before_regstack): ...new function.
13161 (pass_split_before_sched2): Convert from a global struct to a subclass
13162 of rtl_opt_pass along with...
13163 (pass_data_split_before_sched2): ...new pass_data instance and...
13164 (make_pass_split_before_sched2): ...new function.
13165 (pass_split_for_shorten_branches): Convert from a global struct to a
13166 subclass of rtl_opt_pass along with...
13167 (pass_data_split_for_shorten_branches): ...new pass_data instance
13169 (make_pass_split_for_shorten_branches): ...new function.
13170 * ree.c (pass_ree): Convert from a global struct to a subclass of
13171 rtl_opt_pass along with...
13172 (pass_data_ree): ...new pass_data instance and...
13173 (make_pass_ree): ...new function.
13174 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
13175 subclass of rtl_opt_pass along with...
13176 (pass_data_stack_regs): ...new pass_data instance and...
13177 (make_pass_stack_regs): ...new function.
13178 (pass_stack_regs_run): Convert from a global struct to a subclass of
13179 rtl_opt_pass along with...
13180 (pass_data_stack_regs_run): ...new pass_data instance and...
13181 (make_pass_stack_regs_run): ...new function.
13182 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
13183 subclass of rtl_opt_pass along with...
13184 (pass_data_cprop_hardreg): ...new pass_data instance and...
13185 (make_pass_cprop_hardreg): ...new function.
13186 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
13187 subclass of rtl_opt_pass along with...
13188 (pass_data_reginfo_init): ...new pass_data instance and...
13189 (make_pass_reginfo_init): ...new function.
13190 * regmove.c (pass_regmove): Convert from a global struct to a subclass
13191 of rtl_opt_pass along with...
13192 (pass_data_regmove): ...new pass_data instance and...
13193 (make_pass_regmove): ...new function.
13194 * regrename.c (pass_regrename): Convert from a global struct to a
13195 subclass of rtl_opt_pass along with...
13196 (pass_data_regrename): ...new pass_data instance and...
13197 (make_pass_regrename): ...new function.
13198 * reorg.c (pass_delay_slots): Convert from a global struct to a
13199 subclass of rtl_opt_pass along with...
13200 (pass_data_delay_slots): ...new pass_data instance and...
13201 (make_pass_delay_slots): ...new function.
13202 (pass_machine_reorg): Convert from a global struct to a subclass of
13203 rtl_opt_pass along with...
13204 (pass_data_machine_reorg): ...new pass_data instance and...
13205 (make_pass_machine_reorg): ...new function.
13206 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
13207 of rtl_opt_pass along with...
13208 (pass_data_sched): ...new pass_data instance and...
13209 (make_pass_sched): ...new function.
13210 (pass_sched2): Convert from a global struct to a subclass of
13211 rtl_opt_pass along with...
13212 (pass_data_sched2): ...new pass_data instance and...
13213 (make_pass_sched2): ...new function.
13214 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
13215 to a subclass of rtl_opt_pass along with...
13216 (pass_data_stack_ptr_mod): ...new pass_data instance and...
13217 (make_pass_stack_ptr_mod): ...new function.
13218 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
13219 to a subclass of rtl_opt_pass along with...
13220 (pass_data_rtl_store_motion): ...new pass_data instance and...
13221 (make_pass_rtl_store_motion): ...new function.
13222 * tracer.c (pass_tracer): Convert from a global struct to a subclass
13223 of gimple_opt_pass along with...
13224 (pass_data_tracer): ...new pass_data instance and...
13225 (make_pass_tracer): ...new function.
13226 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
13227 to a subclass of gimple_opt_pass along with...
13228 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
13229 (make_pass_diagnose_tm_blocks): ...new function.
13230 (pass_lower_tm): Convert from a global struct to a subclass of
13231 gimple_opt_pass along with...
13232 (pass_data_lower_tm): ...new pass_data instance and...
13233 (make_pass_lower_tm): ...new function.
13234 (pass_tm_init): Convert from a global struct to a subclass of
13235 gimple_opt_pass along with...
13236 (pass_data_tm_init): ...new pass_data instance and...
13237 (make_pass_tm_init): ...new function.
13238 (pass_tm_mark): Convert from a global struct to a subclass of
13239 gimple_opt_pass along with...
13240 (pass_data_tm_mark): ...new pass_data instance and...
13241 (make_pass_tm_mark): ...new function.
13242 (pass_tm_edges): Convert from a global struct to a subclass of
13243 gimple_opt_pass along with...
13244 (pass_data_tm_edges): ...new pass_data instance and...
13245 (make_pass_tm_edges): ...new function.
13246 (pass_tm_memopt): Convert from a global struct to a subclass of
13247 gimple_opt_pass along with...
13248 (pass_data_tm_memopt): ...new pass_data instance and...
13249 (make_pass_tm_memopt): ...new function.
13250 (pass_ipa_tm): Convert from a global struct to a subclass of
13251 simple_ipa_opt_pass along with...
13252 (pass_data_ipa_tm): ...new pass_data instance and...
13253 (make_pass_ipa_tm): ...new function.
13254 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
13255 subclass of gimple_opt_pass along with...
13256 (pass_data_call_cdce): ...new pass_data instance and...
13257 (make_pass_call_cdce): ...new function.
13258 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
13259 subclass of gimple_opt_pass along with...
13260 (pass_data_build_cfg): ...new pass_data instance and...
13261 (make_pass_build_cfg): ...new function.
13262 (pass_split_crit_edges): Convert from a global struct to a subclass of
13263 gimple_opt_pass along with...
13264 (pass_data_split_crit_edges): ...new pass_data instance and...
13265 (make_pass_split_crit_edges): ...new function.
13266 (pass_warn_function_return): Convert from a global struct to a
13267 subclass of gimple_opt_pass along with...
13268 (pass_data_warn_function_return): ...new pass_data instance and...
13269 (make_pass_warn_function_return): ...new function.
13270 (pass_warn_function_noreturn): Convert from a global struct to a
13271 subclass of gimple_opt_pass along with...
13272 (pass_data_warn_function_noreturn): ...new pass_data instance and...
13273 (make_pass_warn_function_noreturn): ...new function.
13274 (pass_warn_unused_result): Convert from a global struct to a subclass
13275 of gimple_opt_pass along with...
13276 (pass_data_warn_unused_result): ...new pass_data instance and...
13277 (make_pass_warn_unused_result): ...new function.
13278 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
13279 a subclass of gimple_opt_pass along with...
13280 (pass_data_merge_phi): ...new pass_data instance and...
13281 (make_pass_merge_phi): ...new function.
13282 * tree-complex.c (pass_lower_complex): Convert from a global struct to
13283 a subclass of gimple_opt_pass along with...
13284 (pass_data_lower_complex): ...new pass_data instance and...
13285 (make_pass_lower_complex): ...new function.
13286 (pass_lower_complex_O0): Convert from a global struct to a subclass of
13287 gimple_opt_pass along with...
13288 (pass_data_lower_complex_O0): ...new pass_data instance and...
13289 (make_pass_lower_complex_O0): ...new function.
13290 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
13291 subclass of gimple_opt_pass along with...
13292 (pass_data_lower_eh): ...new pass_data instance and...
13293 (make_pass_lower_eh): ...new function.
13294 (pass_refactor_eh): Convert from a global struct to a subclass of
13295 gimple_opt_pass along with...
13296 (pass_data_refactor_eh): ...new pass_data instance and...
13297 (make_pass_refactor_eh): ...new function.
13298 (pass_lower_resx): Convert from a global struct to a subclass of
13299 gimple_opt_pass along with...
13300 (pass_data_lower_resx): ...new pass_data instance and...
13301 (make_pass_lower_resx): ...new function.
13302 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
13303 of gimple_opt_pass along with...
13304 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
13305 (make_pass_lower_eh_dispatch): ...new function.
13306 (pass_cleanup_eh): Convert from a global struct to a subclass of
13307 gimple_opt_pass along with...
13308 (pass_data_cleanup_eh): ...new pass_data instance and...
13309 (make_pass_cleanup_eh): ...new function.
13310 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
13311 to a subclass of simple_ipa_opt_pass along with...
13312 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
13313 (make_pass_ipa_lower_emutls): ...new function.
13314 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
13315 a subclass of gimple_opt_pass along with...
13316 (pass_data_if_conversion): ...new pass_data instance and...
13317 (make_pass_if_conversion): ...new function.
13318 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
13319 subclass of gimple_opt_pass along with...
13320 (pass_data_build_ssa): ...new pass_data instance and...
13321 (make_pass_build_ssa): ...new function.
13322 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
13323 global struct to a subclass of gimple_opt_pass along with...
13324 (pass_data_loop_distribution): ...new pass_data instance and...
13325 (make_pass_loop_distribution): ...new function.
13326 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
13327 subclass of gimple_opt_pass along with...
13328 (pass_data_mudflap_1): ...new pass_data instance and...
13329 (make_pass_mudflap_1): ...new function.
13330 (pass_mudflap_2): Convert from a global struct to a subclass of
13331 gimple_opt_pass along with...
13332 (pass_data_mudflap_2): ...new pass_data instance and...
13333 (make_pass_mudflap_2): ...new function.
13334 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
13335 subclass of gimple_opt_pass along with...
13336 (pass_data_mudflap_1): ...new pass_data instance and...
13337 (make_pass_mudflap_1): ...new function.
13338 (pass_mudflap_2): Convert from a global struct to a subclass of
13339 gimple_opt_pass along with...
13340 (pass_data_mudflap_2): ...new pass_data instance and...
13341 (make_pass_mudflap_2): ...new function.
13342 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
13343 gimple_opt_pass along with...
13344 (pass_data_nrv): ...new pass_data instance and...
13345 (make_pass_nrv): ...new function.
13346 (pass_return_slot): Convert from a global struct to a subclass of
13347 gimple_opt_pass along with...
13348 (pass_data_return_slot): ...new pass_data instance and...
13349 (make_pass_return_slot): ...new function.
13350 * tree-object-size.c (pass_object_sizes): Convert from a global struct
13351 to a subclass of gimple_opt_pass along with...
13352 (pass_data_object_sizes): ...new pass_data instance and...
13353 (make_pass_object_sizes): ...new function.
13354 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
13355 global struct to a subclass of gimple_opt_pass along with...
13356 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
13358 (make_pass_cleanup_cfg_post_optimizing): ...new function.
13359 (pass_fixup_cfg): Convert from a global struct to a subclass of
13360 gimple_opt_pass along with...
13361 (pass_data_fixup_cfg): ...new pass_data instance and...
13362 (make_pass_fixup_cfg): ...new function.
13363 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
13364 (make_pass_mudflap_1): ...new function.
13365 (pass_mudflap_2): Replace declaration with that of...
13366 (make_pass_mudflap_2): ...new function.
13367 (pass_asan): Replace declaration with that of...
13368 (make_pass_asan): ...new function.
13369 (pass_asan_O0): Replace declaration with that of...
13370 (make_pass_asan_O0): ...new function.
13371 (pass_tsan): Replace declaration with that of...
13372 (make_pass_tsan): ...new function.
13373 (pass_tsan_O0): Replace declaration with that of...
13374 (make_pass_tsan_O0): ...new function.
13375 (pass_lower_cf): Replace declaration with that of...
13376 (make_pass_lower_cf): ...new function.
13377 (pass_refactor_eh): Replace declaration with that of...
13378 (make_pass_refactor_eh): ...new function.
13379 (pass_lower_eh): Replace declaration with that of...
13380 (make_pass_lower_eh): ...new function.
13381 (pass_lower_eh_dispatch): Replace declaration with that of...
13382 (make_pass_lower_eh_dispatch): ...new function.
13383 (pass_lower_resx): Replace declaration with that of...
13384 (make_pass_lower_resx): ...new function.
13385 (pass_build_cfg): Replace declaration with that of...
13386 (make_pass_build_cfg): ...new function.
13387 (pass_early_tree_profile): Replace declaration with that of...
13388 (make_pass_early_tree_profile): ...new function.
13389 (pass_cleanup_eh): Replace declaration with that of...
13390 (make_pass_cleanup_eh): ...new function.
13391 (pass_sra): Replace declaration with that of...
13392 (make_pass_sra): ...new function.
13393 (pass_sra_early): Replace declaration with that of...
13394 (make_pass_sra_early): ...new function.
13395 (pass_early_ipa_sra): Replace declaration with that of...
13396 (make_pass_early_ipa_sra): ...new function.
13397 (pass_tail_recursion): Replace declaration with that of...
13398 (make_pass_tail_recursion): ...new function.
13399 (pass_tail_calls): Replace declaration with that of...
13400 (make_pass_tail_calls): ...new function.
13401 (pass_tree_loop): Replace declaration with that of...
13402 (make_pass_tree_loop): ...new function.
13403 (pass_tree_loop_init): Replace declaration with that of...
13404 (make_pass_tree_loop_init): ...new function.
13405 (pass_lim): Replace declaration with that of...
13406 (make_pass_lim): ...new function.
13407 (pass_tree_unswitch): Replace declaration with that of...
13408 (make_pass_tree_unswitch): ...new function.
13409 (pass_predcom): Replace declaration with that of...
13410 (make_pass_predcom): ...new function.
13411 (pass_iv_canon): Replace declaration with that of...
13412 (make_pass_iv_canon): ...new function.
13413 (pass_scev_cprop): Replace declaration with that of...
13414 (make_pass_scev_cprop): ...new function.
13415 (pass_empty_loop): Replace declaration with that of...
13416 (make_pass_empty_loop): ...new function.
13417 (pass_record_bounds): Replace declaration with that of...
13418 (make_pass_record_bounds): ...new function.
13419 (pass_graphite): Replace declaration with that of...
13420 (make_pass_graphite): ...new function.
13421 (pass_graphite_transforms): Replace declaration with that of...
13422 (make_pass_graphite_transforms): ...new function.
13423 (pass_if_conversion): Replace declaration with that of...
13424 (make_pass_if_conversion): ...new function.
13425 (pass_loop_distribution): Replace declaration with that of...
13426 (make_pass_loop_distribution): ...new function.
13427 (pass_vectorize): Replace declaration with that of...
13428 (make_pass_vectorize): ...new function.
13429 (pass_slp_vectorize): Replace declaration with that of...
13430 (make_pass_slp_vectorize): ...new function.
13431 (pass_complete_unroll): Replace declaration with that of...
13432 (make_pass_complete_unroll): ...new function.
13433 (pass_complete_unrolli): Replace declaration with that of...
13434 (make_pass_complete_unrolli): ...new function.
13435 (pass_parallelize_loops): Replace declaration with that of...
13436 (make_pass_parallelize_loops): ...new function.
13437 (pass_loop_prefetch): Replace declaration with that of...
13438 (make_pass_loop_prefetch): ...new function.
13439 (pass_iv_optimize): Replace declaration with that of...
13440 (make_pass_iv_optimize): ...new function.
13441 (pass_tree_loop_done): Replace declaration with that of...
13442 (make_pass_tree_loop_done): ...new function.
13443 (pass_ch): Replace declaration with that of...
13444 (make_pass_ch): ...new function.
13445 (pass_ccp): Replace declaration with that of...
13446 (make_pass_ccp): ...new function.
13447 (pass_phi_only_cprop): Replace declaration with that of...
13448 (make_pass_phi_only_cprop): ...new function.
13449 (pass_build_ssa): Replace declaration with that of...
13450 (make_pass_build_ssa): ...new function.
13451 (pass_build_alias): Replace declaration with that of...
13452 (make_pass_build_alias): ...new function.
13453 (pass_build_ealias): Replace declaration with that of...
13454 (make_pass_build_ealias): ...new function.
13455 (pass_dominator): Replace declaration with that of...
13456 (make_pass_dominator): ...new function.
13457 (pass_dce): Replace declaration with that of...
13458 (make_pass_dce): ...new function.
13459 (pass_dce_loop): Replace declaration with that of...
13460 (make_pass_dce_loop): ...new function.
13461 (pass_cd_dce): Replace declaration with that of...
13462 (make_pass_cd_dce): ...new function.
13463 (pass_call_cdce): Replace declaration with that of...
13464 (make_pass_call_cdce): ...new function.
13465 (pass_merge_phi): Replace declaration with that of...
13466 (make_pass_merge_phi): ...new function.
13467 (pass_split_crit_edges): Replace declaration with that of...
13468 (make_pass_split_crit_edges): ...new function.
13469 (pass_pre): Replace declaration with that of...
13470 (make_pass_pre): ...new function.
13471 (pass_profile): Replace declaration with that of...
13472 (make_pass_profile): ...new function.
13473 (pass_strip_predict_hints): Replace declaration with that of...
13474 (make_pass_strip_predict_hints): ...new function.
13475 (pass_lower_complex_O0): Replace declaration with that of...
13476 (make_pass_lower_complex_O0): ...new function.
13477 (pass_lower_complex): Replace declaration with that of...
13478 (make_pass_lower_complex): ...new function.
13479 (pass_lower_vector): Replace declaration with that of...
13480 (make_pass_lower_vector): ...new function.
13481 (pass_lower_vector_ssa): Replace declaration with that of...
13482 (make_pass_lower_vector_ssa): ...new function.
13483 (pass_lower_omp): Replace declaration with that of...
13484 (make_pass_lower_omp): ...new function.
13485 (pass_diagnose_omp_blocks): Replace declaration with that of...
13486 (make_pass_diagnose_omp_blocks): ...new function.
13487 (pass_expand_omp): Replace declaration with that of...
13488 (make_pass_expand_omp): ...new function.
13489 (pass_expand_omp_ssa): Replace declaration with that of...
13490 (make_pass_expand_omp_ssa): ...new function.
13491 (pass_object_sizes): Replace declaration with that of...
13492 (make_pass_object_sizes): ...new function.
13493 (pass_strlen): Replace declaration with that of...
13494 (make_pass_strlen): ...new function.
13495 (pass_fold_builtins): Replace declaration with that of...
13496 (make_pass_fold_builtins): ...new function.
13497 (pass_stdarg): Replace declaration with that of...
13498 (make_pass_stdarg): ...new function.
13499 (pass_early_warn_uninitialized): Replace declaration with that of...
13500 (make_pass_early_warn_uninitialized): ...new function.
13501 (pass_late_warn_uninitialized): Replace declaration with that of...
13502 (make_pass_late_warn_uninitialized): ...new function.
13503 (pass_cse_reciprocals): Replace declaration with that of...
13504 (make_pass_cse_reciprocals): ...new function.
13505 (pass_cse_sincos): Replace declaration with that of...
13506 (make_pass_cse_sincos): ...new function.
13507 (pass_optimize_bswap): Replace declaration with that of...
13508 (make_pass_optimize_bswap): ...new function.
13509 (pass_optimize_widening_mul): Replace declaration with that of...
13510 (make_pass_optimize_widening_mul): ...new function.
13511 (pass_warn_function_return): Replace declaration with that of...
13512 (make_pass_warn_function_return): ...new function.
13513 (pass_warn_function_noreturn): Replace declaration with that of...
13514 (make_pass_warn_function_noreturn): ...new function.
13515 (pass_cselim): Replace declaration with that of...
13516 (make_pass_cselim): ...new function.
13517 (pass_phiopt): Replace declaration with that of...
13518 (make_pass_phiopt): ...new function.
13519 (pass_forwprop): Replace declaration with that of...
13520 (make_pass_forwprop): ...new function.
13521 (pass_phiprop): Replace declaration with that of...
13522 (make_pass_phiprop): ...new function.
13523 (pass_tree_ifcombine): Replace declaration with that of...
13524 (make_pass_tree_ifcombine): ...new function.
13525 (pass_dse): Replace declaration with that of...
13526 (make_pass_dse): ...new function.
13527 (pass_nrv): Replace declaration with that of...
13528 (make_pass_nrv): ...new function.
13529 (pass_rename_ssa_copies): Replace declaration with that of...
13530 (make_pass_rename_ssa_copies): ...new function.
13531 (pass_sink_code): Replace declaration with that of...
13532 (make_pass_sink_code): ...new function.
13533 (pass_fre): Replace declaration with that of...
13534 (make_pass_fre): ...new function.
13535 (pass_check_data_deps): Replace declaration with that of...
13536 (make_pass_check_data_deps): ...new function.
13537 (pass_copy_prop): Replace declaration with that of...
13538 (make_pass_copy_prop): ...new function.
13539 (pass_vrp): Replace declaration with that of...
13540 (make_pass_vrp): ...new function.
13541 (pass_uncprop): Replace declaration with that of...
13542 (make_pass_uncprop): ...new function.
13543 (pass_return_slot): Replace declaration with that of...
13544 (make_pass_return_slot): ...new function.
13545 (pass_reassoc): Replace declaration with that of...
13546 (make_pass_reassoc): ...new function.
13547 (pass_rebuild_cgraph_edges): Replace declaration with that of...
13548 (make_pass_rebuild_cgraph_edges): ...new function.
13549 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
13550 (make_pass_remove_cgraph_callee_edges): ...new function.
13551 (pass_build_cgraph_edges): Replace declaration with that of...
13552 (make_pass_build_cgraph_edges): ...new function.
13553 (pass_local_pure_const): Replace declaration with that of...
13554 (make_pass_local_pure_const): ...new function.
13555 (pass_tracer): Replace declaration with that of...
13556 (make_pass_tracer): ...new function.
13557 (pass_warn_unused_result): Replace declaration with that of...
13558 (make_pass_warn_unused_result): ...new function.
13559 (pass_diagnose_tm_blocks): Replace declaration with that of...
13560 (make_pass_diagnose_tm_blocks): ...new function.
13561 (pass_lower_tm): Replace declaration with that of...
13562 (make_pass_lower_tm): ...new function.
13563 (pass_tm_init): Replace declaration with that of...
13564 (make_pass_tm_init): ...new function.
13565 (pass_tm_mark): Replace declaration with that of...
13566 (make_pass_tm_mark): ...new function.
13567 (pass_tm_memopt): Replace declaration with that of...
13568 (make_pass_tm_memopt): ...new function.
13569 (pass_tm_edges): Replace declaration with that of...
13570 (make_pass_tm_edges): ...new function.
13571 (pass_split_functions): Replace declaration with that of...
13572 (make_pass_split_functions): ...new function.
13573 (pass_feedback_split_functions): Replace declaration with that of...
13574 (make_pass_feedback_split_functions): ...new function.
13575 (pass_strength_reduction): Replace declaration with that of...
13576 (make_pass_strength_reduction): ...new function.
13577 (pass_ipa_lower_emutls): Replace declaration with that of...
13578 (make_pass_ipa_lower_emutls): ...new function.
13579 (pass_ipa_function_and_variable_visibility): Replace declaration with
13581 (make_pass_ipa_function_and_variable_visibility): ...new function.
13582 (pass_ipa_tree_profile): Replace declaration with that of...
13583 (make_pass_ipa_tree_profile): ...new function.
13584 (pass_early_local_passes): Replace declaration with that of...
13585 (make_pass_early_local_passes): ...new function.
13586 (pass_ipa_whole_program_visibility): Replace declaration with that
13588 (make_pass_ipa_whole_program_visibility): ...new function.
13589 (pass_ipa_lto_gimple_out): Replace declaration with that of...
13590 (make_pass_ipa_lto_gimple_out): ...new function.
13591 (pass_ipa_increase_alignment): Replace declaration with that of...
13592 (make_pass_ipa_increase_alignment): ...new function.
13593 (pass_ipa_inline): Replace declaration with that of...
13594 (make_pass_ipa_inline): ...new function.
13595 (pass_ipa_free_lang_data): Replace declaration with that of...
13596 (make_pass_ipa_free_lang_data): ...new function.
13597 (pass_ipa_free_inline_summary): Replace declaration with that of...
13598 (make_pass_ipa_free_inline_summary): ...new function.
13599 (pass_ipa_cp): Replace declaration with that of...
13600 (make_pass_ipa_cp): ...new function.
13601 (pass_ipa_reference): Replace declaration with that of...
13602 (make_pass_ipa_reference): ...new function.
13603 (pass_ipa_pure_const): Replace declaration with that of...
13604 (make_pass_ipa_pure_const): ...new function.
13605 (pass_ipa_pta): Replace declaration with that of...
13606 (make_pass_ipa_pta): ...new function.
13607 (pass_ipa_lto_finish_out): Replace declaration with that of...
13608 (make_pass_ipa_lto_finish_out): ...new function.
13609 (pass_ipa_tm): Replace declaration with that of...
13610 (make_pass_ipa_tm): ...new function.
13611 (pass_ipa_profile): Replace declaration with that of...
13612 (make_pass_ipa_profile): ...new function.
13613 (pass_ipa_cdtor_merge): Replace declaration with that of...
13614 (make_pass_ipa_cdtor_merge): ...new function.
13615 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
13617 (make_pass_cleanup_cfg_post_optimizing): ...new function.
13618 (pass_init_datastructures): Replace declaration with that of...
13619 (make_pass_init_datastructures): ...new function.
13620 (pass_fixup_cfg): Replace declaration with that of...
13621 (make_pass_fixup_cfg): ...new function.
13622 (pass_expand): Replace declaration with that of...
13623 (make_pass_expand): ...new function.
13624 (pass_instantiate_virtual_regs): Replace declaration with that of...
13625 (make_pass_instantiate_virtual_regs): ...new function.
13626 (pass_rtl_fwprop): Replace declaration with that of...
13627 (make_pass_rtl_fwprop): ...new function.
13628 (pass_rtl_fwprop_addr): Replace declaration with that of...
13629 (make_pass_rtl_fwprop_addr): ...new function.
13630 (pass_jump): Replace declaration with that of...
13631 (make_pass_jump): ...new function.
13632 (pass_jump2): Replace declaration with that of...
13633 (make_pass_jump2): ...new function.
13634 (pass_lower_subreg): Replace declaration with that of...
13635 (make_pass_lower_subreg): ...new function.
13636 (pass_cse): Replace declaration with that of...
13637 (make_pass_cse): ...new function.
13638 (pass_fast_rtl_dce): Replace declaration with that of...
13639 (make_pass_fast_rtl_dce): ...new function.
13640 (pass_ud_rtl_dce): Replace declaration with that of...
13641 (make_pass_ud_rtl_dce): ...new function.
13642 (pass_rtl_dce): Replace declaration with that of...
13643 (make_pass_rtl_dce): ...new function.
13644 (pass_rtl_dse1): Replace declaration with that of...
13645 (make_pass_rtl_dse1): ...new function.
13646 (pass_rtl_dse2): Replace declaration with that of...
13647 (make_pass_rtl_dse2): ...new function.
13648 (pass_rtl_dse3): Replace declaration with that of...
13649 (make_pass_rtl_dse3): ...new function.
13650 (pass_rtl_cprop): Replace declaration with that of...
13651 (make_pass_rtl_cprop): ...new function.
13652 (pass_rtl_pre): Replace declaration with that of...
13653 (make_pass_rtl_pre): ...new function.
13654 (pass_rtl_hoist): Replace declaration with that of...
13655 (make_pass_rtl_hoist): ...new function.
13656 (pass_rtl_store_motion): Replace declaration with that of...
13657 (make_pass_rtl_store_motion): ...new function.
13658 (pass_cse_after_global_opts): Replace declaration with that of...
13659 (make_pass_cse_after_global_opts): ...new function.
13660 (pass_rtl_ifcvt): Replace declaration with that of...
13661 (make_pass_rtl_ifcvt): ...new function.
13662 (pass_into_cfg_layout_mode): Replace declaration with that of...
13663 (make_pass_into_cfg_layout_mode): ...new function.
13664 (pass_outof_cfg_layout_mode): Replace declaration with that of...
13665 (make_pass_outof_cfg_layout_mode): ...new function.
13666 (pass_loop2): Replace declaration with that of...
13667 (make_pass_loop2): ...new function.
13668 (pass_rtl_loop_init): Replace declaration with that of...
13669 (make_pass_rtl_loop_init): ...new function.
13670 (pass_rtl_move_loop_invariants): Replace declaration with that of...
13671 (make_pass_rtl_move_loop_invariants): ...new function.
13672 (pass_rtl_unswitch): Replace declaration with that of...
13673 (make_pass_rtl_unswitch): ...new function.
13674 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
13675 (make_pass_rtl_unroll_and_peel_loops): ...new function.
13676 (pass_rtl_doloop): Replace declaration with that of...
13677 (make_pass_rtl_doloop): ...new function.
13678 (pass_rtl_loop_done): Replace declaration with that of...
13679 (make_pass_rtl_loop_done): ...new function.
13680 (pass_web): Replace declaration with that of...
13681 (make_pass_web): ...new function.
13682 (pass_cse2): Replace declaration with that of...
13683 (make_pass_cse2): ...new function.
13684 (pass_df_initialize_opt): Replace declaration with that of...
13685 (make_pass_df_initialize_opt): ...new function.
13686 (pass_df_initialize_no_opt): Replace declaration with that of...
13687 (make_pass_df_initialize_no_opt): ...new function.
13688 (pass_reginfo_init): Replace declaration with that of...
13689 (make_pass_reginfo_init): ...new function.
13690 (pass_inc_dec): Replace declaration with that of...
13691 (make_pass_inc_dec): ...new function.
13692 (pass_stack_ptr_mod): Replace declaration with that of...
13693 (make_pass_stack_ptr_mod): ...new function.
13694 (pass_initialize_regs): Replace declaration with that of...
13695 (make_pass_initialize_regs): ...new function.
13696 (pass_combine): Replace declaration with that of...
13697 (make_pass_combine): ...new function.
13698 (pass_if_after_combine): Replace declaration with that of...
13699 (make_pass_if_after_combine): ...new function.
13700 (pass_ree): Replace declaration with that of...
13701 (make_pass_ree): ...new function.
13702 (pass_partition_blocks): Replace declaration with that of...
13703 (make_pass_partition_blocks): ...new function.
13704 (pass_match_asm_constraints): Replace declaration with that of...
13705 (make_pass_match_asm_constraints): ...new function.
13706 (pass_regmove): Replace declaration with that of...
13707 (make_pass_regmove): ...new function.
13708 (pass_split_all_insns): Replace declaration with that of...
13709 (make_pass_split_all_insns): ...new function.
13710 (pass_fast_rtl_byte_dce): Replace declaration with that of...
13711 (make_pass_fast_rtl_byte_dce): ...new function.
13712 (pass_lower_subreg2): Replace declaration with that of...
13713 (make_pass_lower_subreg2): ...new function.
13714 (pass_mode_switching): Replace declaration with that of...
13715 (make_pass_mode_switching): ...new function.
13716 (pass_sms): Replace declaration with that of...
13717 (make_pass_sms): ...new function.
13718 (pass_sched): Replace declaration with that of...
13719 (make_pass_sched): ...new function.
13720 (pass_ira): Replace declaration with that of...
13721 (make_pass_ira): ...new function.
13722 (pass_reload): Replace declaration with that of...
13723 (make_pass_reload): ...new function.
13724 (pass_clean_state): Replace declaration with that of...
13725 (make_pass_clean_state): ...new function.
13726 (pass_branch_prob): Replace declaration with that of...
13727 (make_pass_branch_prob): ...new function.
13728 (pass_value_profile_transformations): Replace declaration with that
13730 (make_pass_value_profile_transformations): ...new function.
13731 (pass_postreload_cse): Replace declaration with that of...
13732 (make_pass_postreload_cse): ...new function.
13733 (pass_gcse2): Replace declaration with that of...
13734 (make_pass_gcse2): ...new function.
13735 (pass_split_after_reload): Replace declaration with that of...
13736 (make_pass_split_after_reload): ...new function.
13737 (pass_branch_target_load_optimize1): Replace declaration with that
13739 (make_pass_branch_target_load_optimize1): ...new function.
13740 (pass_thread_prologue_and_epilogue): Replace declaration with that
13742 (make_pass_thread_prologue_and_epilogue): ...new function.
13743 (pass_stack_adjustments): Replace declaration with that of...
13744 (make_pass_stack_adjustments): ...new function.
13745 (pass_peephole2): Replace declaration with that of...
13746 (make_pass_peephole2): ...new function.
13747 (pass_if_after_reload): Replace declaration with that of...
13748 (make_pass_if_after_reload): ...new function.
13749 (pass_regrename): Replace declaration with that of...
13750 (make_pass_regrename): ...new function.
13751 (pass_cprop_hardreg): Replace declaration with that of...
13752 (make_pass_cprop_hardreg): ...new function.
13753 (pass_reorder_blocks): Replace declaration with that of...
13754 (make_pass_reorder_blocks): ...new function.
13755 (pass_branch_target_load_optimize2): Replace declaration with that
13757 (make_pass_branch_target_load_optimize2): ...new function.
13758 (pass_leaf_regs): Replace declaration with that of...
13759 (make_pass_leaf_regs): ...new function.
13760 (pass_split_before_sched2): Replace declaration with that of...
13761 (make_pass_split_before_sched2): ...new function.
13762 (pass_compare_elim_after_reload): Replace declaration with that of...
13763 (make_pass_compare_elim_after_reload): ...new function.
13764 (pass_sched2): Replace declaration with that of...
13765 (make_pass_sched2): ...new function.
13766 (pass_stack_regs): Replace declaration with that of...
13767 (make_pass_stack_regs): ...new function.
13768 (pass_stack_regs_run): Replace declaration with that of...
13769 (make_pass_stack_regs_run): ...new function.
13770 (pass_df_finish): Replace declaration with that of...
13771 (make_pass_df_finish): ...new function.
13772 (pass_compute_alignments): Replace declaration with that of...
13773 (make_pass_compute_alignments): ...new function.
13774 (pass_duplicate_computed_gotos): Replace declaration with that of...
13775 (make_pass_duplicate_computed_gotos): ...new function.
13776 (pass_variable_tracking): Replace declaration with that of...
13777 (make_pass_variable_tracking): ...new function.
13778 (pass_free_cfg): Replace declaration with that of...
13779 (make_pass_free_cfg): ...new function.
13780 (pass_machine_reorg): Replace declaration with that of...
13781 (make_pass_machine_reorg): ...new function.
13782 (pass_cleanup_barriers): Replace declaration with that of...
13783 (make_pass_cleanup_barriers): ...new function.
13784 (pass_delay_slots): Replace declaration with that of...
13785 (make_pass_delay_slots): ...new function.
13786 (pass_split_for_shorten_branches): Replace declaration with that of...
13787 (make_pass_split_for_shorten_branches): ...new function.
13788 (pass_split_before_regstack): Replace declaration with that of...
13789 (make_pass_split_before_regstack): ...new function.
13790 (pass_convert_to_eh_region_ranges): Replace declaration with that
13792 (make_pass_convert_to_eh_region_ranges): ...new function.
13793 (pass_shorten_branches): Replace declaration with that of...
13794 (make_pass_shorten_branches): ...new function.
13795 (pass_set_nothrow_function_flags): Replace declaration with that of...
13796 (make_pass_set_nothrow_function_flags): ...new function.
13797 (pass_dwarf2_frame): Replace declaration with that of...
13798 (make_pass_dwarf2_frame): ...new function.
13799 (pass_final): Replace declaration with that of...
13800 (make_pass_final): ...new function.
13801 (pass_rtl_seqabstr): Replace declaration with that of...
13802 (make_pass_rtl_seqabstr): ...new function.
13803 (pass_release_ssa_names): Replace declaration with that of...
13804 (make_pass_release_ssa_names): ...new function.
13805 (pass_early_inline): Replace declaration with that of...
13806 (make_pass_early_inline): ...new function.
13807 (pass_inline_parameters): Replace declaration with that of...
13808 (make_pass_inline_parameters): ...new function.
13809 (pass_update_address_taken): Replace declaration with that of...
13810 (make_pass_update_address_taken): ...new function.
13811 (pass_convert_switch): Replace declaration with that of...
13812 (make_pass_convert_switch): ...new function.
13813 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
13814 to a subclass of simple_ipa_opt_pass along with...
13815 (pass_data_ipa_tree_profile): ...new pass_data instance and...
13816 (make_pass_ipa_tree_profile): ...new function.
13817 * tree-sra.c (pass_sra_early): Convert from a global struct to a
13818 subclass of gimple_opt_pass along with...
13819 (pass_data_sra_early): ...new pass_data instance and...
13820 (make_pass_sra_early): ...new function.
13821 (pass_sra): Convert from a global struct to a subclass of
13822 gimple_opt_pass along with...
13823 (pass_data_sra): ...new pass_data instance and...
13824 (make_pass_sra): ...new function.
13825 (pass_early_ipa_sra): Convert from a global struct to a subclass of
13826 gimple_opt_pass along with...
13827 (pass_data_early_ipa_sra): ...new pass_data instance and...
13828 (make_pass_early_ipa_sra): ...new function.
13829 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
13830 subclass of gimple_opt_pass along with...
13831 (pass_data_ccp): ...new pass_data instance and...
13832 (make_pass_ccp): ...new function.
13833 (pass_fold_builtins): Convert from a global struct to a subclass of
13834 gimple_opt_pass along with...
13835 (pass_data_fold_builtins): ...new pass_data instance and...
13836 (make_pass_fold_builtins): ...new function.
13837 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
13838 subclass of gimple_opt_pass along with...
13839 (pass_data_copy_prop): ...new pass_data instance and...
13840 (make_pass_copy_prop): ...new function.
13841 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
13842 global struct to a subclass of gimple_opt_pass along with...
13843 (pass_data_rename_ssa_copies): ...new pass_data instance and...
13844 (make_pass_rename_ssa_copies): ...new function.
13845 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
13846 subclass of gimple_opt_pass along with...
13847 (pass_data_dce): ...new pass_data instance and...
13848 (make_pass_dce): ...new function.
13849 (pass_dce_loop): Convert from a global struct to a subclass of
13850 gimple_opt_pass along with...
13851 (pass_data_dce_loop): ...new pass_data instance and...
13852 (make_pass_dce_loop): ...new function.
13853 (pass_cd_dce): Convert from a global struct to a subclass of
13854 gimple_opt_pass along with...
13855 (pass_data_cd_dce): ...new pass_data instance and...
13856 (make_pass_cd_dce): ...new function.
13857 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
13858 subclass of gimple_opt_pass along with...
13859 (pass_data_dominator): ...new pass_data instance and...
13860 (make_pass_dominator): ...new function.
13861 (pass_phi_only_cprop): Convert from a global struct to a subclass of
13862 gimple_opt_pass along with...
13863 (pass_data_phi_only_cprop): ...new pass_data instance and...
13864 (make_pass_phi_only_cprop): ...new function.
13865 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
13866 subclass of gimple_opt_pass along with...
13867 (pass_data_dse): ...new pass_data instance and...
13868 (make_pass_dse): ...new function.
13869 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
13870 a subclass of gimple_opt_pass along with...
13871 (pass_data_forwprop): ...new pass_data instance and...
13872 (make_pass_forwprop): ...new function.
13873 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
13874 struct to a subclass of gimple_opt_pass along with...
13875 (pass_data_tree_ifcombine): ...new pass_data instance and...
13876 (make_pass_tree_ifcombine): ...new function.
13877 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
13878 subclass of gimple_opt_pass along with...
13879 (pass_data_ch): ...new pass_data instance and...
13880 (make_pass_ch): ...new function.
13881 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
13882 subclass of gimple_opt_pass along with...
13883 (pass_data_tree_loop): ...new pass_data instance and...
13884 (make_pass_tree_loop): ...new function.
13885 (pass_tree_loop_init): Convert from a global struct to a subclass of
13886 gimple_opt_pass along with...
13887 (pass_data_tree_loop_init): ...new pass_data instance and...
13888 (make_pass_tree_loop_init): ...new function.
13889 (pass_lim): Convert from a global struct to a subclass of
13890 gimple_opt_pass along with...
13891 (pass_data_lim): ...new pass_data instance and...
13892 (make_pass_lim): ...new function.
13893 (pass_tree_unswitch): Convert from a global struct to a subclass of
13894 gimple_opt_pass along with...
13895 (pass_data_tree_unswitch): ...new pass_data instance and...
13896 (make_pass_tree_unswitch): ...new function.
13897 (pass_predcom): Convert from a global struct to a subclass of
13898 gimple_opt_pass along with...
13899 (pass_data_predcom): ...new pass_data instance and...
13900 (make_pass_predcom): ...new function.
13901 (pass_vectorize): Convert from a global struct to a subclass of
13902 gimple_opt_pass along with...
13903 (pass_data_vectorize): ...new pass_data instance and...
13904 (make_pass_vectorize): ...new function.
13905 (pass_graphite): Convert from a global struct to a subclass of
13906 gimple_opt_pass along with...
13907 (pass_data_graphite): ...new pass_data instance and...
13908 (make_pass_graphite): ...new function.
13909 (pass_graphite_transforms): Convert from a global struct to a subclass
13910 of gimple_opt_pass along with...
13911 (pass_data_graphite_transforms): ...new pass_data instance and...
13912 (make_pass_graphite_transforms): ...new function.
13913 (pass_check_data_deps): Convert from a global struct to a subclass of
13914 gimple_opt_pass along with...
13915 (pass_data_check_data_deps): ...new pass_data instance and...
13916 (make_pass_check_data_deps): ...new function.
13917 (pass_iv_canon): Convert from a global struct to a subclass of
13918 gimple_opt_pass along with...
13919 (pass_data_iv_canon): ...new pass_data instance and...
13920 (make_pass_iv_canon): ...new function.
13921 (pass_scev_cprop): Convert from a global struct to a subclass of
13922 gimple_opt_pass along with...
13923 (pass_data_scev_cprop): ...new pass_data instance and...
13924 (make_pass_scev_cprop): ...new function.
13925 (pass_record_bounds): Convert from a global struct to a subclass of
13926 gimple_opt_pass along with...
13927 (pass_data_record_bounds): ...new pass_data instance and...
13928 (make_pass_record_bounds): ...new function.
13929 (pass_complete_unroll): Convert from a global struct to a subclass of
13930 gimple_opt_pass along with...
13931 (pass_data_complete_unroll): ...new pass_data instance and...
13932 (make_pass_complete_unroll): ...new function.
13933 (pass_complete_unrolli): Convert from a global struct to a subclass of
13934 gimple_opt_pass along with...
13935 (pass_data_complete_unrolli): ...new pass_data instance and...
13936 (make_pass_complete_unrolli): ...new function.
13937 (pass_parallelize_loops): Convert from a global struct to a subclass
13938 of gimple_opt_pass along with...
13939 (pass_data_parallelize_loops): ...new pass_data instance and...
13940 (make_pass_parallelize_loops): ...new function.
13941 (pass_loop_prefetch): Convert from a global struct to a subclass of
13942 gimple_opt_pass along with...
13943 (pass_data_loop_prefetch): ...new pass_data instance and...
13944 (make_pass_loop_prefetch): ...new function.
13945 (pass_iv_optimize): Convert from a global struct to a subclass of
13946 gimple_opt_pass along with...
13947 (pass_data_iv_optimize): ...new pass_data instance and...
13948 (make_pass_iv_optimize): ...new function.
13949 (pass_tree_loop_done): Convert from a global struct to a subclass of
13950 gimple_opt_pass along with...
13951 (pass_data_tree_loop_done): ...new pass_data instance and...
13952 (make_pass_tree_loop_done): ...new function.
13953 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
13954 struct to a subclass of gimple_opt_pass along with...
13955 (pass_data_cse_reciprocals): ...new pass_data instance and...
13956 (make_pass_cse_reciprocals): ...new function.
13957 (pass_cse_sincos): Convert from a global struct to a subclass of
13958 gimple_opt_pass along with...
13959 (pass_data_cse_sincos): ...new pass_data instance and...
13960 (make_pass_cse_sincos): ...new function.
13961 (pass_optimize_bswap): Convert from a global struct to a subclass of
13962 gimple_opt_pass along with...
13963 (pass_data_optimize_bswap): ...new pass_data instance and...
13964 (make_pass_optimize_bswap): ...new function.
13965 (pass_optimize_widening_mul): Convert from a global struct to a
13966 subclass of gimple_opt_pass along with...
13967 (pass_data_optimize_widening_mul): ...new pass_data instance and...
13968 (make_pass_optimize_widening_mul): ...new function.
13969 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
13970 subclass of gimple_opt_pass along with...
13971 (pass_data_phiopt): ...new pass_data instance and...
13972 (make_pass_phiopt): ...new function.
13973 (pass_cselim): Convert from a global struct to a subclass of
13974 gimple_opt_pass along with...
13975 (pass_data_cselim): ...new pass_data instance and...
13976 (make_pass_cselim): ...new function.
13977 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
13978 subclass of gimple_opt_pass along with...
13979 (pass_data_phiprop): ...new pass_data instance and...
13980 (make_pass_phiprop): ...new function.
13981 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
13982 subclass of gimple_opt_pass along with...
13983 (pass_data_pre): ...new pass_data instance and...
13984 (make_pass_pre): ...new function.
13985 (pass_fre): Convert from a global struct to a subclass of
13986 gimple_opt_pass along with...
13987 (pass_data_fre): ...new pass_data instance and...
13988 (make_pass_fre): ...new function.
13989 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
13990 subclass of gimple_opt_pass along with...
13991 (pass_data_reassoc): ...new pass_data instance and...
13992 (make_pass_reassoc): ...new function.
13993 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
13994 subclass of gimple_opt_pass along with...
13995 (pass_data_sink_code): ...new pass_data instance and...
13996 (make_pass_sink_code): ...new function.
13997 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
13998 subclass of gimple_opt_pass along with...
13999 (pass_data_strlen): ...new pass_data instance and...
14000 (make_pass_strlen): ...new function.
14001 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
14002 struct to a subclass of gimple_opt_pass along with...
14003 (pass_data_build_alias): ...new pass_data instance and...
14004 (make_pass_build_alias): ...new function.
14005 (pass_build_ealias): Convert from a global struct to a subclass of
14006 gimple_opt_pass along with...
14007 (pass_data_build_ealias): ...new pass_data instance and...
14008 (make_pass_build_ealias): ...new function.
14009 (pass_ipa_pta): Convert from a global struct to a subclass of
14010 simple_ipa_opt_pass along with...
14011 (pass_data_ipa_pta): ...new pass_data instance and...
14012 (make_pass_ipa_pta): ...new function.
14013 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
14014 subclass of gimple_opt_pass along with...
14015 (pass_data_uncprop): ...new pass_data instance and...
14016 (make_pass_uncprop): ...new function.
14017 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
14018 global struct to a subclass of gimple_opt_pass along with...
14019 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
14020 (make_pass_late_warn_uninitialized): ...new function.
14021 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
14022 to a subclass of gimple_opt_pass along with...
14023 (pass_data_init_datastructures): ...new pass_data instance and...
14024 (make_pass_init_datastructures): ...new function.
14025 (pass_early_warn_uninitialized): Convert from a global struct to a
14026 subclass of gimple_opt_pass along with...
14027 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
14028 (make_pass_early_warn_uninitialized): ...new function.
14029 (pass_update_address_taken): Convert from a global struct to a
14030 subclass of gimple_opt_pass along with...
14031 (pass_data_update_address_taken): ...new pass_data instance and...
14032 (make_pass_update_address_taken): ...new function.
14033 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
14034 struct to a subclass of gimple_opt_pass along with...
14035 (pass_data_release_ssa_names): ...new pass_data instance and...
14036 (make_pass_release_ssa_names): ...new function.
14037 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
14038 subclass of gimple_opt_pass along with...
14039 (pass_data_stdarg): ...new pass_data instance and...
14040 (make_pass_stdarg): ...new function.
14041 * tree-switch-conversion.c (pass_convert_switch): Convert from a
14042 global struct to a subclass of gimple_opt_pass along with...
14043 (pass_data_convert_switch): ...new pass_data instance and...
14044 (make_pass_convert_switch): ...new function.
14045 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
14046 to a subclass of gimple_opt_pass along with...
14047 (pass_data_tail_recursion): ...new pass_data instance and...
14048 (make_pass_tail_recursion): ...new function.
14049 (pass_tail_calls): Convert from a global struct to a subclass of
14050 gimple_opt_pass along with...
14051 (pass_data_tail_calls): ...new pass_data instance and...
14052 (make_pass_tail_calls): ...new function.
14053 * tree-vect-generic.c (pass_lower_vector): Convert from a global
14054 struct to a subclass of gimple_opt_pass along with...
14055 (pass_data_lower_vector): ...new pass_data instance and...
14056 (make_pass_lower_vector): ...new function.
14057 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
14058 gimple_opt_pass along with...
14059 (pass_data_lower_vector_ssa): ...new pass_data instance and...
14060 (make_pass_lower_vector_ssa): ...new function.
14061 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
14062 to a subclass of gimple_opt_pass along with...
14063 (pass_data_slp_vectorize): ...new pass_data instance and...
14064 (make_pass_slp_vectorize): ...new function.
14065 (pass_ipa_increase_alignment): Convert from a global struct to a
14066 subclass of simple_ipa_opt_pass along with...
14067 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
14068 (make_pass_ipa_increase_alignment): ...new function.
14069 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
14070 gimple_opt_pass along with...
14071 (pass_data_vrp): ...new pass_data instance and...
14072 (make_pass_vrp): ...new function.
14073 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
14074 subclass of simple_ipa_opt_pass along with...
14075 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
14076 (make_pass_ipa_free_lang_data): ...new function.
14077 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
14078 gimple_opt_pass along with...
14079 (pass_data_tsan): ...new pass_data instance and...
14080 (make_pass_tsan): ...new function.
14081 (pass_tsan_O0): Convert from a global struct to a subclass of
14082 gimple_opt_pass along with...
14083 (pass_data_tsan_O0): ...new pass_data instance and...
14084 (make_pass_tsan_O0): ...new function.
14085 * var-tracking.c (pass_variable_tracking): Convert from a global
14086 struct to a subclass of rtl_opt_pass along with...
14087 (pass_data_variable_tracking): ...new pass_data instance and...
14088 (make_pass_variable_tracking): ...new function.
14089 * web.c (pass_web): Convert from a global struct to a subclass of
14090 rtl_opt_pass along with...
14091 (pass_data_web): ...new pass_data instance and...
14092 (make_pass_web): ...new function.
14093 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
14094 declaration with that of...
14095 (make_pass_mode_switch_use): ...new function.
14096 (pass_resolve_sw_modes): Replace declaration with that of...
14097 (make_pass_resolve_sw_modes): ...new function.
14098 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
14099 from a global struct to a subclass of rtl_opt_pass along with...
14100 (pass_data_mode_switch_use): ...new pass_data instance and...
14101 (make_pass_mode_switch_use): ...new function.
14102 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
14103 from a global struct to a subclass of rtl_opt_pass along with...
14104 (pass_data_resolve_sw_modes): ...new pass_data instance and...
14105 (make_pass_resolve_sw_modes): ...new function.
14106 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
14107 struct to a subclass of rtl_opt_pass along with...
14108 (pass_data_insert_vzeroupper): ...new pass_data instance and...
14109 (make_pass_insert_vzeroupper): ...new function.
14110 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
14111 global struct to a subclass of rtl_opt_pass along with...
14112 (pass_data_work_around_errata): ...new pass_data instance and...
14113 (make_pass_work_around_errata): ...new function.
14114 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
14115 struct to a subclass of rtl_opt_pass along with...
14116 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
14117 (make_pass_mips_machine_reorg2): ...new function.
14119 2013-08-05 David Malcolm <dmalcolm@redhat.com>
14121 * passes.c (pass_manager::operator new): New.
14123 2013-08-05 David Malcolm <dmalcolm@redhat.com>
14125 Handwritten part of conversion of passes to C++ classes.
14127 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
14128 (toplev.o): Add dep on PASS_MANAGER_H.
14129 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
14130 of early local pases to reflect this moving from a global to a
14131 member of gcc::pass_manager.
14132 (cgraph_add_new_function): Likewise.
14133 * lto-cgraph.c (lto_output_node): Update for conversion of
14134 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
14135 * passes.c (opt_pass::clone): New.
14136 (opt_pass::gate): New.
14137 (opt_pass::execute): New.
14138 (opt_pass::opt_pass): New.
14139 (pass_manager::execute_early_local_passes): New.
14140 (pass_manager::execute_pass_mode_switching): new.
14141 (finish_optimization_passes): Convert to...
14142 (pass_manager::finish_optimization_passes): ...this.
14143 (finish_optimization_passes): Update for conversion of passes to
14145 (register_dump_files_1): Use has_gate since we cannot portably
14146 check a vtable entry against NULL.
14147 (dump_one_pass): Likewise.
14148 (ipa_write_summaries_2): Likewise.
14149 (ipa_write_optimization_summaries_1): Likewise.
14150 (ipa_read_summaries_1): Likewise.
14151 (ipa_read_optimization_summaries_1): Likewise.
14152 (execute_ipa_stmt_fixups): Likewise.
14153 (pass_manager::pass_manager): Rewrite pass-creation, invoking
14154 pass-creation functions rather than wiring up globals, and
14155 storing the results in fields of pass_manager generated using
14156 pass-instances.def.
14157 (pass_manager::dump_profile_report): Update for conversion of
14158 passes to C++ classes.
14159 (pass_manager::execute_ipa_summary_passes): Likewise.
14160 (execute_one_ipa_transform_pass): Likewise.
14161 (execute_one_pass): Use has_gate and has_execute since we cannot
14162 portably check a vtable entry against NULL.
14163 * pass_manager.h (pass_manager::finish_optimization_passes): New.
14164 (pass_manager): Use pass-instances.def to add fields for the
14165 various pass instances.
14166 * toplev.c (finalize): Update for move of
14167 finish_optimization_passes to a method of gcc::pass_manager.
14168 * toplev.h (finish_optimization_passes): Move to method of class
14170 * tree-pass.h (struct pass_data): New.
14171 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
14172 (opt_pass::gate): Convert to virtual function.
14173 (opt_pass::~opt_pass): New.
14174 (opt_pass::clone): New.
14175 (opt_pass::execute): Convert to virtual function.
14176 (opt_pass::opt_pass): New.
14177 (opt_pass::ctxt_): new.
14178 (gimple_opt_pass): Convert to subclass of opt_pass.
14179 (gimple_opt_pass::gimple_opt_pass): New.
14180 (rtl_opt_pass): Convert to subclass of opt_pass.
14181 (rtl_opt_pass::rtl_opt_pass): New.
14182 (ipa_opt_pass_d): Convert to subclass of opt_pass.
14183 (ipa_opt_pass_d::ipa_opt_pass_d): New.
14184 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
14185 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
14186 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
14187 invocation of pass_mode_switching to reflect this moving from a
14188 global to a member of gcc::pass_manager.
14189 (ix86_option_override): Rework how pass_insert_vzeroupper is
14190 added to the pass_manager to reflect autogenerated changes.
14191 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
14193 2013-08-05 Richard Earnshaw <rearnsha@arm.com>
14195 PR rtl-optimization/57708
14196 * recog.c (peep2_find_free_register): Validate all regs in a
14199 2013-08-05 Jan Hubicka <jh@suse.cz>
14202 * cgraph.c (verify_cgraph_node): Accept local flags from other
14204 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
14205 (function_and_variable_visibility): Likewise.
14206 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
14208 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
14210 * graph.c (init_graph_slim_pretty_print): Remove.
14211 (print_graph_cfg): Do not call it. Use local pretty printer.
14212 (start_graph_dump): Likewise.
14214 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
14216 * gimple-pretty-print.c (buffer): Remove.
14217 (initialized): Likewise.
14218 (maybe_init_pretty_print): Likewise.
14219 (print_gimple_stmt): Do not call it. Use non-static local
14220 pretty_printer variable.
14221 (print_gimple_expr): Likewise.
14222 (print_gimple_seq): Likewise.
14223 (gimple_dump_bb): Likewise.
14225 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
14227 * asan.c (asan_pp): Remove.
14228 (asan_pp_initialized): Likewise.
14229 (asan_pp_initialize): Likewise.
14230 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
14231 (asan_emit_stack_protection): Tidy. Use local pretty printer.
14232 (asan_add_global): Likewise.
14234 2013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
14236 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
14237 * diagnostic.h (diagnostic_flush_buffer): Adjust.
14238 * pretty-print.c (pp_formatted_text_data): Likewise.
14239 (pp_indent): Rename from pp_base_indent.
14240 (pp_format): Rename from pp_base_format.
14241 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
14242 (pp_format_verbatim): Rename from pp_base_format_verbatim.
14243 (pp_flush): Rename from pp_base_flush.
14244 (pp_set_line_maximum_length): Rename from
14245 pp_base_set_line_maximum_length.
14246 (pp_clear_output_area): Rename from pp_base_clear_output_area.
14247 (pp_set_prefix): Rename from pp_base_set_prefix.
14248 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
14249 (pp_emit_prefix): Rename from pp_base_emit_prefix.
14250 (pp_append_text): Rename from pp_base_append_text.
14251 (pp_formatted_text): Rename from pp_base_formatted_text.
14252 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
14253 (pp_remaining_character_count_for_line): Rename from
14254 pp_base_remaining_character_count_for_line.
14255 (pp_newline): Rename from pp_base_newline.
14256 (pp_character): Rename from pp_base_character.
14257 (pp_string): Rename from pp_base_string.
14258 (pp_maybe_space): Rename from pp_base_maybe_space.
14259 * asan.c (asan_pp_string): Adjust.
14260 (asan_emit_stack_protection): Likewise.
14261 (asan_add_global): Likewise.
14262 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
14263 * tree-mudflap.c (mf_varname_tree): Likewise.
14264 * tree-pretty-print.c (pp_tree_identifier): Rename from
14265 pp_base_tree_identifier.
14266 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
14267 Declare as function.
14269 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
14271 * pretty-print.h (pp_bar_bar): New.
14272 (pp_ampersand_ampersand): Likewise.
14273 (pp_less_equal): Likewise.
14274 (pp_greater_equal): Likewise.
14275 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
14276 printer functions instead of pp_string or operators and punctuators.
14277 (dump_gimple_call): Likewise.
14278 (dump_gimple_omp_for): Likewise.
14279 (dump_gimple_transaction): Likewise.
14280 (dump_gimple_phi): Likewise.
14281 (pp_gimple_stmt_1): Likewise.
14282 * sched-vis.c (print_insn): Likewise.
14283 * tree-mudflap.c (mf_varname_tree): Likewise.
14284 * tree-pretty-print.c (dump_block_node): Likewise.
14285 (dump_generic_node): Likewise.
14287 2013-08-02 Jan Hubicka <jh@suse.cz>
14289 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
14291 * lto-streamer-out.c (tree_is_indexable): Results decls and
14292 parm decls are not indexable.
14293 (DFS_write_tree_body): Do not follow args and results.
14294 (hash_tree): Likewise.
14295 (output_functions): Rearrange so struct function is needed
14296 only when real body is output; be able to also ouptut abstract
14297 functions; output DECL_ARGUMENTS and DECL_RESULT.
14298 (lto_output): When not in WPA, ale store abstract functions.
14299 (write_symbol): Do not care about RESULT_DECL.
14300 (output_symbol_p): Handle correctly sbtract decls.
14301 * lto-streamer-in.c (input_function): Rearrange so struct
14302 function can be NULL at entry; allow streaming of
14303 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
14304 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
14305 sanity check during LTO.
14306 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
14307 RESULT_DECl and DECL_ARGUMENTS.
14308 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
14311 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
14313 * pretty-print.h (pp_underscore): New.
14315 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
14316 printer functions instead of pp_character.
14317 (dump_binary_rhs): Likewise.
14318 (dump_ternary_rhs): Likewise.
14319 (dump_gimple_call_args): Likewise.
14320 (pp_points_to_solution): Likewise.
14321 (dump_gimple_call): Likewise.
14322 (dump_gimple_switch): Likewise.
14323 (dump_gimple_cond): Likewise.
14324 (dump_gimple_bind): Likewise.
14325 (dump_gimple_try): Likewise.
14326 (dump_gimple_omp_for): Likewise.
14327 (dump_gimple_omp_continue): Likewise.
14328 (dump_gimple_omp_single): Likewise.
14329 (dump_gimple_omp_sections): Likewise.
14330 (dump_gimple_omp_block): Likewise.
14331 (dump_gimple_omp_critical): Likewise.
14332 (dump_gimple_transaction): Likewise.
14333 (dump_gimple_asm): Likewise.
14334 (dump_gimple_phi): Likewise.
14335 (dump_gimple_omp_parallel): Likewise.
14336 (dump_gimple_omp_task): Likewise.
14337 (dump_gimple_omp_atomic_load): Likewise.
14338 (dump_gimple_omp_atomic_store): Likewise.
14339 (dump_gimple_mem_ops): Likewise.
14340 (pp_gimple_stmt_1): Likewise.
14341 (pp_cfg_jump): Likewise.
14342 (dump_implicit_edges): Likewise.
14343 (gimple_dump_bb_for_graph): Likewise.
14344 * graph.c (draw_cfg_node): Likewise.
14345 * langhooks.c (lhd_print_error_function): Likewise.
14346 * sched-vis.c (print_exp): Likewise.
14347 (print_value): Likewise.
14348 (print_pattern): Likewise.
14349 (print_insn): Likewise.
14350 (rtl_dump_bb_for_graph): Likewise.
14351 * tree-pretty-print.c (dump_function_declaration): Likewise.
14352 (dump_array_domain): Likewise.
14353 (dump_omp_clause): Likewise.
14354 (dump_location): Likewise.
14355 (dump_generic_node): Likewise.
14356 (print_struct_decl): Likewise.
14357 * diagnostic.c (diagnostic_show_locus): Use pp_space.
14359 2013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
14361 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
14362 candidate table when replacing a candidate statement.
14363 (replace_rhs_if_not_dup): Likewise.
14364 (replace_one_candidate): Likewise.
14366 2013-08-02 Jan Hubicka <jh@suse.cz>
14367 Martin Liska <marxin.liska@gmail.com>
14369 * cgraphunit.c (add_new_function): Fix logic when adding from
14371 (assemble_thunk): Rename to ...
14372 (expand_thunk); .. this one; export; get it working with
14373 general functions; make produced gimple valid.
14374 * cgraph.h (expand_thunk): Declare.
14376 2013-08-02 Jan Hubicka <jh@suse.cz>
14378 * ipa-cp.c (gather_context_independent_values): Use
14379 ipa_get_param_move_cost.
14380 (get_replacement_map): Remove PARAM; move parameter folding
14382 (create_specialized_node): Update.
14383 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
14384 assert that we have gimple body; update move_cost.
14385 (count_formal_params): Assert that we have gimple body.
14386 (ipa_dump_param): New function.
14387 (ipa_alloc_node_params): Break out from ...
14388 (ipa_initialize_node_params): ... here.
14389 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
14390 (ipa_write_node_info): Stream move costs.
14391 (ipa_read_node_info): Read move costs.
14392 (ipa_update_after_lto_read): Do not recompute node params.
14393 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
14394 (ipa_get_param): Check we are not in WPA.
14395 (ipa_get_param_move_cost): New.
14396 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
14397 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
14398 parm numbers to be present.
14400 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
14402 PR rtl-optimization/58048
14403 * lra-constraints.c (process_alt_operands): Don't check asm
14404 operand on register.
14406 2013-08-02 Eric Botcazou <ebotcazou@adacore.com>
14408 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
14409 the implied StoreLoad barrier for atomic operations if before.
14411 2013-08-02 Jan Hubicka <jh@suse.cz>
14412 Martin Liska <marxin.liska@gmail.com>
14414 * cgraph.c (cgraph_function_body_availability): Do not check
14416 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
14417 symtab_node_availability): Declare.
14418 * ipa.c (can_replace_by_local_alias): New.
14419 (function_and_variable_visibility): Use it.
14420 * symtab.c (symtab_for_node_and_aliases,
14421 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
14423 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
14425 PR rtl-optimization/57963
14426 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
14427 (lra_constraints): Use them.
14429 2013-08-02 Sofiane Naci <sofiane.naci@arm.com>
14431 * config/arm/types.md (define_attr "type"): Add "load_acq"
14433 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
14435 (cortex_a53_store1): Likewise.
14437 2013-08-01 Jan Hubicka <jh@suse.cz>
14439 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
14440 partitions are not needed.
14442 2013-08-01 Uros Bizjak <ubizjak@gmail.com>
14444 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
14445 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
14446 MAYBE_NON_Q_CLASS_P where appropriate.
14448 2013-08-01 Jan Hubicka <jh@suse.cz>
14450 * cgraph.h (release_function_body): Declare.
14451 * tree.c (free_lang_data_in_decl): Free, parameters and return values
14452 of unused delcarations.
14454 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14456 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
14457 RTL form when subtracting a constant.
14459 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14461 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
14462 Generate canonical plus rtx with negated immediate instead of minus
14464 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
14466 2013-08-01 Jan Hubicka <jh@suse.cz>
14468 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
14469 (cgraph_release_function_body): Likewise.
14470 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
14471 * cgraph.h (cgrpah_node): Rename abstract_and_needed
14472 to used_as_abstract_origin.
14473 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
14474 symbols used as abstract origins.
14475 * cgraphunit.c (analyze_functions): Update.
14476 * ipa.c (symtab_remove_unreachable_nodes): Recompute
14477 used_as_abstract_origin.
14478 * tree-inline.c (tree_function_versioning): Update
14479 used_as_abstract_origin; be ready for DECL_RESULT and
14480 DECL_ARGUMENTS to be NULL.
14482 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
14483 for abstract functions.
14484 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
14487 2013-08-01 Jan Hubicka <jh@suse.cz>
14489 * profile.c (compute_value_histograms): Fix thinko.
14491 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
14493 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
14494 aarch-common-protos.h to extra_headers.
14495 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
14496 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
14497 * config/aarch64/t-aarch64 (aarch-common.o): Define.
14499 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
14501 * config/aarch64/aarch64.md (define_attr "type"): Delete.
14502 Include "../arm/types.md". Define "type" attribute for all patterns.
14503 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
14506 2013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
14508 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
14509 to support power8 load fusion.
14510 (fusion_gpr_mem_load): Likewise.
14512 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
14514 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
14515 declarations for power8 load fusion.
14516 (emit_fusion_gpr_load): Likewise.
14518 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
14519 tuning for power8, turn on fusion mode by default. Turn on sign
14520 extending fusion mode if normal fusion mode is on, and we are at
14522 (fusion_gpr_load_p): New function, return true if we can fuse an
14523 addis instruction with a dependent load to a GPR.
14524 (emit_fusion_gpr_load): Emit the instructions for power8 load
14527 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
14528 (VSX load fusion peepholes): New peepholes to fuse together an
14529 addi instruction with a VSX load instruction.
14531 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
14532 peepholes to fuse an addis instruction with a load to a GPR base
14533 register. If we are supporting sign extending fusions, convert
14534 sign extending loads to zero extending loads and add an explicit
14537 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
14539 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
14540 aarch-common-protos.h to extra_headers.
14541 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
14542 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
14543 (arm_early_store_addr_dep): Likewise.
14544 (arm_no_early_alu_shift_dep): Likewise.
14545 (arm_no_early_alu_shift_value_dep): Likewise.
14546 (arm_no_early_mul_dep): Likewise.
14547 (arm_no_early_store_addr_dep): Likewise.
14548 (arm_mac_accumulator_is_mul_result): Likewise.
14549 (arm_mac_accumulator_is_result): Likewise.
14550 * config/arm/aarch-common.c: ... here. New file.
14551 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
14553 (arm_early_store_addr_dep): Likewise.
14554 (arm_no_early_alu_shift_dep): Likewise.
14555 (arm_no_early_alu_shift_value_dep): Likewise.
14556 (arm_no_early_mul_dep): Likewise.
14557 (arm_no_early_store_addr_dep): Likewise.
14558 (arm_mac_accumulator_is_mul_result): Likewise.
14559 (arm_mac_accumulator_is_result): Likewise.
14560 * config/arm/aarch-common-protos.h: ... here. New file.
14561 * config/arm/t-arm (aarch-common.o): Define.
14563 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
14565 * config/arm/arm.md: Include new file "types.md".
14566 (define_attr "type"): Move from here to ...
14567 (define_attr "mul32"): Likewise.
14568 (define_attr "mul64"): Likewise.
14569 * config/arm/types.md: ... here. New file.
14571 2013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
14573 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
14574 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
14576 2013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
14578 * gen-pass-instances.awk: Fix offset of substr().
14580 2013-07-31 David Malcolm <dmalcolm@redhat.com>
14582 * Makefile.in (pass-instances.def): New.
14583 (passes.o): Replace dependency on passes.def with one on
14586 * gen-pass-instances.awk: New.
14588 * passes.c (pass_manager::pass_manager): Use pass-instances.def
14589 rather than passes.def, updating local definition of NEXT_PASS
14590 macro to add an extra NUM parameter (currently unused).
14592 2013-07-30 David Malcolm <dmalcolm@redhat.com>
14594 * Makefile.in (PASS_MANAGER_H): New.
14595 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
14596 (passes.o): Likewise.
14597 (statistics.o): Likewise.
14598 (cgraphunit.o): Likewise.
14599 (context.o): Depend on PASS_MANAGER_H.
14601 * pass_manager.h: New.
14603 * cgraphunit.c (cgraph_add_new_function): Update for moves
14604 of globals to fields of pass_manager.
14605 (analyze_function): Likewise.
14606 (expand_function): Likewise.
14607 (ipa_passes): Likewise.
14608 (compile): Likewise.
14610 * context.c (context::context): New.
14611 * context.h (context::context): New.
14612 (context::get_passes): New.
14613 (context::passes_): New.
14615 * lto-cgraph.c (input_node): Update for moves of globals to
14616 fields of pass_manager.
14618 * passes.c (all_passes): Remove, in favor of a field of the
14619 same name within the new class pass_manager.
14620 (all_small_ipa_passes): Likewise.
14621 (all_lowering_passes): Likewise.
14622 (all_regular_ipa_passes): Likewise.
14623 (all_late_ipa_passes): Likewise.
14624 (all_lto_gen_passes): Likewise.
14625 (passes_by_id): Likewise.
14626 (passes_by_id_size): Likewise.
14627 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
14628 the new class pass_manager.
14629 (set_pass_for_id): Convert to...
14630 (pass_manager::set_pass_for_id): ...method.
14631 (get_pass_for_id): Convert to...
14632 (pass_manager::get_pass_for_id): ...method.
14633 (register_one_dump_file): Move body of implementation into...
14634 (pass_manager::register_one_dump_file): ...here.
14635 (register_dump_files_1): Convert to...
14636 (pass_manager::register_dump_files_1): ...method.
14637 (register_dump_files): Convert to...
14638 (pass_manager::register_dump_files): ...method.
14639 (create_pass_tab): Update for moves of globals to fields of
14641 (dump_passes): Move body of implementation into...
14642 (pass_manager::dump_passes): ...here.
14643 (register_pass): Move body of implementation into...
14644 (pass_manager::register_pass): ...here.
14645 (init_optimization_passes): Convert into...
14646 (pass_manager::pass_manager): ...constructor for new
14647 pass_manager class, and initialize the pass_lists array.
14648 (check_profile_consistency): Update for moves of globals to
14649 fields of pass_manager.
14650 (dump_profile_report): Move body of implementation into...
14651 (pass_manager::dump_profile_report): ...here.
14652 (ipa_write_summaries_1): Update for moves of pass lists from
14653 being globals to fields of pass_manager.
14654 (ipa_write_optimization_summaries): Likewise.
14655 (ipa_read_summaries): Likewise.
14656 (ipa_read_optimization_summaries): Likewise.
14657 (execute_all_ipa_stmt_fixups): Likewise.
14659 * statistics.c (statistics_fini): Update for moves of globals to
14660 fields of pass_manager.
14662 * toplev.c (general_init): Replace call to
14663 init_optimization_passes with construction of the pass_manager
14666 * tree-pass.h (all_passes): Remove, in favor of a field of the
14667 same name within the new class pass_manager.
14668 (all_small_ipa_passes): Likewise.
14669 (all_lowering_passes): Likewise.
14670 (all_regular_ipa_passes): Likewise.
14671 (all_lto_gen_passes): Likewise.
14672 (all_late_ipa_passes): Likewise.
14673 (passes_by_id): Likewise.
14674 (passes_by_id_size): Likewise.
14675 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
14676 the new class pass_manager.
14677 (get_pass_for_id): Remove.
14679 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
14681 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
14684 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
14686 * arm.md (mulhi3): New expand pattern.
