1 2013-11-08 H.J. Lu <hongjiu.lu@intel.com>
4 * doc/extend.texi: Move Cilk Plus Builtins node before Other
7 2013-11-08 Andrew MacLeod <amacleod@redhat.com>
8 Joseph Myers <joseph@codesourcery.com>
10 * ginclude/stdatomic.h: New file.
11 * Makefile.in (USER_H): Add stdatomic.h.
13 2013-11-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15 * config/arm/arm.c (arm_new_rtx_costs): Break after handling
18 2013-11-08 Jeff Law <law@redhat.com>
20 * tree-ssa-threadupdate.h (delete_thread_path): Declare.
21 * tree-ssa-threadupdate.c (delete_thread_path): New function.
22 (ssa_redirect_edges, thread_block_1): Use it.
23 (thread_through_loop_header, mark_threaded_blocks): Likewise.
24 (thread_through_all_blocks, register_jump_thread): Likewise.
25 * tree-ssa-threadedge.c (thread_across_edge): Likewise.
27 2013-11-08 James Greenhalgh <james.greenhalgh@arm.com>
29 * config/arm/aarch-common.c
30 (search_term): New typedef.
31 (shift_rtx_costs): New array.
32 (arm_rtx_shift_left_p): New.
33 (arm_find_sub_rtx_with_search_term): Likewise.
34 (arm_find_sub_rtx_with_code): Likewise.
35 (arm_early_load_addr_dep): Add sanity checking.
36 (arm_no_early_alu_shift_dep): Likewise.
37 (arm_no_early_alu_shift_value_dep): Likewise.
38 (arm_no_early_mul_dep): Likewise.
39 (arm_no_early_store_addr_dep): Likewise.
41 2013-11-08 Richard Biener <rguenther@suse.de>
43 PR tree-optimization/59047
44 * tree-predcom.c (ref_at_iteration): Handle bitfield accesses
47 2013-11-08 Ilya Enkovich <ilya.enkovich@intel.com>
49 * common.opt (fcheck-pointer-bounds): Move to ...
50 * c-family/c.opt: ... here.
51 * langhooks-def.h (LANG_HOOKS_CHKP_SUPPORTED): Remove.
52 (LANG_HOOKS_INITIALIZER): Remove LANG_HOOKS_CHKP_SUPPORTED.
53 * langhooks.h (lang_hooks): Remove chkp_supported field.
54 * toplev.c (process_options): Remove chkp_supported check.
56 2013-11-08 Richard Biener <rguenther@suse.de>
58 PR tree-optimization/59038
59 PR tree-optimization/58955
60 * tree-loop-distribution.c (pg_add_dependence_edges): Revert
61 previous change. Handle known dependences correctly.
63 2013-11-08 Tom de Vries <tom@codesourcery.com>
65 * config/rs6000/t-xilinx: Remove duplicate contents.
67 2013-11-07 Andrew MacLeod <amacleod@redhat.com>
68 Joseph Myers <joseph@codesourcery.com>
70 * tree-core.h (enum cv_qualifier): Add TYPE_QUAL_ATOMIC.
71 (enum tree_index): Add TI_ATOMICQI_TYPE, TI_ATOMICHI_TYPE,
72 TI_ATOMICSI_TYPE, TI_ATOMICDI_TYPE and TI_ATOMICTI_TYPE.
73 (struct tree_base): Add atomic_flag field.
74 * tree.h (TYPE_ATOMIC): New accessor macro.
75 (TYPE_QUALS, TYPE_QUALS_NO_ADDR_SPACE): Add TYPE_QUAL_ATOMIC.
76 (TYPE_QUALS_NO_ADDR_SPACE_NO_ATOMIC): New macro.
77 (atomicQI_type_node, atomicHI_type_node, atomicSI_type_node)
78 (atomicDI_type_node, atomicTI_type_node): New macros for type
80 * tree.c (set_type_quals): Set TYPE_ATOMIC.
81 (find_atomic_core_type): New function.
82 (build_qualified_type): Adjust alignment for qualified types.
83 (build_atomic_base): New function
84 (build_common_tree_nodes): Build atomicQI_type_node,
85 atomicHI_type_node, atomicSI_type_node, atomicDI_type_node and
87 * print-tree.c (print_node): Print atomic qualifier.
88 * tree-pretty-print.c (dump_generic_node): Print atomic type
90 * target.def (atomic_assign_expand_fenv): New hook.
91 * doc/tm.texi.in (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): New @hook.
92 * doc/tm.texi: Regenerate.
93 * targhooks.c (default_atomic_assign_expand_fenv): New function.
94 * targhooks.h (default_atomic_assign_expand_fenv): Declare.
95 * sync-builtins.def (__atomic_feraiseexcept): New built-in
97 * config/i386/i386-builtin-types.def (VOID_FTYPE_PUSHORT): New
99 * config/i386/i386.c (enum ix86_builtins): Add
100 IX86_BUILTIN_FNSTENV, IX86_BUILTIN_FLDENV, IX86_BUILTIN_FNSTSW and
102 (bdesc_special_args): Add __builtin_ia32_fnstenv,
103 __builtin_ia32_fldenv, __builtin_ia32_fnstsw and
104 __builtin_ia32_fnclex.
105 (ix86_expand_builtin): Handle the new built-in functions.
106 (ix86_atomic_assign_expand_fenv): New function.
107 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): New macro.
108 * config/i386/i386.md (UNSPECV_FNSTENV, UNSPECV_FLDENV)
109 (UNSPECV_FNSTSW, UNSPECV_FNCLEX): New unspecs.
110 (fnstenv, fldenv, fnstsw, fnclex): New insns.
112 2013-11-07 Steve Ellcey <sellcey@mips.com>
114 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add fp64 directory.
115 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add -mfp64 flag.
116 (MULTILIB_DIRNAMES): Add fp64 directory.
117 (MULTILIB_EXCEPTIONS): Add new exclusions.
119 2013-11-07 Aldy Hernandez <aldyh@redhat.com>
121 * gimplify.c (gimple_regimplify_operands): Do not set
123 * graphite-sese-to-poly.c (remove_simple_copy_phi): Same.
124 (rewrite_close_phi_out_of_ssa): Same.
125 (rewrite_phi_out_of_ssa): Same.
126 (rewrite_degenerate_phi): Same.
127 (handle_scalar_deps_crossing_scop_limits): Same.
128 * tree-if-conv.c (predicate_scalar_phi): Same.
129 * tree-parloops.c (create_loads_for_reductions): Same.
130 (create_final_loads_for_reduction): Same.
131 (create_loads_and_stores_for_name): Same.
132 (transform_to_exit_first_loop): Same.
133 (create_parallel_loop): Same.
135 (move_computations_dom_walker::before_dom_children): Same.
136 * tree-ssa-loop-manip.c (rewrite_phi_with_iv): Same.
137 * tree-ssa-pre.c (eliminate_dom_walker::before_dom_children):
139 * tree-ssa-propagate.c (substitute_and_fold): Same.
140 * tree-vect-loop.c (vect_finalize_reduction): Same.
141 * tree-vect-stmts.c (vectorizable_call): Same.
143 2013-11-07 Mike Stump <mikestump@comcast.net>
145 * config/pdp11/pdp11.c: Include dbxout.h.
146 * config/picochip/picochip.c: Likewise.
148 2013-11-07 Cong Hou <congh@google.com>
150 PR tree-optimization/56764
151 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
152 Combine alias checks if it is possible to amortize the runtime
153 overhead. Return the number of alias checks after merging.
154 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
155 Use the function vect_create_cond_for_alias_checks () to check
156 the number of alias checks.
158 2013-11-07 Jeff Law <law@redhat.com>
160 * varpool.c (ctor_for_folding): Fix typo in comment.
162 2013-11-07 Joern Rennecke <joern.rennecke@embecosm.com>
164 * config/arc/arc.c (arc_ifcvt): Use commutativity, e.g.:
165 reg_a := reg_b + reg_a ==> reg_a := reg_a + reg_b
167 2013-11-07 Jeff Law <law@redhat.com>
169 * doc/invoke.texi (-fisolate-erroneous-paths): Document.
171 * gimple-ssa-isolate-paths.c (gate_isolate_erroneous_paths):
172 No longer check if we have __builtin_trap, assume it's
175 2013-11-07 Thomas Schwinge <thomas@codesourcery.com>
177 * doc/sourcebuild.texi (Top Level) <lto-plugin>: GNU ld can use
180 * config/arc/arc.h (LINK_COMMAND_SPEC): For
181 -ftree-parallelize-loops=*, link to libgomp and its dependencies.
182 * config/ia64/hpux.h (LIB_SPEC): Likewise.
183 * config/pa/pa-hpux11.h (LIB_SPEC): Likewise.
184 * config/pa/pa64-hpux.h (LIB_SPEC): Likewise.
185 * gcc.c (GOMP_SELF_SPECS): Update comment about libgomp's
188 2013-11-07 Jakub Jelinek <jakub@redhat.com>
190 * tree-ssa-loop-niter.c: Include tree-ssanames.h.
191 (determine_value_range): Add loop argument. Use get_range_info to
193 (bound_difference): Adjust caller.
195 2013-11-07 Richard Biener <rguenther@suse.de>
196 Jakub Jelinek <jakub@redhat.com>
198 * tree-vrp.c (find_assert_locations): Pre-seed live bitmaps for loop
199 latches from header PHI arguments from the latch edge.
201 2013-11-07 Paolo Carlini <paolo.carlini@oracle.com>
204 * varasm.c (output_constant): Handle NULLPTR_TYPE.
206 2013-11-07 H.J. Lu <hongjiu.lu@intel.com>
208 * config/i386/i386.c (ix86_expand_set_or_movmem): Don't set
209 misaligned_prologue_used when it has been set.
211 2013-11-07 Yury Gribov <y.gribov@samsung.com>
212 Jakub Jelinek <jakub@redhat.com>
215 * asan.c (get_mem_refs_of_builtin_call): Allow
216 integer literals as addresses in instrumented builtins.
218 2013-11-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
220 * config/aarch64/aarch64.c (aarch64_legitimize_reload_address):
221 Explain why plus_constant is not used.
223 2013-11-07 Richard Biener <rguenther@suse.de>
225 * tree-ssa-ccp.c (canonicalize_float_value): Rename to ...
226 (canonicalize_value): ... this. Also handle stripping of
228 (get_value, set_lattice_value, get_value_for_expr): Adjust.
229 * gimple-fold.c (canonicalize_constructor_val): Strip
231 * tree-ssa-threadedge.c (set_ssa_name_value): Likewise.
233 2013-11-07 Richard Biener <rguenther@suse.de>
235 * tree-dfa.c (get_ref_base_and_extent): Fix casting.
237 2013-11-07 H.J. Lu <hongjiu.lu@intel.com>
240 * config/i386/i386.md (push peepholer/splitter): Use Pmode
241 with stack_pointer_rtx.
243 2013-11-07 Bin Cheng <bin.cheng@arm.com>
245 * tree-ssa-loop-ivopts.c (get_shiftadd_cost): Check equality
246 using operand_equal_p.
248 2013-11-07 Bin Cheng <bin.cheng@arm.com>
250 * tree-ssa-loop-ivopts.c (alloc_iv): Lower address expressions.
251 * tree-affine.c (get_inner_reference_aff): Return base.
252 * tree-affine.h (get_inner_reference_aff): Change prototype.
254 2013-11-06 Tobias Burnus <burnus@net-b.de>
256 * doc/invoke.texi (Wdate-time): Fix typo.
258 2013-11-06 Oleg Endo <olegendo@gcc.gnu.org>
260 * config/sh/sh.md (addsf3, divsf3, divsf3_i, rsqrtsf2, cmpgtdf_t,
261 cmpeqdf_t, *ieee_ccmpeqdf_t, negdf2, sqrtdf2, absdf2): Use
262 fp_arith_reg_operand instead of arith_reg_operand.
264 2013-11-06 Oleg Endo <olegendo@gcc.gnu.org>
266 * config/sh/sh.md (adddi3): Remove empty constraints.
267 Remove can_create_pseudo_p and arith_reg_operand check.
268 (adddi3_compact, subdi3_compact, *negdi2): Remove constraints.
271 2013-11-06 Jeff Law <law@redhat.com>
272 Tom Tromey <tromey@redhat.com>
274 * gdbinit.in: Disable strict type checking.
276 2013-11-06 Vladimir Makarov <vmakarov@redhat.com>
278 * tree-pass.h (make_pass_live_range_shrinkage): New external.
279 * timevar.def (TV_LIVE_RANGE_SHRINKAGE): New.
280 * sched-rgn.c (gate_handle_live_range_shrinkage): New.
281 (rest_of_handle_live_range_shrinkage): Ditto
282 (class pass_live_range_shrinkage): Ditto.
283 (pass_data_live_range_shrinkage): Ditto.
284 (make_pass_live_range_shrinkage): Ditto.
285 * sched-int.h (initialize_live_range_shrinkage): New prototype.
286 (finish_live_range_shrinkage): Ditto.
287 * sched-deps.c (create_insn_reg_set): Make void return value.
288 * passes.def: Add pass_live_range_shrinkage.
289 * ira.c (update_equiv_regs): Don't move if
290 flag_live_range_shrinkage.
291 * haifa-sched.c (live_range_shrinkage_p): New.
292 (initialize_live_range_shrinkage, finish_live_range_shrinkage):
294 (rank_for_schedule): Add code for pressure relief through live
296 (schedule_insn): Print more debug info.
297 (sched_init): Setup SCHED_PRESSURE_WEIGHTED for pressure relief
298 through live range shrinkage.
299 * doc/invoke.texi (-flive-range-shrinkage): New.
300 * common.opt (flive-range-shrinkage): New.
302 2013-11-06 Uros Bizjak <ubizjak@gmail.com>
305 * config/i386/i386.c (ix86_avx_u128_mode_needed): Require
306 AVX_U128_DIRTY mode for call_insn RTXes that use AVX256 registers.
307 (ix86_avx_u128_mode_needed): Return AVX_U128_DIRTY mode for call_insn
308 RTXes that return in AVX256 register.
310 2013-11-06 Richard Biener <rguenther@suse.de>
312 PR tree-optimization/58653
313 * tree-predcom.c (ref_at_iteration): Rewrite to generate a MEM_REF.
314 (prepare_initializers_chain): Adjust.
316 2013-11-06 Andrew MacLeod <amacleod@redhat.com>
318 * gimple.h (block_in_transaction): Move to basic-block.h and rename.
319 (gimple_in_transaction): Use bb_in_transaction.
320 * basic-block.h (bb_in_transaction): Relocate here and rename.
321 * tree-ssa-loop-im.c (execute_sm): Use bb_in_transaction.
323 2013-11-06 Richard Biener <rguenther@suse.de>
325 * tree.c (drop_tree_overflow): New function.
326 * tree.h (drop_tree_overflow): Declare.
327 * gimplify.c (gimplify_expr): Drop TREE_OVERFLOW.
328 * tree-vrp.c (range_int_cst_singleton_p): Use
329 is_overflow_infinity instead of testing TREE_OVERFLOW.
330 (extract_range_from_assert): Likewise.
331 (zero_nonzero_bits_from_vr): Likewise.
332 (extract_range_basic): Likewise.
333 (register_new_assert_for): Use drop_tree_overflow.
334 (vrp_visit_phi_node): Likewise.
336 2013-11-06 Eric Botcazou <ebotcazou@adacore.com>
338 * config/i386/i386.c (ix86_expand_prologue): Optimize stack
339 checking for leaf functions without dynamic stack allocation.
340 * config/ia64/ia64.c (ia64_emit_probe_stack_range): Adjust.
341 (ia64_expand_prologue): Likewise.
342 * config/mips/mips.c (mips_expand_prologue): Likewise.
343 * config/rs6000/rs6000.c (rs6000_emit_prologue): Likewise.
344 * config/sparc/sparc.c (sparc_expand_prologue): Likewise.
345 (sparc_flat_expand_prologue): Likewise.
347 2013-11-06 James Greenhalgh <james.greenhalgh@arm.com>
349 * config/aarch64/arm_neon.h (__ST2_LANE_FUNC): Better model data size.
350 (__ST3_LANE_FUNC): Likewise.
351 (__ST4_LANE_FUNC): Likewise.
353 2013-11-06 Nick Clifton <nickc@redhat.com>
355 * config/msp430/msp430.h (TARGET_CPU_CPP_BUILTINS): Define the
356 name returned by msp430_mcu_name.
357 (LIB_SPEC): If a -T option has not been specified then set a
358 default, mcu-specific, linker script.
359 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more mcu names.
360 * config/msp430/msp430.c (msp430x_names): Likewise.
361 Alpha sort the names for ease of comparison.
362 (msp430_mcu_name): New function: Returns a string suitable for
363 use as a C preprocessor symbol based upon the name of the MCU
365 (msp430_option_override): Accept msp430x and msp430xv2 as generic
367 * config/msp430/msp430-protos.h (msp430_mcu_name): Prototype.
369 * gcc.c (do_spec_1): Do not insert a space after a %* substitution
370 unless it is the last part of a spec substring.
371 * doc/invoke.texi (Spec Files): Document space insertion
374 2013-11-06 Christian Bruel <christian.bruel@st.com>
376 * config/sh/sh-mem.cc (sh_expand_cmpnstr, sh_expand_cmpstr):
377 Factorize probabilities, Use adjust_address instead of
378 adjust_automodify_address when possible. Enable for optimize.
379 (sh_expand_strlen): New function.
380 * config/sh/sh-protos.h (sh_expand_strlen): Declare.
381 * config/sh/sh.md (strlensi): New pattern.
382 (UNSPEC_BUILTIN_STRLEN): Define.
384 2013-11-06 Jakub Jelinek <jakub@redhat.com>
387 * expr.c (get_bit_range): Handle *offset == NULL_TREE.
388 (expand_assignment): If *bitpos is negative, set *offset
389 and adjust *bitpos, so that it is not negative.
391 2013-11-06 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
393 * config/i386/bdver3.md : Added two additional decoder units
394 to support issue rate of 4 and remodeled vector unit.
395 * config/i386/i386.c (ix86_issue_rate): Issue rate for BD
396 architectures is set to 4.
397 * config/i386/i386.c (ia32_multipass_dfa_lookahead): DFA
398 lookahead is set to 4 for BD architectures.
400 2013-11-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
402 * config/rs6000/rs6000.c (rs6000_option_override_internal):
403 Remove restriction against use of VSX instructions when generating
404 code for little endian mode.
406 2013-11-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
408 * config/rs6000/altivec.md (mulv4si3): Ensure we generate vmulouh
409 for both big and little endian.
410 (mulv8hi3): Swap input operands for merge high and merge low
411 instructions for little endian.
413 2013-11-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
415 * config/rs6000/altivec.md (vec_widen_umult_even_v16qi): Change
416 define_insn to define_expand that uses even patterns for big
417 endian and odd patterns for little endian.
418 (vec_widen_smult_even_v16qi): Likewise.
419 (vec_widen_umult_even_v8hi): Likewise.
420 (vec_widen_smult_even_v8hi): Likewise.
421 (vec_widen_umult_odd_v16qi): Likewise.
422 (vec_widen_smult_odd_v16qi): Likewise.
423 (vec_widen_umult_odd_v8hi): Likewise.
424 (vec_widen_smult_odd_v8hi): Likewise.
425 (altivec_vmuleub): New define_insn.
426 (altivec_vmuloub): Likewise.
427 (altivec_vmulesb): Likewise.
428 (altivec_vmulosb): Likewise.
429 (altivec_vmuleuh): Likewise.
430 (altivec_vmulouh): Likewise.
431 (altivec_vmulesh): Likewise.
432 (altivec_vmulosh): Likewise.
434 2013-11-05 Mike Stump <mikestump@comcast.net>
436 * Makefile.in (mostlyclean): Remove c-family objects.
438 2013-11-05 Ian Lance Taylor <iant@google.com>
440 * config/i386/sync.md (atomic_compare_and_swap<dwi>_doubleword):
441 If possible, add .cfi directives to record change to bx.
442 * config/i386/i386.c (ix86_emit_cfi): New function.
443 * config/i386/i386-protos.h (ix86_emit_cfi): Declare.
445 2013-11-05 Steven Bosscher <steven@gcc.gnu.org>
447 * rtlanal.c (tablejump_p): Expect a JUMP_TABLE_DATA to always follow
448 immediately after a label for a tablejump pattern.
450 * config/arm/arm.c (is_jump_table): Remove.
451 (create_fix_barrier): Use tablejump_p instead.
452 (arm_reorg): Likewise.
453 (thumb1_output_casesi): Expect JUMP_TABLE_DATA to always be NEXT_INSN.
454 (thumb2_output_casesi): Likewise.
455 * config/aarch64/aarch64.c (aarch64_output_casesi): Likewise.
456 * config/sh/sh.md (casesi_worker_1, casesi_worker_2,
457 casesi_shift_media, casesi_load_media): Likewise.
458 * config/iq2000/iq2000.md: Likewise (in anonymous define_insn).
459 * config/microblaze/microblaze.md: Likewise.
461 2013-11-05 Tobias Burnus <burnus@net-b.de>
463 * doc/invoke.texi (-Wdate-time): Document.
465 2013-11-05 Richard Sandiford <rdsandiford@googlemail.com>
467 * double-int.c (lshift_double, rshift_double): Remove
468 SHIFT_COUNT_TRUNCATED handling.
470 2013-11-05 Jeff Law <law@redhat.com>
472 * Makefile.in (OBJS): Add gimple-ssa-isolate-paths.o
473 * common.opt (-fisolate-erroneous-paths): Add option and documentation.
474 * gimple-ssa-isolate-paths.c: New file.
475 * gimple.c (check_loadstore): New function.
476 (infer_nonnull_range): Moved into gimple.c from tree-vrp.c
477 Verify OP is in the argument list and the argument corresponding
478 to OP is a pointer type. Use operand_equal_p rather than
479 pointer equality when testing if OP is on the nonnull list.
480 Use check_loadstore rather than count_ptr_derefs. Handle
481 GIMPLE_RETURN statements.
482 * tree-vrp.c (infer_nonnull_range): Remove.
483 * gimple.h (infer_nonnull_range): Declare.
484 * opts.c (default_options_table): Add OPT_fisolate_erroneous_paths.
485 * passes.def: Add pass_isolate_erroneous_paths.
486 * timevar.def (TV_ISOLATE_ERRONEOUS_PATHS): New timevar.
487 * tree-pass.h (make_pass_isolate_erroneous_paths): Declare.
488 * tree-ssa.c (struct count_ptr_d): Remove.
489 (count_ptr_derefs, count_uses_and_derefs): Remove.
490 * tree-ssa.h (count_uses_and_derefs): Remove.
492 2013-11-05 Jakub Jelinek <jakub@redhat.com>
494 PR rtl-optimization/58997
495 * loop-iv.c (iv_subreg): For IV_UNKNOWN_EXTEND, expect
496 get_iv_value to be in iv->mode rather than iv->extend_mode.
497 (iv_extend): Likewise. Otherwise, if iv->extend != extend,
498 use lowpart_subreg on get_iv_value before calling simplify_gen_unary.
499 * loop-unswitch.c (may_unswitch_on): Make sure op[i] is in the right
502 2013-11-05 Andrew MacLeod <amacleod@redhat.com>
504 * gimple.h: Move some prototypes to gimple-expr.h and add to include
506 (extract_ops_from_tree, gimple_call_addr_fndecl, is_gimple_reg_type):
507 Move to gimple-expr.h.
508 * gimple-expr.h: New file. Relocate some prototypes from gimple.h.
509 (types_compatible_p, is_gimple_reg_type, is_gimple_variable,
510 is_gimple_id, virtual_operand_p, is_gimple_addressable,
511 is_gimple_constant, extract_ops_from_tree, gimple_call_addr_fndecl):
513 * gimple.c (extract_ops_from_tree_1, gimple_cond_get_ops_from_tree,
514 gimple_set_body, gimple_body, gimple_has_body_p, is_gimple_lvalue,
515 is_gimple_condexpr, is_gimple_addressable, is_gimple_constant,
516 is_gimple_address, is_gimple_invariant_address,
517 is_gimple_ip_invariant_address, is_gimple_min_invariant,
518 is_gimple_ip_invariant, is_gimple_variable, is_gimple_id,
519 virtual_operand_p, is_gimple_reg, is_gimple_val, is_gimple_asm_val,
520 is_gimple_min_lval, is_gimple_call_addr, is_gimple_mem_ref_addr,
521 gimple_decl_printable_name, useless_type_conversion_p,
522 types_compatible_p, gimple_can_coalesce_p, copy_var_decl): Move to
524 * gimple-expr.c: New File.
525 (useless_type_conversion_p, gimple_set_body, gimple_body,
526 gimple_has_body_p, gimple_decl_printable_name, copy_var_decl,
527 gimple_can_coalesce_p, extract_ops_from_tree_1,
528 gimple_cond_get_ops_from_tree, is_gimple_lvalue, is_gimple_condexpr,
529 is_gimple_address, is_gimple_invariant_address,
530 is_gimple_ip_invariant_address, is_gimple_min_invariant,
531 is_gimple_ip_invariant, is_gimple_reg, is_gimple_val,
532 is_gimple_asm_val, is_gimple_min_lval, is_gimple_call_addr,
533 is_gimple_mem_ref_addr): Relocate here.
534 * Makefile.in (OBJS): Add gimple-expr.o.
536 2013-11-05 David Malcolm <dmalcolm@redhat.com>
538 * gengtype-parse.c (struct_field_seq): Support empty structs.
540 2013-11-05 Uros Bizjak <ubizjak@gmail.com>
542 * config/i386/t-rtems (MULTILIB_MATCHES): Fix option typos.
544 2013-11-05 Uros Bizjak <ubizjak@gmail.com>
546 * config/i386/i386-c.c (ix86_target_macros): Define _SOFT_FLOAT
548 * config/i386/rtemself.h (TARGET_OS_CPP_BUILTINS): Do not define
550 (LONG_DOUBLE_TYPE_SIZE): New define.
551 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Ditto.
553 2013-11-05 Paolo Carlini <paolo.carlini@oracle.com>
556 * doc/extend.texi [visibility ("visibility_type")]: Add example
557 about visibility attribute on namespace declaration.
559 2013-11-05 Richard Biener <rguenther@suse.de>
562 * passes.def (all_passes): Start with pass_fixup_cfg again.
564 2013-11-05 Richard Biener <rguenther@suse.de>
566 PR tree-optimization/58955
567 * tree-loop-distribution.c (pg_add_dependence_edges): Fix
570 2013-11-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
572 * config/rs6000/vector.md (vec_pack_sfix_trunc_v2df): Adjust for
574 (vec_pack_ufix_trunc_v2df): Likewise.
576 2013-11-05 H.J. Lu <hongjiu.lu@intel.com>
579 * doc/md.texi (@code{movmem@var{m}}): Specify Pmode as mode of
580 pattern, instead of word_mode.
582 * expr.c (emit_block_move_via_movmem): Don't use mode wider than
584 (set_storage_via_setmem): Likewise.
586 2013-11-05 Andrew MacLeod <amacleod@redhat.com>
588 * tree-outof-ssa.c (queue_phi_copy_p): Combine phi_ssa_name_p from
589 gimple.h and the rest of the condition in eliminate_build.
590 (eliminate_build): Call new routine.
591 * gimple.h (phi_ssa_name_p): Delete.
593 2013-11-05 Trevor Saunders <tsaunders@mozilla.com>
595 * vec.c (vec_prefix::calculate_allocation): Don't try to handle the
596 case of no prefix and reserving zero slots, because when that's the
597 case we'll never get here.
598 * vec.h (va_heap::reserve): Don't try and handle
599 vec_prefix::calculate_allocation returning zero because that should
602 2013-11-05 Richard Biener <rguenther@suse.de>
605 * tree-dfa.c (get_ref_base_and_extent): Merge common code
606 in MEM_REF and TARGET_MEM_REF handling. Make sure to
607 process trailing array detection before diving into the
608 view-converted object (and possibly apply some extra offset).
610 2013-11-05 Joseph Myers <joseph@codesourcery.com>
612 * config/i386/i386.c (ix86_float_exceptions_rounding_supported_p):
614 (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P): Define.
616 2013-11-05 Marc Glisse <marc.glisse@inria.fr>
618 PR tree-optimization/58958
619 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Use
620 get_addr_base_and_unit_offset instead of get_ref_base_and_extent.
622 2013-11-05 Marc Glisse <marc.glisse@inria.fr>
624 * tree-ssa-alias.h (ranges_overlap_p): Handle negative offsets.
625 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Likewise.
627 2013-11-05 Jakub Jelinek <jakub@redhat.com>
629 PR tree-optimization/58984
630 * ipa-prop.c (ipa_load_from_parm_agg_1): Add SIZE_P argument,
631 set *SIZE_P if non-NULL on success.
632 (ipa_load_from_parm_agg, ipa_analyze_indirect_call_uses): Adjust
634 (ipcp_transform_function): Likewise. Punt if size of access
635 is different from TYPE_SIZE on v->value's type.
637 2013-11-05 Tobias Burnus <burnus@net-b.de>
639 * doc/invoke.texi (-fopenmp-simd): Document new option.
640 * gimplify.c (gimplify_body): Accept -fopenmp-simd.
641 * omp-low.c (execute_expand_omp, execute_lower_omp): Ditto.
642 * tree.c (attribute_value_equal): Ditto.
644 2013-11-04 Wei Mi <wmi@google.com>
646 * sched-rgn.c (add_branch_dependences): Keep insns in
647 a SCHED_GROUP at the end of BB to remain their location.
649 2013-11-04 Wei Mi <wmi@google.com>
651 * config/i386/i386.c (memory_address_length): Extract a part
652 of code to rip_relative_addr_p.
653 (rip_relative_addr_p): New Function.
654 (ix86_macro_fusion_p): Ditto.
655 (ix86_macro_fusion_pair_p): Ditto.
656 * config/i386/i386.h: Add new tune features about macro-fusion.
657 * config/i386/x86-tune.def (DEF_TUNE): Ditto.
658 * doc/tm.texi: Generated.
659 * doc/tm.texi.in: Ditto.
660 * haifa-sched.c (try_group_insn): New Function.
661 (group_insns_for_macro_fusion): Ditto.
662 (sched_init): Call group_insns_for_macro_fusion.
663 * target.def: Add two hooks: macro_fusion_p and
666 2013-11-04 Kostya Serebryany <kcc@google.com>
668 Update to match the changed asan API.
669 * asan.c (asan_function_start): New function.
670 (asan_emit_stack_protection): Update the string stored in the
671 stack red zone to match new API. Store the PC of the current
672 function in the red zone.
673 (asan_global_struct): Update the __asan_global definition to match
675 (asan_add_global): Ditto.
676 * asan.h (asan_function_start): New prototype.
677 * final.c (final_start_function): Call asan_function_start.
678 * sanitizer.def (BUILT_IN_ASAN_INIT): Rename __asan_init_v1
681 2013-11-04 Wei Mi <wmi@google.com>
683 * config/i386/i386-c.c (ix86_target_macros_internal): Separate
684 PROCESSOR_COREI7_AVX out from PROCESSOR_COREI7.
685 * config/i386/i386.c (ix86_option_override_internal): Ditto.
686 (ix86_issue_rate): Ditto.
687 (ix86_adjust_cost): Ditto.
688 (ia32_multipass_dfa_lookahead): Ditto.
689 (ix86_sched_init_global): Ditto.
690 (get_builtin_code_for_version): Ditto.
691 * config/i386/i386.h (enum target_cpu_default): Ditto.
692 (enum processor_type): Ditto.
693 * config/i386/x86-tune.def (DEF_TUNE): Ditto.
695 2013-11-04 Vladimir Makarov <vmakarov@redhat.com>
697 PR rtl-optimization/58967
698 * config/rs6000/rs6000.c (legitimate_lo_sum_address_p): Remove
699 !lra_in_progress for mode sizes bigger word.
701 2013-11-04 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
703 * config/rs6000/altivec.md (vec_widen_umult_hi_v16qi): Swap
704 arguments to merge instruction for little endian.
705 (vec_widen_umult_lo_v16qi): Likewise.
706 (vec_widen_smult_hi_v16qi): Likewise.
707 (vec_widen_smult_lo_v16qi): Likewise.
708 (vec_widen_umult_hi_v8hi): Likewise.
709 (vec_widen_umult_lo_v8hi): Likewise.
710 (vec_widen_smult_hi_v8hi): Likewise.
711 (vec_widen_smult_lo_v8hi): Likewise.
713 2013-11-04 Ian Lance Taylor <iant@google.com>
715 * builtins.def (ATTR_NOTHROWCALL_LEAF_LIST): Define.
716 * sync-builtins.def: Use ATTR_NOTHROWCALL_LEAF_LIST for all sync
717 builtins that take pointers.
718 * lto-opts.c (lto_write_options): Write -fnon-call-exceptions if set.
719 * lto-wrapper.c (merge_and_complain): Collect OPT_fnon_call_exceptions.
720 (run_gcc): Pass -fnon-call-exceptions.
722 2013-11-04 Jakub Jelinek <jakub@redhat.com>
724 * optabs.c (expand_vec_perm): Revert one incorrect line from
727 PR tree-optimization/58978
728 * tree-vrp.c (all_imm_uses_in_stmt_or_feed_cond): Don't modify
729 use_stmt by single_imm_use directly. Only call single_imm_use
732 2013-11-04 Vladimir Makarov <vmakarov@redhat.com>
734 PR rtl-optimization/58968
735 * lra-spills.c (return_regno_p): New function.
736 (lra_final_code_change): Use it.
738 2013-11-04 Joseph Myers <joseph@codesourcery.com>
740 * doc/cpp.texi (__GCC_IEC_559, __GCC_IEC_559_COMPLEX): Document macros.
741 * target.def (float_exceptions_rounding_supported_p): New hook.
742 * targhooks.c (default_float_exceptions_rounding_supported_p): New
744 * targhooks.h (default_float_exceptions_rounding_supported_p): Declare.
745 * doc/tm.texi.in (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P):
747 * doc/tm.texi: Regenerate.
748 * config.gcc (powerpc*-*-linux*): Set extra_objs.
749 * config/rs6000/rs6000-linux.c: New file.
750 * config/rs6000/rs6000-protos.h
751 (rs6000_linux_float_exceptions_rounding_supported_p): Declare.
752 * config/rs6000/linux.h
753 (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P): New macro.
754 * config/rs6000/linux64.h
755 (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P): Likewise.
756 * config/rs6000/t-linux (rs6000-linux.o): New rule.
757 * config/rs6000/t-linux64 (rs6000-linux.o): Likewise.
759 2013-11-04 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
761 * config/rs6000/vsx.md (*vsx_le_perm_store_<mode> for VSX_D):
762 Replace the define_insn_and_split with a define_insn and two
763 define_splits, with the split after reload re-permuting the source
764 register to its original value.
765 (*vsx_le_perm_store_<mode> for VSX_W): Likewise.
766 (*vsx_le_perm_store_v8hi): Likewise.
767 (*vsx_le_perm_store_v16qi): Likewise.
769 2013-11-04 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
771 * config/rs6000/vector.md (vec_pack_trunc_v2df): Adjust for
774 2013-11-04 Jakub Jelinek <jakub@redhat.com>
776 PR tree-optimization/58946
777 * tree-ssa-reassoc.c (maybe_optimize_range_tests): Update all
778 bbs with bbinfo[idx].op != NULL before all blocks with
779 bbinfo[idx].op == NULL.
781 2013-11-04 Richard Sandiford <rdsandiford@googlemail.com>
783 * config/avr/avr-log.c (avr_double_int_pop_digit): Delete.
784 (avr_dump_double_int_hex): Likewise.
785 (avr_log_vadump): Remove %D and %X handling.
786 * config/avr/avr.c (avr_double_int_push_digit): Delete.
787 (avr_map_op_t): Change map from double_int to unsigned int.
788 (avr_map_op): Update accordingly.
789 (avr_map, avr_map_metric, avr_has_nibble_0xf, avr_map_decompose)
790 (avr_move_bits, avr_out_insert_bits, avr_fold_builtin): Operate on
791 unsigned ints rather than double_ints.
793 2013-11-03 Marek Polacek <polacek@redhat.com>
795 Implement -fsanitize=vla-bound.
796 * opts.c (common_handle_option): Handle vla-bound.
797 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE): Define.
798 * flag-types.h (enum sanitize_code): Add SANITIZE_VLA.
799 * asan.c (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR.
801 2013-11-02 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
803 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Adjust for
806 2013-11-02 Uros Bizjak <ubizjak@gmail.com>
808 * config/i386/constraints.md (Ts, Tv): New address constrains.
809 * config/i386/i386.md (*lea<mode>, *<mode>_<bndcheck>): Use Ts
810 constraint for address_no_seg_operand.
811 * config/i386/sse.md (*avx512pf_gatherpf<mode>_mask)
812 (*avx512pf_gatherpf<mode>, *avx512pf_scatterpf<mode>_mask)
813 (*avx512pf_scatterpf<mode>, *avx2_gathersi<mode>)
814 (*avx2_gathersi<mode>_2, *avx2_gatherdi<mode>, *avx2_gatherdi<mode>_2)
815 (*avx2_gatherdi<mode>_3, *avx2_gatherdi<mode>_4)
816 (*avx512f_gathersi<mode>, *avx512f_gathersi<mode>_2)
817 (*avx512f_gatherdi<mode>, *avx512f_gatherdi<mode>_2)
818 (*avx512f_scattersi<mode> *avx512f_scatterdi<mode>): Use Tv
819 constraint for vsib_address_operand.
821 2013-11-02 Steven Bosscher <steven@gcc.gnu.org>
823 * gcse.c (pre_delete): Remove references to regmove from comments.
824 * recog.c: (validate_replace_rtx_1): Likewise.
825 * config/rl78/rl78.c: Likewise.
826 * config/v850/v850.h: Likewise, and remove unused ENABLE_REGMOVE_PASS.
827 * common/config/m32r/m32r-common.c: Don't manipulate OPT_fregmove.
828 * common/config/mmix/mmix-common.c: Likewise.
830 2013-11-01 Trevor Saunders <tsaunders@mozilla.com>
832 * function.c (reorder_blocks): Convert block_stack to a stack_vec.
833 * gimplify.c (gimplify_compound_lval): Likewise.
834 * graphite-clast-to-gimple.c (gloog): Likewise.
835 * graphite-dependences.c (loop_is_parallel_p): Likewise.
836 * graphite-scop-detection.c (scopdet_basic_block_info): Likewise.
837 (limit_scop); Likewise.
838 (build_scops): Likewise.
839 (dot_scop): Likewise.
840 * graphite-sese-to-poly.c (sese_dom_walker): Likewise.
841 (build_scop_drs): Likewise.
842 (insert_stmts): Likewise.
843 (insert_out_of_ssa_copy): Likewise.
844 (remove_phi): Likewise.
845 (rewrite_commutative_reductions_out_of_ssa_close_phi): Likewise.
846 * hw-doloop.c (discover_loop): Likewise.
847 * tree-call-cdce.c (shrink_wrap_one_built_in_call): Likewise.
848 * tree-dfa.c (dump_enumerated_decls): Likewise.
849 * tree-if-conv.c (if_convertable_loop_p): Likewise.
850 * tree-inline.c (tree_function_versioning): Likewise.
851 * tree-loop-distribution.c (build_rdg): Likewise.
852 (rdg_flag_vertex_and_dependent): Likewise.
853 (distribute_loop): Likewise.
854 * tree-parloops.c (loop_parallel_p): Likewise.
855 (eliminate_local_variables): Likewise.
856 (separate_decls_in_region): Likewise.
857 * tree-predcom.c (tree_predictive_commoning_loop): Likewise.
858 * tree-ssa-phiopt.c (cond_if_else_store_replacement): Likewise.
859 * tree-ssa-uncprop.c (uncprop_dom_walker): Likewise.
860 * tree-vect-loop.c (vect_analyze_scaler_cycles_1): Likewise.
861 * tree-vect-patterns.c (vect_pattern_recog): Likewise.
862 * tree-vect-stmts.c (vect_mark_stmts_to_be_vectorized): Likewise.
863 (vectorizable_condition): Likewise.
865 2013-11-01 Uros Bizjak <ubizjak@gmail.com>
867 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): Always define as 0/1.
868 * configure: Regenerate.
869 * config/i386/i386.md (*movdi_internal): Change
870 HAVE_AS_IX86_INTERUNIT_MOVQ to runtime check.
871 (*movdf_internal): Ditto.
872 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
873 * config/i386/sse.md (vec_concatv2di): Output interunit movq
874 for HAVE_AS_IX86_INTERUNIT_MOVQ targets.
876 2013-10-31 Robert Suchanek <Robert.Suchanek@imgtec.com>
878 * lra-spills.c (assign_spill_hard_regs): Remove statement terminator
881 2013-10-31 David Malcolm <dmalcolm@redhat.com>
883 Automated part of renaming of symtab_node_base to symtab_node.
885 Patch autogenerated by rename_symtab.py from
886 https://github.com/davidmalcolm/gcc-refactoring-scripts
887 revision 58bb219cc090b2f4516a9297d868c245495ee622
888 with ChangeLog entry fixed up by hand.
890 * cgraph.c (x_cgraph_nodes_queue): Rename symtab_node_base to
892 (cgraph_node_for_asm): Likewise.
893 * cgraph.h (symtab_node_base): Likewise.
894 (cgraph_node): Likewise.
895 (varpool_node): Likewise.
896 (is_a_helper <cgraph_node>::test): Likewise.
897 (is_a_helper <varpool_node>::test): Likewise.
898 (symtab_nodes): Likewise.
899 (symtab_register_node): Likewise.
900 (symtab_unregister_node): Likewise.
901 (symtab_remove_node): Likewise.
902 (symtab_get_node): Likewise.
903 (symtab_node_for_asm): Likewise.
904 (symtab_node_asm_name): Likewise.
905 (symtab_node_name): Likewise.
906 (symtab_insert_node_to_hashtable): Likewise.
907 (symtab_add_to_same_comdat_group): Likewise.
908 (symtab_dissolve_same_comdat_group_list): Likewise.
909 (dump_symtab_node): Likewise.
910 (debug_symtab_node): Likewise.
911 (dump_symtab_base): Likewise.
912 (verify_symtab_node): Likewise.
913 (verify_symtab_base): Likewise.
914 (symtab_used_from_object_file_p): Likewise.
915 (symtab_alias_ultimate_target): Likewise.
916 (symtab_resolve_alias): Likewise.
917 (fixup_same_cpp_alias_visibility): Likewise.
918 (symtab_for_node_and_aliases): Likewise.
919 (symtab_nonoverwritable_alias): Likewise.
920 (availability symtab_node_availability): Likewise.
921 (symtab_semantically_equivalent_p): Likewise.
922 (fixup_same_cpp_alias_visibility): Likewise.
923 (symtab_prevail_in_asm_name_hash): Likewise.
926 (varpool_first_variable): Likewise.
927 (varpool_next_variable): Likewise.
928 (varpool_first_static_initializer): Likewise.
929 (varpool_next_static_initializer): Likewise.
930 (varpool_first_defined_variable): Likewise.
931 (varpool_next_defined_variable): Likewise.
932 (cgraph_first_defined_function): Likewise.
933 (cgraph_next_defined_function): Likewise.
934 (cgraph_first_function): Likewise.
935 (cgraph_next_function): Likewise.
936 (cgraph_first_function_with_gimple_body): Likewise.
937 (cgraph_next_function_with_gimple_body): Likewise.
938 (symtab_alias_target): Likewise.
939 (symtab_real_symbol_p): Likewise.
940 (symtab_can_be_discarded): Likewise.
941 * cgraphbuild.c (mark_address): Likewise.
942 (mark_load): Likewise.
943 (mark_store): Likewise.
944 * cgraphunit.c (decide_is_symbol_needed): Likewise.
946 (enqueue_node): Likewise.
947 (referred_to_p): Likewise.
948 (cgraph_process_same_body_aliases): Likewise.
949 (analyze_functions): Likewise.
950 (handle_alias_pairs): Likewise.
951 (output_weakrefs): Likewise.
953 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
954 * ipa-inline-analysis.c (inline_write_summary): Likewise.
955 * ipa-prop.c (remove_described_reference): Likewise.
956 (try_decrement_rdesc_refcount): Likewise.
957 (ipa_edge_duplication_hook): Likewise.
958 * ipa-ref.c (ipa_record_reference): Likewise.
959 (ipa_maybe_record_reference): Likewise.
960 (ipa_clone_ref): Likewise.
961 (ipa_clone_references): Likewise.
962 (ipa_clone_referring): Likewise.
963 (ipa_find_reference): Likewise.
964 (ipa_remove_stmt_references): Likewise.
965 (ipa_clear_stmts_in_references): Likewise.
966 * ipa-ref.h (symtab_node_base): Likewise.
968 (ipa_record_reference): Likewise.
969 (ipa_maybe_record_reference): Likewise.
970 (ipa_clone_references): Likewise.
971 (ipa_clone_referring): Likewise.
972 (ipa_clone_ref): Likewise.
973 (ipa_find_reference): Likewise.
974 (ipa_remove_stmt_references): Likewise.
975 (ipa_clear_stmts_in_references): Likewise.
976 * ipa-reference.c (ipa_reference_write_optimization_summary):
978 * ipa.c (enqueue_node): Likewise.
979 (process_references): Likewise.
980 (walk_polymorphic_call_targets): Likewise.
981 (symtab_remove_unreachable_nodes): Likewise.
982 (address_taken_from_non_vtable_p): Likewise.
983 (comdat_can_be_unshared_p_1): Likewise.
984 (comdat_can_be_unshared_p): Likewise.
985 (can_replace_by_local_alias): Likewise.
986 (function_and_variable_visibility): Likewise.
987 * is-a.h: Likewise (within example in comment).
988 * lto-cgraph.c (input_cgraph_opt_summary): Likewise.
989 (lto_symtab_encoder_encode): Likewise.
990 (lto_symtab_encoder_delete_node): Likewise.
991 (lto_symtab_encoder_in_partition_p): Likewise.
992 (lto_set_symtab_encoder_in_partition): Likewise.
993 (output_refs): Likewise.
994 (compute_ltrans_boundary): Likewise.
995 (output_symtab): Likewise.
996 (input_node): Likewise.
997 (input_ref): Likewise.
998 (input_edge): Likewise.
999 (input_cgraph_1): Likewise.
1000 (input_refs): Likewise.
1001 (output_cgraph_opt_summary): Likewise.
1002 (input_node_opt_summary): Likewise.
1003 (input_cgraph_opt_section): Likewise.
1004 * lto-section-in.c (lto_free_function_in_decl_state_for_node):
1006 * lto-streamer-out.c (lto_output): Likewise.
1007 (output_symbol_p): Likewise.
1008 (produce_symtab): Likewise.
1009 * lto-streamer.h (lto_encoder_entry): Likewise.
1010 (lto_free_function_in_decl_state_for_node): Likewise.
1011 (lto_symtab_encoder_encode): Likewise.
1012 (lto_symtab_encoder_delete_node): Likewise.
1013 (lto_symtab_encoder_in_partition_p): Likewise.
1014 (lto_set_symtab_encoder_in_partition): Likewise.
1015 (lto_symtab_encoder_lookup): Likewise.
1016 (lsei_node): Likewise.
1017 (lto_symtab_encoder_deref): Likewise.
1018 * symtab.c (symtab_hash): Likewise.
1019 (assembler_name_hash): Likewise.
1020 (symtab_nodes): Likewise.
1021 (hash_node): Likewise.
1022 (eq_node): Likewise.
1023 (hash_node_by_assembler_name): Likewise.
1024 (eq_assembler_name): Likewise.
1025 (insert_to_assembler_name_hash): Likewise.
1026 (unlink_from_assembler_name_hash): Likewise.
1027 (symtab_prevail_in_asm_name_hash): Likewise.
1028 (symtab_register_node): Likewise.
1029 (symtab_insert_node_to_hashtable): Likewise.
1030 (symtab_unregister_node): Likewise.
1031 (symtab_get_node): Likewise.
1032 (symtab_remove_node): Likewise.
1033 (symtab_initialize_asm_name_hash): Likewise.
1034 (symtab_node_for_asm): Likewise.
1035 (symtab_add_to_same_comdat_group): Likewise.
1036 (symtab_dissolve_same_comdat_group_list): Likewise.
1037 (symtab_node_asm_name): Likewise.
1038 (symtab_node_name): Likewise.
1039 (dump_symtab_base): Likewise.
1040 (dump_symtab_node): Likewise.
1041 (dump_symtab): Likewise.
1042 (debug_symtab_node): Likewise.
1043 (verify_symtab_base): Likewise.
1044 (verify_symtab_node): Likewise.
1045 (verify_symtab): Likewise.
1046 (symtab_used_from_object_file_p): Likewise.
1047 (symtab_node_availability): Likewise.
1048 (symtab_alias_ultimate_target): Likewise.
1049 (fixup_same_cpp_alias_visibility): Likewise.
1050 (symtab_resolve_alias): Likewise.
1051 (symtab_for_node_and_aliases): Likewise.
1052 (symtab_for_node_and_aliases): Likewise.
1053 (symtab_nonoverwritable_alias_1): Likewise.
1054 (symtab_nonoverwritable_alias): Likewise.
1055 (symtab_semantically_equivalent_p): Likewise.
1056 * value-prof.c (init_node_map): Likewise.
1057 * varasm.c (find_decl): Likewise.
1058 * varpool.c (varpool_node_for_asm): Likewise.
1059 (varpool_remove_unreferenced_decls): Likewise.
1061 2013-10-31 David Malcolm <dmalcolm@redhat.com>
1063 Manual part of renaming of symtab_node_base to symtab_node.
1065 * ipa-ref.h (symtab_node): Remove typedef to pointer type, as it
1066 clashes with the preferred name for the base class.
1067 (const_symtab_node): Remove redundant typedef.
1069 2013-10-31 Jakub Jelinek <jakub@redhat.com>
1071 * optabs.c (expand_vec_perm): Avoid vector mode punning
1072 SUBREGs in SET_DEST.
1073 * expmed.c (store_bit_field_1): Likewise.
1074 * config/i386/sse.md (movdi_to_sse, vec_pack_sfix_trunc_v2df,
1075 vec_pack_sfix_v2df, vec_shl_<mode>, vec_shr_<mode>,
1076 vec_interleave_high<mode>, vec_interleave_low<mode>): Likewise.
1077 * config/i386/i386.c (ix86_expand_vector_move_misalign,
1078 ix86_expand_sse_movcc, ix86_expand_int_vcond, ix86_expand_vec_perm,
1079 ix86_expand_sse_unpack, ix86_expand_args_builtin,
1080 ix86_expand_vector_init_duplicate, ix86_expand_vector_set,
1081 emit_reduc_half, expand_vec_perm_blend, expand_vec_perm_pshufb,
1082 expand_vec_perm_interleave2, expand_vec_perm_pshufb2,
1083 expand_vec_perm_vpshufb2_vpermq,
1084 expand_vec_perm_vpshufb2_vpermq_even_odd, expand_vec_perm_even_odd_1,
1085 expand_vec_perm_broadcast_1, expand_vec_perm_vpshufb4_vpermq2,
1086 ix86_expand_sse2_mulv4si3, ix86_expand_pinsr): Likewise.
1087 (expand_vec_perm_palignr): Likewise. Modify a copy of *d rather
1090 2013-10-31 Uros Bizjak <ubizjak@gmail.com>
1092 * config/i386/i386.c (ix86_expand_sse2_abs): Rename function arguments.
1093 Use gcc_unreachable for unhandled modes. Do not check results of
1094 expand_simple_binop. If not expanded to target, move the result.
1096 2013-10-31 Chung-Ju Wu <jasonwucj@gmail.com>
1097 Shiva Chen <shiva0217@gmail.com>
1099 * config.gcc (nds32*-*-*): Add nds32 target.
1100 * config/nds32/nds32.c: New file.
1101 * config/nds32/nds32.h: New file.
1102 * config/nds32/nds32.md: New file.
1103 * config/nds32/constants.md: New file.
1104 * config/nds32/constraints.md: New file.
1105 * config/nds32/iterators.md: New file.
1106 * config/nds32/nds32-doubleword.md: New file.
1107 * config/nds32/nds32-intrinsic.md: New file.
1108 * config/nds32/nds32_intrinsic.h: New file.
1109 * config/nds32/nds32-modes.def: New file.
1110 * config/nds32/nds32-multiple.md: New file.
1111 * config/nds32/nds32.opt: New file.
1112 * config/nds32/nds32-opts.h: New file.
1113 * config/nds32/nds32-protos.h: New file.
1114 * config/nds32/nds32-peephole2.md: New file.
1115 * config/nds32/pipelines.md: New file.
1116 * config/nds32/predicates.md: New file.
1117 * config/nds32/t-mlibs: New file.
1118 * common/config/nds32: New directory and files.
1120 * doc/invoke.texi (NDS32 options): Document nds32 specific options.
1121 * doc/md.texi (NDS32 family): Document nds32 specific constraints.
1122 * doc/install.texi (Cross-Compiler-Specific Options): Document
1123 --with-nds32-lib for nds32 target.
1124 * doc/extend.texi (Function Attributes, Target Builtins): Document
1125 nds32 specific attributes.
1127 2013-10-31 Vladimir Makarov <vmakarov@redhat.com>
1129 * lra-constraints (process_alt_operands): Use the result
1130 elimination register for operand when matching constraints.
1132 2013-10-31 Jakub Jelinek <jakub@redhat.com>
1134 * tree-vrp.c (maybe_set_nonzero_bits): New function.
1135 (remove_range_assertions): Call it.
1137 * tree.c (tree_ctz): New function.
1138 * tree.h (tree_ctz): New prototype.
1139 * tree-ssanames.h (get_range_info, get_nonzero_bits): Change
1140 first argument from tree to const_tree.
1141 * tree-ssanames.c (get_range_info, get_nonzero_bits): Likewise.
1142 * tree-vectorizer.h (vect_generate_tmps_on_preheader): New prototype.
1143 * tree-vect-loop-manip.c (vect_generate_tmps_on_preheader): No longer
1145 * expr.c (highest_pow2_factor): Reimplemented using tree_ctz.
1146 * tree-vect-loop.c (vect_analyze_loop_operations,
1147 vect_transform_loop): Don't force scalar loop for bound just because
1148 number of iterations is unknown, only do it if it is not known to be
1149 a multiple of vectorization_factor.
1150 * builtins.c (get_object_alignment_2): Use tree_ctz on offset.
1152 * gimple-pretty-print.c (dump_ssaname_info): Print newline also
1153 in case of VR_VARYING. Print get_nonzero_bits if not all ones.
1154 * tree-ssanames.h (struct range_info_def): Add nonzero_bits field.
1155 (set_nonzero_bits, get_nonzero_bits): New prototypes.
1156 * tree-ssa-ccp.c (get_default_value): Use get_range_info to see if
1157 a default def isn't partially constant.
1158 (ccp_finalize): If after IPA, set_range_info if integral SSA_NAME
1159 is known to be partially zero.
1160 (evaluate_stmt): If we'd return otherwise VARYING, use get_range_info
1161 to see if a default def isn't partially constant.
1162 * tree-ssanames.c (set_range_info): Initialize nonzero_bits upon
1163 creation of a range, if VR_RANGE, try to improve nonzero_bits from
1165 (set_nonzero_bits, get_nonzero_bits): New functions.
1167 * tree-cfg.c (assert_unreachable_fallthru_edge_p): New function.
1168 * tree-cfg.h (assert_unreachable_fallthru_edge_p): New prototype.
1169 * tree-vrp.c (all_imm_uses_in_stmt_or_feed_cond): New function.
1170 (remove_range_assertions): If ASSERT_EXPR_VAR has no other immediate
1171 uses but in the condition and ASSERT_EXPR and the other successor of
1172 the predecessor bb is __builtin_unreachable (), set_range_info of the
1173 ASSERT_EXPR_VAR to the range info of the ASSERT_EXPR's lhs.
1175 2013-10-31 Martin Jambor <mjambor@suse.cz>
1177 PR rtl-optimization/58934
1179 2013-10-30 Martin Jambor <mjambor@suse.cz>
1180 PR rtl-optimization/10474
1181 * ira.c (find_moveable_pseudos): Do not calculate dominance info
1183 (interesting_dest_for_shprep): New function.
1184 (split_live_ranges_for_shrink_wrap): Likewise.
1185 (ira): Calculate dominance info and df analysis. Call
1186 split_live_ranges_for_shrink_wrap.
1188 2013-10-31 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
1189 Yury Gribov <y.gribov@samsung.com>
1192 * asan.c (asan_clear_shadow): Allocate a new vreg for temporary
1193 shadow pointer to avoid clobbering the main one.
1195 2013-10-31 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1197 * lower-subreg.c (resolve_simple_move): Copy REG_INC note.
1199 2013-10-30 Vladimir Makarov <vmakarov@redhat.com>
1202 * ira-color.c (update_costs_from_copies): Add new parameter. Use
1203 it for calling update_costs_from_allocno.
1204 (assign_hard_reg): Call restore_costs_from_copies only for
1205 !retry_p. Pass new argument to update_costs_from_copies.
1206 (color_pass): Pass new argument to update_costs_from_copies.
1207 (ira_mark_allocation_change): Ditto.
1209 2013-10-30 Sharad Singhai <singhai@google.com>
1212 * opts.c (common_handle_option): Remove deprecated option
1213 -ftree-vectorizer-verbose.
1214 * doc/invoke.texi (Debugging Options): Ditto.
1215 * opts-global.c (handle_common_deferred_options): Ditto.
1216 (dump_remap_tree_vectorizer_verbose): Delete.
1217 * common.opt: Set -ftree-vectorizer-verbose as an ignored option.
1219 2013-10-30 DJ Delorie <dj@redhat.com>
1221 * config/rx/rx.c (ADD_RX_BUILTIN0): New macro, used for builtins
1222 that take no arguments.
1224 2013-10-30 Joern Rennecke <joern.rennecke@embecosm.com>
1227 * reload1.c (update_eliminables_and_spill): New function, broken
1229 (reload): Use it. Check for frame size change after frame size
1230 alignment, and call update_eliminables_and_spill first if continue-ing.
1232 2013-10-30 Cong Hou <congh@google.com>
1235 * config/i386/i386-protos.h (ix86_expand_sse2_abs): New function.
1236 * config/i386/i386.c (ix86_expand_sse2_abs): New function.
1237 * config/i386/sse.md: Add SSE2 support to abs (8/16/32-bit-int).
1239 2013-10-18 Mikael Pettersson <mikpelinux@gmail.com>
1241 PR rtl-optimization/58369
1242 * reload1.c (compute_reload_subreg_offset): New function.
1243 (choose_reload_regs): Use it to pass endian-correct
1244 offset to subreg_regno_offset.
1246 2013-10-30 Tobias Burnus <burnus@net-b.de>
1249 * tree-cfg.c (replace_loop_annotate): Replace warning by
1252 2013-10-30 Jason Merrill <jason@redhat.com>
1254 * configure.ac (loose_warn): Add -Wno-format if
1255 --disable-build-format-warnings.
1257 2013-10-30 David Malcolm <dmalcolm@redhat.com>
1259 * cgraphunit.c (analyze_functions): Split symtab_node declarations
1260 onto multiple lines to make things easier for rename_symtab.py.
1262 * symtab.c (symtab_dissolve_same_comdat_group_list): Likewise.
1263 (symtab_semantically_equivalent_p): Likewise.
1265 2013-10-30 Vladimir Makarov <vmakarov@redhat.com>
1268 * lra.c (check_rtl): Remove address check before LRA work.
1270 2013-10-30 Marc Glisse <marc.glisse@inria.fr>
1272 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Look for a
1273 POINTER_PLUS_EXPR in the defining statement.
1275 2013-10-30 Vladimir Makarov <vmakarov@redhat.com>
1277 * regmove.c: Remove.
1278 * tree-pass.h (make_pass_regmove): Remove.
1279 * timevar.def (TV_REGMOVE): Remove.
1280 * passes.def (pass_regmove): Remove.
1281 * opts.c (default_options_table): Remove entry for regmove.
1282 * doc/passes.texi: Remove regmove pass description.
1283 * doc/invoke.texi (-foptimize-register-move, -fregmove): Remove
1285 (-fdump-rtl-regmove): Ditto.
1286 * common.opt (foptimize-register-move, fregmove): Ignore.
1287 * Makefile.in (OBJS): Remove regmove.o.
1288 * regmove.c: Remove.
1289 * ira-int.h (struct ira_allocno_pref, ira_pref_t): New structure
1291 (struct ira_allocno) New member allocno_prefs.
1292 (ALLOCNO_PREFS): New macro.
1293 (ira_prefs, ira_prefs_num): New external vars.
1294 (ira_setup_alts, ira_get_dup_out_num, ira_debug_pref): New prototypes.
1295 (ira_debug_prefs, ira_debug_allocno_prefs, ira_create_pref): Ditto.
1296 (ira_add_allocno_pref, ira_remove_pref, ira_remove_allocno_prefs):
1298 (ira_add_allocno_copy_to_list): Remove prototype.
1299 (ira_swap_allocno_copy_ends_if_necessary): Ditto.
1300 (ira_pref_iterator): New type.
1301 (ira_pref_iter_init, ira_pref_iter_cond): New functions.
1302 (FOR_EACH_PREF): New macro.
1303 * ira.c (commutative_constraint_p): Move from ira-conflicts.c.
1304 (ira_get_dup_out_num): Ditto. Rename from get_dup_num. Modify the
1306 (ira_setup_alts): New function.
1307 (decrease_live_ranges_number): New function.
1308 (ira): Call the above function.
1309 * ira-build.c (ira_prefs, ira_prefs_num): New global vars.
1310 (ira_create_allocno): Initialize allocno prefs.
1311 (pref_pool, pref_vec): New static vars.
1312 (initiate_prefs, find_allocno_pref, ira_create_pref): New
1314 (add_allocno_pref_to_list, ira_add_allocno_pref, print_pref): Ditto.
1315 (ira_debug_pref, print_prefs, ira_debug_prefs): Ditto.
1316 (print_allocno_prefs, ira_debug_allocno_prefs, finish_pref): Ditto.
1317 (ira_remove_pref, ira_remove_allocno_prefs, finish_prefs): Ditto.
1318 (ira_add_allocno_copy_to_list): Make static. Rename to
1319 add_allocno_copy_to_list.
1320 (ira_swap_allocno_copy_ends_if_necessary): Make static. Rename to
1321 swap_allocno_copy_ends_if_necessary.
1322 (remove_unnecessary_allocnos, remove_low_level_allocnos): Call
1323 ira_remove_allocno_prefs.
1324 (ira_flattening): Ditto.
1325 (ira_build): Call initiate_prefs, print_prefs.
1326 (ira_destroy): Call finish_prefs.
1327 * ira-color.c (struct update_cost_record): New.
1328 (struct allocno_color_data): Add new member update_cost_records.
1329 (update_cost_record_pool): New static var.
1330 (init_update_cost_records, get_update_cost_record): New functions.
1331 (free_update_cost_record_list, finish_update_cost_records): Ditto.
1332 (struct update_cost_queue_elem): Add member from.
1333 (initiate_cost_update): Call init_update_cost_records.
1334 (finish_cost_update): Call finish_update_cost_records.
1335 (queue_update_cost, get_next_update_cost): Add new param from.
1336 (Update_allocno_cost, update_costs_from_allocno): New functions.
1337 (update_costs_from_prefs): Ditto.
1338 (update_copy_costs): Rename to update_costs_from_copies.
1339 (restore_costs_from_copies): New function.
1340 (update_conflict_hard_regno_costs): Don't go back.
1341 (assign_hard_reg): Call restore_costs_from_copies. Add printing
1343 (pop_allocnos): Add priniting more debug info.
1344 (color_allocnos): Remove prefs for conflicting hard regs.
1345 Call update_costs_from_prefs.
1346 * ira-conflicts.c (commutative_constraint_p): Move to ira.c
1347 (get_dup_num): Rename, modify, and move to ira.c
1348 (process_regs_for_copy): Add prefs.
1349 (add_insn_allocno_copies): Put src as first arg of
1350 process_regs_for_copy. Remove dead code. Call ira_setup_alts.
1351 * ira-costs.c (record_reg_classes): Modify and move code into
1352 record_operands_costs.
1353 (find_costs_and_classes): Create prefs for the hard reg of small
1355 (process_bb_node_for_hard_reg_moves): Add prefs.
1357 2013-10-30 Richard Biener <rguenther@suse.de>
1360 * basic-block.h (pre_and_rev_post_order_compute_fn): New function.
1361 * cfganal.c (pre_and_rev_post_order_compute_fn): New function
1363 (pre_and_rev_post_order_compute): ... which now wraps it.
1364 * graph.c (draw_cfg_nodes_no_loops): Use
1365 pre_and_rev_post_order_compute_fn to avoid ICEing and dependence
1368 2013-10-30 Christian Bruel <christian.bruel@st.com>
1370 * config/sh/sh-mem.cc (sh_expand_cmpnstr): New function.
1371 (sh_expand_cmpstr): Handle known align and schedule improvements.
1372 * config/sh/sh-protos.h (sh_expand_cmpstrn): Declare.
1373 * config/sh/sh.md (cmpstrnsi): New pattern.
1375 2013-10-30 Martin Jambor <mjambor@suse.cz>
1377 PR rtl-optimization/10474
1378 * ira.c (find_moveable_pseudos): Do not calculate dominance info
1380 (interesting_dest_for_shprep): New function.
1381 (split_live_ranges_for_shrink_wrap): Likewise.
1382 (ira): Calculate dominance info and df analysis. Call
1383 split_live_ranges_for_shrink_wrap.
1385 2013-10-30 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1388 * config/arm/arm.c (arm_expand_epilogue_apcs_frame): Emit blockage.
1390 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
1392 * tree-core.h (tree_index): Add TI_POINTER_BOUNDS_TYPE.
1393 * tree.h (POINTER_BOUNDS_P): New.
1394 (BOUNDED_TYPE_P): New.
1396 (pointer_bounds_type_node): New.
1397 * tree.c (build_common_tree_nodes): Initialize
1398 pointer_bounds_type_node.
1399 * gimple.h (gimple_call_get_nobnd_arg_index): New.
1400 (gimple_call_num_nobnd_args): New.
1401 (gimple_call_nobnd_arg): New.
1402 (gimple_return_retbnd): New.
1403 (gimple_return_set_retbnd): New
1404 * gimple.c (gimple_build_return): Increase number of ops
1405 for return statement.
1406 (gimple_call_get_nobnd_arg_index): New.
1407 * gimple-pretty-print.c (dump_gimple_return): Print second op.
1409 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
1411 * ipa.c (cgraph_build_static_cdtor_1): Support contructors
1412 with "chkp ctor" and "bnd_legacy" attributes.
1413 * gimplify.c (gimplify_init_constructor): Avoid infinite
1414 loop during gimplification of bounds initializer.
1416 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
1418 * c-family/c-common.c (handle_bnd_variable_size_attribute): New.
1419 (handle_bnd_legacy): New.
1420 (c_common_attribute_table): Add bnd_variable_size and bnd_legacy.
1421 * doc/extend.texi: Document bnd_variable_size and bnd_legacy
1424 2013-10-29 Ilya Enkovich <ilya.enkovich@intel.com>
1426 * builtin-types.def (BT_FN_VOID_CONST_PTR): New.
1427 (BT_FN_PTR_CONST_PTR): New.
1428 (BT_FN_CONST_PTR_CONST_PTR): New.
1429 (BT_FN_PTR_CONST_PTR_SIZE): New.
1430 (BT_FN_PTR_CONST_PTR_CONST_PTR): New.
1431 (BT_FN_VOID_PTRPTR_CONST_PTR): New.
1432 (BT_FN_VOID_CONST_PTR_SIZE): New.
1433 (BT_FN_PTR_CONST_PTR_CONST_PTR_SIZE): New.
1434 * chkp-builtins.def: New.
1435 * builtins.def: include chkp-builtins.def.
1436 (DEF_CHKP_BUILTIN): New.
1437 * builtins.c (expand_builtin): Support BUILT_IN_CHKP_INIT_PTR_BOUNDS,
1438 BUILT_IN_CHKP_NULL_PTR_BOUNDS, BUILT_IN_CHKP_COPY_PTR_BOUNDS,
1439 BUILT_IN_CHKP_CHECK_PTR_LBOUNDS, BUILT_IN_CHKP_CHECK_PTR_UBOUNDS,
1440 BUILT_IN_CHKP_CHECK_PTR_BOUNDS, BUILT_IN_CHKP_SET_PTR_BOUNDS,
1441 BUILT_IN_CHKP_NARROW_PTR_BOUNDS, BUILT_IN_CHKP_STORE_PTR_BOUNDS,
1442 BUILT_IN_CHKP_GET_PTR_LBOUND, BUILT_IN_CHKP_GET_PTR_UBOUND,
1443 BUILT_IN_CHKP_BNDMK, BUILT_IN_CHKP_BNDSTX, BUILT_IN_CHKP_BNDCL,
1444 BUILT_IN_CHKP_BNDCU, BUILT_IN_CHKP_BNDLDX, BUILT_IN_CHKP_BNDRET,
1445 BUILT_IN_CHKP_INTERSECT, BUILT_IN_CHKP_ARG_BND, BUILT_IN_CHKP_NARROW,
1446 BUILT_IN_CHKP_EXTRACT_LOWER, BUILT_IN_CHKP_EXTRACT_UPPER.
1447 * common.opt (fcheck-pointer-bounds): New.
1448 * toplev.c (process_options): Check Pointer Bounds Checker is
1450 * doc/extend.texi: Document Pointer Bounds Checker built-in functions.
1452 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
1454 * target.def (builtin_chkp_function): New.
1455 (chkp_bound_type): New.
1456 (chkp_bound_mode): New.
1457 (fn_abi_va_list_bounds_size): New.
1458 (load_bounds_for_arg): New.
1459 (store_bounds_for_arg): New.
1460 * targhooks.h (default_load_bounds_for_arg): New.
1461 (default_store_bounds_for_arg): New.
1462 (default_fn_abi_va_list_bounds_size): New.
1463 (default_chkp_bound_type): New.
1464 (default_chkp_bound_mode): New.
1465 (default_builtin_chkp_function): New.
1466 * targhooks.c (default_load_bounds_for_arg): New.
1467 (default_store_bounds_for_arg): New.
1468 (default_fn_abi_va_list_bounds_size): New.
1469 (default_chkp_bound_type): New.
1470 (default_chkp_bound_mode); New.
1471 (default_builtin_chkp_function): New.
1472 * doc/tm.texi.in (TARGET_FN_ABI_VA_LIST_BOUNDS_SIZE): New.
1473 (TARGET_LOAD_BOUNDS_FOR_ARG): New.
1474 (TARGET_STORE_BOUNDS_FOR_ARG): New.
1475 (TARGET_BUILTIN_CHKP_FUNCTION): New.
1476 (TARGET_CHKP_BOUND_TYPE): New.
1477 (TARGET_CHKP_BOUND_MODE): New.
1478 * doc/tm.texi: Regenerated.
1479 * langhooks.h (lang_hooks): Add chkp_supported field.
1480 * langhooks-def.h (LANG_HOOKS_CHKP_SUPPORTED): New.
1481 (LANG_HOOKS_INITIALIZER); Add LANG_HOOKS_CHKP_SUPPORTED.
1483 2013-10-29 Andrew Pinski <apinski@cavium.com>
1485 * tree-ssa-ifcombine.c: Include rtl.h and tm_p.h.
1486 (ifcombine_ifandif): Handle cases where maybe_fold_and_comparisons
1487 fails, combining the branches anyways.
1488 (tree_ssa_ifcombine): Inverse the order of the basic block walk,
1489 increases the number of combinings.
1490 * gimple.h (gsi_start_nondebug_after_labels_bb): New function.
1492 2013-10-29 Mike Stump <mikestump@comcast.net>
1494 * machmode.def (PARTIAL_INT_MODE): Add precision and name.
1495 * genmodes.c (PARTIAL_INT_MODE): Add precision and name.
1496 (make_vector_mode): Increase namebuf to 16.
1497 (emit_insn_modes_h): When processing BImode, don't
1498 also match partial int modes.
1499 (emit_class_narrowest_mode): Likewise.
1501 * config/bfin/bfin-modes.def: Add precision to PDI.
1502 * config/m32c/m32c-modes.def: Add precision to PSI.
1503 * config/msp430/msp430-modes.def: Add precision to PSI.
1504 * config/rs6000/rs6000-modes.def: Add precision to PTI.
1505 * config/sh/sh-modes.def: Add precision to PSI and PDI.
1507 2013-10-29 Oleg Endo <olegendo@gcc.gnu.org>
1510 * config/sh/sh.md (*addc): Rename existing variations to ...
1511 (*addc_r_r_1, *addc_2r_1, *addc_r_1): ... these.
1512 (*addc_r_lsb, *addc_r_r_lsb, *addc_r_lsb_r, *addc_2r_lsb, *addc_r_msb,
1513 *addc_r_r_msb, *addc_2r_msb): New insn_and_split patterns.
1514 * config/sh/sh.c (addsubcosts): Handle some addc special cases.
1516 2013-10-29 Teresa Johnson <tejohnson@google.com>
1519 * tree-ssa-tail-merge.c (replace_block_by): Tolerate profile
1520 insanities when updating probabilities.
1522 2013-10-29 David Malcolm <dmalcolm@redhat.com>
1524 * gdbhooks.py (CGraphNodePrinter.to_string): Update gdb
1525 prettyprinter for cgraph_node to reflect the conversion of the
1526 symtable types to a C++ class hierarchy: it now *is* a
1527 symtab_node_base, rather than having one (named "symbol").
1529 2013-10-29 Balaji V. Iyer <balaji.v.iyer@intel.com>
1531 * builtins.c (is_builtin_name): Added a check for __cilkrts_detach and
1532 __cilkrts_pop_frame. If matched, then return true for built-in
1534 (expand_builtin): Added BUILT_IN_CILK_DETACH and
1535 BUILT_IN_CILK_POP_FRAME case.
1536 * langhooks-def.h (lhd_install_body_with_frame_cleanup): New prototype.
1537 (lhs_cilk_detect_spawn): Likewise.
1538 (LANG_HOOKS_DECLS): Added LANG_HOOKS_CILKPLUS.
1539 (LANG_HOOKS_CILKPLUS_DETECT_SPAWN_AND_UNWRAP): New #define.
1540 (LANG_HOOKS_CILKPLUS_FRAME_CLEANUP): Likewise.
1541 (LANG_HOOKS_CILKPLUS_GIMPLIFY_SPAWN): Likewise.
1542 (LANG_HOOKS_CILKPLUS): Likewise.
1543 * tree.h (CILK_SPAWN_FN): Likewise.
1544 * builtin.def (DEF_CILK_BUILTIN_STUB): Likewise.
1545 * Makefile.in (C_COMMON_OBJS): Added c-family/cilk.o.
1546 (OBJS): Added cilk-common.o.
1547 (BUILTINS_DEF): Added cilk-builtins.def.
1548 * langhooks.c (lhd_install_body_with_frame_cleanup): New function.
1549 (lhd_cilk_detect_spawn): Likewise.
1550 * langhooks.h (lang_hooks_for_cilkplus): New struct.
1551 (struct lang_hooks): Added new field called "cilkplus."
1552 * cilk-common.c: New file.
1554 * cilk-builtins.def: Likewise.
1555 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Added
1556 "__cilk" macro and set it to 200.
1557 * function.h (struct function::cilk_frame_decl): New field.
1558 (struct function::is_cilk_function): Likewise.
1559 (struct function::calls_cilk_spawn): Likewise.
1560 * gimplify.c (gimplify_call_expr): Added a check if the function call
1561 being gimplified is a spawn detach point. If so, then add pop_frame
1562 and detach function calls.
1563 (gimplify_expr): Added a CILK_SPAWN_STMT and CILK_SYNC_STMT case
1564 for gimplifying _Cilk_spawn and _Cilk_sync statements.
1565 (gimplify_return_expr): Added a check for _Cilk_spawn usage in
1566 function. If so, added a _Cilk_sync and gimplified it.
1567 (gimplify_modify_expr): Added a check for _Cilk_spawn in MODIFY and
1568 INIT_EXPRs. If so, then call gimplify_cilk_spawn.
1569 * ipa-inline-analysis (initialize_inline_failed): Prevent inlining of
1571 (can_inline_edge_p): Prevent inling of spawnee function.
1572 * ira.c (ira_setup_eliminable_regset): Force usage of frame pointer
1573 for functions that use Cilk keywords.
1574 * tree-inline.h (struct copy_body_data::remap_var_for_cilk): New field.
1575 * tree-pretty-print.c (dump_generic_node): Added CILK_SPAWN_STMT and
1576 CILK_SYNC_STMT cases.
1577 * tree.def (DEFTREECODE): Added CILK_SPAWN_STMT and CILK_SYNC_STMT
1579 * generic.texi (CILK_SPAWN_STMT): Added documentation for _Cilk_spawn.
1580 (CILK_SYNC_STMT): Added documentation for _Cilk_sync.
1581 * passes.texi (Cilk Keywords): New section that describes the compiler
1582 code changes for handling Cilk Keywords.
1584 2013-10-29 David Malcolm <dmalcolm@redhat.com>
1586 Patch autogenerated by refactor_symtab.py from
1587 https://github.com/davidmalcolm/gcc-refactoring-scripts
1588 revision 58bb219cc090b2f4516a9297d868c245495ee622
1590 * asan.c (asan_finish_file): Update for conversion of symtab types to
1591 a true class hierarchy.
1592 * cfgexpand.c (estimated_stack_frame_size): Likewise.
1593 * cgraph.c (cgraph_get_body): Likewise.
1594 (cgraph_get_create_real_symbol_node): Likewise.
1595 (verify_cgraph_node): Likewise.
1596 (verify_edge_corresponds_to_fndecl): Likewise.
1597 (verify_edge_count_and_frequency): Likewise.
1598 (cgraph_will_be_removed_from_program_if_no_direct_calls): Likewise.
1599 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
1600 (cgraph_can_remove_if_no_direct_calls_and_refs_p): Likewise.
1601 (cgraph_node_cannot_return): Likewise.
1602 (cgraph_set_pure_flag_1): Likewise.
1603 (cgraph_set_const_flag_1): Likewise.
1604 (cgraph_set_nothrow_flag_1): Likewise.
1605 (cgraph_make_node_local_1): Likewise.
1606 (cgraph_for_node_and_aliases): Likewise.
1607 (cgraph_for_node_thunks_and_aliases): Likewise.
1608 (cgraph_node_can_be_local_p): Likewise.
1609 (cgraph_node_cannot_be_local_p_1): Likewise.
1610 (cgraph_function_body_availability): Likewise.
1611 (dump_cgraph_node): Likewise.
1612 (cgraph_rtl_info): Likewise.
1613 (cgraph_mark_address_taken_node): Likewise.
1614 (cgraph_remove_node): Likewise.
1615 (cgraph_release_function_body): Likewise.
1616 (cgraph_update_edges_for_call_stmt_node): Likewise.
1617 (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
1618 (cgraph_make_edge_direct): Likewise.
1619 (cgraph_resolve_speculation): Likewise.
1620 (cgraph_speculative_call_info): Likewise.
1621 (cgraph_turn_edge_to_speculative): Likewise.
1622 (cgraph_create_edge_1): Likewise.
1623 (cgraph_set_call_stmt): Likewise.
1624 (cgraph_node_for_asm): Likewise.
1625 (cgraph_add_thunk): Likewise.
1626 (cgraph_same_body_alias): Likewise.
1627 (cgraph_create_function_alias): Likewise.
1628 (cgraph_create_node): Likewise.
1629 (cgraph_create_empty_node): Likewise.
1630 (record_function_versions): Likewise.
1631 (used_from_object_file_p): Likewise.
1632 * cgraph.h (symtab_can_be_discarded): Likewise.
1633 (symtab_real_symbol_p): Likewise.
1634 (cgraph_mark_force_output_node): Likewise.
1635 (cgraph_edge_recursive_p): Likewise.
1636 (symtab_alias_target): Likewise.
1637 (varpool_all_refs_explicit_p): Likewise.
1638 (varpool_can_remove_if_no_refs): Likewise.
1639 (cgraph_only_called_directly_or_aliased_p): Likewise.
1640 (cgraph_next_function_with_gimple_body): Likewise.
1641 (cgraph_first_function_with_gimple_body): Likewise.
1642 (cgraph_function_with_gimple_body_p): Likewise.
1643 (cgraph_next_function): Likewise.
1644 (cgraph_first_function): Likewise.
1645 (cgraph_next_defined_function): Likewise.
1646 (cgraph_first_defined_function): Likewise.
1647 (varpool_next_defined_variable): Likewise.
1648 (varpool_first_defined_variable): Likewise.
1649 (varpool_next_static_initializer): Likewise.
1650 (varpool_first_static_initializer): Likewise.
1651 (varpool_next_variable): Likewise.
1652 (varpool_first_variable): Likewise.
1653 (varpool_node_name): Likewise.
1654 (varpool): Likewise.
1656 (is_a_helper <varpool_node>::test): Likewise.
1657 (is_a_helper <cgraph_node>::test): Likewise.
1658 (varpool_variable_node): Likewise.
1659 (cgraph_function_or_thunk_node): Likewise.
1660 (varpool_alias_target): Likewise.
1661 (cgraph_alias_target): Likewise.
1662 (cgraph_node_name): Likewise.
1663 (varpool_node_asm_name): Likewise.
1664 (cgraph_node_asm_name): Likewise.
1665 * cgraphbuild.c (remove_cgraph_callee_edges): Likewise.
1666 (cgraph_rebuild_references): Likewise.
1667 (rebuild_cgraph_edges): Likewise.
1668 (record_eh_tables): Likewise.
1669 (build_cgraph_edges): Likewise.
1670 (mark_store): Likewise.
1671 (mark_load): Likewise.
1672 (mark_address): Likewise.
1673 (record_type_list): Likewise.
1674 (record_reference): Likewise.
1675 * cgraphclones.c (cgraph_materialize_all_clones): Likewise.
1676 (cgraph_materialize_clone): Likewise.
1677 (cgraph_function_versioning): Likewise.
1678 (cgraph_copy_node_for_versioning): Likewise.
1679 (update_call_expr): Likewise.
1680 (cgraph_find_replacement_node): Likewise.
1681 (cgraph_create_virtual_clone): Likewise.
1682 (cgraph_clone_node): Likewise.
1683 * cgraphunit.c (compile): Likewise.
1684 (output_weakrefs): Likewise.
1685 (output_in_order): Likewise.
1686 (expand_function): Likewise.
1687 (assemble_thunks_and_aliases): Likewise.
1688 (expand_thunk): Likewise.
1689 (mark_functions_to_output): Likewise.
1690 (handle_alias_pairs): Likewise.
1691 (analyze_functions): Likewise.
1692 (walk_polymorphic_call_targets): Likewise.
1693 (varpool_finalize_decl): Likewise.
1694 (process_function_and_variable_attributes): Likewise.
1695 (cgraph_process_same_body_aliases): Likewise.
1696 (analyze_function): Likewise.
1697 (cgraph_add_new_function): Likewise.
1698 (cgraph_finalize_function): Likewise.
1699 (referred_to_p): Likewise.
1700 (cgraph_reset_node): Likewise.
1701 (cgraph_process_new_functions): Likewise.
1702 (enqueue_node): Likewise.
1703 (decide_is_symbol_needed): Likewise.
1704 * coverage.c (coverage_compute_profile_id): Likewise.
1705 * dbxout.c (dbxout_expand_expr): Likewise.
1706 * dwarf2out.c (premark_types_used_by_global_vars_helper): Likewise.
1707 (reference_to_unused): Likewise.
1708 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
1709 * gimplify.c (unvisit_body): Likewise.
1710 (unshare_body): Likewise.
1711 * ipa-cp.c (ipcp_generate_summary): Likewise.
1712 (ipcp_decision_stage): Likewise.
1713 (identify_dead_nodes): Likewise.
1714 (decide_whether_version_node): Likewise.
1715 (decide_about_value): Likewise.
1716 (perhaps_add_new_callers): Likewise.
1717 (create_specialized_node): Likewise.
1718 (update_profiling_info): Likewise.
1719 (ipcp_propagate_stage): Likewise.
1720 (estimate_local_effects): Likewise.
1721 (good_cloning_opportunity_p): Likewise.
1722 (devirtualization_time_bonus): Likewise.
1723 (propagate_constants_accross_call): Likewise.
1724 (initialize_node_lattices): Likewise.
1725 (ipcp_cloning_candidate_p): Likewise.
1726 (determine_versionability): Likewise.
1727 (print_all_lattices): Likewise.
1728 (print_lattice): Likewise.
1729 (ipcp_discover_new_direct_edges): Likewise.
1730 * ipa-devirt.c (ipa_devirt): Likewise.
1731 (likely_target_p): Likewise.
1732 (update_type_inheritance_graph): Likewise.
1733 (possible_polymorphic_call_target_p): Likewise.
1734 (dump_possible_polymorphic_call_targets): Likewise.
1735 (devirt_variable_node_removal_hook): Likewise.
1736 (record_binfo): Likewise.
1737 (maybe_record_node): Likewise.
1738 (build_type_inheritance_graph): Likewise.
1739 * ipa-inline-analysis.c (inline_write_summary): Likewise.
1740 (inline_generate_summary): Likewise.
1741 (inline_analyze_function): Likewise.
1742 (do_estimate_growth): Likewise.
1743 (simple_edge_hints): Likewise.
1744 (estimate_node_size_and_time): Likewise.
1745 (estimate_edge_devirt_benefit): Likewise.
1746 (compute_inline_parameters): Likewise.
1747 (estimate_function_body_sizes): Likewise.
1748 (compute_bb_predicates): Likewise.
1749 (initialize_inline_failed): Likewise.
1750 (dump_inline_summary): Likewise.
1751 (dump_inline_edge_summary): Likewise.
1752 * ipa-inline-transform.c (inline_transform): Likewise.
1753 (preserve_function_body_p): Likewise.
1754 (save_inline_function_body): Likewise.
1755 (inline_call): Likewise.
1756 (clone_inlined_nodes): Likewise.
1757 (can_remove_node_now_p): Likewise.
1758 (can_remove_node_now_p_1): Likewise.
1759 * ipa-inline.c (early_inliner): Likewise.
1760 (early_inline_small_functions): Likewise.
1761 (inline_always_inline_functions): Likewise.
1762 (ipa_inline): Likewise.
1763 (flatten_function): Likewise.
1764 (inline_small_functions): Likewise.
1765 (speculation_useful_p): Likewise.
1766 (recursive_inlining): Likewise.
1767 (update_caller_keys): Likewise.
1768 (reset_edge_caches): Likewise.
1769 (update_edge_key): Likewise.
1770 (edge_badness): Likewise.
1771 (relative_time_benefit): Likewise.
1772 (want_inline_self_recursive_call_p): Likewise.
1773 (want_inline_small_function_p): Likewise.
1774 (want_early_inline_function_p): Likewise.
1775 (num_calls): Likewise.
1776 (can_early_inline_edge_p): Likewise.
1777 (can_inline_edge_p): Likewise.
1778 (report_inline_failed_reason): Likewise.
1779 * ipa-profile.c (ipa_profile): Likewise.
1780 (ipa_propagate_frequency): Likewise.
1781 (ipa_propagate_frequency_1): Likewise.
1782 (ipa_profile_generate_summary): Likewise.
1783 * ipa-prop.c (ipcp_transform_function): Likewise.
1784 (read_replacements_section): Likewise.
1785 (ipa_prop_read_section): Likewise.
1786 (ipa_modify_call_arguments): Likewise.
1787 (ipa_print_node_params): Likewise.
1788 (propagate_controlled_uses): Likewise.
1789 (update_indirect_edges_after_inlining): Likewise.
1790 (remove_described_reference): Likewise.
1791 (ipa_make_edge_direct_to_target): Likewise.
1792 (ipa_analyze_node): Likewise.
1793 (ipa_analyze_params_uses): Likewise.
1794 (ipa_compute_jump_functions): Likewise.
1795 (ipa_get_callee_param_type): Likewise.
1796 (ipa_print_node_jump_functions): Likewise.
1797 (ipa_initialize_node_params): Likewise.
1798 (ipa_populate_param_decls): Likewise.
1799 (ipa_func_spec_opts_forbid_analysis_p): Likewise.
1800 (write_agg_replacement_chain): Likewise.
1801 (ipa_write_node_info): Likewise.
1802 (ipa_edge_duplication_hook): Likewise.
1803 (try_decrement_rdesc_refcount): Likewise.
1804 * ipa-pure-const.c (propagate_nothrow): Likewise.
1805 (propagate_pure_const): Likewise.
1806 (pure_const_read_summary): Likewise.
1807 (pure_const_write_summary): Likewise.
1808 (analyze_function): Likewise.
1809 * ipa-ref-inline.h (ipa_ref_referred_ref_list): Likewise.
1810 (ipa_ref_referring_ref_list): Likewise.
1811 * ipa-ref.c (ipa_clear_stmts_in_references): Likewise.
1812 (ipa_remove_stmt_references): Likewise.
1813 (ipa_find_reference): Likewise.
1814 (ipa_dump_referring): Likewise.
1815 (ipa_dump_references): Likewise.
1816 (ipa_record_reference): Likewise.
1817 * ipa-reference.c (ipa_reference_read_optimization_summary): Likewise.
1818 (ipa_reference_write_optimization_summary): Likewise.
1819 (write_node_summary_p): Likewise.
1820 (propagate): Likewise.
1821 (read_write_all_from_decl): Likewise.
1822 (generate_summary): Likewise.
1823 (analyze_function): Likewise.
1824 (propagate_bits): Likewise.
1825 (ipa_reference_get_not_written_global): Likewise.
1826 (ipa_reference_get_not_read_global): Likewise.
1827 * ipa-split.c (execute_split_functions): Likewise.
1828 (split_function): Likewise.
1829 * ipa-utils.c (ipa_merge_profiles): Likewise.
1830 (dump_cgraph_node_set): Likewise.
1831 (ipa_reverse_postorder): Likewise.
1832 (ipa_edge_within_scc): Likewise.
1833 (ipa_get_nodes_in_cycle): Likewise.
1834 (ipa_free_postorder_info): Likewise.
1835 (ipa_reduced_postorder): Likewise.
1836 (searchc): Likewise.
1837 (recursive_call_p): Likewise.
1838 * ipa.c (ipa_cdtor_merge): Likewise.
1839 (record_cdtor_fn): Likewise.
1840 (function_and_variable_visibility): Likewise.
1841 (varpool_externally_visible_p): Likewise.
1842 (cgraph_externally_visible_p): Likewise.
1843 (comdat_can_be_unshared_p): Likewise.
1844 (comdat_can_be_unshared_p_1): Likewise.
1845 (address_taken_from_non_vtable_p): Likewise.
1846 (ipa_discover_readonly_nonaddressable_vars): Likewise.
1847 (symtab_remove_unreachable_nodes): Likewise.
1848 (walk_polymorphic_call_targets): Likewise.
1849 (process_references): Likewise.
1850 (enqueue_node): Likewise.
1851 (has_addr_references_p): Likewise.
1852 (cgraph_non_local_node_p_1): Likewise.
1853 * is-a.h (varpool_analyze_node): Likewise.
1854 * lto-cgraph.c (input_symtab): Likewise.
1855 (merge_profile_summaries): Likewise.
1856 (input_cgraph_1): Likewise.
1857 (input_edge): Likewise.
1858 (input_varpool_node): Likewise.
1859 (input_node): Likewise.
1860 (input_overwrite_node): Likewise.
1861 (compute_ltrans_boundary): Likewise.
1862 (output_refs): Likewise.
1863 (lto_output_varpool_node): Likewise.
1864 (lto_output_node): Likewise.
1865 (reachable_from_other_partition_p): Likewise.
1866 (referenced_from_other_partition_p): Likewise.
1867 (lto_output_edge): Likewise.
1868 (output_node_opt_summary): Likewise.
1869 (add_node_to): Likewise.
1870 (reachable_from_this_partition_p): Likewise.
1871 (lto_set_symtab_encoder_in_partition): Likewise.
1872 (lto_symtab_encoder_in_partition_p): Likewise.
1873 (lto_set_symtab_encoder_encode_initializer): Likewise.
1874 (lto_symtab_encoder_encode_initializer_p): Likewise.
1875 (lto_set_symtab_encoder_encode_body): Likewise.
1876 (lto_symtab_encoder_encode_body_p): Likewise.
1877 * lto-section-in.c (lto_free_function_in_decl_state_for_node):
1879 * lto-streamer-in.c (lto_read_body): Likewise.
1880 (fixup_call_stmt_edges): Likewise.
1881 (fixup_call_stmt_edges_1): Likewise.
1882 * lto-streamer-out.c (produce_symtab): Likewise.
1883 (output_symbol_p): Likewise.
1884 (write_symbol): Likewise.
1885 (lto_output): Likewise.
1886 (copy_function): Likewise.
1887 (output_function): Likewise.
1888 * passes.c (function_called_by_processed_nodes_p): Likewise.
1889 (ipa_write_optimization_summaries): Likewise.
1890 (ipa_write_summaries): Likewise.
1891 (do_per_function_toporder): Likewise.
1892 (do_per_function): Likewise.
1893 (dump_passes): Likewise.
1894 * symtab.c (symtab_semantically_equivalent_p): Likewise.
1895 (symtab_nonoverwritable_alias): Likewise.
1896 (symtab_nonoverwritable_alias_1): Likewise.
1897 (symtab_for_node_and_aliases): Likewise.
1898 (symtab_resolve_alias): Likewise.
1899 (fixup_same_cpp_alias_visibility): Likewise.
1900 (symtab_alias_ultimate_target): Likewise.
1901 (symtab_used_from_object_file_p): Likewise.
1902 (verify_symtab_base): Likewise.
1903 (dump_symtab_base): Likewise.
1904 (symtab_node_name): Likewise.
1905 (symtab_node_asm_name): Likewise.
1906 (symtab_dissolve_same_comdat_group_list): Likewise.
1907 (symtab_add_to_same_comdat_group): Likewise.
1908 (symtab_unregister_node): Likewise.
1909 (symtab_insert_node_to_hashtable): Likewise.
1910 (symtab_register_node): Likewise.
1911 (unlink_from_assembler_name_hash): Likewise.
1912 (insert_to_assembler_name_hash): Likewise.
1913 (eq_assembler_name): Likewise.
1914 (hash_node_by_assembler_name): Likewise.
1915 (eq_node): Likewise.
1916 (hash_node): Likewise.
1917 * toplev.c (wrapup_global_declaration_2): Likewise.
1918 * trans-mem.c (ipa_tm_execute): Likewise.
1919 (ipa_tm_transform_clone): Likewise.
1920 (ipa_tm_transform_transaction): Likewise.
1921 (ipa_tm_transform_calls_redirect): Likewise.
1922 (ipa_tm_insert_gettmclone_call): Likewise.
1923 (ipa_tm_insert_irr_call): Likewise.
1924 (ipa_tm_create_version): Likewise.
1925 (ipa_tm_create_version_alias): Likewise.
1926 (ipa_tm_mark_forced_by_abi_node): Likewise.
1927 (ipa_tm_mark_force_output_node): Likewise.
1928 (ipa_tm_diagnose_tm_safe): Likewise.
1929 (ipa_tm_mayenterirr_function): Likewise.
1930 (ipa_tm_scan_irr_function): Likewise.
1931 (ipa_tm_note_irrevocable): Likewise.
1932 (ipa_tm_scan_calls_clone): Likewise.
1933 (get_cg_data): Likewise.
1934 * tree-eh.c (tree_could_trap_p): Likewise.
1935 * tree-emutls.c (ipa_lower_emutls): Likewise.
1936 (create_emultls_var): Likewise.
1937 (lower_emutls_function_body): Likewise.
1938 (gen_emutls_addr): Likewise.
1939 (emutls_decl): Likewise.
1940 (new_emutls_decl): Likewise.
1941 * tree-inline.c (tree_function_versioning): Likewise.
1942 (optimize_inline_calls): Likewise.
1943 (expand_call_inline): Likewise.
1944 (estimate_num_insns): Likewise.
1945 (copy_bb): Likewise.
1946 (delete_unreachable_blocks_update_callgraph): Likewise.
1947 * tree-nested.c (gimplify_all_functions): Likewise.
1948 (create_nesting_tree): Likewise.
1949 (check_for_nested_with_variably_modified): Likewise.
1950 * tree-pretty-print.c (dump_function_header): Likewise.
1951 * tree-profile.c (tree_profiling): Likewise.
1952 * tree-sra.c (ipa_sra_preliminary_function_checks): Likewise.
1953 (modify_function): Likewise.
1954 (convert_callers): Likewise.
1955 (convert_callers_for_node): Likewise.
1956 * tree-ssa-structalias.c (ipa_pta_execute): Likewise.
1957 (associate_varinfo_to_alias): Likewise.
1958 (create_variable_info_for): Likewise.
1959 (get_constraint_for_ssa_var): Likewise.
1960 * tree-vectorizer.c (increase_alignment): Likewise.
1961 * tree.c (find_decls_types_in_var): Likewise.
1962 (find_decls_types_in_node): Likewise.
1963 (free_lang_data_in_decl): Likewise.
1964 * value-prof.c (gimple_ic_transform): Likewise.
1965 (gimple_ic): Likewise.
1966 (check_ic_target): Likewise.
1967 (init_node_map): Likewise.
1968 * varasm.c (decl_binds_to_current_def_p): Likewise.
1969 (default_binds_local_p_1): Likewise.
1970 (dump_tm_clone_pairs): Likewise.
1971 (assemble_alias): Likewise.
1972 (find_decl): Likewise.
1973 (mark_decl_referenced): Likewise.
1974 * varpool.c (varpool_for_node_and_aliases): Likewise.
1975 (varpool_extra_name_alias): Likewise.
1976 (varpool_create_variable_alias): Likewise.
1977 (add_new_static_var): Likewise.
1978 (varpool_finalize_named_section_flags): Likewise.
1979 (varpool_remove_unreferenced_decls): Likewise.
1980 (enqueue_node): Likewise.
1981 (varpool_assemble_decl): Likewise.
1982 (assemble_aliases): Likewise.
1983 (varpool_analyze_node): Likewise.
1984 (cgraph_variable_initializer_availability): Likewise.
1985 (varpool_add_new_variable): Likewise.
1986 (ctor_for_folding): Likewise.
1987 (dump_varpool_node): Likewise.
1988 (varpool_remove_initializer): Likewise.
1989 (varpool_remove_node): Likewise.
1990 (varpool_node_for_decl): Likewise.
1991 (varpool_create_empty_node): Likewise.
1992 * config/i386/i386.c (ix86_generate_version_dispatcher_body): Likewise.
1993 (ix86_get_function_versions_dispatcher): Likewise.
1995 2013-10-29 David Malcolm <dmalcolm@redhat.com>
1997 * cgraph.h (symtab_node_base): Convert to a class;
1998 add GTY((desc ("%h.type"), tag ("SYMTAB_SYMBOL"))), and take
1999 chain_next/prev from symtab_node_def.
2000 (cgraph_node): Inherit from symtab_node; add GTY option
2001 tag ("SYMTAB_FUNCTION").
2002 (varpool_node): Inherit from symtab_node; add GTY option
2003 tag ("SYMTAB_VARIABLE").
2004 (symtab_node_def): Remove.
2005 (is_a_helper <cgraph_node>::test (symtab_node_def *)): Convert to...
2006 (is_a_helper <cgraph_node>::test (symtab_node_base *)): ...this.
2007 (is_a_helper <varpool_node>::test (symtab_node_def *)): Convert to...
2008 (is_a_helper <varpool_node>::test (symtab_node_base *)): ...this.
2010 * ipa-ref.h (symtab_node_def): Drop.
2011 (symtab_node): Change underlying type from symtab_node_def to
2013 (const_symtab_node): Likwise.
2015 * is-a.h: Update examples in comment.
2017 * symtab.c (symtab_hash): Change symtab_node_def to symtab_node_base.
2018 (assembler_name_hash): Likewise.
2020 2013-10-29 Martin Liska <marxin.liska@gmail.com>
2022 * doc/tree-ssa.texi (gimple_phi_result): Document.
2023 (gimple_phi_num_args, gimple_phi_arg): Likewise.
2024 (gimple_phi_arg_edge, gimple_phi_arg_def): Likewise.
2025 (PHI_RESULT, PHI_NUM_ARGS): Remove.
2026 (PHI_ARG_ELT, PHI_ARG_EDGE, PHI_ARG_DEF): Likewise.
2028 2013-10-29 Andrew MacLeod <amacleod@redhat.com>
2030 * expr.h: Revert change and include tree-core.h.
2031 * rtl.h: Revert change and don't include tree-core.h.
2033 2013-10-29 Andrew MacLeod <amacleod@redhat.com>
2035 * config/darwin.c: Include gimple.h.
2036 * config/i386/winnt.c: Likewise.
2038 2013-10-29 Marc Glisse <marc.glisse@inria.fr>
2040 PR tree-optimization/19831
2041 * tree-ssa-alias.c (stmt_kills_ref_p_1): Handle BUILT_IN_FREE.
2043 2013-10-29 Andrew MacLeod <amacleod@redhat.com>
2045 * tree-outof-ssa.h: Remove include files.
2046 * tree-outof-ssa.c: Add required include files from tree-outof-ssa.h.
2048 * tree-ssa-coalesce.c: Likewise.
2049 * cfgexpand.c: Likewise.
2050 * tree-ssa-ter.c: Likewise.
2051 * ipa-prop.h: Remove gimple.h and tree-core.h from include list.
2052 * lto-streamer.h: Likewise.
2053 * cgraphbuild.c: Add gimple.h to include list.
2054 * data-streamer-in.c: Likewise.
2055 * ipa-cp.c: Likewise.
2056 * tree-streamer.c: Likewise.
2057 * lto-compress.c: Likewise.
2058 * ipa-reference.c: Likewise.
2059 * data-streamer-out.c: Likewise.
2060 * lto-cgraph.c: Likewise.
2061 * cgraphclones.c: Likewise.
2062 * ipa-utils.c: Likewise.
2063 * data-streamer.c: Likewise.
2064 * ipa-split.c: Likewise.
2065 * lto-section-in.c: Likewise.
2066 * tree-streamer-out.c: Likewise.
2067 * ipa-prop.c: Likewise.
2068 * tree-streamer-in.c: Likewise.
2069 * symtab.c: Likewise.
2070 * opts-global.c: Likewise.
2071 * lto-opts.c: Likewise.
2072 * lto-section-out.c: Likewise.
2073 * lto-streamer.c: Likewise.
2074 * rtl.h: Add tree-core.h to include list.
2075 * expr.h: Remove tree-core.h from include list.
2076 * gimple.h: Likewise.
2077 * ipa-utils.h: Likewise.
2078 * streamer-hooks.h: Likewise.
2079 * streamer-hooks.c: Include input.h.
2081 2013-10-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2083 * config/arm/arm.c (cortexa7_extra_costs): New table.
2084 (arm_cortex_a7_tune): New.
2085 * config/arm/arm-cores.def: Use cortex_a7 tuning for cortex-a7.
2087 2013-10-29 Eric Botcazou <ebotcazou@adacore.com>
2089 * expr.c (expand_expr_real_1) <MEM_EXPR>: Eliminate small redundancy.
2091 2013-10-29 David Malcolm <dmalcolm@redhat.com>
2093 * doc/gty.texi ("Inheritance and GTY"): Make it clear that
2094 to use autogenerated markers for a class-hierarchy, every class
2095 must have a GTY marker.
2096 * gengtype.h (struct type): Add linked list of subclasses to
2097 the "s" member of the union.
2098 (add_subclass): New decl.
2099 * gengtype-state.c (read_state_struct_type): Set up subclass
2101 * gengtype.c (get_ultimate_base_class): New.
2102 (add_subclass): New.
2103 (new_structure): Set up subclass linked list.
2104 (set_gc_used_type): Propagate usage information to subclasses.
2105 (output_mangled_typename): Use get_ultimate_base_class.
2106 (walk_subclasses): Use the subclass linked list, avoiding an
2107 O(N^2) when writing out all types.
2108 (walk_type): Issue an error if the base class is missing a tag,
2109 rather than generating bogus C code. Add a gcc_unreachable
2110 default case, in case people omit tags from concrete subclasses,
2111 or get the values wrong.
2112 (write_func_for_structure): Issue an error for subclasses for
2113 which the base doesn't have a "desc", since otherwise the
2114 autogenerated routines for the base would silently fail to visit
2115 any subclass fields.
2116 (write_root): Use get_ultimate_base_class, tweaking constness of
2117 tp to match that function's signature.
2119 2013-10-29 David Malcolm <dmalcolm@redhat.com>
2121 * doc/gty.texi (GTY Options): Add note about inheritance to
2122 description of desc and tag.
2123 (Inheritance and GTY): New.
2125 2013-10-29 David Malcolm <dmalcolm@redhat.com>
2127 * gengtype-parse.c (opts_have): Drop "static" so that
2128 we can use this from gengtype.c.
2129 * gengtype.c (set_gc_used_type): Mark any base class as used;
2130 update field traversal to visit inherited fields.
2131 (output_mangled_typename): Convert references to classes within
2132 an inheritance hierarchy to reference the ultimate base class,
2133 since only it will have gt_ functions.
2134 (get_string_option): New.
2135 (walk_subclasses): New.
2136 (walk_type): Treat GTY structs that have a "desc" as being the
2137 root of an inheritance hierarchy. Generate a switch on it
2138 within the marking function which walks all subclasses, adding
2139 cases for them via walk_subclasses. For subclasses, visit all
2140 fields of the type (including inherited ones).
2141 (write_func_for_structure): Don't write fns for subclasses, only
2142 for the ultimate base class within an inheritance hierarchy.
2143 Subclasses-marking will be handled by the base class marking functions.
2144 (write_types): Likewise.
2145 (write_local_func_for_structure): Likewise.
2146 (USED_BY_TYPED_GC_P): Emit allocators for subclasses that have
2147 a "tag" option (and are thus concrete subclasses).
2148 (write_root): Use the marker function for the ultimate base class.
2149 * gengtype.h (FOR_ALL_INHERITED_FIELDS): New.
2150 (opts_have): Add declaration.
2152 2013-10-28 Vladimir Makarov <vmakarov@redhat.com>
2154 * lra-spills.c (lra_final_code_change): Remove useless move insns
2155 originated from moves of pseudos.
2157 2013-10-28 Jeff Law <law@redhat.com>
2159 * lower-subreg.c (resolve_simple_move): Fix comment typo.
2161 2013-10-28 Trevor Saunders <tsaunders@mozilla.com>
2163 * df-scan.c (df_collection_rec): Adjust.
2164 (copy_defs): New constant.
2165 (copy_uses): Likewise.
2166 (copy_eq_uses): Likewise.
2167 (copy_mw): Likewise.
2168 (copy_all): Likewise.
2169 (df_insn_rescan): Adjust.
2170 (df_notes_rescan): Likewise.
2171 (df_swap_refs): Likewise.
2172 (df_sort_and_compress_refs): Likewise.
2173 (df_sort_and_compress_mws): Likewise.
2174 (df_install_refs): Likewise.
2175 (df_install_mws): Likewise.
2176 (df_refs_add_to_chains): Add flags parameter controlling which vectors
2178 (df_bb_refs_record): Adjust.
2179 (df_record_entry_block_defs): Likewise.
2180 (df_record_exit_block_defs): Likewise.
2181 (df_refs_verify): Likewise.
2182 (df_mws_verify): Likewise.
2183 (df_insn_refs_verify): Likewise.
2184 (df_bb_verify): Likewise.
2185 * ipa-pure-const.c (finish_state): Remove.
2186 (propagate): Adjust.
2187 * tree-data-ref.c tree-ssa-alias.c tree-ssa-loop-ivcanon.c
2188 tree-ssa-threadedge.c tree-vect-loop-manip.c tree-vect-slp.c
2189 var-tracking.c: Adjust.
2190 * vec.c (stack_vecs): Remove.
2191 (register_stack_vec): Likewise.
2192 (stack_vec_register_index): Likewise.
2193 (unregister_stack_vec): Likewise.
2194 * vec.h (struct va_stack): Remove.
2195 (struct vec<T, A, vl_ptr>): Specialize as
2196 struct vec<T, va_heap, vl_ptr> instead since va_heap is the only
2197 allocation strategy compatable with the vl_ptr layout.
2198 (struct vec<T, va_gc, vl_ptr>): Remove because it now gets an empty
2199 specialization anyway.
2200 (class stack_vec): New class.
2201 (vec_stack_alloc): Remove.
2202 (vec<T, va_heap, vl_ptr>::using_auto_storage): New method.
2204 2013-10-28 Alexander Ivchenko <alexander.ivchenko@intel.com>
2205 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2206 Sergey Lega <sergey.s.lega@intel.com>
2207 Anna Tikhonova <anna.tikhonova@intel.com>
2208 Ilya Tocar <ilya.tocar@intel.com>
2209 Andrey Turetskiy <andrey.turetskiy@intel.com>
2210 Ilya Verbin <ilya.verbin@intel.com>
2211 Kirill Yukhin <kirill.yukhin@intel.com>
2212 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2214 * config/i386/i386.md (prefetch): Allow TARGET_AVX512PF.
2215 (*prefetch_avx512pf_<mode>): New.
2216 * config/i386/sse.md (avx512f_vmcmp<mode>3): Ditto.
2217 (avx512f_maskcmp<mode>3): Ditto.
2218 (vashrv16si3): Ditto.
2220 2013-10-28 Alexander Ivchenko <alexander.ivchenko@intel.com>
2221 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2222 Sergey Lega <sergey.s.lega@intel.com>
2223 Anna Tikhonova <anna.tikhonova@intel.com>
2224 Ilya Tocar <ilya.tocar@intel.com>
2225 Andrey Turetskiy <andrey.turetskiy@intel.com>
2226 Ilya Verbin <ilya.verbin@intel.com>
2227 Kirill Yukhin <kirill.yukhin@intel.com>
2228 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2230 * config/i386/i386.md (any_truncate): New.
2231 (trunsuffix): Ditto.
2232 * config/i386/predicates.md (const_8_to_9_operand): New.
2233 (const_10_to_11_operand): Ditto.
2234 (const_12_to_13_operand): Ditto.
2235 (const_14_to_15_operand): Ditto.
2236 (const_16_to_19_operand): Ditto.
2237 (const_20_to_23_operand): Ditto.
2238 (const_24_to_27_operand): Ditto.
2239 (const_28_to_31_operand): Ditto.
2240 * config/i386/sse.md (unspec): Add UNSPEC_UNSIGNED_FIX_NOTRUNC.
2241 (cvtusi2<ssescalarmodesuffix>32): New.
2242 (cvtusi2<ssescalarmodesuffix>64): Ditto.
2243 (ufloatv16siv16sf2): Ditto.
2244 (avx512f_fix_notruncv16sfv16si): Ditto.
2245 (avx512f_ufix_notruncv16sfv16si): Ditto.
2246 (avx512f_vcvtss2usi): Ditto.
2247 (avx512f_vcvtss2usiq): Ditto.
2248 (avx512f_vcvttss2usi): Ditto.
2249 (avx512f_vcvttss2usiq): Ditto.
2250 (avx512f_vcvtsd2usi): Ditto.
2251 (avx512f_vcvtsd2usiq): Ditto.
2252 (avx512f_vcvttsd2usi): Ditto.
2253 (avx512f_vcvttsd2usiq): Ditto.
2254 (ufloatv8siv8df): Ditto.
2255 (avx512f_cvtdq2pd512_2): Ditto.
2256 (avx512f_cvtpd2dq512): Ditto.
2257 (avx512f_ufix_notruncv8dfv8si): Ditto.
2258 (avx512f_cvtpd2ps512): Ditto.
2259 (vec_unpacks_lo_v16sf): Ditto.
2260 (vec_unpacks_hi_v16sf): Ditto.
2261 (vec_unpacks_float_hi_v16si): Ditto.
2262 (vec_unpacks_float_lo_v16si): Ditto.
2263 (avx512f_unpckhps512): Ditto.
2264 (avx512f_unpcklps512): Ditto.
2265 (avx512f_movshdup512): Ditto.
2266 (avx512f_movsldup512): Ditto.
2267 (vec_extract_lo_v32hi): Ditto.
2268 (vec_extract_hi_v32hi): Ditto.
2269 (vec_extract_lo_v64qi): Ditto.
2270 (vec_extract_hi_v64qi): Ditto.
2271 (avx512f_unpckhpd512): Ditto.
2272 (avx512f_movddup512): Ditto.
2273 (avx512f_unpcklpd512): Ditto.
2274 (*avx512f_unpcklpd512): Ditto.
2275 (avx512f_shufps512_1): Ditto.
2276 (avx512f_shufpd512_1): Ditto.
2277 (avx512f_interleave_highv8di): Ditto.
2278 (avx512f_interleave_lowv8di): Ditto.
2279 (PMOV_DST_MODE): Ditto.
2280 (pmov_src_mode): Ditto.
2281 (pmov_src_lower): Ditto.
2283 (*avx512f_<code><pmov_src_lower><mode>2): Ditto.
2284 (*avx512f_<code>v8div16qi2): Ditto.
2285 (*avx512f_<code>v8div16qi2_store): Ditto.
2286 (vec_widen_umult_even_v16si): Ditto.
2287 (*vec_widen_umult_even_v16si): Ditto.
2288 (vec_widen_smult_even_v16si): Ditto.
2289 (*vec_widen_smult_even_v16si): Ditto.
2290 (avx512f_interleave_highv16si): Ditto.
2291 (avx512f_interleave_lowv16si): Ditto.
2292 (avx512f_<code>v16qiv16si2): Ditto.
2293 (avx512f_<code>v16hiv16si2): Ditto.
2294 (avx512f_<code>v8qiv8di2): Ditto.
2295 (avx512f_<code>v8hiv8di2): Ditto.
2296 (avx512f_<code>v8siv8di2): Ditto.
2297 (avx512cd_maskb_vec_dupv8di): Ditto.
2298 (avx512cd_maskw_vec_dupv16si): Ditto.
2299 (avx512f_vcvtph2ps512): Ditto.
2300 (avx512f_vcvtps2ph512): Ditto.
2301 (VEC_EXTRACT_MODE): Extened with wider modes.
2302 (VEC_PERM_AVX2): Ditto.
2303 (VEC_PERM_CONST): Ditto.
2305 2013-10-28 Joern Rennecke <joern.rennecke@embecosm.com>
2307 * config/arc/arc.c (arc_ccfsm_post_advance):
2308 Add comment about TYPE_RETURN.
2310 2013-10-28 Bin Cheng <bin.cheng@arm.com>
2312 * tree-ssa-loop-ivopts.c (strip_offset_1): Change parameter type.
2313 Count DECL_FIELD_BIT_OFFSET for COMPONENT_REF.
2314 (strip_offset): Convert offset to unsigned number.
2316 2013-10-27 Tom de Vries <tom@codesourcery.com>
2318 * cfgexpand.c (gimple_expand_cfg): Remove test for parm_birth_insn.
2319 Don't commit insertions after NOTE_INSN_FUNCTION_BEG.
2321 2013-10-27 Oleg Endo <olegendo@gcc.gnu.org>
2323 * config/sh/sh.c (MSW, LSW): Move and rename macros to...
2324 * config/sh/sh.h (SH_REG_MSW_OFFSET, SH_REG_LSW_OFFSET): ... here.
2325 (TARGET_BIG_ENDIAN): New macro.
2326 * config/sh/sh.md: Use it instead of !TARGET_LITTLE_ENDIAN.
2327 Use SH_REG_MSW_OFFSET and SH_REG_LSW_OFFSET.
2328 * config/sh/sh.c: Likewise.
2329 * config/sh/sh.h: Likewise.
2331 2013-10-27 Hans-Peter Nilsson <hp@axis.com>
2333 * config/cris/cris.c (cris_emit_trap_for_misalignment): Replace the
2334 removed PRED_MUDFLAP with PRED_NORETURN. Correct file-path in comment.
2336 2013-10-26 Oleg Endo <olegendo@gcc.gnu.org>
2338 * config/sh/sh.md (movmemsi): Remove empty constraints and predicates.
2340 (cmpstr_t, cmpstrsi): Fix formatting.
2342 2013-10-26 Oleg Endo <olegendo@gcc.gnu.org>
2345 * config/sh/predicates.md (general_movdst_operand): Allow reg+reg
2346 addressing, do not use general_operand for memory operands.
2348 2013-10-26 Vladimir Makarov <vmakarov@redhat.com>
2351 2013-10-25 Vladimir Makarov <vmakarov@redhat.com>
2352 * lra-spills.c (lra_final_code_change): Remove useless move insns.
2354 2013-10-26 Jeff Law <law@redhat.com>
2356 * predict.c (PRED_MUDFLAP): Remove.
2357 * targhooks.c (build_va_arg_indirect_ref): Remove mudflap support.
2359 * Makefile.in (C_COMMON_OBJS): Remove tree-mudflap.
2360 (OBJS): Remove tree-nomudflap.o
2361 (GTFILES): Remove tree-mudflap.c
2362 * builtins.c (expand_builtin_alloc): Remove mudflap support.
2363 * gcc.c (MFWRAP_SPEC, MFLIB_SPEC): Likewise.
2364 (mfwrap_spec, mflib_spec): Likewise.
2365 (cpp_unique_options, cc1_options, static_specs): Likewise.
2366 * gimplify (gimplify_vla_decl, build_va_arg_indirect_ref): Likewise.
2367 * passes.def: Likewise.
2368 * toplev.c (compile_file, process_options): Likewise.
2369 * tree-inline.c (copy_tree_r): Likewise.
2370 * tree-pass.,h (make_pass_mudflap_1, make_pass_mudflap_2): Likewise.
2371 * varasm.c (make_decl_rtl, make_decl_rtl_for_debug): Likewise.
2372 (build_constant_desc, output_constant_def_contents): Likewise.
2373 (categorize_decl_for_section): Likewise.
2374 * tree-mudflap.c: Removed.
2375 * tree-mudflap.h: Removed.
2376 * tree-nomudflap.c: Removed.
2377 * bfin/uclinux.h (MFWRAP_SPEC): Remove.
2378 * moxie/uclinux.h (MFWRAP_SPEC): Likewise.
2379 * rs6000/aix.h (MFWRAP_SPEC, MFLIB_SPEC): Likewise.
2380 * config/sol2.h (MFLIB_SPEC): Likewise.
2381 * doc/install.texi: Remove mudflap references.
2382 * doc/passes.texi: Similarly.
2383 * doc/sourcebuild.texi: Similarly.
2384 * doc/invoke.texi: Remove mudlfap related options.
2386 2013-10-25 Vladimir Makarov <vmakarov@redhat.com>
2388 PR rtl-optimization/58759
2389 * lra-constraints.c (lra_constraints): Remove wrong condition to
2390 remove insn setting up an equivalent pseudo.
2392 2013-10-25 Vladimir Makarov <vmakarov@redhat.com>
2394 * config/rs6000/rs6000-protos.h
2395 (rs6000_secondary_memory_needed_mode): New prototype.
2396 * config/rs6000/rs6000.c: Include ira.h.
2397 (TARGET_LRA_P): Redefine.
2398 (rs6000_legitimate_offset_address_p): Call
2399 legitimate_constant_pool_address_p in strict mode for LRA.
2400 (rs6000_legitimate_address_p): Ditto.
2401 (legitimate_lo_sum_address_p): Add code for LRA. Use lra_in_progress.
2402 (rs6000_emit_move): Add LRA version of code to generate load/store
2404 (rs6000_secondary_memory_needed_mode): New.
2405 (rs6000_alloc_sdmode_stack_slot): Do nothing for LRA.
2406 (rs6000_secondary_reload_class): Return NO_REGS for LRA for
2407 constants, memory, and FP registers.
2408 (rs6000_lra_p): New.
2409 * config/rs6000/rs6000.h (SECONDARY_MEMORY_NEEDED_MODE): New macro.
2410 * config/rs6000/rs6000.opt (mlra): New option.
2411 * lra-spills.c (lra_final_code_change): Remove useless move insns.
2413 2013-10-25 Yufeng Zhang <yufeng.zhang@arm.com>
2415 * tree-ssa-math-opts.c (convert_plusminus_to_widen): Call
2416 has_single_use () and not do the conversion if has_single_use ()
2417 returns false for the multiplication result.
2419 2013-10-25 David Malcolm <dmalcolm@redhat.com>
2421 * tree.h (EXCEPTIONAL_CLASS_P): Rename parameter from "CODE"
2422 to "NODE", since this works on a "tree", not an
2424 (CONSTANT_CLASS_P): Likewise.
2427 (INDIRECT_REF_P): Likewise.
2428 (REFERENCE_CLASS_P): Likewise.
2429 (COMPARISON_CLASS_P): Likewise.
2430 (UNARY_CLASS_P): Likewise.
2431 (BINARY_CLASS_P): Likewise.
2432 (STATEMENT_CLASS_P): Likewise.
2433 (VL_EXP_CLASS_P): Likewise.
2434 (EXPRESSION_CLASS_P): Likewise.
2435 (IS_TYPE_OR_DECL_P): Likewise.
2437 2013-10-25 Marc Glisse <marc.glisse@inria.fr>
2439 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Look for an
2440 ADDR_EXPR in the defining statement.
2442 2013-10-25 Richard Biener <rguenther@suse.de>
2444 PR tree-optimization/58626
2445 * tree-loop-distribution.c (enum rdg_dep_type): Remove
2446 anti_dd, output_dd and input_dd.
2447 (struct rdg_edge): Remove level and relation members.
2448 (RDGE_LEVEL, RDGE_RELATION): Remove.
2449 (dot_rdg_1): Adjust.
2450 (create_rdg_edge_for_ddr): Remove.
2451 (create_rdg_edges_for_scalar): Adjust.
2452 (create_edge_for_control_dependence): Likewise.
2453 (create_rdg_edges): Split into ...
2454 (create_rdg_flow_edges): ... this
2455 (create_rdg_cd_edges): ... and this.
2457 (build_rdg): Likewise, do not compute data dependences or
2459 (pg_add_dependence_edges): New function.
2461 (distribute_loop): First apply all non-dependence based
2462 partition mergings. Then compute dependences between partitions
2463 and merge and order partitions according to them.
2465 2013-10-25 Eric Botcazou <ebotcazou@adacore.com>
2467 PR rtl-optimization/58831
2468 * alias.c (init_alias_analysis): At the beginning of each iteration,
2469 set the reg_seen[N] bit if static_reg_base_value[N] is non-null.
2471 2013-10-25 Eric Botcazou <ebotcazou@adacore.com>
2473 * recog.c (search_ofs): New static variable moved from...
2474 (peep2_find_free_register): ...here.
2475 (peephole2_optimize): Initialize it.
2477 2013-10-25 Tobias Burnus <burnus@net-b.de>
2479 * doc/invoke.texi (fopenmp): Change supported OpenMP version to 4.0.
2481 2013-10-25 Uros Bizjak <ubizjak@gmail.com>
2483 * config/i386/i386.h (TARGET_MPX): New define.
2484 (TARGET_MPX_P): Ditto.
2486 2013-10-24 Ilya Enkovich <ilya.enkovich@intel.com>
2488 * config/i386/constraints.md (B): New.
2491 * config/i386/i386-c.c (ix86_target_macros_internal): Add __MPX__.
2492 * config/i386/i386-modes.def (BND32): New.
2494 * config/i386/i386-protos.h (ix86_bnd_prefixed_insn_p): New.
2495 * config/i386/i386.c (isa_opts): Add mmpx.
2496 (regclass_map): Add bound registers.
2497 (dbx_register_map): Likewise.
2498 (dbx64_register_map): Likewise.
2499 (svr4_dbx_register_map): Likewise.
2501 (ix86_option_override_internal): Support MPX ISA.
2502 (ix86_conditional_register_usage): Support bound registers.
2503 (print_reg): Likewise.
2504 (ix86_code_end): Add MPX bnd prefix.
2505 (output_set_got): Likewise.
2506 (ix86_output_call_insn): Likewise.
2507 (ix86_print_operand): Add '!' (MPX bnd) print prefix support.
2508 (ix86_print_operand_punct_valid_p): Likewise.
2509 (ix86_print_operand_address): Support UNSPEC_BNDMK_ADDR and
2511 (ix86_class_likely_spilled_p): Add bound regs support.
2512 (ix86_hard_regno_mode_ok): Likewise.
2513 (x86_order_regs_for_local_alloc): Likewise.
2514 (ix86_bnd_prefixed_insn_p): New.
2515 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Fix to new value.
2516 (FIXED_REGISTERS): Add bound registers.
2517 (CALL_USED_REGISTERS): Likewise.
2518 (REG_ALLOC_ORDER): Likewise.
2519 (HARD_REGNO_NREGS): Likewise.
2521 (VALID_BND_REG_MODE): New.
2522 (FIRST_BND_REG): New.
2523 (LAST_BND_REG): New.
2524 (reg_class): Add BND_REGS.
2525 (REG_CLASS_NAMES): Likewise.
2526 (REG_CLASS_CONTENTS): Likewise.
2528 (ANY_BND_REG_P): New.
2530 (HI_REGISTER_NAMES): Add bound registers.
2531 * config/i386/i386.md (UNSPEC_BNDMK): New.
2532 (UNSPEC_BNDMK_ADDR): New.
2533 (UNSPEC_BNDSTX): New.
2534 (UNSPEC_BNDLDX): New.
2535 (UNSPEC_BNDLDX_ADDR): New.
2536 (UNSPEC_BNDCL): New.
2537 (UNSPEC_BNDCU): New.
2538 (UNSPEC_BNDCN): New.
2539 (UNSPEC_MPX_FENCE): New.
2542 (type): Add mpxmov, mpxmk, mpxchk, mpxld, mpxst.
2543 (length_immediate): Likewise.
2544 (prefix_0f): Likewise.
2546 (prefix_rep): Check for bnd prefix.
2547 (length_nobnd): New.
2548 (length): Use length_nobnd if specified.
2553 (*jcc_1): Add bnd prefix and rename length attr to length_nobnd.
2556 (simple_return_internal): Likewise.
2557 (simple_return_pop_internal): Likewise.
2558 (*indirect_jump): Add MPX bnd prefix.
2559 (*tablejump_1): Likewise.
2560 (simple_return_internal_long): Likewise.
2561 (simple_return_indirect_internal): Likewise.
2565 (*mov<mode>_internal_mpx): New.
2566 (<mode>_<bndcheck>): New.
2567 (*<mode>_<bndcheck>): New.
2572 * config/i386/predicates.md (lea_address_operand): Rename to...
2573 (address_no_seg_operand): ... this.
2574 (address_mpx_no_base_operand): New.
2575 (address_mpx_no_index_operand): New.
2576 (bnd_mem_operator): New.
2577 * config/i386/i386.opt (mmpx): New.
2578 * doc/invoke.texi: Add documentation for the flags -mmpx, -mno-mpx.
2579 * doc/rtl.texi Add documentation for BND32mode and BND64mode.
2581 2013-10-24 Ilya Enkovich <ilya.enkovich@intel.com>
2583 * mode-classes.def (MODE_POINTER_BOUNDS): New.
2584 * tree.def (POINTER_BOUNDS_TYPE): New.
2585 * genmodes.c (complete_mode): Support MODE_POINTER_BOUNDS.
2586 (POINTER_BOUNDS_MODE): New.
2587 (make_pointer_bounds_mode): New.
2588 * machmode.h (POINTER_BOUNDS_MODE_P): New.
2589 * stor-layout.c (int_mode_for_mode): Support MODE_POINTER_BOUNDS.
2590 (layout_type): Support POINTER_BOUNDS_TYPE.
2591 * tree-pretty-print.c (dump_generic_node): Support POINTER_BOUNDS_TYPE.
2592 * tree.c (build_int_cst_wide): Support POINTER_BOUNDS_TYPE.
2593 (type_contains_placeholder_1): Likewise.
2594 * tree.h (POINTER_BOUNDS_TYPE_P): New.
2595 * varasm.c (output_constant): Support POINTER_BOUNDS_TYPE.
2596 * doc/rtl.texi (MODE_POINTER_BOUNDS): New.
2598 2013-10-24 Igor Shevlyakov <igor.shevlyakov@gmail.com>
2600 * expr.c (expand_expr_real_1): Use mode of memory reference rather than
2601 mode of address computation when calling memory_address_addr_space.
2603 2013-08-24 Richard Henderson <rth@twiddle.net>
2606 * optabs.c (maybe_emit_atomic_exchange): Use create_input_operand
2607 instead of create_convert_operand_to.
2608 (maybe_emit_sync_lock_test_and_set): Likewise.
2609 (expand_atomic_compare_and_swap): Likewise.
2610 (maybe_emit_compare_and_swap_exchange_loop): Don't convert_modes.
2612 2013-08-24 Sriraman Tallam <tmsriram@google.com>
2614 * cgraph.c (cgraph_fnver_htab): Move GTY((...)) to be before htab_t.
2615 Change param_is to use the struct and not the pointer to the struct.
2617 2013-10-24 Andrew MacLeod <amacleod@redhat.com>
2619 * builtins.c (dummy_object, gimplify_va_arg_expr): Move to gimplify.c.
2620 * gimplify.c (build_va_arg_indirect_ref, std_gimplify_va_arg_expr):
2621 Move to targhooks.c.
2622 (dummy_object, gimplify_va_arg_expr): Relocate from builtins.c.
2623 * targhooks.c (build_va_arg_indirect_ref, std_gimplify_va_arg_expr):
2624 Relocate from gimplify.c.
2625 * targhooks.h: Add 2 prototypes.
2626 * tree.h: Delete 2 prototypes.
2628 2013-10-24 Igor Shevlyakov <igor.shevlyakov@gmail.com>
2630 * tree-ssa-loop-ivopts.c (multiplier_allowed_in_address_p ): Check both
2631 [reg+mult*reg] and [mult*reg] to determine if multiplier is allowed.
2633 2013-10-24 Cong Hou <congh@google.com>
2635 * convert.c (convert_to_real): Guard those unsafe math function
2636 convertions with flag_unsafe_math_optimizations. Handle sqrt()
2639 2013-10-24 Markus Trippelsdorf <markus@trippelsdorf.de>
2642 * cgraph.c (cgraph_create_edge_1): Add indirect_unknown_callee
2644 (cgraph_create_edge): Use the new argument.
2645 (cgraph_create_indirect_edge): Likewise.
2647 2013-10-24 Joern Rennecke <joern.rennecke@embecosm.com>
2649 * config/arc/arc.c (arc_ccfsm_post_advance): Also handle
2651 (arc_ifcvt) <case 1 and 2>: Check that arc_ccfsm_post_advance
2652 changes statep->state.
2654 2013-10-24 Tobias Burnus <burnus@net-b.de>
2657 * tree-cfg.c (replace_loop_annotate): New function.
2658 (execute_build_cfg): Call it.
2659 * gimplify.c (gimple_boolify, gimplify_expr): Handle ANNOTATE_EXPR.
2660 * internal-fn.c (expand_ANNOTATE): New function.
2661 * internal-fn.def (ANNOTATE): Define as new internal function.
2662 * tree-core.h (tree_node_kind): Add annot_expr_ivdep_kind.
2663 * tree-pretty-print.c (dump_generic_node): Handle ANNOTATE_EXPR.
2664 * tree.def (ANNOTATE_EXPR): New DEFTREECODE.
2665 * doc/extend.texi (Pragmas): Document #pragma ivdep.
2666 * doc/generic.texi (Expressions): Document ANNOTATE_EXPR.
2668 2013-10-17 Ian Bolton <ian.bolton@arm.com>
2669 Marcus Shawcroft <marcus.shawcroft@arm.com>
2671 * config/aarch64/aarch64.c (aarch64_preferred_reload_class):
2672 Special case reload SP+C into none GENERAL_REGS.
2674 2013-10-24 Michael Matz <matz@suse.de>
2676 * gengtype.c (is_file_equal): Check that files will be same length.
2678 2013-10-25 Christian Bruel <christian.bruel@st.com>
2680 * config.gcc (sh-*): Add sh-mem.o to extra_obj.
2681 * config/sh/t-sh (sh-mem.o): New rule.
2682 * config/sh/sh-mem.cc (expand_block_move): Moved here.
2683 (sh_expand_cmpstr): New function.
2684 * config/sh/sh.c (force_into, expand_block_move): Move to sh-mem.c.
2685 * config/sh/sh-protos.h (sh_expand_cmpstr): Declare.
2686 * config/sh/sh.md (cmpstrsi, cmpstr_t): New patterns.
2687 (rotlhi3_8): Rename.
2689 2013-10-24 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2691 * configure.ac (ZW_PROG_COMPILER_DEPENDENCIES): Use CXX instead of CC.
2692 * Makefile.in (CXXDEPMODE): Assign and change users.
2693 (CCDEPMODE): Delete.
2694 * configure: Regenerate.
2696 2013-10-23 David Malcolm <dmalcolm@redhat.com>
2698 * gengtype-parse.c (require_without_advance): New.
2699 (type): For GTY-marked types that are not GTY((user)), parse any
2700 base classes, requiring them to be single-inheritance, and not
2701 be templates. For non-GTY-marked types and GTY((user)),
2702 continue to skip over any C++ inheritance specification.
2703 * gengtype-state.c (state_writer::write_state_struct_type):
2704 Write base class of type (if any).
2705 (read_state_struct_type): Read base class of type (if any).
2706 * gengtype.c (new_structure): Add a "base_class" parameter.
2707 (create_optional_field_): Update for new parameter to new_structure.
2708 (adjust_field_rtx_def): Likewise.
2709 (adjust_field_tree_exp): Likewise.
2710 * gengtype.h (struct type): Add "base_class" field to the s
2712 (new_structure): Add "base" parameter.
2714 2013-10-23 Sriraman Tallam <tmsriram@google.com>
2717 * config/i386/i386.c (ix86_option_override_internal):
2718 Change TARGET_SSE2 to TARGET_SSE2_P (opts->...)
2719 (ix86_valid_target_attribute_tree):
2720 Change TARGET_64BIT to TARGET_64BIT_P (opts->...)
2721 Change TARGET_SSE to TARGET_SSE_P (opts->...)
2723 2013-10-23 Andrew MacLeod <amacleod@redhat.com>
2725 * tree-ssa-loop.h: Remove include files.
2726 * gengtype.c (open_base_files): Adjust include list for gtype-desc.c.
2727 * cfgloopmanip.c: Move required includes from tree-ssa-loop.h.
2728 * graphite-clast-to-gimple.c: Likewise.
2729 * graphite-scop-detection.c: Likewise.
2730 * graphite-sese-to-poly.c: Likewise.
2731 * ipa-inline-analysis.c: Likewise.
2732 * ipa-pure-const.c: Likewise.
2733 * loop-init.c: Likewise.
2734 * passes.c: Likewise.
2735 * predict.c: Likewise.
2736 * tree-cfg.c: Likewise.
2737 * tree-cfgcleanup.c: Likewise.
2738 * tree-chrec.c: Likewise.
2739 * tree-data-ref.c: Likewise.
2740 * tree-loop-distribution.c: Likewise.
2741 * tree-parloops.c: Likewise.
2742 * tree-predcom.c: Likewise.
2743 * tree-scalar-evolution.c: Likewise.
2744 * tree-ssa-address.c: Likewise.
2745 * tree-ssa.c: Likewise.
2746 * tree-ssa-dce.c: Likewise.
2747 * tree-ssa-loop.c: Likewise.
2748 * tree-ssa-loop-im.c: Likewise.
2749 * tree-ssa-loop-ivcanon.c: Likewise.
2750 * tree-ssa-loop-ivopts.c: Likewise.
2751 * tree-ssa-loop-manip.c: Likewise.
2752 * tree-ssa-loop-niter.c: Likewise.
2753 * tree-ssa-loop-prefetch.c: Likewise.
2754 * tree-ssa-loop-unswitch.c: Likewise.
2755 * tree-ssa-reassoc.c: Likewise.
2756 * tree-vect-data-refs.c: Likewise.
2757 * tree-vect-loop.c: Likewise.
2758 * tree-vect-loop-manip.c: Likewise.
2759 * tree-vectorizer.c: Likewise.
2760 * tree-vect-stmts.c: Likewise.
2761 * tree-vrp.c: Likewise.
2763 2013-10-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2765 * config/rs6000/altivec.md (mulv8hi3): Adjust for little endian.
2767 2013-10-23 Jakub Jelinek <jakub@redhat.com>
2769 PR tree-optimization/58775
2770 PR tree-optimization/58791
2771 * tree-ssa-reassoc.c (reassoc_stmt_dominates_stmt_p): New function.
2772 (insert_stmt_after): Rewritten, don't move the stmt, but really
2774 (get_stmt_uid_with_default): Remove.
2775 (build_and_add_sum): Use insert_stmt_after and
2776 reassoc_stmt_dominates_stmt_p. Fix up uid if bb contains only labels.
2777 (update_range_test): Set uid on stmts added by
2778 force_gimple_operand_gsi. Don't immediately modify statements
2779 in inter-bb optimization, just update oe->op values.
2780 (optimize_range_tests): Return bool whether any changed have been made.
2781 (update_ops): New function.
2782 (struct inter_bb_range_test_entry): New type.
2783 (maybe_optimize_range_tests): Perform statement changes here.
2784 (not_dominated_by, appears_later_in_bb, get_def_stmt,
2785 ensure_ops_are_available): Remove.
2786 (find_insert_point): Rewritten.
2787 (rewrite_expr_tree): Remove MOVED argument, add CHANGED argument,
2788 return LHS of the (new resp. old) stmt. Don't call
2789 ensure_ops_are_available, don't reuse SSA_NAMEs, recurse first
2790 instead of last, move new stmt at the right place.
2791 (linearize_expr, repropagate_negates): Don't reuse SSA_NAMEs.
2792 (negate_value): Likewise. Set uids.
2793 (break_up_subtract_bb): Initialize uids.
2794 (reassociate_bb): Adjust rewrite_expr_tree caller.
2795 (do_reassoc): Don't call renumber_gimple_stmt_uids.
2797 2013-10-23 David Edelsohn <dje.gcc@gmail.com>
2800 * config/rs6000/rs6000.md (mulsi3_internal1 and splitter): Add
2801 TARGET_32BIT final condition.
2802 (mulsi3_internal2 and splitter): Same.
2804 2013-10-23 Jeff Law <law@redhat.com>
2806 * tree-ssa-threadedge.c (thread_across_edge): Do not allow threading
2807 through joiner blocks with abnormal outgoing edges.
2809 * tree-ssa-threadupdate.c (thread_block_1): Renamed from thread_block.
2810 Add parameter JOINERS, to allow/disallow threading through joiner
2812 (thread_block): New. Call thread_block_1.
2813 (mark_threaded_blocks): Remove code to filter out certain cases
2814 of threading through joiner blocks.
2815 (thread_through_all_blocks): Document how we can have a dangling
2816 edge AUX field and clear it.
2818 2013-10-23 Ian Lance Taylor <iant@google.com>
2820 * doc/invoke.texi (Option Summary): Remove -fno-default-inline.
2821 (C++ Dialect Options): Likewise.
2822 (Optimize Options): Likewise.
2824 2013-10-23 Tom de Vries <tom@codesourcery.com>
2826 PR tree-optimization/58805
2827 * tree-ssa-tail-merge.c (stmt_local_def): Add gimple_vdef check.
2829 2013-10-23 Jakub Jelinek <jakub@redhat.com>
2831 * tree-vect-patterns.c (vect_recog_divmod_pattern): Optimize
2832 sequence based on get_range_info returned range.
2834 2013-10-23 Andrew MacLeod <amacleod@redhat.com>
2836 * tree-ssa.h: Remove all #include's
2837 * gengtype.c (open_base_files): Adjust include list for gtype-desc.c.
2838 * alias.c: Move required includes from tree-ssa.h.
2840 * builtins.c: Likewise.
2841 * calls.c: Likewise.
2842 * cfgexpand.c: Likewise.
2843 * cfghooks.c: Likewise.
2844 * cfgloop.c: Likewise.
2845 * cfgloopmanip.c: Likewise.
2846 * cgraph.c: Likewise.
2847 * cgraphbuild.c: Likewise.
2848 * cgraphclones.c: Likewise.
2849 * cgraphunit.c: Likewise.
2851 * except.c: Likewise.
2853 * final.c: Likewise.
2854 * fold-const.c: Likewise.
2855 * ggc-page.c: Likewise.
2856 * gimple-builder.c: Likewise.
2857 * gimple-fold.c: Likewise.
2858 * gimple-iterator.c: Likewise.
2859 * gimple-low.c: Likewise.
2860 * gimple-pretty-print.c: Likewise.
2861 * gimple-ssa-strength-reduction.c: Likewise.
2862 * gimple-streamer-in.c: Likewise.
2863 * gimple-streamer-out.c: Likewise.
2864 * gimplify.c: Likewise.
2865 * graphite-blocking.c: Likewise.
2866 * graphite-clast-to-gimple.c: Likewise.
2867 * graphite-dependences.c: Likewise.
2868 * graphite-interchange.c: Likewise.
2869 * graphite-optimize-isl.c: Likewise.
2870 * graphite-poly.c: Likewise.
2871 * graphite-scop-detection.c: Likewise.
2872 * graphite-sese-to-poly.c: Likewise.
2873 * graphite.c: Likewise.
2874 * ipa-cp.c: Likewise.
2875 * ipa-inline-analysis.c: Likewise.
2876 * ipa-inline-transform.c: Likewise.
2877 * ipa-inline.c: Likewise.
2878 * ipa-prop.c: Likewise.
2879 * ipa-pure-const.c: Likewise.
2880 * ipa-reference.c: Likewise.
2881 * ipa-split.c: Likewise.
2882 * ipa-utils.c: Likewise.
2883 * loop-init.c: Likewise.
2884 * lto-cgraph.c: Likewise.
2885 * lto-section-in.c: Likewise.
2886 * lto-section-out.c: Likewise.
2887 * lto-streamer-in.c: Likewise.
2888 * lto-streamer-out.c: Likewise.
2889 * lto-streamer.c: Likewise.
2890 * omp-low.c: Likewise.
2891 * passes.c: Likewise.
2892 * predict.c: Likewise.
2893 * print-tree.c: Likewise.
2894 * profile.c: Likewise.
2896 * targhooks.c: Likewise.
2897 * tracer.c: Likewise.
2898 * trans-mem.c: Likewise.
2899 * tree-call-cdce.c: Likewise.
2900 * tree-cfg.c: Likewise.
2901 * tree-cfgcleanup.c: Likewise.
2902 * tree-chrec.c: Likewise.
2903 * tree-complex.c: Likewise.
2904 * tree-data-ref.c: Likewise.
2905 * tree-dfa.c: Likewise.
2906 * tree-eh.c: Likewise.
2907 * tree-emutls.c: Likewise.
2908 * tree-if-conv.c: Likewise.
2909 * tree-inline.c: Likewise.
2910 * tree-into-ssa.c: Likewise.
2911 * tree-loop-distribution.c: Likewise.
2912 * tree-mudflap.c: Likewise.
2913 * tree-nested.c: Likewise.
2914 * tree-nrv.c: Likewise.
2915 * tree-object-size.c: Likewise.
2916 * tree-outof-ssa.c: Likewise.
2917 * tree-parloops.c: Likewise.
2918 * tree-phinodes.c: Likewise.
2919 * tree-predcom.c: Likewise.
2920 * tree-pretty-print.c: Likewise.
2921 * tree-profile.c: Likewise.
2922 * tree-scalar-evolution.c: Likewise.
2923 * tree-sra.c: Likewise.
2924 * tree-ssa-address.c: Likewise.
2925 * tree-ssa-alias.c: Likewise.
2926 * tree-ssa-ccp.c: Likewise.
2927 * tree-ssa-coalesce.c: Likewise.
2928 * tree-ssa-copy.c: Likewise.
2929 * tree-ssa-copyrename.c: Likewise.
2930 * tree-ssa-dce.c: Likewise.
2931 * tree-ssa-dom.c: Likewise.
2932 * tree-ssa-dse.c: Likewise.
2933 * tree-ssa-forwprop.c: Likewise.
2934 * tree-ssa-ifcombine.c: Likewise.
2935 * tree-ssa-live.c: Likewise.
2936 * tree-ssa-loop-ch.c: Likewise.
2937 * tree-ssa-loop-im.c: Likewise.
2938 * tree-ssa-loop-ivcanon.c: Likewise.
2939 * tree-ssa-loop-ivopts.c: Likewise.
2940 * tree-ssa-loop-manip.c: Likewise.
2941 * tree-ssa-loop-niter.c: Likewise.
2942 * tree-ssa-loop-prefetch.c: Likewise.
2943 * tree-ssa-loop-unswitch.c: Likewise.
2944 * tree-ssa-loop.c: Likewise.
2945 * tree-ssa-math-opts.c: Likewise.
2946 * tree-ssa-operands.c: Likewise.
2947 * tree-ssa-phiopt.c: Likewise.
2948 * tree-ssa-phiprop.c: Likewise.
2949 * tree-ssa-pre.c: Likewise.
2950 * tree-ssa-propagate.c: Likewise.
2951 * tree-ssa-reassoc.c: Likewise.
2952 * tree-ssa-sccvn.c: Likewise.
2953 * tree-ssa-sink.c: Likewise.
2954 * tree-ssa-strlen.c: Likewise.
2955 * tree-ssa-structalias.c: Likewise.
2956 * tree-ssa-tail-merge.c: Likewise.
2957 * tree-ssa-ter.c: Likewise.
2958 * tree-ssa-threadedge.c: Likewise.
2959 * tree-ssa-threadupdate.c: Likewise.
2960 * tree-ssa-uncprop.c: Likewise.
2961 * tree-ssa-uninit.c: Likewise.
2962 * tree-ssa.c: Likewise.
2963 * tree-ssanames.c: Likewise.
2964 * tree-stdarg.c: Likewise.
2965 * tree-streamer-in.c: Likewise.
2966 * tree-switch-conversion.c: Likewise.
2967 * tree-tailcall.c: Likewise.
2968 * tree-vect-data-refs.c: Likewise.
2969 * tree-vect-generic.c: Likewise.
2970 * tree-vect-loop-manip.c: Likewise.
2971 * tree-vect-loop.c: Likewise.
2972 * tree-vect-patterns.c: Likewise.
2973 * tree-vect-slp.c: Likewise.
2974 * tree-vect-stmts.c: Likewise.
2975 * tree-vectorizer.c: Likewise.
2976 * tree-vrp.c: Likewise.
2979 * value-prof.c: Likewise.
2980 * var-tracking.c: Likewise.
2981 * varpool.c: Likewise.
2982 * vtable-verify.c: Likewise.
2984 2013-10-23 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2986 * config/tilegx/tilegx.c: Include "tree.h".
2988 2013-10-23 Jakub Jelinek <jakub@redhat.com>
2990 * gimple-pretty-print.c (dump_ssaname_info): Always print "# " before
2991 the info, not after it.
2992 (gump_gimple_phi): Add COMMENT argument, if true, print "# " after
2993 dump_ssaname_info call.
2994 (pp_gimple_stmt_1): Adjust caller.
2995 (dump_phi_nodes): Likewise. Don't print "# " here.