14688 2013-07-30 Jan Hubicka <jh@suse.cz>
14689 Martin Liska <marxin.liska@gmail.com>
14691 * profile.c (compute_value_histograms): Do not ICE when
14692 there is mismatch only on some counters.
14694 2013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
14696 PR rtl-optimization/57637
14697 * function.c (move_insn_for_shrink_wrap): Also check the
14698 GEN set of the LIVE problem for the liveness analysis
14699 if it exists, otherwise give up.
14701 2013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
14703 PR tree-optimization/57993
14704 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
14705 replaced statement in the candidate table.
14706 (phi_add_costs): Return infinite cost when the hidden basis does
14707 not dominate all phis on which the candidate is dependent.
14708 (replace_one_candidate): Record replaced statement in the
14711 2013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
14713 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
14714 (ashlv2si3): New expander.
14715 (*ashlv2si3_i): New define_insn_and_split.
14716 * predicates.md (float_operation): Allow patterns with three
14717 basic sub-patterns.
14719 PR rtl-optimization/58021
14720 * mode-switching.c (create_pre_exit): Always split off preceding
14721 insns if we are not at the basic block head.
14723 2013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
14725 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
14726 (UCLIBC_DYNAMIC_LINKER): New macro.
14727 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
14729 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
14730 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
14732 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
14733 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
14734 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
14735 for SF and DF modes. Use ieee_quad_format for TF mode.
14736 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
14737 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
14738 (mips_option_override): Handle `-mnan=legacy'.
14739 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
14740 `-mabs=2008' and `-mnan=2008'.
14741 (OPTION_DEFAULT_SPECS): Add "nan" default.
14742 (ASM_SPEC): Handle `-mnan='.
14743 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
14744 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
14745 comment accordingly.
14746 (neg<mode>2): Likewise.
14747 * config/mips/mips.opt (mabs, mnan): New options.
14748 * doc/install.texi (Configuration): Document `--with-nan=' option.
14749 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
14751 (MIPS Options): Document them.
14752 * config.gcc <mips*-*-*>: Handle `--with-nan='.
14753 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
14754 * configure: Regenerate.
14755 * config.in: Regenerate.
14757 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
14759 * config/i386/i386.md (float post-reload splitters): Do not check
14760 for subregs of SSE registers.
14762 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
14763 H.J. Lu <hongjiu.lu@intel.com>
14767 * config/i386/i386.md (post-reload splitter
14768 to avoid partial SSE reg dependency stalls): New pattern.
14770 2013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
14772 * config/s390/s390.md ("movcc"): Swap load and store instructions.
14774 2013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
14776 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
14777 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
14779 2013-07-26 Cary Coutant <ccoutant@google.com>
14781 * dwarf2out.c (die_checksum_ordered): Don't include template
14782 instantiations in signature.
14783 (is_template_parameter): New function.
14784 (is_template_instantiation): New function.
14785 (generate_skeleton_bottom_up): Don't include template instantiations
14787 (generate_skeleton): Likewise.
14788 (break_out_comdat_types): Move recursive call to break out nested
14790 (prune_unused_types_mark_generic_parms_dies): Call
14791 is_template_parameter.
14793 2013-07-26 Ian Bolton <ian.bolton@arm.com>
14795 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
14796 uses vector registers.
14797 * config/aarch64/iterators.md: Add attributes rtn and vas.
14799 2013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14800 Richard Earnshaw <richard.earnshaw@arm.com>
14802 * combine.c (simplify_comparison): Re-canonicalize operands
14804 * config/arm/arm.md (movcond_addsi): New splitter.
14806 2013-07-25 Sterling Augustine <saugustine@google.com>
14808 * dwarf2out.c (size_of_pubnames): Move code to...
14809 (include_pubname_in_output): ...here. New.
14810 (want_pubnames): Rearrange.
14811 (output_pubnames): Call include_pubname_in_output. Move assertion.
14813 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
14815 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
14817 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
14820 * doc/extend.texi: Remove obsolete builtins. Fix
14821 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
14823 2013-07-25 Jan Hubicka <jh@suse.cz>
14825 * cgraph.c (release_function_body): Break out from ...
14826 (cgraph_release_function_body): ... this one; also release DECL_RESULT
14827 and DECL_ARGUMENTS.
14828 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
14829 old_tree in the map.
14830 (create_specialized_node): Update.
14831 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
14833 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
14834 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
14835 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
14836 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
14839 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14841 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
14842 addsi3_carryin_alt2_<optab>): Correct output template.
14844 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14846 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
14847 Adjust for arm_restrict_it.
14848 Remove trailing whitespace.
14850 2013-07-25  Mark Kettenis  <kettenis@openbsd.org>
14852 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
14853 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
14855 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
14857 2013-07-25 Vladimir Makarov <vmakarov@redhat.com>
14859 PR rtl-optimization/57960
14860 * lra-constraints.c (process_alt_operands): Use the right mode
14861 when checking strict_low.
14863 2013-07-25 Jan Hubicka <jh@suse.cz>
14865 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
14866 * cgraph.c (cgraph_remove_node): Do not release function body
14867 when in cgraph streaming.
14868 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
14869 in other partitions are not considered reachable; fix handling of
14872 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
14874 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
14876 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
14881 * config/arm/arm.md ("*sibcall_insn): Replace use of
14882 Ss with US. Adjust output for v5 and v4t.
14883 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
14885 * config/arm/constraints.md ("Ss"): Rename to US.
14887 2013-07-25 Terry Guo <terry.guo@arm.com>
14889 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
14890 shift_add/shift_sub0/shift_sub1 RTXs.
14892 2013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
14893 Anton Blanchard <anton@au1.ibm.com>
14895 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
14896 (altivec_vpks<VI_char>ss): Likewise.
14897 (altivec_vpks<VI_char>us): Likewise.
14898 (altivec_vpku<VI_char>us): Likewise.
14899 (altivec_vpku<VI_char>um): Likewise.
14901 2013-07-24 David Malcolm <dmalcolm@redhat.com>
14903 Introduce context class.
14905 * Makefile.in (CONTEXT_H): New.
14906 (OBJS): Add context.o.
14907 (toplev.o): Add CONTEXT_H to dependencies.
14910 * toplev.c (general_init): Create the singleton gcc::context instance.
14916 2013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
14918 PR rtl-optimization/57968
14919 * mode-switching.c (create_pre_exit): Allow instructions that
14920 don't set a return register to need a non-exit mode.
14922 2013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
14923 Anton Blanchard <anton@au1.ibm.com>
14925 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
14926 operands to vperm for little endian.
14927 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
14928 of lvsl to create the control mask for a vperm for little endian.
14930 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14931 Anton Blanchard <anton@au1.ibm.com>
14933 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
14934 two operands for little-endian.
14936 2013-07-23 Steve Ellcey <sellcey@mips.com>
14938 * config/mips/mips.c (mips_case_values_threshold): New.
14939 (TARGET_CASE_VALUES_THRESHOLD): Define.
14941 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14942 Anton Blanchard <anton@au1.ibm.com>
14944 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
14945 selection of field for vector splat in little endian mode.
14947 2013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
14949 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
14950 expanders to rs6000.md.
14951 (ior<mode>3): Likewise.
14952 (and<mode>3): Likewise.
14953 (one_cmpl<mode>2): Likewise.
14954 (nor<mode>3): Likewise.
14955 (andc<mode>3): Likewise.
14956 (eqv<mode>3): Likewise.
14957 (nand<mode>3): Likewise.
14958 (orc<mode>3): Likewise.
14960 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
14963 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
14964 to split multi-word logical operations.
14965 (rs6000_split_logical_di): Likewise.
14966 (rs6000_split_logical): Likewise.
14968 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
14969 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
14970 and allow TImode operations in 32-bit.
14971 (vsx_and<mode>3_64bit): Likewise.
14972 (vsx_ior<mode>3_32bit): Likewise.
14973 (vsx_ior<mode>3_64bit): Likewise.
14974 (vsx_xor<mode>3_32bit): Likewise.
14975 (vsx_xor<mode>3_64bit): Likewise.
14976 (vsx_one_cmpl<mode>2_32bit): Likewise.
14977 (vsx_one_cmpl<mode>2_64bit): Likewise.
14978 (vsx_nor<mode>3_32bit): Likewise.
14979 (vsx_nor<mode>3_64bit): Likewise.
14980 (vsx_andc<mode>3_32bit): Likewise.
14981 (vsx_andc<mode>3_64bit): Likewise.
14982 (vsx_eqv<mode>3_32bit): Likewise.
14983 (vsx_eqv<mode>3_64bit): Likewise.
14984 (vsx_nand<mode>3_32bit): Likewise.
14985 (vsx_nand<mode>3_64bit): Likewise.
14986 (vsx_orc<mode>3_32bit): Likewise.
14987 (vsx_orc<mode>3_64bit): Likewise.
14989 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
14990 logical types in GPRs.
14992 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
14993 logical insns to rs6000.md, and allow TImode operations in
14995 (altivec_ior<mode>3): Likewise.
14996 (altivec_xor<mode>3): Likewise.
14997 (altivec_one_cmpl<mode>2): Likewise.
14998 (altivec_nor<mode>3): Likewise.
14999 (altivec_andc<mode>3): Likewise.
15001 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
15002 attributes for moving the 128-bit logical operations into
15004 (BOOL_REGS_OUTPUT): Likewise.
15005 (BOOL_REGS_OP1): Likewise.
15006 (BOOL_REGS_OP2): Likewise.
15007 (BOOL_REGS_UNARY): Likewise.
15008 (BOOL_REGS_AND_CR0): Likewise.
15009 (one_cmpl<mode>2): Add support for DI logical operations on
15010 32-bit, splitting the operations to 32-bit.
15011 (anddi3): Likewise.
15012 (iordi3): Likewise.
15013 (xordi3): Likewise.
15014 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
15015 changes to combine the 32/64-bit code, allow logical operations on
15016 TI mode in 32-bit, and to use similar match_operator patterns like
15017 scalar mode uses. Combine the Altivec and VSX code for logical
15018 operations, and move it here.
15019 (ior<mode>3, 128-bit types): Likewise.
15020 (xor<mode>3, 128-bit types): Likewise.
15021 (one_cmpl<mode>3, 128-bit types): Likewise.
15022 (nor<mode>3, 128-bit types): Likewise.
15023 (andc<mode>3, 128-bit types): Likewise.
15024 (eqv<mode>3, 128-bit types): Likewise.
15025 (nand<mode>3, 128-bit types): Likewise.
15026 (orc<mode>3, 128-bit types): Likewise.
15027 (and<mode>3_internal): Likewise.
15028 (bool<mode>3_internal): Likewise.
15029 (boolc<mode>3_internal1): Likewise.
15030 (boolc<mode>3_internal2): Likewise.
15031 (boolcc<mode>3_internal1): Likewise.
15032 (boolcc<mode>3_internal2): Likewise.
15033 (eqv<mode>3_internal1): Likewise.
15034 (eqv<mode>3_internal2): Likewise.
15035 (one_cmpl1<mode>3_internal): Likewise.
15037 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
15039 * config/microblaze/microblaze.c (microblaze_expand_prologue):
15040 Rename flag_stack_usage to flag_stack_usage_info.
15042 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
15044 * config/microblaze/sync.md: New file.
15045 * config/microblaze/microblaze.md: Include sync.md
15046 * config/microblaze/microblaze.c: Add print_operand 'y'.
15047 * config/microblaze/constraints.md: Add memory_contraint
15048 'Q' which is a single register.
15050 2013-07-23 Eric Botcazou <ebotcazou@adacore.com>
15052 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
15054 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
15056 * reload.c (find_reloads): Exit loop once we find this operand
15057 cannot be reloaded somehow for this alternative.
15059 * reload.c (find_reloads): Exit loop once we find a hard register.
15061 * rtlanal.c (computed_jump_p): Exit loop once we find label
15064 * i386.c (ix86_pad_returns): Exit loop after setting replace.
15066 * cfgloopmanip.c (remove_path): Exit loop after setting
15069 * gensupport.c (subst_dup): Avoid loop if code is not
15070 MATCH_DUP nor MATCH_OP_DUP.
15072 2013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
15074 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
15076 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
15078 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
15079 true for SP_REGNUM if mode == ptr_mode.
15080 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
15081 with value R0_REGNUM + 31.
15083 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
15085 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
15086 pad pointer-typed argument downward.
15088 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
15090 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
15091 and __ILP32__ when the ILP32 model is in use.
15093 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
15095 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
15096 (aarch64_load_symref_appropriately): In the case of
15097 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
15098 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
15099 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
15100 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
15101 if 'mode' doesn't equal to 'ptr_mode'.
15102 (aarch64_output_mi_thunk): Add an assertion on the alignment of
15103 'vcall_offset'; change to call aarch64_emit_move differently depending
15104 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
15105 to calculate the upper bound of 'vcall_offset'.
15106 (aarch64_cannot_force_const_mem): Change to also return true if
15108 (aarch64_legitimize_reload_address): In the case of large
15109 displacements, add new local variable 'xmode' and an assertion
15110 based on it; change to use 'xmode' to generate the new rtx and
15112 (aarch64_asm_trampoline_template): Change to generate the template
15113 differently depending on TARGET_ILP32 or not; change to use
15114 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
15115 (aarch64_trampoline_size): Removed.
15116 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
15117 and replace immediate literals with it. Change to use 'ptr_mode'
15118 instead of 'DImode' and call convert_memory_address if the mode
15119 of 'fnaddr' doesn't equal to 'ptr_mode'.
15120 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
15122 (aarch64_elf_asm_destructor): Likewise.
15123 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
15124 on TARGET_ILP32 instead of aarch64_trampoline_size.
15125 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
15126 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
15127 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
15128 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
15129 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
15130 (storewb_pair<GPI:mode>_<P:mode>): ... this.
15131 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
15132 depending on the value of 'mode'.
15133 (add_losym_<mode>): New.
15134 (ldr_got_small_<mode>): New, based on ldr_got_small.
15135 (ldr_got_small): Remove.
15136 (ldr_got_small_sidi): New.
15137 * config/aarch64/iterators.md (P): New.
15138 (PTR): Change to 'ptr_mode' in the condition.
15140 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
15142 * config.gcc (aarch64*-*-*): Support --with-abi.
15143 (aarch64*-*-elf): Support --with-multilib-list.
15144 (aarch64*-*-linux*): Likewise.
15145 (supported_defaults): Add abi to aarch64*-*-*.
15146 * configure.ac: Mention AArch64 for --with-multilib-list.
15147 * configure: Re-generated.
15148 * config/aarch64/biarchilp32.h: New file.
15149 * config/aarch64/biarchlp64.h: New file.
15150 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
15152 (MULTILIB_DEFAULTS): Ditto.
15153 (DRIVER_SELF_SPECS): Ditto.
15154 (ASM_SPEC): Update to also substitute -mabi.
15155 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
15156 file whose name depends on -mabi= and -mbig-endian.
15157 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
15159 (POINTER_SIZE): New define.
15160 (POINTERS_EXTEND_UNSIGNED): Ditto.
15161 (enum aarch64_abi_type): New enumeration tag.
15162 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
15163 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
15164 (TARGET_ILP32): New define.
15165 * config/aarch64/aarch64.opt (mabi): New.
15166 (aarch64_abi): New.
15167 (ilp32, lp64): New values for -mabi.
15168 * config/aarch64/t-aarch64 (comma): New define.
15169 (MULTILIB_OPTIONS): Ditto.
15170 (MULTILIB_DIRNAMES): Ditto.
15171 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
15172 * doc/invoke.texi: Document -mabi for AArch64.
15174 2013-07-23 Georg-Johann Lay <avr@gjlay.de>
15176 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
15178 2013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
15179 Anton Blanchard <anton@au1.ibm.com>
15181 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
15182 endianness when selecting field to splat.
15184 2013-07-22 Eric Christopher <echristo@gmail.com>
15186 * dwarf2out.c (die_odr_checksum): New function to use
15187 CHECKSUM_ macros and ULEB128 for DIE tag.
15188 (generate_type_signature): Use.
15190 2013-07-22 Eric Botcazou <ebotcazou@adacore.com>
15192 * config.gcc (sparc*-*-*): Accept leon3 processor.
15193 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
15194 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
15195 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
15196 * config/sparc/sparc.opt (enum processor_type): Add leon3.
15197 (mfix-ut699): Adjust comment.
15198 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
15199 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
15200 (CPP_CPU_SPEC): Likewise.
15201 (ASM_CPU_SPEC): Likewise.
15202 * config/sparc/sparc.c (leon3_cost): New constant.
15203 (sparc_option_override): Add leon3 support.
15204 (mem_ref): New function.
15205 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
15206 (sparc_do_work_around_errata): Look into the instruction in the delay
15207 slot and adjust accordingly. Add fix for the data cache nullify issues
15208 of the UT699. Change insertion position for the NOP.
15209 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
15210 (leon3_load): New reservation.
15211 (leon_store): Bump latency to 2.
15212 (grfpu): New automaton.
15213 (grfpu_alu): New unit.
15214 (grfpu_ds): Likewise.
15215 (leon_fp_alu): Adjust.
15216 (leon_fp_mult): Delete.
15217 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
15218 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
15219 * config/sparc/sparc.md (cpu): Add leon3.
15220 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
15221 (swapsi): Likewise.
15222 (atomic_test_and_set): Likewise.
15223 (ldstub): Likewise.
15225 2013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
15227 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
15228 default for R5900 targets.
15229 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
15230 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
15231 * config/mips/mips.c (mips_option_override): Report an error for
15232 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
15233 for -march=r5900 -mhard-float.
15235 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
15237 * df-problems.c (can_move_insns_across): Exit loop once we
15238 find a non-fixed, non-global register.
15240 * ipa-pure-const.c (propagate_nothrow): Exit loop after
15243 * omega.c (omega_eliminate_red): Break after setting red_found.
15244 (omega_problem_has_red_equations): Similarly after setting found.
15245 (omega_query_variable): Similarly after setting coupled.
15247 2013-07-22 Marek Polacek <polacek@redhat.com>
15249 * gimplify.c: Don't include gimple.h twice.
15251 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15253 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
15254 instead of TARGET_THUMB1.
15255 (Pz): New constraint.
15256 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
15258 (compare_negsi_si): Likewise.
15259 (compare_addsi2_op0): Likewise.
15260 (compare_addsi2_op1): Likewise.
15261 (addsi3_carryin_<optab>): Likewise.
15262 (addsi3_carryin_alt2_<optab>): Likewise.
15263 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
15264 for arm_restrict_it.
15265 (subsi3_carryin): Likewise.
15266 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
15267 (minmax_arithsi): Disable for arm_restrict_it.
15268 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
15269 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
15270 (satsi_<SAT:code>_shift): Likewise.
15271 (arm_shiftsi3): Add alternative for 16-bit encoding.
15272 (arm32_movhf): Disable for arm_restrict_it.
15273 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
15274 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
15276 2013-07-22 Sofiane Naci <sofiane.naci@arm.com>
15278 * config/arm/arm.md (attribute "insn"): Delete.
15279 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
15280 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
15281 (not_shiftsi): Update for attribute change.
15282 (not_shiftsi_compare0): Likewise.
15283 (not_shiftsi_compare0_scratch): Likewise.
15284 (arm_one_cmplsi2): Likewise.
15285 (thumb1_one_cmplsi2): Likewise.
15286 (notsi_compare0): Likewise.
15287 (notsi_compare0_scratch): Likewise.
15288 (thumb1_movdi_insn): Likewise.
15289 (arm_movsi_insn): Likewise.
15290 (movhi_insn_arch4): Likewise.
15291 (movhi_bytes): Likewise.
15292 (arm_movqi_insn): Likewise.
15293 (thumb1_movqi_insn): Likewise.
15294 (arm32_movhf): Likewise.
15295 (thumb1_movhf): Likewise.
15296 (arm_movsf_soft_insn): Likewise.
15297 (thumb1_movsf_insn): Likewise.
15298 (thumb_movdf_insn): Likewise.
15299 (movsicc_insn): Likewise.
15300 (movsfcc_soft_insn): Likewise.
15301 (and_scc): Likewise.
15302 (cond_move): Likewise.
15303 (if_move_not): Likewise.
15304 (if_not_move): Likewise.
15305 (if_shift_move): Likewise.
15306 (if_move_shift): Likewise.
15307 (if_shift_shift): Likewise.
15308 (if_not_arith): Likewise.
15309 (if_arith_not): Likewise.
15310 (cond_move_not): Likewise.
15311 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
15312 (neon_mov<mode>): Likewise.
15313 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
15314 (thumb2_movsi_vfp): Likewise.
15315 (movsf_vfp): Likewise.
15316 (thumb2_movsf_vfp): Likewise.
15317 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
15319 (cortexa7_older_only): Likewise.
15320 (cortexa7_younger): Likewise.
15321 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
15322 (1020alu_shift_op): Likewise.
15323 (1020alu_shift_reg_op): Likewise.
15324 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
15325 (alu_shift_op): Likewise.
15326 (alu_shift_reg_op): Likewise.
15327 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
15328 (11_alu_shift_op): Likewise.
15329 (11_alu_shift_reg_op): Likewise.
15330 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
15331 (9_alu_shift_reg_op): Likewise.
15332 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
15334 (cortex_a15_alu_shift): Likewise.
15335 (cortex_a15_alu_shift_reg): Likewise.
15336 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
15337 (cortex_a5_alu_shift): Likewise.
15338 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
15340 (cortex_a53_alu_shift): Likewise.
15341 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
15343 (cortex_a7_alu_reg): Likewise.
15344 (cortex_a7_alu_shift): Likewise.
15345 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
15346 (cortex_a8_alu_shift): Likewise.
15347 (cortex_a8_alu_shift_reg): Likewise.
15348 (cortex_a8_mov): Likewise.
15349 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
15350 (cortex_a9_dp_shift): Likewise.
15351 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
15352 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
15353 (cortex_r4_mov): Likewise.
15354 (cortex_r4_alu_shift): Likewise.
15355 (cortex_r4_alu_shift_reg): Likewise.
15356 * config/arm/fa526.md (526_alu_op): Update for attribute change.
15357 (526_alu_shift_op): Likewise.
15358 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
15359 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
15360 (626te_alu_shift_op): Likewise.
15361 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
15362 (726te_alu_op): Likewise.
15363 (726te_alu_shift_op): Likewise.
15364 (726te_alu_shift_reg_op): Likewise.
15365 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
15366 (mp626_alu_shift_op): Likewise.
15367 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
15368 (pj4_alu_e1_conds): Likewise.
15369 (pj4_alu): Likewise.
15370 (pj4_alu_conds): Likewise.
15371 (pj4_shift): Likewise.
15372 (pj4_shift_conds): Likewise.
15373 (pj4_alu_shift): Likewise.
15374 (pj4_alu_shift_conds): Likewise.
15376 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15378 * config/arm/predicates.md (shiftable_operator_strict_it):
15380 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
15381 Disable cond_exec version for arm_restrict_it.
15382 (thumb2_smaxsi3): Convert to generate cond_exec.
15383 (thumb2_sminsi3): Likewise.
15384 (thumb32_umaxsi3): Likewise.
15385 (thumb2_uminsi3): Likewise.
15386 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
15387 (thumb2_neg_abssi2): Likewise.
15388 (thumb2_mov_scc): Add alternative for 16-bit encoding.
15389 (thumb2_movsicc_insn): Adjust alternatives.
15390 (thumb2_mov_negscc): Disable for arm_restrict_it.
15391 (thumb2_mov_negscc_strict_it): New pattern.
15392 (thumb2_mov_notscc_strict_it): New pattern.
15393 (thumb2_mov_notscc): Disable for arm_restrict_it.
15394 (thumb2_ior_scc): Likewise.
15395 (thumb2_ior_scc_strict_it): New pattern.
15396 (thumb2_cond_move): Adjust for arm_restrict_it.
15397 (thumb2_cond_arith): Disable for arm_restrict_it.
15398 (thumb2_cond_arith_strict_it): New pattern.
15399 (thumb2_cond_sub): Adjust for arm_restrict_it.
15400 (thumb2_movcond): Likewise.
15401 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
15402 (thumb2_zero_extendhisi2_v6): Likewise.
15403 (thumb2_zero_extendqisi2_v6): Likewise.
15404 (orsi_notsi_si): Likewise.
15405 (orsi_not_shiftsi_si): Likewise.
15407 2013-07-22 Georg-Johann Lay <avr@gjlay.de>
15409 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
15410 instruction sequence is 1 byte shorter.
15412 2013-07-22 Uros Bizjak <ubizjak@gmail.com>
15414 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
15415 it is not needed after split.
15417 2013-07-20 Iain Sandoe <iain@codesourcery.com>
15420 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
15421 second label for nonlocal goto receivers. Don't output pic base labels
15422 unless we're producing PIC; mark that action unreachable().
15423 (ix86_save_reg): If the function contains a nonlocal label, save the
15425 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
15426 * config/darwin.c (emitted_pic_label_num): New GTY.
15427 (update_pic_label_number_if_needed): New.
15428 (machopic_output_function_base_name): Adjust for nonlocal receiver
15430 (machopic_should_output_picbase_label): New.
15431 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
15432 (nonlocal_goto_receiver): New insn and split.
15434 2013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
15436 * config/aarch64/aarch64-builtins.c
15437 (aarch64_fold_builtin): Fold abs in all modes.
15438 * config/aarch64/aarch64-simd-builtins.def
15439 (abs): Enable for all modes.
15440 * config/aarch64/arm_neon.h
15441 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
15442 (vabs_f64): Add missing intrinsic.
15444 2013-07-19 Ian Bolton <ian.bolton@arm.com>
15446 * config/aarch64/arm_neon.h (vabs_s64): New function
15448 2013-07-19 Georg-Johann Lay <avr@gjlay.de>
15451 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
15452 * config/avr/avr.md (adjust_len): Add `round'.
15453 * config/avr/avr-protos.h (avr_out_round): New prototype.
15454 (avr_out_plus): Add `out_label' argument.
15455 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
15456 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
15457 Handle the case where `insn' is just a pattern.
15458 (avr_out_bitop): Handle the case where `insn' is just a pattern.
15459 (avr_out_round): New function.
15460 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
15462 2013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
15464 * config/microblaze/microblaze.c (microblaze_expand_prologue):
15465 Add check for flag_stack_usage to handle -fstack-usage support
15467 2013-07-18 Pat Haugen <pthaugen@us.ibm.com>
15469 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
15470 interaction for new Power8 flags and VSX.
15472 2013-07-18 Sriraman Tallam <tmsriram@google.com>
15474 PR middle-end/57698
15475 * tree-inline.c (expand_call_inline): Emit errors during
15476 early_inlining only if optimization is not turned on.
15478 2013-07-18 David Malcolm <dmalcolm@redhat.com>
15482 * passes.c (init_optimization_passes): Move the construction of
15483 the pass hierarchy into a new passes.def file.
15485 * Makefile.in (passes.o): Add dependency on passes.def.
15487 2013-07-18 David Malcolm <dmalcolm@redhat.com>
15489 * passes.c (init_optimization_passes): Introduce macros for
15490 constructing the tree of passes (INSERT_PASSES_AFTER,
15491 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
15492 TERMINATE_PASS_LIST).
15494 2013-07-18 Vladimir Makarov <vmakarov@redhat.com>
15495 Wei Mi <wmi@google.com>
15497 PR rtl-optimization/57878
15498 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
15500 (reload_pseudo_compare_func): Check nregs first for reload
15503 2013-07-18 David Malcolm <dmalcolm@redhat.com>
15505 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
15507 2013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
15509 * read-rtl.c (validate_const_int): Once an invalid character is
15510 seen, quit the loop.
15512 * gengtype.c (write_roots): Similarly once we find the "deletable"
15513 or "if_marked" option.
15515 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
15517 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
15518 "xtab" and "sat". Move value "clz" from here to ...
15519 (attriubte "type"): ... here.
15520 (satsi_<SAT:code>): Delete "insn" attribute.
15521 (satsi_<SAT:code>_shift): Likewise.
15522 (arm_zero_extendqisi2addsi): Likewise.
15523 (arm_extendqisi2addsi): Likewise.
15524 (clzsi2): Update for attribute changes.
15525 (rbitsi2): Likewise.
15526 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
15528 (arm_usatsihi): Likewise.
15529 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
15531 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
15533 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
15534 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
15535 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
15536 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
15537 in alphabetical order.
15538 (attribute "core_cycles"): Update for attribute changes.
15539 (arm_addsi3): Likewise.
15540 (addsi3_compare0): Likewise.
15541 (addsi3_compare0_scratch): Likewise.
15542 (addsi3_compare_op1): Likewise.
15543 (addsi3_compare_op2): Likewise.
15544 (compare_addsi2_op0): Likewise.
15545 (compare_addsi2_op1): Likewise.
15546 (addsi3_carryin_shift_<optab>): Likewise.
15547 (subsi3_carryin_shift): Likewise.
15548 (rsbsi3_carryin_shift): Likewise.
15549 (arm_subsi3_insn): Likewise.
15550 (subsi3_compare0): Likewise.
15551 (subsi3_compare): Likewise.
15552 (arm_andsi3_insn): Likewise.
15553 (thumb1_andsi3_insn): Likewise.
15554 (andsi3_compare0): Likewise.
15555 (andsi3_compare0_scratch): Likewise.
15556 (zeroextractsi_compare0_scratch
15557 (andsi_not_shiftsi_si): Likewise.
15558 (iorsi3_insn): Likewise.
15559 (iorsi3_compare0): Likewise.
15560 (iorsi3_compare0_scratch): Likewise.
15561 (arm_xorsi3): Likewise.
15562 (thumb1_xorsi3_insn): Likewise.
15563 (xorsi3_compare0): Likewise.
15564 (xorsi3_compare0_scratch): Likewise.
15565 (satsi_<SAT:code>_shift): Likewise.
15567 (arm_shiftsi3): Likewise.
15568 (shiftsi3_compare0): Likewise.
15569 (not_shiftsi): Likewise.
15570 (not_shiftsi_compare0): Likewise.
15571 (not_shiftsi_compare0_scratch): Likewise.
15572 (arm_one_cmplsi2): Likewise.
15573 (thumb_one_complsi2): Likewise.
15574 (notsi_compare0): Likewise.
15575 (notsi_compare0_scratch): Likewise.
15576 (thumb1_zero_extendhisi2): Likewise.
15577 (arm_zero_extendhisi2): Likewise.
15578 (arm_zero_extendhisi2_v6): Likewise.
15579 (arm_zero_extendhisi2addsi): Likewise.
15580 (thumb1_zero_extendqisi2): Likewise.
15581 (thumb1_zero_extendqisi2_v6): Likewise.
15582 (arm_zero_extendqisi2): Likewise.
15583 (arm_zero_extendqisi2_v6): Likewise.
15584 (arm_zero_extendqisi2addsi): Likewise.
15585 (thumb1_extendhisi2): Likewise.
15586 (arm_extendhisi2): Likewise.
15587 (arm_extendhisi2_v6): Likewise.
15588 (arm_extendqisi): Likewise.
15589 (arm_extendqisi_v6): Likewise.
15590 (arm_extendqisi2addsi): Likewise.
15591 (thumb1_extendqisi2): Likewise.
15592 (thumb1_movdi_insn): Likewise.
15593 (arm_movsi_insn): Likewise.
15594 (movsi_compare0): Likewise.
15595 (movhi_insn_arch4): Likewise.
15596 (movhi_bytes): Likewise.
15597 (arm_movqi_insn): Likewise.
15598 (thumb1_movqi_insn): Likewise.
15599 (arm32_movhf): Likewise.
15600 (thumb1_movhf): Likewise.
15601 (arm_movsf_soft_insn): Likewise.
15602 (thumb1_movsf_insn): Likewise.
15603 (movdf_soft_insn): Likewise.
15604 (thumb_movdf_insn): Likewise.
15605 (arm_cmpsi_insn): Likewise.
15606 (cmpsi_shiftsi): Likewise.
15607 (cmpsi_shiftsi_swp): Likewise.
15608 (arm_cmpsi_negshiftsi_si): Likewise.
15609 (movsicc_insn): Likewise.
15610 (movsfcc_soft_insn): Likewise.
15611 (arith_shiftsi): Likewise.
15612 (arith_shiftsi_compare0
15613 (arith_shiftsi_compare0_scratch
15614 (sub_shiftsi): Likewise.
15615 (sub_shiftsi_compare0
15616 (sub_shiftsi_compare0_scratch
15617 (and_scc): Likewise.
15618 (cond_move): Likewise.
15619 (if_plus_move): Likewise.
15620 (if_move_plus): Likewise.
15621 (if_move_not): Likewise.
15622 (if_not_move): Likewise.
15623 (if_shift_move): Likewise.
15624 (if_move_shift): Likewise.
15625 (if_shift_shift): Likewise.
15626 (if_not_arith): Likewise.
15627 (if_arith_not): Likewise.
15628 (cond_move_not): Likewise.
15629 (thumb1_ashlsi3): Set type attribute.
15630 (thumb1_ashrsi3): Likewise.
15631 (thumb1_lshrsi3): Likewise.
15632 (thumb1_rotrsi3): Likewise.
15633 (shiftsi3_compare0_scratch): Likewise.
15634 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
15635 (neon_mov<mode>): Likewise.
15636 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
15638 (thumb2_movsi_insn): Likewise.
15639 (thumb2_cmpsi_neg_shiftsi): Likewise.
15640 (thumb2_extendqisi_v6): Likewise.
15641 (thumb2_zero_extendhisi2_v6): Likewise.
15642 (thumb2_zero_extendqisi2_v6): Likewise.
15643 (thumb2_shiftsi3_short): Likewise.
15644 (thumb2_addsi3_compare0_scratch): Likewise.
15645 (orsi_not_shiftsi_si): Likewise.
15646 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
15647 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
15649 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
15650 (1020alu_shift_op): Likewise.
15651 (1020alu_shift_reg_op): Likewise.
15652 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
15653 (alu_shift_op): Likewise.
15654 (alu_shift_reg_op): Likewise.
15655 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
15656 (11_alu_shift_op): Likewise.
15657 (11_alu_shift_reg_op): Likewise.
15658 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
15659 (9_alu_shift_reg_op): Likewise.
15660 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
15662 (cortex_a15_alu_shift): Likewise.
15663 (cortex_a15_alu_shift_reg): Likewise.
15664 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
15666 (cortex_a5_alu_shift): Likewise.
15667 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
15669 (cortex_a53_alu_shift): Likewise.
15670 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
15672 (cortex_a7_alu_reg): Likewise.
15673 (cortex_a7_alu_shift): Likewise.
15674 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
15676 (cortex_a8_alu_shift): Likewise.
15677 (cortex_a8_alu_shift_reg): Likewise.
15678 (cortex_a8_mov): Likewise.
15679 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
15680 (cortex_a9_dp_shift): Likewise.
15681 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
15683 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
15685 (cortex_r4_mov): Likewise.
15686 (cortex_r4_alu_shift): Likewise.
15687 (cortex_r4_alu_shift_reg): Likewise.
15688 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
15689 (526_alu_shift_op): Likewise.
15690 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
15691 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
15692 (626te_alu_shift_op): Likewise.
15693 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
15694 (726te_alu_op): Likewise.
15695 (726te_alu_shift_op): Likewise.
15696 (726te_alu_shift_reg_op): Likewise.
15697 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
15698 (mp626_alu_shift_op): Likewise.
15699 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
15700 (pj4_alu_e1_conds): Likewise.
15701 (pj4_alu): Likewise.
15702 (pj4_alu_conds): Likewise.
15703 (pj4_shift): Likewise.
15704 (pj4_shift_conds): Likewise.
15705 (pj4_alu_shift): Likewise.
15706 (pj4_alu_shift_conds): Likewise.
15707 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
15709 (cortexa7_older_only): Likewise.
15710 (cortexa7_younger): Likewise.
15712 2013-07-18 David Malcolm <dmalcolm@redhat.com>
15714 * ipa-pure-const.c (generate_summary): Rename to...
15715 (pure_const_generate_summary): ... this.
15717 2013-07-17 Iain Sandoe <iain@codesourcery.com>
15719 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
15721 2013-07-17 Yvan Roux <yvan.roux@linaro.org>
15724 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
15727 2013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
15729 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
15730 enabled without -march=zEC12.
15731 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
15734 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
15736 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
15737 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
15738 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
15739 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
15740 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
15741 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
15742 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
15744 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
15745 and ISA_HAS_NMADD3_NMSUB3.
15746 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
15747 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
15748 (nmsub4<mode>, nmsub3<mode>): Likewise.
15749 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
15751 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
15753 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
15754 TARGET_MIPS5400 checking.
15756 2013-07-16 Jakub Jelinek <jakub@redhat.com>
15757 Peter Bergner <bergner@vnet.ibm.com>
15759 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
15760 registers in the comment.
15761 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
15762 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
15763 rather than FIRST_PSEUDO_REGISTERS.
15765 2013-07-16 Peter Bergner <bergner@vnet.ibm.com>
15767 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
15768 enable extra ISA flags with TARGET_HTM.
15770 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
15772 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
15775 2013-07-15 Cong Hou <congh@google.com>
15777 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
15778 in compare function for sorting.
15780 2013-07-15 Peter Bergner <bergner@vnet.ibm.com>
15782 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
15783 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
15784 * config/rs6000/rs6000.opt: Add -mhtm option.
15785 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
15786 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
15787 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
15788 __HTM__ if the HTM instructions are available.
15789 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
15790 htm_spr_reg_operand): New define_predicates.
15791 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
15792 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
15794 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
15795 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
15796 HTM builtin functions.
15797 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
15798 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
15799 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
15800 (rs6000_builtin_mask_calculate): Likewise.
15801 (rs6000_option_override_internal): Likewise.
15802 (bdesc_htm): Add new HTM builtin support.
15803 (htm_spr_num): New function.
15804 (htm_spr_regno): Likewise.
15805 (rs6000_htm_spr_icode): Likewise.
15806 (htm_expand_builtin): Likewise.
15807 (htm_init_builtins): Likewise.
15808 (rs6000_expand_builtin): Add support for HTM builtin functions.
15809 (rs6000_init_builtins): Likewise.
15810 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
15812 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
15813 (TARGET_HTM, MASK_HTM): Define macros.
15814 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
15815 (FIXED_REGISTERS): Likewise.
15816 (CALL_USED_REGISTERS): Likewise.
15817 (CALL_REALLY_USED_REGISTERS): Likewise.
15818 (REG_ALLOC_ORDER): Likewise.
15819 (enum reg_class): Likewise.
15820 (REG_CLASS_NAMES): Likewise.
15821 (REG_CLASS_CONTENTS): Likewise.
15822 (REGISTER_NAMES): Likewise.
15823 (ADDITIONAL_REGISTER_NAMES): Likewise.
15824 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
15825 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
15826 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
15827 * config/rs6000/htm.md: New file.
15828 * config/rs6000/htmintrin.h: New file.
15829 * config/rs6000/htmxlintrin.h: New file.
15831 2013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
15833 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
15834 Define SYMBOL_TINY_GOT, update comment.
15835 * config/aarch64/aarch64.c
15836 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
15837 (aarch64_expand_mov_immediate): Likewise.
15838 (aarch64_print_operand): Likewise.
15839 (aarch64_classify_symbol): Likewise.
15840 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
15841 (ldr_got_tiny): Define.
15843 2013-07-13 Tobias Grosser <tobias@grosser.es>
15845 PR tree-optimization/54094
15846 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
15847 scheduling dimension for the parallelism check from the polyhedral
15848 information in the AST.
15849 * graphite-dependences.c (carries_deps): Do not assume the schedule is
15852 2013-07-13 Jason Merrill <jason@redhat.com>
15854 * print-tree.c (debug_vec_tree): Use debug_raw.
15855 (debug_raw (vec<tree, va_gc> &)): New.
15856 (debug_raw (vec<tree, va_gc> *)): New.
15857 * tree.h: Declare them.
15859 2013-07-13 Bin Cheng <bin.cheng@arm.com>
15861 * ifcvt.c (ifcvt_after_combine): New static variable.
15862 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
15864 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
15865 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
15866 rest_of_handle_if_after_reload): Pass new argument for if_convert.
15868 2013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
15870 * config/mips/mips.c (mips_expand_call): Remove empty statement.
15872 2013-07-12 Michael Matz <matz@suse.de>
15874 PR middle-end/55771
15875 * convert.c (convert_to_real): Reject non-float inner types.
15877 2013-07-12 Tejas Belagod <tejas.belagod@arm.com>
15879 * config/aarch64/aarch64-protos.h
15880 (aarch64_simd_immediate_valid_for_move): Remove.
15881 * config/aarch64/aarch64.c (simd_immediate_info): New member.
15882 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
15884 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
15886 2013-07-11 Steve Ellcey <sellcey@mips.com>
15888 * config/mips/mips.c (mips_conditional_register_usage): Do not
15889 use t[0-7] registers in MIPS16 mode when optimizing for size.
15891 2013-07-11 Sriraman Tallam <tmsriram@google.com>
15893 * config/i386/i386.c (dispatch_function_versions): Fix array
15894 indexing of function_version_info to match actual_versions.
15896 2013-07-11 Teresa Johnson <tejohnson@google.com>
15898 * vec.h (struct va_gc): Move release out-of-line.
15899 (va_gc::release): Call ggc_free on released vec.
15901 2013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
15903 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
15904 Require GOT register as additional operand in UNSPEC.
15905 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
15906 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
15907 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
15908 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
15909 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
15910 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
15911 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
15913 2013-07-11 Georg-Johann Lay <avr@gjlay.de>
15916 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
15917 name seen by assembler/linker rather if available.
15919 2013-07-11 Andreas Schwab <schwab@suse.de>
15921 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
15923 2013-07-10 Vladimir Makarov <vmakarov@redhat.com>
15925 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
15927 2013-07-10 Joseph Myers <joseph@codesourcery.com>
15929 * doc/tm.texi.in: Move hook documentation to ....
15930 * target.def: ... here.
15932 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
15933 text on @hook line.
15934 * doc/tm.texi: Regenerate.
15936 2013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
15939 * doc/invoke.texi: Document -Wconditionally-supported.
15941 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
15944 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
15947 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
15950 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
15951 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
15952 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
15953 Remove duplicate devices.
15954 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
15955 * config/avr/t-multilib: Regenerate.
15956 * config/avr/avr-tables.opt: Regenerate.
15957 * doc/avr-mmcu.texi: Regenerate.
15959 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
15962 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
15964 2013-07-10 Graham Stott <graham.stott@btinternet.com>
15966 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
15967 the cost of MULT when optimizing for size.
15969 2013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
15971 * config/cr16/cr16-protos.h: Don't include target.h.
15973 2013-07-09 Joseph Myers <joseph@codesourcery.com>
15975 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
15976 adjust register size for TDmode and TFmode for VSX registers.
15978 2013-07-08 Kai Tietz <ktietz@redhat.com>
15981 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
15982 hook_bool_const_tree_true.
15984 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
15986 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
15987 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
15988 * config/s390/s390.md: Define FPR*_REGNUM constants.
15989 Fix FPR2_REGNUM constant (18 -> 17).
15990 ("*trunc<BFP:mode><DFP_ALL:mode>2")
15991 ("*trunc<DFP_ALL:mode><BFP:mode>2")
15992 ("trunc<BFP:mode><DFP_ALL:mode>2")
15993 ("trunc<DFP_ALL:mode><BFP:mode>2")
15994 ("*extend<BFP:mode><DFP_ALL:mode>2")
15995 ("*extend<DFP_ALL:mode><BFP:mode>2")
15996 ("extend<BFP:mode><DFP_ALL:mode>2")
15997 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
16000 2013-07-08 Graham Stott <graham.stott@btinternet.com>
16002 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
16004 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16006 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
16007 and cfun_fpr_bit_p to cfun_fpr_save_p.
16008 (s390_frame_area, s390_register_info, s390_frame_info)
16009 (s390_emit_prologue, s390_emit_epilogue)
16010 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
16012 * config/s390/s390.h: Define *_REGNUM macros for floating point
16015 2013-07-08 Eric Botcazou <ebotcazou@adacore.com>
16017 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
16019 2013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
16021 PR rtl-optimization/57786
16022 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
16023 and break out of the loop when it is set to false.
16025 2013-07-08 Jakub Jelinek <jakub@redhat.com>
16028 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
16029 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
16030 (const_int 63)) 0)).
16031 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
16032 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
16033 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
16035 PR rtl-optimization/57829
16036 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
16037 mask bits outside of mode are just sign-extension from mode to HWI.
16039 2013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
16041 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
16042 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
16043 adjust_address instead of change_address to keep info about alignment.
16044 (emit_strmov): Remove.
16045 (emit_memmov): New function.
16046 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
16047 (expand_movmem_epilogue): Likewise and return updated rtx for
16049 (expand_constant_movmem_prologue): Likewise and return updated rtx for
16050 destination and source.
16051 (decide_alignment): Refactor, handle vector_loop.
16052 (ix86_expand_movmem): Likewise.
16053 (ix86_expand_setmem): Likewise.
16054 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
16056 2013-07-07 Uros Bizjak <ubizjak@gmail.com>
16058 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
16059 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
16061 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
16063 * config/i386/sse.md (sse_movlhps): Change alternative 3
16064 of operand 2 to "m".
16066 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
16069 * config/i386/sse.md (iptr): New mode attribute.
16070 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
16071 (<sse>_vm<plusminus_insn><mode>3): Ditto.
16072 (<sse>_vmmul<mode>3): Ditto.
16073 (<sse>_vmdiv<mode>3): Ditto.
16074 (sse_vmrcpv4sf2): Ditto.
16075 (<sse>_vmsqrt<mode>2): Ditto.
16076 (sse_vmrsqrtv4sf2): Ditto.
16077 (<sse>_vm<code><mode>3): Ditto.
16078 (avx_vmcmp<mode>3): Ditto.
16079 (<sse>_vmmaskcmp<mode>3): Ditto.
16080 (<sse>_comi): Ditto.
16081 (<sse>_ucomi): Ditto.
16082 (*xop_vmfrcz_<mode>): Ditto.
16083 (*fmai_fmadd_<mode>): Ditto.
16084 (*fmai_fmsub_<mode>): Ditto.
16085 (*fmai_fnmadd_<mode>): Ditto.
16086 (*fmai_fnmsub_<mode>): Ditto.
16087 (*fma4i_vmfmadd_<mode>): Ditto.
16088 (*fma4i_vmfmsub_<mode>): Ditto.
16089 (*fma4i_vmfnmadd_<mode>): Ditto.
16090 (*fma4i_vmfnmsub_<mode>): Ditto.
16091 (*xop_vmfrcz_<mode>): Ditto.
16092 (sse_cvtps2pi): Ditto.
16093 (sse_cvttps2pi): Ditto.
16094 (sse_cvtss2si): Ditto.
16095 (sse_cvtss2si_2): Ditto.
16096 (sse_cvtss2siq_2): Ditto.
16097 (sse_cvttss2si): Ditto.
16098 (sse_cvttss2siq): Ditto.
16099 (sse_cvtsd2si): Ditto.
16100 (sse_cvtsd2si_2): Ditto.
16101 (sse_cvtsd2siq_2): Ditto.
16102 (sse_cvttsd2si): Ditto.
16103 (sse_cvttsd2siq): Ditto.
16104 (sse_cvtsd2ss): Ditto.
16105 (sse_cvtss2sd): Ditto.
16106 (avx2_pbroadcast<mode>): Ditto.
16107 (avx2_pbroadcast<mode>_1): Ditto.
16108 (*avx_vperm_broadcast_v4sf): Ditto.
16110 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
16111 (sse_movlhps): Ditto.
16112 (sse_storehps): Ditto.
16113 (sse_loadhps): Ditto.
16114 (sse_storelps): Ditto.
16115 (sse_loadlps): Ditto.
16116 (*vec_concatv4sf): Ditto.
16117 (*vec_interleave_highv2df): Ditto.
16118 (*vec_interleave_lowv2df): Ditto.
16119 (*vec_extractv2df_1_sse): Ditto.
16120 (*vec_extractv2df_0_sse): Ditto.
16121 (sse2_storelpd): Ditto.
16122 (sse2_loadlpd): Ditto.
16123 (sse2_movsd): Ditto.
16124 (*vec_concatv4si): Ditto.
16125 (vec_concatv2di): Ditto.
16127 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
16128 for Intel asm dialect.
16129 (mmx_punpcklwd): Ditto.
16130 (mmx_punpckldq): Ditto.
16132 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
16133 for intel assembler dialect.
16135 2013-07-06 Jakub Jelinek <jakub@redhat.com>
16138 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
16139 for BUILT_IN_C{LZ,LRSB}*.
16140 * tree.h (CASE_INT_FN): Add FN##IMAX case.
16141 * tree-vrp.c (extract_range_basic): Handle
16142 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
16143 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
16144 fall thru to code calling set_value*.
16145 * builtins.c (expand_builtin): Remove *IMAX cases.
16146 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
16147 if width is bigger than 2*HWI.
16149 2013-07-05 Vladimir Makarov <vmakarov@redhat.com>
16151 PR rtl-optimization/55342
16152 * lra-int.h (lra_subreg_reload_pseudos): New.
16153 * lra.c: Add undoing optional reloads to the block diagram.
16154 (lra_subreg_reload_pseudos): New.
16155 (lra_optional_reload_pseudos): Change comments.
16156 (lra): Init and clear lra_subreg_reload_pseudos. Clear
16157 lra_optional_reload_pseudos after undo transformations.
16158 * lra-assigns.c (pseudo_prefix_title): New.
16159 (lra_setup_reg_renumber): Use it.
16160 (spill_for): Ditto. Check subreg reload pseudos too.
16161 (assign_by_spills): Consider subreg reload pseudos too.
16162 * lra-constraints.c (simplify_operand_subreg): Use
16163 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
16164 (curr_insn_transform): Recognize and do optional reloads.
16165 (undo_optional_reloads): New.
16166 (lra_undo_inheritance): Call undo_optional_reloads.
16168 2013-07-05 Thomas Quinot <quinot@adacore.com>
16170 * tree-complex.c (expand_complex_operations_1): Fix typo.
16172 2013-07-04 Tejas Belagod <tejas.belagod@arm.com>
16174 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
16175 (tune_params): New member 'const vec_costs'.
16176 * config/aarch64/aarch64.c (generic_vector_cost): New.
16177 (generic_tunings): New member 'generic_vector_cost'.
16178 (aarch64_builtin_vectorization_cost): New.
16179 (aarch64_add_stmt_cost): New.
16180 (TARGET_VECTORIZE_ADD_STMT_COST): New.
16181 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
16183 2013-07-03 Jakub Jelinek <jakub@redhat.com>
16186 * config/i386/predicates.md (vsib_address_operand): Disallow
16187 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
16189 2013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
16191 PR middle-end/55030
16192 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
16193 expand_builtin_setjmp_receiver.
16194 (expand_label): Adjust, call expand_builtin_setjmp_receiver
16195 with NULL for the label parameter.
16196 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
16197 the frame-pointer. Adjust comments.
16198 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
16199 only if LABEL is non-NULL.
16201 2013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
16203 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
16204 (ARM_ABI_AAPCS64): Ditto.
16206 (ARM_DEFAULT_ABI): Ditto.
16208 2013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
16210 * config/aarch64/aarch64-builtins.c
16211 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
16212 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
16214 * config/aarch64/aarch64-simd.md
16215 (aarch64_ld1<VALL:mode>): New.
16216 (aarch64_st1<VALL:mode>): Likewise.
16217 * config/aarch64/arm_neon.h
16218 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
16220 2013-07-02 Sriraman Tallam <tmsriram@google.com>
16222 * config/i386/i386.c (gate_insert_vzeroupper): Check if
16224 (ix86_option_override_internal):Turn on all -mavx target flags by
16225 default as they are dependent on AVX anyway.
16227 2013-07-02 Cary Coutant <ccoutant@google.com>
16229 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
16230 deterministic hash.
16231 (loc_checksum_ordered): Likewise.
16232 (hash_loc_operands): Remove inline keyword.
16234 2013-07-02 Jakub Jelinek <jakub@redhat.com>
16236 PR tree-optimization/57741
16237 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
16238 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
16239 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
16240 Allow REAL_CST step_exprs if flag_associative_math.
16241 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
16243 2013-07-02 Ian Bolton <ian.bolton@arm.com>
16245 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
16247 2013-07-02 Ian Bolton <ian.bolton@arm.com>
16249 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
16251 2013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16253 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
16255 (iorsi3_insn): Likewise.
16256 (arm_xorsi3): Likewise.
16258 2013-07-01 Sofiane Naci <sofiane.naci@arm.com>
16260 * arm.md (attribute "wtype"): Delete. Move attribute values from here
16262 (attribute "type"): ... here, and prefix with "wmmx_".
16263 (attribute "core_cycles"): Update for attribute changes.
16264 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
16265 (tbcstv4hi): Likewise.
16266 (tbcstv2si): Likewise.
16267 (iwmmxt_iordi3): Likewise.
16268 (iwmmxt_xordi3): Likewise.
16269 (iwmmxt_anddi3): Likewise.
16270 (iwmmxt_nanddi3): Likewise.
16271 (iwmmxt_arm_movdi): Likewise.
16272 (iwmmxt_movsi_insn): Likewise.
16273 (mov<mode>_internal): Likewise.
16274 (and<mode>3_iwmmxt): Likewise.
16275 (ior<mode>3_iwmmxt): Likewise.
16276 (xor<mode>3_iwmmxt): Likewise.
16277 (add<mode>3_iwmmxt): Likewise.
16278 (ssaddv8qi3): Likewise.
16279 (ssaddv4hi3): Likewise.
16280 (ssaddv2si3): Likewise.
16281 (usaddv8qi3): Likewise.
16282 (usaddv4hi3): Likewise.
16283 (usaddv2si3): Likewise.
16284 (sub<mode>3_iwmmxt): Likewise.
16285 (sssubv8qi3): Likewise.
16286 (sssubv4hi3): Likewise.
16287 (sssubv2si3): Likewise.
16288 (ussubv8qi3): Likewise.
16289 (ussubv4hi3): Likewise.
16290 (ussubv2si3): Likewise.
16291 (mulv4hi3_iwmmxt): Likewise.
16292 (smulv4hi3_highpart): Likewise.
16293 (umulv4hi3_highpart): Likewise.
16294 (iwmmxt_wmacs): Likewise.
16295 (iwmmxt_wmacsz): Likewise.
16296 (iwmmxt_wmacu): Likewise.
16297 (iwmmxt_wmacuz): Likewise.
16298 (iwmmxt_clrdi): Likewise.
16299 (iwmmxt_clrv8qi): Likewise.
16300 (iwmmxt_clr4hi): Likewise.
16301 (iwmmxt_clr2si): Likewise.
16302 (iwmmxt_uavgrndv8qi3): Likewise.
16303 (iwmmxt_uavgrndv4hi3): Likewise.
16304 (iwmmxt_uavgv8qi3): Likewise.
16305 (iwmmxt_uavgv4hi3): Likewise.
16306 (iwmmxt_tinsrb): Likewise.
16307 (iwmmxt_tinsrh): Likewise.
16308 (iwmmxt_tinsrw): Likewise.
16309 (iwmmxt_textrmub): Likewise.
16310 (iwmmxt_textrmsb): Likewise.
16311 (iwmmxt_textrmuh): Likewise.
16312 (iwmmxt_textrmsh): Likewise.
16313 (iwmmxt_textrmw): Likewise.
16314 (iwmxxt_wshufh): Likewise.
16315 (eqv8qi3): Likewise.
16316 (eqv4hi3): Likewise.
16317 (eqv2si3): Likewise.
16318 (gtuv8qi3): Likewise.
16319 (gtuv4hi3): Likewise.
16320 (gtuv2si3): Likewise.
16321 (gtv8qi3): Likewise.
16322 (gtv4hi3): Likewise.
16323 (gtv2si3): Likewise.
16324 (smax<mode>3_iwmmxt): Likewise.
16325 (umax<mode>3_iwmmxt): Likewise.
16326 (smin<mode>3_iwmmxt): Likewise.
16327 (umin<mode>3_iwmmxt): Likewise.
16328 (iwmmxt_wpackhss): Likewise.
16329 (iwmmxt_wpackwss): Likewise.
16330 (iwmmxt_wpackdss): Likewise.
16331 (iwmmxt_wpackhus): Likewise.
16332 (iwmmxt_wpackwus): Likewise.
16333 (iwmmxt_wpackdus): Likewise.
16334 (iwmmxt_wunpckihb): Likewise.
16335 (iwmmxt_wunpckihh): Likewise.
16336 (iwmmxt_wunpckihw): Likewise.
16337 (iwmmxt_wunpckilb): Likewise.
16338 (iwmmxt_wunpckilh): Likewise.
16339 (iwmmxt_wunpckilw): Likewise.
16340 (iwmmxt_wunpckehub): Likewise.
16341 (iwmmxt_wunpckehuh): Likewise.
16342 (iwmmxt_wunpckehuw): Likewise.
16343 (iwmmxt_wunpckehsb): Likewise.
16344 (iwmmxt_wunpckehsh): Likewise.
16345 (iwmmxt_wunpckehsw): Likewise.
16346 (iwmmxt_wunpckelub): Likewise.
16347 (iwmmxt_wunpckeluh): Likewise.
16348 (iwmmxt_wunpckeluw): Likewise.
16349 (iwmmxt_wunpckelsb): Likewise.
16350 (iwmmxt_wunpckelsh): Likewise.
16351 (iwmmxt_wunpckelsw): Likewise.
16352 (ror<mode>3): Likewise.
16353 (ashr<mode>3_iwmmxt): Likewise.
16354 (lshr<mode>3_iwmmxt): Likewise.
16355 (ashl<mode>3_iwmmxt): Likewise.
16356 (ror<mode>3_di): Likewise.
16357 (ashr<mode>3_di): Likewise.
16358 (lshr<mode>3_di): Likewise.
16359 (ashl<mode>3_di): Likewise.
16360 (iwmmxt_wmadds): Likewise.
16361 (iwmmxt_wmaddu): Likewise.
16362 (iwmmxt_tmia): Likewise.
16363 (iwmmxt_tmiaph): Likewise.
16364 (iwmmxt_tmiabb): Likewise.
16365 (iwmmxt_tmiatb): Likewise.
16366 (iwmmxt_tmiabt): Likewise.
16367 (iwmmxt_tmiatt): Likewise.
16368 (iwmmxt_tmovmskb): Likewise.
16369 (iwmmxt_tmovmskh): Likewise.
16370 (iwmmxt_tmovmskw): Likewise.
16371 (iwmmxt_waccb): Likewise.
16372 (iwmmxt_wacch): Likewise.
16373 (iwmmxt_waccw): Likewise.
16374 (iwmmxt_waligni): Likewise.
16375 (iwmmxt_walignr): Likewise.
16376 (iwmmxt_walignr0): Likewise.
16377 (iwmmxt_walignr1): Likewise.
16378 (iwmmxt_walignr2): Likewise.
16379 (iwmmxt_walignr3): Likewise.
16380 (iwmmxt_wsadb): Likewise.
16381 (iwmmxt_wsadh): Likewise.
16382 (iwmmxt_wsadbz): Likewise.
16383 (iwmmxt_wsadhz): Likewise.
16384 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
16385 (iwmmxt_wabsdiffb): Likewise.
16386 (iwmmxt_wabsdiffh): Likewise.
16387 (iwmmxt_wabsdiffw): Likewise.
16388 (iwmmxt_waddsubhx): Likewise
16389 (iwmmxt_wsubaddhx): Likewise.
16390 (addc<mode>3): Likewise.
16391 (iwmmxt_avg4): Likewise.
16392 (iwmmxt_avg4r): Likewise.
16393 (iwmmxt_wmaddsx): Likewise.
16394 (iwmmxt_wmaddux): Likewise.
16395 (iwmmxt_wmaddsn): Likewise.
16396 (iwmmxt_wmaddun): Likewise.
16397 (iwmmxt_wmulwsm): Likewise.
16398 (iwmmxt_wmulwum): Likewise.
16399 (iwmmxt_wmulsmr): Likewise.
16400 (iwmmxt_wmulumr): Likewise.
16401 (iwmmxt_wmulwsmr): Likewise.
16402 (iwmmxt_wmulwumr): Likewise.
16403 (iwmmxt_wmulwl): Likewise.
16404 (iwmmxt_wqmulm): Likewise.
16405 (iwmmxt_wqmulwm): Likewise.
16406 (iwmmxt_wqmulmr): Likewise.
16407 (iwmmxt_wqmulwmr): Likewise.
16408 (iwmmxt_waddbhusm): Likewise.
16409 (iwmmxt_waddbhusl): Likewise.
16410 (iwmmxt_wqmiabb): Likewise.
16411 (iwmmxt_wqmiabt): Likewise.
16412 (iwmmxt_wqmiatb): Likewise.
16413 (iwmmxt_wqmiatt): Likewise.
16414 (iwmmxt_wqmiabbn): Likewise.
16415 (iwmmxt_wqmiabtn): Likewise.
16416 (iwmmxt_wqmiatbn): Likewise.
16417 (iwmmxt_wqmiattn): Likewise.
16418 (iwmmxt_wmiabb): Likewise.
16419 (iwmmxt_wmiabt): Likewise.
16420 (iwmmxt_wmiatb): Likewise.
16421 (iwmmxt_wmiatt): Likewise.
16422 (iwmmxt_wmiabbn): Likewise.
16423 (iwmmxt_wmiabtn): Likewise.
16424 (iwmmxt_wmiatbn): Likewise.
16425 (iwmmxt_wmiattn): Likewise.
16426 (iwmmxt_wmiawbb): Likewise.
16427 (iwmmxt_wmiawbt): Likewise.
16428 (iwmmxt_wmiawtb): Likewise.
16429 (iwmmxt_wmiawtt): Likewise.
16430 (iwmmxt_wmiawbbn): Likewise.
16431 (iwmmxt_wmiawbtn): Likewise.
16432 (iwmmxt_wmiawtbn): Likewise.
16433 (iwmmxt_wmiawttn): Likewise.
16434 (iwmmxt_wmerge): Likewise.
16435 (iwmmxt_tandc<mode>3): Likewise.
16436 (iwmmxt_torc<mode>3): Likewise.
16437 (iwmmxt_torvsc<mode>3): Likewise.
16438 (iwmmxt_textrc<mode>3): Likewise.
16439 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
16440 (wmmxt_pack): Likewise.
16441 (wmmxt_mult_c1): Likewise.
16442 (wmmxt_mult_c2): Likewise.
16443 (wmmxt_alu_c1): Likewise.
16444 (wmmxt_alu_c2): Likewise.
16445 (wmmxt_alu_c3): Likewise.
16446 (wmmxt_transfer_c1): Likewise.
16447 (wmmxt_transfer_c2): Likewise.
16448 (wmmxt_transfer_c3): Likewise.
16449 (marvell_f_iwmmxt_wstr): Likewise.
16450 (marvell_f_iwmmxt_wldr): Likewise.
16452 2013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
16454 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
16456 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
16459 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
16460 * lra-constraints.c (need_for_split_p): Check call used hard regs
16461 living through calls.
16463 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
16464 call used regs for call insn.
16466 2013-06-28 Jakub Jelinek <jakub@redhat.com>
16469 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
16470 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
16472 2013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
16474 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
16476 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
16478 * lra-constraints.c (need_for_split_p): Check call used hard regs
16479 living through calls.
16481 2013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
16484 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
16485 to tie with any other modes. Eliminate Altivec vector mode tests,
16486 since these are a subset of ALTIVEC or VSX vector modes. Simplify
16487 code, to return 0 if testing MODE2 for a condition, if we've
16488 already tested MODE1 for the same condition.
16490 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
16492 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
16495 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
16497 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
16498 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
16500 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
16502 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
16504 (aarch64_symbolic_constant_p): Remove.
16505 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
16506 static. Fix line length and white space.
16507 (aarch64_symbolic_constant_p): Remove.
16508 * config/aarch64/predicates.md (aarch64_valid_symref):
16509 Use aarch64_classify_symbol_expression.
16511 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16513 * config/arm/constraints.md (Ts): New constraint.
16514 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
16516 (compare_scc): Use "Ts" constraint for operand 0.
16517 (ior_scc_scc): Likewise.
16518 (and_scc_scc): Likewise.
16519 (and_scc_scc_nodom): Likewise.
16520 (ior_scc_scc_cmp): Likewise for operand 7.
16521 (and_scc_scc_cmp): Likewise.
16522 * config/arm/thumb2.md (thumb2_movsi_insn):
16523 Add alternatives for 16-bit encodings.
16524 (thumb2_movhi_insn): Likewise.
16525 (thumb2_movsicc_insn): Likewise.
16526 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
16527 (thumb2_negscc): Use "Ts" constraint.
16528 Move mvn instruction outside cond_exec block.
16529 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
16530 for 16-bit encodings.
16532 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16534 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
16536 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
16537 (mulsi3subsi): Likewise.
16538 (mulsidi3adddi): Likewise.
16539 (mulsidi3_v6): Likewise.
16540 (umulsidi3_v6): Likewise.
16541 (umulsidi3adddi_v6): Likewise.
16542 (smulsi3_highpart_v6): Likewise.
16543 (umulsi3_highpart_v6): Likewise.
16544 (mulhisi3tb): Likewise.
16545 (mulhisi3bt): Likewise.
16546 (mulhisi3tt): Likewise.
16547 (maddhisi4): Likewise.
16548 (maddhisi4tb): Likewise.
16549 (maddhisi4tt): Likewise.
16550 (maddhidi4): Likewise.
16551 (maddhidi4tb): Likewise.
16552 (maddhidi4tt): Likewise.
16553 (zeroextractsi_compare0_scratch): Likewise.
16554 (insv_zero): Likewise.
16555 (insv_t2): Likewise.
16556 (anddi_notzesidi_di): Likewise.
16557 (anddi_notsesidi_di): Likewise.
16558 (andsi_notsi_si): Likewise.
16559 (iordi_zesidi_di): Likewise.
16560 (xordi_zesidi_di): Likewise.
16561 (andsi_iorsi3_notsi): Likewise.
16562 (smax_0): Likewise.
16563 (smax_m1): Likewise.
16564 (smin_0): Likewise.
16565 (not_shiftsi): Likewise.
16566 (unaligned_loadsi): Likewise.
16567 (unaligned_loadhis): Likewise.
16568 (unaligned_loadhiu): Likewise.
16569 (unaligned_storesi): Likewise.
16570 (unaligned_storehi): Likewise.
16571 (extv_reg): Likewise.
16572 (extzv_t2): Likewise.
16573 (divsi3): Likewise.
16574 (udivsi3): Likewise.
16575 (arm_zero_extendhisi2addsi): Likewise.
16576 (arm_zero_extendqisi2addsi): Likewise.
16577 (compareqi_eq0): Likewise.
16578 (arm_extendhisi2_v6): Likewise.
16579 (arm_extendqisi2addsi): Likewise.
16580 (arm_movt): Likewise.
16581 (thumb2_ldrd): Likewise.
16582 (thumb2_ldrd_base): Likewise.
16583 (thumb2_ldrd_base_neg): Likewise.
16584 (thumb2_strd): Likewise.
16585 (thumb2_strd_base): Likewise.
16586 (thumb2_strd_base_neg): Likewise.
16587 (arm_negsi2): Add alternative for 16-bit encoding.
16588 (arm_one_cmplsi2): Likewise.
16590 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16592 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
16593 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
16594 (movdfcc): Likewise.
16595 * config/arm/vfp.md (*thumb2_movsf_vfp):
16596 Disable predication for arm_restrict_it.
16597 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
16598 (*thumb2_movdfcc_vfp): Likewise.
16599 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
16600 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
16601 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
16602 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
16603 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
16604 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
16605 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
16606 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
16607 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
16608 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
16609 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
16610 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
16611 Disable predication for arm_restrict_it.
16613 2013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
16615 * config/i386/bmiintrin.h (_bextr_u32): New.
16616 (_bextr_u64): Ditto.
16618 2013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
16620 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
16621 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
16622 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
16623 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
16624 * config/mips/n32-elf.h: ...this new file.
16626 2013-06-27 Marc Glisse <marc.glisse@inria.fr>
16629 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
16630 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
16632 2013-06-27 Catherine Moore <clm@codesourcery.com>
16634 * config/mips/mips-tables.opt: Regenerate.
16635 * config/mips/mips-cpus.def: Add m14ke and m14kec.
16636 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
16637 * doc/invoke.texi: Add -m14kc.
16639 2013-06-27 Jakub Jelinek <jakub@redhat.com>
16642 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
16643 constraints of operand 1 and 2.
16646 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
16647 to match RTL canonicalization. Swap predicates and
16648 constraints of operand 1 and 2.
16650 2013-06-27 Vladimir Makarov <vmakarov@redhat.com>
16652 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
16653 Process OP_INOUT regs for splitting too.
16655 2013-06-27 Jakub Jelinek <jakub@redhat.com>
16657 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
16658 decl before the loop, initialize to NULL.
16659 (vectorizable_load): Initialize ptr_incr to NULL.
16661 2013-06-27 Martin Jambor <mjambor@suse.cz>
16664 * ipa-ref.h (ipa_maybe_record_reference): Declare.
16665 * ipa-ref.c (ipa_maybe_record_reference): New function.
16666 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
16667 * ipa-cp.c (create_specialized_node): Record potential references from
16669 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
16671 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
16673 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
16674 parameter 'mode' of type 'enum machine_mode mode'; change to pass
16675 'mode' to force_reg.
16676 (aarch64_add_offset): Update calls to aarch64_force_temporary.
16677 (aarch64_expand_mov_immediate): Likewise.
16679 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
16681 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
16682 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
16684 2013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16686 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
16687 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
16688 (struct machine_function): Add tbegin_p.
16689 (s390_canonicalize_comparison): Fold CC mode compares to
16690 conditional jump if possible.
16691 (s390_emit_jump): Return the emitted jump.
16692 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
16693 Handle CCRAWmode compares.
16694 (s390_option_override): Default to -mhtm if available.
16695 (s390_reg_clobbered_rtx): Handle floating point regs as well.
16696 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
16697 FPRs instead of df_regs_ever_live_p.
16698 (s390_optimize_nonescaping_tx): New function.
16699 (s390_init_frame_layout): Extend clobbered_regs array to cover
16701 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
16702 (s390_expand_tbegin): New function.
16703 (enum s390_builtin): New enum definition.
16704 (code_for_builtin): New array definition.
16705 (s390_init_builtins): New function.
16706 (s390_expand_builtin): New function.
16707 (TARGET_INIT_BUILTINS): Define.
16708 (TARGET_EXPAND_BUILTIN): Define.
16709 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
16710 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
16711 (s390_alc_comparison): Likewise.
16712 * config/s390/s390-modes.def: Add CCRAWmode.
16713 * config/s390/s390.h (processor_flags): Add PF_TX.
16714 (TARGET_CPU_HTM): Define macro.
16715 (TARGET_HTM): Define macro.
16716 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
16717 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
16718 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
16719 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
16720 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
16721 (TBEGIN_MASK, TBEGINC_MASK): New constants.
16722 ("*cc_to_int"): Move up.
16723 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
16724 constants other than 0.
16725 ("*ccraw_to_int"): New insn and splitter definition.
16726 ("tbegin", "tbegin_nofloat", "tbegin_retry")
16727 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
16728 ("tx_assist"): New expander.
16729 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
16730 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
16731 * config/s390/s390.opt: Add -mhtm option.
16732 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
16733 * config/s390/htmxlintrin.h: New file.
16734 * config/s390/htmintrin.h: New file.
16735 * config/s390/s390intrin.h: New file.
16736 * doc/extend.texi: Document htm builtins.
16737 * config.gcc: Add the new header files to extra_headers.
16739 2013-06-26 Thomas Schwinge <thomas@codesourcery.com>
16741 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
16742 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
16744 2013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
16745 Pat Haugen <pthaugen@us.ibm.com>
16746 Peter Bergner <bergner@vnet.ibm.com>
16748 * config/rs6000/power8.md: New.
16749 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
16750 setting for power8 entry.
16751 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
16752 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
16753 test for Power4/Power5 only.
16754 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
16756 (force_new_group): Adjust comment.
16757 * config/rs6000/rs6000.md: Include power8.md.
16759 2013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
16761 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
16762 * config/arm/arm-protos.h (arm_max_conditional_execute): New
16764 (tune_params): Update comment.
16765 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
16766 (arm_max_conditional_execute): New function.
16767 (thumb2_final_prescan_insn): Use max_insn_skipped and
16768 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
16770 2013-06-25 Jakub Jelinek <jakub@redhat.com>
16772 PR tree-optimization/57705
16773 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
16774 SSA_NAME step, provided that it is not defined inside the loop.
16775 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
16776 (get_initial_def_for_induction): Handle SSA_NAME IV step.
16778 2013-06-25 Martin Jambor <mjambor@suse.cz>
16780 PR middle-end/57670
16781 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
16782 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
16784 (ipa_note_param_call): Initialize member_ptr flag.
16785 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
16786 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
16787 (ipa_write_indirect_edge_info): Stream member_ptr flag.
16788 (ipa_read_indirect_edge_info): Likewise.
16790 2013-06-25 Richard Biener <rguenther@suse.de>
16792 PR middle-end/56977
16793 * passes.c (init_optimization_passes): Move pass_fold_builtins
16794 and pass_dce earlier with -Og.
16796 2013-06-25 Eric Botcazou <ebotcazou@adacore.com>
16798 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
16799 <BIT_FIELD_REF>: Remove trailing TAB.
16800 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
16803 2013-06-24 Martin Jambor <mjambor@suse.cz>
16805 PR tree-optimization/57358
16806 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
16807 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
16808 (ipa_analyze_params_uses): Generate pessimistic info when true.
16810 2013-06-24 Martin Jambor <mjambor@suse.cz>
16812 PR tree-optimization/57539
16813 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
16814 global.inlined_to of the new node to it. All callers changed.
16815 * ipa-inline-transform.c (clone_inlined_nodes): New variable
16816 inlining_into, pass it to cgraph_clone_node.
16817 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
16818 ipa_free_edge_args_substructures.
16819 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
16820 rdesc linked list. Do not assert rdesc edges have inlined caller.
16821 Assert we have found an rdesc in the rdesc list.
16823 2013-06-24 Richard Biener <rguenther@suse.de>
16825 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
16826 (pointer_set_lookup): Declare.
16827 (class pointer_map): New template class implementing a
16828 generic pointer to T map.
16829 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
16830 pointer_map<T>::contains, pointer_map<T>::insert,
16831 pointer_map<T>::traverse): New functions.
16832 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
16833 (pointer_set_lookup): New function.
16834 (pointer_set_contains): Use pointer_set_lookup.
16835 (pointer_set_insert): Likewise.
16836 (insert_aux): Remove.
16837 (struct pointer_map_t): Embed a pointer_set_t.
16838 (pointer_map_create): Adjust.
16839 (pointer_map_destroy): Likewise.
16840 (pointer_map_contains): Likewise.
16841 (pointer_map_insert): Likewise.
16842 (pointer_map_traverse): Likewise.
16843 * tree-streamer.h (struct streamer_tree_cache_d): Use a
16844 pointer_map<unsigned> instead of a pointer_map_t.
16845 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
16846 (streamer_tree_cache_lookup): Likewise.
16847 (streamer_tree_cache_create): Likewise.
16848 (streamer_tree_cache_delete): Likewise.
16849 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
16850 pointer_map<unsigned> instead of a pointer_map_t.
16851 (lto_init_tree_ref_encoder): Adjust.
16852 (lto_destroy_tree_ref_encoder): Likewise.
16853 * lto-section-out.c (lto_output_decl_index): Likewise.
16854 (lto_record_function_out_decl_state): Likewise.
16855 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
16857 2013-06-24 Richard Biener <rguenther@suse.de>
16859 PR tree-optimization/57488
16860 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
16862 2013-06-24 Alan Modra <amodra@gmail.com>
16864 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
16865 (gen_easy_altivec_constant): Likewise.
16866 * config/rs6000/predicates.md (easy_vector_constant_add_self,
16867 easy_vector_constant_msb): Likewise.
16869 2013-06-23 Jakub Jelinek <jakub@redhat.com>
16872 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
16873 add missing return true.
16875 2013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
16878 * config/sh/predicates.md (general_extend_operand): Invoke
16879 general_movsrc_operand for memory operands.
16880 (general_movsrc_operand): Allow reg+reg addressing, do not use
16881 general_operand for memory operands.
16883 2013-06-23 Sriraman Tallam <tmsriram@google.com>
16885 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
16886 when current target options does not apply.
16887 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
16888 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
16889 * config/i386/bmiintrin.h: Pass appropriate target
16890 attributes to header.
16891 * config/i386/mmintrin.h: Ditto.
16892 * config/i386/nmmintrin.h: Ditto.
16893 * config/i386/avx2intrin.h: Ditto.
16894 * config/i386/fxsrintrin.h: Ditto.
16895 * config/i386/tbmintrin.h: Ditto.
16896 * config/i386/xsaveintrin.h: Ditto.
16897 * config/i386/f16cintrin.h: Ditto.
16898 * config/i386/xtestintrin.h: Ditto.
16899 * config/i386/xsaveoptintrin.h: Ditto.
16900 * config/i386/bmi2intrin.h: Ditto.
16901 * config/i386/lzcntintrin.h: Ditto.
16902 * config/i386/smmintrin.h: Ditto.
16903 * config/i386/wmmintrin.h: Ditto.
16904 * config/i386/x86intrin.h: Remove all header include guards.
16905 * config/i386/prfchwintrin.h: Ditto.
16906 * config/i386/pmmintrin.h: Ditto.
16907 * config/i386/tmmintrin.h: Ditto.
16908 * config/i386/xmmintrin.h: Ditto.
16909 * config/i386/popcntintrin.h: Ditto.
16910 * config/i386/rdseedintrin.h: Ditto.
16911 * config/i386/ammintrin.h: Ditto.
16912 * config/i386/emmintrin.h: Ditto.
16913 * config/i386/immintrin.h: Remove all header include guards.
16914 * config/i386/fma4intrin.h: Ditto.
16915 * config/i386/lwpintrin.h: Ditto.
16916 * config/i386/xopintrin.h: Ditto.
16917 * config/i386/ia32intrin.h: Ditto.
16918 * config/i386/avxintrin.h: Ditto.
16919 * config/i386/rtmintrin.h: Ditto.
16920 * config/i386/fmaintrin.h: Ditto.
16921 * config/i386/mm3dnow.h: Ditto.
16923 2013-06-22 Sriraman Tallam <tmsriram@google.com>
16925 * common/config/i386/i386-common.c: Handle LZCNT.
16927 2013-06-22 Andi Kleen <ak@linux.intel.com>
16929 * doc/extend.texi: Use __atomic_store_n instead of
16930 __atomic_store in HLE example.
16932 2013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
16934 * config/sh/sh.c: Remove <cstdlib> workaround.
16936 2013-06-21 Andi Kleen <ak@linux.intel.com>
16938 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
16940 2013-06-21 Andi Kleen <ak@linux.intel.com>
16942 * doc/extend.texi: Document that __atomic_clear and
16943 __atomic_test_and_set should only be used with bool.
16945 2013-06-20 Jan Hubicka <jh@suse.cz>
16947 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
16948 types_same_for_odr.
16949 * tree.c (decls_same_for_odr): New function.
16950 (same_for_edr): New function.
16951 (types_same_for_odr): New function.
16952 (get_binfo_at_offset): Use it.
16953 * tree.h (types_same_for_odr): Declare.
16955 2013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
16956 Jason Merrill <jason@redhat.com>
16958 * system.h: Include <cstdlib> as well as <stdlib.h>.
16960 2013-06-20 Uros Bizjak <ubizjak@gmail.com>
16963 * config/i386/i386.c (construct_container): Report error if
16964 long double is used with disabled x87 float returns.
16966 2013-06-20 Jan Hubicka <jh@suse.cz>
16968 * lto-cgraph.c (input_symtab): Do not set cgraph state.
16970 2013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
16972 PR rtl-optimization/57425
16973 PR rtl-optimization/57569
16974 * alias.c (write_dependence_p): Remove parameters mem_mode and
16975 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
16976 Changed all callers.
16977 (canon_anti_dependence): Get comments and semantics in sync.
16978 Add parameter mem_canonicalized. Changed all callers.
16979 * rtl.h (canon_anti_dependence): Update prototype.
16981 2013-06-20 Richard Biener <rguenther@suse.de>
16983 * data-streamer-in.c (streamer_read_uhwi): Optimize single
16984 byte case, inline streamer_read_uchar and defer section
16987 2013-06-20 Richard Biener <rguenther@suse.de>
16989 PR tree-optimization/57584
16990 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
16991 SSA names into the expanded expression that take part in
16992 abnormal coalescing.
16994 2013-06-19 Sharad Singhai <singhai@google.com>
16996 * gcov.c (print_usage): Handle new option.
16997 (process_args): Ditto.
16998 (get_gcov_intermediate_filename): New function.
16999 (output_intermediate_file): New function.
17000 (output_gcov_file): New function
17001 (generate_results): Handle new option.
17002 (release_function): Relase demangled name.
17003 (read_graph_file): Handle demangled name.
17004 (output_lines): Ditto.
17005 * doc/gcov.texi: Document gcov intermediate format.
17007 2013-06-19 Vladimir Makarov <vmakarov@redhat.com>
17010 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
17011 (lra_emit_add): Use the functions. Add comment about Y as an
17014 2013-06-19 David Edelsohn <dje.gcc@gmail.com>
17017 * collect2.c (collect_atexit): New.
17018 (collect_exit): Delete.
17019 (main): Register collect_atexit with atexit.
17020 (collect_wait): Change collect_exit to exit.
17022 * collect2.h (collect_exit): Delete.
17023 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
17025 2013-06-19 Wei Mi <wmi@google.com>
17027 PR rtl-optimization/57518
17028 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
17029 if regno is used in paradoxical subreg.
17030 (update_equiv_regs): Check pdx_subregs[regno] before
17031 set a reg to be equivalent with a mem.
17033 2013-06-19 Matthias Klose <doko@ubuntu.com>
17036 * file-find.h (find_a_file): Add a mode parameter.
17037 * file-find.c (find_a_file): Likewise.
17038 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
17039 with X_OK for the executables.
17040 * collect2.c (main): Call find_a_file with X_OK.
17042 2013-06-19 Steve Ellcey <sellcey@mips.com>
17045 * config/mips/mips.md (casesi_internal_mips16_<mode>):
17046 Use NEXT_INSN instead of next_real_insn.
17048 2013-06-19 Jan Hubicka <jh@suse.cz>
17050 * cgraph.h (const_value_known_p): Replace by ...
17051 (ctor_for_folding): .. this one.
17052 * cgraphunit.c (process_function_and_variable_attributes): Use it.
17053 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
17054 * expr.c (expand_expr_real_1): Likewise.
17055 (string_constant): Likewise.
17056 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
17057 * ipa.c (process_references): Likewise.
17058 (symtab_remove_unreachable_nodes): Likewise.
17059 * ipa-inline-analysis.c (param_change_prob): Likewise.
17060 * gimple-fold.c (canonicalize_constructor_val): Likewise.
17061 (get_base_constructor): Likwise.
17062 * varpool.c (varpool_remove_node): Likewise.
17063 (varpool_remove_initializer): LIkewise.
17064 (dump_varpool_node): LIkwise.
17065 (const_value_known_p): Rewrite to ...
17066 (ctor_for_folding): ... this one.
17068 2013-06-19 Jakub Jelinek <jakub@redhat.com>
17071 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
17072 PERSONALITY in $PATH derived prefixes.
17074 2013-06-19 Jeff Law <law@redhat.com>
17076 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
17079 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
17080 (simplify_bitwise_binary): Use it to simpify certain binary ops on
17083 2013-06-19 Sofiane Naci <sofiane.naci@arm.com>
17085 * config/arm/vfp.md: Move VFP instruction classification documentation
17087 * config/arm/arm.md: ... here. Update instruction classification
17090 2013-06-19 Richard Earnshaw <rearnsha@arm.com>
17092 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
17093 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
17094 pattern. Use more efficient sequences on ARMv5 and Thumb2.
17096 2013-06-19 Steven Bosscher <steven@gcc.gnu.org>
17099 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
17100 with NEXT_INSN. Use tablejump_p to check for jump table data
17103 2013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
17106 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
17107 that now in C++ the value is correct per the C++ standards.
17109 2013-06-19 Richard Biener <rguenther@suse.de>
17111 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
17112 for global context.
17114 2013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17117 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17120 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
17121 with next_active_insn.
17123 2013-06-18 Sriraman Tallam <tmsriram@google.com>
17125 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
17126 functions are inlined during failures to flag an error.
17127 * tree-inline.c (expand_call_inline): Allow the error to be flagged
17128 in early inline pass.
17130 2013-06-18 H.J. Lu <hongjiu.lu@intel.com>
17132 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
17135 2013-06-18 Julian Brown <julian@codesourcery.com>
17137 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
17138 Permit virtual register pre-reload if !strict.
17139 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
17141 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
17143 * config/arm/neon.md (movmisalign<mode>): Use
17144 neon_perm_struct_or_reg_operand instead of
17145 neon_struct_or_register_operand.
17146 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
17147 neon_permissive_struct_operand instead of neon_struct_operand.
17148 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
17149 neon_vector_mem_operand.
17150 * config/arm/predicates.md (neon_struct_operand): Adjust call to
17151 neon_vector_mem_operand.
17152 (neon_permissive_struct_operand): New.
17153 (neon_struct_or_register_operand): Rename to...
17154 (neon_perm_struct_or_reg_operand): This. Adjust call to
17155 neon_vector_mem_operand.
17157 2013-06-18 Richard Biener <rguenther@suse.de>
17159 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
17160 * lto-streamer.h: Include pointer-set.h.
17161 (struct lto_decl_slot): Remove.
17162 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
17163 Remove next_index entry.
17164 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
17165 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
17166 (lto_init_tree_ref_encoder): Adjust.
17167 (lto_destroy_tree_ref_encoder): Likewise.
17168 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
17169 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
17170 (lto_output_decl_index): Adjust.
17171 (lto_new_out_decl_state): Likewise.
17172 (lto_record_function_out_decl_state): Likewise.
17173 * lto-streamer-out.c (copy_function): Likewise.
17175 2013-06-18 Richard Biener <rguenther@suse.de>
17177 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
17178 * cgraphunit.c: Include cfgloop.h.
17179 (init_lowered_empty_function): Initialize the loop tree.
17180 (assemble_thunk): Insert new BBs into loops.
17182 2013-06-18 Richard Biener <rguenther@suse.de>
17184 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
17185 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
17186 the map from cache entry to cache index optional.
17187 (streamer_tree_cache_replace_tree): Adjust accordingly.
17188 (streamer_tree_cache_append): Likewise.
17189 (streamer_tree_cache_delete): Likewise.
17190 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
17191 streamer cache map from cache entry to cache index.
17192 * lto-streamer-out.c (create_output_block): Adjust.
17194 2013-06-18 Sofiane Naci <sofiane.naci@arm.com>
17196 * config/arm/arm.md (attribute "insn"): Move multiplication and
17197 division attributes to...
17198 (attribute "type"): ... here. Remove mult.
17199 (attribute "mul32"): New attribute.
17200 (attribute "mul64"): Add umaal.
17201 (*arm_mulsi3): Update attributes.
17202 (*arm_mulsi3_v6): Likewise.
17203 (*thumb_mulsi3): Likewise.
17204 (*thumb_mulsi3_v6): Likewise.
17205 (*mulsi3_compare0): Likewise.
17206 (*mulsi3_compare0_v6): Likewise.
17207 (*mulsi_compare0_scratch): Likewise.
17208 (*mulsi_compare0_scratch_v6): Likewise.
17209 (*mulsi3addsi): Likewise.
17210 (*mulsi3addsi_v6): Likewise.
17211 (*mulsi3addsi_compare0): Likewise.
17212 (*mulsi3addsi_compare0_v6): Likewise.
17213 (*mulsi3addsi_compare0_scratch): Likewise.
17214 (*mulsi3addsi_compare0_scratch_v6): Likewise.
17215 (*mulsi3subsi): Likewise.
17216 (*mulsidi3adddi): Likewise.
17217 (*mulsi3addsi_v6): Likewise.
17218 (*mulsidi3adddi_v6): Likewise.
17219 (*mulsidi3_nov6): Likewise.
17220 (*mulsidi3_v6): Likewise.
17221 (*umulsidi3_nov6): Likewise.
17222 (*umulsidi3_v6): Likewise.
17223 (*umulsidi3adddi): Likewise.
17224 (*umulsidi3adddi_v6): Likewise.
17225 (*smulsi3_highpart_nov6): Likewise.
17226 (*smulsi3_highpart_v6): Likewise.
17227 (*umulsi3_highpart_nov6): Likewise.
17228 (*umulsi3_highpart_v6): Likewise.
17229 (mulhisi3): Likewise.
17230 (*mulhisi3tb): Likewise.
17231 (*mulhisi3bt): Likewise.
17232 (*mulhisi3tt): Likewise.
17233 (maddhisi4): Likewise.
17234 (*maddhisi4tb): Likewise.
17235 (*maddhisi4tt): Likewise.
17236 (maddhidi4): Likewise.
17237 (*maddhidi4tb): Likewise.
17238 (*maddhidi4tt): Likewise.
17239 (divsi3): Likewise.
17240 (udivsi3): Likewise.
17241 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
17242 (thumb2_mulsi_short_compare0): Likewise.
17243 (thumb2_mulsi_short_compare0_scratch): Likewise.
17244 * config/arm/arm1020e.md (1020mult1): Update attribute change.
17245 (1020mult2): Likewise.
17246 (1020mult3): Likewise.
17247 (1020mult4): Likewise.
17248 (1020mult5): Likewise.
17249 (1020mult6): Likewise.
17250 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
17252 (cortex_a15_mult64): Likewise.
17253 (cortex_a15_sdiv): Likewise.
17254 (cortex_a15_udiv): Likewise.
17255 * config/arm/arm1026ejs.md (mult1): Update attribute change.
17261 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
17262 (pj4_ir_div): Likewise.
17263 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
17264 (11_mult2): Likewise.
17265 (11_mult3): Likewise.
17266 (11_mult4): Likewise.
17267 (11_mult5): Likewise.
17268 (11_mult6): Likewise.
17269 (11_mult7): Likewise.
17270 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
17271 (cortex_a8_mla): Likewise.
17272 (cortex_a8_mull): Likewise.
17273 (cortex_a8_smulwy): Likewise.
17274 (cortex_a8_smlald): Likewise.
17275 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
17276 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
17277 (cortex_r4_mul_3): Likewise.
17278 (cortex_r4_mla_4): Likewise.
17279 (cortex_r4_mla_3): Likewise.
17280 (cortex_r4_smlald): Likewise.
17281 (cortex_r4_mull): Likewise.
17282 (cortex_r4_sdiv): Likewise.
17283 (cortex_r4_udiv): Likewise.
17284 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
17285 (cortex_a7_idiv): Likewise.
17286 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
17287 (9_mult2): Likewise.
17288 (9_mult3): Likewise.
17289 (9_mult4): Likewise.
17290 (9_mult5): Likewise.
17291 (9_mult6): Likewise.
17292 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
17293 (cortex_a53_sdiv): Likewise.
17294 (cortex_a53_udiv): Likewise.
17295 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
17296 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
17297 (mp626_mult2): Likewise.
17298 (mp626_mult3): Likewise.
17299 (mp626_mult4): Likewise.
17300 * config/arm/fa526.md (526_mult1): Update attribute change.
17301 (526_mult2): Likewise.
17302 * config/arm/arm-generic.md (mult): Update attribute change.
17303 (mult_ldsched_strongarm): Likewise.
17304 (mult_ldsched): Likewise.
17305 (multi_cycle): Likewise.
17306 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
17307 * config/arm/fa606te.md (606te_mult1): Update attribute change.
17308 (606te_mult2): Likewise.
17309 (606te_mult3): Likewise.
17310 (606te_mult4): Likewise.
17311 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
17312 (cortex_a9_mac16): Likewise.
17313 (cortex_a9_multiply): Likewise.
17314 (cortex_a9_mac): Likewise.
17315 (cortex_a9_multiply_long): Likewise.
17316 * config/arm/fa626te.md (626te_mult1): Update attribute change.
17317 (626te_mult2): Likewise.
17318 (626te_mult3): Likewise.
17319 (626te_mult4): Likewise.
17321 2013-06-18 Richard Biener <rguenther@suse.de>
17324 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
17326 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17329 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
17330 with next_active_insn.
17332 2013-06-18 Alan Modra <amodra@gmail.com>
17334 * config/rs6000/rs6000.h (enum data_align): New.
17335 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
17336 (DATA_ABI_ALIGNMENT): Define.
17337 (CONSTANT_ALIGNMENT): Correct comment.
17338 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
17339 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
17341 2013-06-17 David Malcolm <dmalcolm@redhat.com>
17343 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
17344 ATTRIBUTE_UNUSED marking.
17346 2013-06-17 Sofiane Naci <sofiane.naci@arm.com>
17348 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
17349 alternative and update.
17350 (aarch64_dup_lanedi): Delete.
17351 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
17352 * config/aarch64/aarch64-simd-builtins.def: Update.
17354 2013-06-17 Richard Biener <rguenther@suse.de>
17356 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
17357 (lto_input_scc): Declare.
17358 (lto_input_tree_1): Likewise.
17359 (struct lto_stats_d): Add num_tree_bodies_output and
17360 num_pickle_refs_output.
17361 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
17362 (lto_read_tree_1): Split out from ...
17363 (lto_read_tree): ... this.
17364 (lto_input_scc): New function.
17365 (lto_input_tree_1): Split out from ...
17366 (lto_input_tree): ... this. Handle LTO_tree_scc.
17367 (lto_data_in_create): Create the streamer cache without hashes.
17368 * lto-streamer-out.c (create_output_block): Create the streamer
17369 cache with hashes when not doing WPA.
17370 (lto_write_tree_1): Split out from ...
17371 (lto_write_tree): ... this.
17372 (get_symbol_initial_value): New function.
17373 (lto_output_tree_1): Split out from ...
17374 (lto_output_tree): ... this. Write trees as series of SCCs
17375 using a DFS walk via DFS_write_tree.
17376 (struct sccs, struct scc_entry): New types.
17377 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
17378 (DFS_write_tree_body): New function.
17379 (DFS_write_tree): Likewise.
17380 (hash_tree): Likewise.
17381 (scc_entry_compare): Likewise.
17382 (hash_scc): Likewise.
17383 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
17384 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
17385 TREE_CHAIN as regular reference.
17386 (streamer_read_integer_cst): Remove.
17387 (streamer_get_pickled_tree): Adjust.
17388 * tree-streamer-out.c (streamer_write_chain): Disable streaming
17389 of DECL_EXTERNALs in BLOCK_VARS for now.
17390 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
17392 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
17393 Add hash value argument and record that if hashes are recorded
17395 (streamer_tree_cache_insert_1): Adjust.
17396 (streamer_tree_cache_insert): Likewise.
17397 (streamer_tree_cache_insert_at): Rename to ...
17398 (streamer_tree_cache_replace_tree): ... this and adjust.
17399 (streamer_tree_cache_append): Adjust.
17400 (record_common_node): Likewise.
17401 (streamer_tree_cache_create): Add argument whether to
17402 record hash values together with trees.
17403 (streamer_tree_cache_delete): Adjust.
17404 * tree-streamer.h (struct streamer_tree_cache_d): Add
17406 (streamer_read_integer_cst): Remove.
17407 (streamer_tree_cache_insert): Adjust.
17408 (streamer_tree_cache_append): Likewise.
17409 (streamer_tree_cache_insert_at): Rename to ...
17410 (streamer_tree_cache_replace_tree): ... this and adjust.
17411 (streamer_tree_cache_create): Add argument whether to record hashes.
17412 (streamer_tree_cache_get): Rename to ...
17413 (streamer_tree_cache_get_tree): ... this.
17414 (streamer_tree_cache_get_hash): New function.
17415 * tree.c (cache_integer_cst): New function.
17416 * tree.h (cache_integer_cst): Declare.
17417 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
17418 * lto-symtab.c (lto_varpool_replace_node): Only release
17419 DECL_INITIAL of non-prevailing decls.
17420 * varpool.c (varpool_remove_initializer): Do not release
17421 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
17423 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
17425 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
17426 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
17427 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
17428 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
17429 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
17430 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
17431 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
17432 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
17433 instead of TARGET_64BIT.
17434 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
17435 Require ISA_HAS_<D>DIV.
17437 2013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
17439 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
17440 (mips*-*-linux*): Move default with_llsc setting to where other
17442 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
17444 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
17445 Likewise. Remove default with_tune setting. Move default float
17446 setting to its own block. Handle with_llsc in the same block as above.
17448 2013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
17450 PR rtl-optimization/57425
17451 PR rtl-optimization/57569
17452 * alias.c (write_dependence_p): Add new parameters mem_mode,
17453 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
17454 Changed all callers.
17455 (canon_anti_dependence): New function.
17456 * cse.c (check_dependence): Use canon_anti_dependence.
17457 * cselib.c (cselib_invalidate_mem): Likewise.
17458 * rtl.h (canon_anti_dependence): Declare.
17460 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
17462 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
17463 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
17464 ".set mips3" for 64-bit targets.
17466 2013-06-15 Dehao Chen <dehao@google.com>
17468 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
17469 * gimple-low.c (gimple_check_call_matching_types): Likewise.
17470 (gimple_check_call_args): Likewise.
17471 * value-prof.c (check_ic_target): Likewise.
17472 * ipa-inline.c (early_inliner): Likewise.
17473 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
17474 * cgraph.c (cgraph_create_edge_1): Likewise.
17475 (cgraph_make_edge_direct): Likewise.
17477 2013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
17480 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
17481 rs6000_output_move_128bit to handle emitting quad memory
17482 operations. Set attribute length to 8 bytes.
17484 2013-06-14 Vidya Praveen <vidyapraveen@arm.com>
17486 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
17488 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
17489 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
17490 (aarch64_<su>mlsl<mode>): Likewise.
17492 2013-06-14 Mike Stump <mikestump@comcast.net>
17494 * Makefile.in (TARGET_H): Add insn-codes.h.
17496 2013-06-14 Alan Modra <amodra@gmail.com>
17498 PR middle-end/57134
17499 PR middle-end/57586
17500 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
17501 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
17502 bitfield expansion when EXPAND_MEMORY.
17503 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
17505 2013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
17507 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
17508 test for clearing quad memory on 32-bit later.
17510 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
17512 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
17513 (fold_negate_expr): Likewise.
17514 (fold_real_zero_addition_p): Handle vectors.
17515 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
17517 2013-06-14 Alan Modra <amodra@gmail.com>
17519 * varasm.c (force_const_mem): Revert 2013-06-07 change.
17521 2013-06-13 Jan Hubicka <jh@suse.cz>
17523 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
17524 Local comdats are not externally visible.
17525 * symtab.c (dump_symtab_base): Dump externally visible.
17526 (verify_symtab_base): Verify back links in the symtab hash.
17528 2013-06-13 Bin Cheng <bin.cheng@arm.com>
17530 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
17531 CONVERT_EXPR as equal nodes.
17533 2013-06-13 Bin Cheng <bin.cheng@arm.com>
17535 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
17537 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
17539 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
17540 Generalize to complex and vector.
17541 * tree.c (build_all_ones_cst): New function.
17542 * tree.h (build_all_ones_cst): Declare it.
17544 2013-06-13 Alan Modra <amodra@gmail.com>
17546 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
17547 * config/rs6000/rs6000.md (signbittf2): New insn.
17548 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
17549 (abstf2_internal, cmptf_internal2): Likewise.
17550 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
17552 2013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
17553 Pat Haugen <pthaugen@us.ibm.com>
17554 Peter Bergner <bergner@vnet.ibm.com>
17556 * config/rs6000/rs6000.c (emit_load_locked): Add support for
17557 power8 byte, half-word, and quad-word atomic instructions.
17558 (emit_store_conditional): Likewise.
17559 (rs6000_expand_atomic_compare_and_swap): Likewise.
17560 (rs6000_expand_atomic_op): Likewise.
17562 * config/rs6000/sync.md (larx): Add new modes for power8.
17564 (AINT): New mode iterator to include TImode as well as normal
17565 integer modes on power8.
17566 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
17567 that VSX registers are not considered. Use AINT mode iterator
17568 instead of INT1 to allow inclusion of quad word atomic operations
17570 (load_locked<mode>): Likewise.
17571 (store_conditional<mode>): Likewise.
17572 (atomic_compare_and_swap<mode>): Likewise.
17573 (atomic_exchange<mode>): Likewise.
17574 (atomic_nand<mode>): Likewise.
17575 (atomic_fetch_<fetchop_name><mode>): Likewise.
17576 (atomic_nand_fetch<mode>): Likewise.
17577 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
17579 (ATOMIC): On power8, add QImode, HImode modes.
17580 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
17581 modes that promote to SImode.
17582 (load_lockedti): Convert TImode arguments to PTImode, so that we
17583 get a guaranteed even/odd register pair.
17584 (load_lockedpti): Likewise.
17585 (store_conditionalti): Likewise.
17586 (store_conditionalpti): Likewise.
17588 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
17589 atomic load/store instructions.
17592 2013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
17594 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
17596 (insn_count): New attribute, with most cases extracted from...
17597 (length): ...here. Redefine most cases in terms of insn_count.
17598 (single_insn): Delete.
17599 (can_delay): Use insn_count to check for single instructions.
17600 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
17601 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
17602 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
17603 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
17604 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
17605 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
17606 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
17607 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
17608 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
17609 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
17610 rather than "length".
17611 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
17612 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
17613 Use "insn_count" rather than "length".
17614 * config/mips/mips-dsp.md
17615 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
17616 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
17619 2013-06-12 Marc Glisse <marc.glisse@inria.fr>
17621 PR tree-optimization/57361
17622 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
17624 2013-06-12 Sofiane Naci <sofiane.naci@arm.com>
17626 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
17628 (aarch64_simd_combine<mode>): New instruction expansion.
17629 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
17630 function prototype.
17631 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
17632 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
17634 2013-06-12 Jan Hubicka <jh@suse.cz>
17636 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
17637 decl has when in streaming stage.
17638 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
17639 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
17641 2013-06-12 Roland Stigge <stigge@antcom.de>
17644 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
17646 2013-06-12 Jakub Jelinek <jakub@redhat.com>
17648 PR tree-optimization/57537
17649 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
17650 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
17652 2013-06-12 Richard Biener <rguenther@suse.de>
17654 * data-streamer.h (streamer_write_char_stream): CSE
17655 obs->current_pointer.
17656 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
17657 streamer_write_char_stream manually and optimize the resulting loop.
17658 (streamer_write_hwi_stream): Likewise.
17660 2013-06-12 Jan Hubicka <jh@suse.cz>
17662 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
17663 * cgraph.h (varpool_create_empty_node): Declare.
17664 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
17666 * symtab.c (symtab_unregister_node): Be lax about missin entries
17668 (symtab_get_node): Update comment.
17669 * varpool.c (varpool_create_empty_node): Break out from ...
17670 (varpool_node_for_decl): ... here.
17671 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
17673 2013-06-12 Eric Botcazou <ebotcazou@adacore.com>
17675 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
17676 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
17677 part. Use straight-line flow at the end.
17678 <COMPONENT_REF>: Remove superfluous else.
17679 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
17681 2013-06-12 Jakub Jelinek <jakub@redhat.com>
17684 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
17685 target hook even for !TREE_PUBLIC decls. If no resolution info
17686 is available, return false for common and external decls.
17688 2013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
17690 * config/rl78/constraints.md (U): New constraint.
17691 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
17694 2013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
17697 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
17698 to allow returning address to AT_PLATFORM name.
17700 2013-06-11 Jan Hubicka <jh@suse.cz>
17702 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
17703 * cgraph.h (symtab_node_base): Add weakref flag.
17704 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
17705 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
17706 (output_weakrefs): Use weakref flag.
17707 * fold-const.c (simple_operand_p): Handle WEAK.
17708 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
17709 * ipa.c (varpool_externally_visible_p): Drop weakref.
17710 (function_and_variable_visibility): Update comment; fix weakref
17711 sanity checks; do not clear DECL_WEAK on them.
17712 * lto-cgraph.c (lto_output_node): update.
17713 (lto_output_varpool_node): Update.
17714 (input_overwrite_node): Update.
17715 (input_node): Update.
17716 (input_varpool_node): Update.
17717 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
17718 (lto_symtab_merge_symbols): Add sanity check.
17719 (lto_symtab_prevailing_decl): Do not special case weakrefs.
17720 * passes.c (rest_of_decl_compilation): Set static flag, too.
17721 * symtab.c (dump_symtab_base): Dump weakref.
17722 (verify_symtab_base): Sanity check weakrefs.
17723 (symtab_make_decl_local): Remove duplicated code.
17724 (symtab_alias_ultimate_target): Simplify.
17725 * varpool.c (varpool_create_variable_alias): Set weakref flag.
17727 2013-06-11 Tom de Vries <tom@codesourcery.com>
17729 * genautomata.c (gen_regexp_sequence): Handle els_num == -1. Handle
17730 sequence_vect == NULL.
17732 2013-06-11 DJ Delorie <dj@redhat.com>
17734 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
17735 (rl78_unwind_word_mode): New.
17737 2013-06-11 David Malcolm <dmalcolm@redhat.com>
17739 * final.c (debug_prefix_maps): Make static.
17741 2013-06-11 David Malcolm <dmalcolm@redhat.com>
17743 * function.c (initial_trampoline): Remove stray copy.
17745 2013-06-11 Sofiane Naci <sofiane.naci@arm.com>
17747 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
17749 2013-06-11 Martin Jambor <mjambor@suse.cz>
17751 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
17752 within bounds at the beginning of the function.
17754 2013-06-11 Alan Modra <amodra@gmail.com>
17756 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
17758 (get_named_section): Don't NULL !DECL_P decl.
17760 2013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
17762 * doc/invoke.texi (core-avx2): Document.
17764 (atom): Updated with MOVBE.
17766 2013-06-11 Richard Biener <rguenther@suse.de>
17768 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
17770 2013-06-11 Anton Blanchard <anton@samba.org>
17772 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
17773 correct shift value in little-endian mode.
17775 2013-06-11 Jakub Jelinek <jakub@redhat.com>
17778 * varasm.c (get_variable_align): Move #endif to the right place.
17780 2013-06-10 Cary Coutant <ccoutant@google.com>
17782 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
17783 for hash so that hash table traversal order is deterministic.
17785 2013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
17786 Pat Haugen <pthaugen@us.ibm.com>
17787 Peter Bergner <bergner@vnet.ibm.com>
17789 * config/rs6000/vector.md (GPR move splitter): Do not split moves
17790 of vectors in GPRS if they are direct moves or quad word load or
17793 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
17795 (direct_move_p): Likewise.
17796 (quad_load_store_p): Likewise.
17798 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
17799 classes into bins based on the physical register type.
17800 (reg_class_to_reg_type): Likewise.
17801 (IS_STD_REG_TYPE): Likewise.
17802 (IS_FP_VECT_REG_TYPE): Likewise.
17803 (reload_fpr_gpr): Arrays to determine what insn to use if we can
17804 use direct move instructions.
17805 (reload_gpr_vsx): Likewise.
17806 (reload_vsx_gpr): Likewise.
17807 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
17808 information that is a simplification of register classes. Also
17809 precalculate direct move reload helpers.
17810 (direct_move_p): New function to return true if the operation can
17811 be done as a direct move instruciton.
17812 (quad_load_store_p): New function to return true if the operation
17813 is a quad memory operation.
17814 (rs6000_legitimize_address): If quad memory, only allow register
17815 indirect for TImode addresses.
17816 (rs6000_legitimate_address_p): Likewise.
17817 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
17818 (rs6000_reload_register_type): Likewise.
17819 (register_to_reg_type): Return register type.
17820 (rs6000_secondary_reload_simple_move): New helper function for
17821 secondary reload and secondary memory needed to identify anything
17822 that is a simple move, and does not need reloading.
17823 (rs6000_secondary_reload_direct_move): New helper function for
17824 secondary reload to identify cases that can be done with several
17825 instructions via the direct move instructions.
17826 (rs6000_secondary_reload_move): New helper function for secondary
17827 reload to identify moves between register types that can be done.
17828 (rs6000_secondary_reload): Add support for quad memory operations
17829 and for direct move.
17830 (rs6000_secondary_memory_needed): Likewise.
17831 (rs6000_debug_secondary_memory_needed): Change argument names.
17832 (rs6000_output_move_128bit): New function to return the move to
17833 use for 128-bit moves, including knowing about the various
17834 limitations of quad memory operations.
17836 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
17837 memory operations. call rs6000_output_move_128bit for the actual
17838 instruciton(s) to generate.
17839 (vsx_movti_64bit): Likewise.
17841 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
17842 (UNSPEC_P8V_MTVSRWZ): Likewise.
17843 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
17844 (UNSPEC_P8V_MTVSRD): Likewise.
17845 (UNSPEC_P8V_XXPERMDI): Likewise.
17846 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
17847 (UNSPEC_FUSION_GPR): Likewise.
17848 (FMOVE128_GPR): New iterator for direct move.
17849 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
17850 (f32_sv): Likewise.
17851 (f32_dm): Likewise.
17852 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
17853 loads and direct move instructions.
17854 (zero_extendsidi2_lfiwzx): Likewise.
17855 (extendsidi2_lfiwax): Likewise.
17856 (extendsidi2_nocell): Likewise.
17857 (floatsi<mode>2_lfiwax): Likewise.
17858 (lfiwax): Likewise.
17859 (floatunssi<mode>2_lfiwzx): Likewise.
17860 (lfiwzx): Likewise.
17861 (fix_trunc<mode>_stfiwx): Likewise.
17862 (fixuns_trunc<mode>_stfiwx): Likewise.
17863 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
17864 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
17865 (parity<mode>2_cmpb): Set length/type attr.
17866 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
17867 for 'mr.' to fast_compare.
17868 (bpermd_<mode>): Change type attr to popcnt.
17869 (p8_fmrgow_<mode>): New insns for power8 direct move support.
17870 (p8_mtvsrwz_1): Likewise.
17871 (p8_mtvsrwz_2): Likewise.
17872 (reload_fpr_from_gpr<mode>): Likewise.
17873 (p8_mtvsrd_1): Likewise.
17874 (p8_mtvsrd_2): Likewise.
17875 (p8_xxpermdi_<mode>): Likewise.
17876 (reload_vsx_from_gpr<mode>): Likewise.
17877 (reload_vsx_from_gprsf): Likewise.
17878 (p8_mfvsrd_3_<mode>): LIkewise.
17879 (reload_gpr_from_vsx<mode>): Likewise.
17880 (reload_gpr_from_vsxsf): Likewise.
17881 (p8_mfvsrd_4_disf): Likewise.
17882 (multi-word GPR splits): Do not split direct moves or quad memory
17885 2013-06-10 David Malcolm <dmalcolm@redhat.com>
17887 * tree-into-ssa.c (interesting_blocks): Make static.
17889 2013-06-10 Jakub Jelinek <jakub@redhat.com>
17892 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
17893 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
17894 Use DATA_ABI_ALIGNMENT for that case instead if defined.
17895 (get_variable_align): New function.
17896 (get_variable_section, emit_bss, emit_common,
17897 assemble_variable_contents, place_block_symbol): Use
17898 get_variable_align instead of DECL_ALIGN.
17899 (assemble_noswitch_variable): Add align argument, use it
17900 instead of DECL_ALIGN.
17901 (assemble_variable): Adjust caller. Use get_variable_align
17902 instead of DECL_ALIGN.
17903 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
17905 (DATA_ABI_ALIGNMENT): Define.
17906 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
17907 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
17908 opt is false, only return the psABI mandated alignment increase.
17909 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
17910 (DATA_ABI_ALIGNMENT): ... this.
17911 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
17912 (DATA_ABI_ALIGNMENT): ... this.
17913 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
17914 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
17915 (DATA_ABI_ALIGNMENT): ... this.
17916 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
17917 * doc/tm.texi: Regenerated.
17919 2013-06-10 Uros Bizjak <ubizjak@gmail.com>
17921 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
17922 cmp_code to construct REG_EQUAL note.
17924 2013-06-09 Jakub Jelinek <jakub@redhat.com>
17927 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
17928 that operands[2] doesn't overlap with operands[0].
17930 2013-06-09 David Edelsohn <dje.gcc@gmail.com>
17931 Jan Hubicka <jh@suse.cz>
17933 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
17934 hack to mark symbols as used.
17936 2013-06-08 Vladimir Makarov <vmakarov@redhat.com>
17938 PR rtl-optimization/57559
17939 * lra-constraints.c (process_alt_operands): Don't discourage
17940 memory with known offset for offsetable memory constraint.
17941 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
17943 2013-06-08 Eric Botcazou <ebotcazou@adacore.com>
17945 * varasm.c (struct oc_local_state): Reorder fields.
17946 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
17947 and adjust accordingly.
17948 (output_constructor): Reorder initialization code and adjust call to
17949 output_constructor_bitfield.
17951 2013-06-07 Jan Hubicka <jh@suse.cz>
17953 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
17955 2013-06-07 David Malcolm <dmalcolm@redhat.com>
17957 * tree-object-size.c (unknown): Make const.
17959 2013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17961 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
17962 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
17963 for last alternative in the cpu_facility attribute.
17965 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17968 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
17969 (xordi3): Change operand 2 constraint to arm_xordi_operand.
17970 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
17971 * config/arm/constraints.md (Dg): New constraint.
17972 * config/arm/neon.md (xordi3_neon): Remove.
17973 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
17974 * config/arm/predicates.md (arm_xordi_operand): New predicate.
17976 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17978 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
17979 Clean up alternatives.
17981 2013-06-07 Alan Modra <amodra@gmail.com>
17983 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
17986 2013-06-07 Alan Modra <amodra@gmail.com>
17988 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
17990 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17992 * config/arm/constraints.md (Df): New constraint.
17993 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
17994 Correct length attribute for last two alternatives.
17996 2013-06-07 Alan Modra <amodra@gmail.com>
17998 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
17999 override user -mfp-in-toc.
18000 (offsettable_ok_by_alignment): Consider just the current access
18001 rather than the whole object, unless BLKmode. Handle
18002 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
18003 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
18004 for -mcmodel=medium.
18005 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
18006 override user -mfp-in-toc or -msum-in-toc. Default to
18007 -mno-fp-in-toc for -mcmodel=medium.
18009 2013-06-06 DJ Delorie <dj@redhat.com>
18011 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
18012 TARGET_VALID_POINTER_MODE.
18014 2013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
18015 Pat Haugen <pthaugen@us.ibm.com>
18016 Peter Bergner <bergner@vnet.ibm.com>
18018 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
18019 Document new power8 builtins.
18021 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
18022 condition code register, to allow 128-bit logical operations to be
18023 done in the VSX or GPR registers.
18024 (nor<mode>3): Use the canonical form for nor.
18025 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
18026 vclz*, and vpopcnt* vector instructions.
18027 (nand<mode>3): Likewise.
18028 (orc<mode>3): Likewise.
18029 (clz<mode>2): LIkewise.
18030 (popcount<mode>2): Likewise.
18032 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
18033 that only the GPRs are recognized.
18035 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
18036 support for new power8 builtins.
18038 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
18040 (xscvdpspn): Likewise.
18046 (vpopcnt): Likewise.
18047 (vpopcntb): Likewise.
18048 (vpopcnth): Likewise.
18049 (vpopcntw): Likewise.
18050 (vpopcntd): Likewise.
18052 (vmrgew): Likewise.
18053 (vmrgow): Likewise.
18055 (eqv_v16qi3): Likewise.
18056 (eqv_v8hi3): Likewise.
18057 (eqv_v4si3): Likewise.
18058 (eqv_v2di3): Likewise.
18059 (eqv_v4sf3): Likewise.
18060 (eqv_v2df3): Likewise.
18062 (nand_v16qi3): Likewise.
18063 (nand_v8hi3): Likewise.
18064 (nand_v4si3): Likewise.
18065 (nand_v2di3): Likewise.
18066 (nand_v4sf3): Likewise.
18067 (nand_v2df3): Likewise.
18069 (orc_v16qi3): Likewise.
18070 (orc_v8hi3): Likewise.
18071 (orc_v4si3): Likewise.
18072 (orc_v2di3): Likewise.
18073 (orc_v4sf3): Likewise.
18074 (orc_v2df3): Likewise.
18076 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
18077 allow power8 quad mode in 64-bit.
18078 (rs6000_builtin_vectorized_function): Add support to vectorize
18079 ISA 2.07 count leading zeros, population count builtins.
18080 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
18081 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
18082 (builtin_function_type): Add vgbbd builtin function which takes an
18084 (altivec_expand_vec_perm_const): Add support for new power8 merge
18087 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
18088 that does not include TImdoe for use with 32-bit.
18089 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
18091 (UNSPEC_VSX_CVDPSPN): Likewise.
18092 (vsx_xscvdpspn): Likewise.
18093 (vsx_xscvspdpn): Likewise.
18094 (vsx_xscvdpspn_scalar): Likewise.
18095 (vsx_xscvspdpn_directmove): Likewise.
18096 (vsx_and<mode>3): Split logical operations into 32-bit and
18097 64-bit. Add support to do logical operations on TImode as well as
18098 VSX vector types. Allow logical operations to be done in either
18099 VSX registers or in general purpose registers in 64-bit mode. Add
18100 splitters if GPRs were used. For AND, add clobber of CCmode to
18101 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
18103 (vsx_and<mode>3_32bit): Likewise.
18104 (vsx_and<mode>3_64bit): Likewise.
18105 (vsx_ior<mode>3): Likewise.
18106 (vsx_ior<mode>3_32bit): Likewise.
18107 (vsx_ior<mode>3_64bit): Likewise.
18108 (vsx_xor<mode>3): Likewise.
18109 (vsx_xor<mode>3_32bit): Likewise.
18110 (vsx_xor<mode>3_64bit): Likewise.
18111 (vsx_one_cmpl<mode>2): Likewise.
18112 (vsx_one_cmpl<mode>2_32bit): Likewise.
18113 (vsx_one_cmpl<mode>2_64bit): Likewise.
18114 (vsx_nor<mode>3): Likewise.
18115 (vsx_nor<mode>3_32bit): Likewise.
18116 (vsx_nor<mode>3_64bit): Likewise.
18117 (vsx_andc<mode>3): Likewise.
18118 (vsx_andc<mode>3_32bit): Likewise.
18119 (vsx_andc<mode>3_64bit): Likewise.
18120 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
18121 and xxlorc instructions.
18122 (vsx_eqv<mode>3_64bit): Likewise.
18123 (vsx_nand<mode>3_32bit): Likewise.
18124 (vsx_nand<mode>3_64bit): Likewise.
18125 (vsx_orc<mode>3_32bit): Likewise.
18126 (vsx_orc<mode>3_64bit): Likewise.
18128 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
18130 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
18132 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
18133 (p8_vmrgow): Likewise.
18134 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
18135 GPRs to be split under VSX.
18136 (p8v_clz<mode>2): Add power8 count leading zero support.
18137 (p8v_popcount<mode>2): Add power8 population count support.
18138 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
18141 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
18144 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
18146 (vec_nand): Likewise.
18147 (vec_vclz): Likewise.
18148 (vec_vclzb): Likewise.
18149 (vec_vclzd): Likewise.
18150 (vec_vclzh): Likewise.
18151 (vec_vclzw): Likewise.
18152 (vec_vgbbd): Likewise.
18153 (vec_vmrgew): Likewise.
18154 (vec_vmrgow): Likewise.
18155 (vec_vpopcnt): Likewise.
18156 (vec_vpopcntb): Likewise.
18157 (vec_vpopcntd): Likewise.
18158 (vec_vpopcnth): Likewise.
18159 (vec_vpopcntw): Likewise.
18161 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
18163 PR rtl-optimization/57468
18164 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
18167 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
18169 PR rtl-optimization/57459
18170 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
18171 type when setting live regs.
18173 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
18175 * config/s390/s390.opt (mlra): New option.
18176 * config/s390/s390.c (s390_decompose_address): Check displacement
18177 for all registers for LRA.
18178 (s390_secondary_reload): Don't used secondary reloads for LRA.
18179 (s390_lra_p): New function.
18180 (TARGET_LRA_P): Define.
18181 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
18182 of attribute cpu_facility to zarch for the last alternative.
18183 (*cmpmem_short): Ditto.
18185 2013-06-06 Eric Botcazou <ebotcazou@adacore.com>
18187 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
18188 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
18189 (arm_expand_prologue): Likewise.
18191 2013-06-06 Teresa Johnson <tejohnson@google.com>
18194 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
18195 as this is now done by redirect_edge_and_branch_force.
18196 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
18197 barriers, and fix interaction with splitting.
18198 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
18199 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
18200 reflect changes made in the routine.
18201 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
18202 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
18203 since this is called in cfglayout mode, and replace partition fixup
18204 with assert as that is now done by force_nonfallthru_and_redirect.
18205 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
18206 already be marked with region crossing note.
18207 (insert_section_boundary_note): Make non-static, gate on flag
18208 has_bb_partition, rewrite to also check for multiple partitions.
18209 (rest_of_handle_reorder_blocks): Remove call to
18210 insert_section_boundary_note, now done later during free_cfg.
18211 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
18212 * bb-reorder.h (insert_section_boundary_note): Declare.
18213 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
18214 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
18215 invoke insert_section_boundary_note.
18216 (try_redirect_by_replacing_jump): Remove unnecessary
18217 check for region crossing note.
18218 (fixup_partition_crossing): New function.
18219 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
18220 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
18221 in non-cfglayout mode.
18222 (force_nonfallthru_and_redirect): Fixup partition boundaries,
18223 remove old code that tried to do this. Emit barrier correctly
18224 when we are in cfglayout mode.
18225 (last_bb_in_partition): New function.
18226 (rtl_split_edge): Correctly fixup partition boundaries.
18227 (commit_one_edge_insertion): Remove old code that tried to
18228 fixup region crossing edge since this is now handled in
18229 split_block, and set up insertion point correctly since
18230 block may now end in a jump.
18231 (verify_hot_cold_block_grouping): Guard against checking when not in
18232 linearized RTL mode.
18233 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
18235 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
18236 rtl_verify_flow_info, so not called in cfglayout mode.
18237 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
18238 (fixup_reorder_chain): Remove old code that attempted to fixup region
18239 crossing note as this is now handled in force_nonfallthru_and_redirect.
18240 (duplicate_insn_chain): Don't duplicate switch section notes.
18241 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
18243 * basic-block.h (emit_barrier_after_bb): Declare.
18245 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18247 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
18248 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
18249 arm_usatsihi): Adjust alternatives for arm_restrict_it.
18251 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18253 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
18255 * config/arm/ldmstm.md: Regenerate.
18257 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18259 * config/arm/sync.md (atomic_loaddi_1):
18260 Disable predication for arm_restrict_it.
18261 (arm_load_exclusive<mode>): Likewise.
18262 (arm_load_exclusivesi): Likewise.
18263 (arm_load_exclusivedi): Likewise.
18264 (arm_load_acquire_exclusive<mode>): Likewise.
18265 (arm_load_acquire_exclusivesi): Likewise.
18266 (arm_load_acquire_exclusivedi): Likewise.
18267 (arm_store_exclusive<mode>): Likewise.
18268 (arm_store_exclusive<mode>): Likewise.
18269 (arm_store_release_exclusivedi): Likewise.
18270 (arm_store_release_exclusive<mode>): Likewise.
18272 2013-06-06 Richard Biener <rguenther@suse.de>
18274 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
18276 (lto_tag_is_tree_code_p): Adjust.
18277 (lto_tag_is_gimple_code_p): Likewise.
18278 (lto_gimple_code_to_tag): Likewise.
18279 (lto_tag_to_gimple_code): Likewise.
18280 (lto_tree_code_to_tag): Likewise.
18281 (lto_tag_to_tree_code): Likewise.
18282 * data-streamer.h (streamer_write_hwi_in_range): Use
18283 uhwi streaming to stream the normalized range.
18284 (streamer_read_hwi_in_range): Likewise.
18286 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18288 * config/arm/arm.md (enabled_for_depr_it): New attribute.
18289 (predicable_short_it): Likewise.
18290 (predicated): Likewise.
18291 (enabled): Handle above.
18292 (define_cond_exec): Set predicated attribute to yes.
18294 2013-06-05 Mike Stump <mikestump@comcast.net>
18296 * gdbinit.in (__FUNCTION__): Add.
18298 2013-06-05 Uros Bizjak <ubizjak@gmail.com>
18300 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
18301 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
18303 2013-06-05 Jan Hubicka <jh@suse.cz>
18305 * varasm.c (mark_decl_referenced): Revert the removal until targets
18308 2013-06-05 David Edelsohn <dje.gcc@gmail.com>
18310 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
18311 instead of mark_decl_referenced.
18313 2013-06-05 Jan Hubicka <jh@suse.cz>
18315 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
18316 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
18317 and symtab_used_from_object_file_p.
18318 (cgraph_make_node_local_1): Clear forced_by_abi.
18319 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
18320 * cgraph.h (symtab_node_base): Add forced_by_abi.
18321 (decide_is_variable_needed): Remove.
18322 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
18323 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
18324 (decide_is_symbol_needed): ... this one; handle symbols in general;
18325 always analyze virtuals; honnor forced_by_abi.
18326 (cgraph_finalize_function): Update.
18327 (varpool_finalize_decl): Update.
18328 (symbol_defined_and_needed): Remove.
18329 (analyze_functions): Update.
18330 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
18331 output_refs, input_overwrite_node): Handle forced_by_abi.
18332 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
18333 (address_taken_from_non_vtable_p): ... this one.
18334 (comdat_can_be_unshared_p_1): New function.
18335 (cgraph_comdat_can_be_unshared_p): Rename to ...
18336 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
18337 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
18338 (function_and_variable_visibility): Clear forced_by_abi as needed.
18339 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
18340 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
18341 * symtab.c (dump_symtab_base): Dump forced_by_abi.
18342 * varpool.c (decide_is_variable_needed): Remove.
18344 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18346 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
18347 (arm_option_override): Override arm_restrict_it where appropriate.
18348 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
18349 * config/arm/arm.opt (mrestrict-it): New command-line option.
18350 * doc/invoke.texi: Document -mrestrict-it.
18352 2013-06-05 David Malcolm <dmalcolm@redhat.com>
18354 * tsan.c (tsan_atomic_table): Make const.
18356 2013-06-05 Richard Biener <rguenther@suse.de>
18358 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
18359 index associated with the tree we are supposed to replace.
18360 * tree-streamer-out.c (pack_ts_base_value_fields): Output
18361 TREE_ASM_WRITTEN as zero for everything but SSA names.
18363 2013-06-05 David Malcolm <dmalcolm@redhat.com>
18365 * tree-ssa-structalias.c (call_stmt_vars): Make static.
18367 2013-06-04 Jan Hubicka <jh@suse.cz>
18369 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
18370 (input_node, input_varpool_node): Handle correctly external same
18372 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
18373 nodes at ltrans stage.
18375 2013-06-04 Jan Hubicka <jh@suse.cz>
18377 * ipa-inline.c (update_caller_keys): Fix availability test.
18378 (update_callee_keys): Likewise.
18379 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
18380 to follow ELF standard.
18382 2013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
18384 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
18385 (mips64r5900el-*-elf*): New configurations.
18386 * config/mips/mips-cpus.def (r5900): New processor.
18387 * config/mips/mips-tables.opt: Regenerate.
18388 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
18389 (mips_issue_rate): Handle PROCESSOR_R5900.
18390 (mips_reorg_process_insns): Force reorder mode for the R5900.
18391 * config/mips/mips.h (TARGET_MIPS5900): Define.
18392 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
18394 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
18396 * config/mips/mips.md (processor): Add r5900.
18397 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
18399 2013-06-04 Ian Bolton <ian.bolton@arm.com>
18401 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
18402 into function to generate MOVI instruction.
18403 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
18404 (aarch64_preferred_simd_mode): Turn into wrapper.
18405 (aarch64_output_scalar_simd_mov_immediate): New function.
18406 * config/aarch64/aarch64-protos.h: Add prototype for above.
18408 2013-06-04 Ian Bolton <ian.bolton@arm.com>
18410 * config/aarch64/aarch64.c (simd_immediate_info): Remove
18411 element_char member.
18412 (sizetochar): Return signed char.
18413 (aarch64_simd_valid_immediate): Remove elchar and other
18414 unnecessary variables.
18415 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
18416 Calculate element_char as required.
18417 * config/aarch64/aarch64-protos.h: Update and move prototype
18418 for aarch64_output_simd_mov_immediate.
18419 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
18422 2013-06-04 Ian Bolton <ian.bolton@arm.com>
18424 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
18425 information completed by aarch64_simd_valid_immediate.
18426 (aarch64_legitimate_constant_p): Update arguments.
18427 (aarch64_simd_valid_immediate): Work with struct rather than many
18429 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
18430 (aarch64_simd_make_constant): Update arguments.
18431 (aarch64_output_simd_mov_immediate): Work with struct rather than
18432 many pointers. Output immediate directly rather than as operand.
18433 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
18435 * config/aarch64/constraints.md (Dn): Update arguments.
18437 2013-06-04 Ian Bolton <ian.bolton@arm.com>
18439 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
18441 (aarch64_simd_immediate_valid_for_move): Remove.
18442 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
18443 (aarch64_simd_make_constant): Update call.
18444 (aarch64_output_simd_mov_immediate): Update call.
18445 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
18447 * config/aarch64/constraints.md (Dn): Update call.
18449 2013-06-04 Ian Bolton <ian.bolton@arm.com>
18451 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
18452 return type to bool for prototype.
18453 (aarch64_legitimate_constant_p): Check for true instead of not -1.
18454 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
18455 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
18457 2013-06-04 Catherine Moore <clm@codesourcery.com>
18459 * config/mips/mips.opt (meva): New.
18460 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
18461 (ASM_SPEC): Handle -meva.
18462 * doc/invoke.texi (meva): Document.
18464 2013-06-04 Alan Modra <amodra@gmail.com>
18466 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
18469 2013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18471 * rtl.def: Add extra fourth optional field to define_cond_exec.
18472 * gensupport.c (process_one_cond_exec): Process attributes from
18474 * doc/md.texi: Document fourth field in define_cond_exec.
18476 2013-06-04 Eric Botcazou <ebotcazou@adacore.com>
18478 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
18479 out the processing order as in store_bit_field_1.
18481 2013-06-04 Jan Hubicka <jh@suse.cz>
18483 PR middle-end/57500
18484 * cgraphunit.c (cgraph_process_same_body_aliases): Create
18485 non-VAR_DECL node if it does not exist yet.
18487 2013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
18489 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
18490 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
18491 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
18492 target_cpu_default setting.
18494 2013-06-03 Teresa Johnson <tejohnson@google.com>
18496 * dumpfile.c (opt_info_switch_p): Change -fopt-info
18497 default to -fopt-info=optimized instead of all.
18498 * doc/invoke.texi: Ditto.
18499 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
18500 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
18501 (execute_vect_slp): Emit BB vectorization success under
18502 MSG_OPTIMIZED_LOCATIONS.
18503 * tree-vect-slp.c (vect_slp_transform_bb): Change
18504 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
18505 * tree-vect-loop.c (vect_transform_loop): Ditto.
18507 2013-06-03 Jason Merrill <jason@redhat.com>
18510 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
18511 Use TARGET_EXPR for C++.
18513 2013-06-03 Jakub Jelinek <jakub@redhat.com>
18515 PR rtl-optimization/57268
18516 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
18517 if DEBUG_INSN_P (insn).
18520 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
18522 PR rtl-optimization/57268
18523 * sched-deps.c (sched_analyze_2): Flush dependence lists if
18524 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
18526 2013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
18528 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
18529 (ix86_avoid_lea_for_addr): Likewise.
18530 (exact_dependency_1): Likewise.
18531 (ix86_adjust_cost): Likewise.
18532 (swap_top_of_ready_list): Fix formatting and !reload_completed check
18534 (do_reorder_for_imul): Fix typo, formatting and
18535 !reload_completed check removed.
18536 (ix86_sched_reorder): Fix typo and formatting.
18537 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
18540 2013-06-03 Sofiane Naci <sofiane.naci@arm.com>
18542 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
18544 2013-06-03 Eric Botcazou <ebotcazou@adacore.com>
18546 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
18547 <STRING_CST>: Likewise.
18548 <VECTOR_CST>: Likewise.
18550 2013-06-01 Janus Weil <janus@gcc.gnu.org>
18551 Mikael Morin <mikael@gcc.gnu.org>
18553 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
18554 * config.in: Regenerated.
18555 * configure: Regenerated.
18557 2013-06-01 Jan Hubicka <jh@suse.cz>
18559 PR middle-end/57366
18560 * cgraphunit.c (compile): When weakref is not supported,
18561 set up transparent aliases before final output pass.
18562 * varasm.c (assemble_alias): Do not try to do it here.
18564 2013-06-01 Jan Hubicka <jh@suse.cz>
18566 PR middle-end/57467
18567 * passes.c (for_per_function): Skip unanalyzed functions.
18569 2013-06-01 Jan Hubicka <jh@suse.cz>
18571 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
18572 (lto_symtab_merge_symbols_1): ... this one.
18573 (lto_symtab_merge_cgraph_nodes): Rename to ...
18574 (lto_symtab_merge_symbols): ... this one; simplify.
18575 * cgraph.c (same_body_aliases_done): Rename to ...
18576 (cpp_implicit_aliases_done): ... this one.
18577 (cgraph_create_function_alias): Update.
18578 (cgraph_same_body_alias): Update.
18579 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
18580 (verify_edge_corresponds_to_fndecl): Simplify.
18581 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
18582 (cgraph_node): Remove same_body_alias.
18583 (varpool_node): Remove alias_of and extra_name_alias.
18584 (same_body_aliases_done): Rename to ..
18585 (cpp_implicit_aliases_done): ... this one.
18586 (symtab_alias_ultimate_target): Add default parameter.
18587 (symtab_resolve_alias): New function.
18588 (fixup_same_cpp_alias_visibility): Declare.
18589 (cgraph_function_node): Add default parameter.
18590 (cgraph_node_asm_name): Likewise.
18591 (cgraph_function_or_thunk_node): Add default parameter; do
18592 not ICE when it is NULL.
18593 (varpool_variable_node): Likewise.
18594 * tree-emutls.c (create_emultls_var): Update.
18595 (ipa_lower_emutls): Update.
18596 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
18597 (cgraph_reset_node): Reset alias info.
18598 (cgraph_finalize_function): Update.
18599 (fixup_same_cpp_alias_visibility): Move to symtab.c.
18600 (analyze_function): Simplify.
18601 (cgraph_process_same_body_aliases): Simplify.
18602 (analyze_functions): Fixup same body aliases.
18603 (handle_alias_pairs): Simplify.
18604 (assemble_thunk): Update.
18605 (assemble_thunks_and_aliases): Update.
18606 (output_weakrefs): Rewrite.
18607 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
18608 (lto_output_varpool_node): Likewise.
18609 (compute_ltrans_boundary): Remve assert.
18610 (get_alias_symbol): New functoin.
18611 (input_node): Rewrite alias handling.
18612 (input_varpool_node): Likewise.
18613 * ipa-pure-const.c (propagate_pure_const): Fix formating.
18614 * ipa.c (process_references): Handle weakrefs correctly.
18615 (symtab_remove_unreachable_nodes): Likewise.
18616 * trans-mem.c (get_cg_data): Update.
18617 (ipa_tm_create_version_alias): Update.
18618 (ipa_tm_execute): Update.
18619 * symtab.c (dump_symtab_base): Dump aliases.
18620 (verify_symtab_base): Verify aliases.
18621 (symtab_node_availability): New function.
18622 (symtab_alias_ultimate_target): Simplify.
18623 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
18624 handle all the fixup cases.
18625 (symtab_resolve_alias): New function.
18626 * passes.c (ipa_write_summaries): Handle weakrefs.
18627 * varpool.c (varpool_analyze_node): Simplify.
18628 (assemble_aliases): Update.
18629 (varpool_create_variable_alias): Simplify.
18630 (varpool_extra_name_alias): Simplify.
18631 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
18632 (lto_symtab_merge_symbols): ... this one.
18634 2013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
18637 PR rtl-optimization/57268
18638 * sched-deps.c (sched_analyze_2): Flush dependence lists if
18639 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
18641 2013-06-01 Tobias Burnus <burnus@net-b.de>
18643 Partially reverted:
18644 2013-05-31 Tobias Burnus <burnus@net-b.de>
18646 PR middle-end/57073
18647 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
18650 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
18652 PR rtl-optimization/57268
18653 * sched-deps.c (sched_analyze_2): Flush dependence lists if
18654 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
18656 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
18658 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
18659 unordered comparison operators when -fno-trapping-math is in effect
18661 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
18662 and implement unordered comparison operators properly on the e500.
18664 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
18666 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
18667 for constant scalar integers.
18668 (simplify_relational_operation_1): Likewise.
18670 2013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
18672 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
18673 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
18676 2013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
18677 Igor Zamyatin <igor.zamyatin@intel.com>
18679 Silvermont (SLM) architecture performance tuning.
18680 * config/i386/i386.h (enum ix86_tune_indices): Add
18681 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
18682 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
18684 * config/i386/i386.c (initial_ix86_tune_features)
18685 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
18686 (ix86_lea_outperforms): Handle Silvermont tuning.
18687 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
18689 (ix86_use_lea_for_mov): Likewise.
18690 (ix86_avoid_lea_for_addr): Likewise.
18691 (ix86_lea_for_add_ok): Likewise.
18692 (exact_dependency_1): New function.
18693 (exact_store_load_dependency): Likewise.
18694 (ix86_adjust_cost): Handle Silvermont tuning.
18695 (do_reoder_for_imul): Likewise.
18696 (swap_top_of_ready_list): New function.
18697 (ix86_sched_reorder): Changed to handle Silvermont tuning.
18699 * config/i386/i386.md (peepholes that split memory operand in fp
18702 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
18704 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
18705 Remove un-necessary braces.
18707 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
18709 * config/aarch64/aarch64.c (aarch64_classify_symbol):
18710 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
18712 2013-05-31 Tobias Burnus <burnus@net-b.de>
18714 PR middle-end/57073
18715 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
18717 2013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18720 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
18721 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
18722 * config/arm/neon.md (iordi3_neon): Remove.
18723 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
18724 * config/arm/predicates.md (imm_for_neon_logic_operand):
18725 Move to earlier in the file.
18726 (neon_logic_op2): Likewise.
18727 (arm_iordi_operand_neon): New predicate.
18729 2013-05-31 Richard Biener <rguenther@suse.de>
18731 PR tree-optimization/57478
18732 PR tree-optimization/57453
18733 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
18736 2013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
18738 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
18739 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
18741 2013-05-30 Tobias Burnus <burnus@net-b.de>
18742 Thomas Koenig <tkoenig@gcc.gnu.org>
18744 PR middle-end/57073
18745 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
18746 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
18748 2013-05-30 Steven Bosscher <steven@gcc.gnu.org>
18750 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
18752 2013-05-30 Vladimir Makarov <vmakarov@redhat.com>
18754 * target.def (register_usage_leveling_p): New hook.
18755 * targhooks.c (default_register_usage_leveling_p): New.
18756 * targhooks.h (default_register_usage_leveling_p): New prototype.
18757 * lra-assigns.c (register_usage_leveling_p): Use the hook.
18758 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
18759 * doc/tm.texi: Update.
18760 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
18762 2013-05-30 Ian Bolton <ian.bolton@arm.com>
18764 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
18765 (*insv_reg<mode>): New define_insn.
18767 2013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
18769 PR rtl-optimization/57439
18770 * postreload.c (move2add_valid_value_p): Check that we have
18771 a zero subreg_regno_offset when accessing the register in
18772 the requested mode.
18774 2013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
18775 Igor Zamyatin <igor.zamyatin@intel.com>
18777 Silvermont (SLM) architecture pipeline model, tuning and
18779 * config.gcc: Add slm config options and target.
18781 * config/i386/slm.md: New.
18783 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
18785 * config/i386/i386-c.c (ix86_target_macros_internal): New case
18787 (ix86_target_macros_internal): Likewise.
18789 * config/i386/i386.c (slm_cost): New cost.
18790 (m_SLM): New macro flag.
18791 (initial_ix86_tune_features): Set m_SLM.
18792 (x86_accumulate_outgoing_args): Likewise.
18793 (x86_arch_always_fancy_math_387): Likewise.
18794 (processor_target_table): Add slm cost.
18795 (cpu_names): Add slm cpu name.
18796 (x86_option_override_internal): Set SLM ISA.
18797 (ix86_issue_rate): New case PROCESSOR_SLM.
18798 (ia32_multipass_dfa_lookahead): Likewise.
18799 (fold_builtin_cpu): Add slm.
18801 * config/i386/i386.h (TARGET_SLM): New target macro.
18802 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
18803 (processor_type): Add PROCESSOR_SLM.
18805 * config/i386/i386.md (cpu): Add new value "slm".
18806 (slm.md): Include slm.md.
18808 2013-05-30 Bernd Schmidt <bernds@codesourcery.com>
18809 Zhenqiang Chen <zhenqiang.chen@linaro.org>
18811 * config/arm/arm-protos.h: Add and update function protos.
18812 * config/arm/arm.c (use_simple_return_p): New added.
18813 (thumb2_expand_return): Check simple_return flag.
18814 * config/arm/arm.md: Add simple_return and conditional simple_return.
18815 * config/arm/iterators.md: Add iterator for return and simple_return.
18817 2013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
18819 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
18820 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
18821 (arm_emit_vfp_multi_reg_pop): Likewise.
18822 (thumb2_emit_ldrd_pop): Likewise.
18823 (arm_expand_epilogue): Add misc REG_CFA notes.
18824 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
18826 2013-05-29 Lawrence Crowl <crowl@google.com>
18828 * config/arm/t-arm: Update for below.
18830 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
18831 Change type to hash_table. Update dependent calls and types.
18833 * config/i386/t-cygming: Update for below.
18835 * config/i386/t-interix: Update for below.
18837 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
18838 Change type to hash_table. Update dependent calls and types.
18839 (i386_find_on_wrapper_list::wrappers): Likewise.
18841 * config/ia64/t-ia64: Update for below.
18843 * config/ia64/ia64.c (bundle_state_table):
18844 Change type to hash_table. Update dependent calls and types.
18846 * config/mips/mips.c (mips_reorg_process_insns::htab):
18847 Change type to hash_table. Update dependent calls and types.
18849 * config/sol2.c (solaris_comdat_htab):
18850 Change type to hash_table. Update dependent calls and types.
18852 * config/t-sol2: Update for above.
18854 2013-05-29 Teresa Johnson <tejohnson@google.com>
18856 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
18857 functions are not yet marked as defined.
18859 2013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
18860 Pat Haugen <pthaugen@us.ibm.com>
18861 Peter Bergner <bergner@vnet.ibm.com>
18863 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
18867 (vrotl<mode>3): Likewise.
18868 (vashl<mode>3): Likewise.
18869 (vlshr<mode>3): Likewise.
18870 (vashr<mode>3): Likewise.
18872 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
18873 support for power8 V2DI builtins.
18875 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
18876 power8 V2DI builtins.
18877 (vupkhsw): Likewise.
18878 (vupklsw): Likewise.
18879 (vaddudm): Likewise.
18880 (vminsd): Likewise.
18881 (vmaxsd): Likewise.
18882 (vminud): Likewise.
18883 (vmaxud): Likewise.
18884 (vpkudum): Likewise.
18885 (vpksdss): Likewise.
18886 (vpkudus): Likewise.
18887 (vpksdus): Likewise.
18892 (vsubudm): Likewise.
18893 (vcmpequd): Likewise.
18894 (vcmpgtsd): Likewise.