2997 2013-10-22 Jan Hubicka <jh@suse.cz>
2999 * i386.h (TARGET_MISALIGNED_MOVE_STRING_PROLOGUES_EPILOGUES): New
3001 * x86-tune.def (TARGET_MISALIGNED_MOVE_STRING_PROLOGUES): Define it.
3002 * i386.c (expand_small_movmem_or_setmem): New function.
3003 (expand_set_or_movmem_prologue_epilogue_by_misaligned_moves): New
3005 (alg_usable_p): Add support for value ranges; cleanup.
3006 (ix86_expand_set_or_movmem): Add support for misaligned moves.
3008 2013-10-22 Sterling Augustine <saugustine@google.com>
3010 * doc/invoke.texi: Document -ggnu-pubnames.
3011 * common.opt: Add new option -ggnu-pubnames and modify -gpubnames
3013 * dwarf2out.c: Include gdb/gdb-index.h.
3014 (DEBUG_PUBNAMES_SECTION, DEBUG_PUBTYPES_SECTION): Handle
3015 debug_generate_pub_sections.
3016 (is_java, output_pubtables, output_pubname): New functions.
3017 (include_pubname_in_output): Handle debug_generate_pub_sections at
3019 (size_of_pubnames): Use new local space_for_flags based on
3020 debug_generate_pub_sections.
3021 (output_pubnames): Unify pubnames and pubtypes output logic.
3022 Genericize comments. Call output_pubname.
3023 (dwarf2out_finish): Move logic to output_pubnames and call it.
3025 2013-10-22 Uros Bizjak <ubizjak@gmail.com>
3028 * config/i386/i386.c (put_condition_code) <case GTU, case LEU>:
3029 Remove CCCmode handling.
3030 <case LTU>: Return 'c' suffix for CCCmode.
3031 <case GEU>: Return 'nc' suffix for CCCmode.
3032 (ix86_cc_mode) <case GTU, case LEU>: Do not generate overflow checks.
3033 * config/i386/i386.md (*sub<mode>3_cconly_overflow): Remove.
3034 (*sub<mode>3_cc_overflow): Ditto.
3035 (*subsi3_zext_cc_overflow): Ditto.
3037 2013-10-22 Steve Ellcey <sellcey@mips.com>
3039 * config/mips/mips.c (mips_rtx_costs): Fix cost estimate for nor
3040 (AND (NOT OP1) (NOT OP2)).
3042 2013-10-22 Bill Schmidt <wschmidt@vnet.ibm.com>
3044 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
3045 meaning of merge-high and merge-low masks for little endian; avoid
3046 use of vector-pack masks for little endian for mismatched modes.
3048 2013-10-22 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3050 * config/tilepro/tilepro.c: Include "tree.h".
3052 2013-10-22 Andreas Schwab <schwab@suse.de>
3054 * config/m68k/m68k.c (notice_update_cc): Handle register conflict
3057 2013-10-22 Paolo Carlini <paolo.carlini@oracle.com>
3059 * doc/contrib.texi ([Fran@,{c}ois Dumont], [Tim Shen],
3060 [Ed Smith-Rowland]): New entries.
3061 ([Stephen M. Webb]): Update.
3063 2013-10-22 Andrew MacLeod <amacleod@redhat.com>
3065 * tree-ssa-ter.h: Remove duplicate copy of file contents.
3067 2013-10-21 Marek Polacek <polacek@redhat.com>
3070 * fold-const.c (fold_range_test): Return 0 if the type is not
3073 2013-10-21 Richard Sandiford <rdsandiford@googlemail.com>
3075 * system.h: Move hwint.h include further down.
3076 * hwint.h (sext_hwi, zext_hwi): Define unconditionally. Add
3077 gcc_checking_asserts.
3078 * hwint.c (sext_hwi, zext_hwi): Delete ENABLE_CHECKING versions.
3080 2013-10-21 Bernd Edlinger <bernd.edlinger@hotmail.de>
3082 Fix volatile issues in optimize_bit_field_compare.
3083 * fold-const.c (optimize_bit_field_compare): Bail out if
3084 lvolatilep or rvolatilep.
3086 2013-10-21 Bernd Edlinger <bernd.edlinger@hotmail.de>
3088 Fix DECL_BIT_FIELD depencency on flag_strict_volatile_bitfields
3089 and get_inner_reference returning different pmode for non-volatile
3090 bit-field members dependent on flag_strict_volatile_bitfields.
3091 * stor-layout.c (layout_decl): Remove special handling of
3092 flag_strict_volatile_bitfields.
3093 * expr.c (get_inner_reference): Don't use DECL_BIT_FIELD
3094 if flag_strict_volatile_bitfields > 0 and TREE_THIS_VOLATILE.
3096 2013-10-21 Paulo Matos <pmatos@broadcom.com>
3098 * ipa-inline.c (edge_badness): Cap edge->count at max_count for badness
3101 2013-10-21 Jeff Law <law@redhat.com>
3103 * tree-ssa-threadedge.c (thread_through_normal_block): New
3104 argument VISITED. Remove VISISTED as a local variable. When we
3105 have a threadable jump, verify the destination of the jump has not
3107 (thread_across_edge): Allocate VISITED bitmap once at function
3108 scope and use it throughout. Make sure to set appropriate bits in
3109 VISITED for E (start of jump thread path).
3110 * tree-ssa-threadupdate.c (mark_threaded_blocks): Reject threading
3111 through a joiner if any edge on the path has a recorded jump thread.
3113 2013-10-21 Ian Lance Taylor <iant@google.com>
3115 * doc/invoke.texi (Optimize Options): For -fno-toplevel-reorder,
3116 don't imply that attributes can solve all problems.
3117 (Directory Options): Fix typo.
3119 2013-10-21 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3121 * config/arm/arm.c (cortexa9_extra_costs): Update mult costs for
3122 extend and extend_add.
3124 2013-10-21 Richard Biener <rguenther@suse.de>
3126 PR tree-optimization/58794
3127 * fold-const.c (operand_equal_p): Compare FIELD_DECL operand
3128 of COMPONENT_REFs with OEP_CONSTANT_ADDRESS_OF left in place.
3130 2013-10-21 Richard Biener <rguenther@suse.de>
3133 * fold-const.c (fold_binary_loc): Fold ((T) (X /[ex] C)) * C
3134 to (T) X for sign-changing conversions (or no conversion).
3136 2013-10-20 Uros Bizjak <ubizjak@gmail.com>
3138 * config/i386/i386.md (kxnor<mode>): Add FLAGS_REG clobber.
3140 2013-10-20 Jan Hubicka <jh@suse.cz>
3142 * config/i386/i386-tune.def: Add comment; organize into categories
3144 2013-10-21 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
3146 * config/i386/i386.c (expand_set_or_movmem_via_loop): Add issetmem
3147 argument. Update function comment.
3148 (expand_set_or_movmem_via_rep): New function combining
3149 expand_movmem_via_rep_mov and expand_setmem_via_rep_stos.
3150 (expand_movmem_via_rep_mov): Remove.
3151 expand_setmem_via_rep_stos): Remove.
3152 (expand_movmem_epilogue): Update calls correspondingly.
3153 (expand_setmem_epilogue_via_loop): Likewise.
3155 (expand_setmem_epilogue): Add VEC_VALUE argument, refactor.
3156 (expand_set_or_movmem_prologue): New function combining
3157 expand_movmem_prologue and expand_setmem_prologue.
3158 (expand_movmem_prologue): Remove.
3159 (expand_setmem_prologue): Remove.
3160 (expand_set_or_movmem_constant_prologue): New function combining
3161 expand_constant_movmem_prologue and expand_constant_setmem_prologue.
3162 (expand_constant_movmem_prologue): Remove.
3163 (expand_constant_setmem_prologue): Remove.
3164 (promote_duplicated_reg): Allow vector-const0 value.
3165 (ix86_expand_set_or_movmem): New function combining ix86_expand_movmem
3166 and ix86_expand_setmem.
3167 (ix86_expand_movmem): Call ix86_expand_set_or_movmem.
3168 (ix86_expand_setmem): Call ix86_expand_set_or_movmem.
3170 2013-10-21 Diego Novillo <dnovillo@google.com>
3172 * asan.c: Include tree.h
3173 * bb-reorder.c: Likewise.
3174 * cfgcleanup.c: Likewise.
3175 * cfgloopmanip.c: Likewise.
3176 * data-streamer-in.c: Likewise.
3177 * data-streamer-out.c: Likewise.
3178 * data-streamer.c: Likewise.
3179 * dwarf2cfi.c: Likewise.
3180 * graphite-blocking.c: Likewise.
3181 * graphite-clast-to-gimple.c: Likewise.
3182 * graphite-dependences.c: Likewise.
3183 * graphite-interchange.c: Likewise.
3184 * graphite-optimize-isl.c: Likewise.
3185 * graphite-poly.c: Likewise.
3186 * graphite-scop-detection.c: Likewise.
3187 * graphite-sese-to-poly.c: Likewise.
3188 * graphite.c: Likewise.
3189 * ipa-devirt.c: Likewise.
3190 * ipa-profile.c: Likewise.
3193 * loop-init.c: Likewise.
3194 * loop-unroll.c: Likewise.
3195 * lower-subreg.c: Likewise.
3196 * lto/lto-object.c: Likewise.
3197 * recog.c: Likewise.
3198 * reginfo.c: Likewise.
3199 * tree-loop-distribution.c: Likewise.
3200 * tree-parloops.c: Likewise.
3201 * tree-ssa-strlen.c: Likewise.
3202 * tree-streamer.c: Likewise.
3203 * value-prof.c: Likewise.
3204 * target-globals.c: Likewise.
3205 * expr.h: Include tree-core.h instead of tree.h.
3206 * gimple.h: Likewise.
3207 * ipa-prop.h: Likewise.
3208 * ipa-utils.h: Likewise.
3209 * lto-streamer.h: Likewise.
3210 * streamer-hooks.h: Likewise.
3211 * ipa-reference.h: Include cgraph.h instead of tree.h.
3212 * cgraph.h: Include basic-block.h instead of tree.h.
3213 * tree-streamer.h: Do not include tree.h.
3214 * genattrtab.c (write_header): Generate inclusion of
3216 * genautomata.c (main): Likewise.
3217 * genemit.c: Likewise.
3218 * genopinit.c: Likewise.
3219 * genoutput.c (output_prologue): Likewise.
3220 * genpeep.c: Likewise.
3222 2013-10-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3224 * config/rs6000/altivec.md (vec_unpacku_hi_v16qi): Adjust for
3226 (vec_unpacku_hi_v8hi): Likewise.
3227 (vec_unpacku_lo_v16qi): Likewise.
3228 (vec_unpacku_lo_v8hi): Likewise.
3230 2013-10-20 Jan Hubicka <jh@suse.cz>
3232 * config/i386/x86-tune.def (X86_TUNE_SLOW_IMUL_IMM32_MEM,
3233 X86_TUNE_SLOW_IMUL_IMM8): Keep enabled only for K8 and AMDFAM10.
3234 (X86_TUNE_USE_VECTOR_FP_CONVERTS): Disable for generic.
3236 2013-10-20 Richard Sandiford <rdsandiford@googlemail.com>
3238 * config/mips/mips.h (ISA_HAS_WSBH): Define.
3239 * config/mips/mips.md (UNSPEC_WSBH, UNSPEC_DSBH, UNSPEC_DSHD): New
3241 (bswaphi2, bswapsi2, bswapdi2, wsbh, dsbh, dshd): New patterns.
3243 2013-10-19 John David Anglin <danglin@gcc.gnu.org>
3246 * system.h: Undef m_slot.
3248 2013-10-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3250 * config/rs6000/rs6000.c (vspltis_constant): Make sure we check
3251 all elements for both endian flavors.
3253 2013-10-19 Uros Bizjak <ubizjak@gmail.com>
3256 * config/i386/i386.c (ix86_function_value_regno): Add DX_REG,
3257 ST1_REG and XMM1_REG for 32bit and 64bit targets. Also add DI_REG
3258 and SI_REG for 64bit SYSV ABI targets.
3260 2013-10-19 Uros Bizjak <ubizjak@gmail.com>
3262 * mode-switching.c (create_pre_exit): Rename maybe_builtin_apply
3263 to multi_reg_return. Clarify that we are skipping USEs of multiple
3264 return registers. Use bool type where appropriate.
3266 2013-10-18 Jan Hubicka <jh@suse.cz>
3268 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Disable accumulation
3270 * x86-tune.def (X86_TUNE_USE_LEAVE): Update comment.
3271 (X86_TUNE_PUSH_MEMORY): Likewise.
3272 (X86_TUNE_AVX256_UNALIGNED_LOAD_OPTIMAL,
3273 X86_TUNE_AVX256_UNALIGNED_STORE_OPTIMAL): New.
3274 (X86_TUNE_ACCUMULATE_OUTGOING_ARGS, X86_TUNE_ALWAYS_FANCY_MATH_387):
3276 * i386.c (x86_accumulate_outgoing_args, x86_arch_always_fancy_math_387,
3277 x86_avx256_split_unaligned_load, x86_avx256_split_unaligned_store):
3279 (ix86_option_override_internal): Update to use tune features instead
3282 2013-10-18 Cong Hou <congh@google.com>
3284 PR tree-optimization/58508
3285 * tree-vect-loop-manip.c (vect_loop_versioning): Hoist loop invariant
3286 statement that contains data refs with zero-step.
3288 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
3290 * tree-ssa.h: Don't include gimple-low.h, tree-ssa-address.h,
3291 sbitmap.h, tree-ssa-threadedge.h, tree-ssa-dom.h and tree-cfgcleanup.h.
3292 * gimple-low.c (gimple_check_call_arg,
3293 gimple_check_call_matching_types): Move to cgraph.c.
3294 * gimple-low.h: Remove prototype.
3295 * cgraph.c: (gimple_check_call_arg, gimple_check_call_matching_types):
3296 Relocate from gimple-low.c.
3297 * cgraph.h: Add prototype. Don't include basic-block.h.
3298 * gimplify.c: Add gimple-low to include list.
3299 * omp-low.c: Add gimple-low and tree-cfgcleanup.h to include list.
3300 * tree-eh.c: Add gimple-low to include list.
3301 * tree-nested.c: Likewise.
3302 * cfgexpand.c: Add tree-ssa-address.h to include list.
3304 * gimple-fold.c: Likewise.
3305 * gimple-ssa-strength-reduction.c: Likewise.
3306 * trans-mem.c: Likewise.
3307 * tree-mudflap.c: Likewise.
3308 * tree-ssa-loop-ivopts.c: Likewise.
3309 * tree-ssa-dom.c: Include tree-ssa-threadedge.h and tree-ssa-dom.h.
3310 (degenerate_phi_result): Move to tree-phinodes.c.
3311 * tree-ssa-loop-ch.c: Include tree-ssa-threadedge.h.
3312 * tree-ssa-threadedge.c: Likewise.
3313 * tree-vrp.c: Likewise.
3314 * tree-phinodes.c (degenerate_phi_result): Relocate here.
3315 * tree-ssa-dom.h (degenerate_phi_result): Remove Prototype.
3316 * tree-phinodes.h (degenerate_phi_result): Add prototype.
3317 * tree-ssa-copy.c: Include tree-ssa-dom.h.
3318 * tree-ssa-forwprop.c: Likewise.
3319 * tree-cfgcleanup.c (execute_cleanup_cfg_post_optimizing,
3320 pass_data_cleanup_cfg_post_optimizing,
3321 make_pass_cleanup_cfg_post_optimizing): Relocate from tree-optimize.c.
3322 * tree-optimize.c: Delete File.
3323 * graphite.c: Include tree-cfgcleanup.h.
3324 * passes.c: Likewise.
3325 * tree-cfg.c: Likewise.
3326 * tree-profile.c: Likewise.
3327 * tree-ssa-dse.c: Likewise.
3328 * tree-ssa-loop-ivcanon.c: Likewise.
3329 * tree-switch-conversion.c: Don't include tree-ssa-operands.h.
3330 * tree-outof-ssa.c: Include sbitmap.h.
3331 * tree-ssa-live.c: Likewise.
3332 * tree-ssa-propagate.c: Likewise.
3333 * tree-ssa-structalias.c: Likewise.
3334 * tree-stdarg.c: Likewise.
3335 * Makefile.in (OBJS): Delete tree-optimize.o.
3336 * basic-block.h (gcov_type, gcov_type_unsigned): Move to coretypes.h.
3337 * coretypes.h (gcov_type, gcov_type_unsigned): Relocate here.
3338 * varasm.c: Include basic-block.h.
3339 * cfgloop.h: Include function.h instead of basic-block.h
3340 (bb_loop_depth): Move to cfgloop.c.
3341 * cfgloop.c (bb_loop_depth): Relocate from cfgloop.h.
3343 2013-10-18 Teresa Johnson <tejohnson@google.com>
3345 * predict.c (probably_never_executed): Compare frequency-based
3346 count to number of training runs.
3347 * params.def (UNLIKELY_BB_COUNT_FRACTION): New parameter.
3349 2013-10-18 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3351 * config/arm/arm.c (cortexa9_extra_costs): New table.
3352 (arm_cortex_a9_tune): Use cortexa9_extra_costs.
3354 2013-10-18 Jeff Law <law@redhat.com>
3356 * tree-ssa-threadupdate.c: Do not include "tm.h" or "tm_p.h".
3358 * tree-ssa-threadupdate.c: Include "dbgcnt.h".
3359 (register_jump_thread): Add "registered_jump_thread" debug
3361 * dbgcnt.def (registered_jump_thread): New debug counter.
3363 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
3365 * config/rs6000/rs6000.c: Include cgraph.h.
3367 2013-10-18 Teresa Johnson <tejohnson@google.com>
3369 * tree-ssa-tail-merge.c (replace_block_by): Update edge
3370 weights during merging.
3372 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
3374 * tree-cfg.h: Rename from tree-flow.h. Remove #includes.
3375 * tree-ssa.h: Relocate required #includes from tree-cfg.h.
3376 * tree-ssa-operands.h: Remove prototype.
3377 * tree-ssa-operands.c (virtual_operand_p): Move to gimple.c.
3378 * gimple.c (virtual_operand_p): Relocate from gimple.c.
3379 * gimple.h: Add prototype.
3380 * gimple-ssa.h: Include tree-ssa-operands.h.
3381 * tree-dump.c: Add tree-cfg.h to include list.
3382 * tree-ssa-alias.c: Add ipa-reference.h to include list.
3383 * config/alpha/alpha.c: Include gimple-ssa.h instead of tree-flow.h.
3384 * config/i386/i386.c: Don't include tree-flow.h.
3385 * config/rs6000/rs6000.c: Likewise.
3387 2013-10-18 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3389 * config/frv/frv.c (frv_init_cumulative_args): Fix wrong cast.
3391 2013-10-18 Richard Biener <rguenther@suse.de>
3393 * stor-layout.c (layout_type): Do not change TYPE_PRECISION
3394 or TYPE_UNSIGNED of integral types.
3395 (set_min_and_max_values_for_integral_type): Leave TYPE_MIN/MAX_VALUE
3396 NULL_TREE for zero-precision integral types.
3398 2013-10-18 James Greenhalgh <james.greenhalgh@arm.com>
3400 * config/aarch64/arm_neon.h
3401 (vcvt<ds>_n_<fsu><32,64>_<fsu><32,64>): Correct argument types.
3403 2013-10-17 Sriraman Tallam <tmsriram@google.com>
3406 * opth-gen.awk: Define target_flags_explicit.
3408 2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
3410 * config/rs6000/rs6000.c (enum rs6000_reload_reg_type): Add new
3411 fields to the reg_addr array that describes the valid addressing
3412 mode for any register, general purpose registers, floating point
3413 registers, and Altivec registers.
3414 (FIRST_RELOAD_REG_CLASS): Likewise.
3415 (LAST_RELOAD_REG_CLASS): Likewise.
3416 (struct reload_reg_map_type): Likewise.
3417 (reload_reg_map_type): Likewise.
3418 (RELOAD_REG_VALID): Likewise.
3419 (RELOAD_REG_MULTIPLE): Likewise.
3420 (RELOAD_REG_INDEXED): Likewise.
3421 (RELOAD_REG_OFFSET): Likewise.
3422 (RELOAD_REG_PRE_INCDEC): Likewise.
3423 (RELOAD_REG_PRE_MODIFY): Likewise.
3424 (reg_addr): Likewise.
3425 (mode_supports_pre_incdec_p): New helper functions to say whether
3426 a given mode supports PRE_INC, PRE_DEC, and PRE_MODIFY.
3427 (mode_supports_pre_modify_p): Likewise.
3428 (rs6000_debug_vector_unit): Rearrange the -mdebug=reg output to
3429 print the valid address mode bits for each mode.
3430 (rs6000_debug_print_mode): Likewise.
3431 (rs6000_debug_reg_global): Likewise.
3432 (rs6000_setup_reg_addr_masks): New function to set up the address
3433 mask bits for each type.
3434 (rs6000_init_hard_regno_mode_ok): Use memset to clear arrays.
3435 Call rs6000_setup_reg_addr_masks to set up the address mask bits.
3436 (rs6000_legitimate_address_p): Use mode_supports_pre_incdec_p and
3437 mode_supports_pre_modify_p to determine if PRE_INC, PRE_DEC, and
3438 PRE_MODIFY are supported.
3439 (rs6000_output_move_128bit): Change to use {src,dest}_vmx_p for altivec
3440 registers, instead of {src,dest}_av_p.
3441 (rs6000_print_options_internal): Tweak the debug output slightly.
3443 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
3445 * config/i386/sse.md (*vec_widen_smult_even_v8si): Remove
3448 2013-10-17 Andrew MacLeod <amacleod@redhat.com>
3450 * tree-flow.h (struct omp_region): Move to omp-low.c.
3451 Remove omp_ prototypes and variables.
3452 * gimple.h (omp_reduction_init): Move prototype to omp-low.h.
3453 (copy_var_decl): Relocate prototype from tree-flow.h.
3454 * gimple.c (copy_var_decl): Relocate from omp-low.c.
3455 * tree.h: Move prototype to omp-low.h.
3456 * omp-low.h: New File. Relocate prototypes here.
3457 * omp-low.c (struct omp_region): Make local here.
3458 (root_omp_region): Make static.
3459 (copy_var_decl) Move to gimple.c.
3460 (new_omp_region): Make static.
3461 (make_gimple_omp_edges): New. Refactored from tree-cfg.c make_edges.
3462 * tree-cfg.c: Include omp-low.h.
3463 (make_edges): Factor out OMP specific bits to make_gimple_omp_edges.
3464 * gimplify.c: Include omp-low.h.
3465 * tree-parloops.c: Likewise.
3467 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
3469 * config/i386/i386.c (ix86_fixup_binary_operands): When both source
3470 operands are in memory, prefer to force non-matched operand 1 to
3473 2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
3476 * config/rs6000/rs6000.c (rs6000_legitimate_address_p): Only
3477 restrict TImode addresses to single indirect registers if both
3478 -mquad-memory and -mvsx-timode are used.
3479 (rs6000_output_move_128bit): Use quad_load_store_p to determine if
3480 we should emit load/store quad. Remove using %y for quad memory
3483 * config/rs6000/rs6000.md (mov<mode>_ppc64, TI/PTImode): Add
3484 constraints to allow load/store quad on machines where TImode is
3485 not allowed in VSX registers. Use 'n' instead of 'F' constraint
3486 for TImode to load integer constants.
3488 2013-10-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3490 * config/aarch64/aarch64.c (aarch64_print_operand): Handle 'c'.
3492 2013-10-17 Marcus Shawcroft <marcus.shawcroft@arm.com>
3494 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Adjust
3495 handling of STACK_REG.
3497 2013-10-17 Richard Biener <rguenther@suse.de>
3499 PR tree-optimization/58143
3500 * tree-ssa-loop-im.c (arith_code_with_undefined_signed_overflow):
3502 (rewrite_to_defined_overflow): Likewise.
3503 (move_computations_dom_walker::before_dom): Rewrite stmts
3504 with undefined signed overflow that are not always executed
3505 into unsigned arithmetic.
3507 2013-10-16 Michael Meissner <meissner@linux.vnet.ibm.com>
3510 * config/rs6000/rs6000.opt (rs6000_isa_flags_explicit): Move the
3511 explicit isa flag to be an options variable, instead of using
3512 global_options_set. Remove define from rs6000.h.
3513 * config/rs6000/rs6000.h (rs6000_isa_flags_explicit): Likewise.
3515 * config/rs6000/rs6000.c (rs6000_option_override_internal):
3516 Initialize rs6000_isa_flags_explicit.
3517 (rs6000_function_specific_save): Add gcc_options* parameter, so
3518 that the powerpc builds after the 2013-10-15 changes.
3519 (rs6000_function_specific_restore): Likewise.
3521 2013-10-16 DJ Delorie <dj@redhat.com>
3523 * config/rl78/rl78.c (rl78_alloc_address_registers_macax): Verify
3524 op is a REG before checking REGNO.
3525 (rl78_alloc_physical_registers): Verify pattern is a SET before
3528 2013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3530 * config/rs6000/vector.md (vec_unpacks_hi_v4sf): Correct for
3532 (vec_unpacks_lo_v4sf): Likewise.
3533 (vec_unpacks_float_hi_v4si): Likewise.
3534 (vec_unpacks_float_lo_v4si): Likewise.
3535 (vec_unpacku_float_hi_v4si): Likewise.
3536 (vec_unpacku_float_lo_v4si): Likewise.
3538 2013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3540 * config/rs6000/vsx.md (vsx_concat_<mode>): Adjust output for LE.
3541 (vsx_concat_v2sf): Likewise.
3543 2013-10-16 James Greenhalgh <james.greenhalgh@arm.com>
3545 * config/aarch64/aarch64.md
3546 (*mov<mode>_aarch64): Fix output template for DUP (element) Scalar.
3548 2013-10-16 Andrew MacLeod <amacleod@redhat.com>
3550 PR tree-optimization/58697
3551 * cfgloop.c (get_estimated_loop_iterations_int): Rename from
3552 estimated_loop_iterations_int.
3553 (max_stmt_executions_int): Call get_max_loop_iterations_int.
3554 (get_max_loop_iterations_int): New. HWINT version of
3555 get_max_loop_iterations.
3556 * cfgloop.h: Add prototypes.
3557 * loop-iv.c (find_simple_exit): call get_estimated_loop_iterations_int.
3558 * loop-unroll.c (decide_peel_once_rolling): Call
3559 get_estimated_loop_iterations_int.
3560 * tree-ssa-loop-niter.c (estimated_loop_iterations_int): Add back.
3561 * tree-ssa-loop-niter.h: Tweak prototypes.
3563 2013-10-16 David Malcolm <dmalcolm@redhat.com>
3565 * gengtype-parse.c (struct_field_seq): Ignore access-control
3566 keywords ("public:" etc).
3568 2013-10-16 Marcus Shawcroft <marcus.shawcroft@arm.com>
3570 * config/aarch64/aarch64.c (aarch64_regno_regclass): Classify
3571 FRAME_POINTER_REGNUM and ARG_POINTER_REGNUM as POINTER_REGS.
3573 2013-10-16 Yvan Roux <yvan.roux@linaro.org>
3575 * config/arm/arm.opt (mlra): New option.
3576 * config/arm/arm.c (arm_lra_p): New function.
3577 (TARGET_LRA_P): Define.
3579 2013-10-16 Paulo Matos <pmatos@broadcom.com>
3581 * tree-core.h (tree_code_name): Remove.
3582 * tree.h (get_tree_code_name): New prototype.
3583 * tree.c (tree_code_name): Make static.
3584 (get_tree_code_name): New function.
3585 (dump_tree_statistics, tree_check_failed, tree_not_check_failed,
3586 tree_class_check_failed, tree_range_check_failed,
3587 tree_not_class_check_failed, omp_clause_check_failed,
3588 tree_contains_struct_check_failed, tree_operand_check_failed): Use new
3589 wrapper get_tree_code_name instead of calling tree_code_name directly.
3590 * tree-vrp.c (dump_asserts_for): Likewise.
3591 * tree-dump.c (dequeue_and_dump): Likewise.
3592 * tree-pretty-print.c (do_niy, dump_generic_node): Likewise.
3593 * tree-pretty-print.h (pp_unsupported_tree): Likewise.
3594 * lto-streamer-out.c (lto_write_tree, DFS_write_tree): Likewise.
3595 * tree-ssa-dom.c (print_expr_hash_elt): Likewise.
3596 * gimple-pretty-print.c (dump_unary_rhs, dump_binary_rhs,
3597 dump_ternary_rhs, dump_gimple_assign, dump_gimple_cond,
3598 dump_gimple_omp_for): Likewise.
3599 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Likewise.
3600 * tree-ssa-pre.c (print_pre_expr): Likewise.
3601 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Likewise.
3602 * print-tree.c (print_node_brief, print_node): Likewise.
3603 * gimple.c (gimple_check_failed): Likewise.
3604 * lto-streamer.c (lto_tag_name, print_lto_report): Likewise.
3605 * config/frv/frv.c (frv_init_cumulative_args): Likewise.
3606 * config/mep/mep.c (mep_validate_vliw): Likewise.
3607 * config/iq2000/iq2000.c (init_cumulative_args): Likewise.
3608 * config/rs6000/rs6000.c (init_cumulative_args): Likewise.
3610 2013-10-16 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
3612 * config/i386/i386.c (ix86_option_override_internal): Enable FMA4
3615 2013-10-16 Hans-Peter Nilsson <hp@axis.com>
3617 * config/cris/t-elfmulti (MULTILIB_OPTIONS, MULTILIB_DIRNAMES)
3618 (MULTILIB_MATCHES): Add multilib for -march=v8.
3620 2013-10-15 Sriraman Tallam <tmsriram@google.com>
3623 * optc-save-gen.awk: Add extra parameter to the save and restore
3625 * opth-gen.awk: Generate new TARGET_* macros to accept a parameter.
3626 * tree.c (build_optimization_node): New parameter. Add extra parameter
3627 to call to cl_optimization_save.
3628 (build_target_option_node): New parameter. Add extra parameter
3629 to call to cl_target_option_save.
3630 * tree.h (build_optimization_node): New parameter.
3631 (build_target_option_node): New parameter.
3632 * c-family/c-common.c (handle_optimize_attribute): Fix calls to
3633 build_optimization_node and build_target_option_node.
3634 * c-family/c-pragma.c (handle_pragma_optimize): Ditto.
3635 (handle_pragma_push_options): Ditto.
3636 * toplev.c (process_options): Ditto.
3637 * opts.c (init_options_struct): Check for opts_set non-null.
3638 * target.def (target_option.save): New parameter.
3639 (target_option.restore): New parameter.
3640 * tm.texi: Generate.
3641 * config/i386/i386-c.c (ix86_target_macros_internal): Ditto.
3642 (ix86_pragma_target_parse): Ditto.
3643 * config/i386/i386-protos.h (ix86_valid_target_attribute_tree): New
3645 * config/rs6000/rs6000.c (rs6000_option_override_internal): Fix calls
3646 to build_optimization_node and build_target_option_node.
3647 (rs6000_valid_attribute_p): Ditto.
3648 (rs6000_pragma_target_parse): Ditto.
3649 * config/i386/i386.opt (x_ix86_target_flags_explicit): New TargetSave
3651 * config/i386/i386.h:
3652 TARGET_64BIT_P: New Macro
3653 TARGET_MMX_P: New Macro.
3654 TARGET_3DNOW_P: New Macro.
3655 TARGET_3DNOW_A_P: New Macro.
3656 TARGET_SSE_P: New Macro.
3657 TARGET_SSE2_P: New Macro.
3658 TARGET_SSE3_P: New Macro.
3659 TARGET_SSSE3_P: New Macro.
3660 TARGET_SSE4_1_P: New Macro.
3661 TARGET_SSE4_2_P: New Macro.
3662 TARGET_AVX_P: New Macro.
3663 TARGET_AVX2_P: New Macro.
3664 TARGET_AVX512F_P: New Macro.
3665 TARGET_AVX512PF_P: New Macro.
3666 TARGET_AVX512ER_P: New Macro.
3667 TARGET_AVX512CD_P: New Macro.
3668 TARGET_FMA_P: New Macro.
3669 TARGET_SSE4A_P: New Macro.
3670 TARGET_FMA4_P: New Macro.
3671 TARGET_XOP_P: New Macro.
3672 TARGET_LWP_P: New Macro.
3673 TARGET_ABM_P: New Macro.
3674 TARGET_BMI_P: New Macro.
3675 TARGET_BMI2_P: New Macro.
3676 TARGET_LZCNT_P: New Macro.
3677 TARGET_TBM_P: New Macro.
3678 TARGET_POPCNT_P: New Macro.
3679 TARGET_SAHF_P: New Macro.
3680 TARGET_MOVBE_P: New Macro.
3681 TARGET_CRC32_P: New Macro.
3682 TARGET_AES_P: New Macro.
3683 TARGET_PCLMUL_P: New Macro.
3684 TARGET_CMPXCHG16B_P: New Macro.
3685 TARGET_FSGSBASE_P: New Macro.
3686 TARGET_RDRND_P: New Macro.
3687 TARGET_F16C_P: New Macro.
3688 TARGET_RTM_P: New Macro.
3689 TARGET_HLE_P: New Macro.
3690 TARGET_RDSEED_P: New Macro.
3691 TARGET_PRFCHW_P: New Macro.
3692 TARGET_ADX_P: New Macro.
3693 TARGET_FXSR_P: New Macro.
3694 TARGET_XSAVE_P: New Macro.
3695 TARGET_XSAVEOPT_P: New Macro.
3696 TARGET_LP64_P: New Macro.
3697 TARGET_X32_P: New Macro.
3698 TARGET_FPMATH_DEFAULT_P: New Macro.
3699 TARGET_FLOAT_RETURNS_IN_80387_P: New Macro.
3700 * config/i386/i386.c (ix86_option_override_internal): New parameters.
3702 Change ix86_tune_string to access opts->x_ix86_tune_string.
3703 Change ix86_isa_flags to access opts->x_ix86_isa_flags.
3704 Change ix86_arch_string to access opts->x_ix86_arch_string.
3705 Change ix86_stringop_alg to access opts->x_ix86_stringop_alg.
3706 Change ix86_pmode to access opts->x_ix86_pmode.
3707 Change ix86_abi to access opts->x_ix86_abi.
3708 Change ix86_cmodel to access opts->x_ix86_cmodel.
3709 Change ix86_asm_dialect to access opts->x_ix86_asm_dialect.
3710 Change ix86_isa_flags_explicit to access
3711 opts->x_ix86_isa_flags_explicit.
3712 Change ix86_dump_tunes to access opts->x_ix86_dump_tunes.
3713 Change ix86_regparm to access opts->x_ix86_regparm.
3714 Change ix86_branch_cost to access opts->x_ix86_branch_cost.
3715 Change ix86_preferred_stack_boundary_arg to access
3716 opts->x_ix86_preferred_stack_boundary_arg.
3717 Change ix86_force_align_arg_pointer to access
3718 opts->x_ix86_force_align_arg_pointer.
3719 Change ix86_incoming_stack_boundar_arg to access
3720 opts->x_ix86_incoming_stack_boundar_arg.
3721 Change ix86_fpmath to access opts->x_ix86_fpmath.
3722 Change ix86_veclibabi_type to access opts->x_ix86_veclibabi_type.
3723 Change ix86_recip_name to access opts->x_ix86_recip_name.
3724 Change ix86_stack_protector_guard to access
3725 opts->x_ix86_stack_protector_guard.
3726 Change ix86_tune_memcpy_strategy to access
3727 opts->x_ix86_tune_memcpy_strategy.
3728 Change ix86_tune_memset_strategy to access
3729 opts->x_ix86_tune_memset_strategy.
3730 Change global_options to access opts.
3731 Change global_options_set to access opts_set.
3732 Change TARGET_64BIT to TARGET_64BIT_P (opts->...).
3733 Change TARGET_MMX to TARGET_MMX_P (opts->...).
3734 Change TARGET_3DNOW to TARGET_3DNOW_P (opts->...).
3735 Change TARGET_3DNOW_A to TARGET_3DNOW_A_P (opts->...).
3736 Change TARGET_SSE to TARGET_SSE_P (opts->...).
3737 Change TARGET_SSE2 to TARGET_SSE2_P (opts->...).
3738 Change TARGET_SSE3 to TARGET_SSE3_P (opts->...).
3739 Change TARGET_SSSE3 to TARGET_SSSE3_P (opts->...).
3740 Change TARGET_SSE4_1 to TARGET_SSE4_1_P (opts->...).
3741 Change TARGET_SSE4_2 to TARGET_SSE4_2_P (opts->...).
3742 Change TARGET_AVX to TARGET_AVX_P (opts->...).
3743 Change TARGET_AVX2 to TARGET_AVX2_P (opts->...).
3744 Change TARGET_AVX512F to TARGET_AVX512F_P (opts->...).
3745 Change TARGET_AVX512PF to TARGET_AVX512PF_P (opts->...).
3746 Change TARGET_AVX512ER to TARGET_AVX512ER_P (opts->...).
3747 Change TARGET_AVX512CD to TARGET_AVX512CD_P (opts->...).
3748 Change TARGET_FMA to TARGET_FMA_P (opts->...).
3749 Change TARGET_SSE4A to TARGET_SSE4A_P (opts->...).
3750 Change TARGET_FMA4 to TARGET_FMA4_P (opts->...).
3751 Change TARGET_XOP to TARGET_XOP_P (opts->...).
3752 Change TARGET_LWP to TARGET_LWP_P (opts->...).
3753 Change TARGET_ABM to TARGET_ABM_P (opts->...).
3754 Change TARGET_BMI to TARGET_BMI_P (opts->...).
3755 Change TARGET_BMI2 to TARGET_BMI2_P (opts->...).
3756 Change TARGET_LZCNT to TARGET_LZCNT_P (opts->...).
3757 Change TARGET_TBM to TARGET_TBM_P (opts->...).
3758 Change TARGET_POPCNT to TARGET_POPCNT_P (opts->...).
3759 Change TARGET_SAHF to TARGET_SAHF_P (opts->...).
3760 Change TARGET_MOVBE to TARGET_MOVBE_P (opts->...).
3761 Change TARGET_CRC32 to TARGET_CRC32_P (opts->...).
3762 Change TARGET_AES to TARGET_AES_P (opts->...).
3763 Change TARGET_PCLMUL to TARGET_PCLMUL_P (opts->...).
3764 Change TARGET_CMPXCHG16B to TARGET_CMPXCHG16B_P (opts->...).
3765 Change TARGET_FSGSBASE to TARGET_FSGSBASE_P (opts->...).
3766 Change TARGET_RDRND to TARGET_RDRND_P (opts->...).
3767 Change TARGET_F16C to TARGET_F16C_P (opts->...).
3768 Change TARGET_RTM to TARGET_RTM_P (opts->...).
3769 Change TARGET_HLE to TARGET_HLE_P (opts->...).
3770 Change TARGET_RDSEED to TARGET_RDSEED_P (opts->...).
3771 Change TARGET_PRFCHW to TARGET_PRFCHW_P (opts->...).
3772 Change TARGET_ADX to TARGET_ADX_P (opts->...).
3773 Change TARGET_FXSR to TARGET_FXSR_P (opts->...).
3774 Change TARGET_XSAVE to TARGET_XSAVE_P (opts->...).
3775 Change TARGET_XSAVEOPT to TARGET_XSAVEOPT_P (opts->...).
3776 Change TARGET_LP64 to TARGET_LP64_P (opts->...).
3777 Change TARGET_X32 to TARGET_X32_P (opts->...).
3778 Change TARGET_FPMATH_DEFAULT to TARGET_FPMATH_DEFAULT_P (opts->...).
3779 Change TARGET_FLOAT_RETURNS_IN_80387 to
3780 TARGET_FLOAT_RETURNS_IN_80387_P (opts->...).
3781 (ix86_function_specific_save): New parameter. Use opts-> fields
3782 to replace global fields.
3783 (ix86_function_specific_restore): Ditto.
3784 (ix86_valid_target_attribute_inner_p): New parameters.
3786 Fix call to ix86_handle_option and set_option.
3787 (ix86_valid_target_attribute_tree): New parameters.
3788 Change global_options to access opts.
3789 Change global_options_set to access opts_set.
3790 Fix call to ix86_valid_target_attribute_inner_p.
3791 Change ix86_tune_string to access opts->x_ix86_tune_string.
3792 Change ix86_arch_string to access opts->x_ix86_arch_string.
3793 Change ix86_fpmath to access opts->x_ix86_fpmath
3794 Fix call to ix86_option_override_internal.
3795 Fix call to ix86_add_new_builtins.
3796 Fix calls to build_optimization_node and build_target_option_node.
3797 (ix86_valid_target_attribute_p): Remove access to global_options.
3798 Use new gcc_options structure func_options.
3799 Fix call to ix86_valid_target_attribute_tree.
3800 Fix call to build_optimization_node.
3801 (get_builtin_code_for_version): Fix call to
3802 ix86_valid_target_attribute_tree.
3804 2013-10-15 David Malcolm <dmalcolm@redhat.com>
3806 * Makefile.in (PICFLAG): New.
3807 (enable_host_shared): New.
3808 (INTERNAL_CFLAGS): Use PICFLAG.
3809 (LIBIBERTY): Use pic build of libiberty.a if configured with
3810 --enable-host-shared.
3811 * configure.ac: Add --enable-host-shared, setting up new
3813 * configure: Regenerate.
3814 * doc/install.texi (--enable-shared): Add note contrasting it with ...
3815 (--enable-host-shared): New option.
3817 2013-10-15 Richard Biener <rguenther@suse.de>
3819 * tree-tailcall.c (find_tail_calls): Don't use tail-call recursion
3820 for built-in functions.
3822 2013-10-15 Zhenqiang Chen <zhenqiang.chen@arm.com>
3824 * tree-ssa-reassoc.c: Include rtl.h and tm_p.h.
3825 (optimize_range_tests_1): New function,
3826 extracted from optimize_range_tests.
3827 (optimize_range_tests_xor): Similarly.
3828 (optimize_range_tests_diff): New function.
3829 (optimize_range_tests): Use optimize_range_tests_1.
3831 2013-10-15 Cong Hou <congh@google.com>
3833 * tree-vect-loop.c (vect_is_simple_reduction_1): Relax the
3834 requirement of the reduction pattern so that one operand of the
3835 reduction operation can come from outside of the loop.
3837 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3839 * config/arm/neon-schedgen.ml: Remove.
3840 * config/arm/cortex-a9-neon.md: Remove comment regarding
3843 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3845 * config/arm/types: Remove old neon types.
3847 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3849 * config/arm/cortex-a7.md
3850 (cortex_a7_neon_type): New.
3851 (cortex_a7_neon_mul): Update for new types.
3852 (cortex_a7_neon_mla): Likewise.
3853 (cortex_a7_neon): Likewise.
3855 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3857 * config/arm/cortex-a15-neon.md
3858 (cortex_a15_neon_type): New,
3860 (cortex_a15_neon_int_1): Remove.
3861 (cortex_a15_neon_int_2): Likewise.
3862 (cortex_a15_neon_int_3): Likewise.
3863 (cortex_a15_neon_int_4): Likewise.
3864 (cortex_a15_neon_int_5): Likewise.
3865 (cortex_a15_neon_vqneg_vqabs): Likewise.
3866 (cortex_a15_neon_vmov): Likewise.
3867 (cortex_a15_neon_vaba): Likewise.
3868 (cortex_a15_neon_vaba_qqq): Likewise.
3869 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3870 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
3871 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
3873 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3874 (cortex_a15_neon_mla_qqq_8_16): Likewise.
3875 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar): Likewise.
3876 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
3877 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
3878 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
3879 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
3880 (cortex_a15_neon_shift_1): Likewise.
3881 (cortex_a15_neon_shift_2): Likewise.
3882 (cortex_a15_neon_shift_3): Likewise.
3883 (cortex_a15_neon_vshl_ddd): Likewise.
3884 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
3885 (cortex_a15_neon_vsra_vrsra): Likewise.
3886 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
3887 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
3888 (cortex_a15_neon_bp_3cycle): Likewise.
3889 (cortex_a15_neon_ldm_2): Likewise.
3890 (cortex_a15_neon_stm_2): Likewise.
3891 (cortex_a15_neon_mcr): Likewise.
3892 (cortex_a15_neon_mrc): Likewise.
3893 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
3894 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
3895 (cortex_a15_neon_fp_vmul_ddd): Likewise.
3896 (cortex_a15_neon_fp_vmul_qqd): Likewise.
3897 (cortex_a15_neon_fp_vmla_ddd): Likewise.
3898 (cortex_a15_neon_fp_vmla_qqq): Likewise.
3899 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
3900 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
3901 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
3902 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
3903 (cortex_a15_neon_bp_simple): Likewise.
3904 (cortex_a15_neon_bp_2cycle): Likewise.
3905 (cortex_a15_neon_bp_3cycle): Likewise.
3906 (cortex_a15_neon_vld1_1_2_regs): Likewise.
3907 (cortex_a15_neon_vld1_3_4_regs): Likewise.
3908 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
3909 (cortex_a15_neon_vld2_4_regs): Likewise.
3910 (cortex_a15_neon_vld3_vld4): Likewise.
3911 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
3912 (cortex_a15_neon_vst1_3_4_regs): Likewise.
3913 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Rename to...
3914 (cortex_a15_neon_vst2_4_regs_vst3): ...This, update for new attributes.
3915 (cortex_a15_neon_vst3_vst4): Rename to...
3916 (cortex_a15_neon_vst4): This, update for new attributes.
3917 (cortex_a15_neon_vld1_vld2_lane): Update for new attributes.
3918 (cortex_a15_neon_vld3_vld4_lane): Likewise.
3919 (cortex_a15_neon_vst1_vst2_lane): Likewise.
3920 (cortex_a15_neon_vst3_vst4_lane): Likewise.
3921 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
3922 (cortex_a15_neon_ldm_2): Likewise.
3923 (cortex_a15_neon_stm_2): Likewise.
3924 (cortex_a15_neon_mcr): Likewise.
3925 (cortex_a15_neon_mcr_2_mcrr): Likewise.
3926 (cortex_a15_neon_mrc): Likewise.
3927 (cortex_a15_neon_mrrc): Likewise.
3929 (cortex_a15_neon_abd): New.
3930 (cortex_a15_neon_abd_q): Likewise.
3931 (cortex_a15_neon_aba): Likewise.
3932 (cortex_a15_neon_aba_q): Likewise.
3933 (cortex_a15_neon_acc): Likewise.
3934 (cortex_a15_neon_acc_q): Likewise.
3935 (cortex_a15_neon_arith_basic): Likewise.
3936 (cortex_a15_neon_arith_complex): Likewise.
3937 (cortex_a15_neon_multiply): Likewise.
3938 (cortex_a15_neon_multiply_q): Likewise.
3939 (cortex_a15_neon_mla): Likewise.
3940 (cortex_a15_neon_mla_q): Likewise.
3941 (cortex_a15_neon_sat_mla_long): Likewise.
3942 (cortex_a15_neon_shift_acc): Likewise.
3943 (cortex_a15_neon_shift_imm_basic): Likewise.
3944 (cortex_a15_neon_shift_imm_complex): Likewise.
3945 (cortex_a15_neon_shift_reg_basic): Likewise.
3946 (cortex_a15_neon_shift_reg_basic_q): Likewise.
3947 (cortex_a15_neon_shift_reg_complex): Likewise.
3948 (cortex_a15_neon_shift_reg_complex_q): Likewise.
3949 (cortex_a15_neon_fp_negabs): Likewise
3950 (cortex_a15_neon_fp_arith): Likewise
3951 (cortex_a15_neon_fp_arith_q): Likewise
3952 (cortex_a15_neon_fp_cvt_int): Likewise
3953 (cortex_a15_neon_fp_cvt_int_q): Likewise
3954 (cortex_a15_neon_fp_cvt_16): Likewise
3955 (cortex_a15_neon_fp_mul): Likewise
3956 (cortex_a15_neon_fp_mul_q): Likewise
3957 (cortex_a15_neon_fp_mla): Likewise
3958 (cortex_a15_neon_fp_mla_q): Likewise
3959 (cortex_a15_neon_fp_recps_rsqrte): Likewise.
3960 (cortex_a15_neon_fp_recps_rsqrte_q): Likewise.
3961 (cortex_a15_neon_bitops): Likewise.
3962 (cortex_a15_neon_bitops_q): Likewise.
3963 (cortex_a15_neon_from_gp): Likewise.
3964 (cortex_a15_neon_from_gp_q): Likewise.
3965 (cortex_a15_neon_tbl3_tbl4): Likewise.
3966 (cortex_a15_neon_zip_q): Likewise.
3967 (cortex_a15_neon_to_gp): Likewise.
3968 (cortex_a15_neon_load_a): Likewise.
3969 (cortex_a15_neon_load_b): Likewise.
3970 (cortex_a15_neon_load_c): Likewise.
3971 (cortex_a15_neon_load_d): Likewise.
3972 (cortex_a15_neon_load_e): Likewise.
3973 (cortex_a15_neon_load_f): Likewise.
3974 (cortex_a15_neon_store_a): Likewise.
3975 (cortex_a15_neon_store_b): Likewise.
3976 (cortex_a15_neon_store_c): Likewise.
3977 (cortex_a15_neon_store_d): Likewise.
3978 (cortex_a15_neon_store_e): Likewise.
3979 (cortex_a15_neon_store_f): Likewise.
3980 (cortex_a15_neon_store_g): Likewise.
3981 (cortex_a15_neon_store_h): Likewise.
3982 (cortex_a15_vfp_to_from_gp): Likewise.
3984 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3986 * config/arm/cortex-a9-neon.md (cortex_a9_neon_type): New.
3988 (cortex_a9_neon_vshl_ddd): Remove.
3989 (cortex_a9_neon_vst3_vst4): Likewise.
3990 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
3992 (cortex_a9_neon_bit_ops_q): New.
3994 (cortex_a9_neon_int_1): Use cortex_a8_neon_type.
3995 (cortex_a9_neon_int_2): Likewise.
3996 (cortex_a9_neon_int_3): Likewise.
3997 (cortex_a9_neon_int_4): Likewise.
3998 (cortex_a9_neon_int_5): Likewise.
3999 (cortex_a9_neon_vqneg_vqabs): Likewise.
4000 (cortex_a9_neon_vmov): Likewise.
4001 (cortex_a9_neon_vaba): Likewise.
4002 (cortex_a9_neon_vaba_qqq): Likewise.
4003 (cortex_a9_neon_shift_1): Likewise.
4004 (cortex_a9_neon_shift_2): Likewise.
4005 (cortex_a9_neon_shift_3): Likewise.
4006 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
4007 (cortex_a9_neon_vsra_vrsra): Likewise.
4008 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
4009 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
4010 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
4012 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
4013 (cortex_a9_neon_mla_qqq_8_16): Likewise.
4014 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
4016 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
4017 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
4018 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
4019 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
4020 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
4021 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
4022 (cortex_a9_neon_fp_vsum): Likewise.
4023 (cortex_a9_neon_fp_vmul_ddd): Likewise.
4024 (cortex_a9_neon_fp_vmul_qqd): Likewise.
4025 (cortex_a9_neon_fp_vmla_ddd): Likewise.
4026 (cortex_a9_neon_fp_vmla_qqq): Likewise.
4027 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
4028 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
4029 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
4030 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
4031 (cortex_a9_neon_bp_simple): Likewise.
4032 (cortex_a9_neon_bp_2cycle): Likewise.
4033 (cortex_a9_neon_bp_3cycle): Likewise.
4034 (cortex_a9_neon_ldr): Likewise.
4035 (cortex_a9_neon_str): Likewise.
4036 (cortex_a9_neon_vld1_1_2_regs): Likewise.
4037 (cortex_a9_neon_vld1_3_4_regs): Likewise.
4038 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
4039 (cortex_a9_neon_vld2_4_regs): Likewise.
4040 (cortex_a9_neon_vld3_vld4): Likewise.
4041 (cortex_a9_neon_vld1_vld2_lane): Likewise.
4042 (cortex_a9_neon_vld3_vld4_lane): Likewise.
4043 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
4044 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
4045 (cortex_a9_neon_vst1_3_4_regs): Likewise.
4046 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
4047 (cortex_a9_neon_vst1_vst2_lane): Likewise.
4048 (cortex_a9_neon_vst3_vst4_lane): Likewise.
4049 (cortex_a9_neon_mcr): Likewise.
4050 (cortex_a9_neon_mcr_2_mcrr): Likewise.
4051 (cortex_a9_neon_mrc): Likewise.
4052 (cortex_a9_neon_mrrc): Likewise.
4054 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
4056 * config/arm/cortex-a8-neon.md (cortex_a8_neon_type): New.
4058 (cortex_a8_neon_vshl_ddd): Remove.
4059 (cortex_a8_neon_vst3_vst4): Likewise.
4060 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
4062 (cortex_a8_neon_bit_ops_q): New.
4064 (cortex_a8_neon_int_1): Use cortex_a8_neon_type.
4065 (cortex_a8_neon_int_2): Likewise..
4066 (cortex_a8_neon_int_3): Likewise.
4067 (cortex_a8_neon_int_5): Likewise.
4068 (cortex_a8_neon_vqneg_vqabs): Likewise.
4069 (cortex_a8_neon_int_4): Likewise.
4070 (cortex_a8_neon_vaba): Likewise.
4071 (cortex_a8_neon_vaba_qqq): Likewise.
4072 (cortex_a8_neon_shift_1): Likewise.
4073 (cortex_a8_neon_shift_2): Likewise.
4074 (cortex_a8_neon_shift_3): Likewise.
4075 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
4076 (cortex_a8_neon_vsra_vrsra): Likewise.
4077 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
4078 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
4079 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
4081 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
4082 (cortex_a8_neon_mla_qqq_8_16): Likewise.
4083 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
4085 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
4086 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
4087 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
4088 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
4089 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
4090 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
4091 (cortex_a8_neon_fp_vsum): Likewise.
4092 (cortex_a8_neon_fp_vmul_ddd): Likewise.
4093 (cortex_a8_neon_fp_vmul_qqd): Likewise.
4094 (cortex_a8_neon_fp_vmla_ddd): Likewise.
4095 (cortex_a8_neon_fp_vmla_qqq): Likewise.
4096 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
4097 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
4098 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
4099 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
4100 (cortex_a8_neon_bp_simple): Likewise.
4101 (cortex_a8_neon_bp_2cycle): Likewise.
4102 (cortex_a8_neon_bp_3cycle): Likewise.
4103 (cortex_a8_neon_ldr): Likewise.
4104 (cortex_a8_neon_str): Likewise.
4105 (cortex_a8_neon_vld1_1_2_regs): Likewise.
4106 (cortex_a8_neon_vld1_3_4_regs): Likewise.
4107 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
4108 (cortex_a8_neon_vld2_4_regs): Likewise.
4109 (cortex_a8_neon_vld3_vld4): Likewise.
4110 (cortex_a8_neon_vld1_vld2_lane): Likewise.
4111 (cortex_a8_neon_vld3_vld4_lane): Likewise.
4112 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
4113 (cortex_a8_neon_vst1_3_4_regs): Likewise.
4114 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
4115 (cortex_a8_neon_vst1_vst2_lane): Likewise.
4116 (cortex_a8_neon_vst3_vst4_lane): Likewise.
4117 (cortex_a8_neon_mcr): Likewise.
4118 (cortex_a8_neon_mcr_2_mcrr): Likewise.
4119 (cortex_a8_neon_mrc): Likewise.
4120 (cortex_a8_neon_mrrc): Likewise.
4122 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
4124 * config/aarch64/iterators.md (Vetype): Add SF and DF modes.
4126 * config/aarch64/aarch64-simd.md (neon_type): Remove.
4127 (aarch64_simd_dup<mode>): Add "type" attribute.
4128 (aarch64_dup_lane<mode>): Likewise.
4129 (aarch64_dup_lane_<vswap_width_name><mode>): Likewise.
4130 (*aarch64_simd_mov<mode>): Likewise.
4131 (aarch64_simd_mov_from_<mode>low): Likewise.
4132 (aarch64_simd_mov_from_<mode>high): Likewise.
4133 (orn<mode>3): Likewise.
4134 (bic<mode>3): Likewise.
4135 (add<mode>3): Likewise.
4136 (sub<mode>3): Likewise.
4137 (mul<mode>3): Likewise.
4138 (*aarch64_mul3_elt<mode>): Likewise.
4139 (*aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
4140 (*aarch64_mul3_elt_to_128df): Likewise.
4141 (*aarch64_mul3_elt_to_64v2df): Likewise.
4142 (neg<mode>2): Likewise.
4143 (abs<mode>2): Likewise.
4144 (abd<mode>_3): Likewise.
4145 (aba<mode>_3): Likewise.
4146 (fabd<mode>_3): Likewise.
4147 (*fabd_scalar<mode>3): Likewise.
4148 (and<mode>3): Likewise.
4149 (ior<mode>3): Likewise.
4150 (xor<mode>3): Likewise.
4151 (one_cmpl<mode>2): Likewise.
4152 (aarch64_simd_vec_set<mode>): Likewise.
4153 (aarch64_simd_lshr<mode>): Likewise.
4154 (aarch64_simd_ashr<mode>): Likewise.
4155 (aarch64_simd_imm_shl<mode>): Likewise.
4156 (aarch64_simd_reg_sshl<mode): Likewise.
4157 (aarch64_simd_reg_shl<mode>_unsigned): Likewise.
4158 (aarch64_simd_reg_shl<mode>_signed): Likewise.
4159 (aarch64_simd_vec_setv2di): Likewise.
4160 (aarch64_simd_vec_set<mode>): Likewise.
4161 (aarch64_mla<mode>): Likewise.
4162 (*aarch64_mla_elt<mode>): Likewise.
4163 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
4164 (aarch64_mls<mode>): Likewise.
4165 (*aarch64_mls_elt<mode>): Likewise.
4166 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
4167 (<su><maxmin><mode>3): Likewise.
4168 (move_lo_quad_<mode>): Likewise.
4169 (aarch64_simd_move_hi_quad_<mode>): Likewise.
4170 (aarch64_simd_vec_pack_trunc_<mode>): Likewise.
4171 (vec_pack_trunc_<mode>): Likewise.
4172 (aarch64_simd_vec_unpack<su>_lo_<mode>): Likewise.
4173 (aarch64_simd_vec_unpack<su>_hi_<mode>): Likewise.
4174 (*aarch64_<su>mlal_lo<mode>): Likewise.
4175 (*aarch64_<su>mlal_hi<mode>): Likewise.
4176 (*aarch64_<su>mlsl_lo<mode>): Likewise.
4177 (*aarch64_<su>mlsl_hi<mode>): Likewise.
4178 (*aarch64_<su>mlal<mode>): Likewise.
4179 (*aarch64_<su>mlsl<mode>): Likewise.
4180 (aarch64_simd_vec_<su>mult_lo_<mode>): Likewise.
4181 (aarch64_simd_vec_<su>mult_hi_<mode>): Likewise.
4182 (add<mode>3): Likewise.
4183 (sub<mode>3): Likewise.
4184 (mul<mode>3): Likewise.
4185 (div<mode>3): Likewise.
4186 (neg<mode>2): Likewise.
4187 (abs<mode>2): Likewise.
4188 (fma<mode>4): Likewise.
4189 (*aarch64_fma4_elt<mode>): Likewise.
4190 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
4191 (*aarch64_fma4_elt_to_128df): Likewise.
4192 (*aarch64_fma4_elt_to_64v2df): Likewise.
4193 (fnma<mode>4): Likewise.
4194 (*aarch64_fnma4_elt<mode>): Likewise.
4195 (*aarch64_fnma4_elt_<vswap_width_name><mode>
4196 (*aarch64_fnma4_elt_to_128df): Likewise.
4197 (*aarch64_fnma4_elt_to_64v2df): Likewise.
4198 (<frint_pattern><mode>2): Likewise.
4199 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): Likewise.
4200 (<optab><fcvt_target><VDQF:VDQF:mode>2): Likewise.
4201 (vec_unpacks_lo_v4sf): Likewise.
4202 (aarch64_float_extend_lo_v2df): Likewise.
4203 (vec_unpacks_hi_v4sf): Likewise.
4204 (aarch64_float_truncate_lo_v2sf): Likewise.
4205 (aarch64_float_truncate_hi_v4sf): Likewise.
4206 (aarch64_vmls<mode>): Likewise.
4207 (<su><maxmin><mode>3): Likewise.
4208 (<maxmin_uns><mode>3): Likewise.
4209 (reduc_<sur>plus_<mode>): Likewise.
4210 (reduc_<sur>plus_v2di): Likewise.
4211 (reduc_<sur>plus_v2si): Likewise.
4212 (reduc_<sur>plus_<mode>): Likewise.
4213 (aarch64_addpv4sf): Likewise.
4214 (clz<mode>2): Likewise.
4215 (reduc_<maxmin_uns>_<mode>): Likewise.
4216 (reduc_<maxmin_uns>_v2di): Likewise.
4217 (reduc_<maxmin_uns>_v2si): Likewise.
4218 (reduc_<maxmin_uns>_<mode>): Likewise.
4219 (reduc_<maxmin_uns>_v4sf): Likewise.
4220 (aarch64_simd_bsl<mode>_internal): Likewise.
4221 (*aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): Likewise.
4222 (*aarch64_get_lane_zero_extendsi<mode>): Likewise.
4223 (aarch64_get_lane<mode>): Likewise.
4224 (*aarch64_combinez<mode>): Likewise.
4225 (aarch64_combine<mode>): Likewise.
4226 (aarch64_simd_combine<mode>): Likewise.
4227 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): Likewise.
4228 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): Likewise.
4229 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>): Likewise.
4230 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w<mode>): Likewise.
4231 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w2<mode>_internal): Likewise.
4232 (aarch64_<sur>h<addsub><mode>): Likewise.
4233 (aarch64_<sur><addsub>hn<mode>): Likewise.
4234 (aarch64_<sur><addsub>hn2<mode>): Likewise.
4235 (aarch64_pmul<mode>): Likewise.
4236 (aarch64_<su_optab><optab><mode>): Likewise.
4237 (aarch64_<sur>qadd<mode>): Likewise.
4238 (aarch64_sqmovun<mode>): Likewise.
4239 (aarch64_<sur>qmovn<mode>): Likewise.
4240 (aarch64_s<optab><mode>): Likewise.
4241 (aarch64_sq<r>dmulh<mode>): Likewise.
4242 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
4243 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
4244 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
4245 (aarch64_sqdml<SBINQOPS:as>l<mode>): Likewise.
4246 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
4247 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
4248 (aarch64_sqdml<SBINQOPS:as>l_n<mode>): Likewise.
4249 (aarch64_sqdml<SBINQOPS:as>l2<mode>_internal): Likewise.
4250 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
4251 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
4252 (aarch64_sqdmull<mode>): Likewise.
4253 (aarch64_sqdmull_lane<mode>_internal): Likewise.
4254 (aarch64_sqdmull_n<mode>): Likewise.
4255 (aarch64_sqdmull2<mode>_internal): Likewise.
4256 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
4257 (aarch64_sqdmull2_n<mode>_internal): Likewise.
4258 (aarch64_<sur>shl<mode>): Likewise.
4259 (aarch64_<sur>q<r>shl<mode>
4260 (aarch64_<sur>shll_n<mode>): Likewise.
4261 (aarch64_<sur>shll2_n<mode>): Likewise.
4262 (aarch64_<sur>shr_n<mode>): Likewise.
4263 (aarch64_<sur>sra_n<mode>): Likewise.
4264 (aarch64_<sur>s<lr>i_n<mode>): Likewise.
4265 (aarch64_<sur>qshl<u>_n<mode>): Likewise.
4266 (aarch64_<sur>q<r>shr<u>n_n<mode>): Likewise.
4267 (aarch64_cm<optab><mode>): Likewise.
4268 (aarch64_cm<optab>di): Likewise.
4269 (aarch64_cm<optab><mode>): Likewise.
4270 (aarch64_cm<optab>di): Likewise.
4271 (aarch64_cmtst<mode>): Likewise.
4272 (aarch64_cmtstdi): Likewise.
4273 (aarch64_cm<optab><mode>): Likewise.
4274 (*aarch64_fac<optab><mode>): Likewise.
4275 (aarch64_addp<mode>): Likewise.
4276 (aarch64_addpdi): Likewise.
4277 (sqrt<mode>2): Likewise.
4278 (vec_load_lanesoi<mode>): Likewise.
4279 (vec_store_lanesoi<mode>): Likewise.
4280 (vec_load_lanesci<mode>): Likewise.
4281 (vec_store_lanesci<mode>): Likewise.
4282 (vec_load_lanesxi<mode>): Likewise.
4283 (vec_store_lanesxi<mode>): Likewise.
4284 (*aarch64_mov<mode>): Likewise.
4285 (aarch64_ld2<mode>_dreg): Likewise.
4286 (aarch64_ld2<mode>_dreg): Likewise.
4287 (aarch64_ld3<mode>_dreg): Likewise.
4288 (aarch64_ld3<mode>_dreg): Likewise.
4289 (aarch64_ld4<mode>_dreg): Likewise.
4290 (aarch64_ld4<mode>_dreg): Likewise.
4291 (aarch64_tbl1<mode>): Likewise.
4292 (aarch64_tbl2v16qi): Likewise.
4293 (aarch64_combinev16qi): Likewise.
4294 (aarch64_<PERMUTE:perm_insn><PERMUTE:perm_hilo><mode>): Likewise.
4295 (aarch64_st2<mode>_dreg): Likewise.
4296 (aarch64_st2<mode>_dreg): Likewise.
4297 (aarch64_st3<mode>_dreg): Likewise.
4298 (aarch64_st3<mode>_dreg): Likewise.
4299 (aarch64_st4<mode>_dreg): Likewise.
4300 (aarch64_st4<mode>_dreg): Likewise.
4301 (*aarch64_simd_ld1r<mode>): Likewise.
4302 (aarch64_frecpe<mode>): Likewise.
4303 (aarch64_frecp<FRECP:frecp_suffix><mode>): Likewise.
4304 (aarch64_frecps<mode>): Likewise.
4306 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
4308 * config/arm/iterators.md (V_elem_ch): New.
4310 (VQH_type): Likewise.
4311 * config/arm/arm.md (is_neon_type): New.
4312 (conds): Use is_neon_type.
4313 (anddi3_insn): Update type attribute.
4314 (xordi3_insn): Likewise.
4315 (one_cmpldi2): Likewise.
4316 * config/arm/vfp.md (movhf_vfp_neon): Update type attribute.
4317 * config/arm/neon.md (neon_mov): Update type attribute.
4318 (*movmisalign<mode>_neon_store): Likewise.
4319 (*movmisalign<mode>_neon_load): Likewise.
4320 (vec_set<mode>_internal): Likewise.
4321 (vec_set<mode>_internal): Likewise.
4322 (vec_setv2di_internal): Likewise.
4323 (vec_extract<mode>): Likewise.
4324 (vec_extract<mode>): Likewise.
4325 (vec_extractv2di): Likewise.
4326 (*add<mode>3_neon): Likewise.
4327 (adddi3_neon): Likewise.
4328 (*sub<mode>3_neon): Likewise.
4329 (subdi3_neon): Likewise.
4330 (fma<VCVTF:mode>4): Likewise.
4331 (fma<VCVTF:mode>4_intrinsic): Likewise.
4332 (*fmsub<VCVTF:mode>4): Likewise.
4333 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
4334 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
4335 (ior<mode>3): Likewise.
4336 (and<mode>3): Likewise.
4337 (orn<mode>3_neon): Likewise.
4338 (orndi3_neon): Likewise.
4339 (bic<mode>3_neon): Likewise.
4340 (bicdi3_neon): Likewise.
4341 (xor<mode>3): Likewise.
4342 (one_cmpl<mode>2): Likewise.
4343 (abs<mode>2): Likewise.
4344 (neg<mode>2): Likewise.
4345 (negdi2_neon): Likewise.
4346 (*umin<mode>3_neon): Likewise.
4347 (*umax<mode>3_neon): Likewise.
4348 (*smin<mode>3_neon): Likewise.
4349 (*smax<mode>3_neon): Likewise.
4350 (vashl<mode>3): Likewise.
4351 (vashr<mode>3_imm): Likewise.
4352 (vlshr<mode>3_imm): Likewise.
4353 (ashl<mode>3_signed): Likewise.
4354 (ashl<mode>3_unsigned): Likewise.
4355 (neon_load_count): Likewise.
4356 (ashldi3_neon_noclobber): Likewise.
4357 (ashldi3_neon): Likewise.
4358 (signed_shift_di3_neon): Likewise.
4359 (unsigned_shift_di3_neon): Likewise.
4360 (ashrdi3_neon_imm_noclobber): Likewise.
4361 (lshrdi3_neon_imm_noclobber): Likewise.
4362 (<shift>di3_neon): Likewise.
4363 (widen_ssum<mode>3): Likewise.
4364 (widen_usum<mode>3): Likewise.
4365 (quad_halves_<code>v4si): Likewise.
4366 (quad_halves_<code>v4sf): Likewise.
4367 (quad_halves_<code>v8hi): Likewise.
4368 (quad_halves_<code>v16qi): Likewise.
4369 (reduc_splus_v2di): Likewise.
4370 (neon_vpadd_internal<mode>): Likewise.
4371 (neon_vpsmin<mode>): Likewise.
4372 (neon_vpsmax<mode>): Likewise.
4373 (neon_vpumin<mode>): Likewise.
4374 (neon_vpumax<mode>): Likewise.
4375 (*ss_add<mode>_neon): Likewise.
4376 (*us_add<mode>_neon): Likewise.
4377 (*ss_sub<mode>_neon): Likewise.
4378 (*us_sub<mode>_neon): Likewise.
4379 (neon_vadd<mode>_unspec): Likewise.
4380 (neon_vaddl<mode>): Likewise.
4381 (neon_vaddw<mode>): Likewise.
4382 (neon_vhadd<mode>): Likewise.
4383 (neon_vqadd<mode>): Likewise.
4384 (neon_vaddhn<mode>): Likewise.
4385 (neon_vmul<mode>): Likewise.
4386 (neon_vfms<VCVTF:mode>): Likewise.
4387 (neon_vmlal<mode>): Likewise.
4388 (neon_vmls<mode>): Likewise.
4389 (neon_vmlsl<mode>): Likewise.
4390 (neon_vqdmulh<mode>): Likewise.
4391 (neon_vqdmlal<mode>): Likewise.
4392 (neon_vqdmlsl<mode>): Likewise.
4393 (neon_vmull<mode>): Likewise.
4394 (neon_vqdmull<mode>): Likewise.
4395 (neon_vsub<mode>_unspec): Likewise.
4396 (neon_vsubl<mode>): Likewise.
4397 (neon_vsubw<mode>): Likewise.
4398 (neon_vqsub<mode>): Likewise.
4399 (neon_vhsub<mode>): Likewise.
4400 (neon_vsubhn<mode>): Likewise.
4401 (neon_vceq<mode>): Likewise.
4402 (neon_vcge<mode>): Likewise.
4403 (neon_vcgeu<mode>): Likewise.
4404 (neon_vcgt<mode>): Likewise.
4405 (neon_vcgtu<mode>): Likewise.
4406 (neon_vcle<mode>): Likewise.
4407 (neon_vclt<mode>): Likewise.
4408 (neon_vcage<mode>): Likewise.
4409 (neon_vcagt<mode>): Likewise.
4410 (neon_vtst<mode>): Likewise.
4411 (neon_vabd<mode>): Likewise.
4412 (neon_vabdl<mode>): Likewise.
4413 (neon_vaba<mode>): Likewise.
4414 (neon_vabal<mode>): Likewise.
4415 (neon_vmax<mode>): Likewise.
4416 (neon_vmin<mode>): Likewise.
4417 (neon_vpaddl<mode>): Likewise.
4418 (neon_vpadal<mode>): Likewise.
4419 (neon_vpmax<mode>): Likewise.
4420 (neon_vpmin<mode>): Likewise.
4421 (neon_vrecps<mode>): Likewise.
4422 (neon_vrsqrts<mode>): Likewise.
4423 (neon_vqabs<mode>): Likewise.
4424 (neon_vqneg<mode>): Likewise.
4425 (neon_vcls<mode>): Likewise.
4426 (clz<mode>2): Likewise.
4427 (popcount<mode>2): Likewise.
4428 (neon_vrecpe<mode>): Likewise.
4429 (neon_vrsqrte<mode>): Likewise.
4430 (neon_vget_lane<mode>_sext_internal): Likewise.
4431 (neon_vget_lane<mode>_zext_internal): Likewise.
4432 (neon_vdup_n<mode>): Likewise.
4433 (neon_vdup_n<mode>): Likewise.
4434 (neon_vdup_nv2di): Likewise.
4435 (neon_vdup_lane<mode>_interal): Likewise.
4436 (*neon_vswp<mode>): Likewise.
4437 (neon_vcombine<mode>): Likewise.
4438 (float<mode><V_cvtto>2): Likewise.
4439 (floatuns<mode><V_cvtto>2): Likewise.
4440 (fix_trunc<mode><V_cvtto>2): Likewise.
4441 (fixuns_trunc<mode><V_cvtto>2
4442 (neon_vcvt<mode>): Likewise.
4443 (neon_vcvt<mode>): Likewise.
4444 (neon_vcvtv4sfv4hf): Likewise.
4445 (neon_vcvtv4hfv4sf): Likewise.
4446 (neon_vcvt_n<mode>): Likewise.
4447 (neon_vcvt_n<mode>): Likewise.
4448 (neon_vmovn<mode>): Likewise.
4449 (neon_vqmovn<mode>): Likewise.
4450 (neon_vqmovun<mode>): Likewise.
4451 (neon_vmovl<mode>): Likewise.
4452 (neon_vmul_lane<mode>): Likewise.
4453 (neon_vmul_lane<mode>): Likewise.
4454 (neon_vmull_lane<mode>): Likewise.
4455 (neon_vqdmull_lane<mode>): Likewise.
4456 (neon_vqdmulh_lane<mode>): Likewise.
4457 (neon_vqdmulh_lane<mode>): Likewise.
4458 (neon_vmla_lane<mode>): Likewise.
4459 (neon_vmla_lane<mode>): Likewise.
4460 (neon_vmlal_lane<mode>): Likewise.
4461 (neon_vqdmlal_lane<mode>): Likewise.
4462 (neon_vmls_lane<mode>): Likewise.
4463 (neon_vmls_lane<mode>): Likewise.
4464 (neon_vmlsl_lane<mode>): Likewise.
4465 (neon_vqdmlsl_lane<mode>): Likewise.
4466 (neon_vext<mode>): Likewise.
4467 (neon_vrev64<mode>): Likewise.
4468 (neon_vrev32<mode>): Likewise.
4469 (neon_vrev16<mode>): Likewise.
4470 (neon_vbsl<mode>_internal): Likewise.
4471 (neon_vshl<mode>): Likewise.
4472 (neon_vqshl<mode>): Likewise.
4473 (neon_vshr_n<mode>): Likewise.
4474 (neon_vshrn_n<mode>): Likewise.
4475 (neon_vqshrn_n<mode>): Likewise.
4476 (neon_vqshrun_n<mode>): Likewise.
4477 (neon_vshl_n<mode>): Likewise.
4478 (neon_vqshl_n<mode>): Likewise.
4479 (neon_vqshlu_n<mode>): Likewise.
4480 (neon_vshll_n<mode>): Likewise.
4481 (neon_vsra_n<mode>): Likewise.
4482 (neon_vsri_n<mode>): Likewise.
4483 (neon_vsli_n<mode>): Likewise.
4484 (neon_vtbl1v8qi): Likewise.
4485 (neon_vtbl2v8qi): Likewise.
4486 (neon_vtbl3v8qi): Likewise.
4487 (neon_vtbl4v8qi): Likewise.
4488 (neon_vtbl1v16qi): Likewise.
4489 (neon_vtbl2v16qi): Likewise.
4490 (neon_vcombinev16qi): Likewise.
4491 (neon_vtbx1v8qi): Likewise.
4492 (neon_vtbx2v8qi): Likewise.
4493 (neon_vtbx3v8qi): Likewise.
4494 (neon_vtbx4v8qi): Likewise.
4495 (*neon_vtrn<mode>_insn): Likewise.
4496 (*neon_vzip<mode>_insn): Likewise.
4497 (*neon_vuzp<mode>_insn): Likewise.
4498 (neon_vld1<mode>): Likewise.
4499 (neon_vld1_lane<mode>): Likewise.
4500 (neon_vld1_lane<mode>): Likewise.
4501 (neon_vld1_dup<mode>): Likewise.
4502 (neon_vld1_dup<mode>): Likewise.
4503 (neon_vld1_dupv2di): Likewise.
4504 (neon_vst1<mode>): Likewise.
4505 (neon_vst1_lane<mode>): Likewise.
4506 (neon_vst1_lane<mode>): Likewise.
4507 (neon_vld2<mode>): Likewise.
4508 (neon_vld2<mode>): Likewise.
4509 (neon_vld2_lane<mode>): Likewise.
4510 (neon_vld2_lane<mode>): Likewise.
4511 (neon_vld2_dup<mode>): Likewise.
4512 (neon_vst2<mode>): Likewise.
4513 (neon_vst2<mode>): Likewise.
4514 (neon_vst2_lane<mode>): Likewise.
4515 (neon_vst2_lane<mode>): Likewise.
4516 (neon_vld3<mode>): Likewise.
4517 (neon_vld3qa<mode>): Likewise.
4518 (neon_vld3qb<mode>): Likewise.
4519 (neon_vld3_lane<mode>): Likewise.
4520 (neon_vld3_lane<mode>): Likewise.
4521 (neon_vld3_dup<mode>): Likewise.
4522 (neon_vst3<mode>): Likewise.
4523 (neon_vst3qa<mode>): Likewise.
4524 (neon_vst3qb<mode>): Likewise.
4525 (neon_vst3_lane<mode>): Likewise.
4526 (neon_vst3_lane<mode>): Likewise.
4527 (neon_vld4<mode>): Likewise.
4528 (neon_vld4qa<mode>): Likewise.
4529 (neon_vld4qb<mode>): Likewise.
4530 (neon_vld4_lane<mode>): Likewise.
4531 (neon_vld4_lane<mode>): Likewise.
4532 (neon_vld4_dup<mode>): Likewise.
4533 (neon_vst4<mode>): Likewise.
4534 (neon_vst4qa<mode>): Likewise.
4535 (neon_vst4qb<mode>): Likewise.
4536 (neon_vst4_lane<mode>): Likewise.
4537 (neon_vst4_lane<mode>): Likewise.
4538 (neon_vec_unpack<US>_lo_<mode>): Likewise.
4539 (neon_vec_unpack<US>_hi_<mode>): Likewise.
4540 (neon_vec_<US>mult_lo_<mode>): Likewise.
4541 (neon_vec_<US>mult_hi_<mode>): Likewise.
4542 (neon_vec_<US>shiftl_<mode>): Likewise.
4543 (neon_unpack<US>_<mode>): Likewise.
4544 (neon_vec_<US>mult_<mode>): Likewise.
4545 (vec_pack_trunc_<mode>): Likewise.
4546 (neon_vec_pack_trunc_<mode>): Likewise.
4547 (neon_vabd<mode>_2): Likewise.
4548 (neon_vabd<mode>_3): Likewise.
4550 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
4552 * config/aarch64/aarch64.md (movtf_aarch64): Update type attribute.
4553 (load_pair): Update type attribute.
4554 (store_pair): Update type attribute.
4555 * config/aarch64/iterators.md (q): New.
4557 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
4559 * config/arm/types.md: Add new types for Neon insns.
4561 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
4562 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4563 Sergey Lega <sergey.s.lega@intel.com>
4564 Anna Tikhonova <anna.tikhonova@intel.com>
4565 Ilya Tocar <ilya.tocar@intel.com>
4566 Andrey Turetskiy <andrey.turetskiy@intel.com>
4567 Ilya Verbin <ilya.verbin@intel.com>
4568 Kirill Yukhin <kirill.yukhin@intel.com>
4569 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4571 * config/i386/sse.md (unspec): Add UNSPEC_RCP14, UNSPEC_RSQRT14,
4572 UNSPEC_FIXUPIMM, UNSPEC_SCALEF, UNSPEC_GETEXP, UNSPEC_GETMANT,
4573 UNSPEC_EXP2, UNSPEC_RCP28, UNSPEC_RSQRT28.
4575 (srcp14<mode>): Ditto.
4576 (rsqrt14<mode>): Ditto.
4577 (rsqrt14<mode>): Ditto.
4578 (avx512f_vmscalef<mode>): Ditto.
4579 (avx512f_scalef<mode>): Ditto.
4580 (avx512f_getexp<mode>): Ditto.
4581 (avx512f_sgetexp<mode>): Ditto.
4582 (avx512f_fixupimm<mode>): Ditto.
4583 (avx512f_sfixupimm<mode>): Ditto.
4584 (avx512f_rndscale<mode>): Ditto.
4585 (*avx512er_exp2<mode>): Ditto.
4586 (*avx512er_rcp28<mode>): Ditto.
4587 (avx512er_rsqrt28<mode>): Ditto.
4588 (avx512f_getmant<mode>): Ditto.
4589 (avx512f_getmant<mode>): Ditto.
4590 (avx512f_rndscale<mode>): Fix formatting.
4592 2013-10-15 Martin Jambor <mjambor@suse.cz>
4594 * ipa-utils.h (ipa_edge_within_scc): Declare.
4595 * ipa-cp.c (edge_within_scc): Moved...
4596 * ipa-utils.c (ipa_edge_within_scc): ...here. Updated all callers.
4598 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
4599 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4600 Sergey Lega <sergey.s.lega@intel.com>
4601 Anna Tikhonova <anna.tikhonova@intel.com>
4602 Ilya Tocar <ilya.tocar@intel.com>
4603 Andrey Turetskiy <andrey.turetskiy@intel.com>
4604 Ilya Verbin <ilya.verbin@intel.com>
4605 Kirill Yukhin <kirill.yukhin@intel.com>
4606 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4608 * config/i386/predicates.md (const_8_to_15_operand): New.
4609 (const_16_to_31_operand): Ditto.
4610 * config/i386/sse.md (V8FI): New.
4612 (reduc_splus_v8df): Ditto.
4613 (reduc_splus_v16sf): Ditto.
4614 (avx512f_vextract<shuffletype>32x4_1): Ditto.
4615 (vec_extract_hi_<mode>): Ditto.
4616 (avx512f_vinsert<shuffletype>32x4_1): Ditto.
4617 (vec_set_lo_<mode>): Ditto.
4618 (vec_set_hi_<mode>): Ditto.
4619 (avx512f_shuf_<shuffletype>64x2_1): Ditto.
4620 (avx512f_shuf_<shuffletype>32x4_1): Ditto.
4621 (avx512f_pshufd_1): Ditto.
4622 (avx512f_broadcast<mode>): Ditto.
4623 (avx512f_broadcast<mode>): Ditto.
4624 (define_split): Split vec_extract_lo into move.
4625 (ssequartermode): Ditto.
4626 (ssedoublemode): Extened with wider modes.
4627 (vec_extract_lo_<mode>): Ditto.
4629 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
4630 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4631 Sergey Lega <sergey.s.lega@intel.com>
4632 Anna Tikhonova <anna.tikhonova@intel.com>
4633 Ilya Tocar <ilya.tocar@intel.com>
4634 Andrey Turetskiy <andrey.turetskiy@intel.com>
4635 Ilya Verbin <ilya.verbin@intel.com>
4636 Kirill Yukhin <kirill.yukhin@intel.com>
4637 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4639 * config/i386/predicates.md (register_or_constm1_operand): New.
4640 * config/i386/sse.md (unspec): Add UNSPEC_UNSIGNED_PCMP, UNSPEC_TESTM,
4641 UNSPEC_TESTNM, UNSPEC_VTERNLOG, UNSPEC_ALIGN, UNSPEC_CONFLICT,
4642 UNSPEC_MASKED_EQ, UNSPEC_MASKED_GT, UNSPEC_GATHER_PREFETCH,
4643 UNSPEC_SCATTER_PREFETCH
4645 (avx512f_ucmp<mode>3): Ditto.
4646 (avx512f_vternlog<mode>): Ditto.
4647 (avx512f_align<mode>): Ditto.
4648 (<shift_insn><mode>3): Ditto.
4649 (avx512f_<rotate>v<mode>): Ditto.
4650 (avx512f_<rotate><mode>): Ditto.
4651 (avx512f_eq<mode>3): Ditto.
4652 (avx512f_eq<mode>3_1): Ditto.
4653 (avx512f_gt<mode>3): Ditto.
4654 (avx512f_testm<mode>3): Ditto.
4655 (avx512f_testnm<mode>3): Ditto.
4656 (avx512pf_gatherpf<mode>): Ditto.
4657 (*avx512pf_gatherpf<mode>_mask): Ditto.
4658 (*avx512pf_gatherpf<mode>): Ditto.
4659 (avx512pf_scatterpf<mode>): Ditto.
4660 (*avx512pf_scatterpf<mode>_mask): Ditto.
4661 (*avx512pf_scatterpf<mode>): Ditto.
4662 (avx512f_vec_dup_gpr<mode>): Ditto.
4663 (clz<mode>2): Ditto.
4664 (conflict<mode>): Ditto.
4665 (REDUC_SMINMAX_MODE): Extened with wider modes.
4666 (reduc_<code>_<mode>): Ditto.
4667 (vlshr<mode>3): Ditto.
4668 (vashl<mode>3): Ditto.
4670 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
4671 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4672 Sergey Lega <sergey.s.lega@intel.com>
4673 Anna Tikhonova <anna.tikhonova@intel.com>
4674 Ilya Tocar <ilya.tocar@intel.com>
4675 Andrey Turetskiy <andrey.turetskiy@intel.com>
4676 Ilya Verbin <ilya.verbin@intel.com>
4677 Kirill Yukhin <kirill.yukhin@intel.com>
4678 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4680 * config/i386/sse.md (unspec): Added UNSPEC_VPERMI2, UNSPEC_VPERMT2,
4683 (avx512fmaskmode): Ditto.
4684 (bcstscalarsuff): Ditto.
4685 (avx512f_blendm<mode>): Ditto.
4686 (cmp_imm_predicate): Ditto.
4687 (avx512f_cmp<mode>3): Ditto.
4688 (avx512f_vec_dup<mode>): Ditto.
4689 (avx512f_vec_dup_mem<mode>): Ditto.
4690 (avx512f_vpermi2var<mode>3): Ditto.
4691 (avx512f_vpermt2var<mode>3): Ditto.
4692 (vec_init<mode>): Ditto.
4693 (avx512f_gathersi<mode>): Ditto.
4694 (*avx512f_gathersi<mode>): Ditto.
4695 (*avx512f_gathersi<mode>_2): Ditto.
4696 (avx512f_gatherdi<mode>): Ditto.
4697 (*avx512f_gatherdi<mode>): Ditto.
4698 (*avx512f_gatherdi<mode>_2): Ditto.
4699 (avx512f_scattersi<mode>): Ditto.
4700 (*avx512f_scattersi<mode>): Ditto.
4701 (avx512f_scatterdi<mode>): Ditto.
4702 (*avx512f_scatterdi<mode>): Ditto.
4703 (sseintprefix): Extened with wider modes.
4704 (VEC_GATHER_IDXSI): Ditto.
4705 (VEC_GATHER_IDXDI): Ditto.
4706 (VEC_GATHER_SRCDI): Ditto.
4708 2013-10-15 Matthew Gretton-Dann <matthew.gretton-dann@arm.com>
4709 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4711 * config/arm/t-aprofile: New file.
4712 * config.gcc: Handle --with-multilib-list option.
4714 2013-10-15 Bernd Schmidt <bernds@codesourcery.com>
4716 * reload1.c (reloads_unique_chain_p): Ensure that r1 is
4719 2013-10-15 Richard Biener <rguenther@suse.de>
4721 * tree-loop-distribution.c (build_empty_rdg): Inline into
4723 (rdg_flag_vertex): Inline into single user.
4724 (rdg_flag_vertex_and_dependent): Likewise.
4725 (build_rdg_partition_for_vertex): Remove processed bitmap.
4726 (rdg_build_partitions): Simplify.
4728 2013-10-15 Richard Biener <rguenther@suse.de>
4730 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1):
4731 Restructure forwarding through conversions and copies to
4732 avoid performing copy-propagation the wrong way. Adjust
4733 recursion invocations.
4734 (forward_propagate_addr_expr): Add argument stating if we
4735 are recursing from a single-use.
4736 (ssa_forward_propagate_and_combine): Adjust.
4738 2013-10-14 David Malcolm <dmalcolm@redhat.com>
4740 * dumpfile.h (gcc::dump_manager): New class, to hold state
4741 relating to dumpfile management.
4742 (get_dump_file_name): Remove in favor of method of dump_manager.
4743 (dump_initialized_p): Likewise.
4744 (dump_start): Likewise.
4745 (dump_finish): Likewise.
4746 (dump_switch_p): Likewise.
4747 (dump_register): Likewise.
4748 (get_dump_file_info): Likewise.
4749 * context.c (gcc::context::context): Construct the dump_manager
4751 * context.h (gcc::context::get_dumps): New.
4752 (gcc::context::m_dumps): New.
4753 * coverage.c (coverage_init): Port to dump_manager API.
4754 * dumpfile.c (extra_dump_files): Convert to field of gcc::dump_manager.
4755 (extra_dump_files_in_use): Likewise.
4756 (extra_dump_files_alloced): Likewise.
4757 (gcc::dump_manager::dump_manager): New.
4758 (dump_register): Convert to...
4759 (gcc::dump_manager::dump_register): ...method, replacing
4760 function-static next_dump with m_next_dump field.
4761 (get_dump_file_info): Convert to...
4762 (gcc::dump_manager::get_dump_file_info): ...method.
4763 (get_dump_file_name): Convert to...
4764 (gcc::dump_manager::get_dump_file_name): ...method.
4765 (dump_start): Convert to...
4766 (gcc::dump_manager::dump_start): ...method.
4767 (dump_finish): Convert to...
4768 (gcc::dump_manager::dump_finish): ...method.
4769 (dump_begin): Replace body with...
4770 (gcc::dump_manager::dump_begin): ...new method.
4771 (dump_phase_enabled_p): Convert to...
4772 (gcc::dump_manager::dump_phase_enabled_p): ...method.
4773 (dump_phase_enabled_p): Convert to...
4774 (gcc::dump_manager::dump_phase_enabled_p): ...method.
4775 (dump_initialized_p): Convert to...
4776 (gcc::dump_manager::dump_initialized_p): ...method.
4777 (dump_flag_name): Replace body with...
4778 (gcc::dump_manager::dump_flag_name): ...new method.
4779 (dump_enable_all): Convert to...
4780 (gcc::dump_manager::dump_enable_all): ...new method.
4781 (opt_info_enable_passes): Convert to...
4782 (gcc::dump_manager::opt_info_enable_passes): ...new method.
4783 (dump_switch_p_1): Convert to...
4784 (gcc::dump_manager::dump_switch_p_1): ...new method.
4785 (dump_switch_p): Convert to...
4786 (gcc::dump_manager::dump_switch_p): ...new method.
4787 (opt_info_switch_p): Port to dump_manager API.
4788 (enable_rtl_dump_file): Likewise.
4789 * opts-global.c (handle_common_deferred_options): Port to new
4791 * passes.c (pass_manager::finish_optimization_passes): Likewise.
4792 (pass_manager::register_one_dump_file): Likewise.
4793 (pass_manager::register_pass): Likewise.
4794 (pass_init_dump_file): Likewise.
4795 (pass_fini_dump_file): Likewise.
4796 * statistics.c (statistics_early_init): Likewise.
4798 2013-10-14 Richard Biener <rguenther@suse.de>
4800 * gimple.c (gimple_canonical_types, canonical_type_hash_cache,
4801 iterative_hash_canonical_type, gimple_canonical_type_hash,
4802 gimple_canonical_types_compatible_p, gimple_canonical_type_eq,
4803 gimple_register_canonical_type, print_gimple_types_stats,
4804 free_gimple_type_tables): Move to lto/lto.c
4805 (gt-gimple.h): Do not include.
4806 * gimple.h (gimple_register_canonical_type,
4807 print_gimple_types_stats, free_gimple_type_tables): Remove.
4808 * Makefile.in (GTFILES): Remove gimple.c.
4810 2013-10-14 Travis Snoozy <quandary@remstate.com>
4813 * config/msp430/msp430.c (msp430_option_override): Correct thinko
4814 scanning for msp430x targets.
4816 2013-10-14 Eric Botcazou <ebotcazou@adacore.com>
4819 * config/sparc/sparc-protos.h (widen_mem_for_ldd_peep): Declare.
4820 (registers_ok_for_ldd_peep): Move around.
4821 * config/sparc/sparc.c (widen_mem_for_ldd_peep): New.
4822 * config/sparc/sparc.md (widening peepholes): Use it.
4824 2013-10-14 Richard Biener <rguenther@suse.de>
4828 * gimple.c (iterative_hash_canonical_type): Make sure to
4829 record the hash into the correct hashtable slot.
4831 2013-10-13 Eric Botcazou <ebotcazou@adacore.com>
4833 PR rtl-optimization/58662
4834 * combine.c (try_combine): Take into account death nodes on I2 when
4835 splitting a PARALLEL of two independent SETs. Fix dump message.
4837 2013-10-12 Oleg Endo <olegendo@gcc.gnu.org>
4840 * config/sh/sh_treg_combine.cc: New SH specific RTL pass.
4841 * config.gcc (SH extra_objs): Add sh_ifcvt.o.
4842 * config/sh/t-sh (sh_treg_combine.o): New entry.
4843 * config/sh/sh.c (sh_fixed_condition_code_regs): New function that
4844 implements the target hook TARGET_FIXED_CONDITION_CODE_REGS.
4845 (register_sh_passes): New function. Register sh_treg_combine pass.
4846 (sh_option_override): Invoke it.
4847 (sh_canonicalize_comparison): Handle op0_preserve_value.
4848 * sh.md (*cbranch_t"): Do not try to optimize missed test and branch
4849 opportunities. Canonicalize branch condition.
4850 (nott): Allow only if pseudos can be created for non-SH2A.
4852 2013-10-12 H.J. Lu <hongjiu.lu@intel.com>
4855 * config/i386/i386.c (ix86_copy_addr_to_reg): New function.
4856 (ix86_expand_movmem): Replace copy_addr_to_reg with
4857 ix86_copy_addr_to_reg.
4858 (ix86_expand_setmem): Likewise.
4860 2013-10-12 Alexander Monakov <amonakov@ispras.ru>
4862 * config/i386/i386.c (ix86_expand_sse_compare_and_jump): Use mode
4863 provided by ix86_fp_compare_mode instead of CCFPUmode.
4865 2013-10-12 James Greenhalgh <james.greenhalgh@arm.com>
4867 * config/aarch64/arm_neon.h
4868 (vtbx<1,3>_<psu>8): Fix register constriants.
4870 2013-10-11 Jeff Law <law@redhat.com>
4872 PR tree-optimization/58640
4873 * tree-ssa-threadupdate.c (mark_threaded_blocks): Truncate jump
4874 threading paths that cross over two loop entry points.
4876 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4878 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): Generalize to
4879 handle vector float as well.
4880 (*vsx_le_perm_load_v4si): Likewise.
4881 (*vsx_le_perm_store_v2di): Likewise.
4882 (*vsx_le_perm_store_v4si): Likewise.
4884 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4886 * config/rs6000/vector.md (vec_realign_load<mode>): Generate vperm
4887 directly to circumvent subtract from splat{31} workaround.
4888 * config/rs6000/rs6000-protos.h (altivec_expand_vec_perm_le): New
4890 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): New.
4891 * config/rs6000/altivec.md (define_c_enum "unspec"): Add
4892 UNSPEC_VPERM_X and UNSPEC_VPERM_UNS_X.
4893 (altivec_vperm_<mode>): Convert to define_insn_and_split to
4894 separate big and little endian logic.
4895 (*altivec_vperm_<mode>_internal): New define_insn.
4896 (altivec_vperm_<mode>_uns): Convert to define_insn_and_split to
4897 separate big and little endian logic.
4898 (*altivec_vperm_<mode>_uns_internal): New define_insn.
4899 (vec_permv16qi): Add little endian logic.
4901 2013-10-11 Marc Glisse <marc.glisse@inria.fr>
4903 * doc/extend.texi (returns_nonnull): Remove arguments.
4905 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4906 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4907 Sergey Lega <sergey.s.lega@intel.com>
4908 Anna Tikhonova <anna.tikhonova@intel.com>
4909 Ilya Tocar <ilya.tocar@intel.com>
4910 Andrey Turetskiy <andrey.turetskiy@intel.com>
4911 Ilya Verbin <ilya.verbin@intel.com>
4912 Kirill Yukhin <kirill.yukhin@intel.com>
4913 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4915 * config/i386/sse.md (VI48F_256_512): New.
4916 (avx2_permvar<mode>): Change to ...
4917 (<avx2_avx512f>_permvar<mode>): This.
4919 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4920 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4921 Sergey Lega <sergey.s.lega@intel.com>
4922 Anna Tikhonova <anna.tikhonova@intel.com>
4923 Ilya Tocar <ilya.tocar@intel.com>
4924 Andrey Turetskiy <andrey.turetskiy@intel.com>
4925 Ilya Verbin <ilya.verbin@intel.com>
4926 Kirill Yukhin <kirill.yukhin@intel.com>
4927 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4929 * config/i386/i386.c (bdesc_args): Change corresponding pattern for
4930 __builtin_ia32_cvtps2dq, __builtin_ia32_cvtps2dq256.
4931 * config/i386/sse.md (VI4_AVX): New.
4932 (sf2simodelower): Ditto.
4933 (sse2_cvtps2dq): Change to ...
4934 (<sse2_avx_avx512f>_fix_notrunc<sf2simodelower><mode>): This.
4936 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4937 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4938 Sergey Lega <sergey.s.lega@intel.com>
4939 Anna Tikhonova <anna.tikhonova@intel.com>
4940 Ilya Tocar <ilya.tocar@intel.com>
4941 Andrey Turetskiy <andrey.turetskiy@intel.com>
4942 Ilya Verbin <ilya.verbin@intel.com>
4943 Kirill Yukhin <kirill.yukhin@intel.com>
4944 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4946 * config/i386/sse.md (V_512): New.
4948 (vcond<V_512:mode><VF_512:mode>): Ditto.
4949 (vcond<V_512:mode><VI_512:mode>): Ditto.
4950 (vcondu<V_512:mode><VI_512:mode>): Ditto.
4952 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4953 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4954 Sergey Lega <sergey.s.lega@intel.com>
4955 Anna Tikhonova <anna.tikhonova@intel.com>
4956 Ilya Tocar <ilya.tocar@intel.com>
4957 Andrey Turetskiy <andrey.turetskiy@intel.com>
4958 Ilya Verbin <ilya.verbin@intel.com>
4959 Kirill Yukhin <kirill.yukhin@intel.com>
4960 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4962 * config/i386/i386.c (ix86_rtx_costs): Enable fma for TARGET_AVX512F.
4963 * config/i386/sse.md (FMAMODEM): Changed modes and conditions.
4965 (fma<mode>4): Removed condition.
4966 (fms<mode>4): Ditto.
4967 (fnma<mode>4): Ditto.
4968 (fnms<mode>4): Ditto.
4969 (fma4i_fmadd_<mode>): Ditto.
4970 (*fma_fmadd_<mode>): Ditto.
4971 (*fma_fmsub_<mode>): Ditto.
4972 (*fma_fnmadd_<mode>): Ditto.
4973 (*fma_fnmsub_<mode>): Ditto.
4974 (fmaddsub_<mode>): Allow for TARGET_AVX512F.
4975 (*fma_fmaddsub_<mode>): Ditto.
4976 (*fma_fmsubadd_<mode>): Ditto.
4977 (*fmai_fmadd_<mode>): Ditto.
4978 (*fmai_fmsub_<mode>): Ditto.
4979 (*fmai_fnmadd_<mode>): Ditto.
4980 (*fmai_fnmsub_<mode>): Ditto.
4982 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4983 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4984 Sergey Lega <sergey.s.lega@intel.com>
4985 Anna Tikhonova <anna.tikhonova@intel.com>
4986 Ilya Tocar <ilya.tocar@intel.com>
4987 Andrey Turetskiy <andrey.turetskiy@intel.com>
4988 Ilya Verbin <ilya.verbin@intel.com>
4989 Kirill Yukhin <kirill.yukhin@intel.com>
4990 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4992 * config/i386/sse.md (VI248_AVX2_8_AVX512F): New.
4993 (VI124_256): Changed to ...
4994 (VI124_256_48_512): This.
4995 (ssepackmode): Extended with wider modes.
4996 (<code><mode>3): Changed iterator.
4997 (*avx2_<code><mode>3): Ditto.
4998 (vec_pack_trunc_<mode>): Ditto.
5000 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5001 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5002 Sergey Lega <sergey.s.lega@intel.com>
5003 Anna Tikhonova <anna.tikhonova@intel.com>
5004 Ilya Tocar <ilya.tocar@intel.com>
5005 Andrey Turetskiy <andrey.turetskiy@intel.com>
5006 Ilya Verbin <ilya.verbin@intel.com>
5007 Kirill Yukhin <kirill.yukhin@intel.com>
5008 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5010 * config/i386/sse.md (VI124_AVX2_48_AVX512F): New.
5011 (VI8F_256_512): Ditto.
5012 (abs<mode>2): Changed iterator.
5013 (avx2_perm<mode>): Changed to ...
5014 (<avx2_avx512f>_perm<mode>): This.
5015 (avx2_perm<mode>_1): Changed to ...
5016 (<avx2_avx512f>_perm<mode>_1): This.
5018 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5019 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5020 Sergey Lega <sergey.s.lega@intel.com>
5021 Anna Tikhonova <anna.tikhonova@intel.com>
5022 Ilya Tocar <ilya.tocar@intel.com>
5023 Andrey Turetskiy <andrey.turetskiy@intel.com>
5024 Ilya Verbin <ilya.verbin@intel.com>
5025 Kirill Yukhin <kirill.yukhin@intel.com>
5026 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5028 * config/i386/sse.md (VI48_AVX512F): New.
5029 (VI48_AVX2): Changed to ...
5030 (VI48_AVX2_48_AVX512F): This.
5031 (avx2_ashrv<mode>): Changed to ...
5032 (<avx2_avx512f>_ashrv<mode>): This.
5033 (avx2_<shift_insn>v<mode>): Changed to ...
5034 (<avx2_avx512f>_<shift_insn>v<mode>): This.
5036 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5037 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5038 Sergey Lega <sergey.s.lega@intel.com>
5039 Anna Tikhonova <anna.tikhonova@intel.com>
5040 Ilya Tocar <ilya.tocar@intel.com>
5041 Andrey Turetskiy <andrey.turetskiy@intel.com>
5042 Ilya Verbin <ilya.verbin@intel.com>
5043 Kirill Yukhin <kirill.yukhin@intel.com>
5044 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5046 * config/i386/sse.md (VI4_AVX512F): New.
5047 (VI8_AVX2_AVX512F): Ditto.
5048 (mul<mode>3): Extended with wider modes.
5049 (*<sse4_1_avx2>_mul<mode>3): Ditto.
5050 (mul<mode>3): Ditto.
5051 (vec_widen_<s>mult_odd_<mode>): Ditto.
5053 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5054 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5055 Sergey Lega <sergey.s.lega@intel.com>
5056 Anna Tikhonova <anna.tikhonova@intel.com>
5057 Ilya Tocar <ilya.tocar@intel.com>
5058 Andrey Turetskiy <andrey.turetskiy@intel.com>
5059 Ilya Verbin <ilya.verbin@intel.com>
5060 Kirill Yukhin <kirill.yukhin@intel.com>
5061 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5063 * config/i386/sse.md (VI2_AVX512F): New.
5064 (VI124_AVX512F): Ditto.
5065 (sseunpackmode): Extended with wider modes.
5066 (sseunpackfltmode): Ditto.
5067 (vec_unpacks_float_hi_<mode>): Ditto.
5068 (vec_unpacks_float_lo_<mode>): Ditto.
5069 (vec_unpacku_float_hi_<mode>): Ditto.
5070 (vec_unpacku_float_lo_<mode>): Ditto.
5071 (vec_unpacks_lo_<mode>): Ditto.
5072 (vec_unpacks_hi_<mode>): Ditto.
5073 (vec_unpacku_lo_<mode>): Ditto.
5074 (vec_unpacku_hi_<mode>): Ditto.
5076 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5077 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5078 Sergey Lega <sergey.s.lega@intel.com>
5079 Anna Tikhonova <anna.tikhonova@intel.com>
5080 Ilya Tocar <ilya.tocar@intel.com>
5081 Andrey Turetskiy <andrey.turetskiy@intel.com>
5082 Ilya Verbin <ilya.verbin@intel.com>
5083 Kirill Yukhin <kirill.yukhin@intel.com>
5084 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5086 * config/i386/i386.md (multdiv): New.
5087 (multdiv_mnemonic): Ditto.
5088 * config/i386/sse.md (<sse>_vmmul<mode>3): Changed to...
5089 (<sse>_vm<multdiv_mnemonic><mode>3): This.
5090 (<sse>_vmdiv<mode>3): Removed.
5092 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5093 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5094 Sergey Lega <sergey.s.lega@intel.com>
5095 Anna Tikhonova <anna.tikhonova@intel.com>
5096 Ilya Tocar <ilya.tocar@intel.com>
5097 Andrey Turetskiy <andrey.turetskiy@intel.com>
5098 Ilya Verbin <ilya.verbin@intel.com>
5099 Kirill Yukhin <kirill.yukhin@intel.com>
5100 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5102 * config/i386/sse.md (V): Extended with wider modes.
5104 (ssehalfvecmode): Ditto.
5106 (ssepackfltmode): Ditto.
5107 (avx_vec_concat<mode>): Ditto.
5108 (V_256_512): New iterator.
5109 (VF2_512_256): Ditto.
5110 (si2dfmode): New attribute.
5111 (si2dfmodelower): Ditto.
5113 (concat_tg_mode): Ditto.
5114 (floatv4siv4df2): Changed to ...
5115 (float<si2dfmodelower><mode>2): This.
5116 (avx_cvtps2pd256): Changed to ...
5117 (<sse2_avx_avx512f>_cvtps2pd<avxsizesuffix>): This.
5118 (vec_pack_trunc_v4df): Changed to ...
5119 (vec_pack_trunc_<mode>): This.
5120 (avx_vpermil<mode>): Changed to ...
5121 (<sse2_avx_avx512f>_vpermil<mode>): This.
5122 (<fixsuffix>fix_truncv8dfv8si2): New.
5123 (vec_pack_sfix_trunc_v8df): Ditto.
5124 (avx512f_rndscale<mode>): Ditto.
5125 (avx512f_roundpd512): Ditto.
5126 (vec_pack_ufix_trunc_<mode>): Updated iterator.
5128 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5129 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5130 Sergey Lega <sergey.s.lega@intel.com>
5131 Anna Tikhonova <anna.tikhonova@intel.com>
5132 Ilya Tocar <ilya.tocar@intel.com>
5133 Andrey Turetskiy <andrey.turetskiy@intel.com>
5134 Ilya Verbin <ilya.verbin@intel.com>
5135 Kirill Yukhin <kirill.yukhin@intel.com>
5136 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5138 * config/i386/i386.md (any_fix): New iterator.
5139 (fixsuffix): New attribute.
5140 * config/i386/sse.md (VF1): Extened with wider modes.
5144 (sseintvecmodelower): Ditto.
5145 (ssescalarmode): Ditto.
5146 (ssescalarnum): Ditto.
5148 (ssexmmmode): Ditto.
5149 (<fixsuffix>fix_truncv16sfv16si2): Ditto.
5150 (<sse>_rcp<mode>2): Change iterator.
5151 (rsqrt<mode>2): Ditto.
5152 (<sse>_rsqrt<mode>2): Ditto.
5153 (avx2_vec_dup<mode>): Ditto.
5154 (<sse4_1>_round<ssemodesuffix>_sfix<avxsizesuffix>): Ditto.
5155 (round<mode>2_sfix): Ditto.
5156 (avx2_pbroadcast<mode>): Ditto.
5157 (*andnot<mode>3): Handle XI mode.
5158 (*<code><mode>3): Ditto.
5159 (AVXTOSSEMODE): Removed.
5160 (avx_vpermil<mode>): Changed to ...
5161 (<sse2_avx_avx512f>_vpermil<mode>): This.
5163 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5164 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5165 Sergey Lega <sergey.s.lega@intel.com>
5166 Anna Tikhonova <anna.tikhonova@intel.com>
5167 Ilya Tocar <ilya.tocar@intel.com>
5168 Andrey Turetskiy <andrey.turetskiy@intel.com>
5169 Ilya Verbin <ilya.verbin@intel.com>
5170 Kirill Yukhin <kirill.yukhin@intel.com>
5171 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5173 * config/i386/sse.md (<sse2>_movnt<mode>): Update constraint to "v".
5174 (<sse>_comi): Ditto.
5175 (<sse>_ucomi): Ditto.
5176 (sse_cvtss2siq_2): Ditto.
5177 (sse2_cvtsd2si): Ditto.
5178 (sse2_cvtsd2siq): Ditto.
5179 (sse2_cvttsd2si): Ditto.
5180 (sse2_cvttsd2siq): Ditto.
5181 (<shift_insn><mode>3): Ditto.
5182 (sse2_cvtsi2sdq): Update constraint and prefix.
5183 (sse_cvtsi2ss): Update prefix.
5184 (sse_cvtsi2ssq): Ditto.
5186 2013-10-11 Jakub Jelinek <jakub@redhat.com>
5188 * tree-vrp.c (infer_nonnull_range): Use is_gimple_call,
5189 ignore internal calls.
5191 2013-10-11 Richard Biener <rguenther@suse.de>
5193 * tree-pretty-print.c (dump_generic_node): Allow to dump both (D)
5194 and (ab) for SSA_NAMEs. Mark INTEGER_CSTs with (OVF) if
5195 TREE_OVERFLOW is set.
5197 2013-10-11 Thomas Schwinge <thomas@codesourcery.com>
5199 * tree.h (OMP_CLAUSE_CODE): Remove duplicate definition.
5201 * gimple.c: GIMPLE statements have subcodes, not sub-codes.
5202 * gimple.h: Likewise.
5204 * doc/generic.texi (OpenMP): OMP_CLAUSE_* are subcodes, not sub-codes.
5206 * doc/generic.texi (Adding new DECL node types): Explain *_CHECK
5209 * doc/gimple.texi (is_gimple_omp): Move into the correct section.
5211 * acinclude.m4 (gcc_GAS_FLAGS): Add more gcc_cv_as_flags overrides.
5212 * configure: Regenerate.
5214 2013-10-11 Jakub Jelinek <jakub@redhat.com>
5216 * tree-pretty-print.c (dump_omp_clause): Handle OMP_CLAUSE__LOOPTEMP_
5217 and new OpenMP 4.0 clauses, handle UDR OMP_CLAUSE_REDUCTION,
5218 formatting fixes, use pp_colon instead of pp_character (..., ':'),
5219 similarly pp_right_paren.
5220 (dump_generic_node): Handle OMP_DISTRIBUTE, OMP_TEAMS,
5221 OMP_TARGET_DATA, OMP_TARGET, OMP_TARGET_UPDATE, OMP_TASKGROUP,
5222 allow OMP_FOR_INIT to be NULL, handle OMP_ATOMIC_SEQ_CST.
5223 * tree.c (omp_clause_num_ops, omp_clause_code_name): Add OpenMP 4.0
5225 (omp_declare_simd_clauses_equal,
5226 omp_remove_redundant_declare_simd_attrs): New functions.
5227 (attribute_value_equal): Use omp_declare_simd_clauses_equal.
5228 (walk_tree_1): Handle new OpenMP 4.0 clauses.
5229 * tree.h (OMP_LOOP_CHECK): Define.
5230 (OMP_FOR_BODY, OMP_FOR_CLAUSES, OMP_FOR_INIT, OMP_FOR_COND,
5231 OMP_FOR_INCR, OMP_FOR_PRE_BODY): Use it.
5232 (OMP_TASKGROUP_BODY, OMP_TEAMS_BODY, OMP_TEAMS_CLAUSES,
5233 OMP_TARGET_DATA_BODY, OMP_TARGET_DATA_CLAUSES, OMP_TARGET_BODY,
5234 OMP_TARGET_CLAUSES, OMP_TARGET_UPDATE_CLAUSES, OMP_CLAUSE_SIZE,
5235 OMP_ATOMIC_SEQ_CST, OMP_CLAUSE_DEPEND_KIND, OMP_CLAUSE_MAP_KIND,
5236 OMP_CLAUSE_MAP_ZERO_BIAS_ARRAY_SECTION, OMP_CLAUSE_PROC_BIND_KIND,
5237 OMP_CLAUSE_REDUCTION_OMP_ORIG_REF, OMP_CLAUSE_ALIGNED_ALIGNMENT,
5238 OMP_CLAUSE_NUM_TEAMS_EXPR, OMP_CLAUSE_THREAD_LIMIT_EXPR,
5239 OMP_CLAUSE_DEVICE_ID, OMP_CLAUSE_DIST_SCHEDULE_CHUNK_EXPR,
5240 OMP_CLAUSE_SIMDLEN_EXPR): Define.
5241 (OMP_CLAUSE_DECL): Change range up to OMP_CLAUSE__LOOPTEMP_.
5242 (omp_remove_redundant_declare_simd_attrs): New prototype.
5243 * gimple.def (GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET,
5244 GIMPLE_OMP_TEAMS): New codes.
5245 (GIMPLE_OMP_RETURN): Use GSS_OMP_ATOMIC_STORE instead of GSS_BASE.
5246 * omp-low.c (struct omp_context): Add cancel_label and cancellable
5248 (target_nesting_level): New variable.
5249 (extract_omp_for_data): Handle GF_OMP_FOR_KIND_DISTRIBUTE and
5250 OMP_CLAUSE_DIST_SCHEDULE. Don't fallback to library implementation
5251 for collapse > 1 static schedule unless ordered.
5252 (get_ws_args_for): Add par_stmt argument. Handle combined loops.
5253 (determine_parallel_type): Adjust get_ws_args_for caller.
5254 (install_var_field): Handle mask & 4 for double indirection.
5255 (scan_sharing_clauses): Ignore shared clause on teams construct.
5256 Handle OMP_CLAUSE__LOOPTEMP_ and new OpenMP 4.0 clauses.
5257 (create_omp_child_function): If inside target or declare target
5258 constructs, set "omp declare target" attribute on the child function.
5259 (find_combined_for): New function.
5260 (scan_omp_parallel): Handle combined loops.
5261 (scan_omp_target, scan_omp_teams): New functions.
5262 (check_omp_nesting_restrictions): Check new OpenMP 4.0 nesting
5263 restrictions and set ctx->cancellable for cancellable constructs.
5264 (scan_omp_1_stmt): Call check_omp_nesting_restrictions also on
5265 selected builtin calls. Handle GIMPLE_OMP_TASKGROUP,
5266 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS.
5267 (build_omp_barrier): Add lhs argument, return gimple rather than tree.
5268 (omp_clause_aligned_alignment): New function.
5269 (lower_rec_simd_input_clauses): Only call SET_DECL_VALUE_EXPR on decls.
5270 (lower_rec_input_clauses): Add FD argument. Ignore shared clauses
5271 on teams constructs. Handle user defined reductions and new
5273 (lower_reduction_clauses): Don't set placeholder to address of ref
5274 if it has already the right type.
5275 (lower_send_clauses): Handle OMP_CLAUSE__LOOPTEMP_.
5276 (expand_parallel_call): Use the new non-_start suffixed builtins,
5277 handle OMP_CLAUSE_PROC_BIND, don't call the outlined function
5278 and GOMP_parallel_end after the call.
5279 (expand_task_call): Handle OMP_CLAUSE_DEPEND.
5280 (expand_omp_for_init_counts): Handle combined loops.
5281 (expand_omp_for_init_vars): Add inner_stmt argument, handle combined
5283 (expand_omp_for_generic): Likewise. Use GOMP_loop_end_cancel at the
5284 end of cancellable loops.
5285 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
5286 Likewise. Handle collapse > 1 loops.
5287 (expand_omp_simd): Handle combined loops.
5288 (expand_omp_for): Add inner_stmt argument, adjust callers of
5289 expand_omp_for* functions, use expand_omp_for_static*chunk even
5290 for collapse > 1 unless ordered.
5291 (expand_omp_sections): Use GOMP_sections_end_cancel at the end
5292 of cancellable sections.
5293 (expand_omp_single): Remove need_barrier variable, just rely on
5294 gimple_omp_return_nowait_p. Adjust build_omp_barrier caller.
5295 (expand_omp_synch): Allow GIMPLE_OMP_TASKGROUP and GIMPLE_OMP_TEAMS.
5296 (expand_omp_atomic_load, expand_omp_atomic_store,
5297 expand_omp_atomic_fetch_op): Handle gimple_omp_atomic_seq_cst_p.
5298 (expand_omp_target): New function.
5299 (expand_omp): Handle combined loops. Handle GIMPLE_OMP_TASKGROUP,
5300 GIMPLE_OMP_TEAMS, GIMPLE_OMP_TARGET.
5301 (build_omp_regions_1): Immediately close region for
5302 GF_OMP_TARGET_KIND_UPDATE.
5303 (maybe_add_implicit_barrier_cancel): New function.
5304 (lower_omp_sections): Adjust lower_rec_input_clauses caller. Handle
5306 (lower_omp_single): Likewise. Add clobber after the barrier.
5307 (lower_omp_taskgroup): New function.
5308 (lower_omp_for): Handle combined loops. Adjust
5309 lower_rec_input_clauses caller. Handle cancellation.
5310 (lower_depend_clauses): New function.
5311 (lower_omp_taskreg): Lower depend clauses. Adjust
5312 lower_rec_input_clauses caller. Add clobber after the call. Handle
5314 (lower_omp_target, lower_omp_teams): New functions.
5315 (lower_omp_1): Handle cancellation. Handle GIMPLE_OMP_TASKGROUP,
5316 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GOMP_barrier, GOMP_cancel
5317 and GOMP_cancellation_point calls.
5318 (lower_omp): Fold stmts inside of target region.
5319 (diagnose_sb_1, diagnose_sb_2): Handle GIMPLE_OMP_TASKGROUP,
5320 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
5321 * builtin-types.def (DEF_FUNCTION_TYPE_8): Document.
5322 (BT_FN_VOID_OMPFN_PTR_UINT,
5323 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG,
5324 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG,
5325 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT): Remove.
5326 (BT_FN_VOID_OMPFN_PTR_UINT_UINT_UINT,
5327 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_UINT,
5328 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG_UINT,
5329 BT_FN_BOOL_INT, BT_FN_BOOL_INT_BOOL, BT_FN_VOID_UINT_UINT,
5330 BT_FN_VOID_INT_PTR_SIZE_PTR_PTR_PTR,
5331 BT_FN_VOID_INT_OMPFN_PTR_SIZE_PTR_PTR_PTR,
5332 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR): New.
5333 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1,
5334 call_may_clobber_ref_p_1): Handle BUILT_IN_GOMP_BARRIER_CANCEL,
5335 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_LOOP_END_CANCEL,
5336 BUILT_IN_GOMP_SECTIONS_END_CANCEL. Don't handle
5337 BUILT_IN_GOMP_PARALLEL_END.
5338 * gimple-low.c (lower_stmt): Handle GIMPLE_OMP_TASKGROUP,
5339 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
5340 * gimple-pretty-print.c (dump_gimple_omp_for): Handle
5341 GF_OMP_FOR_KIND_DISTRIBUTE.
5342 (dump_gimple_omp_target, dump_gimple_omp_teams): New functions.
5343 (dump_gimple_omp_block): Handle GIMPLE_OMP_TASKGROUP.
5344 (dump_gimple_omp_return): Print lhs if it has any.
5345 (dump_gimple_omp_atomic_load, dump_gimple_omp_atomic_store): Handle
5346 gimple_omp_atomic_seq_cst_p.
5347 (pp_gimple_stmt_1): Handle GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET
5348 and GIMPLE_OMP_TEAMS.
5349 * langhooks.c (lhd_omp_mappable_type): New function.
5350 * tree-vectorizer.c (struct simd_array_to_simduid): Fix up comment.
5351 * langhooks.h (struct lang_hooks_for_types): Add omp_mappable_type
5353 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_MAP,
5354 GOVD_ALIGNED and GOVD_MAP_TO_ONLY.
5355 (enum omp_region_type): Add ORT_TEAMS, ORT_TARGET_DATA and ORT_TARGET.
5356 (struct gimplify_omp_ctx): Add combined_loop field.
5357 (gimplify_call_expr, gimplify_modify_expr): Don't call fold_stmt
5358 on stmts inside of target region.
5359 (is_gimple_stmt): Return true for OMP_DISTRIBUTE and OMP_TASKGROUP.
5360 (omp_firstprivatize_variable): Handle GOVD_MAP, GOVD_ALIGNED,
5361 ORT_TARGET and ORT_TARGET_DATA.
5362 (omp_add_variable): Avoid checks on readding var for GOVD_ALIGNED.
5364 (omp_notice_threadprivate_variable): Complain about threadprivate
5365 variables in target region.
5366 (omp_notice_variable): Complain about vars with non-mappable type
5367 in target region. Handle ORT_TEAMS, ORT_TARGET and ORT_TARGET_DATA.
5368 (omp_check_private): Ignore ORT_TARGET* regions.
5369 (gimplify_scan_omp_clauses, gimplify_adjust_omp_clauses_1,
5370 gimplify_adjust_omp_clauses): Handle new OpenMP 4.0 clauses.
5371 (find_combined_omp_for): New function.
5372 (gimplify_omp_for): Handle gimplification of combined loops.
5373 (gimplify_omp_workshare): Gimplify also OMP_TARGET, OMP_TARGET_DATA,
5375 (gimplify_omp_target_update): New function.
5376 (gimplify_omp_atomic): Handle OMP_ATOMIC_SEQ_CST.
5377 (gimplify_expr): Handle OMP_DISTRIBUTE, OMP_TARGET, OMP_TARGET_DATA,
5378 OMP_TARGET_UPDATE, OMP_TEAMS, OMP_TASKGROUP.
5379 (gimplify_body): If fndecl has "omp declare target" attribute, add
5380 implicit ORT_TARGET context around it.
5381 * tree.def (OMP_DISTRIBUTE, OMP_TEAMS, OMP_TARGET_DATA, OMP_TARGET,
5382 OMP_TASKGROUP, OMP_TARGET_UPDATE): New tree codes.
5383 * tree-nested.c (convert_nonlocal_reference_stmt,
5384 convert_local_reference_stmt, convert_gimple_call): Handle
5385 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
5386 * omp-builtins.def (BUILT_IN_GOMP_TASK): Use
5387 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR
5388 instead of BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT.
5389 (BUILT_IN_GOMP_TARGET, BUILT_IN_GOMP_TARGET_DATA,
5390 BUILT_IN_GOMP_TARGET_END_DATA, BUILT_IN_GOMP_TARGET_UPDATE,
5391 BUILT_IN_GOMP_TEAMS, BUILT_IN_BARRIER_CANCEL,
5392 BUILT_IN_GOMP_LOOP_END_CANCEL,
5393 BUILT_IN_GOMP_SECTIONS_END_CANCEL, BUILT_IN_OMP_GET_TEAM_NUM,
5394 BUILT_IN_OMP_GET_NUM_TEAMS, BUILT_IN_GOMP_TASKGROUP_START,
5395 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_PARALLEL_LOOP_STATIC,
5396 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC,
5397 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED,
5398 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME, BUILT_IN_GOMP_PARALLEL,
5399 BUILT_IN_GOMP_PARALLEL_SECTIONS, BUILT_IN_GOMP_CANCEL,
5400 BUILT_IN_GOMP_CANCELLATION_POINT): New built-ins.
5401 (BUILT_IN_GOMP_PARALLEL_LOOP_STATIC_START,
5402 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC_START,
5403 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED_START,
5404 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME_START,
5405 BUILT_IN_GOMP_PARALLEL_START, BUILT_IN_GOMP_PARALLEL_END,
5406 BUILT_IN_GOMP_PARALLEL_SECTIONS_START): Remove.
5407 * tree-inline.c (remap_gimple_stmt, estimate_num_insns):
5408 Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
5409 * gimple.c (gimple_build_omp_taskgroup, gimple_build_omp_target,
5410 gimple_build_omp_teams): New functions.
5411 (walk_gimple_op): Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and
5412 GIMPLE_OMP_TASKGROUP. Walk optional lhs on GIMPLE_OMP_RETURN.
5413 (walk_gimple_stmt, gimple_copy): Handle GIMPLE_OMP_TARGET,
5414 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
5415 * gimple.h (enum gf_mask): GF_OMP_FOR_KIND_DISTRIBUTE,
5416 GF_OMP_FOR_COMBINED, GF_OMP_FOR_COMBINED_INTO,
5417 GF_OMP_TARGET_KIND_MASK, GF_OMP_TARGET_KIND_REGION,
5418 GF_OMP_TARGET_KIND_DATA, GF_OMP_TARGET_KIND_UPDATE,
5419 GF_OMP_ATOMIC_SEQ_CST): New.
5420 (gimple_build_omp_taskgroup, gimple_build_omp_target,
5421 gimple_build_omp_teams): New prototypes.
5422 (gimple_has_substatements): Handle GIMPLE_OMP_TARGET,
5423 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
5424 (gimple_omp_subcode): Use GIMPLE_OMP_TEAMS instead of
5425 GIMPLE_OMP_SINGLE as end of range.
5426 (gimple_omp_return_set_lhs, gimple_omp_return_lhs,
5427 gimple_omp_return_lhs_ptr, gimple_omp_atomic_seq_cst_p,
5428 gimple_omp_atomic_set_seq_cst, gimple_omp_for_combined_p,
5429 gimple_omp_for_set_combined_p, gimple_omp_for_combined_into_p,
5430 gimple_omp_for_set_combined_into_p, gimple_omp_target_clauses,
5431 gimple_omp_target_clauses_ptr, gimple_omp_target_set_clauses,
5432 gimple_omp_target_kind, gimple_omp_target_set_kind,
5433 gimple_omp_target_child_fn, gimple_omp_target_child_fn_ptr,
5434 gimple_omp_target_set_child_fn, gimple_omp_target_data_arg,
5435 gimple_omp_target_data_arg_ptr, gimple_omp_target_set_data_arg,
5436 gimple_omp_teams_clauses, gimple_omp_teams_clauses_ptr,
5437 gimple_omp_teams_set_clauses): New inlines.
5438 (CASE_GIMPLE_OMP): Add GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS
5439 and GIMPLE_OMP_TASKGROUP.
5440 * tree-core.h (enum omp_clause_code): Add new OpenMP 4.0 clause codes.
5441 (enum omp_clause_depend_kind, enum omp_clause_map_kind,
5442 enum omp_clause_proc_bind_kind): New.
5443 (union omp_clause_subcode): Add depend_kind, map_kind and
5444 proc_bind_kind fields.
5445 * tree-cfg.c (make_edges): Handle GIMPLE_OMP_TARGET,
5446 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
5447 * langhooks-def.h (lhd_omp_mappable_type): New prototype.
5448 (LANG_HOOKS_OMP_MAPPABLE_TYPE): Define.
5449 (LANG_HOOKS_FOR_TYPES_INITIALIZER): Add it.
5451 2013-10-10 Teresa Johnson <tejohnson@google.com>
5453 * predict.c (tree_estimate_probability): Add new parameter
5454 for estimate_bb_frequencies.
5455 (estimate_bb_frequencies): Add new parameter to force estimation.
5456 (rebuild_frequencies): When max frequency in function is small,
5457 recompute counts from frequencies.
5458 * predict.h (estimate_bb_frequencies): New parameter.
5460 2013-10-10 David Malcolm <dmalcolm@redhat.com>
5462 * ipa-inline.c (ipa_inline): Fix leak of "order" when
5463 optimizations are disabled.
5465 2013-10-10 David Malcolm <dmalcolm@redhat.com>
5467 * coverage.c (coverage_finish): Fix leak of da_file_name.
5469 2013-10-10 Jan Hubicka <jh@suse.cz>
5471 * config/i386/x86-tune.def: Enable X86_TUNE_SSE_TYPELESS_STORES
5472 for generic, enable X86_TUNE_SSE_LOAD0_BY_PXOR for Bulldozer,
5475 2013-10-10 Jakub Jelinek <jakub@redhat.com>
5478 * stmt.c (expand_asm_operands): Add FALLTHRU_BB argument,
5479 if any labels are in FALLTHRU_BB, use a special label emitted
5480 immediately after the asm goto insn rather than label_rtx
5482 (expand_asm_stmt): Adjust caller.
5483 * cfgrtl.c (commit_one_edge_insertion): Force splitting of
5484 edge if the last insn in predecessor is a jump with single successor,
5485 but it isn't simplejump_p.
5487 2013-10-10 Richard Biener <rguenther@suse.de>
5489 PR tree-optimization/58656
5490 * tree-ssa-pre.c (phi_translate): Do not cache failed translations.
5492 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
5494 * gimplify.c: Include expr.h and tm_p.h for targets with special
5495 va-arg padding requirements.
5497 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
5499 * tree-flow.h: Move some prototypes to gimple.h.
5500 (gimple_fold_indirect_ref): Move prototype to gimple-fold.h.
5501 * gimple.h: Relocate some prototypes from tree-flow.h
5502 * builtins.c (std_gimplify_va_arg_expr, build_va_arg_indirect_ref):
5504 * gimplify.c (gimple_fold_indirect_ref): Move to gimple-fold.c.
5505 (build_va_arg_indirect_ref): Relocate and make static.
5506 (std_gimplify_va_arg_expr): Relocate here.
5507 * gimple-fold.c (gimple_fold_indirect_ref): Relocate here.
5508 * gimple-fold.h (gimple_fold_indirect_ref): Add prototype.
5510 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5512 * doc/md.texi: Document the mnemonic attribute.
5514 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5517 * gensupport.c (gen_mnemonic_attr): Handle (set (attr x) y) and
5518 (set_attr_alternative x ...) when searching for user defined
5521 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
5523 * config/aplha/alpha.c: Add gimple-ssa.h to include list.
5525 2013-10-09 Easwaran Raman <eraman@google.com>
5527 * params.def (PARAM_MIN_SIZE_FOR_STACK_SHARING): New param...
5528 * cfgexpand.c (defer_stack_allocation): ...use here
5529 * doc/invoke.texi: Add documentation for min-size-for-stack-sharing.
5531 2013-10-09 Zhenqiang Chen <zhenqiang.chen@arm.com>
5533 * tree-ssa-phiopts.c (rhs_is_fed_for_value_replacement): New function.
5534 (operand_equal_for_value_replacement): New function, extracted from
5535 value_replacement and enhanced to catch more cases.
5536 (value_replacement): Use operand_equal_for_value_replacement.
5538 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
5540 * loop-doloop.c (doloop_modify, doloop_optimize): Use
5541 get_max_loop_iterations.
5543 2013-10-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5545 * config/arm/aarch-common.c (arm_early_load_addr_dep):
5546 Place comment above function.
5548 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
5550 * tree-flow.h: Remove all remaining prototypes, enums and structs that
5551 are not related to tree-cfg.c.
5552 * tree-ssa-address.h: New file. Relocate prototypes.
5553 * tree-ssa-address.c: (struct mem_address): Relocate from tree-flow.h.
5554 (addr_for_mem_ref): New. Combine call to get_address_description and
5555 return addr_for_mem_ref.
5556 * expr.c (expand_expr_real_1): Use new addr_for_mem_ref routine.
5557 * tree-ssa-live.h: Adjust prototypes.
5558 * passes.c: Include tree-ssa-live.h.
5559 * gimple-pretty-print.h (gimple_dump_bb): Add prototype.
5560 * graphite.c (graphite_transform_loops): Make static.
5561 (graphite_transforms, gate_graphite_transforms, pass_data_graphite,
5562 make_pass_graphite, pass_data_graphite_transforms,
5563 make_pass_graphite_transforms): Relocate here from tree-ssa-loop.c.
5564 * ipa-pure-const.c (warn_function_noreturn): Make static.
5565 (execute_warn_function_noreturn, gate_warn_function_noreturn,
5566 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
5567 Relocate from tree-cfg.c
5568 * tree-cfg.c (tree_node_can_be_shared, gimple_empty_block_p): Make
5570 (execute_warn_function_noreturn, gate_warn_function_noreturn,
5571 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
5572 Move to ipa-pure-const.c.
5573 (execute_fixup_cfg, class pass_fixup_cfg, make_pass_fixup_cfg):
5574 Relocate from tree-optimize.c.
5575 * tree-optimize.c (execute_fixup_cfg, class pass_fixup_cfg,
5576 make_pass_fixup_cfg): Move to tree-cfg.c.
5577 * tree-chrec.h: (enum ev_direction): Relocate here from tree-flow.h.
5578 Relocate some prototypes.
5579 * tree-data-ref.h (tree_check_data_deps) Add prototype.
5580 * tree-dump.c (dump_function_to_file): Remove prototype.
5581 Add tree-flow.h to the include file.
5582 * tree-dump.h: Remove prototype.
5583 * tree-parloops.h: New File. Add prototypes.
5584 * tree-parloops.c (gate_tree_parallelize_loops, tree_parallelize_loops,
5585 pass_data_parallelize_loops, make_pass_parallelize_loops): Relocate
5586 from tree-ssa-loop.c.
5587 * tree-predcom.c (run_tree_predictive_commoning,
5588 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom):
5589 Relocate here from tree-ssa-loop.c.
5590 * tree-ssa-dom.c (tree_ssa_dominator_optimize) Don't call
5591 ssa_name_values.release ().
5592 * tree-ssa-threadedge.h: New File. Relocate prototypes here.
5593 (ssa_name_values): Relocate from tree-flow.h.
5594 * tree-ssa.h: Include tree-ssa-threadedge.h and tree-ssa-address.h.
5595 * tree-ssa-loop.c (run_tree_predictive_commoning,
5596 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom,
5597 graphite_transforms, gate_graphite_transforms, pass_data_graphite,
5598 make_pass_graphite, pass_data_graphite_transforms,
5599 make_pass_graphite_transforms, gate_tree_parallelize_loops,
5600 tree_parallelize_loops, pass_data_parallelize_loops,
5601 make_pass_parallelize_loops): Move to other files.
5602 * tree-vectorizer.h (lpeel_tree_duplicate_loop_to_edge_cfg): Prototype
5604 * tree.h: Remove prototypes from tree-address.c.
5606 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
5608 * tree-flow.h (tm_restart_node, gimple_df): Move to gimple-ssa.h.
5609 (struct int_tree_map): Move to tree-hasher.h
5610 (SCALE, LABEL, PERCENT): Move to gimple.h
5611 * tree-flow-inline.h: Delete. Move functions to other files.
5612 (unmodifiable_var_p, ref_contains_array_ref): Unused, so delete.
5613 * gimple-ssa.h (tm_restart_node, gimple_df): Relocate from tree-flow.h.
5614 (gimple_in_ssa_p, gimple_vop): Relocate from tree-flow-inline.h
5615 * gimple.h (imple_stmt_max_uid, set_gimple_stmt_max_uid,
5616 inc_gimple_stmt_max_uid, get_lineno): Relocate from tree-flow-inline.h.
5617 (SCALE, LABEL, PERCENT): Relocate from tree-flow.h
5618 * tree-hasher.h: Don't include tree-flow.h.
5619 (struct int_tree_map): Relocate from tree-flow.h.
5620 * tree-sra.c (contains_view_convert_expr_p): Relocate from
5621 tree-flow-inline.h and make static.
5622 * tree-ssa-alias.h (ranges_overlap_p): Relocate from
5624 * tree-ssa-operands.c (gimple_ssa_operands): Relocate from
5625 tree-flow-inline.h and make static.
5626 * tree.h (is_global_var, may_be_aliased): Relocate from
5628 * Makefile.in (GTFILES): Remove tree-flow.h and add gimple-ssa.h.
5629 * value-prof.c: No longer include tree-flow-inline.h.
5630 * tree-switch-conversion.c: No longer include tree-flow-inline.h.
5632 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
5634 * tree-flow.h: Move some protoypes. Include new tree-ssa-loop.h.
5635 (struct affine_iv, struct tree_niter_desc): Move to tree-ssa-loop.h.
5636 (enum move_pos): Move to tree-ssa-loop-im.h
5637 * cfgloop.h: Move some prototypes.
5638 (gcov_type_to_double_int): relocate from tree-ssa-loop.niter.c.
5639 * tree-flow-inline.h (loop_containing_stmt): Move to tree-ssa-loop.h.
5640 * tree-ssa-loop.h: New File. Include other tree-ssa-loop-*.h files.
5641 (struct affine_iv, struct tree_niter_desc): Relocate from tree-flow.h.
5642 (loop_containing_stmt): Relocate from tree-flow-inline.h.
5643 * tree-ssa-loop-ch.c: (do_while_loop_p): Make static.
5644 * tree-ssa-loop-im.c (for_each_index): Move to tree-ssa-loop.c.
5645 (enum move_pos): Relocate here.
5646 (lsm_tmp_name_add, gen_lsm_tmp_name, get_lsm_tmp_name): Move to
5648 (execute_sm_if_changed_flag_set): Change get_lsm_tmp_name call.
5649 (tree_ssa_loop_im, gate_tree_ssa_loop_im, pass_data_lim,
5650 make_pass_lim): Relocate here from tree-ssa-loop.c.
5651 * tree-ssa-loop-ivcanon.c (tree_num_loop_insns): Move to
5653 (loop_edge_to_cancel, unloop_loops): Make static.
5654 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
5655 make_pass_iv_canon): Relocate from tree-ssa-loop.c.
5656 (tree_complete_unroll, gate_tree_complete_unroll,
5657 pass_data_complete_unroll, make_pass_complete_unroll): Relocate here.
5658 (tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
5659 pass_data_complete_unrolli, make_pass_complete_unrolli): Relocate here.
5660 * tree-ssa-loop-ivopts.c: Remove local prototypes.
5661 (stmt_invariant_in_loop_p): Remove unused function.
5662 * tree-ssa-loop-ivopts.h: New file. Add prototypes.
5663 * tree-ssa-loop-manip.h: New file. Add prototypes.
5664 * tree-ssa-loop-niter.c (record_niter_bound): Move to cfgloop.c.
5665 (gcov_type_to_double_int): Move to cfgloop.h.
5666 (double_int_cmp, bound_index,
5667 estimate_numbers_of_iterations_loop): Make static.
5668 (estimated_loop_iterations): Factor out get_estimated_loop_iterations.
5669 (max_loop_iterations): Factor out get_max_loop_iterations.
5670 (estimated_loop_iterations_int, max_stmt_executions_int): Move to
5672 * tree-ssa-loop-niter.h: New file. Add prototypes.
5673 * tree-ssa-loop-prefetch.c (tree_ssa_loop_prefetch,
5674 gate_tree_ssa_loop_prefetch, pass_data_loop_prefetch,
5675 make_pass_loop_prefetch): Relocate from tree-ssa-loop.c.
5676 * tree-ssa-loop-unswitch.c (tree_ssa_loop_unswitch,
5677 gate_tree_ssa_loop_unswitch, pass_data_tree_unswitch,
5678 make_pass_tree_unswitch): Relocate from tree-ssa-loop.c.
5679 * tree-ssa-loop.c (tree_ssa_loop_im, gate_tree_ssa_loop_im,
5680 pass_data_lim, make_pass_lim): Move to tree-ssa-loop-im.c.
5681 (tree_ssa_loop_unswitch, gate_tree_ssa_loop_unswitch,
5682 pass_data_tree_unswitch, make_pass_tree_unswitch): Move.
5683 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
5684 make_pass_iv_canon, tree_complete_unroll, gate_tree_complete_unroll,
5685 pass_data_complete_unroll, make_pass_complete_unroll,
5686 tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
5687 pass_data_complete_unrolli, make_pass_complete_unrolli): Move to
5688 tree-ssa-loop-ivcanon.c.
5689 (tree_ssa_loop_prefetch, gate_tree_ssa_loop_prefetch,
5690 pass_data_loop_prefetch, make_pass_loop_prefetch): Move to
5691 tree-ssa-loop-prefetch.c.
5692 (for_each_index, lsm_tmp_name_add, gen_lsm_tmp_name): Relocate from
5694 (get_lsm_tmp_name): Relocate and add suffix parameter.
5695 (tree_num_loop_insns): Relocate from tree-ssa-ivcanon.c.
5696 * tree-scalar-evolution.h (simple_iv): Don't use affive_iv typedef.
5697 * cfgloop.c (record_niter_bound, estimated_loop_iterations_int,
5698 max_stmt_executions_int): Move from tree-ssa-loop-niter.c.
5699 (get_estimated_loop_iterations): Factor out accessor from
5700 estimated_loop_iterations in tree-ssa-loop-niter.c.
5701 (get_max_loop_iterations): Factor out accessor from
5702 _max_loop_iterations in tree-ssa-niter.c.
5703 * loop-unroll.c (decide_unroll_constant_iterations,
5704 decide_unroll_runtime_iterations, decide_peel_simple,
5705 decide_unroll_stupid): Use new get_* accessors.
5707 2013-10-09 Marc Glisse <marc.glisse@inria.fr>
5709 PR tree-optimization/20318
5710 * doc/extend.texi (returns_nonnull): New function attribute.
5711 * fold-const.c (tree_expr_nonzero_warnv_p): Look for returns_nonnull
5713 * tree-vrp.c (gimple_stmt_nonzero_warnv_p): Likewise.
5714 (stmt_interesting_for_vrp): Accept all GIMPLE_CALL.
5716 2013-10-09 Eric Botcazou <ebotcazou@adacore.com>
5719 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): Return
5720 false if both components are bitfields.
5722 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
5724 * config/aarch64/arm_neon.h (vclz_s8, vclz_s16, vclz_s32)
5725 (vclzq_s8, vclzq_s16, vclzq_s32, vclz_u8, vclz_u16, vclz_u32)
5726 (vclzq_u8, vclzq_u16, vclzq_u32): Replace ASM with C.
5727 * config/aarch64/aarch64.h
5728 (CLZ_DEFINED_VALUE_AT_ZERO): Macro fixed for clz.
5729 * config/aarch64/aarch64-simd-builtins.def
5730 (VAR1 (UNOP, clz, 0, v4si)): Replaced with iterator.
5732 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
5734 * config/aarch64/arm_neon.h (vadd_f64, vsub_f64): Implementation added.
5736 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
5738 * config/aarch64/arm_neon.h (vdiv_f64): Added.
5740 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
5742 * config/aarch64/arm_neon.h (vneg_f32): Asm replaced with C.
5743 (vneg_f64): New intrinsic.
5744 (vneg_s8): Asm replaced with C.
5745 (vneg_s16): Likewise.
5746 (vneg_s32): Likewise.
5747 (vneg_s64): New intrinsic.
5748 (vnegq_f32): Asm replaced with C.
5749 (vnegq_f64): Likewise.
5750 (vnegq_s8): Likewise.
5751 (vnegq_s16): Likewise.
5752 (vnegq_s32): Likewise.
5753 (vnegq_s64): Likewise.
5755 2013-10-09 Renlin Li <Renlin.Li@arm.com>
5757 * config/arm/arm.c (arm_output_mi_thunk): Use plus_constant.
5759 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5761 * config/s390/s390.c (s390_register_info_stdarg_fpr): Remove
5762 packed stack special handling.
5763 (s390_frame_info, s390_emit_prologue, s390_emit_epilogue): Switch
5764 back to fixed stack slots for FPRs saved due to stdarg.
5766 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5768 * config/s390/s390.c (s390_frame_info): Restructure function.
5770 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5772 * config/s390/s390.c (struct s390_frame_layout): New field
5774 (cfun_save_arg_fprs_p, cfun_gpr_save_slot): New macros.
5775 (s390_reg_clobbered_rtx, s390_regs_ever_clobbered): Change type of
5776 regs_ever_clobbered to char*.
5777 (s390_regs_ever_clobbered): Check crtl->saves_all_registers instead
5778 of cfun->has_nonlocal_label. Ignore frame related restore INSNs.
5779 (s390_register_info): Enable FPR save slots. Move/Copy some
5780 functionality into ...
5781 (s390_register_info_gprtofpr, s390_register_info_stdarg_fpr)
5782 (s390_register_info_stdarg_gpr, s390_optimize_register_info): New
5784 (s390_frame_info): Do gpr slot allocation here now. stdarg does
5785 not imply a stack frame.
5786 (s390_init_frame_layout): Remove variable clobbered_regs.
5787 (s390_update_register_info): Remove function.
5788 (s390_hard_regno_rename_ok): Call-saved regs without a save slot
5789 cannot be used for register renaming.
5790 (s390_hard_regno_scratch_ok): New function.
5791 (TARGET_HARD_REGNO_SCRATCH_OK): Define target hook.
5792 (s390_initial_elimination_offset): Change offset calculation of
5793 the return address pointer.
5794 (save_gprs): Deal with only r6 being saved from the call-saved regs.
5795 (restore_gprs): Set frame related flag.
5796 (s390_save_gprs_to_fprs, s390_restore_gprs_from_fprs): New functions.
5797 (s390_emit_prologue): Call s390_register_info instead of
5798 s390_update_frame_layout. Call s390_save_gprs_to_fprs.
5799 (s390_emit_epilogue): Call s390_restore_gprs_from_fprs.
5800 (s390_optimize_prologue): Call s390_optimize_register_info.
5801 Try to remove also FPR slot save/restore INSNs. Remove frame
5802 related flags from restore INSNs.
5804 2013-10-08 DJ Delorie <dj@redhat.com>
5806 * config/rl78/rl78-expand.md (movqi): use operands[] not operandN.
5809 * config/rl78/rl78.c (rl78_print_operand_1): Change %c to %C to
5810 avoid conflict with the MI use of %c.
5811 * config/rl78/rl78-real.md: change %c to %C throughout.
5812 * config/rl78/rl78-virt.md: Likewise.
5814 2013-10-08 Jan Hubicka <jh@suse.cz>
5816 * config/i386/i386.c (ix86_option_override_internal): Switch
5817 to SSE math for -ffast-math when target ISA supports SSE2.
5819 2013-10-08 Andrew MacLeod <amacleod@redhat.com>
5821 * tree-flow.h: Remove some prototypes.
5822 * tree.h: Remove some protypes, add a couple.
5823 * tree.c (using_eh_for_cleanups_flag, using_eh_for_cleanups,
5824 using_eh_for_cleanups_p): Add interface routines for front ends.
5825 * tree-eh.h: New file. Add protoptyes.
5826 * tree-eh.c (using_eh_for_cleanups_p, using_eh_for_cleanups): Delete.
5827 (add_stmt_to_eh_lp_fn): Make static.
5828 (lower_try_finally): Use new using_eh_for_cleanups_p.
5829 * emit-rtl.c: Include tree-eh.h.
5830 * gimple.h: Include tree-eh.h.
5832 2013-10-08 Marc Glisse <marc.glisse@inria.fr>
5834 PR tree-optimization/58480
5835 * tree-vrp.c (infer_nonnull_range): New function.
5836 (infer_value_range): Call infer_nonnull_range.