18895 (vcmpgtud): Likewise.
18896 (vcmpequd_p): Likewise.
18897 (vcmpgtsd_p): Likewise.
18898 (vcmpgtud_p): Likewise.
18899 (vupkhsw): Likewise.
18900 (vupklsw): Likewise.
18901 (vaddudm): Likewise.
18902 (vmaxsd): Likewise.
18903 (vmaxud): Likewise.
18904 (vminsd): Likewise.
18905 (vminud): Likewise.
18906 (vpksdss): Likewise.
18907 (vpksdus): Likewise.
18908 (vpkudum): Likewise.
18909 (vpkudus): Likewise.
18914 (vsubudm): Likewise.
18916 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
18917 support for power8 V2DI instructions.
18919 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
18920 power8 V2DI instructions. Combine pack and unpack insns to use an
18921 iterator for each mode. Check whether a particular mode supports
18922 Altivec instructions instead of just checking TARGET_ALTIVEC.
18923 (UNSPEC_VPKUWUM): Likewise.
18924 (UNSPEC_VPKSHSS): Likewise.
18925 (UNSPEC_VPKSWSS): Likewise.
18926 (UNSPEC_VPKUHUS): Likewise.
18927 (UNSPEC_VPKSHUS): Likewise.
18928 (UNSPEC_VPKUWUS): Likewise.
18929 (UNSPEC_VPKSWUS): Likewise.
18930 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
18931 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
18932 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
18933 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
18934 (UNSPEC_VUPKHSB): Likewise.
18935 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
18936 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
18937 (UNSPEC_VUPKHSH): Likewise.
18938 (UNSPEC_VUPKLSB): Likewise.
18939 (UNSPEC_VUPKLSH): Likewise.
18941 (VI_char): Likewise.
18942 (VI_scalar): Likewise.
18943 (VI_unit): Likewise.
18945 (VP_small): Likewise.
18946 (VP_small_lc): Likewise.
18947 (VU_char): Likewise.
18948 (add<mode>3): Likewise.
18949 (altivec_vaddcuw): Likewise.
18950 (altivec_vaddu<VI_char>s): Likewise.
18951 (altivec_vadds<VI_char>s): Likewise.
18952 (sub<mode>3): Likewise.
18953 (altivec_vsubcuw): Likewise.
18954 (altivec_vsubu<VI_char>s): Likewise.
18955 (altivec_vsubs<VI_char>s): Likewise.
18956 (altivec_vavgs<VI_char>): Likewise.
18957 (altivec_vcmpbfp): Likewise.
18958 (altivec_eq<mode>): Likewise.
18959 (altivec_gt<mode>): Likewise.
18960 (altivec_gtu<mode>): Likewise.
18961 (umax<mode>3): Likewise.
18962 (smax<mode>3): Likewise.
18963 (umin<mode>3): Likewise.
18964 (smin<mode>3): Likewise.
18965 (altivec_vpkuhum): Likewise.
18966 (altivec_vpkuwum): Likewise.
18967 (altivec_vpkshss): Likewise.
18968 (altivec_vpkswss): Likewise.
18969 (altivec_vpkuhus): Likewise.
18970 (altivec_vpkshus): Likewise.
18971 (altivec_vpkuwus): Likewise.
18972 (altivec_vpkswus): Likewise.
18973 (altivec_vpks<VI_char>ss): Likewise.
18974 (altivec_vpks<VI_char>us): Likewise.
18975 (altivec_vpku<VI_char>us): Likewise.
18976 (altivec_vpku<VI_char>um): Likewise.
18977 (altivec_vrl<VI_char>): Likewise.
18978 (altivec_vsl<VI_char>): Likewise.
18979 (altivec_vsr<VI_char>): Likewise.
18980 (altivec_vsra<VI_char>): Likewise.
18981 (altivec_vsldoi_<mode>): Likewise.
18982 (altivec_vupkhsb): Likewise.
18983 (altivec_vupkhs<VU_char>): Likewise.
18984 (altivec_vupkls<VU_char>): Likewise.
18985 (altivec_vupkhsh): Likewise.
18986 (altivec_vupklsb): Likewise.
18987 (altivec_vupklsh): Likewise.
18988 (altivec_vcmpequ<VI_char>_p): Likewise.
18989 (altivec_vcmpgts<VI_char>_p): Likewise.
18990 (altivec_vcmpgtu<VI_char>_p): Likewise.
18991 (abs<mode>2): Likewise.
18992 (vec_unpacks_hi_v16qi): Likewise.
18993 (vec_unpacks_hi_v8hi): Likewise.
18994 (vec_unpacks_lo_v16qi): Likewise.
18995 (vec_unpacks_hi_<VP_small_lc>): Likewise.
18996 (vec_unpacks_lo_v8hi): Likewise.
18997 (vec_unpacks_lo_<VP_small_lc>): Likewise.
18998 (vec_pack_trunc_v8h): Likewise.
18999 (vec_pack_trunc_v4si): Likewise.
19000 (vec_pack_trunc_<mode>): Likewise.
19002 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
19004 (vec_vmaxsd): Likewise.
19005 (vec_vmaxud): Likewise.
19006 (vec_vminsd): Likewise.
19007 (vec_vminud): Likewise.
19008 (vec_vpksdss): Likewise.
19009 (vec_vpksdus): Likewise.
19010 (vec_vpkudum): Likewise.
19011 (vec_vpkudus): Likewise.
19012 (vec_vrld): Likewise.
19013 (vec_vsld): Likewise.
19014 (vec_vsrad): Likewise.
19015 (vec_vsrd): Likewise.
19016 (vec_vsubudm): Likewise.
19017 (vec_vupkhsw): Likewise.
19018 (vec_vupklsw): Likewise.
19020 2013-05-29 Jan Hubicka <jh@suse.cz>
19022 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
19023 flags; reorder rest of fields in more consistent way.
19024 (varpool_node): Remove analyzed, finalized and alias.
19025 (cgraph_ndoe): Likewise.
19026 (symtab_alias_ultimate_target): New function.
19027 (cgraph_function_node): Move offline.
19028 (cgraph_reset_node): Declare.
19029 (cgraph_comdat_can_be_unshared_p): Remove.
19030 (varpool_remove_initializer): Declare.
19031 (varpool_first_defined_variable, varpool_next_defined_variable
19032 cgraph_first_defined_function, cgraph_next_defined_function): Update.
19033 (cgraph_function_with_gimple_body_p): Update.
19034 (varpool_all_refs_explicit_p): Update.
19035 (symtab_alias_target): New function.
19036 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
19037 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
19038 (cgraph_function_or_thunk_node): Simplify using
19039 symtab_alias_ultimate_target.
19040 (varpool_variable_node): Likewise.
19041 * cgraph.c (cgraph_create_function_alias): Update.
19042 (cgraph_add_thunk): Update.
19043 (cgraph_remove_node): Update.
19044 (dump_cgraph_node): Do not dump removed flags.
19045 (cgraph_function_body_availability): Update.
19046 (cgraph_propagate_frequency): Update.
19047 (verify_cgraph_node): Check sanity of local flag.
19048 (cgraph_function_node): Move here from cgraph.h; revamp for
19049 cgraph_function_or_thunk_node.
19050 * lto-symtab.c (lto_varpool_replace_node): Update.
19051 (lto_symtab_resolve_can_prevail_p): Update.
19052 (lto_symtab_merge_cgraph_nodes): Update.
19053 * ipa-cp.c (determine_versionability, initialize_node_lattices,
19054 propagate_constants_accross_call, devirtualization_time_bonus,
19055 ipcp_propagate_stage): Update.
19056 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
19057 * ipa-inline-transform.c (clone_inlined_nodes,
19058 preserve_function_body_p): Update.
19059 * ipa-reference.c (propagate): Update.
19060 (write_node_summary_p): Update.
19061 * toplev.c (wrapup_global_declaration_2): Update.
19062 * cgraphunit.c (cgraph_analyze_function): Rename to ...
19063 (analyze_function) ... this one.
19064 (cgraph_process_new_functions): Update.
19065 (cgraph_reset_node): Export.
19066 (cgraph_finalize_function): Update.
19067 (cgraph_add_new_function): Update.
19068 (process_function_and_variable_attributes): Update.
19069 (varpool_finalize_decl): Update.
19070 (symbol_finalized): Remove.
19071 (symbol_finalized_and_needed): Rename to ...
19072 (symbol_defined_and_needed): ... update.
19073 (cgraph_analyze_functions): Update.
19074 (handle_alias_pairs): Update.
19075 (mark_functions_to_output): Update.
19076 (assemble_thunk): Update.
19077 (output_in_order): Update.
19078 (output_weakrefs): Update.
19079 (finalize_compilation_unit): Update.
19080 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
19081 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
19082 input_node, input_varpool_node): Update.
19083 * dbxout.c (dbxout_expand_expr): Update.
19084 * cgraphclones.c (cgraph_clone_node): Update.
19085 (cgraph_copy_node_for_versioning): Update.
19086 (cgraph_materialize_clone): Update.
19087 (cgraph_materialize_all_clones): Update.
19088 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
19089 propagate_pure_const, propagate_nothrow): Update.
19090 * lto-streamer-out.c (lto_output, write_symbol): Update.
19091 * ipa-utils.c (ipa_reverse_postorder): Update.
19092 * ipa-inline.c (can_inline_edge_p): Update.
19093 (update_caller_keys, ipa_inline): Update.
19094 * dwarf2out.c (reference_to_unused,
19095 premark_types_used_by_global_vars_helper): Update.
19096 * tree-eh.c (tree_could_trap_p): Update.
19097 * ipa-split.c (consider_split, execute_split_functions): Update.
19098 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
19099 has_addr_references_p): Update; move ahead in file for better
19101 (process_references): Simplify.
19102 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
19103 bodies are removed.
19104 (cgraph_comdat_can_be_unshared_p): Make static.
19105 (cgraph_externally_visible_p): Update.
19106 (varpool_externally_visible_p): Update.
19107 (function_and_variable_visibility): Update.
19108 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
19109 ipa_tm_mark_force_output_node): Update.
19110 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
19111 estimate_edge_devirt_benefit, inline_generate_summary,
19112 inline_write_summary): Update.
19113 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
19114 * ipa-prop.c (ipa_compute_jump_functions): Update.
19115 (ipa_print_node_params, ipa_prop_read_section,
19116 ipa_update_after_lto_read, read_replacements_section): Update.
19117 * varasm.c (mark_decl_referenced): Update.
19118 (assemble_alias, dump_tm_clone_pairs): Update.
19119 * tree-inline.c (copy_bb): Update.
19120 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
19122 * symtab.c (dump_symtab_base): Print new flags.
19123 (verify_symtab_base): Verify new flags.
19124 (symtab_alias_ultimate_target): New function.
19125 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
19126 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
19128 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
19130 * i386.c (ix86_get_function_versions_dispatcher,
19131 ix86_generate_version_dispatcher_body): Update.
19132 (fold_builtin_cpu): Use varpool_add_new_variable.
19133 * varpool.c (varpool_remove_initializer): Break out from ...
19134 (varpool_remove_node): ... this one.
19135 (dump_varpool_node, varpool_node_for_asm,
19136 cgraph_variable_initializer_availability, varpool_analyze_node,
19137 varpool_assemble_decl, varpool_remove_unreferenced_decls,
19138 varpool_finalize_named_section_flags, varpool_create_variable_alias):
19141 2013-05-29 Jan Hubicka <jh@suse.cz>
19143 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
19145 2013-05-29 Easwaran Raman <eraman@google.com>
19147 PR tree-optimization/57442
19148 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
19149 when control exits the main loop.
19151 2013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
19153 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
19155 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
19156 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
19157 * rx/t-rx: Add rx100 under multi library matches option for nofpu
19160 2013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
19162 PR tree-optimization/57441
19163 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
19164 Don't limit size of incr_vec to number of candidates.
19166 2013-05-29 Steve Ellcey <sellcey@imgtec.com>
19168 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
19169 and mips16 directories.
19170 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
19171 (MULTILIB_DIRNAMES): Ditto.
19172 (MULTILIB_EXCEPTIONS): Add new exceptions.
19173 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
19174 (MULTILIB_DIRNAMES): Ditto.
19175 (MULTILIB_EXCEPTIONS): Add new exceptions.
19177 2012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
19178 Marcus Shawcroft <marcus.shawcroft@arm.com>
19180 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
19181 SYMBOL_TINY_ABSOLUTE.
19182 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
19183 SYMBOL_TINY_ABSOLUTE.
19184 (aarch64_expand_mov_immediate): Likewise.
19185 (aarch64_classify_symbol): Likewise.
19186 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
19187 Permit SYMBOL_TINY_ABSOLUTE.
19188 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
19190 2013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
19191 Marcus Shawcroft <marcus.shawcroft@arm.com>
19193 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
19194 Refactor if/switch. Replace gcc_assert with if.
19196 2013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
19198 * config/i386/i386.c (initial_ix86_tune_features): Enable
19199 FP Reassociation for AMD bdver1 and bdver2.
19201 2013-05-29 Martin Jambor <mjambor@suse.cz>
19203 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
19204 and IMAGPART_EXPR do not occur within other handled_components.
19206 2013-05-29 Richard Biener <rguenther@suse.de>
19208 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
19209 access on whether the use is in the BB we currently try to
19211 (vect_bb_vectorization_profitable_p): Pass the BB we currently
19212 vectorize to vect_bb_slp_scalar_cost.
19214 2013-05-29 Richard Biener <rguenther@suse.de>
19216 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
19217 computing scalar cost offsetted by stmts that are kept live
19219 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
19220 for computation of scalar cost.
19222 2013-05-28 Steve Ellcey <sellcey@mips.com>
19224 * config/mips/mips-cpus.def (mips32r2): Change processor type.
19226 2013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
19228 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
19229 array notation built-in reduction functions.
19230 * doc/passes.texi (Passes): Added documentation about changes done
19232 * doc/invoke.texi (C Dialect Options): Added documentation about
19233 the -fcilkplus flag.
19234 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
19235 (BUILTINS_DEF): Depend on cilkplus.def.
19236 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
19237 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
19238 * cilkplus.def: New file.
19240 2013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
19242 PR rtl-optimization/57439
19243 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
19245 2013-05-28 Easwaran Raman <eraman@google.com>
19247 PR tree-optimization/57337
19248 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
19249 (find_insert_point): Correctly identify the insertion point
19250 when two statements with the same UID is compared.
19252 2013-05-28 Richard Biener <rguenther@suse.de>
19254 PR tree-optimization/56787
19255 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
19256 from the list of data references.
19257 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
19259 (vect_analyze_loop_operations): Likewise.
19260 (vect_transform_loop): Remove clobbers.
19262 2013-05-28 Martin Jambor <mjambor@suse.cz>
19264 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
19265 and REALPART_EXPRs have scalar type.
19267 2013-05-28 Richard Biener <rguenther@suse.de>
19269 PR tree-optimization/57411
19270 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
19272 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
19273 virtual operand propagation.
19275 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
19277 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
19278 destination register for bmasksi_vis.
19279 (vector_init_bshuffle): Likewise.
19280 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
19282 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
19284 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
19285 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
19286 mode if the instruction isn't available in the original mode.
19287 * config/sparc/sparc.opt (mfix-ut699): New option.
19288 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
19289 (divdf3): Turn into expander.
19290 (divdf3_nofix): New insn.
19291 (divdf3_fix): Likewise.
19292 (divsf3): Disable if -mfix-ut699.
19293 (sqrtdf2): Turn into expander.
19294 (sqrtdf2_nofix): New insn.
19295 (sqrtdf2_fix): Likewise.
19296 (sqrtsf2): Disable if -mfix-ut699.
19298 2013-05-27 Richard Biener <rguenther@suse.de>
19300 PR middle-end/57412
19301 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
19302 block for the new loop.
19304 2013-05-27 Richard Biener <rguenther@suse.de>
19306 PR tree-optimization/57343
19307 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
19308 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
19309 (number_of_iterations_cond): Do not build the folded tree.
19311 2013-05-27 Richard Biener <rguenther@suse.de>
19314 PR middle-end/57381
19315 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
19316 OEP_CONSTANT_ADDRESS_OF retained.
19318 PR tree-optimization/57417
19319 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
19320 for unchanged base.
19321 (set_ssa_val_to): Compare addresses using
19322 get_addr_base_and_unit_offset.
19324 2013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
19326 PR rtl-optimization/56833
19327 * postreload.c (move2add_record_mode): New function.
19328 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
19329 (move2add_use_add2_insn): Use move2add_record_sym_value.
19330 (move2add_use_add3_insn): Likewise.
19331 (reload_cse_move2add): Use move2add_valid_value_p and
19332 move2add_record_mode. Invalidate call-clobbered and REG_INC
19333 affected regs by setting reg_mode to VOIDmode.
19334 (move2add_note_store): Don't pretend the inside of a SUBREG is
19335 the actual destination. Invalidate single/leading registers by
19336 setting reg_mode to VOIDmode.
19337 Use move2add_record_sym_value, move2add_valid_value_p and
19338 move2add_record_mode.
19340 2013-05-27 Richard Biener <rguenther@suse.de>
19342 PR tree-optimization/57396
19343 * tree-affine.c (double_int_constant_multiple_p): Properly
19344 return false for val == 0 and div != 0.
19346 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
19348 * config/mips/mips.h: Use #elif in preprocessor conditions.
19350 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
19353 * config/mips/constraints.md (kl): New constraint.
19354 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
19355 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
19356 constraint for operand 0. Split after CSE for MIPS16. Emit a move
19357 from LO for MIPS16.
19358 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
19360 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
19363 * config/mips/mips.c (mips_can_inline_p): New function.
19364 (TARGET_CAN_INLINE_P): Define.
19366 2013-05-25 Steven Bosscher <steven@gcc.gnu.org>
19368 * sched-int.h (ds_t, dw_t): Make unsigned int.
19369 Fix documentation that describes how all the ds_t bits are used.
19370 Reserve the last bit for delayed-branch scheduling.
19371 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
19372 (BITS_PER_DEP_WEAK): Fix definition and documentation.
19373 (gen_dep_weak_1): Remove prototype.
19374 * sched-deps.c (get_dep_weak_1): Make static.
19375 * target.def (speculate_insn, needs_block_p, gen_spec_check,
19376 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
19377 * doc/tm.texi: Regenerate.
19378 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
19380 2013-05-24 Steven Bosscher <steven@gcc.gnu.org>
19383 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
19385 2013-05-24 Nathan Sidwell <nathan@codesourcery.com>
19386 Sandra Loosemore <sandra@codesourcery.com>
19388 * config.gcc (powerpc-*): Allow native for with-cpu.
19390 2013-05-24 Jeff Law <law@redhat.com>
19392 PR tree-optimization/57124
19393 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
19394 conversion feeding a condition if the range has an overflow
19395 if -fstrict-overflow. Add warnings for when we do make the
19398 2013-05-24 Dehao Chen <dehao@google.com>
19400 * tree-cfg.c (locus_discrim_map): Fix the typo.
19401 (locus_discrim_hasher): Likewise.
19402 (locus_discrim_hasher::hash): Likewise.
19403 (locus_discrim_hasher::equal): Likewise.
19405 2013-05-24 Martin Jambor <mjambor@suse.cz>
19407 PR tree-optimization/57294
19408 * cgraph.h (ipa_record_stmt_references): Declare.
19409 * cgraphbuild.c (ipa_record_stmt_references): New function.
19410 (build_cgraph_edges): Use ipa_record_stmt_references.
19411 (rebuild_cgraph_edges): Likewise.
19412 (cgraph_rebuild_references): Likewise.
19413 * ipa-prop.c (ipa_modify_call_arguments): Discard references
19414 associated with the old statement and build references from the
19415 newly built statements.
19416 * ipa-ref.c (ipa_remove_stmt_references): New function.
19417 * ipa-ref.h (ipa_remove_stmt_references): Declare.
19419 2013-05-24 Vladimir Makarov <vmakarov@redhat.com>
19421 * lra-constraints.c (emit_spill_move): Use smaller mode for
19423 (check_and_process_move): Consider mem-reg moves for secondary
19425 (curr_insn_transform): Don't lose insns emitted before for
19426 secondary memory moves.
19427 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
19428 reg set up in the current insn.
19430 2013-05-24 Dehao Chen <dehao@google.com>
19432 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
19434 (locus_descrim_hasher::equal): Likewise.
19435 (build_gimple_cfg): New discriminator assignment algorithm.
19436 (make_edges): Likewise.
19437 (next_discriminator_for_locus): Likewise.
19438 (same_line_p): Likewise.
19439 (assign_discriminators): Likewise.
19440 (make_cond_expr_edges): Likewise.
19441 (make_gimple_switch_edges): Likewise.
19442 (make_goto_expr_edges): Likewise.
19443 (make_gimple_asm_edges): Likewise.
19445 2013-05-24 Ian Bolton <ian.bolton@arm.com>
19447 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
19448 X format specifier to only display bottom 16 bits.
19449 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
19450 immediate to match for operand 2, since it will be masked.
19452 2013-05-24 Richard Biener <rguenther@suse.de>
19454 PR tree-optimization/57287
19455 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
19456 all SSA names that occur in abnormal PHIs.
19458 2013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
19461 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
19462 that index is not negative.
19464 2013-05-24 Eric Botcazou <ebotcazou@adacore.com>
19466 PR rtl-optimization/55177
19467 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
19468 (simplify_byte_swapping_operation): New.
19469 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
19470 (simplify_relational_operation_1): Deal with BSWAP.
19472 2013-05-23 Richard Henderson <rth@redhat.com>
19475 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
19476 (ix86_reorg): Call it.
19478 2013-05-23 Uros Bizjak <ubizjak@gmail.com>
19481 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
19482 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
19483 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
19485 2013-05-23 Christian Bruel <christian.bruel@st.com>
19488 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
19490 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
19491 Marcus Shawcroft <marcus.shawcroft@arm.com>
19493 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
19494 * config/aarch64/constraints.md (Usa): Remove.
19495 * doc/md.texi (AArch64 Usa): Remove.
19497 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
19498 Marcus Shawcroft <marcus.shawcroft@arm.com>
19500 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
19501 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
19502 * config/aarch64/predicates.md (aarch64_const_address): Remove.
19503 (aarch64_mov_operand): Use aarch64_mov_operand_p.
19505 2013-05-23 Vidya Praveen <vidyapraveen@arm.com>
19507 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
19508 instruction (AdvSIMD).
19509 * config/aarch64/aarch64-builtins.c
19510 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
19511 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
19513 2013-05-23 Martin Jambor <mjambor@suse.cz>
19515 PR middle-end/57347
19516 * tree.h (contains_bitfld_component_ref_p): Declare.
19517 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
19518 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
19520 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
19521 not access a bit-field. Assert all final offsets are byte-aligned.
19523 2013-05-23 Richard Biener <rguenther@suse.de>
19525 PR tree-optimization/57380
19526 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
19527 least one invariant or re-used load.
19528 * passes.c (init_optimization_passes): Move pass_phiprop before
19531 2013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
19533 * config/aarch64/aarch64-simd.md
19534 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
19536 2013-05-23 Richard Biener <rguenther@suse.de>
19538 PR middle-end/57381
19539 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
19540 OEP_CONSTANT_ADDRESS_OF retained.
19542 2013-05-23 Jakub Jelinek <jakub@redhat.com>
19544 PR middle-end/57344
19545 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
19546 don't lower unit. Handle unit not being always BITS_PER_WORD.
19548 2013-05-23 Richard Biener <rguenther@suse.de>
19550 PR rtl-optimization/57341
19551 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
19552 instead of true_dependence.
19554 2013-05-22 David Malcolm <dmalcolm@redhat.com>
19556 * bb-reorder.c (branch_threshold): Make const.
19557 (exec_threshold): Ditto.
19559 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
19560 Pat Haugen <pthaugen@us.ibm.com>
19561 Peter Bergner <bergner@vnet.ibm.com>
19563 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
19564 documentation for the power8 crypto builtins.
19566 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
19568 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
19569 macros for defining power8 builtin functions.
19570 (BU_P8V_AV_2): Likewise.
19571 (BU_P8V_AV_P): Likewise.
19572 (BU_P8V_VSX_1): Likewise.
19573 (BU_P8V_OVERLOAD_1): Likewise.
19574 (BU_P8V_OVERLOAD_2): Likewise.
19575 (BU_CRYPTO_1): Likewise.
19576 (BU_CRYPTO_2): Likewise.
19577 (BU_CRYPTO_3): Likewise.
19578 (BU_CRYPTO_OVERLOAD_1): Likewise.
19579 (BU_CRYPTO_OVERLOAD_2): Likewise.
19580 (XSCVSPDP): Fix typo, point to the correct instruction.
19581 (VCIPHER): Add power8 crypto builtins.
19582 (VCIPHERLAST): Likewise.
19583 (VNCIPHER): Likewise.
19584 (VNCIPHERLAST): Likewise.
19585 (VPMSUMB): Likewise.
19586 (VPMSUMH): Likewise.
19587 (VPMSUMW): Likewise.
19588 (VPERMXOR_V2DI): Likewise.
19589 (VPERMXOR_V4SI: Likewise.
19590 (VPERMXOR_V8HI: Likewise.
19591 (VPERMXOR_V16QI: Likewise.
19592 (VSHASIGMAW): Likewise.
19593 (VSHASIGMAD): Likewise.
19594 (VPMSUM): Likewise.
19595 (VPERMXOR): Likewise.
19596 (VSHASIGMA): Likewise.
19598 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
19599 __CRYPTO__ if the crypto instructions are available.
19600 (altivec_overloaded_builtins): Add support for overloaded power8
19603 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
19604 support for power8 crypto builtins.
19605 (builtin_function_type): Likewise.
19606 (altivec_init_builtins): Add support for builtins that take vector
19607 long long (V2DI) arguments.
19609 * config/rs6000/crypto.md: New file, define power8 crypto
19612 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
19613 Pat Haugen <pthaugen@us.ibm.com>
19614 Peter Bergner <bergner@vnet.ibm.com>
19616 * doc/invoke.texi (Option Summary): Add power8 options.
19617 (RS/6000 and PowerPC Options): Likewise.
19619 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
19620 constraints.md instead of rs6000.h. Reorder w* constraints. Add
19621 wm, wn, wr documentation.
19623 * config/rs6000/constraints.md (wm): New constraint for VSX
19624 registers if direct move instructions are enabled.
19625 (wn): New constraint for no registers.
19626 (wq): New constraint for quad word even GPR registers.
19627 (wr): New constraint if 64-bit instructions are enabled.
19628 (wv): New constraint if power8 vector instructions are enabled.
19629 (wQ): New constraint for quad word memory locations.
19631 * config/rs6000/predicates.md (const_0_to_15_operand): New
19632 constraint for 0..15 for crypto instructions.
19633 (gpc_reg_operand): If VSX allow registers in VSX registers as well
19634 as GPR and floating point registers.
19635 (int_reg_operand): New predicate to match only GPR registers.
19636 (base_reg_operand): New predicate to match base registers.
19637 (quad_int_reg_operand): New predicate to match even GPR registers
19638 for quad memory operations.
19639 (vsx_reg_or_cint_operand): New predicate to allow vector logical
19640 operations in both GPR and VSX registers.
19641 (quad_memory_operand): New predicate for quad memory operations.
19642 (reg_or_indexed_operand): New predicate for direct move support.
19644 * config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
19645 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
19646 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
19647 (POWERPC_MASKS): Add power8 options.
19648 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
19651 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
19652 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
19654 * config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
19655 (-mpower8-fusion): New power8 options.
19656 (-mpower8-fusion-sign): Likewise.
19657 (-mpower8-vector): Likewise.
19658 (-mcrypto): Likewise.
19659 (-mdirect-move): Likewise.
19660 (-mquad-memory): Likewise.
19662 * config/rs6000/rs6000.c (power8_cost): Initial definition for power8.
19663 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
19665 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
19666 (rs6000_debug_vector_unit): Add p8_vector.
19667 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
19668 definitions. Also print fusion state.
19669 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
19670 (rs6000_builtin_mask_calculate): Add power8 builtin support.
19671 (rs6000_option_override_internal): Add support for power8.
19672 (rs6000_common_init_builtins): Add debugging for skipped builtins
19673 if -mdebug=builtin.
19674 (rs6000_adjust_cost): Add power8 support.
19675 (rs6000_issue_rate): Likewise.
19676 (insn_must_be_first_in_group): Likewise.
19677 (insn_must_be_last_in_group): Likewise.
19678 (force_new_group): Likewise.
19679 (rs6000_register_move_cost): Likewise.
19680 (rs6000_opt_masks): Likewise.
19682 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
19683 power8 capable assembler, default to power7 options.
19684 (TARGET_DIRECT_MOVE): Likewise.
19685 (TARGET_CRYPTO): Likewise.
19686 (TARGET_P8_VECTOR): Likewise.
19687 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
19688 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
19689 (VECTOR_MEM_P8_VECTOR_P): Likewise.
19690 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
19691 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
19692 (TARGET_XSCVDPSPN): Likewise.
19693 (TARGET_XSCVSPDPN): Likewsie.
19694 (TARGET_SYNC_HI_QI): Likewise.
19695 (TARGET_SYNC_TI): Likewise.
19696 (MASK_CRYPTO): Likewise.
19697 (MASK_DIRECT_MOVE): Likewise.
19698 (MASK_P8_FUSION): Likewise.
19699 (MASK_P8_VECTOR): Likewise.
19700 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
19701 temporary used by some of the direct move instructions to get two FP
19702 temporary registers does not force creation of a stack frame.
19703 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
19704 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
19705 that any VSX registers are tieable, even if they are also an
19706 Altivec vector mode.
19707 (r6000_reg_class_enum): Add wm, wr, wv constraints.
19708 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
19709 (RS6000_BTM_CRYPTO): Likewise.
19710 (RS6000_BTM_COMMON): Likewise.
19712 * config/rs6000/rs6000.md (cpu attribute): Add power8.
19713 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
19714 (enum rs6000_vector): Add power8 vector support.
19716 2013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
19720 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
19721 (any_sibcall_could_use_r3): this and handle indirect calls.
19722 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
19724 2013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
19726 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
19728 2013-05-22 Richard Biener <rguenther@suse.de>
19730 PR middle-end/57349
19731 * profile.c (branch_prob): Do not split blocks that are
19732 abnormally receiving from ECF_RETURNS_TWICE functions.
19734 2013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
19736 * recog.c (offsettable_address_addr_space_p): Fix calculation of
19737 address mode. Move pointer mode initialization to the same place.
19739 2013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
19741 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
19742 while it has any effect.
19744 2013-05-21 Easwaran Raman <eraman@google.com>
19746 PR tree-optimization/57322
19747 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
19748 UID of the statement added to the BB to be 1.
19750 2013-05-21 Jakub Jelinek <jakub@redhat.com>
19752 PR tree-optimization/57331
19753 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
19754 of conversion from pointer type to integral type with integer.
19756 2013-05-21 Martin Jambor <mjambor@suse.cz>
19759 * ipa-prop.c (ipa_read_node_info): Process param_used and
19760 controlled_uses in the same order as when writing.
19762 2013-05-21 Magnus Granberg <baldrick@free.fr>
19765 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
19767 2013-05-21 Richard Biener <rguenther@suse.de>
19769 PR tree-optimization/57318
19770 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
19771 estimate stmts with side-effects as likely eliminated.
19773 2013-05-21 Richard Biener <rguenther@suse.de>
19775 PR tree-optimization/57330
19776 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
19777 preserve the call stmts fntype.
19779 2013-05-21 Richard Biener <rguenther@suse.de>
19781 PR tree-optimization/57303
19782 * tree-ssa-sink.c (statement_sink_location): Improve killing
19783 stmt detection and properly handle self-assignments.
19785 2013-05-21 Christian Bruel <christian.bruel@st.com>
19787 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
19788 spanning registers. LEAF_REG_REMAP is supported only for contiguous
19789 registers. Set register size out of the PARALLEL loop.
19791 2013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
19794 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
19795 (*fmasf4, *fmasf4_media): New insns.
19797 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
19799 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
19800 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
19801 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
19802 (mips_idiv_insns): Update the comments to say that the returned
19803 instruction counts are in units of BASE_INSN_LENGTH.
19804 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
19805 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
19806 using 2 rather than 4 as the length of indirect MIPS16 and
19807 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
19808 length of a NOP. Don't divide MIPS16 lengths by 2.
19809 (mips16_split_long_branches): Assume a branch is long if the
19810 length is greater than 4 rather than 8.
19811 * config/mips/mips.md (length): Give MIPS16 lengths directly,
19812 rather than multiplying them by 2. Multiply instruction counts
19813 by BASE_INSN_LENGTH rather than 4.
19814 (*jump_mips16, tls_get_tp_mips16_<mode>)
19815 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
19817 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
19819 * config/mips/mips.md (extended_mips16): Remove branch case.
19820 (length): Remove duplicated extended_mips16 test.
19822 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
19824 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
19826 2013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
19828 * recog.h (Recog_data): Rename to...
19829 (recog_data_d): ...this.
19830 (recog_data): Update accordingly.
19831 * recog.c (recog_data): Likewise.
19832 * reload.c (save_recog_data): Likewise.
19833 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
19834 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
19836 2013-05-17 Julian Brown <julian@codesourcery.com>
19838 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
19839 found in a REG_EQUAL note, invalidate it.
19841 2013-05-17 Easwaran Raman <eraman@google.com>
19843 * tree-ssa-reassoc.c (find_insert_point): New function.
19844 (insert_stmt_after): Likewise.
19845 (get_def_stmt): Likewise.
19846 (ensure_ops_are_available): Likewise.
19847 (not_dominated_by): Likewise.
19848 (rewrite_expr_tree): Do not move statements beyond what is
19849 necessary. Remove call to swap_ops_for_binary_stmt...
19850 (reassociate_bb): ... and move it here.
19851 (build_and_add_sum): Assign UIDs for new statements.
19852 (linearize_expr): Likewise.
19853 (do_reassoc): Renumber gimple statement UIDs.
19855 2013-05-17 Jan Hubicka <jh@suse.cz>
19857 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
19859 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
19860 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
19862 (output_weakrefs): Update.
19864 2013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
19865 Martin Jambor <mjambor@suse.cz>
19867 PR middle-end/57276
19868 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
19869 value that corresponds to the given aggval is found in values vector.
19871 2013-05-17 Uros Bizjak <ubizjak@gmail.com>
19873 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
19874 sse, sse2, sse3, ssse3 and sse4a flags to options.
19876 2013-05-17 David Malcolm <dmalcolm@redhat.com>
19878 * gengtype-state.c: (s_expr_writer): New class, to handle
19879 prettifying of output layout of s-expressions.
19880 (state_writer): New class, to write out gtype.state.
19881 (state_written_type_count): Move this variable into member data of
19883 (s_expr_writer::s_expr_writer): New code: constructor for new class
19884 (state_writer::state_writer(): ditto
19885 (s_expr_writer::write_new_line): New function
19886 (s_expr_writer::write_any_indent): ditto
19887 (s_expr_writer::begin_s_expr): ditto
19888 (s_expr_writer::end_s_expr): ditto
19889 (write_state_fileloc): convert to method of state_writer...
19890 (state_writer:: write_state_fileloc): ...and use methods of
19891 s_expr_writer to write indentation into the gtype.state output file
19892 to visually represent the hierarchical structure of the list
19894 (write_state_fields): ditto, renaming to...
19895 (state_writer::write_state_fields)
19896 (write_state_a_string): ditto, renaming to...
19897 (state_writer::write_state_a_string)
19898 (write_state_string_option): ditto, renaming to...
19899 (state_writer::write_state_string_option)
19900 (write_state_type_option): ditto, renaming to...
19901 (state_writer::write_state_type_option)
19902 (write_state_nested_option): ditto, renaming to...
19903 (state_writer::write_state_nested_option)
19904 (write_state_option): ditto, renaming to...
19905 (state_writer::write_state_option)
19906 (write_state_options): ditto, renaming to...
19907 (state_writer::write_state_options)
19908 (write_state_lang_bitmap): ditto, renaming to...
19909 (state_writer::write_state_lang_bitmap)
19910 (write_state_version): ditto, renaming to...
19911 (state_writer::write_state_version)
19912 (write_state_scalar_type): ditto, renaming to...
19913 (state_writer::write_state_scalar_type)
19914 (write_state_string_type): ditto, renaming to...
19915 (state_writer::write_state_string_type)
19916 (write_state_undefined_type): ditto, renaming to...
19917 (state_writer::write_state_undefined_type)
19918 (write_state_struct_union_type): ditto, renaming to...
19919 (state_writer::write_state_struct_union_type)
19920 (write_state_struct_type): ditto, renaming to...
19921 (state_writer::write_state_struct_type)
19922 (write_state_user_struct_type): ditto, renaming to...
19923 (state_writer::write_state_user_struct_type)
19924 (write_state_lang_struct_type): ditto, renaming to...
19925 (state_writer::write_state_lang_struct_type)
19926 (write_state_param_struct_type): ditto, renaming to...
19927 (state_writer::write_state_param_struct_type)
19928 (write_state_pointer_type): ditto, renaming to...
19929 (state_writer::write_state_pointer_type)
19930 (write_state_array_type): ditto, renaming to...
19931 (state_writer::write_state_array_type)
19932 (write_state_gc_used): ditto, renaming to...
19933 (state_writer::write_state_gc_used)
19934 (write_state_common_type_content): ditto, renaming to...
19935 (state_writer::write_state_common_type_content)
19936 (write_state_type): ditto, renaming to...
19937 (state_writer::write_state_type)
19938 (write_state_pair_list): ditto, renaming to...
19939 (state_writer::write_state_pair_list)
19940 (write_state_pair): ditto, renaming to...
19941 (state_writer::write_state_pair)
19942 (write_state_typedefs): ditto, renaming to...
19943 (state_writer::write_state_typedefs)
19944 (write_state_structures): ditto, renaming to...
19945 (state_writer::write_state_structures)
19946 (write_state_param_structs): ditto, renaming to...
19947 (state_writer::write_state_param_structs)
19948 (write_state_variables): ditto, renaming to...
19949 (state_writer::write_state_variables)
19950 (write_state_srcdir): ditto, renaming to...
19951 (state_writer::write_state_srcdir)
19952 (write_state_files_list): ditto, renaming to...
19953 (state_writer::write_state_files_list)
19954 (write_state_languages): ditto, renaming to...
19955 (state_writer::write_state_languages)
19956 (write_state): create a state_writer instance and use it when
19957 writing out the state file
19959 2013-05-17 Mike Stump <mikestump@comcast.net>
19961 PR rtl-optimization/57304
19962 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
19963 accessing DF_REF_REAL_LOC.
19965 2013-05-17 Jakub Jelinek <jakub@redhat.com>
19967 PR rtl-optimization/57281
19968 PR rtl-optimization/57300
19969 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
19970 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
19971 what the other splitter did if the registers are dead.
19973 2013-05-17 Richard Biener <rguenther@suse.de>
19975 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
19978 2013-05-17 Jakub Jelinek <jakub@redhat.com>
19980 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
19983 2013-05-17 Marek Polacek <polacek@redhat.com>
19985 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
19986 length when doing non-zero store of storing '\0' to '\0'.
19988 2013-05-17 Jakub Jelinek <jakub@redhat.com>
19990 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
19991 vect_external_def oprnd1 with loop_vinfo, try to emit
19992 optional cast, negation and and stmts on the loop preheader
19993 edge instead of into the pattern def seq.
19995 PR tree-optimization/57051
19996 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
19997 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
19999 2013-05-16 Nick Clifton <nickc@redhat.com>
20001 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
20002 (rl78_is_naked_func): New function.
20003 (rl78_expand_prologue): Skip prologue generation for naked functions.
20004 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
20005 * doc/extend.texi (naked): Add RL78 to the list of processors
20006 that supports this attribute.
20008 2013-05-16 Jeff Law <law@redhat.com>
20010 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
20012 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
20014 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
20015 cache parameters using detect_caches_amd also for CYRIX,
20016 NSC and TM2 signatures.
20018 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
20019 Dzianis Kahanovich <mahatma@eu.by>
20023 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
20024 VIA/Centaur processors and determine their cache parameters
20025 using detect_caches_amd.
20027 2013-05-16 Teresa Johnson <tejohnson@google.com>
20029 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
20030 (rtl_verify_edges): New function.
20031 (rtl_verify_bb_insns): Ditto.
20032 (rtl_verify_bb_pointers): Ditto.
20033 (rtl_verify_bb_insn_chain): Ditto.
20034 (rtl_verify_fallthru): Ditto.
20035 (rtl_verify_bb_layout): Ditto.
20036 (rtl_verify_flow_info_1): Outline checks into new functions.
20037 (rtl_verify_flow_info): Ditto.
20039 2013-05-16 Steve Ellcey <sellcey@imgtec.com>
20041 * cfghooks.c (copy_bbs): Add update_dominance argument.
20042 * cfghooks.h (copy_bbs): Update prototype.
20043 * tree-cfg.c (gimple_duplicate_sese_region):
20044 Add update_dominance argument.
20045 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
20046 * tree-ssa-loop-ch.c (copy_loop_headers): Update
20047 gimple_duplicate_sese_region call.
20048 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
20049 Update copy_bbs call.
20050 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
20051 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
20053 2013-05-16 Jakub Jelinek <jakub@redhat.com>
20055 * tree-vectorizer.h (NUM_PATTERNS): Increment.
20056 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
20057 vect_recog_rotate_pattern.
20058 (vect_recog_rotate_pattern): New function.
20060 2013-05-16 Jason Merrill <jason@redhat.com>
20062 * Makefile.in (LLINKER): New variable.
20063 (mostlyclean): Remove link mutex.
20064 * configure.ac: Handle --enable-link-mutex.
20065 * lock-and-run.sh: New script.
20067 2013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
20070 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
20073 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
20075 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
20077 2013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
20079 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
20080 * config/arm/arm.c (next_consecutive_mem): New function.
20081 (gen_movmem_ldrd_strd): Likewise.
20082 * config/arm/arm.md (movmemqi): Update condition and code.
20083 (unaligned_loaddi, unaligned_storedi): New patterns.
20085 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
20087 * config.gcc: Obsolete *-*-solaris2.9*.
20088 * doc/install.texi (Specific, *-*-solaris2*): Document it.
20090 2013-05-16 Richard Biener <rguenther@suse.de>
20092 * passes.c (init_optimization_passes): Move pass_parallelize_loops
20093 earlier, after GRAPHITE transforms and IV canonicalization.
20095 2013-05-16 Jakub Jelinek <jakub@redhat.com>
20097 * omp-low.c (extract_omp_for_data): For collapsed loops,
20098 if at least one of the loops is known at compile time to
20099 iterate zero times, set count to 0.
20100 (expand_omp_regimplify_p): New function.
20101 (expand_omp_for_generic): For collapsed loops, if at least
20102 one of the loops isn't known to iterate at least once,
20103 add runtime check with setting count to 0.
20104 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
20105 For unsigned types if it isn't known at compile time that
20106 the loop will iterate at least once, add runtime check to bypass
20107 the whole loop if initial condition isn't true.
20109 2013-05-16 Nathan Sidwell <nathan@codesourcery.com>
20111 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
20113 2013-05-16 Marc Glisse <marc.glisse@inria.fr>
20115 PR middle-end/57286
20116 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
20117 transformations to avoid an infinite loop.
20119 2013-05-16 Marek Polacek <polacek@redhat.com>
20121 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
20123 2013-05-15 Leif Ekblad <leif@rdos.net>
20125 * config/i386/i386.c (ix86_decompose_address): Use
20126 DEFAULT_TLS_SEG_REG to access TLS segment register.
20127 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
20128 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
20129 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
20131 2013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
20134 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
20135 sibling calls to functions that would normally be lazily bound,
20136 unless $gp is call-clobbered.
20138 2013-05-15 Uros Bizjak <ubizjak@gmail.com>
20140 * config/i386/i386.c (ix86_option_override_internal): Update
20141 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
20142 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
20143 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
20144 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
20146 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
20148 2013-05-15 Andreas Schwab <schwab@suse.de>
20150 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
20151 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
20154 2013-05-15 Teresa Johnson <tejohnson@google.com>
20156 * loop-unroll.c (report_unroll_peel): Check decision before
20157 emitting unroll/peel message.
20159 2013-05-15 Teresa Johnson <tejohnson@google.com>
20161 * function.h (has_bb_partition): New rtl_data flag.
20162 (bb_reorder_complete): Ditto.
20163 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
20164 instead of flag_reorder_blocks_and_partition.
20165 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
20166 with some enhancements.
20167 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
20168 * bb-reorder.c (connect_traces): Check for has_bb_partition
20169 instead of flag_reorder_blocks_and_partition.
20170 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
20171 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
20172 verify_hot_cold_block_grouping.
20173 (partition_hot_cold_basic_blocks): Set has_bb_partition.
20175 2013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
20178 * config/arm/predicates.md (call_insn_operand): New predicate.
20179 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
20180 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
20181 if insn is not a tail call.
20182 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
20184 * config/arm/arm.h (enum reg_class): New caller save register class.
20185 (REG_CLASS_NAMES): Likewise.
20186 (REG_CLASS_CONTENTS): Likewise.
20187 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
20190 2013-05-15 Richard Biener <rguenther@suse.de>
20192 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
20193 of MSG_OPTIMIZED_LOCATIONS.
20194 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
20195 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
20197 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
20198 of MSG_OPTIMIZED_LOCATIONS.
20199 (execute_vect_slp): Likewise.
20200 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
20201 (vect_create_cond_for_alias_checks): Likewise.
20202 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
20203 (vect_recog_widen_mult_pattern): Likewise.
20204 (vect_recog_widen_sum_pattern): Likewise.
20205 (vect_recog_over_widening_pattern): Likewise.
20206 (vect_recog_widen_shift_pattern): Likewise.
20207 (vect_recog_vector_vector_shift_pattern): Likewise.
20208 (vect_recog_divmod_pattern): Likewise.
20209 (vect_recog_mixed_size_cond_pattern): Likewise.
20210 (vect_recog_bool_pattern): Likewise.
20211 (vect_pattern_recog_1): Likewise.
20213 2013-05-15 Martin Jambor <mjambor@suse.cz>
20215 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
20216 non-functions to builtin_unreachable.
20217 * ipa-inline-transform.c (inline_call): Do not assert estimates were
20218 correct when new direct edges were discovered.
20220 2013-05-15 Martin Jambor <mjambor@suse.cz>
20222 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
20223 header, print symbol order instead of node uid, print more information
20224 about indirect edge targets.
20225 (ipa_make_edge_direct_to_target): Print symbol order instead of node
20227 (ipa_make_edge_direct_to_target): Likewise.
20228 (remove_described_reference): Likewise.
20229 (propagate_controlled_uses): Likewise.
20230 (ipa_print_node_params): Also print symbol order.
20231 (ipcp_transform_function): Print symbol order instead of node uids.
20232 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
20233 (cgraph_get_create_real_symbol_node): Likewise.
20234 * ipa-cp.c (print_lattice): Likewise.
20235 (print_all_lattices): Likewise.
20236 (determine_versionability): Likewise.
20237 (initialize_node_lattices): Likewise.
20238 (estimate_local_effects): Likewise.
20239 (update_profiling_info): Likewise.
20240 (create_specialized_node): Likewise.
20241 (perhaps_add_new_callers): Likewise.
20242 (decide_about_value): Likewise.
20243 (decide_whether_version_node): Likewise.
20244 (identify_dead_nodes): Likewise.
20245 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
20246 (dump_inline_summary): Likewise.
20247 (estimate_node_size_and_time): Likewise.
20248 (inline_analyze_function): Likewise.
20249 * ipa-inline.c (report_inline_failed_reason): Likewise.
20250 (want_early_inline_function_p): Likewise.
20251 (edge_badness): Likewise.
20252 (update_edge_key): Likewise.
20253 (inline_small_functions): Likewise. Add dumping of order to two other
20255 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
20256 instead of node uids.
20257 (propagate_pure_const): Likewise.
20258 (propagate_pure_const): Likewise.
20259 * ipa-utils.c (dump_cgraph_node_set): Likewise.
20260 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
20261 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
20263 * tree-pretty-print.c (dump_function_header): Likewise.
20264 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
20265 Print symbol order instead of node uids.
20267 2013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
20269 * config/s390/s390.c (s390_register_move_cost): Don't impose the
20270 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
20272 2013-05-15 Richard Biener <rguenther@suse.de>
20274 PR tree-optimization/57275
20275 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
20276 return value for fail to do runtime alias checks for gather loads.
20278 2013-05-15 Jan Hubicka <jh@suse.cz>
20282 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
20283 weakrefs are not external.
20284 (lto_symtab_merge_decls): Fix thinko when dealing with
20285 non-lto_symtab decls.
20286 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
20287 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
20288 * varpool.c (dump_varpool_node): Dump more flags.
20290 2013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
20292 * config/i386/i386.c (processor_alias_table): Add instruction
20293 FSGSBASE for AMD bdver3 architecture.
20295 2013-05-14 Jakub Jelinek <jakub@redhat.com>
20297 * tree.c (warn_deprecated_use): Print file:line using locus color.
20298 * diagnostic.c (diagnostic_report_current_module): Print file:line
20299 and file:line:column using locus color.
20301 2013-05-14 Mike Stump <mikestump@comcast.net>
20303 * gdbinit.in: Add __null.
20305 2013-05-14 Mike Stump <mikestump@comcast.net>
20307 * recog.h: Rename struct recog_data to Recog_data.
20308 * recog.c: Likewise.
20309 * reload.c (can_reload_into): Likewise.
20310 * config/picochip/picochip.c: Likewise.
20312 2013-05-14 Mike Stump <mikestump@comcast.net>
20314 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
20316 2013-05-14 Steven Bosscher <steven@gcc.gnu.org>
20318 * resource.h (struct resources): Remove unch_memory member.
20319 (CLEAR_RESOURCE): Don't clear unch_memory.
20320 * resource.c (mark_referenced_resources): Don't set it.
20321 (mark_set_resources): Likewise.
20322 (mark_target_live_regs): Don't clear it.
20323 (init_resource_info): Likewise.
20324 * reorg.c (resource_conflicts_p): Don't compare it.
20325 (redundant_insn): Don't set it.
20327 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
20329 * emit-rtl.c (next_label): Remove unused function.
20330 (skip_consecutive_labels, link_cc0_insns): Move to ...
20331 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
20332 only place where these functions are used, and make them static.
20334 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
20336 * fold-const.c (fold_negate_expr): Handle vectors.
20337 (fold_truth_not_expr): Make it static.
20338 (fold_invert_truthvalue): New static function.
20339 (invert_truthvalue_loc): Handle vectors. Do not call
20340 fold_truth_not_expr directly.
20341 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
20342 <TRUTH_NOT_EXPR>: Do not cast to boolean.
20343 (fold_comparison): Handle vector constants.
20344 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
20345 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
20346 * tree.h (fold_truth_not_expr): Remove declaration.
20348 2013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
20350 * config/aarch64/aarch64-simd.md
20351 (aarch64_vcond_internal<mode>): Rename to...
20352 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
20353 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
20354 float modes. Clarify all iterator modes.
20355 (vcond<mode><mode>): Use new name for vcond expanders.
20356 (vcond<v_cmp_result><mode>): Likewise.
20357 (vcondu<mode><mode>: Likewise.
20358 * config/aarch64/iterators.md (VDQF_COND): New.
20360 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
20363 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
20364 variable for the shift amount. Check that we shift by non-negative
20367 2013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
20370 * config/arm/arm.h (EPILOGUE_USES): Only return true
20371 for LR_REGNUM after epilogue_completed.
20373 2013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
20375 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
20376 is error_mark_node.
20378 2013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
20381 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
20382 and Solaris 11+/x86 with gld.
20383 * configure: Regenerate.
20385 2013-05-14 Jakub Jelinek <jakub@redhat.com>
20387 * expmed.c (expand_shift_1): Canonicalize rotates by
20388 constant bitsize / 2 to bitsize - 1.
20389 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
20390 case ROTATERT>: Likewise.
20393 2013-05-10 Jakub Jelinek <jakub@redhat.com>
20395 * config/i386/i386.md (rotateinv): New code attr.
20396 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
20397 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
20398 roll $31, %eax, etc.
20400 2013-05-14 Richard Biener <rguenther@suse.de>
20402 PR middle-end/57235
20403 * tree-eh.c (sink_clobbers): Give up for successors with
20404 multiple predecessors and no virtual uses.
20406 2013-05-14 Eric Botcazou <ebotcazou@adacore.com>
20408 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
20409 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
20411 2013-05-14 Jakub Jelinek <jakub@redhat.com>
20413 PR middle-end/57251
20414 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
20415 the case when both op0 and op1 have VOIDmode.
20417 2013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
20419 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
20420 in multiply-accumulate mode.
20422 2013-05-13 Guozhi Wei <carrot@google.com>
20424 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
20426 2013-05-13 Kai Tietz <ktietz@redhat.com>
20429 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
20430 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
20431 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
20432 * config/i386/i386.c (ix86_option_override_internal): Likewise.
20433 (ix86_expand_prologue): Likewise.
20434 (ix86_expand_split_stack_prologue): Likewise.
20435 (legitimate_pic_address_disp_p): Likewise.
20436 (legitimize_pic_address): Likewise.
20437 (legitimize_tls_address): Likewise.
20438 (legitimize_pe_coff_symbol): Likewise.
20439 (output_pic_addr_const): Likewise.
20440 (construct_plt_address): Likewise.
20441 (ix86_expand_call): Likewise.
20442 (x86_output_mi_thunk): Likewise.
20443 (x86_function_profiler): Likewise.
20445 2013-05-13 Sofiane Naci <sofiane.naci@arm.com>
20447 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
20448 similar switch cases.
20449 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
20450 (aarch64_simd_mov_to_<mode>low): Delete.
20451 (aarch64_simd_mov_to_<mode>high): Delete.
20452 (move_lo_quad_<mode>): Add w<-r alternative.
20453 (aarch64_simd_move_hi_quad_<mode>): Likewise.
20454 (aarch64_simd_mov_from_*): Update type attribute.
20455 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
20458 2013-05-13 Jan Hubicka <jh@suse.cz>
20460 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
20461 * config/i386/i386.c (ix86_compute_frame_layout,
20462 ix86_expand_epilogue, emit_i387_cw_initialization,
20463 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
20464 ix86_local_alignment): Fix use of size/speed predicates.
20466 2013-05-13 Jakub Jelinek <jakub@redhat.com>
20468 PR tree-optimization/45216
20469 PR tree-optimization/57157
20470 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
20471 the (-Y) & (B - 1) variant if OP is |.
20472 * expmed.c (expand_shift_1): For rotations by const0_rtx just
20473 return shifted. Use (-op1) & (prec - 1) as other_amount
20474 instead of prec - op1.
20476 2013-05-13 Martin Jambor <mjambor@suse.cz>
20478 PR middle-end/42371
20479 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
20480 (ipa_constant_data): New type.
20481 (ipa_jump_func): Use ipa_constant_data to hold information about
20482 constant jump functions.
20483 (ipa_get_jf_constant): Adjust to jump function type changes.
20484 (ipa_get_jf_constant_rdesc): New function.
20485 (ipa_param_descriptor): New field controlled_uses.
20486 (ipa_get_controlled_uses): New function.
20487 (ipa_set_controlled_uses): Likewise.
20488 * ipa-ref.h (ipa_find_reference): Declare.
20489 * ipa-prop.c (ipa_cst_ref_desc): New type.
20490 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
20492 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
20493 New parameter cs. Adjust all callers.
20494 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
20495 (remove_described_reference): New function.
20496 (jfunc_rdesc_usable): Likewise.
20497 (try_make_edge_direct_simple_call): Decrement controlled use count,
20498 attempt to remove reference if it hits zero.
20499 (combine_controlled_uses_counters): New function.
20500 (propagate_controlled_uses): Likewise.
20501 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
20502 (ipa_edge_duplication_hook): Duplicate reference descriptions.
20503 (ipa_print_node_params): Print described use counter.
20504 (ipa_write_jump_function): Adjust to jump function type changes.
20505 (ipa_read_jump_function): New parameter CS, pass it to
20506 ipa_set_jf_constant. Adjust caller.
20507 (ipa_write_node_info): Stream controlled use count
20508 (ipa_read_node_info): Likewise.
20509 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
20511 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
20512 count. Remove cloning-added reference if it reaches zero.
20513 * ipa-ref.c (ipa_find_reference): New function.
20515 2013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
20517 * config/i386/i386.c (processor_target_table): Modified default
20518 alignment values for AMD BD and BT architectures.
20520 2013-05-13 Marc Glisse <marc.glisse@inria.fr>
20522 * tree-vect-generic.c (uniform_vector_p): Move ...
20523 * tree.c (uniform_vector_p): ... here.
20524 * tree.h (uniform_vector_p): Declare it.
20525 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
20528 2013-05-13 Jakub Jelinek <jakub@redhat.com>
20530 PR tree-optimization/57230
20531 * tree-ssa-strlen.c (handle_char_store): Record length for
20532 array store from STRING_CST.
20534 PR tree-optimization/57230
20535 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
20538 2013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
20540 * config/epiphany/epiphany.c (epiphany_init): Check size of
20541 NUM_MODES_FOR_MODE_SWITCHING.
20542 (epiphany_expand_prologue):
20543 Remove CONFIG_REGNUM initial value handling code.
20544 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
20545 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
20546 (emit_set_fp_mode, epiphany_mode_after): Likewise.
20547 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
20548 Don't return 1 for FP_MODE_NONE.
20549 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
20550 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
20551 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
20552 * config/epiphany/epiphany.md (save_config): New pattern.
20554 2013-05-12 Uros Bizjak <ubizjak@gmail.com>
20556 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
20558 2013-05-10 Uros Bizjak <ubizjak@gmail.com>
20560 * config/i386/i386.md (memory): Handle sseishft1.
20561 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
20562 (*vec_extractv2di_1): Ditto.
20564 2013-05-10 Vladimir Makarov <vmakarov@redhat.com>
20566 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
20569 2013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
20571 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
20572 Add mthumb/march=armv7-a multilib.
20573 Add mthumb/march=armv7-r multilib.
20574 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
20576 2013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
20578 * config/v850/t-rtems: Add more multilibs.
20580 2013-05-10 Richard Biener <rguenther@suse.de>
20582 PR tree-optimization/57214
20583 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
20584 not propagate from SSA names that occur in abnormal PHI nodes.
20586 2013-05-10 Marc Glisse <marc.glisse@inria.fr>
20588 * stor-layout.c (element_precision): New function.
20589 * machmode.h (element_precision): Declare it.
20590 * tree.c (build_minus_one_cst): New function.
20591 (element_precision): Likewise.
20592 * tree.h (build_minus_one_cst): Declare new function.
20593 (element_precision): Likewise.
20594 * fold-const.c (operand_equal_p): Use element_precision.
20595 (fold_binary_loc): Handle vector types.
20596 * convert.c (convert_to_integer): Use element_precision.
20597 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
20600 2013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
20602 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
20603 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
20604 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
20605 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
20606 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
20607 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
20608 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
20609 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
20610 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
20611 (Uuw8): New constraints.
20612 (Usb4): Move into alphabetical order.
20613 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
20614 (sd8_operand, ub8_operand, uw8_operand): New predicates.
20615 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
20616 previously unnamed patterns.
20617 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
20618 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
20619 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
20620 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
20621 of set_attr_alternative/if_then_else. Use extended_mips16 instead
20622 of specific lengths.
20624 2013-05-10 Jakub Jelinek <jakub@redhat.com>
20626 * config/i386/i386.md (rotateinv): New code attr.
20627 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
20628 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
20629 roll $31, %eax, etc.
20631 PR tree-optimization/45216
20632 PR tree-optimization/57157
20633 * tree-ssa-forwprop.c (simplify_rotate): New function.
20634 (ssa_forward_propagate_and_combine): Call it.
20636 2013-05-10 Richard Biener <rguenther@suse.de>
20638 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
20639 disable peeling when we version for aliasing.
20640 (vector_alignment_reachable_p): Honor explicit user alignment.
20641 (vect_supportable_dr_alignment): Likewise.
20642 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
20643 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
20644 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
20645 then peeling to arrange for the cost-model check to come first.
20647 2013-05-10 Alan Modra <amodra@gmail.com>
20649 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
20650 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
20651 * configure: Regenerate.
20653 2013-05-10 Alan Modra <amodra@gmail.com>
20656 * varasm.c (default_elf_select_section): Move !DECL_P check..
20657 (get_named_section): ..to here before calling get_section_name.
20659 (default_section_type_flags): Add DECL_P check.
20660 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
20661 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
20663 2013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
20665 * config/epiphany/epiphany.c (epiphany_expand_prologue):
20666 When using gen_stack_adjust_str with a register offset, add a
20667 REG_FRAME_RELATED_EXPR note.
20669 2013-05-09 Uros Bizjak <ubizjak@gmail.com>
20671 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
20672 (*vec_extractv4si_zext_mem): Ditto.
20673 (*vec_extractv2di): Add 0->x and x->x alternatives.
20674 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
20675 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
20677 2013-05-09 Jason Merrill <jason@redhat.com>
20679 N3639 C++1y VLA support
20680 * gimplify.c (gimplify_vla_decl): Don't touch an existing
20683 * tree.c (build_constructor_va): New.
20684 * tree.h: Declare it.
20686 2013-05-09 Martin Jambor <mjambor@suse.cz>
20689 * gimple-fold.c (canonicalize_constructor_val): Call
20690 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
20692 2013-05-09 Jan Hubicka <jh@suse.cz>
20693 Richard Biener <rguenther@suse.de>
20696 * symtab.c (symtab_make_decl_local): Do not add private names.
20698 2013-05-09 Jan Hubicka <jh@suse.cz>
20701 * symtab.c (insert_to_assembler_name_hash): Handle clones.
20702 (unlink_from_assembler_name_hash): Likewise.
20703 (symtab_prevail_in_asm_name_hash, symtab_register_node,
20704 symtab_unregister_node, symtab_initialize_asm_name_hash,
20705 change_decl_assembler_name): Update.
20707 2013-05-09 Sofiane Naci <sofiane.naci@arm.com>
20709 * config/aarch64/aarch64.md: New movtf split.
20710 (*movtf_aarch64): Update.
20711 (aarch64_movdi_tilow): Handle TF modes and rename to
20712 aarch64_movdi_<mode>low.
20713 (aarch64_movdi_tihigh): Handle TF modes and rename to
20714 aarch64_movdi_<mode>high
20715 (aarch64_movtihigh_di): Handle TF modes and rename to
20716 aarch64_mov<mode>high_di
20717 (aarch64_movtilow_di): Handle TF modes and rename to
20718 aarch64_mov<mode>low_di
20719 (aarch64_movtilow_tilow): Remove spurious whitespace.
20720 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
20722 (aarch64_print_operand): Update.
20724 2013-05-09 Alan Modra <amodra@gmail.com>
20726 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
20728 * configure: Regenerate.
20730 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
20732 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
20733 splitter preparation statements.
20734 * config/i386/sse.md (*vec_extract* splitters): Ditto.
20735 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
20738 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20740 * gimple-ssa-strength-reduction.c (count_candidates): Change
20741 return value to int.
20742 (analyze_candidates_and_replace): Change type of length to int.
20744 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
20746 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
20747 (*vec_extract<mode>): Use VI12_128 mode iterator.
20748 (*vec_extract<mode>_mem): Ditto.
20749 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
20752 2013-05-08 Diego Novillo <dnovillo@google.com>
20757 2012-08-17 Diego Novillo <dnovillo@google.com>
20760 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
20761 * config.in: Regenerate.
20762 * configure: Regenerate.
20763 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
20765 2013-05-08 Jan Hubicka <jh@suse.cz>
20768 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
20769 * cgraph.h (symtab_node_base): Add unique_name.
20770 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
20771 input_overwrite_node, input_varpool_node): Stream unique_name.
20772 * cgraphclones.c (cgraph_create_virtual_clone,
20773 cgraph_function_versioning): Set unique_name.
20774 * ipa.c (function_and_variable_visibility): Set unique_name.
20776 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20778 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
20779 (alloc_cand_and_find_basis): Restrict conditional candidate
20780 processing to CAND_MULTs.
20782 2013-05-08 Jan Hubicka <jh@suse.cz>
20785 lto-symtab.c (lto_symtab_symbol_p): New function.
20786 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
20787 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
20788 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
20789 Skip static symbols.
20791 2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
20793 PR tree-optimization/57200
20794 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
20795 Only call inform if the preceding warning_at returns true.
20797 2013-05-07 Han Shen <shenhan@google.com>
20799 * cfgexpand.c (record_or_union_type_has_array_p): New function.
20800 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
20801 * common.opt (fstack-protector-strong): New option.
20802 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
20803 * doc/invoke.texi (Optimization Options): Document
20804 "-fstack-protector-strong".
20805 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
20807 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
20809 * config/mips/mips.c (mips_machine_reorg2): Return 0.
20811 2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
20813 * ira.c (update_equiv_regs): Add insn having equiv memory even if
20814 it is not lhs of the insn.
20815 (setup_reg_equiv): Remove insn having equiv memory which it is not
20817 * lra-constraints.c (process_address): Try to improve generation
20818 code for address base + disp.
20819 (lra_constraints): Make correct the code for checking insn setting
20820 up backward equivalence. Remove insn only if it is in the init
20822 * lra-eliminations.c (update_reg_eliminate): Change return value.
20823 (lra_eliminate): Use the result.
20825 2013-05-07 Uros Bizjak <ubizjak@gmail.com>
20827 * config/i386/sse.md (ssescalarnummask): New mode attribute.
20828 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
20829 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
20830 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
20831 register target operands.
20832 (*vec_extractv8hi_sse2): New pattern.
20833 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
20834 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
20835 (*vec_extract<mode>_mem): New insn and split pattern.
20837 2013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
20839 * config/arm/arm.c (arm_asan_shadow_offset): New function.
20840 (TARGET_ASAN_SHADOW_OFFSET): Define.
20841 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
20842 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
20844 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20846 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
20847 (incr_vec_index): Return -1 if increment not found.
20848 (create_add_on_incoming_edge): Assert if increment not found.
20849 (record_increment): Limit number of increments recorded.
20850 (all_phi_incrs_profitable): Return false if an increment not found.
20851 (replace_profitable_candidates): Don't process increments that were
20853 (analyze_candidates_and_replace): Limit size of incr_vec.
20855 2013-05-07 Richard Biener <rguenther@suse.de>
20857 * calls.c (special_function_p): setjmp-like functions are leaf.
20858 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
20859 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
20861 2013-05-07 Sofiane Naci <sofiane.naci@arm.com>
20863 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
20864 (aarch64_simd_mov<mode>): New expander.
20865 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
20866 (aarch64_simd_mov_to_<mode>high): Likewise.
20867 (aarch64_simd_mov_from_<mode>low): Likewise.
20868 (aarch64_simd_mov_from_<mode>high): Likewise.
20869 (aarch64_dup_lane<mode>): Update.
20870 (aarch64_dup_lanedi): New instruction pattern.
20871 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
20872 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
20874 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20876 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
20877 (replace_mult_candidate): Remove unnecessary argument; remove
20878 unnecessary parameter from call to introduce_cast_before_cand.
20879 (replace_unconditional_candidate): Remove unnecessary parameter
20880 from call to replace_mult_candidate.
20881 (replace_conditional_candidate): Likewise.
20882 (insert_initializers): Use make_temp_ssa_name.
20883 (introduce_cast_before_cand): Remove unnecessary argument; use
20884 make_temp_ssa_name.
20885 (replace_one_candidate): Remove unnecessary argument; remove
20886 unnecessary parameter from calls to introduce_cast_before_cand.
20887 (replace_profitable_candidates): Remove unnecessary parameters
20888 from calls to replace_one_candidate.
20890 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20892 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
20893 phi def as possibly hiding a basis for a CAND_ADD whose operands
20894 have been commuted in the analysis.
20895 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
20897 2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
20899 * config/aarch64/aarch64.md
20900 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
20901 shift value between 0-4.
20903 2013-05-07 Richard Biener <rguenther@suse.de>
20905 * double-int.h (rshift): New overload.
20906 * double-int.c (rshift): New function.
20907 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
20908 (create_reference_ops_from_ref): Remove.
20909 (vn_reference_insert): Use shared ops for constructing the
20910 reference and copy it.
20912 2013-05-07 Richard Biener <rguenther@suse.de>
20914 PR middle-end/57190
20915 * tree-eh.c (sink_clobbers): Properly propagate
20916 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
20918 2013-05-07 Jakub Jelinek <jakub@redhat.com>
20920 PR tree-optimization/57149
20921 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
20922 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
20923 collect_phi_def_edges, execute_late_warn_uninitialized): Use
20924 uninit_undefined_value_p instead of ssa_undefined_value_p.
20927 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
20928 for modifier == EXPAND_INITIALIZER.
20930 2013-05-07 Anton Blanchard <anton@samba.org>
20932 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
20933 for powerpc64 little endian.
20934 * configure: Regenerate.
20936 2013-05-06 Graham Stott <grahams@btinternet.com>
20938 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
20939 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
20940 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
20941 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
20943 2013-05-06 Graham Stott <grahams@btinternet.com>
20945 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
20946 codes which allow non-lvalues.
20948 2013-05-06 Marc Glisse <marc.glisse@inria.fr>
20950 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
20951 components are all 1s.
20952 (integer_minus_onep): New function.
20953 * tree.h (integer_minus_onep): Declare it.
20954 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
20955 integer_minus_onep instead of integer_all_onesp.
20957 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
20960 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
20961 variations of these patterns.
20963 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
20965 * config/i386/i386.md (isa): Add x64_sse4 member.
20966 (enabled): Handle x64_sse4.
20967 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
20968 instruction for 64bit SSE4_1 targets. Update insn attributes.
20969 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
20970 instruction for SSE4_1 targets. Update insn attributes.
20971 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
20972 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
20973 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
20975 (*vec_extractv4si): Rename from *sse4_1_pextrd.
20976 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
20977 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
20979 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
20982 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
20984 2013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
20986 * final.c (do_assembler_dialects): Don't handle curly braces and
20987 vertical bar escaped by % as dialect delimiters.
20988 (output_asm_insn): Print curly braces and vertical bar if escaped
20989 by % and ASSEMBLER_DIALECT defined.
20990 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
20991 * doc/tm.texi: Regenerated.
20993 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
20995 * config/mips/mips.c: Include tree-pass.h.
20996 (mips_reorg): Split in pre- and post-dbr_schedule parts.
20997 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
20998 (pass_mips_machine_reorg2): New machine specific pass.
20999 (insert_pass_mips_machine_reorg2): New pass plugin definition.
21000 (mips_option_override): Register the new pass.
21001 * rtl.h (cleanup_barriers): Remove prototype.
21002 (dbr_schedule): Likewise.
21003 * jump.c (cleanup_barriers): Make static.
21004 * reorg.c (dbr_schedule): Likewise.
21006 2013-05-06 Richard Biener <rguenther@suse.de>
21008 PR tree-optimization/57185
21009 * tree-parloops.c (add_field_for_reduction): Handle anonymous
21010 SSA names properly.
21012 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
21015 * config/i386/i386.c (add_parameter_dependencies): Add dependence
21016 between "first_arg" and "insn", not "last" and "insn".
21018 2013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
21020 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
21021 (find_candidates_in_block): Re-enable slsr_process_phi.
21022 (create_phi_basis): Fix double counting of candidate adjustment.
21024 2013-05-06 Richard Biener <rguenther@suse.de>
21026 PR middle-end/57147
21027 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
21028 the edge is also fallthru, preserve it and just clear the
21030 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
21031 also complex, preserve that and just clear the fallthru flag.
21032 * tree-inline.c (update_ssa_across_abnormal_edges): Also
21033 update virtual operands.
21035 2013-05-06 Alan Modra <amodra@gmail.com>
21037 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
21038 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
21039 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
21040 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
21041 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
21042 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
21044 2013-05-06 Alan Modra <amodra@gmail.com>
21046 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
21047 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
21048 (DEFAULT_ASM_ENDIAN): Define.
21049 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
21050 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
21051 Update -K PIC clause from sysv4.h.
21052 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
21053 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
21055 2013-05-06 Alan Modra <amodra@gmail.com>
21057 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
21058 twice for little-endian.
21059 (ashrdi3_no_power, ashrdi3): Support little-endian.
21061 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
21064 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
21065 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
21067 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
21068 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
21069 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
21072 2013-05-05 Steven Bosscher <steven@gcc.gnu.org>
21073 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
21075 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
21076 * config/pa/pa.opt: Make mbig-switch a no-op.
21077 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
21078 (CASE_VECTOR_MODE): Always return SImode.
21079 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
21080 for the !TARGET_BIG_SWITCH case.
21081 * config/pa/pa-linux.h: Likewise.
21082 * config/pa/pa-openbsd.h: Likewise.
21083 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
21084 * config/pa/pa.md (short_jump): Remove define_insn.
21085 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
21086 (casesi0): Remove define_insn.
21087 (type): Remove btable_branch.
21088 (pa_combine_type): Likewise.
21089 (in_nullified_branch_delay): Likewise.
21090 (in_call_delay): Likewise.
21091 (define_delay): Likewise.
21092 (define_insn_reservation "Z3"): Likewise.
21093 (define_insn_reservation "Z4"): Likewise.
21094 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
21095 (pa_adjust_insn_length): Remove adjustment for btable branches.
21096 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
21099 2013-05-05 Uros Bizjak <ubizjak@gmail.com>
21101 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
21102 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
21103 Add m->r,x alternatives.
21104 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
21105 splitters using SWI48x mode iterator.
21106 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
21107 TARGET_64BIT. Add m->x alternative.
21108 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
21109 Add o->x alternative. Enable for TARGET_SSE.
21110 (sse_storeq): Remove expander.
21111 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
21112 with memory input operand.
21113 (*vec_extractv2di_1 splitter): New.
21114 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
21115 * config/i386/i386.md (ssevecmodelower): New mode attribute.
21117 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
21119 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
21120 (INT_LOWPART): Delete.
21121 (extract_MB): Adjust.
21122 (extract_ME): Adjust.
21123 (print_operand): Adjust.
21125 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
21127 * config/rs6000/predicates.md (reg_or_add_cint_operand,
21128 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
21129 (reg_or_logical_cint_operand, easy_fp_constant,
21130 logical_const_operand): Delete "CONST_DOUBLE" case.
21131 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
21132 "HOST_BITS_PER_WIDE_INT == 64" test.
21133 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
21134 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
21135 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
21136 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
21138 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
21139 CONST_DOUBLE DImode/VOIDmode case.
21140 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
21141 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
21142 CONST_DOUBLE VOIDmode case.
21143 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
21144 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
21145 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
21146 Delete CONST_DOUBLE case.
21147 (splitters for mov FMOVE64 const_double): Delete
21148 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
21149 "HOST_BITS_PER_WIDE_INT >= 64" test.
21150 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
21152 (mov DI const_double): Delete.
21154 2013-05-04 Jakub Jelinek <jakub@redhat.com>
21156 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
21157 on op shows all bits zero in mode of a lowpart subreg, return zero.
21159 2013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
21162 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
21163 to save TFmode registers and DImode to save TImode registers for
21164 caller save operations.
21165 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
21166 mark being partially clobbered since they only use the first
21169 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
21170 and TDmode only use the upper 64-bits of each VSX register.
21172 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
21174 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
21175 (find_candidates_in_block): Disable slsr_process_phi.
21177 2013-05-03 Guozhi Wei <carrot@google.com>
21179 * coverage.c (coverage_obj_init): Move the construction of gcov
21181 (build_init_ctor): ... here.
21183 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
21185 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
21186 (slsr_cand_d): Redefine def_phi.
21187 (stride_status, phi_adjust_status, count_phis_status): New enums.
21188 (find_phi_def): New.
21189 (find_basis_for_base_expr): New.
21190 (find_basis_for_candidate): Handle hidden bases.
21191 (alloc_cand_and_find_basis): Handle phi candidates.
21192 (slsr_process_phi): New.
21193 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
21194 (create_mul_imm_cand): Likewise.
21195 (create_add_ssa_cand): Exclude phi base candidates.
21196 (create_add_imm_cand): Likewise.
21197 (slsr_process_cast): Likewise.
21198 (slsr_process_copy): Likewise.
21199 (find_candidates_in_block): Handle phi candidates.
21200 (dump_candidate): Likewise.
21201 (unconditional_cands): Delete.
21202 (unconditional_cands_with_known_stride_p): Delete.
21203 (phi_dependent_cand_p): New.
21204 (cand_increment): Handle phi-dependent candidates.
21205 (replace_dependent): Delete.
21206 (replace_mult_candidate): New.
21207 (replace_unconditional_candidate): New.
21208 (incr_vec_index): Move to avoid forward reference.
21209 (create_add_on_incoming_edge): New.
21210 (create_phi_basis): New.
21211 (replace_dependents): Delete.
21212 (replace_conditional_candidate): New.
21213 (phi_add_costs): New.
21214 (replace_uncond_cands_and_profitable_phis): New.
21215 (record_increment): Handle phi adjustments.
21216 (record_phi_increments): New.
21217 (record_increments): Handle phi adjustments.
21218 (phi_incr_cost): New.
21219 (lowest_cost_path): Handle phis.
21220 (total_savings): Likewise.
21221 (analyze_increments): Likewise.
21222 (ncd_with_phi): New.
21223 (ncd_of_cand_and_phis): New.
21224 (nearest_common_dominator_for_cands): Handle phi increments.
21225 (all_phi_incrs_profitable): New.
21226 (replace_profitable_candidates): Handle phi-dependent candidates.
21227 (analyze_candidates_and_replace): Likewise.
21229 2013-05-03 Teresa Johnson <tejohnson@google.com>
21232 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
21233 do not exceed REG_BR_PROB_BASE.
21235 2013-05-03 Jeff Law <law@redhat.com>
21237 PR tree-optimization/57144
21238 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
21239 operand of the condition will bit into the new type when eliminating
21240 a cast feeding a condition.
21242 2013-05-03 Jakub Jelinek <jakub@redhat.com>
21244 PR rtl-optimization/57130
21245 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
21246 of COMPARE as in_code to the recursive call if needed.
21248 2013-05-03 Uros Bizjak <ubizjak@gmail.com>
21250 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
21251 (enabled): Handle new members.
21252 * config/i386/sse.md (*vec_concatv2si): Merge from
21253 *vec_concatv2si_sse2 and vec_concatv2si_sse.
21254 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
21256 2013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
21258 PR tree-optimization/57027
21259 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
21260 for fnms opportunity, check we got the prerequisite kind
21261 of tree / gimple before using accessor functions.
21263 2013-05-03 Richard Biener <rguenther@suse.de>
21265 * double-int.h (lshift): New overload without precision
21266 and arith argument.
21267 (operator *=, operator +=, operator -=): Move ...
21268 * double-int.c (operator *=, operator +=, operator -=): ... here
21269 and implement more efficiently.
21270 (mul_double_with_sign): Remove.
21271 (lshift_double): Adjust to take unsinged shift argument, push
21272 dispatching code to callers.
21273 (mul_double_wide_with_sign): Add early out for callers that
21274 are not interested in high parts or overflow.
21275 (lshift): New function.
21276 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
21277 dispatch code here.
21278 (lrotate, rrotate): Use logical shifts.
21279 * expr.c (get_inner_reference): Use lshift.
21280 * fixed-value.c (do_fixed_divide): Likewise.
21281 * tree-dfa.c (get_ref_base_and_extent): Likewise.
21282 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
21283 (indirect_refs_may_alias_p): Likewise.
21284 (stmt_kills_ref_p_1): Likewise.
21286 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
21288 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
21290 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
21292 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
21293 scalar form of FABD instruction.
21295 2013-05-02 Vladimir Makarov <vmakarov@redhat.com>
21297 * lra-constraints.c (process_alt_operands): Add checking alt
21298 number to choose the best alternative.
21300 2013-05-02 Richard Biener <rguenther@suse.de>
21302 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
21303 bitmap and its handling.
21304 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
21306 2013-05-02 Richard Biener <rguenther@suse.de>
21308 PR middle-end/57140
21309 * tree-inline.c (copy_loops): Properly handle removed loops.
21310 (copy_cfg_body): Mark destination loops for fixup if source
21311 loops needed fixup.
21313 2013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
21316 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
21317 generating simple_return for naked functions.
21319 2013-05-02 Martin Jambor <mjambor@suse.cz>
21321 PR middle-end/56988
21322 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
21323 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
21325 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
21326 ipa_agg_replacement_value structures.
21327 (known_aggs_to_agg_replacement_list): Likewise.
21328 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
21329 (read_agg_replacement_chain): Likewise.
21330 (ipcp_transform_function): Also check that by_ref flags match.
21332 2013-05-02 Richard Biener <rguenther@suse.de>
21334 * graphds.h (struct graph): Add obstack member.
21335 * graphds.c (new_graph): Initialize obstack and allocate
21337 (add_edge): Allocate edge from the obstack.
21338 (free_graph): Free the obstack instead of all edges and vertices.
21340 2013-05-02 Teresa Johnson <tejohnson@google.com>
21342 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
21344 * cfg.c (update_bb_profile_for_threading): Ditto.
21345 * tree-inline.c (copy_bb): Ditto.
21346 (copy_edges_for_bb): Ditto.
21347 (initialize_cfun): Ditto.
21348 (copy_cfg_body): Ditto.
21349 (expand_call_inline): Ditto.
21350 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
21351 (estimate_node_size_and_time): Ditto.
21352 (inline_merge_summary): Ditto.
21353 * cgraphclones.c (cgraph_clone_edge): Ditto.
21354 (cgraph_clone_node): Ditto.
21355 * sched-rgn.c (compute_dom_prob_ps): Ditto.
21356 (compute_trg_info): Ditto.
21358 2013-05-02 Ian Bolton <ian.bolton@arm.com>
21360 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
21361 S reg when fp attribute set.
21362 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
21364 2013-05-02 Ian Bolton <ian.bolton@arm.com>
21366 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
21368 (*and_one_cmplsi3_compare0_uxtw): Likewise.
21369 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
21370 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
21372 2013-05-02 Richard Biener <rguenther@suse.de>
21374 * tree-scalar-evolution.c (scev_info_hasher): Remove.
21375 (struct instantiate_cache_entry): New type.
21376 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
21377 (struct instantiate_cache_type): New type.
21378 (set_instantiated_value, get_instantiated_value): Remove.
21379 (get_instantiated_value_entry): New function.
21380 (instantiate_scev_name): Use the new cache and adjust.
21381 (instantiate_scev_poly): Adjust.
21382 (instantiate_scev_binary): Likewise.
21383 (instantiate_array_ref): Likewise.
21384 (instantiate_scev_convert): Likewise.
21385 (instantiate_scev_not): Likewise.
21386 (instantiate_scev_3): Likewise.
21387 (instantiate_scev_2): Likewise.
21388 (instantiate_scev_r): Likewise.
21389 (instantiate_scev): Likewise.
21390 (resolve_mixers): Likewise.
21392 2013-05-01 Vladimir Makarov <vmakarov@redhat.com>
21395 * lra-constraints.c (best_small_class_operands_num): Remove.
21396 (process_alt_operands): Remove small_class_operands_num. Take
21397 small classes operands into losers and only if the operand is not
21398 matched. Modify debugging output.
21399 (curr_insn_transform): Remove best_small_class_operands_num.
21402 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21404 * config/aarch64/aarch64-builtins.c
21405 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
21406 * config/aarch64/aarch64-simd-builtins.def
21407 (reduc_splus_): Add new modes.
21408 (reduc_uplus_): New.
21409 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
21410 (reduc_uplus_v4sf): Likewise.
21411 (reduc_splus_v4sf): Likewise.
21412 (aarch64_addv<mode>): Likewise.
21413 (reduc_uplus_<mode>): Likewise.
21414 (reduc_splus_<mode>): Likewise.
21415 (aarch64_addvv2di): Likewise.
21416 (reduc_uplus_v2di): Likewise.
21417 (reduc_splus_v2di): Likewise.
21418 (aarch64_addvv2si): Likewise.
21419 (reduc_uplus_v2si): Likewise.
21420 (reduc_splus_v2si): Likewise.
21421 (reduc_<sur>plus_<mode>): New.
21422 (reduc_<sur>plus_v2di): Likewise.
21423 (reduc_<sur>plus_v2si): Likewise.
21424 (reduc_<sur>plus_v4sf): Likewise.
21425 (aarch64_addpv4sf): Likewise.
21426 * config/aarch64/arm_neon.h
21427 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
21428 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
21429 add UNSPEC_SADDV, UNSPEC_UADDV.
21431 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
21433 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21435 * config/aarch64/arm_neon.h
21436 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
21438 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21440 * config/aarch64/aarch64-builtins
21441 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
21443 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21445 * config/aarch64/aarch64-simd-builtins.def
21446 (reduc_smax_): New.
21447 (reduc_smin_): Likewise.
21448 (reduc_umax_): Likewise.
21449 (reduc_umin_): Likewise.
21450 (reduc_smax_nan_): Likewise.
21451 (reduc_smin_nan_): Likewise.
21454 (smax): Update for V2SF, V4SF and V2DF modes.
21457 (smin_nan): Likewise.
21458 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
21459 (<su><maxmin><mode>3): ...This, refactor.
21460 (s<maxmin><mode>3): New.
21461 (<maxmin_uns><mode>3): Likewise.
21462 (reduc_<maxmin_uns>_<mode>): Refactor.
21463 (reduc_<maxmin_uns>_v4sf): Likewise.
21464 (reduc_<maxmin_uns>_v2si): Likewise.
21465 (aarch64_<fmaxmin><mode>: Remove.
21466 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
21468 (vmin<q>_f<32,64>): Likewise.
21469 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
21471 (su): Add mappings for smax, smin, umax, umin.
21473 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
21474 (FMAXMIN): Rename as...
21475 (FMAXMIN_UNS): ...This.
21477 (fmaxminv): Likewise.
21478 (fmaxmin): Likewise.
21480 (maxmin_uns_op): Likewise.
21482 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21484 * config/aarch64/arm_neon.h
21485 (vac<ge, gt><sd>_f<32, 64>): Rename to...
21486 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
21487 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
21489 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21491 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
21492 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
21494 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21496 * config/aarch64/aarch64-simd.md
21497 (vcond<mode>_internal): Handle special cases for constant masks.
21498 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
21499 (vcondu<mode><mode>): Likewise.
21500 (vcond<v_cmp_result><mode>): New.
21502 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21504 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
21505 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
21506 * config/aarch64/aarch64-simd-builtins.def
21507 (cmeq): Update to BUILTIN_VALLDI.
21512 * config/aarch64/arm_neon.h
21513 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
21514 to builtins or C as appropriate.
21516 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21518 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
21520 (cmhi): Rename to...
21522 * config/aarch64/aarch64-simd.md
21523 (simd_mode): Add SF.
21524 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
21525 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
21526 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
21527 (cstore<mode>_neg): ...This.
21528 * config/aarch64/iterators.md
21530 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
21531 (COMPARISONS): New.
21532 (UCOMPARISONS): Likewise.
21533 (optab): Add missing comparisons.
21539 (VCMP_S): Likewise.
21540 (VCMP_U): Likewise.
21541 (V_cmp_result): Add DF, SF modes.
21542 (v_cmp_result): Likewise.
21544 (vmtype): Likewise.
21545 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
21547 2013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
21549 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
21550 define_insn to define_insn_and_split.
21551 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
21552 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
21553 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
21554 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
21555 (thumb2_negscc): Likewise.
21557 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
21559 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
21561 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
21563 * config/arm/thumb2.md: Remove trailing whitespaces.
21565 2013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
21567 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
21568 Use gen_int_mode rather than GEN_INT.
21570 2013-04-30 H.J. Lu <hongjiu.lu@intel.com>
21572 * value-prof.c (stream_in_histogram_value): Remove the strayed
21575 2013-04-30 Richard Biener <rguenther@suse.de>
21577 PR middle-end/57122
21578 * cfghooks.c (split_edge): Properly check for the loop latch edge.
21580 2013-04-30 Richard Biener <rguenther@suse.de>
21582 PR middle-end/57107
21583 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
21585 2013-04-30 Andrey Belevantsev <abel@ispras.ru>
21587 PR rtl-optimization/56957
21588 PR rtl-optimization/57105
21589 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
21590 variable. Use just INSN_UID for determining whether an insn
21591 should be only disconnected from the insn stream.
21592 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
21594 2013-04-30 Jakub Jelinek <jakub@redhat.com>
21596 PR tree-optimization/57104
21597 * tsan.c (instrument_expr): Don't instrument accesses to
21598 DECL_HARD_REGISTER VAR_DECLs.
21600 2013-04-30 Richard Biener <rguenther@suse.de>
21602 * function.h (loops_for_fn): New inline function.
21603 (set_loops_for_fn): Likewise.
21604 * cfgloop.h (place_new_loop): Add struct function parameter.
21605 (get_loop): Likewise.
21606 (get_loops): Likewise.
21607 (number_of_loops): Likewise.
21608 (fel_next): Adjust.
21609 (fel_init): Likewise.
21610 * cfg.c (get_loop_copy): Adjust.
21611 * cfgloop.c (flow_loops_dump): Likewise.
21612 (record_loop_exits): Likewise.
21613 (verify_loop_structure): Likewise.
21614 * cfgloopanal.c (mark_irreducible_loops): Likewise.
21615 (estimate_reg_pressure_cost): Likewise.
21616 (mark_loop_exit_edges): Likewise.
21617 * cfgloopmanip.c (place_new_loop): Likewise.
21618 (add_loop): Likewise.
21619 (duplicate_loop): Likewise.
21620 * graph.c (draw_cfg_nodes): Likewise.
21621 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
21622 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
21623 (extract_affine_chrec): Likewise.
21624 (build_scop_iteration_domain): Likewise.
21625 * graphite.c (graphite_initialize): Likewise.
21626 * ira-build.c (create_loop_tree_nodes): Likewise.
21627 (more_one_region_p): Likewise.
21628 (rebuild_regno_allocno_maps): Likewise.
21629 (mark_loops_for_removal): Likewise.
21630 (mark_all_loops_for_removal): Likewise.
21631 (remove_unnecessary_regions): Likewise.
21632 (ira_build): Likewise.
21633 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
21634 * loop-init.c (fix_loop_structure): Likewise.
21635 (gate_rtl_move_loop_invariants): Likewise.
21636 (gate_rtl_unswitch): Likewise.
21637 (gate_rtl_unroll_and_peel_loops): Likewise.
21638 (rtl_doloop): Likewise.
21639 * lto-streamer-in.c (input_cfg): Likewise.
21640 * lto-streamer-out.c (output_cfg): Likewise.
21641 * modulo-sched.c (sms_schedule): Likewise.
21642 * predict.c (tree_estimate_probability): Likewise.
21643 (tree_estimate_probability_driver): Likewise.
21644 (estimate_loops): Likewise.
21645 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
21646 (move_sese_region_to_fn): Likewise.
21647 (debug_loop_num): Likewise.
21648 * tree-chrec.c (chrec_evaluate): Likewise.
21649 (hide_evolution_in_other_loops_than_loop): Likewise.
21650 (chrec_component_in_loop_num): Likewise.
21651 (reset_evolution_in_loop): Likewise.
21652 (evolution_function_is_invariant_rec_p): Likewise.
21653 * tree-if-conv.c (main_tree_if_conversion): Likewise.
21654 * tree-inline.c (copy_loops): Likewise.
21655 (copy_cfg_body): Likewise.
21656 (tree_function_versioning): Likewise.
21657 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
21658 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
21660 (add_to_evolution_1): Likewise.
21661 (scev_const_prop): Likewise.
21662 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
21663 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
21664 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
21665 (tree_ssa_lim_initialize): Likewise.
21666 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
21667 (verify_loop_closed_ssa): Likewise.
21668 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
21669 (tree_ssa_loop_im): Likewise.
21670 (tree_ssa_loop_unswitch): Likewise.
21671 (tree_vectorize): Likewise.
21672 (check_data_deps): Likewise.
21673 (tree_ssa_loop_ivcanon): Likewise.
21674 (tree_ssa_loop_bounds): Likewise.
21675 (tree_complete_unroll): Likewise.
21676 (tree_complete_unroll_inner): Likewise.
21677 (tree_parallelize_loops): Likewise.
21678 (tree_ssa_loop_prefetch): Likewise.
21679 (tree_ssa_loop_ivopts): Likewise.
21680 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
21681 * tree-vectorizer.c (vectorize_loops): Likewise.
21683 2013-04-29 Mike Frysinger <vapier@gentoo.org>
21685 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
21686 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
21687 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
21688 with EABI_LINK_SPEC.
21690 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
21693 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
21696 2013-04-29 Vladimir Makarov <vmakarov@redhat.com>
21699 * lra-constraints.c (process_alt_operands): Discourage a bit more
21700 using memory for pseudos. Print cost dump for alternatives.
21701 Modify cost values for conflicts with early clobbers.
21702 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
21704 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
21707 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
21709 2013-04-29 Ian Bolton <ian.bolton@arm.com>
21711 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
21712 from/to S register.
21713 (movdi_aarch64): Support LDR/STR from/to D register.
21715 2013-04-29 Ian Bolton <ian.bolton@arm.com>
21717 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
21718 or higher by default.
21720 2013-04-29 Richard Biener <rguenther@suse.de>
21722 PR middle-end/57075
21723 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
21724 even if not adding abnormal edges for calls that can make
21727 2013-04-29 Richard Biener <rguenther@suse.de>
21729 PR middle-end/57103
21730 * tree-cfg.c (move_stmt_op): Fix condition under which to update
21732 (move_stmt_r): Remove redundant checking.
21734 2013-04-29 Teresa Johnson <tejohnson@google.com>
21737 * basic-block.h (apply_scale): New function.
21738 (apply_probability): Use apply_scale.
21739 * gimple-streamer-in.c (input_bb): Ditto.
21740 * lto-streamer-in.c (input_cfg): Ditto.
21741 * lto-cgraph.c (merge_profile_summaries): Ditto.
21742 * tree-optimize.c (execute_fixup_cfg): Ditto.
21743 * tree-inline.c (copy_bb): Update comment to use apply_scale.
21744 (copy_edges_for_bb): Ditto.
21745 (copy_cfg_body): Ditto.
21747 2013-04-29 Tom de Vries <tom@codesourcery.com>
21749 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
21750 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
21751 (tail_merge_optimize): Handle current_loops == NULL.
21753 2013-04-26 Jeff Law <law@redhat.com>
21755 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
21756 (simplify_cond_using_ranges): Generalize code to simplify
21757 COND_EXPRs where one argument is a constant and the other
21758 is an SSA_NAME created by an integral type conversion.
21760 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21762 * config/arm/arm.md (store_minmaxsi): Use only when
21763 optimize_insn_for_size_p.
21765 2013-04-29 Christian Bruel <christian.bruel@st.com>
21768 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
21770 2013-04-29 Richard Biener <rguenther@suse.de>
21772 PR middle-end/57089
21773 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
21774 loop tree make sure to schedule a fixup for the child as well.
21775 (expand_omp_for_generic): Properly add loops.
21776 (expand_omp_for_static_nochunk): Likewise.
21777 (expand_omp_for_static_chunk): Likewise.
21778 (expand_omp_for): For the degenerate case fixup loops.
21779 (expand_omp_sections): Fix default bb placement in loops.
21780 (expand_omp_atomic_pipeline): Properly add loops.
21782 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21784 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
21786 2013-04-29 Tom de Vries <tom@codesourcery.com>
21788 * tree-ssa-tail-merge.c: Update header comment.
21790 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21792 * config/aarch64/arm_neon.h
21793 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
21794 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
21795 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
21796 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
21797 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
21798 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
21799 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
21800 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
21802 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21804 * config/aarch64/aarch64-simd.md
21805 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
21806 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
21807 fix_trunc, fixuns_trunc.
21808 (ftrunc<VDQF:mode>2): New.
21809 * config/aarch64/iterators.md (optab): Add fix, fixuns.
21810 (fix_trunc_optab): New.
21812 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21814 * config/aarch64/aarch64-builtins.c
21815 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
21816 iceilf, lround, iroundf.
21818 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
21821 * config/i386/i386.h (enum ix86_tune_indices)
21822 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
21823 New, split from X86_TUNE_INTER_UNIT_MOVES.
21824 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
21825 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
21826 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
21827 (TARGET_INTER_UNIT_MOVES): Remove.
21828 * config/i386/i386.c (initial_ix86_tune_features): Update.
21829 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
21830 (ix86_expand_convert_uns_didf_sse): Use
21831 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
21832 (ix86_expand_vector_init_one_nonzero): Ditto.
21833 (ix86_expand_vector_init_interleave): Ditto.
21834 (inline_secondary_memory_needed): Return true for moves from SSE class
21835 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
21836 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
21837 * config/i386/constraints.md (Yi, Ym): Depend on
21838 TARGET_INTER_UNIT_MOVES_TO_VEC.
21839 (Yj, Yn): New constraints.
21840 * config/i386/i386.md (*movdi_internal): Change constraints of
21841 operand 1 from Yi to Yj and from Ym to Yn.
21842 (*movsi_internal): Ditto.
21843 (*movdf_internal): Ditto.
21844 (*movsf_internal): Ditto.
21845 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
21846 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
21847 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
21848 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
21849 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
21850 * config/i386/sse.md (movdi_to_sse): Ditto.
21851 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
21852 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
21853 TARGET_INTER_UNIT_MOVES.
21854 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
21855 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
21856 instead of TARGET_INTER_UNIT_MOVES.
21857 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
21858 operand 1 from Yi to Yj and from Ym to Yn.
21860 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21862 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
21863 (float_truncate_hi_): Likewise.
21864 (float_extend_lo_): Likewise.
21865 (float_truncate_lo_): Likewise.
21866 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
21867 (aarch64_float_extend_lo_v2df): Likewise.
21868 (vec_unpacks_hi_v4sf): Likewise.
21869 (aarch64_float_truncate_lo_v2sf): Likewise.
21870 (aarch64_float_truncate_hi_v4sf): Likewise.
21871 (vec_pack_trunc_v2df): Likewise.
21872 (vec_pack_trunc_df): Likewise.
21874 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21876 * config/aarch64/aarch64-builtins.c
21877 (aarch64_fold_builtin): Fold float conversions.
21878 * config/aarch64/aarch64-simd-builtins.def
21879 (floatv2si, floatv4si, floatv2di): New.
21880 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
21881 * config/aarch64/aarch64-simd.md
21882 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
21883 * config/aarch64/iterators.md (FLOATUORS): New.
21884 (optab): Add float, floatuns.
21885 (su_optab): Likewise.
21887 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21889 * config/aarch64/aarch64-builtins.c
21890 (aarch64_builtin_vectorized_function): Use new names for
21892 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
21893 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
21894 (fcvtzu): Split as...
21895 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
21896 (fcvtas): Split as...
21897 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
21898 (fcvtau): Split as...
21899 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
21900 (fcvtps): Split as...
21901 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
21902 (fcvtpu): Split as...
21903 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
21904 (fcvtms): Split as...
21905 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
21906 (fcvtmu): Split as...
21907 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
21908 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
21909 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
21910 (lfrintnusf, lfrintnudf): Likewise.
21911 * config/aarch64/aarch64-simd.md
21912 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
21914 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
21915 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
21916 (fcvt_pattern): Likewise.
21918 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21920 * config/aarch64/aarch64-simd.md
21921 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
21922 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
21924 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21926 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
21927 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
21928 (vrnd<a,m,n,p>_f32): Implement using builtins.
21929 (vrnd<i,x><q>_f<32, 64>): New.
21931 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21933 * config/aarch64/aarch64-builtins.c
21934 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
21935 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
21936 (frintz): Rename to...
21938 (frintp): Rename to...
21940 (frintm): Rename to...
21942 (frinti): Rename to...
21943 (nearbyint): ...this.
21944 (frintx): Rename to...
21946 (frinta): Rename to...
21948 * config/aarch64/aarch64-simd.md
21949 (aarch64_frint<frint_suffix><mode>): Delete.
21950 (<frint_pattern><mode>2): Convert to insn.
21951 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
21952 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
21953 (frint_pattern): Likewise.
21954 (frint_suffix): Likewise.
21956 2013-04-29 Richard Biener <rguenther@suse.de>
21958 PR tree-optimization/57081
21959 * loop-init.c: Include tree-flow.h.
21960 (loop_optimizer_finalize): Free number of iteration estimates.
21961 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
21963 2013-04-29 Jakub Jelinek <jakub@redhat.com>
21965 PR tree-optimization/57083
21966 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
21967 non-singleton shift count range, zero extend low_bound for uns case.
21969 * config/i386/predicates.md (general_vector_operand): New predicate.
21970 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
21971 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
21972 if they aren't nonimmediate operands. If their original values
21973 satisfy const_vector_equal_evenodd_p, don't shift them.
21974 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
21975 predicates. For the SSE4.1 case force operands[{1,2}] into registers
21976 if not nonimmediate_operand.
21977 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
21978 instead of register_operand.
21979 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
21981 2013-04-28 Eric Botcazou <ebotcazou@adacore.com>
21983 * stor-layout.c (finalize_size_functions): Allocate a structure and
21984 reset cfun before dumping the functions.
21986 2013-04-27 Jakub Jelinek <jakub@redhat.com>
21988 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
21991 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
21992 use xop_pmacsdqh if uns_p.
21993 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
21994 the immediate rotate count.
21996 2013-04-26 Vladimir Makarov <vmakarov@redhat.com>
21998 * rtl.h (struct rtx_def): Add comment for field jump.
21999 (LRA_SUBREG_P): New macro.
22000 * recog.c (register_operand): Check LRA_SUBREG_P.
22001 * lra.c (lra): Add note at the end of RTL code. Align non-empty
22003 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
22004 (lra_final_code_change): Skip subreg change for operators.
22005 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
22006 if there are no operand changes.
22007 * lra-constraints.c (curr_insn_set): New.
22008 (match_reload): Set LRA_SUBREG_P.
22009 (emit_spill_move): Ditto.
22010 (check_and_process_move): Use curr_insn_set. Process only single
22011 set insns. Don't initialize sec_mem_p and change_p.
22012 (simplify_operand_subreg): Use LRA_SUBREG_P.
22013 (reg_in_class_p): New function.
22014 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
22015 of #ifdef. Add code to remove cycling.
22016 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
22017 non-null disp. Reload inner instead of disp when base and index
22018 are null. Try to put lo_sum into register.
22019 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
22020 (check_and_process_move): Move code for move cost check to
22021 simple_move_p. Remove equiv_substitution.
22022 (simple_move_p): New function.
22023 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
22024 curr_insn_set. Call check_and_process_move only for single set
22025 insns. Use the new function. Move call of check_and_process_move
22026 after operand equiv substitution and address process.
22028 2013-04-26 Jakub Jelinek <jakub@redhat.com>
22031 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
22032 with nonlocal goto receivers or returns twice calls, ignore
22033 unininitialized values from abnormal edges to nl goto receiver
22034 or returns twice call.
22036 2013-04-26 Jakub Jelinek <jakub@redhat.com>
22038 PR tree-optimization/57051
22039 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
22040 and VEC_RSHIFT_EXPR if shift count is a multiple of element
22043 2013-04-26 Richard Biener <rguenther@suse.de>
22045 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
22046 (expand_omp_taskreg): Likewise. Mark loops for fixup.
22047 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
22048 (fixup_loop_arrays_after_move): New function.
22049 (move_sese_region_to_fn): Properly outline the loop tree parts
22050 of the SESE region.
22052 2013-04-26 Uros Bizjak <ubizjak@gmail.com>
22054 * config/i386/i386.md (type, unit): Fix long lines.
22056 2013-04-26 Richard Biener <rguenther@suse.de>
22058 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
22059 (lto-streamer-out.o): Likewise.
22060 * cfgloop.c (init_loops_structure): Export, add struct function
22061 argument and adjust.
22062 (flow_loops_find): Adjust.
22063 * cfgloop.h (enum loop_estimation): Add EST_LAST.
22064 (init_loops_structure): Declare.
22065 * lto-streamer-in.c: Include cfgloop.h.
22066 (input_cfg): Input the loop tree.
22067 * lto-streamer-out.c: Include cfgloop.h.
22068 (output_cfg): Output the loop tree.
22069 (output_struct_function_base): Do not drop PROP_loops.
22071 2013-03-26 Richard Biener <rguenther@suse.de>
22073 * tree-cfg.c (execute_build_cfg): Build the loop tree.
22074 (pass_build_cfg): Provide PROP_loops.
22075 (move_sese_region_to_fn): Remove loops that are outlined into fn
22077 * tree-inline.c: Include cfgloop.h.
22078 (initialize_cfun): Do not drop PROP_loops.
22079 (copy_loops): New function.
22080 (copy_cfg_body): Copy loop structure.
22081 (tree_function_versioning): Initialize destination loop tree.
22082 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
22083 (pass_parallelize_loops): Do IL verification.
22084 * loop-init.c (loop_optimizer_init): Fixup loops if required.
22085 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
22086 the CFG make sure we fixup loops as well.
22087 * tree-ssa-tail-merge.c: Include cfgloop.h.
22088 (replace_block_by): When merging loop latches mark loops for fixup.
22089 * lto-streamer-out.c (output_struct_function_base): Drop
22090 PROP_loops for now.
22091 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
22092 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
22093 * ipa-split.c: Include cfgloop.h.
22094 (split_function): Add the new return block to the loop tree root.
22095 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
22096 whether we have removed the forwarder block.
22097 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
22098 * cfgloop.h (place_new_loop): Declare.
22099 * cfgloopmanip.c (place_new_loop): Export.
22100 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
22101 (tree-switch-conversion.o): Likewise.
22102 (tree-complex.o): Likewise.
22103 (tree-inline.o): Likewise.
22104 (tree-ssa-tailmerge.o): Likewise.
22105 (ipa-split.o): Likewise.
22106 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
22107 (tree-ssa-copy.o): Likewise.
22108 * tree-switch-conversion.c: Include cfgloop.h
22109 (process_switch): If we emit a bit-test cascade, schedule loops
22111 * tree-complex.c: Include cfgloop.h.
22112 (expand_complex_div_wide): Properly add new basic-blocks to loops.
22113 * asan.c: Include cfgloop.h.
22114 (create_cond_insert_point): Properly add new basic-blocks to
22115 loops, schedule loop fixup.
22116 * cfgloop.c (verify_loop_structure): Check that looks are not
22118 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
22120 (expand_omp_for_generic): Likewise.
22121 (expand_omp_sections): Likewise.
22122 (expand_omp_atomic_pipeline): Schedule loops for fixup.
22123 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
22124 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
22125 is initialized, not when loops are present.
22126 * tree-parloops.c (parallelize_loops): Remove checking here.
22127 * passes.c (init_optimization_passes): Schedule a copy-propagation
22128 pass before complete unrolling of inner loops.
22130 2013-04-26 Jakub Jelinek <jakub@redhat.com>
22132 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
22133 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
22134 (colorize_init): Add argument to _WIN32 version.
22135 * toplev.c: Include diagnostic-color.h.
22136 (process_options): Default to -fdiagnostics-color=auto if
22137 GCC_COLORS env var is in the environment.
22138 * common.opt (fdiagnostics-color=): Add Var and Init.
22139 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
22140 env var is in the environment, the default is auto rather than never.
22142 * diagnostic.h (file_name_as_prefix): Add context argument.
22143 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
22144 the string as locus.
22145 * langhooks.c (lhd_print_error_function): Adjust caller.
22147 2013-04-25 Lawrence Crowl <crowl@google.com>
22149 * var-tracking.c (shared_hash_def::htab):
22150 Change type to hash_table. Update dependent calls and types.
22152 2013-04-25 Lawrence Crowl <crowl@google.com>
22154 * Makefile.in: Update as needed below.
22156 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
22157 Move declaration to after the type's method definitons.
22159 * attribs.c (htab_t scoped_attributes::attribute_hash):
22160 Change type to hash_table. Update dependent calls and types.
22162 * bitmap.c (htab_t bitmap_desc_hash):
22163 Change type to hash_table. Update dependent calls and types.
22165 * cselib.c (htab_t cselib_hash_table):
22166 Change type to hash_table. Update dependent calls and types.
22168 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
22169 (hash_string_slot_node): Move implementation into lto-streamer.h
22170 struct string_slot_hasher.
22171 (eq_string_slot_node): Likewise.
22173 * data-streamer-out.c: Update output_block::string_hash_table
22174 dependent calls and types.
22176 * dwarf2cfi.c (htab_t trace_index):
22177 Change type to hash_table. Update dependent calls and types.
22179 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
22180 Change type to hash_table. Update dependent calls and types.
22181 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
22182 (htab_t optimize_external_refs::map): Likewise.
22183 (htab_t output_comp_unit::extern_map): Likewise.
22184 (htab_t output_comdat_type_unit::extern_map): Likewise.
22185 (htab_t output_macinfo::macinfo_htab): Likewise.
22186 (htab_t optimize_location_lists::htab): Likewise.
22187 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
22189 * except.c (htab_t ehspec_hash_type):
22190 Change type to hash_table. Update dependent calls and types.
22191 (assign_filter_values::ttypes): Likewise.
22192 (assign_filter_values::ehspec): Likewise.
22193 (sjlj_assign_call_site_values::ar_hash): Likewise.
22194 (convert_to_eh_region_ranges::ar_hash): Likewise.
22196 * gcse.c (htab_t pre_ldst_table):
22197 Change type to hash_table. Update dependent calls and types.
22199 * ggc-common.c (htab_t saving_htab):
22200 Change type to hash_table. Update dependent calls and types.
22201 (htab_t loc_hash): Likewise.
22202 (htab_t ptr_hash): Likewise.
22203 (call_count): Rename ggc_call_count.
22204 (call_alloc): Rename ggc_call_alloc.
22205 (loc_descriptor): Rename make_loc_descriptor.
22206 (add_statistics): Rename ggc_add_statistics.
22208 * ggc-common.c (saving_htab):
22209 Change type to hash_table. Update dependent calls and types.
22211 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
22212 (push_gimplify_context): Likewise.
22213 (pop_gimplify_context): Likewise.
22214 (struct gimple_temp_hash_elt): Added.
22215 (struct gimplify_hasher): Likewise.
22216 (struct gimplify_ctx.temp_htab):
22217 Change type to hash_table. Update dependent calls and types.
22219 * gimple-fold.c: Include gimplify-ctx.h.
22221 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
22222 Change type to hash_table. Update dependent calls and types.
22223 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
22224 avoid potential global name collision.
22226 * gimplify.c: Include gimplify-ctx.h.
22227 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
22228 (htab_t gimplify_ctx::temp_htab):
22229 Update dependent calls and types for new type hash_table.
22230 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
22231 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
22233 * gimplify-ctx.h: New.
22234 (struct gimple_temp_hash_elt): Move from gimplify.c.
22235 (class gimplify_hasher): New.
22236 (struct gimplify_ctx): Move from gimple.h.
22237 (htab_t gimplify_ctx::temp_htab):
22238 Change type to hash_table. Update dependent calls and types.
22240 * graphite-clast-to-gimple.c: Include graphite-htab.h.
22241 (htab_t ivs_params::newivs_index):
22242 Change type to hash_table. Update dependent calls and types.
22243 (htab_t ivs_params::params_index): Likewise.
22244 (htab_t print_generated_program::params_index): Likewise.
22245 (htab_t gloog::newivs_index): Likewise.
22246 (htab_t gloog::params_index): Likewise.
22248 * graphite.c: Include graphite-htab.h.
22249 4htab_t graphite_transform_loops::bb_pbb_mapping):
22250 Change type to hash_table. Update dependent calls and types.
22252 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
22253 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
22254 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
22256 * graphite-dependences.c: Include graphite-htab.h.
22257 (loop_is_parallel_p): Change hash table type of parameter.
22259 * graphite-htab.h: New.
22260 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
22261 (extern find_pbb_via_hash): Move from graphite-poly.h.
22262 (extern loop_is_parallel_p): Move from graphite-poly.h.
22263 (extern get_loop_body_pbbs): Move from graphite-poly.h.
22265 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
22266 (extern loop_is_parallel_p): Move to graphite-htab.h.
22267 (extern get_loop_body_pbbs): Move to graphite-htab.h.
22269 * haifa-sched.c (htab_t delay_htab):
22270 Change type to hash_table. Update dependent calls and types.
22271 (htab_t delay_htab_i2): Likewise.
22273 * ira-color.c (htab_t allocno_hard_regs_htab):
22274 Change type to hash_table. Update dependent calls and types.
22276 * ira-costs.c (htab_t cost_classes_htab):
22277 Change type to hash_table. Update dependent calls and types.
22279 * loop-invariant.c (htab_t merge_identical_invariants::eq):
22280 Change type to hash_table. Update dependent calls and types.
22282 * loop-iv.c (htab_t bivs):
22283 Change type to hash_table. Update dependent calls and types.
22285 * loop-unroll.c (htab_t opt_info::insns_to_split):
22286 Change type to hash_table. Update dependent calls and types.
22287 (htab_t opt_info::insns_with_var_to_expand): Likewise.
22289 * lto-streamer.h (struct string_slot): Move from data-streamer.h
22290 (struct string_slot_hasher): New.
22291 (htab_t output_block::string_hash_table):
22292 Change type to hash_table. Update dependent calls and types.
22294 * lto-streamer-in.c (freeing_string_slot_hasher): New.
22295 (htab_t file_name_hash_table):
22296 Change type to hash_table. Update dependent calls and types.
22298 * lto-streamer-out.c: Update output_block::string_hash_table dependent
22301 * lto-streamer.c (htab_t tree_htab):
22302 Change type to hash_table. Update dependent calls and types.
22304 * omp-low.c: Include gimplify-ctx.h.
22306 * passes.c (htab_t name_to_pass_map):
22307 Change type to hash_table. Update dependent calls and types.
22308 (pass_traverse): Rename to passes_pass_traverse.
22310 * plugin.c (htab_t event_tab):
22311 Change type to hash_table. Update dependent calls and types.
22313 * postreload-gcse.c (htab_t expr_table):
22314 Change type to hash_table. Update dependent calls and types.
22315 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
22317 * sese.c (debug_rename_map_1): Make extern.
22318 (htab_t copy_bb_and_scalar_dependences::rename_map):
22319 Change type to hash_table. Update dependent calls and types.
22321 * sese.h (extern debug_rename_map): Move to .c file.
22323 * store-motion.c (htab_t store_motion_mems_table):
22324 Change type to hash_table. Update dependent calls and types.
22326 * trans-mem.c (htab_t tm_new_mem_hash):
22327 Change type to hash_table. Update dependent calls and types.
22329 * tree-browser.c (htab_t TB_up_ht):
22330 Change type to hash_table. Update dependent calls and types.
22332 * tree-cfg.c (htab_t discriminator_per_locus):
22333 Change type to hash_table. Update dependent calls and types.
22335 * tree-complex.c: Include tree-hasher.h
22336 (htab_t complex_variable_components):
22337 Change type to hash_table. Update dependent calls and types.
22339 * tree-eh.c (htab_t finally_tree):
22340 Change type to hash_table. Update dependent calls and types.
22342 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
22343 struct int_tree_hasher.
22344 (extern int_tree_map_eq): Likewise.
22345 (uid_decl_map_hash): Removed.
22346 (extern decl_tree_map_eq): Likewise.
22348 * tree-hasher.h: New.
22349 (struct int_tree_hasher): New.
22350 (typedef int_tree_htab_type): New.
22352 * tree-inline.c: Include gimplify-ctx.h.
22354 * tree-mudflap.c: Include gimplify-ctx.h.
22356 * tree-parloops.c: Include tree-hasher.h.
22357 (htab_t eliminate_local_variables_stmt::decl_address):
22358 Change type to hash_table. Update dependent calls and types.
22359 (htab_t separate_decls_in_region::decl_copies): Likewise.
22361 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
22362 Change type to hash_table. Update dependent calls and types.
22364 * tree-sra.c (candidates):
22365 Change type to hash_table. Update dependent calls and types.
22367 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
22369 (int_tree_map_hash): Likewise.
22371 * tree-ssa-dom.c (htab_t avail_exprs):
22372 Change type to hash_table. Update dependent calls and types.
22374 * tree-ssa-live.c (var_map_base_init::tree_to_index):
22375 Change type to hash_table. Update dependent calls and types.
22377 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
22378 Change type to hash_table. Update dependent calls and types.
22380 * tree-ssa-phiopt.c (seen_ssa_names):
22381 Change type to hash_table. Update dependent calls and types.
22383 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
22384 Change type to hash_table. Update dependent calls and types.
22386 * tree-ssa-uncprop.c (equiv):
22387 Change type to hash_table. Update dependent calls and types.
22389 2013-04-25 Jakub Jelinek <jakub@redhat.com>
22391 PR rtl-optimization/57003
22392 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
22393 call note_stores with kill_clobbered_value callback again after
22394 killing regs_invalidated_by_call.
22396 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
22398 * config/aarch64/aarch64-simd.md
22399 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
22400 (aarch64_simd_bsl<mode>): Likewise.
22401 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
22403 2013-04-25 Marek Polacek <polacek@redhat.com>
22405 PR tree-optimization/57066
22406 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
22408 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
22410 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
22412 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
22414 * config/aarch64/aarch64-builtins.c
22415 (aarch64_fold_builtin): New.
22416 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
22417 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
22418 * config/aarch64/aarch64-simd-builtins.def (abs): New.
22419 * config/aarch64/arm_neon.h
22420 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
22422 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
22423 Tejas Belagod <tejas.belagod@arm.com>
22425 * config/aarch64/aarch64-builtins.c
22426 (aarch64_gimple_fold_builtin): New.
22427 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
22428 * config/aarch64/aarch64-simd-builtins.def (addv): New.
22429 * config/aarch64/aarch64-simd.md (addpv4sf): New.
22430 (addvv4sf): Update.
22431 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
22433 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22435 * config/aarch64/aarch64.md
22436 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
22438 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22440 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
22441 (*ngcsi_uxtw): New pattern.
22443 2013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22444 Julian Brown <julian@codesourcery.com>
22446 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
22447 (TB_DREG): Add T_V4HF.
22448 (v4hf_UP): New macro.
22449 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
22450 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
22451 Handle initialisation of V4HF. Adjust initialisation of reinterpret
22453 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
22454 (arm_vector_mode_supported_p): Handle V4HF.
22455 (arm_mangle_map): Handle V4HFmode.
22456 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
22457 * config/arm/arm_neon_builtins.def: Add entries for
22458 vcvtv4hfv4sf, vcvtv4sfv4hf.
22459 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
22460 (neon_vcvtv4hfv4sf): Likewise.
22461 * config/arm/neon-gen.ml: Handle half-precision floating point
22463 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
22464 * config/arm/arm_neon.h: Regenerate.
22465 * config/arm/neon.ml (type elts): Add F16.
22466 (type vectype): Add T_float16x4, T_floatHF.
22467 (type vecmode): Add V4HF.
22468 (type features): Add Requires_FP_bit feature.
22469 (elt_width): Handle F16.
22470 (elt_class): Likewise.
22471 (elt_of_class_width): Likewise.
22472 (mode_of_elt): Refactor.
22473 (type_for_elt): Handle F16, fix error messages.
22474 (vectype_size): Handle T_float16x4.
22475 (vcvt_sh): New function.
22476 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
22477 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
22478 (string_of_mode): Handle V4HF.
22479 * doc/arm-neon-intrinsics.texi: Regenerate.
22481 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
22483 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
22484 format specifier in 'X' case.
22486 2013-04-25 Alan Modra <amodra@gmail.com>
22489 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
22490 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
22491 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
22492 Repeat for many other rotate/shift and mask patterns using subregs.
22493 Name lshiftrt insns.
22494 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
22495 on WORDS_BIG_ENDIAN.
22497 2013-04-25 Alan Modra <amodra@gmail.com>
22499 * config.gcc: Support little-endian powerpc-linux targets.
22500 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
22501 (LINK_OS_LINUX_SPEC): Define.
22502 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
22503 Preserve MASK_LITTLE_ENDIAN.
22504 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
22505 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
22506 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
22507 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
22508 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
22509 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
22510 Correct fp word order for little-endian. Don't shift toc entries
22511 smaller than a word for little-endian.
22512 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
22513 (bswapdi2 splits): Correct low-part subreg for little-endian.
22514 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
22515 low/high where such is correct only for be.
22516 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
22517 little-endian for -mcall-aixdesc.
22519 2013-04-25 Alan Modra <amodra@gmail.com>
22521 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
22522 replace_equiv_address_nv.
22524 2013-04-25 Alan Modra <amodra@gmail.com>
22526 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
22528 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
22531 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
22532 * rtl.h (struct rtx_def): ...
22534 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
22536 PR rtl-optimizations/57046
22537 * lra-constraints (split_reg): Set up lra_risky_transformations_p
22538 for multi-reg splits.
22540 2013-04-24 H.J. Lu <hongjiu.lu@intel.com>
22542 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
22544 2013-04-24 Sterling Augustine <saugustine@google.com>
22546 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
22547 (comp_dir_string, debug_str_dwo_section): New.
22548 (DEBUG_STR_DWO_SECTION): Rename to ...
22549 (DEBUG_DWO_STR_SECTION): ... this.
22550 (DEBUG_NORM_STR_SECTION): Delete.
22551 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
22552 (DEBUG_STR_DWO_SECTION_FLAGS): New.
22553 (find_AT_string): Move most logic to ...
22554 (find_AT_string_in_table): ... here. New.
22555 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
22556 add_skeleton_AT_string. Delete logic.
22557 (output_skeleton_debug_sections): Remove call to
22558 add_top_level_skeleton_die_attrs.
22559 (add_comp_dir_attribute): Move logic to comp_dir_string.
22560 (dwarf2out_init): Initialize debug_str_dwo_section.
22561 (output_indirect_string): Call find_string_form.
22562 (output_indirect_strings): Rewrite.
22563 (prune_unused_types): Empty skeleton_debug_str_hash.
22564 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
22565 (dwarf2out_finish): Call output_indirect_strings.
22567 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
22569 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
22571 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
22573 * rtl.h (struct rtx_def): Add comment for field jump.
22574 (LRA_SUBREG_P): New macro.
22575 * recog.c (register_operand): Check LRA_SUBREG_P.
22576 * lra.c (lra): Add note at the end of RTL code. Align non-empty
22578 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
22579 (lra_final_code_change): Skip subreg change for operators.
22580 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
22581 if there are no operand changes.
22582 * lra-constraints.c (curr_insn_set): New.
22583 (match_reload): Set LRA_SUBREG_P.
22584 (emit_spill_move): Ditto.
22585 (check_and_process_move): Use curr_insn_set. Process only single
22586 set insns. Don't initialize sec_mem_p and change_p.
22587 (simplify_operand_subreg): Use LRA_SUBREG_P.
22588 (reg_in_class_p): New function.
22589 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
22590 of #ifdef. Add code to remove cycling.
22591 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
22592 non-null disp. Reload inner instead of disp when base and index
22593 are null. Try to put lo_sum into register.
22594 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
22595 (check_and_process_move): Move code for move cost check to
22596 simple_move_p. Remove equiv_substitution.
22597 (simple_move_p): New function.
22598 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
22599 curr_insn_set. Call check_and_process_move only for single set
22600 insns. Use the new function. Move call of check_and_process_move
22601 after operand equiv substitution and address process.
22603 2013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
22605 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
22606 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
22607 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
22609 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
22611 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
22613 2013-04-24 Marek Polacek <polacek@redhat.com>
22615 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
22616 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
22617 (select_loops_exit_conditions): Likewise.
22618 (number_of_iterations_for_all_loops): Likewise.
22619 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
22620 (scev_analysis): Likewise.
22622 2013-04-02 Catherine Moore <clm@codesourcery.com>
22623 Chao-ying Fu <fu@mips.com>
22625 * config/mips/micromips.md (jraddiusp): New pattern.
22626 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
22627 instruction if possible.
22629 2013-04-24 Alan Modra <amodra@gmail.com>
22631 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
22633 2013-04-24 Julian Brown <julian@codesourcery.com>
22634 Chung-Lin Tang <cltang@codesourcery.com>
22636 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
22637 dependency behavior in enumeration type DIE generation. Add TODO note
22638 to comments about future DW_FORM_sdata/udata re-work of related code.
22640 2013-04-23 Lawrence Crowl <crowl@google.com>
22642 * Makefile.in: Update as needed below.
22644 * hash-table.h (class hash_table):
22645 Correct many methods with parameter types compare_type to the correct
22646 value_type. (Correct code was unlikely to notice the change.)
22647 (hash_table::elements_with_deleted) New.
22648 (class hashtable::iterator): New.
22649 (hashtable::begin()): New.
22650 (hashtable::end()): New.
22651 (FOR_EACH_HASH_TABLE_ELEMENT): New.
22653 * statistics.c (statistics_hashes):
22654 Change type to hash_table. Update dependent calls and types.
22656 * tree-into-ssa.c (var_infos):
22657 Change type to hash_table. Update dependent calls and types.
22659 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
22660 Change type to hash_table. Update dependent calls and types.
22662 * tree-ssa-loop-im.c (struct mem_ref.refs):
22663 Change type to hash_table. Update dependent calls and types.
22665 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
22666 Change type to hash_table. Update dependent calls and types.
22668 * tree-ssa-sccvn.c (vn_tables_s::nary):
22669 Change type to hash_table. Update dependent calls and types.
22670 (vn_tables_s::phis): Likewise.
22671 (vn_tables_s::references): Likewise.
22673 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
22674 (vn_reference_eq): Update parameter and return types.
22676 * tree-ssa-structalias.c (pointer_equiv_class_table):
22677 Change type to hash_table. Update dependent calls and types.
22678 (location_equiv_class_table): Likewise.
22680 * tree-vect-data-refs.c: Consequential changes for making
22681 peeling a hash_table.
22683 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
22684 (destroy_loop_vec_info): Dependent hash_table update.
22686 * tree-vectorizer.h (peeling_htab):
22687 Change type to hash_table. Update dependent calls and types.
22689 2013-04-23 Shiva Chen <shiva0217@gmail.com>
22691 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
22692 to check the register content is equal or not.
22693 * lra-constraints.c (match_reload): Use lra_assign_reg_val
22694 to assign register content record.
22695 * lra-eliminations.c (update_reg_eliminate): Use
22696 lra_update_reg_val_offset to update register content offset.
22697 * lra-int.h (struct lra_reg): Add offset member.
22698 (lra_reg_val_equal_p): New static inline function.
22699 (lra_update_reg_val_offset): New static inline function.
22700 (lra_assign_reg_val): New static inline function.
22701 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
22702 to assign register content record.
22703 (initialize_lra_reg_info_element): Initial offset to zero.
22705 2013-04-23 Catherine Moore <clm@codesourcery.com>
22707 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
22708 operands. Record compression.
22710 2013-04-23 Xinliang David Li <davidxl@google.com>
22712 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
22714 2013-04-23 Richard Biener <rguenther@suse.de>
22716 PR middle-end/57036
22717 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
22718 parameter, only add abnormal goto edges from the copied body
22719 if the call could perform abnormal gotos.
22720 (copy_cfg_body): Adjust.
22722 2013-04-23 Sofiane Naci <sofiane.naci@arm.com>
22724 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
22726 2013-04-23 Andreas Schwab <schwab@linux-m68k.org>
22728 * coretypes.h (gimple_stmt_iterator): Add struct to make
22731 2013-04-23 Richard Biener <rguenther@suse.de>
22733 PR tree-optimization/57026
22734 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
22735 from SSA names occuring in abnormal PHI nodes.
22737 2013-04-22 Andi Kleen <ak@linux.intel.com>
22739 * lto/lto.c (print_lto_report_1): Fix LTO report names.
22741 2013-04-22 Andi Kleen <ak@linux.intel.com>
22743 * lto/lto.c (print_lto_report_1): Declare early.
22744 (read_cgraph_and_symbols): Call print_lto_report_1 early.
22746 2013-04-22 Andi Kleen <ak@linux.intel.com>
22748 * common.opt (-flto-report-wpa): Add.
22749 * doc/invoke.texi (-flto-report-wpa): Add.
22750 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
22753 2013-04-22 Xinliang David Li <davidxl@google.com>
22755 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
22756 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
22757 * Makefile.in: New dependency
22759 David Daney <ddaney.cavm@gmail.com>
22761 * configure.ac (gcc_cv_as_micromips_support): Use the
22762 --fatal-warnings option.
22763 * configure: Regenerate.
22765 2013-04-22 Marek Polacek <polacek@redhat.com>
22768 * tsan.c (instrument_expr): Don't instrument expression
22769 in case its size is zero.
22771 2013-04-22 Uros Bizjak <ubizjak@gmail.com>
22775 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
22777 * config/alpha/alpha.c (TARGET_LRA_P): New define.
22779 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
22781 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
22782 (gimple_stmt_iterator): New typedef.
22783 * gimple.h (gimple_stmt_iterator): Rename to...
22784 (gimple_stmt_iterator_d): ... This.
22785 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
22786 trees be valid for GIMPLE and GENERIC.
22787 (TARGET_GIMPLE_FOLD_BUILTIN): New.
22788 * gimple-fold.c (gimple_fold_call): Call target hook
22789 gimple_fold_builtin.
22790 * hooks.c (hook_bool_gsiptr_false): New.
22791 * hooks.h (hook_bool_gsiptr_false): New.
22792 * target.def (fold_stmt): New.
22793 * doc/tm.texi: Regenerate.
22795 2013-04-22 Vladimir Makarov <vmakarov@redhat.com>
22798 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
22799 a set sp if no stack realignment.
22801 2013-04-22 Nick Clifton <nickc@redhat.com>
22803 * config.gcc (tilegx-linux): Extend extra_objs rather than
22805 (tilepro-linux): Likewise.
22807 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
22809 * config/aarch64/aarch64-builtins.c
22811 (CF0, CF1, CF2, CF3, CF4, CF10): New.
22812 (VAR<1-12>): Add MAP parameter.
22813 (BUILTIN_*): Likewise.
22814 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
22815 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
22816 (aarch64_ushl_n<mode>): Likewise.
22817 (aarch64_sshr_n<mode>): Likewise.
22818 (aarch64_ushr_n<mode>): Likewise.
22819 (aarch64_<maxmin><mode>): Likewise.
22820 (aarch64_sqrt<mode>): Likewise.
22821 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
22822 (vshr<q>_n_*): Likewise.
22824 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
22826 * config/aarch64/aarch64-builtins.c
22827 (aarch64_simd_builtin_type_mode): Handle SF types.
22829 (BUILTIN_GPF): Define.
22830 (aarch64_init_simd_builtins): Handle SF types.
22831 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
22832 (frecps): Likewise.
22833 (frecpx): Likewise.
22834 * config/aarch64/aarch64-simd.md
22835 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
22836 (aarch64_frecpe<mode>): New.
22837 (aarch64_frecps<mode>): Likewise.
22838 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
22839 (v8type): Add frecp<esx>.
22840 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
22841 (aarch64_frecps<mode>): Likewise.
22842 * config/aarch64/iterators.md (FRECP): New.
22843 (frecp_suffix): Likewise.
22844 * config/aarch64/arm_neon.h
22845 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
22847 2013-04-22 Christian Bruel <christian.bruel@st.com>
22850 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
22851 (REG_CLASS_NAMES): Idem.
22852 (REG_CLASS_CONTENTS): Idem.
22853 (REGCLASS_HAS_FP_REG): Idem.
22854 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
22855 (sh_conditional_register_usage): Idem.
22857 2013-04-21 Jeff Law <law@redhat.com>
22859 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
22860 (ssa_forward_propagate_and_combine): Use it.
22862 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
22864 * lra.c: Update the flow chart diagram.
22866 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
22868 PR rtl-optimization/56847
22869 * lra-constraints.c (process_alt_operands): Discourage alternative
22870 with non-matche doffsettable memory constraint fro memory with
22873 2013-04-19 Richard Biener <rguenther@suse.de>
22875 PR tree-optimization/56982
22876 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
22878 * gimplify.c (gimplify_call_expr): Notice special calls.
22879 (gimplify_modify_expr): Likewise.
22880 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
22881 abnormal control flow receivers.
22882 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
22883 in the same way as cfun->has_nonlocal_labels.
22884 (gimple_purge_dead_abnormal_call_edges): Likewise.
22885 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
22886 receivers start a basic-block.
22888 2013-04-19 Richard Biener <rguenther@suse.de>
22890 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
22892 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
22893 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
22894 (SLP_TREE_LOAD_PERMUTATION): Add.
22895 (vect_transform_slp_perm_load): Adjust prototype.
22896 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
22897 (vect_free_slp_instance): Likewise.
22898 (vect_create_new_slp_node): Likewise.
22899 (vect_supported_slp_permutation_p): Remove.
22900 (vect_slp_rearrange_stmts): Adjust.
22901 (vect_supported_load_permutation_p): Likewise. Inline
22902 vect_supported_slp_permutation_p here.
22903 (vect_analyze_slp_instance): Compute load permutations per
22904 slp node instead of per instance.
22905 (vect_get_slp_defs): Adjust.
22906 (vect_transform_slp_perm_load): Likewise.
22907 (vect_schedule_slp_instance): Remove redundant code.
22908 (vect_schedule_slp): Remove hack for PR56270, add it ...
22909 * tree-vect-stmts.c (vectorizable_load): ... here, do not
22910 CSE loads for SLP. Adjust.
22912 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
22914 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
22915 spelling in two comments.
22917 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
22920 * config/arm/arm.c (load_multiple_sequence): Require SP
22921 as base register for loads if SP is in the register list.
22923 2013-04-19 Martin Jambor <mjambor@suse.cz>
22925 PR tree-optimization/56718
22926 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
22927 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
22928 and made public. Adjusted all callers.
22929 (ipa_intraprocedural_devirtualization): New function.
22930 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
22931 (ipa_intraprocedural_devirtualization): Likewise.
22932 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
22934 2013-04-19 Richard Biener <rguenther@suse.de>
22936 PR tree-optimization/57000
22937 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
22939 2013-04-19 Terry Guo <terry.guo@arm.com>
22941 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
22943 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
22944 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
22945 (cortex_m4_fmacs): Use new reservations.
22946 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
22948 2013-04-18 Vladimir Makarov <vmakarov@redhat.com>
22950 PR rtl-optimization/56999
22951 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
22953 (lra_coalesce): Remove split_origin_bitmap and related code.
22954 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
22955 ranges if necessary.
22957 2013-04-18 Uros Bizjak <ubizjak@gmail.com>
22959 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
22961 (ix86_expand_call): Remove clobbered_registers array and use
22962 x86_64_ms_sysv_extra_clobbered_registers instead.
22963 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
22965 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
22967 * config/i386/i386.md (*call_rex64_ms_sysv): Use
22968 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
22969 (*call_value_rex64_ms_sysv): Ditto.
22971 2013-04-18 Cary Coutant <ccoutant@google.com>
22973 * dwarf2out.c (output_pubnames): Check die_perennial_p of
22974 parent instead of die_mark.
22976 2013-04-18 Diego Novillo <dnovillo@google.com>
22978 * gimple.c (create_gimple_tmp): New.
22979 (get_expr_type): New.
22980 (build_assign): New.
22981 (build_type_cast): New.
22982 * gimple.h (enum ssa_mode): Define.
22983 (gimple_seq_set_location): New.
22984 * asan.c (build_check_stmt): Change some gimple_build_* calls
22985 to use build_assign and build_type_cast.
22987 2013-04-18 Richard Biener <rguenther@suse.de>
22989 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
22990 handle negative step. Remove redundant checks.
22991 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
22992 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
22993 for negative step and grouped loads fail to vectorize.
22995 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
22997 * emit-rtl.c (reset_insn_used_flags): New function.
22998 (reset_all_used_flags): Use it.
22999 (verify_insn_sharing): New function.
23000 (verify_rtl_sharing): Fix verification for SEQUENCEs.
23002 2013-04-18 Jakub Jelinek <jakub@redhat.com>
23004 PR tree-optimization/56984
23005 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
23006 and (x >> M) >= N don't register any assertion if N << M is the
23009 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
23011 * lower-subreg.c (resolve_simple_move): If called self-recursive,
23012 do not delete_insn insns that have not yet been emitted, only
23013 unlink them with remove_insn.
23014 * df-scan.c (df_insn_delete): Revert r197492.
23016 2013-04-17 Steven Bosscher <steven@gcc.gnu.org>
23018 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
23019 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
23021 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
23023 * config/arm/arm.md (movsicc_insn): Convert define_insn into
23024 define_insn_and_split.
23025 (and_scc,ior_scc,negscc): Likewise.
23026 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
23028 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
23030 * config/arm/arm.c (use_return_insn): Return 0 for targets that
23031 can benefit from using a sequence of LDRD instructions in epilogue
23032 instead of a single LDM instruction.
23034 2013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
23037 * doc/extend.texi: Fix typo.
23039 2013-04-17 Richard Biener <rguenther@suse.de>
23041 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
23042 (vect_build_slp_tree): ... here.
23043 (vect_build_slp_tree_1): Compute which stmts of the SLP group
23044 match. Remove special-casing of mismatched complex loads.
23045 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
23046 re-try the match with swapped commutative operands.
23047 (vect_supported_load_permutation_p): Remove special-casing of
23048 mismatched complex loads.
23049 (vect_analyze_slp_instance): Adjust.
23051 2013-04-17 Richard Biener <rguenther@suse.de>
23053 PR rtl-optimization/56921
23054 * cfgloop.h (struct loop): Add simple_loop_desc member.
23055 (struct niter_desc): Mark with GTY(()).
23056 (simple_loop_desc): Do not use aux field but simple_loop_desc.
23057 * loop-iv.c (get_simple_loop_desc): Likewise.
23058 (free_simple_loop_desc): Likewise.
23061 2013-04-16 Richard Biener <rguenther@suse.de>
23063 PR rtl-optimization/56921
23064 * loop-init.c (pass_rtl_move_loop_invariants): Add
23065 TODO_do_not_ggc_collect to todo_flags_finish.
23066 (pass_rtl_unswitch): Same.
23067 (pass_rtl_unroll_and_peel_loops): Same.
23068 (pass_rtl_doloop): Same.
23070 2013-04-17 Eric Botcazou <ebotcazou@adacore.com>
23072 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
23073 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
23074 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
23076 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
23077 * tree-streamer.c (record_common_node): Adjust reference in comment.
23079 2013-04-17 Terry Guo <terry.guo@arm.com>
23081 * config/arm/cortex-m4.md: Add a new bypass.
23083 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
23085 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
23087 (*subs_<optab><mode>_multp2): New pattern.
23088 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
23089 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
23091 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
23093 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
23094 (*subs_mul_imm_<mode>): New pattern.
23096 2013-04-16 David Edelsohn <dje.gcc@gmail.com>
23099 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
23100 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
23101 (vsx_movti_32bit): Change j->wa to O->wa.
23103 2013-04-16 Richard Biener <rguenther@suse.de>
23105 PR rtl-optimization/56921
23106 * loop-init.c (pass_rtl_move_loop_invariants): Add
23107 TODO_do_not_ggc_collect to todo_flags_finish.
23108 (pass_rtl_unswitch): Same.
23109 (pass_rtl_unroll_and_peel_loops): Same.
23110 (pass_rtl_doloop): Same.
23112 2013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
23114 * config/arm/arm.c (emit_multi_reg_push): New declaration
23115 for an existing function.
23116 (arm_emit_strd_push): New function.
23117 (arm_expand_prologue): Used here.
23118 (arm_emit_ldrd_pop): New function.
23119 (arm_expand_epilogue): Used here.
23120 (arm_get_frame_offsets): Update condition.
23121 (arm_emit_multi_reg_pop): Add a special case for load of a single
23122 register with writeback.
23124 2013-04-16 Uros Bizjak <ubizjak@gmail.com>
23126 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
23129 2013-04-16 Richard Biener <rguenther@suse.de>
23131 PR tree-optimization/56756
23132 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
23133 (first_mem_ref_loc): New.
23134 (execute_sm): Place the load temporarily before a previous
23135 access instead of in the latch edge to ensure its SSA dependencies
23136 are defined at points dominating the load.
23138 2013-04-16 Steven Bosscher <steven@gcc.gnu.org>
23140 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
23141 correct fix by moving header and footer insn to the footer of
23142 the merged basic block. Clear BB_END of the merged-away block.
23144 PR middle-end/43631
23145 * emit-rtl.c (make_note_raw): New function.
23146 (link_insn_into_chain): New static inline function.
23147 (add_insn): Use it.
23148 (add_insn_before, add_insn_after): Factor insn chain linking code...
23149 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
23150 using link_insn_into_chain.
23151 (note_outside_basic_block_p): New helper function for emit_note_after
23152 and emit_note_before.
23153 (emit_note_after): Use nobb variant of add_insn_after if the note
23154 should not be contained in a basic block.
23155 (emit_note_before): Use nobb variant of add_insn_before if the note
23156 should not be contained in a basic block.
23157 (emit_note_copy): Use make_note_raw.
23158 (emit_note): Likewise.
23159 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
23160 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
23161 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
23162 the moved barrier the tail of the basic block it follows.
23163 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
23165 2013-04-15 Jakub Jelinek <jakub@redhat.com>
23167 PR tree-optimization/56962
23168 * gimple-ssa-strength-reduction.c (record_increment): Only set
23169 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
23170 either rhs1 or rhs2 is equal to c->base_expr.
23172 2013-04-15 Richard Biener <rguenther@suse.de>
23174 PR tree-optimization/56933
23175 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
23177 (GROUP_READ_WRITE_DEPENDENCE): Remove.
23178 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
23179 * tree-vect-data-refs.c (vect_analyze_group_access): Move
23180 dependence check ...
23181 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
23183 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
23184 GROUP_READ_WRITE_DEPENDENCE.
23186 2013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
23188 * emit-rtl.c (reset_all_used_flags): New function.
23189 (verify_rtl_sharing): Call reset_all_used_flags before and after
23190 performing the checks.
23192 2013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23194 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
23195 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
23196 * config/arm/constraints.md (De): New constraint.
23197 * config/arm/neon.md (anddi3_neon): Delete.
23198 (neon_vand<mode>): Expand to standard anddi3 pattern.
23199 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
23200 Move earlier in the file.
23201 (neon_inv_logic_op2): Likewise.
23202 (arm_anddi_operand_neon): New predicate.
23204 2013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
23206 * configure.ac (gcc_cv_ld_as_needed): Set
23207 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
23208 Use -z ignore, -z record on *-*-solaris2*.
23209 (HAVE_LD_AS_NEEDED): Update comment.
23210 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
23211 * configure: Regenerate.
23212 * config.in: Regenerate.
23213 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
23214 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
23215 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
23216 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
23217 equivalents. Fix markup.
23218 * doc/tm.texi: Regenerate.
23220 2013-04-15 Andrew Hsieh <andrewhsieh.google.com>
23222 * config/i386/i386.opt: New option mstack-protector-guard=.
23223 * config/i386/i386-opts.h: Add enum stack_protector_guard.
23224 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
23225 TARGET_SSP_TLS_GUARD.
23226 * config/i386/i386.c (ix86_option_override_internal): Set
23227 ix86_stack_protector_guard.
23228 * config/i386/i386.md (stack_protect_set): Enable for
23229 TARGET_SSP_TLS_GUARD only.
23230 (stack_protect_set_<mode>): Ditto.
23231 (stack_protect_test): Ditto.
23232 (stack_protect_test_<mode>): Ditto.
23233 * doc/invoke.texi (i386 Option): Document.
23235 2013-04-15 Eric Botcazou <ebotcazou@adacore.com>
23238 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
23239 (S_MODES): Set H_MODE bit.
23240 (SF_MODES): Set only S_MODE and SF_MODE bits.
23241 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
23242 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
23243 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
23244 <MODE_FLOAT>: Likewise.
23246 2013-04-15 Joey Ye <joey.ye@arm.com>
23248 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
23250 2013-04-15 Joey Ye <joey.ye@arm.com>
23252 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
23254 (thumb_far_jump_used_p): Count instruction size and set
23257 2013-04-14 Eric Botcazou <ebotcazou@adacore.com>
23259 * reorg.c (fill_simple_delay_slots): Reindent block of code.