5838 2013-10-08 Dehao Chen <dehao@google.com>
5840 PR tree-optimization/58619
5841 * tree-inline.c (copy_phis_for_bb): Combine location data
5844 2013-10-08 Zhenqiang Chen <zhenqiang.chen@linaro.org>
5847 * config/arm/arm.c (arm_emit_ldrd_pop): Attach
5848 RTX_FRAME_RELATED_P on INSN.
5850 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5852 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const_le): New.
5853 (altivec_expand_vec_perm_const): Call it.
5855 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5857 * config/rs6000/vector.md (mov<mode>): Emit permuted move
5858 sequences for LE VSX loads and stores at expand time.
5859 * config/rs6000/rs6000-protos.h (rs6000_emit_le_vsx_move): New
5861 * config/rs6000/rs6000.c (rs6000_const_vec): New.
5862 (rs6000_gen_le_vsx_permute): New.
5863 (rs6000_gen_le_vsx_load): New.
5864 (rs6000_gen_le_vsx_store): New.
5865 (rs6000_gen_le_vsx_move): New.
5866 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): New.
5867 (*vsx_le_perm_load_v4si): New.
5868 (*vsx_le_perm_load_v8hi): New.
5869 (*vsx_le_perm_load_v16qi): New.
5870 (*vsx_le_perm_store_v2di): New.
5871 (*vsx_le_perm_store_v4si): New.
5872 (*vsx_le_perm_store_v8hi): New.
5873 (*vsx_le_perm_store_v16qi): New.
5874 (*vsx_xxpermdi2_le_<mode>): New.
5875 (*vsx_xxpermdi4_le_<mode>): New.
5876 (*vsx_xxpermdi8_le_V8HI): New.
5877 (*vsx_xxpermdi16_le_V16QI): New.
5878 (*vsx_lxvd2x2_le_<mode>): New.
5879 (*vsx_lxvd2x4_le_<mode>): New.
5880 (*vsx_lxvd2x8_le_V8HI): New.
5881 (*vsx_lxvd2x16_le_V16QI): New.
5882 (*vsx_stxvd2x2_le_<mode>): New.
5883 (*vsx_stxvd2x4_le_<mode>): New.
5884 (*vsx_stxvd2x8_le_V8HI): New.
5885 (*vsx_stxvd2x16_le_V16QI): New.
5887 2013-10-07 Renlin Li <Renlin.Li@arm.com>
5889 * config/arm/arm-cores.def (cortex-a53): Use cortex tuning.
5891 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5893 * config/s390/s390.c (s390_register_info): Make the call-saved FPR
5894 loop to work also for 31bit ABI.
5895 Save the stack pointer for frame_size > 0.
5897 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5899 * config/s390/s390.md ("tbegin", "tbegin_nofloat", "tbegin_retry")
5900 ("tbegin_retry_nofloat", "tend", "tabort", "tx_assist"): Remove
5901 constraint letters from expanders.
5902 ("tbegin_retry", "tbegin_retry_nofloat"): Change predicate of the
5903 retry count to general_operand.
5904 ("tabort"): Give operand 0 a mode.
5905 ("tabort_1"): Add mode and constraint letter for operand 0.
5906 * doc/extend.texi: Fix protoype of __builtin_non_tx_store.
5908 2013-10-04 Jeff Law <law@redhat.com>
5910 * tree-ssa-threadedge.c: Fix some trailing whitespace problems.
5912 * tree-ssa-threadedge.c (thread_through_normal_block): Broken
5914 (thread_across_edge): Here. Call it.
5916 2013-10-04 Cary Coutant <ccoutant@google.com>
5918 * dwarf2out.c (dw_sra_loc_expr): Release addr_table entries when
5919 discarding a location list expression (or a piece of one).
5921 2013-10-03 Jan Hubicka <jh@suse.cz>
5923 * config/i386/i386.c (ix86_issue_rate): Pentium4, Nocona has issue
5924 rate of 2. Core2, Corei7 and Haswell has issue rate of 4.
5925 (ix86_adjust_cost): Remove Atom case; fix core2/corei7/Haswell case.
5927 2013-10-03 Jan Hubicka <jh@suse.cz>
5929 * config/i386/i386.c (ix86_option_override_internal): Do not enable
5930 accumulate-outgoing-args when producing unwind info.
5932 2013-10-03 Wei Mi <wmi@google.com>
5934 * lra-constraints.c (insert_move_for_subreg): New function
5935 extracted from simplify_operand_subreg.
5936 (simplify_operand_subreg): Add reload for paradoxical subreg.
5938 2013-10-03 Rong Xu <xur@google.com>
5940 * ipa-inline-analysis.c (find_foldable_builtin_expect): Find
5941 the candidate of builtin_expect such that we should fix the
5942 size/time estimation.
5943 (estimate_function_body_sizes): Do the acutally size/time fix-up
5946 2013-10-03 Rong Xu <xur@google.com>
5948 * predict.c (tree_predict_by_opcode): Get the probability
5949 for builtin_expect from param builtin_expect_probability.
5950 * params.def (BUILTIN_EXPECT_PROBABILITY): New parameter.
5951 * predict.def (PRED_BUILTIN_EXPECT_RELAXED): Fix comments.
5952 * doc/invoke.texi: Add documentation for builtin-expect-probability.
5954 2013-10-03 Marc Glisse <marc.glisse@inria.fr>
5957 * common.opt (fcheck-new): Moved from c.opt. Make it 'Common'.
5958 * calls.c (alloca_call_p): Use get_callee_fndecl.
5959 * fold-const.c (tree_expr_nonzero_warnv_p): Handle operator new.
5960 * tree-vrp.c (gimple_stmt_nonzero_warnv_p, stmt_interesting_for_vrp):
5962 (vrp_visit_stmt): Remove duplicated code.
5964 2013-10-03 Michael Meissner <meissner@linux.vnet.ibm.com>
5966 * config/rs6000/rs6000-builtin.def (XSRDPIM): Use floatdf2,
5967 ceildf2, btruncdf2, instead of vsx_* name.
5969 * config/rs6000/vsx.md (vsx_add<mode>3): Change arithmetic
5970 iterators to only do V2DF and V4SF here. Move the DF code to
5971 rs6000.md where it is combined with SF mode. Replace <VSv> with
5972 just 'v' since only vector operations are handled with these insns
5973 after moving the DF support to rs6000.md.
5974 (vsx_sub<mode>3): Likewise.
5975 (vsx_mul<mode>3): Likewise.
5976 (vsx_div<mode>3): Likewise.
5977 (vsx_fre<mode>2): Likewise.
5978 (vsx_neg<mode>2): Likewise.
5979 (vsx_abs<mode>2): Likewise.
5980 (vsx_nabs<mode>2): Likewise.
5981 (vsx_smax<mode>3): Likewise.
5982 (vsx_smin<mode>3): Likewise.
5983 (vsx_sqrt<mode>2): Likewise.
5984 (vsx_rsqrte<mode>2): Likewise.
5985 (vsx_fms<mode>4): Likewise.
5986 (vsx_nfma<mode>4): Likewise.
5987 (vsx_copysign<mode>3): Likewise.
5988 (vsx_btrunc<mode>2): Likewise.
5989 (vsx_floor<mode>2): Likewise.
5990 (vsx_ceil<mode>2): Likewise.
5991 (vsx_smaxsf3): Delete scalar ops that were moved to rs6000.md.
5992 (vsx_sminsf3): Likewise.
5993 (vsx_fmadf4): Likewise.
5994 (vsx_fmsdf4): Likewise.
5995 (vsx_nfmadf4): Likewise.
5996 (vsx_nfmsdf4): Likewise.
5997 (vsx_cmpdf_internal1): Likewise.
5999 * config/rs6000/rs6000.h (TARGET_SF_SPE): Define macros to make it
6000 simpler to select whether a target has SPE or traditional floating
6001 point support in iterators.
6002 (TARGET_DF_SPE): Likewise.
6003 (TARGET_SF_FPR): Likewise.
6004 (TARGET_DF_FPR): Likewise.
6005 (TARGET_SF_INSN): Macros to say whether floating point support
6006 exists for a given operation for expanders.
6007 (TARGET_DF_INSN): Likewise.
6009 * config/rs6000/rs6000.c (Ftrad): New mode attributes to allow
6010 combining of SF/DF mode operations, using both traditional and VSX
6018 (abs<mode>2): Combine SF/DF modes using traditional floating point
6019 instructions. Add support for using the upper DF registers with
6020 VSX support, and SF registers with power8-vector support. Update
6021 expanders for operations supported by both the SPE and traditional
6022 floating point units.
6023 (abs<mode>2_fpr): Likewise.
6024 (nabs<mode>2): Likewise.
6025 (nabs<mode>2_fpr): Likewise.
6026 (neg<mode>2): Likewise.
6027 (neg<mode>2_fpr): Likewise.
6028 (add<mode>3): Likewise.
6029 (add<mode>3_fpr): Likewise.
6030 (sub<mode>3): Likewise.
6031 (sub<mode>3_fpr): Likewise.
6032 (mul<mode>3): Likewise.
6033 (mul<mode>3_fpr): Likewise.
6034 (div<mode>3): Likewise.
6035 (div<mode>3_fpr): Likewise.
6036 (sqrt<mode>3): Likewise.
6037 (sqrt<mode>3_fpr): Likewise.
6038 (fre<Fs>): Likewise.
6039 (rsqrt<mode>2): Likewise.
6040 (cmp<mode>_fpr): Likewise.
6041 (smax<mode>3): Likewise.
6042 (smin<mode>3): Likewise.
6043 (smax<mode>3_vsx): Likewise.
6044 (smin<mode>3_vsx): Likewise.
6045 (negsf2): Delete SF operations that are merged with DF.
6052 (fmasf4_fpr): Likewise.
6053 (fmssf4_fpr): Likewise.
6054 (nfmasf4_fpr): Likewise.
6055 (nfmssf4_fpr): Likewise.
6056 (sqrtsf2): Likewise.
6057 (rsqrtsf_internal1): Likewise.
6058 (smaxsf3): Likewise.
6059 (sminsf3): Likewise.
6060 (cmpsf_internal1): Likewise.
6061 (copysign<mode>3_fcpsgn): Add VSX/power8-vector support.
6062 (negdf2): Delete DF operations that are merged with SF.
6064 (nabsdf2): Likewise.
6070 (rsqrtdf_internal1): Likewise.
6071 (fmadf4_fpr): Likewise.
6072 (fmsdf4_fpr): Likewise.
6073 (nfmadf4_fpr): Likewise.
6074 (nfmsdf4_fpr): Likewise.
6075 (sqrtdf2): Likewise.
6076 (smaxdf3): Likewise.
6077 (smindf3): Likewise.
6078 (cmpdf_internal1): Likewise.
6079 (lrint<mode>di2): Use TARGET_<MODE>_FPR macro.
6080 (btrunc<mode>2): Delete separate expander, and combine with the
6081 insn and add VSX instruction support. Use TARGET_<MODE>_FPR.
6082 (btrunc<mode>2_fpr): Likewise.
6083 (ceil<mode>2): Likewise.
6084 (ceil<mode>2_fpr): Likewise.
6085 (floor<mode>2): Likewise.
6086 (floor<mode>2_fpr): Likewise.
6087 (fma<mode>4_fpr): Combine SF and DF fused multiply/add support.
6088 Add support for using the upper registers with VSX and
6089 power8-vector. Move insns to be closer to the define_expands. On
6090 VSX systems, prefer the traditional form of FMA over the VSX
6091 version, since the traditional form allows the target not to
6092 overlap with the inputs.
6093 (fms<mode>4_fpr): Likewise.
6094 (nfma<mode>4_fpr): Likewise.
6095 (nfms<mode>4_fpr): Likewise.
6097 2013-10-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6098 Richard Earnshaw <richard.earnshaw@arm.com>
6100 * config/arm/aarch-common-protos.h (struct alu_cost_table): New.
6101 (struct mult_cost_table): Likewise.
6102 (struct mem_cost_table): Likewise.
6103 (struct fp_cost_table): Likewise.
6104 (struct vector_cost_table): Likewise.
6105 (cpu_cost_table): Likewise.
6106 * config/arm/arm.opt (mold-rts-costs): New option.
6107 (mnew-generic-costs): Likewise.
6108 * config/arm/arm.c (generic_extra_costs): New table.
6109 (cortexa15_extra_costs): Likewise.
6110 (arm_slowmul_tune): Use NULL as new costs.
6111 (arm_fastmul_tune): Likewise.
6112 (arm_strongarm_tune): Likewise.
6113 (arm_xscale_tune): Likewise.
6114 (arm_9e_tune): Likewise.
6115 (arm_v6t2_tune): Likewise.
6116 (arm_cortex_a5_tune): Likewise.
6117 (arm_cortex_a9_tune): Likewise.
6118 (arm_v6m_tune): Likewise.
6119 (arm_fa726te_tune): Likewise.
6120 (arm_cortex_a15_tune): Use cortex15_extra_costs.
6121 (arm_cortex_tune): Use generict_extra_costs.
6122 (shifter_op_p): New function.
6123 (arm_unspec_cost): Likewise.
6124 (LIBCALL_COST): Define.
6125 (arm_new_rtx_costs): New function.
6126 (arm_rtx_costs): Use arm_new_rtx_costs when core-specific
6127 table is available. Use old costs otherwise unless mnew-generic-costs
6129 * config/arm/arm-protos.h (tune_params): Add insn_extra_cost field.
6130 (cpu_cost_table): Declare.
6132 2013-10-03 Marcus Shawcroft <marcus.shawcroft@arm.com>
6135 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>)
6136 (*subs_mul_imm_<mode>)
6137 (*add_<shift>_<mode>, *add_<shift>_si_uxtw,*add_mul_imm_<mode>)
6138 (*sub_<shift>_<mode>)
6139 (*sub_<shift>_si_uxtw,*sub_mul_imm_<mode>, *sub_mul_imm_si_uxtw):
6140 Remove k constraint.
6142 2013-10-03 Ian Bolton <ian.bolton@arm.com>
6144 * config/aarch64/aarch64.c (aarch64_secondary_reload): Remove legacy
6146 * config/aarch64/aarch64.md (reload_sp_immediate): Likewise.
6148 2013-10-02 Teresa Johnson <tejohnson@google.com>
6150 * predict.c (probably_never_executed): New function.
6151 (probably_never_executed_bb_p): Invoke probably_never_executed.
6152 (probably_never_executed_edge_p): Ditto.
6153 * bb-reorder.c (find_rarely_executed_basic_blocks_and_crossing_edges):
6154 Treat profile insanities conservatively.
6156 2013-10-02 John David Anglin <danglin@gcc.gnu.org>
6158 * config.gcc (hppa*64*-*-linux*): Don't add pa/t-linux to tmake_file.
6160 2013-10-02 Vladimir Makarov <vmakarov@redhat.com>
6162 * lra-constraints.c (process_alt_operand): Calculate scratch_p and
6163 use it. Use smaller increase for scratch. Don't increase reject
6164 for early clobber scratch.
6165 * lra-eliminations.c (eliminate_regs_in_insn): Remove all insns
6166 setting eliminated regs except setting fp from hfp.
6167 (lra_eliminate): Check lra_insn_recog_data on NULL.
6169 2013-10-02 Michael Meissner <meissner@linux.vnet.ibm.com>
6172 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Turn off
6173 setting -mvsx-timode by default until the underlying problem is fixed.
6174 (RS6000_CPU, power7 defaults): Likewise.
6176 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
6178 * config/x-linux (host-linux.o): Remove header dependencies.
6179 Use $(COMPILE) and $(POSTCOMPILE).
6180 * config/t-linux-android (linux-android.o): Ditto.
6182 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
6184 * Makefile.in (expmed.o-warn): Remove.
6186 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
6188 * graphite-scop-detection.c: Include tree-ssa-propagate,h.
6189 * graphite-sese-to-poly.c: Include tree-ssa-propagate.h.
6191 2013-10-02 Teresa Johnson <tejohnson@google.com>
6193 * dojump.c (do_jump_1): Divide probability between
6194 both conditions of a TRUTH_ANDIF_EXPR/TRUTH_ORIF_EXPR.
6196 2013-10-02 Tom Tromey <tromey@redhat.com>
6198 * Makefile.in (DRIVER_DEFINES): Use $(if), not $(and).
6200 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
6202 * tree-flow.h: Remove some prototypes.
6203 * tree-ssa-dce.c (mark_virtual_operand_for_renaming,
6204 mark_virtual_phi_result_for_renaming): Move to tree-into-ssa.c.
6205 * tree-into-ssa.c (mark_virtual_operand_for_renaming,
6206 mark_virtual_phi_result_for_renaming): Relocate here.
6207 * tree-into-ssa.h: Add prototypes.
6208 * tree-ssa-phiopt.c: (tree_ssa_phiopt_worker) Use
6209 single_pred_before_succ_order.
6210 (blocks_in_phiopt_order): Rename and move to cfganal.c.
6211 (nonfreeing_call_p) Move to gimple.c.
6212 * cfganal.c (single_pred_before_succ_order): Move and renamed from
6214 * basic-block.h (single_pred_before_succ_order): Add prototype.
6215 * gimple.c (nonfreeing_call_p): Relocate here.
6216 * gimple.h: Add prototype.
6217 * tree-ssa-ifcombine.c: Include tree-ssa-phiopt.h.
6218 * tree-ssa-dom.h: New file. Relocate prototypes here.
6219 * tree-ssa.h: Include tree-ssa-dom.h.
6221 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
6223 * config/i386/x-i386 (driver-i386.o): Remove header dependencies.
6224 Use $(COMPILE) and $(POSTCOMPILE).
6226 * config/alpha/x-alpha (driver-alpha.o): Ditto.
6228 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
6230 * tree-flow.h: Remove some prototypes.
6231 * gimple-fold.h: Add prototypes from gimple.h and tree-flow.h.
6232 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
6233 * tree-ssa-copy.c (may_propagate*, propagate_value, replace_exp,
6234 propagate_tree_value*): Move from here to...
6235 * tree-ssa-propagate.c (may_propagate*, propagate_value, replace_exp,
6236 propagate_tree_value*): Relocate here.
6237 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
6238 * gimple.h: Include gimple-fold.h, move prototypes into gimple-fold.h.
6239 * gimple-fold.c: Remove gimple-fold.h from include list.
6240 * tree-vrp.c: Remove gimple-fold.h from include list.
6241 * tree-ssa-sccvn.c: Remove gimple-fold.h from include list.
6242 * tree-ssa-ccp.c: Remove gimple-fold.h from include list.
6243 * tree-scalar-evolution.c: Add tree-ssa-propagate.h to include list.
6244 * tree-ssa-pre.c: Add tree-ssa-propagate.h to include list.
6245 * sese.c: Add tree-ssa-propagate.h to include list.
6247 2013-10-02 Richard Biener <rguenther@suse.de>
6249 * tree-loop-distribution.c: Include tree-vectorizer.h for
6251 (enum partition_kind): Remove PKIND_REDUCTION.
6252 (struct partition_s): Remove has_writes member, add reduction_p member.
6253 (partition_alloc): Adjust.
6254 (partition_builtin_p): Likewise.
6255 (partition_has_writes): Remove.
6256 (partition_reduction_p): New function.
6257 (partition_merge_into): Likewise.
6258 (generate_code_for_partition): Commonize builtin partition
6260 (rdg_cannot_recompute_vertex_p): Remove.
6261 (already_processed_vertex_p): Likewise.
6262 (rdg_flag_vertex): Do not set has_writes.
6263 (classify_partition): Adjust.
6264 (rdg_build_partitions): Do not set has_writes, treat all
6265 partitions as useful.
6266 (distribute_loop): Record number of library calls generated. Adjust.
6267 (tree_loop_distribution): Report number of loops and library
6268 calls generated as opt-info.
6270 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
6272 * tree-flow.h: Include new .h files. Move prototypes.
6273 * tree-cfgcleanup.h: New file. Add prototypes from tree-flow.h.
6274 * tree-dfa.h: New File. Add prototypes from tree-flow.h.
6275 (get_addr_base_and_unit_offset_1) Move from tree-flow-inline.h.
6276 * tree-pretty-print.h: Add prototypes from tree-flow.h.
6277 * tree-into-ssa.h: New File. Add prototypes from tree-flow.h.
6278 ({debug|dump}*): Move debugging prototypes out of tree-into-ssa.c.
6279 * tree-into-ssa.c ({debug|dump}*): Move prototypes to header file.
6280 * tree.h (get_ref_base_and_extent): Move prototype out.
6281 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Move to
6283 * gimple-low.h: New File. Add prototypes from tree-flow.h.
6284 * gimple-low.c (try_catch_may_fallthru, block_may_fallthru): Move to...
6285 * tree.c (try_catch_may_fallthru, block_may_fallthru): Here.
6286 * tree-scalar-evolution.c: Include tree.h.
6287 * sese.c: Include tree.h.
6288 * dumpfile.c: Move gimple-pretty-print.h include after tree.h.
6289 * dwarf2out.c: Include tree-dfa.h.
6290 * tree-chrec.c: Include tree.h.
6291 * tree-data-ref.c: Include tree.h.
6293 2013-10-02 Yufeng Zhang <yufeng.zhang@arm.com>
6295 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref):
6298 2013-10-02 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
6300 * config/t-sol2 (sol2-c.o): Remove header dependencies.
6301 Use $(COMPILE) and $(POSTCOMPILE).
6302 (sol2-cxx.o): Likewise.
6303 (sol2-stubs.o): Likewise.
6305 * config/x-solaris (host-solaris.o): Likewise.
6307 * config/sparc/t-sparc (sparc.o): Remove.
6308 (sparc-c.o): Remove header dependencies.
6309 Use $(COMPILE) and $(POSTCOMPILE).
6310 * config/sparc/x-sparc: Likewise.
6312 2013-10-02 Joern Rennecke <joern.rennecke@embecosm.com>
6314 * config/arc/arc-opts.h: Add 2013 to Copyright years.
6315 * config/arc/arc700.md: Likewise.
6316 * config/arc/arc-modes.def: Likewise.
6317 * config/arc/arc-simd.h: Likewise.
6318 * config/arc/t-arc-uClibc: Likewise.
6319 * config/arc/t-arc-newlib: Likewise.
6321 2013-10-02 Renlin Li <renlin.li@arm.com>
6323 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
6325 (aarch64_expand_epilogue): Likewise.
6327 2013-10-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6328 Yufeng Zhang <yufeng.zhang@arm.com>
6330 * gimple-ssa-strength-reduction.c (legal_cast_p_1): Forward
6332 (backtrace_base_for_ref): Call get_unwidened with 'base_in' if
6333 'base_in' represent a conversion and legal_cast_p_1 holds; set
6334 'base_in' with the returned value from get_unwidened.
6336 2013-10-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6338 * config/arm/arm.c (arm_legitimize_reload_address): Explain why
6339 plus_constant is not used.
6341 2013-10-01 Wei Mi <wmi@google.com>
6343 * config/i386/x86-tune.def (DEF_TUNE): Remove m_CORE_ALL.
6344 * config/i386/i386.md: Add define_peephole2 to
6345 break partial reg stall for cvtss2sd/cvtsd2ss.
6347 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
6349 * config/arc/arc.c (pass_arc_ifcvt::clone):
6350 Update for ctxt_ -> m_ctxt change.
6352 2013-10-01 Jeff Law <law@redhat.com>
6354 * tree-ssa-threadupdate.c (struct redirection_data): Delete
6355 outgoing_edge and intermediate_edge fields. Instead store the path.
6356 (redirection_data::hash): Hash on the last edge's destination index.
6357 (redirection_data::equal): Check the entire thread path.
6358 (lookup_redirectio_data): Corresponding changes.
6359 (create_edge_and_update_destination_phis): Likewise.
6360 (thread_single_edge): Likewise.
6362 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
6363 Diego Novillo <dnovillo@google.com>
6365 * config/arc/simdext.md (UNSPEC_ARC_SIMD_VLD32WH): Delete.
6366 (UNSPEC_ARC_SIMD_VLD32WL): Likewise.
6367 (vld32wh_insn, vld32wl_insn): Delete commented-out old
6368 versions of these patterns.
6370 * doc/extend.texi (long_call/medium_call/short_call): Typo fix.
6371 (__builtin_arc_aligned): Likewise.
6373 * config/arc/arc.md: Expand adc_0 comment stating the intended
6374 purpose and why it isn't ready.
6375 Replace commented out call_value_via_label_mixed with a
6376 plain comment about bl_s.
6378 * config/arc/arc.c (stdio.h): Don't include directly.
6379 (arc_expand_epilogue): Remove [0]: Remove fp_restored_p.
6380 Remove if (1) condition.
6381 (arc_encode_section_info): Fix comment.
6383 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
6385 * config/arc/arc.c (arc_conditional_register_usage):
6386 Use ARC_FIRST_SIMD_VR_REG / ARC_LAST_SIMD_VR_REG.
6387 Also set reg_alloc_order for DMA config regs.
6389 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
6390 Jeremy Bennett <jeremy.bennett@embecosm.com>
6392 * doc/install.texi (--with-cpu): Mention ARC.
6393 (arc-*-elf32): New paragraph.
6394 (arc-linux-uclibc): Likewise.
6395 * doc/md.texi (Machine Constraints): Add ARC part.
6396 * doc/invoke.texi: (menu): Add ARC Options.
6397 (Machine Dependent Options) <ARC Options>: Add synopsis.
6398 (node ARC Options): Add.
6399 * doc/extend.texi (long_call / short_call attribute): Add ARC.
6400 (ARC Built-in Functions): New section defining
6401 generic ARC built-in functions.
6402 (ARC SIMD Built-in Functions): New section defining SIMD specific
6404 (Declaring Attributes of Functions): Extended
6405 description of short_call and long_call attributes for ARC and
6406 added index entries.
6408 2013-10-01 Saurabh Verma <saurabh.verma@codito.com>
6409 Ramana Radhakrishnan <ramana.radhakrishnan@codito.com>
6410 Joern Rennecke <joern.rennecke@embecosm.com>
6411 Muhammad Khurram Riaz <khurram.riaz@arc.com>
6412 Brendan Kehoe <brendan@zen.org>
6413 Michael Eager <eager@eagercon.com>
6414 Simon Cook <simon.cook@embecosm.com>
6415 Jeremy Bennett <jeremy.bennett@embecosm.com>
6417 * config/arc, common/config/arc: New directories.
6419 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
6420 Brendan Kehoe <brendan@zen.org>
6421 Simon Cook <simon.cook@embecosm.com>
6423 * config.gcc (arc*-*-elf*, arc*-*-linux-uclibc*): New configurations.
6425 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
6427 * tree-ssa-live.h (coalesce_ssa_name): Move Prototype to...
6428 * tree-ssa-coalesce.h: New. Move prototype to here.
6429 * tree-outof-ssa.h: Include tree-ssa-coalesce.h.
6430 * tree-ssa-coalesce.c: Include tree-outof-ssa.h.
6431 (gimple_can_coalesce_p): Move to...
6432 * gimple.c (gimple_can_coalesce_p): Here.
6434 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
6436 * tree-into-ssa.c (enum need_phi_state): Relocate from tree-flow.h.
6437 (dump_decl_set): Move to gimple.c.
6438 * gimple.h: Don't include tree-ssa-operands.h.
6439 (dump_decl_set): Add prototype.
6440 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
6441 Move to gimple-ssa.h.
6442 (phi_ssa_name_p, phi_nodes, phi_nodes_ptr, gimple_phi_arg_def,
6443 gimple_phi_arg_def_ptr, gimple_phi_arg_edge, gimple_phi_arg_location,
6444 gimple_phi_arg_location_from_edge, gimple_phi_arg_set_location,
6445 gimple_phi_arg_has_location): Relocate from tree-flow-inline.h
6446 * gimple.c (walk_stmt_load_store_ops): Use gimple_phi_arg_def rather
6448 (dump_decl_set): Relocate here.
6449 * gimple-ssa.h: New file.
6450 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
6451 Relocate from gimple.h.
6452 * tree-cfg.c (has_zero_uses_1, single_imm_use_1): Move to...
6453 * tree-ssa-operands.c (swap_ssa_operands): Rename from
6454 swap_tree_operands and remove non-ssa path.
6455 (has_zero_uses_1, single_imm_use_1): Relocate from tree-cfg.c.
6456 * tree-ssa-reassoc.c (linearize_expr_tree, repropagate_negates): Use
6458 * tree-vect-loop.c (destroy_loop_vec_info, vect_is_slp_reduction,
6459 vect_is_simple_reduction_1): Use swap_ssa_operands.
6460 * tree-flow.h: Move various prototypes to tree-phinodes.h.
6461 (enum need_phi_state): Move to tree-into-ssa.c.
6462 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
6463 BREAK_FROM_IMM_USE_STMT): Move to ssa-iterators.h.
6464 (swap_tree_operands): Rename and move prototype to tree-ssa-operands.h.
6465 * tree-flow-inline.h (delink_imm_use, link_imm_use_to_list,
6466 link_imm_use, set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
6467 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
6468 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
6469 num_imm_uses): Move to ssa-iterators.h.
6470 (get_use_from_ptr, get_def_from_ptr): Move to tree-ssa-operands.h
6471 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Move to
6473 (op_iter_done, op_iter_next_def, op_iter_next_tree,
6474 clear_and_done_ssa_iter, op_iter_init, op_iter_init_use,
6475 op_iter_init_def, op_iter_init_tree, single_ssa_tree_operand,
6476 single_ssa_use_operand, single_ssa_def_operand, zero_ssa_operands,
6477 num_ssa_operands, delink_stmt_imm_use, single_phi_def,
6478 op_iter_init_phiuse, op_iter_init_phidef, end_imm_use_stmt_p,
6479 end_imm_use_stmt_traverse, move_use_after_head, link_use_stmts_after,
6480 first_imm_use_stmt, next_imm_use_stmt, first_imm_use_on_stmt,
6481 end_imm_use_on_stmt_p, next_imm_use_on_stmt): Move to ssa-iterators.h.
6482 (gimple_phi_arg_def, gimple_phi_arg_def_ptr, gimple_phi_arg_edge,
6483 gimple_phi_arg_location, gimple_phi_arg_location_from_edge,
6484 gimple_phi_arg_set_location, gimple_phi_arg_has_location, phi_nodes,
6485 phi_nodes_ptr, phi_ssa_name_p): Move to gimple.h.
6486 (set_phi_nodes): Move to tree-phinodes.h.
6487 * tree-ssa-operands.h (enum ssa_op_iter_type,
6488 struct ssa_operand_iterator_d, SSA_OP*, FOR_EACH_SSA*, SINGLE_SSA*,
6489 ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS): Move to ssa-iterators.h.
6490 (dump_decl_set): Remove prototype.
6491 (get_use_from_ptr, get_def_from_ptr): Relocate from tree-flow.h.
6492 * tree-phinodes.h: New file. Move some prototypes from tree-flow.h.
6493 (set_phi_nodes): Relocate from tree-flow-inline.h.
6494 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Relocate from
6496 * tree-ssa.h: Add tree-phinodes.h, gimple-ssa.h, ssa-iterators.h to
6497 include list. Temporarily add gimple.h to include list.
6498 * ssa-iterators.h: New file.
6499 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
6500 BREAK_FROM_IMM_USE_STMT): Relocate from tree-flow.h.
6501 (enum ssa_op_iter_type, struct ssa_operand_iterator_d, SSA_OP*,
6502 FOR_EACH_SSA*, SINGLE_SSA*, ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS):
6503 Relocate from tree-ssa-operands.h.
6504 (delink_imm_use, link_imm_use_to_list, link_imm_use,
6505 set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
6506 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
6507 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
6508 num_imm_uses, get_use_from_ptr, get_def_from_ptr,
6509 phi_arg_index_from_use, op_iter_done, op_iter_next_def,
6510 op_iter_next_tree, clear_and_done_ssa_iter, op_iter_init,
6511 op_iter_init_use, op_iter_init_def, op_iter_init_tree,
6512 single_ssa_tree_operand, single_ssa_use_operand, single_ssa_def_operand,
6513 zero_ssa_operands, num_ssa_operands, delink_stmt_imm_use,
6514 single_phi_def, op_iter_init_phiuse, op_iter_init_phidef,
6515 end_imm_use_stmt_p, end_imm_use_stmt_traverse, move_use_after_head,
6516 link_use_stmts_after, first_imm_use_stmt, next_imm_use_stmt,
6517 first_imm_use_on_stmt, end_imm_use_on_stmt_p, next_imm_use_on_stmt):
6518 Relocate from tree-flow-inline.h.
6519 * tree-outof-ssa.h: Change _SSAEXPAND_H macro to GCC_TREE_OUTOF_SSA_H.
6521 2013-10-01 Vidya Praveen <vidyapraveen@arm.com>
6524 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l2<mode>_internal): Rename to ...
6525 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): ... this;
6526 Insert '\t' to output template.
6527 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): New.
6528 (aarch64_saddl2<mode>, aarch64_uaddl2<mode>): Modify to call
6529 gen_aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal() instead.
6530 (aarch64_ssubl2<mode>, aarch64_usubl2<mode>): Ditto.
6532 2013-10-01 Uros Bizjak <ubizjak@gmail.com>
6534 * doc/install.texi (Host/target specific installation notes for GCC):
6535 Put @anchor before @heading.
6536 * doc/gcc.texi (titlepage): Use @uref and http:// prefix for website.
6537 Use @email for email addresses.
6539 2013-10-01 Jeff Law <law@redhat.com>
6541 * tree-ssa-threadedge.c (thread_across_edge): Make path a pointer to
6542 a vec. Only delete the path if we create one without successfully
6543 registering a jump thread.
6544 * tree-ssa-threadupdate.h (register_jump_thread): Pass in path vector
6546 * tree-ssa-threadupdate.c (threaded_edges): Remove. No longer used
6547 (paths): New vector of jump threading paths.
6548 (THREAD_TARGET, THREAD_TARGET2): Remove accessor macros.
6549 (THREAD_PATH): New accessor macro for the entire thread path.
6550 (lookup_redirection_data): Get intermediate and final outgoing edge
6551 from the thread path.
6552 (create_edge_and_update_destination_phis): Copy the threading path.
6553 (ssa_fix_duplicate_block_edges): Get edges and block types from the
6554 jump threading path.
6555 (ssa_redirect_edges): Get edges and block types from the jump threading
6556 path. Free the path vector.
6557 (thread_block): Get edges from the jump threading path. Look at the
6558 entire path to see if we thread to a loop exit. If we cancel a jump
6559 thread request, then free the path vector.
6560 (thread_single_edge): Get edges and block types from the jump threading
6561 path. Free the path vector.
6562 (thread_through_loop_header): Get edges and block types from the jump
6563 threading path. Free the path vector.
6564 (mark_threaded_blocks): Iterate over the vector of paths and store
6565 the path on the appropriate edge. Get edges and block types from the
6566 jump threading path.
6567 (mark_threaded_blocks): Get edges and block types from the jump
6568 threading path. Free the path vector.
6569 (thread_through_all_blocks): Use the vector of paths rather than
6570 a vector of 3-edge sets.
6571 (register_jump_thread): Accept pointer to a path vector rather
6572 than the path vector itself. Store the path vector for later use.
6575 2013-10-01 Jakub Jelinek <jakub@redhat.com>
6576 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6579 * config/s390/s390.c (s390_split_branches): Modify check for table
6581 (s390_chunkify_start): Rearrange table jump insn check in order to
6582 deal with compare and branch insns correctly.
6584 2013-10-01 Kugan Vivekanandarajah <kuganv@linaro.org>
6588 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
6589 * config/arm/arm.md (arm_ashldi3_1bit): define_insn into
6590 define_insn_and_split.
6591 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
6592 (shiftsi3_compare): New pattern.
6594 * config/arm/unspecs.md (UNSPEC_RRX): New.
6596 2013-10-01 Alan Modra <amodra@gmail.com>
6598 * stmt.c (expand_asm_operands): Revert part of 2013-09-24 special
6599 casing inout operands.
6601 2013-10-01 Richard Biener <rguenther@suse.de>
6603 PR tree-optimization/58553
6604 * tree-loop-distribution.c (struct partition_s): Add niter member.
6605 (classify_partition): Populate niter member for the partition
6606 and properly identify whether the relevant store happens before
6607 or after the loop exit.
6608 (generate_memset_builtin): Use niter member from the partition.
6609 (generate_memcpy_builtin): Likewise.
6611 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
6613 * vec.h (vec_prefix, vec): Prefix member names with "m_".
6614 * vec.c (vec_prefix::calculate_allocation): Update accordingly.
6616 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
6618 * basic-block.h (edge_list): Prefix member names with "m_".
6619 * context.h (context): Likewise.
6620 * domwalk.h (dom_walker): Likewise.
6621 * gengtype-state.c (s_expr_writer, state_writer): Likewise.
6622 * graphite-sese-to-poly.c (sese_dom_walker): Likewise.
6623 * hash-table.h (hash_table): Likewise.
6624 * machmode.h (bit_field_mode_iterator): Likewise.
6625 * pass_manager.h (pass_list): Likewise.
6626 * tree-into-ssa.c (mark_def_dom_walker): Likewise.
6627 * tree-pass.h (pass_data): Likewise.
6628 * tree-ssa-dom.c (dom_opt_dom_walker): Likewise.
6629 * tree-ssa-phiopt.c (nontrapping_dom_walker): Likewise,
6630 * tree-ssa-uncprop.c (uncprop_dom_walker): Likewise.
6631 * asan.c (pass_data_asan): Update accordingly.
6632 * cfganal.c (control_dependences::find_control_dependence): Likewise.
6633 (control_dependences::control_dependences): Likewise.
6634 (control_dependences::~control_dependences): Likewise.
6635 (control_dependences::~control_dependences): Likewise.
6636 (control_dependences::get_edges_dependent_on): Likewise.
6637 * cgraphbuild.c (pass_data_rebuild_cgraph_edges::clone): Likewise.
6638 (pass_data_remove_cgraph_callee_edges::clone): Likewise.
6639 * context.c (gcc::context::context): Likewise.
6640 * cprop.c (pass_rtl_cprop::clone): Likewise.
6641 * domwalk.c (dom_walker::walk): Likewise.
6642 * ipa-inline-analysis.c (pass_inline_parameters::clone): Likewise.
6643 * ipa-pure-const.c (pass_local_pure_const::clone): Likewise.
6644 * mode-switching.c (pass_mode_switching::clone): Likewise.
6645 * passes.c (opt_pass::opt_pass): Likewise.
6646 (pass_manager::pass_manager): Likewise.
6647 * predict.c (pass_strip_predict_hints::clone): Likewise.
6648 * recog.c (pass_data pass_data_peephole2::clone): Likewise.
6649 (pass_split_all_insns::clone): Likewise.
6650 * stor-layout.c (bit_field_mode_iterator::bit_field_mode_iterator):
6652 (bit_field_mode_iterator::next_mode): Likewise.
6653 (bit_field_mode_iterator::prefer_smaller_modes): Likewise.
6654 * tree-cfg.c (pass_split_crit_edges::clone): Likewise.
6655 * tree-cfgcleanup.c (pass_merge_phi::clone): Likewise.
6656 * tree-complex.c (pass_lower_complex::clone): Likewise.
6657 * tree-eh.c (pass_cleanup_eh::clone): Likewise.
6658 * tree-object-size.c (pass_object_sizes::clone): Likewise.
6659 * tree-optimize.c (pass_fixup_cfg::clone): Likewise.
6660 * tree-ssa-ccp.c (pass_data_ccp::clone): Likewise.
6661 (pass_fold_builtins::clone): Likewise.
6662 * tree-ssa-copy.c (pass_data_copy_prop::clone): Likewise.
6663 * tree-ssa-copyrename.c (pass_rename_ssa_copies::clone): Likewise.
6664 * tree-ssa-dce.c (pass_dce::clone, pass_dce_loop::clone): Likewise.
6665 (pass_cd_dce::clone): Likewise.
6666 * tree-ssa-dom.c (pass_dominator::clone): Likewise.
6667 (pass_phi_only_cprop::clone): Likewise.
6668 * tree-ssa-dse.c (pass_dse::clone): Likewise.
6669 * tree-ssa-forwprop.c (pass_forwprop::clone): Likewise.
6670 * tree-ssa-loop.c (pass_lim::clone): Likewise.
6671 * tree-ssa-phiopt.c (pass_phiopt::clone): Likewise.
6672 * tree-ssa-pre.c (pass_fre::clone): Likewise.
6673 * tree-ssa-reassoc.c (pass_reassoc::clone): Likewise.
6674 * tree-ssa-uninit.c (pass_late_warn_uninitialized::clone): Likewise.
6675 * tree-tailcall.c (pass_tail_recursion::clone): Likewise.
6676 * tree-vect-generic.c (pass_lower_vector_ssa::clone): Likewise.
6677 * tree-vrp.c (pass_vrp::clone): Likewise.
6678 * tsan.c (pass_tsan::clone): Likewise.
6680 2013-09-30 Jakub Jelinek <jakub@redhat.com>
6683 * fold-const.c (tree_unary_nonnegative_warnv_p): Use
6684 INTEGRAL_TYPE_P (t) instead of TREE_CODE (t) == INTEGER_TYPE.
6687 * fold-const.c (fold_ternary_loc): For A < 0 : <sign bit of A> : 0
6688 optimization, punt if sign_bit_p looked through any zero extension.
6690 2013-09-30 Teresa Johnson <tejohnson@google.com>
6692 * tree-ssa-threadupdate.c (ssa_fix_duplicate_block_edges):
6693 Update redirected out edge count in joiner case.
6694 (ssa_redirect_edges): Common the joiner and non-joiner cases
6695 so that joiner case gets profile updates.
6697 2013-09-30 Richard Biener <rguenther@suse.de>
6699 PR tree-optimization/58554
6700 * tree-loop-distribution.c (classify_partition): Require
6701 unconditionally executed stores for memcpy and memset recognition.
6702 (tree_loop_distribution): Calculate dominance info.
6704 2013-09-30 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
6706 * config/aarch64/aarch64.h (MCOUNT_NAME): Define.
6707 (NO_PROFILE_COUNTERS): Likewise.
6708 (PROFILE_HOOK): Likewise.
6709 (FUNCTION_PROFILER): Likewise.
6710 * config/aarch64/aarch64.c (aarch64_function_profiler): Remove.
6712 2013-09-30 Iain Sandoe <iain@codesourcery.com>
6714 * config/rs6000/darwin.md (load_macho_picbase_si): Wrap machopic
6715 calls and defines in TARGET_MACHO conditional.
6716 (load_macho_picbase_di): Likewise.
6717 (reload_macho_picbase): Likewise.
6718 (reload_macho_picbase_si): Likewise.
6719 (reload_macho_picbase_di): Likewise.
6720 (nonlocal_goto_receiver): Likewise.
6722 2013-09-30 Nick Clifton <nickc@redhat.com>
6724 * config/msp430/msp430.c (msp430x_names): New array. Lists MCUs
6725 that use the MSP430X ISA.
6726 (msp430_option_override): Scan -mmcu command line option for any
6727 MCU name that supports the MSP430X ISA.
6728 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add matches for known
6729 -mmcu options which enable the MSP430X ISA.
6731 2013-09-30 Richard Biener <rguenther@suse.de>
6734 * tree-cfg.c (make_abnormal_goto_edges): Skip debug statements
6735 before looking for setjmp-like calls.
6737 2013-09-29 Iain Sandoe <iain@codesourcery.com>
6740 * config/darwin-protos.h (machopic_get_function_picbase): New.
6741 * config/darwin.c (machopic_get_function_picbase): New.
6742 * config/rs6000/darwin.md (load_macho_picbase_si): Update picbase
6743 label for a new func. (load_macho_picbase_di): Likewise.
6744 (reload_macho_picbase): New expand.
6745 (reload_macho_picbase_si): New insn.
6746 (reload_macho_picbase_di): New insn.
6747 (nonlocal_goto_receiver): New define and split.
6748 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_RELD_MPIC.
6749 (unspecv enum): Add UNSPECV_NLGR.
6751 2013-09-29 Iain Sandoe <iain@codesourcery.com>
6753 * config/rs6000/rs6000.c (rs6000_init_dwarf_reg_sizes_extra): Ensure
6754 that altivec registers are correctly sized on Darwin.
6756 2013-09-29 Iain Sandoe <iain@codesourcery.com>
6758 * config/t-darwin (darwin.o, darwin-c.o, darwin-f.o,
6759 darwin-driver.o): Use COMPILE and POSTCOMPILE.
6760 * config/x-darwin (host-darwin.o): Likewise.
6761 * config/i386/x-darwin (host-i386-darwin.o): Likewise.
6762 * config/rs6000/x-darwin (host-ppc-darwin.o): Likewise.
6763 * config/rs6000/x-darwin64 (host-ppc64-darwin.o): Likewise.
6765 2013-09-29 Uros Bizjak <ubizjak@gmail.com>
6767 * doc/invoke.texi: Fix usage of @tie{} command.
6769 2013-09-29 Eric Botcazou <ebotcazou@adacore.com>
6771 * config/sparc/sync.md: Add peephole for consecutive memory barriers.
6773 2013-09-28 Jan Hubicka <jh@suse.cz>
6775 * config/i386/x86-tune.def: Add documentation for each of the options;
6778 2013-09-28 Jan Hubicka <jh@suse.cz>
6780 * x86-tune.def (X86_TUNE_SSE_UNALIGNED_LOAD_OPTIMAL): Enable for
6782 (X86_TUNE_SSE_UNALIGNED_STORE_OPTIMAL): Likewise.
6783 (X86_TUNE_FOUR_JUMP_LIMIT): Drop for generic and buldozer.
6784 (X86_TUNE_PAD_RETURNS): Drop for buldozer chips.
6785 (X86_TUNE_AVOID_VECTOR_DECODE): Drop for generic.
6786 (X86_TUNE_REASSOC_FP_TO_PARALLEL): Enable for generic.
6788 2013-09-28 Richard Sandiford <rdsandiford@googlemail.com>
6790 * alloc-pool.c, asan.c, auto-inc-dec.c, basic-block.h, bb-reorder.c,
6791 bitmap.c, bitmap.h, bt-load.c, builtins.c, calls.c, cfgcleanup.c,
6792 cfgexpand.c, cfghooks.c, cfgloop.c, cfgloopmanip.c, cfgrtl.c, cgraph.c,
6793 cgraph.h, cgraphbuild.c, cgraphclones.c, cgraphunit.c, collect2.c,
6794 combine-stack-adj.c, combine.c, compare-elim.c, context.c, context.h,
6795 cprop.c, cse.c, cselib.c, dbxout.c, dce.c, defaults.h, df-core.c,
6796 df-problems.c, df-scan.c, df.h, diagnostic.c, double-int.c, dse.c,
6797 dumpfile.c, dwarf2asm.c, dwarf2cfi.c, dwarf2out.c, emit-rtl.c,
6798 errors.c, except.c, expmed.c, expr.c, file-find.c, final.c,
6799 fixed-value.c, fold-const.c, function.c, fwprop.c, gcc-ar.c, gcc.c,
6800 gcov-io.c, gcov-io.h, gcov.c, gcse.c, genattr-common.c, genattr.c,
6801 genattrtab.c, genautomata.c, genconfig.c, genemit.c, genextract.c,
6802 genflags.c, gengenrtl.c, gengtype-state.c, gengtype.c, genmodes.c,
6803 genopinit.c, genoutput.c, genpeep.c, genpreds.c, genrecog.c,
6804 gensupport.c, ggc-common.c, ggc-page.c, gimple-fold.c, gimple-low.c,
6805 gimple-pretty-print.c, gimple-ssa-strength-reduction.c, gimple.c,
6806 gimple.h, godump.c, graphite-clast-to-gimple.c,
6807 graphite-optimize-isl.c, graphite-poly.h, graphite-sese-to-poly.c,
6808 graphite.c, haifa-sched.c, hash-table.c, hash-table.h, hwint.c,
6809 hwint.h, ifcvt.c, incpath.c, init-regs.c, input.h, intl.c, intl.h,
6810 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
6811 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c,
6812 ipa-utils.c, ipa.c, ira-build.c, ira.c, jump.c, loop-doloop.c,
6813 loop-init.c, loop-invariant.c, loop-iv.c, lower-subreg.c, lto-cgraph.c,
6814 lto-streamer-in.c, lto-streamer-out.c, lto-wrapper.c, mcf.c,
6815 mode-switching.c, modulo-sched.c, omp-low.c, optabs.c, opts.c,
6816 pass_manager.h, passes.c, plugin.c, postreload-gcse.c, postreload.c,
6817 predict.c, prefix.c, pretty-print.c, print-rtl.c, print-tree.c,
6818 profile.c, read-md.c, real.c, real.h, recog.c, ree.c, reg-stack.c,
6819 regcprop.c, reginfo.c, regmove.c, regrename.c, regs.h, regstat.c,
6820 reload1.c, reorg.c, rtl.c, rtl.h, rtlanal.c, sbitmap.c, sched-rgn.c,
6821 sdbout.c, sel-sched-ir.c, sel-sched.c, sparseset.c, stack-ptr-mod.c,
6822 statistics.c, stmt.c, stor-layout.c, store-motion.c, streamer-hooks.h,
6823 system.h, target-hooks-macros.h, targhooks.c, targhooks.h, toplev.c,
6824 tracer.c, trans-mem.c, tree-browser.c, tree-call-cdce.c, tree-cfg.c,
6825 tree-cfgcleanup.c, tree-complex.c, tree-data-ref.c, tree-data-ref.h,
6826 tree-eh.c, tree-emutls.c, tree-flow.h, tree-if-conv.c, tree-into-ssa.c,
6827 tree-iterator.c, tree-loop-distribution.c, tree-mudflap.c,
6828 tree-nested.c, tree-nomudflap.c, tree-nrv.c, tree-object-size.c,
6829 tree-optimize.c, tree-pass.h, tree-pretty-print.c, tree-profile.c,
6830 tree-scalar-evolution.c, tree-sra.c, tree-ssa-ccp.c,
6831 tree-ssa-coalesce.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
6832 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
6833 tree-ssa-ifcombine.c, tree-ssa-live.c, tree-ssa-loop-ch.c,
6834 tree-ssa-loop-im.c, tree-ssa-loop-ivopts.c, tree-ssa-loop-prefetch.c,
6835 tree-ssa-loop.c, tree-ssa-math-opts.c, tree-ssa-operands.c,
6836 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
6837 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
6838 tree-ssa-structalias.c, tree-ssa-threadedge.c, tree-ssa-threadupdate.c,
6839 tree-ssa-uncprop.c, tree-ssa-uninit.c, tree-ssa.c, tree-ssanames.c,
6840 tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
6841 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
6842 tree-vect-stmts.c, tree-vectorizer.c, tree-vectorizer.h, tree-vrp.c,
6843 tree.c, tree.h, tsan.c, tsystem.h, value-prof.c, var-tracking.c,
6844 varasm.c, vec.h, vmsdbgout.c, vtable-verify.c, web.c: Add missing
6845 whitespace before "(".
6847 2013-09-28 Sandra Loosemore <sandra@codesourcery.com>
6849 * expr.h (extract_bit_field): Remove packedp parameter.
6850 * expmed.c (extract_fixed_bit_field): Remove packedp parameter
6851 from forward declaration.
6852 (store_split_bit_field): Remove packedp arg from calls to
6853 extract_fixed_bit_field.
6854 (extract_bit_field_1): Remove packedp parameter and packedp
6855 argument from recursive calls and calls to extract_fixed_bit_field.
6856 (extract_bit_field): Remove packedp parameter and corresponding
6857 arg to extract_bit_field_1.
6858 (extract_fixed_bit_field): Remove packedp parameter. Remove code
6860 (extract_split_bit_field): Remove packedp arg from call to
6861 extract_fixed_bit_field.
6862 * expr.c (emit_group_load_1): Adjust calls to extract_bit_field.
6863 (copy_blkmode_from_reg): Likewise.
6864 (copy_blkmode_to_reg): Likewise.
6865 (read_complex_part): Likewise.
6866 (store_field): Likewise.
6867 (expand_expr_real_1): Likewise.
6868 * calls.c (store_unaligned_arguments_into_pseudos): Adjust call
6869 to extract_bit_field.
6870 * config/tilegx/tilegx.c (tilegx_expand_unaligned_load): Adjust
6871 call to extract_bit_field.
6872 * config/tilepro/tilepro.c (tilepro_expand_unaligned_load): Adjust
6873 call to extract_bit_field.
6874 * doc/invoke.texi (Code Gen Options): Remove mention of warnings
6875 and special packedp behavior from -fstrict-volatile-bitfields
6878 2013-09-27 Jan-Benedict Glaw <jbglaw@lug-owl.de>
6880 * lra-eliminations.c (init_elim_table): Guard value_p.
6882 2013-09-27 Michael Meissner <meissner@linux.vnet.ibm.com>
6884 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Allow
6885 DFmode, DImode, and SFmode in the upper VSX registers based on the
6886 -mupper-regs-{df,sf} flags. Fix wu constraint to be ALTIVEC_REGS
6887 if -mpower8-vector. Combine -mvsx-timode handling with the rest
6888 of the VSX register handling.
6890 * config/rs6000/rs6000.md (f32_lv): Use %x0 for VSX regsters.
6892 (zero_extendsidi2_lfiwzx): Add support for loading into the
6893 Altivec registers with -mpower8-vector. Use wu/wv constraints to
6894 only do VSX memory options on Altivec registers.
6895 (extendsidi2_lfiwax): Likewise.
6896 (extendsfdf2_fpr): Likewise.
6897 (mov<mode>_hardfloat, SF/SD modes): Likewise.
6898 (mov<mode>_hardfloat32, DF/DD modes): Likewise.
6899 (mov<mode>_hardfloat64, DF/DD modes): Likewise.
6900 (movdi_internal64): Likewise.
6902 2013-09-27 Xinliang David Li <davidxl@google.com>
6904 * opts.c (finish_options): Adjust parameters
6905 according to vect cost model.
6906 (common_handle_option): Set dynamic vect cost
6908 targhooks.c (default_add_stmt_cost): Compute stmt cost
6910 * tree-vect-loop.c (vect_estimate_min_profitable_iters):
6911 Use helper function.
6912 * tree-vectorizer.h (unlimited_cost_model): New function.
6913 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use helper function.
6914 * tree-vect-data-refs.c (vect_peeling_hash_insert): Use helper
6916 (vect_enhance_data_refs_alignment): Ditto.
6917 * flag-types.h: New enum.
6918 * common/config/i386/i386-common.c (ix86_option_init_struct):
6919 No need to initialize vect_cost_model flag.
6920 * config/i386/i386.c (ix86_add_stmt_cost): Compute stmt cost
6923 2013-09-27 Diego Novillo <dnovillo@google.com>
6925 * gimple.h (enum ssa_mode): Remove.
6927 2013-09-27 Paulo Matos <pmatos@broadcom.com>
6929 * cfgloop.h (number_of_loops): Fix typo in check for null.
6931 2013-09-27 Jakub Jelinek <jakub@redhat.com>
6934 * tree-cfg.c (move_sese_region_to_fn): Also move loops that
6935 are children of outermost saved_cfun's loop, and set it up to
6936 be moved to dest_cfun's outermost loop. Fix up num_nodes adjustments
6937 if loop != loop0 and SESE region contains bbs that belong to loop0.
6939 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
6941 * rtlanal.c (must_be_base_p, must_be_index_p): Delete.
6942 (binary_scale_code_p, get_base_term, get_index_term): New functions.
6943 (set_address_segment, set_address_base, set_address_index)
6944 (set_address_disp): Accept the argument unconditionally.
6945 (baseness): Remove must_be_base_p and must_be_index_p checks.
6946 (decompose_normal_address): Classify as much as possible in the
6949 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
6951 * cse.c (count_reg_usage): Handle INT_LIST.
6952 * lra-eliminations.c (lra_eliminate_regs_1): Likewise.
6953 * reginfo.c (reg_scan_mark_refs): Likewise.
6954 * reload1.c (eliminate_regs_1): Likewise.
6956 2013-09-27 Iain Sandoe <iain@codesourcery.com>
6959 * rtlanal.c (lsb_bitfield_op_p): Make both parts of the comparison
6962 2013-09-27 Richard Biener <rguenther@suse.de>
6964 PR tree-optimization/58459
6965 * tree-ssa-forwprop.c (forward_propagate_addr_expr): Remove
6966 restriction not propagating into loops.
6968 2013-09-26 Florian Weimer <fw@deneb.enyo.de>
6970 * tree-ssa.h (walk_use_def_chains_fn, walk_use_def_chains): Delete.
6971 * tree-ssa.c (walk_use_def_chains_1, walk_use_def_chains): Delete.
6972 * doc/tree-ssa.texi (Walking use-def chains): Delete.
6974 2013-09-26 Richard Biener <rguenther@suse.de>
6976 * tree-into-ssa.c (rewrite_into_ssa): Make more SSA names to anonymous.
6978 2013-09-26 Richard Biener <rguenther@suse.de>
6980 * alias.h (component_uses_parent_alias_set): Rename to ...
6981 (component_uses_parent_alias_set_from): ... this.
6982 * alias.c (component_uses_parent_alias_set): Rename to ...
6983 (component_uses_parent_alias_set_from): ... this and return
6985 (reference_alias_ptr_type_1): Use the result from
6986 component_uses_parent_alias_set_from instead of stripping
6987 components one at a time.
6988 * emit-rtl.c (set_mem_attributes_minus_bitpos): Adjust.
6990 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
6992 * tree-ssa-live.h (find_replaceable_exprs, dump_replaceable_exprs):
6993 Move prototypes to...
6994 * tree-ssa-ter.h: New File. Move prototypes here.
6995 * tree-flow.h (stmt_is_replaceable_p): Remove prototype.
6996 * tree-outof-ssa.h: New. Rename ssaexpand.h, include tree-ssa-ter.h.
6997 * tree-outof-ssa.c (ssa_is_replaceable_p): New. Refactor common bits
6998 from is_replaceable_p.
6999 * tree-ssa-ter.c (is_replaceable_p, stmt_is_replaceable_p): Delete.
7000 (ter_is_replaceable_p): New. Use new refactored ssa_is_replaceable_p.
7001 (process_replaceable): Use ter_is_replaceable_p.
7002 (find_replaceable_in_bb): Use ter_is_replaceable_p.
7003 * expr.c (stmt_is_replaceable_p): Relocate from tree-ssa-ter.c. Use
7004 newly refactored ssa_is_replaceable_p.
7005 * cfgexpand.c: Include tree-outof-ssa.h.
7006 * ssaexpand.h: Delete.
7008 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
7010 * gimple.c (gimple_replace_lhs): Move to tree-ssa.c and rename.
7011 (struct count_ptr_d, count_ptr_derefs, count_uses_and_derefs): Move to
7013 (create_gimple_tmp): Delete.
7014 (get_expr_type, build_assign, build_type_cast): Move to...
7015 * gimple-builder.c: New File.
7016 (get_expr_type): Relocate from gimple.c.
7017 (build_assign, build_type_cast): Change to only create ssanames.
7018 * gimple.h: Move prototypes to...
7019 * gimple-builder.h: New File. Here.
7020 * tree-ssa.h: And here.
7021 * tree-ssa.c (struct count_ptr_d, count_ptr_derefs,
7022 count_uses_and_derefs): Relocate from gimple.c.
7023 (gimple_replace_ssa_lhs): Renamed gimple_replace_ssa from gimple.c
7024 * tree-ssa-reassoc.c (repropagate_negates): Use gimple_replace_ssa_lhs.
7025 * tree-ssa-math-opts (execute_cse_reciprocals): Use
7026 gimple_replace_ssa_lhs.
7027 * asan.c: Include gimple-builder.h.
7028 * Makefile.in: Add gimple-builder.o.
7030 2013-09-26 Richard Biener <rguenther@suse.de>
7032 * tree-ssa-live.c (var_map_base_init): Handle SSA names with
7033 DECL_IGNORED_P base VAR_DECLs like anonymous SSA names.
7034 (loe_visit_block): Use gcc_checking_assert.
7035 * tree-ssa-coalesce.c (create_outofssa_var_map): Use
7036 gimple_assign_ssa_name_copy_p.
7037 (gimple_can_coalesce_p): Adjust according to the var_map_base_init
7040 2013-09-26 David Edelsohn <dje.gcc@gmail.com>
7042 * config/rs6000/t-rs6000 (rs6000.o): Remove.
7043 (rs6000-c.o): Use COMPILE and POSTCOMPILE.
7045 2013-09-26 Richard Biener <rguenther@suse.de>
7047 PR tree-optimization/58539
7048 * tree-vect-loop.c (vect_create_epilog_for_reduction): Honor
7049 the fact that debug statements are not taking part in loop-closed
7052 2013-09-26 Nick Clifton <nickc@redhat.com>
7054 * config/msp430/msp430.c (msp430_expand_epilogue): Fix compile
7055 time warning message.
7056 (msp430_print_operand_raw): Delete unused letter parameter.
7057 (TARGET_PRINT_OPERAND_ADDRESS): Define.
7058 (msp430_print_operand_address): New function.
7059 (msp430_print_operand): Move address printing code from here to
7061 * config/msp430/msp430.md (movsipsi2): Add comment in generated
7063 (zero_extendpsisi2): Likewise.
7064 (extendpsisi2): New pattern.
7065 (andneghi3): New pattern.
7067 2013-09-26 Yvan Roux <yvan.roux@linaro.org>
7069 * config/aarch64/aarch64.opt (mlra): New option.
7070 * config/aarch64/aarch64.c (aarch64_lra_p): New function.
7071 (TARGET_LRA_P): Define.
7073 2013-09-26 Eric Botcazou <ebotcazou@adacore.com>
7075 * expr.c (expand_assignment): Remove obsolete comment.
7077 2013-09-25 Jeff Law <law@redhat.com>
7079 * tree-flow.h (thread_through_all_blocks): Prototype moved into
7080 tree-ssa-threadupdate.h.
7081 (register_jump_thread): Similarly.
7082 * tree-ssa-threadupdate.h: New header file.
7083 * tree-ssa-dom.c: Include tree-ssa-threadupdate.h.
7084 * tree-vrp.c: Likewise.
7085 * tree-ssa-threadedge.c: Include tree-ssa-threadupdate.h.
7086 (thread_around_empty_blocks): Change type of path vector argument to
7087 an edge,type pair from just an edge. Initialize both elements when
7088 appending to a jump threading path. Tweak references to elements
7090 (thread_across_edge): Similarly. Release memory for the elements
7092 * tree-ssa-threadupdate.c: Include tree-ssa-threadupdate.h.
7093 (dump_jump_thread_path): New function broken out from
7094 register_jump_thread.
7095 (register_jump_thread): Use dump_jump_thread_path. Change type of
7096 path vector entries. Search the path for NULL edges and dump
7097 the path if one is found. Tweak the conversion of path to 3-edge
7098 form to use the block copy type information embedded in the path.
7100 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
7102 * lra.c (update_inc_notes): Remove all REG_DEAD and REG_UNUSED notes.
7104 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
7105 Vladimir Makarov <vmakarov@redhat.com>
7107 * rtlanal.c (lsb_bitfield_op_p): New predicate for bitfield operations
7108 from the least significant bit.
7109 (strip_address_mutations): Add bitfield operations handling.
7110 (must_be_index_p): Add shifting and rotate operations handling.
7111 (set_address_base): Use must_be_base_p predicate.
7112 (set_address_index): Use must_be_index_p predicate.
7114 2013-09-25 Alexander Ivchenko <alexander.ivchenko@intel.com>
7115 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
7116 Sergey Lega <sergey.s.lega@intel.com>
7117 Anna Tikhonova <anna.tikhonova@intel.com>
7118 Ilya Tocar <ilya.tocar@intel.com>
7119 Andrey Turetskiy <andrey.turetskiy@intel.com>
7120 Ilya Verbin <ilya.verbin@intel.com>
7121 Kirill Yukhin <kirill.yukhin@intel.com>
7122 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
7124 * config/i386/i386.c (ix86_avx256_split_vector_move_misalign):
7126 (ix86_expand_vector_move_misalign): Support new unaligned load and
7127 stores and use new names.
7128 (CODE_FOR_sse2_storedqu): Rename to ...
7129 (CODE_FOR_sse2_storedquv16qi): ... this.
7130 (CODE_FOR_sse2_loaddqu): Rename to ...
7131 (CODE_FOR_sse2_loaddquv16qi): ... this.
7132 (CODE_FOR_avx_loaddqu256): Rename to ...
7133 (CODE_FOR_avx_loaddquv32qi): ... this.
7134 (CODE_FOR_avx_storedqu256): Rename to ...
7135 (CODE_FOR_avx_storedquv32qi): ... this.
7136 * config/i386/i386.md (fpint_logic): New.
7137 * config/i386/sse.md (VMOVE): Extend for AVX512.
7141 (VI_UNALIGNED_LOADSTORE): Ditto.
7142 (sse2_avx_avx512f): Ditto.
7143 (sse2_avx2): Extend for AVX512.
7144 (sse4_1_avx2): Ditto.
7145 (avx2_avx512f): New.
7146 (sse): Extend for AVX512.
7149 (avxsizesuffix): Ditto.
7150 (sseintvecmode): Ditto.
7152 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Ditto.
7153 (<sse>_storeu<ssemodesuffix><avxsizesuffix>): Ditto.
7154 (<sse2>_loaddqu<avxsizesuffix>): Extend for AVX512 and rename to ...
7155 (<sse2_avx_avx512f>_loaddqu<mode>): ... this.
7156 (<sse2>_storedqu<avxsizesuffix>): Extend for AVX512 and rename to ...
7157 (<sse2_avx_avx512f>_storedqu<mode): ... this.
7158 (<sse>_movnt<mode>): Replace constraint "x" with "v".
7159 (STORENT_MODE): Extend for AVX512.
7160 (*absneg<mode>2): Replace constraint "x" with "v".
7161 (*mul<mode>3): Ditto.
7162 (*ieee_smin<mode>3): Ditto.
7163 (*ieee_smax<mode>3): Ditto.
7164 (avx_cmp<mode>3): Replace VF with VF_128_256.
7165 (*<sse>_maskcmp<mode>3_comm): Ditto.
7166 (<sse>_maskcmp<mode>3): Ditto.
7167 (<sse>_andnot<mode>3): Extend for AVX512.
7168 (<code><mode>3, anylogic): Replace VF with VF_128_256.
7169 (<code><mode>3, fpint_logic): New.
7170 (*<code><mode>3): Extend for AVX512.
7171 (avx512flogicsuff): New.
7172 (avx512f_<logic><mode>): Ditto.
7173 (<sse>_movmsk<ssemodesuffix><avxsizesuffix>): Replace VF with
7175 (<sse4_1>_blend<ssemodesuffix><avxsizesuffix>): Ditto.
7176 (<sse4_1>_blendv<ssemodesuffix><avxsizesuffix>): Ditto.
7177 (<sse4_1>_dp<ssemodesuffix><avxsizesuffix>): Ditto.
7178 (avx_vtest<ssemodesuffix><avxsizesuffix>): Ditto.
7179 (<sse4_1>_round<ssemodesuffix><avxsizesuffix>): Ditto.
7180 (xop_vpermil2<mode>3): Ditto.
7181 (*avx_vpermilp<mode>): Extend for AVX512 and rename to ...
7182 (*<sse2_avx_avx512f>_vpermilp<mode>): ... this.
7183 (avx_vpermilvar<mode>3): Extend for AVX512 and rename to ...
7184 (<sse2_avx_avx512f>_vpermilvar<mode>3): ... this.
7186 2013-09-25 Tom Tromey <tromey@redhat.com>
7188 * Makefile.in (PARTITION_H, LTO_SYMTAB_H, COMMON_TARGET_DEF_H)
7189 (RTL_ERROR_H, TRANS_MEM_H, COVERAGE_H, DEMANGLE_H, ALIAS_H)
7190 (SCHED_INT_H, SEL_SCHED_IR_H, SEL_SCHED_DUMP_H, VALTRACK_H, DDG_H)
7191 (GGC_INTERNAL_H, DECNUM_H, BACKTRACE_H, MKDEPS_H, TREE_HASHER_H)
7192 (TREE_SSA_LIVE_H, SSAEXPAND_H, DWARF2OUT_H, SCEV_H, OMEGA_H)
7193 (TREE_DATA_REF_H, IRA_INT_H, LRA_INT_H, DBGCNT_H, DATA_STREAMER_H)
7194 (GIMPLE_STREAMER_H, TREE_STREAMER_H, STREAMER_HOOKS_H)
7195 (TREE_VECTORIZER_H, IPA_INLINE_H, GSTAB_H, LIBFUNCS_H)
7196 (GRAPHITE_HTAB_H): Remove.
7198 2013-09-25 Tom Tromey <tromey@redhat.com>
7200 * config/mcore/t-mcore (CROSS_FLOAT_H): Remove.
7202 2013-09-25 Tom Tromey <tromey@redhat.com>
7204 * config/t-glibc (glibc-c.o): Use COMPILE and POSTCOMPILE.
7206 2013-09-25 Tom Tromey <tromey@redhat.com>
7208 * config/i386/t-i386 (i386.o): Remove.
7209 (i386-c.o): Use COMPILE and POSTCOMPILE.
7211 2013-09-25 Tom Tromey <tromey@redhat.com>
7213 * Makefile.in ($(out_object_file)): Use COMPILE and POSTCOMPILE.
7215 2013-09-25 Tom Tromey <tromey@redhat.com>
7217 * Makefile.in (graph.o, sbitmap.o, sparseset.o, gcc-ar.o)
7218 (gcc-ranlib.o, gcc-nm.o, collect2.o, collect2-aix.o, tlink.o)
7219 (lto-wrapper.o, default-c.o, attribs.o, incpath.o, prefix.o)
7220 (gcc.o, options.o, options-save.o, version.o, gtype-desc.o)
7221 (trans-mem.o, ggc-common.o, ggc-page.o, ggc-none.o, stringpool.o)
7222 (convert.o, double-int.o, lto-compress.o, data-streamer-in.o)
7223 (data-streamer-out.o, data-streamer.o, gimple-streamer-in.o)
7224 (gimple-streamer-out.o, tree-streamer.o, tree-streamer-in.o)
7225 (tree-streamer-out.o, streamer-hooks.o, lto-cgraph.o)
7226 (lto-streamer-in.o, lto-streamer-out.o, lto-section-in.o)
7227 (lto-section-out.o, lto-opts.o, lto-streamer.o, langhooks.o)
7228 (test-dump.o, tree.o, tree-dump.o, tree-inline.o, print-tree.o)
7229 (stor-layout.o, asan.o, tsan.o, ubsan.o, tree-ssa-tail-merge.o)
7230 (tree-ssa-structalias.o, tree-ssa-uninit.o, tree-ssa.o)
7231 (tree-into-ssa.o, tree-ssa-ter.o, tree-ssa-coalesce.o)
7232 (tree-outof-ssa.o, tree-ssa-dse.o, tree-ssa-forwprop.o)
7233 (tree-ssa-phiprop.o, tree-ssa-ifcombine.o, tree-ssa-phiopt.o)
7234 (tree-nrv.o, tree-ssa-copy.o, tree-ssa-propagate.o)
7235 (tree-ssa-dom.o, tree-ssa-uncprop.o, tree-ssa-threadedge.o)
7236 (tree-ssa-threadupdate.o, tree-ssanames.o, tree-phinodes.o)
7237 (domwalk.o, tree-ssa-live.o, tree-ssa-copyrename.o)
7238 (tree-ssa-pre.o, tree-ssa-sccvn.o)
7239 (gimple-ssa-strength-reduction.o, tree-vrp.o, tree-cfg.o)
7240 (tree-cfgcleanup.o, tree-tailcall.o, tree-ssa-sink.o)
7241 (tree-nested.o, tree-if-conv.o, tree-iterator.o, tree-dfa.o)
7242 (tree-ssa-operands.o, tree-eh.o, tree-ssa-loop.o)
7243 (tree-ssa-loop-unswitch.o, tree-ssa-address.o)
7244 (tree-ssa-loop-niter.o, tree-ssa-loop-ivcanon.o)
7245 (tree-ssa-loop-ch.o, tree-ssa-loop-prefetch.o, tree-predcom.o)
7246 (tree-ssa-loop-ivopts.o, tree-affine.o, tree-ssa-loop-manip.o)
7247 (tree-ssa-loop-im.o, tree-ssa-math-opts.o, tree-ssa-alias.o)
7248 (tree-ssa-reassoc.o, tree-optimize.o, gimplify.o)
7249 (gimple-iterator.o, gimple-fold.o, gimple-low.o, omp-low.o)
7250 (tree-browser.o, omega.o, tree-chrec.o, tree-scalar-evolution.o)
7251 (tree-data-ref.o, sese.o, graphite.o, graphite-blocking.o)
7252 (graphite-clast-to-gimple.o, graphite-dependences.o)
7253 (graphite-interchange.o, graphite-poly.o)
7254 (graphite-scop-detection.o, graphite-sese-to-poly.o)
7255 (graphite-optimize-isl.o, tree-vect-loop.o)
7256 (tree-vect-loop-manip.o, tree-vect-patterns.o, tree-vect-slp.o)
7257 (tree-vect-stmts.o, tree-vect-data-refs.o, tree-vectorizer.o)
7258 (vtable-verify.o, tree-loop-distribution.o, tree-parloops.o)
7259 (tree-stdarg.o, tree-object-size.o, internal-fn.o, gimple.o)
7260 (gimple-pretty-print.o, tree-mudflap.o, tree-nomudflap.o)
7261 (tree-pretty-print.o, tree-diagnostic.o, fold-const.o)
7262 (diagnostic.o, diagnostic-color.o, opts.o, opts-global.o)
7263 (opts-common.o, targhooks.o, common/common-targhooks.o, input.o)
7264 (toplev.o, hwint.o, passes.o, plugin.o, main.o, host-default.o)
7265 (rtl-error.o, rtl.o, print-rtl.o, rtlanal.o, varasm.o, function.o)
7266 (statistics.o, stmt.o, except.o, expr.o, dojump.o, builtins.o)
7267 (calls.o, expmed.o, explow.o, optabs.o, dbxout.o, debug.o)
7268 (sdbout.o, dwarf2out.o, dwarf2cfi.o, dwarf2asm.o, vmsdbgout.o)
7269 (xcoffout.o, godump.o, emit-rtl.o, real.o, realmpfr.o, dfp.o)
7270 (fixed-value.o, jump.o, simplify-rtx.o, symtab.o, cgraph.o)
7271 (cgraphunit.o, cgraphclones.o, cgraphbuild.o, varpool.o, ipa.o)
7272 (ipa-profile.o, ipa-devirt.o, ipa-prop.o, ipa-ref.o, ipa-cp.o)
7273 (ipa-split.o, ipa-inline.o, ipa-inline-analysis.o)
7274 (ipa-inline-transform.o, ipa-utils.o, ipa-reference.o)
7275 (ipa-pure-const.o, coverage.o, cselib.o, cse.o, dce.o, dumpfile.o)
7276 (dse.o, fwprop.o, web.o, ree.o, cprop.o, gcse.o, store-motion.o)
7277 (resource.o, lcm.o, mode-switching.o, tree-ssa-dce.o)
7278 (tree-call-cdce.o, tree-ssa-ccp.o, tree-ssa-strlen.o, tree-sra.o)
7279 (tree-switch-conversion.o, tree-complex.o, tree-emutls.o)
7280 (tree-vect-generic.o, df-core.o, df-problems.o, df-scan.o)
7281 (regstat.o, valtrack.o, var-tracking.o, profile.o, mcf.o)
7282 (tree-profile.o, value-prof.o, loop-doloop.o, alloc-pool.o)
7283 (auto-inc-dec.o, cfg.o, cfghooks.o, cfgexpand.o, cfgrtl.o)
7284 (cfganal.o, cfgbuild.o, cfgcleanup.o, cfgloop.o, cfgloopanal.o)
7285 (graphds.o, loop-iv.o, loop-invariant.o, cfgloopmanip.o)
7286 (loop-init.o, loop-unswitch.o, loop-unroll.o, dominance.o)
7287 (et-forest.o, combine.o, reginfo.o, bitmap.o, vec.o, hash-table.o)
7288 (reload.o, reload1.o, rtlhooks.o, postreload.o, postreload-gcse.o)
7289 (caller-save.o, bt-load.o, reorg.o, alias.o, stack-ptr-mod.o)
7290 (init-regs.o, ira-build.o, ira-costs.o, ira-conflicts.o)
7291 (ira-color.o, ira-emit.o, ira-lives.o, ira.o, lra.o)
7292 (lra-assigns.o, lra-coalesce.o, lra-constraints.o)
7293 (lra-eliminations.o, lra-lives.o, lra-spills.o, regmove.o)
7294 (combine-stack-adj.o, compare-elim.o, ddg.o, modulo-sched.o)
7295 (haifa-sched.o, sched-deps.o, sched-rgn.o, sched-ebb.o)
7296 (sched-vis.o, sel-sched.o, sel-sched-dump.o, sel-sched-ir.o)
7297 (final.o, recog.o, reg-stack.o, sreal.o, predict.o, lists.o)
7298 (bb-reorder.o, tracer.o, timevar.o, regcprop.o, regrename.o)
7299 (ifcvt.o, params.o, pointer-set.o, hooks.o, pretty-print.o)
7300 (errors.o, dbgcnt.o, lower-subreg.o, target-globals.o)
7301 (hw-doloop.o, file-find.o, context.o, $(common_out_object_file))
7302 (insn-attrtab.o, insn-automata.o, insn-dfatab.o, insn-emit.o)
7303 (insn-enums.o, insn-extract.o, insn-latencytab.o, insn-modes.o)
7304 (insn-opinit.o, insn-output.o, insn-peep.o, insn-preds.o)
7305 (insn-recog.o, intl.o, cppbuiltin.o, cppdefault.o, gcov.o)
7306 (gcov-dump.o): Remove.
7307 (default-c.o): Use COMPILE and POSTCOMPILE.
7308 (CFLAGS-gcc.o): New variable.
7309 ($(common_out_object_file)): Use COMPILE and POSTCOMPILE.
7311 2013-09-25 Tom Tromey <tromey@redhat.com>
7313 * Makefile.in (c-family/cppspec.o, c-family/c-common.o)
7314 (c-family/c-cppbuiltin.o, c-family/c-dump.o, c-family/c-format.o)
7315 (c-family/c-gimplify.o, c-family/c-lex.o, c-family/c-omp.o)
7316 (c-family/c-opts.o, c-family/c-pch.o, c-family/c-ppoutput.o)
7317 (c-family/c-pragma.o, c-family/c-pretty-print.o)
7318 (c-family/c-semantics.o, c-family/c-ada-spec.o)
7319 (c-family/array-notation-common.o, c-family/stub-objc.o)
7320 (c-family/c-ubsan.o): Remove.
7322 2013-09-25 Tom Tromey <tromey@redhat.com>
7324 * Makefile.in (C_TREE_H): Reference c/c-tree.h.
7326 2013-09-25 Tom Tromey <tromey@redhat.com>
7328 * Makefile.in (DRIVER_DEFINES): Use $(and), not shell code,
7329 to add -DENABLE_SHARED_LIBGCC.
7330 (gcc.o): Don't use subshell.
7332 2013-09-25 Tom Tromey <tromey@redhat.com>
7334 * Makefile.in (OUTPUT_OPTION): Define as "-o $@".
7335 * configure.ac: Don't invoke AM_PROG_CC_C_O.
7336 (NO_MINUS_C_MINUS_O, OUTPUT_OPTION): Don't subst.
7337 * configure, config.in: Rebuild.
7339 2013-09-25 Tom Tromey <tromey@redhat.com>
7341 * Makefile.in (CCDEPMODE, DEPDIR, depcomp, COMPILE.base)
7342 (COMPILE, POSTCOMPILE): New variables.
7343 (.cc.o .c.o): Use COMPILE, POSTCOMPILE.
7344 (DEPFILES): New variable.
7345 Include ".Po" files.
7346 * configure.ac: Add checks for dependency checking.
7347 * configure, aclocal.m4: Regenerate.
7349 2013-09-25 Tom Tromey <tromey@redhat.com>
7351 * Makefile.in (ALL_HOST_BACKEND_OBJS): Add lto-wrapper.o.
7352 ($(ALL_HOST_OBJS)): Move order-only dependency to end of file.
7354 2013-09-25 Tom Tromey <tromey@redhat.com>
7356 * Makefile.in (generated_files): Add options.h,
7357 target-hooks-def.h, insn-opinit.h,
7358 common/common-target-hooks-def.h, pass-instances.def,
7359 c-family/c-target-hooks-def.h.
7361 2013-09-25 Jeff Law <law@redhat.com>
7363 * tree-ssa-threadedge.c (thread_across_edge): Use foo.last () rather
7364 than foo[foo.length () - 1] to access last member in a vec.
7365 * tree-ssa-threadupdate.c (register_jump_thread): Similarly.
7367 2013-09-25 Richard Biener <rguenther@suse.de>
7370 * tree.c (iterative_hash_expr): Remove MEM_REF special handling.
7372 2013-09-25 Jan Hubicka <jh@suse.cz>
7374 * cgraph.c (cgraph_resolve_speculation): Use semantical equivalency
7377 2013-09-25 Marek Polacek <polacek@redhat.com>
7380 * ubsan.c (ubsan_type_descriptor): Handle IDENTIFIER_NODEs
7381 when determining the type name.
7383 2013-09-24 Oleg Endo <olegendo@gcc.gnu.org>
7385 * config/sh/sh.md: Fix formatting.
7387 2013-09-24 Xinliang David Li <davidxl@google.com>
7389 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Check
7390 max peel iterations parameter.
7391 * param.def: New parameter.
7392 * doc/invoke.texi: Document New parameter.
7394 2013-09-24 Christophe Lyon <christophe.lyon@linaro.org>
7396 * gimple-pretty-print.c: Various whitespace tweaks.
7397 * tree-core.h: Likewise.
7398 * tree-pretty-print.c: Likewise.
7399 * tree-ssa-alias.c: Likewise.
7400 * tree-ssa-copy.c: Likewise.
7401 * tree-ssanames.c: Likewise.
7402 * tree-ssanames.h: Likewise.
7403 * tree-vrp.c: Likewise.
7405 2013-09-24 Alan Modra <amodra@gmail.com>
7409 * stmt.c (expand_asm_operands): Call expand_expr with EXPAND_MEMORY
7410 for output operands that disallow regs. Don't use EXPAND_WRITE on
7413 2013-09-24 Richard Biener <rguenther@suse.de>
7416 * tree.c (reference_alias_ptr_type): Move ...
7417 * alias.c (reference_alias_ptr_type): ... here and implement
7418 in terms of the new reference_alias_ptr_type_1.
7419 (ref_all_alias_ptr_type_p): New helper.
7420 (get_deref_alias_set_1): Drop flag_strict_aliasing here,
7421 use ref_all_alias_ptr_type_p.
7422 (get_deref_alias_set): Add flag_strict_aliasing check here.
7423 (reference_alias_ptr_type_1): New function, split out from ...
7424 (get_alias_set): ... here.
7425 (alias_ptr_types_compatible_p): New function.
7426 * alias.h (reference_alias_ptr_type): Declare.
7427 (alias_ptr_types_compatible_p): Likewise.
7428 * tree.h (reference_alias_ptr_type): Remove.
7429 * fold-const.c (operand_equal_p): Use alias_ptr_types_compatible_p
7430 to compare MEM_REF alias types.
7432 2013-09-24 Richard Biener <rguenther@suse.de>
7434 * tree-vrp.c (vrp_finalize): Check for SSA name presence.
7436 2013-09-23 Michael Meissner <meissner@linux.vnet.ibm.com>
7438 * config/rs6000/rs6000.c (rs6000_vector_reload): Delete, combine
7439 reload helper function arrays into a single array reg_addr.
7440 (reload_fpr_gpr): Likewise.
7441 (reload_gpr_vsx): Likewise.
7442 (reload_vsx_gpr): Likewise.
7443 (struct rs6000_reg_addr): Likewise.
7444 (reg_addr): Likewise.
7445 (rs6000_debug_reg_global): Change rs6000_vector_reload,
7446 reload_fpr_gpr, reload_gpr_vsx, reload_vsx_gpr uses to reg_addr.
7447 (rs6000_init_hard_regno_mode_ok): Likewise.
7448 (rs6000_secondary_reload_direct_move): Likewise.
7449 (rs6000_secondary_reload): Likewise.
7451 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add new
7452 constraints: wu, ww, and wy. Repurpose wv constraint added during
7453 power8 changes. Put wg constraint in alphabetical order.
7455 * config/rs6000/rs6000.opt (-mvsx-scalar-float): New debug switch
7456 for future work to add ISA 2.07 VSX single precision support.
7457 (-mvsx-scalar-double): Change default from -1 to 1, update
7458 documentation comment.
7459 (-mvsx-scalar-memory): Rename debug switch to -mupper-regs-df.
7460 (-mupper-regs-df): New debug switch to control whether DF values
7461 can go in the traditional Altivec registers.
7462 (-mupper-regs-sf): New debug switch to control whether SF values
7463 can go in the traditional Altivec registers.
7465 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print wu, ww,
7467 (rs6000_init_hard_regno_mode_ok): Use ssize_t instead of int for
7468 loop variables. Rename -mvsx-scalar-memory to -mupper-regs-df.
7469 Add new constraints, wu/ww/wy. Repurpose wv constraint.
7470 (rs6000_debug_legitimate_address_p): Print if we are running
7471 before, during, or after reload.
7472 (rs6000_secondary_reload): Add a comment.
7473 (rs6000_opt_masks): Add -mupper-regs-df, -mupper-regs-sf.
7475 * config/rs6000/constraints.md (wa constraint): Sort w<x>
7476 constraints. Update documentation string.
7477 (wd constraint): Likewise.
7478 (wf constraint): Likewise.
7479 (wg constraint): Likewise.
7480 (wn constraint): Likewise.
7481 (ws constraint): Likewise.
7482 (wt constraint): Likewise.
7483 (wx constraint): Likewise.
7484 (wz constraint): Likewise.
7485 (wu constraint): New constraint for ISA 2.07 SFmode scalar
7487 (ww constraint): Likewise.
7488 (wy constraint): Likewise.
7489 (wv constraint): Repurpose ISA 2.07 constraint that we did not use
7490 in the previous submissions.
7491 * doc/md.texi (PowerPC and IBM RS6000): Likewise.
7493 2013-09-23 Richard Sandiford <rdsandiford@googlemail.com>
7495 * doc/rtl.texi (REG_NOTES): Say that int_list can also be used.
7496 (REG_BR_PROB): Say that the probability is stored in an int_list.
7497 * reg-notes.def: Update commentary to mention INT_LIST.
7498 * rtl.def (EXPR_LIST, INSN_LIST): Capitalize comments.
7499 (INT_LIST): New rtx.
7500 * rtl.h (add_int_reg_note, add_shallow_copy_of_reg_note): Declare.
7501 * rtlanal.c (int_reg_note_p): New function.
7502 (alloc_reg_note): Assert that the note does not have an int argument.
7503 (add_int_reg_note, add_shallow_copy_of_reg_note): New functions.
7504 * combine.c (distribute_notes): Use add_shallow_copy_of_rtx.
7505 * cse.c (cse_process_notes_1): Expect REG_EQUAL to be an EXPR_LIST
7506 rather than an INSN_LIST. Handle INT_LIST.
7507 * ifcvt.c (cond_exec_process_insns): Take the probability as an int
7508 rather than an rtx. Use gen_rtx_INT_LIST to create a REG_BR_PROB note.
7509 (cond_exec_process_if_block): Use XINT to extract REG_BR_PROB values.
7510 Manipulate them as ints rather than rtxes.
7511 * reg-stack.c (subst_asm_stack_regs): Only handle EXPR_LIST notes.
7512 * regmove.c (copy_src_to_dest): Likewise.
7513 * sched-vis.c (print_insn_with_notes): Handle INT_LIST.
7515 * config/i386/winnt.c (i386_pe_seh_unwind_emit): Sink pat assignment
7516 into the cases that need it.
7517 * config/arm/arm.c (arm_unwind_emit): Likewise.
7519 * asan.c (asan_clear_shadow): Use add_int_reg_note for REG_BR_PROB.
7520 * emit-rtl.c (try_split, emit_copy_of_insn_after): Likewise.
7521 * loop-doloop.c (add_test, doloop_modify): Likewise.
7522 * loop-unswitch.c (compare_and_jump_seq): Likewise.
7523 * optabs.c (emit_cmp_and_jump_insn_1): Likewise.
7524 * predict.c (combine_predictions_for_insn): Likewise.
7525 * print-rtl.c (print_rtx): Handle INT_LIST.
7526 * config/aarch64/aarch64.c (aarch64_emit_unlikely_jump): Likewise.
7527 * config/alpha/alpha.c (emit_unlikely_jump): Likewise.
7528 * config/arm/arm.c (emit_unlikely_jump): Likewise.
7529 * config/i386/i386.c (ix86_expand_split_stack_prologue): Likewise.
7530 (ix86_split_fp_branch, predict_jump): Likewise.
7531 * config/rs6000/rs6000.c (emit_unlikely_jump): Likewise.
7532 * config/sh/sh.c (expand_cbranchsi4): Likewise.
7533 * config/spu/spu.c (ea_load_store_inline): Likewise.
7535 * cfgbuild.c (compute_outgoing_frequencies): Use XINT to access the
7536 value of a REG_BR_PROB note.
7537 * cfgrtl.c (force_nonfallthru_and_redirect): Likewise.
7538 (update_br_prob_note, rtl_verify_edges, purge_dead_edges): Likewise.
7539 * emit-rtl.c (try_split): Likewise.
7540 * predict.c (br_prob_note_reliable_p): Likewise.
7541 (invert_br_probabilities, combine_predictions_for_insn): Likewise.
7542 * reorg.c (mostly_true_jump): Likewise.
7543 * config/bfin/bfin.c (cbranch_predicted_taken_p): Likewise.
7544 * config/frv/frv.c (frv_print_operand_jump_hint): Likewise.
7545 * config/i386/i386.c (ix86_print_operand): Likewise.
7546 * config/ia64/ia64.c (ia64_print_operand): Likewise.
7547 * config/mmix/mmix.c (mmix_print_operand): Likewise.
7548 * config/rs6000/rs6000.c (output_cbranch): Likewise.
7549 * config/s390/s390.c (s390_expand_tbegin): Likewise.
7550 * config/sh/sh.c (sh_print_operand, sh_adjust_cost): Likewise.
7551 * config/sparc/sparc.c (output_cbranch): Likewise.
7552 * config/spu/spu.c (get_branch_target): Likewise.
7553 * config/tilegx/tilegx.c (cbranch_predicted_p): Likewise.
7554 * config/tilepro/tilepro.c (cbranch_predicted_p): Likewise.
7556 2013-09-23 Jan Hubicka <jh@suse.cz>
7558 * ipa-cp.c (ipa_get_indirect_edge_target_1): Add sanity check
7560 * ipa-utils.h (possible_polymorphic_call_target_p): New function.
7561 * ipa-devirt.c (possible_polymorphic_call_target_p): Be tolerant
7563 * gimple-fold.c: Include ipa-utils.h and gimple-pretty-print.h
7564 (gimple_fold_call): Dump inconsistent devirtualizations; add
7565 sanity check for type based devirtualizations.
7566 * ipa-prop.c: Include ipa-utils.h
7567 (ipa_intraprocedural_devirtualization): Add sanity check.
7568 (try_make_edge_direct_virtual_call): Likewise.
7570 2013-09-23 Eric Botcazou <ebotcazou@adacore.com>
7572 * tree-ssa-ccp.c (insert_clobber_before_stack_restore): Recurse on copy
7573 assignment statements.
7575 2013-09-23 Kugan Vivekanandarajah <kuganv@linaro.org>
7577 * gimple-pretty-print.c (dump_ssaname_info): New function.
7578 (dump_gimple_phi): Call it.
7579 (pp_gimple_stmt_1): Likewise.
7580 * tree-core.h (tree_ssa_name): New union ssa_name_info_type field.
7581 (range_info_def): Declare.
7582 * tree-pretty-print.c (pp_double_int): New function.
7583 (dump_generic_node): Call it.
7584 * tree-pretty-print.h (pp_double_int): Declare.
7585 * tree-ssa-alias.c (dump_alias_info): Check pointer type.
7586 * tree-ssanames.h (range_info_def): New structure.
7587 (value_range_type): Move definition here.
7588 (set_range_info, value_range_type, duplicate_ssa_name_range_info):
7590 * tree-ssanames.c (make_ssa_name_fn): Check pointer type at
7592 (set_range_info): New function.
7593 (get_range_info): Likewise.
7594 (duplicate_ssa_name_range_info): Likewise.
7595 (duplicate_ssa_name_fn): Check pointer type and call
7596 duplicate_ssa_name_range_info.
7597 * tree-ssa-copy.c (fini_copy_prop): Likewise.
7598 * tree-vrp.c (value_range_type): Remove definition, now in
7600 (vrp_finalize): Call set_range_info to update value range of SSA_NAMEs.
7601 * tree.h (SSA_NAME_PTR_INFO): Macro changed to access via union.
7602 (SSA_NAME_RANGE_INFO): New macro.
7604 2013-09-23 Richard Biener <rguenther@suse.de>
7606 PR tree-optimization/58464
7607 * tree-ssa-pre.c (phi_trans_lookup): Remove.
7608 (phi_trans_add): Change to add conditionally on being not present.
7609 (phi_translate_1): Remove recursion detection here.
7610 (phi_translate): Pre-seed the cache with NULL to catch
7611 recursion here in a more generic way.
7612 (bitmap_find_leader): Adjust comment.
7613 (get_representative_for): Dump value-numbers.
7614 (create_expression_by_pieces): Likewise.
7615 (insert_into_preds_of_block): Likewise.
7617 2013-09-23 Christian Bruel <christian.bruel@st.com>
7620 * config/sh/sh.md (movsf_ie): Allow fpul_operand.
7621 * config/sh/predicate.md (arith_reg_operand): Disallow FPUL_REG.
7623 2013-09-23 James Greenhalgh <james.greenhalgh@arm.com>
7626 2013-09-20 Renlin Li <renlin.li@arm.com>
7628 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
7630 (aarch64_expand_epilogue): Likewise.
7631 (aarch64_legitimize_reload_address): Likewise.
7633 2013-09-22 Eric Botcazou <ebotcazou@adacore.com>
7635 * gimplify.c (gimplify_asm_expr): Reset the TREE_CHAIN of clobbers to
7636 NULL_TREE before pushing them onto the vector. Likewise for labels.
7638 2013-09-21 Eric Botcazou <ebotcazou@adacore.com>
7640 * config/ia64/predicates.md (ia64_cbranch_operator): Accept unordered
7641 comparison operators when -fno-trapping-math is in effect.
7642 * config/ia64/ia64.c (ia64_expand_compare): Add support for unordered
7643 comparison operators in TFmode and assert that unsupported operators
7645 (ia64_print_operand): Likewise.
7647 2013-09-21 Jan Hubicka <jh@suse.cz>
7649 * x86-tune.def (partial_reg_stall): Disable for CoreI7 and newer.
7650 (sse_typeless_stores): Enable for core
7651 (sse_load0_by_pxor): Likewise.
7652 (four_jump_limit): Disable for core.
7653 (pad_returns): Likewise.
7654 (avoid_vector_decode): Likewise.
7655 (fuse_cmp_and_branch): Enable for cores.
7656 * i386.c (x86_accumulate_outgoing_args): Disable for cores.
7658 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
7661 * config/pa/pa.c (pa_option_override): Disable auto increment and
7662 decrement instructions until reload is completed.
7664 * config/pa/pa-linux.h (TARGET_OS_CPP_BUILTINS): Define
7665 __GCC_HAVE_SYNC_COMPARE_AND_SWAP_1, __GCC_HAVE_SYNC_COMPARE_AND_SWAP_2,
7666 and __GCC_HAVE_SYNC_COMPARE_AND_SWAP_4.
7668 2013-09-20 DJ Delorie <dj@redhat.com>
7669 Nick Clifton <nickc@redhat.com>
7671 * config/rl78/rl78.c: Various whitespace and comment tweaks.
7672 (need_to_save): Save bank 0 on interrupts.
7673 (characterize_address): Strip far address wrappers.
7674 (rl78_as_legitimate_address): Likewise.
7675 (transcode_memory_rtx): Likewise.
7676 (rl78_peep_movhi_p): Disable this peephole after devirt.
7677 (rl78_propogate_register_origins): Forget all origins when a
7679 * config/rl78/rl78-virt.md: Various whitespace tweaks.
7680 * config/rl78/rl78-real.md: Various whitespace tweaks. Additional
7682 * config/rl78/rl78.md (sel_rb): Disable for G10 just in case.
7683 * config/rl78/rl78-expand.md (movqi): Check for subregs of consts.
7684 * config/rl78/rl78.h (LINK_SPEC): Pass -gc-sections unless
7686 * config/rl78/constraints.md: Various whitespace and paren tweaks.
7688 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
7690 * config/pa/pa.md: In "scc" insn patterns, change output template to
7691 handle const0_rtx in reg_or_0_operand operands.
7693 2013-09-20 Martin Husemann <martin@NetBSD.org>
7696 * config/vax/vax.c (vax_output_int_move): Use D format specifier.
7697 * config/vax/vax.md (ashldi3, <unnamed>): Ditto.
7699 2013-09-20 Richard Biener <rguenther@suse.de>
7702 * tree-scalar-evolution.c (struct scev_info_str): Shrink by
7703 remembering SSA name version and block index.
7704 (new_scev_info_str): Adjust.
7705 (hash_scev_info): Likewise. Also hash the block index.
7706 (eq_scev_info): Adjust.
7707 (find_var_scev_info): Likewise.
7708 (struct instantiate_cache_entry): Remove.
7709 (struct instantiate_cache_type): Use a htab to map name, block
7711 (instantiate_cache_type::~instantiate_cache_type): Adjust.
7712 (get_instantiated_value_entry): Likewise.
7713 (hash_idx_scev_info, eq_idx_scev_info): New functions.
7714 (instantiate_scev_name): Adjust.
7716 2013-09-20 Jeff Law <law@redhat.com>
7718 * tree-ssa-dom.c (record_temporary_equivalences): Add comment.
7720 2013-09-20 Yufeng Zhang <yufeng.zhang@arm.com>
7722 * config/aarch64/aarch64-builtins.c (aarch64_simd_expand_args):
7723 Call aarch64_simd_expand_args to update op[argc].
7725 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
7727 * plugin.c (parse_plugin_arg_opt): Accept equal sign inside
7730 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
7732 * gengtype.c (file_rules): Added rule for *.cc files.
7733 (get_output_file_with_visibility): Give fatal message when no
7736 2013-09-20 Renlin Li <renlin.li@arm.com>
7738 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use plus_constant.
7739 (aarch64_expand_epilogue): Likewise.
7740 (aarch64_legitimize_reload_address): Likewise.
7742 2013-09-20 Bernd Edlinger <bernd.edlinger@hotmail.de>
7745 * expr.c (expand_assignment): Remove misalignp code path.
7747 2013-09-20 Marek Polacek <polacek@redhat.com>
7750 * ubsan.c (get_ubsan_type_info_for_type): Use TYPE_SIZE instead of
7751 TYPE_PRECISION. Add asserts.
7753 2013-09-20 Richard Biener <rguenther@suse.de>
7755 PR tree-optimization/58453
7756 * tree-loop-distribution.c (distribute_loop): Apply the cost
7757 model for -ftree-loop-distribute-patterns, too.
7759 2013-09-20 Richard Biener <rguenther@suse.de>
7762 * tree-chrec.h (build_polynomial_chrec): Use gcc_checking_assert,
7763 make type comparison less strict.
7765 2013-09-20 Alan Modra <amodra@gmail.com>
7767 * configure: Regenerate.
7768 * aclocal.m4: Regenerate.
7770 2013-09-20 Marek Polacek <polacek@redhat.com>
7773 * doc/extend.texi: Document that attribute used is meant to be used
7774 on variables with static storage duration.
7776 2013-09-19 Jakub Jelinek <jakub@redhat.com>
7778 PR tree-optimization/58472
7779 * tree-vect-stmts.c (vectorizable_store, vectorizable_load): For
7780 simd_lane_access set inv_p = false.
7781 * omp-low.c (lower_rec_input_clauses): Set TREE_NO_WARNING on
7782 the simduid magic VAR_DECL.
7784 2013-09-19 Jan Hubicka <jh@suse.cz>
7786 * i386.c (generic_memcpy, generic_memset): Fix 32bit template.
7788 2013-09-17 Jeff Law <law@redhat.com>
7790 * tree-ssa-dom.c (record_temporary_equivalences): New function
7791 split out of dom_opt_dom_walker::after_dom_children.
7792 (dom_opt_dom_walker::thread_across_edge): Move common code
7793 in here from dom_opt_dom_walker::after_dom_children.
7794 (dom_opt_dom_walker::after_dom_children): Corresponding simplifictions.
7796 2013-09-19 Jan Hubicka <jh@suse.cz>
7798 * i386.h (TARGET_GENERIC32, TARGET_GENERIC64): Remove.
7799 (TARGET_GENERIC): Use PROCESOR_GENERIC
7800 (enum processor_type): Unify generic32 and 64.
7801 * i386.md (cpu): Likewise.
7802 * x86-tune.def (use_leave): Enable for generic32.
7803 (avoid_vector_decode, slow_imul_imm32_mem, slow_imul_imm8): Likewise.
7804 * athlon.md: Change generic64 to generic in all occurences.
7805 * i386-c.c (ix86_target_macros_internal): Unify generic64 and 32.
7806 (ix86_target_macros_internal): Likewise.
7807 * driver-i386.c (host_detect_local_cpu): Likewise.
7808 * i386.c (generic64_memcpy, generic64_memset, generic64_cost): Rename
7810 (generic_memcpy, generic_memset, generic_cost): This one.
7811 (generic32_memcpy, generic32_memset, generic32_cost): Remove.
7812 (m_GENERIC32, m_GENERIC64): Remove.
7813 (m_GENERIC): Turn into one flag.
7814 (processor_target): Unify generic tunnings.
7815 (ix86_option_override_internal): Replace generic32/64 by generic.
7816 (ix86_issue_rate): Likewise.
7817 (ix86_adjust_cost): Likewise.
7819 2013-09-19 Jan Hubicka <jh@suse.cz>
7821 * cgraph.c (cgraph_create_edge_1): Avoid uninitialized read
7822 of speculative flag.
7824 2013-09-19 Jakub Jelinek <jakub@redhat.com>
7826 * omp-low.c (expand_omp_sections): Always pass len - 1 to
7827 GOMP_sections_start, even if !exit_reachable.
7829 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
7831 * lra-constraints.c (need_for_all_save_p): Use macro
7832 HARD_REGNO_CALL_PART_CLOBBERED.
7833 * lra-lives.c (check_pseudos_live_through_calls): Use the macro to
7834 set up pseudo conflict hard regs.
7836 2013-09-18 Michael Meissner <meissner@linux.vnet.ibm.com>
7839 * config/rs6000/paired.md (movmisalignv2sf): Fix to allow memory
7842 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
7844 PR rtl-optimization/58438
7845 * lra.c (lra): Clear lra_optional_reload_pseudos in upper loop.
7846 * lra-constraints.c (undo_optional_reloads): Keep optional reloads
7847 from previous subpasses.
7849 2013-09-18 Richard Earnshaw <rearnsha@arm.com>
7851 * arm.c (arm_get_frame_offsets): Validate architecture supports
7852 LDRD/STRD before accepting the tuning preference.
7853 (arm_expand_prologue): Likewise.
7854 (arm_expand_epilogue): Likewise.
7856 2013-09-18 Richard Biener <rguenther@suse.de>
7858 PR tree-optimization/58417
7859 * tree-chrec.c (chrec_fold_plus_1): Assert that we do not
7860 have chrecs with symbols defined in the loop as operands.
7861 (chrec_fold_multiply): Likewise.
7862 * tree-scalar-evolution.c (interpret_rhs_expr): Instantiate
7863 parameters before folding binary operations.
7864 (struct instantiate_cache_entry_hasher): Remove.
7865 (struct instantiate_cache_type): Use a pointer-map.
7866 (instantiate_cache_type::instantiate_cache_type): New function.
7867 (instantiate_cache_type::get): Likewise.
7868 (instantiate_cache_type::set): Likewise.
7869 (instantiate_cache_type::~instantiate_cache_type): Adjust.
7870 (get_instantiated_value_entry): Likewise.
7871 (global_cache): New global.
7872 (instantiate_scev_r, instantiate_scev_poly, instantiate_scev_binary,
7873 instantiate_array_ref, instantiate_scev_convert, instantiate_scev_3,
7874 instantiate_scev_2, instantiate_scev_1): Do not pass along cache.
7875 (instantiate_scev_name): Adjust.
7876 (instantiate_scev): Construct global instead of local cache.
7877 (resolve_mixers): Likewise.
7879 2013-09-18 Daniel Morris <danielm@ecoscentric.com>
7880 Paolo Carlini <paolo.carlini@oracle.com>
7883 * doc/implement-cxx.texi: Fix references to the C++ standards.
7885 2013-09-18 Jakub Jelinek <jakub@redhat.com>
7887 * omp-low.c (copy_var_decl): Copy DECL_ATTRIBUTES.
7888 * tree-vect-data-refs.c (vect_analyze_data_refs): For
7889 simd_lane_access drs, update also DR_ALIGNED_TO.
7891 2013-09-18 Marek Polacek <polacek@redhat.com>
7894 * doc/extend.texi: Document no_sanitize_undefined attribute.
7895 * builtins.c (fold_builtin_0): Don't sanitize function if it has the
7896 no_sanitize_undefined attribute.
7898 2013-09-18 Nick Clifton <nickc@redhat.com>
7900 * config/msp430/msp430.h (ASM_SPEC): Pass -md on to the assembler.
7901 (ASM_DECLARE_FUNCTION_NAME): Define.
7903 2013-09-17 Trevor Saunders <tsaunders@mozilla.com>
7905 * compare-elim.c (find_comparison_dom_walker): New class
7906 (find_comparisons_in_bb): Rename to
7907 find_comparison_dom_walker::before_dom_children
7908 (find_comparisons): Adjust
7909 * domwalk.c (walk_dominator_tree): Rename to dom_walker::walk, and
7911 (init_walk_dominator_tree, fini_walk_dominator_tree): Remove
7912 * domwalk.h (dom_walk_data): Convert it To a class dom_walker.
7913 (init_walk_dominator_tree): Remove declaration.
7914 (fini_walk_dominator_tree): Remove declaration.
7915 * fwprop.c (single_def_use_dom_walker): New class
7916 (single_def_use_enter_block): Convert to
7917 single_def_use_dom_walker::before_dom_children.
7918 (single_def_use_leave_block): Convert to
7919 single_def_use_dom_walker::after_dom_children.
7920 (build_single_def_use_links): Adjust.
7921 * gimple-ssa-strength-reduction.c (find_candidates_dom_walker): New
7923 (find_candidates_in_block): Convert to
7924 find_candidates_dom_walker::before_dom_children.
7925 (execute_strength_reduction): Adjust.
7926 * graphite-sese-to-poly.c (struct bsc, build_sese_conditions): Remove.
7927 (sese_dom_walker): New class.
7928 (sese_dom_walker::sese_dom_walker): New constructor.
7929 (sese_dom_walker::~sese_dom_walker): New destructor.
7930 (build_sese_conditions_before): Convert to
7931 sese_dom_walker::before_dom_children.
7932 (build_sese_conditions_after): Convert to
7933 sese_dom_walker::after_dom_children.
7934 (build_poly_scop): Adjust
7935 * tree-into-ssa.c (rewrite_dom_walker): New class
7936 (rewrite_enter_block): Convert to
7937 rewrite_dom_walker::before_dom_children.
7938 (rewrite_leave_block): Convert to
7939 rewrite_dom_walker::after_dom_children.
7940 (rewrite_update_dom_walker): New class.
7941 (rewrite_update_enter_block): Convert to
7942 rewrite_update_dom_walker::before_dom_children.
7943 (rewrite_update_leave_block): Convert to
7944 rewrite_update_dom_walker::after_dom_children.
7945 (rewrite_blocks, rewrite_into_ssa): Adjust.
7946 (mark_def_dom_walker): New class.
7947 (mark_def_dom_walker::mark_def_dom_walker): New constructor.
7948 (mark_def_dom_walker::~mark_def_dom_walker): New destructor.
7949 (mark_def_sites_blocks): Convert to
7950 mark_def_dom_walker::before_dom_children.
7951 (mark_def_site_blocks): Remove.
7952 * tree-ssa-dom.c (dom_opt_dom_walker): New class.
7953 (tree_ssa_dominator_optimize): Adjust.
7954 (dom_thread_across_edge): Convert to method
7955 dom_opt_dom_walker::thread_across_edge.
7956 (dom_opt_enter_block): Convert to member function
7957 dom_opt_dom_walker::before_dom_children.
7958 (dom_opt_leave_block): Convert to member function
7959 dom_opt_dom_walker::after_dom_children.
7960 * tree-ssa-dse.c (dse_dom_walker): New class.
7961 (dse_enter_block): Convert to member function
7962 dse_dom_walker::before_dom_children.
7963 (tree_ssa_dse): Adjust.
7964 * tree-ssa-loop-im.c (invariantness_dom_walker): New class.
7965 (determine_invariantness_stmt): Convert to method
7966 invariantness_dom_walker::before_dom_children.
7967 (determine_invariantness): Remove
7968 (move_computations_dom_walker): New class.
7969 (move_computations_stmt): Convert to method
7970 move_computations_dom_walker::before_dom_children.
7971 (move_computations, tree_ssa_lim): Adjust.
7972 * tree-ssa-phiopt.c (nontrapping_dom_walker): New class.
7973 (nt_init_block): Convert to method
7974 notrappping_dom_walker::before_dom_children.
7975 (nt_fini_block): Convert to method
7976 method nontrapping_dom_walker::after_dom_children.
7977 (get_non_trapping): Adjust.
7978 * tree-ssa-pre.c (eliminate_dom_walker): New class.
7979 (eliminate_bb): Convert to method
7980 eliminate_dom_walker::before_dom_children.
7981 (eliminate_leave_block): Convert to method
7982 eliminate_dom_walker::after_dom_children.
7983 (eliminate): Adjust.
7984 * tree-ssa-strlen.c (strlen_dom_walker): New class.
7985 (strlen_enter_block): Convert to method
7986 strlen_dom_walker::before_dom_children.
7987 (strlen_leave_block): Convert to method
7988 method strlen_dom_walker::after_dom_children.
7989 (tree_ssa_strlen): Adjust.
7990 * tree-ssa-uncprop.c (uncprop_dom_walker): New class.
7991 (tree_ssa_uncprop): Adjust.
7992 (uncprop_leave_block): Convert to method
7993 uncprop_dom_walker::after_dom_children.
7994 (uncprop_leave_block): Convert to method
7995 uncprop_dom_walker::before_dom_children.
7997 2013-09-18 Bin Cheng <bin.cheng@arm.com>
7999 * config/arm/arm.c (thumb1_reorg): Search for flag setting insn before
8000 branch in same basic block. Check both src and dest of the move insn.
8002 2013-09-17 Nick Clifton <nickc@redhat.com>
8004 * config/rl78/rl78-real.md (bf): New pattern.
8006 * config/rl78/rl78.c (rl78_print_operand_1): Handle %B.
8007 (rl78_print_operand): Do not put a # before a %B.
8008 * config/rl78/rl78.opt: Tweak doc strings.
8010 2013-09-17 DJ Delorie <dj@redhat.com>
8012 * config/rl78/constraints.md (Wcv): Allow up to $r31.
8013 * config/rl78/rl78.c (rl78_asm_file_start: Likewise.
8014 (rl78_option_override): Likewise, if -mallregs.
8015 (is_virtual_register): Likewise.
8016 * config/rl78/rl78.h (reg_class): Extend VREGS to $r31.
8017 (REGNO_OK_FOR_BASE_P): Likewise.
8018 * config/rl78/rl78.opt (-mallregs): New.
8020 2013-09-17 Nick Clifton <nickc@redhat.com>
8022 * config/rl78/rl78.c (need_to_save): Change return type to bool.
8023 For interrupt functions: save all call clobbered registers if the
8024 interrupt handler is not a leaf function.
8025 (rl78_expand_prologue): Always recompute the frame information.
8026 For interrupt functions: only select bank 0 if one of the bank 0
8027 registers is going to be psuhed.
8029 2013-09-17 DJ Delorie <dj@redhat.com>
8031 * config/rl78/constraints.md: For each W* constraint, rename to C*
8032 and create a W* constraint that checks for an optional ES: prefix
8034 * config/rl78/rl78.md (UNS_ES_ADDR): New.
8035 (es_addr): New. Used to wrap far addresses.
8036 * config/rl78/rl78-protos.h (rl78_es_addr): New.
8037 (rl78_es_base): New.
8038 * config/rl78/rl78.c (rl78_as_legitimate_address): Accept "unspec"
8039 wrapped far addresses.
8040 (rl78_print_operand_1): Unwrap far addresses before processing.
8041 (rl78_lo16): Wrap far addresses in unspecs.
8042 (rl78_es_addr): New.
8043 (rl78_es_base): New.
8044 (insn_ok_now): Check for not-yet-wrapped far addresses.
8045 (transcode_memory_rtx): Properly re-wrap far addresses.
8047 2013-09-17 Sebastian Huber <sebastian.huber@embedded-brains.de>
8049 * config/sparc/t-rtems: Add leon3 multilibs.
8051 2013-09-17 Cong Hou <congh@google.com>
8053 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Fix a bug
8054 when checking the dot production pattern. The type of rhs operand
8055 of multiply is now checked correctly.
8057 2013-09-17 Jeff Law <law@redhat.com>
8059 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
8060 edge implied equivalences into successor phis.
8061 * tree-ssa-threadupdate.c (phi_args_equal_on_edges): Moved into
8062 here from tree-ssa-threadedge.c.
8063 (mark_threaded_blocks): When threading through a joiner, if both
8064 successors of the joiner's clone reach the same block, verify the
8065 PHI arguments are equal. If not, cancel the jump threading request.
8066 * tree-ssa-threadedge.c (phi_args_equal_on_edges): Moved into
8067 tree-ssa-threadupdate.c
8068 (thread_across_edge): Don't check PHI argument equality when
8069 threading through joiner block here.
8071 2013-09-17 Andrew MacLeod <amacleod@redhat.com>
8073 * tree-flow.h (ssa_undefined_value_p): Remove prototype.
8074 * tree-ssa.c (ssa_undefined_value_p): Move pass independent parts here.
8075 (warn_uninit, warn_uninitialized_vars,
8076 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
8077 Move to tree-ssa-uninit.c.
8078 * tree-ssa-uninit.c (ssa_undefined_value_p): Move to tree-ssa.c.
8079 (has_undefined_value_p): New. Pass dependant parts of
8080 ssa_undefined_value_p.
8081 (uninit_undefined_value_p): Use has_undefined_value_p.
8082 (warn_uninit, warn_uninitialized_vars,
8083 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
8084 Move from tree-ssa.c.
8085 * tree-ssa.h: Adjust prototypes.
8087 2013-09-17 Jan Hubicka <jh@suse.cz>
8090 * cif-code.def (FUNCTION_NOT_OPTIMIZED): New CIF code.
8091 * ipa-inline.c (can_inline_edge_p): Do not downgrade
8092 FUNCTION_NOT_OPTIMIZED.
8093 * ipa-inline-analysis.c (compute_inline_parameters): Function
8094 not optimized is not inlinable unless it is alwaysinline.
8095 (inline_analyze_function): Force calls in not optimized
8096 function not inlinable.
8098 2013-09-17 Jan Hubicka <jh@suse.cz>
8101 * ipa-devirt.c (ipa_devirt): Be ready for symtab_nonoverwritable_alias
8103 * ipa.c (function_and_variable_visibility): Likewise.
8104 * ipa-profile.c (ipa_profile): Likewise.
8106 2013-09-17 Bernd Edlinger <bernd.edlinger@hotmail.de>
8109 * cgraph.c (cgraph_function_body_availability): Check for ifunc
8110 attribute, and don't inline the resolver in this case.
8112 2013-09-17 Teresa Johnson <tejohnson@google.com>
8114 * coverage.c (get_coverage_counts): Add missing newline.
8116 2013-09-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8118 PR tree-optimization/58088
8119 * fold-const.c (mask_with_trailing_zeros): New function.
8120 (fold_binary_loc): Make sure we don't recurse infinitely
8121 when the X in (X & C1) | C2 is a tree of the form (Y * K1) & K2.
8122 Use mask_with_trailing_zeros where appropriate.
8124 2013-09-17 Yuri Rumyantsev <ysrumyan@gmail.com>
8126 * config/i386/i386.c (distance_agu_use_in_bb) : Proper initialization
8127 of 'prev' var to get better distance estimation.
8129 2013-09-17 Eric Botcazou <ebotcazou@adacore.com>
8131 * tree-inline.h (struct copy_body_data): Add transform_parameter.
8132 * tree-inline.c (is_parameter_of): New predicate.
8133 (remap_gimple_op_r): Do not propagate TREE_THIS_NOTRAP on MEM_REF if
8134 a parameter has been remapped.
8135 (copy_tree_body_r): Likewise on INDIRECT_REF and MEM_REF.
8136 (optimize_inline_calls): Initialize transform_parameter.
8137 (copy_gimple_seq_and_replace_locals): Likewise.
8138 (tree_function_versioning): Likewise.
8139 (maybe_inline_call_in_expr): Likewise.
8141 2013-09-17 Nick Clifton <nickc@redhat.com>
8143 * config/msp430/msp430-protos.h: Add prototypes for new functions.
8144 * config/msp430/msp430.c (msp430_preserve_reg_p): Add support for
8146 (is_attr_func): New function.
8147 (msp430_is_interrupt_func): New function.
8148 (is_naked_func): New function.
8149 (is_reentrant_func): New function.
8150 (is_critical_func): New function.
8151 (msp430_start_function): Add annotations for function attributes.
8152 (msp430_attr): New function.
8153 (msp430_attribute_table): New.
8154 (msp430_function_section): New function.
8155 (TARGET_ASM_FUNCTION_SECTION): Define.
8156 (msp430_builtin): New enum.
8157 (msp430_init_builtins): New function.
8158 (msp430_builtin_devl): New function.
8159 (msp430_expand_builtin): New function.
8160 (TARGET_INIT_BUILTINS): Define.
8161 (TARGET_EXPAND_BUILTINS): Define.
8162 (TARGET_BUILTIN_DECL): Define.
8163 (msp430_expand_prologue): Add support for naked, interrupt,
8164 critical and reentrant functions.
8165 (msp430_expand_epilogue): Likewise.
8166 (msp430_print_operand): Handle 'O' character.
8167 * config/msp430/msp430.h (TARGET_CPU_CPP_BUILTINS): Define
8169 * config/msp430/msp430.md (unspec): Add UNS_DINT, UNS_EINT,
8170 UNS_PUSH_INTR, UNS_POP_INTR, UNS_BIC_SR, UNS_BIS_SR.
8171 (pushm): Use a 'n' rather than an 'i' constraint.
8172 (msp_return): Add generation of the interrupt return instruction.
8173 (disable_interrupts): New pattern.
8174 (enable_interrupts): New pattern.
8175 (push_intr_state): New pattern.
8176 (pop_intr_state): New pattern.
8177 (bic_SR): New pattern.
8178 (bis_SR): New pattern.
8179 * doc/extend.texi: Document MSP430 function attributes and builtin
8182 2013-09-17 Richard Biener <rguenther@suse.de>
8184 PR tree-optimization/58432
8185 * tree-loop-distribution.c (tree_loop_distribution): Also
8186 scan PHIs for outside loop uses and seed a partition from them.
8188 2013-09-17 Bin Cheng <bin.cheng@arm.com>
8190 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref): New.
8191 (restructure_reference): Call backtrace_base_for_ref.
8193 2013-09-17 Alan Modra <amodra@gmail.com>
8196 * config/rs6000/driver-rs6000.c (elf_platform): Revert 2013-06-11
8199 2013-09-16 DJ Delorie <dj@redhat.com>
8201 * config/rl78/rl78.c (rl78_asm_file_start): Specify alternate
8202 vregs location for RL78/G10.
8203 (rl78_expand_prologue): Avoid SEL on G10.
8204 (rl78_expand_epilogue): Likewise.
8205 (rl78_peep_movhi_p): Can't move a constant to memory in HImode.
8206 * config/rl78/rl78.h (TARGET_CPU_CPP_BUILTINS): Define
8207 __RL78_G10__ when appropriate.
8208 (ASM_SPEC): Pass -mg10 along to the assembler.
8209 * config/rl78/rl78.md (sel_rb): Disable for G10.
8210 * config/rl78/rl78.opt: Add -mg10 option.
8211 * config/rl78/t-rl78: Add -mg10 multilib.
8213 2013-09-16 Xinliang David Li <davidxl@google.com>
8215 * tree-if-conv.c (main_tree_if_conversion): Check new flag.
8216 * omp-low.c (omp_max_vf): Ditto.
8217 (expand_omp_simd): Ditto.
8218 * tree-vectorizer.c (vectorize_loops): Ditto.
8219 (gate_vect_slp): Ditto.
8220 (gate_increase_alignment): Ditto.
8221 * tree-ssa-pre.c (inhibit_phi_insertion): Ditto.
8222 * tree-ssa-loop.c (gate_tree_vectorize): Ditto.
8223 (gate_tree_vectorize): Name change.
8224 (tree_vectorize): Ditto.
8225 (pass_vectorize::gate): Call new function.
8226 (pass_vectorize::execute): Ditto.
8227 * opts.c: O3 default setting change.
8228 (finish_options): Check new flag.
8229 * doc/invoke.texi: Document new flags.
8230 * common.opt: New flags.
8232 2013-09-16 Andreas Schwab <schwab@linux-m68k.org>
8234 * doc/tm.texi.in (Cond Exec Macros): Remove node.
8235 (Condition Code): Don't reference it.
8236 * doc/tm.texi: Regenerate.
8238 2013-09-16 Vladimir Makarov <vmakarov@redhat.com>
8241 * lra-constraints.c (undo_optional_reloads): Consider all optional
8242 reload even if it did not get a hard reg.
8244 2013-09-16 Teresa Johnson <tejohnson@google.com>
8246 * dumpfile.c (dump_loc): Remove newline emission.
8247 * tree-vect-data-refs.c (vect_lanes_optab_supported_p): Add newline
8248 emission to dump_printf_loc calls where missing.
8249 (vect_mark_for_runtime_alias_test): Ditto.
8250 (vect_analyze_data_ref_dependence): Ditto.
8251 (vect_analyze_data_ref_dependences): Ditto.
8252 (vect_slp_analyze_data_ref_dependence): Ditto.
8253 (vect_slp_analyze_data_ref_dependences): Ditto.
8254 (vect_compute_data_ref_alignment): Ditto.
8255 (vect_update_misalignment_for_peel): Ditto.
8256 (vect_verify_datarefs_alignment): Ditto.
8257 (vector_alignment_reachable_p): Ditto.
8258 (vect_get_data_access_cost): Ditto.
8259 (vect_enhance_data_refs_alignment): Ditto.
8260 (vect_find_same_alignment_drs): Ditto.
8261 (vect_analyze_data_refs_alignment): Ditto.
8262 (vect_analyze_group_access): Ditto.
8263 (vect_analyze_data_ref_access): Ditto.
8264 (vect_analyze_data_ref_accesses): Ditto.
8265 (vect_prune_runtime_alias_test_list): Ditto.
8266 (vect_analyze_data_refs): Ditto.
8267 (vect_create_addr_base_for_vector_ref): Ditto.
8268 (vect_create_data_ref_ptr): Ditto.
8269 (vect_grouped_store_supported): Ditto.
8270 (vect_grouped_load_supported): Ditto.
8271 * value-prof.c (check_counter): Ditto.
8272 (check_ic_target): Ditto.
8273 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Ditto.
8274 (vect_recog_widen_mult_pattern): Ditto.
8275 (vect_recog_widen_sum_pattern): Ditto.
8276 (vect_recog_over_widening_pattern): Ditto.
8277 (vect_recog_widen_shift_pattern): Ditto.
8278 (vect_recog_rotate_pattern): Ditto.
8279 (vect_recog_vector_vector_shift_pattern): Ditto.
8280 (vect_recog_divmod_pattern): Ditto.
8281 (vect_recog_mixed_size_cond_pattern): Ditto.
8282 (vect_recog_bool_pattern): Ditto.
8283 (vect_pattern_recog_1): Ditto.
8284 (vect_pattern_recog): Ditto.
8285 * tree-vect-loop.c (vect_determine_vectorization_factor): Ditto.
8286 (vect_is_simple_iv_evolution): Ditto.
8287 (vect_analyze_scalar_cycles_1): Ditto.
8288 (vect_get_loop_niters): Ditto.
8289 (vect_analyze_loop_1): Ditto.
8290 (vect_analyze_loop_form): Ditto.
8291 (vect_analyze_loop_operations): Ditto.
8292 (vect_analyze_loop_2): Ditto.
8293 (vect_analyze_loop): Ditto.
8294 (report_vect_op): Ditto.
8295 (vect_is_slp_reduction): Ditto.
8296 (vect_is_simple_reduction_1): Ditto.
8297 (vect_get_known_peeling_cost): Ditto.
8298 (vect_estimate_min_profitable_iters): Ditto.
8299 (vect_model_reduction_cost): Ditto.
8300 (vect_model_induction_cost): Ditto.
8301 (get_initial_def_for_induction): Ditto.
8302 (vect_create_epilog_for_reduction): Ditto.
8303 (vectorizable_reduction): Ditto.
8304 (vectorizable_induction): Ditto.
8305 (vectorizable_live_operation): Ditto.
8306 (vect_loop_kill_debug_uses): Ditto.
8307 (vect_transform_loop): Ditto.
8308 * tree-vect-stmts.c (vect_mark_relevant): Ditto.
8309 (vect_stmt_relevant_p): Ditto.
8310 (process_use): Ditto.
8311 (vect_mark_stmts_to_be_vectorized): Ditto.
8312 (vect_model_simple_cost): Ditto.
8313 (vect_model_promotion_demotion_cost): Ditto.
8314 (vect_model_store_cost): Ditto.
8315 (vect_get_store_cost): Ditto.
8316 (vect_model_load_cost): Ditto.
8317 (vect_get_load_cost): Ditto.
8318 (vect_init_vector_1): Ditto.
8319 (vect_get_vec_def_for_operand): Ditto.
8320 (vect_finish_stmt_generation): Ditto.
8321 (vectorizable_call): Ditto.
8322 (vectorizable_conversion): Ditto.
8323 (vectorizable_assignment): Ditto.
8324 (vectorizable_shift): Ditto.
8325 (vectorizable_operation): Ditto.
8326 (vectorizable_store): Ditto.
8327 (vectorizable_load): Ditto.
8328 (vectorizable_condition): Ditto.
8329 (vect_analyze_stmt): Ditto.
8330 (vect_transform_stmt): Ditto.
8331 (vect_is_simple_use): Ditto.
8332 * tree-vect-loop-manip.c (slpeel_make_loop_iterate_ntimes): Ditto.
8333 (vect_can_advance_ivs_p): Ditto.
8334 (vect_update_ivs_after_vectorizer): Ditto.
8335 (vect_do_peeling_for_loop_bound): Ditto.
8336 (vect_gen_niters_for_prolog_loop): Ditto.
8337 (vect_update_inits_of_drs): Ditto.
8338 (vect_create_cond_for_alias_checks): Ditto.
8339 * tree-vect-slp.c (vect_get_and_check_slp_defs): Ditto.
8340 (vect_build_slp_tree_1): Ditto.
8341 (vect_supported_load_permutation_p): Ditto.
8342 (vect_analyze_slp_instance): Ditto.
8343 (vect_analyze_slp): Ditto.
8344 (vect_make_slp_decision): Ditto.
8345 (vect_detect_hybrid_slp): Ditto.
8346 (vect_bb_vectorization_profitable_p): Ditto.
8347 (vect_slp_analyze_bb_1): Ditto.
8348 (vect_update_slp_costs_according_to_vf): Ditto.
8349 (vect_get_mask_element): Ditto.
8350 (vect_transform_slp_perm_load): Ditto.
8351 (vect_schedule_slp_instance): Ditto.
8352 (vect_schedule_slp): Ditto.
8353 (vect_slp_transform_bb): Ditto.
8354 * profile.c (read_profile_edge_counts): Ditto.
8355 (compute_branch_probabilities): Ditto.
8356 * coverage.c (get_coverage_counts): Ditto.
8358 2013-09-16 Diego Novillo <dnovillo@google.com>
8360 * tree-core.h: Add missing comment lines from refactoring of tree.h.
8362 2013-09-16 Jan Hubicka <jh@suse.cz>
8364 * gimple-fold.c (can_refer_decl_in_current_unit_p): Do not accept
8365 abstract functions; for static functions check the presence of body.
8367 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
8369 * config/aarch64/aarch64-simd-builtins.def (fma): New.
8370 * config/aarch64/aarch64-simd.md
8371 (aarch64_mla_elt<mode>): New.
8372 (aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
8373 (aarch64_mls_elt<mode>): Likewise.
8374 (aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
8375 (aarch64_fma4_elt<mode>): Likewise.
8376 (aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
8377 (aarch64_fma4_elt_to_128v2df): Likewise.
8378 (aarch64_fma4_elt_to_64df): Likewise.
8379 (fnma<mode>4): Likewise.
8380 (aarch64_fnma4_elt<mode>): Likewise.
8381 (aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
8382 (aarch64_fnma4_elt_to_128v2df): Likewise.
8383 (aarch64_fnma4_elt_to_64df): Likewise.
8384 * config/aarch64/iterators.md (VDQSF): New.
8385 * config/aarch64/arm_neon.h
8386 (vfm<as><sdq>_lane<q>_f<32, 64>): Convert to C implementation.
8387 (vml<sa><q>_lane<q>_<fsu><16, 32, 64>): Likewise.
8389 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
8391 * config/aarch64/aarch64-simd.md (aarch64_mul3_elt<mode>): New.
8392 (aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
8393 (aarch64_mul3_elt_to_128df): Likewise.
8394 (aarch64_mul3_elt_to_64v2df): Likewise.
8395 * config/aarch64/iterators.md (VEL): Also handle DFmode.
8397 (VMUL_CHANGE_NLANES) Likewise.
8400 * config/aarch64/arm_neon.h
8401 (vmul<q>_lane<q>_<suf><16,32,64>): Convert to C implementation.
8403 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
8405 * config/aarch64/arm_neon.h
8406 (vcvtx_high_f32_f64): Fix parameters.
8408 2013-09-16 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8409 Uros Bizjak <ubizjak@gmail.com>
8411 * config/alpha.c: Include tree-ssanames.h.
8413 2013-09-16 Richard Biener <rguenther@suse.de>
8415 * tree-loop-distribution.c (enum rdg_dep_type): Add control_dd.
8416 (dot_rdg_1): Handle control_dd.
8417 (create_edge_for_control_dependence): New function.
8418 (create_rdg_edges): Add control dependences if asked for.
8419 (build_rdg): Likewise.
8420 (generate_loops_for_partition): If there are not necessary
8421 control stmts remove all their dependencies.
8422 (collect_condition_stmts, rdg_flag_loop_exits): Remove.
8423 (distribute_loop): Pass on control dependences.
8424 (tree_loop_distribution): Compute control dependences and remove
8425 restriction on number of loop nodes.
8427 2013-09-16 Jakub Jelinek <jakub@redhat.com>
8429 * ipa-prop.c (ipa_compute_jump_functions_for_edge): Return early
8432 2013-09-16 Richard Sandiford <rdsandiford@googlemail.com>
8434 * cse.c (try_const_anchors): Punt on CC modes.
8436 2013-09-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8438 * config/vax/constraints.md (T): Add missing CONSTANT_P check.
8440 2013-09-14 John David Anglin <danglin@gcc.gnu.org>
8443 * config/pa/pa.c (pa_expand_prologue): Change mode in gen_rtx_POST_INC
8446 2013-09-14 Iain Sandoe <iain@codesourcery.com>
8449 * config/darwin.c (darwin_objc2_section): Note if ObjC Metadata is
8451 (darwin_objc1_section): Likewise.
8452 (darwin_file_end): Emit Image Info section when required.
8454 2013-09-14 Jan Hubicka <jh@suse.cz>
8456 * tree-into-ssa.c (gate_into_ssa): New.
8457 (pass_data_build_ssa): Use it.
8458 * cgraph.h (expand_thunk): Update prototype.
8459 * cgraphunit.c (analyze_function): Expand thunks early.
8460 (expand_thunk): Fix DECL_CONTEXT of reust_decl;
8461 build proper cgraph; set in_ssa_p; clear bogus TREE_ASM_WRITTEN;
8462 set lowered flag; do not add new function.
8463 (assemble_thunks_and_aliases): Update.
8464 * tree-ssa.c (gate_init_datastructures): New gate.
8465 (pass_data_init_datastructures): Use it.
8467 2013-09-14 Iain Sandoe <iain@codesourcery.com>
8470 * config/i386/i386.c (ix86_function_arg_regno_p): Make Darwin use the
8471 xmm register set described in the psABI.
8473 2013-09-13 Evgeny Gavrin <e.gavrin@samsung.com>
8475 * dwarf2out.c (should_emit_struct_debug): Add check
8476 for type_decl variable is not NULL.
8478 2013-09-13 Jacek Caban <jacek@codeweavers.com>
8480 * config.gcc: Use new winnt-c.c target hooks
8481 * config/t-winnt: New file
8482 * config/winnt-c.c: New file
8483 * doc/tm.texi.in: Document new hook
8484 * doc/tm.texi: Regenerated
8486 2013-09-13 Jan Hubicka <jh@suse.cz>
8489 * ipa-inline.c (check_callers): New function.
8490 (check_caller_edge): Remove.
8491 (want_inline_function_to_all_callers_p): Also permit alises that are
8493 (inline_to_all_callers): Terminate the walk when devirtualization
8494 introduce new calls.
8496 2013-09-13 Jan Hubicka <jh@suse.cz>
8498 * ipa-inline-analysis.c (struct growth_data): Add node.
8499 (do_estimate_growth_1): Fix detection of recursion.
8501 2013-09-13 Jakub Jelinek <jakub@redhat.com>
8503 PR tree-optimization/58392
8504 * tree-cfg.c (move_sese_region_to_fn): Rename loop variable
8505 to avoid shadowing of outer loop variable. If
8506 saved_cfun->has_simduid_loops or saved_cfun->has_force_vect_loops,
8507 replace_by_duplicate_decl simduid of loops that have it set and
8508 set dest_cfun->has_simduid_loops and/or
8509 dest_cfun->has_force_vect_loops.
8510 * omp-low.c (build_outer_var_ref): Call maybe_lookup_decl_in_outer_ctx
8511 instead of maybe_lookup_decl.
8512 * tree-inline.c (copy_loops): Change blocks_to_copy argument to id.
8513 Use id->blocks_to_copy instead of blocks_to_copy. Adjust recursive
8514 call. Copy over force_vect and copy and remap simduid. Set
8515 cfun->has_simduid_loops and/or cfun->has_force_vect_loops.
8516 (copy_cfg_body): Remove blocks_to_copy argument. Use
8517 id->blocks_to_copy instead of blocks_to_copy. Adjust copy_loops
8518 caller. Don't set cfun->has_simduid_loops and/or
8519 cfun->has_force_vect_loops here.
8520 (copy_body): Remove blocks_to_copy argument. Adjust copy_cfg_body
8522 (expand_call_inline, tree_function_versioning): Adjust copy_body
8525 2013-09-13 Martin Jambor <mjambor@suse.cz>
8528 * ipa-prop.c (try_make_edge_direct_simple_call): Be less strict in
8529 the assert if the edge was a speculative one.
8531 2013-09-13 Richard Biener <rguenther@suse.de>
8533 * tree-data-ref.h (known_dependences_p): Move here ...
8534 * tree-loop-distribution.c (known_dependences_p): ... from here.
8535 (dump_rdg_component, debug_rdg_component): Remove.
8537 (generate_loops_for_partition): Use gimple_uid instead of
8538 relying on matching stmt visit order.
8539 (rdg_build_partitions): Take starting stmt vector.
8540 (ldist_gen): Merge into ...
8541 (distribute_loop): ... this function. Do not compute starting
8543 * tree-cfg.c (gimple_duplicate_bb): Copy UID for PHIs.
8545 2013-09-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8547 * config/arm/arm.md (arm_cmpsi_insn): Split rI alternative.
8548 Set type attribute correctly. Set predicable_short_it attribute.
8549 (cmpsi_shiftsi): Remove %? from output template.
8551 2013-09-13 Richard Biener <rguenther@suse.de>
8553 * tree-loop-distribution.c (struct rdg_component,
8554 rdg_defs_used_in_other_loops_p, free_rdg_components,
8555 rdg_build_components): Remove.
8556 (stmts_from_loop): Do not record virtual PHIs.
8557 (generate_loops_for_partition): Skip virtual PHIs.
8558 (build_rdg_partition_for_component): Rename to ...
8559 (build_rdg_partition_for_vertex): ... this and adjust.
8560 (rdg_build_partitions): Take a vector of starting vertices
8561 instead of components. Remove unnecessary leftover handling.
8562 (ldist_gen): Do not build components or record other stores.
8563 (distribute_loop): Do not distribute loops containing stmts
8566 2013-09-13 Christian Bruel <christian.bruel@st.com>
8569 * config/sh/sh.md (mov<mode>_reg_reg): Allow memory reloads.
8571 2013-09-13 Kai Tietz <ktietz@redhat.com>
8573 * config.gcc: Separate cases for mingw and cygwin targets,
8574 and add 64-bit cygwin target case.
8576 * config/i386/winnt-cxx.c (i386_pe_type_dllexport_p): Don't
8577 dll-export inline-functions.
8578 * config/i386/winnt.c (i386_pe_determine_dllexport_p): Likewise.
8580 2013-09-13 Jeff Law <law@redhat.com>
8584 2013-09-06 Jeff Law <law@redhat.com>
8586 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
8587 edge implied equivalences into successor phis.
8589 2013-09-12 DJ Delorie <dj@redhat.com>
8591 * config/rl78/rl78-virt.md: Change from | to \; for asm line
8594 2013-09-12 Brooks Moses <bmoses@google.com>
8597 * Makefile.in: Do not install driver binaries in $(target)/bin.
8599 2013-09-12 DJ Delorie <dj@redhat.com>
8601 * config/rl78/rl78.opt (mrelax): New.
8602 * config/rl78/rl78.h (ASM_SPEC): New, pass on -mrelax to gas.
8603 * config/rl78/rl78.h (LINK_SPEC): New, pass on -mrelax to ld.
8605 * config/rl78/rl78.c (rl78_expand_prologue): Use AX to copy
8607 (rl78_expand_epilogue): Likewise.
8609 2013-09-12 Vladimir Makarov <vmakarov@redhat.com>
8612 * lra-eliminations.c (remove_reg_equal_offset_note): New.
8613 (eliminate_regs_in_insn): Rewrite frame pointer to hard frame
8614 pointer elimination with using remove_reg_equal_offset_note.
8616 2013-09-12 DJ Delorie <dj@redhat.com>
8618 * config/msp430/: New port.
8619 * config.gcc (msp430): Added.
8620 * doc/invoke.texi: Document MSP430 options.
8621 * doc/install.texi: Document msp430-elf
8622 * doc/md.texi: Document msp430-elf
8623 * doc/contrib.texi: Document msp430-elf
8625 * cfgexpand.c (expand_debug_expr): Avoid sign-extending SImode to
8628 2013-09-12 Martin Jambor <mjambor@suse.cz>
8631 * ipa-prop.c (remove_described_reference): Give up if the edge in the
8632 reference descriptor is NULL.
8633 (ipa_edge_removal_hook): If owning a reference descriptor, set its
8636 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
8638 * tree-flow.h (FREE_SSANAMES): Move to tree-ssanames.c
8639 (SSANAMES, MODIFIED_NORETURN_CALLS, DEFAULT_DEFS, ptr_info_def,
8640 num_ssa_names, ssa_name): Move to tree-ssanames.h + prototypes.
8641 * tree-flow-inline.h (make_ssa_name, copy_ssa_name, duplicate_ssa_name,
8642 make_temp_ssa_name): move to tree-ssanames.h
8643 * tree-ssa-alias.h: Move prototype.
8644 * tree-ssa.h: Include tree-ssanames.h.
8645 * tree-ssanames.c (FREE_SSANAMES): Move to here.
8646 * tree-ssanames.h: New. Move items from tree-flow*.h
8647 * Makefile.in (tree-ssanames.h): Add to tree-ssanames.o and GTFILES.
8649 2013-09-12 Richard Biener <rguenther@suse.de>
8651 PR tree-optimization/58404
8652 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Also
8653 propagate non-invariant addresses into dereferences wrapped
8654 in component references.
8656 2013-09-12 Richard Biener <rguenther@suse.de>
8658 PR tree-optimization/58402
8659 * passes.def: Move pass_late_warn_uninitialized later.
8661 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
8663 * tree-ssa.h: New. Move content from tree-flow.h and
8665 * tree-flow.h (_edge_var_map, edge_var_map_vector): Move to tree-ssa.h.
8666 Move prototypes belonging to tree-ssa.c.
8667 * tree-flow-inline.h (redirect_edge_var_map_def,
8668 redirect_edge_var_map_result, redirect_edge_var_map_location): Move to
8670 * gimple.h: Adjust prototypes.
8671 * tree-ssa.c (useless_type_conversion_p, types_compatible_p): Move
8673 * gimple.c (useless_type_conversion_p, types_compatible_p): Here.
8674 * tree.h: Move prototype to tree-ssa.h.
8675 * gengtype.c (open_base_files): Replace tree-flow.h with tree-ssa.h.
8676 * Makefile.in: (TREE_SSA_H, TREE_FLOW_H): Adjust dependencies.
8677 * alias.c, asan.c, builtins.c, calls.c, cfgexpand.c, cfghooks.c,
8678 cfgloop.c, cfgloopmanip.c, cgraph.c, cgraphbuild.c, cgraphclones.c,
8679 cgraphunit.c, dse.c, except.c, expr.c, final.c, fold-const.c,
8680 ggc-page.c, gimple-fold.c, gimple-iterator.c, gimple-low.c,
8681 gimple-pretty-print.c, gimple-ssa-strength-reduction.c,
8682 gimple-streamer-in.c, gimple-streamer-out.c, gimple.c, gimplify.c,
8683 graphite-blocking.c, graphite-clast-to-gimple.c,
8684 graphite-dependences.c, graphite-interchange.c,
8685 graphite-optimize-isl.c, graphite-poly.c, graphite-scop-detection.c,
8686 graphite-sese-to-poly.c, graphite.c, ipa-cp.c, ipa-inline-analysis.c,
8687 ipa-inline-transform.c, ipa-inline.c, ipa-prop.c, ipa-pure-const.c,
8688 ipa-reference.c, ipa-split.c, ipa-utils.c,
8689 loop-init.c, lto-cgraph.c, lto-section-in.c, lto-section-out.c,
8690 lto-streamer-in.c, lto-streamer-out.c, lto-streamer.c, omp-low.c,
8691 passes.c, predict.c, print-tree.c, profile.c, sese.c, targhooks.c,
8692 tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
8693 tree-chrec.c, tree-complex.c, tree-data-ref.c, tree-dfa.c, tree-eh.c,
8694 tree-emutls.c, tree-if-conv.c, tree-inline.c, tree-into-ssa.c,
8695 tree-loop-distribution.c, tree-mudflap.c, tree-nested.c, tree-nrv.c,
8696 tree-object-size.c, tree-optimize.c, tree-outof-ssa.c, tree-parloops.c,
8697 tree-phinodes.c, tree-predcom.c, tree-pretty-print.c, tree-profile.c,
8698 tree-scalar-evolution.c, tree-sra.c, tree-ssa*.c, tree-stdarg.c,
8699 tree-streamer-in.c, tree-switch-conversion.c, tree-tailcall.c,
8700 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
8701 tree-vect-loop.c, tree-vect-patterns.c, tree-vect-slp.c,
8702 tree-vect-stmts.c, tree-vectorizer.c, tree-vrp.c, tsan.c,
8703 value-prof.c, var-tracking.c,
8704 varpool.c, vtable-verify.c: Replace tree-flow.h with tree-ssa.h
8706 2013-09-12 Richard Biener <rguenther@suse.de>
8708 PR tree-optimization/58396
8709 * tree-loop-distribution.c (create_rdg_edges): Free unused DDRs.
8710 (build_rdg): Take a loop-nest parameter, fix memleaks.
8711 (distribute_loop): Compute loop-nest here and pass it to build_rdg.
8713 2013-09-12 Yuri Rumyantsev <ysrumyan@gmail.com>
8715 * config/i386/x86-tune.def: Turn on X86_TUNE_AVOID_MEM_OPND_FOR_CMOVE
8718 2013-09-12 Cameron McInally <cameron.mcinally@nyu.edu>
8720 * doc/extend.texi: Fix errors in x86 FMA builtin naming.
8721 The FMA instruction names should have a 'v' prefix.
8723 2013-09-12 Richard Biener <rguenther@suse.de>
8725 * tree-loop-distribution.c (dot_rdg_1): Make graph prettier.
8726 (dot_rdg): Use popen instead of system in optional code.
8727 (remaining_stmts, upstream_mem_writes): Remove global bitmaps.
8728 (already_processed_vertex_p): Adjust.
8729 (has_anti_or_output_dependence, predecessor_has_mem_write,
8730 mark_nodes_having_upstream_mem_writes, has_upstream_mem_writes,
8731 rdg_flag_uses): Remove.
8732 (rdg_flag_vertex): Simplify.
8733 (rdg_flag_vertex_and_dependent): Rely on a correct RDG and
8735 (build_rdg_partition_for_component): Process the first vertex
8736 of a component only.
8737 (ldist_gen): Do not compute remaining_stmts or upstream_mem_writes.
8739 2013-09-12 Alan Modra <amodra@gmail.com>
8741 * config/rs6000/rs6000.c (toc_relative_expr_p): Use add_cint_operand.
8743 2013-09-11 DJ Delorie <dj@redhat.com>
8744 Nick Clifton <nickc@redhat.com>
8746 * config/rl78/predicates.md (rl78_cmp_operator_signed): New.
8747 (rl78_stack_based_mem): New.
8748 * config/rl78/constraints.md (Iv08): New.
8757 * config/rl78/rl78-expand.md (movqi): Reject more SUBREG operands.
8759 (movsi): Change from expand to insn-and-split.
8760 (ashrsi3): Clobber AX.
8764 * config/rl78/rl78.md (CC_REG): Fix.
8765 (addsi3): Allow memory and immediate operands.
8766 (addsi3_internal): Split into...
8767 (addsi3_internal_virt): ...new, and ...
8768 (addsi3_internal_real): ...new.
8770 (subsi3_internal_virt): New.
8771 (subsi3_internal_real): New.
8772 (mulsi3): Add memory operand.
8773 (mulsi3_rl78): Likewise.
8774 (mulsi3_g13): Likewise.
8775 * config/rl78/rl78-real.md (cbranchqi4_real_signed): New.
8776 (cbranchqi4_real): Add more constraint options.
8777 (cbranchhi4_real): Expand pattern.
8778 (cbranchhi4_real_signed): New.
8779 (cbranchhi4_real_inverted): New.
8780 (cbranchsi4_real_lt): New.
8781 (cbranchsi4_real_ge): New.