23260 * resource.c (mark_target_live_regs): Reformat conditional block.
23262 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
23264 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
23265 notes, they are emitted only just before final.
23266 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
23268 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
23270 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
23271 * cfgrtl.c (delete_insn): Call it here instead.
23272 * lra-spills.c (lra_final_code_change): Use delete_insn.
23273 * haifa-sched.c (sched_remove_insn): Likewise.
23274 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
23275 returning to the nop pool.
23276 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
23277 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
23279 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
23281 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
23282 * doc/tm.texi: Regenerated.
23284 2013-04-12 Uros Bizjak <ubizjak@gmail.com>
23286 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
23289 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
23291 * df-core.c (df_find_def): Compare register numbers.
23292 (df_find_use): Likewise.
23294 2013-04-12 Vladimir Makarov <vmakarov@redhat.com>
23297 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
23298 lra_in_progress for return.
23300 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
23302 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
23303 define_insn into define_insn_and_split and emit movsicc patterns.
23305 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
23307 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
23309 2013-04-12 Richard Biener <rguenther@suse.de>
23311 * tree-pass.h (TODO_do_not_ggc_collect): New.
23312 * passes.c (execute_one_ipa_transform_pass): Honor
23313 TODO_do_not_ggc_collect.
23314 (execute_one_pass): Likewise.
23317 2013-04-10 Richard Biener <rguenther@suse.de>
23319 * passes.c (init_optimization_passes): Remove reload pass.
23320 * ira.c (do_reload): Merge into ...
23322 (rest_of_handle_reload): Remove.
23323 (pass_reload): Likewise.
23324 * config/i386/i386.c (ix86_option_override): Refer to ira instead
23325 of reload for vzeroupper pass placement.
23327 2013-04-12 Jakub Jelinek <jakub@redhat.com>
23329 PR tree-optimization/56918
23330 PR tree-optimization/56920
23331 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
23332 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
23333 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
23334 use wide_mul_with_sign method.
23336 2013-04-12 Richard Biener <rguenther@suse.de>
23338 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
23339 not be considered a gimple constant.
23341 2013-04-12 Marc Glisse <marc.glisse@inria.fr>
23343 * fold-const.c (const_binop): Handle vector shifts by a scalar.
23344 (fold_binary_loc): Call const_binop also for mixed vector-scalar
23347 2013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
23348 Jakub Jelinek <jakub@redhat.com>
23350 * opts.c: Include diagnostic-color.h.
23351 (common_handle_option): Handle OPT_fdiagnostics_color_.
23352 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
23353 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
23354 (diagnostic-color.o): New.
23355 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
23356 (diagnostic_color_rule): New enum.
23357 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
23358 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
23359 the location string.
23360 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
23361 either NULL, or color kind.
23362 * diagnostic-color.c: New file.
23363 * diagnostic-color.h: New file.
23364 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
23366 * doc/invoke.texi (-fdiagnostics-color): Document.
23367 * pretty-print.h (pp_show_color): Define.
23368 (struct pretty_print_info): Add show_color field.
23369 * diagnostic.c: Include diagnostic-color.h.
23370 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
23371 macros. Colorize error:, warning: etc. strings and also the location
23373 (diagnostic_show_locus): Colorize the caret line.
23374 * pretty-print.c: Include diagnostic-color.h.
23375 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
23376 inside of %< %> quotes or quoted through q format modifier.
23378 2013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
23380 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
23382 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
23384 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
23385 code in CC_NZ mode.
23386 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
23389 2013-04-11 Marek Polacek <polacek@redhat.com>
23391 PR tree-optimization/48184
23392 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
23394 2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
23396 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
23397 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
23398 (skip_simple_arithmetic): Tidy up.
23399 * tree.h (skip_simple_constant_arithmetic): Declare.
23401 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
23403 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
23405 2013-04-11 Richard Biener <rguenther@suse.de>
23407 * tree-vect-loop.c (get_initial_def_for_induction): Properly
23408 generate vector constants.
23410 2013-04-11 Richard Biener <rguenther@suse.de>
23412 PR tree-optimization/56878
23413 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
23414 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
23416 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
23417 Prefer to align the DR with the most invariant base address.
23419 2013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
23421 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
23424 2013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
23426 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
23427 floating-point vector comparisons against 0.
23429 2013-04-11 Jakub Jelinek <jakub@redhat.com>
23431 PR tree-optimization/56899
23432 * fold-const.c (extract_muldiv_1): Apply distributive law
23433 only if TYPE_OVERFLOW_WRAPS (ctype).
23435 2013-04-11 Bin Cheng <bin.cheng@arm.com>
23438 * ira-costs.c (scan_one_insn): Check whether the source rtx of
23439 loading has side effect.
23441 2013-04-10 Steven Bosscher <steven@gcc.gnu.org>
23443 * config/sparc/sparc.c: Include tree-pass.h.
23444 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
23445 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
23446 head of file. Change return type. Split off gate function.
23447 (sparc_gate_work_around_errata): New function.
23448 (pass_work_around_errata): New pass definition.
23449 (insert_pass_work_around_errata) New pass insert definition to
23450 insert pass_work_around_errata just after delayed-branch scheduling.
23451 (sparc_option_override): Insert the pass.
23452 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
23454 2013-04-10 David S. Miller <davem@davemloft.net>
23456 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
23457 or -mcpu=hypersparc.
23459 * target.def (cstore_mode): New hook.
23460 * target.h: Include insn-codes.h
23461 * targhooks.c: Likewise.
23462 (default_cstore_mode): New function.
23463 * targhooks.h: Declare it.
23464 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
23465 * doc/tm.texi: Rebuild.
23466 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
23467 target hook, rather than inspecting the insn_data.
23468 * config/sparc/sparc.c (sparc_cstore_mode): New function.
23469 (TARGET_CSTORE_MODE): Redefine.
23470 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
23472 * config/sparc/predicates.md (cstore_result_operand): New special
23474 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
23475 Use it for operand 0.
23476 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
23477 (*snesi_special): Likewise.
23478 (*snesi_zero): Likewise.
23479 (*seqsi_zero): Likewise.
23480 (*sltu_insn): Likewise.
23481 (*sgeu_insn): Likewise.
23482 (*seqdi_special): Make operand 0 and comparison operation be of
23484 (*snedi_special): Likewise.
23485 (*snedi_special_vis3): Likewise.
23486 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
23487 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
23488 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
23489 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
23490 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
23491 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
23492 (*sltu_extend_sp64): Likewise.
23493 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
23494 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
23495 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
23496 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
23497 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
23499 2013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
23501 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
23502 (aarch64_start_file): Use the new function.
23504 2013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
23505 Jason Merrill <jason@redhat.com>
23507 * common.opt: Add -gdwarf.
23508 * opts.c (common_handle_option): Handle it.
23509 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
23511 2013-04-10 Richard Biener <rguenther@suse.de>
23513 * passes.c (execute_todo): Do not call ggc_collect conditional here.
23514 (execute_one_ipa_transform_pass): But unconditionally here.
23515 (execute_one_pass): And here.
23516 (init_optimization_passes): Remove reload pass.
23517 * tree-pass.h (TODO_ggc_collect): Remove.
23518 (pass_reload): Likewise.
23519 * ira.c (do_reload): Merge into ...
23521 (rest_of_handle_reload): Remove.
23522 (pass_reload): Likewise.
23523 * config/i386/i386.c (ix86_option_override): Refer to ira instead
23524 of reload for vzeroupper pass placement.
23525 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
23526 and todo_flags_finish of all passes.
23528 2013-04-10 Richard Biener <rguenther@suse.de>
23530 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
23531 first_const_oprnd field, rename first_def_type to first_op_type.
23532 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
23533 (vect_get_and_check_slp_defs): Always use the type of the
23534 operand. Allow mixed vect_external_def, vect_constant_def types.
23535 (vect_get_constant_vectors): Handle mixed vect_external_def,
23536 vect_constant_def types.
23538 2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
23540 PR tree-optimization/55524
23541 * tree-ssa-math-opts.c
23542 (convert_mult_to_fma): Don't use an fms construct
23543 when we don't have an fms operation, but fnma, and it looks
23544 likely that we'll be able to use the latter.
23546 2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
23548 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
23550 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
23551 inline fail caused by overwritable functions.
23553 2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
23555 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
23556 unnecessary bits in the constant power of two case.
23558 2013-04-10 Richard Biener <rguenther@suse.de>
23560 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
23561 broken code swapping operands.
23562 (vect_build_slp_tree): Do not compute load permutations here.
23563 (vect_analyze_slp_instance): Compute load permutations here,
23564 after building the SLP tree.
23566 2013-04-09 Christian Bruel <christian.bruel@st.com>
23568 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
23569 of next/prev_real_insn.
23571 2013-04-09 Jan Hubicka <jh@suse.cz>
23573 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
23574 Drop aliased parameter.
23575 (function_and_variable_visibility): Do not handle alias pairs.
23576 * cgraph.c (varpool_externally_visible_p): Update prototype.
23577 * varpool.c (varpool_add_new_variable): Update.
23579 2013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23581 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
23583 2013-04-09 Steven Bosscher <steven@gcc.gnu.org>
23585 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
23587 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
23589 2013-04-09 Marek Polacek <polacek@redhat.com>
23591 PR tree-optimization/48762
23592 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
23594 2013-04-09 Richard Biener <rguenther@suse.de>
23596 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
23598 (vect_build_slp_tree): Likewise.
23599 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
23600 calculating the cost of a SLP instance.
23601 (vect_analyze_slp_instance): Use it from here, after building
23604 2013-04-09 Jakub Jelinek <jakub@redhat.com>
23606 PR middle-end/56883
23607 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
23608 expand_omp_for_static_chunk): Use simple_p = true in
23609 force_gimple_operand_gsi calls when assigning to addressable decls.
23611 2013-04-09 Jeff Law <law@redhat.com>
23613 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
23614 when the boolean was created by converting a wider object which
23615 had a boolean range.
23617 2013-04-09 Richard Biener <rguenther@suse.de>
23619 * tree-vectorizer.h (slp_void_p): Remove.
23620 (slp_tree): Typedef before _slp_tree declaration.
23621 (struct _slp_tree): Use a vector of slp_tree as children.
23622 (vect_get_place_in_interleaving_chain): Remove.
23623 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
23625 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
23627 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
23628 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
23629 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
23630 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
23631 Use slp_node instead of slp_void_p and adjust.
23633 2013-04-09 Richard Biener <rguenther@suse.de>
23635 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
23636 work that is not necessary.
23638 2013-04-09 Jakub Jelinek <jakub@redhat.com>
23640 PR tree-optimization/56854
23641 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
23642 forward into clobber stmts if it would change MEM_REF lhs into
23645 2013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
23647 * tree.c (type_hash_lookup, type_hash_add): Make static.
23648 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
23650 2013-04-09 Richard Biener <rguenther@suse.de>
23652 * tree.h (unsave_expr_now): Remove.
23653 * tree-inline.c (mark_local_for_remap_r): Remove.
23654 (unsave_expr_1): Likewise.
23655 (unsave_r): Likewise.
23656 (unsave_expr_now): Likewise.
23657 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
23658 (propagate_tree_value): Likewise.
23660 2013-04-08 Steven Bosscher <steven@gcc.gnu.org>
23662 * doc/rtl.texi (sequence): Rewrite documentation to match the
23663 current use of SEQUENCE rtl objects.
23664 * rtl.def (SEQUENCE): Likewise.
23666 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
23667 Update documentation.
23668 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
23669 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
23671 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
23673 2013-04-08 Teresa Johnson <tejohnson@google.com>
23675 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
23676 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
23678 (estimate_edge_size_and_time): Add comment to suggest using rounding
23680 (estimate_node_size_and_time): Ditto.
23681 (remap_edge_change_prob): Use helper rounding divide methods.
23682 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
23683 (gimple_mod_pow2_value_transform): Ditto.
23684 (gimple_mod_subtract_transform): Ditto.
23685 (gimple_ic_transform): Ditto.
23686 (gimple_stringops_transform): Ditto.
23687 * stmt.c (conditional_probability): Ditto.
23688 (emit_case_dispatch_table): Ditto.
23689 * lto-cgraph.c (merge_profile_summaries): Ditto.
23690 * tree-optimize.c (execute_fixup_cfg): Ditto.
23691 * cfgcleanup.c (try_forward_edges): Ditto.
23692 * cfgloopmanip.c (scale_loop_profile): Ditto.
23694 (duplicate_loop_to_header_edge): Ditto.
23695 (lv_adjust_loop_entry_edge): Ditto.
23696 * tree-vect-loop.c (vect_transform_loop): Ditto.
23697 * profile.c (compute_branch_probabilities): Ditto.
23698 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
23699 * lto-streamer-in.c (input_cfg): Ditto.
23700 * gimple-streamer-in.c (input_bb): Ditto.
23701 * ipa-cp.c (update_profiling_info): Ditto.
23702 (update_specialized_profile): Ditto.
23703 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
23704 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
23706 * sched-rgn.c (compute_dom_prob_ps): Ditto.
23707 (compute_trg_info): Ditto.
23708 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
23709 (purge_dead_edges): Ditto.
23710 * loop-unswitch.c (unswitch_loop): Ditto.
23711 * cgraphclones.c (cgraph_clone_edge): Ditto.
23712 (cgraph_clone_node): Ditto.
23713 * tree-inline.c (copy_bb): Ditto.
23714 (copy_edges_for_bb): Ditto.
23715 (initialize_cfun): Ditto.
23716 (copy_cfg_body): Ditto.
23717 (expand_call_inline): Ditto.
23719 2013-04-08 Kai Tietz <ktietz@redhat.com>
23721 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
23722 TARGET_CYGWIN64 by TARGET_64BIT.
23724 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
23726 * config/epiphany/epiphany.md (GPR_1): New constant.
23727 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
23728 * config/epiphany/epiphany.c (gen_compare_reg):
23729 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
23730 is already in place.
23731 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
23732 Don't require being called during rtl expansion; If y operlaps r0,
23734 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
23735 (epiphany_expand_epilogue): Likewise.
23737 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
23738 Don't use CC_FPmode for ORDERED / UNORDERED.
23739 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
23741 * config/epiphany/constraints.md (CnL): New constraint.
23742 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
23743 * config/epiphany/predicates.md (add_operand): Allow 1024.
23745 * config/epiphany/epiphany.md (logical_op): New code iterator.
23746 (op_mnc): New code attribute.
23747 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
23748 (mov_f+1, mov_f+2): New peephole2 patterns.
23750 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
23751 (cstoresi4): Also allow re-use of zero result when doing a NE
23752 comparison to a non-zero operand.
23753 Use (clobber (scratch)) for first insn if the gpr output is not needed.
23755 * config/epiphany/epiphany.md (<insn_opname>v2si3):
23756 Use gen_addsi3_i / gen_subsi3_i.
23758 2013-04-08 Jakub Jelinek <jakub@redhat.com>
23762 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
23763 contain anything but clobbers, at most one __builtin_stack_restore,
23764 optionally debug stmts and final resx, and if it has at least one
23765 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
23766 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
23767 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
23768 which isn't defaut definition, remove them.
23769 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
23770 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
23771 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
23772 with MEM_REF LHS with SSA_NAME address.
23774 2013-04-08 Jeff Law <law@redhat.com>
23776 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
23778 2013-04-08 Richard Biener <rguenther@suse.de>
23780 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
23782 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
23783 determined vector type.
23784 (vect_analyze_data_refs): Likewise.
23785 (vect_get_new_vect_var): Adjust.
23786 (vect_create_destination_var): Preserve SSA name versions.
23787 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
23788 not dump anything here.
23790 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
23792 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
23793 Add member lr_slot_known.
23794 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
23796 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
23797 Remove code that sets lr_slot_offset according to what a previous
23798 version of epiphany_emit_save_restore used to do.
23799 (epiphany_emit_save_restore): When doing an lr save or restore,
23800 set/verify lr_slot_known and lr_slot_offset.
23802 2013-04-08 Xinyu Qi <xyqi@marvell.com>
23805 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
23808 2013-04-08 Richard Biener <rguenther@suse.de>
23810 * alias.c (find_base_term): Fix thinko in previous change.
23812 2013-04-08 Jakub Jelinek <jakub@redhat.com>
23814 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
23815 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
23816 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
23817 if possible to compute val.
23818 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
23819 For QImode integers don't require anything about precision. Use
23820 const_with_all_bytes_same to find out if the constant doesn't have
23821 repeated bytes in it.
23823 2013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
23825 * config/s390/s390.c (s390_expand_insv): Only accept insertions
23828 2013-04-08 Marek Polacek <polacek@redhat.com>
23830 PR rtl-optimization/48182
23831 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
23834 2013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
23837 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
23838 nuses, make sure we have a label.
23840 2013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
23843 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
23844 (rs6000_emit_swdiv_low_precision): Remove.
23845 (rs6000_emit_swdiv): Rewrite to handle between one and four
23846 iterations of Newton-Raphson generally; modify required number of
23847 iterations for some cases.
23848 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
23850 2013-04-05 Steven Bosscher <steven@gcc.gnu.org>
23852 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
23853 set-but-unused variable.
23855 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
23856 basic blocks of released function bodies garbage-collectable.
23858 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
23859 (struct rtl_opt_pass): Add TODO_df_finish.
23861 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
23863 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23865 * config/arm/constraints.md (q): New constraint.
23866 * config/arm/ldrdstrd.md: New file.
23867 * config/arm/arm.md (ldrdstrd.md) New include.
23868 (arm_movdi): Use "q" instead of "r" constraint
23869 for double-word memory access.
23870 (movdf_soft_insn): Likewise.
23871 * config/arm/vfp.md (movdi_vfp): Likewise.
23872 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
23873 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
23874 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
23875 (mem_ok_for_ldrd_strd): Likewise.
23876 (output_move_double): Update assertion.
23878 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23880 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
23882 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23884 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
23885 define_insn_and_split.
23886 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
23888 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23890 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
23891 define_insn_and_split.
23892 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
23893 (shiftsi3_compare): New pattern.
23894 (rrx): New pattern.
23895 * config/arm/unspecs.md (UNSPEC_RRX): New.
23897 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23899 * config/arm/arm.md (negdi_extendsidi): New pattern.
23900 (negdi_zero_extendsidi): Likewise.
23902 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23904 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
23905 define_insn_and_split.
23906 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
23907 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
23909 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23911 * config/arm/arm.md (arm_subdi3): Convert define_insn into
23912 define_insn_and_split.
23913 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
23914 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
23916 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23918 * config/arm/arm.md (subsi3_carryin): New pattern.
23919 (subsi3_carryin_const): Likewise.
23920 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
23921 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
23923 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23925 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
23927 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23929 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
23930 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
23932 2013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23934 * config/arm/arm.c (arm_expand_builtin): Change fcode
23935 type to unsigned int.
23937 2013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
23939 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
23941 2013-04-04 Ian Lance Taylor <iant@google.com>
23943 * doc/standards.texi (Standards): The Go frontend supports the Go 1
23946 2013-04-04 Steven Bosscher <steven@gcc.gnu.org>
23948 PR middle-end/56729
23949 * df-scan.c (df_insn_delete): Disable failing assert.
23951 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23953 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
23954 New function prototype.
23955 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
23956 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
23957 (arm_builtin_vectorized_function): New function.
23959 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23961 * config/arm/arm_neon_builtins.def: New file.
23962 * config/arm/arm.c (neon_builtin_data): Move contents to
23963 arm_neon_builtins.def.
23964 (enum arm_builtins): Include neon builtin definitions.
23965 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
23966 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
23968 2013-04-04 Marek Polacek <polacek@redhat.com>
23970 PR tree-optimization/48186
23971 * predict.c (maybe_hot_frequency_p): Return false if
23972 HOT_BB_FREQUENCY_FRACTION is 0.
23973 (cgraph_maybe_hot_edge_p): Likewise.
23975 2013-04-04 Richard Biener <rguenther@suse.de>
23977 PR tree-optimization/56826
23978 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
23981 2013-04-04 Richard Biener <rguenther@suse.de>
23983 PR tree-optimization/56213
23984 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
23985 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
23987 2013-04-04 Richard Biener <rguenther@suse.de>
23989 PR tree-optimization/56837
23990 * tree-loop-distribution.c (classify_partition): For non-zero
23991 values require that the value has the same precision as its
23992 mode to be useful as memset value.
23994 2013-04-03 Nick Clifton <nickc@redhat.com>
23996 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
23997 (fmssf4): Use fmsf.s on E3V5 architectures.
23998 (fnmasf4): Use fnmaf.s on E3V5 architectures.
23999 (fnmssf4): Use fnmsf.s on E3V5 architectures.
24001 2013-04-03 Jeff Law <law@redhat.com>
24003 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
24004 (lra-eliminations.o): Likewise.
24006 2013-04-03 Teresa Johnson <tejohnson@google.com>
24008 * gcov-io.c (compute_working_sets): Moved most of body of old
24009 compute_working_sets here from profile.c.
24010 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
24011 (gcov_working_set_t): Moved typedef here from basic-block.h
24012 (compute_working_set): Declare.
24013 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
24014 (get_working_sets): Renamed from compute_working_set,
24015 replace most of body with call to new compute_working_sets.
24016 (get_exec_counts): Replace call to compute_working_sets
24017 to get_working_sets.
24018 * profile.h (get_working_sets): Renamed from compute_working_set.
24019 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
24020 to get_working_sets.
24021 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
24022 * gcov-dump.c (dump_working_sets): New function.
24024 2013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
24026 * hwint.c (sext_hwi, zext_hwi): New functions.
24027 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
24028 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
24029 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
24030 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
24031 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
24032 (sext_hwi, zext_hwi): New functions.
24034 2013-04-03 Jeff Law <law@redhat.com>
24036 PR tree-optimization/56799
24037 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
24038 back test for widening conversion erroneously dropped in prior change.
24040 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24043 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
24044 instead of next_real_insn.
24046 2013-04-03 Marek Polacek <polacek@redhat.com>
24049 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
24051 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24054 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
24056 (thumb1_output_casesi): Likewise.
24057 (thumb2_output_casesi): Likewise.
24059 2013-04-03 Richard Biener <rguenther@suse.de>
24061 PR tree-optimization/56817
24062 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
24064 (tree_unroll_loops_completely_1): ... new function to manually
24065 walk the loop tree, properly defering outer loops of unrolled
24066 loops to later iterations.
24068 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
24070 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
24071 (vectorizable_load): Likewise.
24072 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
24073 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
24075 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
24077 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
24080 2013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
24082 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
24084 2013-04-03 Bin Cheng <bin.cheng@arm.com>
24086 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
24088 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
24090 PR tree-optimization/56790
24091 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
24094 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
24096 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
24098 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
24099 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
24102 2013-04-03 Jakub Jelinek <jakub@redhat.com>
24105 * tree.h (force_folding_builtin_constant_p): New decl.
24106 * builtins.c (force_folding_builtin_constant_p): New variable.
24107 (fold_builtin_constant_p): Fold immediately also if
24108 force_folding_builtin_constant_p.
24110 2013-04-03 Richard Biener <rguenther@suse.de>
24112 PR tree-optimization/56812
24113 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
24114 DRs of the same interleaving chain are independent.
24116 2013-04-02 Jason Merrill <jason@redhat.com>
24118 * gdbinit.in (pbb): Use debug fn.
24120 2013-04-02 Lawrence Crowl <crowl@google.com>
24122 * sese.h (struct ivtype_map_elt_s): Remove unused.
24123 (extern debug_ivtype_map): Remove unused.
24124 (extern eq_ivtype_map_elts): Remove unused.
24125 * sese.c (debug_ivtype_map): Removed unused.
24126 (debug_ivtype_map_1): Removed unused.
24127 (debug_ivtype_elt): Remove unused.
24128 (eq_ivtype_map_elts): Remove unused.
24130 2013-04-02 Kai Tietz <ktietz@redhat.com>
24133 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
24134 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
24135 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
24137 (legitimize_pe_coff_symbol): Likewise.
24138 (is_imported_p): New helper-function.
24139 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
24140 for Windows x64 targets.
24141 (ix86_expand_prologue): Optimize for pe-coff targets.
24142 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
24143 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
24144 medium/large code-model.
24145 (legitimize_pic_address): Likewise.
24146 (legitimize_tls_address): Likewise.
24147 (ix86_expand_call): Likewise.
24148 (x86_output_mi_thunk): Likewise.
24149 (get_dllimport_decl): Add new beimport argument.
24150 (construct_plt_address): Don't assert for x64 pe-coff targets.
24151 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
24153 (SYMBOL_FLAG_STUBVAR): New macro.
24154 (SYMBOL_REF_STUBVAR_P): Likewise.
24155 * config/i386/winnt.c (stub_list): New structure.
24156 (stub_head): New local variable.
24157 (i386_pe_record_stub): New function.
24158 (i386_pe_file_end): Emit refptr-stubs.
24160 2013-04-02 Jakub Jelinek <jakub@redhat.com>
24162 PR rtl-optimization/56745
24163 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
24164 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
24167 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
24168 and both of them are MEM_REFs, just compare first argument for
24169 equality and attempt to deal even with differing offsets.
24172 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
24173 of gimple_clobber_p to be MEM_REF.
24174 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
24175 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
24176 after gimplification.
24177 * asan.c (get_mem_ref_of_assignment): Don't instrument
24178 gimple_clobber_p stmts.
24179 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
24180 gimple_clobber_p stmt if they have MEM_REF lhs and
24181 are dead because of another gimple_clobber_p stmt.
24182 * tree-ssa-live.c (clear_unused_block_pointer): Treat
24183 gimple_clobber_p stmts like debug stmts.
24184 (remove_unused_locals): Remove clobbers with MEM_REF lhs
24185 that refer to unused VAR_DECLs or uninitialized values.
24186 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
24187 gimple_clobber_p stmts if they refer to removed parameters.
24188 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
24191 2013-04-02 Uros Bizjak <ubizjak@gmail.com>
24193 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
24194 using SWI48 mode attribute.
24196 2013-04-02 Wei Mi <wmi@google.com>
24198 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
24199 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
24200 *<rotate_insn><mode>3_mask in i386.md.
24202 2013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
24204 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
24206 2013-04-02 Richard Biener <rguenther@suse.de>
24208 PR tree-optimization/56778
24209 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
24210 Runtime alias tests are not supported for gather loads.
24211 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
24212 stmts referenced from SSA operands before updating SSA form.
24214 2013-04-02 Ian Caulfield <ian.caulfield@arm.com>
24215 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
24217 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
24218 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
24219 * config/arm/cortex-a53.md: New file.
24220 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
24221 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
24222 * config/arm/arm.c (arm_issue_rate): Likewise.
24223 * config/arm/arm-tune.md: Regenerate
24224 * config/arm/arm-tables.opt: Regenerate.
24225 * config/arm/arm-cores.def: Add cortex-a53.
24227 2013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
24229 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
24232 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
24234 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
24235 scalar load/store operations using B/H registers.
24236 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
24238 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
24240 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
24242 * config/aarch64/aarch64.c
24243 (aarch64_simd_scalar_immediate_valid_for_move): New.
24244 * config/aarch64/aarch64-protos.h
24245 (aarch64_simd_scalar_immediate_valid_for_move): New.
24246 * config/aarch64/constraints.md (Dh, Dq): New.
24247 * config/aarch64/iterators.md (hq): New.
24249 2013-04-02 Eric Botcazou <ebotcazou@adacore.com>
24251 * reorg.c (get_branch_condition): Deal with conditional returns.
24252 (fill_simple_delay_slots): Remove dead code dealing with jumps.
24254 2013-04-01 Wei Mi <wmi@google.com>
24256 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
24257 Truncate operand 2 using %b asm operand modifier.
24258 (*<shift_insn><mode>3_mask): Ditto.
24259 (*<rotate_insn><mode>3_mask): Ditto.
24261 2013-04-01 Steven Bosscher <steven@gcc.gnu.org>
24263 PR middle-end/56798
24264 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
24266 2013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
24268 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
24270 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
24272 2013-03-30 Lawrence Crowl <crowl@google.com>
24274 * dse.c (clear_alias_sets): Remove never set.
24275 (disqualified_clear_alias_sets): Remove never set.
24276 (clear_alias_mode_pool): Remove never set.
24277 (dse_step0): Remove condition that is never true.
24278 (canon_address): Remove condition that is never true.
24279 (dse_step7): Remove condition that is never true.
24280 (rest_of_handle_dse): Remove condition that is never true.
24281 (rest_of_handle_dse::did_global): Remove never read from above.
24282 (dse_step2_spill): Remove never called from above.
24283 (dse_step5_spill): Remove never called from above.
24285 2013-03-30 Steven Bosscher <steven@gcc.gnu.org>
24287 * doc/md.texi (Standard Names) <casesi>: Update documentation for
24288 JUMP_TABLE_DATA changes.
24289 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
24290 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
24291 (Insns) <jump_table_data>: New entry.
24292 * doc/tm.texi: Regenerate.
24294 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
24296 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
24297 for table jump at the end of a basic block using tablejump_p.
24298 * targhooks.c (default_invalid_within_doloop): Likewise.
24299 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
24300 target hook implementation that is identical to the default hook.
24301 (rs6000_invalid_within_doloop): Remove.
24303 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
24304 unused variable from tablejump_p call.
24306 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
24307 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
24308 (INSN_DELETED_P): Likewise.
24309 (emit_jump_table_data): New prototype.
24310 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
24311 after 4th as unused.
24312 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
24313 * sched-vis.c (print_insn): Likewise.
24314 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
24315 insn for compatibility with back ends that use next_active_insn to
24316 identify jump table data.
24317 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
24318 (remove_insn): Likewise.
24319 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
24321 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
24322 (emit_jump_table_data): New function.
24324 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
24325 basic block, a JUMP_TABLE_DATA never is.
24326 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
24327 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
24328 off from code handling real insns.
24329 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
24330 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
24331 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
24332 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
24333 is not a NONDEBUG_INSN_P.
24334 * ira-costs.c (scan_one_insn): Likewise.
24335 * jump.c (mark_all_labels): Likewise.
24336 (mark_jump_label_1): Likewise.
24337 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
24338 * lra.c (get_insn_freq): Expect all insns reaching here to be in
24340 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
24341 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
24342 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
24343 JUMP_TABLE_DATA_P insns.
24344 (calculate_elim_costs_all_insns): Likewise.
24345 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
24346 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
24347 (delete_output_reload): Code style fixups.
24348 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
24349 insn flags on this non-insn.
24350 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
24351 as scheduling barriers, for pre-change compatibility.
24352 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
24353 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
24355 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
24356 redundant JUMP_TABLE_DATA_P test.
24357 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
24358 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
24359 (frv_for_each_packet): Likewise.
24360 * config/i386/i386.c (min_insn_size): Likewise.
24361 (ix86_avoid_jump_mispredicts): Likewise.
24362 * config/m32r/m32r.c (m32r_is_insn): Likewise.
24363 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
24364 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
24365 (mips16_insn_length): Robustify.
24366 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
24367 (mips16_split_long_branches): Likewise.
24368 * config/pa/pa.c (pa_combine_instructions): Likewise.
24369 * config/rs6000/rs6000.c (get_next_active_insn): Treat
24370 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
24371 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
24372 as contributing to pool range lengths.
24373 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
24374 Remove redundant JUMP_TABLE_DATA_P test.
24375 (sh_loop_align): Likewise.
24376 (split_branches): Likewise.
24377 (sh_insn_length_adjustment): Likewise.
24378 * config/spu/spu.c (get_branch_target): Likewise.
24380 2013-03-29 Jan Hubicka <jh@suse.cz>
24382 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
24383 gcov streaming; stream hot bb threshold to ltrans.
24384 * predict.c (get_hot_bb_threshold): Break out from ....
24385 (maybe_hot_count_p): ... here.
24386 (set_hot_bb_threshold): New function.
24387 * lto-section-in.c (lto_section_name): Add profile.
24388 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
24389 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
24390 and data-streamer.h
24391 (histogram_entry): New structure.
24392 (histogram, histogram_pool): New global vars.
24393 (histogram_hash): New structure.
24394 (histogram_hash::hash): New method.
24395 (histogram_hash::equal): Likewise.
24396 (account_time_size): New function.
24397 (cmp_counts): New function.
24398 (dump_histogram): New function.
24399 (ipa_profile_generate_summary): New function.
24400 (ipa_profile_write_summary): New function.
24401 (ipa_profile_read_summary): New function.
24402 (ipa_profile): Decide on threshold.
24403 (pass_ipa_profile): Add ipa_profile_write_summary and
24404 ipa_profile_read_summary.
24405 * Makefile.in (ipa.o): Update dependencies.
24406 * lto-streamer.h (LTO_section_ipa_profile): New section.
24408 2013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
24410 * tree.h (VAR_P): New.
24412 2013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
24415 * doc/invoke.texi ([-fwhole-program]): Fix typo.
24417 2013-03-29 Steven Bosscher <steven@gcc.gnu.org>
24419 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
24420 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
24421 (control_flow_insn_p): Likewise.
24422 * cfgrtl.c (duplicate_insn_chain): Likewise.
24423 * final.c (get_attr_length_1): Likewise.
24424 (shorten_branches): Likewise.
24425 (final_scan_insn): Likewise.
24426 * function.c (instantiate_virtual_regs): Likewise.
24427 * gcse.c (insert_insn_end_basic_block): Likewise.
24428 * ira-costs.c (scan_one_insn): Likewise.
24429 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
24430 * lra.c (check_rtl): Likewise.
24431 * reload1.c (elimination_costs_in_insn): Likewise.
24432 * reorg.c (follow_jumps): Likewise.
24434 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
24435 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
24436 (thumb_far_jump_used_p): Likewise.
24437 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
24438 (workaround_speculation): Likewise.
24439 (add_sched_insns_for_speculation): Likewise.
24440 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
24441 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
24442 (frv_for_each_packet): Likewise.
24443 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
24444 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
24445 (final_emit_insn_group_barriers): Likewise.
24446 * config/m32r/m32r.c (m32r_is_insn): Likewise.
24447 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
24448 (mips16_insn_length): Likewise.
24449 * config/pa/pa.c (pa_reorg): Likewise.
24450 (pa_combine_instructions): Likewise.
24451 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
24452 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
24453 (sh_reorg): Likewise.
24454 (split_branches): Likewise.
24455 * config/spu/spu.c (get_branch_target): Likewise.
24457 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
24460 2013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
24462 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
24463 Fix declaration name.
24465 2013-03-28 Lawrence Crowl <crowl@google.com>
24467 * graphds.h (struct graph.indicies): Remove unused.
24468 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
24469 (SCOP_ORIGINAL_PDDRS): Remove unused.
24470 * sese.h (extern insert_loop_close_phis): Removed unused.
24471 (extern insert_guard_phis): Removed unused.
24472 (extern ivtype_map_elt_info): Removed unused.
24473 (new_ivtype_map_elt): Removed unused.
24474 * sese.c (ivtype_map_elt_info): Removed unused.
24476 2013-03-28 Lawrence Crowl <crowl@google.com>
24478 * Makefile.in: Add several missing include dependences.
24480 (test-dump.o): New. This object is not added to any executable,
24481 but is present for ad-hoc testing.
24483 (debug (const bitmap_head_def &)): New.
24484 (debug (const bitmap_head_def *)): New.
24486 (extern debug (const bitmap_head_def &)): New.
24487 (extern debug (const bitmap_head_def *)): New.
24489 (debug (edge_def &)): New.
24490 (debug (edge_def *)): New.
24492 (debug (basic_block_def &)): New.
24493 (debug (basic_block_def *)): New.
24495 (dump_node (const_tree, int, FILE *)): Correct source file.
24497 (debug (die_struct &)): New.
24498 (debug (die_struct *)): New.
24500 (extern debug (die_struct &)): New.
24501 (extern debug (die_struct *)): New.
24502 * gimple-pretty-print.c
24503 (debug (gimple_statement_d &)): New.
24504 (debug (gimple_statement_d *)): New.
24505 * gimple-pretty-print.h
24506 (extern debug (gimple_statement_d &)): New.
24507 (extern debug (gimple_statement_d *)): New.
24509 (debug (ira_allocno_copy &)): New.
24510 (debug (ira_allocno_copy *)): New.
24511 (debug (ira_allocno &)): New.
24512 (debug (ira_allocno *)): New.
24514 (extern debug (ira_allocno_copy &)): New.
24515 (extern debug (ira_allocno_copy *)): New.
24516 (extern debug (ira_allocno &)): New.
24517 (extern debug (ira_allocno *)): New.
24519 (debug (live_range &)): New.
24520 (debug (live_range *)): New.
24522 (debug (lra_live_range &)): New.
24523 (debug (lra_live_range *)): New.
24525 (debug (lra_live_range &)): New.
24526 (debug (lra_live_range *)): New.
24528 (debug (omega_pb_d &)): New.
24529 (debug (omega_pb_d *)): New.
24531 (extern debug (omega_pb_d &)): New.
24532 (extern debug (omega_pb_d *)): New.
24534 (debug (const rtx_def &)): New.
24535 (debug (const rtx_def *)): New.
24537 (debug_tree (tree): Move within file.
24538 (debug_raw (const tree_node &)): New.
24539 (debug_raw (const tree_node *)): New.
24540 (dump_tree_via_hooks (const tree_node *, int)): New.
24541 (debug (const tree_node &)): New.
24542 (debug (const tree_node *)): New.
24543 (debug_verbose (const tree_node &)): New.
24544 (debug_verbose (const tree_node *)): New.
24545 (debug_head (const tree_node &)): New.
24546 (debug_head (const tree_node *)): New.
24547 (debug_body (const tree_node &)): New.
24548 (debug_body (const tree_node *)): New.
24549 (debug_vec_tree (tree): Move and reimplement in terms of dump.
24550 (debug (vec<tree, va_gc> &)): New.
24551 (debug (vec<tree, va_gc> *)): New.
24553 (extern debug (const rtx_def &)): New.
24554 (extern debug (const rtx_def *)): New.
24556 (debug_raw (simple_bitmap_def &)): New.
24557 (debug_raw (simple_bitmap_def *)): New.
24558 (debug (simple_bitmap_def &)): New.
24559 (debug (simple_bitmap_def *)): New.
24561 (extern debug (simple_bitmap_def &)): New.
24562 (extern debug (simple_bitmap_def *)): New.
24563 (extern debug_raw (simple_bitmap_def &)): New.
24564 (extern debug_raw (simple_bitmap_def *)): New.
24566 (debug (vinsn_def &)): New.
24567 (debug (vinsn_def *)): New.
24568 (debug_verbose (vinsn_def &)): New.
24569 (debug_verbose (vinsn_def *)): New.
24570 (debug (expr_def &)): New.
24571 (debug (expr_def *)): New.
24572 (debug_verbose (expr_def &)): New.
24573 (debug_verbose (expr_def *)): New.
24574 (debug (vec<rtx> &)): New.
24575 (debug (vec<rtx> *)): New.
24577 (extern debug (vinsn_def &)): New.
24578 (extern debug (vinsn_def *)): New.
24579 (extern debug_verbose (vinsn_def &)): New.
24580 (extern debug_verbose (vinsn_def *)): New.
24581 (extern debug (expr_def &)): New.
24582 (extern debug (expr_def *)): New.
24583 (extern debug_verbose (expr_def &)): New.
24584 (extern debug_verbose (expr_def *)): New.
24585 (extern debug (vec<rtx> &)): New.
24586 (extern debug (vec<rtx> *)): New.
24588 (_list_iter_cond_expr): Make inline instead of static.
24590 (debug (sreal &)): New.
24591 (debug (sreal *)): New.
24593 (extern debug (sreal &)): New.
24594 (extern debug (sreal *)): New.
24596 (extern debug_raw (const tree_node &)): New.
24597 (extern debug_raw (const tree_node *)): New.
24598 (extern debug (const tree_node &)): New.
24599 (extern debug (const tree_node *)): New.
24600 (extern debug_verbose (const tree_node &)): New.
24601 (extern debug_verbose (const tree_node *)): New.
24602 (extern debug_head (const tree_node &)): New.
24603 (extern debug_head (const tree_node *)): New.
24604 (extern debug_body (const tree_node &)): New.
24605 (extern debug_body (const tree_node *)): New.
24606 (extern debug (vec<tree, va_gc> &)): New.
24607 (extern debug (vec<tree, va_gc> *)): New.
24609 (debug (struct loop &)): New.
24610 (debug (struct loop *)): New.
24611 (debug_verbose (struct loop &)): New.
24612 (debug_verbose (struct loop *)): New.
24613 * tree-dump.c: Add header dependence.
24615 (extern debug (struct loop &)): New.
24616 (extern debug (struct loop *)): New.
24617 (extern debug_verbose (struct loop &)): New.
24618 (extern debug_verbose (struct loop *)): New.
24620 (debug (data_reference &)): New.
24621 (debug (data_reference *)): New.
24622 (debug (vec<data_reference_p> &)): New.
24623 (debug (vec<data_reference_p> *)): New.
24624 (debug (vec<ddr_p> &)): New.
24625 (debug (vec<ddr_p> *)): New.
24627 (extern debug (data_reference &)): New.
24628 (extern debug (data_reference *)): New.
24629 (extern debug (vec<data_reference_p> &)): New.
24630 (extern debug (vec<data_reference_p> *)): New.
24631 (extern debug (vec<ddr_p> &)): New.
24632 (extern debug (vec<ddr_p> *)): New.
24634 (debug (pt_solution &)): New.
24635 (debug (pt_solution *)): New.
24637 (extern debug (pt_solution &)): New.
24638 (extern debug (pt_solution *)): New.
24640 (debug (_var_map &)): New.
24641 (debug (_var_map *)): New.
24642 (debug (tree_live_info_d &)): New.
24643 (debug (tree_live_info_d *)): New.
24645 (extern debug (_var_map &)): New.
24646 (extern debug (_var_map *)): New.
24647 (extern debug (tree_live_info_d &)): New.
24648 (extern debug (tree_live_info_d *)): New.
24650 2013-03-28 Jan Hubicka <jh@suse.cz>
24652 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
24654 2013-03-28 Ian Bolton <ian.bolton@arm.com>
24656 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
24657 record only when desired or required.
24659 2013-03-28 Uros Bizjak <ubizjak@gmail.com>
24661 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
24662 *vec_extractv2di_1_rex64. Use x64 isa attribute.
24664 2013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
24666 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
24667 (*andsi3_compare0_uxtw): New pattern.
24668 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
24669 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
24671 2013-03-28 Jan Hubicka <jh@suse.cz>
24673 * data-streamer-in.c (streamer_read_gcov_count): New function.
24674 * gimple-streamer-out.c: Include value-prof.h.
24675 (output_gimple_stmt): Output histogram.
24676 (output_bb): Use streamer_write_gcov_count.
24677 * value-prof.c: Include data-streamer.h
24678 (dump_histogram_value): Add HIST_TYPE_MAX.
24679 (stream_out_histogram_value): New function.
24680 (stream_in_histogram_value): New function.
24681 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
24682 (stream_out_histogram_value, stream_in_histogram_value): Declare.
24683 * data-streamer-out.c (streamer_write_gcov_count): New function.
24684 (streamer_write_gcov_count_stream): New function.
24685 * lto-cgraph.c (lto_output_edge): Update counter streaming.
24686 (lto_output_node): Likewise.
24687 (input_node, input_edge): Likewise.
24688 * lto-streamer-out.c (output_cfg): Update streaming.
24689 * lto-streamer-in.c (input_cfg): Likewise.
24690 * data-streamer.h (streamer_write_gcov_count,
24691 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
24692 * gimple-streamer-in.c: Include value-prof.h
24693 (input_gimple_stmt): Input histograms.
24694 (input_bb): Update profile streaming.
24696 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
24698 * genmodes.c (emit_max_int): New function.
24699 (emit_insn_modes_h): Added call to emit_max_function.
24700 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
24702 * machmode.def: Fixed comment.
24704 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
24706 * combine.c (try_combine): Removed useless assert.
24707 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
24709 2013-03-28 Marek Polacek <polacek@redhat.com>
24710 Richard Biener <rguenther@suse.de>
24712 PR tree-optimization/56695
24713 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
24714 build signed result of a vector comparison.
24715 * tree-cfg.c (verify_gimple_comparison): Check that a result
24716 of a vector comparison has signed type.
24718 2013-03-28 Richard Biener <rguenther@suse.de>
24720 PR tree-optimization/37021
24721 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
24722 do not restrict gaps between groups.
24723 * tree-vect-stmts.c (vectorizable_load): Properly account for
24724 a gap between groups.
24726 2013-03-28 Eric Botcazou <ebotcazou@adacore.com>
24728 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
24729 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
24732 2013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
24734 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
24735 * doc/extend.texi (Named Address Spaces): Ditto.
24736 (Variable Attributes): Ditto.
24738 2013-03-27 Kai Tietz <ktietz@redhat.com>
24740 * config.build: Add support for cygwin x64 target.
24741 * config.gcc: Likewise.
24742 * config.host: Likewise.
24743 * configure.ac: Likewise
24744 * configure: Regenerated.
24746 2013-03-27 Kai Tietz <ktietz@redhat.com>
24748 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
24749 * config/i386/t-cygwin-w64: New file.
24750 * config/i386/cygwin-w64.h: New file.
24751 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
24752 and add support for x64-cygwin target.
24753 (CPP_SPEC): Likewise.
24754 (CXX_WRAP_SPEC_LIST): Undefine before define.
24755 (LIBGCJ_SONAME): Use 15 as version.
24757 2013-03-27 Richard Biener <rguenther@suse.de>
24759 PR tree-optimization/56716
24760 * tree-ssa-structalias.c (perform_var_substitution): Adjust
24761 dumping for ref nodes.
24763 2013-03-27 Martin Jambor <mjambor@suse.cz>
24765 PR tree-optimization/55334
24766 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
24767 restricted pointers to arrays.
24769 2013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
24771 * Makefile.in (.SUFFIXES): Add .cc.
24772 (.c.o): Apply same recipe for implicit rule .cc.o.
24774 2013-03-27 Richard Biener <rguenther@suse.de>
24776 PR tree-optimization/37021
24777 * tree-vect-data-refs.c (vect_check_strided_load): Allow
24778 REALPART/IMAGPART_EXPRs around the supported refs.
24779 * tree-ssa-structalias.c (find_func_aliases): Assume that
24780 floating-point values are not used to transfer pointers.
24782 2013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
24784 * target.def (TARGET_HAS_IFUNC_P): New target hook.
24785 * doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
24786 * doc/tm.texi: Regenerate.
24787 * targhooks.h (default_has_ifunc_p): New.
24788 * targhooks.c (default_has_ifunc_p): Ditto.
24789 * config/linux-protos.h: New file.
24790 * config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of this
24791 hook for linux which disables support of indirect functions in android.
24792 * config/linux-android.c: New file.
24793 * config/t-linux-android.c: Ditto.
24794 * config.gcc: Added new object file linux-android.o.
24795 * config/i386/i386.c (ix86_get_function_versions_dispatcher):
24796 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
24797 * varasm.c (do_assemble_alias): Likewise.
24798 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
24799 doesn't support indirect functions.
24800 * configure: Regenerate.
24802 2013-03-27 Bin Cheng <bin.cheng@arm.com>
24805 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
24806 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
24809 2013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
24811 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
24813 2013-03-27 Terry Guo <terry.guo@arm.com>
24815 * config/arm/arm-cores.def: Added core cortex-r7.
24816 * config/arm/arm-tune.md: Regenerated.
24817 * config/arm/arm-tables.opt: Regenerated.
24818 * doc/invoke.texi: Added entry for core cortex-r7.
24820 2013-03-27 Walter Lee <walt@tilera.com>
24822 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
24823 double-decrement of next_scratch_regno.
24825 2013-03-27 Walter Lee <walt@tilera.com>
24827 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
24829 (insn_v1mulus): Ditto.
24830 (insn_v2muls): Ditto.
24832 2013-03-27 Walter Lee <walt@tilera.com>
24834 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
24835 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
24837 2013-03-27 Walter Lee <walt@tilera.com>
24839 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
24840 (*sibcall_value): Ditto.
24842 2013-03-27 Walter Lee <walt@tilera.com>
24844 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
24845 (insn_mnz_v8qi): ... this ...
24846 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
24848 (insn_v<n>mnz): Replaced by ...
24849 (insn_v1mnz): ... this ...
24850 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
24852 (insn_mz_<mode>): Replaced by ...
24853 (insn_mz_v8qi): ... this ...
24854 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
24856 (insn_v<n>mz): Replaced by ...
24857 (insn_v1mz): ... this ...
24858 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
24861 2013-03-26 Eric Botcazou <ebotcazou@adacore.com>
24863 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
24865 2013-03-26 Roland McGrath <mcgrathr@google.com>
24867 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
24868 than fprintf with a non-constant, non-format string.
24870 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
24872 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
24873 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
24874 operand 0 predicate.
24875 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
24876 attribute. Use general_x64nomem_operand as operand 1 predicate.
24877 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
24878 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
24879 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
24880 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
24881 (mov<mode>_insv_1): Remove expander. Merge insn with
24882 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
24883 Use general_x64nomem_operand as operand 1 predicate.
24884 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
24885 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
24886 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
24887 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
24888 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
24889 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
24890 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
24891 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
24892 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
24893 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
24894 (general_x64nomem_operand): Ditto.
24896 2013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
24898 * config/rtems.opt: Add -pthread option.
24900 2013-03-26 Richard Biener <rguenther@suse.de>
24902 * alias.c (find_base_term): Avoid redundant and not used recursion.
24903 (base_alias_check): Get the initial base term from the caller.
24904 (true_dependence_1): Compute and pass base terms to base_alias_check.
24905 (write_dependence_p): Likewise.
24906 (may_alias_p): Likewise.
24908 2013-03-26 Sofiane Naci <sofiane.naci@arm.com>
24910 * config/aarch64/aarch64.c (aarch64_classify_address): Support
24911 PC-relative load in SI modes and above only.
24913 2013-03-26 Xinyu Qi <xyqi@marvell.com>
24915 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
24916 * config/arm/iwmmxt.md (WCGR0): Update.
24917 (WCGR1, WCGR2, WCGR3): Likewise.
24919 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
24921 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
24922 Use x64 and nox64 isa attributes.
24924 2013-03-26 Richard Biener <rguenther@suse.de>
24926 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
24927 alignment computations and rely on get_object_alignment_1
24928 for the !TYPE_P case.
24929 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
24931 2013-03-26 Walter Lee <walt@tilera.com>
24933 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
24934 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
24936 2013-03-25 Jeff Law <law@redhat.com>
24938 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
24939 check for INTEGRAL_TYPE_P that was missing due to checking in
24940 wrong version of prior patch.
24942 2013-03-25 Walter Lee <walt@tilera.com>
24944 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
24945 TILEGX_INSN_SHUFFLEBYTES1.
24946 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
24948 (tilegx_builtins): Ditto.
24949 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
24951 2013-03-25 Walter Lee <walt@tilera.com>
24953 * config/tilegx/tilegx.md (floatsisf2): New pattern.
24954 (floatunssisf2): New pattern.
24955 (floatsidf2): New pattern.
24956 (floatunssidf2): New pattern.
24958 2013-03-25 Walter Lee <walt@tilera.com>
24960 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
24961 tests for constraint J, K, N, P.
24963 2013-03-25 Walter Lee <walt@tilera.com>
24965 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
24966 Use indirect/pcrel encoding.
24967 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
24970 2013-03-25 Steve Ellcey <sellcey@mips.com>
24972 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
24973 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
24974 * config/mips/mips.c (mips_option_override): Set IMADD default.
24975 * config/mips/mips.h (PTF_AVOID_IMADD): New.
24976 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
24977 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
24978 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
24979 * doc/invoke.texi (-mimadd/-mno-imadd): New.
24981 2013-03-25 Jeff Law <law@redhat.com>
24983 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
24984 slightly to avoid creating and folding useless trees. Simplify
24985 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
24987 2013-03-25 Uros Bizjak <ubizjak@gmail.com>
24989 * config/i386/i386.md (*zero_extendsidi2): Merge with
24990 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
24991 * config/i386/predicates.md (x86_64_zext_operand): Rename from
24992 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
24993 targets. Clarify comment.
24995 2013-03-25 Martin Jambor <mjambor@suse.cz>
24997 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
24998 pass-through jump functions differently.
24999 (ipa_read_jump_function): Likewise. Also use setter functions to set
25002 2013-03-25 Martin Jambor <mjambor@suse.cz>
25004 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
25005 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
25007 (ipa_get_indirect_edge_target): New function.
25008 (devirtualization_time_bonus): New parameter known_aggs, pass it to
25009 ipa_get_indirect_edge_target. Update all callers.
25010 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
25011 ipa_get_indirect_edge_target_1 instead of calling
25012 ipa_get_indirect_edge_target.
25013 (create_specialized_node): Pass aggvlas to
25014 ipcp_discover_new_direct_edges.
25016 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
25018 * config/arm/arm.md (f_sels, f_seld): New types.
25019 (*cmov<mode>): New pattern.
25020 * config/arm/predicates.md (arm_vsel_comparison_operator): New
25023 2013-03-25 Kai Tietz <ktietz@redhat.com>
25025 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
25026 POSIX-printf for mingw-hosted builds.
25028 2013-03-25 Richard Biener <rguenther@suse.de>
25030 PR middle-end/56694
25031 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
25032 must-not-throw stmt location.
25034 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
25036 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
25037 Emit load-acquire versions when acq is true.
25038 (arm_emit_store_exclusive): Add rel parameter.
25039 Emit store-release versions when rel is true.
25040 (arm_split_compare_and_swap): Use acquire-release instructions
25042 of barriers when appropriate.
25043 (arm_split_atomic_op): Likewise.
25044 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
25045 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
25046 (VUNSPEC_SLX): Likewise.
25047 (VUNSPEC_LDA): Likewise.
25048 (VUNSPEC_STL): Likewise.
25049 * config/arm/sync.md (atomic_load<mode>): New pattern.
25050 (atomic_store<mode>): Likewise.
25051 (arm_load_acquire_exclusive<mode>): Likewise.
25052 (arm_load_acquire_exclusivesi): Likewise.
25053 (arm_load_acquire_exclusivedi): Likewise.
25054 (arm_store_release_exclusive<mode>): Likewise.
25056 2013-03-25 Catherine Moore <clm@codesourcery.com>
25058 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
25059 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
25060 * config/mip/predicates.md (lwsp_swsp_operand,
25061 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
25062 sb16_operand, db4_operand, db7_operand, ib3_operand,
25063 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
25064 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
25065 andi16_operand): New predicates.
25066 * config/mips/mips.md (compression): New attribute.
25067 (enabled): New attribute.
25068 (length): Consider compression in computing length.
25069 (shift_compression): New code attribute.
25070 (*add<mode>3): New operands. Record compression.
25071 (sub<mode>3): Likewise.
25072 (one_cmpl<mode>2): Likewise.
25073 (*and<mode>3): Likewise.
25074 (*ior<mode>3): Likewise.
25075 (unnamed pattern for xor): Likewise.
25076 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
25077 (*<optab><mode>3): Likewise.
25078 (*mov<mode>_internal: Likewise.
25079 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
25080 (mips_unsigned_immediate_p): New.
25081 (umips_lwsp_swsp_address_p): New.
25082 (m16_based_address_p): New.
25083 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
25084 (mips_unsigned_immediate_p): New prototype.
25085 (lwsp_swsp_address_p): New prototype.
25086 (m16_based_address_p): New prototype.
25087 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
25088 (mips_signed_immediate_p): New function.
25089 (m16_based_address_p): New function.
25090 (lwsp_swsp_address_p): New function.
25091 (mips_print_operand_punctuation): Recognize short delay slot insns
25092 for microMIPS.add<mode>3"
25094 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
25097 * config/arm/iterators.md (v_cmp_result): New mode attribute.
25098 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
25100 2013-03-25 Richard Biener <rguenther@suse.de>
25102 PR tree-optimization/56689
25103 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
25106 2013-03-25 Richard Biener <rguenther@suse.de>
25108 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
25110 (memory_references): Likewise.
25111 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
25112 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
25113 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
25114 (gather_mem_refs_in_loops): Fold into ...
25115 (analyze_memory_references): ... this. Move initialization
25116 to tree_ssa_lim_initialize.
25117 (fill_always_executed_in): Rename to ...
25118 (fill_always_executed_in_1): ... this.
25119 (fill_always_executed_in): Move contains_call computation to
25120 this new function from ...
25121 (tree_ssa_lim_initialize): ... here.
25122 (tree_ssa_lim): Call fill_always_executed_in.
25124 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
25126 * postreload.c (reload_combine): Fix code detecting returns.
25128 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
25130 * function.c (emit_use_return_register_into_block): On cc0 targets,
25131 do not emit the sequence between cc0 setter and user.
25133 2013-03-25 Kai Tietz <ktietz@redhat.com>
25135 * config/i386/predicates.md (local_symbolic_operand): Interpret
25136 dll-imported symbols as none-local.
25138 2013-03-25 Richard Biener <rguenther@suse.de>
25140 * tree-ssa-loop-im.c (struct depend): Remove.
25141 (struct lim_aux_data): Make depends a vec of gimples.
25142 (free_lim_aux_data): Adjust.
25143 (add_dependency): Likewise.
25144 (set_level): Likewise.
25146 2013-03-25 Richard Biener <rguenther@suse.de>
25148 PR middle-end/56434
25149 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
25150 the pointer returned by calls with ECF_MALLOC set.
25152 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
25154 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
25156 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
25158 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
25159 using MMXMODE mode iterator.
25160 (*move<mode>_internal): Merge with *movv2sf_internal and
25161 *movv2sf_internal_rex64 using MMXMODE mode iterator.
25163 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
25165 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
25166 (record_last_mem_set_info): Likewise.
25168 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
25169 of XNEWVEC followed by memset.
25170 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
25172 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
25174 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
25175 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
25176 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
25177 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
25178 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
25179 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
25180 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
25181 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
25182 BARRIER_P instead of GET_CODE.
25184 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
25186 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
25187 inaccuracy in the probing code.
25189 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
25190 (ctrapdi4): Likewise.
25192 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
25194 * calls.c (expand_call): Add missing guard to code handling return
25195 of non-BLKmode structures in MSB.
25196 * function.c (expand_function_end): Likewise.
25198 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
25200 * combine.c (try_combine): Adjust comment. Do not add the set of
25201 insn #0 if the destination indirectly is set or dies in insn #2.
25202 Tidy up code to distribute a new note.
25204 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
25206 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
25207 also for alternatives 16 and 17.
25209 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
25211 * config/i386/sse.md (*mov<mode>_internal): Merge with
25212 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
25213 Emit insn template depending on type attribute. Use
25214 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
25215 movd instead of movq mnemonic for interunit moves. Rewrite mode
25216 attribute calculation. Remove unit attribute calculation.
25217 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
25218 Set prefix_data16 attribute for DImode ssemov types.
25219 Use Ym instead of y for SSE-MMX conversion alternatives.
25220 Reorder operand constraints.
25222 2013-03-22 Steven Bosscher <steven@gcc.gnu.org>
25224 * df.h (df_insn_delete): Adjust prototype.
25225 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
25226 and let it decide whether mark the basic block dirty.
25227 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
25228 * df-scan.c (df_insn_info_delete): New helper function, split
25229 off from df_insn_delete.
25230 (df_scan_free_bb_info): Use it.
25231 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
25233 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
25234 that the insn is actually an insn and it has a non-NULL basic block.
25235 Do not mark basic block dirty if only deleting a DEBUG_INSN.
25237 2013-03-22 Richard Biener <rguenther@suse.de>
25239 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
25241 (mem_ref_alloc): Do not allocate them.
25242 (refs_independent_p): Do not query or maintain a cache.
25244 2013-03-22 Richard Biener <rguenther@suse.de>
25246 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
25247 (gather_mem_refs_in_loops): Do not compute it.
25248 (analyze_memory_references): Do not allocate it.
25249 (tree_ssa_lim_finalize): Do not free it.
25250 (for_all_locs_in_loop): Do not query all_refs_in_loop.
25252 2013-03-22 Richard Biener <rguenther@suse.de>
25254 * is-a.h (as_a): Use gcc_checking_assert.
25256 2013-03-22 Ian Bolton <ian.bolton@arm.com>
25258 * config/aarch64/aarch64.c (aarch64_print_operand): New
25259 format specifier for printing a constant in hex.
25260 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
25261 format specifier for printing second operand.
25263 2013-03-22 Richard Biener <rguenther@suse.de>
25265 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
25267 (gather_mem_refs_in_loops): Perform store accumulation here.
25268 (create_vop_ref_mapping_loop): Remove.
25269 (create_vop_ref_mapping): Likewise.
25270 (analyze_memory_references): Initialize refs_stored_in_loop.
25271 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
25272 (record_indep_loop): Remove.
25273 (record_dep_loop): New function.
25274 (ref_indep_loop_p_1): Adjust to only walk over references
25275 in the loop, not its subloops.
25276 (ref_indep_loop_p): Rename to ...
25277 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
25278 maintaining a more fine-grained cache.
25279 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
25280 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
25282 2013-03-22 Richard Biener <rguenther@suse.de>
25284 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
25285 (struct mem_ref): Make accesses_in_loop a vec of a vec of
25286 aggregate mem_ref_loc.
25287 (free_mem_ref_locs): Inline into ...
25288 (memref_free): ... this and adjust.
25289 (mem_ref_alloc): Adjust.
25290 (mem_ref_locs_alloc): Remove.
25291 (record_mem_ref_loc): Adjust.
25292 (get_all_locs_in_loop): Rewrite into ...
25293 (for_all_locs_in_loop): ... this iterator.
25294 (rewrite_mem_ref_loc): New functor.
25295 (rewrite_mem_refs): Use for_all_locs_in_loop.
25296 (sm_set_flag_if_changed): New functor.
25297 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
25298 (ref_always_accessed): New functor.
25299 (ref_always_accessed_p): Use for_all_locs_in_loop.
25301 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
25303 * tree-pass.h (PROP_gimple_lvec): New.
25304 * passes.c (dump_properties): Handle PROP_gimple_lvec.
25305 (init_optimization_passes): Move pass_lower_vector.
25306 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
25308 (pass_lower_vector): Provide PROP_gimple_lvec.
25309 (pass_lower_vector_ssa): Likewise.
25310 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
25312 2013-03-21 Mark Wielaard <mjw@redhat.com>
25314 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
25316 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
25318 * config/i386/i386.md (*movdi_internal): Disparage slightly
25319 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
25320 conversion alternatives.
25322 2013-03-21 Jakub Jelinek <jakub@redhat.com>
25324 PR middle-end/48087
25325 * diagnostic.def (DK_WERROR): New kind.
25326 * diagnostic.h (werrorcount): Define.
25327 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
25328 promoted to DK_ERROR, increment DK_WERROR counter instead of
25330 * toplev.c (toplev_main): Call print_ignored_options even if
25331 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
25332 even if just werrorcount is non-zero.
25335 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
25337 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
25338 (string_cst_pool_decl): New function.
25339 (optimize_one_addr_into_implicit_ptr): New function.
25340 (resolve_addr_in_expr): Optimize DWARF location expression
25341 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
25342 which doesn't live in memory, but has DW_AT_location or
25343 DW_AT_const_value, or refers to a string literal, into
25344 DW_OP_GNU_implicit_pointer.
25345 (optimize_location_into_implicit_ptr): New function.
25346 (resolve_addr): If removing DW_AT_location of a variable because
25347 it was DW_OP_addr of address of the variable, but the variable doesn't
25348 live in memory, try to emit const value attribute for the initializer.
25350 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
25352 * tree.h (VECTOR_TYPE_P): New macro.
25353 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
25354 TYPE_MODE): Use it.
25355 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
25356 VEC_COND_EXPR cannot be lvalues.
25357 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
25359 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
25361 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
25362 Restrict the transformation to equal modes.
25364 2013-03-21 Richard Biener <rguenther@suse.de>
25366 PR tree-optimization/39326
25367 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
25368 (MEM_ANALYZABLE): Adjust.
25369 (record_mem_ref_loc): Move bitmap ops ...
25370 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
25371 unanalyzable refs, do not record locations for it.
25372 (analyze_memory_references): Allocate ref zero as shared
25374 (refs_independent_p): Do not test for unanalyzed mems here.
25375 (ref_indep_loop_p_1): Special-case disambiguation against
25376 the unanalyzed ref.
25377 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
25379 2013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
25381 * config/arm/arm-protos.h (tune_params): Add
25382 prefer_neon_for_64bits field.
25383 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
25384 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
25385 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
25386 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
25387 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
25388 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
25389 (arm_option_override): Handle -mneon-for-64bits new option.
25390 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
25391 (prefer_neon_for_64bits): Declare new variable.
25392 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
25393 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
25394 (arch_enabled): Handle new arch types. Remove support for onlya8
25396 (one_cmpldi2): Use new arch names.
25397 (zero_extend<mode>di2, extend<mode>di2): Ditto.
25398 * config/arm/arm.opt (mneon-for-64bits): Add option.
25399 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
25400 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
25401 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
25403 * doc/invoke.texi (-mneon-for-64bits): Document.
25405 2013-03-21 Richard Biener <rguenther@suse.de>
25407 PR tree-optimization/39326
25408 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
25409 (sort_bbs_in_loop_postorder_cmp): New function.
25410 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
25412 2013-03-21 Richard Biener <rguenther@suse.de>
25414 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
25415 (vect_insert_into_interleaving_chain): Likewise.
25416 (vect_drs_dependent_in_basic_block): Inline ...
25417 (vect_slp_analyze_data_ref_dependence): ... here. New function,
25419 (vect_analyze_data_ref_dependence): ... here. Simplify.
25420 (vect_check_interleaving): Simplify.
25421 (vect_analyze_data_ref_dependences): Likewise. Split out ...
25422 (vect_slp_analyze_data_ref_dependences): ... this new function.
25423 (dr_group_sort_cmp): New function.
25424 (vect_analyze_data_ref_accesses): Compute data-reference groups
25425 here instead of in vect_analyze_data_ref_dependence. Use
25426 a more efficient algorithm.
25427 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
25428 vect_slp_analyze_data_ref_dependences. Call
25429 vect_analyze_data_ref_accesses earlier.
25430 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
25431 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
25432 (vect_slp_analyze_data_ref_dependences): New prototype.
25434 2013-03-21 Richard Biener <rguenther@suse.de>
25436 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
25437 ref is stored in the loop.
25438 (find_refs_for_sm): Walk only over all stores.
25439 (store_motion_loop): Allocate from lim_bitmap_obstack.
25440 (store_motion): Likewise.
25442 2013-03-21 Richard Biener <rguenther@suse.de>
25444 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
25445 Update virtual SSA form.
25447 2013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
25449 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
25450 * configure: Regenerate.
25451 * config.in: Regenerate.
25452 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
25453 if !HAVE_LD_EH_FRAME_CIEV3.
25455 2013-03-21 Richard Biener <rguenther@suse.de>
25457 * tree-cfg.c (verify_expr_no_block): New function.
25458 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
25459 nor DECL_VALUE_EXPR have locations with associated blocks.
25460 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
25461 (clear_unused_block_pointer): Remove code dealing with
25462 blocks in DECL_DEBUG_EXPR locations.
25464 2013-03-21 Richard Biener <rguenther@suse.de>
25466 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
25467 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
25468 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
25469 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
25470 instead of DECL_DEBUG_EXPR_IS_FROM.
25471 * gimplify.c (gimplify_modify_expr): Likewise.
25472 * tree-cfg.c (verify_expr_location_1): Likewise.
25473 * tree-complex.c (create_one_component_var): Likewise.
25474 * tree-sra.c (create_access_replacement): Likewise.
25475 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
25476 (clear_unused_block_pointer): Likewise.
25477 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
25478 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
25479 * var-tracking.c (var_debug_decl): Likewise.
25480 (track_expr_p): Likewise.
25481 * tree-inline.c (add_local_variables): Likewise. Set
25482 DECL_HAS_DEBUG_EXPR_P after copying it.
25483 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
25484 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
25486 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
25489 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
25490 * configure: Regenerate.
25491 * config.in: Regenerate.
25492 * config/i386/i386.md (*movdf_internal): Use
25493 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
25494 movd instead of movq mnemonic for interunit moves.
25495 (*movdi_internal): Ditto.
25497 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
25499 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
25500 (abd<mode>_3): New pattern.
25501 (aba<mode>_3): New pattern.
25502 (fabd<mode>_3): New pattern.
25504 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
25506 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
25507 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
25508 occurrence of REGISTER_PREFIX as its empty string.
25510 2013-03-20 Jeff Law <law@redhat.com>
25512 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
25513 addititional equivalences for equality comparisons between an SSA_NAME
25514 and a constant where the SSA_NAME was set from a widening conversion.
25516 2013-03-20 Walter Lee <walt@tilera.com>
25518 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
25520 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
25522 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
25523 depending on type attribute.
25524 (*movti_internal): Ditto.
25525 (*movtf_internal): Ditto.
25526 (*movxf_internal): Ditto.
25527 (*movdf_internal): Ditto.
25528 (*movsf_internal): Ditto.
25530 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
25532 * config/i386/i386.md (*movti_internal): Set prefix attribute to
25533 maybe_vex for sselog1 and ssemov types.
25534 (*movdi_internal): Reorder operand constraints.
25535 (*movsi_internal): Ditto. Set prefix attribute to
25536 maybe_vex for sselog1 and ssemov types.
25537 (*movtf_internal): Set prefix attribute to maybe_vex
25538 for sselog1 and ssemov types.
25539 (*movdf_internal): Ditto. Set prefix_data16 attribute for
25540 DImode ssemov types. Reorder operand constraints.
25541 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
25542 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
25543 attribute for SImode ssemov types. Reorder operand constraints.
25545 2013-03-20 Martin Jambor <mjambor@suse.cz>
25547 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
25548 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
25550 2013-03-20 Pat Haugen <pthaugen@us.ibm.com>
25552 * config/rs6000/predicates.md (indexed_address, update_address_mem
25553 update_indexed_address_mem): New predicates.
25554 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
25555 attribute for load/store instructions.
25556 * config/rs6000/dfp.md (movsd_store): Likewise.
25557 (movsd_load): Likewise.
25558 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
25559 (unnamed HI->DI extend define_insn): Likewise.
25560 (unnamed SI->DI extend define_insn): Likewise.
25561 (unnamed QI->SI extend define_insn): Likewise.
25562 (unnamed QI->HI extend define_insn): Likewise.
25563 (unnamed HI->SI extend define_insn): Likewise.
25564 (unnamed HI->SI extend define_insn): Likewise.
25565 (extendsfdf2_fpr): Likewise.
25566 (movsi_internal1): Likewise.
25567 (movsi_internal1_single): Likewise.
25568 (movhi_internal): Likewise.
25569 (movqi_internal): Likewise.
25570 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
25571 attribute for load/store instructions.
25572 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
25574 (mov<mode>_softfloat): Likewise.
25575 (mov<mode>_hardfloat32): Likewise.
25576 (mov<mode>_hardfloat64): Likewise.
25577 (mov<mode>_softfloat64): Likewise.
25578 (movdi_internal32): Likewise.
25579 (movdi_internal64): Likewise.
25580 (probe_stack_<mode>): Likewise.
25582 2013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
25584 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
25585 floating point, and decimal floating point to reload iterator.
25587 * config/rs6000/constraints.md (wl constraint): New constraints to
25588 return FLOAT_REGS if certain options are used to reduce the number
25589 of separate patterns that exist in the file.
25590 (wx constraint): Likewise.
25591 (wz constraint): Likewise.
25593 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
25594 -mdebug=reg, print wg, wl, wx, and wz constraints.
25595 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
25596 Initialize the reload functions for 64-bit binary/decimal floating
25598 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
25599 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
25600 create the buffer on the stack to overcome not having a 32-bit
25602 (rs6000_emit_move): Likewise.
25603 (rs6000_secondary_memory_needed_rtx): Likewise.
25604 (rs6000_alloc_sdmode_stack_slot): Likewise.
25605 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
25606 via xxlxor, just like DFmode 0.0.
25608 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
25609 define as 1 if we are running on a power7 or newer.
25610 (enum r6000_reg_class_enum): Add new constraints.
25612 * config/rs6000/dfp.md (movsd): Delete, combine with binary
25613 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
25614 with other moves by using conditional constraits (wg). Use LFIWZX
25615 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
25616 (movsd splitter): Likewise.
25617 (movsd_hardfloat): Likewise.
25618 (movsd_softfloat): Likewise.
25620 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
25621 binary and decimal floating point moves.
25622 (fmove_ok): New attributes to combine binary and decimal floating
25623 point moves, and to combine power6x (mfpgpr) moves along normal
25625 (real_value_to_target): Likewise.
25626 (f32_lr): Likewise.
25627 (f32_lm): Likewise.
25628 (f32_li): Likewise.
25629 (f32_sr): Likewise.
25630 (f32_sm): Likewise.
25631 (f32_si): Likewise.
25632 (movsf): Combine binary and decimal floating point moves. Combine
25633 power6x (mfpgpr) moves with other moves by using conditional
25634 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
25635 (mov<mode> for SFmode/SDmode); Likewise.
25636 (SFmode/SDmode splitters): Likewise.
25637 (movsf_hardfloat): Likewise.
25638 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
25639 (movsf_softfloat): Likewise.
25640 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
25642 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
25643 wx and wz constraints.
25645 * config/rs6000/constraints.md (wg constraint): New constraint to
25646 return FLOAT_REGS if -mmfpgpr (power6x) was used.
25648 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
25651 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
25652 -mdebug=reg, print wg, wl, wx, and wz constraints.
25653 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
25654 Initialize the reload functions for 64-bit binary/decimal floating
25656 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
25657 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
25658 create the buffer on the stack to overcome not having a 32-bit
25660 (rs6000_emit_move): Likewise.
25661 (rs6000_secondary_memory_needed_rtx): Likewise.
25662 (rs6000_alloc_sdmode_stack_slot): Likewise.
25663 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
25664 via xxlxor, just like DFmode 0.0.
25666 * config/rs6000/dfp.md (movdd): Delete, combine with binary
25667 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
25668 with other moves by using conditional constraits (wg). Use LFIWZX
25669 and STFIWX for loading SDmode on power7.
25670 (movdd splitters): Likewise.
25671 (movdd_hardfloat32): Likewise.
25672 (movdd_softfloat32): Likewise.
25673 (movdd_hardfloat64_mfpgpr): Likewise.
25674 (movdd_hardfloat64): Likewise.
25675 (movdd_softfloat64): Likewise.
25677 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
25678 64-bit binary and decimal floating point moves.
25679 (FMOVE64X): Likewise.
25680 (movdf): Combine 64-bit binary and decimal floating point moves.
25681 Combine power6x (mfpgpr) moves with other moves by using
25682 conditional constraits (wg).
25683 (mov<mode> for DFmode/DDmode): Likewise.
25684 (DFmode/DDmode splitters): Likewise.
25685 (movdf_hardfloat32): Likewise.
25686 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
25687 (movdf_softfloat32): Likewise.
25688 (movdf_hardfloat64_mfpgpr): Likewise.
25689 (movdf_hardfloat64): Likewise.
25690 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
25691 (movdf_softfloat64): Likewise.
25692 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
25693 (reload_<mode>_load): Move to later in the file so they aren't in
25694 the middle of the floating point move insns.
25695 (reload_<mode>_store): Likewise.
25697 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
25700 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
25701 constraint if -mdebug=reg.
25702 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
25703 Enable using dd reload support if needed.
25705 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
25706 binary and decimal floating point moves in rs6000.md.
25707 (movtd_internal): Likewise.
25709 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
25710 decimal floating point moves.
25712 (movtf_internal): Likewise.
25713 (mov<mode>_internal, TDmode/TFmode): Likewise.
25714 (movtf_softfloat): Likewise.
25715 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
25717 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
25718 movdi_internal64, using wg constraint for move direct operations.
25719 (movdi_internal64): Likewise.
25721 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
25722 MODES_TIEABLE_P for selected modes. Print the numerical value of
25723 the various virtual registers. Use GPR/FPR first/last values,
25724 instead of hard coding the register numbers. Print which modes
25725 have reload functions registered.
25726 (rs6000_option_override_internal): If -mdebug=reg, trace the options
25727 settings before/after setting cpu, target and subtarget settings.
25728 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
25729 and for secondary reload failures in rs6000_secondary_reload_inner.
25730 (rs6000_secondary_reload_fail): Likewise.
25731 (rs6000_secondary_reload_inner): Likewise.
25733 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
25734 macros for first/last GPR and FPR registers.
25735 (LAST_GPR_REGNO): Likewise.
25736 (FIRST_FPR_REGNO): Likewise.
25737 (LAST_FPR_REGNO): Likewise.
25739 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
25740 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
25741 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
25742 (vcond<mode><mode>): Likewise.
25743 (vcondu<mode><mode>): Likewise.
25744 (vector_gtu<mode>): Likewise.
25745 (vector_gte<mode>): Likewise.
25746 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
25747 to prevent the compiler from converting DImode operations to TImode.
25748 (ior<mode>3): Likewise.
25749 (and<mode>3): Likewise.
25750 (one_cmpl<mode>2): Likewise.
25751 (nor<mode>3): Likewise.
25752 (andc<mode>3): Likewise.
25754 * config/rs6000/constraints.md (wt constraint): New constraint
25755 that returns VSX_REGS if TImode is allowed in VSX registers.
25757 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
25758 constant under VSX.
25760 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
25761 similar to TImode, but it is restricted to being in the GPRs.
25763 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
25764 TImode to occupy a single VSX register.
25766 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
25767 -mvsx-timode for power7/power8.
25768 (power7 cpu): Likewise.
25769 (power8 cpu): Likewise.
25771 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
25772 sure that TFmode/TDmode take up two registers if they are ever
25773 allowed in the upper VSX registers.
25774 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
25776 (rs6000_init_hard_regno_mode_ok): Likewise.
25777 (rs6000_debug_reg_global): Add debugging for PTImode and wt
25778 constraint. Print if LRA is turned on.
25779 (rs6000_option_override_internal): Give an error if -mvsx-timode
25780 and VSX is not enabled.
25781 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
25782 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
25783 to reg+offset addressing. Use PTImode when checking offset
25784 addresses for validity.
25785 (reg_offset_addressing_ok_p): Likewise.
25786 (rs6000_legitimate_offset_address_p): Likewise.
25787 (rs6000_legitimize_address): Likewise.
25788 (rs6000_legitimize_reload_address): Likewise.
25789 (rs6000_legitimate_address_p): Likewise.
25790 (rs6000_eliminate_indexed_memrefs): Likewise.
25791 (rs6000_emit_move): Likewise.
25792 (rs6000_secondary_reload): Likewise.
25793 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
25794 reloads to fpr registers to continue to use reg+offset addressing,
25795 but 64-bit reloads to altivec registers need reg+reg addressing.
25796 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
25797 it. Treat LO_SUM like a PLUS operation.
25798 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
25799 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
25800 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
25801 registers to share a register with a smaller sized type, since VSX
25802 puts scalars in the upper 64-bits.
25803 (print_operand): Add support for PTImode.
25804 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
25805 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
25806 registers, but don't have arithmetic support.
25807 (rs6000_memory_move_cost): Add test for VSX.
25808 (rs6000_opt_masks): Add -mvsx-timode.
25810 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
25813 (VSr): Use wt constraint for TImode.
25814 (VSv): Drop TImode support.
25815 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
25816 (vsx_movti_64bit): Likewise.
25817 (vsx_movti_32bit): Likewise.
25818 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
25819 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
25820 one '?' on the appropriate output constraint. Do not allow TImode
25821 logical operations on 32-bit systems.
25822 (vsx_ior<mode>3): Likewise.
25823 (vsx_xor<mode>3): Likewise.
25824 (vsx_one_cmpl<mode>2): Likewise.
25825 (vsx_nor<mode>3): Likewise.
25826 (vsx_andc<mode>3): Likewise.
25827 (vsx_concat_<mode>): Likewise.
25828 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
25830 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
25831 OPTION_MASK_VSX_TIMODE.
25832 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
25833 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
25835 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
25836 (TI2 iterator): New iterator for TImode, PTImode.
25837 (wd mode attribute): Add values for vector types.
25838 (movti_string): Replace TI move operations with operations for TImode
25839 and PTImode. Add support for TImode being allowed in VSX registers.
25840 (mov<mode>_string, TImode/PTImode): Likewise.
25841 (movti_ppc64): Likewise.
25842 (mov<mode>_ppc64, TImode/PTImode): Likewise.
25843 (TI mode splitters): Likewise.
25845 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
25848 2013-03-20 Marc Glisse <marc.glisse@inria.fr>
25850 PR tree-optimization/56355
25851 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
25852 Also handle integers with undefined overflow.
25854 2013-03-20 Catherine Moore <clm@codesourcery.com>
25855 Maciej W. Rozycki <macro@codesourcery.com>
25856 Tom de Vries <tom@codesourcery.com>
25857 Nathan Sidwell <nathan@codesourcery.com>
25858 Iain Sandoe <iain@codesourcery.com>
25859 Nathan Froyd <froydnj@codesourcery.com>
25860 Chao-ying Fu <fu@mips.com>
25862 * doc/extend.texi: (micromips, nomicromips, nocompression):
25863 Document new function attributes.
25864 * doc/invoke.texi (minterlink-compressed, mmicromips,
25865 m14k, m14ke, m14kec): Document new options.
25866 (minterlink-mips16): Update documentation.
25867 * doc/md.texi (ZC, ZD): Document new constraints.
25868 * configure.ac (gcc_cv_as_micromips): Check if linker
25869 supports the .set micromips directive.
25870 * configure: Regenerate.
25871 * config.in: Regenerate.
25872 * config/mips/mips-tables.opt: Regenerate.
25873 * config/mips/micromips.md: New file.
25874 * constraints.md (ZC, ZD): New constraints.
25875 * config/mips/predicates.md (movep_src_register): New predicate.
25876 (movep_src_operand): New predicate.
25877 (non_volatile_mem_operand): New predicate.
25878 * config/mips/mips.md (multimem): New type.
25879 (length): Differentiate between 17-bit and 18-bit branch offsets.
25880 (MOVEP1, MOVEP2): New mode iterator.
25881 (mov_<load>l): Use ZC constraint.
25882 (mov_<load>r): Likewise.
25883 (mov_<store>l): Likewise.
25884 (mov_<store>r): Likewise.
25885 (*branch_equality<mode>_inverted): Add microMIPS support.
25886 (*branch_equality<mode>): Likewise.
25887 (*jump_absolute): Likewise.
25888 (indirect_jump_<mode>): Likewise.
25889 (tablejump_<mode>): Likewise.
25890 (<optab>_internal): Likewise.
25891 (sibcall_internal): Likewise.
25892 (sibcall_value_internal): Likewise.
25893 (prefetch): Use constraint ZD.
25894 * config/mips/mips.opt (minterlink-compressed): New option.
25895 (minterlink-mips16): Now an alias for minterlink-compressed.
25896 (mmicromips): New option.
25897 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
25898 (compare_and_swap_12): Likewise.
25899 (sync_add<mode>): Likewise.
25900 (sync_<optab>_12): Likewise.
25901 (sync_old_<optab>_12): Likewise.
25902 (sync_new_<optab>_12): Likewise.
25903 (sync_nand_12): Likewise.
25904 (sync_old_nand_12): Likewise.
25905 (sync_new_nand_12): Likewise.
25906 (sync_sub<mode>): Likewise.
25907 (sync_old_add<mode>): Likewise.
25908 (sync_old_sub<mode>): Likewise.
25909 (sync_new_add<mode>): Likewise.
25910 (sync_new_sub<mode>): Likewise.
25911 (sync_<optab><mode>): Likewise.
25912 (sync_old_<optab><mode>): Likewise.
25913 (sync_new_<optab><mode>): Likewise.
25914 (sync_nand<mode>): Likewise.
25915 (sync_old_nand<mode>): Likewise.
25916 (sync_new_nand<mode>): Likewise.
25917 (sync_lock_test_and_set<mode>): Likewise.
25918 (test_and_set_12): Likewise.
25919 (atomic_compare_and_swap<mode>): Likewise.
25920 (atomic_exchange<mode>_llsc): Likewise.
25921 (atomic_fetch_add<mode>_llsc): Likewise.
25922 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
25923 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
25924 (umips_save_restore_pattern_p): Likewise.
25925 (umips_load_store_pair_p): Likewise.
25926 (umips_output_load_store_pair): Likewise.
25927 (umips_movep_target_p): Likewise.
25928 (umips_12bit_offset_address_p): Likewise.
25929 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
25930 (mips_base_mips16): Rename this...
25931 (mips_base_compression_flags): ...to this. Update all uses.
25932 (mips_attribute_table): Add micromips, nomicromips and nocompression.
25933 (mips_mips16_decl_p): Delete.
25934 (mips_nomips16_decl_p): Delete.
25935 (mips_get_compress_on_flags): New function.
25936 (mips_get_compress_off_flags): New function.
25937 (mips_get_compress_mode): New function.
25938 (mips_get_compress_on_name): New function.
25939 (mips_get_compress_off_name): New function.
25940 (mips_insert_attributes): Support multiple compression types.
25941 (mips_merge_decl_attributes): Likewise.
25942 (umips_12bit_offset_address_p): New function.
25943 (mips_start_function_definition): Emit .set micromips directive.
25944 (mips_call_may_need_jalx_p): New function.
25945 (mips_function_ok_for_sibcall): Add microMIPS support.
25946 (mips_print_operand_punctuation): Support short delay slots and
25948 (umips_swm_mask, umips_swm_encoding): New.
25949 (umips_build_save_restore): New function.
25950 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
25951 (was_mips16_p): Remove.
25952 (old_compression_mode): New.
25953 (mips_set_compression_mode): New function.
25954 (mips_set_current_function): Add microMIPS support.
25955 (mips_option_override): Likewise.
25956 (umips_save_restore_pattern_p): New function.
25957 (umips_output_save_restore): New function.
25958 (umips_load_store_pair_p_1): New function.
25959 (umips_load_store_pair_p): New function.
25960 (umips_output_load_store_pair_1): New function.
25961 (umips_output_load_store_pair): New function.
25962 (umips_movep_target_p) New function.
25963 (mips_prepare_pch_save): Add microMIPS support.
25964 * config/mips/mips.h (TARGET_COMPRESSION): New.
25965 (TARGET_CPU_CPP_BUILTINS): Update macro
25966 to use new compression flags and to support microMIPS.
25967 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
25968 (MIPS_ARCH_FLOAT_SPEC): Likewise.
25969 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
25970 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
25971 (ASM_SPEC): Support mmicromips and mno-micromips.
25972 (M16STORE_REG_P): New macro.
25973 (MIPS_CALL): Support TARGET_MICROMIPS.
25974 (MICROMIPS_J): New macro.
25975 (mips_base_mips16): Rename this...
25976 (mips_base_compression_flags): ...to this.
25977 (UMIPS_12BIT_OFFSET_P): New macro.
25978 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
25979 (MULTILIB_DIRNAMES): Likewise.
25980 2013-03-20 Richard Biener <rguenther@suse.de>
25982 PR tree-optimization/56661
25983 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
25984 the result does not have to be distinct.
25986 2013-03-20 Richard Biener <rguenther@suse.de>
25988 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
25991 2013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
25992 Steven Bosscher <steven@gcc.gnu.org>
25994 PR rtl-optimization/56605
25995 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
25997 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
26000 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
26001 that require movd instead of movq.
26003 2013-03-20 Richard Biener <rguenther@suse.de>
26005 * tree-ssa-structalias.c (struct variable_info): Add pointer
26006 to the first field of an aggregate with sub-vars. Make
26007 this and the pointer to the next subfield its ID.
26008 (vi_next): New function.
26009 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
26010 storedanything_id, integer_id): Increment by one.
26011 (new_var_info, get_call_vi, lookup_call_clobber_vi,
26012 get_call_clobber_vi): Adjust.
26013 (solution_set_expand): Simplify and speedup.
26014 (solution_set_add): Inline into ...
26015 (set_union_with_increment): ... this. Adjust accordingly.
26016 (do_sd_constraint): Likewise.
26017 (do_ds_constraint): Likewise.
26018 (do_complex_constraint): Simplify.
26019 (build_pred_graph): Adjust.
26020 (solve_graph): Likewise. Simplify and speedup.
26021 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
26022 get_constraint_for_component_ref, get_constraint_for_1,
26023 first_vi_for_offset, first_or_preceding_vi_for_offset,
26024 create_function_info_for, create_variable_info_for_1,
26025 create_variable_info_for, intra_create_variable_infos): Adjust.
26026 (init_base_vars): Push NULL for ID zero.
26027 (compute_points_to_sets): Adjust.
26029 2013-03-20 Richard Biener <rguenther@suse.de>
26031 * cfgloop.c (verify_loop_structure): Streamline and avoid
26032 ICEing on corrupt loop tree.
26033 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
26036 2013-03-20 Richard Biener <rguenther@suse.de>
26038 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
26039 check whether an SSA update is needed.
26041 2013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
26043 * config/mips/constraints.md (T): Rename to...
26047 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
26048 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
26050 2013-03-19 Ian Bolton <ian.bolton@arm.com>
26052 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
26053 (*subsi3_carryin_uxtw): Likewise.
26055 2013-03-19 Ian Bolton <ian.bolton@arm.com>
26057 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
26058 (*rorsi3_insn_uxtw): Likewise.
26060 2013-03-19 Ian Bolton <ian.bolton@arm.com>
26062 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
26063 (*extrsi5_insn_uxtw): Likewise.
26065 2013-03-19 Richard Biener <rguenther@suse.de>
26067 PR tree-optimization/56273
26068 * passes.c (init_optimization_passes): Move second VRP after DOM.
26070 2013-03-19 Uros Bizjak <ubizjak@gmail.com>
26072 * config/i386/i386.md (*movti_internal): Merge from
26073 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
26074 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
26075 nox64 isa attributes.
26077 2013-03-18 Richard Biener <rguenther@suse.de>
26079 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
26081 (merge_node_constraints): Likewise.
26082 (build_succ_graph): Likewise.
26083 (valid_graph_edge): Inline into single caller.
26084 (unify_nodes): Likewise. Use bitmap_set_bit return value
26086 (scc_visit): Fix formatting and variable use.
26087 (do_sd_constraint): Use gcc_checking_assert.
26088 (do_ds_constraint): Likewise.
26089 (do_complex_constraint): Likewise.
26090 (condense_visit): Likewise. Cleanup.
26091 (dump_pred_graph): New function.
26092 (perform_var_substitution): Dump the pred-graph before
26093 variable substitution.
26094 (find_equivalent_node): Use gcc_checking_assert.
26095 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
26097 2013-03-18 Richard Biener <rguenther@suse.de>
26099 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
26100 Remove cond_expr_stmt_list argument and do not gimplify the
26102 (vect_loop_versioning): Adjust.
26103 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
26104 Cleanup to use less temporaries.
26105 (vect_create_data_ref_ptr): Cleanup.
26107 2013-03-18 Jakub Jelinek <jakub@redhat.com>
26109 PR tree-optimization/56635
26110 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
26111 require types_compatible_p types.
26113 2013-03-18 Nick Clifton <nickc@redhat.com>
26115 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
26116 spurious backslash.
26118 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
26119 Add missing line to comment describing function.
26121 2013-03-18 Richard Biener <rguenther@suse.de>
26123 PR tree-optimization/56210
26124 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
26125 Handle string / character search functions.
26126 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
26128 2013-03-18 Richard Biener <rguenther@suse.de>
26130 PR middle-end/56483
26131 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
26132 and implement properly.
26133 * gimple.h (gimple_cond_single_var_p): Remove.
26135 2013-03-18 Richard Biener <rguenther@suse.de>
26137 * tree-data-ref.h (find_data_references_in_loop): Declare.
26138 * tree-data-ref.c (get_references_in_stmt): Use a stack
26139 vector pre-allocated in the callers.
26140 (find_data_references_in_stmt): Adjust.
26141 (graphite_find_data_references_in_stmt): Likewise.
26142 (create_rdg_vertices): Likewise.
26143 (find_data_references_in_loop): Export.
26144 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
26145 Compute dependences here...
26146 (vect_analyze_data_refs): ...not here. When we encounter
26147 a non-vectorizable data reference in basic-block vectorization
26148 truncate the data reference vector. Do not bother to
26149 fixup data-dependence information for gather loads.
26150 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
26151 of data references, as reported.
26153 2013-03-18 Richard Biener <rguenther@suse.de>
26155 PR tree-optimization/3713
26156 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
26157 has_constants and expr.
26158 (stmt_has_constants): Properly valueize SSA names when deciding
26159 whether the stmt has constants.
26161 2013-03-18 Richard Biener <rguenther@suse.de>
26163 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
26164 whole function when there is nothing to do.
26165 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
26166 * tree-vectorizer.c (vectorize_loops): Update virtual and
26167 loop-closed SSA once.
26168 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
26170 2013-03-18 Richard Biener <rguenther@suse.de>
26172 PR middle-end/56113
26173 * domwalk.c (bb_postorder): New global static.
26174 (cmp_bb_postorder): New function.
26175 (walk_dominator_tree): Replace scheme imposing an order for
26176 visiting dominator sons by one sorting them at the time they
26177 are pushed on the stack.
26179 2013-03-18 Richard Biener <rguenther@suse.de>
26181 PR tree-optimization/39326
26182 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
26183 (struct mem_ref): Replace mem member with ao_ref typed member.
26184 (MEM_ANALYZABLE): Adjust.
26185 (memref_eq): Likewise.
26186 (mem_ref_alloc): Likewise.
26187 (gather_mem_refs_stmt): Likewise.
26188 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
26189 (execute_sm_if_changed_flag_set): Adjust.
26190 (execute_sm): Likewise.
26191 (ref_always_accessed_p): Likewise.
26192 (refs_independent_p): Likewise.
26193 (can_sm_ref_p): Likewise.
26195 2013-03-18 Jakub Jelinek <jakub@redhat.com>
26198 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
26199 return 1 even for !unsignedp.
26201 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
26203 * config/i386/i386.md (isa): Add x64 and nox64.
26204 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
26205 (*pushtf): Enable *roF alternative for x64 isa only.
26206 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
26207 mode attribute of integer alternatives to DImode for TARGET_64BIT.
26208 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
26209 (*movtf_internal): Merge from *movtf_internal_rex64 and
26210 *movtf_internal_sse. Use x64 and nox64 isa attributes.
26211 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
26212 nox64 isa attributes.
26213 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
26214 nox64 isa attributes.
26215 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
26217 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
26219 * config/alpha/alpha.c (TARGET_LRA_P): New define.
26221 2013-03-17 Jakub Jelinek <jakub@redhat.com>
26224 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
26225 class names. Remove trailing comma after "ALL_REGS".
26227 2013-03-16 Jan Hubicka <jh@suse.cz>
26229 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
26230 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
26231 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
26232 of cgraph_get_create_node.
26233 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
26235 2013-03-16 Jason Merrill <jason@redhat.com>
26238 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
26239 with DW_AT_default_value.
26241 2013-03-16 Jakub Jelinek <jakub@redhat.com>
26243 * BASE-VER: Set to 4.9.0.
26245 2013-03-14 Andi Kleen <ak@linux.intel.com>
26248 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
26249 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
26250 Document _x* TSX intrinsics.
26252 2013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
26253 David Holsgrove <david.holsgrove@xilinx.com>
26255 * configure.ac: Add MicroBlaze TLS support detection.
26256 * configure: Regenerate.
26257 * config/microblaze/microblaze-protos.h
26258 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
26259 symbol_mentioned_p, label_mentioned_p): Add prototypes.
26260 * config/microblaze/microblaze.c (microblaze_address_type): Add
26261 ADDRESS_TLS and tls_reloc address types.
26262 (microblaze_address_info): Add tls_reloc.
26263 (TARGET_HAVE_TLS): Define.
26264 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
26265 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
26266 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
26267 load_tls_operand, microblaze_call_tls_get_addr,
26268 microblaze_legitimize_tls_address): New functions.
26269 (microblaze_classify_unspec): Handle UNSPEC_TLS.
26270 (get_base_reg): Use microblaze_tls_symbol_p.
26271 (microblaze_classify_address): Handle TLS.
26272 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
26273 label_mentioned_p and microblaze_tls_referenced_p.
26274 (microblaze_legitimize_address): Handle TLS.
26275 (microblaze_address_insns): Handle ADDRESS_TLS.
26276 (pic_address_needs_scratch): Handle TLS.
26277 (print_operand_address): Handle TLS.
26278 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
26279 (microblaze_expand_move): Handle TLS.
26280 (microblaze_legitimate_constant_p): Check
26281 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
26282 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
26283 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
26284 (PIC_OFFSET_TABLE_REGNUM): Set.
26285 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
26286 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
26287 (addsi3, movsi_internal2, movdf_internal): Update constraints
26288 * config/microblaze/predicates.md (arith_plus_operand): Define
26289 (move_operand): Redefine as move_src_operand,
26290 check microblaze_tls_referenced_p.
26292 2013-03-14 Ian Bolton <ian.bolton@arm.com>
26294 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
26295 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
26297 2013-03-14 Ian Bolton <ian.bolton@arm.com>
26299 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
26302 2013-03-14 Jakub Jelinek <jakub@redhat.com>
26304 PR tree-optimization/53265
26305 * common.opt (Waggressive-loop-optimizations): New option.
26306 * tree-ssa-loop-niter.c: Include tree-pass.h.
26307 (do_warn_aggressive_loop_optimizations): New function.
26308 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
26309 if number_of_latch_executions returned constant.
26310 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
26311 early. If number_of_latch_executions returned constant, set
26312 nb_iterations_upper_bound back to it.
26313 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
26315 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
26316 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
26318 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
26319 (MULTILIB_OSDIRNAMES): Set.
26320 * genmultilib: If defaultosdirname doesn't start with :: , set
26321 defaultosdirname2 instead, clear it and emit two . multilib_raw
26322 entries instead of just one.
26324 2013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
26326 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
26327 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
26328 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
26329 (SUBTARGET_OVERRIDE_OPTIONS): New.
26331 2013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
26334 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
26335 (musermode): Convert to Var(TARGET_USERMODE).
26336 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
26337 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
26338 * config/sh/sh.c (sh_option_override): Use
26339 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
26340 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
26342 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
26344 (udivsi3_i4_single, divsi3_i4_single): Use
26345 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
26347 2013-03-13 Dave Korn <dave.korn.cygwin@gmail.com>
26349 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
26352 2013-03-13 Richard Biener <rguenther@suse.de>
26354 PR tree-optimization/56608
26355 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
26356 calls when vectorizing basic-blocks.
26358 2013-03-13 Jakub Jelinek <jakub@redhat.com>
26361 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
26364 2013-03-12 Jakub Jelinek <jakub@redhat.com>
26366 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
26368 2013-03-11 Jan Hubicka <jh@suse.cz>
26371 * lto-streamer-out.c (output_symbol_p): Skip references from
26372 constructors of external variables.
26374 2013-03-11 Jan Hubicka <jh@suse.cz>
26376 PR middle-end/56571
26377 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
26379 * emit-rtl.c (verify_rtx_sharing): Likewise.
26380 (copy_insn_1): Likewise.
26381 * rtl.c (copy_rtx): Likewise.
26383 2013-03-11 Georg-Johann Lay <avr@gjlay.de>
26386 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
26387 output_operand_lossage message.
26389 2013-03-11 Richard Earnshaw <rearnsha@arm.com>
26392 * arm.c (shift_op): Validate RTL pattern on the fly.
26393 (arm_print_operand, case 'S'): Don't use shift_operator to validate
26396 2013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26399 * config/pa/pa.md (call_value): Check for calls to powf and direct to
26400 new call patterns that clobber %fr12.
26401 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
26402 split and postreload patterns.
26403 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
26404 registers %fr12 and %fr12R as call used.
26406 2013-03-09 Steven Bosscher <steven@gcc.gnu.org>
26408 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
26409 (canon_address, record_store, replace_read, check_mem_read_rtx,
26410 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
26411 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
26412 rest_of_handle_dse): Likewise.
26414 2013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
26416 PR middle-end/56524
26417 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
26419 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
26420 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
26421 (save_optabs_if_changed): Replace with...
26422 (init_tree_optimization_optabs): ...this.
26423 * optabs.c (save_optabs_if_changed): Rename to...
26424 (init_tree_optimization_optabs): ...this. Take the optimization node
26425 as argument. Do nothing if the base optabs are already correct.
26426 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
26427 to recompute optabs.
26428 * function.h (function): Remove optabs field.
26429 * function.c (invoke_set_current_function_hook): Call
26430 init_tree_optimization_optabs. Use the result to initialize
26433 2013-02-27 Aldy Hernandez <aldyh@redhat.com>
26435 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
26436 if GTMA_HAS_NO_INSTRUMENTATION.
26437 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
26438 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
26439 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
26440 * gimple-pretty-print.c (dump_gimple_transaction): Handle
26441 GTMA_HAS_NO_INSTRUMENTATION.
26443 2013-03-08 Jakub Jelinek <jakub@redhat.com>
26445 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
26448 2013-03-08 Marek Polacek <polacek@redhat.com>
26449 Jakub Jelinek <jakub@redhat.com>
26451 PR tree-optimization/56478
26452 * predict.c (is_comparison_with_loop_invariant_p): Change the
26453 type of loop_step to tree.
26454 (predict_loops): Adjust.
26455 (predict_iv_comparison): Perform the computations on double_ints.
26457 2013-03-08 Richard Biener <rguenther@suse.de>
26459 PR tree-optimization/56570
26460 * tree-cfg.c (verify_expr_location_1): Verify locations for
26462 * tree-sra.c (create_access_replacement): Strip locations
26463 from DECL_DEBUG_EXPRs.
26465 2013-03-08 Richard Biener <rguenther@suse.de>
26467 * tree-inline.c (expand_call_inline): Do not associate
26468 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
26469 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
26471 2013-03-08 Richard Biener <rguenther@suse.de>
26473 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
26474 or block changes with -Og. Fix for location / block encoding
26475 changes and PHI arguments with locations.
26477 2013-03-07 Steven Bosscher <steven@gcc.gnu.org>
26479 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
26481 (struct output_info): Likewise.
26482 (register_overhead): Remove bad gcc_assert.
26483 (bitmap_find_bit): If there is only a single bitmap element, do not
26484 count a miss as a search.
26485 (print_statistics): Update for counter type changes.
26486 (dump_bitmap_statistics): Likewise. Print headers such that they
26487 are properly lined up with the printed counters.
26489 2013-03-07 Jakub Jelinek <jakub@redhat.com>
26491 PR tree-optimization/56559
26492 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
26493 check that it has only a single use.
26495 2013-03-07 Richard Biener <rguenther@suse.de>
26497 * doc/invoke.texi (fwhole-program): Discourage use in combination
26500 2013-03-06 Jakub Jelinek <jakub@redhat.com>
26502 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
26504 PR tree-optimization/56539
26505 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
26506 instead of GSI_CONTINUE_LINKING as last argument to
26507 force_gimple_operand_gsi. Adjust function comment.
26509 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
26512 PR middle-end/56548
26513 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
26514 promoted mode, convert the result back to the original mode.
26516 2013-03-06 Richard Biener <rguenther@suse.de>
26518 PR middle-end/56294
26519 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
26520 (insert_updated_phi_nodes_compare_uids): New function.
26521 (update_ssa): Sort symbols_to_rename after UID before
26522 traversing it to insert PHI nodes.
26524 2013-03-06 Richard Biener <rguenther@suse.de>
26526 PR middle-end/50494
26527 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
26528 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
26531 2013-02-13 Richard Biener <rguenther@suse.de>
26534 * varasm.c (output_constant_def_1): Get the decl representing
26535 the constant as argument.
26536 (output_constant_def): Wrap output_constant_def_1.
26537 (make_decl_rtl): Use output_constant_def_1 with the decl
26538 representing the constant.
26539 (build_constant_desc): Optionally re-use a decl already
26540 representing the constant.
26541 (tree_output_constant_def): Adjust.
26543 2013-03-06 Joey Ye <joey.ye@arm.com>
26546 * gcc.c (convert_white_space): New function.
26547 (main): Handles white space in function name.
26549 2013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
26552 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
26553 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
26554 to SH_DIV_CALL_TABLE for TARGET_SH2.
26555 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
26557 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
26558 call-table options.
26560 2013-03-05 Sterling Augustine <saugustine@google.com>
26561 Cary Coutant <ccoutant@google.com>
26564 * dwarf2out.c (resolve_addr): Don't call
26565 remove_loc_list_addr_table_entries a second time for the same
26568 2013-03-05 Jakub Jelinek <jakub@redhat.com>
26571 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
26572 (avoid_complex_debug_insns): New function.
26573 (expand_debug_locations): Call it.
26575 PR rtl-optimization/56484
26576 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
26577 lifetimes of hard registers on small register class machines.
26579 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
26581 * config/microblaze/microblaze-protos.h: Rename
26582 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
26583 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
26585 (microblaze_fast_interrupt_function_p): New function.
26586 (microblaze_is_interrupt_handler): Rename to
26587 microblaze_is_interrupt_variant and add fast_interrupt check.
26588 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
26589 (save_restore_insns): Likewise.
26590 (compute_frame_size): Likewise.
26591 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
26592 (microblaze_globalize_label): Likewise.
26593 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
26594 * config/microblaze/microblaze.md: Use wrapper
26595 microblaze_is_interrupt_variant.
26597 2013-03-05 Kai Tietz <ktietz@redhat.com>
26599 * sdbout.c (sdbout_one_type): Switch to current function's section
26600 supporting cold/hot.
26602 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
26604 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
26607 2013-03-05 Jakub Jelinek <jakub@redhat.com>
26609 PR middle-end/56461
26610 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
26611 if VALGRIND_GET_VBITS is defined, temporarily make object
26612 memory all defined, and restore previous valgrind addressability
26613 and definability afterwards. Free this_object at the end.
26615 PR middle-end/56461
26616 * lra.c (lra): Call lra_clear_live_ranges if live_p,
26617 right before calling lra_create_live_ranges, also call it
26618 when clearing live_p. Only call lra_clear_live_ranges
26619 at the end if live_p.
26621 PR middle-end/56461
26622 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
26624 2013-03-05 Richard Biener <rguenther@suse.de>
26626 PR tree-optimization/56521
26627 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
26630 2013-03-05 Steven Bosscher <steven@gcc.gnu.org>
26633 * except.h (remove_unreachable_eh_regions): New prototype.
26634 * except.c (remove_eh_handler_splicer): New function, split out
26635 of remove_eh_handler.
26636 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
26637 warning about running it on many EH regions one at a time.
26638 (remove_unreachable_eh_regions_worker): New function, walk the
26639 EH tree in depth-first order and remove non-marked regions.
26640 (remove_unreachable_eh_regions): New function.
26641 * tree-eh.c (mark_reachable_handlers): New function, split out
26642 from remove_unreachable_handlers.
26643 (remove_unreachable_handlers): Use mark_reachable_handlers and
26644 remove_unreachable_eh_regions.
26645 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
26646 and remove_unreachable_eh_regions.
26648 2013-03-05 Richard Biener <rguenther@suse.de>
26650 PR middle-end/56525
26651 * loop-init.c (fix_loop_structure): Remove loops in two stages,
26652 not freeing them until the end.
26654 2013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
26656 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
26658 2013-03-05 Richard Biener <rguenther@suse.de>
26660 PR tree-optimization/56270
26661 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
26662 of loads after scheduling an SLP instance.
26664 2013-03-05 Jakub Jelinek <jakub@redhat.com>
26666 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
26668 (check_gcc_parallelize): Run guality.exp as a separate job from
26669 vect.exp with unsorted.exp and $(dg_target_exps) separately from
26670 struct-layout-1.exp with stackalign.exp.
26672 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
26674 PR middle-end/56461
26675 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
26676 load_index sbitmap even if some bit in it isn't set.
26678 PR middle-end/56461
26679 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
26680 (discover_iteration_bound_by_body_walk): Change queues to
26681 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
26682 spelling in comment. Call safe_push on queues[bound_index] directly.
26683 Release queues[queue_index] in every iteration unconditionally.
26684 Release bounds vector.
26686 PR middle-end/56461
26687 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
26688 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
26689 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
26692 2013-03-05 Richard Biener <rguenther@suse.de>
26695 * tree-inline.c (remap_blocks_to_null): New function.
26696 (expand_call_inline): When expanding a call stmt without
26697 an associated block inline remap all callee blocks to NULL.
26699 2013-03-05 Jakub Jelinek <jakub@redhat.com>
26701 PR rtl-optimization/56494
26702 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
26703 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
26704 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
26706 PR middle-end/56461
26707 * sel-sched-ir.c (free_sched_pools): Release
26708 succs_info_pool.stack[succs_info_pool.max_top] vectors too
26709 if succs_info_pool.max_top isn't -1.
26712 * opts.c (opts_obstack, opts_concat): Moved to...
26713 * opts-common.c (opts_obstack, opts_concat): ... here.
26715 2013-03-04 Jakub Jelinek <jakub@redhat.com>
26717 PR middle-end/56461
26718 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
26720 2013-03-04 Martin Jambor <mjambor@suse.cz>
26722 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
26723 all appropriate places.
26725 2013-01-04 Eric Botcazou <ebotcazou@adacore.com>
26727 PR tree-optimization/56424
26728 * ipa-split.c (split_function): Do not set the RSO flag if result is
26729 not by reference and its type is a register type.
26731 2013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
26733 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
26734 (microblaze_legitimate_pic_operand): Likewise
26735 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
26736 new function microblaze_legitimate_pic_operand
26737 * config/microblaze/microblaze-protos.h
26738 (microblaze_legitimate_pic_operand): Declare.
26740 2013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
26742 * config/microblaze/predicates.md (call_insn_simple_operand):
26743 New predicate for supported rtx code types.
26744 * config/microblaze/microblaze.md (call_internal1): Use
26745 call_insn_simple_operand predicate.
26747 2013-03-04 Jakub Jelinek <jakub@redhat.com>
26749 PR middle-end/56461
26750 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
26751 partitions.ordered_remove.
26753 PR middle-end/56461
26754 * tree-vect-stmts.c (vectorizable_conversion): Don't call
26755 vec_oprnds0.create (1) for modifier == NONE.
26757 PR middle-end/56461
26758 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
26759 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
26760 vec_oprnds1 right before pushing anything to it for
26763 PR middle-end/56461
26764 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
26765 set nbbs to 0 instead of having separate code path.
26766 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
26767 instead of false as last argument if returning NULL.
26769 2013-03-03 Sandra Loosemore <sandra@codesourcery.com>
26771 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
26772 the attribute is now called "target" instead of "option".
26773 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
26774 * doc/tm.texi.in (Target Attributes): Likewise document the correct
26775 attribute/pragma name for TARGET_OPTION_VALID_P and
26776 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
26777 * doc/tm.texi: Regenerated.
26779 2013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
26781 * config/microblaze/microblaze.c:
26782 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
26783 * config/microblaze/microblaze.h: Add -mxl-reorder to
26785 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
26786 instructions emitted if TARGET_REORDER.
26787 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
26788 or 0 for -m/-mno case, but initialises as 2 to detect default use case
26791 2013-03-01 Xinliang David Li <davidxl@google.com>
26793 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
26796 2013-03-01 Jakub Jelinek <jakub@redhat.com>
26798 PR middle-end/56461
26799 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
26800 vector even when returning true. Fix up function comment formatting.
26802 PR middle-end/56461
26803 * ira-build.c (ira_loop_nodes_count): New variable.
26804 (create_loop_tree_nodes): Initialize it.
26805 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
26807 PR middle-end/56461
26808 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
26809 method on dr_chain and result_chain.
26810 * tree-vect-stmts.c (vectorizable_store): Only call
26811 result_chain.create if j == 0.
26813 PR middle-end/56461
26814 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
26815 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
26816 before overwriting it.
26818 2013-03-01 Tobias Burnus <burnus@net-b.de>
26820 * doc/extended.texi (C Extensions): Change order in @menu
26822 (Other MIPS Built-in Functions): Move last MIPS entry before
26823 "picoChip Built-in Functions".
26824 (SH Built-in Functions): Move after RX Built-in Functions.
26825 * doc/gcc.texi (Introduction): Change order in @menu
26827 * doc/md.texi (Constraints): Ditto.
26828 * gty.texi (Type Information): Ditto.
26829 (User-provided marking routines for template types): Make
26831 * doc/invoke.texi (AArch64 Options): Move before
26832 "Adapteva Epiphany Options".
26834 2013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
26835 Jakub Jelinek <jakub@redhat.com>
26838 * asan.c (gate_asan): Lookup no_sanitize_address instead of
26839 no_address_safety_analysis attribute.
26840 * doc/extend.texi (no_address_safety_attribute): Rename to
26841 no_sanitize_address attribute, mention no_address_safety_analysis
26842 attribute as deprecated alias.
26844 2013-02-28 Jakub Jelinek <jakub@redhat.com>
26846 PR middle-end/56461
26847 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
26848 type to vec<vec<tree> > *.
26849 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
26850 to be vec<tree> instead of vec<tree> *, set vec_defs
26851 to vNULL and call vec_defs.create (number_of_vects), adjust other
26853 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
26854 vectorizable_condition): Adjust vect_get_slp_defs callers.
26856 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
26858 * config/aarch64/aarch64.c
26859 (aarch64_float_const_representable): Remove unused variable.
26861 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
26863 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
26865 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
26867 * config/aarch64/aarch64-builtins.c
26868 (aarch64_init_simd_builtins): Make static.
26870 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
26872 * config/aarch64/aarch64.c
26873 (aarch64_simd_make_constant): Make static.
26875 2013-02-28 Martin Jambor <mjambor@suse.cz>
26877 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
26878 with no initialization to the RHS of debug statements.
26880 2013-02-28 Martin Jambor <mjambor@suse.cz>
26882 PR tree-optimization/56294
26883 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
26885 (get_access_replacement): Do not call create_access_replacement.
26886 Assert a replacement exists.
26887 (get_repl_default_def_ssa_name): Create the replacement declaration
26890 2013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
26892 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
26893 final_end_function.
26895 2013-02-28 Marek Polacek <polacek@redhat.com>
26897 PR rtl-optimization/56466
26898 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
26899 if we're changing a loop.
26900 (peel_loops_completely): Likewise.
26902 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
26905 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
26907 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
26910 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
26911 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
26912 INTX_FTYPE_FX, FX_FTYPE_INTX.
26913 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
26915 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
26917 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
26918 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
26919 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
26920 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
26921 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
26922 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
26923 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
26924 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
26925 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
26926 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
26927 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
26928 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
26929 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
26930 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
26931 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
26932 (avrxmega6): Increase max flash segments from 5 to 6.
26933 * config/avr/t-multilib: Regenerate.
26934 * config/avr/avr-tables.opt: Regenerate.
26935 * doc/avr-mmcu.texi: Regenerate.
26937 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
26939 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
26940 (avr_device_to_arch): Rename to avr_device_to_ld.
26941 (avr_device_to_as): New prototype.
26942 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
26943 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
26944 * config/avr/driver-avr.c (avr_device_to_as): New.
26945 (avr_device_to_arch): Rename to avr_device_to_ld.
26947 2013-02-27 Jakub Jelinek <jakub@redhat.com>
26949 PR middle-end/56461
26950 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
26951 method on dr_chain and result_chain.
26953 PR middle-end/56461
26954 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
26955 pointer_set_destroy on not_executed_last_iteration.
26957 PR middle-end/56461
26958 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
26960 PR middle-end/56461
26961 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
26962 FOR_EACH_DEFINED_FUNCTION when freeing state.
26964 PR middle-end/56461
26965 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
26967 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
26970 PR middle-end/56461
26971 * ipa-cp.c (decide_whether_version_node): Call vec_free on
26972 known_aggs[i].items and release known_aggs vector.
26974 PR middle-end/56461
26975 * ipa-reference.c (propagate): Free node_info even for alias nodes.
26977 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
26979 * config/microblaze/microblaze.c (microblaze_emit_compare):
26980 Use xor for EQ/NE comparisions.
26981 * config/microblaze/microblaze.md (cstoresf4): Add constraints
26982 (cbranchsf4): Adjust operator to comparison_operator.
26984 2013-02-27 Jakub Jelinek <jakub@redhat.com>
26986 PR middle-end/56461
26987 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
26989 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
26990 vec_safe_push, always update *slot.
26991 (redirect_edge_var_map_clear): Use vec_free.
26992 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
26993 (free_var_map_entry): Use vec_free.
26994 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
26995 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
26997 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
26999 PR middle-end/45472
27000 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
27001 when the may_trap_p bit of the exprs being merged differs.
27002 Reorder tests for speculativeness in the logical and operator.
27004 2013-02-27 Jakub Jelinek <jakub@redhat.com>
27006 * incpath.c (add_standard_paths): Use reconcat instead of concat
27007 where appropriate and avoid leaking memory.
27009 * opts.h: Include obstack.h.
27010 (opts_concat): New prototype.
27011 (opts_obstack): New declaration.
27012 * opts.c (opts_concat): New function.
27013 (opts_obstack): New variable.
27014 (init_options_struct): Call gcc_init_obstack on opts_obstack.
27015 (finish_options): Use opts_concat instead of concat
27016 and XOBNEWVEC instead of XNEWVEC.
27017 * opts-common.c (generate_canonical_option, decode_cmdline_option,
27018 generate_option): Likewise.
27019 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
27020 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
27023 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
27024 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
27026 2013-02-26 Jakub Jelinek <jakub@redhat.com>
27028 PR middle-end/56461
27029 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
27031 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
27033 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
27034 (arm_block_move_unaligned_straight): Likewise.
27035 (arm_adjust_block_mem): Likewise.
27037 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
27040 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
27041 temp, cond and label.
27042 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
27045 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
27046 * config/c6x/c6x.h (dbx_register_map): Update declaration.
27049 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
27050 of prologue/epilogue functions.
27053 * config/tilegx/tilegx.c (tilegx_expand_prologue):
27054 Remove unused variable cfa_offset.
27055 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
27058 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
27059 type promotion to unsigned.
27062 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
27063 for HOST_WIDE_INT of 32 bit / same size as int.
27064 (arm_block_move_unaligned_straight): Likewise.
27065 (arm_adjust_block_mem): Likewise.
27068 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
27071 2013-02-26 Marek Polacek <polacek@redhat.com>
27073 PR tree-optimization/56426
27074 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
27076 2013-02-26 Richard Biener <rguenther@suse.de>
27079 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
27080 unused variable loops.
27082 2013-02-26 Jakub Jelinek <jakub@redhat.com>
27084 PR tree-optimization/56448
27085 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
27086 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
27087 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
27088 later operands of the references, or even first operand for
27089 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
27091 PR tree-optimization/56443
27092 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
27093 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
27094 to type_for_mode langhook.
27096 2013-02-25 Matt Turner <mattst88@gmail.com>
27098 * doc/invoke.texi: Document r4700.
27100 2013-02-25 Richard Biener <rguenther@suse.de>
27102 PR tree-optimization/56175
27103 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
27105 (simplify_bitwise_binary): ... here. Also guard the conversion
27106 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
27108 2013-02-25 Catherine Moore <clm@codesourcery.com>
27111 2013-02-24 Catherine Moore <clm@codesourcery.com>
27112 Maciej W. Rozycki <macro@codesourcery.com>
27113 Tom de Vries <tom@codesourcery.com>
27114 Nathan Sidwell <nathan@codesourcery.com>
27115 Iain Sandoe <iain@codesourcery.com>
27116 Nathan Froyd <froydnj@codesourcery.com>
27117 Chao-ying Fu <fu@mips.com>
27119 * doc/extend.texi: (micromips, nomicromips, nocompression):
27120 Document new function attributes.
27121 * doc/invoke.texi (minterlink-compressed, mmicromips,
27122 m14k, m14ke, m14kec): Document new options.
27123 (minterlink-mips16): Update documentation.
27124 * doc/md.texi (ZC, ZD): Document new constraints.
27125 * configure.ac (gcc_cv_as_micromips): Check if linker
27126 supports the .set micromips directive.
27127 * configure: Regenerate.
27128 * config.in: Regenerate.
27129 * config/mips/mips-tables.opt: Regenerate.
27130 * config/mips/micromips.md: New file.
27131 * constraints.md (ZC, AD): New constraints.
27132 * config/mips/predicates.md (movep_src_register): New predicate.
27133 (movep_src_operand): New predicate.
27134 (non_volatile_mem_operand): New predicate.
27135 * config/mips/mips.md (multimem): New type.
27136 (length): Differentiate between 17-bit and 18-bit branch offsets.
27137 (MOVEP1, MOVEP2): New mode iterator.
27138 (mov_<load>l): Use ZC constraint.
27139 (mov_<load>r): Likewise.
27140 (mov_<store>l): Likewise.
27141 (mov_<store>r): Likewise.
27142 (*branch_equality<mode>_inverted): Add microMIPS support.
27143 (*branch_equality<mode>): Likewise.
27144 (*jump_absolute): Likewise.
27145 (indirect_jump_<mode>): Likewise.
27146 (tablejump_<mode>): Likewise.
27147 (<optab>_internal): Likewise.
27148 (sibcall_internal): Likewise.
27149 (sibcall_value_internal): Likewise.
27150 (prefetch): Use constraint ZD.
27151 * config/mips/mips.opt (minterlink-compressed): New option.
27152 (minterlink-mips16): Now an alias for minterlink-compressed.
27153 (mmicromips): New option.
27154 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
27155 (compare_and_swap_12): Likewise.
27156 (sync_add<mode>): Likewise.
27157 (sync_<optab>_12): Likewise.
27158 (sync_old_<optab>_12): Likewise.
27159 (sync_new_<optab>_12): Likewise.
27160 (sync_nand_12): Likewise.
27161 (sync_old_nand_12): Likewise.
27162 (sync_new_nand_12): Likewise.
27163 (sync_sub<mode>): Likewise.
27164 (sync_old_add<mode>): Likewise.
27165 (sync_old_sub<mode>): Likewise.
27166 (sync_new_add<mode>): Likewise.
27167 (sync_new_sub<mode>): Likewise.
27168 (sync_<optab><mode>): Likewise.
27169 (sync_old_<optab><mode>): Likewise.
27170 (sync_new_<optab><mode>): Likewise.
27171 (sync_nand<mode>): Likewise.
27172 (sync_old_nand<mode>): Likewise.
27173 (sync_new_nand<mode>): Likewise.
27174 (sync_lock_test_and_set<mode>): Likewise.
27175 (test_and_set_12): Likewise.
27176 (atomic_compare_and_swap<mode>): Likewise.
27177 (atomic_exchange<mode>_llsc): Likewise.
27178 (atomic_fetch_add<mode>_llsc): Likewise.
27179 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
27180 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
27181 (umips_save_restore_pattern_p): Likewise.
27182 (umips_load_store_pair_p): Likewise.
27183 (umips_output_load_store_pair): Likewise.
27184 (umips_movep_target_p): Likewise.
27185 (umips_12bit_offset_address_p): Likewise.
27186 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
27187 (mips_base_mips16): Rename this...
27188 (mips_base_compression_flags): ...to this. Update all uses.
27189 (mips_attribute_table): Add micromips, nomicromips and nocompression.
27190 (mips_mips16_decl_p): Delete.
27191 (mips_nomips16_decl_p): Delete.
27192 (mips_get_compress_on_flags): New function.
27193 (mips_get_compress_off_flags): New function.
27194 (mips_get_compress_mode): New function.
27195 (mips_get_compress_on_name): New function.
27196 (mips_get_compress_off_name): New function.
27197 (mips_insert_attributes): Support multiple compression types.
27198 (mips_merge_decl_attributes): Likewise.
27199 (umips_12bit_offset_address_p): New function.
27200 (mips_start_function_definition): Emit .set micromips directive.
27201 (mips_call_may_need_jalx_p): New function.
27202 (mips_function_ok_for_sibcall): Add microMIPS support.
27203 (mips_print_operand_punctuation): Support short delay slots and
27205 (umips_swm_mask, umips_swm_encoding): New.
27206 (umips_build_save_restore): New function.
27207 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
27208 (was_mips16_p): Remove.
27209 (old_compression_mode): New.
27210 (mips_set_compression_mode): New function.
27211 (mips_set_current_function): Add microMIPS support.
27212 (mips_option_override): Likewise.
27213 (umips_save_restore_pattern_p): New function.
27214 (umips_output_save_restore): New function.
27215 (umips_load_store_pair_p_1): New function.
27216 (umips_load_store_pair_p): New function.
27217 (umips_output_load_store_pair_1): New function.
27218 (umips_output_load_store_pair): New function.
27219 (umips_movep_target_p) New function.
27220 (mips_prepare_pch_save): Add microMIPS support.
27221 * config/mips/mips.h (TARGET_COMPRESSION): New.
27222 (TARGET_CPU_CPP_BUILTINS): Update macro
27223 to use new compression flags and to support microMIPS.
27224 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
27225 (MIPS_ARCH_FLOAT_SPEC): Likewise.
27226 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
27227 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
27228 (ASM_SPEC): Support mmicromips and mno-micromips.
27229 (M16STORE_REG_P): New macro.
27230 (MIPS_CALL): Support TARGET_MICROMIPS.
27231 (MICROMIPS_J): New macro.
27232 (mips_base_mips16): Rename this...
27233 (mips_base_compression_flags): ...to this.
27234 (UMIPS_12BIT_OFFSET_P): New macro.
27235 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
27236 (MULTILIB_DIRNAMES): Likewise.
27238 2013-02-25 Tom de Vries <tom@codesourcery.com>
27240 PR rtl-optimization/56131
27241 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
27242 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
27243 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
27245 2013-02-25 Tobias Burnus <burnus@net-b.de>
27247 * doc/invoke.texi (-fsanitize=): Move from optimization
27248 to debugging options.
27250 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
27252 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
27254 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
27255 Alexander Monakov <amonakov@ispras.ru>
27257 PR middle-end/56077
27258 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
27259 flush pending lists also on non-jumps. Adjust comment.
27261 2013-02-24 Catherine Moore <clm@codesourcery.com>
27262 Maciej W. Rozycki <macro@codesourcery.com>
27263 Tom de Vries <tom@codesourcery.com>
27264 Nathan Sidwell <nathan@codesourcery.com>
27265 Iain Sandoe <iain@codesourcery.com>
27266 Nathan Froyd <froydnj@codesourcery.com>
27267 Chao-ying Fu <fu@mips.com>
27269 * doc/extend.texi: (micromips, nomicromips, nocompression):
27270 Document new function attributes.
27271 * doc/invoke.texi (minterlink-compressed, mmicromips,
27272 m14k, m14ke, m14kec): Document new options.
27273 (minterlink-mips16): Update documentation.
27274 * doc/md.texi (ZC, ZD): Document new constraints.
27275 * configure.ac (gcc_cv_as_micromips): Check if linker
27276 supports the .set micromips directive.
27277 * configure: Regenerate.
27278 * config.in: Regenerate.
27279 * config/mips/mips-tables.opt: Regenerate.
27280 * config/mips/micromips.md: New file.
27281 * constraints.md (ZC, AD): New constraints.
27282 * config/mips/predicates.md (movep_src_register): New predicate.
27283 (movep_src_operand): New predicate.
27284 (non_volatile_mem_operand): New predicate.
27285 * config/mips/mips.md (multimem): New type.
27286 (length): Differentiate between 17-bit and 18-bit branch offsets.
27287 (MOVEP1, MOVEP2): New mode iterator.
27288 (mov_<load>l): Use ZC constraint.
27289 (mov_<load>r): Likewise.
27290 (mov_<store>l): Likewise.
27291 (mov_<store>r): Likewise.
27292 (*branch_equality<mode>_inverted): Add microMIPS support.
27293 (*branch_equality<mode>): Likewise.
27294 (*jump_absolute): Likewise.
27295 (indirect_jump_<mode>): Likewise.
27296 (tablejump_<mode>): Likewise.
27297 (<optab>_internal): Likewise.
27298 (sibcall_internal): Likewise.
27299 (sibcall_value_internal): Likewise.
27300 (prefetch): Use constraint ZD.
27301 * config/mips/mips.opt (minterlink-compressed): New option.
27302 (minterlink-mips16): Now an alias for minterlink-compressed.
27303 (mmicromips): New option.
27304 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
27305 (compare_and_swap_12): Likewise.
27306 (sync_add<mode>): Likewise.
27307 (sync_<optab>_12): Likewise.
27308 (sync_old_<optab>_12): Likewise.
27309 (sync_new_<optab>_12): Likewise.
27310 (sync_nand_12): Likewise.
27311 (sync_old_nand_12): Likewise.
27312 (sync_new_nand_12): Likewise.
27313 (sync_sub<mode>): Likewise.
27314 (sync_old_add<mode>): Likewise.
27315 (sync_old_sub<mode>): Likewise.
27316 (sync_new_add<mode>): Likewise.
27317 (sync_new_sub<mode>): Likewise.
27318 (sync_<optab><mode>): Likewise.
27319 (sync_old_<optab><mode>): Likewise.
27320 (sync_new_<optab><mode>): Likewise.
27321 (sync_nand<mode>): Likewise.
27322 (sync_old_nand<mode>): Likewise.
27323 (sync_new_nand<mode>): Likewise.
27324 (sync_lock_test_and_set<mode>): Likewise.
27325 (test_and_set_12): Likewise.
27326 (atomic_compare_and_swap<mode>): Likewise.
27327 (atomic_exchange<mode>_llsc): Likewise.
27328 (atomic_fetch_add<mode>_llsc): Likewise.
27329 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
27330 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
27331 (umips_save_restore_pattern_p): Likewise.
27332 (umips_load_store_pair_p): Likewise.
27333 (umips_output_load_store_pair): Likewise.
27334 (umips_movep_target_p): Likewise.
27335 (umips_12bit_offset_address_p): Likewise.
27336 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
27337 (mips_base_mips16): Rename this...
27338 (mips_base_compression_flags): ...to this. Update all uses.
27339 (mips_attribute_table): Add micromips, nomicromips and nocompression.
27340 (mips_mips16_decl_p): Delete.
27341 (mips_nomips16_decl_p): Delete.
27342 (mips_get_compress_on_flags): New function.
27343 (mips_get_compress_off_flags): New function.
27344 (mips_get_compress_mode): New function.
27345 (mips_get_compress_on_name): New function.
27346 (mips_get_compress_off_name): New function.
27347 (mips_insert_attributes): Support multiple compression types.
27348 (mips_merge_decl_attributes): Likewise.
27349 (umips_12bit_offset_address_p): New function.
27350 (mips_start_function_definition): Emit .set micromips directive.
27351 (mips_call_may_need_jalx_p): New function.
27352 (mips_function_ok_for_sibcall): Add microMIPS support.
27353 (mips_print_operand_punctuation): Support short delay slots and
27355 (umips_swm_mask, umips_swm_encoding): New.
27356 (umips_build_save_restore): New function.
27357 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
27358 (was_mips16_p): Remove.
27359 (old_compression_mode): New.
27360 (mips_set_compression_mode): New function.
27361 (mips_set_current_function): Add microMIPS support.
27362 (mips_option_override): Likewise.
27363 (umips_save_restore_pattern_p): New function.
27364 (umips_output_save_restore): New function.
27365 (umips_load_store_pair_p_1): New function.
27366 (umips_load_store_pair_p): New function.
27367 (umips_output_load_store_pair_1): New function.
27368 (umips_output_load_store_pair): New function.
27369 (umips_movep_target_p) New function.
27370 (mips_prepare_pch_save): Add microMIPS support.
27371 * config/mips/mips.h (TARGET_COMPRESSION): New.
27372 (TARGET_CPU_CPP_BUILTINS): Update macro
27373 to use new compression flags and to support microMIPS.
27374 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
27375 (MIPS_ARCH_FLOAT_SPEC): Likewise.
27376 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
27377 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
27378 (ASM_SPEC): Support mmicromips and mno-micromips.
27379 (M16STORE_REG_P): New macro.
27380 (MIPS_CALL): Support TARGET_MICROMIPS.
27381 (MICROMIPS_J): New macro.
27382 (mips_base_mips16): Rename this...
27383 (mips_base_compression_flags): ...to this.
27384 (UMIPS_12BIT_OFFSET_P): New macro.
27385 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
27386 (MULTILIB_DIRNAMES): Likewise.
27388 2013-02-24 Jakub Jelinek <jakub@redhat.com>
27391 * target-globals.c (save_target_globals): For init_reg_sets and
27392 target_reinit remporarily set this_fn_optabs to this_target_optabs.
27394 2013-02-22 James Greenhalgh <james.greenhalgh@arm.com>
27396 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
27397 * config/aarch64/t-aarch64
27398 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
27400 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
27402 PR inline-asm/56148
27403 * lra-constraints.c (process_alt_operands): Reload operand
27404 conflicting with earlier clobber only if no more other conflicting
27407 2013-02-22 Jakub Jelinek <jakub@redhat.com>
27410 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
27411 if not linking a shared library.
27413 2013-02-22 Seth LaForge <sethml@google.com>
27415 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
27417 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
27419 * config/arm/arm.md (split for extendsidi): Update condition.
27420 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
27421 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
27422 (qhs_zextenddi_cstr): Likewise.
27424 2013-02-21 Jakub Jelinek <jakub@redhat.com>
27426 PR middle-end/56420
27427 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
27428 avoid signed wrapping.
27429 (expand_mult): Handle properly multiplication by
27430 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
27431 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
27432 in the compiler if coeff is HOST_WIDE_INT_MIN.
27433 (expand_divmod): Don't make ext_op1 static, change it's type to
27434 uhwi. Avoid undefined behavior in -INTVAL (op1).
27436 PR rtl-optimization/50339
27437 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
27439 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
27440 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
27441 into splitting_ashiftrt field.
27442 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
27444 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
27447 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
27449 PR middle-end/56108
27450 * trans-mem.c (execute_tm_mark): Do not expand transactions that
27451 are sure to go irrevocable.
27453 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
27455 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
27456 scalars are valid operands.
27458 2013-02-21 Martin Jambor <mjambor@suse.cz>
27460 PR tree-optimization/56310
27461 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
27462 only matching indices and non-negative final offsets.
27463 (intersect_aggregates_with_edge): Pass src_idx to
27464 agg_replacements_to_vector. Pass src_idx insstead of index to
27465 intersect_with_agg_replacements.
27467 2013-02-21 Martin Jambor <mjambor@suse.cz>
27469 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
27470 instead of hard-wired defaults.
27472 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
27474 * doc/invoke.texi (MIPS Options): Update documentation of the
27475 floating-point multiply-accumulate instruction restrictions.
27477 2013-02-21 Kostya Serebryany <kcc@google.com>
27479 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
27480 asan_shadow_offset on x86_64 linux.
27482 2013-02-21 Richard Biener <rguenther@suse.de>
27484 PR tree-optimization/56415
27486 2013-02-11 Richard Biener <rguenther@suse.de>
27488 PR tree-optimization/56273
27489 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
27492 2013-02-21 Jakub Jelinek <jakub@redhat.com>
27495 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
27498 PR inline-asm/56405
27499 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
27500 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
27502 2013-02-20 Jan Hubicka <jh@suse.cz>
27504 PR tree-optimization/56265
27505 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
27506 when target is referenced for first time.
27508 2013-02-20 Richard Biener <rguenther@suse.de>
27510 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
27511 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
27512 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
27513 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
27514 not return anything.
27515 (rename_ssa_copies): Do not remove unused locals.
27516 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
27517 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
27518 * passes.c (execute_function_todo): Do not schedule unused locals
27519 removal if cleanup_tree_cfg did something.
27520 * tree-ssa-live.c (remove_unused_locals): Dump statistics
27521 about the number of removed locals.
27523 2013-02-20 Richard Biener <rguenther@suse.de>
27525 PR tree-optimization/56398
27526 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
27528 2013-02-20 Martin Jambor <mjambor@suse.cz>
27530 PR tree-optimization/55334
27531 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
27532 restricted pointers to arrays.
27534 2013-02-20 Richard Biener <rguenther@suse.de>
27535 Jakub Jelinek <jakub@redhat.com>
27537 PR tree-optimization/56396
27538 * tree-ssa-ccp.c (n_const_val): New static variable.
27539 (get_value): Return NULL for SSA names we don't have a lattice
27541 (ccp_initialize): Initialize n_const_val.
27542 * tree-ssa-copy.c (n_copy_of): New static variable.
27543 (init_copy_prop): Initialize n_copy_of.
27544 (get_value): Return NULL_TREE for SSA names we don't have a
27547 2013-02-20 Martin Jambor <mjambor@suse.cz>
27549 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
27551 2013-02-20 Richard Biener <rguenther@suse.de>
27553 * genpreds.c (write_lookup_constraint): Do not compare first
27554 letter of the constraint again.
27556 2013-02-20 Richard Biener <rguenther@suse.de>
27558 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
27560 (get_use_iv_cost): Terminate hashtable walk when coming across
27563 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
27565 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
27566 reassociation for avx2 targets.
27568 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27570 * config/microblaze/microblaze.c: microblaze_has_clz = 0
27571 Add version check for v8.10.a to enable microblaze_has_clz
27572 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
27573 version and TARGET_PATTERN_COMPARE check
27574 * config/microblaze/microblaze.md: New clzsi2 instruction
27576 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27578 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
27579 function before branching.
27581 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
27583 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
27585 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
27587 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
27589 PR middle-end/55889
27590 * sel-sched.c: Include ira.h.
27591 (implicit_clobber_conflict_p): New function.
27592 (moveup_expr): Use it.
27593 * Makefile.in (sel-sched.o): Depend on ira.h.
27595 2013-02-19 Richard Biener <rguenther@suse.de>
27597 PR tree-optimization/56384
27598 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
27599 (vn_hash_type): Split out from ...
27600 (vn_hash_constant_with_type): ... here.
27601 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
27602 (vn_phi_eq): Compare types from vn_phi_s structure.
27603 (vn_phi_lookup): Populate vn_phi_s type.
27604 (vn_phi_insert): Likewise.
27606 2013-02-19 Jakub Jelinek <jakub@redhat.com>
27608 PR tree-optimization/56350
27609 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
27610 if haven't found reduction or nested cycle operand, rather than
27611 asserting we must find it.
27613 PR tree-optimization/56381
27614 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
27617 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
27618 Jakub Jelinek <jakub@redhat.com>
27621 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
27622 (swap_optab_enable): Same.
27623 (init_all_optabs): Use argument instead of global.
27624 * tree.h (struct tree_optimization_option): New field target_optabs.
27625 * expr.h (init_all_optabs): Add argument to prototype.
27626 (TREE_OPTIMIZATION_OPTABS): New.
27627 (save_optabs_if_changed): Protoize.
27628 * optabs.h: Declare this_fn_optabs.
27629 * optabs.c (save_optabs_if_changed): New.
27630 Declare this_fn_optabs.
27631 (init_optabs): Add argument to init_all_optabs() call.
27632 * function.c (invoke_set_current_function_hook): Handle per
27634 * function.h (struct function): New field optabs.
27635 * config/mips/mips.c (mips_set_mips16_mode): Handle when
27636 optimization_current_node has changed.
27637 * target-globals.h (save_target_globals_default_opts): Protoize.
27638 * target-globals.c (save_target_globals_default_opts): New.
27640 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
27643 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
27644 registers %fr12 and %fr12R as call used.
27647 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
27648 and HImode, require all displacements to be an integer multiple of
27650 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
27651 only allow QImode and HImode when reload is in progress and strict is
27652 true. Likewise for symbolic addresses. Use base14_operand to check
27653 displacements in REG+BASE addresses.
27655 2013-02-18 Richard Biener <rguenther@suse.de>
27657 PR tree-optimization/56366
27658 * tree-vect-loop.c (get_initial_def_for_induction): Properly
27659 handle sign-conversion of outer-loop initial induction value.
27661 2013-02-18 Richard Biener <rguenther@suse.de>
27663 PR middle-end/56349
27664 * cfghooks.c (merge_blocks): If we merge a latch into another
27665 block adjust references to it.
27666 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
27667 (verify_loop_structure): Verify that a recorded latch is in fact
27670 2013-02-18 Richard Biener <rguenther@suse.de>
27672 PR tree-optimization/56321
27673 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
27674 order SSA name release and virtual operand unlinking.
27676 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27678 * config/microblaze/microblaze.md (save_stack_block): Define.
27679 (restore_stack_block): Likewise.
27681 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27683 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
27684 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
27685 * config/microblaze/microblaze.c (microblaze_option_override):
27686 Bail out early for PIC modes when target does not support PIC.
27688 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27690 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
27691 Replace with a microblaze version.
27692 (microblaze_trampoline_init): Adapt for microblaze.
27693 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
27696 2013-02-16 Jakub Jelinek <jakub@redhat.com>
27697 Dodji Seketeli <dodji@redhat.com>
27700 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
27701 (instrument_mem_region_access): Do not forget to always put
27702 instrumentation of the of 'base' and 'base + len' in a "if (len !=
27703 0) statement, even for cases where either 'base' or 'base + len'
27704 are not instrumented -- because they have been previously
27705 instrumented. Simplify the logic by putting all the statements
27706 instrument 'base + len' inside a sequence, and then insert that
27707 sequence right before the current insertion point. Then, to
27708 instrument 'base + len', just get an iterator on that statement.
27709 And do not forget to update the pointer to iterator the function
27710 received as argument.
27712 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
27714 PR rtl-optimization/56348
27715 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
27717 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
27719 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
27720 (clean_graph_dump_file): Pass base to start_graph_dump.
27722 2013-02-14 Richard Henderson <rth@redhat.com>
27725 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
27727 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
27729 * collect2-aix.h: Define F_LOADONLY.
27731 2013-02-14 Richard Biener <rguenther@suse.de>
27734 * varasm.c (output_constant_def_1): Get the decl representing
27735 the constant as argument.
27736 (output_constant_def): Wrap output_constant_def_1.
27737 (make_decl_rtl): Use output_constant_def_1 with the decl
27738 representing the constant.
27739 (build_constant_desc): Optionally re-use a decl already
27740 representing the constant.
27741 (tree_output_constant_def): Adjust.
27743 2013-02-14 Dodji Seketeli <dodji@redhat.com>
27746 * asan.c (instrument_builtin_call): Really put the length of the
27747 second source argument into src1_len.
27749 2013-02-13 Jakub Jelinek <jakub@redhat.com>
27751 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
27752 argument. If it is false, don't create edge from then_bb to
27754 (insert_if_then_before_iter): Pass true to it.
27755 (build_check_stmt): Pass false to it.
27756 (transform_statements): Flush hash table only on extended basic
27757 block boundaries, rather than at the beginning of every bb.
27758 Don't flush hash table on nonfreeing_call_p calls.
27759 * tree-flow.h (nonfreeing_call_p): New prototype.
27760 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
27762 2013-02-13 David S. Miller <davem@davemloft.net>
27764 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
27766 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
27769 * ira.c (max_regno_before_ira): Move from ...
27771 (fix_reg_equiv_init): Use max_regno_before_ira instead of
27774 2013-02-13 Jakub Jelinek <jakub@redhat.com>
27776 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
27778 2013-02-13 Richard Biener <rguenther@suse.de>
27781 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
27782 globals in MEM_REFs.
27784 2013-02-13 Richard Biener <rguenther@suse.de>
27786 * loop-init.c (loop_optimizer_init): Clear loop state when
27787 re-initializing preserved loops.
27788 * loop-unswitch.c (unswitch_single_loop): Return whether
27789 we unswitched the loop. Do not verify loop state here.
27790 (unswitch_loops): When we unswitched a loop discover new loops.
27792 2013-02-13 Kostya Serebryany <kcc@google.com>
27794 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
27796 * sanitizer.def: Rename __asan_init to __asan_init_v1.