8782 (cbranchsi4_real_signed): New.
8783 (cbranchsi4_real): New.
8785 * config/rl78/rl78-virt.md (ashrsi3_virt): Add custom cases for
8787 (lshrsi3_virt): Likewise.
8788 (ashlsi3_virt): Likewise.
8789 (cbranchqi4_virt_signed): New.
8790 (cbranchhi4_virt_signed): New.
8791 (cbranchsi4_virt): New.
8792 * config/rl78/rl78.c: Whitespace fixes throughout.
8793 (move_elim_pass): New.
8794 (pass_data_rl78_move_elim): New.
8795 (pass_rl78_move_elim): New.
8796 (make_pass_rl78_move_elim): New.
8797 (rl78_devirt_info): Run devirt earlier.
8798 (rl78_move_elim_info): New.
8799 (rl78_asm_file_start): Register it.
8800 (rl78_split_movsi): New.
8801 (rl78_as_legitimate_address): Allow virtual base registers when
8803 (rl78_addr_space_convert): Remove spurious debug stuff.
8804 (rl78_print_operand_1): Add z,s,S,r,E modifiers.
8805 (rl78_print_operand): More cases for not printing '#'.
8806 (rl78_expand_compare): Remove most of the logic.
8807 (content_memory): New.
8808 (clear_content_memory): New.
8809 (get_content_index): New.
8810 (get_content_name): New.
8811 (display_content_memory): New.
8812 (update_content): New.
8813 (record_content): New.
8814 (already_contains): New.
8815 (insn_ok_now): Re-recog insns with virtual registers.
8816 (add_postponed_content_update): New.
8817 (process_postponed_content_update): New.
8818 (gen_and_emit_move): New.
8819 (transcode_memory_rtx): Record new location content.
8820 Use gen_and_emit_move.
8821 (force_into_acc): New.
8822 (move_to_acc): Use gen_and_emit_move.
8823 (move_from_acc): Likewise.
8824 (move_acc_to_reg): Likewise.
8825 (move_to_x): Likewise.
8826 (move_to_hl): Likewise.
8827 (move_to_de): Likewise.
8828 (rl78_alloc_physical_registers_op1): Record location content.
8829 (has_constraint): New.
8830 (rl78_alloc_physical_registers_op2): Record location content.
8832 (rl78_alloc_physical_registers_ro1): Likewise.
8833 (rl78_alloc_physical_registers_cmp): Likewise.
8834 (rl78_alloc_physical_registers_umul): Likewise.
8835 (rl78_alloc_address_registers_macax): New.
8836 (rl78_alloc_physical_registers): Initialize and set location
8837 content memory as needed.
8838 (rl78_reorg): Make sure split2 is called.
8839 (rl78_rtx_costs): New.
8841 2013-09-11 Richard Sandiford <rdsandiford@googlemail.com>
8843 * simplify-rtx.c (simplify_unary_operation_1): Use simplify_gen_binary
8844 for (not (neg ...)) and (neg (not ...)) cases.
8846 2013-09-11 Richard Biener <rguenther@suse.de>
8849 * passes.def: Split critical edges before late uninit warning passes.
8850 * tree-cfg.c (pass_split_crit_edges): Implement clone method.
8852 2013-09-11 Jakub Jelinek <jakub@redhat.com>
8854 PR tree-optimization/58385
8855 * fold-const.c (build_range_check): If both low and high are NULL,
8856 use omit_one_operand_loc to preserve exp side-effects.
8858 2013-09-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8860 * config/arm/arm.md (arm_shiftsi3): New alternative l/l/M.
8862 2013-09-11 Richard Biener <rguenther@suse.de>
8864 * tree-data-ref.c (dump_rdg_vertex, debug_rdg_vertex,
8865 dump_rdg_component, debug_rdg_component, dump_rdg, debug_rdg,
8866 dot_rdg_1, dot_rdg, rdg_vertex_for_stmt, create_rdg_edge_for_ddr,
8867 create_rdg_edges_for_scalar, create_rdg_edges, create_rdg_vertices,
8868 stmts_from_loop, known_dependences_p, build_empty_rdg,
8869 build_rdg, free_rdg, rdg_defs_used_in_other_loops_p): Move ...
8870 * tree-loop-distribution.c: ... here.
8871 * tree-data-ref.h (struct rdg_vertex, RDGV_STMT, RDGV_DATAREFS,
8872 RDGV_HAS_MEM_WRITE, RDGV_HAS_MEM_READS, RDG_STMT, RDG_DATAREFS,
8873 RDG_MEM_WRITE_STMT, RDG_MEM_READS_STMT, enum rdg_dep_type,
8874 struct rdg_edge, RDGE_TYPE, RDGE_LEVEL, RDGE_RELATION): Move ...
8875 * tree-loop-distribution.c: ... here.
8876 * tree-loop-distribution.c: Include gimple-pretty-print.h.
8877 (struct partition_s): Add loops member.
8878 (partition_alloc, partition_free, rdg_flag_uses, rdg_flag_vertex,
8879 rdg_flag_vertex_and_dependent, rdg_flag_loop_exits,
8880 build_rdg_partition_for_component, rdg_build_partitions): Adjust.
8882 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
8883 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
8884 Sergey Lega <sergey.s.lega@intel.com>
8885 Anna Tikhonova <anna.tikhonova@intel.com>
8886 Ilya Tocar <ilya.tocar@intel.com>
8887 Andrey Turetskiy <andrey.turetskiy@intel.com>
8888 Ilya Verbin <ilya.verbin@intel.com>
8889 Kirill Yukhin <kirill.yukhin@intel.com>
8890 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
8892 * config/i386/constraints.md (k): New.
8894 * config/i386/i386.c (const regclass_map): Add new mask registers.
8895 (dbx_register_map): Ditto.
8896 (dbx64_register_map): Ditto.
8897 (svr4_dbx_register_map): Ditto.
8898 (ix86_conditional_register_usage): Squash mask registers if AVX512F is
8900 (ix86_preferred_reload_class): Disable constants for mask registers.
8901 (ix86_secondary_reload): Do spill of mask register using 32-bit insn.
8902 (ix86_hard_regno_mode_ok): Support new mask registers.
8903 (x86_order_regs_for_local_alloc): Ditto.
8904 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Update.
8905 (FIXED_REGISTERS): Add new mask registers.
8906 (CALL_USED_REGISTERS): Ditto.
8907 (REG_ALLOC_ORDER): Ditto.
8908 (VALID_MASK_REG_MODE): New.
8909 (FIRST_MASK_REG): Ditto.
8910 (LAST_MASK_REG): Ditto.
8911 (reg_class): Add MASK_EVEX_REGS, MASK_REGS.
8912 (MAYBE_MASK_CLASS_P): New.
8913 (REG_CLASS_NAMES): Add MASK_EVEX_REGS, MASK_REGS.
8914 (REG_CLASS_CONTENTS): Ditto.
8915 (MASK_REGNO_P): New.
8916 (ANY_MASK_REG_P): Ditto.
8917 (HI_REGISTER_NAMES): Add new mask registers.
8918 * config/i386/i386.md (MASK0_REG, MASK1_REG, MASK2_REG, MASK3_REG,
8919 MASK4_REG, MASK5_REG, MASK6_REG, MASK7_REG): Constants for new
8921 (attribute "type"): Add mskmov, msklog.
8922 (attribute "length_immediate"): Support them.
8923 (attribute "memory"): Ditto.
8924 (attribute "prefix_0f"): Ditto.
8925 (*movhi_internal): Support new mask registers.
8926 (*movqi_internal): Ditto.
8927 (define_split): Split out clobber pattern is a logic
8928 insn on mask registers.
8929 (*k<logic><mode>): New.
8930 (*andhi_1): Extend to support mask regs.
8931 (*andqi_1): Extend to support mask regs.
8933 (define_split): Split and-not to and and not if operands
8935 (*<code><mode>_1): Separate HI mode to new pattern...
8936 (*<code>hi_1): This.
8937 (*<code>qi_1): Extend to support mask regs.
8939 (kortestzhi): Ditto.
8940 (kortestchi): Ditto.
8942 (*one_cmpl<mode>2_1): Remove HImode and handle it...
8943 (*one_cmplhi2_1): ...Here, now with mask registers support.
8944 (*one_cmplqi2_1): Support new mask registers.
8945 (HI/QImode arithmetics splitter): Don't split if mask registers
8947 (HI/QImode not splitter): Ditto.
8948 * config/i386/predicated.md (mask_reg_operand): New.
8949 (general_reg_operand): Ditto.
8951 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
8953 * doc/invoke.texi: Document fxsr, xsave and xsaveopt options.
8954 * doc/extend.texi: Document fxsr, xsave and xsaveopt builtins.
8956 2013-09-10 Jeff Law <law@redhat.com>
8958 PR tree-optimization/58380
8959 * tree-ssa-threadupdate.c (thread_block): Recognize another case
8960 of threading through a buried loop header.
8962 * tree-ssa-threadedge.c (thread_around_empty_blocks): Correct
8963 return value for single successor case.
8965 2013-09-10 Jan Hubicka <jh@suse.cz>
8967 * ipa-devirt.c (ipa_devirt): Enable with LTO.
8969 2013-09-10 Richard Earnshaw <rearnsha@arm.com>
8972 * arm/vfp.md (combine_vcvt_f32_<FCVTI32typename>): Fix pattern to
8973 support conditional execution.
8974 (combine_vcvt_f64_<FCVTI32typename>): Likewise.
8976 2013-09-10 Vladimir Makarov <vmakarov@redhat.com>
8978 * lra.c (lra): Clear lra_optional_reload_pseudos before every
8980 * lra-constraints.c (curr_insn_transform): Switch on optional reloads.
8981 Check destination too to check move insn.
8982 (undo_optional_reloads): Add check that the original peudo did not
8983 changed its allocation and the optional reload was inherited on last
8984 inheritance pass. Break loop after deciding to keep optional reload.
8985 (lra_undo_inheritance): Add check that inherited pseudo still in
8988 2013-09-10 James Greenhalgh <james.greenhalgh@arm.com>
8990 * config/aarch64/aarch64.md (generic_sched): New.
8991 * config/aarch64/aarch64-generic.md (load): Make conditional
8992 on generic_sched attribute.
8993 (nonload): Likewise.
8995 2013-09-10 Jan Hubicka <jh@suse.cz>
8997 * lto-cgraph.c: Include ipa-utils.h.
8998 (compute_ltrans_boundary): Also add possible targets into the boundary.
9000 2013-09-10 Jan Hubicka <jh@suse.cz>
9002 * gimple-fold.c (gimple_get_virt_method_for_binfo): Pass real
9003 VAR_DECL of vtable rather than full expression.
9005 2013-09-10 Jan Hubicka <jh@suse.cz>
9006 Paolo Carlini <paolo.carlini@oracle.com>
9008 * cgraphunit.c (analyze_functions): Save input_location, set it
9009 to UNKNOWN_LOCATION and restore it at the end.
9011 2013-09-10 Martin Jambor <mjambor@suse.cz>
9013 * ipa-cp.c (propagate_constants_topo): Do not ignore SCC
9014 represented by a thunk.
9016 2013-09-10 Jeff Law <law@redhat.com>
9018 PR tree-optimization/58343
9019 * tree-ssa-threadupdate.c (thread_block): Identify and disable
9020 jump threading requests through loop headers buried in the middle
9021 of a jump threading path.
9023 * tree-ssa-threadedge.c (thread_around_empty_blocks): Fix thinko
9024 in return value/type.
9026 2013-09-10 Jakub Jelinek <jakub@redhat.com>
9028 PR rtl-optimization/58365
9029 * cfgcleanup.c (merge_memattrs): Also clear MEM_READONLY_P
9030 resp. MEM_NOTRAP_P if they differ, or set MEM_VOLATILE_P if
9033 2013-09-10 Richard Biener <rguenther@suse.de>
9035 * tree-data-ref.h (build_rdg): Drop all parameters but loop.
9036 * tree-data-ref.c (create_rdg_vertices): Collect all data
9037 references, signal failure to the caller, use data-ref API.
9038 (build_rdg): Compute data references only once. Maintain lifetime
9039 of data references and data dependences from within RDG.
9040 (free_rdg): Free dependence relations.
9041 * tree-loop-distribution.c (rdg_flag_uses): Drop weird code
9042 inventing extra dependences.
9043 (distribute_loop): Update for RDG API changes.
9045 2013-09-10 Kai Tietz <ktietz@redhat.com>
9047 * doc/invoke.texi (fms-extensions): Document changed
9048 behavior for ms-abi targets.
9049 * config/i386/i386.c (ix86_option_override_internal):
9050 Set default value of option -fms-extension for ms-abi targets.
9052 2013-09-10 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
9054 * config/i386/i386.c (ix86_expand_movmem): Fix epilogue generation.
9056 2013-09-10 Alan Modra <amodra@gmail.com>
9059 * config/rs6000/rs6000.md (bswapdi2_64bit): Disable for volatile mems.
9061 2013-09-10 Alan Modra <amodra@gmail.com>
9063 * config/rs6000/predicates.md (add_cint_operand): New.
9064 (reg_or_add_cint_operand, small_toc_ref): Use add_cint_operand.
9065 * config/rs6000/rs6000.md (largetoc_high_plus): Restrict offset
9066 using add_cint_operand.
9067 (largetoc_high_plus_aix): Likewise.
9069 2013-09-09 Jakub Jelinek <jakub@redhat.com>
9071 PR tree-optimization/58364
9072 * tree-ssa-reassoc.c (init_range_entry): For BIT_NOT_EXPR on
9073 BOOLEAN_TYPE, only invert in_p and continue with arg0 if
9074 the current range can't be an unconditional true or false.
9076 2013-09-09 James Greenhalgh <james.greenhalgh@arm.com>
9078 * config/aarch64/arm_neon.h (vrsqrte_f64): Fix parameter type.
9080 2013-09-09 Uros Bizjak <ubizjak@gmail.com>
9082 * ipa-prop.c (ipa_modify_call_arguments): Initialize deref_align.
9084 2013-09-09 Paolo Carlini <paolo.carlini@oracle.com>
9087 * doc/invoke.texi (-Wdelete-incomplete): Document it.
9089 2013-09-09 Ian Bolton <ian.bolton@arm.com>
9091 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Return
9092 NO_REGS for immediate that can't be moved directly into FP_REGS.
9094 2013-09-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9096 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return CC_SWP for
9097 comparison with negated operand.
9098 * config/aarch64/aarch64.md (compare_neg<mode>): Match canonical
9101 2013-09-09 Richard Biener <rguenther@suse.de>
9104 * cfgloopmanip.c (fix_bb_placements): When fixing the placement
9105 of a subloop record all its block as affecting loop-closed SSA form.
9107 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
9109 * expmed.c (lshift_value): Take an unsigned HOST_WIDE_INT instead
9110 of an rtx/bitpos pair.
9111 (store_fixed_bit_field): Update accordingly.
9113 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
9115 * asan.c (asan_emit_stack_protection): Use gen_int_mode instead of
9117 * builtins.c (expand_errno_check): Likewise.
9118 * dwarf2cfi.c (init_return_column_size): Likewise.
9119 * except.c (sjlj_mark_call_sites): Likewise.
9120 * expr.c (move_by_pieces_1, store_by_pieces_2): Likewise.
9121 * lra-constraints.c (emit_inc): Likewise.
9122 * ree.c (combine_set_extension): Likewise.
9123 * regmove.c (fixup_match_2): Likewise.
9124 * reload1.c (inc_for_reload): Likewise.
9126 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
9128 * combine.c (simplify_set, expand_field_assignment, extract_left_shift)
9129 (force_to_mode, simplify_shift_const_1, simplify_comparison):
9130 Use gen_int_mode with the mode of the associated simplify_* call.
9131 * explow.c (probe_stack_range, anti_adjust_stack_and_probe): Likewise.
9132 * expmed.c (expand_shift_1): Likewise.
9133 * function.c (instantiate_virtual_regs_in_insn): Likewise.
9134 * loop-iv.c (iv_number_of_iterations): Likewise.
9135 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
9136 * simplify-rtx.c (simplify_binary_operation_1): Likewise.
9138 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
9140 * asan.c (asan_clear_shadow): Use gen_int_mode with the mode
9141 of the associated expand_* call.
9142 (asan_emit_stack_protection): Likewise.
9143 * builtins.c (round_trampoline_addr): Likewise.
9144 * explow.c (allocate_dynamic_stack_space, probe_stack_range): Likewise.
9145 * expmed.c (expand_smod_pow2, expand_sdiv_pow2, expand_divmod)
9146 (emit_store_flag): Likewise.
9147 * expr.c (emit_move_resolve_push, push_block, emit_single_push_insn_1)
9148 (emit_push_insn, optimize_bitfield_assignment_op, expand_expr_real_1):
9150 * function.c (instantiate_virtual_regs_in_insn): Likewise.
9151 * ifcvt.c (noce_try_store_flag_constants): Likewise.
9152 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
9153 * modulo-sched.c (generate_prolog_epilog): Likewise.
9154 * optabs.c (expand_binop, widen_leading, expand_doubleword_clz)
9155 (expand_ctz, expand_ffs, expand_unop): Likewise.
9157 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
9159 * alias.c (addr_side_effect_eval): Use gen_int_mode with the mode
9160 of the associated gen_rtx_* call.
9161 * caller-save.c (init_caller_save): Likewise.
9162 * combine.c (find_split_point, make_extraction): Likewise.
9163 (make_compound_operation): Likewise.
9164 * dwarf2out.c (mem_loc_descriptor): Likewise.
9165 * explow.c (plus_constant, probe_stack_range): Likewise.
9166 * expmed.c (expand_mult_const): Likewise.
9167 * expr.c (emit_single_push_insn_1, do_tablejump): Likewise.
9168 * reload1.c (init_reload): Likewise.
9169 * valtrack.c (cleanup_auto_inc_dec): Likewise.
9170 * var-tracking.c (adjust_mems): Likewise.
9171 * modulo-sched.c (sms_schedule): Likewise, but use gen_rtx_GT
9172 rather than gen_rtx_fmt_ee.
9174 2013-09-09 Jan Hubicka <jh@suse.cz>
9177 * value-prof.c (gimple_ic): Copy also abnormal edges.
9179 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
9181 * asan.c (asan_shadow_cst): Use gen_int_mode.
9183 2013-09-08 Jan Hubicka <jh@suse.cz>
9185 * ipa-profile.c: Add toplevel comment.
9186 (ipa_propagate_frequency_1): Be more conservative when profile is read.
9187 (contains_hot_call_p): New function.
9188 (ipa_propagate_frequency): Set frequencies based on counts when
9190 * predict.c (compute_function_frequency): Use PROFILE_READ gueard for
9191 profile; do not tamper with profile after inlining if it is read.
9193 2013-09-08 Jan Hubicka <jh@suse.cz>
9195 * ipa-prop.c (try_make_edge_direct_simple_call): Do not special case
9198 2013-09-08 Jan Hubicka <jh@suse.cz>
9200 * ipa.c (walk_polymorphic_call_targets): Fix redirection before IPA
9203 2013-09-08 Jeff Law <law@redhat.com>
9206 * tree-ssa-threadedge.c (thread_across_edge): Fix initialization
9209 2013-09-08 Andi Kleen <ak@linux.intel.com>
9211 * tree-inline.c (estimate_num_insns): Limit asm cost to 1000.
9213 2013-09-08 Jan Hubicka <jh@suse.cz>
9215 * ipa.c (walk_polymorphic_call_targets): Fix inliner summary update.
9217 2013-09-08 Richard Sandiford <rdsandiford@googlemail.com>
9219 * ira.c (update_equiv_regs): Only call set_paradoxical_subreg
9220 for non-debug insns.
9221 * lra.c (new_insn_reg): Take the containing insn as a parameter.
9222 Only modify lra_reg_info[].biggest_mode if it's non-debug insn.
9223 (collect_non_operand_hard_regs, add_regs_to_insn_regno_info): Update
9226 2013-09-08 Jan Hubicka <jh@suse.cz>
9228 * cgraphunit.c (walk_polymorphic_call_targets): Permit 0 possible
9229 targets and devirtualize to BUILT_IN_UNREACHABLE.
9230 * timevar.def (TV_IPA_UNREACHABLE): New timevar.
9231 * ipa.c (walk_polymorphic_call_targets): New function.
9232 (symtab_remove_unreachable_nodes): Use it; do not keep all virtual
9233 functions; use the new timevar.
9234 * ipa-devirt.c (maybe_record_node): Do not insert static nodes that
9235 was removed from the program.
9236 (record_binfo): If BINFO corresponds to an anonymous namespace, we may
9237 not consider it in the walk when its vtable is dead.
9238 (possible_polymorphic_call_targets_1): Pass anonymous flag to
9240 (devirt_variable_node_removal_hook): New function.
9241 (possible_polymorphic_call_targets): Also register
9242 devirt_variable_node_removal_hook.
9243 (ipa_devirt): Do not do non-speculative devirtualization.
9244 (gate_ipa_devirt): One execute if devirtualizing speculatively.
9246 2013-09-08 Jan Hubicka <jh@suse.cz>
9248 * cgraph.h (varpool_node_hook, varpool_node_hook_list,
9249 varpool_add_node_removal_hook, varpool_add_variable_insertion_hook,
9250 varpool_remove_variable_insertion_hook): Declare.
9251 * varpool.c (varpool_node_hook_list): New structure.
9252 (first_varpool_node_removal_hook,
9253 first_varpool_variable_insertion_hook): New variables.
9254 (varpool_add_node_removal_hook, varpool_remove_node_removal_hook,
9255 varpool_call_node_removal_hooks, varpool_add_variable_insertion_hook,
9256 varpool_remove_variable_insertion_hook,
9257 varpool_call_variable_insertion_hooks): New functions.
9258 (varpool_remove_node): Use it.
9260 2013-09-08 Paolo Carlini <paolo.carlini@oracle.com>
9263 * diagnostic.c (diagnostic_build_prefix): When s.file is
9264 "<built-in>" don't output line and column numbers.
9266 2013-09-06 Jan Hubicka <jh@suse.cz>
9268 * cgraphunit.c (expand_thunk): Get body before touching arguments.
9269 * lto-streamer-out.c: Stream thunks, too.
9270 * lto-streamer-in.c (input_function): Pop cfun here
9271 (lto_read_body): Instead of here.
9273 2013-09-06 Caroline Tice <cmtice@google.com>
9275 * doc/install.texi: Add documentation for the --enable-vtable-verify
9276 and the --disable-libvtv configure options.
9278 2013-09-06 Jeff Law <law@redhat.com>
9280 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
9281 edge implied equivalences into successor phis.
9283 2013-09-06 Joern Rennecke <joern.rennecke@embecosm.com>
9285 * resource.c (mark_referenced_resources): Handle COND_EXEC.
9287 2013-09-06 Claudiu Zissulescu <claziss@synopsys.com>
9289 * resource.c (mark_target_live_regs): Compute resources taking
9290 into account if a call is predicated or not.
9292 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
9294 * toplev.c (output_stack_usage): Be prepared for suffixes created by
9295 the compiler in the function names.
9297 2013-09-06 Jan Hubicka <jh@suse.cz>
9300 * ipa-inline.c (has_caller_p): New function.
9301 (want_inline_function_to_all_callers_p): Use it.
9302 (sum_callers, inline_to_all_callers): Break out from ...
9303 (ipa_inline): ... here.
9305 2013-09-06 Jan Hubicka <jh@suse.cz>
9307 * config/i386/i386.c (ix86_hard_regno_mode_ok): AVX modes are valid
9308 only when AVX is enabled.
9310 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9312 * config/aarch64/aarch64.md
9313 (*movtf_aarch64): Use neon_<ls>dm_2 as type where v8type
9314 is fpsimd_<load/store>2.
9315 (load_pair<mode>): Likewise.
9316 (store_pair<mode>): Likewise.
9318 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9320 * config/arm/types.md (type): Add "mrs" type.
9321 * config/aarch64/aarch64.md
9322 (aarch64_load_tp_hard): Make type "mrs".
9324 (load_tp_hard): Make type "mrs".
9325 * config/arm/cortex-a15.md: Update with new attributes.
9326 * config/arm/cortex-a5.md: Update with new attributes.
9327 * config/arm/cortex-a53.md: Update with new attributes.
9328 * config/arm/cortex-a7.md: Update with new attributes.
9329 * config/arm/cortex-a8.md: Update with new attributes.
9330 * config/arm/cortex-a9.md: Update with new attributes.
9331 * config/arm/cortex-m4.md: Update with new attributes.
9332 * config/arm/cortex-r4.md: Update with new attributes.
9333 * config/arm/fa526.md: Update with new attributes.
9334 * config/arm/fa606te.md: Update with new attributes.
9335 * config/arm/fa626te.md: Update with new attributes.
9336 * config/arm/fa726te.md: Update with new attributes.
9338 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9340 * config/aarch64/aarch64.md
9341 (*movti_aarch64): Use "multiple" for type where v8type is "move2".
9342 (*movtf_aarch64): Likewise.
9344 (thumb1_movdi_insn): Use "multiple" for type where more than one
9345 instruction is used for a move.
9346 (*arm32_movhf): Likewise.
9347 (*thumb_movdf_insn): Likewise.
9349 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9351 * config/arm/types.md (type): Rename fcpys to fmov.
9353 (*arm_movsi_vfp): Rename type fcpys as fmov.
9354 (*thumb2_movsi_vfp): Likewise
9355 (*movhf_vfp_neon): Likewise
9356 (*movhf_vfp): Likewise
9357 (*movsf_vfp): Likewise
9358 (*thumb2_movsf_vfp): Likewise
9359 (*movsfcc_vfp): Likewise
9360 (*thumb2_movsfcc_vfp): Likewise
9361 * config/aarch64/aarch64-simd.md
9362 (move_lo_quad_<mode>): Replace type mov_reg with fmovs.
9363 * config/aarch64/aarch64.md
9364 (*movsi_aarch64): Replace type mov_reg with fmovs.
9365 (*movdi_aarch64): Likewise
9366 (*movsf_aarch64): Likewise
9367 (*movdf_aarch64): Likewise
9369 (cortexa7_older_only): Rename TYPE_FCPYS to TYPE_FMOV.
9370 * config/arm/iwmmxt.md
9371 (*iwmmxt_movsi_insn): Rename type fcpys as fmov.
9372 * config/arm/arm1020e.md: Update with new attributes.
9373 * config/arm/cortex-a15-neon.md: Update with new attributes.
9374 * config/arm/cortex-a5.md: Update with new attributes.
9375 * config/arm/cortex-a53.md: Update with new attributes.
9376 * config/arm/cortex-a7.md: Update with new attributes.
9377 * config/arm/cortex-a8-neon.md: Update with new attributes.
9378 * config/arm/cortex-a9.md: Update with new attributes.
9379 * config/arm/cortex-m4-fpu.md: Update with new attributes.
9380 * config/arm/cortex-r4f.md: Update with new attributes.
9381 * config/arm/marvell-pj4.md: Update with new attributes.
9382 * config/arm/vfp11.md: Update with new attributes.
9384 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9386 * config/aarch64/aarch64.md
9387 (*madd<mode>): Fix type attribute.
9388 (*maddsi_uxtw): Likewise.
9389 (*msub<mode>): Likewise.
9390 (*msubsi_uxtw): Likewise.
9391 (<su_optab>maddsidi4): Likewise.
9392 (<su_optab>msubsidi4): Likewise.
9394 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9396 * config/arm/types.md: Split fdiv<sd> as fsqrt<sd>, fdiv<sd>.
9397 * config/arm/arm.md (core_cycles): Remove fdiv.
9398 * config/arm/vfp.md:
9399 (*sqrtsf2_vfp): Update for attribute changes.
9400 (*sqrtdf2_vfp): Likewise.
9401 * config/aarch64/aarch64.md:
9402 (sqrt<mode>2): Update for attribute changes.
9403 * config/arm/arm1020e.md: Update with new attributes.
9404 * config/arm/cortex-a15-neon.md: Update with new attributes.
9405 * config/arm/cortex-a5.md: Update with new attributes.
9406 * config/arm/cortex-a53.md: Update with new attributes.
9407 * config/arm/cortex-a7.md: Update with new attributes.
9408 * config/arm/cortex-a8-neon.md: Update with new attributes.
9409 * config/arm/cortex-a9.md: Update with new attributes.
9410 * config/arm/cortex-m4-fpu.md: Update with new attributes.
9411 * config/arm/cortex-r4f.md: Update with new attributes.
9412 * config/arm/marvell-pj4.md: Update with new attributes.
9413 * config/arm/vfp11.md: Update with new attributes.
9415 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9417 * config/arm/types.md
9418 (type): Split f_cvt as f_cvt, f_cvtf2i, f_cvti2f.
9419 * config/aarch64/aarch64.md
9420 (l<fcvt_pattern><su_optab><GPF:mode><GPI:mode>2): Update with
9422 (fix_trunc<GPF:mode><GPI:mode>2): Likewise.
9423 (fixuns_trunc<GPF:mode><GPI:mode>2): Likewise.
9424 (float<GPI:mode><GPF:mode>2): Likewise.
9426 (*truncsisf2_vfp): Update with new attributes.
9427 (*truncsidf2_vfp): Likewise.
9428 (fixuns_truncsfsi2): Likewise.
9429 (fixuns_truncdfsi2): Likewise.
9430 (*floatsisf2_vfp): Likewise.
9431 (*floatsidf2_vfp): Likewise.
9432 (floatunssisf2): Likewise.
9433 (floatunssidf2): Likewise.
9434 (*combine_vcvt_f32_<FCVTI32typename>): Likewise.
9435 (*combine_vcvt_f64_<FCVTI32typename>): Likewise.
9436 * config/arm/arm1020e.md: Update with new attributes.
9437 * config/arm/cortex-a15-neon.md: Update with new attributes.
9438 * config/arm/cortex-a5.md: Update with new attributes.
9439 * config/arm/cortex-a53.md: Update with new attributes.
9440 * config/arm/cortex-a7.md: Update with new attributes.
9441 * config/arm/cortex-a8-neon.md: Update with new attributes.
9442 * config/arm/cortex-a9.md: Update with new attributes.
9443 * config/arm/cortex-m4-fpu.md: Update with new attributes.
9444 * config/arm/cortex-r4f.md: Update with new attributes.
9445 * config/arm/marvell-pj4.md: Update with new attributes.
9446 * config/arm/vfp11.md: Update with new attributes.
9448 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9450 * config/aarch64/arm_neon.h
9451 (vqtbl<1,2,3,4><q>_s8): Fix control vector parameter type.
9452 (vqtbx<1,2,3,4><q>_s8): Likewise.
9454 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9456 * config/arm/types.md: Add "no_insn", "multiple" and "untyped" types.
9457 * config/arm/arm-fixed.md: Add type attribute to all insn patterns.
9458 (add<mode>3): Add type attribute.
9459 (add<mode>3): Likewise.
9460 (usadd<mode>3): Likewise.
9461 (ssadd<mode>3): Likewise.
9462 (sub<mode>3): Likewise.
9463 (sub<mode>3): Likewise.
9464 (ussub<mode>3): Likewise.
9465 (sssub<mode>3): Likewise.
9466 (ssmulsa3): Likewise.
9467 (usmulusa3): Likewise.
9468 (arm_usatsihi): Likewise.
9470 (*movdi_vfp): Add types for all instructions.
9471 (*movdi_vfp_cortexa8): Likewise.
9472 (*movhf_vfp_neon): Likewise.
9473 (*movhf_vfp): Likewise.
9474 (*movdf_vfp): Likewise.
9475 (*thumb2_movdf_vfp): Likewise.
9476 (*thumb2_movdfcc_vfp): Likewise.
9477 * config/arm/arm.md: Add type attribute to all insn patterns.
9478 (*thumb1_adddi3): Add type attribute.
9479 (*arm_adddi3): Likewise.
9480 (*adddi_sesidi_di): Likewise.
9481 (*adddi_zesidi_di): Likewise.
9482 (*thumb1_addsi3): Likewise.
9483 (addsi3_compare0): Likewise.
9484 (*addsi3_compare0_scratch): Likewise.
9485 (*compare_negsi_si): Likewise.
9486 (cmpsi2_addneg): Likewise.
9487 (*addsi3_carryin_<optab>): Likewise.
9488 (*addsi3_carryin_alt2_<optab>): Likewise.
9489 (*addsi3_carryin_clobercc_<optab>): Likewise.
9490 (*subsi3_carryin): Likewise.
9491 (*subsi3_carryin_const): Likewise.
9492 (*subsi3_carryin_compare): Likewise.
9493 (*subsi3_carryin_compare_const): Likewise.
9494 (*arm_subdi3): Likewise.
9495 (*thumb_subdi3): Likewise.
9496 (*subdi_di_zesidi): Likewise.
9497 (*subdi_di_sesidi): Likewise.
9498 (*subdi_zesidi_di): Likewise.
9499 (*subdi_sesidi_di): Likewise.
9500 (*subdi_zesidi_ze): Likewise.
9501 (thumb1_subsi3_insn): Likewise.
9502 (*arm_subsi3_insn): Likewise.
9503 (*anddi3_insn): Likewise.
9504 (*anddi_zesidi_di): Likewise.
9505 (*anddi_sesdi_di): Likewise.
9506 (*ne_zeroextracts): Likewise.
9507 (*ne_zeroextracts): Likewise.
9508 (*ite_ne_zeroextr): Likewise.
9509 (*ite_ne_zeroextr): Likewise.
9510 (*anddi_notdi_di): Likewise.
9511 (*anddi_notzesidi): Likewise.
9512 (*anddi_notsesidi): Likewise.
9513 (andsi_notsi_si): Likewise.
9514 (thumb1_bicsi3): Likewise.
9515 (*iordi3_insn): Likewise.
9516 (*iordi_zesidi_di): Likewise.
9517 (*iordi_sesidi_di): Likewise.
9518 (*thumb1_iorsi3_insn): Likewise.
9519 (*xordi3_insn): Likewise.
9520 (*xordi_zesidi_di): Likewise.
9521 (*xordi_sesidi_di): Likewise.
9522 (*arm_xorsi3): Likewise.
9523 (*andsi_iorsi3_no): Likewise.
9524 (*smax_0): Likewise.
9525 (*smax_m1): Likewise.
9526 (*arm_smax_insn): Likewise.
9527 (*smin_0): Likewise.
9528 (*arm_smin_insn): Likewise.
9529 (*arm_umaxsi3): Likewise.
9530 (*arm_uminsi3): Likewise.
9531 (*minmax_arithsi): Likewise.
9532 (*minmax_arithsi_): Likewise.
9533 (*satsi_<SAT:code>): Likewise.
9534 (arm_ashldi3_1bit): Likewise.
9535 (arm_ashrdi3_1bit): Likewise.
9536 (arm_lshrdi3_1bit): Likewise.
9537 (*arm_negdi2): Likewise.
9538 (*thumb1_negdi2): Likewise.
9539 (*arm_negsi2): Likewise.
9540 (*thumb1_negsi2): Likewise.
9541 (*negdi_extendsid): Likewise.
9542 (*negdi_zero_extend): Likewise.
9543 (*arm_abssi2): Likewise.
9544 (*thumb1_abssi2): Likewise.
9545 (*arm_neg_abssi2): Likewise.
9546 (*thumb1_neg_abss): Likewise.
9547 (one_cmpldi2): Likewise.
9548 (extend<mode>di2): Likewise.
9549 (*compareqi_eq0): Likewise.
9550 (*arm_extendhisi2addsi): Likewise.
9551 (*arm_movdi): Likewise.
9552 (*thumb1_movdi_insn): Likewise.
9553 (*arm_movt): Likewise.
9554 (*thumb1_movsi_insn): Likewise.
9555 (pic_add_dot_plus_four): Likewise.
9556 (pic_add_dot_plus_eight): Likewise.
9557 (tls_load_dot_plus_eight): Likewise.
9558 (*thumb1_movhi_insn): Likewise.
9559 (*thumb1_movsf_insn): Likewise.
9560 (*movdf_soft_insn): Likewise.
9561 (*thumb_movdf_insn): Likewise.
9562 (cbranchsi4_insn): Likewise.
9563 (cbranchsi4_scratch): Likewise.
9564 (*negated_cbranchsi4): Likewise.
9565 (*tbit_cbranch): Likewise.
9566 (*tlobits_cbranch): Likewise.
9567 (*tstsi3_cbranch): Likewise.
9568 (*cbranchne_decr1): Likewise.
9569 (*addsi3_cbranch): Likewise.
9570 (*addsi3_cbranch_scratch): Likewise.
9571 (*arm_cmpdi_insn): Likewise.
9572 (*arm_cmpdi_unsig): Likewise.
9573 (*arm_cmpdi_zero): Likewise.
9574 (*thumb_cmpdi_zero): Likewise.
9575 (*deleted_compare): Likewise.
9576 (*mov_scc): Likewise.
9577 (*mov_negscc): Likewise.
9578 (*mov_notscc): Likewise.
9579 (*cstoresi_eq0_thumb1_insn): Likewise.
9580 (cstoresi_nltu_thumb1): Likewise.
9581 (cstoresi_ltu_thu): Likewise.
9582 (thumb1_addsi3_addgeu): Likewise.
9583 (*arm_jump): Likewise.
9584 (*thumb_jump): Likewise.
9585 (*check_arch2): Likewise.
9586 (arm_casesi_internal): Likewise.
9587 (thumb1_casesi_dispatch): Likewise.
9588 (*arm_indirect_jump): Likewise.
9589 (*thumb1_indirect_jump): Likewise.
9591 (*and_scc): Likewise.
9592 (*ior_scc): Likewise.
9593 (*compare_scc): Likewise.
9594 (*cond_move): Likewise.
9595 (*cond_arith): Likewise.
9596 (*cond_sub): Likewise.
9597 (*cmp_ite0): Likewise.
9598 (*cmp_ite1): Likewise.
9599 (*cmp_and): Likewise.
9600 (*cmp_ior): Likewise.
9601 (*ior_scc_scc): Likewise.
9602 (*ior_scc_scc_cmp): Likewise.
9603 (*and_scc_scc): Likewise.
9604 (*and_scc_scc_cmp): Likewise.
9605 (*and_scc_scc_nod): Likewise.
9606 (*negscc): Likewise.
9607 (movcond_addsi): Likewise.
9608 (movcond): Likewise.
9609 (*ifcompare_plus_move): Likewise.
9610 (*if_plus_move): Likewise.
9611 (*ifcompare_move_plus): Likewise.
9612 (*if_move_plus): Likewise.
9613 (*ifcompare_arith_arith): Likewise.
9614 (*if_arith_arith): Likewise.
9615 (*ifcompare_arith_move): Likewise.
9616 (*if_arith_move): Likewise.
9617 (*ifcompare_move_arith): Likewise.
9618 (*if_move_arith): Likewise.
9619 (*ifcompare_move_not): Likewise.
9620 (*if_move_not): Likewise.
9621 (*ifcompare_not_move): Likewise.
9622 (*if_not_move): Likewise.
9623 (*ifcompare_shift_move): Likewise.
9624 (*if_shift_move): Likewise.
9625 (*ifcompare_move_shift): Likewise.
9626 (*if_move_shift): Likewise.
9627 (*ifcompare_shift_shift): Likewise.
9628 (*ifcompare_not_arith): Likewise.
9629 (*ifcompare_arith_not): Likewise.
9630 (*if_arith_not): Likewise.
9631 (*ifcompare_neg_move): Likewise.
9632 (*if_neg_move): Likewise.
9633 (*ifcompare_move_neg): Likewise.
9634 (*if_move_neg): Likewise.
9635 (prologue_thumb1_interwork): Likewise.
9636 (*cond_move_not): Likewise.
9637 (*sign_extract_onebit): Likewise.
9638 (*not_signextract_onebit): Likewise.
9639 (stack_tie): Likewise.
9640 (align_4): Likewise.
9641 (align_8): Likewise.
9642 (consttable_end): Likewise.
9643 (consttable_1): Likewise.
9644 (consttable_2): Likewise.
9645 (consttable_4): Likewise.
9646 (consttable_8): Likewise.
9647 (consttable_16): Likewise.
9648 (*thumb1_tablejump): Likewise.
9649 (prefetch): Likewise.
9650 (force_register_use): Likewise.
9651 (thumb_eh_return): Likewise.
9652 (load_tp_hard): Likewise.
9653 (load_tp_soft): Likewise.
9654 (tlscall): Likewise.
9655 (*arm_movtas_ze): Likewise.
9656 (*arm_rev): Likewise.
9657 (*arm_revsh): Likewise.
9658 (*arm_rev16): Likewise.
9659 * config/arm/thumb2.md
9660 (*thumb2_smaxsi3): Likewise.
9661 (*thumb2_sminsi3): Likewise.
9662 (*thumb32_umaxsi3): Likewise.
9663 (*thumb2_uminsi3): Likewise.
9664 (*thumb2_negdi2): Likewise.
9665 (*thumb2_abssi2): Likewise.
9666 (*thumb2_neg_abss): Likewise.
9667 (*thumb2_movsi_insn): Likewise.
9668 (tls_load_dot_plus_four): Likewise.
9669 (*thumb2_movhi_insn): Likewise.
9670 (*thumb2_mov_scc): Likewise.
9671 (*thumb2_mov_negs): Likewise.
9672 (*thumb2_mov_negs): Likewise.
9673 (*thumb2_mov_nots): Likewise.
9674 (*thumb2_mov_nots): Likewise.
9675 (*thumb2_movsicc_): Likewise.
9676 (*thumb2_movsfcc_soft_insn): Likewise.
9677 (*thumb2_indirect_jump): Likewise.
9678 (*thumb2_and_scc): Likewise.
9679 (*thumb2_ior_scc): Likewise.
9680 (*thumb2_ior_scc_strict_it): Likewise.
9681 (*thumb2_cond_move): Likewise.
9682 (*thumb2_cond_arith): Likewise.
9683 (*thumb2_cond_ari): Likewise.
9684 (*thumb2_cond_sub): Likewise.
9685 (*thumb2_negscc): Likewise.
9686 (*thumb2_movcond): Likewise.
9687 (thumb2_casesi_internal): Likewise.
9688 (thumb2_casesi_internal_pic): Likewise.
9689 (*thumb2_alusi3_short): Likewise.
9690 (*thumb2_mov<mode>_shortim): Likewise.
9691 (*thumb2_addsi_short): Likewise.
9692 (*thumb2_subsi_short): Likewise.
9693 (thumb2_addsi3_compare0): Likewise.
9694 (*thumb2_cbz): Likewise.
9695 (*thumb2_cbnz): Likewise.
9696 (*thumb2_one_cmplsi2_short): Likewise.
9697 (*thumb2_negsi2_short): Likewise.
9698 (*orsi_notsi_si): Likewise.
9699 * config/arm/arm1020e.md: Update with new attributes.
9700 * config/arm/arm1026ejs.md: Update with new attributes.
9701 * config/arm/arm1136jfs.md: Update with new attributes.
9702 * config/arm/arm926ejs.md: Update with new attributes.
9703 * config/arm/cortex-a15.md: Update with new attributes.
9704 * config/arm/cortex-a5.md: Update with new attributes.
9705 * config/arm/cortex-a53.md: Update with new attributes.
9706 * config/arm/cortex-a7.md: Update with new attributes.
9707 * config/arm/cortex-a8.md: Update with new attributes.
9708 * config/arm/cortex-a9.md: Update with new attributes.
9709 * config/arm/cortex-m4.md: Update with new attributes.
9710 * config/arm/cortex-r4.md: Update with new attributes.
9711 * config/arm/fa526.md: Update with new attributes.
9712 * config/arm/fa606te.md: Update with new attributes.
9713 * config/arm/fa626te.md: Update with new attributes.
9714 * config/arm/fa726te.md: Update with new attributes.
9716 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9718 * config/aarch64/aarch64-simd.md
9719 (aarch64_sqdml<SBINQOPS:as>l_n<mode>_internal): Use
9720 <vwx> iterator to ensure correct register choice.
9721 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
9722 (aarch64_sqdmull_n<mode>): Likewise.
9723 (aarch64_sqdmull2_n<mode>_internal): Likewise.
9724 * config/aarch64/arm_neon.h
9725 (vml<as><q>_lane<q>_<su>16): Use 'x' constraint for element vector.
9726 (vml<as><q>_n_<su>16): Likewise.
9727 (vml<as>l_high_lane<q>_<su>16): Likewise.
9728 (vml<as>l_high_n_<su>16): Likewise.
9729 (vml<as>l_lane<q>_<su>16): Likewise.
9730 (vml<as>l_n_<su>16): Likewise.
9731 (vmul<q>_lane<q>_<su>16): Likewise.
9732 (vmul<q>_n_<su>16): Likewise.
9733 (vmull_lane<q>_<su>16): Likewise.
9734 (vmull_n_<su>16): Likewise.
9735 (vmull_high_lane<q>_<su>16): Likewise.
9736 (vmull_high_n_<su>16): Likewise.
9737 (vqrdmulh<q>_n_s16): Likewise.
9739 2013-09-06 Tejas Belagod <tejas.belagod@arm.com>
9741 * config/aarch64/arm_neon.h: Fix all vdup<bhsd_lane<q> intrinsics to
9742 have the correct lane parameter.
9744 2013-09-06 Richard Biener <rguenther@suse.de>
9746 * cfganal.c (control_dependences::~control_dependences):
9747 Properly free all of the vector.
9749 2013-09-06 Kirill Yukhin <kirill.yukhin@intel.com>
9752 * config/i386/i386.c (ix86_conditional_register_usage):
9753 Proper initialize extended SSE registers.
9755 2013-09-06 Jan Hubicka <jh@suse.cz>
9757 PR tree-optimization/58311
9758 * ipa-devirt.c (gate_ipa_devirt): Only execute when optimizing.
9760 2013-09-06 Jan Hubicka <jh@suse.cz>
9762 * Makefile.in (tree-sra.o): Update dependencies.
9763 * tree-sra.c: Include ipa-utils.h
9764 (scan_function): Use recursive_call_p.
9765 (has_caller_p): New function.
9766 (cgraph_for_node_and_aliases): Count also callers of aliases.
9768 2013-09-06 Jan Hubicka <jh@suse.cz>
9771 * cgraph.h (symtab_semantically_equivalent_p): Declare.
9772 * tree-tailcall.c: Include ipa-utils.h.
9773 (find_tail_calls): Use it.
9774 * ipa-pure-const.c (check_call): Likewise.
9775 * ipa-utils.c (recursive_call_p): New function.
9776 * ipa-utils.h (recursive_call_p): Dclare.
9777 * symtab.c (symtab_nonoverwritable_alias): Fix formatting.
9778 (symtab_semantically_equivalent_p): New function.
9779 * Makefile.in (tree-tailcall.o): Update dependencies.
9781 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
9783 * ipa-split.c (split_function): Set DECL_NO_INLINE_WARNING_P on the
9786 2013-09-06 Richard Biener <rguenther@suse.de>
9788 * lto-streamer.h (lto_global_var_decls): Remove.
9789 * Makefile.in (OBJS): Remove lto-symtab.o.
9790 (lto-symtab.o): Remove.
9791 (GTFILES): Remove lto-symtab.c
9792 * lto-symtab.c: Move to lto/
9794 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9796 * config/s390/s390.md (UNSPEC_FPINT_FLOOR, UNSPEC_FPINT_BTRUNC)
9797 (UNSPEC_FPINT_ROUND, UNSPEC_FPINT_CEIL, UNSPEC_FPINT_NEARBYINT)
9798 (UNSPEC_FPINT_RINT): New constant definitions.
9799 (FPINT, fpint_name, fpint_roundingmode): New integer iterator
9800 definition with 2 attributes.
9801 ("<FPINT:fpint_name><BFP:mode>2", "rint<BFP:mode>2")
9802 ("<FPINT:fpint_name><DFP:mode>2", "rint<DFP:mode>2"): New pattern
9805 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9807 * config/s390/s390.md: Add "bcr_flush" value to mnemonic attribute.
9808 ("mem_thread_fence_1"): Use bcr 14,0 for z196 and later.
9809 Set the mnemonic attribute to "bcr_flush". Set the "z196prop"
9810 attribute to "z196_alone".
9811 * config/s390/2827.md: Add "bcr_flush" to "ooo_groupalone" and
9814 2013-09-06 Richard Biener <rguenther@suse.de>
9816 * basic-block.h (class control_dependences): New.
9817 * tree-ssa-dce.c (control_dependence_map): Remove.
9819 (EXECUTE_IF_CONTROL_DEPENDENT): Remove.
9820 (set_control_dependence_map_bit, clear_control_dependence_bitmap,
9821 find_pdom, find_control_dependence, find_all_control_dependences):
9823 (mark_control_dependent_edges_necessary,
9824 find_obviously_necessary_stmts, propagate_necessity, tree_dce_init,
9825 tree_dce_done, perform_tree_ssa_dce): Adjust.
9826 * cfganal.c (set_control_dependence_map_bit,
9827 clear_control_dependence_bitmap, find_pdom, find_control_dependence,
9828 find_all_control_dependences): Move from tree-ssa-dce.c and
9829 implement as methods of control_dependences class.
9830 (control_dependences::control_dependences): New.
9831 (control_dependences::~control_dependences): Likewise.
9832 (control_dependences::get_edges_dependent_on): Likewise.
9833 (control_dependences::get_edge): Likewise.
9835 2013-09-04 Jan Hubicka <jh@suse.cz>
9837 * tree.c (types_same_for_odr): Drop overactive check.
9838 * ipa-devirt.c (hash_type_name): Likewise.
9840 2013-09-04 Jan Hubicka <jh@suse.cz>
9842 * cgraphunit.c (walk_polymorphic_call_targets): Break out from ...
9843 (analyze_functions): ... here.
9845 2013-09-04 Jan Hubicka <jh@suse.cz>
9848 * cgraphunit.c (analyze_functions): Clear AUX fields
9849 after processing; initialize assembler name has.
9851 2013-09-05 Jeff Law <law@redhat.com>
9853 * tree-ssa-threadedge.c (thread_around_empty_blocks): Renamed
9854 from thread_around_empty_block. Record threading path into PATH.
9855 Recurse if threading through the initial block is successful.
9856 (thread_across_edge): Corresponding changes to slightly simplify.
9858 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
9860 * config/aarch64/aarch64.md
9861 (type): Remove frecpe, frecps, frecpx.
9862 (aarch64_frecp<FRECP:frecp_suffix><mode>): Move to aarch64-simd.md,
9863 fix to be a TARGET_SIMD instruction.
9864 (aarch64_frecps): Remove.
9865 * config/aarch64/aarch64-simd.md
9866 (aarch64_frecp<FRECP:frecp_suffix><mode>): New, moved from aarch64.md
9867 (aarch64_frecps<mode>): Handle all float/vector of float modes.
9869 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
9870 Sofiane Naci <sofiane.naci@arm.com>
9872 * config/arm/types.md (define_attr "type"): Expand "arlo_imm"
9873 into "adr", "alu_imm", "alus_imm", "logic_imm", "logics_imm".
9874 Expand "arlo_reg" into "adc_reg", "adc_imm", "adcs_reg", "adcs_imm",
9875 "alu_ext", "alu_reg", "alus_ext", "alus_reg", "bfm", "csel",
9876 "logic_reg", "logics_reg", "rev". Expand "arlo_shift" into
9877 "alu_shift_imm", "alus_shift_imm", "logic_shift_imm",
9878 "logics_shift_imm". Expand "arlo_shift_reg" into "alu_shift_reg",
9879 "alus_shift_reg", "logic_shift_reg", "logics_shift_reg". Expand "clz"
9880 into "clz, "rbit". Rename "shift" to "shift_imm".
9881 * config/arm/arm.md (define_attr "core_cycles"): Update for attribute
9882 changes. Update for attribute changes all occurrences of arlo_* and
9884 * config/arm/arm-fixed.md: Update for attribute changes
9885 all occurrences of arlo_* types.
9886 * config/arm/thumb2.md: Update for attribute changes all occurrences
9888 * config/arm/arm.c (xscale_sched_adjust_cost): (rtx insn, rtx
9889 (cortexa7_older_only): Likewise.
9890 (cortexa7_younger): Likewise.
9891 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
9892 (1020alu_shift_op): Likewise.
9893 (1020alu_shift_reg_op): Likewise.
9894 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
9895 (alu_shift_op): Likewise.
9896 (alu_shift_reg_op): Likewise.
9897 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
9898 (11_alu_shift_op): Likewise.
9899 (11_alu_shift_reg_op): Likewise.
9900 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
9901 (9_alu_shift_reg_op): Likewise.
9902 * config/arm/cortex-a15.md (cortex_a15_alu): Update for
9904 (cortex_a15_alu_shift): Likewise.
9905 (cortex_a15_alu_shift_reg): Likewise.
9906 * config/arm/cortex-a5.md (cortex_a5_alu): Update for
9908 (cortex_a5_alu_shift): Likewise.
9909 * config/arm/cortex-a53.md (cortex_a53_alu): Update for
9911 (cortex_a53_alu_shift): Likewise.
9912 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for
9914 (cortex_a7_alu_reg): Likewise.
9915 (cortex_a7_alu_shift): Likewise.
9916 * config/arm/cortex-a8.md (cortex_a8_alu): Update for
9918 (cortex_a8_alu_shift): Likewise.
9919 (cortex_a8_alu_shift_reg): Likewise.
9920 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
9921 (cortex_a9_dp_shift): Likewise.
9922 * config/arm/cortex-m4.md (cortex_m4_alu): Update for
9924 * config/arm/cortex-r4.md
9925 (cortex_r4_alu): Update for attribute changes.
9926 (cortex_r4_mov): Likewise.
9927 (cortex_r4_alu_shift_reg): Likewise.
9928 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
9929 (526_alu_shift_op): Likewise.
9930 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
9931 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
9932 (626te_alu_shift_op): Likewise.
9933 * config/arm/fa726te.md (726te_alu_op): Update for attribute changes.
9934 (726te_alu_shift_op): Likewise.
9935 (726te_alu_shift_reg_op): Likewise.
9936 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
9937 (mp626_alu_shift_op): Likewise.
9938 * config/arm/marvell-pj4.md (pj4_alu): Update for attribute changes.
9939 (pj4_alu_conds): Likewise.
9940 (pj4_shift): Likewise.
9941 (pj4_shift_conds): Likewise.
9942 (pj4_alu_shift): Likewise.
9943 (pj4_alu_shift_conds): Likewise.
9944 * config/aarch64/aarch64.md: Update for attribute change
9945 all occurrences of arlo_* and shift* types.
9947 2013-09-05 Mike Stump <mikestump@comcast.net>
9949 * tree.h: Move documentation for tree_function_decl to tree-core.h
9950 with the declaration.
9952 2013-09-05 Peter Bergner <bergner@vnet.ibm.com>
9955 * reginfo.c (choose_hard_reg_mode): Scan through all mode classes
9956 looking for widest mode.
9958 2013-09-05 Eric Botcazou <ebotcazou@adacore.com>
9960 * config.gcc (*-*-vxworks*): Do not override an existing extra_objs.
9962 2013-09-05 Richard Biener <rguenther@suse.de>
9964 PR tree-optimization/58137
9965 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size):
9966 Do not create vectors of pointers.
9967 * tree-vect-loop.c (get_initial_def_for_induction): Use proper
9968 types for the components of the vector initializer.
9969 * tree-cfg.c (verify_gimple_assign_binary): Remove special-casing
9970 allowing pointer vectors with PLUS_EXPR/MINUS_EXPR.
9972 2013-09-05 Martin Jambor <mjambor@suse.cz>
9974 * ipa-prop.c (remove_described_reference): Accept missing references,
9975 return false if that hppens, otherwise return true.
9976 (cgraph_node_for_jfunc): New function.
9977 (try_decrement_rdesc_refcount): Likewise.
9978 (try_make_edge_direct_simple_call): Use them.
9979 (ipa_edge_removal_hook): Remove references from rdescs.
9980 (ipa_edge_duplication_hook): Clone rdescs and their references
9981 when the new edge has the same caller as the old one.
9982 * cgraph.c (cgraph_resolve_speculation): Remove speculative
9983 reference before removing any edges.
9985 2013-09-05 Richard Earnshaw <rearnsha@arm.com>
9987 * arm.c (thumb2_emit_strd_push): Rewrite to use pre-decrement on
9989 * thumb2.md (thumb2_storewb_parisi): New pattern.
9991 2013-09-05 Yufeng Zhang <yufeng.zhang@arm.com>
9993 * config/aarch64/aarch64-option-extensions.def: Add
9994 AARCH64_OPT_EXTENSION of 'crc'.
9995 * config/aarch64/aarch64.h (AARCH64_FL_CRC): New define.
9996 (AARCH64_ISA_CRC): Ditto.
9997 * doc/invoke.texi (-march and -mcpu feature modifiers): Add
9998 description of the CRC extension.
10000 2013-09-05 Alexander Ivchenko <alexander.ivchenko@intel.com>
10002 * config/rs6000/linux64.h: Define OPTION_BIONIC and OPTION_UCLIBC.
10003 * config/rs6000/linux.h: Ditto.
10004 * alpha/linux.h: Ditto.
10005 * config/bfin/uclinux.h: Define TARGET_LIBC_HAS_FUNCTION as
10006 no_c99_libc_has_function.
10007 * config/c6x/uclinux-elf.h: Ditto.
10008 * config/lm32/uclinux-elf.h: Ditto.
10009 * config/m68k/uclinux.h: Ditto.
10010 * config/moxie/uclinux.h: Ditto.
10011 * config.gcc (bfin*-linux-uclibc*): Add t-linux-android to tmake_file.
10012 (crisv32-*-linux*, cris-*-linux*): Ditto.
10013 * config/bfin/bfin.c: Include "tm_p.h".
10015 2013-09-05 Richard Biener <rguenther@suse.de>
10017 * tree-vect-loop.c (vect_analyze_loop_operations): Properly
10018 check for a definition without a basic-block.
10020 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
10021 Sofiane Naci <sofiane.naci@arm.com>
10023 * config/aarch64/aarch64.md
10024 (*movti_aarch64): Rename r_2_f and f_2_r.
10025 (*movsf_aarch64): Likewise.
10026 (*movdf_aarch64): Likewise.
10027 (*movtf_aarch64): Likewise.
10028 (aarch64_movdi_<mode>low): Likewise.
10029 (aarch64_movdi_<mode>high): Likewise.
10030 (aarch64_mov<mode>high_di): Likewise.
10031 (aarch64_mov<mode>low_di): Likewise.
10032 (aarch64_movtilow_tilow): Likewise.
10033 * config/arm/arm.md (attribute "neon_type"): Delete. Move attribute
10034 values to config/arm/types.md
10035 (attribute "conds"): Update for attribute change.
10036 (anddi3_insn): Likewise.
10037 (iordi3_insn): Likewise.
10038 (xordi3_insn): Likewise.
10039 (one_cmpldi2): Likewise.
10040 * config/arm/types.md (type): Add Neon types.
10041 * config/arm/neon.md (neon_mov<mode>): Remove "neon_type" attribute,
10042 use "type" attribute.
10043 (movmisalign<mode>_neon_store): Likewise.
10044 (movmisalign<mode>_neon_load): Likewise.
10045 (vec_set<mode>_internal): Likewise.
10046 (vec_setv2di_internal): Likewise.
10047 (vec_extract<mode>): Likewise.
10048 (vec_extractv2di): Likewise.
10049 (add<mode>3_neon): Likewise.
10050 (adddi3_neon): Likewise.
10051 (sub<mode>3_neon): Likewise.
10052 (subdi3_neon): Likewise.
10053 (mul<mode>3_neon): Likewise.
10054 (mul<mode>3add<mode>_neon): Likewise.
10055 (mul<mode>3neg<mode>add<mode>_neon): Likewise.
10056 (fma<VCVTF:mode>4)): Likewise.
10057 (fma<VCVTF:mode>4_intrinsic): Likewise.
10058 (fmsub<VCVTF:mode>4)): Likewise.
10059 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
10060 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
10061 (ior<mode>3): Likewise.
10062 (and<mode>3): Likewise.
10063 (anddi3_neon): Likewise.
10064 (orn<mode>3_neon): Likewise.
10065 (orndi3_neon): Likewise.
10066 (bic<mode>3_neon): Likewise.
10067 (bicdi3_neon): Likewise.
10068 (xor<mode>3): Likewise.
10069 (one_cmpl<mode>2): Likewise.
10070 (abs<mode>2): Likewise.
10071 (neg<mode>2): Likewise.
10072 (umin<mode>3_neon): Likewise.
10073 (umax<mode>3_neon): Likewise.
10074 (smin<mode>3_neon): Likewise.
10075 (smax<mode>3_neon): Likewise.
10076 (vashl<mode>3): Likewise.
10077 (vashr<mode>3_imm): Likewise.
10078 (vlshr<mode>3_imm): Likewise.
10079 (ashl<mode>3_signed): Likewise.
10080 (ashl<mode>3_unsigned): Likewise.
10081 (neon_load_count): Likewise.
10082 (ashldi3_neon_noclobber): Likewise.
10083 (signed_shift_di3_neon): Likewise.
10084 (unsigned_shift_di3_neon): Likewise.
10085 (ashrdi3_neon_imm_noclobber): Likewise.
10086 (lshrdi3_neon_imm_noclobber): Likewise.
10087 (widen_ssum<mode>3): Likewise.
10088 (widen_usum<mode>3): Likewise.
10089 (quad_halves_<code>v4si): Likewise.
10090 (quad_halves_<code>v4sf): Likewise.
10091 (quad_halves_<code>v8hi): Likewise.
10092 (quad_halves_<code>v16qi): Likewise.
10093 (reduc_splus_v2di): Likewise.
10094 (neon_vpadd_internal<mode>): Likewise.
10095 (neon_vpsmin<mode>): Likewise.
10096 (neon_vpsmax<mode>): Likewise.
10097 (neon_vpumin<mode>): Likewise.
10098 (neon_vpumax<mode>): Likewise.
10099 (ss_add<mode>_neon): Likewise.
10100 (us_add<mode>_neon): Likewise.
10101 (ss_sub<mode>_neon): Likewise.
10102 (us_sub<mode>_neon): Likewise.
10103 (neon_vadd<mode>_unspec): Likewise.
10104 (neon_vaddl<mode>): Likewise.
10105 (neon_vaddw<mode>): Likewise.
10106 (neon_vhadd<mode>): Likewise.
10107 (neon_vqadd<mode>): Likewise.
10108 (neon_vaddhn<mode>): Likewise.
10109 (neon_vmul<mode>): Likewise.
10110 (neon_vmla<mode>): Likewise.
10111 (neon_vmlal<mode>): Likewise.
10112 (neon_vmls<mode>): Likewise.
10113 (neon_vmlsl<mode>): Likewise.
10114 (neon_vqdmulh<mode>): Likewise.
10115 (neon_vqdmlal<mode>): Likewise.
10116 (neon_vqdmlsl<mode>): Likewise.
10117 (neon_vmull<mode>): Likewise.
10118 (neon_vqdmull<mode>): Likewise.
10119 (neon_vsub<mode>_unspec): Likewise.
10120 (neon_vsubl<mode>): Likewise.
10121 (neon_vsubw<mode>): Likewise.
10122 (neon_vqsub<mode>): Likewise.
10123 (neon_vhsub<mode>): Likewise.
10124 (neon_vsubhn<mode>): Likewise.
10125 (neon_vceq<mode>): Likewise.
10126 (neon_vcge<mode>): Likewise.
10127 (neon_vcgeu<mode>): Likewise.
10128 (neon_vcgt<mode>): Likewise.
10129 (neon_vcgtu<mode>): Likewise.
10130 (neon_vcle<mode>): Likewise.
10131 (neon_vclt<mode>): Likewise.
10132 (neon_vcage<mode>): Likewise.
10133 (neon_vcagt<mode>): Likewise.
10134 (neon_vtst<mode>): Likewise.
10135 (neon_vabd<mode>): Likewise.
10136 (neon_vabdl<mode>): Likewise.
10137 (neon_vaba<mode>): Likewise.
10138 (neon_vabal<mode>): Likewise.
10139 (neon_vmax<mode>): Likewise.
10140 (neon_vmin<mode>): Likewise.
10141 (neon_vpaddl<mode>): Likewise.
10142 (neon_vpadal<mode>): Likewise.
10143 (neon_vpmax<mode>): Likewise.
10144 (neon_vpmin<mode>): Likewise.
10145 (neon_vrecps<mode>): Likewise.
10146 (neon_vrsqrts<mode>): Likewise.
10147 (neon_vqabs<mode>): Likewise.
10148 (neon_vqneg<mode>): Likewise.
10149 (neon_vcls<mode>): Likewise.
10150 (clz<mode>2): Likewise.
10151 (popcount<mode>2): Likewise.
10152 (neon_vrecpe): Likewise.
10153 (neon_vrsqrte): Likewise.
10154 (neon_vget_lane<mode>_sext_internal): Likewise.
10155 (neon_vget_lane<mode>_zext_internal): Likewise.
10156 (neon_vdup_n<mode>): Likewise.
10157 (neon_vdup_nv2di): Likewise.
10158 (neon_vdpu_lane<mode>_internal): Likewise.
10159 (neon_vswp<mode>): Likewise.
10160 (float<mode><V_cvtto>2): Likewise.
10161 (floatuns<mode><V_cvtto>2): Likewise.
10162 (fix_trunc<mode><V_cvtto>)2): Likewise
10163 (fixuns_trunc<mode><V_cvtto)2): Likewise.
10164 (neon_vcvt<mode>): Likewise.
10165 (neon_vcvtv4sfv4hf): Likewise.
10166 (neon_vcvtv4hfv4sf): Likewise.
10167 (neon_vcvt_n<mode>): Likewise.
10168 (neon_vmovn<mode>): Likewise.
10169 (neon_vqmovn<mode>): Likewise.
10170 (neon_vqmovun<mode>): Likewise.
10171 (neon_vmovl<mode>): Likewise.
10172 (neon_vmul_lane<mode>): Likewise.
10173 (neon_vmull_lane<mode>): Likewise.
10174 (neon_vqdmull_lane<mode>): Likewise.
10175 (neon_vqdmulh_lane<mode>): Likewise.
10176 (neon_vmla_lane<mode>): Likewise.
10177 (neon_vmlal_lane<mode>): Likewise.
10178 (neon_vqdmlal_lane<mode>): Likewise.
10179 (neon_vmls_lane<mode>): Likewise.
10180 (neon_vmlsl_lane<mode>): Likewise.
10181 (neon_vqdmlsl_lane<mode>): Likewise.
10182 (neon_vext<mode>): Likewise.
10183 (neon_vrev64<mode>): Likewise.
10184 (neon_vrev32<mode>): Likewise.
10185 (neon_vrev16<mode>): Likewise.
10186 (neon_vbsl<mode>_internal): Likewise.
10187 (neon_vshl<mode>): Likewise.
10188 (neon_vqshl<mode>): Likewise.
10189 (neon_vshr_n<mode>): Likewise.
10190 (neon_vshrn_n<mode>): Likewise.
10191 (neon_vqshrn_n<mode>): Likewise.
10192 (neon_vqshrun_n<mode>): Likewise.
10193 (neon_vshl_n<mode>): Likewise.
10194 (neon_vqshl_n<mode>): Likewise.
10195 (neon_vqshlu_n<mode>): Likewise.
10196 (neon_vshll_n<mode>): Likewise.
10197 (neon_vsra_n<mode>): Likewise.
10198 (neon_vsri_n<mode>): Likewise.
10199 (neon_vsli_n<mode>): Likewise.
10200 (neon_vtbl1v8qi): Likewise.
10201 (neon_vtbl2v8qi): Likewise.
10202 (neon_vtbl3v8qi): Likewise.
10203 (neon_vtbl4v8qi): Likewise.
10204 (neon_vtbx1v8qi): Likewise.
10205 (neon_vtbx2v8qi): Likewise.
10206 (neon_vtbx3v8qi): Likewise.
10207 (neon_vtbx4v8qi): Likewise.
10208 (neon_vtrn<mode>_internal): Likewise.
10209 (neon_vzip<mode>_internal): Likewise.
10210 (neon_vuzp<mode>_internal): Likewise.
10211 (neon_vld1<mode>): Likewise.
10212 (neon_vld1_lane<mode>): Likewise.
10213 (neon_vld1_dup<mode>): Likewise.
10214 (neon_vld1_dupv2di): Likewise.
10215 (neon_vst1<mode>): Likewise.
10216 (neon_vst1_lane<mode>): Likewise.
10217 (neon_vld2<mode>): Likewise.
10218 (neon_vld2_lane<mode>): Likewise.
10219 (neon_vld2_dup<mode>): Likewise.
10220 (neon_vst2<mode>): Likewise.
10221 (neon_vst2_lane<mode>): Likewise.
10222 (neon_vld3<mode>): Likewise.
10223 (neon_vld3qa<mode>): Likewise.
10224 (neon_vld3qb<mode>): Likewise.
10225 (neon_vld3_lane<mode>): Likewise.
10226 (neon_vld3_dup<mode>): Likewise.
10227 (neon_vst3<mode>): Likewise.
10228 (neon_vst3qa<mode>): Likewise.
10229 (neon_vst3qb<mode>): Likewise.
10230 (neon_vst3_lane<mode>): Likewise.
10231 (neon_vld4<mode>): Likewise.
10232 (neon_vld4qa<mode>): Likewise.
10233 (neon_vld4qb<mode>): Likewise.
10234 (neon_vld4_lane<mode>): Likewise.
10235 (neon_vld4_dup<mode>): Likewise.
10236 (neon_vst4<mode>): Likewise.
10237 (neon_vst4qa<mode>): Likewise.
10238 (neon_vst4qb<mode>): Likewise.
10239 (neon_vst4_lane<mode>): Likewise.
10240 (neon_vec_unpack<US>_lo_<mode>): Likewise.
10241 (neon_vec_unpack<US>_hi_<mode>): Likewise.
10242 (neon_vec_<US>mult_lo_<mode>): Likewise.
10243 (neon_vec_<US>mult_hi_<mode>): Likewise.
10244 (neon_vec_<US>shiftl_<mode>): Likewise.
10245 (neon_unpack<US>_<mode>): Likewise.
10246 (neon_vec_<US>mult_<mode>): Likewise.
10247 (vec_pack_trunc_<mode>): Likewise.
10248 (neon_vec_pack_trunk_<mode>): Likewise.
10249 (neon_vabd<mode>_2): Likewise.
10250 (neon_vabd<mode>_3): Likewise.
10251 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
10252 (thumb2_movsi_vfp): Likewise.
10253 (movdi_vfp): Likewise.
10254 (movdi_vfp_cortexa8): Likewise.
10255 (movhf_vfp_neon): Likewise.
10256 (movhf_vfp): Likewiwse.
10257 (movsf_vfp): Likewiwse.
10258 (thumb2_movsf_vfp): Likewiwse.
10259 (movdf_vfp): Likewise.
10260 (thumb2_movdf_vfp): Likewise.
10261 (movsfcc_vfp): Likewise.
10262 (thumb2_movsfcc_vfp): Likewise.
10263 (movdfcc_vfp): Likewise.
10264 (thumb2_movdfcc_vfp): Likewise.
10265 * config/arm/arm.c (cortexa7_older_only): Update for attribute change.
10266 * config/arm/arm1020e.md (v10_c2v): Update for attribute change.
10267 (v10_v2c): Likewise.
10268 * config/arm/cortex-a15-neon.md (cortex_a15_neon_int_1): Update for
10270 (cortex_a15_neon_int_2): Likewise.
10271 (cortex_a15_neon_int_3): Likewise.
10272 (cortex_a15_neon_int_4): Likewise.
10273 (cortex_a15_neon_int_5): Likewise.
10274 (cortex_a15_neon_vqneg_vqabs): Likewise.
10275 (cortex_a15_neon_vmov): Likewise.
10276 (cortex_a15_neon_vaba): Likewise.
10277 (cortex_a15_neon_vaba_qqq): Likewise.
10278 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
10279 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
10280 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_\
10281 scalar_64_32_long_scalar): Likewise.
10282 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
10283 (cortex_a15_neon_mla_qqq_8_16): Likewise.
10284 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
10285 lotype_qdd_64_32_long): Likewise.
10286 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
10287 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
10288 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
10289 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
10290 (cortex_a15_neon_shift_1): Likewise.
10291 (cortex_a15_neon_shift_2): Likewise.
10292 (cortex_a15_neon_shift_3): Likewise.
10293 (cortex_a15_neon_vshl_ddd): Likewise.
10294 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
10295 (cortex_a15_neon_vsra_vrsra): Likewise.
10296 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
10297 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
10298 (cortex_a15_neon_fp_vmul_ddd): Likewise.
10299 (cortex_a15_neon_fp_vmul_qqd): Likewise.
10300 (cortex_a15_neon_fp_vmla_ddd): Likewise.
10301 (cortex_a15_neon_fp_vmla_qqq): Likewise.
10302 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
10303 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
10304 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
10305 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
10306 (cortex_a15_neon_bp_simple): Likewise.
10307 (cortex_a15_neon_bp_2cycle): Likewise.
10308 (cortex_a15_neon_bp_3cycle): Likewise.
10309 (cortex_a15_neon_vld1_1_2_regs): Likewise.
10310 (cortex_a15_neon_vld1_3_4_regs): Likewise.
10311 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
10312 (cortex_a15_neon_vld2_4_regs): Likewise.
10313 (cortex_a15_neon_vld3_vld4): Likewise.
10314 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
10315 (cortex_a15_neon_vst1_3_4_regs): Likewise.
10316 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Likewise.
10317 (cortex_a15_neon_vst3_vst4): Likewise.
10318 (cortex_a15_neon_vld1_vld2_lane): Likewise.
10319 (cortex_a15_neon_vld3_vld4_lane" 10
10320 (cortex_a15_neon_vst1_vst2_lane): Likewise.
10321 (cortex_a15_neon_vst3_vst4_lane): Likewise.
10322 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
10323 (cortex_a15_neon_ldm_2): Likewise.0
10324 (cortex_a15_neon_stm_2): Likewise.
10325 (cortex_a15_neon_mcr): Likewise.
10326 (cortex_a15_neon_mcr_2_mcrr): Likewise.
10327 (cortex_a15_neon_mrc): Likewise.
10328 (cortex_a15_neon_mrrc): Likewise.
10329 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
10331 (cortex_a15_alu_shift): Likewise.
10332 (cortex_a15_alu_shift_reg): Likewise.
10333 (cortex_a15_mult32): Likewise.
10334 (cortex_a15_mult64): Likewise.
10335 (cortex_a15_block): Likewise.
10336 (cortex_a15_branch): Likewise.
10337 (cortex_a15_load1): Likewise.
10338 (cortex_a15_load3): Likewise.
10339 (cortex_a15_store1): Likewise.
10340 (cortex_a15_store3): Likewise.
10341 (cortex_a15_call): Likewise.
10342 * config/arm/cortex-a5.md (cortex_a5_r2f): Update for attribute change.
10343 (cortex_a5_f2r): Likewise.
10344 * config/arm/cortex-a53.md (cortex_a53_r2f): Update for attribute
10346 (cortex_a53_f2r): Likewise.
10347 * config/arm/cortex-a7.md
10348 (cortex_a7_branch): Update for attribute change.
10349 (cortex_a7_call): Likewise.
10350 (cortex_a7_alu_imm): Likewise.
10351 (cortex_a7_alu_reg): Likewise.
10352 (cortex_a7_alu_shift): Likewise.
10353 (cortex_a7_mul): Likewise.
10354 (cortex_a7_load1): Likewise.
10355 (cortex_a7_store1): Likewise.
10356 (cortex_a7_load2): Likewise.
10357 (cortex_a7_store2): Likewise.
10358 (cortex_a7_load3): Likewise.
10359 (cortex_a7_store3): Likewise.
10360 (cortex_a7_load4): Likewise.
10361 (cortex_a7_store4): Likewise.
10362 (cortex_a7_fpalu): Likewise.
10363 (cortex_a7_fconst): Likewise.
10364 (cortex_a7_fpmuls): Likewise.
10365 (cortex_a7_neon_mul): Likewise.
10366 (cortex_a7_fpmacs): Likewise.
10367 (cortex_a7_neon_mla: Likewise.
10368 (cortex_a7_fpmuld: Likewise.
10369 (cortex_a7_fpmacd: Likewise.
10370 (cortex_a7_fpfmad: Likewise.
10371 (cortex_a7_fdivs: Likewise.
10372 (cortex_a7_fdivd: Likewise.
10373 (cortex_a7_r2f: Likewise.
10374 (cortex_a7_f2r: Likewise.
10375 (cortex_a7_f_flags: Likewise.
10376 (cortex_a7_f_loads: Likewise.
10377 (cortex_a7_f_loadd: Likewise.
10378 (cortex_a7_f_stores: Likewise.
10379 (cortex_a7_f_stored: Likewise.
10380 (cortex_a7_neon): Likewise.
10381 * config/arm/cortex-a8-neon.md
10382 (cortex_a8_neon_mrc): Update for attribute change.
10383 (cortex_a8_neon_mrrc): Likewise.
10384 (cortex_a8_neon_int_1): Likewise.
10385 (cortex_a8_neon_int_2): Likewise.
10386 (cortex_a8_neon_int_3): Likewise.
10387 (cortex_a8_neon_int_4): Likewise.
10388 (cortex_a8_neon_int_5): Likewise.
10389 (cortex_a8_neon_vqneg_vqabs): Likewise.
10390 (cortex_a8_neon_vmov): Likewise.
10391 (cortex_a8_neon_vaba): Likewise.
10392 (cortex_a8_neon_vaba_qqq): Likewise.
10393 (cortex_a8_neon_vsma): Likewise.
10394 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
10395 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
10396 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
10397 long_scalar): Likewise.
10398 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
10399 (cortex_a8_neon_mla_qqq_8_16): Likewise.
10400 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
10401 long_scalar_qdd_64_32_long): Likewise.
10402 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
10403 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
10404 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
10405 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
10406 (cortex_a8_neon_shift_1): Likewise.
10407 (cortex_a8_neon_shift_2): Likewise.
10408 (cortex_a8_neon_shift_3): Likewise.
10409 (cortex_a8_neon_vshl_ddd): Likewise.
10410 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
10411 (cortex_a8_neon_vsra_vrsra): Likewise.
10412 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
10413 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
10414 (cortex_a8_neon_fp_vsum): Likewise.
10415 (cortex_a8_neon_fp_vmul_ddd): Likewise.
10416 (cortex_a8_neon_fp_vmul_qqd): Likewise.
10417 (cortex_a8_neon_fp_vmla_ddd): Likewise.
10418 (cortex_a8_neon_fp_vmla_qqq): Likewise.
10419 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
10420 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
10421 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
10422 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
10423 (cortex_a8_neon_bp_simple): Likewise.
10424 (cortex_a8_neon_bp_2cycle): Likewise.
10425 (cortex_a8_neon_bp_3cycle): Likewise.
10426 (cortex_a8_neon_ldr): Likewise.
10427 (cortex_a8_neon_str): Likewise.
10428 (cortex_a8_neon_vld1_1_2_regs): Likewise.
10429 (cortex_a8_neon_vld1_3_4_regs): Likewise.
10430 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
10431 (cortex_a8_neon_vld2_4_regs): Likewise.
10432 (cortex_a8_neon_vld3_vld4): Likewise.
10433 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
10434 (cortex_a8_neon_vst1_3_4_regs): Likewise.
10435 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
10436 (cortex_a8_neon_vst3_vst4): Likewise.
10437 (cortex_a8_neon_vld1_vld2_lane): Likewise.
10438 (cortex_a8_neon_vld3_vld4_lane): Likewise.
10439 (cortex_a8_neon_vst1_vst2_lane): Likewise.
10440 (cortex_a8_neon_vst3_vst4_lane): Likewise.
10441 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
10442 (cortex_a8_neon_mcr): Likewise.
10443 (cortex_a8_neon_mcr_2_mcrr): Likewise.
10444 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
10445 * config/arm/cortex-a9-neon.md (ca9_neon_mrc): Update for attribute
10447 (ca9_neon_mrrc): Likewise.
10448 (cortex_a9_neon_int_1): Likewise.
10449 (cortex_a9_neon_int_2): Likewise.
10450 (cortex_a9_neon_int_3): Likewise.
10451 (cortex_a9_neon_int_4): Likewise.
10452 (cortex_a9_neon_int_5): Likewise.
10453 (cortex_a9_neon_vqneg_vqabs): Likewise.
10454 (cortex_a9_neon_vmov): Likewise.
10455 (cortex_a9_neon_vaba): Likewise.
10456 (cortex_a9_neon_vaba_qqq): Likewise.
10457 (cortex_a9_neon_vsma): Likewise.
10458 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
10459 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
10460 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
10461 long_scalar): Likewise.
10462 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
10463 (cortex_a9_neon_mla_qqq_8_16): Likewise.
10464 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
10465 long_scalar_qdd_64_32_long): Likewise.
10466 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
10467 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
10468 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
10469 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
10470 (cortex_a9_neon_shift_1): Likewise.
10471 (cortex_a9_neon_shift_2): Likewise.
10472 (cortex_a9_neon_shift_3): Likewise.
10473 (cortex_a9_neon_vshl_ddd): Likewise.
10474 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
10475 (cortex_a9_neon_vsra_vrsra): Likewise.
10476 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
10477 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
10478 (cortex_a9_neon_fp_vsum): Likewise.
10479 (cortex_a9_neon_fp_vmul_ddd): Likewise.
10480 (cortex_a9_neon_fp_vmul_qqd): Likewise.
10481 (cortex_a9_neon_fp_vmla_ddd): Likewise.
10482 (cortex_a9_neon_fp_vmla_qqq): Likewise.
10483 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
10484 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
10485 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
10486 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
10487 (cortex_a9_neon_bp_simple): Likewise.
10488 (cortex_a9_neon_bp_2cycle): Likewise.
10489 (cortex_a9_neon_bp_3cycle): Likewise.
10490 (cortex_a9_neon_ldr): Likewise.
10491 (cortex_a9_neon_str): Likewise.
10492 (cortex_a9_neon_vld1_1_2_regs): Likewise.
10493 (cortex_a9_neon_vld1_3_4_regs): Likewise.
10494 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
10495 (cortex_a9_neon_vld2_4_regs): Likewise.
10496 (cortex_a9_neon_vld3_vld4): Likewise.
10497 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
10498 (cortex_a9_neon_vst1_3_4_regs): Likewise.
10499 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
10500 (cortex_a9_neon_vst3_vst4): Likewise.
10501 (cortex_a9_neon_vld1_vld2_lane): Likewise.
10502 (cortex_a9_neon_vld3_vld4_lane): Likewise.
10503 (cortex_a9_neon_vst1_vst2_lane): Likewise.
10504 (cortex_a9_neon_vst3_vst4_lane): Likewise.
10505 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
10506 (cortex_a9_neon_mcr): Likewise.
10507 (cortex_a9_neon_mcr_2_mcrr): Likewise.
10508 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
10509 (cortex_a9_fps): Likewise.
10510 * config/arm/cortex-m4-fpu.md (cortex_m4_vmov_2): Update for attribute
10512 (cortex_m4_fmuls): Likewise.
10513 * config/arm/cortex-r4f.md (cortex_r4_mcr): Update for attribute
10515 (cortex_r4_mrc): Likewise.
10516 * config/arm/iterators.md: Update comment referring to neon_type.
10517 * config/arm/iwmmxt.md (iwmmxt_arm_movdi): Update for attribute change.
10518 (iwmmxt_movsi_insn): Likewise.
10519 * config/arm/marvell-pj4.md (pj4_vfp_to_core): Update for
10521 (pj4_core_to_vfp): Likewise.
10522 * config/arm/neon-schedgen.ml (emit_insn_reservations): Update for
10524 * config/arm/vfp11.md (vfp_fload): Update for attribute change.
10525 (vfp_fstore): Likewise.
10526 * doc/md.texi: Change references to neon_type to refer to type.
10528 2013-09-04 Dodji Seketeli <dodji@redhat.com>
10530 * tree.h (DECL_BUILT_IN): Fix typo in comment.
10532 2013-09-04 David Edelsohn <dje.gcc@gmail.com>
10534 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Only emit
10535 lglobl if not weak.
10537 2013-09-04 Easwaran Raman <eraman@google.com>
10539 PR middle-end/57370
10540 * tree-ssa-reassoc.c (get_stmt_uid_with_default): New function,
10541 (build_and_add_sum): Use it.
10542 (appears_later_in_bb): Simplify code.
10544 2013-09-04 Teresa Johnson <tejohnson@google.com>
10546 * dumpfile.c (dump_finish): Don't close stderr/stdout.
10548 2013-09-04 James Greenhalgh <james.greenhalgh@arm.com>
10550 * config/aarch64/arm_neon.h (vaddvq_<su>64): Fix return types.
10552 2013-09-04 Jan Hubicka <jh@suse.cz>
10554 * Makefile.in (ipa-devirt.o): Add dependency on diagnostic.h
10555 * ipa-devirt.c: Include diganostic.h
10556 (odr_type_d): Add types and types_set.
10557 (hash_type_name): Work for types with vtables during LTO.
10558 (odr_hasher::remove): Fix comment; destroy types_set.
10559 (add_type_duplicate): New function,
10560 (get_odr_type): Use it.
10561 (dump_type_inheritance_graph): Dump type duplicates.
10562 * ipa.c (symtab_remove_unreachable_nodes): Build type inheritance
10564 * tree.c (types_same_for_odr): Give exact answers on types with
10567 2013-09-04 Dodji Seketeli <dodji@redhat.com>
10569 * tree.h (DECL_BUILT_IN, DECL_IS_BUILTIN): Add more comments
10570 explaining their differences.
10572 2013-09-04 Sandeep Kumar Singh<Sandeep.Singh2@kpitcummins.com>
10574 * config/rx/rx.h: Add option -mcpu for target variants RX100 and RX200.
10576 2013-09-03 Jeff Law <law@redhat.com>
10578 * tree-ssa-threadedge.c (thread_across_edge): Record entire path
10579 when not threading through a joiner block. Pass joiner/no joiner
10580 state to register_jump_thread.
10581 * tree-ssa-threadupdate.c (register_jump_thread): Get joiner/no joiner
10582 state from argument rather than implying on path length.
10583 Dump the entire jump thread path into debugging dump.
10584 * tree-flow.h (register_jump_thread): Update prototype.
10586 2013-08-29 Xinliang David Li <davidxl@google.com>
10588 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
10589 Remove a trivial gcc_assert.
10591 2013-08-29 Xinliang David Li <davidxl@google.com>
10593 * tree-vect-slp.c (destroy_bb_vec_info): Data ref cleanup.
10594 * tree-vect-loop.c (destroy_bb_vec_info): Ditto.
10595 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
10596 Delay base decl alignment adjustment.
10597 * tree-vectorizer.c (vect_destroy_datarefs): New function.
10598 * tree-vectorizer.h: New data structure.
10599 (set_dr_misalignment): New function.
10600 (dr_misalignment): Ditto.
10601 * tree-vect-stmts.c (vectorizable_store): Ensure alignment.
10602 (vectorizable_load): Ditto.
10603 (ensure_base_align): New function.
10604 (vectorize_loops): Add dbg_cnt support.
10605 (execute_vect_slp): Ditto.
10606 * dbgcnt.def: New debug counter.
10607 * Makefile: New dependency.
10609 2013-09-03 Meador Inge <meadori@codesourcery.com>
10613 2013-08-30 Meador Inge <meadori@codesourcery.com>
10615 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
10617 2013-09-03 David Edelsohn <dje.gcc@gmail.com>
10619 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Emit lglobl for
10620 function descriptor.
10622 2013-09-03 Richard Biener <rguenther@suse.de>
10624 * tree-affine.c (add_elt_to_tree): Fix association issue,
10625 avoid useless converts and make sure to always return a
10626 properly typed result.
10628 2013-09-03 Richard Biener <rguenther@suse.de>
10630 PR middle-end/57656
10631 * fold-const.c (negate_expr_p): Fix division case.
10632 (negate_expr): Likewise.
10634 2013-09-03 Richard Biener <rguenther@suse.de>
10637 * tree-streamer-out.c: Include tm.h.
10638 * Makefile.in (tree-streamer-out.o): Depend on $(TM_H).
10640 2013-09-03 Jan Hubicka <jh@suse.cz>
10642 * tree-profile.c (tree_profiling): Cleanup CFG when done.
10644 2013-09-03 Alan Modra <amodra@gmail.com>
10646 * config.gcc (powerpc*-*-linux*): Add support for little-endian
10647 multilibs to big-endian target and vice versa.
10648 * config/rs6000/t-linux64: Use := assignment on all vars.
10649 (MULTILIB_EXTRA_OPTS): Remove fPIC.
10650 (MULTILIB_OSDIRNAMES): Specify using mapping from multilib_options.
10651 * config/rs6000/t-linux64le: New file.
10652 * config/rs6000/t-linux64bele: New file.
10653 * config/rs6000/t-linux64lebe: New file.
10655 2013-09-02 Jan Hubicka <jh@suse.cz>
10657 * ipa-inline-transform.c (inline_transform): Do not
10658 optimize_inline_calls when not optimizing.
10660 2013-09-02 Jan Hubicka <jh@suse.cz>
10662 * lto-symtab.c (lto_symtab_merge_symbols): Add comments; merge
10663 duplicated nodes for assembler names.
10664 * symtab.c (symtab_unregister_node): Do not attempt to unlink
10665 hard registers from assembler name hash.
10667 2013-09-02 Jan Hubicka <jh@suse.cz>
10669 * ipa-split.c (execute_split_functions): Split externally visible
10670 functions called once.
10672 2013-09-02 Martin Jambor <mjambor@suse.cz>
10675 * ipa-prop.c (ipa_edge_duplication_hook): Always put new rdesc to the
10676 linked list. When finding the correct duplicate, also consider also
10677 the caller in additon to its inlined_to node.
10679 2013-09-02 James Greenhalgh <james.greenhalgh@arm.com>
10681 * config/aarch64/aarch64-simd-builtins.def
10682 (dup_lane_scalar): Remove.
10683 * config/aarch64/aarch64-simd.md
10684 (aarch64_simd_dup): Add 'w->w' alternative.
10685 (aarch64_dup_lane<mode>): Allow for VALL.
10686 (aarch64_dup_lane_scalar<mode>): Remove.
10687 (aarch64_dup_lane_<vswap_width_name><mode>): New.
10688 (aarch64_get_lane_signed<mode>): Add w->w altenative.
10689 (aarch64_get_lane_unsigned<mode>): Likewise.
10690 (aarch64_get_lane<mode>): Likewise.
10691 * config/aarch64/aarch64.c (aarch64_evpc_dup): New.
10692 (aarch64_expand_vec_perm_const_1): Use aarch64_evpc_dup.
10693 * config/aarch64/iterators.md (VSWAP_WIDTH): New.
10694 (VCON): Change container of V2SF.
10695 (vswap_width_name): Likewise.
10696 * config/aarch64/arm_neon.h
10697 (__aarch64_vdup_lane_any): New.
10698 (__aarch64_vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
10699 (vdup<q>_n_<psuf><8,16,32,64>): Convert to C implementation.
10700 (vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
10702 2013-09-02 Eric Botcazou <ebotcazou@adacore.com>
10704 PR middle-end/56382
10705 * expr.c (emit_move_complex): Do not move complex FP values as parts if
10706 the source or the destination is a single hard register.
10708 2013-09-02 Richard Biener <rguenther@suse.de>
10710 PR middle-end/57511
10711 * tree-scalar-evolution.c (instantiate_scev_name): Allow
10714 2013-09-02 Richard Biener <rguenther@suse.de>
10716 * tree-affine.c (add_elt_to_tree): Avoid converting all pointer
10717 arithmetic to sizetype.
10719 2013-09-02 Bin Cheng <bin.cheng@arm.com>
10721 * tree-ssa-loop-ivopts.c (set_autoinc_for_original_candidates):
10722 Find auto-increment use both before and after candidate.
10724 2013-09-02 Marek Polacek <polacek@redhat.com>
10726 * Makefile.in (ubsan.o): Add $(TM_P_H) dependency.
10728 2013-09-01 Jan Hubicka <jh@suse.cz>
10730 * Makefile.in: Add ipa-profile.o
10731 (ipa.o, ipa-devrit.o, ipa-inline-analysis.o): Adjust dependencies.
10732 * cgraph.c (struct cgraph_propagate_frequency_data,
10733 cgraph_propagate_frequency_1, cgraph_propagate_frequency): Move to
10734 ipa-profile.c; replace cgraph_ by ipa_ prefix.
10735 * cgraph.h (cgraph_propagate_frequency): Remove.
10736 * ipa-inline-analysis.c: Include ipa-utils.h;
10737 drop duplicated cfgloop.h.
10738 (inline_update_callee_summaries): Update.
10739 * ipa-profile.c: New file.
10740 * ipa-utils.h (ipa_propagate_frequency): Declare.
10741 * ipa.c: Do not include pointer-set.h, hash-table.h, lto-streamer.h,
10742 data-streamer.h, value-prof.h.
10743 (symtab_remove_unreachable_nodes): Update profile.
10744 (struct histogram_entry, histogram, histogram_pool, histogram_hash,
10745 account_time_size, cmp_counts, dump_histogram,
10746 ipa_profile_generate_summary, ipa_profile_write_summary,
10747 ipa_profile_read_summary, ipa_profile, gate_ipa_profile,
10748 pass_data_ipa_profile, pass_ipa_profile, make_pass_ipa_profile):
10749 Move to ipa-profile.c.
10751 2013-09-01 John David Anglin <danglin@gcc.gnu.org>
10753 * config/pa/pa.md: Allow "const 0" operand 1 in "scc" insns.
10755 2013-09-01 Jan Hubicka <jh@suse.cz>
10757 * common.opt (fdevirtualize-speculatively): New function.
10758 * invoke.texi (fdevirtualize-speculatively): Document.
10759 * ipa-devirt.c: Include ipa-inline.h
10760 (likely_target_p): New function.
10761 (ipa_devirt): New function.
10762 (gate_ipa_devirt): New function.
10763 (pass_data_ipa_devirt): New static var.
10764 (pass_ipa_devirt): Likewise.
10765 (make_pass_ipa_devirt): New function.
10766 * opts.c (default_options): Add OPT_fdevirtualize_speculatively.
10767 (common_handle_option): Disable devirtualization when
10768 value range profiling is available.
10769 * passes.def (pass_ipa_devirt): Add.
10770 * timever.def (TV_IPA_DEVIRT): New timevar.
10771 * tree-pass.h (make_pass_ipa_devirt):
10773 2013-09-01 Iain Sandoe <iain@codesourcery.com>
10775 * config/darwin.h (LINK_COMMAND_SPEC_A): Revise sanitizer specs to
10776 include sanitize(undefined).
10778 2013-08-31 Diego Novillo <dnovillo@google.com>
10780 * Makefile.in (TREE_CORE_H): Define.
10782 (GTFILES): Add tree-core.h.
10783 * builtins.c (built_in_class_names): Use BUILT_IN_LAST to
10785 * tree-core.h: New file.
10786 Move all data structures, enum, typedefs, global
10787 declarations and constants from ...
10788 * tree.h: ... here.
10790 2013-08-31 Jan Hubicka <jh@suse.cz>
10792 * bulitins.c (expand_builtin): Do not early exit for gcov
10793 instrumented functions.
10795 2013-08-31 Marek Polacek <polacek@redhat.com>
10797 * ubsan.c: Include tm_p.h.
10799 2013-08-31 Jan Hubicka <jh@suse.cz>
10801 * gimple-streamer-in.c (input_gimple_stmt): Silence parameter unused
10804 * cgraph.c (cgraph_get_body): Update call of lto_input_function_body.
10805 * gimple-streamer-in.c (input_gimple_stmt): Move sanity check to ...
10806 * tree-cfg.c (verify_gimple_label): ... here.
10807 * ipa-utils.c: Include lto-streamer.h, ipa-inline.h
10808 (ipa_merge_profiles): New function.
10809 * lto-streamer-in.c (lto_read_body): Take node instead of fn_decl.
10810 (lto_input_function_body): Likewise.
10811 * ipa-utils.h (ipa_merge_profiles): Declare.
10812 * lto-streamer.h (lto_input_function_body): Update prototype.
10813 (emit_label_in_global_context_p): Remove.
10814 * lto-symtab.c: Include ipa-utils.h
10815 (lto_cgraph_replace_node): Use ipa_merge_profiles.
10817 2013-08-31 Jan Hubicka <jh@suse.cz>
10819 * cgraph.c (cgraph_speculative_call_info): Fix ref lookup
10821 2013-08-31 Jan Hubicka <jh@suse.cz>
10823 * basic-block.h (apply_scale): Make scale parmeter gcov_type.
10825 2013-08-31 Uros Bizjak <ubizjak@gmail.com>
10827 * config/alpha/alpha.c (alpha_emit_conditional_move): Update
10828 "cmp" RTX before signed_comparison_operator check to account
10829 for "code" changes.
10831 2013-08-30 Jan Hubicka <jh@suse.cz>
10833 * ipa-prop.c (ipa_set_jf_known_type): Check that we add only records.
10834 (detect_type_change_1): Rename to ...
10835 (detect_type_change): ... this one; early return on non-polymorphic
10837 (detect_type_change_ssa): Add comp_type parameter; update
10838 use of detect_type_change.
10839 (compute_complex_assign_jump_func): Add param_type parameter;
10840 update use of detect_type_change_ssa.
10841 (compute_complex_ancestor_jump_func): Likewise.
10842 (ipa_get_callee_param_type): New function.
10843 (ipa_compute_jump_functions_for_edge): Compute parameter type;
10844 update calls to the jump function computation functions.
10846 2013-08-30 Teresa Johnson <tejohnson@google.com>
10847 Steven Bosscher <steven@gcc.gnu.org>
10849 * cfgrtl.c (fixup_new_cold_bb): New routine.
10850 (commit_edge_insertions): Invoke fixup_partitions.
10851 (find_partition_fixes): New routine.
10852 (fixup_partitions): Ditto.
10853 (verify_hot_cold_block_grouping): Update comments.
10854 (rtl_verify_edges): Invoke find_partition_fixes.
10855 (rtl_verify_bb_pointers): Update comments.
10856 (rtl_verify_bb_layout): Ditto.
10857 * basic-block.h (probably_never_executed_edge_p): Declare.
10858 (fixup_partitions): Ditto.
10859 * cfgcleanup.c (try_optimize_cfg): Invoke fixup_partitions.
10860 * bb-reorder.c (sanitize_hot_paths): New function.
10861 (find_rarely_executed_basic_blocks_and_crossing_edges): Invoke
10862 sanitize_hot_paths.
10863 * predict.c (probably_never_executed_edge_p): New routine.
10864 * cfg.c (check_bb_profile): Add partition insanity warnings.
10866 2013-08-30 Meador Inge <meadori@codesourcery.com>
10868 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
10870 2013-08-30 Marek Polacek <polacek@redhat.com>
10872 * Makefile.in (ubsan.o): Add.
10873 (c-family/c-ubsan.o): Add.
10874 (builtins.o): Add ubsan.h dependency.
10875 * ubsan.h: New file.
10876 * ubsan.c: New file.
10877 * common.opt: Add -fsanitize=undefined option.
10878 (flag_sanitize): Add variable.
10879 (fsanitize=): Add option. Add Driver.
10880 (fsanitize=thread): Remove option.
10881 (fsanitize=address): Likewise.
10882 (static-libubsan): New option.
10883 * doc/invoke.texi: Document the new flag and -static-libubsan.
10884 * sanitizer.def (DEF_SANITIZER_BUILTIN): Define.
10885 (BUILT_IN_UBSAN_HANDLE_BUILTIN_UNREACHABLE): Define.
10886 * builtin-attrs.def (ATTR_COLD): Define.
10887 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
10888 * builtins.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW,
10889 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS): Define.
10890 * flag-types.h (sanitize_code): New enum.
10891 * opts.c (common_handle_option): Parse command line arguments
10892 of -fsanitize=. Add -fsanitize=unreachable option.
10893 * varasm.c (get_variable_section): Adjust.
10894 (assemble_noswitch_variable): Likewise.
10895 (assemble_variable): Likewise.
10896 (output_constant_def_contents): Likewise.
10897 (categorize_decl_for_section): Likewise.
10898 (place_block_symbol): Likewise.
10899 (output_object_block): Likewise.
10900 * builtins.def: Likewise.
10901 * toplev.c (compile_file): Likewise.
10902 (process_options): Likewise.
10903 * cppbuiltin.c: Likewise.
10904 * tsan.c (tsan_pass): Likewise.
10905 (tsan_gate): Likewise.
10906 (tsan_gate_O0): Likewise.
10907 * cfgexpand.c (partition_stack_vars): Likewise.
10908 (expand_stack_vars): Likewise.
10909 (defer_stack_allocation): Likewise.
10910 (expand_used_vars): Likewise.
10911 * cfgcleanup.c (old_insns_match_p): Likewise.
10912 * asan.c (asan_finish_file): Likewise.
10913 (asan_instrument): Likewise.
10914 (gate_asan): Likewise.
10915 (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR_PTR.
10916 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
10917 (asan_global_struct): Use pointer_sized_int_node instead
10918 calling build_nonstandard_integer_type.
10919 (initialize_sanitizer_builtins): Likewise.
10920 (asan_finish_file): Likewise.
10921 * gcc.c: Document %{%:function(args):X}.
10922 (static_spec_functions): Add sanitize.
10923 (handle_spec_function): Add retval_nonnull argument and if non-NULL,
10924 store funcval != NULL there.
10925 (do_spec_1): Adjust handle_spec_function caller.
10926 (handle_braces): Allow %:function(args) as condition.
10927 (sanitize_spec_function): New function.
10928 (ADD_STATIC_LIBUBSAN_LIBS): Define.
10929 (LIBUBSAN_SPEC): Likewise.
10930 (LIBUBSAN_EARLY_SPEC): Likewise.
10931 (SANITIZER_SPEC): Handle libubsan.
10932 (SANITIZER_EARLY_SPEC): Likewise.
10933 * config/darwin.h (LINK_COMMAND_SPEC_A): Use %:sanitize(address)
10934 instead of fsanitize=address.
10935 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Use %:sanitize(address)
10936 instead of fsanitize=address*.
10937 * builtins.c: Include ubsan.h.
10938 (fold_builtin_0): Instrument __builtin_unreachable.
10939 * config/rs6000/rs6000.h (FRAME_GROWS_DOWNWARD): Use flag_sanitize
10940 instead of flag_asan.
10941 * tree.h (enum tree_index): Add TI_POINTER_SIZED_TYPE.
10942 (pointer_sized_int_node): Define.
10943 * tree.c (build_common_tree_nodes): Initialize pointer_sized_int_node.
10945 2013-08-30 Mike Stump <mikestump@comcast.net>
10947 * doc/install.texi (Prerequisites): Note regression in Tcl 8.6
10950 2013-08-29 Jan Hubicka <jh@suse.cz>
10952 * cgraph.c (cgraph_function_body_availability): Handle weakref
10954 * passes.def: Remove pass_fixup_cfg.
10955 * ipa-inline.c (ipa_inline): When not optimizing, do not inline;
10956 track when we need to remove functions.
10957 (gate_ipa_inline): Execute inlining always; add comment why.
10958 (pass_data_ipa_inline): Remove TODO_remove_functions.
10959 * ipa-inline-analysis.c (inline_generate_summary): When not optimizing
10960 do not produce summaries.
10961 * symtab.c (change_decl_assembler_name): Handle renaming of weakrefs.
10962 (symtab_nonoverwritable_alias): Assert we are not called on weakref.
10963 * varpool.c (cgraph_variable_initializer_availability): Fix weakrefs,
10964 constant pool and vtable.
10966 2013-08-30 Tejas Belagod <tejas.belagod@arm.com>
10968 * config/aarch64/arm_neon.h (__AARCH64_UINT64_C, __AARCH64_INT64_C):
10969 New arm_neon.h's internal macros to specify 64-bit constants.
10970 Avoid using stdint.h's macros.
10972 2013-08-30 Joern Rennecke <joern.rennecke@embecosm.com>
10974 * recog.c (verify_changes): Verify that changes[i].old is non-zero
10975 before applying REG_P.
10977 2013-08-30 Jakub Jelinek <jakub@redhat.com>
10979 PR tree-optimization/58277
10980 * tree-ssa-strlen.c (strlen_enter_block): If do_invalidate gave up
10981 after seeing too many stmts with vdef in between dombb and current
10982 bb, invalidate everything.
10984 2013-08-30 Richard Biener <rguenther@suse.de>
10986 * fold-const.c (fold_single_bit_test): Fix overflow test.
10988 2013-08-30 Eric Botcazou <ebotcazou@adacore.com>
10990 * function.c (assign_parm_setup_reg): For a parameter passed by pointer
10991 and which can live in a register, always retrieve the value on entry.
10992 * var-tracking.c (add_stores): Treat the copy on entry for a parameter
10993 passed by invisible reference specially.
10994 (emit_notes_in_bb) <MO_VAL_USE>: Emit notes before the instruction.
10995 (vt_add_function_parameter): Correctly deal with a parameter passed by
10996 invisible reference.
10998 2013-08-30 Jan Hubicka <jh@suse.cz>
11000 * tree.c (set_call_expr_flags): Fix handling of TM_PURE.
11002 2013-08-30 Richard Biener <rguenther@suse.de>
11004 PR tree-optimization/58228
11005 * tree-vect-data-refs.c (vect_analyze_data_ref_access): Do not
11006 allow invariant loads in nested loop vectorization.
11008 2013-08-30 Richard Biener <rguenther@suse.de>
11010 PR tree-optimization/58223
11011 * tree-loop-distribution.c (has_anti_dependence): Rename to ...
11012 (has_anti_or_output_dependence): ... this and adjust to also
11013 look for output dependences.
11014 (mark_nodes_having_upstream_mem_writes): Adjust.
11015 (rdg_flag_uses): Likewise.
11017 2013-08-30 Richard Biener <rguenther@suse.de>
11019 PR tree-optimization/58010
11020 * tree-vect-loop.c (vect_create_epilog_for_reduction): Remove
11021 assert that we have a loop-closed PHI.
11023 2013-08-29 Jan Hubicka <jh@suse.cz>
11025 * lto-symtab.c (lto_cgraph_replace_node): Free decl_in_state.
11026 * cgraph.c (cgraph_release_function_body): Free decl_in_state.
11027 * lto-section-in.c (lto_free_function_in_decl_state): New function.
11028 (lto_free_function_in_decl_state_for_node): New function.
11030 2013-08-29 Xinliang David Li <davidxl@google.com>
11032 * loop-unroll.c (report_unroll_peel): Minor message change.
11033 * tree-vect-loop-manip.c (vect_do_peeling_for_alignment):
11034 Emit alignment peeling message with default -fopt-info.
11035 (vect_loop_versioning): Emit loop version info message.
11036 * tree-vectorizer.c (vectorize_loops): Minor message change.
11037 (execute_vect_slp): Ditto.
11039 2013-08-29 Eric Botcazou <ebotcazou@adacore.com>
11041 * cgraphclones.c (cgraph_create_virtual_clone): Compute the DECL_NAME
11042 of the clone from the DECL_NAME of the original function.
11044 2013-08-29 Oleg Endo <olegendo@gcc.gnu.org>
11046 * passes.c (register_pass): Add overload.
11047 * tree-pass.h (register_pass): Forward declare it. Add comment.
11049 2013-08-29 Jan Hubicka <jh@suse.cz>
11051 * lto-streamer-out.c (hash_tree): Stream DECL_FINAL_P,
11052 DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P and TYPE_FINAL_P.
11053 * lto-streamer-in.c (unpack_ts_decl_with_vis_value_fields): Stream
11054 DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
11055 (unpack_ts_type_common_value_fields): Stream TYPE_FINAL_P.
11056 * tree-streamer-out.c (pack_ts_decl_with_vis_value_fields):
11057 Add DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
11058 (pack_ts_type_common_value_fields): Add TYPE_FINAL_P.
11060 2013-08-29 Teresa Johnson <tejohnson@google.com>
11062 * dumpfile.c (dump_loc): Output column number.
11063 * dumpfile.h (OPTGROUP_OTHER): Add and enable under OPTGROUP_ALL.
11064 * doc/invoke.texi: Document optall -fopt-info flag.
11065 * profile.c (read_profile_edge_counts): Use new dump framework.
11066 (compute_branch_probabilities): Ditto.
11067 * passes.c (pass_manager::register_one_dump_file): Use OPTGROUP_OTHER
11068 when pass not in any opt group.
11069 * pass_manager.h (pass_manager::get_pass_profile): New method.
11070 * value-prof.c (check_counter): Use new dump framework.
11071 (check_ic_target): Ditto.
11072 * coverage.c (get_coverage_counts): Ditto.
11073 (coverage_init): Setup new dump framework.
11075 2013-08-29 Richard Biener <rguenther@suse.de>
11077 PR tree-optimization/58246
11078 * tree-ssa-dce.c (mark_aliased_reaching_defs_necessary_1): Properly
11079 handle the dominance check inside a basic-block.
11081 2013-08-29 Richard Biener <rguenther@suse.de>
11083 PR middle-end/57287
11084 * tree-ssa-copy.c (may_propagate_copy): Allow propagating
11085 of default defs that appear in abnormal PHI nodes.
11087 2013-08-29 Richard Biener <rguenther@suse.de>
11089 PR tree-optimization/57685
11090 * tree-vrp.c (register_edge_assert_for_1): Recurse only for
11091 single-use operands to avoid exponential complexity.
11093 2013-08-28 Dehao Chen <dehao@google.com>
11095 * ipa-inline.c (edge_badness): Fix integer underflow.
11097 2013-08-28 Uros Bizjak <ubizjak@gmail.com>
11099 * gtm-builtins.def (_ITM_free): Declare leaf.
11101 2013-08-28 Jakub Jelinek <jakub@redhat.com>
11104 * config/i386/i386.md (*tls_global_dynamic_64_largepic): New insn.
11105 (*tls_local_dynamic_base_64_largepic): Likewise.
11106 (tls_global_dynamic_64_<mode>, tls_local_dynamic_base_64_<mode>):
11107 Remove predicate from call operand.
11108 * config/i386/i386.c (ix86_tls_get_addr): For -mcmodel=large -fpic
11109 return sum of pic_offset_table_rtx and UNSPEC_PLTOFF of the symbol.
11111 2013-08-28 Jeff Law <law@redhat.com>
11113 * tree-ssa-threadedge.c (thread_around_empty_block): Remove
11114 checks for the number of predecessors and successors allowed.
11115 * tree-ssa-threadupdate.c (mark_threaded_blocks): Ignore requests
11116 which require copying a joiner block if there is a request which
11117 is a subpath that requires no joiner block copying.
11119 2013-08-28 Jan Hubicka <jh@suse.cz>
11121 * lto-streamer-out.c (DFS_write_tree_body): Drop
11122 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX and BINFO_VPTR_INDEX.
11123 (hash_tree): Do not hash DECL_DEFER_OUTPUT, BINFO_INHERITANCE_CHAIN,
11124 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX, DECL_IN_TEXT_SECTION.
11125 * tree-streamer-in.c (unpack_ts_decl_common_value_fields):
11126 Do not read DECL_ERROR_ISSUED.
11127 (unpack_ts_decl_with_vis_value_fields): Do not read
11129 (lto_input_ts_binfo_tree_pointers): Do not read
11130 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX
11131 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Do not
11132 write DECL_ERROR_ISSUED..
11133 (pack_ts_decl_with_vis_value_fields): Do not write
11135 (write_ts_binfo_tree_pointers): Do not read BINFO_INHERITANCE_CHAIN,
11136 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX.
11137 * print-tree.c (print_node): Do not print DECL_ERROR_ISSUED.
11138 * tree.h (tree_decl_common): Update comment.
11139 (DECL_ERROR_ISSUED): Remove.
11141 2013-08-28 Jakub Jelinek <jakub@redhat.com>
11143 PR middle-end/58257
11144 * omp-low.c (copy_var_decl): Copy over TREE_NO_WARNING flag.
11146 2013-08-28 Jan Hubicka <jh@suse.cz>
11148 * builtins.def (free): Declare leaf.
11150 2013-08-27 David Malcolm <dmalcolm@redhat.com>
11152 * gdbhooks.py: New.
11153 * configure.ac (gdbinit.in): Add import of gcc/gdbhooks.py.
11154 * configure: Regenerate.
11156 2013-08-27 Martin Jambor <mjambor@suse.cz>
11158 * ipa-prop.h (ipa_pass_through_data): New field type_preserved.
11159 (ipa_ancestor_jf_data): Likewise.
11160 (ipa_get_jf_pass_through_agg_preserved): Fix comment typo.
11161 (ipa_get_jf_pass_through_type_preserved): New function.
11162 (ipa_get_jf_ancestor_agg_preserved): Fix comment typo.
11163 (ipa_get_jf_ancestor_type_preserved): New function.
11164 * ipa-cp.c (ipa_get_jf_pass_through_result): Honor type_preserved flag.
11165 (ipa_get_jf_ancestor_result): Likewise.
11166 (propagate_vals_accross_pass_through): Use
11167 ipa_get_jf_pass_through_result to do all the value mappings.
11168 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Dump the
11169 type_preserved flag.
11170 (ipa_set_jf_cst_copy): New function.
11171 (ipa_set_jf_simple_pass_through): Set the type_preserved flag.
11172 (ipa_set_jf_arith_pass_through): Likewise.
11173 (ipa_set_ancestor_jf): Likewise.
11174 (compute_complex_assign_jump_func): Set type_preserved instead of
11176 (ipa_compute_jump_functions_for_edge): Likewise.
11177 (combine_known_type_and_ancestor_jfs): Honor type_preserved.
11178 (update_jump_functions_after_inlining): Update type_preserved.
11179 Explicitely create jump functions when combining one with pass_through.
11180 (ipa_write_jump_function): Stream the type_preserved flags.
11181 (ipa_read_jump_function): Likewise.
11183 2013-08-27 Jakub Jelinek <jakub@redhat.com>
11184 Aldy Hernandez <aldyh@redhat.com>
11186 * Makefile.in (omp-low.o): Depend on $(TARGET_H).
11187 * cfgloop.h (struct loop): Add safelen, force_vect, simduid.
11188 * function.h (struct function): Add has_force_vect_loops and
11190 * gimple-pretty-print.c (dump_gimple_omp_for): Handle GF_OMP_FOR_KIND*.
11191 * gimple.c (gimple_build_omp_critical): Add KIND argument and
11193 * gimple.def: Update CLAUSES comments.
11194 * gimple.h (enum gf_mask): Add GF_OMP_FOR_KIND_{FOR,SIMD}.
11195 (gimple_build_omp_for): Add argument to prototype.
11196 (gimple_omp_for_kind): New.
11197 (gimple_omp_for_set_kind): New.
11198 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_LINEAR to
11199 GOVD_DATA_SHARE_CLASS.
11200 (enum omp_region_type): Add ORT_SIMD.
11201 (gimple_add_tmp_var): Handle ORT_SIMD.
11202 (gimplify_var_or_parm_decl): Same.
11203 (is_gimple_stmt): Same.
11204 (omp_firstprivatize_variable): Same.
11205 (omp_add_variable): Only use splay_tree_insert if lookup failed.
11206 (omp_notice_variable): Handle ORT_SIMD.
11207 (omp_is_private): Add SIMD argument and handle it as well as ORT_SIMD.
11208 (omp_check_private): Handle ORT_SIMD.
11209 (gimplify_scan_omp_clauses): Handle OMP_CLAUSE_LINEAR and
11210 OMP_CLAUSE_SAFELEN.
11211 (gimplify_adjust_omp_clauses_1): Handle GOVD_LINEAR.
11212 Handle OMP_CLAUSE_LASTPRIVATE.
11213 (gimplify_adjust_omp_clauses): Handle OMP_CLAUSE_LINEAR and
11214 OMP_CLAUSE_SAFELEN.
11215 (gimplify_omp_for): Handle OMP_SIMD and OMP_CLAUSE_LINEAR.
11216 (gimplify_expr): Handle OMP_SIMD.
11217 * internal-fn.c (expand_GOMP_SIMD_LANE): New.
11218 (expand_GOMP_SIMD_VF): New.
11219 (expand_GOMP_SIMD_LAST_LANE): New.
11220 * internal-fn.def (GOMP_SIMD_LANE): New.
11221 (GOMP_SIMD_VF): New.
11222 (GOMP_SIMD_LAST_LANE): New.
11223 * omp-low.c: Include target.h.
11224 (extract_omp_for_data): Handle OMP_SIMD, OMP_CLAUSE_LINEAR,
11225 OMP_CLAUSE_SAFELEN.
11226 (check_omp_nesting_restrictions): Same.
11228 (lower_rec_simd_input_clauses): New.
11229 (lower_rec_input_clauses): Handle OMP_SIMD, GF_OMP_FOR_KIND_SIMD,
11231 (lower_lastprivate_clauses): Handle OMP_CLAUSE_LINEAR,
11232 GF_OMP_FOR_KIND_SIMD, OMP_SIMD.
11233 (expand_omp_build_assign): New.
11234 (expand_omp_for_init_counts): New.
11235 (expand_omp_for_init_vars): New.
11236 (extract_omp_for_update_vars): New.
11237 (expand_omp_for_generic): Use expand_omp_for_{init,update}_vars
11238 and rewrite accordingly.
11239 (expand_omp_simd): New.
11240 (expand_omp_for): Use expand_omp_simd.
11241 (lower_omp_for_lastprivate): Unshare vinit when appropriate.
11242 (lower_omp_for): Do not lower the body.
11243 * tree-data-ref (get_references_in_stmt): Allow IFN_GOMP_SIMD_LANE
11244 in their own loops.
11245 * tree-flow.h (find_omp_clause): Remove prototype.
11246 * tree-if-conv.c (main_tree_if_conversion): Run if doing if conversion,
11247 forcing vectorization of the loop, or if flag_tree_vectorize.
11248 (gate_tree_if_conversion): Similarly.
11249 * tree-inline.c (remap_gimple_stmt): Pass for kind argument to
11250 gimple_build_omp_for.
11251 (copy_cfg_body): set has_force_vect_loops and has_simduid_loops.
11252 * tree-parloops (create_parallel_loop): Pass kind argument to
11253 gimple_build_omp_for.
11254 * tree-pretty-print.c (dump_omp_clause): Add cases for
11255 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_LINEAR, OMP_CLAUSE_SAFELEN,
11256 OMP_CLAUSE__SIMDUID_.
11257 (dump_generic_node): Handle OMP_SIMD.
11258 * tree-ssa-ccp.c (likely_value): Handle IFN_GOMP_SIMD*.
11259 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Do not
11260 unroll OMP_SIMD loops here.
11261 * tree-ssa-loop.c (gate_tree_vectorize): Run if has_force_vect_loops.
11262 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Handle
11264 (vect_analyze_data_refs): Handle simd loops.
11265 * tree-vect-loop.c (vectorizable_live_operation): Handle
11267 * tree-vect-stmts.c (vectorizable_call): Handle IFN_GOMP_SIMD_LANE.
11268 (vectorizable_store): Handle STMT_VINFO_SIMD_LANE_ACCESS_P.
11269 (vectorizable_load): Same.
11270 * tree-vectorizer.c: Include hash-table.h and tree-ssa-propagate.h.
11271 (struct simduid_to_vf): New.
11272 (simduid_to_vf::hash): New.
11273 (simduid_to-vf::equal): New.
11274 (struct simd_array_to_simduid): New.
11275 (simd_array_to_simduid::hash): New.
11276 (simd_array_to_simduid::equal): New.
11277 (adjust_simduid_builtins): New.
11278 (struct note_simd_array_uses_struct): New.
11279 (note_simd_array_uses_cb): New.
11280 (note_simd_array_uses): New.
11281 (vectorize_loops): Handle simd hints and adjust simd builtins
11283 * tree-vectorizer.h (struct _stmt_vec_info): Add
11284 simd_lane_access_p field.
11285 (STMT_VINFO_SIMD_LANE_ACCESS_P): New macro.
11286 * tree.c (omp_clause_num_ops): Add entries for OMP_CLAUSE_LINEAR,
11287 OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_UNIFORM.
11288 (omp_clause_code_name): Same.
11289 (walk_tree_1): Handle OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN,
11290 OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_LINEAR.
11291 * tree.def (OMP_SIMD): New entry.
11292 * tree.h (enum omp_clause_code): Add entries for OMP_CLAUSE_LINEAR,
11293 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_.
11294 (OMP_CLAUSE_DECL): Adjust range for new clauses.
11295 (OMP_CLAUSE_LINEAR_NO_COPYIN): New.
11296 (OMP_CLAUSE_LINEAR_NO_COPYOUT): New.
11297 (OMP_CLAUSE_LINEAR_STEP): New.
11298 (OMP_CLAUSE_SAFELEN_EXPR): New.
11299 (OMP_CLAUSE__SIMDUID__DECL): New.
11300 (find_omp_clause): New prototype.
11302 2013-08-27 H.J. Lu <hongjiu.lu@intel.com>
11304 * config/i386/driver-i386.c (host_detect_local_cpu): Update
11305 Haswell processor detection.
11307 2013-08-27 Christian Widmer <shadow@umbrox.de>
11310 * config/i386/driver-i386.c (host_detect_local_cpu): Add detection
11311 of Ivy Bridge and Haswell processors. Assume core-avx2 for unknown
11312 AVX2 capable processors.
11314 2013-08-27 Tejas Belagod <tejas.belagod@arm.com>
11316 * config/aarch64/arm_neon.h: Replace all inline asm implementations
11317 of vget_low_* with implementations in terms of other intrinsics.
11319 2013-08-27 Marc Glisse <marc.glisse@inria.fr>
11321 PR middle-end/57219
11322 * doc/extend.texi (__builtin_isinf_sign): Restrict the return
11323 values to -1, 0 and 1.
11325 2013-08-27 Vidya Praveen <vidyapraveen@arm.com>
11327 * config/aarch64/aarch64.md (unspec): Add UNSPEC_SISD_SSHL,
11328 UNSPEC_SISD_USHL, UNSPEC_USHL_2S, UNSPEC_SSHL_2S, UNSPEC_SISD_NEG.
11329 (<optab><mode>3_insn): Remove.
11330 (aarch64_ashl_sisd_or_int_<mode>3): New Pattern.
11331 (aarch64_lshr_sisd_or_int_<mode>3): Likewise.
11332 (aarch64_ashr_sisd_or_int_<mode>3): Likewise.
11333 (define_split for aarch64_lshr_sisd_or_int_di3): Likewise.
11334 (define_split for aarch64_lshr_sisd_or_int_si3): Likewise.
11335 (define_split for aarch64_ashr_sisd_or_int_di3): Likewise.
11336 (define_split for aarch64_ashr_sisd_or_int_si3): Likewise.
11337 (aarch64_sisd_ushl, aarch64_sisd_sshl): Likewise.
11338 (aarch64_ushl_2s, aarch64_sshl_2s, aarch64_sisd_neg_qi): Likewise.
11339 (ror<mode>3_insn): Likewise.
11340 * config/aarch64/predicates.md (aarch64_simd_register): New.
11342 2013-08-27 Richard Biener <rguenther@suse.de>
11344 PR tree-optimization/57521
11345 * tree-if-conv.c (if_convertible_bb_p): Verify that at least
11346 one edge is non-critical.
11347 (find_phi_replacement_condition): Make sure to use a non-critical
11348 edge. Cleanup and remove old bug workarounds.
11349 (bb_postdominates_preds): Remove.
11350 (if_convertible_loop_p_1): Do not compute post-dominators.
11351 (combine_blocks): Do not free post-dominators.
11352 (main_tree_if_conversion): Likewise.
11353 (pass_data_if_conversion): Add TODO_verify_ssa.
11355 2013-08-27 DJ Delorie <dj@redhat.com>
11357 * config/i386/djgpp.h (ASM_DECLARE_FUNCTION_NAME): New.
11359 2013-08-27 Yufeng Zhang <yufeng.zhang@arm.com>
11361 * function.c (assign_parm_find_data_types): Set passed_mode and
11362 nominal_mode to the TYPE_MODE of nominal_type for the built
11363 pointer type in case of the struct-pass-by-reference.
11365 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
11367 * config/avr/avr-stdint.h (INT16_TYPE): Change default to "int".
11368 (UINT16_TYPE): Change default to "unsigned int".
11370 * config/avr/avr.opt (mfract-convert-truncate): New option.
11371 * config/avr/avr.c (avr_out_fract): Unless TARGET_FRACT_CONV_TRUNC
11372 is set, round negative fractional integers according to n1169
11373 when converting to integer types.
11375 2013-08-26 Jan Hubicka <jh@suse.cz>
11377 * cgraph.c (cgraph_propagate_frequency): Do not assume that virtual
11378 methods can not be called indirectly when their address is not taken.
11380 2013-08-26 Jan Hubicka <jh@suse.cz>
11382 * gimple-fold.c (gimple_get_virt_method_for_binfo): Use
11385 2013-08-26 Jan Hubicka <jh@suse.cz>
11387 * ipa.c (comdat_can_be_unshared_p_1): C++ constructors and destructors
11390 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
11392 * reload.c (find_valid_class): Allow classes that do not include
11393 FIRST_PSEUDO_REGISTER - 1.
11395 2013-08-26 Jan Hubicka <jh@suse.cz>
11397 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix formatting;
11398 fix edge count/frequency when speculation failed; fix type check
11399 for the direct call.
11401 2013-08-26 Jan Hubicka <jh@suse.cz>
11403 * ipa-prop.c (ipa_print_node_params): Do not ICE during WPA.
11405 2013-08-26 Jan Hubicka <jh@suse.cz>
11407 * ipa-inline-transform.c (inline_transform): Be ready for basic block
11408 to be changed by edge redirection.
11410 2013-08-26 Jan Hubicka <jh@suse.cz>
11412 * cgraph.c (cgraph_speculative_call_info): Fix parameter order and
11413 formating; add sanity check.
11414 (cgraph_resolve_speculation): Add FIXME about scaling profiles.
11415 (cgraph_redirect_edge_call_stmt_to_callee): Fix ICE in debug dump.
11416 * ipa-inline.c (heap_edge_removal_hook): Reset node growth cache.
11417 (resolve_noninline_speculation): Update callee keys, too.
11419 2013-08-26 Jan Hubicka <jh@suse.cz>
11421 * tree.h (tree_decl_with_vis): Add cxx_constructor, cxx_destructor.
11422 (DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P): New macros.
11424 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
11426 * config/i386/i386.c (x86_64_elf_select_section): Put ATTRIBUTE_UNUSED
11429 2013-08-26 Uros Bizjak <ubizjak@gmail.com>
11431 * config/i386/i386.c (ix86_debug_options): Remove prototype.
11432 (x86_64_elf_select_section): Ditto.
11433 (ix86_handle_tm_regparm_attribute): Remove ATTRIBUTE_UNUSED on used
11435 (ix86_pass_by_reference): Ditto.
11436 (output_set_got): Ditto.
11437 (ix86_unary_operator_ok): Ditto.
11438 (ix86_expand_builtin): Ditto.
11440 2013-08-23 Jan Hubicka <jh@suse.cz>
11442 * cgraph.c (cgraph_turn_edge_to_speculative): Fix debug output.
11444 2013-08-23 Jan Hubicka <jh@suse.cz>
11446 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
11447 (tree_decl_with_vis): Add FINAL field.
11449 2013-08-23 Jeff Law <law@redhat.com>
11451 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
11452 the debugging dump when the expression is fully redundant.
11454 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
11456 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
11457 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
11458 * pretty-print.c (pp_formatted_text_data): Likewise.
11459 (pp_write_text_to_stream): Likewise.
11460 (pp_write_text_as_dot_label_to_stream): Likewise.
11461 (pp_append_r): Likewise.
11462 (pp_format): Likewise.
11463 (pp_flush): Likewise.
11464 (pp_clear_output_area): Likewise.
11465 (pp_append_text): Likewise.
11466 (pp_formatted_text): Likewise.
11467 (pp_remaining_character_count_for_line): Likewise.
11468 (pp_newline): Likewise.
11469 (pp_character): Likewise.
11470 (output_buffer::~output_buffer): Define.
11471 (pretty_printer::~pretty_printer): Destruct output buffer.
11472 * pretty-print.h (output_buffer::~output_buffer): Declare.
11473 (pretty_printer::~pretty_printer): Declare virtual.
11475 2013-08-24 Marc Glisse <marc.glisse@inria.fr>
11478 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
11479 HOST_WIDE_INT_M1U): New macros.
11480 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
11481 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
11482 unsigned -1 for lshift.
11483 * cse.c (cse_insn): Likewise.
11484 * double-int.c (rshift_double, lshift_double): Likewise.
11485 * builtins.c (fold_builtin_bitop): Likewise.
11486 * combine.c (force_to_mode): Likewise.
11487 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
11488 * simplify-rtx.c (simplify_const_unary_operation,
11489 simplify_const_binary_operation): Likewise.
11490 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
11491 check_va_list_escapes): Likewise.
11492 * rtlanal.c (nonzero_bits1): Likewise.
11493 * expmed.c (expand_smod_pow2): Likewise.
11494 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
11496 2013-08-23 Jan Hubicka <jh@suse.cz>
11498 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
11499 as having address taken.
11501 2013-08-23 Jan Hubicka <jh@suse.cz>
11503 * ipa-utils.h (method_class_type): Declare.
11504 * ipa-devirt.c (method_class_type): Export.
11506 * cgraphunit.c (analyze_functions): Do basic devirtualization;
11507 do not walk base classes of anonymous types.
11509 2013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
11511 PR rtl-optimization/58220
11512 PR regression/58221
11513 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
11514 handle SEQUENCE insns properly.
11516 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
11518 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
11520 (pp_newline_and_indent): Likewise.
11521 (pp_separate_with): Likewise.
11522 * pretty-print.c (pp_newline_and_flush): Define.
11523 (pp_newline_and_indent): Likewise.
11524 (pp_separate_with): Likewise.
11526 2013-08-23 Jakub Jelinek <jakub@redhat.com>
11529 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
11530 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
11532 2013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
11534 * config/i386/predicates.md (ext_sse_reg_operand): New.
11535 * config/i386/i386.md (*movti_internal): Use
11536 predicate to determine if EVEX is needed.
11537 (*movsi_internal): Ditto.
11538 (*movdf_internal): Ditto.
11539 (*movsf_internal): Ditto.
11540 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
11542 2013-08-23 Jakub Jelinek <jakub@redhat.com>
11544 PR tree-optimization/58209
11545 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
11546 (find_tail_calls): Give up for pointer result types if m is non-NULL.
11547 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
11548 emit POINTER_PLUS_EXPR.
11549 (create_tailcall_accumulator): For pointer result type accumulate in
11552 2013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
11554 * configure.ac: Add backslashes missing from the last change.
11555 * configure: Regenerate.
11557 2013-08-22 Jan Hubicka <jh@susue.cz>
11559 * ipa.c (function_and_variable_visibility): First remember function
11560 was global and then make it local.
11562 2013-08-22 Julian Brown <julian@codesourcery.com>
11564 * configure.ac: Add aarch64 to list of arches which use "nop" in
11566 * configure: Regenerate.
11568 2013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11570 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
11571 gnu_libc_has_function.
11572 * config/s390/tpf.h: Likewise.
11574 2013-08-22 Jan Hubicka <jh@susue.cz>
11576 * timevar.c (validate_phases): Add cast.
11578 2013-08-22 Jan Hubicka <jh@susue.cz>
11580 * timevar.c (validate_phases): Use size_t for memory.
11581 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
11583 2013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
11585 * pretty-print.h (output_buffer::output_buffer): Declare.
11586 (pretty_printer::pretty_printer): Likewise.
11587 (pp_construct): Remove.
11588 * pretty-print.c (output_buffer::output_buffer): Define.
11589 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
11590 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
11591 (print_gimple_expr): Likewise.
11592 (print_gimple_seq): Likewise.
11593 (gimple_dump_bb): Likewise.
11594 * sched-vis.c (dump_value_slim): Likewise.
11595 (dump_insn_slim): Likewise.
11596 (dump_rtl_slim): Likewise.
11597 (str_pattern_slim): Likewise.
11598 * tree-mudflap.c (mf_varname_tree): Likewise.
11599 * graph.c (print_graph_cfg): Likewise.
11600 (start_graph_dump): Likewise.
11601 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
11603 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
11604 pretty printer initialization.
11605 * coretypes.h (diagnostic_context): Remove superflous type alias
11607 (pretty_printer): Likewise. Declare directly as a class.
11608 (pretty_print_info): Remove declaration as class.
11609 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
11610 and pp_clear_output_area.
11611 (asan_add_global): Likewise.
11613 2013-08-22 Jan Hubicka <jh@suse.cz>
11615 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
11616 * ipa-utils.h (update_type_inheritance_graph): Declare.
11617 (possible_polymorphic_call_target_p): Declare.
11618 (possible_polymorphic_call_target_p): New.
11619 * ipa-devirt.c: Update toplevel comments.
11620 (cached_polymorphic_call_targets): Move up.
11621 (odr_type_d): Move ID down.
11622 (polymorphic_type_binfo_p): Update comment.
11623 (odr_hasher::remove): Likewise;
11624 (get_odr_type): Set anonymous_namespace.
11625 (dump_odr_type): Dump it.
11626 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
11627 (maybe_record_node): Record node in cached_polymorphic_call_targets.
11628 (record_binfo): Add comment.
11629 (free_polymorphic_call_targets_hash): Do not ICE when cache is not
11631 (devirt_node_removal_hook): Do not iCE when cache is freed.
11632 (possible_polymorphic_call_target_p): New predicate.
11633 (update_type_inheritance_graph): New function.
11635 2013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
11636 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
11637 Sergey Lega <sergey.s.lega@intel.com>
11638 Anna Tikhonova <anna.tikhonova@intel.com>
11639 Ilya Tocar <ilya.tocar@intel.com>
11640 Andrey Turetskiy <andrey.turetskiy@intel.com>
11641 Ilya Verbin <ilya.verbin@intel.com>
11642 Kirill Yukhin <kirill.yukhin@intel.com>
11643 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
11645 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
11646 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
11647 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
11648 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
11649 (OPTION_MASK_ISA_AVX2_UNSET): Update.
11650 (OPTION_MASK_ISA_AVX512F_UNSET): New.
11651 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
11652 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
11653 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
11654 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
11655 OPT_mavx512pf, OPT_mavx512er cases.
11656 * config/i386/constraints.md (v): New constraint.
11657 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
11658 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
11659 (bit_AVX512CD): New.
11660 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
11661 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
11662 * config/i386/i386-c.c (ix86_target_macros_internal):
11663 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
11665 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
11666 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
11667 * config/i386/i386.c (regclass_map, dbx_register_map)
11668 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
11669 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
11670 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
11671 -mavx512pf options.
11672 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
11673 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
11674 -mavx512pf options. Fix formatting.
11675 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
11676 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
11677 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
11678 -mavx512cd, -mavx512pf options.
11679 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
11680 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
11681 (ix86_preferred_output_reload_class): Replace SSE_REGS with
11683 (ix86_hard_regno_mode_ok): Support 512-bit registers.
11684 (ix86_set_reg_reg_cost): Ditto.
11685 (x86_order_regs_for_local_alloc): Ditto.
11686 (MAX_VECT_LEN): Extend to 64-byte.
11687 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
11688 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
11689 (TARGET_AVX512ER, TARGET_AVX512CD): New.
11690 (BIGGEST_ALIGNMENT): Extend to 512-bits.
11691 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
11692 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
11693 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
11694 (SSE_REG_MODE_P): Support new modes.
11695 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
11696 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
11697 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
11698 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
11699 (REG_CLASS_CONTENTS): Add new registers.
11700 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
11701 (EXT_REX_SSE_REGNO_P): New.
11702 (HI_REGISTER_NAMES): Add new registers.
11703 * config/i386/i386.md: Define constants for new registers.
11704 (mode): Add new 512-bit modes.
11705 (prefix): Support evex prefix.
11706 (isa): Support avx512f, noavx512f, fma_avx512f.
11707 (ssemodesuffix): Add new 512-bit modes.
11709 (*movxi_internal_avx512f): Ditto.
11710 (*movdi_internal): Replace constraint "x" with the new constraint "v".
11712 (*movsi_internal): Likewise.
11713 (*movdf_internal): Likewise.
11714 (*movsf_internal): Likewise.
11715 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
11716 (<code><mode>3): Likewise.
11717 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd):
11719 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
11720 with the new constraint "v".
11721 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
11723 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
11724 with the new constraint "v".
11725 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
11726 (<sse2>_storedqu<avxsizesuffix>): Likewise.
11727 (*<plusminus_insn><mode>3): Likewise.
11728 (<sse>_vm<plusminus_insn><mode>3): Likewise.
11729 (*mul<mode>3): Likewise.
11730 (<sse>_vmmul<mode>3): Likewise.
11731 (<sse>_div<mode>3): Likewise.
11732 (<sse>_vmdiv<mode>3): Likewise.
11733 (<sse>_sqrt<mode>2): Likewise.
11734 (<sse>_vmsqrt<mode>2): Likewise.
11735 (*<code><mode>3_finite): Likewise.
11736 (*<code><mode>3) <smaxmin>: Likewise.
11737 (<sse>_vm<code><mode>3): Likewise.
11738 (*<code><mode>3) <any_logic>: Likewise.
11739 (*fma_fmadd_<mode>): Likewise.
11740 (*fma_fmsub_<mode>): Likewise.
11741 (*fma_fnmadd_<mode>): Likewise.
11742 (*fma_fnmsub_<mode>): Likewise.
11743 (*fma_fmaddsub_<mode>): Likewise.
11744 (*fma_fmsubadd_<mode>): Likewise.
11745 (*fmai_fmadd_<mode>): Likewise.
11746 (*fmai_fmsub_<mode>): Likewise.
11747 (*fmai_fnmadd_<mode>): Likewise.
11748 (*fmai_fnmsub_<mode>): Likewise.
11749 (sse_cvtsi2ss): Likewise.
11750 (sse_cvtsi2ssq): Likewise.
11751 (sse_cvtss2si): Likewise.
11752 (sse_cvtss2si_2): Likewise.
11753 (sse_cvtss2siq): Likewise.
11754 (sse_cvtss2siq_2): Likewise.
11755 (sse_cvttss2si): Likewise.
11756 (sse_cvtss2siq_2): Likewise.
11757 (float<sseintvecmodelower><mode>2): Likewise.
11758 (sse2_cvtsd2si_2): Likewise.
11759 (sse2_cvtsd2siq_2): Likewise.
11760 (*<plusminus_insn><mode>3): Likewise.
11761 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
11762 (*<sse4_1_avx2>_mul<mode>3): Likewise.
11763 (ashr<mode>3): Likewise.
11764 (<shift_insn><mode>3): Likewise.
11765 (avx2_<code><mode>3): Likewise.
11766 (*avx2_<code><mode>3): Likewise.
11767 (*andnot<mode>3): Likewise.
11768 (*<code><mode>3) <any_logic>: Likewise.
11769 (abs<mode>2): Likewise.
11770 (avx2_permvar<mode>): Likewise.
11771 (avx2_perm<mode>_1): Likewise.
11772 (*avx_vpermilp<mode>): Likewise.
11773 (avx_vpermilvar<mode>3): Likewise.
11774 (avx2_ashrv<mode>): Likewise.
11775 (avx2_<shift_insn>v<mode>): Likewise.
11776 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
11778 * doc/rtl.texi: Document XImode.
11780 2013-08-21 Jeff Law <law@redhat.com>
11782 * tree-flow.h (register_jump_thread): Pass vector of edges
11783 instead of each important edge.
11784 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
11785 thread path into a vector and pass that to register_jump_thread.
11786 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
11787 passed in edge vector to the current 3-edge form.
11790 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
11792 * dce.c (fini_dce): Call df_analyze again just in case
11793 delete_unmarked_insns removed anything.
11795 2013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
11797 * reload.h (struct reg_equivs): Rename to ..
11798 (struct reg_equivs_s): .. this.
11800 2013-08-20 Martin Liska <marxin.liska@gmail.com>
11802 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
11804 2013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
11806 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest comment.
11808 2013-08-21 Jeff Law <law@redhat.com>
11810 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
11811 simplify assignments too. If the RHS collapses to a singleton
11812 range, then return the value for the range.
11814 2013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
11816 * config/i386/sse.md (V16): Rename to...
11818 (mov<mode>): Update iterator name.
11819 (*mov<mode>_internal): Ditto.
11820 (push<mode>1): Ditto.
11821 (movmisalign<mode>): Ditto.
11823 2013-08-20 Jan Hubicka <jh@suse.cz>
11826 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
11827 entry for direct edges.
11828 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
11830 2013-08-20 David Malcolm <dmalcolm@redhat.com>
11832 Revert my last two changes, r201865 and r201864:
11835 2013-08-20 David Malcolm <dmalcolm@redhat.com>
11837 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
11838 instances can own GC refs.
11840 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
11841 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
11842 (gcc::context::gt_pch_nx): Likewise.
11843 (gcc::context::gt_pch_nx): Likewise.
11844 * ggc.h (gt_ggc_mx <T>): New.
11845 (gt_pch_nx_with_op <T>): New.
11846 (gt_pch_nx <T>): New.
11847 * passes.c (opt_pass::gt_ggc_mx): New.
11848 (opt_pass::gt_pch_nx): New.
11849 (opt_pass::gt_pch_nx_with_op): New.
11850 (pass_manager::gt_ggc_mx): New.
11851 (pass_manager::gt_pch_nx): New.
11852 (pass_manager::gt_pch_nx_with_op): New.
11853 (pass_manager::operator new): Use
11854 ggc_internal_cleared_alloc_stat rather than xcalloc.
11855 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
11856 (pass_manager::gt_ggc_mx): New.
11857 (pass_manager::gt_pch_nx): New.
11858 (pass_manager::gt_pch_nx_with_op): New.
11859 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
11860 (opt_pass::operator new): New.
11861 (opt_pass::gt_ggc_mx): New.
11862 (opt_pass::gt_pch_nx): New.
11863 (opt_pass::gt_pch_nx_with_op): New.
11866 2013-08-20 David Malcolm <dmalcolm@redhat.com>
11868 * Makefile.in (GTFILES): Add context.h.
11869 * context.c (gcc::context::operator new): New.
11870 (gcc::context::gt_ggc_mx): New.
11871 (gcc::context::gt_pch_nx): New.
11872 (gcc::context::gt_pch_nx): New.
11873 * context.h (gcc::context): Add GTY((user)) marking.
11874 (gcc::context::operator new): New.
11875 (gcc::context::gt_ggc_mx): New.
11876 (gcc::context::gt_pch_nx): New.
11877 (gcc::context::gt_pch_nx): New.
11878 (g): Add GTY marking.
11879 (gt_ggc_mx (gcc::context *)): New.
11880 (gt_pch_nx (gcc::context *)): New.
11881 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
11882 void *cookie)): New.
11883 * gengtype.c (open_base_files) <ifiles>: Add context.h.
11885 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
11887 * dce.c (fini_dce): Call df_analyze again just in case
11888 delete_unmarked_insns removed anything.
11890 2013-08-20 Teresa Johnson <tejohnson@google.com>
11892 PR rtl-optimizations/57451
11893 * final.c (reemit_insn_block_notes): Prevent lexical blocks
11894 from crossing split section boundaries.
11896 2013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
11898 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
11899 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
11900 with MULTLIB_DEFAULTS.
11902 2013-08-20 Nick Clifton <nickc@redhat.com>
11904 * target.def (narrow_volatile_bitfield): Note that the default
11905 value is false, not !TARGET_STRICT_ALIGN.
11906 * doc/tm.texi: Regenerate.
11908 2013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
11910 Fix LIB_SPEC for systems without libpthread.
11912 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
11913 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
11915 * config/i386/linux-common.h: Likewise.
11916 * config/mips/linux-common.h: Likewise.
11918 2013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
11920 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
11923 2013-08-20 David Malcolm <dmalcolm@redhat.com>
11925 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
11926 instances can own GC refs.
11928 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
11929 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
11930 (gcc::context::gt_pch_nx): Likewise.
11931 (gcc::context::gt_pch_nx): Likewise.
11932 * ggc.h (gt_ggc_mx <T>): New.
11933 (gt_pch_nx_with_op <T>): New.
11934 (gt_pch_nx <T>): New.
11935 * passes.c (opt_pass::gt_ggc_mx): New.
11936 (opt_pass::gt_pch_nx): New.
11937 (opt_pass::gt_pch_nx_with_op): New.
11938 (pass_manager::gt_ggc_mx): New.
11939 (pass_manager::gt_pch_nx): New.
11940 (pass_manager::gt_pch_nx_with_op): New.
11941 (pass_manager::operator new): Use
11942 ggc_internal_cleared_alloc_stat rather than xcalloc.
11943 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
11944 (pass_manager::gt_ggc_mx): New.
11945 (pass_manager::gt_pch_nx): New.
11946 (pass_manager::gt_pch_nx_with_op): New.
11947 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
11948 (opt_pass::operator new): New.
11949 (opt_pass::gt_ggc_mx): New.
11950 (opt_pass::gt_pch_nx): New.
11951 (opt_pass::gt_pch_nx_with_op): New.
11953 2013-08-20 David Malcolm <dmalcolm@redhat.com>
11955 * Makefile.in (GTFILES): Add context.h.
11956 * context.c (gcc::context::operator new): New.
11957 (gcc::context::gt_ggc_mx): New.
11958 (gcc::context::gt_pch_nx): New.
11959 (gcc::context::gt_pch_nx): New.
11960 * context.h (gcc::context): Add GTY((user)) marking.
11961 (gcc::context::operator new): New.
11962 (gcc::context::gt_ggc_mx): New.
11963 (gcc::context::gt_pch_nx): New.
11964 (gcc::context::gt_pch_nx): New.
11965 (g): Add GTY marking.
11966 (gt_ggc_mx (gcc::context *)): New.
11967 (gt_pch_nx (gcc::context *)): New.
11968 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
11969 void *cookie)): New.
11970 * gengtype.c (open_base_files) <ifiles>: Add context.h.
11972 2013-08-20 Alan Modra <amodra@gmail.com>
11975 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
11976 (rs6000_emit_epilogue): Likewise.
11978 2013-08-19 Dehao Chen <dehao@google.com>
11980 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
11982 2013-08-19 Peter Bergner <bergner@vnet.ibm.com>
11983 Jakub Jelinek <jakub@redhat.com>
11985 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
11986 (BUILT_IN_FABSD64): Likewise.
11987 (BUILT_IN_FABSD128): Likewise.
11988 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
11989 (fold_builtin_1): Likewise.
11990 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping
11991 destination and source operands.
11992 (*abstd2_fpr): Likewise.
11993 (*nabstd2_fpr): Likewise.
11995 2013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
11997 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
12000 2013-08-19 Aldy Hernandez <aldyh@redhat.com>
12002 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
12005 2013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
12007 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
12008 * builtins.c (default_libc_has_function): New.
12009 (gnu_libc_has_function): Ditto.
12010 (no_c99_libc_has_function): Ditto.
12011 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
12012 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
12013 (fold_builtin_sincos): Likewise.
12014 (fold_builtin_cexp): Likewise.
12015 * builtins.def (DEF_C94_BUILTIN): Likewise.
12016 (DEF_C99_BUILTIN): Likewise.
12017 (DEF_C99_C90RES_BUILTIN): Likewise.