27798 2013-02-12 Dodji Seketeli <dodji@redhat.com>
27800 Avoid instrumenting duplicated memory access in the same basic block
27801 * Makefile.in (asan.o): Add new dependency on hash-table.h
27802 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
27803 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
27804 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
27805 (free_mem_ref_resources, has_mem_ref_been_instrumented)
27806 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
27807 (get_mem_ref_of_assignment): New functions.
27808 (get_mem_refs_of_builtin_call): Extract from
27809 instrument_builtin_call and tweak a little bit to make it fit with
27811 (instrument_builtin_call): Use the new
27812 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
27813 of is_gimple_builtin_call.
27814 (instrument_derefs, instrument_mem_region_access): Insert the
27815 instrumented memory reference into the hash table.
27816 (maybe_instrument_assignment): Renamed instrument_assignment into
27817 this, and change it to advance the iterator when instrumentation
27818 actually happened and return true in that case. This makes it
27819 homogeneous with maybe_instrument_assignment, and thus give a
27820 chance to callers to be more 'regular'.
27821 (transform_statements): Clear the memory reference hash table
27822 whenever we enter a new BB, when we cross a function call, or when
27823 we are done transforming statements. Use
27824 maybe_instrument_assignment instead of instrumentation. No more
27825 need to special case maybe_instrument_assignment and advance the
27826 iterator after calling it; it's now handled just like
27827 maybe_instrument_call. Update comment.
27829 2013-02-13 Richard Biener <rguenther@suse.de>
27831 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
27832 Fix loop discovery code.
27834 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
27836 PR inline-asm/56148
27837 * lra-constraints.c (process_alt_operands): Match early clobber
27838 operand with itself. Check conflicts with earlyclobber only if
27839 the operand is not reloaded. Prefer to reload conflicting operand
27840 if earlyclobber and matching operands are the same.
27842 2013-02-12 Richard Biener <rguenther@suse.de>
27845 * lto-streamer-out.c (write_symbol): Do not output symbols
27846 for hard register variables.
27848 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
27851 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
27852 (umulsidi3_insn, mulsidi3_insn): New insns.
27854 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
27856 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
27857 (struct tune_params): Add vec_costs field.
27858 * config/arm/arm.c (arm_builtin_vectorization_cost)
27859 (arm_add_stmt_cost): New functions.
27860 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
27861 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
27862 (arm_default_vec_cost): New struct of type cpu_vec_costs.
27863 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
27864 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
27865 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
27866 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
27868 2013-02-12 Richard Biener <rguenther@suse.de>
27871 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
27872 decls again if possible.
27874 2013-02-12 Richard Biener <rguenther@suse.de>
27876 PR middle-end/56288
27877 * tree-ssa.c (verify_ssa_name): Fix check, move
27878 SSA_NAME_IN_FREE_LIST check up.
27880 2013-02-12 Jakub Jelinek <jakub@redhat.com>
27881 Steven Bosscher <steven@gcc.gnu.org>
27883 PR rtl-optimization/56151
27884 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
27885 equal to op0 or op1, and last_insn pattern is CODE operation
27886 with MEM dest and one of the operands matches that MEM.
27888 2013-02-11 Sriraman Tallam <tmsriram@google.com>
27890 * doc/extend.texi: Document Function Multiversioning and "default"
27891 parameter string to target attribute.
27892 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
27893 target attribute parameter is "default".
27894 (ix86_compare_version_priority): Remove checks for target attribute.
27895 (ix86_mangle_function_version_assembler_name): Change error to sorry.
27896 Remove check for target attribute equal to NULL. Add assert.
27897 (ix86_generate_version_dispatcher_body): Change error to sorry.
27899 2013-02-11 Iain Sandoe <iain@codesourcery.com>
27900 Jack Howarth <howarth@bromo.med.uc.edu>
27901 Patrick Marlier <patrick.marlier@gmail.com>
27904 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
27905 define ENDFILE_SPEC as TM_DESTRUCTOR.
27906 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
27908 2013-02-11 Alexander Potapenko <glider@google.com>
27909 Jack Howarth <howarth@bromo.med.uc.edu>
27910 Jakub Jelinek <jakub@redhat.com>
27913 * config/darwin.c (cdtor_record): Rename ctor_record.
27914 (sort_cdtor_records): Rename sort_ctor_records.
27915 (finalize_dtors): New routine to sort destructors by
27916 priority before use in assemble_integer.
27917 (machopic_asm_out_destructor): Use finalize_dtors if needed.
27919 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
27921 PR rtl-optimization/56275
27922 * simplify-rtx.c (avoid_constant_pool_reference): Check that
27923 offset is non-negative and less than cmode size before
27924 calling simplify_subreg.
27926 2013-02-11 Richard Biener <rguenther@suse.de>
27928 PR tree-optimization/56264
27929 * cfgloop.h (fix_loop_structure): Adjust prototype.
27930 * loop-init.c (fix_loop_structure): Return the number of
27931 newly discovered loops.
27932 * tree-cfgcleanup.c (repair_loop_structures): When new loops
27933 are discovered, do a full loop-closed SSA rewrite.
27935 2013-02-11 Richard Biener <rguenther@suse.de>
27937 PR tree-optimization/56273
27938 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
27940 (check_array_ref): Fix missing newline in dumps.
27941 (search_for_addr_array): Likewise.
27943 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
27945 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
27947 2013-02-09 Jakub Jelinek <jakub@redhat.com>
27950 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
27952 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
27954 PR rtl-optimization/56246
27955 * lra-constraints.c (simplify_operand_subreg): Try to reuse
27957 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
27958 constraints are satisfied.
27960 2013-02-08 Jeff Law <law@redhat.com>
27963 * emit-rtl.c (reg_is_parm_p): New function.
27964 * regs.h (reg_is_parm_p): New prototype.
27965 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
27966 callee-clobbered registers.
27968 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
27971 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
27972 If there is no implicit builtin declaration, just return NULL.
27974 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
27976 * config/i386/sse.md (FMAMODEM): New mode iterator.
27977 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
27978 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
27980 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
27982 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
27983 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
27984 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
27986 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27988 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
27989 (microblaze*-*-elf): Likewise.
27990 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
27992 * config/microblaze/microblaze-c.c: Add builtin defines for
27993 _LITTLE_ENDIAN and _BIG_ENDIAN.
27994 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
27995 add to TARGET_DEFAULT flags.
27996 Expand ASM_SPEC and LINK_SPEC.
27997 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
27998 * config/microblaze/microblaze.md: Update extendsidi2 and
27999 movdi_internal instructions to use low-order / high-order reg
28001 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
28002 options and inversemask / mask of LITTLE_ENDIAN.
28003 * config/microblaze/t-microblaze: Expand multilib options to
28004 include mlittle-endian (le) and update exceptions patterns.
28006 2013-02-08 Jakub Jelinek <jakub@redhat.com>
28008 PR rtl-optimization/56195
28009 * lra-constraints.c (get_reload_reg): Don't reuse regs
28010 if they have smaller mode than requested, if they have
28011 wider mode than requested, try to return a SUBREG.
28013 PR tree-optimization/56250
28014 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
28015 if type is unsigned and code isn't MULT_EXPR.
28017 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
28019 PR tree-optimization/56064
28020 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
28021 bits according to mode.
28022 * fixed-value.h (fixed_from_double_int)
28023 (const_fixed_from_double_int): Adjust comments.
28025 2013-02-08 Richard Biener <rguenther@suse.de>
28028 * lto-streamer.h (struct data_in): Remove current_file, current_line
28029 and current_col members.
28030 * lto-streamer-out.c (lto_output_location): Stream changed bits
28031 en-block for efficiency.
28032 * lto-streamer-in.c (clear_line_info): Remove.
28033 (lto_input_location): Cache current file, line and column
28034 globally via local statics. Read changed bits en-block.
28035 (input_function): Do not call clear_line_info.
28036 (lto_read_body): Likewise.
28037 (lto_input_toplevel_asms): Likewise.
28039 2013-02-08 Michael Matz <matz@suse.de>
28041 PR tree-optimization/52448
28042 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
28043 (nt_call_phase): New static.
28044 (add_or_mark_expr): Only mark accesses with newer phase than any
28046 (nonfreeing_call_p): New.
28047 (nt_init_block): Update nt_call_phase, mark blocks as visited.
28048 (nt_fini_block): Keep blocks marked as visited.
28049 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
28051 2013-02-08 Richard Biener <rguenther@suse.de>
28053 * ira.c (ira): Free broken dominator information.
28055 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
28057 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
28059 2013-02-08 Marek Polacek <polacek@redhat.com>
28061 * cfgloop.c (verify_loop_structure): Add more checking of headers.
28063 2013-02-08 Richard Biener <rguenther@suse.de>
28065 PR middle-end/56181
28066 * cfgloop.h (flow_loops_find): Adjust.
28067 (bb_loop_header_p): Declare.
28068 * cfgloop.c (bb_loop_header_p): New function split out from ...
28069 (flow_loops_find): ... here. Adjust function signature,
28070 support incremental loop structure update.
28071 (verify_loop_structure): Cleanup. Verify a loop is a loop.
28072 * cfgloopmanip.c (fix_loop_structure): Move ...
28073 * loop-init.c (fix_loop_structure): ... here.
28074 (apply_loop_flags): Split out from ...
28075 (loop_optimizer_init): ... here.
28076 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
28077 in incremental mode, only remove dead loops here.
28079 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
28082 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
28083 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
28084 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
28085 (*round<mode>3.libgcc): New insns for fixed-modes.
28086 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
28087 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
28088 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
28089 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
28090 implementations. Define to __builtin_avr_absFX,
28091 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
28092 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
28093 __builtin_avr_countlsFX, respectively.
28094 * config/avr/avr-c.c (target.h): Include it.
28095 (enum avr_builtin_id): New enum.
28096 (avr_resolve_overloaded_builtin): New static function.
28097 (avr_register_target_pragmas): Use it to set
28098 targetm.resolve_overloaded_builtin.
28099 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
28100 tree nodes used by DEF_BUILTIN.
28101 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
28102 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
28103 <AVR_BUILTIN_xxBITS>: Same.
28105 2013-02-08 Richard Biener <rguenther@suse.de>
28107 * cfgloop.c (verify_loop_structure): Properly handle
28108 a loop exiting to another loop header.
28109 * ira-int.h (ira_loops): Remove.
28110 * ira.c (ira_loops): Remove.
28111 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
28112 (do_reload): Use loop_optimizer_finalize.
28113 * ira-build.c (create_loop_tree_nodes): Use get_loops and
28114 number_of_loops to access the loop tree.
28115 (more_one_region_p): Likewise.
28116 (finish_loop_tree_nodes): Likewise.
28117 (rebuild_regno_allocno_maps): Likewise.
28118 (mark_loops_for_removal): Likewise.
28119 (mark_all_loops_for_removal): Likewise.
28120 (remove_unnecessary_regions): Likewise.
28121 (ira_build): Likewise.
28122 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
28124 2013-02-08 Richard Biener <rguenther@suse.de>
28126 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
28127 * ipa-pure-const.c (analyze_function): Avoid calling
28128 mark_irreducible_loops twice.
28129 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
28131 2013-02-07 David S. Miller <davem@davemloft.net>
28133 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
28135 * var-tracking.c (vt_add_function_parameter): Test the presence of
28136 HAVE_window_save properly and do not remap argument registers when
28137 we have a leaf function.
28139 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
28142 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
28144 * config/i386/i386.c (ix86_print_operand): Ditto.
28146 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
28148 * lra-constraints.c (process_alt_operands): Fix recently added comment.
28150 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
28152 PR rtl-optimization/56225
28153 * lra-constraints.c (process_alt_operands): Check that reload hard
28154 reg can hold value for strict_low_part.
28156 2013-02-07 Jakub Jelinek <jakub@redhat.com>
28159 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
28160 dwarf2out_end_function.
28161 (in_first_function_p, maybe_at_text_label_p,
28162 first_loclabel_num_not_at_text_label): New variables.
28163 (dwarf2out_var_location): In the first function find out
28164 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
28165 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
28168 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
28170 PR rtl-optimization/56178
28171 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
28172 SUBREG of a register. Tidy up related block of code.
28173 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
28174 note if the source is a register or a SUBREG of a register.
28176 2013-02-07 Jakub Jelinek <jakub@redhat.com>
28179 * config/rs6000/rs6000.md (ptrm): New mode attr.
28180 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
28181 call_value_indirect_aix<pttrsize>,
28182 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
28185 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
28187 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
28188 if -bnortl. Convert to strcmp and strncmp.
28190 2013-02-07 Alan Modra <amodra@gmail.com>
28193 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
28194 addresses won't wrap when offsetting.
28195 (rs6000_secondary_reload): Provide secondary reloads needed for
28196 wrapping LO_SUM addresses.
28198 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
28200 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
28201 MACH, just __MACH__.
28203 2013-02-06 Richard Biener <rguenther@suse.de>
28205 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
28206 instead of calling fix_loop_structure.
28208 2013-02-06 Jakub Jelinek <jakub@redhat.com>
28210 PR middle-end/56217
28211 * omp-low.c (use_pointer_for_field): Return false if
28212 lower_send_shared_vars doesn't generate any copy-out code.
28214 2013-02-06 Tom de Vries <tom@codesourcery.com>
28216 PR rtl-optimization/56131
28217 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
28218 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
28219 of the label is NULL. Add comment.
28221 2013-02-05 Jakub Jelinek <jakub@redhat.com>
28223 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
28226 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
28227 (STATIC_LIBTSAN_LIBS): Likewise.
28228 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
28229 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
28230 is defined, don't add anything else beyond that.
28231 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
28232 (LINK_COMMAND_SPEC): Use them.
28234 PR tree-optimization/56205
28235 * tree-stdarg.c (check_all_va_list_escapes): Return true if
28236 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
28237 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
28239 2013-02-05 Richard Biener <rguenther@suse.de>
28241 PR tree-optimization/53342
28242 PR tree-optimization/53185
28243 * tree-vectorizer.h (vect_check_strided_load): Remove.
28244 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
28245 not disallow peeling for vectorized strided loads.
28246 (vect_check_strided_load): Make static and simplify.
28247 (vect_analyze_data_refs): Adjust.
28248 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
28249 correctly when vectorizing strided loads.
28251 2013-02-05 Richard Biener <rguenther@suse.de>
28253 * doc/install.texi: Refer to ISL, not PPL.
28255 2013-02-05 Jan Hubicka <jh@suse.cz>
28257 PR tree-optimization/55789
28258 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
28260 2013-02-05 Jan Hubicka <jh@suse.cz>
28262 PR tree-optimization/55789
28263 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
28264 the dead call anyway.
28266 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
28269 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
28271 2013-02-04 Alexander Potapenko <glider@google.com>
28272 Jack Howarth <howarth@bromo.med.uc.edu>
28273 Jakub Jelinek <jakub@redhat.com>
28276 * config/darwin.c (sort_ctor_records): Stabilized qsort
28277 on constructor priority by using original position.
28278 (finalize_ctors): New routine to sort constructors by
28279 priority before use in assemble_integer.
28280 (machopic_asm_out_constructor): Use finalize_ctors if needed.
28282 2013-02-04 Jakub Jelinek <jakub@redhat.com>
28285 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
28286 about visibility on artificial decls.
28287 * config/sol2.c (solaris_assemble_visibility): Likewise.
28289 2013-02-04 Kai Tietz <ktietz@redhat.com>
28292 * config/i386/i386.c (function_value_ms_64): Add additional valtype
28293 argument and improve checking of return-argument types for 16-byte
28295 (ix86_function_value_1): Add additional valtype argument on call
28296 of function_value_64.
28297 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
28298 handling infunction_value_64 function.
28300 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
28302 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
28304 2013-02-04 Richard Biener <rguenther@suse.de>
28306 PR tree-optimization/56188
28307 * tree-ssa-structalias.c (label_visit): Consider case with
28308 initially non-empty points-to set.
28309 (perform_var_substitution): Dump node mapping and clean up.
28311 2013-02-04 Richard Guenther <rguenther@suse.de>
28314 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
28315 node prevail as last resort.
28316 (lto_symtab_merge_decls): Remove guard on LTRANS here.
28317 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
28319 2013-02-04 Richard Biener <rguenther@suse.de>
28321 PR tree-optimization/56113
28322 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
28324 (equiv_class_lookup_or_add): ... this.
28325 (label_visit): Adjust and fix error in previous patch.
28326 (perform_var_substitution): Adjust.
28328 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
28330 * config/sh/divtab.c: Fix formatting and comments throughout the file.
28331 * config/sh/sh4-300.md: Likewise.
28332 * config/sh/sh4a.md: Likewise.
28333 * config/sh/constraints.md: Likewise.
28334 * config/sh/sh.md: Likewise.
28335 * config/sh/netbsd-elf.h: Likewise.
28336 * config/sh/predicates.md: Likewise.
28337 * config/sh/sh-protos.h: Likewise.
28338 * config/sh/ushmedia.h: Likewise.
28339 * config/sh/linux.h: Likewise.
28340 * config/sh/sh.c: Likewise.
28341 * config/sh/superh.h: Likewise.
28342 * config/sh/elf.h: Likewise.
28343 * config/sh/sh4.md: Likewise.
28344 * config/sh/sh.h: Likewise.
28346 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
28348 * config/pa/constraints.md: Adjust unused letters. Change "T"
28349 constraint to match_test floating_point_store_memory_operand().
28350 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
28351 (base14_operand): New.
28352 (floating_point_store_memory_operand): New.
28353 (integer_store_memory_operand): Revise to use base14_operand and
28354 reg_plus_base_memory_operand.
28355 (move_dest_operand): Allow symbolic_memory_operands.
28356 (symbolic_memory_operand): Check for LO_SOM.
28357 (symbolic_operand): Change default case to break.
28358 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
28359 CONST_DOUBLE values to be reloaded by putting them into memory when
28360 the destination is a floating point register.
28361 (movdf): Remove code to handle CONST_DOUBLE.
28363 (reload_indf_r1): New.
28364 (reload_insf_r1): New.
28365 Consistently use "Q" and "T" constraints with integer and floating
28366 point move instructions, respectively.
28367 (movdi): Remove FAIL.
28368 Change predicate for source operand unamed DImode move from
28369 general_operand to move_src_operand.
28370 (umulsidi3): Change predicate for destination operand to
28372 Likewise for similar unamed patterns.
28373 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
28374 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
28375 (hppa_legitimize_address): Simplify mask calculation.
28376 (pa_emit_move_sequence): Revised handling of secondary reloads from
28377 REG+D addresses for floating point loads and stores. Directly handle
28378 loading CONST0_RTX (mode) to a floating point register.
28379 (pa_secondary_reload): Handle reloading DF and SFmode constant values
28380 to floating point registers. Don't restrict secondary reloads to
28381 floating point registers to integer modes. Revise some comments and
28383 (TARGET_LEGITIMATE_ADDRESS_P): Define.
28384 (pa_legitimate_address_p): New.
28385 (pa_legitimize_reload_address): New.
28386 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
28387 (STRICT_REG_OK_FOR_BASE_P): New.
28388 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
28389 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
28391 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
28392 Andrew Dixie <andrewd@gentrack.com>
28394 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
28397 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
28399 * expmed.c (extract_bit_field_1): Pass the full width of the
28400 structure to get_best_reg_extraction_insn.
28402 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
28405 * configure.ac (use_cxa_atexit): Add AIX.
28406 * configure: Regenerate.
28408 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
28410 2013-02-01 Jakub Jelinek <jakub@redhat.com>
28413 * final.c (need_profile_function): New variable.
28414 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
28415 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
28416 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
28417 notes, targetm.asm_out.function_prologue doesn't emit anything,
28418 HAVE_prologue and profiler should be emitted before prologue,
28419 set need_profile_function instead of emitting it.
28420 (final_scan_insn): If need_profile_function, emit
28421 profile_function on the first NOTE_INSN_BASIC_BLOCK or
28422 NOTE_INSN_FUNCTION_BEG note.
28424 2013-02-01 Richard Henderson <rth@redhat.com>
28426 * config/rs6000/rs6000.md (smulditi3): New.
28429 * config/alpha/alpha.md (umulditi3): New.
28431 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
28433 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
28434 (ASM_OUTPUT_ALIGNED_LOCAL): New.
28436 2013-02-01 Richard Biener <rguenther@suse.de>
28438 PR tree-optimization/56113
28439 * tree-ssa-structalias.c (label_visit): Reduce work for
28440 single-predecessor nodes.
28442 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
28444 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
28445 range isn't testing for zero.
28447 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
28449 PR middle-end/56113
28450 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
28452 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
28453 Nick Clifton <nickc@redhat.com>
28455 * config/v850/constraints.md (Q): Define as a memory constraint.
28456 * config/v850/predicates.md (label_ref_operand): New predicate.
28457 (e3v5_shift_operand): New predicate.
28458 (ior_operator): New predicate.
28459 * config/v850/t-v850: Add e3v5 multilib.
28460 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
28461 (v850_gen_movdi): Prototype.
28462 * config/v850/v850.c: Add support for e3v5 architecture.
28463 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
28465 (construct_save_jarl): Add e3v5 long JARL support.
28466 (v850_adjust_insn_length): New function. Adjust length of call
28467 insns when using e3v5 instructions.
28468 (v850_gen_movdi): New function: Generate instructions to move a
28470 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
28471 (CPP_SPEC): Define __v850e3v5__ as appropriate.
28472 (TARGET_USE_FPU): Enable for e3v5.
28473 (CONST_OK_FOR_W): New macro.
28474 (ADJUST_INSN_LENGTH): Define.
28475 * config/v850/v850.md (UNSPEC_LOOP): Define.
28476 (attr cpu): Add v850e3v5.
28477 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
28478 (movdi): New pattern.
28479 (movdi_internal): New pattern.
28480 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
28481 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
28482 (cstoresf4): Likewise.
28483 (cstoredf4): Likewise.
28484 (insv): New pattern.
28485 (rotlso3_a): New pattern.
28486 (rotlsi3_b): New pattern
28487 (rotlsi3_v850e3v5): New pattern.
28488 (doloop_begin): New pattern.
28489 (fix_loop_counter): New pattern.
28490 (doloop_end): New pattern.
28491 (branch_normal): Add e3v5 long branch support.
28492 (branch_invert): Likewise.
28493 (branch_z_normal): Likewise.
28494 (branch_z_invert): Likewise.
28495 (branch_nz_normal): Likewise.
28496 (branch_nz_invert): Likewise.
28497 (call_internal_short): Add e3v5 register-indirect JARL support.
28498 (call_internal_long): Likewise.
28499 (call_value_internal_short): Likewise.
28500 (call_value_internal_long): Likewise.
28501 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
28502 (mloop): New option.
28503 * config.gcc: Add support for configuring v840e3v5 target.
28504 * doc/invoke.texi: Document new v850 specific command line options.
28506 2013-01-31 Paul Koning <ni1d@arrl.net>
28510 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
28511 children if parent is a class.
28512 (prune_unused_types_prune): Don't add DW_AT_declaration.
28514 2013-01-31 Richard Biener <rguenther@suse.de>
28516 PR tree-optimization/56157
28517 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
28518 match up operand with SLP child.
28520 2013-01-31 Jason Merrill <jason@redhat.com>
28523 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
28524 parameters the first time.
28525 (gen_scheduled_generic_parms_dies): Check completeness here.
28527 2013-01-31 Richard Biener <rguenther@suse.de>
28529 PR middle-end/53073
28530 * common.opt (faggressive-loop-optimizations): New flag,
28531 enabled by default.
28532 * doc/invoke.texi (faggressive-loop-optimizations): Document.
28533 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
28534 infer_loop_bounds_from_undefined by it.
28536 2013-01-31 Richard Biener <rguenther@suse.de>
28538 PR tree-optimization/56150
28539 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
28540 visit virtual operands.
28541 (find_uses_to_rename_bb): Likewise.
28543 2013-01-31 Richard Biener <rguenther@suse.de>
28545 PR tree-optimization/56150
28546 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
28547 mixed store non-store stmts.
28549 2013-01-30 Jakub Jelinek <jakub@redhat.com>
28552 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
28553 LIBASAN_EARLY_SPEC is defined.
28554 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
28555 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
28557 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
28560 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
28561 invalid args instead of ICEing on it.
28562 (ix86_valid_target_attribute_tree): Return error_mark_node if
28563 ix86_valid_target_attribute_inner_p failed.
28564 (ix86_valid_target_attribute_p): Return false only if
28565 ix86_valid_target_attribute_tree returned error_mark_node. Allow
28566 target("default") attribute.
28567 (sorted_attr_string): Change argument from const char * to tree,
28568 merge in all target attribute arguments rather than just one.
28569 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
28570 instead of free. Avoid using strcat.
28571 (ix86_mangle_function_version_assembler_name): Mangle
28572 target("default") as if no target attribute is present. Adjust
28573 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
28574 instead of xmalloc and XDELETEVEC instead of free.
28575 (ix86_function_versions): Don't return true if one of the decls
28576 doesn't have target attribute. If they don't and one of the decls
28577 is DECL_FUNCTION_VERSIONED, report an error. Adjust
28578 sorted_attr_string caller. Use XDELETEVEC instead of free.
28579 (ix86_supports_function_versions): Remove.
28580 (make_name): Fix up formatting.
28581 (make_dispatcher_decl): Remove resolver_name and its initialization.
28582 Avoid leaking memory.
28583 (is_function_default_version): Return true if there is
28584 target("default") attribute rather than no target attribute at all.
28585 (make_resolver_func): Avoid leaking memory.
28586 (ix86_generate_version_dispatcher_body): Likewise.
28587 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
28588 * target.def (supports_function_versions): Remove.
28589 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
28590 * doc/tm.texi: Regenerated.
28592 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
28594 PR rtl-optimization/56144
28595 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
28596 for values with side effects.
28598 2013-01-30 Richard Biener <rguenther@suse.de>
28600 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
28601 (sparseset_pop): Likewise.
28602 * cfganal.c (compute_idf): Likewise. Increase work-stack size
28603 to be able to use quick_push in the worker loop.
28605 2013-01-30 Marek Polacek <polacek@redhat.com>
28607 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
28609 2013-01-30 Richard Biener <rguenther@suse.de>
28612 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
28614 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
28616 PR tree-optimization/56064
28617 * fixed-value.c (fixed_from_double_int): New function.
28618 * fixed-value.h (fixed_from_double_int): New prototype.
28619 (const_fixed_from_double_int): New static inline function.
28620 * fold-const.c (native_interpret_fixed): New static function.
28621 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
28622 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
28623 (native_encode_fixed): New static function.
28624 (native_encode_expr) <FIXED_CST>: Use it.
28625 (native_interpret_int): Move double_int worker code to...
28626 * double-int.c (double_int::from_buffer): ...this new static method.
28627 * double-int.h (double_int::from_buffer): Prototype it.
28629 2013-01-30 Richard Biener <rguenther@suse.de>
28631 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
28632 New pointer-map and obstack.
28633 (init_alias_vars): Allocate pointer-map and obstack.
28634 (delete_points_to_sets): Free them.
28635 (find_what_var_points_to): Cache result.
28636 (find_what_p_points_to): Adjust for changed interface of
28637 find_what_var_points_to.
28638 (compute_points_to_sets): Likewise.
28639 (ipa_pta_execute): Likewise.
28641 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
28643 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
28644 * configure: Regenerate.
28645 * config.in: Regenerate.
28646 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
28647 #nobits/#progbits if supported.
28649 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
28652 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
28653 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
28654 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
28656 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
28658 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
28659 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
28661 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
28663 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
28664 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
28666 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
28668 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
28670 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
28671 * config/arm/cortex-a7.md: New bypasses using
28672 arm_mac_accumulator_is_result.
28674 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
28676 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
28677 (cortex_a7_neon_mla): Likewise.
28678 (cortex_a7_fpfmad): New reservation.
28679 (cortex_a7_fpmacs): Use ffmas and update required units.
28680 (cortex_a7_fpmuld): Update required units and latency.
28681 (cortex_a7_fpmacd): Likewise.
28682 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
28683 (cortex_a7_neon). Likewise.
28684 (bypass) Update participating units.
28686 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
28688 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
28689 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
28691 * config/arm/vfp11.md (vfp_farith): Use ffmas.
28692 (vfp_fmul): Use ffmad.
28693 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
28694 (cortex_r4_fmacd): Use ffmad.
28695 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
28696 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
28697 (cortex_a9_fmacd): Use ffmad.
28698 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
28699 (cortex_a8_vfp_macd): Use ffmad.
28700 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
28701 (cortex_a5_fpmacd): Use ffmad.
28702 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
28703 (cortex_a15_vfp_macd): Use ffmad.
28704 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
28706 2013-01-29 Jason Merrill <jason@redhat.com>
28709 * varasm.c (default_assemble_visibility): Don't warn about
28710 visibility on artificial decls.
28712 2013-01-29 Richard Biener <rguenther@suse.de>
28714 PR tree-optimization/56113
28715 * tree-ssa-structalias.c (equiv_class_lookup): Also return
28717 (label_visit): Free duplicate bitmaps and record the leader instead.
28718 (perform_var_substitution): Adjust.
28720 2013-01-29 Richard Biener <rguenther@suse.de>
28722 PR tree-optimization/55270
28723 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
28724 the CFG, schedule loops for fixup.
28726 2013-01-29 Nick Clifton <nickc@redhat.com>
28728 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
28731 2013-01-28 Leif Ekblad <leif@rdos.net>
28733 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
28734 * config/i386/i386.h (TARGET_RDOS): New macro.
28735 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
28736 * config/i386/i386.c (ix86_option_override_internal): For 64bit
28737 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
28738 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
28739 DEFAULT_LARGE_SECTION_THRESHOLD.
28740 * config/i386/i386.md (R14_REG, R15_REG): New constants.
28741 * config/i386/rdos.h: New file.
28742 * config/i386/rdos64.h: New file.
28744 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
28747 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
28750 2013-01-28 Jakub Jelinek <jakub@redhat.com>
28752 PR rtl-optimization/56117
28753 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
28754 call cselib_lookup_from_insn on the MEM before calling
28755 add_insn_mem_dependence.
28757 2013-01-28 Richard Biener <rguenther@suse.de>
28759 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
28760 to a stmt that didn't have one.
28761 (copy_phis_for_bb): Likewise for PHI arguments.
28762 (copy_debug_stmt): Likewise for debug stmts.
28764 2013-01-28 Richard Biener <rguenther@suse.de>
28766 PR tree-optimization/56034
28767 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
28768 (partition_builtin_p): Adjust.
28769 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
28770 it is the last partition.
28771 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
28772 up the vertex for the definition.
28773 (classify_partition): Classify whether a partition is a
28774 PKIND_REDUCTION, thus has uses outside of the loop.
28775 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
28776 Merge all PKIND_REDUCTION partitions into the last partition.
28777 (tree_loop_distribution): Seed partitions from reductions as well.
28779 2013-01-28 Jakub Jelinek <jakub@redhat.com>
28781 PR tree-optimization/56125
28782 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
28783 pow(x,c) into sqrt(x) * powi(x, n/2) or
28784 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
28785 optimizing for size.
28786 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
28787 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
28790 PR tree-optimization/56094
28791 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
28792 to UNKNOWN_LOCATION while gimplifying expr.
28794 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
28797 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
28798 operand 0 in movabs insn template for -masm=intel asm alternative.
28799 (*movabs<mode>_2): Ditto for operand 1.
28801 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
28804 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
28807 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
28809 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
28812 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
28814 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
28815 Undef to avoid warning.
28817 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
28819 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
28820 * configure: Regenerate.
28822 2013-01-25 Jakub Jelinek <jakub@redhat.com>
28824 PR tree-optimization/56098
28825 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
28826 for stmts with volatile ops.
28827 (cond_store_replacement): Don't optimize if assign has volatile ops.
28828 (cond_if_else_store_replacement_1): Don't optimize if either
28829 then_assign or else_assign have volatile ops.
28830 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
28833 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
28835 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
28837 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
28839 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
28840 missing ':' in asm example.
28842 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
28844 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
28845 entries into lane and laneq entries.
28846 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
28847 Remove AdvSIMD scalar modes.
28848 (aarch64_sq<r>dmulh_laneq<mode>): New.
28849 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
28851 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
28852 builtin implementations to relfect changes in RTL in aarch64-simd.md.
28853 * config/aarch64/iterators.md (VCOND): New.
28856 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
28859 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
28860 Add NULL LIBNAME argument to existing definitions.
28861 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
28862 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
28863 * config/avr/avr.c (DEF_BUILTIN): Same.
28864 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
28865 (avr_expand_builtin): Expand to a vanilla call if a libgcc
28866 implementation is available (DECL_ASSEMBLER_NAME is set).
28867 (avr_fold_absfx): New static function.
28868 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
28869 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
28870 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
28871 AVR_BUILTIN_ABSLLK.
28872 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
28873 (abshk, absk, abslk, absllk): Provide as static inline functions.
28875 2013-01-25 Marek Polacek <polacek@redhat.com>
28877 PR tree-optimization/56035
28878 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
28880 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
28882 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
28883 (*movtf_internal_rex64): Add (!o,C) alternative
28884 (*movxf_internal_rex64): Ditto.
28885 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
28887 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
28889 * doc/invoke.texi: fix typo.
28890 * doc/objc.texi: fix typo.
28892 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
28894 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
28895 for the first two alternatives.
28897 2013-01-24 Diego Novillo <dnovillo@google.com>
28899 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
28900 (ggc-zone.o): Remove.
28901 * configure.ac: Remove option --with-gc.
28902 * configure: Re-generate.
28903 * doc/install.texi: Remove documentation for --with-gc.
28904 * gengtype.c (write_enum_defn): Remove. Update all users.
28905 (write_Types_process_field): Remove generation of gt_e_* argument.
28906 (output_type_enum): Remove. Update all users.
28907 (write_enum_defn): Remove. Update all users.
28908 (enum alloc_zone): Remove. Update all users.
28909 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
28910 * ggc-common.c (ggc_splay_alloc): Remove first argument.
28911 Update all callers.
28912 (struct ptr_data): Remove field TYPE. Update all users.
28913 (gt_pch_note_object): Remove argument TYPE. Update all users.
28914 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
28916 * ggc-none.c (ggc_alloc_typed_stat): Remove.
28917 (struct alloc_zone): Remove.
28918 (ggc_internal_alloc_zone_stat): Remove.
28919 (ggc_internal_cleared_alloc_zone_stat): Remove.
28920 * ggc-page.c (ggc_alloc_typed_stat): Remove.
28921 (ggc_pch_count_object): Remove last argument. Update all users.
28922 (ggc_pch_alloc_object): Remove last argument. Update all users.
28923 (struct alloc_zone): Remove.
28924 * ggc-zone.c: Remove.
28925 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
28926 (struct alloc_zone): Remove.
28927 (ggc_alloc_typed_stat): Remove.
28928 (ggc_alloc_typed): Remove.
28929 (ggc_splay_alloc): Remove first argument.
28930 (rtl_zone): Remove. Update all users.
28931 (tree_zone): Remove. Update all users.
28932 (tree_id_zone): Remove. Update all users.
28933 (ggc_internal_zone_alloc_stat): Remove. Update all users.
28934 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
28935 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
28936 * tree-ssanames.c: Remove references to zone allocator in comments.
28938 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
28940 * config/avr/avr.c (avr_out_fract): Make register numbers that
28941 might be outside of source operand signed.
28943 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
28945 * config/i386/constraints.md (Yf): New constraint.
28946 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
28947 of f constraint to conditionaly disable x87 register preferences.
28948 (*movdf_internal): Ditto.
28949 (*movsf_internal): Ditto.
28951 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
28953 PR inline-asm/55934
28954 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
28955 that have operands with impossible constraints.
28956 Add a FIXME for a speed-up opportunity.
28957 * lra-constraints.c (process_alt_operands): Verify that a class
28958 selected from constraints on asms is valid for the operand mode.
28959 (curr_insn_transform): Remove incorrect comment.
28961 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
28963 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
28964 TOC operand is a valid symbol ref in the constant pool.
28966 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
28968 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
28970 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
28973 * config/avr/stdfix.h: New file.
28974 * t-avr (stdfix-gcc.h): New rule to build it.
28975 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
28977 2013-01-23 Kostya Serebryany <kcc@google.com>
28979 * config/darwin.h: remove dependency on
28980 CoreFoundation (asan on Mac OS).
28982 2013-01-23 Jakub Jelinek <jakub@redhat.com>
28985 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
28986 instead of cmpdi_operand for first comparison operand.
28987 Don't assert that comparison operands aren't both constants.
28989 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
28991 * doc/install.texi (Downloading the Source): Update references to
28992 downloading separate components.
28994 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
28996 * doc/extend.texi (__int128): Improve grammar.
28998 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
29001 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
29002 alternative to (o,r).
29003 (*movdi_internal_rex64): Remove (!o,n) alternative.
29004 (DImode immediate->memory splitter): Remove.
29005 (DImode immediate->memory peephole2): Remove.
29006 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
29007 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
29008 alternative to (!o,*r).
29009 (*movtf_internal_sse): New pattern.
29010 (*movxf_internal_rex64): New pattern.
29011 (*movxf_internal): Disable for TARGET_64BIT.
29012 (*movdf_internal_rex64): Remove (!o,F) alternative.
29014 2013-01-22 Jakub Jelinek <jakub@redhat.com>
29016 PR middle-end/56074
29017 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
29018 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
29019 * tree-vect-loop-manip.c (find_loop_location): Also ignore
29020 stmt locations where LOCATION_LOCUS of the stmt location is
29021 UNKNOWN_LOCATION or BUILTINS_LOCATION.
29024 * config/i386/i386.md (UNSPEC_STOS): New.
29025 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
29026 *strsetqi_1): Add UNSPEC_STOS.
29028 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
29031 * doc/invoke.texi: Remove left over -Wsynth example.
29033 2013-01-21 Jakub Jelinek <jakub@redhat.com>
29035 PR tree-optimization/56051
29036 * fold-const.c (fold_binary_loc): Don't fold
29037 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
29038 a narrowing conversion, or widening conversion from signed
29041 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
29043 PR rtl-optimization/56023
29044 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
29045 dependent on debug instruction.
29047 2013-01-21 Martin Jambor <mjambor@suse.cz>
29049 PR middle-end/56022
29050 * function.c (allocate_struct_function): Call
29051 invoke_set_current_function_hook earlier.
29053 2013-01-21 Jakub Jelinek <jakub@redhat.com>
29055 * reload1.c (init_reload): Only initialize reload_obstack
29056 during the first call.
29058 2013-01-21 Marek Polacek <polacek@redhat.com>
29060 * cfgloop.c (verify_loop_structure): Fix up grammar.
29062 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
29064 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
29065 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
29067 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
29070 * config/arm/marvell-pj4.md: Update copyright year.
29071 Fix up use of alu to alu_reg and simple_alu_imm.
29073 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
29075 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
29077 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
29080 * lra-constraints.c (curr_insn_transform): Don't reuse original
29081 insn for secondary memory move when memory mode should be different.
29083 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
29085 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
29086 atomic_storedi_1): New patterns.
29088 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
29090 btver2 pipeline descriptions.
29091 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
29093 * config/i386/i386.md (btver2_decode): New type attributes.
29094 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
29096 * config/i386/btver2.md: New file describing btver2 pipelines.
29098 2013-01-19 Andrew Pinski <apinski@cavium.com>
29100 PR tree-optimization/52631
29101 * tree-ssa-sccvn (visit_use): Before looking up the original
29102 statement, try looking up the simplified expression.
29104 2013-01-19 Anthony Green <green@moxielogic.com>
29106 * config/moxie/moxie.c (moxie_expand_prologue): Set
29107 current_function_static_stack_size.
29109 2013-01-18 Jakub Jelinek <jakub@redhat.com>
29111 PR tree-optimization/56029
29112 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
29113 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
29115 2013-01-18 Sharad Singhai <singhai@google.com>
29117 PR tree-optimization/55995
29118 * dumpfile.c (dump_loc): Print location only if available.
29119 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
29121 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
29124 * lra-constraints.c (curr_insn_transform): Reuse original insn for
29125 secondary memory move.
29126 (inherit_reload_reg): Use rclass instead of cl for
29127 check_secondary_memory_needed_p.
29129 2013-01-18 Jakub Jelinek <jakub@redhat.com>
29131 PR middle-end/56015
29132 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
29133 the case where writing real complex part of target modifies op1.
29135 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
29137 * config/aarch64/aarch64-simd.md
29138 (aarch64_vcond_internal<mode>): Handle unordered cases.
29139 * config/aarch64/iterators.md (v_cmp_result): New.
29141 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
29142 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
29144 * config/arm/marvell-pj4.md: New file.
29145 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
29146 * config/arm/arm.md (generic_sched): Add marvell_pj4.
29147 (generic_vfp): Likewise.
29148 * config/arm/arm-cores.def: Add marvell-pj4.
29149 * config/arm/arm-tune.md: Regenerate.
29150 * config/arm/arm-tables.opt: Regenerate.
29151 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
29152 * doc/invoke.texi: Document marvell-pj4.
29154 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
29156 * config/aarch64/arm_neon.h: Map scalar types to standard types.
29158 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
29163 * var-tracking.c (negative_power_of_two_p): New.
29164 (global_get_addr_cache, local_get_addr_cache): New.
29165 (get_addr_from_global_cache, get_addr_from_local_cache): New.
29166 (vt_canonicalize_addr): Rewrite using the above. Adjust the
29168 (vt_stack_offset_p): Remove.
29169 (vt_canon_true_dep): Always canonicalize loc's address.
29170 (clobber_overlapping_mems): Make sure we have a MEM.
29171 (local_get_addr_clear_given_value): New.
29172 (val_reset): Clear local cached entries.
29173 (compute_bb_dataflow): Create and release the local cache.
29174 Disable duplicate MEMs clobbering.
29175 (emit_notes_in_bb): Clobber MEMs likewise.
29176 (vt_emit_notes): Create and release the local cache.
29177 (vt_initialize, vt_finalize): Create and release the global
29178 cache, respectively.
29179 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
29181 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
29183 PR libmudflap/53359
29184 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
29185 not found in the symtab.
29187 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
29190 PR rtl-optimization/55547
29191 PR rtl-optimization/53827
29194 * alias.c (offset_overlap_p): New, factored out of...
29195 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
29196 the conservative special case for symbolic constants. Don't
29197 adjust zero sizes on alignment.
29199 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
29201 PR rtl-optimization/52573
29202 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
29203 REG_UNUSED for the same register.
29205 2013-01-17 Richard Biener <rguenther@suse.de>
29206 Marek Polacek <polacek@redhat.com>
29208 PR rtl-optimization/55833
29209 * loop-unswitch.c (unswitch_loops): Move loop verification...
29210 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
29211 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
29212 Set it to true when we're removing a loop from hierarchy tree in
29213 an irreducible region.
29214 (fix_bb_placements): Adjust caller.
29215 (fix_loop_placements): Likewise.
29217 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
29219 * config/avr/builtins.def (DEF_BUILTIN): Factor out
29220 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
29221 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
29222 Remove ID. Adjust comments.
29223 * config/avr/avr-c.c (avr_builtin_name): Remove.
29224 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
29225 * config/avr/avr.c (avr_tolower): New static function.
29226 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
29227 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
29228 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
29231 2013-01-17 Jan Hubicka <jh@suse.cz>
29233 PR tree-optimization/55273
29234 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
29236 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
29239 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
29240 store through atomic_store<mode>_1.
29241 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
29243 2013-01-17 Martin Jambor <mjambor@suse.cz>
29245 PR tree-optimizations/55264
29246 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
29247 for virtual methods.
29248 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
29249 virtual methods before inlining is over.
29250 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
29252 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
29255 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
29257 PR rtl-optimization/56005
29258 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
29259 pending reads for prefetch.
29261 2013-01-16 Ian Bolton <ian.bolton@arm.com>
29263 * config/aarch64/aarch64.md
29264 (*cstoresi_neg_uxtw): New pattern.
29265 (*cmovsi_insn_uxtw): New pattern.
29266 (*<optab>si3_uxtw): New pattern.
29267 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
29268 (*<optab>si3_insn_uxtw): New pattern.
29269 (*bswapsi2_uxtw): New pattern.
29271 2013-01-16 Richard Biener <rguenther@suse.de>
29273 * tree-inline.c (tree_function_versioning): Remove set but
29274 never used variable.
29276 2013-01-16 Richard Biener <rguenther@suse.de>
29278 PR tree-optimization/55964
29279 * tree-flow.h (rename_variables_in_loop): Remove.
29280 (rename_variables_in_bb): Likewise.
29281 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
29282 (copy_loop_before): Adjust and delete update-ssa status.
29283 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
29284 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
29285 (rename_variables_in_loop): Remove.
29286 (slpeel_update_phis_for_duplicate_loop): Likewise.
29287 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
29288 use available cfg machinery instead of duplicating it.
29289 Update PHI nodes and perform poor-mans SSA update here.
29290 (slpeel_tree_peel_loop_to_edge): Adjust.
29292 2013-01-16 Richard Biener <rguenther@suse.de>
29294 PR tree-optimization/54767
29295 PR tree-optimization/53465
29296 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
29297 (vrp_visit_phi_node): For PHI arguments coming via backedges
29298 drop all symbolical range information.
29299 (execute_vrp): Compute backedges.
29301 2013-01-16 Richard Biener <rguenther@suse.de>
29303 * doc/install.texi: Update CLooG and ISL requirements to
29306 2013-01-16 Christian Bruel <christian.bruel@st.com>
29309 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
29310 (broken_move): Handle UNSPECV_SP_SWITCH_B.
29311 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
29313 2013-01-16 DJ Delorie <dj@redhat.com>
29315 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
29316 (UNSPECV_SP_SWITCH_E): New.
29317 (sp_switch_1): Change to an unspec.
29318 (sp_switch_2): Change to an unspec. Don't use post-inc when we
29321 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
29323 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
29324 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
29325 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
29326 (expand_mem_thread_fence): Ditto.
29327 (expand_mem_signal_fence): Ditto.
29328 (expand_atomic_load): Ditto.
29329 (expand_atomic_store): Ditto.
29331 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
29333 PR rtl-optimization/55547
29334 PR rtl-optimization/53827
29337 * alias.c (memrefs_conflict_p): Set sizes to negative after
29340 2013-01-15 Jakub Jelinek <jakub@redhat.com>
29343 * function.c (thread_prologue_and_epilogue_insns): Always
29344 add crtl->drap_reg to set_up_by_prologue.set, even if
29345 stack_realign_drap is false.
29347 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
29349 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
29350 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
29351 *call): Fix indention.
29353 2013-01-15 Tom de Vries <tom@codesourcery.com>
29356 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
29359 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
29361 PR rtl-optimization/55153
29362 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
29364 2013-01-15 Martin Jambor <mjambor@suse.cz>
29366 PR tree-optimization/55920
29367 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
29368 accesses as grp_to_be_debug_replaced.
29370 2013-01-15 Jakub Jelinek <jakub@redhat.com>
29372 PR tree-optimization/55920
29373 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
29374 there is non-useless type conversion needed from debug rhs to lhs,
29375 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
29377 2013-01-15 Joseph Myers <joseph@codesourcery.com>
29378 Mikael Pettersson <mikpe@it.uu.se>
29381 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
29383 (ASM_OUTPUT_CASE_LABEL): Remove.
29384 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
29385 * final.c (shorten_branches): Update alignment of labels before
29386 jump tables if CASE_VECTOR_SHORTEN_MODE.
29388 2013-01-15 Richard Biener <rguenther@suse.de>
29391 * system.h: Do not include gmp.h for building host tools.
29393 2013-01-15 Richard Biener <rguenther@suse.de>
29395 PR middle-end/55882
29396 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
29397 account for bitpos when computing alignment.
29399 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
29401 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
29402 (ix86_target_macros_internal): Likewise.
29404 * config/i386/i386.c (m_CORE2I7): Removed.
29405 (m_CORE_HASWELL): New macro.
29406 (m_CORE_ALL): Likewise.
29407 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
29408 (initial_ix86_arch_features): Likewise.
29409 (processor_target_table): Initializations for Core avx2.
29410 (cpu_names): New names "core-avx2".
29411 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
29412 PROCESSOR_CORE_HASWELL.
29413 (ix86_issue_rate): New case.
29414 (ia32_multipass_dfa_lookahead): Likewise.
29415 (ix86_sched_init_global): Likewise.
29417 * config/i386/i386.h (TARGET_HASWELL): New macro.
29418 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
29419 (processor_type): New PROCESSOR_HASWELL.
29421 2013-01-15 Jakub Jelinek <jakub@redhat.com>
29423 PR tree-optimization/55955
29424 * tree-vect-loop.c (vectorizable_reduction): Give up early on
29425 *SHIFT_EXPR and *ROTATE_EXPR codes.
29427 PR tree-optimization/48766
29428 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
29429 -ftrapv disable -fwrapv.
29431 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
29434 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
29435 etc. to 1 and not to __flash.
29436 Use LL suffix for __INT24_MAX__ with -mint8.
29437 Use ULL suffix for __UINT24_MAX__ with -mint8.
29439 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
29441 * config/avr/avr-arch.h
29442 (struct base_arch_s): Use typedef avr_arch_t instead.
29443 (struct arch_info_s): Use typedef avr_arch_info_t instead.
29444 (struct mcu_type_s): Use typedef avr_mcu_t instead.
29445 * config/avr/avr.c: Same.
29446 * config/avr/avr-devices.c: Same.
29447 * config/avr/driver-avr.c: Same.
29448 * config/avr/gen-avr-mmcu-texi.c: Same.
29449 * config/avr/avr-mcus.def: Adjust comment.
29451 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
29453 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
29454 * config/aarch64/iterators.md (VALLDI): New.
29456 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
29457 Andi Kleen <ak@linux.intel.com>
29460 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
29461 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
29464 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
29466 * config/avr/avr-stdint.h: Remove trailing blanks.
29467 * config/avr/avr-log.h: Same.
29468 * config/avr/avr-arch.h: Same.
29469 * config/avr/avr-devices.c: Same.
29470 * config/avr/avr-dimode.md: Same.
29471 * config/avr/predicates.md: Same.
29472 * config/avr/avr-c.c: Same. And fix typo.
29474 * config/avr/avr-protos.h: Same. And:
29475 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
29476 (init_cumulative_args): Rename to avr_init_cumulative_args.
29477 (expand_prologue): Rename to avr_expand_prologue.
29478 (expand_epilogue): Rename to avr_expand_epilogue.
29479 (adjust_insn_length): Rename to avr_adjust_insn_length.
29480 (notice_update_cc): Rename to avr_notice_update_cc.
29481 (final_prescan_insn): Rename to avr_final_prescan_insn.
29482 * config/avr/avr.c: Same.
29483 * config/avr/avr.h: Same.
29484 * config/avr/avr.md: Remove trailing blanks.
29485 (prologue): Use avr_expand_prologue.
29486 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
29488 2013-01-14 Richard Biener <rguenther@suse.de>
29490 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
29491 verify_location, collect_subblocks): New functions.
29492 (verify_gimple_in_cfg): Verify that locations only reference
29493 BLOCKs in the functions BLOCK tree.
29495 2013-01-14 Richard Biener <rguenther@suse.de>
29497 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
29499 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
29501 (insert_out_of_ssa_copy_on_edge): Likewise.
29502 (rewrite_close_phi_out_of_ssa): Likewise.
29503 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
29505 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
29506 propagated constants.
29507 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
29510 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
29512 * config/avr/avr-modes.def: Add GPL copyright notice.
29514 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
29516 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
29517 MEMMODEL_MASK to determine memory model.
29518 (atomic_store<mode>): Ditto from operands[2].
29519 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
29521 2013-01-13 Jakub Jelinek <jakub@redhat.com>
29524 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
29525 (fold_gimple_assign): Don't call unshare_expr here.
29526 (fold_ctor_reference): Call unshare_expr.
29528 2013-01-13 Terry Guo <terry.guo@arm.com>
29530 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
29531 * doc/fragments.texi: Document MULTILIB_REUSE.
29532 * gcc.c (multilib_reuse): New internal spec.
29533 (set_multilib_dir): Also search multilib from multilib_reuse.
29534 * genmultilib (tmpmultilib3): Refactor code.
29535 (tmpmultilib4): Ditto.
29536 (multilib_reuse): New multilib argument.
29538 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
29540 * Makefile.in: Update copyright.
29542 2013-01-12 Tom de Vries <tom@codesourcery.com>
29544 PR middle-end/55890
29545 * calls.c (expand_call): Check if arg_nr is valid.
29547 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
29549 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
29550 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
29551 documentation. Add missing '__' in front of
29552 __builtin_ia32_packssdw256.
29554 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
29557 * config/s390/s390.c (s390_preferred_reload_class): Do not return
29558 NO_REGS for larl operands.
29559 (s390_reload_larl_operand): Use s390_load_address instead of
29562 2013-01-11 Richard Biener <rguenther@suse.de>
29564 * tree-cfg.c (verify_node_sharing_1): Split out from ...
29565 (verify_node_sharing): ... here.
29566 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
29568 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
29570 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
29571 Substitute TREECHECKING.
29572 * configure: Regenerate.
29573 * Makefile.in (TREECHECKING): New.
29575 2013-01-11 Richard Guenther <rguenther@suse.de>
29577 PR tree-optimization/44061
29578 * tree-vrp.c (extract_range_basic): Compute zero as
29579 value-range for __builtin_constant_p of function parameters.
29581 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
29583 Update copyright years.
29585 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
29587 PR rtl-optimization/55672
29588 * lra-eliminations.c (mark_not_eliminable): Permit addition with
29589 const to be eliminable.
29591 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
29593 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
29594 * configure: Regenerate.
29596 2013-01-10 Richard Biener <rguenther@suse.de>
29598 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
29600 2013-01-10 Richard Biener <rguenther@suse.de>
29603 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
29604 locations for virtual PHI arguments.
29605 (rewrite_update_phi_arguments): Likewise.
29607 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
29609 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
29612 2013-01-10 Jakub Jelinek <jakub@redhat.com>
29614 PR tree-optimization/55921
29615 * tree-complex.c (expand_complex_asm): New function.
29616 (expand_complex_operations_1): Call it for GIMPLE_ASM.
29618 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
29621 * config/s390/s390.c (s390_symref_operand_p)
29622 (s390_loadrelative_operand_p): Merge the two functions.
29623 (s390_check_qrst_address, print_operand_address): Add parameters
29624 to s390_loadrelative_operand_p invokation.
29625 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
29626 (s390_reload_larl_operand, s390_secondary_reload): Use
29627 s390_loadrelative_operand_p instead of s390_symref_operand_p.
29628 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
29630 2013-01-09 Mike Stump <mikestump@comcast.net>
29632 * dse.c (record_store): Remove unnecessary assert.
29634 2013-01-09 Jan Hubicka <jh@suse.cz>
29636 PR tree-optimization/55569
29637 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
29638 * cfgloop.h (scale_loop_profile): Likewise.
29640 2013-01-09 Jan Hubicka <jh@suse.cz>
29643 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
29645 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
29647 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
29649 PR middle-end/55114
29650 * expr.h (maybe_emit_group_store): Declare.
29651 * expr.c (maybe_emit_group_store): New function.
29652 * builtins.c (expand_builtin_int_roundingfn): Call it.
29653 (expand_builtin_int_roundingfn_2): Likewise.
29655 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
29657 PR rtl-optimization/55829
29658 * lra-constraints.c (match_reload): Add code for absent output.
29659 (curr_insn_transform): Add code for reloads of matched inputs
29662 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
29664 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
29665 attribute of movddup insn to DF.
29666 (*vec_interleave_lowv2df): Ditto.
29667 (vec_dupv2df): Ditto.
29669 2013-01-09 Jan Hubicka <jh@suse.cz>
29671 PR tree-optimiation/55875
29672 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
29673 EVERY_ITERATION parameter.
29674 (number_of_iterations_exit): Check if exit is executed every iteration.
29675 (idx_infer_loop_bounds): Similarly here.
29676 (n_of_executions_at_most): Simplify
29677 to only test for cases where statement is dominated by the
29678 particular bound; handle correctly the "postdominance" test.
29679 (scev_probably_wraps_p): Use max loop iterations info
29680 as a global bound first.
29682 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
29683 Nick Clifton <nickc@redhat.com>
29685 * config/v850/v850.md (cbranchsf4): New pattern.
29686 (cstoresf4): New pattern.
29687 (cbranchdf4): New pattern.
29688 (cstoredf4): New pattern.
29689 (movsicc): Disallow floating point comparisons.
29690 (cmpsf_le_insn): Fix order of operators.
29691 (cmpsf_lt_insn): Likewise.
29692 (cmpsf_eq_insn): Likewise.
29693 (cmpdf_le_insn): Likewise.
29694 (cmpdf_lt_insn): Likewise.
29695 (cmpdf_eq_insn): Likewise.
29696 (cmpsf_ge_insn): Use LE comparison.
29697 (cmpdf_ge_insn): Likewise.
29698 (cmpsf_gt_insn): Use LT comparison.
29699 (cmpdf_gt_insn): Likewise.
29700 (cmpsf_ne_insn): Delete pattern.
29701 (cmpdf_ne_insn): Delete pattern.
29702 * config/v850/v850.c (v850_gen_float_compare): Use
29703 gen_cmpdf_eq_insn for NE comparison.
29704 (v850_float_z_comparison_operator)
29705 (v850_float_nz_comparison_operator): Move from here ...
29706 * config/v850/predicates.md: ... to here. Move GT and GE
29707 comparisons into v850_float_z_comparison_operator.
29708 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
29710 (v850_float_nz_comparison_operator): Likewise.
29712 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
29714 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
29715 with calls to gen_insvsi/gen_insvdi.
29717 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
29719 * config/i386/i386.c (initial_ix86_tune_features): Set up
29720 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
29722 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
29723 Jakub Jelinek <jakub@redhat.com>
29725 PR tree-optimization/48189
29726 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
29727 If nitercst is 0, don't predict the exit edge.
29729 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
29731 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
29732 in asm_fprintf with reg_names.
29733 (aarch64_print_operand_address): Likewise.
29734 (aarch64_return_addr): Likewise.
29735 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
29737 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
29739 * config/pa/pa.h (VAL_U6_BITS_P): Define.
29740 (INT_U6_BITS): Likewise.
29741 * config/pa/predicates.md (uint6_operand): New predicate.
29742 (shift5_operand, shift6_operand): Likewise.
29743 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
29745 (lshrdi3): Use shift6_operand.
29746 (shrpsi4, shrpdi4): New insn patterns.
29747 (extzv): Delete expander.
29748 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
29749 predicates in unamed zero extract patterns. Tighten common constraint.
29750 (extv): Delete expander.
29751 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
29752 predicates in unamed sign extract patterns. Tighten common constraint.
29753 (insv): Delete expander.
29754 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
29755 predicates in unamed insert patterns. Tighten common constraint.
29756 Change uint32_operand predicate to uint6_operand predicate in unamed
29757 DImode pattern to insert constant values of type 1...1xxxx.
29759 2013-01-04 Jan Hubicka <jh@suse.cz>
29761 PR tree-optimization/55823
29762 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
29765 2013-01-08 Jakub Jelinek <jakub@redhat.com>
29766 Uros Bizjak <ubizjak@gmail.com>
29768 PR rtl-optimization/55845
29769 * df-problems.c (can_move_insns_across): Stop scanning at
29770 volatile_insn_p source instruction or give up if
29771 across_from .. across_to range contains any volatile_insn_p
29774 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
29776 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
29777 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
29779 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
29780 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
29782 2013-01-08 Jakub Jelinek <jakub@redhat.com>
29785 * asan.c (asan_clear_shadow): New function.
29786 (asan_emit_stack_protection): Use it.
29788 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
29790 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
29791 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
29792 with tab instead of space.
29794 2013-01-08 Nick Clifton <nickc@redhat.com>
29796 * config/rl78/rl78.c (rl78_expand_prologue): Always select
29797 register bank 0 at the start of an interrupt handler.
29798 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
29801 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
29803 * config/aarch64/aarch64-simd.md
29804 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
29805 (aarch64_simd_bsl): Likewise.
29806 (aarch64_vcond_internal<mode>): Likewise.
29807 (vcond<mode><mode>): Likewise.
29808 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
29809 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
29811 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
29813 * config/aarch64/aarch64-builtins.c
29814 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
29816 2013-01-08 Martin Jambor <mjambor@suse.cz>
29819 * tree-sra.c (analyze_access_subtree): Return true also after
29820 potentially creating a debug-only replacement.
29822 2013-01-08 Jakub Jelinek <jakub@redhat.com>
29824 PR middle-end/55890
29825 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
29827 PR tree-optimization/54120
29828 * tree-vrp.c (range_fits_type_p): Don't allow
29829 src_precision < precision from signed vr to unsigned_p
29830 if vr->min or vr->max is negative.
29831 (simplify_float_conversion_using_ranges): Test can_float_p
29832 against CODE_FOR_nothing.
29834 2013-01-08 Jakub Jelinek <jakub@redhat.com>
29835 Richard Biener <rguenther@suse.de>
29837 PR middle-end/55851
29838 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
29839 types instead of just INTEGER_TYPE types.
29841 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
29843 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
29846 2013-01-07 Steve Ellcey <sellcey@mips.com>
29849 * config/mips/mips.opt: Change mad to mmad to match documentation.
29851 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
29854 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
29855 .progmemx.data now.
29857 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
29860 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
29861 (avr_addrspace_t): Add .section_name field.
29862 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
29864 (avr_addrspace): Same. Initialize .section_name. Remove last
29865 NULL entry. Put __memx into .progmemx.data.
29866 (progmem_section_prefix): Remove.
29867 (avr_asm_init_sections): No need to initialize progmem_section.
29868 (avr_asm_named_section): Use avr_addrspace[].section_name to get
29869 section name prefix.
29870 (avr_asm_select_section): Ditto. And use get_unnamed_section to
29871 retrieve the progmem section.
29872 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
29873 boundary to run over avr_addrspace[].
29874 (avr_register_target_pragmas): Ditto.
29876 2013-01-06 Jakub Jelinek <jakub@redhat.com>
29878 * varasm.c (output_constant_def_contents): For asan_protect_global
29879 protected strings, adjust DECL_ALIGN if needed, before testing for
29881 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
29882 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
29884 (output_object_block): For asan protected decls, emit asan padding
29885 after their contents.
29886 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
29887 (asan_finish_file): Test it here instead.
29889 2013-01-07 Nick Clifton <nickc@redhat.com>
29890 Matthias Klose <doko@debian.org>
29891 Doug Kwan <dougkwan@google.com>
29892 H.J. Lu <hongjiu.lu@intel.com>
29895 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
29897 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
29899 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
29901 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
29903 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
29905 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
29908 * doc/install.texi (Cross-Compiler-Specific Options): Document
29911 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
29913 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
29914 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
29915 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
29916 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
29917 vqmovun_high_s64): Fix source operand number and update copyright.
29919 2013-01-07 Richard Biener <rguenther@suse.de>
29921 PR middle-end/55890
29922 * gimple.h (gimple_call_builtin_p): New overload.
29923 * gimple.c (validate_call): New function.
29924 (gimple_call_builtin_p): Likewise.
29925 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
29926 Use gimple_call_builtin_p.
29927 (find_func_clobbers): Likewise.
29928 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
29929 (strlen_optimize_stmt): Likewise.
29931 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
29933 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
29934 (vld1q_dup_*): Likewise.
29935 (vld1_*): Likewise.
29936 (vld1q_*): Likewise.
29937 (vld1_lane_*): Likewise.
29938 (vld1q_lane_*): Likewise.
29940 2013-01-07 Richard Biener <rguenther@suse.de>
29942 * lto-streamer.h (LTO_minor_version): Bump to 2.
29944 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
29946 * config/aarch64/aarch64-protos.h
29947 (aarch64_const_double_zero_rtx_p): Rename to...
29948 (aarch64_float_const_zero_rtx_p): ...this.
29949 (aarch64_float_const_representable_p): New.
29950 (aarch64_output_simd_mov_immediate): Likewise.
29951 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
29952 move immediate case.
29953 * config/aarch64/aarch64.c
29954 (aarch64_const_double_zero_rtx_p): Rename to...
29955 (aarch64_float_const_zero_rtx_p): ...this.
29956 (aarch64_print_operand): Allow printing of new constants.
29957 (aarch64_valid_floating_const): New.
29958 (aarch64_legitimate_constant_p): Check for valid floating-point
29960 (aarch64_simd_valid_immediate): Likewise.
29961 (aarch64_vect_float_const_representable_p): New.
29962 (aarch64_float_const_representable_p): Likewise.
29963 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
29964 (aarch64_output_simd_mov_immediate): New.
29965 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
29966 (*movdf_aarch64): Likewise.
29967 * config/aarch64/constraints.md (Ufc): New.
29968 (Y): call aarch64_float_const_zero_rtx.
29969 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
29971 2013-01-07 Richard Biener <rguenther@suse.de>
29973 PR tree-optimization/55888
29974 PR tree-optimization/55862
29975 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
29976 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
29977 not if it is contained therein.
29979 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
29981 * config/avr/t-avr: Typo.
29983 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
29986 * config/avr/t-avr: Don't automatically rebuild
29987 $(srcdir)/config/avr/t-multilib
29988 $(srcdir)/config/avr/avr-tables.opt
29989 $(srcdir)/doc/avr-mmcu.texi
29990 (avr-mcus): New phony target to build them on request.
29991 (s-avr-mlib, s-avr-mmcu-texi): Remove.
29992 * avr/avr-mcus.def: Adjust comments.
29994 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
29996 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
29998 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
30000 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
30002 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
30004 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
30006 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
30008 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
30009 to generate profiling.
30010 * config/rs6000/aix64.h (LIB_SPEC): Same.
30012 2013-01-04 Andrew Pinski <apinski@cavium.com>
30014 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
30016 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
30018 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
30020 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
30022 (ix86_expand_move): Ditto.
30023 (ix86_zero_extend_to_Pmode): Ditto.
30024 (ix86_expand_call): Ditto.
30025 (ix86_expand_special_args_builtin): Ditto.
30026 (ix86_expand_builtin): Ditto.
30028 2013-01-04 Richard Biener <rguenther@suse.de>
30030 PR tree-optimization/55862
30031 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
30032 translating them through PHI nodes.
30034 2013-01-04 Martin Jambor <mjambor@suse.cz>
30036 PR tree-optimization/55755
30037 * tree-sra.c (sra_modify_assign): Do not check that an access has no
30038 children when trying to avoid producing a VIEW_CONVERT_EXPR.
30040 2013-01-04 Marek Polacek <polacek@redhat.com>
30042 PR middle-end/55859
30043 * opts.c (default_options_optimization): Clarify error message.
30045 2013-01-04 Richard Biener <rguenther@suse.de>
30047 PR middle-end/55863
30048 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
30051 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
30054 * config/pa/pa.md (movsi): Revert previous change.
30055 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
30058 2013-01-03 Richard Henderson <rth@redhat.com>
30060 * config/i386/i386.c (ix86_expand_move): Always assign to op1
30061 after eliminating TLS symbols.
30063 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
30066 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
30067 * graphite-poly.c (debug_gmp_value): Likewise.
30069 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
30072 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
30073 selected code model, define __code_mode_small__, __code_model_medium__,
30074 __code_model_large__, __code_model_32__ or __code_model_kernel__.
30075 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
30076 xchg temporary register with %k. Declare temporary register as
30078 [__x86_64__]: For medium and large code models, preserve %rbx register.
30080 2013-01-03 Richard Biener <rguenther@suse.de>
30082 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
30083 (dump_subscript): Adjust.
30084 (finalize_ddr_dependent): Do not dump redundant info.
30085 (analyze_siv_subscript): Adjust.
30086 (subscript_dependence_tester): Likewise.
30087 (compute_affine_dependence): Likewise.
30089 2013-01-03 Richard Biener <rguenther@suse.de>
30092 2013-01-03 Richard Biener <rguenther@suse.de>
30094 PR tree-optimization/55857
30095 * tree-vect-stmts.c (vectorizable_load): Do not setup
30096 re-alignment for invariant loads.
30098 2013-01-02 Richard Biener <rguenther@suse.de>
30100 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
30101 invariant load do not generate a vector load from the scalar location.
30103 2013-01-03 Richard Biener <rguenther@suse.de>
30105 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
30106 for not vectorizing.
30107 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
30108 not build INDIRECT_REFs, call get_name once only.
30109 (vect_create_data_ref_ptr): Likewise. Dump base object kind
30110 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
30112 2013-01-03 Richard Biener <rguenther@suse.de>
30114 PR tree-optimization/55857
30115 * tree-vect-stmts.c (vectorizable_load): Do not setup
30116 re-alignment for invariant loads.
30118 2013-01-03 Richard Biener <rguenther@suse.de>
30121 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
30122 prefer a built-in decl.
30124 2013-01-03 Jakub Jelinek <jakub@redhat.com>
30126 * gcc.c (process_command): Update copyright notice dates.
30127 * gcov.c (print_version): Likewise.
30128 * gcov-dump.c (print_version): Likewise.
30130 PR rtl-optimization/55838
30131 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
30132 iv0.step, iv1.step and step.
30134 2013-01-03 Jakub Jelinek <jakub@redhat.com>
30135 Marc Glisse <marc.glisse@inria.fr>
30137 PR tree-optimization/55832
30138 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
30139 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
30140 integer_{one,zero}_node.
30142 2013-01-03 Jakub Jelinek <jakub@redhat.com>
30145 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
30146 * var-tracking.c (reverse_op): Don't add reverse ops to
30147 VALUEs that have already
30148 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
30150 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
30152 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
30154 2013-01-02 Teresa Johnson <tejohnson@google.com>
30156 * dumpfile.c (dump_loc): Print filename with location.
30157 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
30158 new location_t parameter to emit complete unroll message with
30159 new dump framework.
30160 (canonicalize_loop_induction_variables): Compute loops location
30161 and pass to try_unroll_loop_completely.
30162 * loop-unroll.c (report_unroll_peel): New function.
30163 (peel_loops_completely): Use new dump format with location
30164 for main dumpfile message, and invoke report_unroll_peel on success.
30165 (decide_unrolling_and_peeling): Ditto.
30166 (decide_peel_once_rolling): Remove old dumpfile message subsumed
30167 by report_unroll_peel.
30168 (decide_peel_completely): Ditto.
30169 (decide_unroll_constant_iterations): Ditto.
30170 (decide_unroll_runtime_iterations): Ditto.
30171 (decide_peel_simple): Ditto.
30172 (decide_unroll_stupid): Ditto.
30173 * cfgloop.c (get_loop_location): New function.
30174 * cfgloop.h (get_loop_location): Declare.
30176 2013-01-02 Sriraman Tallam <tmsriram@google.com>
30178 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
30181 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
30183 PR middle-end/55198
30184 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
30185 BLKmode objects when EXPAND_MEMORY is specified.
30187 2013-01-02 Sriraman Tallam <tmsriram@google.com>
30189 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
30191 (fold_builtin_cpu): Do not share cpu model decls across statements.
30193 2013-01-02 Jason Merrill <jason@redhat.com>
30196 * tree.c (build_array_type_1): Revert earlier change.
30198 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
30200 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
30202 * config/aarch64/aarch64-tune.md: Re-generate.
30204 2013-01-02 Richard Biener <rguenther@suse.de>
30206 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
30207 invariant load do not generate a vector load from the scalar location.
30209 2013-01-02 Richard Biener <rguenther@suse.de>
30212 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
30213 * configure: Regenerate.
30215 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
30217 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
30218 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
30219 (expand_builtin_int_roundingfn_2): Keep the original target around
30220 for the fallback case.
30222 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
30224 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
30225 to be clear for sign changes.
30227 2013-01-01 Jan Hubicka <jh@suse.cz>
30229 * ipa-inline-analysis.c: Fix formatting.
30231 2013-01-01 Jakub Jelinek <jakub@redhat.com>
30233 PR tree-optimization/55831
30234 * tree-vect-loop.c (get_initial_def_for_induction): Use
30235 gsi_after_labels instead of gsi_start_bb.
30237 Copyright (C) 2013 Free Software Foundation, Inc.
30239 Copying and distribution of this file, with or without modification,
30240 are permitted in any medium without royalty provided the copyright
30241 notice and this notice are preserved.