12018 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
12019 definitions to using this define.
12020 * config/darwin-protos.h (darwin_libc_has_function): New.
12021 * config/darwin.c (darwin_libc_has_function): Ditto.
12022 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
12023 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
12024 * config/darwin.h: Ditto.
12025 * config/elfos.h: Ditto.
12026 * config/freebsd.h: Ditto.
12027 * config/i386/cygming.h: Ditto.
12028 * config/i386/djgpp.h: Ditto.
12029 * config/i386/i386-interix.h: Ditto.
12030 * config/microblaze/microblaze.h: Ditto.
12031 * config/mmix/mmix.h: Ditto.
12032 * config/gnu-user.h: Ditto.
12033 * config/ia64/hpux.h: Ditto.
12034 * config/pa/pa-hpux.h: Ditto.
12035 * config/pdp11/pdp11.h: Ditto.
12036 * config/picochip/picochip.h: Ditto.
12037 * config/linux.h: Ditto.
12038 * config/netbsd.h: Ditto.
12039 * config/openbsd.h: Ditto.
12040 * config/rs6000/aix43.h: Ditto.
12041 * config/rs6000/aix51.h: Ditto.
12042 * config/rs6000/aix52.h: Ditto.
12043 * config/rs6000/aix53.h: Ditto.
12044 * config/rs6000/aix61.h: Ditto.
12045 * config/rs6000/darwin.h: Ditto.
12046 * config/rs6000/linux.h: Ditto.
12047 * config/rs6000/linux64.h: Ditto.
12048 * config/s390/tpf.h: Ditto.
12049 * config/sol2-10.h: Ditto.
12050 * config/sol2.h: Ditto.
12051 * config/vms/vms.h: Ditto.
12052 * config/vxworks.h: Ditto.
12053 * config/linux-android.c (linux_android_libc_has_function):
12054 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
12055 * config/linux-protos.h (linux_android_libc_has_function):
12057 * config/i386/i386.c (ix86_libc_has_function): New.
12058 * config/i386/i386-protos.h
12059 (ix86_libc_has_function): New declaration.
12060 * config/i386/i386.md
12061 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
12062 ("isinf<mode>2): Likewise.
12063 * convert.c (convert_to_integer): Using new target hook
12064 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
12065 TARGET_C99_FUNCTIONS.
12066 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
12067 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
12068 * coretypes.h (function_class): New enum for different
12069 classes of functions.
12070 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
12071 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
12072 (TARGET_HAS_SINCOS): Likewise.
12073 (TARGET_LIBC_HAS_FUNCTION): New.
12074 * doc/tm.texi: Regenerated.
12075 * targhooks.h (default_libc_has_function): New declaration.
12076 (no_c99_libc_has_function): Ditto.
12077 (gnu_libc_has_function): Ditto.
12078 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
12079 and TARGET_HAS_SINCOS.
12081 2013-08-18 Jan Hubicka <jh@suse.cz>
12083 * Makeifle-in (ipa-devirt.o): New.
12084 (GTFILES): Add ipa-utils.h and ipa-devirt.c
12085 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
12086 (analyze_functions): Look into possible targets of polymorphic call.
12087 * dumpfile.c (dump_files): Add type-inheritance dump.
12088 * dumpfile.h (TDI_inheritance): New.
12089 * ipa-devirt.c: New file.
12090 * ipa-utils.h (odr_type_d): Forward declare.
12091 (odr_type): New type.
12092 (build_type_inheritance_graph): Declare.
12093 (possible_polymorphic_call_targets): Declare and introduce inline
12094 variant when only edge is pased.
12095 (dump_possible_polymorphic_call_targets): Likewise.
12096 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
12097 * tree.c (type_in_anonymous_namespace_p): Break out from ...
12098 (types_same_for_odr): ... here.
12099 * tree.h (type_in_anonymous_namespace_p): Declare.
12101 2013-08-18 Jakub Jelinek <jakub@redhat.com>
12103 PR tree-optimization/58006
12104 * tree-parloops.c (take_address_of): Don't ICE if get_name
12106 (eliminate_local_variables_stmt): Remove clobber stmts.
12108 2013-08-18 Eric Botcazou <ebotcazou@adacore.com>
12110 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
12111 error message is issued for an alias to undefined symbol.
12113 2013-08-18 Jan Hubicka <jh@suse.cz>
12115 * cgraph.c (cgraph_create_indirect_edge): Discover
12116 polymorphic calls and record basic info into indirect_info.
12117 * gimple-fold.c (gimple_fold_call): When doing BINFO based
12118 devirtualization, ignore objc function calls.
12119 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
12120 call with no parm index info.
12121 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
12122 * tree.c (virtual_method_call_p): New function.
12123 * tree.h (virtual_method_call_p): Declare.
12125 2013-08-16 Jan Hubicka <jh@suse.cz>
12127 PR middle-end/58179
12128 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
12130 2013-08-16 David Edelsohn <dje.gcc@gmail.com>
12132 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
12135 2013-08-16 David Malcolm <dmalcolm@redhat.com>
12137 * gengtype.c (type_for_name): Add special-case support for
12138 locating types within the "gcc::" namespace.
12139 (open_base_files): Emit a "using namespace gcc" directive.
12141 2013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
12144 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
12145 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
12147 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
12148 array instead of each individual operand as a separate argument.
12149 (emit_fusion_gpr_load): Likewise.
12150 (expand_fusion_gpr_load): Add new function declaration.
12152 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
12153 signature to have the operands passed as an array, instead of as
12154 separate arguments. Allow ZERO_EXTEND to be in the memory
12155 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
12156 depend on the register live/dead flags when peepholes are run.
12157 (expand_fusion_gpr_load): New function to be called from the
12158 peephole2 pass, to change the register that addis sets to be the
12160 (emit_fusion_gpr_load): Change the calling signature to have the
12161 operands passed as an array, instead of as separate arguments.
12162 Allow ZERO_EXTEND to be in the memory address, and also
12163 SIGN_EXTEND if -mpower8-fusion-sign.
12165 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
12166 unspec enumeration.
12167 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
12168 adjust the register addis loads up in the peephole2 pass. Do not
12169 depend on the register live/dead state when the peephole pass is done.
12171 2013-08-16 David Malcolm <dmalcolm@redhat.com>
12173 * gengtype.c (create_user_defined_type): Ensure that the kind
12174 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
12175 declaration is seen before the GTY((user)) marking.
12177 2013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
12180 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
12182 2013-08-16 Jan Hubicka <jh@suse.cz>
12184 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
12185 arugment expected_type.
12186 (gimple_fold_call): Use it.
12187 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
12188 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
12189 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
12190 (try_make_edge_direct_virtual_call): Likewise.
12191 * tree.c (obj_type_ref_class): New.
12192 * tree.h (obj_type_ref_class): Use it.
12194 2013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
12196 * sched-vis.c (rtl_slim_pp_initialized): Remove.
12197 (rtl_slim_pp): Likewise.
12198 (init_rtl_slim_pretty_print): Likewise.
12199 (dump_value_slim): Don't call it. Use local pretty printer.
12200 (dump_insn_slim): Likewise.
12201 (dump_rtl_slim): Likewise.
12202 (str_pattern_slim): Likewise.
12203 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
12206 2013-08-16 Jakub Jelinek <jakub@redhat.com>
12208 PR tree-optimization/58164
12209 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
12210 walk gimple_goto_dest of GIMPLE_GOTO.
12212 PR tree-optimization/58165
12213 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
12214 bi_call must be the last stmt in a bb, don't split_block, instead
12215 use fallthru edge from it and give up if there is none.
12216 Release conds vector when returning early.
12218 2013-08-14 Xinliang David Li <davidxl@google.com>
12220 * config/i386/i386.c (ix86_option_override_internal):
12221 Remove unused variable and field.
12223 2013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12226 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
12227 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
12228 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
12229 and Linux, correct BLKmode alignment when 128-bit alignment is
12230 required and compatibility flag is not set.
12231 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
12232 for zero-size arguments when compatibility flag is not set.
12234 2013-08-14 Jakub Jelinek <jakub@redhat.com>
12236 PR tree-optimization/58145
12237 * tree-sra.c (build_ref_for_offset): If prev_base has
12238 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
12240 2013-08-14 Xinliang David Li <davidxl@google.com>
12242 * config/i386/i386.c (ix86_option_override_internal):
12243 Fix uninitialized variable error.
12245 2013-08-14 Xinliang David Li <davidxl@google.com>
12247 * config/i386/i386.opt: Define two new options.
12248 * config/i386/x86-tune.def: Add arch selector field in macros.
12249 * config/i386/i386.h: Adjust macro definition.
12250 * config/i386/i386.c (ix86_option_override_internal):
12252 (parse_mtune_ctrl_str): New function.
12253 (set_ix86_tune_features): New function.
12254 (ix86_function_specific_restore): Call the new helper function.
12256 2013-08-14 Andrey Belevantsev <abel@ispras.ru>
12258 PR rtl-optimization/57662
12259 * sel-sched.c (code_motion_process_successors): When the current insn
12260 is removed after the recursive traversal, break from the loop.
12261 Add comments and debug printouts.
12263 2013-08-14 Jakub Jelinek <jakub@redhat.com>
12264 Alexandre Oliva <aoliva@redhat.com>
12267 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
12268 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
12269 there also UNSPEC_PLTOFF.
12271 2013-08-14 Marek Polacek <polacek@redhat.com>
12273 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
12274 NUM_CONDITIONS bit positions.
12276 2013-08-13 Cary Coutant <ccoutant@google.com>
12278 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
12279 (attr_checksum): Hash vector contents instead of pointer.
12280 (attr_checksum_ordered): Likewise.
12282 2013-08-13 Uros Bizjak <ubizjak@gmail.com>
12284 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
12285 when Pmode != word_mode. Add length_address attribute.
12286 (sse3_monitor_<mode>): Merge from sse3_monitor and
12287 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
12288 Pmode != word_mode. Update insn length attribute.
12289 * config/i386/i386.c (ix86_option_override_internal): Update
12290 ix86_gen_monitor selection for merged sse3_monitor insn.
12292 2013-08-13 Julian Brown <julian@codesourcery.com>
12294 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
12295 perform invalid legitimization on greater-than-word-size modes for
12296 TARGET_E500_DOUBLE.
12298 2013-08-13 Vladimir Makarov <vmakarov@redhat.com>
12300 * ira.c (setup_class_translate_array): Use aclass instead of cl
12301 for classes not fully covered by allocno classes.
12303 2013-08-13 Jakub Jelinek <jakub@redhat.com>
12305 PR tree-optimization/57661
12306 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
12307 * tree-inline.c (tree_function_versioning): Initialize it.
12308 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
12309 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
12310 that is not being copied.
12313 * asan.c (instrument_strlen_call): Fix typo in comment.
12314 Use char * type even for the lhs of POINTER_PLUS_EXPR.
12316 2013-08-13 Steve Ellcey <sellcey@mips.com>
12318 * config/mips/mips.md (prefetch): Use lw instead of ld on
12319 loongson in 32bit mode.
12321 2013-08-13 Nick Clifton <nickc@redhat.com>
12323 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
12325 2013-08-13 Jan Hubicka <jh@suse.cz>
12327 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
12328 introduced edge; fix typo in sanity check.
12329 (cgraph_resolve_speculation): Export; improve diagnostic.
12330 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
12331 speculation at type mismatch.
12332 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
12333 (cgraph_resolve_speculation): Declare.
12334 (symtab_can_be_discarded): New function.
12335 * value-prof.c (gimple_ic_transform): Remove actual transform code.
12336 * ipa-inline-transform.c (speculation_removed): New global var.
12337 (clone_inlined_nodes): See if speculation can be removed.
12338 (inline_call): If speculations was removed, we growths may not match.
12339 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
12340 (speculation_useful_p): New function.
12341 (resolve_noninline_speculation): New function.
12342 (inline_small_functions): Resolve useless speculations.
12343 * ipa-inline.h (speculation_useful_p): Declare
12344 * ipa.c (can_replace_by_local_alias): Simplify.
12345 (ipa_profile): Produce speculative calls in non-lto, too;
12346 add simple cost model; produce local aliases.
12348 2013-08-13 David Malcolm <dmalcolm@redhat.com>
12350 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
12353 2013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
12355 * config/i386/i386.c (ix86_function_versions): Use error + inform.
12357 2013-08-12 Uros Bizjak <ubizjak@gmail.com>
12359 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
12360 iterator instead of X87MODEF.
12362 2013-08-12 Perez Read <netfirewall@gmail.com>
12365 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
12366 operand 0 for intel asm alternative.
12367 (*movabs<mode>_2): Ditto for operand 1.
12369 2013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
12371 * config/aarch64/arm_none.h
12372 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
12374 2013-08-12 Nick Clifton <nickc@redhat.com>
12376 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
12378 2013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
12380 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
12381 expand for QI/HImode operand to produce more effictive code for
12382 unsigned char(short) --> float(double) conversion.
12384 2013-08-12 Alexander Monakov <amonakov@ispras.ru>
12386 * doc/invoke.texi: Mention that -ftls-model does not force the final
12389 2013-08-12 Marek Polacek <polacek@redhat.com>
12390 Marc Glisse <marc.glisse@inria.fr>
12392 PR tree-optimization/57980
12393 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
12394 when creating -1 constant.
12396 2013-08-10 Jan Hubicka <jh@suse.cz>
12398 Workaround binutils PR14342.
12399 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
12400 (gimple_init_edge_profiler): Likewise.
12401 (gimple_gen_ic_func_profiler): Likewise.
12403 2013-08-09 Jan Hubicka <jh@suse.cz>
12405 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
12407 2013-08-09 Xinliang David Li <davidxl@google.com>
12409 * config/i386/stringop.def: New file.
12410 * config/i386/stringop.opt: New file.
12411 * config/i386/i386-opts.h: Include stringopt.def.
12412 * config/i386/i386.opt: Include stringopt.opt.
12413 * config/i386/i386.c (ix86_option_override_internal):
12414 Override default size based stringop inline strategies with options.
12415 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
12418 2013-08-09 Jan Hubicka <jh@suse.cz>
12420 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
12422 2013-08-09 Jan Hubicka <jh@suse.cz>
12424 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
12426 (dump_cgraph_node): Dump profile-id.
12427 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
12428 and common_target_probability.
12429 * lto-cgraph.c (lto_output_edge): Stream common targets.
12430 (lto_output_node): Stream profile ids.
12431 (input_node): Stream profile ids.
12432 (input_edge): Stream common targets.
12433 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
12434 * ipa.c: Include value-prof.h
12435 (ipa_profile_generate_summary): Turn indirect call statement histograms
12436 into common targets.
12437 (ipa_profile): Turn common targets into speculative edges.
12439 2013-08-09 Jan Hubicka <jh@suse.cz>
12441 * cgraph.h (cgraph_node): Add profile_id.
12442 * value-prof.c (cgraph_node_map): Turn into pointer_map.
12443 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
12444 (del_node_map): Update.
12445 (find_func_by_funcdef_no): Replace by ...
12446 (find_func_by_profile_id): ... this one.
12447 (gimple_ic_transform): Do not remove useful histograms when
12448 speculation is not done; dump info when indirect call removal
12450 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
12451 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
12452 (__gcov_indirect_call_profiler_v2): .. this one.
12453 * profile.h (init_node_map): Update.
12454 * coverage.c (coverage_compute_profile_id): New function.
12455 * coverage.h (coverage_compute_profile_id): Declare.
12456 * tree-profile.c (init_ic_make_global_vars): Make
12457 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
12458 (gimple_init_edge_profiler): Update prototype of
12459 __gcov_indirect_call_profiler.
12460 (gimple_gen_ic_func_profiler): Simplify.
12461 (tree_profiling): Use init_node_map
12463 2013-08-09 Jan Hubicka <jh@suse.cz>
12465 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
12466 non-speculative refs.
12467 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
12468 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
12469 (cgraph_set_call_stmt): Likewise.
12470 (cgraph_create_edge_1): Fix release checking compilatoin;
12471 clear lto_stmt_uid.
12472 (cgraph_free_edge): Free indirect info.
12473 (cgraph_turn_edge_to_speculative): New function.
12474 (cgraph_speculative_call_info): New function.
12475 (cgraph_make_edge_direct): Return direct edge; handle speculation.
12476 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
12477 (dump_cgraph_node): Dump speculation.
12478 (verify_edge_count_and_frequency): Accept speculative edges.
12479 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
12480 (verify_cgraph_node): Handle speculation.
12481 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
12482 (cgraph_set_call_stmt): Update prototype.
12483 (cgraph_make_edge_direct): Update prototype.
12484 (cgraph_speculative_call_info): Declare.
12485 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
12486 to change; update call of ipa_find_references.
12487 * ipa-ref.c (ipa_record_reference): Fix return value; clear
12488 lto_stmt_uid and speculative flags.
12489 (ipa_dump_references): Dump speculation.
12490 (ipa_clone_references): Clone speculative flag.
12491 (ipa_clone_referring): Likewise.
12492 (ipa_clone_ref): New function.
12493 (ipa_find_reference): Look into lto_stmt_uids
12494 (ipa_clear_stmts_in_references): Do not clear speculative calls.
12495 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
12496 (ipa_find_reference): Update declaration.
12497 (ipa_clone_ref): Declare.
12498 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
12499 stream speculative flag.
12500 (lto_output_ref): Stream statements uids and speculation.
12501 (input_ref): Likewise.
12502 (input_edge): Stream speuclation.
12503 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
12504 (cgraph_set_call_stmt_including_clones): Handle speculation.
12505 * ipa-inline.c (heap_edge_removal_hook): New function.
12506 (inline_small_functions): Register it.
12507 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
12508 also initialize refs.
12509 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
12511 (try_make_edge_direct_simple_call): Likewise.
12512 (try_make_edge_direct_simple_call): Likewise.
12513 (update_indirect_edges_after_inlining): Likewise.
12514 (remove_described_reference): Look proper lto_stmt_uid.
12515 (propagate_controlled_uses): Likewise.
12516 (propagate_controlled_uses): Liekwise.
12517 * tree-inline.c (copy_bb): Copy speculative edges.
12518 (redirect_all_calls): New function.
12519 (copy_cfg_body): Do redirection after loop info is updated.
12520 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
12522 2013-08-09 Jan Hubicka <jh@suse.cz>
12524 * lto-streamer-out.c (output_function): Renumber PHIs.
12525 * lto-streamer-in.c (input_function): Likewise.
12527 2013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
12529 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
12530 (get_lane_unsigned): Likewise.
12531 (dup_lane_scalar): Likewise.
12532 (get_lane): enable for VALL.
12533 * config/aarch64/aarch64-simd.md
12534 (aarch64_dup_lane_scalar<mode>): Remove.
12535 (aarch64_get_lane_signed<mode>): Likewise.
12536 (aarch64_get_lane_unsigned<mode>): Likewise.
12537 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
12538 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
12539 (aarch64_get_lane<mode>): Enable for all vector modes.
12540 (aarch64_get_lanedi): Remove misleading constraints.
12541 * config/aarch64/arm_neon.h
12542 (__aarch64_vget_lane_any): Define.
12543 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
12544 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
12545 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
12546 * config/aarch64/iterators.md (VDQQH): New.
12547 (VDQQHS): Likewise.
12548 (vwcore): Likewise.
12550 2013-08-09 Eric Botcazou <ebotcazou@adacore.com>
12552 * configure.ac: Add GAS check for LEON instructions on SPARC.
12553 * configure: Regenerate.
12554 * config.in: Likewise.
12555 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
12557 * config/sparc/sparc.opt (LEON, LEON3): New masks.
12558 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
12560 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
12561 (AS_LEON_FLAG): New macro.
12562 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
12563 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
12564 Deal with LEON and LEON3 for the memory model.
12565 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
12566 (atomic_compare_and_swap<mode>_1): Likewise.
12567 (*atomic_compare_and_swap<mode>_1): Likewise.
12569 2013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
12571 * config/arm/neon.md (vcond): Fix floating-point vector
12572 comparisons against 0.
12574 2013-08-08 Vladimir Makarov <vmakarov@redhat.com>
12576 * lra-constraints.c (emit_spill_move): Remove assert.
12577 (process_alt_operands): Add more debugging
12578 output. Increase reject for spilling into memory. Decrease
12579 reject for reloading scratch.
12580 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
12582 2013-08-08 Steve Ellcey <sellcey@mips.com>
12584 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
12585 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
12586 micromips incompatible. Add nan2008.
12587 (MULTILIB_DIRNAMES): Add nan2008.
12588 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
12589 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
12590 and micromips incompatible. Add nan2008.
12591 (MULTILIB_DIRNAMES): Add nan2008.
12592 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
12594 2013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
12596 PR rtl-optimization/58079
12597 * combine.c (combine_simplify_rtx): Avoid using SUBST if
12598 simplify_comparison has widened a comparison with an integer.
12600 2013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12602 * config/arm/neon.md (movmisalign<mode>): Disable when we
12603 don't allow unaligned accesses.
12604 (*movmisalign<mode>_neon_store): Likewise.
12605 (*movmisalign<mode>_neon_load): Likewise.
12606 (*movmisalign<mode>_neon_store): Likewise.
12607 (*movmisalign<mode>_neon_load): Likewise.
12609 2013-08-08 Jan Hubicka <jh@suse.cz>
12611 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
12612 (make_pass_rebuild_cgraph_edges): Also clear references.
12613 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
12614 * ipa-inline-transform.c (inline_transform): Remove all references
12616 * cgraphunit.c (expand_function): Remove all references after
12618 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
12619 (ipa_find_reference): Rewrite to iterator.
12620 (remove_stmt_references): Likewise.
12621 (ipa_clear_stmts_in_references): New function.
12622 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
12623 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
12625 * ipa-split.c (split_function): Remove references in split function.
12627 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
12630 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
12631 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
12633 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
12636 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
12637 suggested mode for the assignment isn't compatible with the
12638 registers required.
12640 2013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
12643 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
12645 2013-08-07 Xinliang David Li <davidxl@google.com>
12647 * config/i386/i386.opt: New option -mtune-ctrl=.
12648 * config/i386/x86-tune.def: New file.
12649 * config/i386/i386.h: include x86-tune.def.
12650 * config/i386/i386.c (ix86_option_override_internal):
12651 Parsing -mtune-ctrl= option and set tune features.
12653 2013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
12656 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
12657 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
12660 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
12662 * rtl.h (update_alignments): Declare.
12663 * final.c (grow_label_align): New function extracted from...
12664 (shorten_branches): ...here. Call it.
12665 (update_alignments): New function.
12666 * reorg.c (sibling_labels): New variable.
12667 (get_label_before): Add SIBLING parameter. If it is non-zero, push
12668 the new label along with it onto the sibling_labels vector.
12669 (fill_simple_delay_slots): Adjust call to get_label_before.
12670 (fill_slots_from_thread): Likewise.
12671 (relax_delay_slots): Likewise.
12672 (make_return_insns): Likewise.
12673 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
12675 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
12677 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
12678 document its semantics.
12679 (diagnostic_report_diagnostic): Adjust accordingly.
12681 2013-08-07 David Malcolm <dmalcolm@redhat.com>
12683 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
12684 (sparc_option_override): ...and port to new C++ pass API.
12685 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
12687 2013-08-07 Peter Bergner <bergner@vnet.ibm.com>
12689 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
12691 2013-08-06 Caroline Tice <cmtice@google.com>
12693 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
12694 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
12695 * tree-pass.h: Add pass_vtable_verify.
12696 * varasm.c (assemble_variable): Add code to properly set the comdat
12697 section and name for the .vtable_map_vars section.
12698 (assemble_vtyv_preinit_initializer): New function.
12699 (default_sectin_type_flags): Make sure .vtable_map_vars section has
12701 * output.h: Add function decl for assemble_vtv_preinit_initializer.
12702 * vtable-verify.c: New file.
12703 * vtable-verify.h: New file.
12704 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
12705 initialiation levels.
12706 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
12707 * passes.def: Insert pass_vtable_verify.
12708 * aclocal.m4: Reorder includes.
12709 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
12710 -fvtv-counts options.
12711 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
12712 as appropriate, if -fvtable-verify=... is used.
12713 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
12714 -fvtable-verify=... is used.
12715 * Makefile.in (OBJS): Add vtable-verify.o to list.
12716 (vtable-verify.o): Add new build rule.
12717 (GTFILES): Add vtable-verify.c to list.
12718 * common.opt (fvtable-verify=): New flag.
12719 (vtv_priority): Values for fvtable-verify= flag.
12720 (fvtv-counts): New flag.
12721 (fvtv-debug): New flag.
12722 * tree.h (save_vtable_map_decl): New extern function decl.
12724 2013-08-07 David Malcolm <dmalcolm@redhat.com>
12726 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
12727 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
12728 (pass_data_rl78_devirt): ...new pass_data instance and...
12729 (make_pass_rl78_devirt): ...new function.
12730 (rl78_asm_file_start): Port pass registration to new C++ API.
12732 2013-08-07 David Malcolm <dmalcolm@redhat.com>
12734 * coretypes.h (rtl_opt_pass): Add.
12735 (gcc::context): Add.
12736 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
12737 (epiphany_init): Port to new C++ pass API.
12738 (epiphany_optimize_mode_switching): Likewise.
12739 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
12740 (pass_manager::get_pass_mode_switching): New.
12741 (pass_manager::get_pass_peephole2): New.
12742 * mode-switching.c (pass_mode_switching): Add clone method.
12743 * recog.c (pass_peephole2): Add clone method.
12744 (pass_split_all_insns): Add clone method.
12746 2013-08-06 David Malcolm <dmalcolm@redhat.com>
12748 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
12749 (mips_option_override): ...here, porting to new C++ API for passes.
12751 2013-08-06 Jan Hubicka <jh@suse.cz>
12753 * cgraph.c (cgraph_get_body): New function based on lto.c
12755 * cgraph.h (cgraph_get_body): Declare.
12756 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
12758 * cgraphunit.c (expand_function): Get body prior expanding.
12759 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
12760 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
12762 * passes.c (do_per_function_toporder): Get body.
12763 * tree-inline.c (expand_call_inline): Get body prior inlining it.
12764 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
12766 2013-08-06 Martin Jambor <mjambor@suse.cz>
12769 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
12770 is not re-finalized. Rename second parameter to no_collect.
12772 2013-08-06 Martin Jambor <mjambor@suse.cz>
12774 PR middle-end/58041
12775 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
12776 MEM_REF has proper alignment information.
12778 2013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
12781 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
12783 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
12784 rtx (*) (rtx, ...) with insn_gen_fn.
12785 * genoutput.c (output_insn_data): Cast gen_? function pointers to
12786 insn_gen_fn::stored_funcptr. Add initializer braces.
12788 2013-08-05 David Malcolm <dmalcolm@redhat.com>
12790 Rewrite how instances of passes are cloned to remove assumptions
12791 about their sizes (thus allowing pass subclasses to have
12792 additional data fields, albeit non-GC-managed ones at this point).
12794 * passes.c (make_pass_instance): Now that passes have clone
12795 methods, rewrite this function to eliminate XNEW and memcpy
12796 calls that used hardcoded sizes. Since this function no longer
12797 creates pass instances, rename it to...
12798 (add_pass_instance): ...this. Document the old way that passes were
12799 numbered and flagged, and rework this function to continue using it.
12800 (next_pass_1): Add an initial_pass argument for use by
12802 (position_pass): When adding multiple instances of a pass, use
12803 the pass's clone method, rather than relying on the XNEW/memcpy
12804 within the former make_pass_instance (now add_pass_instance).
12805 (pass_manager::pass_manager): When invoking next_pass_1, also supply
12806 the initial instance of the current pass within the pass manager.
12808 2013-08-05 David Malcolm <dmalcolm@redhat.com>
12810 This is the automated part of the conversion of passes from C
12811 structs to C++ classes.
12813 Patch autogenerated by refactor_passes.py from
12814 https://github.com/davidmalcolm/gcc-refactoring-scripts
12815 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
12817 * asan.c (pass_asan): Convert from a global struct to a subclass of
12818 gimple_opt_pass along with...
12819 (pass_data_asan): ...new pass_data instance and...
12820 (make_pass_asan): ...new function.
12821 (pass_asan_O0): Convert from a global struct to a subclass of
12822 gimple_opt_pass along with...
12823 (pass_data_asan_O0): ...new pass_data instance and...
12824 (make_pass_asan_O0): ...new function.
12825 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
12826 subclass of rtl_opt_pass along with...
12827 (pass_data_inc_dec): ...new pass_data instance and...
12828 (make_pass_inc_dec): ...new function.
12829 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
12830 a subclass of rtl_opt_pass along with...
12831 (pass_data_reorder_blocks): ...new pass_data instance and...
12832 (make_pass_reorder_blocks): ...new function.
12833 (pass_duplicate_computed_gotos): Convert from a global struct to a
12834 subclass of rtl_opt_pass along with...
12835 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
12836 (make_pass_duplicate_computed_gotos): ...new function.
12837 (pass_partition_blocks): Convert from a global struct to a subclass of
12838 rtl_opt_pass along with...
12839 (pass_data_partition_blocks): ...new pass_data instance and...
12840 (make_pass_partition_blocks): ...new function.
12841 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
12842 struct to a subclass of rtl_opt_pass along with...
12843 (pass_data_branch_target_load_optimize1): ...new pass_data instance
12845 (make_pass_branch_target_load_optimize1): ...new function.
12846 (pass_branch_target_load_optimize2): Convert from a global struct to a
12847 subclass of rtl_opt_pass along with...
12848 (pass_data_branch_target_load_optimize2): ...new pass_data instance
12850 (make_pass_branch_target_load_optimize2): ...new function.
12851 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
12852 of rtl_opt_pass along with...
12853 (pass_data_jump): ...new pass_data instance and...
12854 (make_pass_jump): ...new function.
12855 (pass_jump2): Convert from a global struct to a subclass of
12856 rtl_opt_pass along with...
12857 (pass_data_jump2): ...new pass_data instance and...
12858 (make_pass_jump2): ...new function.
12859 * cfgexpand.c (pass_expand): Convert from a global struct to a
12860 subclass of rtl_opt_pass along with...
12861 (pass_data_expand): ...new pass_data instance and...
12862 (make_pass_expand): ...new function.
12863 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
12864 of rtl_opt_pass along with...
12865 (pass_data_free_cfg): ...new pass_data instance and...
12866 (make_pass_free_cfg): ...new function.
12867 (pass_into_cfg_layout_mode): Convert from a global struct to a
12868 subclass of rtl_opt_pass along with...
12869 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
12870 (make_pass_into_cfg_layout_mode): ...new function.
12871 (pass_outof_cfg_layout_mode): Convert from a global struct to a
12872 subclass of rtl_opt_pass along with...
12873 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
12874 (make_pass_outof_cfg_layout_mode): ...new function.
12875 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
12876 struct to a subclass of gimple_opt_pass along with...
12877 (pass_data_build_cgraph_edges): ...new pass_data instance and...
12878 (make_pass_build_cgraph_edges): ...new function.
12879 (pass_rebuild_cgraph_edges): Convert from a global struct to a
12880 subclass of gimple_opt_pass along with...
12881 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
12882 (make_pass_rebuild_cgraph_edges): ...new function.
12883 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
12884 subclass of gimple_opt_pass along with...
12885 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
12887 (make_pass_remove_cgraph_callee_edges): ...new function.
12888 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
12889 struct to a subclass of rtl_opt_pass along with...
12890 (pass_data_stack_adjustments): ...new pass_data instance and...
12891 (make_pass_stack_adjustments): ...new function.
12892 * combine.c (pass_combine): Convert from a global struct to a subclass
12893 of rtl_opt_pass along with...
12894 (pass_data_combine): ...new pass_data instance and...
12895 (make_pass_combine): ...new function.
12896 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
12897 global struct to a subclass of rtl_opt_pass along with...
12898 (pass_data_compare_elim_after_reload): ...new pass_data instance
12900 (make_pass_compare_elim_after_reload): ...new function.
12901 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
12902 of rtl_opt_pass along with...
12903 (pass_data_rtl_cprop): ...new pass_data instance and...
12904 (make_pass_rtl_cprop): ...new function.
12905 * cse.c (pass_cse): Convert from a global struct to a subclass of
12906 rtl_opt_pass along with...
12907 (pass_data_cse): ...new pass_data instance and...
12908 (make_pass_cse): ...new function.
12909 (pass_cse2): Convert from a global struct to a subclass of
12910 rtl_opt_pass along with...
12911 (pass_data_cse2): ...new pass_data instance and...
12912 (make_pass_cse2): ...new function.
12913 (pass_cse_after_global_opts): Convert from a global struct to a
12914 subclass of rtl_opt_pass along with...
12915 (pass_data_cse_after_global_opts): ...new pass_data instance and...
12916 (make_pass_cse_after_global_opts): ...new function.
12917 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
12918 of rtl_opt_pass along with...
12919 (pass_data_ud_rtl_dce): ...new pass_data instance and...
12920 (make_pass_ud_rtl_dce): ...new function.
12921 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
12922 rtl_opt_pass along with...
12923 (pass_data_fast_rtl_dce): ...new pass_data instance and...
12924 (make_pass_fast_rtl_dce): ...new function.
12925 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
12926 a subclass of rtl_opt_pass along with...
12927 (pass_data_df_initialize_opt): ...new pass_data instance and...
12928 (make_pass_df_initialize_opt): ...new function.
12929 (pass_df_initialize_no_opt): Convert from a global struct to a
12930 subclass of rtl_opt_pass along with...
12931 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
12932 (make_pass_df_initialize_no_opt): ...new function.
12933 (pass_df_finish): Convert from a global struct to a subclass of
12934 rtl_opt_pass along with...
12935 (pass_data_df_finish): ...new pass_data instance and...
12936 (make_pass_df_finish): ...new function.
12937 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
12938 rtl_opt_pass along with...
12939 (pass_data_rtl_dse1): ...new pass_data instance and...
12940 (make_pass_rtl_dse1): ...new function.
12941 (pass_rtl_dse2): Convert from a global struct to a subclass of
12942 rtl_opt_pass along with...
12943 (pass_data_rtl_dse2): ...new pass_data instance and...
12944 (make_pass_rtl_dse2): ...new function.
12945 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
12946 subclass of rtl_opt_pass along with...
12947 (pass_data_dwarf2_frame): ...new pass_data instance and...
12948 (make_pass_dwarf2_frame): ...new function.
12949 * except.c (pass_set_nothrow_function_flags): Convert from a global
12950 struct to a subclass of rtl_opt_pass along with...
12951 (pass_data_set_nothrow_function_flags): ...new pass_data instance
12953 (make_pass_set_nothrow_function_flags): ...new function.
12954 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
12955 subclass of rtl_opt_pass along with...
12956 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
12958 (make_pass_convert_to_eh_region_ranges): ...new function.
12959 * final.c (pass_compute_alignments): Convert from a global struct to a
12960 subclass of rtl_opt_pass along with...
12961 (pass_data_compute_alignments): ...new pass_data instance and...
12962 (make_pass_compute_alignments): ...new function.
12963 (pass_final): Convert from a global struct to a subclass of
12964 rtl_opt_pass along with...
12965 (pass_data_final): ...new pass_data instance and...
12966 (make_pass_final): ...new function.
12967 (pass_shorten_branches): Convert from a global struct to a subclass of
12968 rtl_opt_pass along with...
12969 (pass_data_shorten_branches): ...new pass_data instance and...
12970 (make_pass_shorten_branches): ...new function.
12971 (pass_clean_state): Convert from a global struct to a subclass of
12972 rtl_opt_pass along with...
12973 (pass_data_clean_state): ...new pass_data instance and...
12974 (make_pass_clean_state): ...new function.
12975 * function.c (pass_instantiate_virtual_regs): Convert from a global
12976 struct to a subclass of rtl_opt_pass along with...
12977 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
12978 (make_pass_instantiate_virtual_regs): ...new function.
12979 (pass_leaf_regs): Convert from a global struct to a subclass of
12980 rtl_opt_pass along with...
12981 (pass_data_leaf_regs): ...new pass_data instance and...
12982 (make_pass_leaf_regs): ...new function.
12983 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
12984 subclass of rtl_opt_pass along with...
12985 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
12987 (make_pass_thread_prologue_and_epilogue): ...new function.
12988 (pass_match_asm_constraints): Convert from a global struct to a
12989 subclass of rtl_opt_pass along with...
12990 (pass_data_match_asm_constraints): ...new pass_data instance and...
12991 (make_pass_match_asm_constraints): ...new function.
12992 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
12993 subclass of rtl_opt_pass along with...
12994 (pass_data_rtl_fwprop): ...new pass_data instance and...
12995 (make_pass_rtl_fwprop): ...new function.
12996 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
12997 rtl_opt_pass along with...
12998 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
12999 (make_pass_rtl_fwprop_addr): ...new function.
13000 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
13001 rtl_opt_pass along with...
13002 (pass_data_rtl_pre): ...new pass_data instance and...
13003 (make_pass_rtl_pre): ...new function.
13004 (pass_rtl_hoist): Convert from a global struct to a subclass of
13005 rtl_opt_pass along with...
13006 (pass_data_rtl_hoist): ...new pass_data instance and...
13007 (make_pass_rtl_hoist): ...new function.
13008 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
13009 subclass of gimple_opt_pass along with...
13010 (pass_data_lower_cf): ...new pass_data instance and...
13011 (make_pass_lower_cf): ...new function.
13012 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
13013 from a global struct to a subclass of gimple_opt_pass along with...
13014 (pass_data_strength_reduction): ...new pass_data instance and...
13015 (make_pass_strength_reduction): ...new function.
13016 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
13017 of rtl_opt_pass along with...
13018 (pass_data_rtl_ifcvt): ...new pass_data instance and...
13019 (make_pass_rtl_ifcvt): ...new function.
13020 (pass_if_after_combine): Convert from a global struct to a subclass of
13021 rtl_opt_pass along with...
13022 (pass_data_if_after_combine): ...new pass_data instance and...
13023 (make_pass_if_after_combine): ...new function.
13024 (pass_if_after_reload): Convert from a global struct to a subclass of
13025 rtl_opt_pass along with...
13026 (pass_data_if_after_reload): ...new pass_data instance and...
13027 (make_pass_if_after_reload): ...new function.
13028 * init-regs.c (pass_initialize_regs): Convert from a global struct to
13029 a subclass of rtl_opt_pass along with...
13030 (pass_data_initialize_regs): ...new pass_data instance and...
13031 (make_pass_initialize_regs): ...new function.
13032 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
13033 of ipa_opt_pass_d along with...
13034 (pass_data_ipa_cp): ...new pass_data instance and...
13035 (make_pass_ipa_cp): ...new function.
13036 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
13037 global struct to a subclass of gimple_opt_pass along with...
13038 (pass_data_inline_parameters): ...new pass_data instance and...
13039 (make_pass_inline_parameters): ...new function.
13040 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
13041 subclass of gimple_opt_pass along with...
13042 (pass_data_early_inline): ...new pass_data instance and...
13043 (make_pass_early_inline): ...new function.
13044 (pass_ipa_inline): Convert from a global struct to a subclass of
13045 ipa_opt_pass_d along with...
13046 (pass_data_ipa_inline): ...new pass_data instance and...
13047 (make_pass_ipa_inline): ...new function.
13048 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
13049 struct to a subclass of gimple_opt_pass along with...
13050 (pass_data_local_pure_const): ...new pass_data instance and...
13051 (make_pass_local_pure_const): ...new function.
13052 (pass_ipa_pure_const): Convert from a global struct to a subclass of
13053 ipa_opt_pass_d along with...
13054 (pass_data_ipa_pure_const): ...new pass_data instance and...
13055 (make_pass_ipa_pure_const): ...new function.
13056 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
13057 to a subclass of ipa_opt_pass_d along with...
13058 (pass_data_ipa_reference): ...new pass_data instance and...
13059 (make_pass_ipa_reference): ...new function.
13060 * ipa-split.c (pass_split_functions): Convert from a global struct to
13061 a subclass of gimple_opt_pass along with...
13062 (pass_data_split_functions): ...new pass_data instance and...
13063 (make_pass_split_functions): ...new function.
13064 (pass_feedback_split_functions): Convert from a global struct to a
13065 subclass of gimple_opt_pass along with...
13066 (pass_data_feedback_split_functions): ...new pass_data instance and...
13067 (make_pass_feedback_split_functions): ...new function.
13068 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
13069 global struct to a subclass of simple_ipa_opt_pass along with...
13070 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
13072 (make_pass_ipa_function_and_variable_visibility): ...new function.
13073 (pass_ipa_free_inline_summary): Convert from a global struct to a
13074 subclass of simple_ipa_opt_pass along with...
13075 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
13076 (make_pass_ipa_free_inline_summary): ...new function.
13077 (pass_ipa_whole_program_visibility): Convert from a global struct to a
13078 subclass of ipa_opt_pass_d along with...
13079 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
13081 (make_pass_ipa_whole_program_visibility): ...new function.
13082 (pass_ipa_profile): Convert from a global struct to a subclass of
13083 ipa_opt_pass_d along with...
13084 (pass_data_ipa_profile): ...new pass_data instance and...
13085 (make_pass_ipa_profile): ...new function.
13086 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
13087 ipa_opt_pass_d along with...
13088 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
13089 (make_pass_ipa_cdtor_merge): ...new function.
13090 * ira.c (pass_ira): Convert from a global struct to a subclass of
13091 rtl_opt_pass along with...
13092 (pass_data_ira): ...new pass_data instance and...
13093 (make_pass_ira): ...new function.
13094 (pass_reload): Convert from a global struct to a subclass of
13095 rtl_opt_pass along with...
13096 (pass_data_reload): ...new pass_data instance and...
13097 (make_pass_reload): ...new function.
13098 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
13099 subclass of rtl_opt_pass along with...
13100 (pass_data_cleanup_barriers): ...new pass_data instance and...
13101 (make_pass_cleanup_barriers): ...new function.
13102 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
13103 of rtl_opt_pass along with...
13104 (pass_data_loop2): ...new pass_data instance and...
13105 (make_pass_loop2): ...new function.
13106 (pass_rtl_loop_init): Convert from a global struct to a subclass of
13107 rtl_opt_pass along with...
13108 (pass_data_rtl_loop_init): ...new pass_data instance and...
13109 (make_pass_rtl_loop_init): ...new function.
13110 (pass_rtl_loop_done): Convert from a global struct to a subclass of
13111 rtl_opt_pass along with...
13112 (pass_data_rtl_loop_done): ...new pass_data instance and...
13113 (make_pass_rtl_loop_done): ...new function.
13114 (pass_rtl_move_loop_invariants): Convert from a global struct to a
13115 subclass of rtl_opt_pass along with...
13116 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
13117 (make_pass_rtl_move_loop_invariants): ...new function.
13118 (pass_rtl_unswitch): Convert from a global struct to a subclass of
13119 rtl_opt_pass along with...
13120 (pass_data_rtl_unswitch): ...new pass_data instance and...
13121 (make_pass_rtl_unswitch): ...new function.
13122 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
13123 subclass of rtl_opt_pass along with...
13124 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
13126 (make_pass_rtl_unroll_and_peel_loops): ...new function.
13127 (pass_rtl_doloop): Convert from a global struct to a subclass of
13128 rtl_opt_pass along with...
13129 (pass_data_rtl_doloop): ...new pass_data instance and...
13130 (make_pass_rtl_doloop): ...new function.
13131 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
13132 a subclass of rtl_opt_pass along with...
13133 (pass_data_lower_subreg): ...new pass_data instance and...
13134 (make_pass_lower_subreg): ...new function.
13135 (pass_lower_subreg2): Convert from a global struct to a subclass of
13136 rtl_opt_pass along with...
13137 (pass_data_lower_subreg2): ...new pass_data instance and...
13138 (make_pass_lower_subreg2): ...new function.
13139 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
13140 struct to a subclass of ipa_opt_pass_d along with...
13141 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
13142 (make_pass_ipa_lto_gimple_out): ...new function.
13143 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
13144 of ipa_opt_pass_d along with...
13145 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
13146 (make_pass_ipa_lto_finish_out): ...new function.
13147 * mode-switching.c (pass_mode_switching): Convert from a global struct
13148 to a subclass of rtl_opt_pass along with...
13149 (pass_data_mode_switching): ...new pass_data instance and...
13150 (make_pass_mode_switching): ...new function.
13151 * modulo-sched.c (pass_sms): Convert from a global struct to a
13152 subclass of rtl_opt_pass along with...
13153 (pass_data_sms): ...new pass_data instance and...
13154 (make_pass_sms): ...new function.
13155 * omp-low.c (pass_expand_omp): Convert from a global struct to a
13156 subclass of gimple_opt_pass along with...
13157 (pass_data_expand_omp): ...new pass_data instance and...
13158 (make_pass_expand_omp): ...new function.
13159 (pass_lower_omp): Convert from a global struct to a subclass of
13160 gimple_opt_pass along with...
13161 (pass_data_lower_omp): ...new pass_data instance and...
13162 (make_pass_lower_omp): ...new function.
13163 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
13164 of gimple_opt_pass along with...
13165 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
13166 (make_pass_diagnose_omp_blocks): ...new function.
13167 * passes.c (pass_early_local_passes): Convert from a global struct to
13168 a subclass of simple_ipa_opt_pass along with...
13169 (pass_data_early_local_passes): ...new pass_data instance and...
13170 (make_pass_early_local_passes): ...new function.
13171 (pass_all_early_optimizations): Convert from a global struct to a
13172 subclass of gimple_opt_pass along with...
13173 (pass_data_all_early_optimizations): ...new pass_data instance and...
13174 (make_pass_all_early_optimizations): ...new function.
13175 (pass_all_optimizations): Convert from a global struct to a subclass
13176 of gimple_opt_pass along with...
13177 (pass_data_all_optimizations): ...new pass_data instance and...
13178 (make_pass_all_optimizations): ...new function.
13179 (pass_all_optimizations_g): Convert from a global struct to a subclass
13180 of gimple_opt_pass along with...
13181 (pass_data_all_optimizations_g): ...new pass_data instance and...
13182 (make_pass_all_optimizations_g): ...new function.
13183 (pass_rest_of_compilation): Convert from a global struct to a subclass
13184 of rtl_opt_pass along with...
13185 (pass_data_rest_of_compilation): ...new pass_data instance and...
13186 (make_pass_rest_of_compilation): ...new function.
13187 (pass_postreload): Convert from a global struct to a subclass of
13188 rtl_opt_pass along with...
13189 (pass_data_postreload): ...new pass_data instance and...
13190 (make_pass_postreload): ...new function.
13191 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
13192 subclass of rtl_opt_pass along with...
13193 (pass_data_gcse2): ...new pass_data instance and...
13194 (make_pass_gcse2): ...new function.
13195 * postreload.c (pass_postreload_cse): Convert from a global struct to
13196 a subclass of rtl_opt_pass along with...
13197 (pass_data_postreload_cse): ...new pass_data instance and...
13198 (make_pass_postreload_cse): ...new function.
13199 * predict.c (pass_profile): Convert from a global struct to a subclass
13200 of gimple_opt_pass along with...
13201 (pass_data_profile): ...new pass_data instance and...
13202 (make_pass_profile): ...new function.
13203 (pass_strip_predict_hints): Convert from a global struct to a subclass
13204 of gimple_opt_pass along with...
13205 (pass_data_strip_predict_hints): ...new pass_data instance and...
13206 (make_pass_strip_predict_hints): ...new function.
13207 * recog.c (pass_peephole2): Convert from a global struct to a subclass
13208 of rtl_opt_pass along with...
13209 (pass_data_peephole2): ...new pass_data instance and...
13210 (make_pass_peephole2): ...new function.
13211 (pass_split_all_insns): Convert from a global struct to a subclass of
13212 rtl_opt_pass along with...
13213 (pass_data_split_all_insns): ...new pass_data instance and...
13214 (make_pass_split_all_insns): ...new function.
13215 (pass_split_after_reload): Convert from a global struct to a subclass
13216 of rtl_opt_pass along with...
13217 (pass_data_split_after_reload): ...new pass_data instance and...
13218 (make_pass_split_after_reload): ...new function.
13219 (pass_split_before_regstack): Convert from a global struct to a
13220 subclass of rtl_opt_pass along with...
13221 (pass_data_split_before_regstack): ...new pass_data instance and...
13222 (make_pass_split_before_regstack): ...new function.
13223 (pass_split_before_sched2): Convert from a global struct to a subclass
13224 of rtl_opt_pass along with...
13225 (pass_data_split_before_sched2): ...new pass_data instance and...
13226 (make_pass_split_before_sched2): ...new function.
13227 (pass_split_for_shorten_branches): Convert from a global struct to a
13228 subclass of rtl_opt_pass along with...
13229 (pass_data_split_for_shorten_branches): ...new pass_data instance
13231 (make_pass_split_for_shorten_branches): ...new function.
13232 * ree.c (pass_ree): Convert from a global struct to a subclass of
13233 rtl_opt_pass along with...
13234 (pass_data_ree): ...new pass_data instance and...
13235 (make_pass_ree): ...new function.
13236 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
13237 subclass of rtl_opt_pass along with...
13238 (pass_data_stack_regs): ...new pass_data instance and...
13239 (make_pass_stack_regs): ...new function.
13240 (pass_stack_regs_run): Convert from a global struct to a subclass of
13241 rtl_opt_pass along with...
13242 (pass_data_stack_regs_run): ...new pass_data instance and...
13243 (make_pass_stack_regs_run): ...new function.
13244 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
13245 subclass of rtl_opt_pass along with...
13246 (pass_data_cprop_hardreg): ...new pass_data instance and...
13247 (make_pass_cprop_hardreg): ...new function.
13248 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
13249 subclass of rtl_opt_pass along with...
13250 (pass_data_reginfo_init): ...new pass_data instance and...
13251 (make_pass_reginfo_init): ...new function.
13252 * regmove.c (pass_regmove): Convert from a global struct to a subclass
13253 of rtl_opt_pass along with...
13254 (pass_data_regmove): ...new pass_data instance and...
13255 (make_pass_regmove): ...new function.
13256 * regrename.c (pass_regrename): Convert from a global struct to a
13257 subclass of rtl_opt_pass along with...
13258 (pass_data_regrename): ...new pass_data instance and...
13259 (make_pass_regrename): ...new function.
13260 * reorg.c (pass_delay_slots): Convert from a global struct to a
13261 subclass of rtl_opt_pass along with...
13262 (pass_data_delay_slots): ...new pass_data instance and...
13263 (make_pass_delay_slots): ...new function.
13264 (pass_machine_reorg): Convert from a global struct to a subclass of
13265 rtl_opt_pass along with...
13266 (pass_data_machine_reorg): ...new pass_data instance and...
13267 (make_pass_machine_reorg): ...new function.
13268 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
13269 of rtl_opt_pass along with...
13270 (pass_data_sched): ...new pass_data instance and...
13271 (make_pass_sched): ...new function.
13272 (pass_sched2): Convert from a global struct to a subclass of
13273 rtl_opt_pass along with...
13274 (pass_data_sched2): ...new pass_data instance and...
13275 (make_pass_sched2): ...new function.
13276 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
13277 to a subclass of rtl_opt_pass along with...
13278 (pass_data_stack_ptr_mod): ...new pass_data instance and...
13279 (make_pass_stack_ptr_mod): ...new function.
13280 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
13281 to a subclass of rtl_opt_pass along with...
13282 (pass_data_rtl_store_motion): ...new pass_data instance and...
13283 (make_pass_rtl_store_motion): ...new function.
13284 * tracer.c (pass_tracer): Convert from a global struct to a subclass
13285 of gimple_opt_pass along with...
13286 (pass_data_tracer): ...new pass_data instance and...
13287 (make_pass_tracer): ...new function.
13288 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
13289 to a subclass of gimple_opt_pass along with...
13290 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
13291 (make_pass_diagnose_tm_blocks): ...new function.
13292 (pass_lower_tm): Convert from a global struct to a subclass of
13293 gimple_opt_pass along with...
13294 (pass_data_lower_tm): ...new pass_data instance and...
13295 (make_pass_lower_tm): ...new function.
13296 (pass_tm_init): Convert from a global struct to a subclass of
13297 gimple_opt_pass along with...
13298 (pass_data_tm_init): ...new pass_data instance and...
13299 (make_pass_tm_init): ...new function.
13300 (pass_tm_mark): Convert from a global struct to a subclass of
13301 gimple_opt_pass along with...
13302 (pass_data_tm_mark): ...new pass_data instance and...
13303 (make_pass_tm_mark): ...new function.
13304 (pass_tm_edges): Convert from a global struct to a subclass of
13305 gimple_opt_pass along with...
13306 (pass_data_tm_edges): ...new pass_data instance and...
13307 (make_pass_tm_edges): ...new function.
13308 (pass_tm_memopt): Convert from a global struct to a subclass of
13309 gimple_opt_pass along with...
13310 (pass_data_tm_memopt): ...new pass_data instance and...
13311 (make_pass_tm_memopt): ...new function.
13312 (pass_ipa_tm): Convert from a global struct to a subclass of
13313 simple_ipa_opt_pass along with...
13314 (pass_data_ipa_tm): ...new pass_data instance and...
13315 (make_pass_ipa_tm): ...new function.
13316 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
13317 subclass of gimple_opt_pass along with...
13318 (pass_data_call_cdce): ...new pass_data instance and...
13319 (make_pass_call_cdce): ...new function.
13320 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
13321 subclass of gimple_opt_pass along with...
13322 (pass_data_build_cfg): ...new pass_data instance and...
13323 (make_pass_build_cfg): ...new function.
13324 (pass_split_crit_edges): Convert from a global struct to a subclass of
13325 gimple_opt_pass along with...
13326 (pass_data_split_crit_edges): ...new pass_data instance and...
13327 (make_pass_split_crit_edges): ...new function.
13328 (pass_warn_function_return): Convert from a global struct to a
13329 subclass of gimple_opt_pass along with...
13330 (pass_data_warn_function_return): ...new pass_data instance and...
13331 (make_pass_warn_function_return): ...new function.
13332 (pass_warn_function_noreturn): Convert from a global struct to a
13333 subclass of gimple_opt_pass along with...
13334 (pass_data_warn_function_noreturn): ...new pass_data instance and...
13335 (make_pass_warn_function_noreturn): ...new function.
13336 (pass_warn_unused_result): Convert from a global struct to a subclass
13337 of gimple_opt_pass along with...
13338 (pass_data_warn_unused_result): ...new pass_data instance and...
13339 (make_pass_warn_unused_result): ...new function.
13340 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
13341 a subclass of gimple_opt_pass along with...
13342 (pass_data_merge_phi): ...new pass_data instance and...
13343 (make_pass_merge_phi): ...new function.
13344 * tree-complex.c (pass_lower_complex): Convert from a global struct to
13345 a subclass of gimple_opt_pass along with...
13346 (pass_data_lower_complex): ...new pass_data instance and...
13347 (make_pass_lower_complex): ...new function.
13348 (pass_lower_complex_O0): Convert from a global struct to a subclass of
13349 gimple_opt_pass along with...
13350 (pass_data_lower_complex_O0): ...new pass_data instance and...
13351 (make_pass_lower_complex_O0): ...new function.
13352 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
13353 subclass of gimple_opt_pass along with...
13354 (pass_data_lower_eh): ...new pass_data instance and...
13355 (make_pass_lower_eh): ...new function.
13356 (pass_refactor_eh): Convert from a global struct to a subclass of
13357 gimple_opt_pass along with...
13358 (pass_data_refactor_eh): ...new pass_data instance and...
13359 (make_pass_refactor_eh): ...new function.
13360 (pass_lower_resx): Convert from a global struct to a subclass of
13361 gimple_opt_pass along with...
13362 (pass_data_lower_resx): ...new pass_data instance and...
13363 (make_pass_lower_resx): ...new function.
13364 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
13365 of gimple_opt_pass along with...
13366 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
13367 (make_pass_lower_eh_dispatch): ...new function.
13368 (pass_cleanup_eh): Convert from a global struct to a subclass of
13369 gimple_opt_pass along with...
13370 (pass_data_cleanup_eh): ...new pass_data instance and...
13371 (make_pass_cleanup_eh): ...new function.
13372 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
13373 to a subclass of simple_ipa_opt_pass along with...
13374 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
13375 (make_pass_ipa_lower_emutls): ...new function.
13376 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
13377 a subclass of gimple_opt_pass along with...
13378 (pass_data_if_conversion): ...new pass_data instance and...
13379 (make_pass_if_conversion): ...new function.
13380 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
13381 subclass of gimple_opt_pass along with...
13382 (pass_data_build_ssa): ...new pass_data instance and...
13383 (make_pass_build_ssa): ...new function.
13384 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
13385 global struct to a subclass of gimple_opt_pass along with...
13386 (pass_data_loop_distribution): ...new pass_data instance and...
13387 (make_pass_loop_distribution): ...new function.
13388 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
13389 subclass of gimple_opt_pass along with...
13390 (pass_data_mudflap_1): ...new pass_data instance and...
13391 (make_pass_mudflap_1): ...new function.
13392 (pass_mudflap_2): Convert from a global struct to a subclass of
13393 gimple_opt_pass along with...
13394 (pass_data_mudflap_2): ...new pass_data instance and...
13395 (make_pass_mudflap_2): ...new function.
13396 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
13397 subclass of gimple_opt_pass along with...
13398 (pass_data_mudflap_1): ...new pass_data instance and...
13399 (make_pass_mudflap_1): ...new function.
13400 (pass_mudflap_2): Convert from a global struct to a subclass of
13401 gimple_opt_pass along with...
13402 (pass_data_mudflap_2): ...new pass_data instance and...
13403 (make_pass_mudflap_2): ...new function.
13404 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
13405 gimple_opt_pass along with...
13406 (pass_data_nrv): ...new pass_data instance and...
13407 (make_pass_nrv): ...new function.
13408 (pass_return_slot): Convert from a global struct to a subclass of
13409 gimple_opt_pass along with...
13410 (pass_data_return_slot): ...new pass_data instance and...
13411 (make_pass_return_slot): ...new function.
13412 * tree-object-size.c (pass_object_sizes): Convert from a global struct
13413 to a subclass of gimple_opt_pass along with...
13414 (pass_data_object_sizes): ...new pass_data instance and...
13415 (make_pass_object_sizes): ...new function.
13416 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
13417 global struct to a subclass of gimple_opt_pass along with...
13418 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
13420 (make_pass_cleanup_cfg_post_optimizing): ...new function.
13421 (pass_fixup_cfg): Convert from a global struct to a subclass of
13422 gimple_opt_pass along with...
13423 (pass_data_fixup_cfg): ...new pass_data instance and...
13424 (make_pass_fixup_cfg): ...new function.
13425 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
13426 (make_pass_mudflap_1): ...new function.
13427 (pass_mudflap_2): Replace declaration with that of...
13428 (make_pass_mudflap_2): ...new function.
13429 (pass_asan): Replace declaration with that of...
13430 (make_pass_asan): ...new function.
13431 (pass_asan_O0): Replace declaration with that of...
13432 (make_pass_asan_O0): ...new function.
13433 (pass_tsan): Replace declaration with that of...
13434 (make_pass_tsan): ...new function.
13435 (pass_tsan_O0): Replace declaration with that of...
13436 (make_pass_tsan_O0): ...new function.
13437 (pass_lower_cf): Replace declaration with that of...
13438 (make_pass_lower_cf): ...new function.
13439 (pass_refactor_eh): Replace declaration with that of...
13440 (make_pass_refactor_eh): ...new function.
13441 (pass_lower_eh): Replace declaration with that of...
13442 (make_pass_lower_eh): ...new function.
13443 (pass_lower_eh_dispatch): Replace declaration with that of...
13444 (make_pass_lower_eh_dispatch): ...new function.
13445 (pass_lower_resx): Replace declaration with that of...
13446 (make_pass_lower_resx): ...new function.
13447 (pass_build_cfg): Replace declaration with that of...
13448 (make_pass_build_cfg): ...new function.
13449 (pass_early_tree_profile): Replace declaration with that of...
13450 (make_pass_early_tree_profile): ...new function.
13451 (pass_cleanup_eh): Replace declaration with that of...
13452 (make_pass_cleanup_eh): ...new function.
13453 (pass_sra): Replace declaration with that of...
13454 (make_pass_sra): ...new function.
13455 (pass_sra_early): Replace declaration with that of...
13456 (make_pass_sra_early): ...new function.
13457 (pass_early_ipa_sra): Replace declaration with that of...
13458 (make_pass_early_ipa_sra): ...new function.
13459 (pass_tail_recursion): Replace declaration with that of...
13460 (make_pass_tail_recursion): ...new function.
13461 (pass_tail_calls): Replace declaration with that of...
13462 (make_pass_tail_calls): ...new function.
13463 (pass_tree_loop): Replace declaration with that of...
13464 (make_pass_tree_loop): ...new function.
13465 (pass_tree_loop_init): Replace declaration with that of...
13466 (make_pass_tree_loop_init): ...new function.
13467 (pass_lim): Replace declaration with that of...
13468 (make_pass_lim): ...new function.
13469 (pass_tree_unswitch): Replace declaration with that of...
13470 (make_pass_tree_unswitch): ...new function.
13471 (pass_predcom): Replace declaration with that of...
13472 (make_pass_predcom): ...new function.
13473 (pass_iv_canon): Replace declaration with that of...
13474 (make_pass_iv_canon): ...new function.
13475 (pass_scev_cprop): Replace declaration with that of...
13476 (make_pass_scev_cprop): ...new function.
13477 (pass_empty_loop): Replace declaration with that of...
13478 (make_pass_empty_loop): ...new function.
13479 (pass_record_bounds): Replace declaration with that of...
13480 (make_pass_record_bounds): ...new function.
13481 (pass_graphite): Replace declaration with that of...
13482 (make_pass_graphite): ...new function.
13483 (pass_graphite_transforms): Replace declaration with that of...
13484 (make_pass_graphite_transforms): ...new function.
13485 (pass_if_conversion): Replace declaration with that of...
13486 (make_pass_if_conversion): ...new function.
13487 (pass_loop_distribution): Replace declaration with that of...
13488 (make_pass_loop_distribution): ...new function.
13489 (pass_vectorize): Replace declaration with that of...
13490 (make_pass_vectorize): ...new function.
13491 (pass_slp_vectorize): Replace declaration with that of...
13492 (make_pass_slp_vectorize): ...new function.
13493 (pass_complete_unroll): Replace declaration with that of...
13494 (make_pass_complete_unroll): ...new function.
13495 (pass_complete_unrolli): Replace declaration with that of...
13496 (make_pass_complete_unrolli): ...new function.
13497 (pass_parallelize_loops): Replace declaration with that of...
13498 (make_pass_parallelize_loops): ...new function.
13499 (pass_loop_prefetch): Replace declaration with that of...
13500 (make_pass_loop_prefetch): ...new function.
13501 (pass_iv_optimize): Replace declaration with that of...
13502 (make_pass_iv_optimize): ...new function.
13503 (pass_tree_loop_done): Replace declaration with that of...
13504 (make_pass_tree_loop_done): ...new function.
13505 (pass_ch): Replace declaration with that of...
13506 (make_pass_ch): ...new function.
13507 (pass_ccp): Replace declaration with that of...
13508 (make_pass_ccp): ...new function.
13509 (pass_phi_only_cprop): Replace declaration with that of...
13510 (make_pass_phi_only_cprop): ...new function.
13511 (pass_build_ssa): Replace declaration with that of...
13512 (make_pass_build_ssa): ...new function.
13513 (pass_build_alias): Replace declaration with that of...
13514 (make_pass_build_alias): ...new function.
13515 (pass_build_ealias): Replace declaration with that of...
13516 (make_pass_build_ealias): ...new function.
13517 (pass_dominator): Replace declaration with that of...
13518 (make_pass_dominator): ...new function.
13519 (pass_dce): Replace declaration with that of...
13520 (make_pass_dce): ...new function.
13521 (pass_dce_loop): Replace declaration with that of...
13522 (make_pass_dce_loop): ...new function.
13523 (pass_cd_dce): Replace declaration with that of...
13524 (make_pass_cd_dce): ...new function.
13525 (pass_call_cdce): Replace declaration with that of...
13526 (make_pass_call_cdce): ...new function.
13527 (pass_merge_phi): Replace declaration with that of...
13528 (make_pass_merge_phi): ...new function.
13529 (pass_split_crit_edges): Replace declaration with that of...
13530 (make_pass_split_crit_edges): ...new function.
13531 (pass_pre): Replace declaration with that of...
13532 (make_pass_pre): ...new function.
13533 (pass_profile): Replace declaration with that of...
13534 (make_pass_profile): ...new function.
13535 (pass_strip_predict_hints): Replace declaration with that of...
13536 (make_pass_strip_predict_hints): ...new function.
13537 (pass_lower_complex_O0): Replace declaration with that of...
13538 (make_pass_lower_complex_O0): ...new function.
13539 (pass_lower_complex): Replace declaration with that of...
13540 (make_pass_lower_complex): ...new function.
13541 (pass_lower_vector): Replace declaration with that of...
13542 (make_pass_lower_vector): ...new function.
13543 (pass_lower_vector_ssa): Replace declaration with that of...
13544 (make_pass_lower_vector_ssa): ...new function.
13545 (pass_lower_omp): Replace declaration with that of...
13546 (make_pass_lower_omp): ...new function.
13547 (pass_diagnose_omp_blocks): Replace declaration with that of...
13548 (make_pass_diagnose_omp_blocks): ...new function.
13549 (pass_expand_omp): Replace declaration with that of...
13550 (make_pass_expand_omp): ...new function.
13551 (pass_expand_omp_ssa): Replace declaration with that of...
13552 (make_pass_expand_omp_ssa): ...new function.
13553 (pass_object_sizes): Replace declaration with that of...
13554 (make_pass_object_sizes): ...new function.
13555 (pass_strlen): Replace declaration with that of...
13556 (make_pass_strlen): ...new function.
13557 (pass_fold_builtins): Replace declaration with that of...
13558 (make_pass_fold_builtins): ...new function.
13559 (pass_stdarg): Replace declaration with that of...
13560 (make_pass_stdarg): ...new function.
13561 (pass_early_warn_uninitialized): Replace declaration with that of...
13562 (make_pass_early_warn_uninitialized): ...new function.
13563 (pass_late_warn_uninitialized): Replace declaration with that of...
13564 (make_pass_late_warn_uninitialized): ...new function.
13565 (pass_cse_reciprocals): Replace declaration with that of...
13566 (make_pass_cse_reciprocals): ...new function.
13567 (pass_cse_sincos): Replace declaration with that of...
13568 (make_pass_cse_sincos): ...new function.
13569 (pass_optimize_bswap): Replace declaration with that of...
13570 (make_pass_optimize_bswap): ...new function.
13571 (pass_optimize_widening_mul): Replace declaration with that of...
13572 (make_pass_optimize_widening_mul): ...new function.
13573 (pass_warn_function_return): Replace declaration with that of...
13574 (make_pass_warn_function_return): ...new function.
13575 (pass_warn_function_noreturn): Replace declaration with that of...
13576 (make_pass_warn_function_noreturn): ...new function.
13577 (pass_cselim): Replace declaration with that of...
13578 (make_pass_cselim): ...new function.
13579 (pass_phiopt): Replace declaration with that of...
13580 (make_pass_phiopt): ...new function.
13581 (pass_forwprop): Replace declaration with that of...
13582 (make_pass_forwprop): ...new function.
13583 (pass_phiprop): Replace declaration with that of...
13584 (make_pass_phiprop): ...new function.
13585 (pass_tree_ifcombine): Replace declaration with that of...
13586 (make_pass_tree_ifcombine): ...new function.
13587 (pass_dse): Replace declaration with that of...
13588 (make_pass_dse): ...new function.
13589 (pass_nrv): Replace declaration with that of...
13590 (make_pass_nrv): ...new function.
13591 (pass_rename_ssa_copies): Replace declaration with that of...
13592 (make_pass_rename_ssa_copies): ...new function.
13593 (pass_sink_code): Replace declaration with that of...
13594 (make_pass_sink_code): ...new function.
13595 (pass_fre): Replace declaration with that of...
13596 (make_pass_fre): ...new function.
13597 (pass_check_data_deps): Replace declaration with that of...
13598 (make_pass_check_data_deps): ...new function.
13599 (pass_copy_prop): Replace declaration with that of...
13600 (make_pass_copy_prop): ...new function.
13601 (pass_vrp): Replace declaration with that of...
13602 (make_pass_vrp): ...new function.
13603 (pass_uncprop): Replace declaration with that of...
13604 (make_pass_uncprop): ...new function.
13605 (pass_return_slot): Replace declaration with that of...
13606 (make_pass_return_slot): ...new function.
13607 (pass_reassoc): Replace declaration with that of...
13608 (make_pass_reassoc): ...new function.
13609 (pass_rebuild_cgraph_edges): Replace declaration with that of...
13610 (make_pass_rebuild_cgraph_edges): ...new function.
13611 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
13612 (make_pass_remove_cgraph_callee_edges): ...new function.
13613 (pass_build_cgraph_edges): Replace declaration with that of...
13614 (make_pass_build_cgraph_edges): ...new function.
13615 (pass_local_pure_const): Replace declaration with that of...
13616 (make_pass_local_pure_const): ...new function.
13617 (pass_tracer): Replace declaration with that of...
13618 (make_pass_tracer): ...new function.
13619 (pass_warn_unused_result): Replace declaration with that of...
13620 (make_pass_warn_unused_result): ...new function.
13621 (pass_diagnose_tm_blocks): Replace declaration with that of...
13622 (make_pass_diagnose_tm_blocks): ...new function.
13623 (pass_lower_tm): Replace declaration with that of...
13624 (make_pass_lower_tm): ...new function.
13625 (pass_tm_init): Replace declaration with that of...
13626 (make_pass_tm_init): ...new function.
13627 (pass_tm_mark): Replace declaration with that of...
13628 (make_pass_tm_mark): ...new function.
13629 (pass_tm_memopt): Replace declaration with that of...
13630 (make_pass_tm_memopt): ...new function.
13631 (pass_tm_edges): Replace declaration with that of...
13632 (make_pass_tm_edges): ...new function.
13633 (pass_split_functions): Replace declaration with that of...
13634 (make_pass_split_functions): ...new function.
13635 (pass_feedback_split_functions): Replace declaration with that of...
13636 (make_pass_feedback_split_functions): ...new function.
13637 (pass_strength_reduction): Replace declaration with that of...
13638 (make_pass_strength_reduction): ...new function.
13639 (pass_ipa_lower_emutls): Replace declaration with that of...
13640 (make_pass_ipa_lower_emutls): ...new function.
13641 (pass_ipa_function_and_variable_visibility): Replace declaration with
13643 (make_pass_ipa_function_and_variable_visibility): ...new function.
13644 (pass_ipa_tree_profile): Replace declaration with that of...
13645 (make_pass_ipa_tree_profile): ...new function.
13646 (pass_early_local_passes): Replace declaration with that of...
13647 (make_pass_early_local_passes): ...new function.
13648 (pass_ipa_whole_program_visibility): Replace declaration with that
13650 (make_pass_ipa_whole_program_visibility): ...new function.
13651 (pass_ipa_lto_gimple_out): Replace declaration with that of...
13652 (make_pass_ipa_lto_gimple_out): ...new function.
13653 (pass_ipa_increase_alignment): Replace declaration with that of...
13654 (make_pass_ipa_increase_alignment): ...new function.
13655 (pass_ipa_inline): Replace declaration with that of...
13656 (make_pass_ipa_inline): ...new function.
13657 (pass_ipa_free_lang_data): Replace declaration with that of...
13658 (make_pass_ipa_free_lang_data): ...new function.
13659 (pass_ipa_free_inline_summary): Replace declaration with that of...
13660 (make_pass_ipa_free_inline_summary): ...new function.
13661 (pass_ipa_cp): Replace declaration with that of...
13662 (make_pass_ipa_cp): ...new function.
13663 (pass_ipa_reference): Replace declaration with that of...
13664 (make_pass_ipa_reference): ...new function.
13665 (pass_ipa_pure_const): Replace declaration with that of...
13666 (make_pass_ipa_pure_const): ...new function.
13667 (pass_ipa_pta): Replace declaration with that of...
13668 (make_pass_ipa_pta): ...new function.
13669 (pass_ipa_lto_finish_out): Replace declaration with that of...
13670 (make_pass_ipa_lto_finish_out): ...new function.
13671 (pass_ipa_tm): Replace declaration with that of...
13672 (make_pass_ipa_tm): ...new function.
13673 (pass_ipa_profile): Replace declaration with that of...
13674 (make_pass_ipa_profile): ...new function.
13675 (pass_ipa_cdtor_merge): Replace declaration with that of...
13676 (make_pass_ipa_cdtor_merge): ...new function.
13677 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
13679 (make_pass_cleanup_cfg_post_optimizing): ...new function.
13680 (pass_init_datastructures): Replace declaration with that of...
13681 (make_pass_init_datastructures): ...new function.
13682 (pass_fixup_cfg): Replace declaration with that of...
13683 (make_pass_fixup_cfg): ...new function.
13684 (pass_expand): Replace declaration with that of...
13685 (make_pass_expand): ...new function.
13686 (pass_instantiate_virtual_regs): Replace declaration with that of...
13687 (make_pass_instantiate_virtual_regs): ...new function.
13688 (pass_rtl_fwprop): Replace declaration with that of...
13689 (make_pass_rtl_fwprop): ...new function.
13690 (pass_rtl_fwprop_addr): Replace declaration with that of...
13691 (make_pass_rtl_fwprop_addr): ...new function.
13692 (pass_jump): Replace declaration with that of...
13693 (make_pass_jump): ...new function.
13694 (pass_jump2): Replace declaration with that of...
13695 (make_pass_jump2): ...new function.
13696 (pass_lower_subreg): Replace declaration with that of...
13697 (make_pass_lower_subreg): ...new function.
13698 (pass_cse): Replace declaration with that of...
13699 (make_pass_cse): ...new function.
13700 (pass_fast_rtl_dce): Replace declaration with that of...
13701 (make_pass_fast_rtl_dce): ...new function.
13702 (pass_ud_rtl_dce): Replace declaration with that of...
13703 (make_pass_ud_rtl_dce): ...new function.
13704 (pass_rtl_dce): Replace declaration with that of...
13705 (make_pass_rtl_dce): ...new function.
13706 (pass_rtl_dse1): Replace declaration with that of...
13707 (make_pass_rtl_dse1): ...new function.
13708 (pass_rtl_dse2): Replace declaration with that of...
13709 (make_pass_rtl_dse2): ...new function.
13710 (pass_rtl_dse3): Replace declaration with that of...
13711 (make_pass_rtl_dse3): ...new function.
13712 (pass_rtl_cprop): Replace declaration with that of...
13713 (make_pass_rtl_cprop): ...new function.
13714 (pass_rtl_pre): Replace declaration with that of...
13715 (make_pass_rtl_pre): ...new function.
13716 (pass_rtl_hoist): Replace declaration with that of...
13717 (make_pass_rtl_hoist): ...new function.
13718 (pass_rtl_store_motion): Replace declaration with that of...
13719 (make_pass_rtl_store_motion): ...new function.
13720 (pass_cse_after_global_opts): Replace declaration with that of...
13721 (make_pass_cse_after_global_opts): ...new function.
13722 (pass_rtl_ifcvt): Replace declaration with that of...
13723 (make_pass_rtl_ifcvt): ...new function.
13724 (pass_into_cfg_layout_mode): Replace declaration with that of...
13725 (make_pass_into_cfg_layout_mode): ...new function.
13726 (pass_outof_cfg_layout_mode): Replace declaration with that of...
13727 (make_pass_outof_cfg_layout_mode): ...new function.
13728 (pass_loop2): Replace declaration with that of...
13729 (make_pass_loop2): ...new function.
13730 (pass_rtl_loop_init): Replace declaration with that of...
13731 (make_pass_rtl_loop_init): ...new function.
13732 (pass_rtl_move_loop_invariants): Replace declaration with that of...
13733 (make_pass_rtl_move_loop_invariants): ...new function.
13734 (pass_rtl_unswitch): Replace declaration with that of...
13735 (make_pass_rtl_unswitch): ...new function.
13736 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
13737 (make_pass_rtl_unroll_and_peel_loops): ...new function.
13738 (pass_rtl_doloop): Replace declaration with that of...
13739 (make_pass_rtl_doloop): ...new function.
13740 (pass_rtl_loop_done): Replace declaration with that of...
13741 (make_pass_rtl_loop_done): ...new function.
13742 (pass_web): Replace declaration with that of...
13743 (make_pass_web): ...new function.
13744 (pass_cse2): Replace declaration with that of...
13745 (make_pass_cse2): ...new function.
13746 (pass_df_initialize_opt): Replace declaration with that of...
13747 (make_pass_df_initialize_opt): ...new function.
13748 (pass_df_initialize_no_opt): Replace declaration with that of...
13749 (make_pass_df_initialize_no_opt): ...new function.
13750 (pass_reginfo_init): Replace declaration with that of...
13751 (make_pass_reginfo_init): ...new function.
13752 (pass_inc_dec): Replace declaration with that of...
13753 (make_pass_inc_dec): ...new function.
13754 (pass_stack_ptr_mod): Replace declaration with that of...
13755 (make_pass_stack_ptr_mod): ...new function.
13756 (pass_initialize_regs): Replace declaration with that of...
13757 (make_pass_initialize_regs): ...new function.
13758 (pass_combine): Replace declaration with that of...
13759 (make_pass_combine): ...new function.
13760 (pass_if_after_combine): Replace declaration with that of...
13761 (make_pass_if_after_combine): ...new function.
13762 (pass_ree): Replace declaration with that of...
13763 (make_pass_ree): ...new function.
13764 (pass_partition_blocks): Replace declaration with that of...
13765 (make_pass_partition_blocks): ...new function.
13766 (pass_match_asm_constraints): Replace declaration with that of...
13767 (make_pass_match_asm_constraints): ...new function.
13768 (pass_regmove): Replace declaration with that of...
13769 (make_pass_regmove): ...new function.
13770 (pass_split_all_insns): Replace declaration with that of...
13771 (make_pass_split_all_insns): ...new function.
13772 (pass_fast_rtl_byte_dce): Replace declaration with that of...
13773 (make_pass_fast_rtl_byte_dce): ...new function.
13774 (pass_lower_subreg2): Replace declaration with that of...
13775 (make_pass_lower_subreg2): ...new function.
13776 (pass_mode_switching): Replace declaration with that of...
13777 (make_pass_mode_switching): ...new function.
13778 (pass_sms): Replace declaration with that of...
13779 (make_pass_sms): ...new function.
13780 (pass_sched): Replace declaration with that of...
13781 (make_pass_sched): ...new function.
13782 (pass_ira): Replace declaration with that of...
13783 (make_pass_ira): ...new function.
13784 (pass_reload): Replace declaration with that of...
13785 (make_pass_reload): ...new function.
13786 (pass_clean_state): Replace declaration with that of...
13787 (make_pass_clean_state): ...new function.
13788 (pass_branch_prob): Replace declaration with that of...
13789 (make_pass_branch_prob): ...new function.
13790 (pass_value_profile_transformations): Replace declaration with that
13792 (make_pass_value_profile_transformations): ...new function.
13793 (pass_postreload_cse): Replace declaration with that of...
13794 (make_pass_postreload_cse): ...new function.
13795 (pass_gcse2): Replace declaration with that of...
13796 (make_pass_gcse2): ...new function.
13797 (pass_split_after_reload): Replace declaration with that of...
13798 (make_pass_split_after_reload): ...new function.
13799 (pass_branch_target_load_optimize1): Replace declaration with that
13801 (make_pass_branch_target_load_optimize1): ...new function.
13802 (pass_thread_prologue_and_epilogue): Replace declaration with that
13804 (make_pass_thread_prologue_and_epilogue): ...new function.
13805 (pass_stack_adjustments): Replace declaration with that of...
13806 (make_pass_stack_adjustments): ...new function.
13807 (pass_peephole2): Replace declaration with that of...
13808 (make_pass_peephole2): ...new function.
13809 (pass_if_after_reload): Replace declaration with that of...
13810 (make_pass_if_after_reload): ...new function.
13811 (pass_regrename): Replace declaration with that of...
13812 (make_pass_regrename): ...new function.
13813 (pass_cprop_hardreg): Replace declaration with that of...
13814 (make_pass_cprop_hardreg): ...new function.
13815 (pass_reorder_blocks): Replace declaration with that of...
13816 (make_pass_reorder_blocks): ...new function.
13817 (pass_branch_target_load_optimize2): Replace declaration with that
13819 (make_pass_branch_target_load_optimize2): ...new function.
13820 (pass_leaf_regs): Replace declaration with that of...
13821 (make_pass_leaf_regs): ...new function.
13822 (pass_split_before_sched2): Replace declaration with that of...
13823 (make_pass_split_before_sched2): ...new function.
13824 (pass_compare_elim_after_reload): Replace declaration with that of...
13825 (make_pass_compare_elim_after_reload): ...new function.
13826 (pass_sched2): Replace declaration with that of...
13827 (make_pass_sched2): ...new function.
13828 (pass_stack_regs): Replace declaration with that of...
13829 (make_pass_stack_regs): ...new function.
13830 (pass_stack_regs_run): Replace declaration with that of...
13831 (make_pass_stack_regs_run): ...new function.
13832 (pass_df_finish): Replace declaration with that of...
13833 (make_pass_df_finish): ...new function.
13834 (pass_compute_alignments): Replace declaration with that of...
13835 (make_pass_compute_alignments): ...new function.
13836 (pass_duplicate_computed_gotos): Replace declaration with that of...
13837 (make_pass_duplicate_computed_gotos): ...new function.
13838 (pass_variable_tracking): Replace declaration with that of...
13839 (make_pass_variable_tracking): ...new function.
13840 (pass_free_cfg): Replace declaration with that of...
13841 (make_pass_free_cfg): ...new function.
13842 (pass_machine_reorg): Replace declaration with that of...
13843 (make_pass_machine_reorg): ...new function.
13844 (pass_cleanup_barriers): Replace declaration with that of...
13845 (make_pass_cleanup_barriers): ...new function.
13846 (pass_delay_slots): Replace declaration with that of...
13847 (make_pass_delay_slots): ...new function.
13848 (pass_split_for_shorten_branches): Replace declaration with that of...
13849 (make_pass_split_for_shorten_branches): ...new function.
13850 (pass_split_before_regstack): Replace declaration with that of...
13851 (make_pass_split_before_regstack): ...new function.
13852 (pass_convert_to_eh_region_ranges): Replace declaration with that
13854 (make_pass_convert_to_eh_region_ranges): ...new function.
13855 (pass_shorten_branches): Replace declaration with that of...
13856 (make_pass_shorten_branches): ...new function.
13857 (pass_set_nothrow_function_flags): Replace declaration with that of...
13858 (make_pass_set_nothrow_function_flags): ...new function.
13859 (pass_dwarf2_frame): Replace declaration with that of...
13860 (make_pass_dwarf2_frame): ...new function.
13861 (pass_final): Replace declaration with that of...
13862 (make_pass_final): ...new function.
13863 (pass_rtl_seqabstr): Replace declaration with that of...
13864 (make_pass_rtl_seqabstr): ...new function.
13865 (pass_release_ssa_names): Replace declaration with that of...
13866 (make_pass_release_ssa_names): ...new function.
13867 (pass_early_inline): Replace declaration with that of...
13868 (make_pass_early_inline): ...new function.
13869 (pass_inline_parameters): Replace declaration with that of...
13870 (make_pass_inline_parameters): ...new function.
13871 (pass_update_address_taken): Replace declaration with that of...
13872 (make_pass_update_address_taken): ...new function.
13873 (pass_convert_switch): Replace declaration with that of...
13874 (make_pass_convert_switch): ...new function.
13875 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
13876 to a subclass of simple_ipa_opt_pass along with...
13877 (pass_data_ipa_tree_profile): ...new pass_data instance and...
13878 (make_pass_ipa_tree_profile): ...new function.
13879 * tree-sra.c (pass_sra_early): Convert from a global struct to a
13880 subclass of gimple_opt_pass along with...
13881 (pass_data_sra_early): ...new pass_data instance and...
13882 (make_pass_sra_early): ...new function.
13883 (pass_sra): Convert from a global struct to a subclass of
13884 gimple_opt_pass along with...
13885 (pass_data_sra): ...new pass_data instance and...
13886 (make_pass_sra): ...new function.
13887 (pass_early_ipa_sra): Convert from a global struct to a subclass of
13888 gimple_opt_pass along with...
13889 (pass_data_early_ipa_sra): ...new pass_data instance and...
13890 (make_pass_early_ipa_sra): ...new function.
13891 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
13892 subclass of gimple_opt_pass along with...
13893 (pass_data_ccp): ...new pass_data instance and...
13894 (make_pass_ccp): ...new function.
13895 (pass_fold_builtins): Convert from a global struct to a subclass of
13896 gimple_opt_pass along with...
13897 (pass_data_fold_builtins): ...new pass_data instance and...
13898 (make_pass_fold_builtins): ...new function.
13899 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
13900 subclass of gimple_opt_pass along with...
13901 (pass_data_copy_prop): ...new pass_data instance and...
13902 (make_pass_copy_prop): ...new function.
13903 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
13904 global struct to a subclass of gimple_opt_pass along with...
13905 (pass_data_rename_ssa_copies): ...new pass_data instance and...
13906 (make_pass_rename_ssa_copies): ...new function.
13907 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
13908 subclass of gimple_opt_pass along with...
13909 (pass_data_dce): ...new pass_data instance and...
13910 (make_pass_dce): ...new function.
13911 (pass_dce_loop): Convert from a global struct to a subclass of
13912 gimple_opt_pass along with...
13913 (pass_data_dce_loop): ...new pass_data instance and...
13914 (make_pass_dce_loop): ...new function.
13915 (pass_cd_dce): Convert from a global struct to a subclass of
13916 gimple_opt_pass along with...
13917 (pass_data_cd_dce): ...new pass_data instance and...
13918 (make_pass_cd_dce): ...new function.
13919 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
13920 subclass of gimple_opt_pass along with...
13921 (pass_data_dominator): ...new pass_data instance and...
13922 (make_pass_dominator): ...new function.
13923 (pass_phi_only_cprop): Convert from a global struct to a subclass of
13924 gimple_opt_pass along with...
13925 (pass_data_phi_only_cprop): ...new pass_data instance and...
13926 (make_pass_phi_only_cprop): ...new function.
13927 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
13928 subclass of gimple_opt_pass along with...
13929 (pass_data_dse): ...new pass_data instance and...
13930 (make_pass_dse): ...new function.
13931 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
13932 a subclass of gimple_opt_pass along with...
13933 (pass_data_forwprop): ...new pass_data instance and...
13934 (make_pass_forwprop): ...new function.
13935 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
13936 struct to a subclass of gimple_opt_pass along with...
13937 (pass_data_tree_ifcombine): ...new pass_data instance and...
13938 (make_pass_tree_ifcombine): ...new function.
13939 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
13940 subclass of gimple_opt_pass along with...
13941 (pass_data_ch): ...new pass_data instance and...
13942 (make_pass_ch): ...new function.
13943 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
13944 subclass of gimple_opt_pass along with...
13945 (pass_data_tree_loop): ...new pass_data instance and...
13946 (make_pass_tree_loop): ...new function.
13947 (pass_tree_loop_init): Convert from a global struct to a subclass of
13948 gimple_opt_pass along with...
13949 (pass_data_tree_loop_init): ...new pass_data instance and...
13950 (make_pass_tree_loop_init): ...new function.
13951 (pass_lim): Convert from a global struct to a subclass of
13952 gimple_opt_pass along with...
13953 (pass_data_lim): ...new pass_data instance and...
13954 (make_pass_lim): ...new function.
13955 (pass_tree_unswitch): Convert from a global struct to a subclass of
13956 gimple_opt_pass along with...
13957 (pass_data_tree_unswitch): ...new pass_data instance and...
13958 (make_pass_tree_unswitch): ...new function.
13959 (pass_predcom): Convert from a global struct to a subclass of
13960 gimple_opt_pass along with...
13961 (pass_data_predcom): ...new pass_data instance and...
13962 (make_pass_predcom): ...new function.
13963 (pass_vectorize): Convert from a global struct to a subclass of
13964 gimple_opt_pass along with...
13965 (pass_data_vectorize): ...new pass_data instance and...
13966 (make_pass_vectorize): ...new function.
13967 (pass_graphite): Convert from a global struct to a subclass of
13968 gimple_opt_pass along with...
13969 (pass_data_graphite): ...new pass_data instance and...
13970 (make_pass_graphite): ...new function.
13971 (pass_graphite_transforms): Convert from a global struct to a subclass
13972 of gimple_opt_pass along with...
13973 (pass_data_graphite_transforms): ...new pass_data instance and...
13974 (make_pass_graphite_transforms): ...new function.
13975 (pass_check_data_deps): Convert from a global struct to a subclass of
13976 gimple_opt_pass along with...
13977 (pass_data_check_data_deps): ...new pass_data instance and...
13978 (make_pass_check_data_deps): ...new function.
13979 (pass_iv_canon): Convert from a global struct to a subclass of
13980 gimple_opt_pass along with...
13981 (pass_data_iv_canon): ...new pass_data instance and...
13982 (make_pass_iv_canon): ...new function.
13983 (pass_scev_cprop): Convert from a global struct to a subclass of
13984 gimple_opt_pass along with...
13985 (pass_data_scev_cprop): ...new pass_data instance and...
13986 (make_pass_scev_cprop): ...new function.
13987 (pass_record_bounds): Convert from a global struct to a subclass of
13988 gimple_opt_pass along with...
13989 (pass_data_record_bounds): ...new pass_data instance and...
13990 (make_pass_record_bounds): ...new function.
13991 (pass_complete_unroll): Convert from a global struct to a subclass of
13992 gimple_opt_pass along with...
13993 (pass_data_complete_unroll): ...new pass_data instance and...
13994 (make_pass_complete_unroll): ...new function.
13995 (pass_complete_unrolli): Convert from a global struct to a subclass of
13996 gimple_opt_pass along with...
13997 (pass_data_complete_unrolli): ...new pass_data instance and...
13998 (make_pass_complete_unrolli): ...new function.
13999 (pass_parallelize_loops): Convert from a global struct to a subclass
14000 of gimple_opt_pass along with...
14001 (pass_data_parallelize_loops): ...new pass_data instance and...
14002 (make_pass_parallelize_loops): ...new function.
14003 (pass_loop_prefetch): Convert from a global struct to a subclass of
14004 gimple_opt_pass along with...
14005 (pass_data_loop_prefetch): ...new pass_data instance and...
14006 (make_pass_loop_prefetch): ...new function.
14007 (pass_iv_optimize): Convert from a global struct to a subclass of
14008 gimple_opt_pass along with...
14009 (pass_data_iv_optimize): ...new pass_data instance and...
14010 (make_pass_iv_optimize): ...new function.
14011 (pass_tree_loop_done): Convert from a global struct to a subclass of
14012 gimple_opt_pass along with...
14013 (pass_data_tree_loop_done): ...new pass_data instance and...
14014 (make_pass_tree_loop_done): ...new function.
14015 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
14016 struct to a subclass of gimple_opt_pass along with...
14017 (pass_data_cse_reciprocals): ...new pass_data instance and...
14018 (make_pass_cse_reciprocals): ...new function.
14019 (pass_cse_sincos): Convert from a global struct to a subclass of
14020 gimple_opt_pass along with...
14021 (pass_data_cse_sincos): ...new pass_data instance and...
14022 (make_pass_cse_sincos): ...new function.
14023 (pass_optimize_bswap): Convert from a global struct to a subclass of
14024 gimple_opt_pass along with...
14025 (pass_data_optimize_bswap): ...new pass_data instance and...
14026 (make_pass_optimize_bswap): ...new function.
14027 (pass_optimize_widening_mul): Convert from a global struct to a
14028 subclass of gimple_opt_pass along with...
14029 (pass_data_optimize_widening_mul): ...new pass_data instance and...
14030 (make_pass_optimize_widening_mul): ...new function.
14031 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
14032 subclass of gimple_opt_pass along with...
14033 (pass_data_phiopt): ...new pass_data instance and...
14034 (make_pass_phiopt): ...new function.
14035 (pass_cselim): Convert from a global struct to a subclass of
14036 gimple_opt_pass along with...
14037 (pass_data_cselim): ...new pass_data instance and...
14038 (make_pass_cselim): ...new function.
14039 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
14040 subclass of gimple_opt_pass along with...
14041 (pass_data_phiprop): ...new pass_data instance and...
14042 (make_pass_phiprop): ...new function.
14043 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
14044 subclass of gimple_opt_pass along with...
14045 (pass_data_pre): ...new pass_data instance and...
14046 (make_pass_pre): ...new function.
14047 (pass_fre): Convert from a global struct to a subclass of
14048 gimple_opt_pass along with...
14049 (pass_data_fre): ...new pass_data instance and...
14050 (make_pass_fre): ...new function.
14051 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
14052 subclass of gimple_opt_pass along with...
14053 (pass_data_reassoc): ...new pass_data instance and...
14054 (make_pass_reassoc): ...new function.
14055 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
14056 subclass of gimple_opt_pass along with...
14057 (pass_data_sink_code): ...new pass_data instance and...
14058 (make_pass_sink_code): ...new function.
14059 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
14060 subclass of gimple_opt_pass along with...
14061 (pass_data_strlen): ...new pass_data instance and...
14062 (make_pass_strlen): ...new function.
14063 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
14064 struct to a subclass of gimple_opt_pass along with...
14065 (pass_data_build_alias): ...new pass_data instance and...
14066 (make_pass_build_alias): ...new function.
14067 (pass_build_ealias): Convert from a global struct to a subclass of
14068 gimple_opt_pass along with...
14069 (pass_data_build_ealias): ...new pass_data instance and...
14070 (make_pass_build_ealias): ...new function.
14071 (pass_ipa_pta): Convert from a global struct to a subclass of
14072 simple_ipa_opt_pass along with...
14073 (pass_data_ipa_pta): ...new pass_data instance and...
14074 (make_pass_ipa_pta): ...new function.
14075 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
14076 subclass of gimple_opt_pass along with...
14077 (pass_data_uncprop): ...new pass_data instance and...
14078 (make_pass_uncprop): ...new function.
14079 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
14080 global struct to a subclass of gimple_opt_pass along with...
14081 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
14082 (make_pass_late_warn_uninitialized): ...new function.
14083 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
14084 to a subclass of gimple_opt_pass along with...
14085 (pass_data_init_datastructures): ...new pass_data instance and...
14086 (make_pass_init_datastructures): ...new function.
14087 (pass_early_warn_uninitialized): Convert from a global struct to a
14088 subclass of gimple_opt_pass along with...
14089 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
14090 (make_pass_early_warn_uninitialized): ...new function.
14091 (pass_update_address_taken): Convert from a global struct to a
14092 subclass of gimple_opt_pass along with...
14093 (pass_data_update_address_taken): ...new pass_data instance and...
14094 (make_pass_update_address_taken): ...new function.
14095 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
14096 struct to a subclass of gimple_opt_pass along with...
14097 (pass_data_release_ssa_names): ...new pass_data instance and...
14098 (make_pass_release_ssa_names): ...new function.
14099 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
14100 subclass of gimple_opt_pass along with...
14101 (pass_data_stdarg): ...new pass_data instance and...
14102 (make_pass_stdarg): ...new function.
14103 * tree-switch-conversion.c (pass_convert_switch): Convert from a
14104 global struct to a subclass of gimple_opt_pass along with...
14105 (pass_data_convert_switch): ...new pass_data instance and...
14106 (make_pass_convert_switch): ...new function.
14107 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
14108 to a subclass of gimple_opt_pass along with...
14109 (pass_data_tail_recursion): ...new pass_data instance and...
14110 (make_pass_tail_recursion): ...new function.
14111 (pass_tail_calls): Convert from a global struct to a subclass of
14112 gimple_opt_pass along with...
14113 (pass_data_tail_calls): ...new pass_data instance and...
14114 (make_pass_tail_calls): ...new function.
14115 * tree-vect-generic.c (pass_lower_vector): Convert from a global
14116 struct to a subclass of gimple_opt_pass along with...
14117 (pass_data_lower_vector): ...new pass_data instance and...
14118 (make_pass_lower_vector): ...new function.
14119 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
14120 gimple_opt_pass along with...
14121 (pass_data_lower_vector_ssa): ...new pass_data instance and...
14122 (make_pass_lower_vector_ssa): ...new function.
14123 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
14124 to a subclass of gimple_opt_pass along with...
14125 (pass_data_slp_vectorize): ...new pass_data instance and...
14126 (make_pass_slp_vectorize): ...new function.
14127 (pass_ipa_increase_alignment): Convert from a global struct to a
14128 subclass of simple_ipa_opt_pass along with...
14129 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
14130 (make_pass_ipa_increase_alignment): ...new function.
14131 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
14132 gimple_opt_pass along with...
14133 (pass_data_vrp): ...new pass_data instance and...
14134 (make_pass_vrp): ...new function.
14135 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
14136 subclass of simple_ipa_opt_pass along with...
14137 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
14138 (make_pass_ipa_free_lang_data): ...new function.
14139 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
14140 gimple_opt_pass along with...
14141 (pass_data_tsan): ...new pass_data instance and...
14142 (make_pass_tsan): ...new function.
14143 (pass_tsan_O0): Convert from a global struct to a subclass of
14144 gimple_opt_pass along with...
14145 (pass_data_tsan_O0): ...new pass_data instance and...
14146 (make_pass_tsan_O0): ...new function.
14147 * var-tracking.c (pass_variable_tracking): Convert from a global
14148 struct to a subclass of rtl_opt_pass along with...
14149 (pass_data_variable_tracking): ...new pass_data instance and...
14150 (make_pass_variable_tracking): ...new function.
14151 * web.c (pass_web): Convert from a global struct to a subclass of
14152 rtl_opt_pass along with...
14153 (pass_data_web): ...new pass_data instance and...
14154 (make_pass_web): ...new function.
14155 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
14156 declaration with that of...
14157 (make_pass_mode_switch_use): ...new function.
14158 (pass_resolve_sw_modes): Replace declaration with that of...
14159 (make_pass_resolve_sw_modes): ...new function.
14160 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
14161 from a global struct to a subclass of rtl_opt_pass along with...
14162 (pass_data_mode_switch_use): ...new pass_data instance and...
14163 (make_pass_mode_switch_use): ...new function.
14164 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
14165 from a global struct to a subclass of rtl_opt_pass along with...
14166 (pass_data_resolve_sw_modes): ...new pass_data instance and...
14167 (make_pass_resolve_sw_modes): ...new function.
14168 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
14169 struct to a subclass of rtl_opt_pass along with...
14170 (pass_data_insert_vzeroupper): ...new pass_data instance and...
14171 (make_pass_insert_vzeroupper): ...new function.
14172 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
14173 global struct to a subclass of rtl_opt_pass along with...
14174 (pass_data_work_around_errata): ...new pass_data instance and...
14175 (make_pass_work_around_errata): ...new function.
14176 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
14177 struct to a subclass of rtl_opt_pass along with...
14178 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
14179 (make_pass_mips_machine_reorg2): ...new function.
14181 2013-08-05 David Malcolm <dmalcolm@redhat.com>
14183 * passes.c (pass_manager::operator new): New.
14185 2013-08-05 David Malcolm <dmalcolm@redhat.com>
14187 Handwritten part of conversion of passes to C++ classes.
14189 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
14190 (toplev.o): Add dep on PASS_MANAGER_H.
14191 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
14192 of early local pases to reflect this moving from a global to a
14193 member of gcc::pass_manager.
14194 (cgraph_add_new_function): Likewise.
14195 * lto-cgraph.c (lto_output_node): Update for conversion of
14196 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
14197 * passes.c (opt_pass::clone): New.
14198 (opt_pass::gate): New.
14199 (opt_pass::execute): New.
14200 (opt_pass::opt_pass): New.
14201 (pass_manager::execute_early_local_passes): New.
14202 (pass_manager::execute_pass_mode_switching): new.
14203 (finish_optimization_passes): Convert to...
14204 (pass_manager::finish_optimization_passes): ...this.
14205 (finish_optimization_passes): Update for conversion of passes to
14207 (register_dump_files_1): Use has_gate since we cannot portably
14208 check a vtable entry against NULL.
14209 (dump_one_pass): Likewise.
14210 (ipa_write_summaries_2): Likewise.
14211 (ipa_write_optimization_summaries_1): Likewise.
14212 (ipa_read_summaries_1): Likewise.
14213 (ipa_read_optimization_summaries_1): Likewise.
14214 (execute_ipa_stmt_fixups): Likewise.
14215 (pass_manager::pass_manager): Rewrite pass-creation, invoking
14216 pass-creation functions rather than wiring up globals, and
14217 storing the results in fields of pass_manager generated using
14218 pass-instances.def.
14219 (pass_manager::dump_profile_report): Update for conversion of
14220 passes to C++ classes.
14221 (pass_manager::execute_ipa_summary_passes): Likewise.
14222 (execute_one_ipa_transform_pass): Likewise.
14223 (execute_one_pass): Use has_gate and has_execute since we cannot
14224 portably check a vtable entry against NULL.
14225 * pass_manager.h (pass_manager::finish_optimization_passes): New.
14226 (pass_manager): Use pass-instances.def to add fields for the
14227 various pass instances.
14228 * toplev.c (finalize): Update for move of
14229 finish_optimization_passes to a method of gcc::pass_manager.
14230 * toplev.h (finish_optimization_passes): Move to method of class
14232 * tree-pass.h (struct pass_data): New.
14233 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
14234 (opt_pass::gate): Convert to virtual function.
14235 (opt_pass::~opt_pass): New.
14236 (opt_pass::clone): New.
14237 (opt_pass::execute): Convert to virtual function.
14238 (opt_pass::opt_pass): New.
14239 (opt_pass::ctxt_): new.
14240 (gimple_opt_pass): Convert to subclass of opt_pass.
14241 (gimple_opt_pass::gimple_opt_pass): New.
14242 (rtl_opt_pass): Convert to subclass of opt_pass.
14243 (rtl_opt_pass::rtl_opt_pass): New.
14244 (ipa_opt_pass_d): Convert to subclass of opt_pass.
14245 (ipa_opt_pass_d::ipa_opt_pass_d): New.
14246 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
14247 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
14248 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
14249 invocation of pass_mode_switching to reflect this moving from a
14250 global to a member of gcc::pass_manager.
14251 (ix86_option_override): Rework how pass_insert_vzeroupper is
14252 added to the pass_manager to reflect autogenerated changes.
14253 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
14255 2013-08-05 Richard Earnshaw <rearnsha@arm.com>
14257 PR rtl-optimization/57708
14258 * recog.c (peep2_find_free_register): Validate all regs in a
14261 2013-08-05 Jan Hubicka <jh@suse.cz>
14264 * cgraph.c (verify_cgraph_node): Accept local flags from other
14266 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
14267 (function_and_variable_visibility): Likewise.
14268 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
14270 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
14272 * graph.c (init_graph_slim_pretty_print): Remove.
14273 (print_graph_cfg): Do not call it. Use local pretty printer.
14274 (start_graph_dump): Likewise.
14276 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
14278 * gimple-pretty-print.c (buffer): Remove.
14279 (initialized): Likewise.
14280 (maybe_init_pretty_print): Likewise.
14281 (print_gimple_stmt): Do not call it. Use non-static local
14282 pretty_printer variable.
14283 (print_gimple_expr): Likewise.
14284 (print_gimple_seq): Likewise.
14285 (gimple_dump_bb): Likewise.
14287 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
14289 * asan.c (asan_pp): Remove.
14290 (asan_pp_initialized): Likewise.
14291 (asan_pp_initialize): Likewise.
14292 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
14293 (asan_emit_stack_protection): Tidy. Use local pretty printer.
14294 (asan_add_global): Likewise.
14296 2013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
14298 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
14299 * diagnostic.h (diagnostic_flush_buffer): Adjust.
14300 * pretty-print.c (pp_formatted_text_data): Likewise.
14301 (pp_indent): Rename from pp_base_indent.
14302 (pp_format): Rename from pp_base_format.
14303 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
14304 (pp_format_verbatim): Rename from pp_base_format_verbatim.
14305 (pp_flush): Rename from pp_base_flush.
14306 (pp_set_line_maximum_length): Rename from
14307 pp_base_set_line_maximum_length.
14308 (pp_clear_output_area): Rename from pp_base_clear_output_area.
14309 (pp_set_prefix): Rename from pp_base_set_prefix.
14310 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
14311 (pp_emit_prefix): Rename from pp_base_emit_prefix.
14312 (pp_append_text): Rename from pp_base_append_text.
14313 (pp_formatted_text): Rename from pp_base_formatted_text.
14314 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
14315 (pp_remaining_character_count_for_line): Rename from
14316 pp_base_remaining_character_count_for_line.
14317 (pp_newline): Rename from pp_base_newline.
14318 (pp_character): Rename from pp_base_character.
14319 (pp_string): Rename from pp_base_string.
14320 (pp_maybe_space): Rename from pp_base_maybe_space.
14321 * asan.c (asan_pp_string): Adjust.
14322 (asan_emit_stack_protection): Likewise.
14323 (asan_add_global): Likewise.
14324 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
14325 * tree-mudflap.c (mf_varname_tree): Likewise.
14326 * tree-pretty-print.c (pp_tree_identifier): Rename from
14327 pp_base_tree_identifier.
14328 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
14329 Declare as function.
14331 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
14333 * pretty-print.h (pp_bar_bar): New.
14334 (pp_ampersand_ampersand): Likewise.
14335 (pp_less_equal): Likewise.
14336 (pp_greater_equal): Likewise.
14337 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
14338 printer functions instead of pp_string or operators and punctuators.
14339 (dump_gimple_call): Likewise.
14340 (dump_gimple_omp_for): Likewise.
14341 (dump_gimple_transaction): Likewise.
14342 (dump_gimple_phi): Likewise.
14343 (pp_gimple_stmt_1): Likewise.
14344 * sched-vis.c (print_insn): Likewise.
14345 * tree-mudflap.c (mf_varname_tree): Likewise.
14346 * tree-pretty-print.c (dump_block_node): Likewise.
14347 (dump_generic_node): Likewise.
14349 2013-08-02 Jan Hubicka <jh@suse.cz>
14351 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
14353 * lto-streamer-out.c (tree_is_indexable): Results decls and
14354 parm decls are not indexable.
14355 (DFS_write_tree_body): Do not follow args and results.
14356 (hash_tree): Likewise.
14357 (output_functions): Rearrange so struct function is needed
14358 only when real body is output; be able to also ouptut abstract
14359 functions; output DECL_ARGUMENTS and DECL_RESULT.
14360 (lto_output): When not in WPA, ale store abstract functions.
14361 (write_symbol): Do not care about RESULT_DECL.
14362 (output_symbol_p): Handle correctly sbtract decls.
14363 * lto-streamer-in.c (input_function): Rearrange so struct
14364 function can be NULL at entry; allow streaming of
14365 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
14366 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
14367 sanity check during LTO.
14368 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
14369 RESULT_DECl and DECL_ARGUMENTS.
14370 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
14373 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
14375 * pretty-print.h (pp_underscore): New.
14377 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
14378 printer functions instead of pp_character.
14379 (dump_binary_rhs): Likewise.
14380 (dump_ternary_rhs): Likewise.
14381 (dump_gimple_call_args): Likewise.
14382 (pp_points_to_solution): Likewise.
14383 (dump_gimple_call): Likewise.
14384 (dump_gimple_switch): Likewise.
14385 (dump_gimple_cond): Likewise.
14386 (dump_gimple_bind): Likewise.
14387 (dump_gimple_try): Likewise.
14388 (dump_gimple_omp_for): Likewise.
14389 (dump_gimple_omp_continue): Likewise.
14390 (dump_gimple_omp_single): Likewise.
14391 (dump_gimple_omp_sections): Likewise.
14392 (dump_gimple_omp_block): Likewise.
14393 (dump_gimple_omp_critical): Likewise.
14394 (dump_gimple_transaction): Likewise.
14395 (dump_gimple_asm): Likewise.
14396 (dump_gimple_phi): Likewise.
14397 (dump_gimple_omp_parallel): Likewise.
14398 (dump_gimple_omp_task): Likewise.
14399 (dump_gimple_omp_atomic_load): Likewise.
14400 (dump_gimple_omp_atomic_store): Likewise.
14401 (dump_gimple_mem_ops): Likewise.
14402 (pp_gimple_stmt_1): Likewise.
14403 (pp_cfg_jump): Likewise.
14404 (dump_implicit_edges): Likewise.
14405 (gimple_dump_bb_for_graph): Likewise.
14406 * graph.c (draw_cfg_node): Likewise.
14407 * langhooks.c (lhd_print_error_function): Likewise.
14408 * sched-vis.c (print_exp): Likewise.
14409 (print_value): Likewise.
14410 (print_pattern): Likewise.
14411 (print_insn): Likewise.
14412 (rtl_dump_bb_for_graph): Likewise.
14413 * tree-pretty-print.c (dump_function_declaration): Likewise.
14414 (dump_array_domain): Likewise.
14415 (dump_omp_clause): Likewise.
14416 (dump_location): Likewise.
14417 (dump_generic_node): Likewise.
14418 (print_struct_decl): Likewise.
14419 * diagnostic.c (diagnostic_show_locus): Use pp_space.
14421 2013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
14423 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
14424 candidate table when replacing a candidate statement.
14425 (replace_rhs_if_not_dup): Likewise.
14426 (replace_one_candidate): Likewise.
14428 2013-08-02 Jan Hubicka <jh@suse.cz>
14429 Martin Liska <marxin.liska@gmail.com>
14431 * cgraphunit.c (add_new_function): Fix logic when adding from
14433 (assemble_thunk): Rename to ...
14434 (expand_thunk); .. this one; export; get it working with
14435 general functions; make produced gimple valid.
14436 * cgraph.h (expand_thunk): Declare.
14438 2013-08-02 Jan Hubicka <jh@suse.cz>
14440 * ipa-cp.c (gather_context_independent_values): Use
14441 ipa_get_param_move_cost.
14442 (get_replacement_map): Remove PARAM; move parameter folding
14444 (create_specialized_node): Update.
14445 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
14446 assert that we have gimple body; update move_cost.
14447 (count_formal_params): Assert that we have gimple body.
14448 (ipa_dump_param): New function.
14449 (ipa_alloc_node_params): Break out from ...
14450 (ipa_initialize_node_params): ... here.
14451 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
14452 (ipa_write_node_info): Stream move costs.
14453 (ipa_read_node_info): Read move costs.
14454 (ipa_update_after_lto_read): Do not recompute node params.
14455 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
14456 (ipa_get_param): Check we are not in WPA.
14457 (ipa_get_param_move_cost): New.
14458 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
14459 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
14460 parm numbers to be present.
14462 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
14464 PR rtl-optimization/58048
14465 * lra-constraints.c (process_alt_operands): Don't check asm
14466 operand on register.
14468 2013-08-02 Eric Botcazou <ebotcazou@adacore.com>
14470 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
14471 the implied StoreLoad barrier for atomic operations if before.
14473 2013-08-02 Jan Hubicka <jh@suse.cz>
14474 Martin Liska <marxin.liska@gmail.com>
14476 * cgraph.c (cgraph_function_body_availability): Do not check
14478 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
14479 symtab_node_availability): Declare.
14480 * ipa.c (can_replace_by_local_alias): New.
14481 (function_and_variable_visibility): Use it.
14482 * symtab.c (symtab_for_node_and_aliases,
14483 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
14485 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
14487 PR rtl-optimization/57963
14488 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
14489 (lra_constraints): Use them.
14491 2013-08-02 Sofiane Naci <sofiane.naci@arm.com>
14493 * config/arm/types.md (define_attr "type"): Add "load_acq"
14495 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
14497 (cortex_a53_store1): Likewise.
14499 2013-08-01 Jan Hubicka <jh@suse.cz>
14501 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
14502 partitions are not needed.
14504 2013-08-01 Uros Bizjak <ubizjak@gmail.com>
14506 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
14507 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
14508 MAYBE_NON_Q_CLASS_P where appropriate.
14510 2013-08-01 Jan Hubicka <jh@suse.cz>
14512 * cgraph.h (release_function_body): Declare.
14513 * tree.c (free_lang_data_in_decl): Free, parameters and return values
14514 of unused delcarations.
14516 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14518 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
14519 RTL form when subtracting a constant.
14521 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14523 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
14524 Generate canonical plus rtx with negated immediate instead of minus
14526 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
14528 2013-08-01 Jan Hubicka <jh@suse.cz>
14530 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
14531 (cgraph_release_function_body): Likewise.
14532 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
14533 * cgraph.h (cgrpah_node): Rename abstract_and_needed
14534 to used_as_abstract_origin.
14535 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
14536 symbols used as abstract origins.
14537 * cgraphunit.c (analyze_functions): Update.
14538 * ipa.c (symtab_remove_unreachable_nodes): Recompute
14539 used_as_abstract_origin.
14540 * tree-inline.c (tree_function_versioning): Update
14541 used_as_abstract_origin; be ready for DECL_RESULT and
14542 DECL_ARGUMENTS to be NULL.
14544 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
14545 for abstract functions.
14546 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
14549 2013-08-01 Jan Hubicka <jh@suse.cz>
14551 * profile.c (compute_value_histograms): Fix thinko.
14553 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
14555 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
14556 aarch-common-protos.h to extra_headers.
14557 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
14558 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
14559 * config/aarch64/t-aarch64 (aarch-common.o): Define.
14561 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
14563 * config/aarch64/aarch64.md (define_attr "type"): Delete.
14564 Include "../arm/types.md". Define "type" attribute for all patterns.
14565 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
14568 2013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
14570 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
14571 to support power8 load fusion.
14572 (fusion_gpr_mem_load): Likewise.
14574 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
14576 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
14577 declarations for power8 load fusion.
14578 (emit_fusion_gpr_load): Likewise.
14580 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
14581 tuning for power8, turn on fusion mode by default. Turn on sign
14582 extending fusion mode if normal fusion mode is on, and we are at
14584 (fusion_gpr_load_p): New function, return true if we can fuse an
14585 addis instruction with a dependent load to a GPR.
14586 (emit_fusion_gpr_load): Emit the instructions for power8 load
14589 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
14590 (VSX load fusion peepholes): New peepholes to fuse together an
14591 addi instruction with a VSX load instruction.
14593 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
14594 peepholes to fuse an addis instruction with a load to a GPR base
14595 register. If we are supporting sign extending fusions, convert
14596 sign extending loads to zero extending loads and add an explicit
14599 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
14601 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
14602 aarch-common-protos.h to extra_headers.
14603 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
14604 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
14605 (arm_early_store_addr_dep): Likewise.
14606 (arm_no_early_alu_shift_dep): Likewise.
14607 (arm_no_early_alu_shift_value_dep): Likewise.
14608 (arm_no_early_mul_dep): Likewise.
14609 (arm_no_early_store_addr_dep): Likewise.
14610 (arm_mac_accumulator_is_mul_result): Likewise.
14611 (arm_mac_accumulator_is_result): Likewise.
14612 * config/arm/aarch-common.c: ... here. New file.
14613 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
14615 (arm_early_store_addr_dep): Likewise.
14616 (arm_no_early_alu_shift_dep): Likewise.
14617 (arm_no_early_alu_shift_value_dep): Likewise.
14618 (arm_no_early_mul_dep): Likewise.
14619 (arm_no_early_store_addr_dep): Likewise.
14620 (arm_mac_accumulator_is_mul_result): Likewise.
14621 (arm_mac_accumulator_is_result): Likewise.
14622 * config/arm/aarch-common-protos.h: ... here. New file.
14623 * config/arm/t-arm (aarch-common.o): Define.
14625 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
14627 * config/arm/arm.md: Include new file "types.md".
14628 (define_attr "type"): Move from here to ...
14629 (define_attr "mul32"): Likewise.
14630 (define_attr "mul64"): Likewise.
14631 * config/arm/types.md: ... here. New file.
14633 2013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
14635 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
14636 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
14638 2013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
14640 * gen-pass-instances.awk: Fix offset of substr().
14642 2013-07-31 David Malcolm <dmalcolm@redhat.com>
14644 * Makefile.in (pass-instances.def): New.
14645 (passes.o): Replace dependency on passes.def with one on
14648 * gen-pass-instances.awk: New.
14650 * passes.c (pass_manager::pass_manager): Use pass-instances.def
14651 rather than passes.def, updating local definition of NEXT_PASS
14652 macro to add an extra NUM parameter (currently unused).
14654 2013-07-30 David Malcolm <dmalcolm@redhat.com>
14656 * Makefile.in (PASS_MANAGER_H): New.
14657 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
14658 (passes.o): Likewise.
14659 (statistics.o): Likewise.
14660 (cgraphunit.o): Likewise.
14661 (context.o): Depend on PASS_MANAGER_H.
14663 * pass_manager.h: New.
14665 * cgraphunit.c (cgraph_add_new_function): Update for moves
14666 of globals to fields of pass_manager.
14667 (analyze_function): Likewise.
14668 (expand_function): Likewise.
14669 (ipa_passes): Likewise.
14670 (compile): Likewise.
14672 * context.c (context::context): New.
14673 * context.h (context::context): New.
14674 (context::get_passes): New.
14675 (context::passes_): New.
14677 * lto-cgraph.c (input_node): Update for moves of globals to
14678 fields of pass_manager.
14680 * passes.c (all_passes): Remove, in favor of a field of the
14681 same name within the new class pass_manager.
14682 (all_small_ipa_passes): Likewise.
14683 (all_lowering_passes): Likewise.
14684 (all_regular_ipa_passes): Likewise.
14685 (all_late_ipa_passes): Likewise.
14686 (all_lto_gen_passes): Likewise.
14687 (passes_by_id): Likewise.
14688 (passes_by_id_size): Likewise.
14689 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
14690 the new class pass_manager.
14691 (set_pass_for_id): Convert to...
14692 (pass_manager::set_pass_for_id): ...method.
14693 (get_pass_for_id): Convert to...
14694 (pass_manager::get_pass_for_id): ...method.
14695 (register_one_dump_file): Move body of implementation into...
14696 (pass_manager::register_one_dump_file): ...here.
14697 (register_dump_files_1): Convert to...
14698 (pass_manager::register_dump_files_1): ...method.
14699 (register_dump_files): Convert to...
14700 (pass_manager::register_dump_files): ...method.
14701 (create_pass_tab): Update for moves of globals to fields of
14703 (dump_passes): Move body of implementation into...
14704 (pass_manager::dump_passes): ...here.
14705 (register_pass): Move body of implementation into...
14706 (pass_manager::register_pass): ...here.
14707 (init_optimization_passes): Convert into...
14708 (pass_manager::pass_manager): ...constructor for new
14709 pass_manager class, and initialize the pass_lists array.
14710 (check_profile_consistency): Update for moves of globals to
14711 fields of pass_manager.
14712 (dump_profile_report): Move body of implementation into...
14713 (pass_manager::dump_profile_report): ...here.
14714 (ipa_write_summaries_1): Update for moves of pass lists from
14715 being globals to fields of pass_manager.
14716 (ipa_write_optimization_summaries): Likewise.
14717 (ipa_read_summaries): Likewise.
14718 (ipa_read_optimization_summaries): Likewise.
14719 (execute_all_ipa_stmt_fixups): Likewise.
14721 * statistics.c (statistics_fini): Update for moves of globals to
14722 fields of pass_manager.
14724 * toplev.c (general_init): Replace call to
14725 init_optimization_passes with construction of the pass_manager
14728 * tree-pass.h (all_passes): Remove, in favor of a field of the
14729 same name within the new class pass_manager.
14730 (all_small_ipa_passes): Likewise.
14731 (all_lowering_passes): Likewise.
14732 (all_regular_ipa_passes): Likewise.
14733 (all_lto_gen_passes): Likewise.
14734 (all_late_ipa_passes): Likewise.
14735 (passes_by_id): Likewise.
14736 (passes_by_id_size): Likewise.
14737 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
14738 the new class pass_manager.
14739 (get_pass_for_id): Remove.
14741 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
14743 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
14746 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
14748 * arm.md (mulhi3): New expand pattern.
14750 2013-07-30 Jan Hubicka <jh@suse.cz>
14751 Martin Liska <marxin.liska@gmail.com>
14753 * profile.c (compute_value_histograms): Do not ICE when
14754 there is mismatch only on some counters.
14756 2013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
14758 PR rtl-optimization/57637
14759 * function.c (move_insn_for_shrink_wrap): Also check the
14760 GEN set of the LIVE problem for the liveness analysis
14761 if it exists, otherwise give up.
14763 2013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
14765 PR tree-optimization/57993
14766 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
14767 replaced statement in the candidate table.
14768 (phi_add_costs): Return infinite cost when the hidden basis does
14769 not dominate all phis on which the candidate is dependent.
14770 (replace_one_candidate): Record replaced statement in the
14773 2013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
14775 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
14776 (ashlv2si3): New expander.
14777 (*ashlv2si3_i): New define_insn_and_split.
14778 * predicates.md (float_operation): Allow patterns with three
14779 basic sub-patterns.
14781 PR rtl-optimization/58021
14782 * mode-switching.c (create_pre_exit): Always split off preceding
14783 insns if we are not at the basic block head.
14785 2013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
14787 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
14788 (UCLIBC_DYNAMIC_LINKER): New macro.
14789 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
14791 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
14792 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
14794 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
14795 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
14796 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
14797 for SF and DF modes. Use ieee_quad_format for TF mode.
14798 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
14799 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
14800 (mips_option_override): Handle `-mnan=legacy'.
14801 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
14802 `-mabs=2008' and `-mnan=2008'.
14803 (OPTION_DEFAULT_SPECS): Add "nan" default.
14804 (ASM_SPEC): Handle `-mnan='.
14805 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
14806 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
14807 comment accordingly.
14808 (neg<mode>2): Likewise.
14809 * config/mips/mips.opt (mabs, mnan): New options.
14810 * doc/install.texi (Configuration): Document `--with-nan=' option.
14811 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
14813 (MIPS Options): Document them.
14814 * config.gcc <mips*-*-*>: Handle `--with-nan='.
14815 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
14816 * configure: Regenerate.
14817 * config.in: Regenerate.
14819 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
14821 * config/i386/i386.md (float post-reload splitters): Do not check
14822 for subregs of SSE registers.
14824 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
14825 H.J. Lu <hongjiu.lu@intel.com>
14829 * config/i386/i386.md (post-reload splitter
14830 to avoid partial SSE reg dependency stalls): New pattern.
14832 2013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
14834 * config/s390/s390.md ("movcc"): Swap load and store instructions.
14836 2013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
14838 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
14839 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
14841 2013-07-26 Cary Coutant <ccoutant@google.com>
14843 * dwarf2out.c (die_checksum_ordered): Don't include template
14844 instantiations in signature.
14845 (is_template_parameter): New function.
14846 (is_template_instantiation): New function.
14847 (generate_skeleton_bottom_up): Don't include template instantiations
14849 (generate_skeleton): Likewise.
14850 (break_out_comdat_types): Move recursive call to break out nested
14852 (prune_unused_types_mark_generic_parms_dies): Call
14853 is_template_parameter.
14855 2013-07-26 Ian Bolton <ian.bolton@arm.com>
14857 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
14858 uses vector registers.
14859 * config/aarch64/iterators.md: Add attributes rtn and vas.
14861 2013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14862 Richard Earnshaw <richard.earnshaw@arm.com>
14864 * combine.c (simplify_comparison): Re-canonicalize operands
14866 * config/arm/arm.md (movcond_addsi): New splitter.
14868 2013-07-25 Sterling Augustine <saugustine@google.com>
14870 * dwarf2out.c (size_of_pubnames): Move code to...
14871 (include_pubname_in_output): ...here. New.
14872 (want_pubnames): Rearrange.
14873 (output_pubnames): Call include_pubname_in_output. Move assertion.
14875 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
14877 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
14879 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
14882 * doc/extend.texi: Remove obsolete builtins. Fix
14883 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
14885 2013-07-25 Jan Hubicka <jh@suse.cz>
14887 * cgraph.c (release_function_body): Break out from ...
14888 (cgraph_release_function_body): ... this one; also release DECL_RESULT
14889 and DECL_ARGUMENTS.
14890 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
14891 old_tree in the map.
14892 (create_specialized_node): Update.
14893 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
14895 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
14896 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
14897 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
14898 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
14901 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14903 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
14904 addsi3_carryin_alt2_<optab>): Correct output template.
14906 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14908 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
14909 Adjust for arm_restrict_it.
14910 Remove trailing whitespace.
14912 2013-07-25  Mark Kettenis  <kettenis@openbsd.org>
14914 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
14915 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
14917 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
14919 2013-07-25 Vladimir Makarov <vmakarov@redhat.com>
14921 PR rtl-optimization/57960
14922 * lra-constraints.c (process_alt_operands): Use the right mode
14923 when checking strict_low.
14925 2013-07-25 Jan Hubicka <jh@suse.cz>
14927 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
14928 * cgraph.c (cgraph_remove_node): Do not release function body
14929 when in cgraph streaming.
14930 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
14931 in other partitions are not considered reachable; fix handling of
14934 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
14936 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
14938 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
14943 * config/arm/arm.md ("*sibcall_insn): Replace use of
14944 Ss with US. Adjust output for v5 and v4t.
14945 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
14947 * config/arm/constraints.md ("Ss"): Rename to US.
14949 2013-07-25 Terry Guo <terry.guo@arm.com>
14951 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
14952 shift_add/shift_sub0/shift_sub1 RTXs.
14954 2013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
14955 Anton Blanchard <anton@au1.ibm.com>
14957 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
14958 (altivec_vpks<VI_char>ss): Likewise.
14959 (altivec_vpks<VI_char>us): Likewise.
14960 (altivec_vpku<VI_char>us): Likewise.
14961 (altivec_vpku<VI_char>um): Likewise.
14963 2013-07-24 David Malcolm <dmalcolm@redhat.com>
14965 Introduce context class.
14967 * Makefile.in (CONTEXT_H): New.
14968 (OBJS): Add context.o.
14969 (toplev.o): Add CONTEXT_H to dependencies.
14972 * toplev.c (general_init): Create the singleton gcc::context instance.
14978 2013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
14980 PR rtl-optimization/57968
14981 * mode-switching.c (create_pre_exit): Allow instructions that
14982 don't set a return register to need a non-exit mode.
14984 2013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
14985 Anton Blanchard <anton@au1.ibm.com>
14987 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
14988 operands to vperm for little endian.
14989 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
14990 of lvsl to create the control mask for a vperm for little endian.
14992 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14993 Anton Blanchard <anton@au1.ibm.com>
14995 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
14996 two operands for little-endian.
14998 2013-07-23 Steve Ellcey <sellcey@mips.com>
15000 * config/mips/mips.c (mips_case_values_threshold): New.
15001 (TARGET_CASE_VALUES_THRESHOLD): Define.
15003 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
15004 Anton Blanchard <anton@au1.ibm.com>
15006 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
15007 selection of field for vector splat in little endian mode.
15009 2013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
15011 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
15012 expanders to rs6000.md.
15013 (ior<mode>3): Likewise.
15014 (and<mode>3): Likewise.
15015 (one_cmpl<mode>2): Likewise.
15016 (nor<mode>3): Likewise.
15017 (andc<mode>3): Likewise.
15018 (eqv<mode>3): Likewise.
15019 (nand<mode>3): Likewise.
15020 (orc<mode>3): Likewise.
15022 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
15025 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
15026 to split multi-word logical operations.
15027 (rs6000_split_logical_di): Likewise.
15028 (rs6000_split_logical): Likewise.
15030 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
15031 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
15032 and allow TImode operations in 32-bit.
15033 (vsx_and<mode>3_64bit): Likewise.
15034 (vsx_ior<mode>3_32bit): Likewise.
15035 (vsx_ior<mode>3_64bit): Likewise.
15036 (vsx_xor<mode>3_32bit): Likewise.
15037 (vsx_xor<mode>3_64bit): Likewise.
15038 (vsx_one_cmpl<mode>2_32bit): Likewise.
15039 (vsx_one_cmpl<mode>2_64bit): Likewise.
15040 (vsx_nor<mode>3_32bit): Likewise.
15041 (vsx_nor<mode>3_64bit): Likewise.
15042 (vsx_andc<mode>3_32bit): Likewise.
15043 (vsx_andc<mode>3_64bit): Likewise.
15044 (vsx_eqv<mode>3_32bit): Likewise.
15045 (vsx_eqv<mode>3_64bit): Likewise.
15046 (vsx_nand<mode>3_32bit): Likewise.
15047 (vsx_nand<mode>3_64bit): Likewise.
15048 (vsx_orc<mode>3_32bit): Likewise.
15049 (vsx_orc<mode>3_64bit): Likewise.
15051 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
15052 logical types in GPRs.
15054 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
15055 logical insns to rs6000.md, and allow TImode operations in
15057 (altivec_ior<mode>3): Likewise.
15058 (altivec_xor<mode>3): Likewise.
15059 (altivec_one_cmpl<mode>2): Likewise.
15060 (altivec_nor<mode>3): Likewise.
15061 (altivec_andc<mode>3): Likewise.
15063 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
15064 attributes for moving the 128-bit logical operations into
15066 (BOOL_REGS_OUTPUT): Likewise.
15067 (BOOL_REGS_OP1): Likewise.
15068 (BOOL_REGS_OP2): Likewise.
15069 (BOOL_REGS_UNARY): Likewise.
15070 (BOOL_REGS_AND_CR0): Likewise.
15071 (one_cmpl<mode>2): Add support for DI logical operations on
15072 32-bit, splitting the operations to 32-bit.
15073 (anddi3): Likewise.
15074 (iordi3): Likewise.
15075 (xordi3): Likewise.
15076 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
15077 changes to combine the 32/64-bit code, allow logical operations on
15078 TI mode in 32-bit, and to use similar match_operator patterns like
15079 scalar mode uses. Combine the Altivec and VSX code for logical
15080 operations, and move it here.
15081 (ior<mode>3, 128-bit types): Likewise.
15082 (xor<mode>3, 128-bit types): Likewise.
15083 (one_cmpl<mode>3, 128-bit types): Likewise.
15084 (nor<mode>3, 128-bit types): Likewise.
15085 (andc<mode>3, 128-bit types): Likewise.
15086 (eqv<mode>3, 128-bit types): Likewise.
15087 (nand<mode>3, 128-bit types): Likewise.
15088 (orc<mode>3, 128-bit types): Likewise.
15089 (and<mode>3_internal): Likewise.
15090 (bool<mode>3_internal): Likewise.
15091 (boolc<mode>3_internal1): Likewise.
15092 (boolc<mode>3_internal2): Likewise.
15093 (boolcc<mode>3_internal1): Likewise.
15094 (boolcc<mode>3_internal2): Likewise.
15095 (eqv<mode>3_internal1): Likewise.
15096 (eqv<mode>3_internal2): Likewise.
15097 (one_cmpl1<mode>3_internal): Likewise.
15099 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
15101 * config/microblaze/microblaze.c (microblaze_expand_prologue):
15102 Rename flag_stack_usage to flag_stack_usage_info.
15104 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
15106 * config/microblaze/sync.md: New file.
15107 * config/microblaze/microblaze.md: Include sync.md
15108 * config/microblaze/microblaze.c: Add print_operand 'y'.
15109 * config/microblaze/constraints.md: Add memory_contraint
15110 'Q' which is a single register.
15112 2013-07-23 Eric Botcazou <ebotcazou@adacore.com>
15114 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
15116 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
15118 * reload.c (find_reloads): Exit loop once we find this operand
15119 cannot be reloaded somehow for this alternative.
15121 * reload.c (find_reloads): Exit loop once we find a hard register.
15123 * rtlanal.c (computed_jump_p): Exit loop once we find label
15126 * i386.c (ix86_pad_returns): Exit loop after setting replace.
15128 * cfgloopmanip.c (remove_path): Exit loop after setting
15131 * gensupport.c (subst_dup): Avoid loop if code is not
15132 MATCH_DUP nor MATCH_OP_DUP.
15134 2013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
15136 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
15138 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
15140 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
15141 true for SP_REGNUM if mode == ptr_mode.
15142 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
15143 with value R0_REGNUM + 31.
15145 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
15147 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
15148 pad pointer-typed argument downward.
15150 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
15152 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
15153 and __ILP32__ when the ILP32 model is in use.
15155 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
15157 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
15158 (aarch64_load_symref_appropriately): In the case of
15159 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
15160 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
15161 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
15162 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
15163 if 'mode' doesn't equal to 'ptr_mode'.
15164 (aarch64_output_mi_thunk): Add an assertion on the alignment of
15165 'vcall_offset'; change to call aarch64_emit_move differently depending
15166 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
15167 to calculate the upper bound of 'vcall_offset'.
15168 (aarch64_cannot_force_const_mem): Change to also return true if
15170 (aarch64_legitimize_reload_address): In the case of large
15171 displacements, add new local variable 'xmode' and an assertion
15172 based on it; change to use 'xmode' to generate the new rtx and
15174 (aarch64_asm_trampoline_template): Change to generate the template
15175 differently depending on TARGET_ILP32 or not; change to use
15176 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
15177 (aarch64_trampoline_size): Removed.
15178 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
15179 and replace immediate literals with it. Change to use 'ptr_mode'
15180 instead of 'DImode' and call convert_memory_address if the mode
15181 of 'fnaddr' doesn't equal to 'ptr_mode'.
15182 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
15184 (aarch64_elf_asm_destructor): Likewise.
15185 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
15186 on TARGET_ILP32 instead of aarch64_trampoline_size.
15187 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
15188 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
15189 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
15190 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
15191 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
15192 (storewb_pair<GPI:mode>_<P:mode>): ... this.
15193 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
15194 depending on the value of 'mode'.
15195 (add_losym_<mode>): New.
15196 (ldr_got_small_<mode>): New, based on ldr_got_small.
15197 (ldr_got_small): Remove.
15198 (ldr_got_small_sidi): New.
15199 * config/aarch64/iterators.md (P): New.
15200 (PTR): Change to 'ptr_mode' in the condition.
15202 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
15204 * config.gcc (aarch64*-*-*): Support --with-abi.
15205 (aarch64*-*-elf): Support --with-multilib-list.
15206 (aarch64*-*-linux*): Likewise.
15207 (supported_defaults): Add abi to aarch64*-*-*.
15208 * configure.ac: Mention AArch64 for --with-multilib-list.
15209 * configure: Re-generated.
15210 * config/aarch64/biarchilp32.h: New file.
15211 * config/aarch64/biarchlp64.h: New file.
15212 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
15214 (MULTILIB_DEFAULTS): Ditto.
15215 (DRIVER_SELF_SPECS): Ditto.
15216 (ASM_SPEC): Update to also substitute -mabi.
15217 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
15218 file whose name depends on -mabi= and -mbig-endian.
15219 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
15221 (POINTER_SIZE): New define.
15222 (POINTERS_EXTEND_UNSIGNED): Ditto.
15223 (enum aarch64_abi_type): New enumeration tag.
15224 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
15225 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
15226 (TARGET_ILP32): New define.
15227 * config/aarch64/aarch64.opt (mabi): New.
15228 (aarch64_abi): New.
15229 (ilp32, lp64): New values for -mabi.
15230 * config/aarch64/t-aarch64 (comma): New define.
15231 (MULTILIB_OPTIONS): Ditto.
15232 (MULTILIB_DIRNAMES): Ditto.
15233 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
15234 * doc/invoke.texi: Document -mabi for AArch64.
15236 2013-07-23 Georg-Johann Lay <avr@gjlay.de>
15238 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
15240 2013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
15241 Anton Blanchard <anton@au1.ibm.com>
15243 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
15244 endianness when selecting field to splat.
15246 2013-07-22 Eric Christopher <echristo@gmail.com>
15248 * dwarf2out.c (die_odr_checksum): New function to use
15249 CHECKSUM_ macros and ULEB128 for DIE tag.
15250 (generate_type_signature): Use.
15252 2013-07-22 Eric Botcazou <ebotcazou@adacore.com>
15254 * config.gcc (sparc*-*-*): Accept leon3 processor.
15255 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
15256 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
15257 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
15258 * config/sparc/sparc.opt (enum processor_type): Add leon3.
15259 (mfix-ut699): Adjust comment.
15260 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
15261 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
15262 (CPP_CPU_SPEC): Likewise.
15263 (ASM_CPU_SPEC): Likewise.
15264 * config/sparc/sparc.c (leon3_cost): New constant.
15265 (sparc_option_override): Add leon3 support.
15266 (mem_ref): New function.
15267 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
15268 (sparc_do_work_around_errata): Look into the instruction in the delay
15269 slot and adjust accordingly. Add fix for the data cache nullify issues
15270 of the UT699. Change insertion position for the NOP.
15271 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
15272 (leon3_load): New reservation.
15273 (leon_store): Bump latency to 2.
15274 (grfpu): New automaton.
15275 (grfpu_alu): New unit.
15276 (grfpu_ds): Likewise.
15277 (leon_fp_alu): Adjust.
15278 (leon_fp_mult): Delete.
15279 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
15280 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
15281 * config/sparc/sparc.md (cpu): Add leon3.
15282 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
15283 (swapsi): Likewise.
15284 (atomic_test_and_set): Likewise.
15285 (ldstub): Likewise.
15287 2013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
15289 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
15290 default for R5900 targets.
15291 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
15292 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
15293 * config/mips/mips.c (mips_option_override): Report an error for
15294 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
15295 for -march=r5900 -mhard-float.
15297 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
15299 * df-problems.c (can_move_insns_across): Exit loop once we
15300 find a non-fixed, non-global register.
15302 * ipa-pure-const.c (propagate_nothrow): Exit loop after
15305 * omega.c (omega_eliminate_red): Break after setting red_found.
15306 (omega_problem_has_red_equations): Similarly after setting found.
15307 (omega_query_variable): Similarly after setting coupled.
15309 2013-07-22 Marek Polacek <polacek@redhat.com>
15311 * gimplify.c: Don't include gimple.h twice.
15313 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15315 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
15316 instead of TARGET_THUMB1.
15317 (Pz): New constraint.
15318 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
15320 (compare_negsi_si): Likewise.
15321 (compare_addsi2_op0): Likewise.
15322 (compare_addsi2_op1): Likewise.
15323 (addsi3_carryin_<optab>): Likewise.
15324 (addsi3_carryin_alt2_<optab>): Likewise.
15325 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
15326 for arm_restrict_it.
15327 (subsi3_carryin): Likewise.
15328 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
15329 (minmax_arithsi): Disable for arm_restrict_it.
15330 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
15331 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
15332 (satsi_<SAT:code>_shift): Likewise.
15333 (arm_shiftsi3): Add alternative for 16-bit encoding.
15334 (arm32_movhf): Disable for arm_restrict_it.
15335 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
15336 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
15338 2013-07-22 Sofiane Naci <sofiane.naci@arm.com>
15340 * config/arm/arm.md (attribute "insn"): Delete.
15341 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
15342 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
15343 (not_shiftsi): Update for attribute change.
15344 (not_shiftsi_compare0): Likewise.
15345 (not_shiftsi_compare0_scratch): Likewise.
15346 (arm_one_cmplsi2): Likewise.
15347 (thumb1_one_cmplsi2): Likewise.
15348 (notsi_compare0): Likewise.
15349 (notsi_compare0_scratch): Likewise.
15350 (thumb1_movdi_insn): Likewise.
15351 (arm_movsi_insn): Likewise.
15352 (movhi_insn_arch4): Likewise.
15353 (movhi_bytes): Likewise.
15354 (arm_movqi_insn): Likewise.
15355 (thumb1_movqi_insn): Likewise.
15356 (arm32_movhf): Likewise.
15357 (thumb1_movhf): Likewise.
15358 (arm_movsf_soft_insn): Likewise.
15359 (thumb1_movsf_insn): Likewise.
15360 (thumb_movdf_insn): Likewise.
15361 (movsicc_insn): Likewise.
15362 (movsfcc_soft_insn): Likewise.
15363 (and_scc): Likewise.
15364 (cond_move): Likewise.
15365 (if_move_not): Likewise.
15366 (if_not_move): Likewise.
15367 (if_shift_move): Likewise.
15368 (if_move_shift): Likewise.
15369 (if_shift_shift): Likewise.
15370 (if_not_arith): Likewise.
15371 (if_arith_not): Likewise.
15372 (cond_move_not): Likewise.
15373 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
15374 (neon_mov<mode>): Likewise.
15375 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
15376 (thumb2_movsi_vfp): Likewise.
15377 (movsf_vfp): Likewise.
15378 (thumb2_movsf_vfp): Likewise.
15379 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
15381 (cortexa7_older_only): Likewise.
15382 (cortexa7_younger): Likewise.
15383 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
15384 (1020alu_shift_op): Likewise.
15385 (1020alu_shift_reg_op): Likewise.
15386 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
15387 (alu_shift_op): Likewise.
15388 (alu_shift_reg_op): Likewise.
15389 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
15390 (11_alu_shift_op): Likewise.
15391 (11_alu_shift_reg_op): Likewise.
15392 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
15393 (9_alu_shift_reg_op): Likewise.
15394 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
15396 (cortex_a15_alu_shift): Likewise.
15397 (cortex_a15_alu_shift_reg): Likewise.
15398 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
15399 (cortex_a5_alu_shift): Likewise.
15400 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
15402 (cortex_a53_alu_shift): Likewise.
15403 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
15405 (cortex_a7_alu_reg): Likewise.
15406 (cortex_a7_alu_shift): Likewise.
15407 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
15408 (cortex_a8_alu_shift): Likewise.
15409 (cortex_a8_alu_shift_reg): Likewise.
15410 (cortex_a8_mov): Likewise.
15411 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
15412 (cortex_a9_dp_shift): Likewise.
15413 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
15414 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
15415 (cortex_r4_mov): Likewise.
15416 (cortex_r4_alu_shift): Likewise.
15417 (cortex_r4_alu_shift_reg): Likewise.
15418 * config/arm/fa526.md (526_alu_op): Update for attribute change.
15419 (526_alu_shift_op): Likewise.
15420 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
15421 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
15422 (626te_alu_shift_op): Likewise.
15423 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
15424 (726te_alu_op): Likewise.
15425 (726te_alu_shift_op): Likewise.
15426 (726te_alu_shift_reg_op): Likewise.
15427 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
15428 (mp626_alu_shift_op): Likewise.
15429 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
15430 (pj4_alu_e1_conds): Likewise.
15431 (pj4_alu): Likewise.
15432 (pj4_alu_conds): Likewise.
15433 (pj4_shift): Likewise.
15434 (pj4_shift_conds): Likewise.
15435 (pj4_alu_shift): Likewise.
15436 (pj4_alu_shift_conds): Likewise.
15438 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15440 * config/arm/predicates.md (shiftable_operator_strict_it):
15442 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
15443 Disable cond_exec version for arm_restrict_it.
15444 (thumb2_smaxsi3): Convert to generate cond_exec.
15445 (thumb2_sminsi3): Likewise.
15446 (thumb32_umaxsi3): Likewise.
15447 (thumb2_uminsi3): Likewise.
15448 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
15449 (thumb2_neg_abssi2): Likewise.
15450 (thumb2_mov_scc): Add alternative for 16-bit encoding.
15451 (thumb2_movsicc_insn): Adjust alternatives.
15452 (thumb2_mov_negscc): Disable for arm_restrict_it.
15453 (thumb2_mov_negscc_strict_it): New pattern.
15454 (thumb2_mov_notscc_strict_it): New pattern.
15455 (thumb2_mov_notscc): Disable for arm_restrict_it.
15456 (thumb2_ior_scc): Likewise.
15457 (thumb2_ior_scc_strict_it): New pattern.
15458 (thumb2_cond_move): Adjust for arm_restrict_it.
15459 (thumb2_cond_arith): Disable for arm_restrict_it.
15460 (thumb2_cond_arith_strict_it): New pattern.
15461 (thumb2_cond_sub): Adjust for arm_restrict_it.
15462 (thumb2_movcond): Likewise.
15463 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
15464 (thumb2_zero_extendhisi2_v6): Likewise.
15465 (thumb2_zero_extendqisi2_v6): Likewise.
15466 (orsi_notsi_si): Likewise.
15467 (orsi_not_shiftsi_si): Likewise.
15469 2013-07-22 Georg-Johann Lay <avr@gjlay.de>
15471 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
15472 instruction sequence is 1 byte shorter.
15474 2013-07-22 Uros Bizjak <ubizjak@gmail.com>
15476 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
15477 it is not needed after split.
15479 2013-07-20 Iain Sandoe <iain@codesourcery.com>
15482 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
15483 second label for nonlocal goto receivers. Don't output pic base labels
15484 unless we're producing PIC; mark that action unreachable().
15485 (ix86_save_reg): If the function contains a nonlocal label, save the
15487 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
15488 * config/darwin.c (emitted_pic_label_num): New GTY.
15489 (update_pic_label_number_if_needed): New.
15490 (machopic_output_function_base_name): Adjust for nonlocal receiver
15492 (machopic_should_output_picbase_label): New.
15493 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
15494 (nonlocal_goto_receiver): New insn and split.
15496 2013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
15498 * config/aarch64/aarch64-builtins.c
15499 (aarch64_fold_builtin): Fold abs in all modes.
15500 * config/aarch64/aarch64-simd-builtins.def
15501 (abs): Enable for all modes.
15502 * config/aarch64/arm_neon.h
15503 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
15504 (vabs_f64): Add missing intrinsic.
15506 2013-07-19 Ian Bolton <ian.bolton@arm.com>
15508 * config/aarch64/arm_neon.h (vabs_s64): New function
15510 2013-07-19 Georg-Johann Lay <avr@gjlay.de>
15513 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
15514 * config/avr/avr.md (adjust_len): Add `round'.
15515 * config/avr/avr-protos.h (avr_out_round): New prototype.
15516 (avr_out_plus): Add `out_label' argument.
15517 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
15518 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
15519 Handle the case where `insn' is just a pattern.
15520 (avr_out_bitop): Handle the case where `insn' is just a pattern.
15521 (avr_out_round): New function.
15522 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
15524 2013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
15526 * config/microblaze/microblaze.c (microblaze_expand_prologue):
15527 Add check for flag_stack_usage to handle -fstack-usage support
15529 2013-07-18 Pat Haugen <pthaugen@us.ibm.com>
15531 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
15532 interaction for new Power8 flags and VSX.
15534 2013-07-18 Sriraman Tallam <tmsriram@google.com>
15536 PR middle-end/57698
15537 * tree-inline.c (expand_call_inline): Emit errors during
15538 early_inlining only if optimization is not turned on.
15540 2013-07-18 David Malcolm <dmalcolm@redhat.com>
15544 * passes.c (init_optimization_passes): Move the construction of
15545 the pass hierarchy into a new passes.def file.
15547 * Makefile.in (passes.o): Add dependency on passes.def.
15549 2013-07-18 David Malcolm <dmalcolm@redhat.com>
15551 * passes.c (init_optimization_passes): Introduce macros for
15552 constructing the tree of passes (INSERT_PASSES_AFTER,
15553 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
15554 TERMINATE_PASS_LIST).
15556 2013-07-18 Vladimir Makarov <vmakarov@redhat.com>
15557 Wei Mi <wmi@google.com>
15559 PR rtl-optimization/57878
15560 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
15562 (reload_pseudo_compare_func): Check nregs first for reload
15565 2013-07-18 David Malcolm <dmalcolm@redhat.com>
15567 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
15569 2013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
15571 * read-rtl.c (validate_const_int): Once an invalid character is
15572 seen, quit the loop.
15574 * gengtype.c (write_roots): Similarly once we find the "deletable"
15575 or "if_marked" option.
15577 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
15579 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
15580 "xtab" and "sat". Move value "clz" from here to ...
15581 (attriubte "type"): ... here.
15582 (satsi_<SAT:code>): Delete "insn" attribute.
15583 (satsi_<SAT:code>_shift): Likewise.
15584 (arm_zero_extendqisi2addsi): Likewise.
15585 (arm_extendqisi2addsi): Likewise.
15586 (clzsi2): Update for attribute changes.
15587 (rbitsi2): Likewise.
15588 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
15590 (arm_usatsihi): Likewise.
15591 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
15593 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
15595 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
15596 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
15597 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
15598 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
15599 in alphabetical order.
15600 (attribute "core_cycles"): Update for attribute changes.
15601 (arm_addsi3): Likewise.
15602 (addsi3_compare0): Likewise.
15603 (addsi3_compare0_scratch): Likewise.
15604 (addsi3_compare_op1): Likewise.
15605 (addsi3_compare_op2): Likewise.
15606 (compare_addsi2_op0): Likewise.
15607 (compare_addsi2_op1): Likewise.
15608 (addsi3_carryin_shift_<optab>): Likewise.
15609 (subsi3_carryin_shift): Likewise.
15610 (rsbsi3_carryin_shift): Likewise.
15611 (arm_subsi3_insn): Likewise.
15612 (subsi3_compare0): Likewise.
15613 (subsi3_compare): Likewise.
15614 (arm_andsi3_insn): Likewise.
15615 (thumb1_andsi3_insn): Likewise.
15616 (andsi3_compare0): Likewise.
15617 (andsi3_compare0_scratch): Likewise.
15618 (zeroextractsi_compare0_scratch
15619 (andsi_not_shiftsi_si): Likewise.
15620 (iorsi3_insn): Likewise.
15621 (iorsi3_compare0): Likewise.
15622 (iorsi3_compare0_scratch): Likewise.
15623 (arm_xorsi3): Likewise.
15624 (thumb1_xorsi3_insn): Likewise.
15625 (xorsi3_compare0): Likewise.
15626 (xorsi3_compare0_scratch): Likewise.
15627 (satsi_<SAT:code>_shift): Likewise.
15629 (arm_shiftsi3): Likewise.
15630 (shiftsi3_compare0): Likewise.
15631 (not_shiftsi): Likewise.
15632 (not_shiftsi_compare0): Likewise.
15633 (not_shiftsi_compare0_scratch): Likewise.
15634 (arm_one_cmplsi2): Likewise.
15635 (thumb_one_complsi2): Likewise.
15636 (notsi_compare0): Likewise.
15637 (notsi_compare0_scratch): Likewise.
15638 (thumb1_zero_extendhisi2): Likewise.
15639 (arm_zero_extendhisi2): Likewise.
15640 (arm_zero_extendhisi2_v6): Likewise.
15641 (arm_zero_extendhisi2addsi): Likewise.
15642 (thumb1_zero_extendqisi2): Likewise.
15643 (thumb1_zero_extendqisi2_v6): Likewise.
15644 (arm_zero_extendqisi2): Likewise.
15645 (arm_zero_extendqisi2_v6): Likewise.
15646 (arm_zero_extendqisi2addsi): Likewise.
15647 (thumb1_extendhisi2): Likewise.
15648 (arm_extendhisi2): Likewise.
15649 (arm_extendhisi2_v6): Likewise.
15650 (arm_extendqisi): Likewise.
15651 (arm_extendqisi_v6): Likewise.
15652 (arm_extendqisi2addsi): Likewise.
15653 (thumb1_extendqisi2): Likewise.
15654 (thumb1_movdi_insn): Likewise.
15655 (arm_movsi_insn): Likewise.
15656 (movsi_compare0): Likewise.
15657 (movhi_insn_arch4): Likewise.
15658 (movhi_bytes): Likewise.
15659 (arm_movqi_insn): Likewise.
15660 (thumb1_movqi_insn): Likewise.
15661 (arm32_movhf): Likewise.
15662 (thumb1_movhf): Likewise.
15663 (arm_movsf_soft_insn): Likewise.
15664 (thumb1_movsf_insn): Likewise.
15665 (movdf_soft_insn): Likewise.
15666 (thumb_movdf_insn): Likewise.
15667 (arm_cmpsi_insn): Likewise.
15668 (cmpsi_shiftsi): Likewise.
15669 (cmpsi_shiftsi_swp): Likewise.
15670 (arm_cmpsi_negshiftsi_si): Likewise.
15671 (movsicc_insn): Likewise.
15672 (movsfcc_soft_insn): Likewise.
15673 (arith_shiftsi): Likewise.
15674 (arith_shiftsi_compare0
15675 (arith_shiftsi_compare0_scratch
15676 (sub_shiftsi): Likewise.
15677 (sub_shiftsi_compare0
15678 (sub_shiftsi_compare0_scratch
15679 (and_scc): Likewise.
15680 (cond_move): Likewise.
15681 (if_plus_move): Likewise.
15682 (if_move_plus): Likewise.
15683 (if_move_not): Likewise.
15684 (if_not_move): Likewise.
15685 (if_shift_move): Likewise.
15686 (if_move_shift): Likewise.
15687 (if_shift_shift): Likewise.
15688 (if_not_arith): Likewise.
15689 (if_arith_not): Likewise.
15690 (cond_move_not): Likewise.
15691 (thumb1_ashlsi3): Set type attribute.
15692 (thumb1_ashrsi3): Likewise.
15693 (thumb1_lshrsi3): Likewise.
15694 (thumb1_rotrsi3): Likewise.
15695 (shiftsi3_compare0_scratch): Likewise.
15696 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
15697 (neon_mov<mode>): Likewise.
15698 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
15700 (thumb2_movsi_insn): Likewise.
15701 (thumb2_cmpsi_neg_shiftsi): Likewise.
15702 (thumb2_extendqisi_v6): Likewise.
15703 (thumb2_zero_extendhisi2_v6): Likewise.
15704 (thumb2_zero_extendqisi2_v6): Likewise.
15705 (thumb2_shiftsi3_short): Likewise.
15706 (thumb2_addsi3_compare0_scratch): Likewise.
15707 (orsi_not_shiftsi_si): Likewise.
15708 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
15709 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
15711 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
15712 (1020alu_shift_op): Likewise.
15713 (1020alu_shift_reg_op): Likewise.
15714 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
15715 (alu_shift_op): Likewise.
15716 (alu_shift_reg_op): Likewise.
15717 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
15718 (11_alu_shift_op): Likewise.
15719 (11_alu_shift_reg_op): Likewise.
15720 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
15721 (9_alu_shift_reg_op): Likewise.
15722 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
15724 (cortex_a15_alu_shift): Likewise.
15725 (cortex_a15_alu_shift_reg): Likewise.
15726 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
15728 (cortex_a5_alu_shift): Likewise.
15729 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
15731 (cortex_a53_alu_shift): Likewise.
15732 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
15734 (cortex_a7_alu_reg): Likewise.
15735 (cortex_a7_alu_shift): Likewise.
15736 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
15738 (cortex_a8_alu_shift): Likewise.
15739 (cortex_a8_alu_shift_reg): Likewise.
15740 (cortex_a8_mov): Likewise.
15741 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
15742 (cortex_a9_dp_shift): Likewise.
15743 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
15745 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
15747 (cortex_r4_mov): Likewise.
15748 (cortex_r4_alu_shift): Likewise.
15749 (cortex_r4_alu_shift_reg): Likewise.
15750 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
15751 (526_alu_shift_op): Likewise.
15752 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
15753 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
15754 (626te_alu_shift_op): Likewise.
15755 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
15756 (726te_alu_op): Likewise.
15757 (726te_alu_shift_op): Likewise.
15758 (726te_alu_shift_reg_op): Likewise.
15759 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
15760 (mp626_alu_shift_op): Likewise.
15761 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
15762 (pj4_alu_e1_conds): Likewise.
15763 (pj4_alu): Likewise.
15764 (pj4_alu_conds): Likewise.
15765 (pj4_shift): Likewise.
15766 (pj4_shift_conds): Likewise.
15767 (pj4_alu_shift): Likewise.
15768 (pj4_alu_shift_conds): Likewise.
15769 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
15771 (cortexa7_older_only): Likewise.
15772 (cortexa7_younger): Likewise.
15774 2013-07-18 David Malcolm <dmalcolm@redhat.com>
15776 * ipa-pure-const.c (generate_summary): Rename to...
15777 (pure_const_generate_summary): ... this.
15779 2013-07-17 Iain Sandoe <iain@codesourcery.com>
15781 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
15783 2013-07-17 Yvan Roux <yvan.roux@linaro.org>
15786 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
15789 2013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
15791 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
15792 enabled without -march=zEC12.
15793 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
15796 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
15798 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
15799 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
15800 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
15801 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
15802 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
15803 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
15804 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
15806 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
15807 and ISA_HAS_NMADD3_NMSUB3.
15808 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
15809 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
15810 (nmsub4<mode>, nmsub3<mode>): Likewise.
15811 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
15813 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
15815 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
15816 TARGET_MIPS5400 checking.
15818 2013-07-16 Jakub Jelinek <jakub@redhat.com>
15819 Peter Bergner <bergner@vnet.ibm.com>
15821 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
15822 registers in the comment.
15823 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
15824 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
15825 rather than FIRST_PSEUDO_REGISTERS.
15827 2013-07-16 Peter Bergner <bergner@vnet.ibm.com>
15829 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
15830 enable extra ISA flags with TARGET_HTM.
15832 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
15834 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
15837 2013-07-15 Cong Hou <congh@google.com>
15839 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
15840 in compare function for sorting.
15842 2013-07-15 Peter Bergner <bergner@vnet.ibm.com>
15844 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
15845 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
15846 * config/rs6000/rs6000.opt: Add -mhtm option.
15847 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
15848 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
15849 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
15850 __HTM__ if the HTM instructions are available.
15851 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
15852 htm_spr_reg_operand): New define_predicates.
15853 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
15854 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
15856 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
15857 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
15858 HTM builtin functions.
15859 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
15860 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
15861 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
15862 (rs6000_builtin_mask_calculate): Likewise.
15863 (rs6000_option_override_internal): Likewise.
15864 (bdesc_htm): Add new HTM builtin support.
15865 (htm_spr_num): New function.
15866 (htm_spr_regno): Likewise.
15867 (rs6000_htm_spr_icode): Likewise.
15868 (htm_expand_builtin): Likewise.
15869 (htm_init_builtins): Likewise.
15870 (rs6000_expand_builtin): Add support for HTM builtin functions.
15871 (rs6000_init_builtins): Likewise.
15872 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
15874 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
15875 (TARGET_HTM, MASK_HTM): Define macros.
15876 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
15877 (FIXED_REGISTERS): Likewise.
15878 (CALL_USED_REGISTERS): Likewise.
15879 (CALL_REALLY_USED_REGISTERS): Likewise.
15880 (REG_ALLOC_ORDER): Likewise.
15881 (enum reg_class): Likewise.
15882 (REG_CLASS_NAMES): Likewise.
15883 (REG_CLASS_CONTENTS): Likewise.
15884 (REGISTER_NAMES): Likewise.
15885 (ADDITIONAL_REGISTER_NAMES): Likewise.
15886 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
15887 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
15888 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
15889 * config/rs6000/htm.md: New file.
15890 * config/rs6000/htmintrin.h: New file.
15891 * config/rs6000/htmxlintrin.h: New file.
15893 2013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
15895 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
15896 Define SYMBOL_TINY_GOT, update comment.
15897 * config/aarch64/aarch64.c
15898 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
15899 (aarch64_expand_mov_immediate): Likewise.
15900 (aarch64_print_operand): Likewise.
15901 (aarch64_classify_symbol): Likewise.
15902 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
15903 (ldr_got_tiny): Define.
15905 2013-07-13 Tobias Grosser <tobias@grosser.es>
15907 PR tree-optimization/54094
15908 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
15909 scheduling dimension for the parallelism check from the polyhedral
15910 information in the AST.
15911 * graphite-dependences.c (carries_deps): Do not assume the schedule is
15914 2013-07-13 Jason Merrill <jason@redhat.com>
15916 * print-tree.c (debug_vec_tree): Use debug_raw.
15917 (debug_raw (vec<tree, va_gc> &)): New.
15918 (debug_raw (vec<tree, va_gc> *)): New.
15919 * tree.h: Declare them.
15921 2013-07-13 Bin Cheng <bin.cheng@arm.com>
15923 * ifcvt.c (ifcvt_after_combine): New static variable.
15924 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
15926 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
15927 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
15928 rest_of_handle_if_after_reload): Pass new argument for if_convert.
15930 2013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
15932 * config/mips/mips.c (mips_expand_call): Remove empty statement.
15934 2013-07-12 Michael Matz <matz@suse.de>
15936 PR middle-end/55771
15937 * convert.c (convert_to_real): Reject non-float inner types.
15939 2013-07-12 Tejas Belagod <tejas.belagod@arm.com>
15941 * config/aarch64/aarch64-protos.h
15942 (aarch64_simd_immediate_valid_for_move): Remove.
15943 * config/aarch64/aarch64.c (simd_immediate_info): New member.
15944 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
15946 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
15948 2013-07-11 Steve Ellcey <sellcey@mips.com>
15950 * config/mips/mips.c (mips_conditional_register_usage): Do not
15951 use t[0-7] registers in MIPS16 mode when optimizing for size.
15953 2013-07-11 Sriraman Tallam <tmsriram@google.com>
15955 * config/i386/i386.c (dispatch_function_versions): Fix array
15956 indexing of function_version_info to match actual_versions.
15958 2013-07-11 Teresa Johnson <tejohnson@google.com>
15960 * vec.h (struct va_gc): Move release out-of-line.
15961 (va_gc::release): Call ggc_free on released vec.
15963 2013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
15965 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
15966 Require GOT register as additional operand in UNSPEC.
15967 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
15968 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
15969 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
15970 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
15971 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
15972 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
15973 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
15975 2013-07-11 Georg-Johann Lay <avr@gjlay.de>
15978 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
15979 name seen by assembler/linker rather if available.
15981 2013-07-11 Andreas Schwab <schwab@suse.de>
15983 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
15985 2013-07-10 Vladimir Makarov <vmakarov@redhat.com>
15987 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
15989 2013-07-10 Joseph Myers <joseph@codesourcery.com>
15991 * doc/tm.texi.in: Move hook documentation to ....
15992 * target.def: ... here.
15994 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
15995 text on @hook line.
15996 * doc/tm.texi: Regenerate.
15998 2013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
16001 * doc/invoke.texi: Document -Wconditionally-supported.
16003 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
16006 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
16009 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
16012 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
16013 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
16014 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
16015 Remove duplicate devices.
16016 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
16017 * config/avr/t-multilib: Regenerate.
16018 * config/avr/avr-tables.opt: Regenerate.
16019 * doc/avr-mmcu.texi: Regenerate.
16021 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
16024 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
16026 2013-07-10 Graham Stott <graham.stott@btinternet.com>
16028 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
16029 the cost of MULT when optimizing for size.
16031 2013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
16033 * config/cr16/cr16-protos.h: Don't include target.h.
16035 2013-07-09 Joseph Myers <joseph@codesourcery.com>
16037 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
16038 adjust register size for TDmode and TFmode for VSX registers.
16040 2013-07-08 Kai Tietz <ktietz@redhat.com>
16043 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
16044 hook_bool_const_tree_true.
16046 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16048 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
16049 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
16050 * config/s390/s390.md: Define FPR*_REGNUM constants.
16051 Fix FPR2_REGNUM constant (18 -> 17).
16052 ("*trunc<BFP:mode><DFP_ALL:mode>2")
16053 ("*trunc<DFP_ALL:mode><BFP:mode>2")
16054 ("trunc<BFP:mode><DFP_ALL:mode>2")
16055 ("trunc<DFP_ALL:mode><BFP:mode>2")
16056 ("*extend<BFP:mode><DFP_ALL:mode>2")
16057 ("*extend<DFP_ALL:mode><BFP:mode>2")
16058 ("extend<BFP:mode><DFP_ALL:mode>2")
16059 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
16062 2013-07-08 Graham Stott <graham.stott@btinternet.com>
16064 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
16066 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16068 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
16069 and cfun_fpr_bit_p to cfun_fpr_save_p.
16070 (s390_frame_area, s390_register_info, s390_frame_info)
16071 (s390_emit_prologue, s390_emit_epilogue)
16072 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
16074 * config/s390/s390.h: Define *_REGNUM macros for floating point
16077 2013-07-08 Eric Botcazou <ebotcazou@adacore.com>
16079 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
16081 2013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
16083 PR rtl-optimization/57786
16084 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
16085 and break out of the loop when it is set to false.
16087 2013-07-08 Jakub Jelinek <jakub@redhat.com>
16090 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
16091 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
16092 (const_int 63)) 0)).
16093 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
16094 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
16095 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
16097 PR rtl-optimization/57829
16098 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
16099 mask bits outside of mode are just sign-extension from mode to HWI.
16101 2013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
16103 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
16104 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
16105 adjust_address instead of change_address to keep info about alignment.
16106 (emit_strmov): Remove.
16107 (emit_memmov): New function.
16108 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
16109 (expand_movmem_epilogue): Likewise and return updated rtx for
16111 (expand_constant_movmem_prologue): Likewise and return updated rtx for
16112 destination and source.
16113 (decide_alignment): Refactor, handle vector_loop.
16114 (ix86_expand_movmem): Likewise.
16115 (ix86_expand_setmem): Likewise.
16116 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
16118 2013-07-07 Uros Bizjak <ubizjak@gmail.com>
16120 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
16121 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
16123 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
16125 * config/i386/sse.md (sse_movlhps): Change alternative 3
16126 of operand 2 to "m".
16128 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
16131 * config/i386/sse.md (iptr): New mode attribute.
16132 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
16133 (<sse>_vm<plusminus_insn><mode>3): Ditto.
16134 (<sse>_vmmul<mode>3): Ditto.
16135 (<sse>_vmdiv<mode>3): Ditto.
16136 (sse_vmrcpv4sf2): Ditto.
16137 (<sse>_vmsqrt<mode>2): Ditto.
16138 (sse_vmrsqrtv4sf2): Ditto.
16139 (<sse>_vm<code><mode>3): Ditto.
16140 (avx_vmcmp<mode>3): Ditto.
16141 (<sse>_vmmaskcmp<mode>3): Ditto.
16142 (<sse>_comi): Ditto.
16143 (<sse>_ucomi): Ditto.
16144 (*xop_vmfrcz_<mode>): Ditto.
16145 (*fmai_fmadd_<mode>): Ditto.
16146 (*fmai_fmsub_<mode>): Ditto.
16147 (*fmai_fnmadd_<mode>): Ditto.
16148 (*fmai_fnmsub_<mode>): Ditto.
16149 (*fma4i_vmfmadd_<mode>): Ditto.
16150 (*fma4i_vmfmsub_<mode>): Ditto.
16151 (*fma4i_vmfnmadd_<mode>): Ditto.
16152 (*fma4i_vmfnmsub_<mode>): Ditto.
16153 (*xop_vmfrcz_<mode>): Ditto.
16154 (sse_cvtps2pi): Ditto.
16155 (sse_cvttps2pi): Ditto.
16156 (sse_cvtss2si): Ditto.
16157 (sse_cvtss2si_2): Ditto.
16158 (sse_cvtss2siq_2): Ditto.
16159 (sse_cvttss2si): Ditto.
16160 (sse_cvttss2siq): Ditto.
16161 (sse_cvtsd2si): Ditto.
16162 (sse_cvtsd2si_2): Ditto.
16163 (sse_cvtsd2siq_2): Ditto.
16164 (sse_cvttsd2si): Ditto.
16165 (sse_cvttsd2siq): Ditto.
16166 (sse_cvtsd2ss): Ditto.
16167 (sse_cvtss2sd): Ditto.
16168 (avx2_pbroadcast<mode>): Ditto.
16169 (avx2_pbroadcast<mode>_1): Ditto.
16170 (*avx_vperm_broadcast_v4sf): Ditto.
16172 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
16173 (sse_movlhps): Ditto.
16174 (sse_storehps): Ditto.
16175 (sse_loadhps): Ditto.
16176 (sse_storelps): Ditto.
16177 (sse_loadlps): Ditto.
16178 (*vec_concatv4sf): Ditto.
16179 (*vec_interleave_highv2df): Ditto.
16180 (*vec_interleave_lowv2df): Ditto.
16181 (*vec_extractv2df_1_sse): Ditto.
16182 (*vec_extractv2df_0_sse): Ditto.
16183 (sse2_storelpd): Ditto.
16184 (sse2_loadlpd): Ditto.
16185 (sse2_movsd): Ditto.
16186 (*vec_concatv4si): Ditto.
16187 (vec_concatv2di): Ditto.
16189 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
16190 for Intel asm dialect.
16191 (mmx_punpcklwd): Ditto.
16192 (mmx_punpckldq): Ditto.
16194 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
16195 for intel assembler dialect.
16197 2013-07-06 Jakub Jelinek <jakub@redhat.com>
16200 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
16201 for BUILT_IN_C{LZ,LRSB}*.
16202 * tree.h (CASE_INT_FN): Add FN##IMAX case.
16203 * tree-vrp.c (extract_range_basic): Handle
16204 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
16205 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
16206 fall thru to code calling set_value*.
16207 * builtins.c (expand_builtin): Remove *IMAX cases.
16208 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
16209 if width is bigger than 2*HWI.
16211 2013-07-05 Vladimir Makarov <vmakarov@redhat.com>
16213 PR rtl-optimization/55342
16214 * lra-int.h (lra_subreg_reload_pseudos): New.
16215 * lra.c: Add undoing optional reloads to the block diagram.
16216 (lra_subreg_reload_pseudos): New.
16217 (lra_optional_reload_pseudos): Change comments.
16218 (lra): Init and clear lra_subreg_reload_pseudos. Clear
16219 lra_optional_reload_pseudos after undo transformations.
16220 * lra-assigns.c (pseudo_prefix_title): New.
16221 (lra_setup_reg_renumber): Use it.
16222 (spill_for): Ditto. Check subreg reload pseudos too.
16223 (assign_by_spills): Consider subreg reload pseudos too.
16224 * lra-constraints.c (simplify_operand_subreg): Use
16225 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
16226 (curr_insn_transform): Recognize and do optional reloads.
16227 (undo_optional_reloads): New.
16228 (lra_undo_inheritance): Call undo_optional_reloads.
16230 2013-07-05 Thomas Quinot <quinot@adacore.com>
16232 * tree-complex.c (expand_complex_operations_1): Fix typo.
16234 2013-07-04 Tejas Belagod <tejas.belagod@arm.com>
16236 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
16237 (tune_params): New member 'const vec_costs'.
16238 * config/aarch64/aarch64.c (generic_vector_cost): New.
16239 (generic_tunings): New member 'generic_vector_cost'.
16240 (aarch64_builtin_vectorization_cost): New.
16241 (aarch64_add_stmt_cost): New.
16242 (TARGET_VECTORIZE_ADD_STMT_COST): New.
16243 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
16245 2013-07-03 Jakub Jelinek <jakub@redhat.com>
16248 * config/i386/predicates.md (vsib_address_operand): Disallow
16249 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
16251 2013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
16253 PR middle-end/55030
16254 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
16255 expand_builtin_setjmp_receiver.
16256 (expand_label): Adjust, call expand_builtin_setjmp_receiver
16257 with NULL for the label parameter.
16258 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
16259 the frame-pointer. Adjust comments.
16260 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
16261 only if LABEL is non-NULL.
16263 2013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
16265 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
16266 (ARM_ABI_AAPCS64): Ditto.
16268 (ARM_DEFAULT_ABI): Ditto.
16270 2013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
16272 * config/aarch64/aarch64-builtins.c
16273 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
16274 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
16276 * config/aarch64/aarch64-simd.md
16277 (aarch64_ld1<VALL:mode>): New.
16278 (aarch64_st1<VALL:mode>): Likewise.
16279 * config/aarch64/arm_neon.h
16280 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
16282 2013-07-02 Sriraman Tallam <tmsriram@google.com>
16284 * config/i386/i386.c (gate_insert_vzeroupper): Check if
16286 (ix86_option_override_internal):Turn on all -mavx target flags by
16287 default as they are dependent on AVX anyway.
16289 2013-07-02 Cary Coutant <ccoutant@google.com>
16291 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
16292 deterministic hash.
16293 (loc_checksum_ordered): Likewise.
16294 (hash_loc_operands): Remove inline keyword.
16296 2013-07-02 Jakub Jelinek <jakub@redhat.com>
16298 PR tree-optimization/57741
16299 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
16300 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
16301 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
16302 Allow REAL_CST step_exprs if flag_associative_math.
16303 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
16305 2013-07-02 Ian Bolton <ian.bolton@arm.com>
16307 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
16309 2013-07-02 Ian Bolton <ian.bolton@arm.com>
16311 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
16313 2013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16315 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
16317 (iorsi3_insn): Likewise.
16318 (arm_xorsi3): Likewise.
16320 2013-07-01 Sofiane Naci <sofiane.naci@arm.com>
16322 * arm.md (attribute "wtype"): Delete. Move attribute values from here
16324 (attribute "type"): ... here, and prefix with "wmmx_".
16325 (attribute "core_cycles"): Update for attribute changes.
16326 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
16327 (tbcstv4hi): Likewise.
16328 (tbcstv2si): Likewise.
16329 (iwmmxt_iordi3): Likewise.
16330 (iwmmxt_xordi3): Likewise.
16331 (iwmmxt_anddi3): Likewise.
16332 (iwmmxt_nanddi3): Likewise.
16333 (iwmmxt_arm_movdi): Likewise.
16334 (iwmmxt_movsi_insn): Likewise.
16335 (mov<mode>_internal): Likewise.
16336 (and<mode>3_iwmmxt): Likewise.
16337 (ior<mode>3_iwmmxt): Likewise.
16338 (xor<mode>3_iwmmxt): Likewise.
16339 (add<mode>3_iwmmxt): Likewise.
16340 (ssaddv8qi3): Likewise.
16341 (ssaddv4hi3): Likewise.
16342 (ssaddv2si3): Likewise.
16343 (usaddv8qi3): Likewise.
16344 (usaddv4hi3): Likewise.
16345 (usaddv2si3): Likewise.
16346 (sub<mode>3_iwmmxt): Likewise.
16347 (sssubv8qi3): Likewise.
16348 (sssubv4hi3): Likewise.
16349 (sssubv2si3): Likewise.
16350 (ussubv8qi3): Likewise.
16351 (ussubv4hi3): Likewise.
16352 (ussubv2si3): Likewise.
16353 (mulv4hi3_iwmmxt): Likewise.
16354 (smulv4hi3_highpart): Likewise.
16355 (umulv4hi3_highpart): Likewise.
16356 (iwmmxt_wmacs): Likewise.
16357 (iwmmxt_wmacsz): Likewise.
16358 (iwmmxt_wmacu): Likewise.
16359 (iwmmxt_wmacuz): Likewise.
16360 (iwmmxt_clrdi): Likewise.
16361 (iwmmxt_clrv8qi): Likewise.
16362 (iwmmxt_clr4hi): Likewise.
16363 (iwmmxt_clr2si): Likewise.
16364 (iwmmxt_uavgrndv8qi3): Likewise.
16365 (iwmmxt_uavgrndv4hi3): Likewise.
16366 (iwmmxt_uavgv8qi3): Likewise.
16367 (iwmmxt_uavgv4hi3): Likewise.
16368 (iwmmxt_tinsrb): Likewise.
16369 (iwmmxt_tinsrh): Likewise.
16370 (iwmmxt_tinsrw): Likewise.
16371 (iwmmxt_textrmub): Likewise.
16372 (iwmmxt_textrmsb): Likewise.
16373 (iwmmxt_textrmuh): Likewise.
16374 (iwmmxt_textrmsh): Likewise.
16375 (iwmmxt_textrmw): Likewise.
16376 (iwmxxt_wshufh): Likewise.
16377 (eqv8qi3): Likewise.
16378 (eqv4hi3): Likewise.
16379 (eqv2si3): Likewise.
16380 (gtuv8qi3): Likewise.
16381 (gtuv4hi3): Likewise.
16382 (gtuv2si3): Likewise.
16383 (gtv8qi3): Likewise.
16384 (gtv4hi3): Likewise.
16385 (gtv2si3): Likewise.
16386 (smax<mode>3_iwmmxt): Likewise.
16387 (umax<mode>3_iwmmxt): Likewise.
16388 (smin<mode>3_iwmmxt): Likewise.
16389 (umin<mode>3_iwmmxt): Likewise.
16390 (iwmmxt_wpackhss): Likewise.
16391 (iwmmxt_wpackwss): Likewise.
16392 (iwmmxt_wpackdss): Likewise.
16393 (iwmmxt_wpackhus): Likewise.
16394 (iwmmxt_wpackwus): Likewise.
16395 (iwmmxt_wpackdus): Likewise.
16396 (iwmmxt_wunpckihb): Likewise.
16397 (iwmmxt_wunpckihh): Likewise.
16398 (iwmmxt_wunpckihw): Likewise.
16399 (iwmmxt_wunpckilb): Likewise.
16400 (iwmmxt_wunpckilh): Likewise.
16401 (iwmmxt_wunpckilw): Likewise.
16402 (iwmmxt_wunpckehub): Likewise.
16403 (iwmmxt_wunpckehuh): Likewise.
16404 (iwmmxt_wunpckehuw): Likewise.
16405 (iwmmxt_wunpckehsb): Likewise.
16406 (iwmmxt_wunpckehsh): Likewise.
16407 (iwmmxt_wunpckehsw): Likewise.
16408 (iwmmxt_wunpckelub): Likewise.
16409 (iwmmxt_wunpckeluh): Likewise.
16410 (iwmmxt_wunpckeluw): Likewise.
16411 (iwmmxt_wunpckelsb): Likewise.
16412 (iwmmxt_wunpckelsh): Likewise.
16413 (iwmmxt_wunpckelsw): Likewise.
16414 (ror<mode>3): Likewise.
16415 (ashr<mode>3_iwmmxt): Likewise.
16416 (lshr<mode>3_iwmmxt): Likewise.
16417 (ashl<mode>3_iwmmxt): Likewise.
16418 (ror<mode>3_di): Likewise.
16419 (ashr<mode>3_di): Likewise.
16420 (lshr<mode>3_di): Likewise.
16421 (ashl<mode>3_di): Likewise.
16422 (iwmmxt_wmadds): Likewise.
16423 (iwmmxt_wmaddu): Likewise.
16424 (iwmmxt_tmia): Likewise.
16425 (iwmmxt_tmiaph): Likewise.
16426 (iwmmxt_tmiabb): Likewise.
16427 (iwmmxt_tmiatb): Likewise.
16428 (iwmmxt_tmiabt): Likewise.
16429 (iwmmxt_tmiatt): Likewise.
16430 (iwmmxt_tmovmskb): Likewise.
16431 (iwmmxt_tmovmskh): Likewise.
16432 (iwmmxt_tmovmskw): Likewise.
16433 (iwmmxt_waccb): Likewise.
16434 (iwmmxt_wacch): Likewise.
16435 (iwmmxt_waccw): Likewise.
16436 (iwmmxt_waligni): Likewise.
16437 (iwmmxt_walignr): Likewise.
16438 (iwmmxt_walignr0): Likewise.
16439 (iwmmxt_walignr1): Likewise.
16440 (iwmmxt_walignr2): Likewise.
16441 (iwmmxt_walignr3): Likewise.
16442 (iwmmxt_wsadb): Likewise.
16443 (iwmmxt_wsadh): Likewise.
16444 (iwmmxt_wsadbz): Likewise.
16445 (iwmmxt_wsadhz): Likewise.
16446 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
16447 (iwmmxt_wabsdiffb): Likewise.
16448 (iwmmxt_wabsdiffh): Likewise.
16449 (iwmmxt_wabsdiffw): Likewise.
16450 (iwmmxt_waddsubhx): Likewise
16451 (iwmmxt_wsubaddhx): Likewise.
16452 (addc<mode>3): Likewise.
16453 (iwmmxt_avg4): Likewise.
16454 (iwmmxt_avg4r): Likewise.
16455 (iwmmxt_wmaddsx): Likewise.
16456 (iwmmxt_wmaddux): Likewise.
16457 (iwmmxt_wmaddsn): Likewise.
16458 (iwmmxt_wmaddun): Likewise.
16459 (iwmmxt_wmulwsm): Likewise.
16460 (iwmmxt_wmulwum): Likewise.
16461 (iwmmxt_wmulsmr): Likewise.
16462 (iwmmxt_wmulumr): Likewise.
16463 (iwmmxt_wmulwsmr): Likewise.
16464 (iwmmxt_wmulwumr): Likewise.
16465 (iwmmxt_wmulwl): Likewise.
16466 (iwmmxt_wqmulm): Likewise.
16467 (iwmmxt_wqmulwm): Likewise.
16468 (iwmmxt_wqmulmr): Likewise.
16469 (iwmmxt_wqmulwmr): Likewise.
16470 (iwmmxt_waddbhusm): Likewise.
16471 (iwmmxt_waddbhusl): Likewise.
16472 (iwmmxt_wqmiabb): Likewise.
16473 (iwmmxt_wqmiabt): Likewise.
16474 (iwmmxt_wqmiatb): Likewise.
16475 (iwmmxt_wqmiatt): Likewise.
16476 (iwmmxt_wqmiabbn): Likewise.
16477 (iwmmxt_wqmiabtn): Likewise.
16478 (iwmmxt_wqmiatbn): Likewise.
16479 (iwmmxt_wqmiattn): Likewise.
16480 (iwmmxt_wmiabb): Likewise.
16481 (iwmmxt_wmiabt): Likewise.
16482 (iwmmxt_wmiatb): Likewise.
16483 (iwmmxt_wmiatt): Likewise.
16484 (iwmmxt_wmiabbn): Likewise.
16485 (iwmmxt_wmiabtn): Likewise.
16486 (iwmmxt_wmiatbn): Likewise.
16487 (iwmmxt_wmiattn): Likewise.
16488 (iwmmxt_wmiawbb): Likewise.
16489 (iwmmxt_wmiawbt): Likewise.
16490 (iwmmxt_wmiawtb): Likewise.
16491 (iwmmxt_wmiawtt): Likewise.
16492 (iwmmxt_wmiawbbn): Likewise.
16493 (iwmmxt_wmiawbtn): Likewise.
16494 (iwmmxt_wmiawtbn): Likewise.
16495 (iwmmxt_wmiawttn): Likewise.
16496 (iwmmxt_wmerge): Likewise.
16497 (iwmmxt_tandc<mode>3): Likewise.
16498 (iwmmxt_torc<mode>3): Likewise.
16499 (iwmmxt_torvsc<mode>3): Likewise.
16500 (iwmmxt_textrc<mode>3): Likewise.
16501 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
16502 (wmmxt_pack): Likewise.
16503 (wmmxt_mult_c1): Likewise.
16504 (wmmxt_mult_c2): Likewise.
16505 (wmmxt_alu_c1): Likewise.
16506 (wmmxt_alu_c2): Likewise.
16507 (wmmxt_alu_c3): Likewise.
16508 (wmmxt_transfer_c1): Likewise.
16509 (wmmxt_transfer_c2): Likewise.
16510 (wmmxt_transfer_c3): Likewise.
16511 (marvell_f_iwmmxt_wstr): Likewise.
16512 (marvell_f_iwmmxt_wldr): Likewise.
16514 2013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
16516 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
16518 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
16521 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
16522 * lra-constraints.c (need_for_split_p): Check call used hard regs
16523 living through calls.
16525 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
16526 call used regs for call insn.
16528 2013-06-28 Jakub Jelinek <jakub@redhat.com>
16531 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
16532 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
16534 2013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
16536 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
16538 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
16540 * lra-constraints.c (need_for_split_p): Check call used hard regs
16541 living through calls.
16543 2013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
16546 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
16547 to tie with any other modes. Eliminate Altivec vector mode tests,
16548 since these are a subset of ALTIVEC or VSX vector modes. Simplify
16549 code, to return 0 if testing MODE2 for a condition, if we've
16550 already tested MODE1 for the same condition.
16552 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
16554 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
16557 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
16559 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
16560 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
16562 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
16564 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
16566 (aarch64_symbolic_constant_p): Remove.
16567 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
16568 static. Fix line length and white space.
16569 (aarch64_symbolic_constant_p): Remove.
16570 * config/aarch64/predicates.md (aarch64_valid_symref):
16571 Use aarch64_classify_symbol_expression.
16573 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16575 * config/arm/constraints.md (Ts): New constraint.
16576 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
16578 (compare_scc): Use "Ts" constraint for operand 0.
16579 (ior_scc_scc): Likewise.
16580 (and_scc_scc): Likewise.
16581 (and_scc_scc_nodom): Likewise.
16582 (ior_scc_scc_cmp): Likewise for operand 7.
16583 (and_scc_scc_cmp): Likewise.
16584 * config/arm/thumb2.md (thumb2_movsi_insn):
16585 Add alternatives for 16-bit encodings.
16586 (thumb2_movhi_insn): Likewise.
16587 (thumb2_movsicc_insn): Likewise.
16588 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
16589 (thumb2_negscc): Use "Ts" constraint.
16590 Move mvn instruction outside cond_exec block.
16591 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
16592 for 16-bit encodings.
16594 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16596 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
16598 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
16599 (mulsi3subsi): Likewise.
16600 (mulsidi3adddi): Likewise.
16601 (mulsidi3_v6): Likewise.
16602 (umulsidi3_v6): Likewise.
16603 (umulsidi3adddi_v6): Likewise.
16604 (smulsi3_highpart_v6): Likewise.
16605 (umulsi3_highpart_v6): Likewise.
16606 (mulhisi3tb): Likewise.
16607 (mulhisi3bt): Likewise.
16608 (mulhisi3tt): Likewise.
16609 (maddhisi4): Likewise.
16610 (maddhisi4tb): Likewise.
16611 (maddhisi4tt): Likewise.
16612 (maddhidi4): Likewise.
16613 (maddhidi4tb): Likewise.
16614 (maddhidi4tt): Likewise.
16615 (zeroextractsi_compare0_scratch): Likewise.
16616 (insv_zero): Likewise.
16617 (insv_t2): Likewise.
16618 (anddi_notzesidi_di): Likewise.
16619 (anddi_notsesidi_di): Likewise.
16620 (andsi_notsi_si): Likewise.
16621 (iordi_zesidi_di): Likewise.
16622 (xordi_zesidi_di): Likewise.
16623 (andsi_iorsi3_notsi): Likewise.
16624 (smax_0): Likewise.
16625 (smax_m1): Likewise.
16626 (smin_0): Likewise.
16627 (not_shiftsi): Likewise.
16628 (unaligned_loadsi): Likewise.
16629 (unaligned_loadhis): Likewise.
16630 (unaligned_loadhiu): Likewise.
16631 (unaligned_storesi): Likewise.
16632 (unaligned_storehi): Likewise.
16633 (extv_reg): Likewise.
16634 (extzv_t2): Likewise.
16635 (divsi3): Likewise.
16636 (udivsi3): Likewise.
16637 (arm_zero_extendhisi2addsi): Likewise.
16638 (arm_zero_extendqisi2addsi): Likewise.
16639 (compareqi_eq0): Likewise.
16640 (arm_extendhisi2_v6): Likewise.
16641 (arm_extendqisi2addsi): Likewise.
16642 (arm_movt): Likewise.
16643 (thumb2_ldrd): Likewise.
16644 (thumb2_ldrd_base): Likewise.
16645 (thumb2_ldrd_base_neg): Likewise.
16646 (thumb2_strd): Likewise.
16647 (thumb2_strd_base): Likewise.
16648 (thumb2_strd_base_neg): Likewise.
16649 (arm_negsi2): Add alternative for 16-bit encoding.
16650 (arm_one_cmplsi2): Likewise.
16652 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16654 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
16655 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
16656 (movdfcc): Likewise.
16657 * config/arm/vfp.md (*thumb2_movsf_vfp):
16658 Disable predication for arm_restrict_it.
16659 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
16660 (*thumb2_movdfcc_vfp): Likewise.
16661 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
16662 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
16663 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
16664 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
16665 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
16666 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
16667 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
16668 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
16669 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
16670 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
16671 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
16672 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
16673 Disable predication for arm_restrict_it.
16675 2013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
16677 * config/i386/bmiintrin.h (_bextr_u32): New.
16678 (_bextr_u64): Ditto.
16680 2013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
16682 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
16683 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
16684 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
16685 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
16686 * config/mips/n32-elf.h: ...this new file.
16688 2013-06-27 Marc Glisse <marc.glisse@inria.fr>
16691 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
16692 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
16694 2013-06-27 Catherine Moore <clm@codesourcery.com>
16696 * config/mips/mips-tables.opt: Regenerate.
16697 * config/mips/mips-cpus.def: Add m14ke and m14kec.
16698 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
16699 * doc/invoke.texi: Add -m14kc.
16701 2013-06-27 Jakub Jelinek <jakub@redhat.com>
16704 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
16705 constraints of operand 1 and 2.
16708 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
16709 to match RTL canonicalization. Swap predicates and
16710 constraints of operand 1 and 2.
16712 2013-06-27 Vladimir Makarov <vmakarov@redhat.com>
16714 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
16715 Process OP_INOUT regs for splitting too.
16717 2013-06-27 Jakub Jelinek <jakub@redhat.com>
16719 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
16720 decl before the loop, initialize to NULL.
16721 (vectorizable_load): Initialize ptr_incr to NULL.
16723 2013-06-27 Martin Jambor <mjambor@suse.cz>
16726 * ipa-ref.h (ipa_maybe_record_reference): Declare.
16727 * ipa-ref.c (ipa_maybe_record_reference): New function.
16728 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
16729 * ipa-cp.c (create_specialized_node): Record potential references from
16731 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
16733 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
16735 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
16736 parameter 'mode' of type 'enum machine_mode mode'; change to pass
16737 'mode' to force_reg.
16738 (aarch64_add_offset): Update calls to aarch64_force_temporary.
16739 (aarch64_expand_mov_immediate): Likewise.
16741 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
16743 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
16744 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
16746 2013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16748 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
16749 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
16750 (struct machine_function): Add tbegin_p.
16751 (s390_canonicalize_comparison): Fold CC mode compares to
16752 conditional jump if possible.
16753 (s390_emit_jump): Return the emitted jump.
16754 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
16755 Handle CCRAWmode compares.
16756 (s390_option_override): Default to -mhtm if available.
16757 (s390_reg_clobbered_rtx): Handle floating point regs as well.
16758 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
16759 FPRs instead of df_regs_ever_live_p.
16760 (s390_optimize_nonescaping_tx): New function.
16761 (s390_init_frame_layout): Extend clobbered_regs array to cover
16763 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
16764 (s390_expand_tbegin): New function.
16765 (enum s390_builtin): New enum definition.
16766 (code_for_builtin): New array definition.
16767 (s390_init_builtins): New function.
16768 (s390_expand_builtin): New function.
16769 (TARGET_INIT_BUILTINS): Define.
16770 (TARGET_EXPAND_BUILTIN): Define.
16771 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
16772 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
16773 (s390_alc_comparison): Likewise.
16774 * config/s390/s390-modes.def: Add CCRAWmode.
16775 * config/s390/s390.h (processor_flags): Add PF_TX.
16776 (TARGET_CPU_HTM): Define macro.
16777 (TARGET_HTM): Define macro.
16778 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
16779 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
16780 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
16781 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
16782 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
16783 (TBEGIN_MASK, TBEGINC_MASK): New constants.
16784 ("*cc_to_int"): Move up.
16785 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
16786 constants other than 0.
16787 ("*ccraw_to_int"): New insn and splitter definition.
16788 ("tbegin", "tbegin_nofloat", "tbegin_retry")
16789 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
16790 ("tx_assist"): New expander.
16791 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
16792 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
16793 * config/s390/s390.opt: Add -mhtm option.
16794 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
16795 * config/s390/htmxlintrin.h: New file.
16796 * config/s390/htmintrin.h: New file.
16797 * config/s390/s390intrin.h: New file.
16798 * doc/extend.texi: Document htm builtins.
16799 * config.gcc: Add the new header files to extra_headers.
16801 2013-06-26 Thomas Schwinge <thomas@codesourcery.com>
16803 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
16804 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
16806 2013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
16807 Pat Haugen <pthaugen@us.ibm.com>
16808 Peter Bergner <bergner@vnet.ibm.com>
16810 * config/rs6000/power8.md: New.
16811 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
16812 setting for power8 entry.
16813 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
16814 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
16815 test for Power4/Power5 only.
16816 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
16818 (force_new_group): Adjust comment.
16819 * config/rs6000/rs6000.md: Include power8.md.
16821 2013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
16823 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
16824 * config/arm/arm-protos.h (arm_max_conditional_execute): New
16826 (tune_params): Update comment.
16827 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
16828 (arm_max_conditional_execute): New function.
16829 (thumb2_final_prescan_insn): Use max_insn_skipped and
16830 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
16832 2013-06-25 Jakub Jelinek <jakub@redhat.com>
16834 PR tree-optimization/57705
16835 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
16836 SSA_NAME step, provided that it is not defined inside the loop.
16837 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
16838 (get_initial_def_for_induction): Handle SSA_NAME IV step.
16840 2013-06-25 Martin Jambor <mjambor@suse.cz>
16842 PR middle-end/57670
16843 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
16844 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
16846 (ipa_note_param_call): Initialize member_ptr flag.
16847 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
16848 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
16849 (ipa_write_indirect_edge_info): Stream member_ptr flag.
16850 (ipa_read_indirect_edge_info): Likewise.
16852 2013-06-25 Richard Biener <rguenther@suse.de>
16854 PR middle-end/56977
16855 * passes.c (init_optimization_passes): Move pass_fold_builtins
16856 and pass_dce earlier with -Og.
16858 2013-06-25 Eric Botcazou <ebotcazou@adacore.com>
16860 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
16861 <BIT_FIELD_REF>: Remove trailing TAB.
16862 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
16865 2013-06-24 Martin Jambor <mjambor@suse.cz>
16867 PR tree-optimization/57358
16868 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
16869 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
16870 (ipa_analyze_params_uses): Generate pessimistic info when true.
16872 2013-06-24 Martin Jambor <mjambor@suse.cz>
16874 PR tree-optimization/57539
16875 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
16876 global.inlined_to of the new node to it. All callers changed.
16877 * ipa-inline-transform.c (clone_inlined_nodes): New variable
16878 inlining_into, pass it to cgraph_clone_node.
16879 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
16880 ipa_free_edge_args_substructures.
16881 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
16882 rdesc linked list. Do not assert rdesc edges have inlined caller.
16883 Assert we have found an rdesc in the rdesc list.
16885 2013-06-24 Richard Biener <rguenther@suse.de>
16887 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
16888 (pointer_set_lookup): Declare.
16889 (class pointer_map): New template class implementing a
16890 generic pointer to T map.
16891 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
16892 pointer_map<T>::contains, pointer_map<T>::insert,
16893 pointer_map<T>::traverse): New functions.
16894 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
16895 (pointer_set_lookup): New function.
16896 (pointer_set_contains): Use pointer_set_lookup.
16897 (pointer_set_insert): Likewise.
16898 (insert_aux): Remove.
16899 (struct pointer_map_t): Embed a pointer_set_t.
16900 (pointer_map_create): Adjust.
16901 (pointer_map_destroy): Likewise.
16902 (pointer_map_contains): Likewise.
16903 (pointer_map_insert): Likewise.
16904 (pointer_map_traverse): Likewise.
16905 * tree-streamer.h (struct streamer_tree_cache_d): Use a
16906 pointer_map<unsigned> instead of a pointer_map_t.
16907 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
16908 (streamer_tree_cache_lookup): Likewise.
16909 (streamer_tree_cache_create): Likewise.
16910 (streamer_tree_cache_delete): Likewise.
16911 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
16912 pointer_map<unsigned> instead of a pointer_map_t.
16913 (lto_init_tree_ref_encoder): Adjust.
16914 (lto_destroy_tree_ref_encoder): Likewise.
16915 * lto-section-out.c (lto_output_decl_index): Likewise.
16916 (lto_record_function_out_decl_state): Likewise.
16917 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
16919 2013-06-24 Richard Biener <rguenther@suse.de>
16921 PR tree-optimization/57488
16922 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
16924 2013-06-24 Alan Modra <amodra@gmail.com>
16926 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
16927 (gen_easy_altivec_constant): Likewise.
16928 * config/rs6000/predicates.md (easy_vector_constant_add_self,
16929 easy_vector_constant_msb): Likewise.
16931 2013-06-23 Jakub Jelinek <jakub@redhat.com>
16934 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
16935 add missing return true.
16937 2013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
16940 * config/sh/predicates.md (general_extend_operand): Invoke
16941 general_movsrc_operand for memory operands.
16942 (general_movsrc_operand): Allow reg+reg addressing, do not use
16943 general_operand for memory operands.
16945 2013-06-23 Sriraman Tallam <tmsriram@google.com>
16947 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
16948 when current target options does not apply.
16949 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
16950 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
16951 * config/i386/bmiintrin.h: Pass appropriate target
16952 attributes to header.
16953 * config/i386/mmintrin.h: Ditto.
16954 * config/i386/nmmintrin.h: Ditto.
16955 * config/i386/avx2intrin.h: Ditto.
16956 * config/i386/fxsrintrin.h: Ditto.
16957 * config/i386/tbmintrin.h: Ditto.
16958 * config/i386/xsaveintrin.h: Ditto.
16959 * config/i386/f16cintrin.h: Ditto.
16960 * config/i386/xtestintrin.h: Ditto.
16961 * config/i386/xsaveoptintrin.h: Ditto.
16962 * config/i386/bmi2intrin.h: Ditto.
16963 * config/i386/lzcntintrin.h: Ditto.
16964 * config/i386/smmintrin.h: Ditto.
16965 * config/i386/wmmintrin.h: Ditto.
16966 * config/i386/x86intrin.h: Remove all header include guards.
16967 * config/i386/prfchwintrin.h: Ditto.
16968 * config/i386/pmmintrin.h: Ditto.
16969 * config/i386/tmmintrin.h: Ditto.
16970 * config/i386/xmmintrin.h: Ditto.
16971 * config/i386/popcntintrin.h: Ditto.
16972 * config/i386/rdseedintrin.h: Ditto.
16973 * config/i386/ammintrin.h: Ditto.
16974 * config/i386/emmintrin.h: Ditto.
16975 * config/i386/immintrin.h: Remove all header include guards.
16976 * config/i386/fma4intrin.h: Ditto.
16977 * config/i386/lwpintrin.h: Ditto.
16978 * config/i386/xopintrin.h: Ditto.
16979 * config/i386/ia32intrin.h: Ditto.
16980 * config/i386/avxintrin.h: Ditto.
16981 * config/i386/rtmintrin.h: Ditto.
16982 * config/i386/fmaintrin.h: Ditto.
16983 * config/i386/mm3dnow.h: Ditto.
16985 2013-06-22 Sriraman Tallam <tmsriram@google.com>
16987 * common/config/i386/i386-common.c: Handle LZCNT.
16989 2013-06-22 Andi Kleen <ak@linux.intel.com>
16991 * doc/extend.texi: Use __atomic_store_n instead of
16992 __atomic_store in HLE example.
16994 2013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
16996 * config/sh/sh.c: Remove <cstdlib> workaround.
16998 2013-06-21 Andi Kleen <ak@linux.intel.com>
17000 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
17002 2013-06-21 Andi Kleen <ak@linux.intel.com>
17004 * doc/extend.texi: Document that __atomic_clear and
17005 __atomic_test_and_set should only be used with bool.
17007 2013-06-20 Jan Hubicka <jh@suse.cz>
17009 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
17010 types_same_for_odr.
17011 * tree.c (decls_same_for_odr): New function.
17012 (same_for_edr): New function.
17013 (types_same_for_odr): New function.
17014 (get_binfo_at_offset): Use it.
17015 * tree.h (types_same_for_odr): Declare.
17017 2013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
17018 Jason Merrill <jason@redhat.com>
17020 * system.h: Include <cstdlib> as well as <stdlib.h>.
17022 2013-06-20 Uros Bizjak <ubizjak@gmail.com>
17025 * config/i386/i386.c (construct_container): Report error if
17026 long double is used with disabled x87 float returns.
17028 2013-06-20 Jan Hubicka <jh@suse.cz>
17030 * lto-cgraph.c (input_symtab): Do not set cgraph state.
17032 2013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
17034 PR rtl-optimization/57425
17035 PR rtl-optimization/57569
17036 * alias.c (write_dependence_p): Remove parameters mem_mode and
17037 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
17038 Changed all callers.
17039 (canon_anti_dependence): Get comments and semantics in sync.
17040 Add parameter mem_canonicalized. Changed all callers.
17041 * rtl.h (canon_anti_dependence): Update prototype.
17043 2013-06-20 Richard Biener <rguenther@suse.de>
17045 * data-streamer-in.c (streamer_read_uhwi): Optimize single
17046 byte case, inline streamer_read_uchar and defer section
17049 2013-06-20 Richard Biener <rguenther@suse.de>
17051 PR tree-optimization/57584
17052 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
17053 SSA names into the expanded expression that take part in
17054 abnormal coalescing.
17056 2013-06-19 Sharad Singhai <singhai@google.com>
17058 * gcov.c (print_usage): Handle new option.
17059 (process_args): Ditto.
17060 (get_gcov_intermediate_filename): New function.
17061 (output_intermediate_file): New function.
17062 (output_gcov_file): New function
17063 (generate_results): Handle new option.
17064 (release_function): Relase demangled name.
17065 (read_graph_file): Handle demangled name.
17066 (output_lines): Ditto.
17067 * doc/gcov.texi: Document gcov intermediate format.
17069 2013-06-19 Vladimir Makarov <vmakarov@redhat.com>
17072 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
17073 (lra_emit_add): Use the functions. Add comment about Y as an
17076 2013-06-19 David Edelsohn <dje.gcc@gmail.com>
17079 * collect2.c (collect_atexit): New.
17080 (collect_exit): Delete.
17081 (main): Register collect_atexit with atexit.
17082 (collect_wait): Change collect_exit to exit.
17084 * collect2.h (collect_exit): Delete.
17085 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
17087 2013-06-19 Wei Mi <wmi@google.com>
17089 PR rtl-optimization/57518
17090 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
17091 if regno is used in paradoxical subreg.
17092 (update_equiv_regs): Check pdx_subregs[regno] before
17093 set a reg to be equivalent with a mem.
17095 2013-06-19 Matthias Klose <doko@ubuntu.com>
17098 * file-find.h (find_a_file): Add a mode parameter.
17099 * file-find.c (find_a_file): Likewise.
17100 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
17101 with X_OK for the executables.
17102 * collect2.c (main): Call find_a_file with X_OK.
17104 2013-06-19 Steve Ellcey <sellcey@mips.com>
17107 * config/mips/mips.md (casesi_internal_mips16_<mode>):
17108 Use NEXT_INSN instead of next_real_insn.
17110 2013-06-19 Jan Hubicka <jh@suse.cz>
17112 * cgraph.h (const_value_known_p): Replace by ...
17113 (ctor_for_folding): .. this one.
17114 * cgraphunit.c (process_function_and_variable_attributes): Use it.
17115 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
17116 * expr.c (expand_expr_real_1): Likewise.
17117 (string_constant): Likewise.
17118 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
17119 * ipa.c (process_references): Likewise.
17120 (symtab_remove_unreachable_nodes): Likewise.
17121 * ipa-inline-analysis.c (param_change_prob): Likewise.
17122 * gimple-fold.c (canonicalize_constructor_val): Likewise.
17123 (get_base_constructor): Likwise.
17124 * varpool.c (varpool_remove_node): Likewise.
17125 (varpool_remove_initializer): LIkewise.
17126 (dump_varpool_node): LIkwise.
17127 (const_value_known_p): Rewrite to ...
17128 (ctor_for_folding): ... this one.
17130 2013-06-19 Jakub Jelinek <jakub@redhat.com>
17133 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
17134 PERSONALITY in $PATH derived prefixes.
17136 2013-06-19 Jeff Law <law@redhat.com>
17138 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
17141 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
17142 (simplify_bitwise_binary): Use it to simpify certain binary ops on
17145 2013-06-19 Sofiane Naci <sofiane.naci@arm.com>
17147 * config/arm/vfp.md: Move VFP instruction classification documentation
17149 * config/arm/arm.md: ... here. Update instruction classification
17152 2013-06-19 Richard Earnshaw <rearnsha@arm.com>
17154 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
17155 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
17156 pattern. Use more efficient sequences on ARMv5 and Thumb2.
17158 2013-06-19 Steven Bosscher <steven@gcc.gnu.org>
17161 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
17162 with NEXT_INSN. Use tablejump_p to check for jump table data
17165 2013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
17168 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
17169 that now in C++ the value is correct per the C++ standards.
17171 2013-06-19 Richard Biener <rguenther@suse.de>
17173 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
17174 for global context.
17176 2013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17179 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17182 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
17183 with next_active_insn.
17185 2013-06-18 Sriraman Tallam <tmsriram@google.com>
17187 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
17188 functions are inlined during failures to flag an error.
17189 * tree-inline.c (expand_call_inline): Allow the error to be flagged
17190 in early inline pass.
17192 2013-06-18 H.J. Lu <hongjiu.lu@intel.com>
17194 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
17197 2013-06-18 Julian Brown <julian@codesourcery.com>
17199 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
17200 Permit virtual register pre-reload if !strict.
17201 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
17203 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
17205 * config/arm/neon.md (movmisalign<mode>): Use
17206 neon_perm_struct_or_reg_operand instead of
17207 neon_struct_or_register_operand.
17208 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
17209 neon_permissive_struct_operand instead of neon_struct_operand.
17210 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
17211 neon_vector_mem_operand.
17212 * config/arm/predicates.md (neon_struct_operand): Adjust call to
17213 neon_vector_mem_operand.
17214 (neon_permissive_struct_operand): New.
17215 (neon_struct_or_register_operand): Rename to...
17216 (neon_perm_struct_or_reg_operand): This. Adjust call to
17217 neon_vector_mem_operand.
17219 2013-06-18 Richard Biener <rguenther@suse.de>
17221 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
17222 * lto-streamer.h: Include pointer-set.h.
17223 (struct lto_decl_slot): Remove.
17224 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
17225 Remove next_index entry.
17226 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
17227 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
17228 (lto_init_tree_ref_encoder): Adjust.
17229 (lto_destroy_tree_ref_encoder): Likewise.
17230 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
17231 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
17232 (lto_output_decl_index): Adjust.
17233 (lto_new_out_decl_state): Likewise.
17234 (lto_record_function_out_decl_state): Likewise.
17235 * lto-streamer-out.c (copy_function): Likewise.
17237 2013-06-18 Richard Biener <rguenther@suse.de>
17239 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
17240 * cgraphunit.c: Include cfgloop.h.
17241 (init_lowered_empty_function): Initialize the loop tree.
17242 (assemble_thunk): Insert new BBs into loops.
17244 2013-06-18 Richard Biener <rguenther@suse.de>
17246 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
17247 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
17248 the map from cache entry to cache index optional.
17249 (streamer_tree_cache_replace_tree): Adjust accordingly.
17250 (streamer_tree_cache_append): Likewise.
17251 (streamer_tree_cache_delete): Likewise.
17252 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
17253 streamer cache map from cache entry to cache index.
17254 * lto-streamer-out.c (create_output_block): Adjust.
17256 2013-06-18 Sofiane Naci <sofiane.naci@arm.com>
17258 * config/arm/arm.md (attribute "insn"): Move multiplication and
17259 division attributes to...
17260 (attribute "type"): ... here. Remove mult.
17261 (attribute "mul32"): New attribute.
17262 (attribute "mul64"): Add umaal.
17263 (*arm_mulsi3): Update attributes.
17264 (*arm_mulsi3_v6): Likewise.
17265 (*thumb_mulsi3): Likewise.
17266 (*thumb_mulsi3_v6): Likewise.
17267 (*mulsi3_compare0): Likewise.
17268 (*mulsi3_compare0_v6): Likewise.
17269 (*mulsi_compare0_scratch): Likewise.
17270 (*mulsi_compare0_scratch_v6): Likewise.
17271 (*mulsi3addsi): Likewise.
17272 (*mulsi3addsi_v6): Likewise.
17273 (*mulsi3addsi_compare0): Likewise.
17274 (*mulsi3addsi_compare0_v6): Likewise.
17275 (*mulsi3addsi_compare0_scratch): Likewise.
17276 (*mulsi3addsi_compare0_scratch_v6): Likewise.
17277 (*mulsi3subsi): Likewise.
17278 (*mulsidi3adddi): Likewise.
17279 (*mulsi3addsi_v6): Likewise.
17280 (*mulsidi3adddi_v6): Likewise.
17281 (*mulsidi3_nov6): Likewise.
17282 (*mulsidi3_v6): Likewise.
17283 (*umulsidi3_nov6): Likewise.
17284 (*umulsidi3_v6): Likewise.
17285 (*umulsidi3adddi): Likewise.
17286 (*umulsidi3adddi_v6): Likewise.
17287 (*smulsi3_highpart_nov6): Likewise.
17288 (*smulsi3_highpart_v6): Likewise.
17289 (*umulsi3_highpart_nov6): Likewise.
17290 (*umulsi3_highpart_v6): Likewise.
17291 (mulhisi3): Likewise.
17292 (*mulhisi3tb): Likewise.
17293 (*mulhisi3bt): Likewise.
17294 (*mulhisi3tt): Likewise.
17295 (maddhisi4): Likewise.
17296 (*maddhisi4tb): Likewise.
17297 (*maddhisi4tt): Likewise.
17298 (maddhidi4): Likewise.
17299 (*maddhidi4tb): Likewise.
17300 (*maddhidi4tt): Likewise.
17301 (divsi3): Likewise.
17302 (udivsi3): Likewise.
17303 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
17304 (thumb2_mulsi_short_compare0): Likewise.
17305 (thumb2_mulsi_short_compare0_scratch): Likewise.
17306 * config/arm/arm1020e.md (1020mult1): Update attribute change.
17307 (1020mult2): Likewise.
17308 (1020mult3): Likewise.
17309 (1020mult4): Likewise.
17310 (1020mult5): Likewise.
17311 (1020mult6): Likewise.
17312 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
17314 (cortex_a15_mult64): Likewise.
17315 (cortex_a15_sdiv): Likewise.
17316 (cortex_a15_udiv): Likewise.
17317 * config/arm/arm1026ejs.md (mult1): Update attribute change.
17323 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
17324 (pj4_ir_div): Likewise.
17325 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
17326 (11_mult2): Likewise.
17327 (11_mult3): Likewise.
17328 (11_mult4): Likewise.
17329 (11_mult5): Likewise.
17330 (11_mult6): Likewise.
17331 (11_mult7): Likewise.
17332 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
17333 (cortex_a8_mla): Likewise.
17334 (cortex_a8_mull): Likewise.
17335 (cortex_a8_smulwy): Likewise.
17336 (cortex_a8_smlald): Likewise.
17337 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
17338 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
17339 (cortex_r4_mul_3): Likewise.
17340 (cortex_r4_mla_4): Likewise.
17341 (cortex_r4_mla_3): Likewise.
17342 (cortex_r4_smlald): Likewise.
17343 (cortex_r4_mull): Likewise.
17344 (cortex_r4_sdiv): Likewise.
17345 (cortex_r4_udiv): Likewise.
17346 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
17347 (cortex_a7_idiv): Likewise.
17348 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
17349 (9_mult2): Likewise.
17350 (9_mult3): Likewise.
17351 (9_mult4): Likewise.
17352 (9_mult5): Likewise.
17353 (9_mult6): Likewise.
17354 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
17355 (cortex_a53_sdiv): Likewise.
17356 (cortex_a53_udiv): Likewise.
17357 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
17358 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
17359 (mp626_mult2): Likewise.
17360 (mp626_mult3): Likewise.
17361 (mp626_mult4): Likewise.
17362 * config/arm/fa526.md (526_mult1): Update attribute change.
17363 (526_mult2): Likewise.
17364 * config/arm/arm-generic.md (mult): Update attribute change.
17365 (mult_ldsched_strongarm): Likewise.
17366 (mult_ldsched): Likewise.
17367 (multi_cycle): Likewise.
17368 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
17369 * config/arm/fa606te.md (606te_mult1): Update attribute change.
17370 (606te_mult2): Likewise.
17371 (606te_mult3): Likewise.
17372 (606te_mult4): Likewise.
17373 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
17374 (cortex_a9_mac16): Likewise.
17375 (cortex_a9_multiply): Likewise.
17376 (cortex_a9_mac): Likewise.
17377 (cortex_a9_multiply_long): Likewise.
17378 * config/arm/fa626te.md (626te_mult1): Update attribute change.
17379 (626te_mult2): Likewise.
17380 (626te_mult3): Likewise.
17381 (626te_mult4): Likewise.
17383 2013-06-18 Richard Biener <rguenther@suse.de>
17386 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
17388 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17391 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
17392 with next_active_insn.
17394 2013-06-18 Alan Modra <amodra@gmail.com>
17396 * config/rs6000/rs6000.h (enum data_align): New.
17397 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
17398 (DATA_ABI_ALIGNMENT): Define.
17399 (CONSTANT_ALIGNMENT): Correct comment.
17400 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
17401 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
17403 2013-06-17 David Malcolm <dmalcolm@redhat.com>
17405 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
17406 ATTRIBUTE_UNUSED marking.
17408 2013-06-17 Sofiane Naci <sofiane.naci@arm.com>
17410 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
17411 alternative and update.
17412 (aarch64_dup_lanedi): Delete.
17413 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
17414 * config/aarch64/aarch64-simd-builtins.def: Update.
17416 2013-06-17 Richard Biener <rguenther@suse.de>
17418 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
17419 (lto_input_scc): Declare.
17420 (lto_input_tree_1): Likewise.
17421 (struct lto_stats_d): Add num_tree_bodies_output and
17422 num_pickle_refs_output.
17423 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
17424 (lto_read_tree_1): Split out from ...
17425 (lto_read_tree): ... this.
17426 (lto_input_scc): New function.
17427 (lto_input_tree_1): Split out from ...
17428 (lto_input_tree): ... this. Handle LTO_tree_scc.
17429 (lto_data_in_create): Create the streamer cache without hashes.
17430 * lto-streamer-out.c (create_output_block): Create the streamer
17431 cache with hashes when not doing WPA.
17432 (lto_write_tree_1): Split out from ...
17433 (lto_write_tree): ... this.
17434 (get_symbol_initial_value): New function.
17435 (lto_output_tree_1): Split out from ...
17436 (lto_output_tree): ... this. Write trees as series of SCCs
17437 using a DFS walk via DFS_write_tree.
17438 (struct sccs, struct scc_entry): New types.
17439 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
17440 (DFS_write_tree_body): New function.
17441 (DFS_write_tree): Likewise.
17442 (hash_tree): Likewise.
17443 (scc_entry_compare): Likewise.
17444 (hash_scc): Likewise.
17445 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
17446 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
17447 TREE_CHAIN as regular reference.
17448 (streamer_read_integer_cst): Remove.
17449 (streamer_get_pickled_tree): Adjust.
17450 * tree-streamer-out.c (streamer_write_chain): Disable streaming
17451 of DECL_EXTERNALs in BLOCK_VARS for now.
17452 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
17454 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
17455 Add hash value argument and record that if hashes are recorded
17457 (streamer_tree_cache_insert_1): Adjust.
17458 (streamer_tree_cache_insert): Likewise.
17459 (streamer_tree_cache_insert_at): Rename to ...
17460 (streamer_tree_cache_replace_tree): ... this and adjust.
17461 (streamer_tree_cache_append): Adjust.
17462 (record_common_node): Likewise.
17463 (streamer_tree_cache_create): Add argument whether to
17464 record hash values together with trees.
17465 (streamer_tree_cache_delete): Adjust.
17466 * tree-streamer.h (struct streamer_tree_cache_d): Add
17468 (streamer_read_integer_cst): Remove.
17469 (streamer_tree_cache_insert): Adjust.
17470 (streamer_tree_cache_append): Likewise.
17471 (streamer_tree_cache_insert_at): Rename to ...
17472 (streamer_tree_cache_replace_tree): ... this and adjust.
17473 (streamer_tree_cache_create): Add argument whether to record hashes.
17474 (streamer_tree_cache_get): Rename to ...
17475 (streamer_tree_cache_get_tree): ... this.
17476 (streamer_tree_cache_get_hash): New function.
17477 * tree.c (cache_integer_cst): New function.
17478 * tree.h (cache_integer_cst): Declare.
17479 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
17480 * lto-symtab.c (lto_varpool_replace_node): Only release
17481 DECL_INITIAL of non-prevailing decls.
17482 * varpool.c (varpool_remove_initializer): Do not release
17483 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
17485 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
17487 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
17488 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
17489 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
17490 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
17491 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
17492 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
17493 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
17494 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
17495 instead of TARGET_64BIT.
17496 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
17497 Require ISA_HAS_<D>DIV.
17499 2013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
17501 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
17502 (mips*-*-linux*): Move default with_llsc setting to where other
17504 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
17506 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
17507 Likewise. Remove default with_tune setting. Move default float
17508 setting to its own block. Handle with_llsc in the same block as above.
17510 2013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
17512 PR rtl-optimization/57425
17513 PR rtl-optimization/57569
17514 * alias.c (write_dependence_p): Add new parameters mem_mode,
17515 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
17516 Changed all callers.
17517 (canon_anti_dependence): New function.
17518 * cse.c (check_dependence): Use canon_anti_dependence.
17519 * cselib.c (cselib_invalidate_mem): Likewise.
17520 * rtl.h (canon_anti_dependence): Declare.
17522 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
17524 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
17525 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
17526 ".set mips3" for 64-bit targets.
17528 2013-06-15 Dehao Chen <dehao@google.com>
17530 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
17531 * gimple-low.c (gimple_check_call_matching_types): Likewise.
17532 (gimple_check_call_args): Likewise.
17533 * value-prof.c (check_ic_target): Likewise.
17534 * ipa-inline.c (early_inliner): Likewise.
17535 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
17536 * cgraph.c (cgraph_create_edge_1): Likewise.
17537 (cgraph_make_edge_direct): Likewise.
17539 2013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
17542 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
17543 rs6000_output_move_128bit to handle emitting quad memory
17544 operations. Set attribute length to 8 bytes.
17546 2013-06-14 Vidya Praveen <vidyapraveen@arm.com>
17548 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
17550 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
17551 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
17552 (aarch64_<su>mlsl<mode>): Likewise.
17554 2013-06-14 Mike Stump <mikestump@comcast.net>
17556 * Makefile.in (TARGET_H): Add insn-codes.h.
17558 2013-06-14 Alan Modra <amodra@gmail.com>
17560 PR middle-end/57134
17561 PR middle-end/57586
17562 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
17563 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
17564 bitfield expansion when EXPAND_MEMORY.
17565 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
17567 2013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
17569 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
17570 test for clearing quad memory on 32-bit later.
17572 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
17574 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
17575 (fold_negate_expr): Likewise.
17576 (fold_real_zero_addition_p): Handle vectors.
17577 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
17579 2013-06-14 Alan Modra <amodra@gmail.com>
17581 * varasm.c (force_const_mem): Revert 2013-06-07 change.
17583 2013-06-13 Jan Hubicka <jh@suse.cz>
17585 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
17586 Local comdats are not externally visible.
17587 * symtab.c (dump_symtab_base): Dump externally visible.
17588 (verify_symtab_base): Verify back links in the symtab hash.
17590 2013-06-13 Bin Cheng <bin.cheng@arm.com>
17592 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
17593 CONVERT_EXPR as equal nodes.
17595 2013-06-13 Bin Cheng <bin.cheng@arm.com>
17597 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
17599 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
17601 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
17602 Generalize to complex and vector.
17603 * tree.c (build_all_ones_cst): New function.
17604 * tree.h (build_all_ones_cst): Declare it.
17606 2013-06-13 Alan Modra <amodra@gmail.com>
17608 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
17609 * config/rs6000/rs6000.md (signbittf2): New insn.
17610 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
17611 (abstf2_internal, cmptf_internal2): Likewise.
17612 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
17614 2013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
17615 Pat Haugen <pthaugen@us.ibm.com>
17616 Peter Bergner <bergner@vnet.ibm.com>
17618 * config/rs6000/rs6000.c (emit_load_locked): Add support for
17619 power8 byte, half-word, and quad-word atomic instructions.
17620 (emit_store_conditional): Likewise.
17621 (rs6000_expand_atomic_compare_and_swap): Likewise.
17622 (rs6000_expand_atomic_op): Likewise.
17624 * config/rs6000/sync.md (larx): Add new modes for power8.
17626 (AINT): New mode iterator to include TImode as well as normal
17627 integer modes on power8.
17628 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
17629 that VSX registers are not considered. Use AINT mode iterator
17630 instead of INT1 to allow inclusion of quad word atomic operations
17632 (load_locked<mode>): Likewise.
17633 (store_conditional<mode>): Likewise.
17634 (atomic_compare_and_swap<mode>): Likewise.
17635 (atomic_exchange<mode>): Likewise.
17636 (atomic_nand<mode>): Likewise.
17637 (atomic_fetch_<fetchop_name><mode>): Likewise.
17638 (atomic_nand_fetch<mode>): Likewise.
17639 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
17641 (ATOMIC): On power8, add QImode, HImode modes.
17642 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
17643 modes that promote to SImode.
17644 (load_lockedti): Convert TImode arguments to PTImode, so that we
17645 get a guaranteed even/odd register pair.
17646 (load_lockedpti): Likewise.
17647 (store_conditionalti): Likewise.
17648 (store_conditionalpti): Likewise.
17650 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
17651 atomic load/store instructions.
17654 2013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
17656 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
17658 (insn_count): New attribute, with most cases extracted from...
17659 (length): ...here. Redefine most cases in terms of insn_count.
17660 (single_insn): Delete.
17661 (can_delay): Use insn_count to check for single instructions.
17662 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
17663 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
17664 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
17665 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
17666 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
17667 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
17668 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
17669 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
17670 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
17671 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
17672 rather than "length".
17673 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
17674 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
17675 Use "insn_count" rather than "length".
17676 * config/mips/mips-dsp.md
17677 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
17678 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
17681 2013-06-12 Marc Glisse <marc.glisse@inria.fr>
17683 PR tree-optimization/57361
17684 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
17686 2013-06-12 Sofiane Naci <sofiane.naci@arm.com>
17688 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
17690 (aarch64_simd_combine<mode>): New instruction expansion.
17691 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
17692 function prototype.
17693 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
17694 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
17696 2013-06-12 Jan Hubicka <jh@suse.cz>
17698 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
17699 decl has when in streaming stage.
17700 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
17701 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
17703 2013-06-12 Roland Stigge <stigge@antcom.de>
17706 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
17708 2013-06-12 Jakub Jelinek <jakub@redhat.com>
17710 PR tree-optimization/57537
17711 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
17712 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
17714 2013-06-12 Richard Biener <rguenther@suse.de>
17716 * data-streamer.h (streamer_write_char_stream): CSE
17717 obs->current_pointer.
17718 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
17719 streamer_write_char_stream manually and optimize the resulting loop.
17720 (streamer_write_hwi_stream): Likewise.
17722 2013-06-12 Jan Hubicka <jh@suse.cz>
17724 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
17725 * cgraph.h (varpool_create_empty_node): Declare.
17726 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
17728 * symtab.c (symtab_unregister_node): Be lax about missin entries
17730 (symtab_get_node): Update comment.
17731 * varpool.c (varpool_create_empty_node): Break out from ...
17732 (varpool_node_for_decl): ... here.
17733 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
17735 2013-06-12 Eric Botcazou <ebotcazou@adacore.com>
17737 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
17738 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
17739 part. Use straight-line flow at the end.
17740 <COMPONENT_REF>: Remove superfluous else.
17741 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
17743 2013-06-12 Jakub Jelinek <jakub@redhat.com>
17746 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
17747 target hook even for !TREE_PUBLIC decls. If no resolution info
17748 is available, return false for common and external decls.
17750 2013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
17752 * config/rl78/constraints.md (U): New constraint.
17753 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
17756 2013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
17759 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
17760 to allow returning address to AT_PLATFORM name.
17762 2013-06-11 Jan Hubicka <jh@suse.cz>
17764 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
17765 * cgraph.h (symtab_node_base): Add weakref flag.
17766 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
17767 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
17768 (output_weakrefs): Use weakref flag.
17769 * fold-const.c (simple_operand_p): Handle WEAK.
17770 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
17771 * ipa.c (varpool_externally_visible_p): Drop weakref.
17772 (function_and_variable_visibility): Update comment; fix weakref
17773 sanity checks; do not clear DECL_WEAK on them.
17774 * lto-cgraph.c (lto_output_node): update.
17775 (lto_output_varpool_node): Update.
17776 (input_overwrite_node): Update.
17777 (input_node): Update.
17778 (input_varpool_node): Update.
17779 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
17780 (lto_symtab_merge_symbols): Add sanity check.
17781 (lto_symtab_prevailing_decl): Do not special case weakrefs.
17782 * passes.c (rest_of_decl_compilation): Set static flag, too.
17783 * symtab.c (dump_symtab_base): Dump weakref.
17784 (verify_symtab_base): Sanity check weakrefs.
17785 (symtab_make_decl_local): Remove duplicated code.
17786 (symtab_alias_ultimate_target): Simplify.
17787 * varpool.c (varpool_create_variable_alias): Set weakref flag.
17789 2013-06-11 Tom de Vries <tom@codesourcery.com>
17791 * genautomata.c (gen_regexp_sequence): Handle els_num == -1. Handle
17792 sequence_vect == NULL.
17794 2013-06-11 DJ Delorie <dj@redhat.com>
17796 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
17797 (rl78_unwind_word_mode): New.
17799 2013-06-11 David Malcolm <dmalcolm@redhat.com>
17801 * final.c (debug_prefix_maps): Make static.
17803 2013-06-11 David Malcolm <dmalcolm@redhat.com>
17805 * function.c (initial_trampoline): Remove stray copy.
17807 2013-06-11 Sofiane Naci <sofiane.naci@arm.com>
17809 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
17811 2013-06-11 Martin Jambor <mjambor@suse.cz>
17813 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
17814 within bounds at the beginning of the function.
17816 2013-06-11 Alan Modra <amodra@gmail.com>
17818 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
17820 (get_named_section): Don't NULL !DECL_P decl.
17822 2013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
17824 * doc/invoke.texi (core-avx2): Document.
17826 (atom): Updated with MOVBE.
17828 2013-06-11 Richard Biener <rguenther@suse.de>
17830 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
17832 2013-06-11 Anton Blanchard <anton@samba.org>
17834 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
17835 correct shift value in little-endian mode.
17837 2013-06-11 Jakub Jelinek <jakub@redhat.com>
17840 * varasm.c (get_variable_align): Move #endif to the right place.
17842 2013-06-10 Cary Coutant <ccoutant@google.com>
17844 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
17845 for hash so that hash table traversal order is deterministic.
17847 2013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
17848 Pat Haugen <pthaugen@us.ibm.com>
17849 Peter Bergner <bergner@vnet.ibm.com>
17851 * config/rs6000/vector.md (GPR move splitter): Do not split moves
17852 of vectors in GPRS if they are direct moves or quad word load or
17855 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
17857 (direct_move_p): Likewise.
17858 (quad_load_store_p): Likewise.
17860 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
17861 classes into bins based on the physical register type.
17862 (reg_class_to_reg_type): Likewise.
17863 (IS_STD_REG_TYPE): Likewise.
17864 (IS_FP_VECT_REG_TYPE): Likewise.
17865 (reload_fpr_gpr): Arrays to determine what insn to use if we can
17866 use direct move instructions.
17867 (reload_gpr_vsx): Likewise.
17868 (reload_vsx_gpr): Likewise.
17869 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
17870 information that is a simplification of register classes. Also
17871 precalculate direct move reload helpers.
17872 (direct_move_p): New function to return true if the operation can
17873 be done as a direct move instruciton.
17874 (quad_load_store_p): New function to return true if the operation
17875 is a quad memory operation.
17876 (rs6000_legitimize_address): If quad memory, only allow register
17877 indirect for TImode addresses.
17878 (rs6000_legitimate_address_p): Likewise.
17879 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
17880 (rs6000_reload_register_type): Likewise.
17881 (register_to_reg_type): Return register type.
17882 (rs6000_secondary_reload_simple_move): New helper function for
17883 secondary reload and secondary memory needed to identify anything
17884 that is a simple move, and does not need reloading.
17885 (rs6000_secondary_reload_direct_move): New helper function for
17886 secondary reload to identify cases that can be done with several
17887 instructions via the direct move instructions.
17888 (rs6000_secondary_reload_move): New helper function for secondary
17889 reload to identify moves between register types that can be done.
17890 (rs6000_secondary_reload): Add support for quad memory operations
17891 and for direct move.
17892 (rs6000_secondary_memory_needed): Likewise.
17893 (rs6000_debug_secondary_memory_needed): Change argument names.
17894 (rs6000_output_move_128bit): New function to return the move to
17895 use for 128-bit moves, including knowing about the various
17896 limitations of quad memory operations.
17898 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
17899 memory operations. call rs6000_output_move_128bit for the actual
17900 instruciton(s) to generate.
17901 (vsx_movti_64bit): Likewise.
17903 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
17904 (UNSPEC_P8V_MTVSRWZ): Likewise.
17905 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
17906 (UNSPEC_P8V_MTVSRD): Likewise.
17907 (UNSPEC_P8V_XXPERMDI): Likewise.
17908 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
17909 (UNSPEC_FUSION_GPR): Likewise.
17910 (FMOVE128_GPR): New iterator for direct move.
17911 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
17912 (f32_sv): Likewise.
17913 (f32_dm): Likewise.
17914 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
17915 loads and direct move instructions.
17916 (zero_extendsidi2_lfiwzx): Likewise.
17917 (extendsidi2_lfiwax): Likewise.
17918 (extendsidi2_nocell): Likewise.
17919 (floatsi<mode>2_lfiwax): Likewise.
17920 (lfiwax): Likewise.
17921 (floatunssi<mode>2_lfiwzx): Likewise.
17922 (lfiwzx): Likewise.
17923 (fix_trunc<mode>_stfiwx): Likewise.
17924 (fixuns_trunc<mode>_stfiwx): Likewise.
17925 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
17926 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
17927 (parity<mode>2_cmpb): Set length/type attr.
17928 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
17929 for 'mr.' to fast_compare.
17930 (bpermd_<mode>): Change type attr to popcnt.
17931 (p8_fmrgow_<mode>): New insns for power8 direct move support.
17932 (p8_mtvsrwz_1): Likewise.
17933 (p8_mtvsrwz_2): Likewise.
17934 (reload_fpr_from_gpr<mode>): Likewise.
17935 (p8_mtvsrd_1): Likewise.
17936 (p8_mtvsrd_2): Likewise.
17937 (p8_xxpermdi_<mode>): Likewise.
17938 (reload_vsx_from_gpr<mode>): Likewise.
17939 (reload_vsx_from_gprsf): Likewise.
17940 (p8_mfvsrd_3_<mode>): LIkewise.
17941 (reload_gpr_from_vsx<mode>): Likewise.
17942 (reload_gpr_from_vsxsf): Likewise.
17943 (p8_mfvsrd_4_disf): Likewise.
17944 (multi-word GPR splits): Do not split direct moves or quad memory
17947 2013-06-10 David Malcolm <dmalcolm@redhat.com>
17949 * tree-into-ssa.c (interesting_blocks): Make static.
17951 2013-06-10 Jakub Jelinek <jakub@redhat.com>
17954 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
17955 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
17956 Use DATA_ABI_ALIGNMENT for that case instead if defined.
17957 (get_variable_align): New function.
17958 (get_variable_section, emit_bss, emit_common,
17959 assemble_variable_contents, place_block_symbol): Use
17960 get_variable_align instead of DECL_ALIGN.
17961 (assemble_noswitch_variable): Add align argument, use it
17962 instead of DECL_ALIGN.
17963 (assemble_variable): Adjust caller. Use get_variable_align
17964 instead of DECL_ALIGN.
17965 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
17967 (DATA_ABI_ALIGNMENT): Define.
17968 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
17969 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
17970 opt is false, only return the psABI mandated alignment increase.
17971 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
17972 (DATA_ABI_ALIGNMENT): ... this.
17973 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
17974 (DATA_ABI_ALIGNMENT): ... this.
17975 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
17976 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
17977 (DATA_ABI_ALIGNMENT): ... this.
17978 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
17979 * doc/tm.texi: Regenerated.
17981 2013-06-10 Uros Bizjak <ubizjak@gmail.com>
17983 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
17984 cmp_code to construct REG_EQUAL note.
17986 2013-06-09 Jakub Jelinek <jakub@redhat.com>
17989 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
17990 that operands[2] doesn't overlap with operands[0].
17992 2013-06-09 David Edelsohn <dje.gcc@gmail.com>
17993 Jan Hubicka <jh@suse.cz>
17995 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
17996 hack to mark symbols as used.
17998 2013-06-08 Vladimir Makarov <vmakarov@redhat.com>
18000 PR rtl-optimization/57559
18001 * lra-constraints.c (process_alt_operands): Don't discourage
18002 memory with known offset for offsetable memory constraint.
18003 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
18005 2013-06-08 Eric Botcazou <ebotcazou@adacore.com>
18007 * varasm.c (struct oc_local_state): Reorder fields.
18008 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
18009 and adjust accordingly.
18010 (output_constructor): Reorder initialization code and adjust call to
18011 output_constructor_bitfield.
18013 2013-06-07 Jan Hubicka <jh@suse.cz>
18015 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
18017 2013-06-07 David Malcolm <dmalcolm@redhat.com>
18019 * tree-object-size.c (unknown): Make const.
18021 2013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
18023 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
18024 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
18025 for last alternative in the cpu_facility attribute.
18027 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18030 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
18031 (xordi3): Change operand 2 constraint to arm_xordi_operand.
18032 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
18033 * config/arm/constraints.md (Dg): New constraint.
18034 * config/arm/neon.md (xordi3_neon): Remove.
18035 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
18036 * config/arm/predicates.md (arm_xordi_operand): New predicate.
18038 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18040 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
18041 Clean up alternatives.
18043 2013-06-07 Alan Modra <amodra@gmail.com>
18045 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
18048 2013-06-07 Alan Modra <amodra@gmail.com>
18050 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
18052 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18054 * config/arm/constraints.md (Df): New constraint.
18055 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
18056 Correct length attribute for last two alternatives.
18058 2013-06-07 Alan Modra <amodra@gmail.com>
18060 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
18061 override user -mfp-in-toc.
18062 (offsettable_ok_by_alignment): Consider just the current access
18063 rather than the whole object, unless BLKmode. Handle
18064 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
18065 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
18066 for -mcmodel=medium.
18067 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
18068 override user -mfp-in-toc or -msum-in-toc. Default to
18069 -mno-fp-in-toc for -mcmodel=medium.
18071 2013-06-06 DJ Delorie <dj@redhat.com>
18073 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
18074 TARGET_VALID_POINTER_MODE.
18076 2013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
18077 Pat Haugen <pthaugen@us.ibm.com>
18078 Peter Bergner <bergner@vnet.ibm.com>
18080 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
18081 Document new power8 builtins.
18083 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
18084 condition code register, to allow 128-bit logical operations to be
18085 done in the VSX or GPR registers.
18086 (nor<mode>3): Use the canonical form for nor.
18087 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
18088 vclz*, and vpopcnt* vector instructions.
18089 (nand<mode>3): Likewise.
18090 (orc<mode>3): Likewise.
18091 (clz<mode>2): LIkewise.
18092 (popcount<mode>2): Likewise.
18094 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
18095 that only the GPRs are recognized.
18097 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
18098 support for new power8 builtins.
18100 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
18102 (xscvdpspn): Likewise.
18108 (vpopcnt): Likewise.
18109 (vpopcntb): Likewise.
18110 (vpopcnth): Likewise.
18111 (vpopcntw): Likewise.
18112 (vpopcntd): Likewise.
18114 (vmrgew): Likewise.
18115 (vmrgow): Likewise.
18117 (eqv_v16qi3): Likewise.
18118 (eqv_v8hi3): Likewise.
18119 (eqv_v4si3): Likewise.
18120 (eqv_v2di3): Likewise.
18121 (eqv_v4sf3): Likewise.
18122 (eqv_v2df3): Likewise.
18124 (nand_v16qi3): Likewise.
18125 (nand_v8hi3): Likewise.
18126 (nand_v4si3): Likewise.
18127 (nand_v2di3): Likewise.
18128 (nand_v4sf3): Likewise.
18129 (nand_v2df3): Likewise.
18131 (orc_v16qi3): Likewise.
18132 (orc_v8hi3): Likewise.
18133 (orc_v4si3): Likewise.
18134 (orc_v2di3): Likewise.
18135 (orc_v4sf3): Likewise.
18136 (orc_v2df3): Likewise.
18138 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
18139 allow power8 quad mode in 64-bit.
18140 (rs6000_builtin_vectorized_function): Add support to vectorize
18141 ISA 2.07 count leading zeros, population count builtins.
18142 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
18143 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
18144 (builtin_function_type): Add vgbbd builtin function which takes an
18146 (altivec_expand_vec_perm_const): Add support for new power8 merge
18149 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
18150 that does not include TImdoe for use with 32-bit.
18151 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
18153 (UNSPEC_VSX_CVDPSPN): Likewise.
18154 (vsx_xscvdpspn): Likewise.
18155 (vsx_xscvspdpn): Likewise.
18156 (vsx_xscvdpspn_scalar): Likewise.
18157 (vsx_xscvspdpn_directmove): Likewise.
18158 (vsx_and<mode>3): Split logical operations into 32-bit and
18159 64-bit. Add support to do logical operations on TImode as well as
18160 VSX vector types. Allow logical operations to be done in either
18161 VSX registers or in general purpose registers in 64-bit mode. Add
18162 splitters if GPRs were used. For AND, add clobber of CCmode to
18163 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
18165 (vsx_and<mode>3_32bit): Likewise.
18166 (vsx_and<mode>3_64bit): Likewise.
18167 (vsx_ior<mode>3): Likewise.
18168 (vsx_ior<mode>3_32bit): Likewise.
18169 (vsx_ior<mode>3_64bit): Likewise.
18170 (vsx_xor<mode>3): Likewise.
18171 (vsx_xor<mode>3_32bit): Likewise.
18172 (vsx_xor<mode>3_64bit): Likewise.
18173 (vsx_one_cmpl<mode>2): Likewise.
18174 (vsx_one_cmpl<mode>2_32bit): Likewise.
18175 (vsx_one_cmpl<mode>2_64bit): Likewise.
18176 (vsx_nor<mode>3): Likewise.
18177 (vsx_nor<mode>3_32bit): Likewise.
18178 (vsx_nor<mode>3_64bit): Likewise.
18179 (vsx_andc<mode>3): Likewise.
18180 (vsx_andc<mode>3_32bit): Likewise.
18181 (vsx_andc<mode>3_64bit): Likewise.
18182 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
18183 and xxlorc instructions.
18184 (vsx_eqv<mode>3_64bit): Likewise.
18185 (vsx_nand<mode>3_32bit): Likewise.
18186 (vsx_nand<mode>3_64bit): Likewise.
18187 (vsx_orc<mode>3_32bit): Likewise.
18188 (vsx_orc<mode>3_64bit): Likewise.
18190 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
18192 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
18194 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
18195 (p8_vmrgow): Likewise.
18196 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
18197 GPRs to be split under VSX.
18198 (p8v_clz<mode>2): Add power8 count leading zero support.
18199 (p8v_popcount<mode>2): Add power8 population count support.
18200 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
18203 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
18206 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
18208 (vec_nand): Likewise.
18209 (vec_vclz): Likewise.
18210 (vec_vclzb): Likewise.
18211 (vec_vclzd): Likewise.
18212 (vec_vclzh): Likewise.
18213 (vec_vclzw): Likewise.
18214 (vec_vgbbd): Likewise.
18215 (vec_vmrgew): Likewise.
18216 (vec_vmrgow): Likewise.
18217 (vec_vpopcnt): Likewise.
18218 (vec_vpopcntb): Likewise.
18219 (vec_vpopcntd): Likewise.
18220 (vec_vpopcnth): Likewise.
18221 (vec_vpopcntw): Likewise.
18223 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
18225 PR rtl-optimization/57468
18226 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
18229 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
18231 PR rtl-optimization/57459
18232 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
18233 type when setting live regs.
18235 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
18237 * config/s390/s390.opt (mlra): New option.
18238 * config/s390/s390.c (s390_decompose_address): Check displacement
18239 for all registers for LRA.
18240 (s390_secondary_reload): Don't used secondary reloads for LRA.
18241 (s390_lra_p): New function.
18242 (TARGET_LRA_P): Define.
18243 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
18244 of attribute cpu_facility to zarch for the last alternative.
18245 (*cmpmem_short): Ditto.
18247 2013-06-06 Eric Botcazou <ebotcazou@adacore.com>
18249 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
18250 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
18251 (arm_expand_prologue): Likewise.
18253 2013-06-06 Teresa Johnson <tejohnson@google.com>
18256 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
18257 as this is now done by redirect_edge_and_branch_force.
18258 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
18259 barriers, and fix interaction with splitting.
18260 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
18261 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
18262 reflect changes made in the routine.
18263 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
18264 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
18265 since this is called in cfglayout mode, and replace partition fixup
18266 with assert as that is now done by force_nonfallthru_and_redirect.
18267 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
18268 already be marked with region crossing note.
18269 (insert_section_boundary_note): Make non-static, gate on flag
18270 has_bb_partition, rewrite to also check for multiple partitions.
18271 (rest_of_handle_reorder_blocks): Remove call to
18272 insert_section_boundary_note, now done later during free_cfg.
18273 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
18274 * bb-reorder.h (insert_section_boundary_note): Declare.
18275 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
18276 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
18277 invoke insert_section_boundary_note.
18278 (try_redirect_by_replacing_jump): Remove unnecessary
18279 check for region crossing note.
18280 (fixup_partition_crossing): New function.
18281 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
18282 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
18283 in non-cfglayout mode.
18284 (force_nonfallthru_and_redirect): Fixup partition boundaries,
18285 remove old code that tried to do this. Emit barrier correctly
18286 when we are in cfglayout mode.
18287 (last_bb_in_partition): New function.
18288 (rtl_split_edge): Correctly fixup partition boundaries.
18289 (commit_one_edge_insertion): Remove old code that tried to
18290 fixup region crossing edge since this is now handled in
18291 split_block, and set up insertion point correctly since
18292 block may now end in a jump.
18293 (verify_hot_cold_block_grouping): Guard against checking when not in
18294 linearized RTL mode.
18295 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
18297 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
18298 rtl_verify_flow_info, so not called in cfglayout mode.
18299 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
18300 (fixup_reorder_chain): Remove old code that attempted to fixup region
18301 crossing note as this is now handled in force_nonfallthru_and_redirect.
18302 (duplicate_insn_chain): Don't duplicate switch section notes.
18303 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
18305 * basic-block.h (emit_barrier_after_bb): Declare.
18307 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18309 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
18310 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
18311 arm_usatsihi): Adjust alternatives for arm_restrict_it.
18313 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18315 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
18317 * config/arm/ldmstm.md: Regenerate.
18319 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18321 * config/arm/sync.md (atomic_loaddi_1):
18322 Disable predication for arm_restrict_it.
18323 (arm_load_exclusive<mode>): Likewise.
18324 (arm_load_exclusivesi): Likewise.
18325 (arm_load_exclusivedi): Likewise.
18326 (arm_load_acquire_exclusive<mode>): Likewise.
18327 (arm_load_acquire_exclusivesi): Likewise.
18328 (arm_load_acquire_exclusivedi): Likewise.
18329 (arm_store_exclusive<mode>): Likewise.
18330 (arm_store_exclusive<mode>): Likewise.
18331 (arm_store_release_exclusivedi): Likewise.
18332 (arm_store_release_exclusive<mode>): Likewise.
18334 2013-06-06 Richard Biener <rguenther@suse.de>
18336 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
18338 (lto_tag_is_tree_code_p): Adjust.
18339 (lto_tag_is_gimple_code_p): Likewise.
18340 (lto_gimple_code_to_tag): Likewise.
18341 (lto_tag_to_gimple_code): Likewise.
18342 (lto_tree_code_to_tag): Likewise.
18343 (lto_tag_to_tree_code): Likewise.
18344 * data-streamer.h (streamer_write_hwi_in_range): Use
18345 uhwi streaming to stream the normalized range.
18346 (streamer_read_hwi_in_range): Likewise.
18348 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18350 * config/arm/arm.md (enabled_for_depr_it): New attribute.
18351 (predicable_short_it): Likewise.
18352 (predicated): Likewise.
18353 (enabled): Handle above.
18354 (define_cond_exec): Set predicated attribute to yes.
18356 2013-06-05 Mike Stump <mikestump@comcast.net>
18358 * gdbinit.in (__FUNCTION__): Add.
18360 2013-06-05 Uros Bizjak <ubizjak@gmail.com>
18362 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
18363 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
18365 2013-06-05 Jan Hubicka <jh@suse.cz>
18367 * varasm.c (mark_decl_referenced): Revert the removal until targets
18370 2013-06-05 David Edelsohn <dje.gcc@gmail.com>
18372 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
18373 instead of mark_decl_referenced.
18375 2013-06-05 Jan Hubicka <jh@suse.cz>
18377 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
18378 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
18379 and symtab_used_from_object_file_p.
18380 (cgraph_make_node_local_1): Clear forced_by_abi.
18381 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
18382 * cgraph.h (symtab_node_base): Add forced_by_abi.
18383 (decide_is_variable_needed): Remove.
18384 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
18385 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
18386 (decide_is_symbol_needed): ... this one; handle symbols in general;
18387 always analyze virtuals; honnor forced_by_abi.
18388 (cgraph_finalize_function): Update.
18389 (varpool_finalize_decl): Update.
18390 (symbol_defined_and_needed): Remove.
18391 (analyze_functions): Update.
18392 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
18393 output_refs, input_overwrite_node): Handle forced_by_abi.
18394 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
18395 (address_taken_from_non_vtable_p): ... this one.
18396 (comdat_can_be_unshared_p_1): New function.
18397 (cgraph_comdat_can_be_unshared_p): Rename to ...
18398 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
18399 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
18400 (function_and_variable_visibility): Clear forced_by_abi as needed.
18401 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
18402 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
18403 * symtab.c (dump_symtab_base): Dump forced_by_abi.
18404 * varpool.c (decide_is_variable_needed): Remove.
18406 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18408 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
18409 (arm_option_override): Override arm_restrict_it where appropriate.
18410 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
18411 * config/arm/arm.opt (mrestrict-it): New command-line option.
18412 * doc/invoke.texi: Document -mrestrict-it.
18414 2013-06-05 David Malcolm <dmalcolm@redhat.com>
18416 * tsan.c (tsan_atomic_table): Make const.
18418 2013-06-05 Richard Biener <rguenther@suse.de>
18420 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
18421 index associated with the tree we are supposed to replace.
18422 * tree-streamer-out.c (pack_ts_base_value_fields): Output
18423 TREE_ASM_WRITTEN as zero for everything but SSA names.
18425 2013-06-05 David Malcolm <dmalcolm@redhat.com>
18427 * tree-ssa-structalias.c (call_stmt_vars): Make static.
18429 2013-06-04 Jan Hubicka <jh@suse.cz>
18431 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
18432 (input_node, input_varpool_node): Handle correctly external same
18434 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
18435 nodes at ltrans stage.
18437 2013-06-04 Jan Hubicka <jh@suse.cz>
18439 * ipa-inline.c (update_caller_keys): Fix availability test.
18440 (update_callee_keys): Likewise.
18441 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
18442 to follow ELF standard.
18444 2013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
18446 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
18447 (mips64r5900el-*-elf*): New configurations.
18448 * config/mips/mips-cpus.def (r5900): New processor.
18449 * config/mips/mips-tables.opt: Regenerate.
18450 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
18451 (mips_issue_rate): Handle PROCESSOR_R5900.
18452 (mips_reorg_process_insns): Force reorder mode for the R5900.
18453 * config/mips/mips.h (TARGET_MIPS5900): Define.
18454 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
18456 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
18458 * config/mips/mips.md (processor): Add r5900.
18459 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
18461 2013-06-04 Ian Bolton <ian.bolton@arm.com>
18463 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
18464 into function to generate MOVI instruction.
18465 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
18466 (aarch64_preferred_simd_mode): Turn into wrapper.
18467 (aarch64_output_scalar_simd_mov_immediate): New function.
18468 * config/aarch64/aarch64-protos.h: Add prototype for above.
18470 2013-06-04 Ian Bolton <ian.bolton@arm.com>
18472 * config/aarch64/aarch64.c (simd_immediate_info): Remove
18473 element_char member.
18474 (sizetochar): Return signed char.
18475 (aarch64_simd_valid_immediate): Remove elchar and other
18476 unnecessary variables.
18477 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
18478 Calculate element_char as required.
18479 * config/aarch64/aarch64-protos.h: Update and move prototype
18480 for aarch64_output_simd_mov_immediate.
18481 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
18484 2013-06-04 Ian Bolton <ian.bolton@arm.com>
18486 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
18487 information completed by aarch64_simd_valid_immediate.
18488 (aarch64_legitimate_constant_p): Update arguments.
18489 (aarch64_simd_valid_immediate): Work with struct rather than many
18491 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
18492 (aarch64_simd_make_constant): Update arguments.
18493 (aarch64_output_simd_mov_immediate): Work with struct rather than
18494 many pointers. Output immediate directly rather than as operand.
18495 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
18497 * config/aarch64/constraints.md (Dn): Update arguments.
18499 2013-06-04 Ian Bolton <ian.bolton@arm.com>
18501 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
18503 (aarch64_simd_immediate_valid_for_move): Remove.
18504 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
18505 (aarch64_simd_make_constant): Update call.
18506 (aarch64_output_simd_mov_immediate): Update call.
18507 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
18509 * config/aarch64/constraints.md (Dn): Update call.
18511 2013-06-04 Ian Bolton <ian.bolton@arm.com>
18513 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
18514 return type to bool for prototype.
18515 (aarch64_legitimate_constant_p): Check for true instead of not -1.
18516 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
18517 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
18519 2013-06-04 Catherine Moore <clm@codesourcery.com>
18521 * config/mips/mips.opt (meva): New.
18522 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
18523 (ASM_SPEC): Handle -meva.
18524 * doc/invoke.texi (meva): Document.
18526 2013-06-04 Alan Modra <amodra@gmail.com>
18528 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
18531 2013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18533 * rtl.def: Add extra fourth optional field to define_cond_exec.
18534 * gensupport.c (process_one_cond_exec): Process attributes from
18536 * doc/md.texi: Document fourth field in define_cond_exec.
18538 2013-06-04 Eric Botcazou <ebotcazou@adacore.com>
18540 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
18541 out the processing order as in store_bit_field_1.
18543 2013-06-04 Jan Hubicka <jh@suse.cz>
18545 PR middle-end/57500
18546 * cgraphunit.c (cgraph_process_same_body_aliases): Create
18547 non-VAR_DECL node if it does not exist yet.
18549 2013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
18551 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
18552 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
18553 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
18554 target_cpu_default setting.
18556 2013-06-03 Teresa Johnson <tejohnson@google.com>
18558 * dumpfile.c (opt_info_switch_p): Change -fopt-info
18559 default to -fopt-info=optimized instead of all.
18560 * doc/invoke.texi: Ditto.
18561 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
18562 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
18563 (execute_vect_slp): Emit BB vectorization success under
18564 MSG_OPTIMIZED_LOCATIONS.
18565 * tree-vect-slp.c (vect_slp_transform_bb): Change
18566 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
18567 * tree-vect-loop.c (vect_transform_loop): Ditto.
18569 2013-06-03 Jason Merrill <jason@redhat.com>
18572 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
18573 Use TARGET_EXPR for C++.
18575 2013-06-03 Jakub Jelinek <jakub@redhat.com>
18577 PR rtl-optimization/57268
18578 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
18579 if DEBUG_INSN_P (insn).
18582 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
18584 PR rtl-optimization/57268
18585 * sched-deps.c (sched_analyze_2): Flush dependence lists if
18586 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
18588 2013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
18590 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
18591 (ix86_avoid_lea_for_addr): Likewise.
18592 (exact_dependency_1): Likewise.
18593 (ix86_adjust_cost): Likewise.
18594 (swap_top_of_ready_list): Fix formatting and !reload_completed check
18596 (do_reorder_for_imul): Fix typo, formatting and
18597 !reload_completed check removed.
18598 (ix86_sched_reorder): Fix typo and formatting.
18599 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
18602 2013-06-03 Sofiane Naci <sofiane.naci@arm.com>
18604 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
18606 2013-06-03 Eric Botcazou <ebotcazou@adacore.com>
18608 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
18609 <STRING_CST>: Likewise.
18610 <VECTOR_CST>: Likewise.
18612 2013-06-01 Janus Weil <janus@gcc.gnu.org>
18613 Mikael Morin <mikael@gcc.gnu.org>
18615 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
18616 * config.in: Regenerated.
18617 * configure: Regenerated.
18619 2013-06-01 Jan Hubicka <jh@suse.cz>
18621 PR middle-end/57366
18622 * cgraphunit.c (compile): When weakref is not supported,
18623 set up transparent aliases before final output pass.
18624 * varasm.c (assemble_alias): Do not try to do it here.
18626 2013-06-01 Jan Hubicka <jh@suse.cz>
18628 PR middle-end/57467
18629 * passes.c (for_per_function): Skip unanalyzed functions.
18631 2013-06-01 Jan Hubicka <jh@suse.cz>
18633 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
18634 (lto_symtab_merge_symbols_1): ... this one.
18635 (lto_symtab_merge_cgraph_nodes): Rename to ...
18636 (lto_symtab_merge_symbols): ... this one; simplify.
18637 * cgraph.c (same_body_aliases_done): Rename to ...
18638 (cpp_implicit_aliases_done): ... this one.
18639 (cgraph_create_function_alias): Update.
18640 (cgraph_same_body_alias): Update.
18641 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
18642 (verify_edge_corresponds_to_fndecl): Simplify.
18643 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
18644 (cgraph_node): Remove same_body_alias.
18645 (varpool_node): Remove alias_of and extra_name_alias.
18646 (same_body_aliases_done): Rename to ..
18647 (cpp_implicit_aliases_done): ... this one.
18648 (symtab_alias_ultimate_target): Add default parameter.
18649 (symtab_resolve_alias): New function.
18650 (fixup_same_cpp_alias_visibility): Declare.
18651 (cgraph_function_node): Add default parameter.
18652 (cgraph_node_asm_name): Likewise.
18653 (cgraph_function_or_thunk_node): Add default parameter; do
18654 not ICE when it is NULL.
18655 (varpool_variable_node): Likewise.
18656 * tree-emutls.c (create_emultls_var): Update.
18657 (ipa_lower_emutls): Update.
18658 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
18659 (cgraph_reset_node): Reset alias info.
18660 (cgraph_finalize_function): Update.
18661 (fixup_same_cpp_alias_visibility): Move to symtab.c.
18662 (analyze_function): Simplify.
18663 (cgraph_process_same_body_aliases): Simplify.
18664 (analyze_functions): Fixup same body aliases.
18665 (handle_alias_pairs): Simplify.
18666 (assemble_thunk): Update.
18667 (assemble_thunks_and_aliases): Update.
18668 (output_weakrefs): Rewrite.
18669 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
18670 (lto_output_varpool_node): Likewise.
18671 (compute_ltrans_boundary): Remve assert.
18672 (get_alias_symbol): New functoin.
18673 (input_node): Rewrite alias handling.
18674 (input_varpool_node): Likewise.
18675 * ipa-pure-const.c (propagate_pure_const): Fix formating.
18676 * ipa.c (process_references): Handle weakrefs correctly.
18677 (symtab_remove_unreachable_nodes): Likewise.
18678 * trans-mem.c (get_cg_data): Update.
18679 (ipa_tm_create_version_alias): Update.
18680 (ipa_tm_execute): Update.
18681 * symtab.c (dump_symtab_base): Dump aliases.
18682 (verify_symtab_base): Verify aliases.
18683 (symtab_node_availability): New function.
18684 (symtab_alias_ultimate_target): Simplify.
18685 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
18686 handle all the fixup cases.
18687 (symtab_resolve_alias): New function.
18688 * passes.c (ipa_write_summaries): Handle weakrefs.
18689 * varpool.c (varpool_analyze_node): Simplify.
18690 (assemble_aliases): Update.
18691 (varpool_create_variable_alias): Simplify.
18692 (varpool_extra_name_alias): Simplify.
18693 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
18694 (lto_symtab_merge_symbols): ... this one.
18696 2013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
18699 PR rtl-optimization/57268
18700 * sched-deps.c (sched_analyze_2): Flush dependence lists if
18701 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
18703 2013-06-01 Tobias Burnus <burnus@net-b.de>
18705 Partially reverted:
18706 2013-05-31 Tobias Burnus <burnus@net-b.de>
18708 PR middle-end/57073
18709 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
18712 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
18714 PR rtl-optimization/57268
18715 * sched-deps.c (sched_analyze_2): Flush dependence lists if
18716 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
18718 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
18720 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
18721 unordered comparison operators when -fno-trapping-math is in effect
18723 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
18724 and implement unordered comparison operators properly on the e500.
18726 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
18728 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
18729 for constant scalar integers.
18730 (simplify_relational_operation_1): Likewise.
18732 2013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
18734 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
18735 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
18738 2013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
18739 Igor Zamyatin <igor.zamyatin@intel.com>
18741 Silvermont (SLM) architecture performance tuning.
18742 * config/i386/i386.h (enum ix86_tune_indices): Add
18743 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
18744 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
18746 * config/i386/i386.c (initial_ix86_tune_features)
18747 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
18748 (ix86_lea_outperforms): Handle Silvermont tuning.
18749 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
18751 (ix86_use_lea_for_mov): Likewise.
18752 (ix86_avoid_lea_for_addr): Likewise.
18753 (ix86_lea_for_add_ok): Likewise.
18754 (exact_dependency_1): New function.
18755 (exact_store_load_dependency): Likewise.
18756 (ix86_adjust_cost): Handle Silvermont tuning.
18757 (do_reoder_for_imul): Likewise.
18758 (swap_top_of_ready_list): New function.
18759 (ix86_sched_reorder): Changed to handle Silvermont tuning.
18761 * config/i386/i386.md (peepholes that split memory operand in fp
18764 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
18766 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
18767 Remove un-necessary braces.
18769 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
18771 * config/aarch64/aarch64.c (aarch64_classify_symbol):
18772 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
18774 2013-05-31 Tobias Burnus <burnus@net-b.de>
18776 PR middle-end/57073
18777 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
18779 2013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18782 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
18783 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
18784 * config/arm/neon.md (iordi3_neon): Remove.
18785 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
18786 * config/arm/predicates.md (imm_for_neon_logic_operand):
18787 Move to earlier in the file.
18788 (neon_logic_op2): Likewise.
18789 (arm_iordi_operand_neon): New predicate.
18791 2013-05-31 Richard Biener <rguenther@suse.de>
18793 PR tree-optimization/57478
18794 PR tree-optimization/57453
18795 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
18798 2013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
18800 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
18801 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
18803 2013-05-30 Tobias Burnus <burnus@net-b.de>
18804 Thomas Koenig <tkoenig@gcc.gnu.org>
18806 PR middle-end/57073
18807 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
18808 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
18810 2013-05-30 Steven Bosscher <steven@gcc.gnu.org>
18812 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
18814 2013-05-30 Vladimir Makarov <vmakarov@redhat.com>
18816 * target.def (register_usage_leveling_p): New hook.
18817 * targhooks.c (default_register_usage_leveling_p): New.
18818 * targhooks.h (default_register_usage_leveling_p): New prototype.
18819 * lra-assigns.c (register_usage_leveling_p): Use the hook.
18820 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
18821 * doc/tm.texi: Update.
18822 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
18824 2013-05-30 Ian Bolton <ian.bolton@arm.com>
18826 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
18827 (*insv_reg<mode>): New define_insn.
18829 2013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
18831 PR rtl-optimization/57439
18832 * postreload.c (move2add_valid_value_p): Check that we have
18833 a zero subreg_regno_offset when accessing the register in
18834 the requested mode.
18836 2013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
18837 Igor Zamyatin <igor.zamyatin@intel.com>
18839 Silvermont (SLM) architecture pipeline model, tuning and
18841 * config.gcc: Add slm config options and target.
18843 * config/i386/slm.md: New.
18845 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
18847 * config/i386/i386-c.c (ix86_target_macros_internal): New case
18849 (ix86_target_macros_internal): Likewise.
18851 * config/i386/i386.c (slm_cost): New cost.
18852 (m_SLM): New macro flag.
18853 (initial_ix86_tune_features): Set m_SLM.
18854 (x86_accumulate_outgoing_args): Likewise.
18855 (x86_arch_always_fancy_math_387): Likewise.
18856 (processor_target_table): Add slm cost.
18857 (cpu_names): Add slm cpu name.
18858 (x86_option_override_internal): Set SLM ISA.
18859 (ix86_issue_rate): New case PROCESSOR_SLM.
18860 (ia32_multipass_dfa_lookahead): Likewise.
18861 (fold_builtin_cpu): Add slm.
18863 * config/i386/i386.h (TARGET_SLM): New target macro.
18864 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
18865 (processor_type): Add PROCESSOR_SLM.
18867 * config/i386/i386.md (cpu): Add new value "slm".
18868 (slm.md): Include slm.md.
18870 2013-05-30 Bernd Schmidt <bernds@codesourcery.com>
18871 Zhenqiang Chen <zhenqiang.chen@linaro.org>
18873 * config/arm/arm-protos.h: Add and update function protos.
18874 * config/arm/arm.c (use_simple_return_p): New added.
18875 (thumb2_expand_return): Check simple_return flag.
18876 * config/arm/arm.md: Add simple_return and conditional simple_return.
18877 * config/arm/iterators.md: Add iterator for return and simple_return.
18879 2013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
18881 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
18882 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
18883 (arm_emit_vfp_multi_reg_pop): Likewise.
18884 (thumb2_emit_ldrd_pop): Likewise.
18885 (arm_expand_epilogue): Add misc REG_CFA notes.
18886 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
18888 2013-05-29 Lawrence Crowl <crowl@google.com>
18890 * config/arm/t-arm: Update for below.
18892 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
18893 Change type to hash_table. Update dependent calls and types.
18895 * config/i386/t-cygming: Update for below.
18897 * config/i386/t-interix: Update for below.
18899 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
18900 Change type to hash_table. Update dependent calls and types.
18901 (i386_find_on_wrapper_list::wrappers): Likewise.
18903 * config/ia64/t-ia64: Update for below.
18905 * config/ia64/ia64.c (bundle_state_table):
18906 Change type to hash_table. Update dependent calls and types.
18908 * config/mips/mips.c (mips_reorg_process_insns::htab):
18909 Change type to hash_table. Update dependent calls and types.
18911 * config/sol2.c (solaris_comdat_htab):
18912 Change type to hash_table. Update dependent calls and types.
18914 * config/t-sol2: Update for above.
18916 2013-05-29 Teresa Johnson <tejohnson@google.com>
18918 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
18919 functions are not yet marked as defined.
18921 2013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
18922 Pat Haugen <pthaugen@us.ibm.com>
18923 Peter Bergner <bergner@vnet.ibm.com>
18925 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
18929 (vrotl<mode>3): Likewise.
18930 (vashl<mode>3): Likewise.
18931 (vlshr<mode>3): Likewise.
18932 (vashr<mode>3): Likewise.
18934 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
18935 support for power8 V2DI builtins.
18937 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
18938 power8 V2DI builtins.
18939 (vupkhsw): Likewise.
18940 (vupklsw): Likewise.
18941 (vaddudm): Likewise.
18942 (vminsd): Likewise.
18943 (vmaxsd): Likewise.
18944 (vminud): Likewise.
18945 (vmaxud): Likewise.
18946 (vpkudum): Likewise.
18947 (vpksdss): Likewise.
18948 (vpkudus): Likewise.
18949 (vpksdus): Likewise.
18954 (vsubudm): Likewise.
18955 (vcmpequd): Likewise.
18956 (vcmpgtsd): Likewise.
18957 (vcmpgtud): Likewise.
18958 (vcmpequd_p): Likewise.
18959 (vcmpgtsd_p): Likewise.
18960 (vcmpgtud_p): Likewise.
18961 (vupkhsw): Likewise.
18962 (vupklsw): Likewise.
18963 (vaddudm): Likewise.
18964 (vmaxsd): Likewise.
18965 (vmaxud): Likewise.
18966 (vminsd): Likewise.
18967 (vminud): Likewise.
18968 (vpksdss): Likewise.
18969 (vpksdus): Likewise.
18970 (vpkudum): Likewise.
18971 (vpkudus): Likewise.
18976 (vsubudm): Likewise.
18978 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
18979 support for power8 V2DI instructions.
18981 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
18982 power8 V2DI instructions. Combine pack and unpack insns to use an
18983 iterator for each mode. Check whether a particular mode supports
18984 Altivec instructions instead of just checking TARGET_ALTIVEC.
18985 (UNSPEC_VPKUWUM): Likewise.
18986 (UNSPEC_VPKSHSS): Likewise.
18987 (UNSPEC_VPKSWSS): Likewise.
18988 (UNSPEC_VPKUHUS): Likewise.
18989 (UNSPEC_VPKSHUS): Likewise.
18990 (UNSPEC_VPKUWUS): Likewise.
18991 (UNSPEC_VPKSWUS): Likewise.
18992 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
18993 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
18994 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
18995 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
18996 (UNSPEC_VUPKHSB): Likewise.
18997 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
18998 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
18999 (UNSPEC_VUPKHSH): Likewise.
19000 (UNSPEC_VUPKLSB): Likewise.
19001 (UNSPEC_VUPKLSH): Likewise.
19003 (VI_char): Likewise.
19004 (VI_scalar): Likewise.
19005 (VI_unit): Likewise.
19007 (VP_small): Likewise.
19008 (VP_small_lc): Likewise.
19009 (VU_char): Likewise.
19010 (add<mode>3): Likewise.
19011 (altivec_vaddcuw): Likewise.
19012 (altivec_vaddu<VI_char>s): Likewise.
19013 (altivec_vadds<VI_char>s): Likewise.
19014 (sub<mode>3): Likewise.
19015 (altivec_vsubcuw): Likewise.
19016 (altivec_vsubu<VI_char>s): Likewise.
19017 (altivec_vsubs<VI_char>s): Likewise.
19018 (altivec_vavgs<VI_char>): Likewise.
19019 (altivec_vcmpbfp): Likewise.
19020 (altivec_eq<mode>): Likewise.
19021 (altivec_gt<mode>): Likewise.
19022 (altivec_gtu<mode>): Likewise.
19023 (umax<mode>3): Likewise.
19024 (smax<mode>3): Likewise.
19025 (umin<mode>3): Likewise.
19026 (smin<mode>3): Likewise.
19027 (altivec_vpkuhum): Likewise.
19028 (altivec_vpkuwum): Likewise.
19029 (altivec_vpkshss): Likewise.
19030 (altivec_vpkswss): Likewise.
19031 (altivec_vpkuhus): Likewise.
19032 (altivec_vpkshus): Likewise.
19033 (altivec_vpkuwus): Likewise.
19034 (altivec_vpkswus): Likewise.
19035 (altivec_vpks<VI_char>ss): Likewise.
19036 (altivec_vpks<VI_char>us): Likewise.
19037 (altivec_vpku<VI_char>us): Likewise.
19038 (altivec_vpku<VI_char>um): Likewise.
19039 (altivec_vrl<VI_char>): Likewise.
19040 (altivec_vsl<VI_char>): Likewise.
19041 (altivec_vsr<VI_char>): Likewise.
19042 (altivec_vsra<VI_char>): Likewise.
19043 (altivec_vsldoi_<mode>): Likewise.
19044 (altivec_vupkhsb): Likewise.
19045 (altivec_vupkhs<VU_char>): Likewise.
19046 (altivec_vupkls<VU_char>): Likewise.
19047 (altivec_vupkhsh): Likewise.
19048 (altivec_vupklsb): Likewise.
19049 (altivec_vupklsh): Likewise.
19050 (altivec_vcmpequ<VI_char>_p): Likewise.
19051 (altivec_vcmpgts<VI_char>_p): Likewise.
19052 (altivec_vcmpgtu<VI_char>_p): Likewise.
19053 (abs<mode>2): Likewise.
19054 (vec_unpacks_hi_v16qi): Likewise.
19055 (vec_unpacks_hi_v8hi): Likewise.
19056 (vec_unpacks_lo_v16qi): Likewise.
19057 (vec_unpacks_hi_<VP_small_lc>): Likewise.
19058 (vec_unpacks_lo_v8hi): Likewise.
19059 (vec_unpacks_lo_<VP_small_lc>): Likewise.
19060 (vec_pack_trunc_v8h): Likewise.
19061 (vec_pack_trunc_v4si): Likewise.
19062 (vec_pack_trunc_<mode>): Likewise.
19064 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
19066 (vec_vmaxsd): Likewise.
19067 (vec_vmaxud): Likewise.
19068 (vec_vminsd): Likewise.
19069 (vec_vminud): Likewise.
19070 (vec_vpksdss): Likewise.
19071 (vec_vpksdus): Likewise.
19072 (vec_vpkudum): Likewise.
19073 (vec_vpkudus): Likewise.
19074 (vec_vrld): Likewise.
19075 (vec_vsld): Likewise.
19076 (vec_vsrad): Likewise.
19077 (vec_vsrd): Likewise.
19078 (vec_vsubudm): Likewise.
19079 (vec_vupkhsw): Likewise.
19080 (vec_vupklsw): Likewise.
19082 2013-05-29 Jan Hubicka <jh@suse.cz>
19084 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
19085 flags; reorder rest of fields in more consistent way.
19086 (varpool_node): Remove analyzed, finalized and alias.
19087 (cgraph_ndoe): Likewise.
19088 (symtab_alias_ultimate_target): New function.
19089 (cgraph_function_node): Move offline.
19090 (cgraph_reset_node): Declare.
19091 (cgraph_comdat_can_be_unshared_p): Remove.
19092 (varpool_remove_initializer): Declare.
19093 (varpool_first_defined_variable, varpool_next_defined_variable
19094 cgraph_first_defined_function, cgraph_next_defined_function): Update.
19095 (cgraph_function_with_gimple_body_p): Update.
19096 (varpool_all_refs_explicit_p): Update.
19097 (symtab_alias_target): New function.
19098 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
19099 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
19100 (cgraph_function_or_thunk_node): Simplify using
19101 symtab_alias_ultimate_target.
19102 (varpool_variable_node): Likewise.
19103 * cgraph.c (cgraph_create_function_alias): Update.
19104 (cgraph_add_thunk): Update.
19105 (cgraph_remove_node): Update.
19106 (dump_cgraph_node): Do not dump removed flags.
19107 (cgraph_function_body_availability): Update.
19108 (cgraph_propagate_frequency): Update.
19109 (verify_cgraph_node): Check sanity of local flag.
19110 (cgraph_function_node): Move here from cgraph.h; revamp for
19111 cgraph_function_or_thunk_node.
19112 * lto-symtab.c (lto_varpool_replace_node): Update.
19113 (lto_symtab_resolve_can_prevail_p): Update.
19114 (lto_symtab_merge_cgraph_nodes): Update.
19115 * ipa-cp.c (determine_versionability, initialize_node_lattices,
19116 propagate_constants_accross_call, devirtualization_time_bonus,
19117 ipcp_propagate_stage): Update.
19118 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
19119 * ipa-inline-transform.c (clone_inlined_nodes,
19120 preserve_function_body_p): Update.
19121 * ipa-reference.c (propagate): Update.
19122 (write_node_summary_p): Update.
19123 * toplev.c (wrapup_global_declaration_2): Update.
19124 * cgraphunit.c (cgraph_analyze_function): Rename to ...
19125 (analyze_function) ... this one.
19126 (cgraph_process_new_functions): Update.
19127 (cgraph_reset_node): Export.
19128 (cgraph_finalize_function): Update.
19129 (cgraph_add_new_function): Update.
19130 (process_function_and_variable_attributes): Update.
19131 (varpool_finalize_decl): Update.
19132 (symbol_finalized): Remove.
19133 (symbol_finalized_and_needed): Rename to ...
19134 (symbol_defined_and_needed): ... update.
19135 (cgraph_analyze_functions): Update.
19136 (handle_alias_pairs): Update.
19137 (mark_functions_to_output): Update.
19138 (assemble_thunk): Update.
19139 (output_in_order): Update.
19140 (output_weakrefs): Update.
19141 (finalize_compilation_unit): Update.
19142 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
19143 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
19144 input_node, input_varpool_node): Update.
19145 * dbxout.c (dbxout_expand_expr): Update.
19146 * cgraphclones.c (cgraph_clone_node): Update.
19147 (cgraph_copy_node_for_versioning): Update.
19148 (cgraph_materialize_clone): Update.
19149 (cgraph_materialize_all_clones): Update.
19150 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
19151 propagate_pure_const, propagate_nothrow): Update.
19152 * lto-streamer-out.c (lto_output, write_symbol): Update.
19153 * ipa-utils.c (ipa_reverse_postorder): Update.
19154 * ipa-inline.c (can_inline_edge_p): Update.
19155 (update_caller_keys, ipa_inline): Update.
19156 * dwarf2out.c (reference_to_unused,
19157 premark_types_used_by_global_vars_helper): Update.
19158 * tree-eh.c (tree_could_trap_p): Update.
19159 * ipa-split.c (consider_split, execute_split_functions): Update.
19160 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
19161 has_addr_references_p): Update; move ahead in file for better
19163 (process_references): Simplify.
19164 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
19165 bodies are removed.
19166 (cgraph_comdat_can_be_unshared_p): Make static.
19167 (cgraph_externally_visible_p): Update.
19168 (varpool_externally_visible_p): Update.
19169 (function_and_variable_visibility): Update.
19170 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
19171 ipa_tm_mark_force_output_node): Update.
19172 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
19173 estimate_edge_devirt_benefit, inline_generate_summary,
19174 inline_write_summary): Update.
19175 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
19176 * ipa-prop.c (ipa_compute_jump_functions): Update.
19177 (ipa_print_node_params, ipa_prop_read_section,
19178 ipa_update_after_lto_read, read_replacements_section): Update.
19179 * varasm.c (mark_decl_referenced): Update.
19180 (assemble_alias, dump_tm_clone_pairs): Update.
19181 * tree-inline.c (copy_bb): Update.
19182 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
19184 * symtab.c (dump_symtab_base): Print new flags.
19185 (verify_symtab_base): Verify new flags.
19186 (symtab_alias_ultimate_target): New function.
19187 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
19188 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
19190 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
19192 * i386.c (ix86_get_function_versions_dispatcher,
19193 ix86_generate_version_dispatcher_body): Update.
19194 (fold_builtin_cpu): Use varpool_add_new_variable.
19195 * varpool.c (varpool_remove_initializer): Break out from ...
19196 (varpool_remove_node): ... this one.
19197 (dump_varpool_node, varpool_node_for_asm,
19198 cgraph_variable_initializer_availability, varpool_analyze_node,
19199 varpool_assemble_decl, varpool_remove_unreferenced_decls,
19200 varpool_finalize_named_section_flags, varpool_create_variable_alias):
19203 2013-05-29 Jan Hubicka <jh@suse.cz>
19205 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
19207 2013-05-29 Easwaran Raman <eraman@google.com>
19209 PR tree-optimization/57442
19210 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
19211 when control exits the main loop.
19213 2013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
19215 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
19217 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
19218 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
19219 * rx/t-rx: Add rx100 under multi library matches option for nofpu
19222 2013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
19224 PR tree-optimization/57441
19225 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
19226 Don't limit size of incr_vec to number of candidates.
19228 2013-05-29 Steve Ellcey <sellcey@imgtec.com>
19230 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
19231 and mips16 directories.
19232 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
19233 (MULTILIB_DIRNAMES): Ditto.
19234 (MULTILIB_EXCEPTIONS): Add new exceptions.
19235 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
19236 (MULTILIB_DIRNAMES): Ditto.
19237 (MULTILIB_EXCEPTIONS): Add new exceptions.
19239 2012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
19240 Marcus Shawcroft <marcus.shawcroft@arm.com>
19242 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
19243 SYMBOL_TINY_ABSOLUTE.
19244 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
19245 SYMBOL_TINY_ABSOLUTE.
19246 (aarch64_expand_mov_immediate): Likewise.
19247 (aarch64_classify_symbol): Likewise.
19248 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
19249 Permit SYMBOL_TINY_ABSOLUTE.
19250 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
19252 2013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
19253 Marcus Shawcroft <marcus.shawcroft@arm.com>
19255 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
19256 Refactor if/switch. Replace gcc_assert with if.
19258 2013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
19260 * config/i386/i386.c (initial_ix86_tune_features): Enable
19261 FP Reassociation for AMD bdver1 and bdver2.
19263 2013-05-29 Martin Jambor <mjambor@suse.cz>
19265 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
19266 and IMAGPART_EXPR do not occur within other handled_components.
19268 2013-05-29 Richard Biener <rguenther@suse.de>
19270 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
19271 access on whether the use is in the BB we currently try to
19273 (vect_bb_vectorization_profitable_p): Pass the BB we currently
19274 vectorize to vect_bb_slp_scalar_cost.
19276 2013-05-29 Richard Biener <rguenther@suse.de>
19278 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
19279 computing scalar cost offsetted by stmts that are kept live
19281 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
19282 for computation of scalar cost.
19284 2013-05-28 Steve Ellcey <sellcey@mips.com>
19286 * config/mips/mips-cpus.def (mips32r2): Change processor type.
19288 2013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
19290 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
19291 array notation built-in reduction functions.
19292 * doc/passes.texi (Passes): Added documentation about changes done
19294 * doc/invoke.texi (C Dialect Options): Added documentation about
19295 the -fcilkplus flag.
19296 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
19297 (BUILTINS_DEF): Depend on cilkplus.def.
19298 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
19299 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
19300 * cilkplus.def: New file.
19302 2013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
19304 PR rtl-optimization/57439
19305 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
19307 2013-05-28 Easwaran Raman <eraman@google.com>
19309 PR tree-optimization/57337
19310 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
19311 (find_insert_point): Correctly identify the insertion point
19312 when two statements with the same UID is compared.
19314 2013-05-28 Richard Biener <rguenther@suse.de>
19316 PR tree-optimization/56787
19317 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
19318 from the list of data references.
19319 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
19321 (vect_analyze_loop_operations): Likewise.
19322 (vect_transform_loop): Remove clobbers.
19324 2013-05-28 Martin Jambor <mjambor@suse.cz>
19326 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
19327 and REALPART_EXPRs have scalar type.
19329 2013-05-28 Richard Biener <rguenther@suse.de>
19331 PR tree-optimization/57411
19332 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
19334 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
19335 virtual operand propagation.
19337 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
19339 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
19340 destination register for bmasksi_vis.
19341 (vector_init_bshuffle): Likewise.
19342 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
19344 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
19346 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
19347 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
19348 mode if the instruction isn't available in the original mode.
19349 * config/sparc/sparc.opt (mfix-ut699): New option.
19350 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
19351 (divdf3): Turn into expander.
19352 (divdf3_nofix): New insn.
19353 (divdf3_fix): Likewise.
19354 (divsf3): Disable if -mfix-ut699.
19355 (sqrtdf2): Turn into expander.
19356 (sqrtdf2_nofix): New insn.
19357 (sqrtdf2_fix): Likewise.
19358 (sqrtsf2): Disable if -mfix-ut699.
19360 2013-05-27 Richard Biener <rguenther@suse.de>
19362 PR middle-end/57412
19363 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
19364 block for the new loop.
19366 2013-05-27 Richard Biener <rguenther@suse.de>
19368 PR tree-optimization/57343
19369 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
19370 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
19371 (number_of_iterations_cond): Do not build the folded tree.
19373 2013-05-27 Richard Biener <rguenther@suse.de>
19376 PR middle-end/57381
19377 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
19378 OEP_CONSTANT_ADDRESS_OF retained.
19380 PR tree-optimization/57417
19381 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
19382 for unchanged base.
19383 (set_ssa_val_to): Compare addresses using
19384 get_addr_base_and_unit_offset.
19386 2013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
19388 PR rtl-optimization/56833
19389 * postreload.c (move2add_record_mode): New function.
19390 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
19391 (move2add_use_add2_insn): Use move2add_record_sym_value.
19392 (move2add_use_add3_insn): Likewise.
19393 (reload_cse_move2add): Use move2add_valid_value_p and
19394 move2add_record_mode. Invalidate call-clobbered and REG_INC
19395 affected regs by setting reg_mode to VOIDmode.
19396 (move2add_note_store): Don't pretend the inside of a SUBREG is
19397 the actual destination. Invalidate single/leading registers by
19398 setting reg_mode to VOIDmode.
19399 Use move2add_record_sym_value, move2add_valid_value_p and
19400 move2add_record_mode.
19402 2013-05-27 Richard Biener <rguenther@suse.de>
19404 PR tree-optimization/57396
19405 * tree-affine.c (double_int_constant_multiple_p): Properly
19406 return false for val == 0 and div != 0.
19408 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
19410 * config/mips/mips.h: Use #elif in preprocessor conditions.
19412 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
19415 * config/mips/constraints.md (kl): New constraint.
19416 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
19417 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
19418 constraint for operand 0. Split after CSE for MIPS16. Emit a move
19419 from LO for MIPS16.
19420 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
19422 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
19425 * config/mips/mips.c (mips_can_inline_p): New function.
19426 (TARGET_CAN_INLINE_P): Define.
19428 2013-05-25 Steven Bosscher <steven@gcc.gnu.org>
19430 * sched-int.h (ds_t, dw_t): Make unsigned int.
19431 Fix documentation that describes how all the ds_t bits are used.
19432 Reserve the last bit for delayed-branch scheduling.
19433 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
19434 (BITS_PER_DEP_WEAK): Fix definition and documentation.
19435 (gen_dep_weak_1): Remove prototype.
19436 * sched-deps.c (get_dep_weak_1): Make static.
19437 * target.def (speculate_insn, needs_block_p, gen_spec_check,
19438 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
19439 * doc/tm.texi: Regenerate.
19440 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
19442 2013-05-24 Steven Bosscher <steven@gcc.gnu.org>
19445 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
19447 2013-05-24 Nathan Sidwell <nathan@codesourcery.com>
19448 Sandra Loosemore <sandra@codesourcery.com>
19450 * config.gcc (powerpc-*): Allow native for with-cpu.
19452 2013-05-24 Jeff Law <law@redhat.com>
19454 PR tree-optimization/57124
19455 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
19456 conversion feeding a condition if the range has an overflow
19457 if -fstrict-overflow. Add warnings for when we do make the
19460 2013-05-24 Dehao Chen <dehao@google.com>
19462 * tree-cfg.c (locus_discrim_map): Fix the typo.
19463 (locus_discrim_hasher): Likewise.
19464 (locus_discrim_hasher::hash): Likewise.
19465 (locus_discrim_hasher::equal): Likewise.
19467 2013-05-24 Martin Jambor <mjambor@suse.cz>
19469 PR tree-optimization/57294
19470 * cgraph.h (ipa_record_stmt_references): Declare.
19471 * cgraphbuild.c (ipa_record_stmt_references): New function.
19472 (build_cgraph_edges): Use ipa_record_stmt_references.
19473 (rebuild_cgraph_edges): Likewise.
19474 (cgraph_rebuild_references): Likewise.
19475 * ipa-prop.c (ipa_modify_call_arguments): Discard references
19476 associated with the old statement and build references from the
19477 newly built statements.
19478 * ipa-ref.c (ipa_remove_stmt_references): New function.
19479 * ipa-ref.h (ipa_remove_stmt_references): Declare.
19481 2013-05-24 Vladimir Makarov <vmakarov@redhat.com>
19483 * lra-constraints.c (emit_spill_move): Use smaller mode for
19485 (check_and_process_move): Consider mem-reg moves for secondary
19487 (curr_insn_transform): Don't lose insns emitted before for
19488 secondary memory moves.
19489 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
19490 reg set up in the current insn.
19492 2013-05-24 Dehao Chen <dehao@google.com>
19494 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
19496 (locus_descrim_hasher::equal): Likewise.
19497 (build_gimple_cfg): New discriminator assignment algorithm.
19498 (make_edges): Likewise.
19499 (next_discriminator_for_locus): Likewise.
19500 (same_line_p): Likewise.
19501 (assign_discriminators): Likewise.
19502 (make_cond_expr_edges): Likewise.
19503 (make_gimple_switch_edges): Likewise.
19504 (make_goto_expr_edges): Likewise.
19505 (make_gimple_asm_edges): Likewise.
19507 2013-05-24 Ian Bolton <ian.bolton@arm.com>
19509 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
19510 X format specifier to only display bottom 16 bits.
19511 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
19512 immediate to match for operand 2, since it will be masked.
19514 2013-05-24 Richard Biener <rguenther@suse.de>
19516 PR tree-optimization/57287
19517 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
19518 all SSA names that occur in abnormal PHIs.
19520 2013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
19523 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
19524 that index is not negative.
19526 2013-05-24 Eric Botcazou <ebotcazou@adacore.com>
19528 PR rtl-optimization/55177
19529 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
19530 (simplify_byte_swapping_operation): New.
19531 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
19532 (simplify_relational_operation_1): Deal with BSWAP.
19534 2013-05-23 Richard Henderson <rth@redhat.com>
19537 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
19538 (ix86_reorg): Call it.
19540 2013-05-23 Uros Bizjak <ubizjak@gmail.com>
19543 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
19544 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
19545 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
19547 2013-05-23 Christian Bruel <christian.bruel@st.com>
19550 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
19552 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
19553 Marcus Shawcroft <marcus.shawcroft@arm.com>
19555 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
19556 * config/aarch64/constraints.md (Usa): Remove.
19557 * doc/md.texi (AArch64 Usa): Remove.
19559 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
19560 Marcus Shawcroft <marcus.shawcroft@arm.com>
19562 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
19563 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
19564 * config/aarch64/predicates.md (aarch64_const_address): Remove.
19565 (aarch64_mov_operand): Use aarch64_mov_operand_p.
19567 2013-05-23 Vidya Praveen <vidyapraveen@arm.com>
19569 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
19570 instruction (AdvSIMD).
19571 * config/aarch64/aarch64-builtins.c
19572 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
19573 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
19575 2013-05-23 Martin Jambor <mjambor@suse.cz>
19577 PR middle-end/57347
19578 * tree.h (contains_bitfld_component_ref_p): Declare.
19579 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
19580 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
19582 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
19583 not access a bit-field. Assert all final offsets are byte-aligned.
19585 2013-05-23 Richard Biener <rguenther@suse.de>
19587 PR tree-optimization/57380
19588 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
19589 least one invariant or re-used load.
19590 * passes.c (init_optimization_passes): Move pass_phiprop before
19593 2013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
19595 * config/aarch64/aarch64-simd.md
19596 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
19598 2013-05-23 Richard Biener <rguenther@suse.de>
19600 PR middle-end/57381
19601 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
19602 OEP_CONSTANT_ADDRESS_OF retained.
19604 2013-05-23 Jakub Jelinek <jakub@redhat.com>
19606 PR middle-end/57344
19607 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
19608 don't lower unit. Handle unit not being always BITS_PER_WORD.
19610 2013-05-23 Richard Biener <rguenther@suse.de>
19612 PR rtl-optimization/57341
19613 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
19614 instead of true_dependence.
19616 2013-05-22 David Malcolm <dmalcolm@redhat.com>
19618 * bb-reorder.c (branch_threshold): Make const.
19619 (exec_threshold): Ditto.
19621 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
19622 Pat Haugen <pthaugen@us.ibm.com>
19623 Peter Bergner <bergner@vnet.ibm.com>
19625 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
19626 documentation for the power8 crypto builtins.
19628 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
19630 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
19631 macros for defining power8 builtin functions.
19632 (BU_P8V_AV_2): Likewise.
19633 (BU_P8V_AV_P): Likewise.
19634 (BU_P8V_VSX_1): Likewise.
19635 (BU_P8V_OVERLOAD_1): Likewise.
19636 (BU_P8V_OVERLOAD_2): Likewise.
19637 (BU_CRYPTO_1): Likewise.
19638 (BU_CRYPTO_2): Likewise.
19639 (BU_CRYPTO_3): Likewise.
19640 (BU_CRYPTO_OVERLOAD_1): Likewise.
19641 (BU_CRYPTO_OVERLOAD_2): Likewise.
19642 (XSCVSPDP): Fix typo, point to the correct instruction.
19643 (VCIPHER): Add power8 crypto builtins.
19644 (VCIPHERLAST): Likewise.
19645 (VNCIPHER): Likewise.
19646 (VNCIPHERLAST): Likewise.
19647 (VPMSUMB): Likewise.
19648 (VPMSUMH): Likewise.
19649 (VPMSUMW): Likewise.
19650 (VPERMXOR_V2DI): Likewise.
19651 (VPERMXOR_V4SI: Likewise.
19652 (VPERMXOR_V8HI: Likewise.
19653 (VPERMXOR_V16QI: Likewise.
19654 (VSHASIGMAW): Likewise.
19655 (VSHASIGMAD): Likewise.
19656 (VPMSUM): Likewise.
19657 (VPERMXOR): Likewise.
19658 (VSHASIGMA): Likewise.
19660 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
19661 __CRYPTO__ if the crypto instructions are available.
19662 (altivec_overloaded_builtins): Add support for overloaded power8
19665 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
19666 support for power8 crypto builtins.
19667 (builtin_function_type): Likewise.
19668 (altivec_init_builtins): Add support for builtins that take vector
19669 long long (V2DI) arguments.
19671 * config/rs6000/crypto.md: New file, define power8 crypto
19674 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
19675 Pat Haugen <pthaugen@us.ibm.com>
19676 Peter Bergner <bergner@vnet.ibm.com>
19678 * doc/invoke.texi (Option Summary): Add power8 options.
19679 (RS/6000 and PowerPC Options): Likewise.
19681 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
19682 constraints.md instead of rs6000.h. Reorder w* constraints. Add
19683 wm, wn, wr documentation.
19685 * config/rs6000/constraints.md (wm): New constraint for VSX
19686 registers if direct move instructions are enabled.
19687 (wn): New constraint for no registers.
19688 (wq): New constraint for quad word even GPR registers.
19689 (wr): New constraint if 64-bit instructions are enabled.
19690 (wv): New constraint if power8 vector instructions are enabled.
19691 (wQ): New constraint for quad word memory locations.
19693 * config/rs6000/predicates.md (const_0_to_15_operand): New
19694 constraint for 0..15 for crypto instructions.
19695 (gpc_reg_operand): If VSX allow registers in VSX registers as well
19696 as GPR and floating point registers.
19697 (int_reg_operand): New predicate to match only GPR registers.
19698 (base_reg_operand): New predicate to match base registers.
19699 (quad_int_reg_operand): New predicate to match even GPR registers
19700 for quad memory operations.
19701 (vsx_reg_or_cint_operand): New predicate to allow vector logical
19702 operations in both GPR and VSX registers.
19703 (quad_memory_operand): New predicate for quad memory operations.
19704 (reg_or_indexed_operand): New predicate for direct move support.
19706 * config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
19707 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
19708 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
19709 (POWERPC_MASKS): Add power8 options.
19710 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
19713 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
19714 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
19716 * config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
19717 (-mpower8-fusion): New power8 options.
19718 (-mpower8-fusion-sign): Likewise.
19719 (-mpower8-vector): Likewise.
19720 (-mcrypto): Likewise.
19721 (-mdirect-move): Likewise.
19722 (-mquad-memory): Likewise.
19724 * config/rs6000/rs6000.c (power8_cost): Initial definition for power8.
19725 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
19727 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
19728 (rs6000_debug_vector_unit): Add p8_vector.
19729 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
19730 definitions. Also print fusion state.
19731 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
19732 (rs6000_builtin_mask_calculate): Add power8 builtin support.
19733 (rs6000_option_override_internal): Add support for power8.
19734 (rs6000_common_init_builtins): Add debugging for skipped builtins
19735 if -mdebug=builtin.
19736 (rs6000_adjust_cost): Add power8 support.
19737 (rs6000_issue_rate): Likewise.
19738 (insn_must_be_first_in_group): Likewise.
19739 (insn_must_be_last_in_group): Likewise.
19740 (force_new_group): Likewise.
19741 (rs6000_register_move_cost): Likewise.
19742 (rs6000_opt_masks): Likewise.
19744 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
19745 power8 capable assembler, default to power7 options.
19746 (TARGET_DIRECT_MOVE): Likewise.
19747 (TARGET_CRYPTO): Likewise.
19748 (TARGET_P8_VECTOR): Likewise.
19749 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
19750 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
19751 (VECTOR_MEM_P8_VECTOR_P): Likewise.
19752 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
19753 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
19754 (TARGET_XSCVDPSPN): Likewise.
19755 (TARGET_XSCVSPDPN): Likewsie.
19756 (TARGET_SYNC_HI_QI): Likewise.
19757 (TARGET_SYNC_TI): Likewise.
19758 (MASK_CRYPTO): Likewise.
19759 (MASK_DIRECT_MOVE): Likewise.
19760 (MASK_P8_FUSION): Likewise.
19761 (MASK_P8_VECTOR): Likewise.
19762 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
19763 temporary used by some of the direct move instructions to get two FP
19764 temporary registers does not force creation of a stack frame.
19765 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
19766 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
19767 that any VSX registers are tieable, even if they are also an
19768 Altivec vector mode.
19769 (r6000_reg_class_enum): Add wm, wr, wv constraints.
19770 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
19771 (RS6000_BTM_CRYPTO): Likewise.
19772 (RS6000_BTM_COMMON): Likewise.
19774 * config/rs6000/rs6000.md (cpu attribute): Add power8.
19775 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
19776 (enum rs6000_vector): Add power8 vector support.
19778 2013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
19782 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
19783 (any_sibcall_could_use_r3): this and handle indirect calls.
19784 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
19786 2013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
19788 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
19790 2013-05-22 Richard Biener <rguenther@suse.de>
19792 PR middle-end/57349
19793 * profile.c (branch_prob): Do not split blocks that are
19794 abnormally receiving from ECF_RETURNS_TWICE functions.
19796 2013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
19798 * recog.c (offsettable_address_addr_space_p): Fix calculation of
19799 address mode. Move pointer mode initialization to the same place.
19801 2013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
19803 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
19804 while it has any effect.
19806 2013-05-21 Easwaran Raman <eraman@google.com>
19808 PR tree-optimization/57322
19809 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
19810 UID of the statement added to the BB to be 1.
19812 2013-05-21 Jakub Jelinek <jakub@redhat.com>
19814 PR tree-optimization/57331
19815 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
19816 of conversion from pointer type to integral type with integer.
19818 2013-05-21 Martin Jambor <mjambor@suse.cz>
19821 * ipa-prop.c (ipa_read_node_info): Process param_used and
19822 controlled_uses in the same order as when writing.
19824 2013-05-21 Magnus Granberg <baldrick@free.fr>
19827 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
19829 2013-05-21 Richard Biener <rguenther@suse.de>
19831 PR tree-optimization/57318
19832 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
19833 estimate stmts with side-effects as likely eliminated.
19835 2013-05-21 Richard Biener <rguenther@suse.de>
19837 PR tree-optimization/57330
19838 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
19839 preserve the call stmts fntype.
19841 2013-05-21 Richard Biener <rguenther@suse.de>
19843 PR tree-optimization/57303
19844 * tree-ssa-sink.c (statement_sink_location): Improve killing
19845 stmt detection and properly handle self-assignments.
19847 2013-05-21 Christian Bruel <christian.bruel@st.com>
19849 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
19850 spanning registers. LEAF_REG_REMAP is supported only for contiguous
19851 registers. Set register size out of the PARALLEL loop.
19853 2013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
19856 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
19857 (*fmasf4, *fmasf4_media): New insns.
19859 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
19861 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
19862 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
19863 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
19864 (mips_idiv_insns): Update the comments to say that the returned
19865 instruction counts are in units of BASE_INSN_LENGTH.
19866 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
19867 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
19868 using 2 rather than 4 as the length of indirect MIPS16 and
19869 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
19870 length of a NOP. Don't divide MIPS16 lengths by 2.
19871 (mips16_split_long_branches): Assume a branch is long if the
19872 length is greater than 4 rather than 8.
19873 * config/mips/mips.md (length): Give MIPS16 lengths directly,
19874 rather than multiplying them by 2. Multiply instruction counts
19875 by BASE_INSN_LENGTH rather than 4.
19876 (*jump_mips16, tls_get_tp_mips16_<mode>)
19877 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
19879 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
19881 * config/mips/mips.md (extended_mips16): Remove branch case.
19882 (length): Remove duplicated extended_mips16 test.
19884 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
19886 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
19888 2013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
19890 * recog.h (Recog_data): Rename to...
19891 (recog_data_d): ...this.
19892 (recog_data): Update accordingly.
19893 * recog.c (recog_data): Likewise.
19894 * reload.c (save_recog_data): Likewise.
19895 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
19896 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
19898 2013-05-17 Julian Brown <julian@codesourcery.com>
19900 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
19901 found in a REG_EQUAL note, invalidate it.
19903 2013-05-17 Easwaran Raman <eraman@google.com>
19905 * tree-ssa-reassoc.c (find_insert_point): New function.
19906 (insert_stmt_after): Likewise.
19907 (get_def_stmt): Likewise.
19908 (ensure_ops_are_available): Likewise.
19909 (not_dominated_by): Likewise.
19910 (rewrite_expr_tree): Do not move statements beyond what is
19911 necessary. Remove call to swap_ops_for_binary_stmt...
19912 (reassociate_bb): ... and move it here.
19913 (build_and_add_sum): Assign UIDs for new statements.
19914 (linearize_expr): Likewise.
19915 (do_reassoc): Renumber gimple statement UIDs.
19917 2013-05-17 Jan Hubicka <jh@suse.cz>
19919 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
19921 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
19922 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
19924 (output_weakrefs): Update.
19926 2013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
19927 Martin Jambor <mjambor@suse.cz>
19929 PR middle-end/57276
19930 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
19931 value that corresponds to the given aggval is found in values vector.
19933 2013-05-17 Uros Bizjak <ubizjak@gmail.com>
19935 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
19936 sse, sse2, sse3, ssse3 and sse4a flags to options.
19938 2013-05-17 David Malcolm <dmalcolm@redhat.com>
19940 * gengtype-state.c: (s_expr_writer): New class, to handle
19941 prettifying of output layout of s-expressions.
19942 (state_writer): New class, to write out gtype.state.
19943 (state_written_type_count): Move this variable into member data of
19945 (s_expr_writer::s_expr_writer): New code: constructor for new class
19946 (state_writer::state_writer(): ditto
19947 (s_expr_writer::write_new_line): New function
19948 (s_expr_writer::write_any_indent): ditto
19949 (s_expr_writer::begin_s_expr): ditto
19950 (s_expr_writer::end_s_expr): ditto
19951 (write_state_fileloc): convert to method of state_writer...
19952 (state_writer:: write_state_fileloc): ...and use methods of
19953 s_expr_writer to write indentation into the gtype.state output file
19954 to visually represent the hierarchical structure of the list
19956 (write_state_fields): ditto, renaming to...
19957 (state_writer::write_state_fields)
19958 (write_state_a_string): ditto, renaming to...
19959 (state_writer::write_state_a_string)
19960 (write_state_string_option): ditto, renaming to...
19961 (state_writer::write_state_string_option)
19962 (write_state_type_option): ditto, renaming to...
19963 (state_writer::write_state_type_option)
19964 (write_state_nested_option): ditto, renaming to...
19965 (state_writer::write_state_nested_option)
19966 (write_state_option): ditto, renaming to...
19967 (state_writer::write_state_option)
19968 (write_state_options): ditto, renaming to...
19969 (state_writer::write_state_options)
19970 (write_state_lang_bitmap): ditto, renaming to...
19971 (state_writer::write_state_lang_bitmap)
19972 (write_state_version): ditto, renaming to...
19973 (state_writer::write_state_version)
19974 (write_state_scalar_type): ditto, renaming to...
19975 (state_writer::write_state_scalar_type)
19976 (write_state_string_type): ditto, renaming to...
19977 (state_writer::write_state_string_type)
19978 (write_state_undefined_type): ditto, renaming to...
19979 (state_writer::write_state_undefined_type)
19980 (write_state_struct_union_type): ditto, renaming to...
19981 (state_writer::write_state_struct_union_type)
19982 (write_state_struct_type): ditto, renaming to...
19983 (state_writer::write_state_struct_type)
19984 (write_state_user_struct_type): ditto, renaming to...
19985 (state_writer::write_state_user_struct_type)
19986 (write_state_lang_struct_type): ditto, renaming to...
19987 (state_writer::write_state_lang_struct_type)
19988 (write_state_param_struct_type): ditto, renaming to...
19989 (state_writer::write_state_param_struct_type)
19990 (write_state_pointer_type): ditto, renaming to...
19991 (state_writer::write_state_pointer_type)
19992 (write_state_array_type): ditto, renaming to...
19993 (state_writer::write_state_array_type)
19994 (write_state_gc_used): ditto, renaming to...
19995 (state_writer::write_state_gc_used)
19996 (write_state_common_type_content): ditto, renaming to...
19997 (state_writer::write_state_common_type_content)
19998 (write_state_type): ditto, renaming to...
19999 (state_writer::write_state_type)
20000 (write_state_pair_list): ditto, renaming to...
20001 (state_writer::write_state_pair_list)
20002 (write_state_pair): ditto, renaming to...
20003 (state_writer::write_state_pair)
20004 (write_state_typedefs): ditto, renaming to...
20005 (state_writer::write_state_typedefs)
20006 (write_state_structures): ditto, renaming to...
20007 (state_writer::write_state_structures)
20008 (write_state_param_structs): ditto, renaming to...
20009 (state_writer::write_state_param_structs)
20010 (write_state_variables): ditto, renaming to...
20011 (state_writer::write_state_variables)
20012 (write_state_srcdir): ditto, renaming to...
20013 (state_writer::write_state_srcdir)
20014 (write_state_files_list): ditto, renaming to...
20015 (state_writer::write_state_files_list)
20016 (write_state_languages): ditto, renaming to...
20017 (state_writer::write_state_languages)
20018 (write_state): create a state_writer instance and use it when
20019 writing out the state file
20021 2013-05-17 Mike Stump <mikestump@comcast.net>
20023 PR rtl-optimization/57304
20024 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
20025 accessing DF_REF_REAL_LOC.
20027 2013-05-17 Jakub Jelinek <jakub@redhat.com>
20029 PR rtl-optimization/57281
20030 PR rtl-optimization/57300
20031 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
20032 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
20033 what the other splitter did if the registers are dead.
20035 2013-05-17 Richard Biener <rguenther@suse.de>
20037 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
20040 2013-05-17 Jakub Jelinek <jakub@redhat.com>
20042 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
20045 2013-05-17 Marek Polacek <polacek@redhat.com>
20047 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
20048 length when doing non-zero store of storing '\0' to '\0'.
20050 2013-05-17 Jakub Jelinek <jakub@redhat.com>
20052 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
20053 vect_external_def oprnd1 with loop_vinfo, try to emit
20054 optional cast, negation and and stmts on the loop preheader
20055 edge instead of into the pattern def seq.
20057 PR tree-optimization/57051
20058 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
20059 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
20061 2013-05-16 Nick Clifton <nickc@redhat.com>
20063 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
20064 (rl78_is_naked_func): New function.
20065 (rl78_expand_prologue): Skip prologue generation for naked functions.
20066 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
20067 * doc/extend.texi (naked): Add RL78 to the list of processors
20068 that supports this attribute.
20070 2013-05-16 Jeff Law <law@redhat.com>
20072 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
20074 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
20076 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
20077 cache parameters using detect_caches_amd also for CYRIX,
20078 NSC and TM2 signatures.
20080 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
20081 Dzianis Kahanovich <mahatma@eu.by>
20085 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
20086 VIA/Centaur processors and determine their cache parameters
20087 using detect_caches_amd.
20089 2013-05-16 Teresa Johnson <tejohnson@google.com>
20091 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
20092 (rtl_verify_edges): New function.
20093 (rtl_verify_bb_insns): Ditto.
20094 (rtl_verify_bb_pointers): Ditto.
20095 (rtl_verify_bb_insn_chain): Ditto.
20096 (rtl_verify_fallthru): Ditto.
20097 (rtl_verify_bb_layout): Ditto.
20098 (rtl_verify_flow_info_1): Outline checks into new functions.
20099 (rtl_verify_flow_info): Ditto.
20101 2013-05-16 Steve Ellcey <sellcey@imgtec.com>
20103 * cfghooks.c (copy_bbs): Add update_dominance argument.
20104 * cfghooks.h (copy_bbs): Update prototype.
20105 * tree-cfg.c (gimple_duplicate_sese_region):
20106 Add update_dominance argument.
20107 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
20108 * tree-ssa-loop-ch.c (copy_loop_headers): Update
20109 gimple_duplicate_sese_region call.
20110 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
20111 Update copy_bbs call.
20112 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
20113 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
20115 2013-05-16 Jakub Jelinek <jakub@redhat.com>
20117 * tree-vectorizer.h (NUM_PATTERNS): Increment.
20118 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
20119 vect_recog_rotate_pattern.
20120 (vect_recog_rotate_pattern): New function.
20122 2013-05-16 Jason Merrill <jason@redhat.com>
20124 * Makefile.in (LLINKER): New variable.
20125 (mostlyclean): Remove link mutex.
20126 * configure.ac: Handle --enable-link-mutex.
20127 * lock-and-run.sh: New script.
20129 2013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
20132 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
20135 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
20137 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
20139 2013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
20141 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
20142 * config/arm/arm.c (next_consecutive_mem): New function.
20143 (gen_movmem_ldrd_strd): Likewise.
20144 * config/arm/arm.md (movmemqi): Update condition and code.
20145 (unaligned_loaddi, unaligned_storedi): New patterns.
20147 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
20149 * config.gcc: Obsolete *-*-solaris2.9*.
20150 * doc/install.texi (Specific, *-*-solaris2*): Document it.
20152 2013-05-16 Richard Biener <rguenther@suse.de>
20154 * passes.c (init_optimization_passes): Move pass_parallelize_loops
20155 earlier, after GRAPHITE transforms and IV canonicalization.
20157 2013-05-16 Jakub Jelinek <jakub@redhat.com>
20159 * omp-low.c (extract_omp_for_data): For collapsed loops,
20160 if at least one of the loops is known at compile time to
20161 iterate zero times, set count to 0.
20162 (expand_omp_regimplify_p): New function.
20163 (expand_omp_for_generic): For collapsed loops, if at least
20164 one of the loops isn't known to iterate at least once,
20165 add runtime check with setting count to 0.
20166 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
20167 For unsigned types if it isn't known at compile time that
20168 the loop will iterate at least once, add runtime check to bypass
20169 the whole loop if initial condition isn't true.
20171 2013-05-16 Nathan Sidwell <nathan@codesourcery.com>
20173 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
20175 2013-05-16 Marc Glisse <marc.glisse@inria.fr>
20177 PR middle-end/57286
20178 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
20179 transformations to avoid an infinite loop.
20181 2013-05-16 Marek Polacek <polacek@redhat.com>
20183 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
20185 2013-05-15 Leif Ekblad <leif@rdos.net>
20187 * config/i386/i386.c (ix86_decompose_address): Use
20188 DEFAULT_TLS_SEG_REG to access TLS segment register.
20189 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
20190 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
20191 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
20193 2013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
20196 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
20197 sibling calls to functions that would normally be lazily bound,
20198 unless $gp is call-clobbered.
20200 2013-05-15 Uros Bizjak <ubizjak@gmail.com>
20202 * config/i386/i386.c (ix86_option_override_internal): Update
20203 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
20204 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
20205 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
20206 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
20208 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
20210 2013-05-15 Andreas Schwab <schwab@suse.de>
20212 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
20213 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
20216 2013-05-15 Teresa Johnson <tejohnson@google.com>
20218 * loop-unroll.c (report_unroll_peel): Check decision before
20219 emitting unroll/peel message.
20221 2013-05-15 Teresa Johnson <tejohnson@google.com>
20223 * function.h (has_bb_partition): New rtl_data flag.
20224 (bb_reorder_complete): Ditto.
20225 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
20226 instead of flag_reorder_blocks_and_partition.
20227 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
20228 with some enhancements.
20229 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
20230 * bb-reorder.c (connect_traces): Check for has_bb_partition
20231 instead of flag_reorder_blocks_and_partition.
20232 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
20233 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
20234 verify_hot_cold_block_grouping.
20235 (partition_hot_cold_basic_blocks): Set has_bb_partition.
20237 2013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
20240 * config/arm/predicates.md (call_insn_operand): New predicate.
20241 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
20242 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
20243 if insn is not a tail call.
20244 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
20246 * config/arm/arm.h (enum reg_class): New caller save register class.
20247 (REG_CLASS_NAMES): Likewise.
20248 (REG_CLASS_CONTENTS): Likewise.
20249 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
20252 2013-05-15 Richard Biener <rguenther@suse.de>
20254 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
20255 of MSG_OPTIMIZED_LOCATIONS.
20256 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
20257 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
20259 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
20260 of MSG_OPTIMIZED_LOCATIONS.
20261 (execute_vect_slp): Likewise.
20262 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
20263 (vect_create_cond_for_alias_checks): Likewise.
20264 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
20265 (vect_recog_widen_mult_pattern): Likewise.
20266 (vect_recog_widen_sum_pattern): Likewise.
20267 (vect_recog_over_widening_pattern): Likewise.
20268 (vect_recog_widen_shift_pattern): Likewise.
20269 (vect_recog_vector_vector_shift_pattern): Likewise.
20270 (vect_recog_divmod_pattern): Likewise.
20271 (vect_recog_mixed_size_cond_pattern): Likewise.
20272 (vect_recog_bool_pattern): Likewise.
20273 (vect_pattern_recog_1): Likewise.
20275 2013-05-15 Martin Jambor <mjambor@suse.cz>
20277 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
20278 non-functions to builtin_unreachable.
20279 * ipa-inline-transform.c (inline_call): Do not assert estimates were
20280 correct when new direct edges were discovered.
20282 2013-05-15 Martin Jambor <mjambor@suse.cz>
20284 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
20285 header, print symbol order instead of node uid, print more information
20286 about indirect edge targets.
20287 (ipa_make_edge_direct_to_target): Print symbol order instead of node
20289 (ipa_make_edge_direct_to_target): Likewise.
20290 (remove_described_reference): Likewise.
20291 (propagate_controlled_uses): Likewise.
20292 (ipa_print_node_params): Also print symbol order.
20293 (ipcp_transform_function): Print symbol order instead of node uids.
20294 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
20295 (cgraph_get_create_real_symbol_node): Likewise.
20296 * ipa-cp.c (print_lattice): Likewise.
20297 (print_all_lattices): Likewise.
20298 (determine_versionability): Likewise.
20299 (initialize_node_lattices): Likewise.
20300 (estimate_local_effects): Likewise.
20301 (update_profiling_info): Likewise.
20302 (create_specialized_node): Likewise.
20303 (perhaps_add_new_callers): Likewise.
20304 (decide_about_value): Likewise.
20305 (decide_whether_version_node): Likewise.
20306 (identify_dead_nodes): Likewise.
20307 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
20308 (dump_inline_summary): Likewise.
20309 (estimate_node_size_and_time): Likewise.
20310 (inline_analyze_function): Likewise.
20311 * ipa-inline.c (report_inline_failed_reason): Likewise.
20312 (want_early_inline_function_p): Likewise.
20313 (edge_badness): Likewise.
20314 (update_edge_key): Likewise.
20315 (inline_small_functions): Likewise. Add dumping of order to two other
20317 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
20318 instead of node uids.
20319 (propagate_pure_const): Likewise.
20320 (propagate_pure_const): Likewise.
20321 * ipa-utils.c (dump_cgraph_node_set): Likewise.
20322 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
20323 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
20325 * tree-pretty-print.c (dump_function_header): Likewise.
20326 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
20327 Print symbol order instead of node uids.
20329 2013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
20331 * config/s390/s390.c (s390_register_move_cost): Don't impose the
20332 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
20334 2013-05-15 Richard Biener <rguenther@suse.de>
20336 PR tree-optimization/57275
20337 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
20338 return value for fail to do runtime alias checks for gather loads.
20340 2013-05-15 Jan Hubicka <jh@suse.cz>
20344 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
20345 weakrefs are not external.
20346 (lto_symtab_merge_decls): Fix thinko when dealing with
20347 non-lto_symtab decls.
20348 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
20349 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
20350 * varpool.c (dump_varpool_node): Dump more flags.
20352 2013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
20354 * config/i386/i386.c (processor_alias_table): Add instruction
20355 FSGSBASE for AMD bdver3 architecture.
20357 2013-05-14 Jakub Jelinek <jakub@redhat.com>
20359 * tree.c (warn_deprecated_use): Print file:line using locus color.
20360 * diagnostic.c (diagnostic_report_current_module): Print file:line
20361 and file:line:column using locus color.
20363 2013-05-14 Mike Stump <mikestump@comcast.net>
20365 * gdbinit.in: Add __null.
20367 2013-05-14 Mike Stump <mikestump@comcast.net>
20369 * recog.h: Rename struct recog_data to Recog_data.
20370 * recog.c: Likewise.
20371 * reload.c (can_reload_into): Likewise.
20372 * config/picochip/picochip.c: Likewise.
20374 2013-05-14 Mike Stump <mikestump@comcast.net>
20376 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
20378 2013-05-14 Steven Bosscher <steven@gcc.gnu.org>
20380 * resource.h (struct resources): Remove unch_memory member.
20381 (CLEAR_RESOURCE): Don't clear unch_memory.
20382 * resource.c (mark_referenced_resources): Don't set it.
20383 (mark_set_resources): Likewise.
20384 (mark_target_live_regs): Don't clear it.
20385 (init_resource_info): Likewise.
20386 * reorg.c (resource_conflicts_p): Don't compare it.
20387 (redundant_insn): Don't set it.
20389 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
20391 * emit-rtl.c (next_label): Remove unused function.
20392 (skip_consecutive_labels, link_cc0_insns): Move to ...
20393 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
20394 only place where these functions are used, and make them static.
20396 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
20398 * fold-const.c (fold_negate_expr): Handle vectors.
20399 (fold_truth_not_expr): Make it static.
20400 (fold_invert_truthvalue): New static function.
20401 (invert_truthvalue_loc): Handle vectors. Do not call
20402 fold_truth_not_expr directly.
20403 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
20404 <TRUTH_NOT_EXPR>: Do not cast to boolean.
20405 (fold_comparison): Handle vector constants.
20406 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
20407 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
20408 * tree.h (fold_truth_not_expr): Remove declaration.
20410 2013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
20412 * config/aarch64/aarch64-simd.md
20413 (aarch64_vcond_internal<mode>): Rename to...
20414 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
20415 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
20416 float modes. Clarify all iterator modes.
20417 (vcond<mode><mode>): Use new name for vcond expanders.
20418 (vcond<v_cmp_result><mode>): Likewise.
20419 (vcondu<mode><mode>: Likewise.
20420 * config/aarch64/iterators.md (VDQF_COND): New.
20422 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
20425 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
20426 variable for the shift amount. Check that we shift by non-negative
20429 2013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
20432 * config/arm/arm.h (EPILOGUE_USES): Only return true
20433 for LR_REGNUM after epilogue_completed.
20435 2013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
20437 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
20438 is error_mark_node.
20440 2013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
20443 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
20444 and Solaris 11+/x86 with gld.
20445 * configure: Regenerate.
20447 2013-05-14 Jakub Jelinek <jakub@redhat.com>
20449 * expmed.c (expand_shift_1): Canonicalize rotates by
20450 constant bitsize / 2 to bitsize - 1.
20451 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
20452 case ROTATERT>: Likewise.
20455 2013-05-10 Jakub Jelinek <jakub@redhat.com>
20457 * config/i386/i386.md (rotateinv): New code attr.
20458 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
20459 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
20460 roll $31, %eax, etc.
20462 2013-05-14 Richard Biener <rguenther@suse.de>
20464 PR middle-end/57235
20465 * tree-eh.c (sink_clobbers): Give up for successors with
20466 multiple predecessors and no virtual uses.
20468 2013-05-14 Eric Botcazou <ebotcazou@adacore.com>
20470 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
20471 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
20473 2013-05-14 Jakub Jelinek <jakub@redhat.com>
20475 PR middle-end/57251
20476 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
20477 the case when both op0 and op1 have VOIDmode.
20479 2013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
20481 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
20482 in multiply-accumulate mode.
20484 2013-05-13 Guozhi Wei <carrot@google.com>
20486 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
20488 2013-05-13 Kai Tietz <ktietz@redhat.com>
20491 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
20492 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
20493 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
20494 * config/i386/i386.c (ix86_option_override_internal): Likewise.
20495 (ix86_expand_prologue): Likewise.
20496 (ix86_expand_split_stack_prologue): Likewise.
20497 (legitimate_pic_address_disp_p): Likewise.
20498 (legitimize_pic_address): Likewise.
20499 (legitimize_tls_address): Likewise.
20500 (legitimize_pe_coff_symbol): Likewise.
20501 (output_pic_addr_const): Likewise.
20502 (construct_plt_address): Likewise.
20503 (ix86_expand_call): Likewise.
20504 (x86_output_mi_thunk): Likewise.
20505 (x86_function_profiler): Likewise.
20507 2013-05-13 Sofiane Naci <sofiane.naci@arm.com>
20509 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
20510 similar switch cases.
20511 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
20512 (aarch64_simd_mov_to_<mode>low): Delete.
20513 (aarch64_simd_mov_to_<mode>high): Delete.
20514 (move_lo_quad_<mode>): Add w<-r alternative.
20515 (aarch64_simd_move_hi_quad_<mode>): Likewise.
20516 (aarch64_simd_mov_from_*): Update type attribute.
20517 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
20520 2013-05-13 Jan Hubicka <jh@suse.cz>
20522 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
20523 * config/i386/i386.c (ix86_compute_frame_layout,
20524 ix86_expand_epilogue, emit_i387_cw_initialization,
20525 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
20526 ix86_local_alignment): Fix use of size/speed predicates.
20528 2013-05-13 Jakub Jelinek <jakub@redhat.com>
20530 PR tree-optimization/45216
20531 PR tree-optimization/57157
20532 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
20533 the (-Y) & (B - 1) variant if OP is |.
20534 * expmed.c (expand_shift_1): For rotations by const0_rtx just
20535 return shifted. Use (-op1) & (prec - 1) as other_amount
20536 instead of prec - op1.
20538 2013-05-13 Martin Jambor <mjambor@suse.cz>
20540 PR middle-end/42371
20541 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
20542 (ipa_constant_data): New type.
20543 (ipa_jump_func): Use ipa_constant_data to hold information about
20544 constant jump functions.
20545 (ipa_get_jf_constant): Adjust to jump function type changes.
20546 (ipa_get_jf_constant_rdesc): New function.
20547 (ipa_param_descriptor): New field controlled_uses.
20548 (ipa_get_controlled_uses): New function.
20549 (ipa_set_controlled_uses): Likewise.
20550 * ipa-ref.h (ipa_find_reference): Declare.
20551 * ipa-prop.c (ipa_cst_ref_desc): New type.
20552 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
20554 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
20555 New parameter cs. Adjust all callers.
20556 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
20557 (remove_described_reference): New function.
20558 (jfunc_rdesc_usable): Likewise.
20559 (try_make_edge_direct_simple_call): Decrement controlled use count,
20560 attempt to remove reference if it hits zero.
20561 (combine_controlled_uses_counters): New function.
20562 (propagate_controlled_uses): Likewise.
20563 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
20564 (ipa_edge_duplication_hook): Duplicate reference descriptions.
20565 (ipa_print_node_params): Print described use counter.
20566 (ipa_write_jump_function): Adjust to jump function type changes.
20567 (ipa_read_jump_function): New parameter CS, pass it to
20568 ipa_set_jf_constant. Adjust caller.
20569 (ipa_write_node_info): Stream controlled use count
20570 (ipa_read_node_info): Likewise.
20571 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
20573 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
20574 count. Remove cloning-added reference if it reaches zero.
20575 * ipa-ref.c (ipa_find_reference): New function.
20577 2013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
20579 * config/i386/i386.c (processor_target_table): Modified default
20580 alignment values for AMD BD and BT architectures.
20582 2013-05-13 Marc Glisse <marc.glisse@inria.fr>
20584 * tree-vect-generic.c (uniform_vector_p): Move ...
20585 * tree.c (uniform_vector_p): ... here.
20586 * tree.h (uniform_vector_p): Declare it.
20587 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
20590 2013-05-13 Jakub Jelinek <jakub@redhat.com>
20592 PR tree-optimization/57230
20593 * tree-ssa-strlen.c (handle_char_store): Record length for
20594 array store from STRING_CST.
20596 PR tree-optimization/57230
20597 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
20600 2013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
20602 * config/epiphany/epiphany.c (epiphany_init): Check size of
20603 NUM_MODES_FOR_MODE_SWITCHING.
20604 (epiphany_expand_prologue):
20605 Remove CONFIG_REGNUM initial value handling code.
20606 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
20607 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
20608 (emit_set_fp_mode, epiphany_mode_after): Likewise.
20609 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
20610 Don't return 1 for FP_MODE_NONE.
20611 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
20612 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
20613 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
20614 * config/epiphany/epiphany.md (save_config): New pattern.
20616 2013-05-12 Uros Bizjak <ubizjak@gmail.com>
20618 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
20620 2013-05-10 Uros Bizjak <ubizjak@gmail.com>
20622 * config/i386/i386.md (memory): Handle sseishft1.
20623 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
20624 (*vec_extractv2di_1): Ditto.
20626 2013-05-10 Vladimir Makarov <vmakarov@redhat.com>
20628 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
20631 2013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
20633 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
20634 Add mthumb/march=armv7-a multilib.
20635 Add mthumb/march=armv7-r multilib.
20636 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
20638 2013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
20640 * config/v850/t-rtems: Add more multilibs.
20642 2013-05-10 Richard Biener <rguenther@suse.de>
20644 PR tree-optimization/57214
20645 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
20646 not propagate from SSA names that occur in abnormal PHI nodes.
20648 2013-05-10 Marc Glisse <marc.glisse@inria.fr>
20650 * stor-layout.c (element_precision): New function.
20651 * machmode.h (element_precision): Declare it.
20652 * tree.c (build_minus_one_cst): New function.
20653 (element_precision): Likewise.
20654 * tree.h (build_minus_one_cst): Declare new function.
20655 (element_precision): Likewise.
20656 * fold-const.c (operand_equal_p): Use element_precision.
20657 (fold_binary_loc): Handle vector types.
20658 * convert.c (convert_to_integer): Use element_precision.
20659 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
20662 2013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
20664 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
20665 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
20666 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
20667 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
20668 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
20669 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
20670 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
20671 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
20672 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
20673 (Uuw8): New constraints.
20674 (Usb4): Move into alphabetical order.
20675 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
20676 (sd8_operand, ub8_operand, uw8_operand): New predicates.
20677 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
20678 previously unnamed patterns.
20679 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
20680 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
20681 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
20682 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
20683 of set_attr_alternative/if_then_else. Use extended_mips16 instead
20684 of specific lengths.
20686 2013-05-10 Jakub Jelinek <jakub@redhat.com>
20688 * config/i386/i386.md (rotateinv): New code attr.
20689 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
20690 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
20691 roll $31, %eax, etc.
20693 PR tree-optimization/45216
20694 PR tree-optimization/57157
20695 * tree-ssa-forwprop.c (simplify_rotate): New function.
20696 (ssa_forward_propagate_and_combine): Call it.
20698 2013-05-10 Richard Biener <rguenther@suse.de>
20700 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
20701 disable peeling when we version for aliasing.
20702 (vector_alignment_reachable_p): Honor explicit user alignment.
20703 (vect_supportable_dr_alignment): Likewise.
20704 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
20705 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
20706 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
20707 then peeling to arrange for the cost-model check to come first.
20709 2013-05-10 Alan Modra <amodra@gmail.com>
20711 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
20712 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
20713 * configure: Regenerate.
20715 2013-05-10 Alan Modra <amodra@gmail.com>
20718 * varasm.c (default_elf_select_section): Move !DECL_P check..
20719 (get_named_section): ..to here before calling get_section_name.
20721 (default_section_type_flags): Add DECL_P check.
20722 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
20723 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
20725 2013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
20727 * config/epiphany/epiphany.c (epiphany_expand_prologue):
20728 When using gen_stack_adjust_str with a register offset, add a
20729 REG_FRAME_RELATED_EXPR note.
20731 2013-05-09 Uros Bizjak <ubizjak@gmail.com>
20733 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
20734 (*vec_extractv4si_zext_mem): Ditto.
20735 (*vec_extractv2di): Add 0->x and x->x alternatives.
20736 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
20737 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
20739 2013-05-09 Jason Merrill <jason@redhat.com>
20741 N3639 C++1y VLA support
20742 * gimplify.c (gimplify_vla_decl): Don't touch an existing
20745 * tree.c (build_constructor_va): New.
20746 * tree.h: Declare it.
20748 2013-05-09 Martin Jambor <mjambor@suse.cz>
20751 * gimple-fold.c (canonicalize_constructor_val): Call
20752 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
20754 2013-05-09 Jan Hubicka <jh@suse.cz>
20755 Richard Biener <rguenther@suse.de>
20758 * symtab.c (symtab_make_decl_local): Do not add private names.
20760 2013-05-09 Jan Hubicka <jh@suse.cz>
20763 * symtab.c (insert_to_assembler_name_hash): Handle clones.
20764 (unlink_from_assembler_name_hash): Likewise.
20765 (symtab_prevail_in_asm_name_hash, symtab_register_node,
20766 symtab_unregister_node, symtab_initialize_asm_name_hash,
20767 change_decl_assembler_name): Update.
20769 2013-05-09 Sofiane Naci <sofiane.naci@arm.com>
20771 * config/aarch64/aarch64.md: New movtf split.
20772 (*movtf_aarch64): Update.
20773 (aarch64_movdi_tilow): Handle TF modes and rename to
20774 aarch64_movdi_<mode>low.
20775 (aarch64_movdi_tihigh): Handle TF modes and rename to
20776 aarch64_movdi_<mode>high
20777 (aarch64_movtihigh_di): Handle TF modes and rename to
20778 aarch64_mov<mode>high_di
20779 (aarch64_movtilow_di): Handle TF modes and rename to
20780 aarch64_mov<mode>low_di
20781 (aarch64_movtilow_tilow): Remove spurious whitespace.
20782 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
20784 (aarch64_print_operand): Update.
20786 2013-05-09 Alan Modra <amodra@gmail.com>
20788 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
20790 * configure: Regenerate.
20792 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
20794 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
20795 splitter preparation statements.
20796 * config/i386/sse.md (*vec_extract* splitters): Ditto.
20797 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
20800 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20802 * gimple-ssa-strength-reduction.c (count_candidates): Change
20803 return value to int.
20804 (analyze_candidates_and_replace): Change type of length to int.
20806 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
20808 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
20809 (*vec_extract<mode>): Use VI12_128 mode iterator.
20810 (*vec_extract<mode>_mem): Ditto.
20811 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
20814 2013-05-08 Diego Novillo <dnovillo@google.com>
20819 2012-08-17 Diego Novillo <dnovillo@google.com>
20822 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
20823 * config.in: Regenerate.
20824 * configure: Regenerate.
20825 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
20827 2013-05-08 Jan Hubicka <jh@suse.cz>
20830 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
20831 * cgraph.h (symtab_node_base): Add unique_name.
20832 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
20833 input_overwrite_node, input_varpool_node): Stream unique_name.
20834 * cgraphclones.c (cgraph_create_virtual_clone,
20835 cgraph_function_versioning): Set unique_name.
20836 * ipa.c (function_and_variable_visibility): Set unique_name.
20838 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20840 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
20841 (alloc_cand_and_find_basis): Restrict conditional candidate
20842 processing to CAND_MULTs.
20844 2013-05-08 Jan Hubicka <jh@suse.cz>
20847 lto-symtab.c (lto_symtab_symbol_p): New function.
20848 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
20849 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
20850 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
20851 Skip static symbols.
20853 2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
20855 PR tree-optimization/57200
20856 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
20857 Only call inform if the preceding warning_at returns true.
20859 2013-05-07 Han Shen <shenhan@google.com>
20861 * cfgexpand.c (record_or_union_type_has_array_p): New function.
20862 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
20863 * common.opt (fstack-protector-strong): New option.
20864 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
20865 * doc/invoke.texi (Optimization Options): Document
20866 "-fstack-protector-strong".
20867 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
20869 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
20871 * config/mips/mips.c (mips_machine_reorg2): Return 0.
20873 2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
20875 * ira.c (update_equiv_regs): Add insn having equiv memory even if
20876 it is not lhs of the insn.
20877 (setup_reg_equiv): Remove insn having equiv memory which it is not
20879 * lra-constraints.c (process_address): Try to improve generation
20880 code for address base + disp.
20881 (lra_constraints): Make correct the code for checking insn setting
20882 up backward equivalence. Remove insn only if it is in the init
20884 * lra-eliminations.c (update_reg_eliminate): Change return value.
20885 (lra_eliminate): Use the result.
20887 2013-05-07 Uros Bizjak <ubizjak@gmail.com>
20889 * config/i386/sse.md (ssescalarnummask): New mode attribute.
20890 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
20891 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
20892 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
20893 register target operands.
20894 (*vec_extractv8hi_sse2): New pattern.
20895 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
20896 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
20897 (*vec_extract<mode>_mem): New insn and split pattern.
20899 2013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
20901 * config/arm/arm.c (arm_asan_shadow_offset): New function.
20902 (TARGET_ASAN_SHADOW_OFFSET): Define.
20903 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
20904 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
20906 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20908 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
20909 (incr_vec_index): Return -1 if increment not found.
20910 (create_add_on_incoming_edge): Assert if increment not found.
20911 (record_increment): Limit number of increments recorded.
20912 (all_phi_incrs_profitable): Return false if an increment not found.
20913 (replace_profitable_candidates): Don't process increments that were
20915 (analyze_candidates_and_replace): Limit size of incr_vec.
20917 2013-05-07 Richard Biener <rguenther@suse.de>
20919 * calls.c (special_function_p): setjmp-like functions are leaf.
20920 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
20921 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
20923 2013-05-07 Sofiane Naci <sofiane.naci@arm.com>
20925 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
20926 (aarch64_simd_mov<mode>): New expander.
20927 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
20928 (aarch64_simd_mov_to_<mode>high): Likewise.
20929 (aarch64_simd_mov_from_<mode>low): Likewise.
20930 (aarch64_simd_mov_from_<mode>high): Likewise.
20931 (aarch64_dup_lane<mode>): Update.
20932 (aarch64_dup_lanedi): New instruction pattern.
20933 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
20934 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
20936 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20938 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
20939 (replace_mult_candidate): Remove unnecessary argument; remove
20940 unnecessary parameter from call to introduce_cast_before_cand.
20941 (replace_unconditional_candidate): Remove unnecessary parameter
20942 from call to replace_mult_candidate.
20943 (replace_conditional_candidate): Likewise.
20944 (insert_initializers): Use make_temp_ssa_name.
20945 (introduce_cast_before_cand): Remove unnecessary argument; use
20946 make_temp_ssa_name.
20947 (replace_one_candidate): Remove unnecessary argument; remove
20948 unnecessary parameter from calls to introduce_cast_before_cand.
20949 (replace_profitable_candidates): Remove unnecessary parameters
20950 from calls to replace_one_candidate.
20952 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20954 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
20955 phi def as possibly hiding a basis for a CAND_ADD whose operands
20956 have been commuted in the analysis.
20957 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
20959 2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
20961 * config/aarch64/aarch64.md
20962 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
20963 shift value between 0-4.
20965 2013-05-07 Richard Biener <rguenther@suse.de>
20967 * double-int.h (rshift): New overload.
20968 * double-int.c (rshift): New function.
20969 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
20970 (create_reference_ops_from_ref): Remove.
20971 (vn_reference_insert): Use shared ops for constructing the
20972 reference and copy it.
20974 2013-05-07 Richard Biener <rguenther@suse.de>
20976 PR middle-end/57190
20977 * tree-eh.c (sink_clobbers): Properly propagate
20978 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
20980 2013-05-07 Jakub Jelinek <jakub@redhat.com>
20982 PR tree-optimization/57149
20983 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
20984 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
20985 collect_phi_def_edges, execute_late_warn_uninitialized): Use
20986 uninit_undefined_value_p instead of ssa_undefined_value_p.
20989 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
20990 for modifier == EXPAND_INITIALIZER.
20992 2013-05-07 Anton Blanchard <anton@samba.org>
20994 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
20995 for powerpc64 little endian.
20996 * configure: Regenerate.
20998 2013-05-06 Graham Stott <grahams@btinternet.com>
21000 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
21001 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
21002 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
21003 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
21005 2013-05-06 Graham Stott <grahams@btinternet.com>
21007 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
21008 codes which allow non-lvalues.
21010 2013-05-06 Marc Glisse <marc.glisse@inria.fr>
21012 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
21013 components are all 1s.
21014 (integer_minus_onep): New function.
21015 * tree.h (integer_minus_onep): Declare it.
21016 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
21017 integer_minus_onep instead of integer_all_onesp.
21019 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
21022 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
21023 variations of these patterns.
21025 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
21027 * config/i386/i386.md (isa): Add x64_sse4 member.
21028 (enabled): Handle x64_sse4.
21029 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
21030 instruction for 64bit SSE4_1 targets. Update insn attributes.
21031 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
21032 instruction for SSE4_1 targets. Update insn attributes.
21033 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
21034 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
21035 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
21037 (*vec_extractv4si): Rename from *sse4_1_pextrd.
21038 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
21039 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
21041 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
21044 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
21046 2013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
21048 * final.c (do_assembler_dialects): Don't handle curly braces and
21049 vertical bar escaped by % as dialect delimiters.
21050 (output_asm_insn): Print curly braces and vertical bar if escaped
21051 by % and ASSEMBLER_DIALECT defined.
21052 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
21053 * doc/tm.texi: Regenerated.
21055 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
21057 * config/mips/mips.c: Include tree-pass.h.
21058 (mips_reorg): Split in pre- and post-dbr_schedule parts.
21059 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
21060 (pass_mips_machine_reorg2): New machine specific pass.
21061 (insert_pass_mips_machine_reorg2): New pass plugin definition.
21062 (mips_option_override): Register the new pass.
21063 * rtl.h (cleanup_barriers): Remove prototype.
21064 (dbr_schedule): Likewise.
21065 * jump.c (cleanup_barriers): Make static.
21066 * reorg.c (dbr_schedule): Likewise.
21068 2013-05-06 Richard Biener <rguenther@suse.de>
21070 PR tree-optimization/57185
21071 * tree-parloops.c (add_field_for_reduction): Handle anonymous
21072 SSA names properly.
21074 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
21077 * config/i386/i386.c (add_parameter_dependencies): Add dependence
21078 between "first_arg" and "insn", not "last" and "insn".
21080 2013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
21082 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
21083 (find_candidates_in_block): Re-enable slsr_process_phi.
21084 (create_phi_basis): Fix double counting of candidate adjustment.
21086 2013-05-06 Richard Biener <rguenther@suse.de>
21088 PR middle-end/57147
21089 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
21090 the edge is also fallthru, preserve it and just clear the
21092 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
21093 also complex, preserve that and just clear the fallthru flag.
21094 * tree-inline.c (update_ssa_across_abnormal_edges): Also
21095 update virtual operands.
21097 2013-05-06 Alan Modra <amodra@gmail.com>
21099 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
21100 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
21101 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
21102 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
21103 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
21104 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
21106 2013-05-06 Alan Modra <amodra@gmail.com>
21108 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
21109 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
21110 (DEFAULT_ASM_ENDIAN): Define.
21111 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
21112 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
21113 Update -K PIC clause from sysv4.h.
21114 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
21115 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
21117 2013-05-06 Alan Modra <amodra@gmail.com>
21119 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
21120 twice for little-endian.
21121 (ashrdi3_no_power, ashrdi3): Support little-endian.
21123 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
21126 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
21127 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
21129 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
21130 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
21131 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
21134 2013-05-05 Steven Bosscher <steven@gcc.gnu.org>
21135 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
21137 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
21138 * config/pa/pa.opt: Make mbig-switch a no-op.
21139 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
21140 (CASE_VECTOR_MODE): Always return SImode.
21141 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
21142 for the !TARGET_BIG_SWITCH case.
21143 * config/pa/pa-linux.h: Likewise.
21144 * config/pa/pa-openbsd.h: Likewise.
21145 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
21146 * config/pa/pa.md (short_jump): Remove define_insn.
21147 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
21148 (casesi0): Remove define_insn.
21149 (type): Remove btable_branch.
21150 (pa_combine_type): Likewise.
21151 (in_nullified_branch_delay): Likewise.
21152 (in_call_delay): Likewise.
21153 (define_delay): Likewise.
21154 (define_insn_reservation "Z3"): Likewise.
21155 (define_insn_reservation "Z4"): Likewise.
21156 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
21157 (pa_adjust_insn_length): Remove adjustment for btable branches.
21158 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
21161 2013-05-05 Uros Bizjak <ubizjak@gmail.com>
21163 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
21164 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
21165 Add m->r,x alternatives.
21166 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
21167 splitters using SWI48x mode iterator.
21168 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
21169 TARGET_64BIT. Add m->x alternative.
21170 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
21171 Add o->x alternative. Enable for TARGET_SSE.
21172 (sse_storeq): Remove expander.
21173 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
21174 with memory input operand.
21175 (*vec_extractv2di_1 splitter): New.
21176 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
21177 * config/i386/i386.md (ssevecmodelower): New mode attribute.
21179 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
21181 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
21182 (INT_LOWPART): Delete.
21183 (extract_MB): Adjust.
21184 (extract_ME): Adjust.
21185 (print_operand): Adjust.
21187 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
21189 * config/rs6000/predicates.md (reg_or_add_cint_operand,
21190 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
21191 (reg_or_logical_cint_operand, easy_fp_constant,
21192 logical_const_operand): Delete "CONST_DOUBLE" case.
21193 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
21194 "HOST_BITS_PER_WIDE_INT == 64" test.
21195 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
21196 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
21197 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
21198 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
21200 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
21201 CONST_DOUBLE DImode/VOIDmode case.
21202 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
21203 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
21204 CONST_DOUBLE VOIDmode case.
21205 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
21206 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
21207 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
21208 Delete CONST_DOUBLE case.
21209 (splitters for mov FMOVE64 const_double): Delete
21210 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
21211 "HOST_BITS_PER_WIDE_INT >= 64" test.
21212 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
21214 (mov DI const_double): Delete.
21216 2013-05-04 Jakub Jelinek <jakub@redhat.com>
21218 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
21219 on op shows all bits zero in mode of a lowpart subreg, return zero.
21221 2013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
21224 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
21225 to save TFmode registers and DImode to save TImode registers for
21226 caller save operations.
21227 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
21228 mark being partially clobbered since they only use the first
21231 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
21232 and TDmode only use the upper 64-bits of each VSX register.
21234 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
21236 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
21237 (find_candidates_in_block): Disable slsr_process_phi.
21239 2013-05-03 Guozhi Wei <carrot@google.com>
21241 * coverage.c (coverage_obj_init): Move the construction of gcov
21243 (build_init_ctor): ... here.
21245 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
21247 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
21248 (slsr_cand_d): Redefine def_phi.
21249 (stride_status, phi_adjust_status, count_phis_status): New enums.
21250 (find_phi_def): New.
21251 (find_basis_for_base_expr): New.
21252 (find_basis_for_candidate): Handle hidden bases.
21253 (alloc_cand_and_find_basis): Handle phi candidates.
21254 (slsr_process_phi): New.
21255 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
21256 (create_mul_imm_cand): Likewise.
21257 (create_add_ssa_cand): Exclude phi base candidates.
21258 (create_add_imm_cand): Likewise.
21259 (slsr_process_cast): Likewise.
21260 (slsr_process_copy): Likewise.
21261 (find_candidates_in_block): Handle phi candidates.
21262 (dump_candidate): Likewise.
21263 (unconditional_cands): Delete.
21264 (unconditional_cands_with_known_stride_p): Delete.
21265 (phi_dependent_cand_p): New.
21266 (cand_increment): Handle phi-dependent candidates.
21267 (replace_dependent): Delete.
21268 (replace_mult_candidate): New.
21269 (replace_unconditional_candidate): New.
21270 (incr_vec_index): Move to avoid forward reference.
21271 (create_add_on_incoming_edge): New.
21272 (create_phi_basis): New.
21273 (replace_dependents): Delete.
21274 (replace_conditional_candidate): New.
21275 (phi_add_costs): New.
21276 (replace_uncond_cands_and_profitable_phis): New.
21277 (record_increment): Handle phi adjustments.
21278 (record_phi_increments): New.
21279 (record_increments): Handle phi adjustments.
21280 (phi_incr_cost): New.
21281 (lowest_cost_path): Handle phis.
21282 (total_savings): Likewise.
21283 (analyze_increments): Likewise.
21284 (ncd_with_phi): New.
21285 (ncd_of_cand_and_phis): New.
21286 (nearest_common_dominator_for_cands): Handle phi increments.
21287 (all_phi_incrs_profitable): New.
21288 (replace_profitable_candidates): Handle phi-dependent candidates.
21289 (analyze_candidates_and_replace): Likewise.
21291 2013-05-03 Teresa Johnson <tejohnson@google.com>
21294 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
21295 do not exceed REG_BR_PROB_BASE.
21297 2013-05-03 Jeff Law <law@redhat.com>
21299 PR tree-optimization/57144
21300 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
21301 operand of the condition will bit into the new type when eliminating
21302 a cast feeding a condition.
21304 2013-05-03 Jakub Jelinek <jakub@redhat.com>
21306 PR rtl-optimization/57130
21307 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
21308 of COMPARE as in_code to the recursive call if needed.
21310 2013-05-03 Uros Bizjak <ubizjak@gmail.com>
21312 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
21313 (enabled): Handle new members.
21314 * config/i386/sse.md (*vec_concatv2si): Merge from
21315 *vec_concatv2si_sse2 and vec_concatv2si_sse.
21316 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
21318 2013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
21320 PR tree-optimization/57027
21321 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
21322 for fnms opportunity, check we got the prerequisite kind
21323 of tree / gimple before using accessor functions.
21325 2013-05-03 Richard Biener <rguenther@suse.de>
21327 * double-int.h (lshift): New overload without precision
21328 and arith argument.
21329 (operator *=, operator +=, operator -=): Move ...
21330 * double-int.c (operator *=, operator +=, operator -=): ... here
21331 and implement more efficiently.
21332 (mul_double_with_sign): Remove.
21333 (lshift_double): Adjust to take unsinged shift argument, push
21334 dispatching code to callers.
21335 (mul_double_wide_with_sign): Add early out for callers that
21336 are not interested in high parts or overflow.
21337 (lshift): New function.
21338 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
21339 dispatch code here.
21340 (lrotate, rrotate): Use logical shifts.
21341 * expr.c (get_inner_reference): Use lshift.
21342 * fixed-value.c (do_fixed_divide): Likewise.
21343 * tree-dfa.c (get_ref_base_and_extent): Likewise.
21344 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
21345 (indirect_refs_may_alias_p): Likewise.
21346 (stmt_kills_ref_p_1): Likewise.
21348 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
21350 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
21352 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
21354 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
21355 scalar form of FABD instruction.
21357 2013-05-02 Vladimir Makarov <vmakarov@redhat.com>
21359 * lra-constraints.c (process_alt_operands): Add checking alt
21360 number to choose the best alternative.
21362 2013-05-02 Richard Biener <rguenther@suse.de>
21364 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
21365 bitmap and its handling.
21366 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
21368 2013-05-02 Richard Biener <rguenther@suse.de>
21370 PR middle-end/57140
21371 * tree-inline.c (copy_loops): Properly handle removed loops.
21372 (copy_cfg_body): Mark destination loops for fixup if source
21373 loops needed fixup.
21375 2013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
21378 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
21379 generating simple_return for naked functions.
21381 2013-05-02 Martin Jambor <mjambor@suse.cz>
21383 PR middle-end/56988
21384 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
21385 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
21387 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
21388 ipa_agg_replacement_value structures.
21389 (known_aggs_to_agg_replacement_list): Likewise.
21390 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
21391 (read_agg_replacement_chain): Likewise.
21392 (ipcp_transform_function): Also check that by_ref flags match.
21394 2013-05-02 Richard Biener <rguenther@suse.de>
21396 * graphds.h (struct graph): Add obstack member.
21397 * graphds.c (new_graph): Initialize obstack and allocate
21399 (add_edge): Allocate edge from the obstack.
21400 (free_graph): Free the obstack instead of all edges and vertices.
21402 2013-05-02 Teresa Johnson <tejohnson@google.com>
21404 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
21406 * cfg.c (update_bb_profile_for_threading): Ditto.
21407 * tree-inline.c (copy_bb): Ditto.
21408 (copy_edges_for_bb): Ditto.
21409 (initialize_cfun): Ditto.
21410 (copy_cfg_body): Ditto.
21411 (expand_call_inline): Ditto.
21412 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
21413 (estimate_node_size_and_time): Ditto.
21414 (inline_merge_summary): Ditto.
21415 * cgraphclones.c (cgraph_clone_edge): Ditto.
21416 (cgraph_clone_node): Ditto.
21417 * sched-rgn.c (compute_dom_prob_ps): Ditto.
21418 (compute_trg_info): Ditto.
21420 2013-05-02 Ian Bolton <ian.bolton@arm.com>
21422 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
21423 S reg when fp attribute set.
21424 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
21426 2013-05-02 Ian Bolton <ian.bolton@arm.com>
21428 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
21430 (*and_one_cmplsi3_compare0_uxtw): Likewise.
21431 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
21432 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
21434 2013-05-02 Richard Biener <rguenther@suse.de>
21436 * tree-scalar-evolution.c (scev_info_hasher): Remove.
21437 (struct instantiate_cache_entry): New type.
21438 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
21439 (struct instantiate_cache_type): New type.
21440 (set_instantiated_value, get_instantiated_value): Remove.
21441 (get_instantiated_value_entry): New function.
21442 (instantiate_scev_name): Use the new cache and adjust.
21443 (instantiate_scev_poly): Adjust.
21444 (instantiate_scev_binary): Likewise.
21445 (instantiate_array_ref): Likewise.
21446 (instantiate_scev_convert): Likewise.
21447 (instantiate_scev_not): Likewise.
21448 (instantiate_scev_3): Likewise.
21449 (instantiate_scev_2): Likewise.
21450 (instantiate_scev_r): Likewise.
21451 (instantiate_scev): Likewise.
21452 (resolve_mixers): Likewise.
21454 2013-05-01 Vladimir Makarov <vmakarov@redhat.com>
21457 * lra-constraints.c (best_small_class_operands_num): Remove.
21458 (process_alt_operands): Remove small_class_operands_num. Take
21459 small classes operands into losers and only if the operand is not
21460 matched. Modify debugging output.
21461 (curr_insn_transform): Remove best_small_class_operands_num.
21464 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21466 * config/aarch64/aarch64-builtins.c
21467 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
21468 * config/aarch64/aarch64-simd-builtins.def
21469 (reduc_splus_): Add new modes.
21470 (reduc_uplus_): New.
21471 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
21472 (reduc_uplus_v4sf): Likewise.
21473 (reduc_splus_v4sf): Likewise.
21474 (aarch64_addv<mode>): Likewise.
21475 (reduc_uplus_<mode>): Likewise.
21476 (reduc_splus_<mode>): Likewise.
21477 (aarch64_addvv2di): Likewise.
21478 (reduc_uplus_v2di): Likewise.
21479 (reduc_splus_v2di): Likewise.
21480 (aarch64_addvv2si): Likewise.
21481 (reduc_uplus_v2si): Likewise.
21482 (reduc_splus_v2si): Likewise.
21483 (reduc_<sur>plus_<mode>): New.
21484 (reduc_<sur>plus_v2di): Likewise.
21485 (reduc_<sur>plus_v2si): Likewise.
21486 (reduc_<sur>plus_v4sf): Likewise.
21487 (aarch64_addpv4sf): Likewise.
21488 * config/aarch64/arm_neon.h
21489 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
21490 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
21491 add UNSPEC_SADDV, UNSPEC_UADDV.
21493 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
21495 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21497 * config/aarch64/arm_neon.h
21498 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
21500 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21502 * config/aarch64/aarch64-builtins
21503 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
21505 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21507 * config/aarch64/aarch64-simd-builtins.def
21508 (reduc_smax_): New.
21509 (reduc_smin_): Likewise.
21510 (reduc_umax_): Likewise.
21511 (reduc_umin_): Likewise.
21512 (reduc_smax_nan_): Likewise.
21513 (reduc_smin_nan_): Likewise.
21516 (smax): Update for V2SF, V4SF and V2DF modes.
21519 (smin_nan): Likewise.
21520 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
21521 (<su><maxmin><mode>3): ...This, refactor.
21522 (s<maxmin><mode>3): New.
21523 (<maxmin_uns><mode>3): Likewise.
21524 (reduc_<maxmin_uns>_<mode>): Refactor.
21525 (reduc_<maxmin_uns>_v4sf): Likewise.
21526 (reduc_<maxmin_uns>_v2si): Likewise.
21527 (aarch64_<fmaxmin><mode>: Remove.
21528 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
21530 (vmin<q>_f<32,64>): Likewise.
21531 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
21533 (su): Add mappings for smax, smin, umax, umin.
21535 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
21536 (FMAXMIN): Rename as...
21537 (FMAXMIN_UNS): ...This.
21539 (fmaxminv): Likewise.
21540 (fmaxmin): Likewise.
21542 (maxmin_uns_op): Likewise.
21544 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21546 * config/aarch64/arm_neon.h
21547 (vac<ge, gt><sd>_f<32, 64>): Rename to...
21548 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
21549 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
21551 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21553 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
21554 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
21556 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21558 * config/aarch64/aarch64-simd.md
21559 (vcond<mode>_internal): Handle special cases for constant masks.
21560 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
21561 (vcondu<mode><mode>): Likewise.
21562 (vcond<v_cmp_result><mode>): New.
21564 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21566 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
21567 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
21568 * config/aarch64/aarch64-simd-builtins.def
21569 (cmeq): Update to BUILTIN_VALLDI.
21574 * config/aarch64/arm_neon.h
21575 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
21576 to builtins or C as appropriate.
21578 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21580 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
21582 (cmhi): Rename to...
21584 * config/aarch64/aarch64-simd.md
21585 (simd_mode): Add SF.
21586 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
21587 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
21588 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
21589 (cstore<mode>_neg): ...This.
21590 * config/aarch64/iterators.md
21592 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
21593 (COMPARISONS): New.
21594 (UCOMPARISONS): Likewise.
21595 (optab): Add missing comparisons.
21601 (VCMP_S): Likewise.
21602 (VCMP_U): Likewise.
21603 (V_cmp_result): Add DF, SF modes.
21604 (v_cmp_result): Likewise.
21606 (vmtype): Likewise.
21607 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
21609 2013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
21611 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
21612 define_insn to define_insn_and_split.
21613 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
21614 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
21615 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
21616 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
21617 (thumb2_negscc): Likewise.
21619 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
21621 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
21623 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
21625 * config/arm/thumb2.md: Remove trailing whitespaces.
21627 2013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
21629 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
21630 Use gen_int_mode rather than GEN_INT.
21632 2013-04-30 H.J. Lu <hongjiu.lu@intel.com>
21634 * value-prof.c (stream_in_histogram_value): Remove the strayed
21637 2013-04-30 Richard Biener <rguenther@suse.de>
21639 PR middle-end/57122
21640 * cfghooks.c (split_edge): Properly check for the loop latch edge.
21642 2013-04-30 Richard Biener <rguenther@suse.de>
21644 PR middle-end/57107
21645 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
21647 2013-04-30 Andrey Belevantsev <abel@ispras.ru>
21649 PR rtl-optimization/56957
21650 PR rtl-optimization/57105
21651 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
21652 variable. Use just INSN_UID for determining whether an insn
21653 should be only disconnected from the insn stream.
21654 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
21656 2013-04-30 Jakub Jelinek <jakub@redhat.com>
21658 PR tree-optimization/57104
21659 * tsan.c (instrument_expr): Don't instrument accesses to
21660 DECL_HARD_REGISTER VAR_DECLs.
21662 2013-04-30 Richard Biener <rguenther@suse.de>
21664 * function.h (loops_for_fn): New inline function.
21665 (set_loops_for_fn): Likewise.
21666 * cfgloop.h (place_new_loop): Add struct function parameter.
21667 (get_loop): Likewise.
21668 (get_loops): Likewise.
21669 (number_of_loops): Likewise.
21670 (fel_next): Adjust.
21671 (fel_init): Likewise.
21672 * cfg.c (get_loop_copy): Adjust.
21673 * cfgloop.c (flow_loops_dump): Likewise.
21674 (record_loop_exits): Likewise.
21675 (verify_loop_structure): Likewise.
21676 * cfgloopanal.c (mark_irreducible_loops): Likewise.
21677 (estimate_reg_pressure_cost): Likewise.
21678 (mark_loop_exit_edges): Likewise.
21679 * cfgloopmanip.c (place_new_loop): Likewise.
21680 (add_loop): Likewise.
21681 (duplicate_loop): Likewise.
21682 * graph.c (draw_cfg_nodes): Likewise.
21683 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
21684 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
21685 (extract_affine_chrec): Likewise.
21686 (build_scop_iteration_domain): Likewise.
21687 * graphite.c (graphite_initialize): Likewise.
21688 * ira-build.c (create_loop_tree_nodes): Likewise.
21689 (more_one_region_p): Likewise.
21690 (rebuild_regno_allocno_maps): Likewise.
21691 (mark_loops_for_removal): Likewise.
21692 (mark_all_loops_for_removal): Likewise.
21693 (remove_unnecessary_regions): Likewise.
21694 (ira_build): Likewise.
21695 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
21696 * loop-init.c (fix_loop_structure): Likewise.
21697 (gate_rtl_move_loop_invariants): Likewise.
21698 (gate_rtl_unswitch): Likewise.
21699 (gate_rtl_unroll_and_peel_loops): Likewise.
21700 (rtl_doloop): Likewise.
21701 * lto-streamer-in.c (input_cfg): Likewise.
21702 * lto-streamer-out.c (output_cfg): Likewise.
21703 * modulo-sched.c (sms_schedule): Likewise.
21704 * predict.c (tree_estimate_probability): Likewise.
21705 (tree_estimate_probability_driver): Likewise.
21706 (estimate_loops): Likewise.
21707 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
21708 (move_sese_region_to_fn): Likewise.
21709 (debug_loop_num): Likewise.
21710 * tree-chrec.c (chrec_evaluate): Likewise.
21711 (hide_evolution_in_other_loops_than_loop): Likewise.
21712 (chrec_component_in_loop_num): Likewise.
21713 (reset_evolution_in_loop): Likewise.
21714 (evolution_function_is_invariant_rec_p): Likewise.
21715 * tree-if-conv.c (main_tree_if_conversion): Likewise.
21716 * tree-inline.c (copy_loops): Likewise.
21717 (copy_cfg_body): Likewise.
21718 (tree_function_versioning): Likewise.
21719 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
21720 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
21722 (add_to_evolution_1): Likewise.
21723 (scev_const_prop): Likewise.
21724 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
21725 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
21726 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
21727 (tree_ssa_lim_initialize): Likewise.
21728 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
21729 (verify_loop_closed_ssa): Likewise.
21730 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
21731 (tree_ssa_loop_im): Likewise.
21732 (tree_ssa_loop_unswitch): Likewise.
21733 (tree_vectorize): Likewise.
21734 (check_data_deps): Likewise.
21735 (tree_ssa_loop_ivcanon): Likewise.
21736 (tree_ssa_loop_bounds): Likewise.
21737 (tree_complete_unroll): Likewise.
21738 (tree_complete_unroll_inner): Likewise.
21739 (tree_parallelize_loops): Likewise.
21740 (tree_ssa_loop_prefetch): Likewise.
21741 (tree_ssa_loop_ivopts): Likewise.
21742 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
21743 * tree-vectorizer.c (vectorize_loops): Likewise.
21745 2013-04-29 Mike Frysinger <vapier@gentoo.org>
21747 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
21748 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
21749 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
21750 with EABI_LINK_SPEC.
21752 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
21755 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
21758 2013-04-29 Vladimir Makarov <vmakarov@redhat.com>
21761 * lra-constraints.c (process_alt_operands): Discourage a bit more
21762 using memory for pseudos. Print cost dump for alternatives.
21763 Modify cost values for conflicts with early clobbers.
21764 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
21766 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
21769 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
21771 2013-04-29 Ian Bolton <ian.bolton@arm.com>
21773 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
21774 from/to S register.
21775 (movdi_aarch64): Support LDR/STR from/to D register.
21777 2013-04-29 Ian Bolton <ian.bolton@arm.com>
21779 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
21780 or higher by default.
21782 2013-04-29 Richard Biener <rguenther@suse.de>
21784 PR middle-end/57075
21785 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
21786 even if not adding abnormal edges for calls that can make
21789 2013-04-29 Richard Biener <rguenther@suse.de>
21791 PR middle-end/57103
21792 * tree-cfg.c (move_stmt_op): Fix condition under which to update
21794 (move_stmt_r): Remove redundant checking.
21796 2013-04-29 Teresa Johnson <tejohnson@google.com>
21799 * basic-block.h (apply_scale): New function.
21800 (apply_probability): Use apply_scale.
21801 * gimple-streamer-in.c (input_bb): Ditto.
21802 * lto-streamer-in.c (input_cfg): Ditto.
21803 * lto-cgraph.c (merge_profile_summaries): Ditto.
21804 * tree-optimize.c (execute_fixup_cfg): Ditto.
21805 * tree-inline.c (copy_bb): Update comment to use apply_scale.
21806 (copy_edges_for_bb): Ditto.
21807 (copy_cfg_body): Ditto.
21809 2013-04-29 Tom de Vries <tom@codesourcery.com>
21811 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
21812 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
21813 (tail_merge_optimize): Handle current_loops == NULL.
21815 2013-04-26 Jeff Law <law@redhat.com>
21817 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
21818 (simplify_cond_using_ranges): Generalize code to simplify
21819 COND_EXPRs where one argument is a constant and the other
21820 is an SSA_NAME created by an integral type conversion.
21822 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21824 * config/arm/arm.md (store_minmaxsi): Use only when
21825 optimize_insn_for_size_p.
21827 2013-04-29 Christian Bruel <christian.bruel@st.com>
21830 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
21832 2013-04-29 Richard Biener <rguenther@suse.de>
21834 PR middle-end/57089
21835 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
21836 loop tree make sure to schedule a fixup for the child as well.
21837 (expand_omp_for_generic): Properly add loops.
21838 (expand_omp_for_static_nochunk): Likewise.
21839 (expand_omp_for_static_chunk): Likewise.
21840 (expand_omp_for): For the degenerate case fixup loops.
21841 (expand_omp_sections): Fix default bb placement in loops.
21842 (expand_omp_atomic_pipeline): Properly add loops.
21844 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21846 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
21848 2013-04-29 Tom de Vries <tom@codesourcery.com>
21850 * tree-ssa-tail-merge.c: Update header comment.
21852 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21854 * config/aarch64/arm_neon.h
21855 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
21856 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
21857 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
21858 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
21859 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
21860 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
21861 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
21862 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
21864 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21866 * config/aarch64/aarch64-simd.md
21867 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
21868 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
21869 fix_trunc, fixuns_trunc.
21870 (ftrunc<VDQF:mode>2): New.
21871 * config/aarch64/iterators.md (optab): Add fix, fixuns.
21872 (fix_trunc_optab): New.
21874 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21876 * config/aarch64/aarch64-builtins.c
21877 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
21878 iceilf, lround, iroundf.
21880 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
21883 * config/i386/i386.h (enum ix86_tune_indices)
21884 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
21885 New, split from X86_TUNE_INTER_UNIT_MOVES.
21886 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
21887 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
21888 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
21889 (TARGET_INTER_UNIT_MOVES): Remove.
21890 * config/i386/i386.c (initial_ix86_tune_features): Update.
21891 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
21892 (ix86_expand_convert_uns_didf_sse): Use
21893 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
21894 (ix86_expand_vector_init_one_nonzero): Ditto.
21895 (ix86_expand_vector_init_interleave): Ditto.
21896 (inline_secondary_memory_needed): Return true for moves from SSE class
21897 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
21898 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
21899 * config/i386/constraints.md (Yi, Ym): Depend on
21900 TARGET_INTER_UNIT_MOVES_TO_VEC.
21901 (Yj, Yn): New constraints.
21902 * config/i386/i386.md (*movdi_internal): Change constraints of
21903 operand 1 from Yi to Yj and from Ym to Yn.
21904 (*movsi_internal): Ditto.
21905 (*movdf_internal): Ditto.
21906 (*movsf_internal): Ditto.
21907 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
21908 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
21909 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
21910 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
21911 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
21912 * config/i386/sse.md (movdi_to_sse): Ditto.
21913 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
21914 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
21915 TARGET_INTER_UNIT_MOVES.
21916 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
21917 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
21918 instead of TARGET_INTER_UNIT_MOVES.
21919 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
21920 operand 1 from Yi to Yj and from Ym to Yn.
21922 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21924 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
21925 (float_truncate_hi_): Likewise.
21926 (float_extend_lo_): Likewise.
21927 (float_truncate_lo_): Likewise.
21928 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
21929 (aarch64_float_extend_lo_v2df): Likewise.
21930 (vec_unpacks_hi_v4sf): Likewise.
21931 (aarch64_float_truncate_lo_v2sf): Likewise.
21932 (aarch64_float_truncate_hi_v4sf): Likewise.
21933 (vec_pack_trunc_v2df): Likewise.
21934 (vec_pack_trunc_df): Likewise.
21936 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21938 * config/aarch64/aarch64-builtins.c
21939 (aarch64_fold_builtin): Fold float conversions.
21940 * config/aarch64/aarch64-simd-builtins.def
21941 (floatv2si, floatv4si, floatv2di): New.
21942 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
21943 * config/aarch64/aarch64-simd.md
21944 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
21945 * config/aarch64/iterators.md (FLOATUORS): New.
21946 (optab): Add float, floatuns.
21947 (su_optab): Likewise.
21949 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21951 * config/aarch64/aarch64-builtins.c
21952 (aarch64_builtin_vectorized_function): Use new names for
21954 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
21955 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
21956 (fcvtzu): Split as...
21957 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
21958 (fcvtas): Split as...
21959 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
21960 (fcvtau): Split as...
21961 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
21962 (fcvtps): Split as...
21963 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
21964 (fcvtpu): Split as...
21965 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
21966 (fcvtms): Split as...
21967 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
21968 (fcvtmu): Split as...
21969 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
21970 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
21971 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
21972 (lfrintnusf, lfrintnudf): Likewise.
21973 * config/aarch64/aarch64-simd.md
21974 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
21976 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
21977 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
21978 (fcvt_pattern): Likewise.
21980 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21982 * config/aarch64/aarch64-simd.md
21983 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
21984 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
21986 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21988 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
21989 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
21990 (vrnd<a,m,n,p>_f32): Implement using builtins.
21991 (vrnd<i,x><q>_f<32, 64>): New.
21993 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21995 * config/aarch64/aarch64-builtins.c
21996 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
21997 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
21998 (frintz): Rename to...
22000 (frintp): Rename to...
22002 (frintm): Rename to...
22004 (frinti): Rename to...
22005 (nearbyint): ...this.
22006 (frintx): Rename to...
22008 (frinta): Rename to...
22010 * config/aarch64/aarch64-simd.md
22011 (aarch64_frint<frint_suffix><mode>): Delete.
22012 (<frint_pattern><mode>2): Convert to insn.
22013 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
22014 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
22015 (frint_pattern): Likewise.
22016 (frint_suffix): Likewise.
22018 2013-04-29 Richard Biener <rguenther@suse.de>
22020 PR tree-optimization/57081
22021 * loop-init.c: Include tree-flow.h.
22022 (loop_optimizer_finalize): Free number of iteration estimates.
22023 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
22025 2013-04-29 Jakub Jelinek <jakub@redhat.com>
22027 PR tree-optimization/57083
22028 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
22029 non-singleton shift count range, zero extend low_bound for uns case.
22031 * config/i386/predicates.md (general_vector_operand): New predicate.
22032 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
22033 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
22034 if they aren't nonimmediate operands. If their original values
22035 satisfy const_vector_equal_evenodd_p, don't shift them.
22036 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
22037 predicates. For the SSE4.1 case force operands[{1,2}] into registers
22038 if not nonimmediate_operand.
22039 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
22040 instead of register_operand.
22041 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
22043 2013-04-28 Eric Botcazou <ebotcazou@adacore.com>
22045 * stor-layout.c (finalize_size_functions): Allocate a structure and
22046 reset cfun before dumping the functions.
22048 2013-04-27 Jakub Jelinek <jakub@redhat.com>
22050 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
22053 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
22054 use xop_pmacsdqh if uns_p.
22055 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
22056 the immediate rotate count.
22058 2013-04-26 Vladimir Makarov <vmakarov@redhat.com>
22060 * rtl.h (struct rtx_def): Add comment for field jump.
22061 (LRA_SUBREG_P): New macro.
22062 * recog.c (register_operand): Check LRA_SUBREG_P.
22063 * lra.c (lra): Add note at the end of RTL code. Align non-empty
22065 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
22066 (lra_final_code_change): Skip subreg change for operators.
22067 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
22068 if there are no operand changes.
22069 * lra-constraints.c (curr_insn_set): New.
22070 (match_reload): Set LRA_SUBREG_P.
22071 (emit_spill_move): Ditto.
22072 (check_and_process_move): Use curr_insn_set. Process only single
22073 set insns. Don't initialize sec_mem_p and change_p.
22074 (simplify_operand_subreg): Use LRA_SUBREG_P.
22075 (reg_in_class_p): New function.
22076 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
22077 of #ifdef. Add code to remove cycling.
22078 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
22079 non-null disp. Reload inner instead of disp when base and index
22080 are null. Try to put lo_sum into register.
22081 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
22082 (check_and_process_move): Move code for move cost check to
22083 simple_move_p. Remove equiv_substitution.
22084 (simple_move_p): New function.
22085 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
22086 curr_insn_set. Call check_and_process_move only for single set
22087 insns. Use the new function. Move call of check_and_process_move
22088 after operand equiv substitution and address process.
22090 2013-04-26 Jakub Jelinek <jakub@redhat.com>
22093 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
22094 with nonlocal goto receivers or returns twice calls, ignore
22095 unininitialized values from abnormal edges to nl goto receiver
22096 or returns twice call.
22098 2013-04-26 Jakub Jelinek <jakub@redhat.com>
22100 PR tree-optimization/57051
22101 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
22102 and VEC_RSHIFT_EXPR if shift count is a multiple of element
22105 2013-04-26 Richard Biener <rguenther@suse.de>
22107 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
22108 (expand_omp_taskreg): Likewise. Mark loops for fixup.
22109 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
22110 (fixup_loop_arrays_after_move): New function.
22111 (move_sese_region_to_fn): Properly outline the loop tree parts
22112 of the SESE region.
22114 2013-04-26 Uros Bizjak <ubizjak@gmail.com>
22116 * config/i386/i386.md (type, unit): Fix long lines.
22118 2013-04-26 Richard Biener <rguenther@suse.de>
22120 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
22121 (lto-streamer-out.o): Likewise.
22122 * cfgloop.c (init_loops_structure): Export, add struct function
22123 argument and adjust.
22124 (flow_loops_find): Adjust.
22125 * cfgloop.h (enum loop_estimation): Add EST_LAST.
22126 (init_loops_structure): Declare.
22127 * lto-streamer-in.c: Include cfgloop.h.
22128 (input_cfg): Input the loop tree.
22129 * lto-streamer-out.c: Include cfgloop.h.
22130 (output_cfg): Output the loop tree.
22131 (output_struct_function_base): Do not drop PROP_loops.
22133 2013-03-26 Richard Biener <rguenther@suse.de>
22135 * tree-cfg.c (execute_build_cfg): Build the loop tree.
22136 (pass_build_cfg): Provide PROP_loops.
22137 (move_sese_region_to_fn): Remove loops that are outlined into fn
22139 * tree-inline.c: Include cfgloop.h.
22140 (initialize_cfun): Do not drop PROP_loops.
22141 (copy_loops): New function.
22142 (copy_cfg_body): Copy loop structure.
22143 (tree_function_versioning): Initialize destination loop tree.
22144 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
22145 (pass_parallelize_loops): Do IL verification.
22146 * loop-init.c (loop_optimizer_init): Fixup loops if required.
22147 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
22148 the CFG make sure we fixup loops as well.
22149 * tree-ssa-tail-merge.c: Include cfgloop.h.
22150 (replace_block_by): When merging loop latches mark loops for fixup.
22151 * lto-streamer-out.c (output_struct_function_base): Drop
22152 PROP_loops for now.
22153 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
22154 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
22155 * ipa-split.c: Include cfgloop.h.
22156 (split_function): Add the new return block to the loop tree root.
22157 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
22158 whether we have removed the forwarder block.
22159 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
22160 * cfgloop.h (place_new_loop): Declare.
22161 * cfgloopmanip.c (place_new_loop): Export.
22162 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
22163 (tree-switch-conversion.o): Likewise.
22164 (tree-complex.o): Likewise.
22165 (tree-inline.o): Likewise.
22166 (tree-ssa-tailmerge.o): Likewise.
22167 (ipa-split.o): Likewise.
22168 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
22169 (tree-ssa-copy.o): Likewise.
22170 * tree-switch-conversion.c: Include cfgloop.h
22171 (process_switch): If we emit a bit-test cascade, schedule loops
22173 * tree-complex.c: Include cfgloop.h.
22174 (expand_complex_div_wide): Properly add new basic-blocks to loops.
22175 * asan.c: Include cfgloop.h.
22176 (create_cond_insert_point): Properly add new basic-blocks to
22177 loops, schedule loop fixup.
22178 * cfgloop.c (verify_loop_structure): Check that looks are not
22180 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
22182 (expand_omp_for_generic): Likewise.
22183 (expand_omp_sections): Likewise.
22184 (expand_omp_atomic_pipeline): Schedule loops for fixup.
22185 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
22186 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
22187 is initialized, not when loops are present.
22188 * tree-parloops.c (parallelize_loops): Remove checking here.
22189 * passes.c (init_optimization_passes): Schedule a copy-propagation
22190 pass before complete unrolling of inner loops.
22192 2013-04-26 Jakub Jelinek <jakub@redhat.com>
22194 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
22195 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
22196 (colorize_init): Add argument to _WIN32 version.
22197 * toplev.c: Include diagnostic-color.h.
22198 (process_options): Default to -fdiagnostics-color=auto if
22199 GCC_COLORS env var is in the environment.
22200 * common.opt (fdiagnostics-color=): Add Var and Init.
22201 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
22202 env var is in the environment, the default is auto rather than never.
22204 * diagnostic.h (file_name_as_prefix): Add context argument.
22205 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
22206 the string as locus.
22207 * langhooks.c (lhd_print_error_function): Adjust caller.
22209 2013-04-25 Lawrence Crowl <crowl@google.com>
22211 * var-tracking.c (shared_hash_def::htab):
22212 Change type to hash_table. Update dependent calls and types.
22214 2013-04-25 Lawrence Crowl <crowl@google.com>
22216 * Makefile.in: Update as needed below.
22218 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
22219 Move declaration to after the type's method definitons.
22221 * attribs.c (htab_t scoped_attributes::attribute_hash):
22222 Change type to hash_table. Update dependent calls and types.
22224 * bitmap.c (htab_t bitmap_desc_hash):
22225 Change type to hash_table. Update dependent calls and types.
22227 * cselib.c (htab_t cselib_hash_table):
22228 Change type to hash_table. Update dependent calls and types.
22230 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
22231 (hash_string_slot_node): Move implementation into lto-streamer.h
22232 struct string_slot_hasher.
22233 (eq_string_slot_node): Likewise.
22235 * data-streamer-out.c: Update output_block::string_hash_table
22236 dependent calls and types.
22238 * dwarf2cfi.c (htab_t trace_index):
22239 Change type to hash_table. Update dependent calls and types.
22241 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
22242 Change type to hash_table. Update dependent calls and types.
22243 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
22244 (htab_t optimize_external_refs::map): Likewise.
22245 (htab_t output_comp_unit::extern_map): Likewise.
22246 (htab_t output_comdat_type_unit::extern_map): Likewise.
22247 (htab_t output_macinfo::macinfo_htab): Likewise.
22248 (htab_t optimize_location_lists::htab): Likewise.
22249 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
22251 * except.c (htab_t ehspec_hash_type):
22252 Change type to hash_table. Update dependent calls and types.
22253 (assign_filter_values::ttypes): Likewise.
22254 (assign_filter_values::ehspec): Likewise.
22255 (sjlj_assign_call_site_values::ar_hash): Likewise.
22256 (convert_to_eh_region_ranges::ar_hash): Likewise.
22258 * gcse.c (htab_t pre_ldst_table):
22259 Change type to hash_table. Update dependent calls and types.
22261 * ggc-common.c (htab_t saving_htab):
22262 Change type to hash_table. Update dependent calls and types.
22263 (htab_t loc_hash): Likewise.
22264 (htab_t ptr_hash): Likewise.
22265 (call_count): Rename ggc_call_count.
22266 (call_alloc): Rename ggc_call_alloc.
22267 (loc_descriptor): Rename make_loc_descriptor.
22268 (add_statistics): Rename ggc_add_statistics.
22270 * ggc-common.c (saving_htab):
22271 Change type to hash_table. Update dependent calls and types.
22273 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
22274 (push_gimplify_context): Likewise.
22275 (pop_gimplify_context): Likewise.
22276 (struct gimple_temp_hash_elt): Added.
22277 (struct gimplify_hasher): Likewise.
22278 (struct gimplify_ctx.temp_htab):
22279 Change type to hash_table. Update dependent calls and types.
22281 * gimple-fold.c: Include gimplify-ctx.h.
22283 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
22284 Change type to hash_table. Update dependent calls and types.
22285 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
22286 avoid potential global name collision.
22288 * gimplify.c: Include gimplify-ctx.h.
22289 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
22290 (htab_t gimplify_ctx::temp_htab):
22291 Update dependent calls and types for new type hash_table.
22292 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
22293 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
22295 * gimplify-ctx.h: New.
22296 (struct gimple_temp_hash_elt): Move from gimplify.c.
22297 (class gimplify_hasher): New.
22298 (struct gimplify_ctx): Move from gimple.h.
22299 (htab_t gimplify_ctx::temp_htab):
22300 Change type to hash_table. Update dependent calls and types.
22302 * graphite-clast-to-gimple.c: Include graphite-htab.h.
22303 (htab_t ivs_params::newivs_index):
22304 Change type to hash_table. Update dependent calls and types.
22305 (htab_t ivs_params::params_index): Likewise.
22306 (htab_t print_generated_program::params_index): Likewise.
22307 (htab_t gloog::newivs_index): Likewise.
22308 (htab_t gloog::params_index): Likewise.
22310 * graphite.c: Include graphite-htab.h.
22311 4htab_t graphite_transform_loops::bb_pbb_mapping):
22312 Change type to hash_table. Update dependent calls and types.
22314 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
22315 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
22316 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
22318 * graphite-dependences.c: Include graphite-htab.h.
22319 (loop_is_parallel_p): Change hash table type of parameter.
22321 * graphite-htab.h: New.
22322 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
22323 (extern find_pbb_via_hash): Move from graphite-poly.h.
22324 (extern loop_is_parallel_p): Move from graphite-poly.h.
22325 (extern get_loop_body_pbbs): Move from graphite-poly.h.
22327 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
22328 (extern loop_is_parallel_p): Move to graphite-htab.h.
22329 (extern get_loop_body_pbbs): Move to graphite-htab.h.
22331 * haifa-sched.c (htab_t delay_htab):
22332 Change type to hash_table. Update dependent calls and types.
22333 (htab_t delay_htab_i2): Likewise.
22335 * ira-color.c (htab_t allocno_hard_regs_htab):
22336 Change type to hash_table. Update dependent calls and types.
22338 * ira-costs.c (htab_t cost_classes_htab):
22339 Change type to hash_table. Update dependent calls and types.
22341 * loop-invariant.c (htab_t merge_identical_invariants::eq):
22342 Change type to hash_table. Update dependent calls and types.
22344 * loop-iv.c (htab_t bivs):
22345 Change type to hash_table. Update dependent calls and types.
22347 * loop-unroll.c (htab_t opt_info::insns_to_split):
22348 Change type to hash_table. Update dependent calls and types.
22349 (htab_t opt_info::insns_with_var_to_expand): Likewise.
22351 * lto-streamer.h (struct string_slot): Move from data-streamer.h
22352 (struct string_slot_hasher): New.
22353 (htab_t output_block::string_hash_table):
22354 Change type to hash_table. Update dependent calls and types.
22356 * lto-streamer-in.c (freeing_string_slot_hasher): New.
22357 (htab_t file_name_hash_table):
22358 Change type to hash_table. Update dependent calls and types.
22360 * lto-streamer-out.c: Update output_block::string_hash_table dependent
22363 * lto-streamer.c (htab_t tree_htab):
22364 Change type to hash_table. Update dependent calls and types.
22366 * omp-low.c: Include gimplify-ctx.h.
22368 * passes.c (htab_t name_to_pass_map):
22369 Change type to hash_table. Update dependent calls and types.
22370 (pass_traverse): Rename to passes_pass_traverse.
22372 * plugin.c (htab_t event_tab):
22373 Change type to hash_table. Update dependent calls and types.
22375 * postreload-gcse.c (htab_t expr_table):
22376 Change type to hash_table. Update dependent calls and types.
22377 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
22379 * sese.c (debug_rename_map_1): Make extern.
22380 (htab_t copy_bb_and_scalar_dependences::rename_map):
22381 Change type to hash_table. Update dependent calls and types.
22383 * sese.h (extern debug_rename_map): Move to .c file.
22385 * store-motion.c (htab_t store_motion_mems_table):
22386 Change type to hash_table. Update dependent calls and types.
22388 * trans-mem.c (htab_t tm_new_mem_hash):
22389 Change type to hash_table. Update dependent calls and types.
22391 * tree-browser.c (htab_t TB_up_ht):
22392 Change type to hash_table. Update dependent calls and types.
22394 * tree-cfg.c (htab_t discriminator_per_locus):
22395 Change type to hash_table. Update dependent calls and types.
22397 * tree-complex.c: Include tree-hasher.h
22398 (htab_t complex_variable_components):
22399 Change type to hash_table. Update dependent calls and types.
22401 * tree-eh.c (htab_t finally_tree):
22402 Change type to hash_table. Update dependent calls and types.
22404 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
22405 struct int_tree_hasher.
22406 (extern int_tree_map_eq): Likewise.
22407 (uid_decl_map_hash): Removed.
22408 (extern decl_tree_map_eq): Likewise.
22410 * tree-hasher.h: New.
22411 (struct int_tree_hasher): New.
22412 (typedef int_tree_htab_type): New.
22414 * tree-inline.c: Include gimplify-ctx.h.
22416 * tree-mudflap.c: Include gimplify-ctx.h.
22418 * tree-parloops.c: Include tree-hasher.h.
22419 (htab_t eliminate_local_variables_stmt::decl_address):
22420 Change type to hash_table. Update dependent calls and types.
22421 (htab_t separate_decls_in_region::decl_copies): Likewise.
22423 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
22424 Change type to hash_table. Update dependent calls and types.
22426 * tree-sra.c (candidates):
22427 Change type to hash_table. Update dependent calls and types.
22429 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
22431 (int_tree_map_hash): Likewise.
22433 * tree-ssa-dom.c (htab_t avail_exprs):
22434 Change type to hash_table. Update dependent calls and types.
22436 * tree-ssa-live.c (var_map_base_init::tree_to_index):
22437 Change type to hash_table. Update dependent calls and types.
22439 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
22440 Change type to hash_table. Update dependent calls and types.
22442 * tree-ssa-phiopt.c (seen_ssa_names):
22443 Change type to hash_table. Update dependent calls and types.
22445 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
22446 Change type to hash_table. Update dependent calls and types.
22448 * tree-ssa-uncprop.c (equiv):
22449 Change type to hash_table. Update dependent calls and types.
22451 2013-04-25 Jakub Jelinek <jakub@redhat.com>
22453 PR rtl-optimization/57003
22454 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
22455 call note_stores with kill_clobbered_value callback again after
22456 killing regs_invalidated_by_call.
22458 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
22460 * config/aarch64/aarch64-simd.md
22461 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
22462 (aarch64_simd_bsl<mode>): Likewise.
22463 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
22465 2013-04-25 Marek Polacek <polacek@redhat.com>
22467 PR tree-optimization/57066
22468 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
22470 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
22472 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
22474 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
22476 * config/aarch64/aarch64-builtins.c
22477 (aarch64_fold_builtin): New.
22478 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
22479 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
22480 * config/aarch64/aarch64-simd-builtins.def (abs): New.
22481 * config/aarch64/arm_neon.h
22482 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
22484 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
22485 Tejas Belagod <tejas.belagod@arm.com>
22487 * config/aarch64/aarch64-builtins.c
22488 (aarch64_gimple_fold_builtin): New.
22489 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
22490 * config/aarch64/aarch64-simd-builtins.def (addv): New.
22491 * config/aarch64/aarch64-simd.md (addpv4sf): New.
22492 (addvv4sf): Update.
22493 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
22495 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22497 * config/aarch64/aarch64.md
22498 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
22500 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22502 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
22503 (*ngcsi_uxtw): New pattern.
22505 2013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22506 Julian Brown <julian@codesourcery.com>
22508 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
22509 (TB_DREG): Add T_V4HF.
22510 (v4hf_UP): New macro.
22511 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
22512 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
22513 Handle initialisation of V4HF. Adjust initialisation of reinterpret
22515 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
22516 (arm_vector_mode_supported_p): Handle V4HF.
22517 (arm_mangle_map): Handle V4HFmode.
22518 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
22519 * config/arm/arm_neon_builtins.def: Add entries for
22520 vcvtv4hfv4sf, vcvtv4sfv4hf.
22521 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
22522 (neon_vcvtv4hfv4sf): Likewise.
22523 * config/arm/neon-gen.ml: Handle half-precision floating point
22525 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
22526 * config/arm/arm_neon.h: Regenerate.
22527 * config/arm/neon.ml (type elts): Add F16.
22528 (type vectype): Add T_float16x4, T_floatHF.
22529 (type vecmode): Add V4HF.
22530 (type features): Add Requires_FP_bit feature.
22531 (elt_width): Handle F16.
22532 (elt_class): Likewise.
22533 (elt_of_class_width): Likewise.
22534 (mode_of_elt): Refactor.
22535 (type_for_elt): Handle F16, fix error messages.
22536 (vectype_size): Handle T_float16x4.
22537 (vcvt_sh): New function.
22538 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
22539 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
22540 (string_of_mode): Handle V4HF.
22541 * doc/arm-neon-intrinsics.texi: Regenerate.
22543 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
22545 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
22546 format specifier in 'X' case.
22548 2013-04-25 Alan Modra <amodra@gmail.com>
22551 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
22552 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
22553 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
22554 Repeat for many other rotate/shift and mask patterns using subregs.
22555 Name lshiftrt insns.
22556 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
22557 on WORDS_BIG_ENDIAN.
22559 2013-04-25 Alan Modra <amodra@gmail.com>
22561 * config.gcc: Support little-endian powerpc-linux targets.
22562 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
22563 (LINK_OS_LINUX_SPEC): Define.
22564 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
22565 Preserve MASK_LITTLE_ENDIAN.
22566 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
22567 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
22568 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
22569 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
22570 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
22571 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
22572 Correct fp word order for little-endian. Don't shift toc entries
22573 smaller than a word for little-endian.
22574 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
22575 (bswapdi2 splits): Correct low-part subreg for little-endian.
22576 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
22577 low/high where such is correct only for be.
22578 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
22579 little-endian for -mcall-aixdesc.
22581 2013-04-25 Alan Modra <amodra@gmail.com>
22583 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
22584 replace_equiv_address_nv.
22586 2013-04-25 Alan Modra <amodra@gmail.com>
22588 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
22590 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
22593 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
22594 * rtl.h (struct rtx_def): ...
22596 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
22598 PR rtl-optimizations/57046
22599 * lra-constraints (split_reg): Set up lra_risky_transformations_p
22600 for multi-reg splits.
22602 2013-04-24 H.J. Lu <hongjiu.lu@intel.com>
22604 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
22606 2013-04-24 Sterling Augustine <saugustine@google.com>
22608 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
22609 (comp_dir_string, debug_str_dwo_section): New.
22610 (DEBUG_STR_DWO_SECTION): Rename to ...
22611 (DEBUG_DWO_STR_SECTION): ... this.
22612 (DEBUG_NORM_STR_SECTION): Delete.
22613 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
22614 (DEBUG_STR_DWO_SECTION_FLAGS): New.
22615 (find_AT_string): Move most logic to ...
22616 (find_AT_string_in_table): ... here. New.
22617 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
22618 add_skeleton_AT_string. Delete logic.
22619 (output_skeleton_debug_sections): Remove call to
22620 add_top_level_skeleton_die_attrs.
22621 (add_comp_dir_attribute): Move logic to comp_dir_string.
22622 (dwarf2out_init): Initialize debug_str_dwo_section.
22623 (output_indirect_string): Call find_string_form.
22624 (output_indirect_strings): Rewrite.
22625 (prune_unused_types): Empty skeleton_debug_str_hash.
22626 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
22627 (dwarf2out_finish): Call output_indirect_strings.
22629 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
22631 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
22633 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
22635 * rtl.h (struct rtx_def): Add comment for field jump.
22636 (LRA_SUBREG_P): New macro.
22637 * recog.c (register_operand): Check LRA_SUBREG_P.
22638 * lra.c (lra): Add note at the end of RTL code. Align non-empty
22640 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
22641 (lra_final_code_change): Skip subreg change for operators.
22642 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
22643 if there are no operand changes.
22644 * lra-constraints.c (curr_insn_set): New.
22645 (match_reload): Set LRA_SUBREG_P.
22646 (emit_spill_move): Ditto.
22647 (check_and_process_move): Use curr_insn_set. Process only single
22648 set insns. Don't initialize sec_mem_p and change_p.
22649 (simplify_operand_subreg): Use LRA_SUBREG_P.
22650 (reg_in_class_p): New function.
22651 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
22652 of #ifdef. Add code to remove cycling.
22653 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
22654 non-null disp. Reload inner instead of disp when base and index
22655 are null. Try to put lo_sum into register.
22656 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
22657 (check_and_process_move): Move code for move cost check to
22658 simple_move_p. Remove equiv_substitution.
22659 (simple_move_p): New function.
22660 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
22661 curr_insn_set. Call check_and_process_move only for single set
22662 insns. Use the new function. Move call of check_and_process_move
22663 after operand equiv substitution and address process.
22665 2013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
22667 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
22668 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
22669 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
22671 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
22673 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
22675 2013-04-24 Marek Polacek <polacek@redhat.com>
22677 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
22678 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
22679 (select_loops_exit_conditions): Likewise.
22680 (number_of_iterations_for_all_loops): Likewise.
22681 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
22682 (scev_analysis): Likewise.
22684 2013-04-02 Catherine Moore <clm@codesourcery.com>
22685 Chao-ying Fu <fu@mips.com>
22687 * config/mips/micromips.md (jraddiusp): New pattern.
22688 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
22689 instruction if possible.
22691 2013-04-24 Alan Modra <amodra@gmail.com>
22693 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
22695 2013-04-24 Julian Brown <julian@codesourcery.com>
22696 Chung-Lin Tang <cltang@codesourcery.com>
22698 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
22699 dependency behavior in enumeration type DIE generation. Add TODO note
22700 to comments about future DW_FORM_sdata/udata re-work of related code.
22702 2013-04-23 Lawrence Crowl <crowl@google.com>
22704 * Makefile.in: Update as needed below.
22706 * hash-table.h (class hash_table):
22707 Correct many methods with parameter types compare_type to the correct
22708 value_type. (Correct code was unlikely to notice the change.)
22709 (hash_table::elements_with_deleted) New.
22710 (class hashtable::iterator): New.
22711 (hashtable::begin()): New.
22712 (hashtable::end()): New.
22713 (FOR_EACH_HASH_TABLE_ELEMENT): New.
22715 * statistics.c (statistics_hashes):
22716 Change type to hash_table. Update dependent calls and types.
22718 * tree-into-ssa.c (var_infos):
22719 Change type to hash_table. Update dependent calls and types.
22721 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
22722 Change type to hash_table. Update dependent calls and types.
22724 * tree-ssa-loop-im.c (struct mem_ref.refs):
22725 Change type to hash_table. Update dependent calls and types.
22727 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
22728 Change type to hash_table. Update dependent calls and types.
22730 * tree-ssa-sccvn.c (vn_tables_s::nary):
22731 Change type to hash_table. Update dependent calls and types.
22732 (vn_tables_s::phis): Likewise.
22733 (vn_tables_s::references): Likewise.
22735 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
22736 (vn_reference_eq): Update parameter and return types.
22738 * tree-ssa-structalias.c (pointer_equiv_class_table):
22739 Change type to hash_table. Update dependent calls and types.
22740 (location_equiv_class_table): Likewise.
22742 * tree-vect-data-refs.c: Consequential changes for making
22743 peeling a hash_table.
22745 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
22746 (destroy_loop_vec_info): Dependent hash_table update.
22748 * tree-vectorizer.h (peeling_htab):
22749 Change type to hash_table. Update dependent calls and types.
22751 2013-04-23 Shiva Chen <shiva0217@gmail.com>
22753 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
22754 to check the register content is equal or not.
22755 * lra-constraints.c (match_reload): Use lra_assign_reg_val
22756 to assign register content record.
22757 * lra-eliminations.c (update_reg_eliminate): Use
22758 lra_update_reg_val_offset to update register content offset.
22759 * lra-int.h (struct lra_reg): Add offset member.
22760 (lra_reg_val_equal_p): New static inline function.
22761 (lra_update_reg_val_offset): New static inline function.
22762 (lra_assign_reg_val): New static inline function.
22763 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
22764 to assign register content record.
22765 (initialize_lra_reg_info_element): Initial offset to zero.
22767 2013-04-23 Catherine Moore <clm@codesourcery.com>
22769 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
22770 operands. Record compression.
22772 2013-04-23 Xinliang David Li <davidxl@google.com>
22774 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
22776 2013-04-23 Richard Biener <rguenther@suse.de>
22778 PR middle-end/57036
22779 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
22780 parameter, only add abnormal goto edges from the copied body
22781 if the call could perform abnormal gotos.
22782 (copy_cfg_body): Adjust.
22784 2013-04-23 Sofiane Naci <sofiane.naci@arm.com>
22786 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
22788 2013-04-23 Andreas Schwab <schwab@linux-m68k.org>
22790 * coretypes.h (gimple_stmt_iterator): Add struct to make
22793 2013-04-23 Richard Biener <rguenther@suse.de>
22795 PR tree-optimization/57026
22796 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
22797 from SSA names occuring in abnormal PHI nodes.
22799 2013-04-22 Andi Kleen <ak@linux.intel.com>
22801 * lto/lto.c (print_lto_report_1): Fix LTO report names.
22803 2013-04-22 Andi Kleen <ak@linux.intel.com>
22805 * lto/lto.c (print_lto_report_1): Declare early.
22806 (read_cgraph_and_symbols): Call print_lto_report_1 early.
22808 2013-04-22 Andi Kleen <ak@linux.intel.com>
22810 * common.opt (-flto-report-wpa): Add.
22811 * doc/invoke.texi (-flto-report-wpa): Add.
22812 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
22815 2013-04-22 Xinliang David Li <davidxl@google.com>
22817 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
22818 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
22819 * Makefile.in: New dependency
22821 David Daney <ddaney.cavm@gmail.com>
22823 * configure.ac (gcc_cv_as_micromips_support): Use the
22824 --fatal-warnings option.
22825 * configure: Regenerate.
22827 2013-04-22 Marek Polacek <polacek@redhat.com>
22830 * tsan.c (instrument_expr): Don't instrument expression
22831 in case its size is zero.
22833 2013-04-22 Uros Bizjak <ubizjak@gmail.com>
22837 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
22839 * config/alpha/alpha.c (TARGET_LRA_P): New define.
22841 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
22843 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
22844 (gimple_stmt_iterator): New typedef.
22845 * gimple.h (gimple_stmt_iterator): Rename to...
22846 (gimple_stmt_iterator_d): ... This.
22847 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
22848 trees be valid for GIMPLE and GENERIC.
22849 (TARGET_GIMPLE_FOLD_BUILTIN): New.
22850 * gimple-fold.c (gimple_fold_call): Call target hook
22851 gimple_fold_builtin.
22852 * hooks.c (hook_bool_gsiptr_false): New.
22853 * hooks.h (hook_bool_gsiptr_false): New.
22854 * target.def (fold_stmt): New.
22855 * doc/tm.texi: Regenerate.
22857 2013-04-22 Vladimir Makarov <vmakarov@redhat.com>
22860 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
22861 a set sp if no stack realignment.
22863 2013-04-22 Nick Clifton <nickc@redhat.com>
22865 * config.gcc (tilegx-linux): Extend extra_objs rather than
22867 (tilepro-linux): Likewise.
22869 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
22871 * config/aarch64/aarch64-builtins.c
22873 (CF0, CF1, CF2, CF3, CF4, CF10): New.
22874 (VAR<1-12>): Add MAP parameter.
22875 (BUILTIN_*): Likewise.
22876 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
22877 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
22878 (aarch64_ushl_n<mode>): Likewise.
22879 (aarch64_sshr_n<mode>): Likewise.
22880 (aarch64_ushr_n<mode>): Likewise.
22881 (aarch64_<maxmin><mode>): Likewise.
22882 (aarch64_sqrt<mode>): Likewise.
22883 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
22884 (vshr<q>_n_*): Likewise.
22886 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
22888 * config/aarch64/aarch64-builtins.c
22889 (aarch64_simd_builtin_type_mode): Handle SF types.
22891 (BUILTIN_GPF): Define.
22892 (aarch64_init_simd_builtins): Handle SF types.
22893 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
22894 (frecps): Likewise.
22895 (frecpx): Likewise.
22896 * config/aarch64/aarch64-simd.md
22897 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
22898 (aarch64_frecpe<mode>): New.
22899 (aarch64_frecps<mode>): Likewise.
22900 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
22901 (v8type): Add frecp<esx>.
22902 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
22903 (aarch64_frecps<mode>): Likewise.
22904 * config/aarch64/iterators.md (FRECP): New.
22905 (frecp_suffix): Likewise.
22906 * config/aarch64/arm_neon.h
22907 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
22909 2013-04-22 Christian Bruel <christian.bruel@st.com>
22912 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
22913 (REG_CLASS_NAMES): Idem.
22914 (REG_CLASS_CONTENTS): Idem.
22915 (REGCLASS_HAS_FP_REG): Idem.
22916 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
22917 (sh_conditional_register_usage): Idem.
22919 2013-04-21 Jeff Law <law@redhat.com>
22921 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
22922 (ssa_forward_propagate_and_combine): Use it.
22924 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
22926 * lra.c: Update the flow chart diagram.
22928 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
22930 PR rtl-optimization/56847
22931 * lra-constraints.c (process_alt_operands): Discourage alternative
22932 with non-matche doffsettable memory constraint fro memory with
22935 2013-04-19 Richard Biener <rguenther@suse.de>
22937 PR tree-optimization/56982
22938 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
22940 * gimplify.c (gimplify_call_expr): Notice special calls.
22941 (gimplify_modify_expr): Likewise.
22942 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
22943 abnormal control flow receivers.
22944 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
22945 in the same way as cfun->has_nonlocal_labels.
22946 (gimple_purge_dead_abnormal_call_edges): Likewise.
22947 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
22948 receivers start a basic-block.
22950 2013-04-19 Richard Biener <rguenther@suse.de>
22952 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
22954 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
22955 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
22956 (SLP_TREE_LOAD_PERMUTATION): Add.
22957 (vect_transform_slp_perm_load): Adjust prototype.
22958 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
22959 (vect_free_slp_instance): Likewise.
22960 (vect_create_new_slp_node): Likewise.
22961 (vect_supported_slp_permutation_p): Remove.
22962 (vect_slp_rearrange_stmts): Adjust.
22963 (vect_supported_load_permutation_p): Likewise. Inline
22964 vect_supported_slp_permutation_p here.
22965 (vect_analyze_slp_instance): Compute load permutations per
22966 slp node instead of per instance.
22967 (vect_get_slp_defs): Adjust.
22968 (vect_transform_slp_perm_load): Likewise.
22969 (vect_schedule_slp_instance): Remove redundant code.
22970 (vect_schedule_slp): Remove hack for PR56270, add it ...
22971 * tree-vect-stmts.c (vectorizable_load): ... here, do not
22972 CSE loads for SLP. Adjust.
22974 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
22976 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
22977 spelling in two comments.
22979 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
22982 * config/arm/arm.c (load_multiple_sequence): Require SP
22983 as base register for loads if SP is in the register list.
22985 2013-04-19 Martin Jambor <mjambor@suse.cz>
22987 PR tree-optimization/56718
22988 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
22989 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
22990 and made public. Adjusted all callers.
22991 (ipa_intraprocedural_devirtualization): New function.
22992 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
22993 (ipa_intraprocedural_devirtualization): Likewise.
22994 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
22996 2013-04-19 Richard Biener <rguenther@suse.de>
22998 PR tree-optimization/57000
22999 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
23001 2013-04-19 Terry Guo <terry.guo@arm.com>
23003 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
23005 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
23006 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
23007 (cortex_m4_fmacs): Use new reservations.
23008 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
23010 2013-04-18 Vladimir Makarov <vmakarov@redhat.com>
23012 PR rtl-optimization/56999
23013 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
23015 (lra_coalesce): Remove split_origin_bitmap and related code.
23016 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
23017 ranges if necessary.
23019 2013-04-18 Uros Bizjak <ubizjak@gmail.com>
23021 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
23023 (ix86_expand_call): Remove clobbered_registers array and use
23024 x86_64_ms_sysv_extra_clobbered_registers instead.
23025 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
23027 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
23029 * config/i386/i386.md (*call_rex64_ms_sysv): Use
23030 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
23031 (*call_value_rex64_ms_sysv): Ditto.
23033 2013-04-18 Cary Coutant <ccoutant@google.com>
23035 * dwarf2out.c (output_pubnames): Check die_perennial_p of
23036 parent instead of die_mark.
23038 2013-04-18 Diego Novillo <dnovillo@google.com>
23040 * gimple.c (create_gimple_tmp): New.
23041 (get_expr_type): New.
23042 (build_assign): New.
23043 (build_type_cast): New.
23044 * gimple.h (enum ssa_mode): Define.
23045 (gimple_seq_set_location): New.
23046 * asan.c (build_check_stmt): Change some gimple_build_* calls
23047 to use build_assign and build_type_cast.
23049 2013-04-18 Richard Biener <rguenther@suse.de>
23051 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
23052 handle negative step. Remove redundant checks.
23053 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
23054 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
23055 for negative step and grouped loads fail to vectorize.
23057 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
23059 * emit-rtl.c (reset_insn_used_flags): New function.
23060 (reset_all_used_flags): Use it.
23061 (verify_insn_sharing): New function.
23062 (verify_rtl_sharing): Fix verification for SEQUENCEs.
23064 2013-04-18 Jakub Jelinek <jakub@redhat.com>
23066 PR tree-optimization/56984
23067 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
23068 and (x >> M) >= N don't register any assertion if N << M is the
23071 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
23073 * lower-subreg.c (resolve_simple_move): If called self-recursive,
23074 do not delete_insn insns that have not yet been emitted, only
23075 unlink them with remove_insn.
23076 * df-scan.c (df_insn_delete): Revert r197492.
23078 2013-04-17 Steven Bosscher <steven@gcc.gnu.org>
23080 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
23081 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
23083 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
23085 * config/arm/arm.md (movsicc_insn): Convert define_insn into
23086 define_insn_and_split.
23087 (and_scc,ior_scc,negscc): Likewise.
23088 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
23090 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
23092 * config/arm/arm.c (use_return_insn): Return 0 for targets that
23093 can benefit from using a sequence of LDRD instructions in epilogue
23094 instead of a single LDM instruction.
23096 2013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
23099 * doc/extend.texi: Fix typo.
23101 2013-04-17 Richard Biener <rguenther@suse.de>
23103 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
23104 (vect_build_slp_tree): ... here.
23105 (vect_build_slp_tree_1): Compute which stmts of the SLP group
23106 match. Remove special-casing of mismatched complex loads.
23107 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
23108 re-try the match with swapped commutative operands.
23109 (vect_supported_load_permutation_p): Remove special-casing of
23110 mismatched complex loads.
23111 (vect_analyze_slp_instance): Adjust.
23113 2013-04-17 Richard Biener <rguenther@suse.de>
23115 PR rtl-optimization/56921
23116 * cfgloop.h (struct loop): Add simple_loop_desc member.
23117 (struct niter_desc): Mark with GTY(()).
23118 (simple_loop_desc): Do not use aux field but simple_loop_desc.
23119 * loop-iv.c (get_simple_loop_desc): Likewise.
23120 (free_simple_loop_desc): Likewise.
23123 2013-04-16 Richard Biener <rguenther@suse.de>
23125 PR rtl-optimization/56921
23126 * loop-init.c (pass_rtl_move_loop_invariants): Add
23127 TODO_do_not_ggc_collect to todo_flags_finish.
23128 (pass_rtl_unswitch): Same.
23129 (pass_rtl_unroll_and_peel_loops): Same.
23130 (pass_rtl_doloop): Same.
23132 2013-04-17 Eric Botcazou <ebotcazou@adacore.com>
23134 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
23135 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
23136 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
23138 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
23139 * tree-streamer.c (record_common_node): Adjust reference in comment.
23141 2013-04-17 Terry Guo <terry.guo@arm.com>
23143 * config/arm/cortex-m4.md: Add a new bypass.
23145 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
23147 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
23149 (*subs_<optab><mode>_multp2): New pattern.
23150 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
23151 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
23153 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
23155 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
23156 (*subs_mul_imm_<mode>): New pattern.
23158 2013-04-16 David Edelsohn <dje.gcc@gmail.com>
23161 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
23162 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
23163 (vsx_movti_32bit): Change j->wa to O->wa.
23165 2013-04-16 Richard Biener <rguenther@suse.de>
23167 PR rtl-optimization/56921
23168 * loop-init.c (pass_rtl_move_loop_invariants): Add
23169 TODO_do_not_ggc_collect to todo_flags_finish.
23170 (pass_rtl_unswitch): Same.
23171 (pass_rtl_unroll_and_peel_loops): Same.
23172 (pass_rtl_doloop): Same.
23174 2013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
23176 * config/arm/arm.c (emit_multi_reg_push): New declaration
23177 for an existing function.
23178 (arm_emit_strd_push): New function.
23179 (arm_expand_prologue): Used here.
23180 (arm_emit_ldrd_pop): New function.
23181 (arm_expand_epilogue): Used here.
23182 (arm_get_frame_offsets): Update condition.
23183 (arm_emit_multi_reg_pop): Add a special case for load of a single
23184 register with writeback.
23186 2013-04-16 Uros Bizjak <ubizjak@gmail.com>
23188 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
23191 2013-04-16 Richard Biener <rguenther@suse.de>
23193 PR tree-optimization/56756
23194 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
23195 (first_mem_ref_loc): New.
23196 (execute_sm): Place the load temporarily before a previous
23197 access instead of in the latch edge to ensure its SSA dependencies
23198 are defined at points dominating the load.
23200 2013-04-16 Steven Bosscher <steven@gcc.gnu.org>
23202 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
23203 correct fix by moving header and footer insn to the footer of
23204 the merged basic block. Clear BB_END of the merged-away block.
23206 PR middle-end/43631
23207 * emit-rtl.c (make_note_raw): New function.
23208 (link_insn_into_chain): New static inline function.
23209 (add_insn): Use it.
23210 (add_insn_before, add_insn_after): Factor insn chain linking code...
23211 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
23212 using link_insn_into_chain.
23213 (note_outside_basic_block_p): New helper function for emit_note_after
23214 and emit_note_before.
23215 (emit_note_after): Use nobb variant of add_insn_after if the note
23216 should not be contained in a basic block.
23217 (emit_note_before): Use nobb variant of add_insn_before if the note
23218 should not be contained in a basic block.
23219 (emit_note_copy): Use make_note_raw.
23220 (emit_note): Likewise.
23221 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
23222 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
23223 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
23224 the moved barrier the tail of the basic block it follows.
23225 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
23227 2013-04-15 Jakub Jelinek <jakub@redhat.com>
23229 PR tree-optimization/56962
23230 * gimple-ssa-strength-reduction.c (record_increment): Only set
23231 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
23232 either rhs1 or rhs2 is equal to c->base_expr.
23234 2013-04-15 Richard Biener <rguenther@suse.de>
23236 PR tree-optimization/56933
23237 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
23239 (GROUP_READ_WRITE_DEPENDENCE): Remove.
23240 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
23241 * tree-vect-data-refs.c (vect_analyze_group_access): Move
23242 dependence check ...
23243 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
23245 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
23246 GROUP_READ_WRITE_DEPENDENCE.
23248 2013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
23250 * emit-rtl.c (reset_all_used_flags): New function.
23251 (verify_rtl_sharing): Call reset_all_used_flags before and after
23252 performing the checks.
23254 2013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23256 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
23257 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
23258 * config/arm/constraints.md (De): New constraint.
23259 * config/arm/neon.md (anddi3_neon): Delete.
23260 (neon_vand<mode>): Expand to standard anddi3 pattern.
23261 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
23262 Move earlier in the file.
23263 (neon_inv_logic_op2): Likewise.
23264 (arm_anddi_operand_neon): New predicate.
23266 2013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
23268 * configure.ac (gcc_cv_ld_as_needed): Set
23269 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
23270 Use -z ignore, -z record on *-*-solaris2*.
23271 (HAVE_LD_AS_NEEDED): Update comment.
23272 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
23273 * configure: Regenerate.
23274 * config.in: Regenerate.
23275 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
23276 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
23277 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
23278 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
23279 equivalents. Fix markup.
23280 * doc/tm.texi: Regenerate.
23282 2013-04-15 Andrew Hsieh <andrewhsieh.google.com>
23284 * config/i386/i386.opt: New option mstack-protector-guard=.
23285 * config/i386/i386-opts.h: Add enum stack_protector_guard.
23286 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
23287 TARGET_SSP_TLS_GUARD.
23288 * config/i386/i386.c (ix86_option_override_internal): Set
23289 ix86_stack_protector_guard.
23290 * config/i386/i386.md (stack_protect_set): Enable for
23291 TARGET_SSP_TLS_GUARD only.
23292 (stack_protect_set_<mode>): Ditto.
23293 (stack_protect_test): Ditto.
23294 (stack_protect_test_<mode>): Ditto.
23295 * doc/invoke.texi (i386 Option): Document.
23297 2013-04-15 Eric Botcazou <ebotcazou@adacore.com>
23300 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
23301 (S_MODES): Set H_MODE bit.
23302 (SF_MODES): Set only S_MODE and SF_MODE bits.
23303 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
23304 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
23305 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
23306 <MODE_FLOAT>: Likewise.
23308 2013-04-15 Joey Ye <joey.ye@arm.com>
23310 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
23312 2013-04-15 Joey Ye <joey.ye@arm.com>
23314 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
23316 (thumb_far_jump_used_p): Count instruction size and set
23319 2013-04-14 Eric Botcazou <ebotcazou@adacore.com>
23321 * reorg.c (fill_simple_delay_slots): Reindent block of code.
23322 * resource.c (mark_target_live_regs): Reformat conditional block.
23324 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
23326 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
23327 notes, they are emitted only just before final.
23328 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
23330 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
23332 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
23333 * cfgrtl.c (delete_insn): Call it here instead.
23334 * lra-spills.c (lra_final_code_change): Use delete_insn.
23335 * haifa-sched.c (sched_remove_insn): Likewise.
23336 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
23337 returning to the nop pool.
23338 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
23339 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
23341 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
23343 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
23344 * doc/tm.texi: Regenerated.
23346 2013-04-12 Uros Bizjak <ubizjak@gmail.com>
23348 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
23351 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
23353 * df-core.c (df_find_def): Compare register numbers.
23354 (df_find_use): Likewise.
23356 2013-04-12 Vladimir Makarov <vmakarov@redhat.com>
23359 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
23360 lra_in_progress for return.
23362 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
23364 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
23365 define_insn into define_insn_and_split and emit movsicc patterns.
23367 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
23369 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
23371 2013-04-12 Richard Biener <rguenther@suse.de>
23373 * tree-pass.h (TODO_do_not_ggc_collect): New.
23374 * passes.c (execute_one_ipa_transform_pass): Honor
23375 TODO_do_not_ggc_collect.
23376 (execute_one_pass): Likewise.
23379 2013-04-10 Richard Biener <rguenther@suse.de>
23381 * passes.c (init_optimization_passes): Remove reload pass.
23382 * ira.c (do_reload): Merge into ...
23384 (rest_of_handle_reload): Remove.
23385 (pass_reload): Likewise.
23386 * config/i386/i386.c (ix86_option_override): Refer to ira instead
23387 of reload for vzeroupper pass placement.
23389 2013-04-12 Jakub Jelinek <jakub@redhat.com>
23391 PR tree-optimization/56918
23392 PR tree-optimization/56920
23393 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
23394 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
23395 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
23396 use wide_mul_with_sign method.
23398 2013-04-12 Richard Biener <rguenther@suse.de>
23400 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
23401 not be considered a gimple constant.
23403 2013-04-12 Marc Glisse <marc.glisse@inria.fr>
23405 * fold-const.c (const_binop): Handle vector shifts by a scalar.
23406 (fold_binary_loc): Call const_binop also for mixed vector-scalar
23409 2013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
23410 Jakub Jelinek <jakub@redhat.com>
23412 * opts.c: Include diagnostic-color.h.
23413 (common_handle_option): Handle OPT_fdiagnostics_color_.
23414 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
23415 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
23416 (diagnostic-color.o): New.
23417 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
23418 (diagnostic_color_rule): New enum.
23419 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
23420 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
23421 the location string.
23422 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
23423 either NULL, or color kind.
23424 * diagnostic-color.c: New file.
23425 * diagnostic-color.h: New file.
23426 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
23428 * doc/invoke.texi (-fdiagnostics-color): Document.
23429 * pretty-print.h (pp_show_color): Define.
23430 (struct pretty_print_info): Add show_color field.
23431 * diagnostic.c: Include diagnostic-color.h.
23432 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
23433 macros. Colorize error:, warning: etc. strings and also the location
23435 (diagnostic_show_locus): Colorize the caret line.
23436 * pretty-print.c: Include diagnostic-color.h.
23437 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
23438 inside of %< %> quotes or quoted through q format modifier.
23440 2013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
23442 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
23444 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
23446 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
23447 code in CC_NZ mode.
23448 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
23451 2013-04-11 Marek Polacek <polacek@redhat.com>
23453 PR tree-optimization/48184
23454 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
23456 2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
23458 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
23459 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
23460 (skip_simple_arithmetic): Tidy up.
23461 * tree.h (skip_simple_constant_arithmetic): Declare.
23463 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
23465 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
23467 2013-04-11 Richard Biener <rguenther@suse.de>
23469 * tree-vect-loop.c (get_initial_def_for_induction): Properly
23470 generate vector constants.
23472 2013-04-11 Richard Biener <rguenther@suse.de>
23474 PR tree-optimization/56878
23475 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
23476 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
23478 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
23479 Prefer to align the DR with the most invariant base address.
23481 2013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
23483 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
23486 2013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
23488 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
23489 floating-point vector comparisons against 0.
23491 2013-04-11 Jakub Jelinek <jakub@redhat.com>
23493 PR tree-optimization/56899
23494 * fold-const.c (extract_muldiv_1): Apply distributive law
23495 only if TYPE_OVERFLOW_WRAPS (ctype).
23497 2013-04-11 Bin Cheng <bin.cheng@arm.com>
23500 * ira-costs.c (scan_one_insn): Check whether the source rtx of
23501 loading has side effect.
23503 2013-04-10 Steven Bosscher <steven@gcc.gnu.org>
23505 * config/sparc/sparc.c: Include tree-pass.h.
23506 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
23507 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
23508 head of file. Change return type. Split off gate function.
23509 (sparc_gate_work_around_errata): New function.
23510 (pass_work_around_errata): New pass definition.
23511 (insert_pass_work_around_errata) New pass insert definition to
23512 insert pass_work_around_errata just after delayed-branch scheduling.
23513 (sparc_option_override): Insert the pass.
23514 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
23516 2013-04-10 David S. Miller <davem@davemloft.net>
23518 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
23519 or -mcpu=hypersparc.
23521 * target.def (cstore_mode): New hook.
23522 * target.h: Include insn-codes.h
23523 * targhooks.c: Likewise.
23524 (default_cstore_mode): New function.
23525 * targhooks.h: Declare it.
23526 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
23527 * doc/tm.texi: Rebuild.
23528 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
23529 target hook, rather than inspecting the insn_data.
23530 * config/sparc/sparc.c (sparc_cstore_mode): New function.
23531 (TARGET_CSTORE_MODE): Redefine.
23532 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
23534 * config/sparc/predicates.md (cstore_result_operand): New special
23536 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
23537 Use it for operand 0.
23538 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
23539 (*snesi_special): Likewise.
23540 (*snesi_zero): Likewise.
23541 (*seqsi_zero): Likewise.
23542 (*sltu_insn): Likewise.
23543 (*sgeu_insn): Likewise.
23544 (*seqdi_special): Make operand 0 and comparison operation be of
23546 (*snedi_special): Likewise.
23547 (*snedi_special_vis3): Likewise.
23548 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
23549 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
23550 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
23551 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
23552 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
23553 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
23554 (*sltu_extend_sp64): Likewise.
23555 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
23556 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
23557 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
23558 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
23559 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
23561 2013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
23563 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
23564 (aarch64_start_file): Use the new function.
23566 2013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
23567 Jason Merrill <jason@redhat.com>
23569 * common.opt: Add -gdwarf.
23570 * opts.c (common_handle_option): Handle it.
23571 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
23573 2013-04-10 Richard Biener <rguenther@suse.de>
23575 * passes.c (execute_todo): Do not call ggc_collect conditional here.
23576 (execute_one_ipa_transform_pass): But unconditionally here.
23577 (execute_one_pass): And here.
23578 (init_optimization_passes): Remove reload pass.
23579 * tree-pass.h (TODO_ggc_collect): Remove.
23580 (pass_reload): Likewise.
23581 * ira.c (do_reload): Merge into ...
23583 (rest_of_handle_reload): Remove.
23584 (pass_reload): Likewise.
23585 * config/i386/i386.c (ix86_option_override): Refer to ira instead
23586 of reload for vzeroupper pass placement.
23587 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
23588 and todo_flags_finish of all passes.
23590 2013-04-10 Richard Biener <rguenther@suse.de>
23592 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
23593 first_const_oprnd field, rename first_def_type to first_op_type.
23594 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
23595 (vect_get_and_check_slp_defs): Always use the type of the
23596 operand. Allow mixed vect_external_def, vect_constant_def types.
23597 (vect_get_constant_vectors): Handle mixed vect_external_def,
23598 vect_constant_def types.
23600 2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
23602 PR tree-optimization/55524
23603 * tree-ssa-math-opts.c
23604 (convert_mult_to_fma): Don't use an fms construct
23605 when we don't have an fms operation, but fnma, and it looks
23606 likely that we'll be able to use the latter.
23608 2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
23610 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
23612 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
23613 inline fail caused by overwritable functions.
23615 2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
23617 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
23618 unnecessary bits in the constant power of two case.
23620 2013-04-10 Richard Biener <rguenther@suse.de>
23622 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
23623 broken code swapping operands.
23624 (vect_build_slp_tree): Do not compute load permutations here.
23625 (vect_analyze_slp_instance): Compute load permutations here,
23626 after building the SLP tree.
23628 2013-04-09 Christian Bruel <christian.bruel@st.com>
23630 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
23631 of next/prev_real_insn.
23633 2013-04-09 Jan Hubicka <jh@suse.cz>
23635 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
23636 Drop aliased parameter.
23637 (function_and_variable_visibility): Do not handle alias pairs.
23638 * cgraph.c (varpool_externally_visible_p): Update prototype.
23639 * varpool.c (varpool_add_new_variable): Update.
23641 2013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23643 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
23645 2013-04-09 Steven Bosscher <steven@gcc.gnu.org>
23647 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
23649 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
23651 2013-04-09 Marek Polacek <polacek@redhat.com>
23653 PR tree-optimization/48762
23654 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
23656 2013-04-09 Richard Biener <rguenther@suse.de>
23658 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
23660 (vect_build_slp_tree): Likewise.
23661 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
23662 calculating the cost of a SLP instance.
23663 (vect_analyze_slp_instance): Use it from here, after building
23666 2013-04-09 Jakub Jelinek <jakub@redhat.com>
23668 PR middle-end/56883
23669 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
23670 expand_omp_for_static_chunk): Use simple_p = true in
23671 force_gimple_operand_gsi calls when assigning to addressable decls.
23673 2013-04-09 Jeff Law <law@redhat.com>
23675 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
23676 when the boolean was created by converting a wider object which
23677 had a boolean range.
23679 2013-04-09 Richard Biener <rguenther@suse.de>
23681 * tree-vectorizer.h (slp_void_p): Remove.
23682 (slp_tree): Typedef before _slp_tree declaration.
23683 (struct _slp_tree): Use a vector of slp_tree as children.
23684 (vect_get_place_in_interleaving_chain): Remove.
23685 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
23687 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
23689 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
23690 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
23691 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
23692 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
23693 Use slp_node instead of slp_void_p and adjust.
23695 2013-04-09 Richard Biener <rguenther@suse.de>
23697 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
23698 work that is not necessary.
23700 2013-04-09 Jakub Jelinek <jakub@redhat.com>
23702 PR tree-optimization/56854
23703 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
23704 forward into clobber stmts if it would change MEM_REF lhs into
23707 2013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
23709 * tree.c (type_hash_lookup, type_hash_add): Make static.
23710 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
23712 2013-04-09 Richard Biener <rguenther@suse.de>
23714 * tree.h (unsave_expr_now): Remove.
23715 * tree-inline.c (mark_local_for_remap_r): Remove.
23716 (unsave_expr_1): Likewise.
23717 (unsave_r): Likewise.
23718 (unsave_expr_now): Likewise.
23719 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
23720 (propagate_tree_value): Likewise.
23722 2013-04-08 Steven Bosscher <steven@gcc.gnu.org>
23724 * doc/rtl.texi (sequence): Rewrite documentation to match the
23725 current use of SEQUENCE rtl objects.
23726 * rtl.def (SEQUENCE): Likewise.
23728 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
23729 Update documentation.
23730 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
23731 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
23733 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
23735 2013-04-08 Teresa Johnson <tejohnson@google.com>
23737 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
23738 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
23740 (estimate_edge_size_and_time): Add comment to suggest using rounding
23742 (estimate_node_size_and_time): Ditto.
23743 (remap_edge_change_prob): Use helper rounding divide methods.
23744 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
23745 (gimple_mod_pow2_value_transform): Ditto.
23746 (gimple_mod_subtract_transform): Ditto.
23747 (gimple_ic_transform): Ditto.
23748 (gimple_stringops_transform): Ditto.
23749 * stmt.c (conditional_probability): Ditto.
23750 (emit_case_dispatch_table): Ditto.
23751 * lto-cgraph.c (merge_profile_summaries): Ditto.
23752 * tree-optimize.c (execute_fixup_cfg): Ditto.
23753 * cfgcleanup.c (try_forward_edges): Ditto.
23754 * cfgloopmanip.c (scale_loop_profile): Ditto.
23756 (duplicate_loop_to_header_edge): Ditto.
23757 (lv_adjust_loop_entry_edge): Ditto.
23758 * tree-vect-loop.c (vect_transform_loop): Ditto.
23759 * profile.c (compute_branch_probabilities): Ditto.
23760 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
23761 * lto-streamer-in.c (input_cfg): Ditto.
23762 * gimple-streamer-in.c (input_bb): Ditto.
23763 * ipa-cp.c (update_profiling_info): Ditto.
23764 (update_specialized_profile): Ditto.
23765 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
23766 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
23768 * sched-rgn.c (compute_dom_prob_ps): Ditto.
23769 (compute_trg_info): Ditto.
23770 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
23771 (purge_dead_edges): Ditto.
23772 * loop-unswitch.c (unswitch_loop): Ditto.
23773 * cgraphclones.c (cgraph_clone_edge): Ditto.
23774 (cgraph_clone_node): Ditto.
23775 * tree-inline.c (copy_bb): Ditto.
23776 (copy_edges_for_bb): Ditto.
23777 (initialize_cfun): Ditto.
23778 (copy_cfg_body): Ditto.
23779 (expand_call_inline): Ditto.
23781 2013-04-08 Kai Tietz <ktietz@redhat.com>
23783 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
23784 TARGET_CYGWIN64 by TARGET_64BIT.
23786 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
23788 * config/epiphany/epiphany.md (GPR_1): New constant.
23789 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
23790 * config/epiphany/epiphany.c (gen_compare_reg):
23791 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
23792 is already in place.
23793 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
23794 Don't require being called during rtl expansion; If y operlaps r0,
23796 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
23797 (epiphany_expand_epilogue): Likewise.
23799 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
23800 Don't use CC_FPmode for ORDERED / UNORDERED.
23801 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
23803 * config/epiphany/constraints.md (CnL): New constraint.
23804 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
23805 * config/epiphany/predicates.md (add_operand): Allow 1024.
23807 * config/epiphany/epiphany.md (logical_op): New code iterator.
23808 (op_mnc): New code attribute.
23809 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
23810 (mov_f+1, mov_f+2): New peephole2 patterns.
23812 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
23813 (cstoresi4): Also allow re-use of zero result when doing a NE
23814 comparison to a non-zero operand.
23815 Use (clobber (scratch)) for first insn if the gpr output is not needed.
23817 * config/epiphany/epiphany.md (<insn_opname>v2si3):
23818 Use gen_addsi3_i / gen_subsi3_i.
23820 2013-04-08 Jakub Jelinek <jakub@redhat.com>
23824 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
23825 contain anything but clobbers, at most one __builtin_stack_restore,
23826 optionally debug stmts and final resx, and if it has at least one
23827 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
23828 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
23829 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
23830 which isn't defaut definition, remove them.
23831 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
23832 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
23833 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
23834 with MEM_REF LHS with SSA_NAME address.
23836 2013-04-08 Jeff Law <law@redhat.com>
23838 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
23840 2013-04-08 Richard Biener <rguenther@suse.de>
23842 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
23844 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
23845 determined vector type.
23846 (vect_analyze_data_refs): Likewise.
23847 (vect_get_new_vect_var): Adjust.
23848 (vect_create_destination_var): Preserve SSA name versions.
23849 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
23850 not dump anything here.
23852 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
23854 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
23855 Add member lr_slot_known.
23856 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
23858 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
23859 Remove code that sets lr_slot_offset according to what a previous
23860 version of epiphany_emit_save_restore used to do.
23861 (epiphany_emit_save_restore): When doing an lr save or restore,
23862 set/verify lr_slot_known and lr_slot_offset.
23864 2013-04-08 Xinyu Qi <xyqi@marvell.com>
23867 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
23870 2013-04-08 Richard Biener <rguenther@suse.de>
23872 * alias.c (find_base_term): Fix thinko in previous change.
23874 2013-04-08 Jakub Jelinek <jakub@redhat.com>
23876 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
23877 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
23878 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
23879 if possible to compute val.
23880 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
23881 For QImode integers don't require anything about precision. Use
23882 const_with_all_bytes_same to find out if the constant doesn't have
23883 repeated bytes in it.
23885 2013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
23887 * config/s390/s390.c (s390_expand_insv): Only accept insertions
23890 2013-04-08 Marek Polacek <polacek@redhat.com>
23892 PR rtl-optimization/48182
23893 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
23896 2013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
23899 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
23900 nuses, make sure we have a label.
23902 2013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
23905 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
23906 (rs6000_emit_swdiv_low_precision): Remove.
23907 (rs6000_emit_swdiv): Rewrite to handle between one and four
23908 iterations of Newton-Raphson generally; modify required number of
23909 iterations for some cases.
23910 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
23912 2013-04-05 Steven Bosscher <steven@gcc.gnu.org>
23914 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
23915 set-but-unused variable.
23917 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
23918 basic blocks of released function bodies garbage-collectable.
23920 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
23921 (struct rtl_opt_pass): Add TODO_df_finish.
23923 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
23925 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23927 * config/arm/constraints.md (q): New constraint.
23928 * config/arm/ldrdstrd.md: New file.
23929 * config/arm/arm.md (ldrdstrd.md) New include.
23930 (arm_movdi): Use "q" instead of "r" constraint
23931 for double-word memory access.
23932 (movdf_soft_insn): Likewise.
23933 * config/arm/vfp.md (movdi_vfp): Likewise.
23934 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
23935 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
23936 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
23937 (mem_ok_for_ldrd_strd): Likewise.
23938 (output_move_double): Update assertion.
23940 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23942 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
23944 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23946 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
23947 define_insn_and_split.
23948 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
23950 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23952 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
23953 define_insn_and_split.
23954 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
23955 (shiftsi3_compare): New pattern.
23956 (rrx): New pattern.
23957 * config/arm/unspecs.md (UNSPEC_RRX): New.
23959 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23961 * config/arm/arm.md (negdi_extendsidi): New pattern.
23962 (negdi_zero_extendsidi): Likewise.
23964 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23966 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
23967 define_insn_and_split.
23968 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
23969 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
23971 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23973 * config/arm/arm.md (arm_subdi3): Convert define_insn into
23974 define_insn_and_split.
23975 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
23976 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
23978 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23980 * config/arm/arm.md (subsi3_carryin): New pattern.
23981 (subsi3_carryin_const): Likewise.
23982 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
23983 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
23985 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23987 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
23989 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23991 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
23992 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
23994 2013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23996 * config/arm/arm.c (arm_expand_builtin): Change fcode
23997 type to unsigned int.
23999 2013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
24001 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
24003 2013-04-04 Ian Lance Taylor <iant@google.com>
24005 * doc/standards.texi (Standards): The Go frontend supports the Go 1
24008 2013-04-04 Steven Bosscher <steven@gcc.gnu.org>
24010 PR middle-end/56729
24011 * df-scan.c (df_insn_delete): Disable failing assert.
24013 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24015 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
24016 New function prototype.
24017 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
24018 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
24019 (arm_builtin_vectorized_function): New function.
24021 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24023 * config/arm/arm_neon_builtins.def: New file.
24024 * config/arm/arm.c (neon_builtin_data): Move contents to
24025 arm_neon_builtins.def.
24026 (enum arm_builtins): Include neon builtin definitions.
24027 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
24028 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
24030 2013-04-04 Marek Polacek <polacek@redhat.com>
24032 PR tree-optimization/48186
24033 * predict.c (maybe_hot_frequency_p): Return false if
24034 HOT_BB_FREQUENCY_FRACTION is 0.
24035 (cgraph_maybe_hot_edge_p): Likewise.
24037 2013-04-04 Richard Biener <rguenther@suse.de>
24039 PR tree-optimization/56826
24040 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
24043 2013-04-04 Richard Biener <rguenther@suse.de>
24045 PR tree-optimization/56213
24046 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
24047 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
24049 2013-04-04 Richard Biener <rguenther@suse.de>
24051 PR tree-optimization/56837
24052 * tree-loop-distribution.c (classify_partition): For non-zero
24053 values require that the value has the same precision as its
24054 mode to be useful as memset value.
24056 2013-04-03 Nick Clifton <nickc@redhat.com>
24058 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
24059 (fmssf4): Use fmsf.s on E3V5 architectures.
24060 (fnmasf4): Use fnmaf.s on E3V5 architectures.
24061 (fnmssf4): Use fnmsf.s on E3V5 architectures.
24063 2013-04-03 Jeff Law <law@redhat.com>
24065 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
24066 (lra-eliminations.o): Likewise.
24068 2013-04-03 Teresa Johnson <tejohnson@google.com>
24070 * gcov-io.c (compute_working_sets): Moved most of body of old
24071 compute_working_sets here from profile.c.
24072 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
24073 (gcov_working_set_t): Moved typedef here from basic-block.h
24074 (compute_working_set): Declare.
24075 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
24076 (get_working_sets): Renamed from compute_working_set,
24077 replace most of body with call to new compute_working_sets.
24078 (get_exec_counts): Replace call to compute_working_sets
24079 to get_working_sets.
24080 * profile.h (get_working_sets): Renamed from compute_working_set.
24081 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
24082 to get_working_sets.
24083 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
24084 * gcov-dump.c (dump_working_sets): New function.
24086 2013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
24088 * hwint.c (sext_hwi, zext_hwi): New functions.
24089 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
24090 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
24091 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
24092 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
24093 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
24094 (sext_hwi, zext_hwi): New functions.
24096 2013-04-03 Jeff Law <law@redhat.com>
24098 PR tree-optimization/56799
24099 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
24100 back test for widening conversion erroneously dropped in prior change.
24102 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24105 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
24106 instead of next_real_insn.
24108 2013-04-03 Marek Polacek <polacek@redhat.com>
24111 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
24113 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24116 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
24118 (thumb1_output_casesi): Likewise.
24119 (thumb2_output_casesi): Likewise.
24121 2013-04-03 Richard Biener <rguenther@suse.de>
24123 PR tree-optimization/56817
24124 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
24126 (tree_unroll_loops_completely_1): ... new function to manually
24127 walk the loop tree, properly defering outer loops of unrolled
24128 loops to later iterations.
24130 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
24132 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
24133 (vectorizable_load): Likewise.
24134 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
24135 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
24137 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
24139 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
24142 2013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
24144 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
24146 2013-04-03 Bin Cheng <bin.cheng@arm.com>
24148 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
24150 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
24152 PR tree-optimization/56790
24153 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
24156 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
24158 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
24160 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
24161 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
24164 2013-04-03 Jakub Jelinek <jakub@redhat.com>
24167 * tree.h (force_folding_builtin_constant_p): New decl.
24168 * builtins.c (force_folding_builtin_constant_p): New variable.
24169 (fold_builtin_constant_p): Fold immediately also if
24170 force_folding_builtin_constant_p.
24172 2013-04-03 Richard Biener <rguenther@suse.de>
24174 PR tree-optimization/56812
24175 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
24176 DRs of the same interleaving chain are independent.
24178 2013-04-02 Jason Merrill <jason@redhat.com>
24180 * gdbinit.in (pbb): Use debug fn.
24182 2013-04-02 Lawrence Crowl <crowl@google.com>
24184 * sese.h (struct ivtype_map_elt_s): Remove unused.
24185 (extern debug_ivtype_map): Remove unused.
24186 (extern eq_ivtype_map_elts): Remove unused.
24187 * sese.c (debug_ivtype_map): Removed unused.
24188 (debug_ivtype_map_1): Removed unused.
24189 (debug_ivtype_elt): Remove unused.
24190 (eq_ivtype_map_elts): Remove unused.
24192 2013-04-02 Kai Tietz <ktietz@redhat.com>
24195 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
24196 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
24197 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
24199 (legitimize_pe_coff_symbol): Likewise.
24200 (is_imported_p): New helper-function.
24201 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
24202 for Windows x64 targets.
24203 (ix86_expand_prologue): Optimize for pe-coff targets.
24204 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
24205 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
24206 medium/large code-model.
24207 (legitimize_pic_address): Likewise.
24208 (legitimize_tls_address): Likewise.
24209 (ix86_expand_call): Likewise.
24210 (x86_output_mi_thunk): Likewise.
24211 (get_dllimport_decl): Add new beimport argument.
24212 (construct_plt_address): Don't assert for x64 pe-coff targets.
24213 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
24215 (SYMBOL_FLAG_STUBVAR): New macro.
24216 (SYMBOL_REF_STUBVAR_P): Likewise.
24217 * config/i386/winnt.c (stub_list): New structure.
24218 (stub_head): New local variable.
24219 (i386_pe_record_stub): New function.
24220 (i386_pe_file_end): Emit refptr-stubs.
24222 2013-04-02 Jakub Jelinek <jakub@redhat.com>
24224 PR rtl-optimization/56745
24225 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
24226 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
24229 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
24230 and both of them are MEM_REFs, just compare first argument for
24231 equality and attempt to deal even with differing offsets.
24234 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
24235 of gimple_clobber_p to be MEM_REF.
24236 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
24237 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
24238 after gimplification.
24239 * asan.c (get_mem_ref_of_assignment): Don't instrument
24240 gimple_clobber_p stmts.
24241 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
24242 gimple_clobber_p stmt if they have MEM_REF lhs and
24243 are dead because of another gimple_clobber_p stmt.
24244 * tree-ssa-live.c (clear_unused_block_pointer): Treat
24245 gimple_clobber_p stmts like debug stmts.
24246 (remove_unused_locals): Remove clobbers with MEM_REF lhs
24247 that refer to unused VAR_DECLs or uninitialized values.
24248 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
24249 gimple_clobber_p stmts if they refer to removed parameters.
24250 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
24253 2013-04-02 Uros Bizjak <ubizjak@gmail.com>
24255 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
24256 using SWI48 mode attribute.
24258 2013-04-02 Wei Mi <wmi@google.com>
24260 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
24261 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
24262 *<rotate_insn><mode>3_mask in i386.md.
24264 2013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
24266 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
24268 2013-04-02 Richard Biener <rguenther@suse.de>
24270 PR tree-optimization/56778
24271 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
24272 Runtime alias tests are not supported for gather loads.
24273 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
24274 stmts referenced from SSA operands before updating SSA form.
24276 2013-04-02 Ian Caulfield <ian.caulfield@arm.com>
24277 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
24279 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
24280 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
24281 * config/arm/cortex-a53.md: New file.
24282 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
24283 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
24284 * config/arm/arm.c (arm_issue_rate): Likewise.
24285 * config/arm/arm-tune.md: Regenerate
24286 * config/arm/arm-tables.opt: Regenerate.
24287 * config/arm/arm-cores.def: Add cortex-a53.
24289 2013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
24291 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
24294 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
24296 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
24297 scalar load/store operations using B/H registers.
24298 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
24300 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
24302 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
24304 * config/aarch64/aarch64.c
24305 (aarch64_simd_scalar_immediate_valid_for_move): New.
24306 * config/aarch64/aarch64-protos.h
24307 (aarch64_simd_scalar_immediate_valid_for_move): New.
24308 * config/aarch64/constraints.md (Dh, Dq): New.
24309 * config/aarch64/iterators.md (hq): New.
24311 2013-04-02 Eric Botcazou <ebotcazou@adacore.com>
24313 * reorg.c (get_branch_condition): Deal with conditional returns.
24314 (fill_simple_delay_slots): Remove dead code dealing with jumps.
24316 2013-04-01 Wei Mi <wmi@google.com>
24318 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
24319 Truncate operand 2 using %b asm operand modifier.
24320 (*<shift_insn><mode>3_mask): Ditto.
24321 (*<rotate_insn><mode>3_mask): Ditto.
24323 2013-04-01 Steven Bosscher <steven@gcc.gnu.org>
24325 PR middle-end/56798
24326 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
24328 2013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
24330 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
24332 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
24334 2013-03-30 Lawrence Crowl <crowl@google.com>
24336 * dse.c (clear_alias_sets): Remove never set.
24337 (disqualified_clear_alias_sets): Remove never set.
24338 (clear_alias_mode_pool): Remove never set.
24339 (dse_step0): Remove condition that is never true.
24340 (canon_address): Remove condition that is never true.
24341 (dse_step7): Remove condition that is never true.
24342 (rest_of_handle_dse): Remove condition that is never true.
24343 (rest_of_handle_dse::did_global): Remove never read from above.
24344 (dse_step2_spill): Remove never called from above.
24345 (dse_step5_spill): Remove never called from above.
24347 2013-03-30 Steven Bosscher <steven@gcc.gnu.org>
24349 * doc/md.texi (Standard Names) <casesi>: Update documentation for
24350 JUMP_TABLE_DATA changes.
24351 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
24352 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
24353 (Insns) <jump_table_data>: New entry.
24354 * doc/tm.texi: Regenerate.
24356 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
24358 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
24359 for table jump at the end of a basic block using tablejump_p.
24360 * targhooks.c (default_invalid_within_doloop): Likewise.
24361 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
24362 target hook implementation that is identical to the default hook.
24363 (rs6000_invalid_within_doloop): Remove.
24365 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
24366 unused variable from tablejump_p call.
24368 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
24369 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
24370 (INSN_DELETED_P): Likewise.
24371 (emit_jump_table_data): New prototype.
24372 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
24373 after 4th as unused.
24374 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
24375 * sched-vis.c (print_insn): Likewise.
24376 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
24377 insn for compatibility with back ends that use next_active_insn to
24378 identify jump table data.
24379 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
24380 (remove_insn): Likewise.
24381 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
24383 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
24384 (emit_jump_table_data): New function.
24386 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
24387 basic block, a JUMP_TABLE_DATA never is.
24388 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
24389 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
24390 off from code handling real insns.
24391 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
24392 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
24393 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
24394 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
24395 is not a NONDEBUG_INSN_P.
24396 * ira-costs.c (scan_one_insn): Likewise.
24397 * jump.c (mark_all_labels): Likewise.
24398 (mark_jump_label_1): Likewise.
24399 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
24400 * lra.c (get_insn_freq): Expect all insns reaching here to be in
24402 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
24403 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
24404 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
24405 JUMP_TABLE_DATA_P insns.
24406 (calculate_elim_costs_all_insns): Likewise.
24407 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
24408 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
24409 (delete_output_reload): Code style fixups.
24410 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
24411 insn flags on this non-insn.
24412 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
24413 as scheduling barriers, for pre-change compatibility.
24414 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
24415 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
24417 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
24418 redundant JUMP_TABLE_DATA_P test.
24419 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
24420 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
24421 (frv_for_each_packet): Likewise.
24422 * config/i386/i386.c (min_insn_size): Likewise.
24423 (ix86_avoid_jump_mispredicts): Likewise.
24424 * config/m32r/m32r.c (m32r_is_insn): Likewise.
24425 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
24426 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
24427 (mips16_insn_length): Robustify.
24428 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
24429 (mips16_split_long_branches): Likewise.
24430 * config/pa/pa.c (pa_combine_instructions): Likewise.
24431 * config/rs6000/rs6000.c (get_next_active_insn): Treat
24432 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
24433 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
24434 as contributing to pool range lengths.
24435 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
24436 Remove redundant JUMP_TABLE_DATA_P test.
24437 (sh_loop_align): Likewise.
24438 (split_branches): Likewise.
24439 (sh_insn_length_adjustment): Likewise.
24440 * config/spu/spu.c (get_branch_target): Likewise.
24442 2013-03-29 Jan Hubicka <jh@suse.cz>
24444 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
24445 gcov streaming; stream hot bb threshold to ltrans.
24446 * predict.c (get_hot_bb_threshold): Break out from ....
24447 (maybe_hot_count_p): ... here.
24448 (set_hot_bb_threshold): New function.
24449 * lto-section-in.c (lto_section_name): Add profile.
24450 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
24451 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
24452 and data-streamer.h
24453 (histogram_entry): New structure.
24454 (histogram, histogram_pool): New global vars.
24455 (histogram_hash): New structure.
24456 (histogram_hash::hash): New method.
24457 (histogram_hash::equal): Likewise.
24458 (account_time_size): New function.
24459 (cmp_counts): New function.
24460 (dump_histogram): New function.
24461 (ipa_profile_generate_summary): New function.
24462 (ipa_profile_write_summary): New function.
24463 (ipa_profile_read_summary): New function.
24464 (ipa_profile): Decide on threshold.
24465 (pass_ipa_profile): Add ipa_profile_write_summary and
24466 ipa_profile_read_summary.
24467 * Makefile.in (ipa.o): Update dependencies.
24468 * lto-streamer.h (LTO_section_ipa_profile): New section.
24470 2013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
24472 * tree.h (VAR_P): New.
24474 2013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
24477 * doc/invoke.texi ([-fwhole-program]): Fix typo.
24479 2013-03-29 Steven Bosscher <steven@gcc.gnu.org>
24481 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
24482 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
24483 (control_flow_insn_p): Likewise.
24484 * cfgrtl.c (duplicate_insn_chain): Likewise.
24485 * final.c (get_attr_length_1): Likewise.
24486 (shorten_branches): Likewise.
24487 (final_scan_insn): Likewise.
24488 * function.c (instantiate_virtual_regs): Likewise.
24489 * gcse.c (insert_insn_end_basic_block): Likewise.
24490 * ira-costs.c (scan_one_insn): Likewise.
24491 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
24492 * lra.c (check_rtl): Likewise.
24493 * reload1.c (elimination_costs_in_insn): Likewise.
24494 * reorg.c (follow_jumps): Likewise.
24496 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
24497 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
24498 (thumb_far_jump_used_p): Likewise.
24499 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
24500 (workaround_speculation): Likewise.
24501 (add_sched_insns_for_speculation): Likewise.
24502 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
24503 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
24504 (frv_for_each_packet): Likewise.
24505 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
24506 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
24507 (final_emit_insn_group_barriers): Likewise.
24508 * config/m32r/m32r.c (m32r_is_insn): Likewise.
24509 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
24510 (mips16_insn_length): Likewise.
24511 * config/pa/pa.c (pa_reorg): Likewise.
24512 (pa_combine_instructions): Likewise.
24513 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
24514 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
24515 (sh_reorg): Likewise.
24516 (split_branches): Likewise.
24517 * config/spu/spu.c (get_branch_target): Likewise.
24519 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
24522 2013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
24524 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
24525 Fix declaration name.
24527 2013-03-28 Lawrence Crowl <crowl@google.com>
24529 * graphds.h (struct graph.indicies): Remove unused.
24530 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
24531 (SCOP_ORIGINAL_PDDRS): Remove unused.
24532 * sese.h (extern insert_loop_close_phis): Removed unused.
24533 (extern insert_guard_phis): Removed unused.
24534 (extern ivtype_map_elt_info): Removed unused.
24535 (new_ivtype_map_elt): Removed unused.
24536 * sese.c (ivtype_map_elt_info): Removed unused.
24538 2013-03-28 Lawrence Crowl <crowl@google.com>
24540 * Makefile.in: Add several missing include dependences.
24542 (test-dump.o): New. This object is not added to any executable,
24543 but is present for ad-hoc testing.
24545 (debug (const bitmap_head_def &)): New.
24546 (debug (const bitmap_head_def *)): New.
24548 (extern debug (const bitmap_head_def &)): New.
24549 (extern debug (const bitmap_head_def *)): New.
24551 (debug (edge_def &)): New.
24552 (debug (edge_def *)): New.
24554 (debug (basic_block_def &)): New.
24555 (debug (basic_block_def *)): New.
24557 (dump_node (const_tree, int, FILE *)): Correct source file.
24559 (debug (die_struct &)): New.
24560 (debug (die_struct *)): New.
24562 (extern debug (die_struct &)): New.
24563 (extern debug (die_struct *)): New.
24564 * gimple-pretty-print.c
24565 (debug (gimple_statement_d &)): New.
24566 (debug (gimple_statement_d *)): New.
24567 * gimple-pretty-print.h
24568 (extern debug (gimple_statement_d &)): New.
24569 (extern debug (gimple_statement_d *)): New.
24571 (debug (ira_allocno_copy &)): New.
24572 (debug (ira_allocno_copy *)): New.
24573 (debug (ira_allocno &)): New.
24574 (debug (ira_allocno *)): New.
24576 (extern debug (ira_allocno_copy &)): New.
24577 (extern debug (ira_allocno_copy *)): New.
24578 (extern debug (ira_allocno &)): New.
24579 (extern debug (ira_allocno *)): New.
24581 (debug (live_range &)): New.
24582 (debug (live_range *)): New.
24584 (debug (lra_live_range &)): New.
24585 (debug (lra_live_range *)): New.
24587 (debug (lra_live_range &)): New.
24588 (debug (lra_live_range *)): New.
24590 (debug (omega_pb_d &)): New.
24591 (debug (omega_pb_d *)): New.
24593 (extern debug (omega_pb_d &)): New.
24594 (extern debug (omega_pb_d *)): New.
24596 (debug (const rtx_def &)): New.
24597 (debug (const rtx_def *)): New.
24599 (debug_tree (tree): Move within file.
24600 (debug_raw (const tree_node &)): New.
24601 (debug_raw (const tree_node *)): New.
24602 (dump_tree_via_hooks (const tree_node *, int)): New.
24603 (debug (const tree_node &)): New.
24604 (debug (const tree_node *)): New.
24605 (debug_verbose (const tree_node &)): New.
24606 (debug_verbose (const tree_node *)): New.
24607 (debug_head (const tree_node &)): New.
24608 (debug_head (const tree_node *)): New.
24609 (debug_body (const tree_node &)): New.
24610 (debug_body (const tree_node *)): New.
24611 (debug_vec_tree (tree): Move and reimplement in terms of dump.
24612 (debug (vec<tree, va_gc> &)): New.
24613 (debug (vec<tree, va_gc> *)): New.
24615 (extern debug (const rtx_def &)): New.
24616 (extern debug (const rtx_def *)): New.
24618 (debug_raw (simple_bitmap_def &)): New.
24619 (debug_raw (simple_bitmap_def *)): New.
24620 (debug (simple_bitmap_def &)): New.
24621 (debug (simple_bitmap_def *)): New.
24623 (extern debug (simple_bitmap_def &)): New.
24624 (extern debug (simple_bitmap_def *)): New.
24625 (extern debug_raw (simple_bitmap_def &)): New.
24626 (extern debug_raw (simple_bitmap_def *)): New.
24628 (debug (vinsn_def &)): New.
24629 (debug (vinsn_def *)): New.
24630 (debug_verbose (vinsn_def &)): New.
24631 (debug_verbose (vinsn_def *)): New.
24632 (debug (expr_def &)): New.
24633 (debug (expr_def *)): New.
24634 (debug_verbose (expr_def &)): New.
24635 (debug_verbose (expr_def *)): New.
24636 (debug (vec<rtx> &)): New.
24637 (debug (vec<rtx> *)): New.
24639 (extern debug (vinsn_def &)): New.
24640 (extern debug (vinsn_def *)): New.
24641 (extern debug_verbose (vinsn_def &)): New.
24642 (extern debug_verbose (vinsn_def *)): New.
24643 (extern debug (expr_def &)): New.
24644 (extern debug (expr_def *)): New.
24645 (extern debug_verbose (expr_def &)): New.
24646 (extern debug_verbose (expr_def *)): New.
24647 (extern debug (vec<rtx> &)): New.
24648 (extern debug (vec<rtx> *)): New.
24650 (_list_iter_cond_expr): Make inline instead of static.
24652 (debug (sreal &)): New.
24653 (debug (sreal *)): New.
24655 (extern debug (sreal &)): New.
24656 (extern debug (sreal *)): New.
24658 (extern debug_raw (const tree_node &)): New.
24659 (extern debug_raw (const tree_node *)): New.
24660 (extern debug (const tree_node &)): New.
24661 (extern debug (const tree_node *)): New.
24662 (extern debug_verbose (const tree_node &)): New.
24663 (extern debug_verbose (const tree_node *)): New.
24664 (extern debug_head (const tree_node &)): New.
24665 (extern debug_head (const tree_node *)): New.
24666 (extern debug_body (const tree_node &)): New.
24667 (extern debug_body (const tree_node *)): New.
24668 (extern debug (vec<tree, va_gc> &)): New.
24669 (extern debug (vec<tree, va_gc> *)): New.
24671 (debug (struct loop &)): New.
24672 (debug (struct loop *)): New.
24673 (debug_verbose (struct loop &)): New.
24674 (debug_verbose (struct loop *)): New.
24675 * tree-dump.c: Add header dependence.
24677 (extern debug (struct loop &)): New.
24678 (extern debug (struct loop *)): New.
24679 (extern debug_verbose (struct loop &)): New.
24680 (extern debug_verbose (struct loop *)): New.
24682 (debug (data_reference &)): New.
24683 (debug (data_reference *)): New.
24684 (debug (vec<data_reference_p> &)): New.
24685 (debug (vec<data_reference_p> *)): New.
24686 (debug (vec<ddr_p> &)): New.
24687 (debug (vec<ddr_p> *)): New.
24689 (extern debug (data_reference &)): New.
24690 (extern debug (data_reference *)): New.
24691 (extern debug (vec<data_reference_p> &)): New.
24692 (extern debug (vec<data_reference_p> *)): New.
24693 (extern debug (vec<ddr_p> &)): New.
24694 (extern debug (vec<ddr_p> *)): New.
24696 (debug (pt_solution &)): New.
24697 (debug (pt_solution *)): New.
24699 (extern debug (pt_solution &)): New.
24700 (extern debug (pt_solution *)): New.
24702 (debug (_var_map &)): New.
24703 (debug (_var_map *)): New.
24704 (debug (tree_live_info_d &)): New.
24705 (debug (tree_live_info_d *)): New.
24707 (extern debug (_var_map &)): New.
24708 (extern debug (_var_map *)): New.
24709 (extern debug (tree_live_info_d &)): New.
24710 (extern debug (tree_live_info_d *)): New.
24712 2013-03-28 Jan Hubicka <jh@suse.cz>
24714 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
24716 2013-03-28 Ian Bolton <ian.bolton@arm.com>
24718 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
24719 record only when desired or required.
24721 2013-03-28 Uros Bizjak <ubizjak@gmail.com>
24723 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
24724 *vec_extractv2di_1_rex64. Use x64 isa attribute.
24726 2013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
24728 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
24729 (*andsi3_compare0_uxtw): New pattern.
24730 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
24731 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
24733 2013-03-28 Jan Hubicka <jh@suse.cz>
24735 * data-streamer-in.c (streamer_read_gcov_count): New function.
24736 * gimple-streamer-out.c: Include value-prof.h.
24737 (output_gimple_stmt): Output histogram.
24738 (output_bb): Use streamer_write_gcov_count.
24739 * value-prof.c: Include data-streamer.h
24740 (dump_histogram_value): Add HIST_TYPE_MAX.
24741 (stream_out_histogram_value): New function.
24742 (stream_in_histogram_value): New function.
24743 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
24744 (stream_out_histogram_value, stream_in_histogram_value): Declare.
24745 * data-streamer-out.c (streamer_write_gcov_count): New function.
24746 (streamer_write_gcov_count_stream): New function.
24747 * lto-cgraph.c (lto_output_edge): Update counter streaming.
24748 (lto_output_node): Likewise.
24749 (input_node, input_edge): Likewise.
24750 * lto-streamer-out.c (output_cfg): Update streaming.
24751 * lto-streamer-in.c (input_cfg): Likewise.
24752 * data-streamer.h (streamer_write_gcov_count,
24753 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
24754 * gimple-streamer-in.c: Include value-prof.h
24755 (input_gimple_stmt): Input histograms.
24756 (input_bb): Update profile streaming.
24758 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
24760 * genmodes.c (emit_max_int): New function.
24761 (emit_insn_modes_h): Added call to emit_max_function.
24762 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
24764 * machmode.def: Fixed comment.
24766 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
24768 * combine.c (try_combine): Removed useless assert.
24769 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
24771 2013-03-28 Marek Polacek <polacek@redhat.com>
24772 Richard Biener <rguenther@suse.de>
24774 PR tree-optimization/56695
24775 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
24776 build signed result of a vector comparison.
24777 * tree-cfg.c (verify_gimple_comparison): Check that a result
24778 of a vector comparison has signed type.
24780 2013-03-28 Richard Biener <rguenther@suse.de>
24782 PR tree-optimization/37021
24783 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
24784 do not restrict gaps between groups.
24785 * tree-vect-stmts.c (vectorizable_load): Properly account for
24786 a gap between groups.
24788 2013-03-28 Eric Botcazou <ebotcazou@adacore.com>
24790 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
24791 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
24794 2013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
24796 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
24797 * doc/extend.texi (Named Address Spaces): Ditto.
24798 (Variable Attributes): Ditto.
24800 2013-03-27 Kai Tietz <ktietz@redhat.com>
24802 * config.build: Add support for cygwin x64 target.
24803 * config.gcc: Likewise.
24804 * config.host: Likewise.
24805 * configure.ac: Likewise
24806 * configure: Regenerated.
24808 2013-03-27 Kai Tietz <ktietz@redhat.com>
24810 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
24811 * config/i386/t-cygwin-w64: New file.
24812 * config/i386/cygwin-w64.h: New file.
24813 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
24814 and add support for x64-cygwin target.
24815 (CPP_SPEC): Likewise.
24816 (CXX_WRAP_SPEC_LIST): Undefine before define.
24817 (LIBGCJ_SONAME): Use 15 as version.
24819 2013-03-27 Richard Biener <rguenther@suse.de>
24821 PR tree-optimization/56716
24822 * tree-ssa-structalias.c (perform_var_substitution): Adjust
24823 dumping for ref nodes.
24825 2013-03-27 Martin Jambor <mjambor@suse.cz>
24827 PR tree-optimization/55334
24828 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
24829 restricted pointers to arrays.
24831 2013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
24833 * Makefile.in (.SUFFIXES): Add .cc.
24834 (.c.o): Apply same recipe for implicit rule .cc.o.
24836 2013-03-27 Richard Biener <rguenther@suse.de>
24838 PR tree-optimization/37021
24839 * tree-vect-data-refs.c (vect_check_strided_load): Allow
24840 REALPART/IMAGPART_EXPRs around the supported refs.
24841 * tree-ssa-structalias.c (find_func_aliases): Assume that
24842 floating-point values are not used to transfer pointers.
24844 2013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
24846 * target.def (TARGET_HAS_IFUNC_P): New target hook.
24847 * doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
24848 * doc/tm.texi: Regenerate.
24849 * targhooks.h (default_has_ifunc_p): New.
24850 * targhooks.c (default_has_ifunc_p): Ditto.
24851 * config/linux-protos.h: New file.
24852 * config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of this
24853 hook for linux which disables support of indirect functions in android.
24854 * config/linux-android.c: New file.
24855 * config/t-linux-android.c: Ditto.
24856 * config.gcc: Added new object file linux-android.o.
24857 * config/i386/i386.c (ix86_get_function_versions_dispatcher):
24858 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
24859 * varasm.c (do_assemble_alias): Likewise.
24860 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
24861 doesn't support indirect functions.
24862 * configure: Regenerate.
24864 2013-03-27 Bin Cheng <bin.cheng@arm.com>
24867 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
24868 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
24871 2013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
24873 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
24875 2013-03-27 Terry Guo <terry.guo@arm.com>
24877 * config/arm/arm-cores.def: Added core cortex-r7.
24878 * config/arm/arm-tune.md: Regenerated.
24879 * config/arm/arm-tables.opt: Regenerated.
24880 * doc/invoke.texi: Added entry for core cortex-r7.
24882 2013-03-27 Walter Lee <walt@tilera.com>
24884 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
24885 double-decrement of next_scratch_regno.
24887 2013-03-27 Walter Lee <walt@tilera.com>
24889 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
24891 (insn_v1mulus): Ditto.
24892 (insn_v2muls): Ditto.
24894 2013-03-27 Walter Lee <walt@tilera.com>
24896 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
24897 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
24899 2013-03-27 Walter Lee <walt@tilera.com>
24901 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
24902 (*sibcall_value): Ditto.
24904 2013-03-27 Walter Lee <walt@tilera.com>
24906 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
24907 (insn_mnz_v8qi): ... this ...
24908 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
24910 (insn_v<n>mnz): Replaced by ...
24911 (insn_v1mnz): ... this ...
24912 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
24914 (insn_mz_<mode>): Replaced by ...
24915 (insn_mz_v8qi): ... this ...
24916 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
24918 (insn_v<n>mz): Replaced by ...
24919 (insn_v1mz): ... this ...
24920 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
24923 2013-03-26 Eric Botcazou <ebotcazou@adacore.com>
24925 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
24927 2013-03-26 Roland McGrath <mcgrathr@google.com>
24929 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
24930 than fprintf with a non-constant, non-format string.
24932 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
24934 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
24935 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
24936 operand 0 predicate.
24937 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
24938 attribute. Use general_x64nomem_operand as operand 1 predicate.
24939 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
24940 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
24941 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
24942 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
24943 (mov<mode>_insv_1): Remove expander. Merge insn with
24944 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
24945 Use general_x64nomem_operand as operand 1 predicate.
24946 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
24947 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
24948 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
24949 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
24950 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
24951 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
24952 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
24953 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
24954 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
24955 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
24956 (general_x64nomem_operand): Ditto.
24958 2013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
24960 * config/rtems.opt: Add -pthread option.
24962 2013-03-26 Richard Biener <rguenther@suse.de>
24964 * alias.c (find_base_term): Avoid redundant and not used recursion.
24965 (base_alias_check): Get the initial base term from the caller.
24966 (true_dependence_1): Compute and pass base terms to base_alias_check.
24967 (write_dependence_p): Likewise.
24968 (may_alias_p): Likewise.
24970 2013-03-26 Sofiane Naci <sofiane.naci@arm.com>
24972 * config/aarch64/aarch64.c (aarch64_classify_address): Support
24973 PC-relative load in SI modes and above only.
24975 2013-03-26 Xinyu Qi <xyqi@marvell.com>
24977 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
24978 * config/arm/iwmmxt.md (WCGR0): Update.
24979 (WCGR1, WCGR2, WCGR3): Likewise.
24981 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
24983 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
24984 Use x64 and nox64 isa attributes.
24986 2013-03-26 Richard Biener <rguenther@suse.de>
24988 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
24989 alignment computations and rely on get_object_alignment_1
24990 for the !TYPE_P case.
24991 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
24993 2013-03-26 Walter Lee <walt@tilera.com>
24995 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
24996 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
24998 2013-03-25 Jeff Law <law@redhat.com>
25000 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
25001 check for INTEGRAL_TYPE_P that was missing due to checking in
25002 wrong version of prior patch.
25004 2013-03-25 Walter Lee <walt@tilera.com>
25006 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
25007 TILEGX_INSN_SHUFFLEBYTES1.
25008 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
25010 (tilegx_builtins): Ditto.
25011 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
25013 2013-03-25 Walter Lee <walt@tilera.com>
25015 * config/tilegx/tilegx.md (floatsisf2): New pattern.
25016 (floatunssisf2): New pattern.
25017 (floatsidf2): New pattern.
25018 (floatunssidf2): New pattern.
25020 2013-03-25 Walter Lee <walt@tilera.com>
25022 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
25023 tests for constraint J, K, N, P.
25025 2013-03-25 Walter Lee <walt@tilera.com>
25027 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
25028 Use indirect/pcrel encoding.
25029 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
25032 2013-03-25 Steve Ellcey <sellcey@mips.com>
25034 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
25035 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
25036 * config/mips/mips.c (mips_option_override): Set IMADD default.
25037 * config/mips/mips.h (PTF_AVOID_IMADD): New.
25038 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
25039 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
25040 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
25041 * doc/invoke.texi (-mimadd/-mno-imadd): New.
25043 2013-03-25 Jeff Law <law@redhat.com>
25045 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
25046 slightly to avoid creating and folding useless trees. Simplify
25047 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
25049 2013-03-25 Uros Bizjak <ubizjak@gmail.com>
25051 * config/i386/i386.md (*zero_extendsidi2): Merge with
25052 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
25053 * config/i386/predicates.md (x86_64_zext_operand): Rename from
25054 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
25055 targets. Clarify comment.
25057 2013-03-25 Martin Jambor <mjambor@suse.cz>
25059 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
25060 pass-through jump functions differently.
25061 (ipa_read_jump_function): Likewise. Also use setter functions to set
25064 2013-03-25 Martin Jambor <mjambor@suse.cz>
25066 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
25067 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
25069 (ipa_get_indirect_edge_target): New function.
25070 (devirtualization_time_bonus): New parameter known_aggs, pass it to
25071 ipa_get_indirect_edge_target. Update all callers.
25072 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
25073 ipa_get_indirect_edge_target_1 instead of calling
25074 ipa_get_indirect_edge_target.
25075 (create_specialized_node): Pass aggvlas to
25076 ipcp_discover_new_direct_edges.
25078 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
25080 * config/arm/arm.md (f_sels, f_seld): New types.
25081 (*cmov<mode>): New pattern.
25082 * config/arm/predicates.md (arm_vsel_comparison_operator): New
25085 2013-03-25 Kai Tietz <ktietz@redhat.com>
25087 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
25088 POSIX-printf for mingw-hosted builds.
25090 2013-03-25 Richard Biener <rguenther@suse.de>
25092 PR middle-end/56694
25093 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
25094 must-not-throw stmt location.
25096 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
25098 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
25099 Emit load-acquire versions when acq is true.
25100 (arm_emit_store_exclusive): Add rel parameter.
25101 Emit store-release versions when rel is true.
25102 (arm_split_compare_and_swap): Use acquire-release instructions
25104 of barriers when appropriate.
25105 (arm_split_atomic_op): Likewise.
25106 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
25107 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
25108 (VUNSPEC_SLX): Likewise.
25109 (VUNSPEC_LDA): Likewise.
25110 (VUNSPEC_STL): Likewise.
25111 * config/arm/sync.md (atomic_load<mode>): New pattern.
25112 (atomic_store<mode>): Likewise.
25113 (arm_load_acquire_exclusive<mode>): Likewise.
25114 (arm_load_acquire_exclusivesi): Likewise.
25115 (arm_load_acquire_exclusivedi): Likewise.
25116 (arm_store_release_exclusive<mode>): Likewise.
25118 2013-03-25 Catherine Moore <clm@codesourcery.com>
25120 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
25121 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
25122 * config/mip/predicates.md (lwsp_swsp_operand,
25123 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
25124 sb16_operand, db4_operand, db7_operand, ib3_operand,
25125 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
25126 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
25127 andi16_operand): New predicates.
25128 * config/mips/mips.md (compression): New attribute.
25129 (enabled): New attribute.
25130 (length): Consider compression in computing length.
25131 (shift_compression): New code attribute.
25132 (*add<mode>3): New operands. Record compression.
25133 (sub<mode>3): Likewise.
25134 (one_cmpl<mode>2): Likewise.
25135 (*and<mode>3): Likewise.
25136 (*ior<mode>3): Likewise.
25137 (unnamed pattern for xor): Likewise.
25138 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
25139 (*<optab><mode>3): Likewise.
25140 (*mov<mode>_internal: Likewise.
25141 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
25142 (mips_unsigned_immediate_p): New.
25143 (umips_lwsp_swsp_address_p): New.
25144 (m16_based_address_p): New.
25145 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
25146 (mips_unsigned_immediate_p): New prototype.
25147 (lwsp_swsp_address_p): New prototype.
25148 (m16_based_address_p): New prototype.
25149 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
25150 (mips_signed_immediate_p): New function.
25151 (m16_based_address_p): New function.
25152 (lwsp_swsp_address_p): New function.
25153 (mips_print_operand_punctuation): Recognize short delay slot insns
25154 for microMIPS.add<mode>3"
25156 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
25159 * config/arm/iterators.md (v_cmp_result): New mode attribute.
25160 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
25162 2013-03-25 Richard Biener <rguenther@suse.de>
25164 PR tree-optimization/56689
25165 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
25168 2013-03-25 Richard Biener <rguenther@suse.de>
25170 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
25172 (memory_references): Likewise.
25173 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
25174 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
25175 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
25176 (gather_mem_refs_in_loops): Fold into ...
25177 (analyze_memory_references): ... this. Move initialization
25178 to tree_ssa_lim_initialize.
25179 (fill_always_executed_in): Rename to ...
25180 (fill_always_executed_in_1): ... this.
25181 (fill_always_executed_in): Move contains_call computation to
25182 this new function from ...
25183 (tree_ssa_lim_initialize): ... here.
25184 (tree_ssa_lim): Call fill_always_executed_in.
25186 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
25188 * postreload.c (reload_combine): Fix code detecting returns.
25190 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
25192 * function.c (emit_use_return_register_into_block): On cc0 targets,
25193 do not emit the sequence between cc0 setter and user.
25195 2013-03-25 Kai Tietz <ktietz@redhat.com>
25197 * config/i386/predicates.md (local_symbolic_operand): Interpret
25198 dll-imported symbols as none-local.
25200 2013-03-25 Richard Biener <rguenther@suse.de>
25202 * tree-ssa-loop-im.c (struct depend): Remove.
25203 (struct lim_aux_data): Make depends a vec of gimples.
25204 (free_lim_aux_data): Adjust.
25205 (add_dependency): Likewise.
25206 (set_level): Likewise.
25208 2013-03-25 Richard Biener <rguenther@suse.de>
25210 PR middle-end/56434
25211 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
25212 the pointer returned by calls with ECF_MALLOC set.
25214 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
25216 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
25218 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
25220 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
25221 using MMXMODE mode iterator.
25222 (*move<mode>_internal): Merge with *movv2sf_internal and
25223 *movv2sf_internal_rex64 using MMXMODE mode iterator.
25225 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
25227 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
25228 (record_last_mem_set_info): Likewise.
25230 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
25231 of XNEWVEC followed by memset.
25232 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
25234 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
25236 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
25237 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
25238 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
25239 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
25240 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
25241 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
25242 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
25243 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
25244 BARRIER_P instead of GET_CODE.
25246 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
25248 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
25249 inaccuracy in the probing code.
25251 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
25252 (ctrapdi4): Likewise.
25254 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
25256 * calls.c (expand_call): Add missing guard to code handling return
25257 of non-BLKmode structures in MSB.
25258 * function.c (expand_function_end): Likewise.
25260 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
25262 * combine.c (try_combine): Adjust comment. Do not add the set of
25263 insn #0 if the destination indirectly is set or dies in insn #2.
25264 Tidy up code to distribute a new note.
25266 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
25268 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
25269 also for alternatives 16 and 17.
25271 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
25273 * config/i386/sse.md (*mov<mode>_internal): Merge with
25274 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
25275 Emit insn template depending on type attribute. Use
25276 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
25277 movd instead of movq mnemonic for interunit moves. Rewrite mode
25278 attribute calculation. Remove unit attribute calculation.
25279 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
25280 Set prefix_data16 attribute for DImode ssemov types.
25281 Use Ym instead of y for SSE-MMX conversion alternatives.
25282 Reorder operand constraints.
25284 2013-03-22 Steven Bosscher <steven@gcc.gnu.org>
25286 * df.h (df_insn_delete): Adjust prototype.
25287 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
25288 and let it decide whether mark the basic block dirty.
25289 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
25290 * df-scan.c (df_insn_info_delete): New helper function, split
25291 off from df_insn_delete.
25292 (df_scan_free_bb_info): Use it.
25293 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
25295 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
25296 that the insn is actually an insn and it has a non-NULL basic block.
25297 Do not mark basic block dirty if only deleting a DEBUG_INSN.
25299 2013-03-22 Richard Biener <rguenther@suse.de>
25301 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
25303 (mem_ref_alloc): Do not allocate them.
25304 (refs_independent_p): Do not query or maintain a cache.
25306 2013-03-22 Richard Biener <rguenther@suse.de>
25308 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
25309 (gather_mem_refs_in_loops): Do not compute it.
25310 (analyze_memory_references): Do not allocate it.
25311 (tree_ssa_lim_finalize): Do not free it.
25312 (for_all_locs_in_loop): Do not query all_refs_in_loop.
25314 2013-03-22 Richard Biener <rguenther@suse.de>
25316 * is-a.h (as_a): Use gcc_checking_assert.
25318 2013-03-22 Ian Bolton <ian.bolton@arm.com>
25320 * config/aarch64/aarch64.c (aarch64_print_operand): New
25321 format specifier for printing a constant in hex.
25322 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
25323 format specifier for printing second operand.
25325 2013-03-22 Richard Biener <rguenther@suse.de>
25327 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
25329 (gather_mem_refs_in_loops): Perform store accumulation here.
25330 (create_vop_ref_mapping_loop): Remove.
25331 (create_vop_ref_mapping): Likewise.
25332 (analyze_memory_references): Initialize refs_stored_in_loop.
25333 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
25334 (record_indep_loop): Remove.
25335 (record_dep_loop): New function.
25336 (ref_indep_loop_p_1): Adjust to only walk over references
25337 in the loop, not its subloops.
25338 (ref_indep_loop_p): Rename to ...
25339 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
25340 maintaining a more fine-grained cache.
25341 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
25342 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
25344 2013-03-22 Richard Biener <rguenther@suse.de>
25346 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
25347 (struct mem_ref): Make accesses_in_loop a vec of a vec of
25348 aggregate mem_ref_loc.
25349 (free_mem_ref_locs): Inline into ...
25350 (memref_free): ... this and adjust.
25351 (mem_ref_alloc): Adjust.
25352 (mem_ref_locs_alloc): Remove.
25353 (record_mem_ref_loc): Adjust.
25354 (get_all_locs_in_loop): Rewrite into ...
25355 (for_all_locs_in_loop): ... this iterator.
25356 (rewrite_mem_ref_loc): New functor.
25357 (rewrite_mem_refs): Use for_all_locs_in_loop.
25358 (sm_set_flag_if_changed): New functor.
25359 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
25360 (ref_always_accessed): New functor.
25361 (ref_always_accessed_p): Use for_all_locs_in_loop.
25363 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
25365 * tree-pass.h (PROP_gimple_lvec): New.
25366 * passes.c (dump_properties): Handle PROP_gimple_lvec.
25367 (init_optimization_passes): Move pass_lower_vector.
25368 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
25370 (pass_lower_vector): Provide PROP_gimple_lvec.
25371 (pass_lower_vector_ssa): Likewise.
25372 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
25374 2013-03-21 Mark Wielaard <mjw@redhat.com>
25376 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
25378 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
25380 * config/i386/i386.md (*movdi_internal): Disparage slightly
25381 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
25382 conversion alternatives.
25384 2013-03-21 Jakub Jelinek <jakub@redhat.com>
25386 PR middle-end/48087
25387 * diagnostic.def (DK_WERROR): New kind.
25388 * diagnostic.h (werrorcount): Define.
25389 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
25390 promoted to DK_ERROR, increment DK_WERROR counter instead of
25392 * toplev.c (toplev_main): Call print_ignored_options even if
25393 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
25394 even if just werrorcount is non-zero.
25397 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
25399 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
25400 (string_cst_pool_decl): New function.
25401 (optimize_one_addr_into_implicit_ptr): New function.
25402 (resolve_addr_in_expr): Optimize DWARF location expression
25403 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
25404 which doesn't live in memory, but has DW_AT_location or
25405 DW_AT_const_value, or refers to a string literal, into
25406 DW_OP_GNU_implicit_pointer.
25407 (optimize_location_into_implicit_ptr): New function.
25408 (resolve_addr): If removing DW_AT_location of a variable because
25409 it was DW_OP_addr of address of the variable, but the variable doesn't
25410 live in memory, try to emit const value attribute for the initializer.
25412 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
25414 * tree.h (VECTOR_TYPE_P): New macro.
25415 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
25416 TYPE_MODE): Use it.
25417 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
25418 VEC_COND_EXPR cannot be lvalues.
25419 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
25421 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
25423 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
25424 Restrict the transformation to equal modes.
25426 2013-03-21 Richard Biener <rguenther@suse.de>
25428 PR tree-optimization/39326
25429 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
25430 (MEM_ANALYZABLE): Adjust.
25431 (record_mem_ref_loc): Move bitmap ops ...
25432 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
25433 unanalyzable refs, do not record locations for it.
25434 (analyze_memory_references): Allocate ref zero as shared
25436 (refs_independent_p): Do not test for unanalyzed mems here.
25437 (ref_indep_loop_p_1): Special-case disambiguation against
25438 the unanalyzed ref.
25439 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
25441 2013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
25443 * config/arm/arm-protos.h (tune_params): Add
25444 prefer_neon_for_64bits field.
25445 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
25446 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
25447 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
25448 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
25449 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
25450 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
25451 (arm_option_override): Handle -mneon-for-64bits new option.
25452 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
25453 (prefer_neon_for_64bits): Declare new variable.
25454 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
25455 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
25456 (arch_enabled): Handle new arch types. Remove support for onlya8
25458 (one_cmpldi2): Use new arch names.
25459 (zero_extend<mode>di2, extend<mode>di2): Ditto.
25460 * config/arm/arm.opt (mneon-for-64bits): Add option.
25461 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
25462 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
25463 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
25465 * doc/invoke.texi (-mneon-for-64bits): Document.
25467 2013-03-21 Richard Biener <rguenther@suse.de>
25469 PR tree-optimization/39326
25470 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
25471 (sort_bbs_in_loop_postorder_cmp): New function.
25472 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
25474 2013-03-21 Richard Biener <rguenther@suse.de>
25476 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
25477 (vect_insert_into_interleaving_chain): Likewise.
25478 (vect_drs_dependent_in_basic_block): Inline ...
25479 (vect_slp_analyze_data_ref_dependence): ... here. New function,
25481 (vect_analyze_data_ref_dependence): ... here. Simplify.
25482 (vect_check_interleaving): Simplify.
25483 (vect_analyze_data_ref_dependences): Likewise. Split out ...
25484 (vect_slp_analyze_data_ref_dependences): ... this new function.
25485 (dr_group_sort_cmp): New function.
25486 (vect_analyze_data_ref_accesses): Compute data-reference groups
25487 here instead of in vect_analyze_data_ref_dependence. Use
25488 a more efficient algorithm.
25489 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
25490 vect_slp_analyze_data_ref_dependences. Call
25491 vect_analyze_data_ref_accesses earlier.
25492 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
25493 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
25494 (vect_slp_analyze_data_ref_dependences): New prototype.
25496 2013-03-21 Richard Biener <rguenther@suse.de>
25498 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
25499 ref is stored in the loop.
25500 (find_refs_for_sm): Walk only over all stores.
25501 (store_motion_loop): Allocate from lim_bitmap_obstack.
25502 (store_motion): Likewise.
25504 2013-03-21 Richard Biener <rguenther@suse.de>
25506 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
25507 Update virtual SSA form.
25509 2013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
25511 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
25512 * configure: Regenerate.
25513 * config.in: Regenerate.
25514 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
25515 if !HAVE_LD_EH_FRAME_CIEV3.
25517 2013-03-21 Richard Biener <rguenther@suse.de>
25519 * tree-cfg.c (verify_expr_no_block): New function.
25520 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
25521 nor DECL_VALUE_EXPR have locations with associated blocks.
25522 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
25523 (clear_unused_block_pointer): Remove code dealing with
25524 blocks in DECL_DEBUG_EXPR locations.
25526 2013-03-21 Richard Biener <rguenther@suse.de>
25528 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
25529 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
25530 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
25531 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
25532 instead of DECL_DEBUG_EXPR_IS_FROM.
25533 * gimplify.c (gimplify_modify_expr): Likewise.
25534 * tree-cfg.c (verify_expr_location_1): Likewise.
25535 * tree-complex.c (create_one_component_var): Likewise.
25536 * tree-sra.c (create_access_replacement): Likewise.
25537 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
25538 (clear_unused_block_pointer): Likewise.
25539 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
25540 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
25541 * var-tracking.c (var_debug_decl): Likewise.
25542 (track_expr_p): Likewise.
25543 * tree-inline.c (add_local_variables): Likewise. Set
25544 DECL_HAS_DEBUG_EXPR_P after copying it.
25545 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
25546 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
25548 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
25551 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
25552 * configure: Regenerate.
25553 * config.in: Regenerate.
25554 * config/i386/i386.md (*movdf_internal): Use
25555 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
25556 movd instead of movq mnemonic for interunit moves.
25557 (*movdi_internal): Ditto.
25559 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
25561 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
25562 (abd<mode>_3): New pattern.
25563 (aba<mode>_3): New pattern.
25564 (fabd<mode>_3): New pattern.
25566 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
25568 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
25569 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
25570 occurrence of REGISTER_PREFIX as its empty string.
25572 2013-03-20 Jeff Law <law@redhat.com>
25574 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
25575 addititional equivalences for equality comparisons between an SSA_NAME
25576 and a constant where the SSA_NAME was set from a widening conversion.
25578 2013-03-20 Walter Lee <walt@tilera.com>
25580 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
25582 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
25584 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
25585 depending on type attribute.
25586 (*movti_internal): Ditto.
25587 (*movtf_internal): Ditto.
25588 (*movxf_internal): Ditto.
25589 (*movdf_internal): Ditto.
25590 (*movsf_internal): Ditto.
25592 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
25594 * config/i386/i386.md (*movti_internal): Set prefix attribute to
25595 maybe_vex for sselog1 and ssemov types.
25596 (*movdi_internal): Reorder operand constraints.
25597 (*movsi_internal): Ditto. Set prefix attribute to
25598 maybe_vex for sselog1 and ssemov types.
25599 (*movtf_internal): Set prefix attribute to maybe_vex
25600 for sselog1 and ssemov types.
25601 (*movdf_internal): Ditto. Set prefix_data16 attribute for
25602 DImode ssemov types. Reorder operand constraints.
25603 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
25604 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
25605 attribute for SImode ssemov types. Reorder operand constraints.
25607 2013-03-20 Martin Jambor <mjambor@suse.cz>
25609 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
25610 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
25612 2013-03-20 Pat Haugen <pthaugen@us.ibm.com>
25614 * config/rs6000/predicates.md (indexed_address, update_address_mem
25615 update_indexed_address_mem): New predicates.
25616 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
25617 attribute for load/store instructions.
25618 * config/rs6000/dfp.md (movsd_store): Likewise.
25619 (movsd_load): Likewise.
25620 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
25621 (unnamed HI->DI extend define_insn): Likewise.
25622 (unnamed SI->DI extend define_insn): Likewise.
25623 (unnamed QI->SI extend define_insn): Likewise.
25624 (unnamed QI->HI extend define_insn): Likewise.
25625 (unnamed HI->SI extend define_insn): Likewise.
25626 (unnamed HI->SI extend define_insn): Likewise.
25627 (extendsfdf2_fpr): Likewise.
25628 (movsi_internal1): Likewise.
25629 (movsi_internal1_single): Likewise.
25630 (movhi_internal): Likewise.
25631 (movqi_internal): Likewise.
25632 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
25633 attribute for load/store instructions.
25634 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
25636 (mov<mode>_softfloat): Likewise.
25637 (mov<mode>_hardfloat32): Likewise.
25638 (mov<mode>_hardfloat64): Likewise.
25639 (mov<mode>_softfloat64): Likewise.
25640 (movdi_internal32): Likewise.
25641 (movdi_internal64): Likewise.
25642 (probe_stack_<mode>): Likewise.
25644 2013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
25646 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
25647 floating point, and decimal floating point to reload iterator.
25649 * config/rs6000/constraints.md (wl constraint): New constraints to
25650 return FLOAT_REGS if certain options are used to reduce the number
25651 of separate patterns that exist in the file.
25652 (wx constraint): Likewise.
25653 (wz constraint): Likewise.
25655 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
25656 -mdebug=reg, print wg, wl, wx, and wz constraints.
25657 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
25658 Initialize the reload functions for 64-bit binary/decimal floating
25660 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
25661 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
25662 create the buffer on the stack to overcome not having a 32-bit
25664 (rs6000_emit_move): Likewise.
25665 (rs6000_secondary_memory_needed_rtx): Likewise.
25666 (rs6000_alloc_sdmode_stack_slot): Likewise.
25667 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
25668 via xxlxor, just like DFmode 0.0.
25670 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
25671 define as 1 if we are running on a power7 or newer.
25672 (enum r6000_reg_class_enum): Add new constraints.
25674 * config/rs6000/dfp.md (movsd): Delete, combine with binary
25675 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
25676 with other moves by using conditional constraits (wg). Use LFIWZX
25677 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
25678 (movsd splitter): Likewise.
25679 (movsd_hardfloat): Likewise.
25680 (movsd_softfloat): Likewise.
25682 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
25683 binary and decimal floating point moves.
25684 (fmove_ok): New attributes to combine binary and decimal floating
25685 point moves, and to combine power6x (mfpgpr) moves along normal
25687 (real_value_to_target): Likewise.
25688 (f32_lr): Likewise.
25689 (f32_lm): Likewise.
25690 (f32_li): Likewise.
25691 (f32_sr): Likewise.
25692 (f32_sm): Likewise.
25693 (f32_si): Likewise.
25694 (movsf): Combine binary and decimal floating point moves. Combine
25695 power6x (mfpgpr) moves with other moves by using conditional
25696 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
25697 (mov<mode> for SFmode/SDmode); Likewise.
25698 (SFmode/SDmode splitters): Likewise.
25699 (movsf_hardfloat): Likewise.
25700 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
25701 (movsf_softfloat): Likewise.
25702 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
25704 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
25705 wx and wz constraints.
25707 * config/rs6000/constraints.md (wg constraint): New constraint to
25708 return FLOAT_REGS if -mmfpgpr (power6x) was used.
25710 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
25713 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
25714 -mdebug=reg, print wg, wl, wx, and wz constraints.
25715 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
25716 Initialize the reload functions for 64-bit binary/decimal floating
25718 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
25719 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
25720 create the buffer on the stack to overcome not having a 32-bit
25722 (rs6000_emit_move): Likewise.
25723 (rs6000_secondary_memory_needed_rtx): Likewise.
25724 (rs6000_alloc_sdmode_stack_slot): Likewise.
25725 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
25726 via xxlxor, just like DFmode 0.0.
25728 * config/rs6000/dfp.md (movdd): Delete, combine with binary
25729 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
25730 with other moves by using conditional constraits (wg). Use LFIWZX
25731 and STFIWX for loading SDmode on power7.
25732 (movdd splitters): Likewise.
25733 (movdd_hardfloat32): Likewise.
25734 (movdd_softfloat32): Likewise.
25735 (movdd_hardfloat64_mfpgpr): Likewise.
25736 (movdd_hardfloat64): Likewise.
25737 (movdd_softfloat64): Likewise.
25739 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
25740 64-bit binary and decimal floating point moves.
25741 (FMOVE64X): Likewise.
25742 (movdf): Combine 64-bit binary and decimal floating point moves.
25743 Combine power6x (mfpgpr) moves with other moves by using
25744 conditional constraits (wg).
25745 (mov<mode> for DFmode/DDmode): Likewise.
25746 (DFmode/DDmode splitters): Likewise.
25747 (movdf_hardfloat32): Likewise.
25748 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
25749 (movdf_softfloat32): Likewise.
25750 (movdf_hardfloat64_mfpgpr): Likewise.
25751 (movdf_hardfloat64): Likewise.
25752 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
25753 (movdf_softfloat64): Likewise.
25754 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
25755 (reload_<mode>_load): Move to later in the file so they aren't in
25756 the middle of the floating point move insns.
25757 (reload_<mode>_store): Likewise.
25759 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
25762 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
25763 constraint if -mdebug=reg.
25764 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
25765 Enable using dd reload support if needed.
25767 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
25768 binary and decimal floating point moves in rs6000.md.
25769 (movtd_internal): Likewise.
25771 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
25772 decimal floating point moves.
25774 (movtf_internal): Likewise.
25775 (mov<mode>_internal, TDmode/TFmode): Likewise.
25776 (movtf_softfloat): Likewise.
25777 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
25779 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
25780 movdi_internal64, using wg constraint for move direct operations.
25781 (movdi_internal64): Likewise.
25783 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
25784 MODES_TIEABLE_P for selected modes. Print the numerical value of
25785 the various virtual registers. Use GPR/FPR first/last values,
25786 instead of hard coding the register numbers. Print which modes
25787 have reload functions registered.
25788 (rs6000_option_override_internal): If -mdebug=reg, trace the options
25789 settings before/after setting cpu, target and subtarget settings.
25790 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
25791 and for secondary reload failures in rs6000_secondary_reload_inner.
25792 (rs6000_secondary_reload_fail): Likewise.
25793 (rs6000_secondary_reload_inner): Likewise.
25795 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
25796 macros for first/last GPR and FPR registers.
25797 (LAST_GPR_REGNO): Likewise.
25798 (FIRST_FPR_REGNO): Likewise.
25799 (LAST_FPR_REGNO): Likewise.
25801 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
25802 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
25803 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
25804 (vcond<mode><mode>): Likewise.
25805 (vcondu<mode><mode>): Likewise.
25806 (vector_gtu<mode>): Likewise.
25807 (vector_gte<mode>): Likewise.
25808 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
25809 to prevent the compiler from converting DImode operations to TImode.
25810 (ior<mode>3): Likewise.
25811 (and<mode>3): Likewise.
25812 (one_cmpl<mode>2): Likewise.
25813 (nor<mode>3): Likewise.
25814 (andc<mode>3): Likewise.
25816 * config/rs6000/constraints.md (wt constraint): New constraint
25817 that returns VSX_REGS if TImode is allowed in VSX registers.
25819 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
25820 constant under VSX.
25822 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
25823 similar to TImode, but it is restricted to being in the GPRs.
25825 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
25826 TImode to occupy a single VSX register.
25828 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
25829 -mvsx-timode for power7/power8.
25830 (power7 cpu): Likewise.
25831 (power8 cpu): Likewise.
25833 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
25834 sure that TFmode/TDmode take up two registers if they are ever
25835 allowed in the upper VSX registers.
25836 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
25838 (rs6000_init_hard_regno_mode_ok): Likewise.
25839 (rs6000_debug_reg_global): Add debugging for PTImode and wt
25840 constraint. Print if LRA is turned on.
25841 (rs6000_option_override_internal): Give an error if -mvsx-timode
25842 and VSX is not enabled.
25843 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
25844 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
25845 to reg+offset addressing. Use PTImode when checking offset
25846 addresses for validity.
25847 (reg_offset_addressing_ok_p): Likewise.
25848 (rs6000_legitimate_offset_address_p): Likewise.
25849 (rs6000_legitimize_address): Likewise.
25850 (rs6000_legitimize_reload_address): Likewise.
25851 (rs6000_legitimate_address_p): Likewise.
25852 (rs6000_eliminate_indexed_memrefs): Likewise.
25853 (rs6000_emit_move): Likewise.
25854 (rs6000_secondary_reload): Likewise.
25855 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
25856 reloads to fpr registers to continue to use reg+offset addressing,
25857 but 64-bit reloads to altivec registers need reg+reg addressing.
25858 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
25859 it. Treat LO_SUM like a PLUS operation.
25860 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
25861 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
25862 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
25863 registers to share a register with a smaller sized type, since VSX
25864 puts scalars in the upper 64-bits.
25865 (print_operand): Add support for PTImode.
25866 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
25867 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
25868 registers, but don't have arithmetic support.
25869 (rs6000_memory_move_cost): Add test for VSX.
25870 (rs6000_opt_masks): Add -mvsx-timode.
25872 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
25875 (VSr): Use wt constraint for TImode.
25876 (VSv): Drop TImode support.
25877 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
25878 (vsx_movti_64bit): Likewise.
25879 (vsx_movti_32bit): Likewise.
25880 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
25881 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
25882 one '?' on the appropriate output constraint. Do not allow TImode
25883 logical operations on 32-bit systems.
25884 (vsx_ior<mode>3): Likewise.
25885 (vsx_xor<mode>3): Likewise.
25886 (vsx_one_cmpl<mode>2): Likewise.
25887 (vsx_nor<mode>3): Likewise.
25888 (vsx_andc<mode>3): Likewise.
25889 (vsx_concat_<mode>): Likewise.
25890 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
25892 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
25893 OPTION_MASK_VSX_TIMODE.
25894 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
25895 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
25897 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
25898 (TI2 iterator): New iterator for TImode, PTImode.
25899 (wd mode attribute): Add values for vector types.
25900 (movti_string): Replace TI move operations with operations for TImode
25901 and PTImode. Add support for TImode being allowed in VSX registers.
25902 (mov<mode>_string, TImode/PTImode): Likewise.
25903 (movti_ppc64): Likewise.
25904 (mov<mode>_ppc64, TImode/PTImode): Likewise.
25905 (TI mode splitters): Likewise.
25907 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
25910 2013-03-20 Marc Glisse <marc.glisse@inria.fr>
25912 PR tree-optimization/56355
25913 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
25914 Also handle integers with undefined overflow.
25916 2013-03-20 Catherine Moore <clm@codesourcery.com>
25917 Maciej W. Rozycki <macro@codesourcery.com>
25918 Tom de Vries <tom@codesourcery.com>
25919 Nathan Sidwell <nathan@codesourcery.com>
25920 Iain Sandoe <iain@codesourcery.com>
25921 Nathan Froyd <froydnj@codesourcery.com>
25922 Chao-ying Fu <fu@mips.com>
25924 * doc/extend.texi: (micromips, nomicromips, nocompression):
25925 Document new function attributes.
25926 * doc/invoke.texi (minterlink-compressed, mmicromips,
25927 m14k, m14ke, m14kec): Document new options.
25928 (minterlink-mips16): Update documentation.
25929 * doc/md.texi (ZC, ZD): Document new constraints.
25930 * configure.ac (gcc_cv_as_micromips): Check if linker
25931 supports the .set micromips directive.
25932 * configure: Regenerate.
25933 * config.in: Regenerate.
25934 * config/mips/mips-tables.opt: Regenerate.
25935 * config/mips/micromips.md: New file.
25936 * constraints.md (ZC, ZD): New constraints.
25937 * config/mips/predicates.md (movep_src_register): New predicate.
25938 (movep_src_operand): New predicate.
25939 (non_volatile_mem_operand): New predicate.
25940 * config/mips/mips.md (multimem): New type.
25941 (length): Differentiate between 17-bit and 18-bit branch offsets.
25942 (MOVEP1, MOVEP2): New mode iterator.
25943 (mov_<load>l): Use ZC constraint.
25944 (mov_<load>r): Likewise.
25945 (mov_<store>l): Likewise.
25946 (mov_<store>r): Likewise.
25947 (*branch_equality<mode>_inverted): Add microMIPS support.
25948 (*branch_equality<mode>): Likewise.
25949 (*jump_absolute): Likewise.
25950 (indirect_jump_<mode>): Likewise.
25951 (tablejump_<mode>): Likewise.
25952 (<optab>_internal): Likewise.
25953 (sibcall_internal): Likewise.
25954 (sibcall_value_internal): Likewise.
25955 (prefetch): Use constraint ZD.
25956 * config/mips/mips.opt (minterlink-compressed): New option.
25957 (minterlink-mips16): Now an alias for minterlink-compressed.
25958 (mmicromips): New option.
25959 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
25960 (compare_and_swap_12): Likewise.
25961 (sync_add<mode>): Likewise.
25962 (sync_<optab>_12): Likewise.
25963 (sync_old_<optab>_12): Likewise.
25964 (sync_new_<optab>_12): Likewise.
25965 (sync_nand_12): Likewise.
25966 (sync_old_nand_12): Likewise.
25967 (sync_new_nand_12): Likewise.
25968 (sync_sub<mode>): Likewise.
25969 (sync_old_add<mode>): Likewise.
25970 (sync_old_sub<mode>): Likewise.
25971 (sync_new_add<mode>): Likewise.
25972 (sync_new_sub<mode>): Likewise.
25973 (sync_<optab><mode>): Likewise.
25974 (sync_old_<optab><mode>): Likewise.
25975 (sync_new_<optab><mode>): Likewise.
25976 (sync_nand<mode>): Likewise.
25977 (sync_old_nand<mode>): Likewise.
25978 (sync_new_nand<mode>): Likewise.
25979 (sync_lock_test_and_set<mode>): Likewise.
25980 (test_and_set_12): Likewise.
25981 (atomic_compare_and_swap<mode>): Likewise.
25982 (atomic_exchange<mode>_llsc): Likewise.
25983 (atomic_fetch_add<mode>_llsc): Likewise.
25984 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
25985 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
25986 (umips_save_restore_pattern_p): Likewise.
25987 (umips_load_store_pair_p): Likewise.
25988 (umips_output_load_store_pair): Likewise.
25989 (umips_movep_target_p): Likewise.
25990 (umips_12bit_offset_address_p): Likewise.
25991 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
25992 (mips_base_mips16): Rename this...
25993 (mips_base_compression_flags): ...to this. Update all uses.
25994 (mips_attribute_table): Add micromips, nomicromips and nocompression.
25995 (mips_mips16_decl_p): Delete.
25996 (mips_nomips16_decl_p): Delete.
25997 (mips_get_compress_on_flags): New function.
25998 (mips_get_compress_off_flags): New function.
25999 (mips_get_compress_mode): New function.
26000 (mips_get_compress_on_name): New function.
26001 (mips_get_compress_off_name): New function.
26002 (mips_insert_attributes): Support multiple compression types.
26003 (mips_merge_decl_attributes): Likewise.
26004 (umips_12bit_offset_address_p): New function.
26005 (mips_start_function_definition): Emit .set micromips directive.
26006 (mips_call_may_need_jalx_p): New function.
26007 (mips_function_ok_for_sibcall): Add microMIPS support.
26008 (mips_print_operand_punctuation): Support short delay slots and
26010 (umips_swm_mask, umips_swm_encoding): New.
26011 (umips_build_save_restore): New function.
26012 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
26013 (was_mips16_p): Remove.
26014 (old_compression_mode): New.
26015 (mips_set_compression_mode): New function.
26016 (mips_set_current_function): Add microMIPS support.
26017 (mips_option_override): Likewise.
26018 (umips_save_restore_pattern_p): New function.
26019 (umips_output_save_restore): New function.
26020 (umips_load_store_pair_p_1): New function.
26021 (umips_load_store_pair_p): New function.
26022 (umips_output_load_store_pair_1): New function.
26023 (umips_output_load_store_pair): New function.
26024 (umips_movep_target_p) New function.
26025 (mips_prepare_pch_save): Add microMIPS support.
26026 * config/mips/mips.h (TARGET_COMPRESSION): New.
26027 (TARGET_CPU_CPP_BUILTINS): Update macro
26028 to use new compression flags and to support microMIPS.
26029 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
26030 (MIPS_ARCH_FLOAT_SPEC): Likewise.
26031 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
26032 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
26033 (ASM_SPEC): Support mmicromips and mno-micromips.
26034 (M16STORE_REG_P): New macro.
26035 (MIPS_CALL): Support TARGET_MICROMIPS.
26036 (MICROMIPS_J): New macro.
26037 (mips_base_mips16): Rename this...
26038 (mips_base_compression_flags): ...to this.
26039 (UMIPS_12BIT_OFFSET_P): New macro.
26040 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
26041 (MULTILIB_DIRNAMES): Likewise.
26042 2013-03-20 Richard Biener <rguenther@suse.de>
26044 PR tree-optimization/56661
26045 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
26046 the result does not have to be distinct.
26048 2013-03-20 Richard Biener <rguenther@suse.de>
26050 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
26053 2013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
26054 Steven Bosscher <steven@gcc.gnu.org>
26056 PR rtl-optimization/56605
26057 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
26059 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
26062 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
26063 that require movd instead of movq.
26065 2013-03-20 Richard Biener <rguenther@suse.de>
26067 * tree-ssa-structalias.c (struct variable_info): Add pointer
26068 to the first field of an aggregate with sub-vars. Make
26069 this and the pointer to the next subfield its ID.
26070 (vi_next): New function.
26071 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
26072 storedanything_id, integer_id): Increment by one.
26073 (new_var_info, get_call_vi, lookup_call_clobber_vi,
26074 get_call_clobber_vi): Adjust.
26075 (solution_set_expand): Simplify and speedup.
26076 (solution_set_add): Inline into ...
26077 (set_union_with_increment): ... this. Adjust accordingly.
26078 (do_sd_constraint): Likewise.
26079 (do_ds_constraint): Likewise.
26080 (do_complex_constraint): Simplify.
26081 (build_pred_graph): Adjust.
26082 (solve_graph): Likewise. Simplify and speedup.
26083 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
26084 get_constraint_for_component_ref, get_constraint_for_1,
26085 first_vi_for_offset, first_or_preceding_vi_for_offset,
26086 create_function_info_for, create_variable_info_for_1,
26087 create_variable_info_for, intra_create_variable_infos): Adjust.
26088 (init_base_vars): Push NULL for ID zero.
26089 (compute_points_to_sets): Adjust.
26091 2013-03-20 Richard Biener <rguenther@suse.de>
26093 * cfgloop.c (verify_loop_structure): Streamline and avoid
26094 ICEing on corrupt loop tree.
26095 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
26098 2013-03-20 Richard Biener <rguenther@suse.de>
26100 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
26101 check whether an SSA update is needed.
26103 2013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
26105 * config/mips/constraints.md (T): Rename to...
26109 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
26110 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
26112 2013-03-19 Ian Bolton <ian.bolton@arm.com>
26114 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
26115 (*subsi3_carryin_uxtw): Likewise.
26117 2013-03-19 Ian Bolton <ian.bolton@arm.com>
26119 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
26120 (*rorsi3_insn_uxtw): Likewise.
26122 2013-03-19 Ian Bolton <ian.bolton@arm.com>
26124 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
26125 (*extrsi5_insn_uxtw): Likewise.
26127 2013-03-19 Richard Biener <rguenther@suse.de>
26129 PR tree-optimization/56273
26130 * passes.c (init_optimization_passes): Move second VRP after DOM.
26132 2013-03-19 Uros Bizjak <ubizjak@gmail.com>
26134 * config/i386/i386.md (*movti_internal): Merge from
26135 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
26136 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
26137 nox64 isa attributes.
26139 2013-03-18 Richard Biener <rguenther@suse.de>
26141 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
26143 (merge_node_constraints): Likewise.
26144 (build_succ_graph): Likewise.
26145 (valid_graph_edge): Inline into single caller.
26146 (unify_nodes): Likewise. Use bitmap_set_bit return value
26148 (scc_visit): Fix formatting and variable use.
26149 (do_sd_constraint): Use gcc_checking_assert.
26150 (do_ds_constraint): Likewise.
26151 (do_complex_constraint): Likewise.
26152 (condense_visit): Likewise. Cleanup.
26153 (dump_pred_graph): New function.
26154 (perform_var_substitution): Dump the pred-graph before
26155 variable substitution.
26156 (find_equivalent_node): Use gcc_checking_assert.
26157 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
26159 2013-03-18 Richard Biener <rguenther@suse.de>
26161 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
26162 Remove cond_expr_stmt_list argument and do not gimplify the
26164 (vect_loop_versioning): Adjust.
26165 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
26166 Cleanup to use less temporaries.
26167 (vect_create_data_ref_ptr): Cleanup.
26169 2013-03-18 Jakub Jelinek <jakub@redhat.com>
26171 PR tree-optimization/56635
26172 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
26173 require types_compatible_p types.
26175 2013-03-18 Nick Clifton <nickc@redhat.com>
26177 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
26178 spurious backslash.
26180 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
26181 Add missing line to comment describing function.
26183 2013-03-18 Richard Biener <rguenther@suse.de>
26185 PR tree-optimization/56210
26186 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
26187 Handle string / character search functions.
26188 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
26190 2013-03-18 Richard Biener <rguenther@suse.de>
26192 PR middle-end/56483
26193 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
26194 and implement properly.
26195 * gimple.h (gimple_cond_single_var_p): Remove.
26197 2013-03-18 Richard Biener <rguenther@suse.de>
26199 * tree-data-ref.h (find_data_references_in_loop): Declare.
26200 * tree-data-ref.c (get_references_in_stmt): Use a stack
26201 vector pre-allocated in the callers.
26202 (find_data_references_in_stmt): Adjust.
26203 (graphite_find_data_references_in_stmt): Likewise.
26204 (create_rdg_vertices): Likewise.
26205 (find_data_references_in_loop): Export.
26206 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
26207 Compute dependences here...
26208 (vect_analyze_data_refs): ...not here. When we encounter
26209 a non-vectorizable data reference in basic-block vectorization
26210 truncate the data reference vector. Do not bother to
26211 fixup data-dependence information for gather loads.
26212 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
26213 of data references, as reported.
26215 2013-03-18 Richard Biener <rguenther@suse.de>
26217 PR tree-optimization/3713
26218 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
26219 has_constants and expr.
26220 (stmt_has_constants): Properly valueize SSA names when deciding
26221 whether the stmt has constants.
26223 2013-03-18 Richard Biener <rguenther@suse.de>
26225 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
26226 whole function when there is nothing to do.
26227 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
26228 * tree-vectorizer.c (vectorize_loops): Update virtual and
26229 loop-closed SSA once.
26230 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
26232 2013-03-18 Richard Biener <rguenther@suse.de>
26234 PR middle-end/56113
26235 * domwalk.c (bb_postorder): New global static.
26236 (cmp_bb_postorder): New function.
26237 (walk_dominator_tree): Replace scheme imposing an order for
26238 visiting dominator sons by one sorting them at the time they
26239 are pushed on the stack.
26241 2013-03-18 Richard Biener <rguenther@suse.de>
26243 PR tree-optimization/39326
26244 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
26245 (struct mem_ref): Replace mem member with ao_ref typed member.
26246 (MEM_ANALYZABLE): Adjust.
26247 (memref_eq): Likewise.
26248 (mem_ref_alloc): Likewise.
26249 (gather_mem_refs_stmt): Likewise.
26250 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
26251 (execute_sm_if_changed_flag_set): Adjust.
26252 (execute_sm): Likewise.
26253 (ref_always_accessed_p): Likewise.
26254 (refs_independent_p): Likewise.
26255 (can_sm_ref_p): Likewise.
26257 2013-03-18 Jakub Jelinek <jakub@redhat.com>
26260 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
26261 return 1 even for !unsignedp.
26263 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
26265 * config/i386/i386.md (isa): Add x64 and nox64.
26266 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
26267 (*pushtf): Enable *roF alternative for x64 isa only.
26268 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
26269 mode attribute of integer alternatives to DImode for TARGET_64BIT.
26270 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
26271 (*movtf_internal): Merge from *movtf_internal_rex64 and
26272 *movtf_internal_sse. Use x64 and nox64 isa attributes.
26273 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
26274 nox64 isa attributes.
26275 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
26276 nox64 isa attributes.
26277 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
26279 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
26281 * config/alpha/alpha.c (TARGET_LRA_P): New define.
26283 2013-03-17 Jakub Jelinek <jakub@redhat.com>
26286 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
26287 class names. Remove trailing comma after "ALL_REGS".
26289 2013-03-16 Jan Hubicka <jh@suse.cz>
26291 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
26292 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
26293 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
26294 of cgraph_get_create_node.
26295 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
26297 2013-03-16 Jason Merrill <jason@redhat.com>
26300 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
26301 with DW_AT_default_value.
26303 2013-03-16 Jakub Jelinek <jakub@redhat.com>
26305 * BASE-VER: Set to 4.9.0.
26307 2013-03-14 Andi Kleen <ak@linux.intel.com>
26310 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
26311 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
26312 Document _x* TSX intrinsics.
26314 2013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
26315 David Holsgrove <david.holsgrove@xilinx.com>
26317 * configure.ac: Add MicroBlaze TLS support detection.
26318 * configure: Regenerate.
26319 * config/microblaze/microblaze-protos.h
26320 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
26321 symbol_mentioned_p, label_mentioned_p): Add prototypes.
26322 * config/microblaze/microblaze.c (microblaze_address_type): Add
26323 ADDRESS_TLS and tls_reloc address types.
26324 (microblaze_address_info): Add tls_reloc.
26325 (TARGET_HAVE_TLS): Define.
26326 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
26327 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
26328 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
26329 load_tls_operand, microblaze_call_tls_get_addr,
26330 microblaze_legitimize_tls_address): New functions.
26331 (microblaze_classify_unspec): Handle UNSPEC_TLS.
26332 (get_base_reg): Use microblaze_tls_symbol_p.
26333 (microblaze_classify_address): Handle TLS.
26334 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
26335 label_mentioned_p and microblaze_tls_referenced_p.
26336 (microblaze_legitimize_address): Handle TLS.
26337 (microblaze_address_insns): Handle ADDRESS_TLS.
26338 (pic_address_needs_scratch): Handle TLS.
26339 (print_operand_address): Handle TLS.
26340 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
26341 (microblaze_expand_move): Handle TLS.
26342 (microblaze_legitimate_constant_p): Check
26343 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
26344 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
26345 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
26346 (PIC_OFFSET_TABLE_REGNUM): Set.
26347 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
26348 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
26349 (addsi3, movsi_internal2, movdf_internal): Update constraints
26350 * config/microblaze/predicates.md (arith_plus_operand): Define
26351 (move_operand): Redefine as move_src_operand,
26352 check microblaze_tls_referenced_p.
26354 2013-03-14 Ian Bolton <ian.bolton@arm.com>
26356 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
26357 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
26359 2013-03-14 Ian Bolton <ian.bolton@arm.com>
26361 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
26364 2013-03-14 Jakub Jelinek <jakub@redhat.com>
26366 PR tree-optimization/53265
26367 * common.opt (Waggressive-loop-optimizations): New option.
26368 * tree-ssa-loop-niter.c: Include tree-pass.h.
26369 (do_warn_aggressive_loop_optimizations): New function.
26370 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
26371 if number_of_latch_executions returned constant.
26372 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
26373 early. If number_of_latch_executions returned constant, set
26374 nb_iterations_upper_bound back to it.
26375 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
26377 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
26378 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
26380 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
26381 (MULTILIB_OSDIRNAMES): Set.
26382 * genmultilib: If defaultosdirname doesn't start with :: , set
26383 defaultosdirname2 instead, clear it and emit two . multilib_raw
26384 entries instead of just one.
26386 2013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
26388 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
26389 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
26390 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
26391 (SUBTARGET_OVERRIDE_OPTIONS): New.
26393 2013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
26396 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
26397 (musermode): Convert to Var(TARGET_USERMODE).
26398 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
26399 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
26400 * config/sh/sh.c (sh_option_override): Use
26401 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
26402 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
26404 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
26406 (udivsi3_i4_single, divsi3_i4_single): Use
26407 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
26409 2013-03-13 Dave Korn <dave.korn.cygwin@gmail.com>
26411 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
26414 2013-03-13 Richard Biener <rguenther@suse.de>
26416 PR tree-optimization/56608
26417 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
26418 calls when vectorizing basic-blocks.
26420 2013-03-13 Jakub Jelinek <jakub@redhat.com>
26423 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
26426 2013-03-12 Jakub Jelinek <jakub@redhat.com>
26428 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
26430 2013-03-11 Jan Hubicka <jh@suse.cz>
26433 * lto-streamer-out.c (output_symbol_p): Skip references from
26434 constructors of external variables.
26436 2013-03-11 Jan Hubicka <jh@suse.cz>
26438 PR middle-end/56571
26439 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
26441 * emit-rtl.c (verify_rtx_sharing): Likewise.
26442 (copy_insn_1): Likewise.
26443 * rtl.c (copy_rtx): Likewise.
26445 2013-03-11 Georg-Johann Lay <avr@gjlay.de>
26448 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
26449 output_operand_lossage message.
26451 2013-03-11 Richard Earnshaw <rearnsha@arm.com>
26454 * arm.c (shift_op): Validate RTL pattern on the fly.
26455 (arm_print_operand, case 'S'): Don't use shift_operator to validate
26458 2013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26461 * config/pa/pa.md (call_value): Check for calls to powf and direct to
26462 new call patterns that clobber %fr12.
26463 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
26464 split and postreload patterns.
26465 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
26466 registers %fr12 and %fr12R as call used.
26468 2013-03-09 Steven Bosscher <steven@gcc.gnu.org>
26470 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
26471 (canon_address, record_store, replace_read, check_mem_read_rtx,
26472 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
26473 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
26474 rest_of_handle_dse): Likewise.
26476 2013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
26478 PR middle-end/56524
26479 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
26481 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
26482 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
26483 (save_optabs_if_changed): Replace with...
26484 (init_tree_optimization_optabs): ...this.
26485 * optabs.c (save_optabs_if_changed): Rename to...
26486 (init_tree_optimization_optabs): ...this. Take the optimization node
26487 as argument. Do nothing if the base optabs are already correct.
26488 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
26489 to recompute optabs.
26490 * function.h (function): Remove optabs field.
26491 * function.c (invoke_set_current_function_hook): Call
26492 init_tree_optimization_optabs. Use the result to initialize
26495 2013-02-27 Aldy Hernandez <aldyh@redhat.com>
26497 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
26498 if GTMA_HAS_NO_INSTRUMENTATION.
26499 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
26500 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
26501 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
26502 * gimple-pretty-print.c (dump_gimple_transaction): Handle
26503 GTMA_HAS_NO_INSTRUMENTATION.
26505 2013-03-08 Jakub Jelinek <jakub@redhat.com>
26507 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
26510 2013-03-08 Marek Polacek <polacek@redhat.com>
26511 Jakub Jelinek <jakub@redhat.com>
26513 PR tree-optimization/56478
26514 * predict.c (is_comparison_with_loop_invariant_p): Change the
26515 type of loop_step to tree.
26516 (predict_loops): Adjust.
26517 (predict_iv_comparison): Perform the computations on double_ints.
26519 2013-03-08 Richard Biener <rguenther@suse.de>
26521 PR tree-optimization/56570
26522 * tree-cfg.c (verify_expr_location_1): Verify locations for
26524 * tree-sra.c (create_access_replacement): Strip locations
26525 from DECL_DEBUG_EXPRs.
26527 2013-03-08 Richard Biener <rguenther@suse.de>
26529 * tree-inline.c (expand_call_inline): Do not associate
26530 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
26531 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
26533 2013-03-08 Richard Biener <rguenther@suse.de>
26535 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
26536 or block changes with -Og. Fix for location / block encoding
26537 changes and PHI arguments with locations.
26539 2013-03-07 Steven Bosscher <steven@gcc.gnu.org>
26541 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
26543 (struct output_info): Likewise.
26544 (register_overhead): Remove bad gcc_assert.
26545 (bitmap_find_bit): If there is only a single bitmap element, do not
26546 count a miss as a search.
26547 (print_statistics): Update for counter type changes.
26548 (dump_bitmap_statistics): Likewise. Print headers such that they
26549 are properly lined up with the printed counters.
26551 2013-03-07 Jakub Jelinek <jakub@redhat.com>
26553 PR tree-optimization/56559
26554 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
26555 check that it has only a single use.
26557 2013-03-07 Richard Biener <rguenther@suse.de>
26559 * doc/invoke.texi (fwhole-program): Discourage use in combination
26562 2013-03-06 Jakub Jelinek <jakub@redhat.com>
26564 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
26566 PR tree-optimization/56539
26567 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
26568 instead of GSI_CONTINUE_LINKING as last argument to
26569 force_gimple_operand_gsi. Adjust function comment.
26571 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
26574 PR middle-end/56548
26575 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
26576 promoted mode, convert the result back to the original mode.
26578 2013-03-06 Richard Biener <rguenther@suse.de>
26580 PR middle-end/56294
26581 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
26582 (insert_updated_phi_nodes_compare_uids): New function.
26583 (update_ssa): Sort symbols_to_rename after UID before
26584 traversing it to insert PHI nodes.
26586 2013-03-06 Richard Biener <rguenther@suse.de>
26588 PR middle-end/50494
26589 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
26590 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
26593 2013-02-13 Richard Biener <rguenther@suse.de>
26596 * varasm.c (output_constant_def_1): Get the decl representing
26597 the constant as argument.
26598 (output_constant_def): Wrap output_constant_def_1.
26599 (make_decl_rtl): Use output_constant_def_1 with the decl
26600 representing the constant.
26601 (build_constant_desc): Optionally re-use a decl already
26602 representing the constant.
26603 (tree_output_constant_def): Adjust.
26605 2013-03-06 Joey Ye <joey.ye@arm.com>
26608 * gcc.c (convert_white_space): New function.
26609 (main): Handles white space in function name.
26611 2013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
26614 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
26615 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
26616 to SH_DIV_CALL_TABLE for TARGET_SH2.
26617 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
26619 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
26620 call-table options.
26622 2013-03-05 Sterling Augustine <saugustine@google.com>
26623 Cary Coutant <ccoutant@google.com>
26626 * dwarf2out.c (resolve_addr): Don't call
26627 remove_loc_list_addr_table_entries a second time for the same
26630 2013-03-05 Jakub Jelinek <jakub@redhat.com>
26633 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
26634 (avoid_complex_debug_insns): New function.
26635 (expand_debug_locations): Call it.
26637 PR rtl-optimization/56484
26638 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
26639 lifetimes of hard registers on small register class machines.
26641 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
26643 * config/microblaze/microblaze-protos.h: Rename
26644 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
26645 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
26647 (microblaze_fast_interrupt_function_p): New function.
26648 (microblaze_is_interrupt_handler): Rename to
26649 microblaze_is_interrupt_variant and add fast_interrupt check.
26650 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
26651 (save_restore_insns): Likewise.
26652 (compute_frame_size): Likewise.
26653 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
26654 (microblaze_globalize_label): Likewise.
26655 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
26656 * config/microblaze/microblaze.md: Use wrapper
26657 microblaze_is_interrupt_variant.
26659 2013-03-05 Kai Tietz <ktietz@redhat.com>
26661 * sdbout.c (sdbout_one_type): Switch to current function's section
26662 supporting cold/hot.
26664 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
26666 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
26669 2013-03-05 Jakub Jelinek <jakub@redhat.com>
26671 PR middle-end/56461
26672 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
26673 if VALGRIND_GET_VBITS is defined, temporarily make object
26674 memory all defined, and restore previous valgrind addressability
26675 and definability afterwards. Free this_object at the end.
26677 PR middle-end/56461
26678 * lra.c (lra): Call lra_clear_live_ranges if live_p,
26679 right before calling lra_create_live_ranges, also call it
26680 when clearing live_p. Only call lra_clear_live_ranges
26681 at the end if live_p.
26683 PR middle-end/56461
26684 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
26686 2013-03-05 Richard Biener <rguenther@suse.de>
26688 PR tree-optimization/56521
26689 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
26692 2013-03-05 Steven Bosscher <steven@gcc.gnu.org>
26695 * except.h (remove_unreachable_eh_regions): New prototype.
26696 * except.c (remove_eh_handler_splicer): New function, split out
26697 of remove_eh_handler.
26698 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
26699 warning about running it on many EH regions one at a time.
26700 (remove_unreachable_eh_regions_worker): New function, walk the
26701 EH tree in depth-first order and remove non-marked regions.
26702 (remove_unreachable_eh_regions): New function.
26703 * tree-eh.c (mark_reachable_handlers): New function, split out
26704 from remove_unreachable_handlers.
26705 (remove_unreachable_handlers): Use mark_reachable_handlers and
26706 remove_unreachable_eh_regions.
26707 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
26708 and remove_unreachable_eh_regions.
26710 2013-03-05 Richard Biener <rguenther@suse.de>
26712 PR middle-end/56525
26713 * loop-init.c (fix_loop_structure): Remove loops in two stages,
26714 not freeing them until the end.
26716 2013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
26718 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
26720 2013-03-05 Richard Biener <rguenther@suse.de>
26722 PR tree-optimization/56270
26723 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
26724 of loads after scheduling an SLP instance.
26726 2013-03-05 Jakub Jelinek <jakub@redhat.com>
26728 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
26730 (check_gcc_parallelize): Run guality.exp as a separate job from
26731 vect.exp with unsorted.exp and $(dg_target_exps) separately from
26732 struct-layout-1.exp with stackalign.exp.
26734 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
26736 PR middle-end/56461
26737 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
26738 load_index sbitmap even if some bit in it isn't set.
26740 PR middle-end/56461
26741 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
26742 (discover_iteration_bound_by_body_walk): Change queues to
26743 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
26744 spelling in comment. Call safe_push on queues[bound_index] directly.
26745 Release queues[queue_index] in every iteration unconditionally.
26746 Release bounds vector.
26748 PR middle-end/56461
26749 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
26750 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
26751 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
26754 2013-03-05 Richard Biener <rguenther@suse.de>
26757 * tree-inline.c (remap_blocks_to_null): New function.
26758 (expand_call_inline): When expanding a call stmt without
26759 an associated block inline remap all callee blocks to NULL.
26761 2013-03-05 Jakub Jelinek <jakub@redhat.com>
26763 PR rtl-optimization/56494
26764 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
26765 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
26766 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
26768 PR middle-end/56461
26769 * sel-sched-ir.c (free_sched_pools): Release
26770 succs_info_pool.stack[succs_info_pool.max_top] vectors too
26771 if succs_info_pool.max_top isn't -1.
26774 * opts.c (opts_obstack, opts_concat): Moved to...
26775 * opts-common.c (opts_obstack, opts_concat): ... here.
26777 2013-03-04 Jakub Jelinek <jakub@redhat.com>
26779 PR middle-end/56461
26780 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
26782 2013-03-04 Martin Jambor <mjambor@suse.cz>
26784 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
26785 all appropriate places.
26787 2013-01-04 Eric Botcazou <ebotcazou@adacore.com>
26789 PR tree-optimization/56424
26790 * ipa-split.c (split_function): Do not set the RSO flag if result is
26791 not by reference and its type is a register type.
26793 2013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
26795 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
26796 (microblaze_legitimate_pic_operand): Likewise
26797 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
26798 new function microblaze_legitimate_pic_operand
26799 * config/microblaze/microblaze-protos.h
26800 (microblaze_legitimate_pic_operand): Declare.
26802 2013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
26804 * config/microblaze/predicates.md (call_insn_simple_operand):
26805 New predicate for supported rtx code types.
26806 * config/microblaze/microblaze.md (call_internal1): Use
26807 call_insn_simple_operand predicate.
26809 2013-03-04 Jakub Jelinek <jakub@redhat.com>
26811 PR middle-end/56461
26812 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
26813 partitions.ordered_remove.
26815 PR middle-end/56461
26816 * tree-vect-stmts.c (vectorizable_conversion): Don't call
26817 vec_oprnds0.create (1) for modifier == NONE.
26819 PR middle-end/56461
26820 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
26821 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
26822 vec_oprnds1 right before pushing anything to it for
26825 PR middle-end/56461
26826 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
26827 set nbbs to 0 instead of having separate code path.
26828 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
26829 instead of false as last argument if returning NULL.
26831 2013-03-03 Sandra Loosemore <sandra@codesourcery.com>
26833 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
26834 the attribute is now called "target" instead of "option".
26835 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
26836 * doc/tm.texi.in (Target Attributes): Likewise document the correct
26837 attribute/pragma name for TARGET_OPTION_VALID_P and
26838 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
26839 * doc/tm.texi: Regenerated.
26841 2013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
26843 * config/microblaze/microblaze.c:
26844 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
26845 * config/microblaze/microblaze.h: Add -mxl-reorder to
26847 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
26848 instructions emitted if TARGET_REORDER.
26849 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
26850 or 0 for -m/-mno case, but initialises as 2 to detect default use case
26853 2013-03-01 Xinliang David Li <davidxl@google.com>
26855 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
26858 2013-03-01 Jakub Jelinek <jakub@redhat.com>
26860 PR middle-end/56461
26861 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
26862 vector even when returning true. Fix up function comment formatting.
26864 PR middle-end/56461
26865 * ira-build.c (ira_loop_nodes_count): New variable.
26866 (create_loop_tree_nodes): Initialize it.
26867 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
26869 PR middle-end/56461
26870 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
26871 method on dr_chain and result_chain.
26872 * tree-vect-stmts.c (vectorizable_store): Only call
26873 result_chain.create if j == 0.
26875 PR middle-end/56461
26876 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
26877 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
26878 before overwriting it.
26880 2013-03-01 Tobias Burnus <burnus@net-b.de>
26882 * doc/extended.texi (C Extensions): Change order in @menu
26884 (Other MIPS Built-in Functions): Move last MIPS entry before
26885 "picoChip Built-in Functions".
26886 (SH Built-in Functions): Move after RX Built-in Functions.
26887 * doc/gcc.texi (Introduction): Change order in @menu
26889 * doc/md.texi (Constraints): Ditto.
26890 * gty.texi (Type Information): Ditto.
26891 (User-provided marking routines for template types): Make
26893 * doc/invoke.texi (AArch64 Options): Move before
26894 "Adapteva Epiphany Options".
26896 2013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
26897 Jakub Jelinek <jakub@redhat.com>
26900 * asan.c (gate_asan): Lookup no_sanitize_address instead of
26901 no_address_safety_analysis attribute.
26902 * doc/extend.texi (no_address_safety_attribute): Rename to
26903 no_sanitize_address attribute, mention no_address_safety_analysis
26904 attribute as deprecated alias.
26906 2013-02-28 Jakub Jelinek <jakub@redhat.com>
26908 PR middle-end/56461
26909 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
26910 type to vec<vec<tree> > *.
26911 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
26912 to be vec<tree> instead of vec<tree> *, set vec_defs
26913 to vNULL and call vec_defs.create (number_of_vects), adjust other
26915 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
26916 vectorizable_condition): Adjust vect_get_slp_defs callers.
26918 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
26920 * config/aarch64/aarch64.c
26921 (aarch64_float_const_representable): Remove unused variable.
26923 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
26925 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
26927 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
26929 * config/aarch64/aarch64-builtins.c
26930 (aarch64_init_simd_builtins): Make static.
26932 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
26934 * config/aarch64/aarch64.c
26935 (aarch64_simd_make_constant): Make static.
26937 2013-02-28 Martin Jambor <mjambor@suse.cz>
26939 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
26940 with no initialization to the RHS of debug statements.
26942 2013-02-28 Martin Jambor <mjambor@suse.cz>
26944 PR tree-optimization/56294
26945 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
26947 (get_access_replacement): Do not call create_access_replacement.
26948 Assert a replacement exists.
26949 (get_repl_default_def_ssa_name): Create the replacement declaration
26952 2013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
26954 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
26955 final_end_function.
26957 2013-02-28 Marek Polacek <polacek@redhat.com>
26959 PR rtl-optimization/56466
26960 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
26961 if we're changing a loop.
26962 (peel_loops_completely): Likewise.
26964 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
26967 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
26969 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
26972 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
26973 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
26974 INTX_FTYPE_FX, FX_FTYPE_INTX.
26975 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
26977 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
26979 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
26980 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
26981 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
26982 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
26983 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
26984 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
26985 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
26986 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
26987 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
26988 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
26989 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
26990 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
26991 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
26992 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
26993 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
26994 (avrxmega6): Increase max flash segments from 5 to 6.
26995 * config/avr/t-multilib: Regenerate.
26996 * config/avr/avr-tables.opt: Regenerate.
26997 * doc/avr-mmcu.texi: Regenerate.
26999 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
27001 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
27002 (avr_device_to_arch): Rename to avr_device_to_ld.
27003 (avr_device_to_as): New prototype.
27004 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
27005 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
27006 * config/avr/driver-avr.c (avr_device_to_as): New.
27007 (avr_device_to_arch): Rename to avr_device_to_ld.
27009 2013-02-27 Jakub Jelinek <jakub@redhat.com>
27011 PR middle-end/56461
27012 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
27013 method on dr_chain and result_chain.
27015 PR middle-end/56461
27016 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
27017 pointer_set_destroy on not_executed_last_iteration.
27019 PR middle-end/56461
27020 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
27022 PR middle-end/56461
27023 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
27024 FOR_EACH_DEFINED_FUNCTION when freeing state.
27026 PR middle-end/56461
27027 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
27029 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
27032 PR middle-end/56461
27033 * ipa-cp.c (decide_whether_version_node): Call vec_free on
27034 known_aggs[i].items and release known_aggs vector.
27036 PR middle-end/56461
27037 * ipa-reference.c (propagate): Free node_info even for alias nodes.
27039 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27041 * config/microblaze/microblaze.c (microblaze_emit_compare):
27042 Use xor for EQ/NE comparisions.
27043 * config/microblaze/microblaze.md (cstoresf4): Add constraints
27044 (cbranchsf4): Adjust operator to comparison_operator.
27046 2013-02-27 Jakub Jelinek <jakub@redhat.com>
27048 PR middle-end/56461
27049 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
27051 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
27052 vec_safe_push, always update *slot.
27053 (redirect_edge_var_map_clear): Use vec_free.
27054 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
27055 (free_var_map_entry): Use vec_free.
27056 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
27057 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
27059 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
27061 PR middle-end/45472
27062 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
27063 when the may_trap_p bit of the exprs being merged differs.
27064 Reorder tests for speculativeness in the logical and operator.
27066 2013-02-27 Jakub Jelinek <jakub@redhat.com>
27068 * incpath.c (add_standard_paths): Use reconcat instead of concat
27069 where appropriate and avoid leaking memory.
27071 * opts.h: Include obstack.h.
27072 (opts_concat): New prototype.
27073 (opts_obstack): New declaration.
27074 * opts.c (opts_concat): New function.
27075 (opts_obstack): New variable.
27076 (init_options_struct): Call gcc_init_obstack on opts_obstack.
27077 (finish_options): Use opts_concat instead of concat
27078 and XOBNEWVEC instead of XNEWVEC.
27079 * opts-common.c (generate_canonical_option, decode_cmdline_option,
27080 generate_option): Likewise.
27081 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
27082 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
27085 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
27086 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
27088 2013-02-26 Jakub Jelinek <jakub@redhat.com>
27090 PR middle-end/56461
27091 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
27093 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
27095 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
27096 (arm_block_move_unaligned_straight): Likewise.
27097 (arm_adjust_block_mem): Likewise.
27099 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
27102 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
27103 temp, cond and label.
27104 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
27107 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
27108 * config/c6x/c6x.h (dbx_register_map): Update declaration.
27111 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
27112 of prologue/epilogue functions.
27115 * config/tilegx/tilegx.c (tilegx_expand_prologue):
27116 Remove unused variable cfa_offset.
27117 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
27120 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
27121 type promotion to unsigned.
27124 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
27125 for HOST_WIDE_INT of 32 bit / same size as int.
27126 (arm_block_move_unaligned_straight): Likewise.
27127 (arm_adjust_block_mem): Likewise.
27130 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
27133 2013-02-26 Marek Polacek <polacek@redhat.com>
27135 PR tree-optimization/56426
27136 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
27138 2013-02-26 Richard Biener <rguenther@suse.de>
27141 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
27142 unused variable loops.
27144 2013-02-26 Jakub Jelinek <jakub@redhat.com>
27146 PR tree-optimization/56448
27147 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
27148 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
27149 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
27150 later operands of the references, or even first operand for
27151 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
27153 PR tree-optimization/56443
27154 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
27155 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
27156 to type_for_mode langhook.
27158 2013-02-25 Matt Turner <mattst88@gmail.com>
27160 * doc/invoke.texi: Document r4700.
27162 2013-02-25 Richard Biener <rguenther@suse.de>
27164 PR tree-optimization/56175
27165 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
27167 (simplify_bitwise_binary): ... here. Also guard the conversion
27168 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
27170 2013-02-25 Catherine Moore <clm@codesourcery.com>
27173 2013-02-24 Catherine Moore <clm@codesourcery.com>
27174 Maciej W. Rozycki <macro@codesourcery.com>
27175 Tom de Vries <tom@codesourcery.com>
27176 Nathan Sidwell <nathan@codesourcery.com>
27177 Iain Sandoe <iain@codesourcery.com>
27178 Nathan Froyd <froydnj@codesourcery.com>
27179 Chao-ying Fu <fu@mips.com>
27181 * doc/extend.texi: (micromips, nomicromips, nocompression):
27182 Document new function attributes.
27183 * doc/invoke.texi (minterlink-compressed, mmicromips,
27184 m14k, m14ke, m14kec): Document new options.
27185 (minterlink-mips16): Update documentation.
27186 * doc/md.texi (ZC, ZD): Document new constraints.
27187 * configure.ac (gcc_cv_as_micromips): Check if linker
27188 supports the .set micromips directive.
27189 * configure: Regenerate.
27190 * config.in: Regenerate.
27191 * config/mips/mips-tables.opt: Regenerate.
27192 * config/mips/micromips.md: New file.
27193 * constraints.md (ZC, AD): New constraints.
27194 * config/mips/predicates.md (movep_src_register): New predicate.
27195 (movep_src_operand): New predicate.
27196 (non_volatile_mem_operand): New predicate.
27197 * config/mips/mips.md (multimem): New type.
27198 (length): Differentiate between 17-bit and 18-bit branch offsets.
27199 (MOVEP1, MOVEP2): New mode iterator.
27200 (mov_<load>l): Use ZC constraint.
27201 (mov_<load>r): Likewise.
27202 (mov_<store>l): Likewise.
27203 (mov_<store>r): Likewise.
27204 (*branch_equality<mode>_inverted): Add microMIPS support.
27205 (*branch_equality<mode>): Likewise.
27206 (*jump_absolute): Likewise.
27207 (indirect_jump_<mode>): Likewise.
27208 (tablejump_<mode>): Likewise.
27209 (<optab>_internal): Likewise.
27210 (sibcall_internal): Likewise.
27211 (sibcall_value_internal): Likewise.
27212 (prefetch): Use constraint ZD.
27213 * config/mips/mips.opt (minterlink-compressed): New option.
27214 (minterlink-mips16): Now an alias for minterlink-compressed.
27215 (mmicromips): New option.
27216 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
27217 (compare_and_swap_12): Likewise.
27218 (sync_add<mode>): Likewise.
27219 (sync_<optab>_12): Likewise.
27220 (sync_old_<optab>_12): Likewise.
27221 (sync_new_<optab>_12): Likewise.
27222 (sync_nand_12): Likewise.
27223 (sync_old_nand_12): Likewise.
27224 (sync_new_nand_12): Likewise.
27225 (sync_sub<mode>): Likewise.
27226 (sync_old_add<mode>): Likewise.
27227 (sync_old_sub<mode>): Likewise.
27228 (sync_new_add<mode>): Likewise.
27229 (sync_new_sub<mode>): Likewise.
27230 (sync_<optab><mode>): Likewise.
27231 (sync_old_<optab><mode>): Likewise.
27232 (sync_new_<optab><mode>): Likewise.
27233 (sync_nand<mode>): Likewise.
27234 (sync_old_nand<mode>): Likewise.
27235 (sync_new_nand<mode>): Likewise.
27236 (sync_lock_test_and_set<mode>): Likewise.
27237 (test_and_set_12): Likewise.
27238 (atomic_compare_and_swap<mode>): Likewise.
27239 (atomic_exchange<mode>_llsc): Likewise.
27240 (atomic_fetch_add<mode>_llsc): Likewise.
27241 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
27242 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
27243 (umips_save_restore_pattern_p): Likewise.
27244 (umips_load_store_pair_p): Likewise.
27245 (umips_output_load_store_pair): Likewise.
27246 (umips_movep_target_p): Likewise.
27247 (umips_12bit_offset_address_p): Likewise.
27248 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
27249 (mips_base_mips16): Rename this...
27250 (mips_base_compression_flags): ...to this. Update all uses.
27251 (mips_attribute_table): Add micromips, nomicromips and nocompression.
27252 (mips_mips16_decl_p): Delete.
27253 (mips_nomips16_decl_p): Delete.
27254 (mips_get_compress_on_flags): New function.
27255 (mips_get_compress_off_flags): New function.
27256 (mips_get_compress_mode): New function.
27257 (mips_get_compress_on_name): New function.
27258 (mips_get_compress_off_name): New function.
27259 (mips_insert_attributes): Support multiple compression types.
27260 (mips_merge_decl_attributes): Likewise.
27261 (umips_12bit_offset_address_p): New function.
27262 (mips_start_function_definition): Emit .set micromips directive.
27263 (mips_call_may_need_jalx_p): New function.
27264 (mips_function_ok_for_sibcall): Add microMIPS support.
27265 (mips_print_operand_punctuation): Support short delay slots and
27267 (umips_swm_mask, umips_swm_encoding): New.
27268 (umips_build_save_restore): New function.
27269 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
27270 (was_mips16_p): Remove.
27271 (old_compression_mode): New.
27272 (mips_set_compression_mode): New function.
27273 (mips_set_current_function): Add microMIPS support.
27274 (mips_option_override): Likewise.
27275 (umips_save_restore_pattern_p): New function.
27276 (umips_output_save_restore): New function.
27277 (umips_load_store_pair_p_1): New function.
27278 (umips_load_store_pair_p): New function.
27279 (umips_output_load_store_pair_1): New function.
27280 (umips_output_load_store_pair): New function.
27281 (umips_movep_target_p) New function.
27282 (mips_prepare_pch_save): Add microMIPS support.
27283 * config/mips/mips.h (TARGET_COMPRESSION): New.
27284 (TARGET_CPU_CPP_BUILTINS): Update macro
27285 to use new compression flags and to support microMIPS.
27286 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
27287 (MIPS_ARCH_FLOAT_SPEC): Likewise.
27288 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
27289 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
27290 (ASM_SPEC): Support mmicromips and mno-micromips.
27291 (M16STORE_REG_P): New macro.
27292 (MIPS_CALL): Support TARGET_MICROMIPS.
27293 (MICROMIPS_J): New macro.
27294 (mips_base_mips16): Rename this...
27295 (mips_base_compression_flags): ...to this.
27296 (UMIPS_12BIT_OFFSET_P): New macro.
27297 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
27298 (MULTILIB_DIRNAMES): Likewise.
27300 2013-02-25 Tom de Vries <tom@codesourcery.com>
27302 PR rtl-optimization/56131
27303 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
27304 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
27305 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
27307 2013-02-25 Tobias Burnus <burnus@net-b.de>
27309 * doc/invoke.texi (-fsanitize=): Move from optimization
27310 to debugging options.
27312 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
27314 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
27316 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
27317 Alexander Monakov <amonakov@ispras.ru>
27319 PR middle-end/56077
27320 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
27321 flush pending lists also on non-jumps. Adjust comment.
27323 2013-02-24 Catherine Moore <clm@codesourcery.com>
27324 Maciej W. Rozycki <macro@codesourcery.com>
27325 Tom de Vries <tom@codesourcery.com>
27326 Nathan Sidwell <nathan@codesourcery.com>
27327 Iain Sandoe <iain@codesourcery.com>
27328 Nathan Froyd <froydnj@codesourcery.com>
27329 Chao-ying Fu <fu@mips.com>
27331 * doc/extend.texi: (micromips, nomicromips, nocompression):
27332 Document new function attributes.
27333 * doc/invoke.texi (minterlink-compressed, mmicromips,
27334 m14k, m14ke, m14kec): Document new options.
27335 (minterlink-mips16): Update documentation.
27336 * doc/md.texi (ZC, ZD): Document new constraints.
27337 * configure.ac (gcc_cv_as_micromips): Check if linker
27338 supports the .set micromips directive.
27339 * configure: Regenerate.
27340 * config.in: Regenerate.
27341 * config/mips/mips-tables.opt: Regenerate.
27342 * config/mips/micromips.md: New file.
27343 * constraints.md (ZC, AD): New constraints.
27344 * config/mips/predicates.md (movep_src_register): New predicate.
27345 (movep_src_operand): New predicate.
27346 (non_volatile_mem_operand): New predicate.
27347 * config/mips/mips.md (multimem): New type.
27348 (length): Differentiate between 17-bit and 18-bit branch offsets.
27349 (MOVEP1, MOVEP2): New mode iterator.
27350 (mov_<load>l): Use ZC constraint.
27351 (mov_<load>r): Likewise.
27352 (mov_<store>l): Likewise.
27353 (mov_<store>r): Likewise.
27354 (*branch_equality<mode>_inverted): Add microMIPS support.
27355 (*branch_equality<mode>): Likewise.
27356 (*jump_absolute): Likewise.
27357 (indirect_jump_<mode>): Likewise.
27358 (tablejump_<mode>): Likewise.
27359 (<optab>_internal): Likewise.
27360 (sibcall_internal): Likewise.
27361 (sibcall_value_internal): Likewise.
27362 (prefetch): Use constraint ZD.
27363 * config/mips/mips.opt (minterlink-compressed): New option.
27364 (minterlink-mips16): Now an alias for minterlink-compressed.
27365 (mmicromips): New option.
27366 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
27367 (compare_and_swap_12): Likewise.
27368 (sync_add<mode>): Likewise.
27369 (sync_<optab>_12): Likewise.
27370 (sync_old_<optab>_12): Likewise.
27371 (sync_new_<optab>_12): Likewise.
27372 (sync_nand_12): Likewise.
27373 (sync_old_nand_12): Likewise.
27374 (sync_new_nand_12): Likewise.
27375 (sync_sub<mode>): Likewise.
27376 (sync_old_add<mode>): Likewise.
27377 (sync_old_sub<mode>): Likewise.
27378 (sync_new_add<mode>): Likewise.
27379 (sync_new_sub<mode>): Likewise.
27380 (sync_<optab><mode>): Likewise.
27381 (sync_old_<optab><mode>): Likewise.
27382 (sync_new_<optab><mode>): Likewise.
27383 (sync_nand<mode>): Likewise.
27384 (sync_old_nand<mode>): Likewise.
27385 (sync_new_nand<mode>): Likewise.
27386 (sync_lock_test_and_set<mode>): Likewise.
27387 (test_and_set_12): Likewise.
27388 (atomic_compare_and_swap<mode>): Likewise.
27389 (atomic_exchange<mode>_llsc): Likewise.
27390 (atomic_fetch_add<mode>_llsc): Likewise.
27391 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
27392 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
27393 (umips_save_restore_pattern_p): Likewise.
27394 (umips_load_store_pair_p): Likewise.
27395 (umips_output_load_store_pair): Likewise.
27396 (umips_movep_target_p): Likewise.
27397 (umips_12bit_offset_address_p): Likewise.
27398 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
27399 (mips_base_mips16): Rename this...
27400 (mips_base_compression_flags): ...to this. Update all uses.
27401 (mips_attribute_table): Add micromips, nomicromips and nocompression.
27402 (mips_mips16_decl_p): Delete.
27403 (mips_nomips16_decl_p): Delete.
27404 (mips_get_compress_on_flags): New function.
27405 (mips_get_compress_off_flags): New function.
27406 (mips_get_compress_mode): New function.
27407 (mips_get_compress_on_name): New function.
27408 (mips_get_compress_off_name): New function.
27409 (mips_insert_attributes): Support multiple compression types.
27410 (mips_merge_decl_attributes): Likewise.
27411 (umips_12bit_offset_address_p): New function.
27412 (mips_start_function_definition): Emit .set micromips directive.
27413 (mips_call_may_need_jalx_p): New function.
27414 (mips_function_ok_for_sibcall): Add microMIPS support.
27415 (mips_print_operand_punctuation): Support short delay slots and
27417 (umips_swm_mask, umips_swm_encoding): New.
27418 (umips_build_save_restore): New function.
27419 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
27420 (was_mips16_p): Remove.
27421 (old_compression_mode): New.
27422 (mips_set_compression_mode): New function.
27423 (mips_set_current_function): Add microMIPS support.
27424 (mips_option_override): Likewise.
27425 (umips_save_restore_pattern_p): New function.
27426 (umips_output_save_restore): New function.
27427 (umips_load_store_pair_p_1): New function.
27428 (umips_load_store_pair_p): New function.
27429 (umips_output_load_store_pair_1): New function.
27430 (umips_output_load_store_pair): New function.
27431 (umips_movep_target_p) New function.
27432 (mips_prepare_pch_save): Add microMIPS support.
27433 * config/mips/mips.h (TARGET_COMPRESSION): New.
27434 (TARGET_CPU_CPP_BUILTINS): Update macro
27435 to use new compression flags and to support microMIPS.
27436 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
27437 (MIPS_ARCH_FLOAT_SPEC): Likewise.
27438 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
27439 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
27440 (ASM_SPEC): Support mmicromips and mno-micromips.
27441 (M16STORE_REG_P): New macro.
27442 (MIPS_CALL): Support TARGET_MICROMIPS.
27443 (MICROMIPS_J): New macro.
27444 (mips_base_mips16): Rename this...
27445 (mips_base_compression_flags): ...to this.
27446 (UMIPS_12BIT_OFFSET_P): New macro.
27447 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
27448 (MULTILIB_DIRNAMES): Likewise.
27450 2013-02-24 Jakub Jelinek <jakub@redhat.com>
27453 * target-globals.c (save_target_globals): For init_reg_sets and
27454 target_reinit remporarily set this_fn_optabs to this_target_optabs.
27456 2013-02-22 James Greenhalgh <james.greenhalgh@arm.com>
27458 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
27459 * config/aarch64/t-aarch64
27460 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
27462 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
27464 PR inline-asm/56148
27465 * lra-constraints.c (process_alt_operands): Reload operand
27466 conflicting with earlier clobber only if no more other conflicting
27469 2013-02-22 Jakub Jelinek <jakub@redhat.com>
27472 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
27473 if not linking a shared library.
27475 2013-02-22 Seth LaForge <sethml@google.com>
27477 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
27479 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
27481 * config/arm/arm.md (split for extendsidi): Update condition.
27482 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
27483 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
27484 (qhs_zextenddi_cstr): Likewise.
27486 2013-02-21 Jakub Jelinek <jakub@redhat.com>
27488 PR middle-end/56420
27489 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
27490 avoid signed wrapping.
27491 (expand_mult): Handle properly multiplication by
27492 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
27493 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
27494 in the compiler if coeff is HOST_WIDE_INT_MIN.
27495 (expand_divmod): Don't make ext_op1 static, change it's type to
27496 uhwi. Avoid undefined behavior in -INTVAL (op1).
27498 PR rtl-optimization/50339
27499 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
27501 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
27502 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
27503 into splitting_ashiftrt field.
27504 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
27506 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
27509 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
27511 PR middle-end/56108
27512 * trans-mem.c (execute_tm_mark): Do not expand transactions that
27513 are sure to go irrevocable.
27515 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
27517 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
27518 scalars are valid operands.
27520 2013-02-21 Martin Jambor <mjambor@suse.cz>
27522 PR tree-optimization/56310
27523 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
27524 only matching indices and non-negative final offsets.
27525 (intersect_aggregates_with_edge): Pass src_idx to
27526 agg_replacements_to_vector. Pass src_idx insstead of index to
27527 intersect_with_agg_replacements.
27529 2013-02-21 Martin Jambor <mjambor@suse.cz>
27531 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
27532 instead of hard-wired defaults.
27534 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
27536 * doc/invoke.texi (MIPS Options): Update documentation of the
27537 floating-point multiply-accumulate instruction restrictions.
27539 2013-02-21 Kostya Serebryany <kcc@google.com>
27541 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
27542 asan_shadow_offset on x86_64 linux.
27544 2013-02-21 Richard Biener <rguenther@suse.de>
27546 PR tree-optimization/56415
27548 2013-02-11 Richard Biener <rguenther@suse.de>
27550 PR tree-optimization/56273
27551 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
27554 2013-02-21 Jakub Jelinek <jakub@redhat.com>
27557 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
27560 PR inline-asm/56405
27561 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
27562 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
27564 2013-02-20 Jan Hubicka <jh@suse.cz>
27566 PR tree-optimization/56265
27567 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
27568 when target is referenced for first time.
27570 2013-02-20 Richard Biener <rguenther@suse.de>
27572 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
27573 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
27574 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
27575 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
27576 not return anything.
27577 (rename_ssa_copies): Do not remove unused locals.
27578 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
27579 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
27580 * passes.c (execute_function_todo): Do not schedule unused locals
27581 removal if cleanup_tree_cfg did something.
27582 * tree-ssa-live.c (remove_unused_locals): Dump statistics
27583 about the number of removed locals.
27585 2013-02-20 Richard Biener <rguenther@suse.de>
27587 PR tree-optimization/56398
27588 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
27590 2013-02-20 Martin Jambor <mjambor@suse.cz>
27592 PR tree-optimization/55334
27593 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
27594 restricted pointers to arrays.
27596 2013-02-20 Richard Biener <rguenther@suse.de>
27597 Jakub Jelinek <jakub@redhat.com>
27599 PR tree-optimization/56396
27600 * tree-ssa-ccp.c (n_const_val): New static variable.
27601 (get_value): Return NULL for SSA names we don't have a lattice
27603 (ccp_initialize): Initialize n_const_val.
27604 * tree-ssa-copy.c (n_copy_of): New static variable.
27605 (init_copy_prop): Initialize n_copy_of.
27606 (get_value): Return NULL_TREE for SSA names we don't have a
27609 2013-02-20 Martin Jambor <mjambor@suse.cz>
27611 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
27613 2013-02-20 Richard Biener <rguenther@suse.de>
27615 * genpreds.c (write_lookup_constraint): Do not compare first
27616 letter of the constraint again.
27618 2013-02-20 Richard Biener <rguenther@suse.de>
27620 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
27622 (get_use_iv_cost): Terminate hashtable walk when coming across
27625 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
27627 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
27628 reassociation for avx2 targets.
27630 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27632 * config/microblaze/microblaze.c: microblaze_has_clz = 0
27633 Add version check for v8.10.a to enable microblaze_has_clz
27634 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
27635 version and TARGET_PATTERN_COMPARE check
27636 * config/microblaze/microblaze.md: New clzsi2 instruction
27638 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27640 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
27641 function before branching.
27643 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
27645 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
27647 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
27649 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
27651 PR middle-end/55889
27652 * sel-sched.c: Include ira.h.
27653 (implicit_clobber_conflict_p): New function.
27654 (moveup_expr): Use it.
27655 * Makefile.in (sel-sched.o): Depend on ira.h.
27657 2013-02-19 Richard Biener <rguenther@suse.de>
27659 PR tree-optimization/56384
27660 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
27661 (vn_hash_type): Split out from ...
27662 (vn_hash_constant_with_type): ... here.
27663 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
27664 (vn_phi_eq): Compare types from vn_phi_s structure.
27665 (vn_phi_lookup): Populate vn_phi_s type.
27666 (vn_phi_insert): Likewise.
27668 2013-02-19 Jakub Jelinek <jakub@redhat.com>
27670 PR tree-optimization/56350
27671 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
27672 if haven't found reduction or nested cycle operand, rather than
27673 asserting we must find it.
27675 PR tree-optimization/56381
27676 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
27679 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
27680 Jakub Jelinek <jakub@redhat.com>
27683 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
27684 (swap_optab_enable): Same.
27685 (init_all_optabs): Use argument instead of global.
27686 * tree.h (struct tree_optimization_option): New field target_optabs.
27687 * expr.h (init_all_optabs): Add argument to prototype.
27688 (TREE_OPTIMIZATION_OPTABS): New.
27689 (save_optabs_if_changed): Protoize.
27690 * optabs.h: Declare this_fn_optabs.
27691 * optabs.c (save_optabs_if_changed): New.
27692 Declare this_fn_optabs.
27693 (init_optabs): Add argument to init_all_optabs() call.
27694 * function.c (invoke_set_current_function_hook): Handle per
27696 * function.h (struct function): New field optabs.
27697 * config/mips/mips.c (mips_set_mips16_mode): Handle when
27698 optimization_current_node has changed.
27699 * target-globals.h (save_target_globals_default_opts): Protoize.
27700 * target-globals.c (save_target_globals_default_opts): New.
27702 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
27705 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
27706 registers %fr12 and %fr12R as call used.
27709 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
27710 and HImode, require all displacements to be an integer multiple of
27712 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
27713 only allow QImode and HImode when reload is in progress and strict is
27714 true. Likewise for symbolic addresses. Use base14_operand to check
27715 displacements in REG+BASE addresses.
27717 2013-02-18 Richard Biener <rguenther@suse.de>
27719 PR tree-optimization/56366
27720 * tree-vect-loop.c (get_initial_def_for_induction): Properly
27721 handle sign-conversion of outer-loop initial induction value.
27723 2013-02-18 Richard Biener <rguenther@suse.de>
27725 PR middle-end/56349
27726 * cfghooks.c (merge_blocks): If we merge a latch into another
27727 block adjust references to it.
27728 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
27729 (verify_loop_structure): Verify that a recorded latch is in fact
27732 2013-02-18 Richard Biener <rguenther@suse.de>
27734 PR tree-optimization/56321
27735 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
27736 order SSA name release and virtual operand unlinking.
27738 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27740 * config/microblaze/microblaze.md (save_stack_block): Define.
27741 (restore_stack_block): Likewise.
27743 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27745 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
27746 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
27747 * config/microblaze/microblaze.c (microblaze_option_override):
27748 Bail out early for PIC modes when target does not support PIC.
27750 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27752 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
27753 Replace with a microblaze version.
27754 (microblaze_trampoline_init): Adapt for microblaze.
27755 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
27758 2013-02-16 Jakub Jelinek <jakub@redhat.com>
27759 Dodji Seketeli <dodji@redhat.com>
27762 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
27763 (instrument_mem_region_access): Do not forget to always put
27764 instrumentation of the of 'base' and 'base + len' in a "if (len !=
27765 0) statement, even for cases where either 'base' or 'base + len'
27766 are not instrumented -- because they have been previously
27767 instrumented. Simplify the logic by putting all the statements
27768 instrument 'base + len' inside a sequence, and then insert that
27769 sequence right before the current insertion point. Then, to
27770 instrument 'base + len', just get an iterator on that statement.
27771 And do not forget to update the pointer to iterator the function
27772 received as argument.
27774 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
27776 PR rtl-optimization/56348
27777 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
27779 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
27781 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
27782 (clean_graph_dump_file): Pass base to start_graph_dump.
27784 2013-02-14 Richard Henderson <rth@redhat.com>
27787 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
27789 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
27791 * collect2-aix.h: Define F_LOADONLY.
27793 2013-02-14 Richard Biener <rguenther@suse.de>
27796 * varasm.c (output_constant_def_1): Get the decl representing
27797 the constant as argument.
27798 (output_constant_def): Wrap output_constant_def_1.
27799 (make_decl_rtl): Use output_constant_def_1 with the decl
27800 representing the constant.
27801 (build_constant_desc): Optionally re-use a decl already
27802 representing the constant.
27803 (tree_output_constant_def): Adjust.
27805 2013-02-14 Dodji Seketeli <dodji@redhat.com>
27808 * asan.c (instrument_builtin_call): Really put the length of the
27809 second source argument into src1_len.
27811 2013-02-13 Jakub Jelinek <jakub@redhat.com>
27813 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
27814 argument. If it is false, don't create edge from then_bb to
27816 (insert_if_then_before_iter): Pass true to it.
27817 (build_check_stmt): Pass false to it.
27818 (transform_statements): Flush hash table only on extended basic
27819 block boundaries, rather than at the beginning of every bb.
27820 Don't flush hash table on nonfreeing_call_p calls.
27821 * tree-flow.h (nonfreeing_call_p): New prototype.
27822 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
27824 2013-02-13 David S. Miller <davem@davemloft.net>
27826 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
27828 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
27831 * ira.c (max_regno_before_ira): Move from ...
27833 (fix_reg_equiv_init): Use max_regno_before_ira instead of
27836 2013-02-13 Jakub Jelinek <jakub@redhat.com>
27838 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
27840 2013-02-13 Richard Biener <rguenther@suse.de>
27843 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
27844 globals in MEM_REFs.
27846 2013-02-13 Richard Biener <rguenther@suse.de>
27848 * loop-init.c (loop_optimizer_init): Clear loop state when
27849 re-initializing preserved loops.
27850 * loop-unswitch.c (unswitch_single_loop): Return whether
27851 we unswitched the loop. Do not verify loop state here.
27852 (unswitch_loops): When we unswitched a loop discover new loops.
27854 2013-02-13 Kostya Serebryany <kcc@google.com>
27856 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
27858 * sanitizer.def: Rename __asan_init to __asan_init_v1.
27860 2013-02-12 Dodji Seketeli <dodji@redhat.com>
27862 Avoid instrumenting duplicated memory access in the same basic block
27863 * Makefile.in (asan.o): Add new dependency on hash-table.h
27864 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
27865 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
27866 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
27867 (free_mem_ref_resources, has_mem_ref_been_instrumented)
27868 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
27869 (get_mem_ref_of_assignment): New functions.
27870 (get_mem_refs_of_builtin_call): Extract from
27871 instrument_builtin_call and tweak a little bit to make it fit with
27873 (instrument_builtin_call): Use the new
27874 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
27875 of is_gimple_builtin_call.
27876 (instrument_derefs, instrument_mem_region_access): Insert the
27877 instrumented memory reference into the hash table.
27878 (maybe_instrument_assignment): Renamed instrument_assignment into
27879 this, and change it to advance the iterator when instrumentation
27880 actually happened and return true in that case. This makes it
27881 homogeneous with maybe_instrument_assignment, and thus give a
27882 chance to callers to be more 'regular'.
27883 (transform_statements): Clear the memory reference hash table
27884 whenever we enter a new BB, when we cross a function call, or when
27885 we are done transforming statements. Use
27886 maybe_instrument_assignment instead of instrumentation. No more
27887 need to special case maybe_instrument_assignment and advance the
27888 iterator after calling it; it's now handled just like
27889 maybe_instrument_call. Update comment.
27891 2013-02-13 Richard Biener <rguenther@suse.de>
27893 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
27894 Fix loop discovery code.
27896 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
27898 PR inline-asm/56148
27899 * lra-constraints.c (process_alt_operands): Match early clobber
27900 operand with itself. Check conflicts with earlyclobber only if
27901 the operand is not reloaded. Prefer to reload conflicting operand
27902 if earlyclobber and matching operands are the same.
27904 2013-02-12 Richard Biener <rguenther@suse.de>
27907 * lto-streamer-out.c (write_symbol): Do not output symbols
27908 for hard register variables.
27910 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
27913 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
27914 (umulsidi3_insn, mulsidi3_insn): New insns.
27916 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
27918 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
27919 (struct tune_params): Add vec_costs field.
27920 * config/arm/arm.c (arm_builtin_vectorization_cost)
27921 (arm_add_stmt_cost): New functions.
27922 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
27923 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
27924 (arm_default_vec_cost): New struct of type cpu_vec_costs.
27925 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
27926 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
27927 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
27928 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
27930 2013-02-12 Richard Biener <rguenther@suse.de>
27933 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
27934 decls again if possible.
27936 2013-02-12 Richard Biener <rguenther@suse.de>
27938 PR middle-end/56288
27939 * tree-ssa.c (verify_ssa_name): Fix check, move
27940 SSA_NAME_IN_FREE_LIST check up.
27942 2013-02-12 Jakub Jelinek <jakub@redhat.com>
27943 Steven Bosscher <steven@gcc.gnu.org>
27945 PR rtl-optimization/56151
27946 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
27947 equal to op0 or op1, and last_insn pattern is CODE operation
27948 with MEM dest and one of the operands matches that MEM.
27950 2013-02-11 Sriraman Tallam <tmsriram@google.com>
27952 * doc/extend.texi: Document Function Multiversioning and "default"
27953 parameter string to target attribute.
27954 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
27955 target attribute parameter is "default".
27956 (ix86_compare_version_priority): Remove checks for target attribute.
27957 (ix86_mangle_function_version_assembler_name): Change error to sorry.
27958 Remove check for target attribute equal to NULL. Add assert.
27959 (ix86_generate_version_dispatcher_body): Change error to sorry.
27961 2013-02-11 Iain Sandoe <iain@codesourcery.com>
27962 Jack Howarth <howarth@bromo.med.uc.edu>
27963 Patrick Marlier <patrick.marlier@gmail.com>
27966 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
27967 define ENDFILE_SPEC as TM_DESTRUCTOR.
27968 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
27970 2013-02-11 Alexander Potapenko <glider@google.com>
27971 Jack Howarth <howarth@bromo.med.uc.edu>
27972 Jakub Jelinek <jakub@redhat.com>
27975 * config/darwin.c (cdtor_record): Rename ctor_record.
27976 (sort_cdtor_records): Rename sort_ctor_records.
27977 (finalize_dtors): New routine to sort destructors by
27978 priority before use in assemble_integer.
27979 (machopic_asm_out_destructor): Use finalize_dtors if needed.
27981 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
27983 PR rtl-optimization/56275
27984 * simplify-rtx.c (avoid_constant_pool_reference): Check that
27985 offset is non-negative and less than cmode size before
27986 calling simplify_subreg.
27988 2013-02-11 Richard Biener <rguenther@suse.de>
27990 PR tree-optimization/56264
27991 * cfgloop.h (fix_loop_structure): Adjust prototype.
27992 * loop-init.c (fix_loop_structure): Return the number of
27993 newly discovered loops.
27994 * tree-cfgcleanup.c (repair_loop_structures): When new loops
27995 are discovered, do a full loop-closed SSA rewrite.
27997 2013-02-11 Richard Biener <rguenther@suse.de>
27999 PR tree-optimization/56273
28000 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
28002 (check_array_ref): Fix missing newline in dumps.
28003 (search_for_addr_array): Likewise.
28005 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
28007 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
28009 2013-02-09 Jakub Jelinek <jakub@redhat.com>
28012 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
28014 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
28016 PR rtl-optimization/56246
28017 * lra-constraints.c (simplify_operand_subreg): Try to reuse
28019 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
28020 constraints are satisfied.
28022 2013-02-08 Jeff Law <law@redhat.com>
28025 * emit-rtl.c (reg_is_parm_p): New function.
28026 * regs.h (reg_is_parm_p): New prototype.
28027 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
28028 callee-clobbered registers.
28030 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
28033 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
28034 If there is no implicit builtin declaration, just return NULL.
28036 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
28038 * config/i386/sse.md (FMAMODEM): New mode iterator.
28039 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
28040 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
28042 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
28044 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
28045 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
28046 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
28048 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
28050 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
28051 (microblaze*-*-elf): Likewise.
28052 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
28054 * config/microblaze/microblaze-c.c: Add builtin defines for
28055 _LITTLE_ENDIAN and _BIG_ENDIAN.
28056 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
28057 add to TARGET_DEFAULT flags.
28058 Expand ASM_SPEC and LINK_SPEC.
28059 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
28060 * config/microblaze/microblaze.md: Update extendsidi2 and
28061 movdi_internal instructions to use low-order / high-order reg
28063 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
28064 options and inversemask / mask of LITTLE_ENDIAN.
28065 * config/microblaze/t-microblaze: Expand multilib options to
28066 include mlittle-endian (le) and update exceptions patterns.
28068 2013-02-08 Jakub Jelinek <jakub@redhat.com>
28070 PR rtl-optimization/56195
28071 * lra-constraints.c (get_reload_reg): Don't reuse regs
28072 if they have smaller mode than requested, if they have
28073 wider mode than requested, try to return a SUBREG.
28075 PR tree-optimization/56250
28076 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
28077 if type is unsigned and code isn't MULT_EXPR.
28079 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
28081 PR tree-optimization/56064
28082 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
28083 bits according to mode.
28084 * fixed-value.h (fixed_from_double_int)
28085 (const_fixed_from_double_int): Adjust comments.
28087 2013-02-08 Richard Biener <rguenther@suse.de>
28090 * lto-streamer.h (struct data_in): Remove current_file, current_line
28091 and current_col members.
28092 * lto-streamer-out.c (lto_output_location): Stream changed bits
28093 en-block for efficiency.
28094 * lto-streamer-in.c (clear_line_info): Remove.
28095 (lto_input_location): Cache current file, line and column
28096 globally via local statics. Read changed bits en-block.
28097 (input_function): Do not call clear_line_info.
28098 (lto_read_body): Likewise.
28099 (lto_input_toplevel_asms): Likewise.
28101 2013-02-08 Michael Matz <matz@suse.de>
28103 PR tree-optimization/52448
28104 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
28105 (nt_call_phase): New static.
28106 (add_or_mark_expr): Only mark accesses with newer phase than any
28108 (nonfreeing_call_p): New.
28109 (nt_init_block): Update nt_call_phase, mark blocks as visited.
28110 (nt_fini_block): Keep blocks marked as visited.
28111 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
28113 2013-02-08 Richard Biener <rguenther@suse.de>
28115 * ira.c (ira): Free broken dominator information.
28117 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
28119 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
28121 2013-02-08 Marek Polacek <polacek@redhat.com>
28123 * cfgloop.c (verify_loop_structure): Add more checking of headers.
28125 2013-02-08 Richard Biener <rguenther@suse.de>
28127 PR middle-end/56181
28128 * cfgloop.h (flow_loops_find): Adjust.
28129 (bb_loop_header_p): Declare.
28130 * cfgloop.c (bb_loop_header_p): New function split out from ...
28131 (flow_loops_find): ... here. Adjust function signature,
28132 support incremental loop structure update.
28133 (verify_loop_structure): Cleanup. Verify a loop is a loop.
28134 * cfgloopmanip.c (fix_loop_structure): Move ...
28135 * loop-init.c (fix_loop_structure): ... here.
28136 (apply_loop_flags): Split out from ...
28137 (loop_optimizer_init): ... here.
28138 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
28139 in incremental mode, only remove dead loops here.
28141 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
28144 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
28145 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
28146 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
28147 (*round<mode>3.libgcc): New insns for fixed-modes.
28148 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
28149 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
28150 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
28151 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
28152 implementations. Define to __builtin_avr_absFX,
28153 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
28154 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
28155 __builtin_avr_countlsFX, respectively.
28156 * config/avr/avr-c.c (target.h): Include it.
28157 (enum avr_builtin_id): New enum.
28158 (avr_resolve_overloaded_builtin): New static function.
28159 (avr_register_target_pragmas): Use it to set
28160 targetm.resolve_overloaded_builtin.
28161 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
28162 tree nodes used by DEF_BUILTIN.
28163 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
28164 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
28165 <AVR_BUILTIN_xxBITS>: Same.
28167 2013-02-08 Richard Biener <rguenther@suse.de>
28169 * cfgloop.c (verify_loop_structure): Properly handle
28170 a loop exiting to another loop header.
28171 * ira-int.h (ira_loops): Remove.
28172 * ira.c (ira_loops): Remove.
28173 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
28174 (do_reload): Use loop_optimizer_finalize.
28175 * ira-build.c (create_loop_tree_nodes): Use get_loops and
28176 number_of_loops to access the loop tree.
28177 (more_one_region_p): Likewise.
28178 (finish_loop_tree_nodes): Likewise.
28179 (rebuild_regno_allocno_maps): Likewise.
28180 (mark_loops_for_removal): Likewise.
28181 (mark_all_loops_for_removal): Likewise.
28182 (remove_unnecessary_regions): Likewise.
28183 (ira_build): Likewise.
28184 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
28186 2013-02-08 Richard Biener <rguenther@suse.de>
28188 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
28189 * ipa-pure-const.c (analyze_function): Avoid calling
28190 mark_irreducible_loops twice.
28191 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
28193 2013-02-07 David S. Miller <davem@davemloft.net>
28195 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
28197 * var-tracking.c (vt_add_function_parameter): Test the presence of
28198 HAVE_window_save properly and do not remap argument registers when
28199 we have a leaf function.
28201 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
28204 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
28206 * config/i386/i386.c (ix86_print_operand): Ditto.
28208 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
28210 * lra-constraints.c (process_alt_operands): Fix recently added comment.
28212 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
28214 PR rtl-optimization/56225
28215 * lra-constraints.c (process_alt_operands): Check that reload hard
28216 reg can hold value for strict_low_part.
28218 2013-02-07 Jakub Jelinek <jakub@redhat.com>
28221 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
28222 dwarf2out_end_function.
28223 (in_first_function_p, maybe_at_text_label_p,
28224 first_loclabel_num_not_at_text_label): New variables.
28225 (dwarf2out_var_location): In the first function find out
28226 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
28227 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
28230 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
28232 PR rtl-optimization/56178
28233 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
28234 SUBREG of a register. Tidy up related block of code.
28235 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
28236 note if the source is a register or a SUBREG of a register.
28238 2013-02-07 Jakub Jelinek <jakub@redhat.com>
28241 * config/rs6000/rs6000.md (ptrm): New mode attr.
28242 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
28243 call_value_indirect_aix<pttrsize>,
28244 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
28247 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
28249 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
28250 if -bnortl. Convert to strcmp and strncmp.
28252 2013-02-07 Alan Modra <amodra@gmail.com>
28255 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
28256 addresses won't wrap when offsetting.
28257 (rs6000_secondary_reload): Provide secondary reloads needed for
28258 wrapping LO_SUM addresses.
28260 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
28262 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
28263 MACH, just __MACH__.
28265 2013-02-06 Richard Biener <rguenther@suse.de>
28267 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
28268 instead of calling fix_loop_structure.
28270 2013-02-06 Jakub Jelinek <jakub@redhat.com>
28272 PR middle-end/56217
28273 * omp-low.c (use_pointer_for_field): Return false if
28274 lower_send_shared_vars doesn't generate any copy-out code.
28276 2013-02-06 Tom de Vries <tom@codesourcery.com>
28278 PR rtl-optimization/56131
28279 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
28280 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
28281 of the label is NULL. Add comment.
28283 2013-02-05 Jakub Jelinek <jakub@redhat.com>
28285 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
28288 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
28289 (STATIC_LIBTSAN_LIBS): Likewise.
28290 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
28291 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
28292 is defined, don't add anything else beyond that.
28293 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
28294 (LINK_COMMAND_SPEC): Use them.
28296 PR tree-optimization/56205
28297 * tree-stdarg.c (check_all_va_list_escapes): Return true if
28298 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
28299 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
28301 2013-02-05 Richard Biener <rguenther@suse.de>
28303 PR tree-optimization/53342
28304 PR tree-optimization/53185
28305 * tree-vectorizer.h (vect_check_strided_load): Remove.
28306 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
28307 not disallow peeling for vectorized strided loads.
28308 (vect_check_strided_load): Make static and simplify.
28309 (vect_analyze_data_refs): Adjust.
28310 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
28311 correctly when vectorizing strided loads.
28313 2013-02-05 Richard Biener <rguenther@suse.de>
28315 * doc/install.texi: Refer to ISL, not PPL.
28317 2013-02-05 Jan Hubicka <jh@suse.cz>
28319 PR tree-optimization/55789
28320 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
28322 2013-02-05 Jan Hubicka <jh@suse.cz>
28324 PR tree-optimization/55789
28325 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
28326 the dead call anyway.
28328 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
28331 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
28333 2013-02-04 Alexander Potapenko <glider@google.com>
28334 Jack Howarth <howarth@bromo.med.uc.edu>
28335 Jakub Jelinek <jakub@redhat.com>
28338 * config/darwin.c (sort_ctor_records): Stabilized qsort
28339 on constructor priority by using original position.
28340 (finalize_ctors): New routine to sort constructors by
28341 priority before use in assemble_integer.
28342 (machopic_asm_out_constructor): Use finalize_ctors if needed.
28344 2013-02-04 Jakub Jelinek <jakub@redhat.com>
28347 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
28348 about visibility on artificial decls.
28349 * config/sol2.c (solaris_assemble_visibility): Likewise.
28351 2013-02-04 Kai Tietz <ktietz@redhat.com>
28354 * config/i386/i386.c (function_value_ms_64): Add additional valtype
28355 argument and improve checking of return-argument types for 16-byte
28357 (ix86_function_value_1): Add additional valtype argument on call
28358 of function_value_64.
28359 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
28360 handling infunction_value_64 function.
28362 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
28364 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
28366 2013-02-04 Richard Biener <rguenther@suse.de>
28368 PR tree-optimization/56188
28369 * tree-ssa-structalias.c (label_visit): Consider case with
28370 initially non-empty points-to set.
28371 (perform_var_substitution): Dump node mapping and clean up.
28373 2013-02-04 Richard Guenther <rguenther@suse.de>
28376 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
28377 node prevail as last resort.
28378 (lto_symtab_merge_decls): Remove guard on LTRANS here.
28379 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
28381 2013-02-04 Richard Biener <rguenther@suse.de>
28383 PR tree-optimization/56113
28384 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
28386 (equiv_class_lookup_or_add): ... this.
28387 (label_visit): Adjust and fix error in previous patch.
28388 (perform_var_substitution): Adjust.
28390 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
28392 * config/sh/divtab.c: Fix formatting and comments throughout the file.
28393 * config/sh/sh4-300.md: Likewise.
28394 * config/sh/sh4a.md: Likewise.
28395 * config/sh/constraints.md: Likewise.
28396 * config/sh/sh.md: Likewise.
28397 * config/sh/netbsd-elf.h: Likewise.
28398 * config/sh/predicates.md: Likewise.
28399 * config/sh/sh-protos.h: Likewise.
28400 * config/sh/ushmedia.h: Likewise.
28401 * config/sh/linux.h: Likewise.
28402 * config/sh/sh.c: Likewise.
28403 * config/sh/superh.h: Likewise.
28404 * config/sh/elf.h: Likewise.
28405 * config/sh/sh4.md: Likewise.
28406 * config/sh/sh.h: Likewise.
28408 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
28410 * config/pa/constraints.md: Adjust unused letters. Change "T"
28411 constraint to match_test floating_point_store_memory_operand().
28412 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
28413 (base14_operand): New.
28414 (floating_point_store_memory_operand): New.
28415 (integer_store_memory_operand): Revise to use base14_operand and
28416 reg_plus_base_memory_operand.
28417 (move_dest_operand): Allow symbolic_memory_operands.
28418 (symbolic_memory_operand): Check for LO_SOM.
28419 (symbolic_operand): Change default case to break.
28420 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
28421 CONST_DOUBLE values to be reloaded by putting them into memory when
28422 the destination is a floating point register.
28423 (movdf): Remove code to handle CONST_DOUBLE.
28425 (reload_indf_r1): New.
28426 (reload_insf_r1): New.
28427 Consistently use "Q" and "T" constraints with integer and floating
28428 point move instructions, respectively.
28429 (movdi): Remove FAIL.
28430 Change predicate for source operand unamed DImode move from
28431 general_operand to move_src_operand.
28432 (umulsidi3): Change predicate for destination operand to
28434 Likewise for similar unamed patterns.
28435 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
28436 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
28437 (hppa_legitimize_address): Simplify mask calculation.
28438 (pa_emit_move_sequence): Revised handling of secondary reloads from
28439 REG+D addresses for floating point loads and stores. Directly handle
28440 loading CONST0_RTX (mode) to a floating point register.
28441 (pa_secondary_reload): Handle reloading DF and SFmode constant values
28442 to floating point registers. Don't restrict secondary reloads to
28443 floating point registers to integer modes. Revise some comments and
28445 (TARGET_LEGITIMATE_ADDRESS_P): Define.
28446 (pa_legitimate_address_p): New.
28447 (pa_legitimize_reload_address): New.
28448 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
28449 (STRICT_REG_OK_FOR_BASE_P): New.
28450 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
28451 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
28453 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
28454 Andrew Dixie <andrewd@gentrack.com>
28456 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
28459 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
28461 * expmed.c (extract_bit_field_1): Pass the full width of the
28462 structure to get_best_reg_extraction_insn.
28464 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
28467 * configure.ac (use_cxa_atexit): Add AIX.
28468 * configure: Regenerate.
28470 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
28472 2013-02-01 Jakub Jelinek <jakub@redhat.com>
28475 * final.c (need_profile_function): New variable.
28476 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
28477 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
28478 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
28479 notes, targetm.asm_out.function_prologue doesn't emit anything,
28480 HAVE_prologue and profiler should be emitted before prologue,
28481 set need_profile_function instead of emitting it.
28482 (final_scan_insn): If need_profile_function, emit
28483 profile_function on the first NOTE_INSN_BASIC_BLOCK or
28484 NOTE_INSN_FUNCTION_BEG note.
28486 2013-02-01 Richard Henderson <rth@redhat.com>
28488 * config/rs6000/rs6000.md (smulditi3): New.
28491 * config/alpha/alpha.md (umulditi3): New.
28493 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
28495 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
28496 (ASM_OUTPUT_ALIGNED_LOCAL): New.
28498 2013-02-01 Richard Biener <rguenther@suse.de>
28500 PR tree-optimization/56113
28501 * tree-ssa-structalias.c (label_visit): Reduce work for
28502 single-predecessor nodes.
28504 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
28506 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
28507 range isn't testing for zero.
28509 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
28511 PR middle-end/56113
28512 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
28514 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
28515 Nick Clifton <nickc@redhat.com>
28517 * config/v850/constraints.md (Q): Define as a memory constraint.
28518 * config/v850/predicates.md (label_ref_operand): New predicate.
28519 (e3v5_shift_operand): New predicate.
28520 (ior_operator): New predicate.
28521 * config/v850/t-v850: Add e3v5 multilib.
28522 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
28523 (v850_gen_movdi): Prototype.
28524 * config/v850/v850.c: Add support for e3v5 architecture.
28525 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
28527 (construct_save_jarl): Add e3v5 long JARL support.
28528 (v850_adjust_insn_length): New function. Adjust length of call
28529 insns when using e3v5 instructions.
28530 (v850_gen_movdi): New function: Generate instructions to move a
28532 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
28533 (CPP_SPEC): Define __v850e3v5__ as appropriate.
28534 (TARGET_USE_FPU): Enable for e3v5.
28535 (CONST_OK_FOR_W): New macro.
28536 (ADJUST_INSN_LENGTH): Define.
28537 * config/v850/v850.md (UNSPEC_LOOP): Define.
28538 (attr cpu): Add v850e3v5.
28539 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
28540 (movdi): New pattern.
28541 (movdi_internal): New pattern.
28542 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
28543 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
28544 (cstoresf4): Likewise.
28545 (cstoredf4): Likewise.
28546 (insv): New pattern.
28547 (rotlso3_a): New pattern.
28548 (rotlsi3_b): New pattern
28549 (rotlsi3_v850e3v5): New pattern.
28550 (doloop_begin): New pattern.
28551 (fix_loop_counter): New pattern.
28552 (doloop_end): New pattern.
28553 (branch_normal): Add e3v5 long branch support.
28554 (branch_invert): Likewise.
28555 (branch_z_normal): Likewise.
28556 (branch_z_invert): Likewise.
28557 (branch_nz_normal): Likewise.
28558 (branch_nz_invert): Likewise.
28559 (call_internal_short): Add e3v5 register-indirect JARL support.
28560 (call_internal_long): Likewise.
28561 (call_value_internal_short): Likewise.
28562 (call_value_internal_long): Likewise.
28563 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
28564 (mloop): New option.
28565 * config.gcc: Add support for configuring v840e3v5 target.
28566 * doc/invoke.texi: Document new v850 specific command line options.
28568 2013-01-31 Paul Koning <ni1d@arrl.net>
28572 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
28573 children if parent is a class.
28574 (prune_unused_types_prune): Don't add DW_AT_declaration.
28576 2013-01-31 Richard Biener <rguenther@suse.de>
28578 PR tree-optimization/56157
28579 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
28580 match up operand with SLP child.
28582 2013-01-31 Jason Merrill <jason@redhat.com>
28585 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
28586 parameters the first time.
28587 (gen_scheduled_generic_parms_dies): Check completeness here.
28589 2013-01-31 Richard Biener <rguenther@suse.de>
28591 PR middle-end/53073
28592 * common.opt (faggressive-loop-optimizations): New flag,
28593 enabled by default.
28594 * doc/invoke.texi (faggressive-loop-optimizations): Document.
28595 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
28596 infer_loop_bounds_from_undefined by it.
28598 2013-01-31 Richard Biener <rguenther@suse.de>
28600 PR tree-optimization/56150
28601 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
28602 visit virtual operands.
28603 (find_uses_to_rename_bb): Likewise.
28605 2013-01-31 Richard Biener <rguenther@suse.de>
28607 PR tree-optimization/56150
28608 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
28609 mixed store non-store stmts.
28611 2013-01-30 Jakub Jelinek <jakub@redhat.com>
28614 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
28615 LIBASAN_EARLY_SPEC is defined.
28616 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
28617 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
28619 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
28622 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
28623 invalid args instead of ICEing on it.
28624 (ix86_valid_target_attribute_tree): Return error_mark_node if
28625 ix86_valid_target_attribute_inner_p failed.
28626 (ix86_valid_target_attribute_p): Return false only if
28627 ix86_valid_target_attribute_tree returned error_mark_node. Allow
28628 target("default") attribute.
28629 (sorted_attr_string): Change argument from const char * to tree,
28630 merge in all target attribute arguments rather than just one.
28631 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
28632 instead of free. Avoid using strcat.
28633 (ix86_mangle_function_version_assembler_name): Mangle
28634 target("default") as if no target attribute is present. Adjust
28635 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
28636 instead of xmalloc and XDELETEVEC instead of free.
28637 (ix86_function_versions): Don't return true if one of the decls
28638 doesn't have target attribute. If they don't and one of the decls
28639 is DECL_FUNCTION_VERSIONED, report an error. Adjust
28640 sorted_attr_string caller. Use XDELETEVEC instead of free.
28641 (ix86_supports_function_versions): Remove.
28642 (make_name): Fix up formatting.
28643 (make_dispatcher_decl): Remove resolver_name and its initialization.
28644 Avoid leaking memory.
28645 (is_function_default_version): Return true if there is
28646 target("default") attribute rather than no target attribute at all.
28647 (make_resolver_func): Avoid leaking memory.
28648 (ix86_generate_version_dispatcher_body): Likewise.
28649 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
28650 * target.def (supports_function_versions): Remove.
28651 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
28652 * doc/tm.texi: Regenerated.
28654 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
28656 PR rtl-optimization/56144
28657 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
28658 for values with side effects.
28660 2013-01-30 Richard Biener <rguenther@suse.de>
28662 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
28663 (sparseset_pop): Likewise.
28664 * cfganal.c (compute_idf): Likewise. Increase work-stack size
28665 to be able to use quick_push in the worker loop.
28667 2013-01-30 Marek Polacek <polacek@redhat.com>
28669 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
28671 2013-01-30 Richard Biener <rguenther@suse.de>
28674 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
28676 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
28678 PR tree-optimization/56064
28679 * fixed-value.c (fixed_from_double_int): New function.
28680 * fixed-value.h (fixed_from_double_int): New prototype.
28681 (const_fixed_from_double_int): New static inline function.
28682 * fold-const.c (native_interpret_fixed): New static function.
28683 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
28684 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
28685 (native_encode_fixed): New static function.
28686 (native_encode_expr) <FIXED_CST>: Use it.
28687 (native_interpret_int): Move double_int worker code to...
28688 * double-int.c (double_int::from_buffer): ...this new static method.
28689 * double-int.h (double_int::from_buffer): Prototype it.
28691 2013-01-30 Richard Biener <rguenther@suse.de>
28693 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
28694 New pointer-map and obstack.
28695 (init_alias_vars): Allocate pointer-map and obstack.
28696 (delete_points_to_sets): Free them.
28697 (find_what_var_points_to): Cache result.
28698 (find_what_p_points_to): Adjust for changed interface of
28699 find_what_var_points_to.
28700 (compute_points_to_sets): Likewise.
28701 (ipa_pta_execute): Likewise.
28703 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
28705 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
28706 * configure: Regenerate.
28707 * config.in: Regenerate.
28708 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
28709 #nobits/#progbits if supported.
28711 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
28714 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
28715 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
28716 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
28718 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
28720 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
28721 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
28723 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
28725 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
28726 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
28728 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
28730 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
28732 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
28733 * config/arm/cortex-a7.md: New bypasses using
28734 arm_mac_accumulator_is_result.
28736 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
28738 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
28739 (cortex_a7_neon_mla): Likewise.
28740 (cortex_a7_fpfmad): New reservation.
28741 (cortex_a7_fpmacs): Use ffmas and update required units.
28742 (cortex_a7_fpmuld): Update required units and latency.
28743 (cortex_a7_fpmacd): Likewise.
28744 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
28745 (cortex_a7_neon). Likewise.
28746 (bypass) Update participating units.
28748 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
28750 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
28751 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
28753 * config/arm/vfp11.md (vfp_farith): Use ffmas.
28754 (vfp_fmul): Use ffmad.
28755 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
28756 (cortex_r4_fmacd): Use ffmad.
28757 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
28758 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
28759 (cortex_a9_fmacd): Use ffmad.
28760 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
28761 (cortex_a8_vfp_macd): Use ffmad.
28762 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
28763 (cortex_a5_fpmacd): Use ffmad.
28764 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
28765 (cortex_a15_vfp_macd): Use ffmad.
28766 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
28768 2013-01-29 Jason Merrill <jason@redhat.com>
28771 * varasm.c (default_assemble_visibility): Don't warn about
28772 visibility on artificial decls.
28774 2013-01-29 Richard Biener <rguenther@suse.de>
28776 PR tree-optimization/56113
28777 * tree-ssa-structalias.c (equiv_class_lookup): Also return
28779 (label_visit): Free duplicate bitmaps and record the leader instead.
28780 (perform_var_substitution): Adjust.
28782 2013-01-29 Richard Biener <rguenther@suse.de>
28784 PR tree-optimization/55270
28785 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
28786 the CFG, schedule loops for fixup.
28788 2013-01-29 Nick Clifton <nickc@redhat.com>
28790 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
28793 2013-01-28 Leif Ekblad <leif@rdos.net>
28795 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
28796 * config/i386/i386.h (TARGET_RDOS): New macro.
28797 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
28798 * config/i386/i386.c (ix86_option_override_internal): For 64bit
28799 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
28800 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
28801 DEFAULT_LARGE_SECTION_THRESHOLD.
28802 * config/i386/i386.md (R14_REG, R15_REG): New constants.
28803 * config/i386/rdos.h: New file.
28804 * config/i386/rdos64.h: New file.
28806 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
28809 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
28812 2013-01-28 Jakub Jelinek <jakub@redhat.com>
28814 PR rtl-optimization/56117
28815 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
28816 call cselib_lookup_from_insn on the MEM before calling
28817 add_insn_mem_dependence.
28819 2013-01-28 Richard Biener <rguenther@suse.de>
28821 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
28822 to a stmt that didn't have one.
28823 (copy_phis_for_bb): Likewise for PHI arguments.
28824 (copy_debug_stmt): Likewise for debug stmts.
28826 2013-01-28 Richard Biener <rguenther@suse.de>
28828 PR tree-optimization/56034
28829 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
28830 (partition_builtin_p): Adjust.
28831 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
28832 it is the last partition.
28833 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
28834 up the vertex for the definition.
28835 (classify_partition): Classify whether a partition is a
28836 PKIND_REDUCTION, thus has uses outside of the loop.
28837 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
28838 Merge all PKIND_REDUCTION partitions into the last partition.
28839 (tree_loop_distribution): Seed partitions from reductions as well.
28841 2013-01-28 Jakub Jelinek <jakub@redhat.com>
28843 PR tree-optimization/56125
28844 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
28845 pow(x,c) into sqrt(x) * powi(x, n/2) or
28846 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
28847 optimizing for size.
28848 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
28849 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
28852 PR tree-optimization/56094
28853 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
28854 to UNKNOWN_LOCATION while gimplifying expr.
28856 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
28859 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
28860 operand 0 in movabs insn template for -masm=intel asm alternative.
28861 (*movabs<mode>_2): Ditto for operand 1.
28863 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
28866 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
28869 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
28871 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
28874 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
28876 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
28877 Undef to avoid warning.
28879 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
28881 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
28882 * configure: Regenerate.
28884 2013-01-25 Jakub Jelinek <jakub@redhat.com>
28886 PR tree-optimization/56098
28887 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
28888 for stmts with volatile ops.
28889 (cond_store_replacement): Don't optimize if assign has volatile ops.
28890 (cond_if_else_store_replacement_1): Don't optimize if either
28891 then_assign or else_assign have volatile ops.
28892 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
28895 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
28897 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
28899 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
28901 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
28902 missing ':' in asm example.
28904 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
28906 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
28907 entries into lane and laneq entries.
28908 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
28909 Remove AdvSIMD scalar modes.
28910 (aarch64_sq<r>dmulh_laneq<mode>): New.
28911 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
28913 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
28914 builtin implementations to relfect changes in RTL in aarch64-simd.md.
28915 * config/aarch64/iterators.md (VCOND): New.
28918 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
28921 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
28922 Add NULL LIBNAME argument to existing definitions.
28923 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
28924 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
28925 * config/avr/avr.c (DEF_BUILTIN): Same.
28926 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
28927 (avr_expand_builtin): Expand to a vanilla call if a libgcc
28928 implementation is available (DECL_ASSEMBLER_NAME is set).
28929 (avr_fold_absfx): New static function.
28930 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
28931 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
28932 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
28933 AVR_BUILTIN_ABSLLK.
28934 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
28935 (abshk, absk, abslk, absllk): Provide as static inline functions.
28937 2013-01-25 Marek Polacek <polacek@redhat.com>
28939 PR tree-optimization/56035
28940 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
28942 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
28944 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
28945 (*movtf_internal_rex64): Add (!o,C) alternative
28946 (*movxf_internal_rex64): Ditto.
28947 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
28949 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
28951 * doc/invoke.texi: fix typo.
28952 * doc/objc.texi: fix typo.
28954 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
28956 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
28957 for the first two alternatives.
28959 2013-01-24 Diego Novillo <dnovillo@google.com>
28961 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
28962 (ggc-zone.o): Remove.
28963 * configure.ac: Remove option --with-gc.
28964 * configure: Re-generate.
28965 * doc/install.texi: Remove documentation for --with-gc.
28966 * gengtype.c (write_enum_defn): Remove. Update all users.
28967 (write_Types_process_field): Remove generation of gt_e_* argument.
28968 (output_type_enum): Remove. Update all users.
28969 (write_enum_defn): Remove. Update all users.
28970 (enum alloc_zone): Remove. Update all users.
28971 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
28972 * ggc-common.c (ggc_splay_alloc): Remove first argument.
28973 Update all callers.
28974 (struct ptr_data): Remove field TYPE. Update all users.
28975 (gt_pch_note_object): Remove argument TYPE. Update all users.
28976 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
28978 * ggc-none.c (ggc_alloc_typed_stat): Remove.
28979 (struct alloc_zone): Remove.
28980 (ggc_internal_alloc_zone_stat): Remove.
28981 (ggc_internal_cleared_alloc_zone_stat): Remove.
28982 * ggc-page.c (ggc_alloc_typed_stat): Remove.
28983 (ggc_pch_count_object): Remove last argument. Update all users.
28984 (ggc_pch_alloc_object): Remove last argument. Update all users.
28985 (struct alloc_zone): Remove.
28986 * ggc-zone.c: Remove.
28987 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
28988 (struct alloc_zone): Remove.
28989 (ggc_alloc_typed_stat): Remove.
28990 (ggc_alloc_typed): Remove.
28991 (ggc_splay_alloc): Remove first argument.
28992 (rtl_zone): Remove. Update all users.
28993 (tree_zone): Remove. Update all users.
28994 (tree_id_zone): Remove. Update all users.
28995 (ggc_internal_zone_alloc_stat): Remove. Update all users.
28996 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
28997 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
28998 * tree-ssanames.c: Remove references to zone allocator in comments.
29000 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
29002 * config/avr/avr.c (avr_out_fract): Make register numbers that
29003 might be outside of source operand signed.
29005 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
29007 * config/i386/constraints.md (Yf): New constraint.
29008 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
29009 of f constraint to conditionaly disable x87 register preferences.
29010 (*movdf_internal): Ditto.
29011 (*movsf_internal): Ditto.
29013 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
29015 PR inline-asm/55934
29016 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
29017 that have operands with impossible constraints.
29018 Add a FIXME for a speed-up opportunity.
29019 * lra-constraints.c (process_alt_operands): Verify that a class
29020 selected from constraints on asms is valid for the operand mode.
29021 (curr_insn_transform): Remove incorrect comment.
29023 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
29025 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
29026 TOC operand is a valid symbol ref in the constant pool.
29028 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
29030 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
29032 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
29035 * config/avr/stdfix.h: New file.
29036 * t-avr (stdfix-gcc.h): New rule to build it.
29037 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
29039 2013-01-23 Kostya Serebryany <kcc@google.com>
29041 * config/darwin.h: remove dependency on
29042 CoreFoundation (asan on Mac OS).
29044 2013-01-23 Jakub Jelinek <jakub@redhat.com>
29047 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
29048 instead of cmpdi_operand for first comparison operand.
29049 Don't assert that comparison operands aren't both constants.
29051 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
29053 * doc/install.texi (Downloading the Source): Update references to
29054 downloading separate components.
29056 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
29058 * doc/extend.texi (__int128): Improve grammar.
29060 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
29063 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
29064 alternative to (o,r).
29065 (*movdi_internal_rex64): Remove (!o,n) alternative.
29066 (DImode immediate->memory splitter): Remove.
29067 (DImode immediate->memory peephole2): Remove.
29068 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
29069 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
29070 alternative to (!o,*r).
29071 (*movtf_internal_sse): New pattern.
29072 (*movxf_internal_rex64): New pattern.
29073 (*movxf_internal): Disable for TARGET_64BIT.
29074 (*movdf_internal_rex64): Remove (!o,F) alternative.
29076 2013-01-22 Jakub Jelinek <jakub@redhat.com>
29078 PR middle-end/56074
29079 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
29080 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
29081 * tree-vect-loop-manip.c (find_loop_location): Also ignore
29082 stmt locations where LOCATION_LOCUS of the stmt location is
29083 UNKNOWN_LOCATION or BUILTINS_LOCATION.
29086 * config/i386/i386.md (UNSPEC_STOS): New.
29087 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
29088 *strsetqi_1): Add UNSPEC_STOS.
29090 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
29093 * doc/invoke.texi: Remove left over -Wsynth example.
29095 2013-01-21 Jakub Jelinek <jakub@redhat.com>
29097 PR tree-optimization/56051
29098 * fold-const.c (fold_binary_loc): Don't fold
29099 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
29100 a narrowing conversion, or widening conversion from signed
29103 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
29105 PR rtl-optimization/56023
29106 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
29107 dependent on debug instruction.
29109 2013-01-21 Martin Jambor <mjambor@suse.cz>
29111 PR middle-end/56022
29112 * function.c (allocate_struct_function): Call
29113 invoke_set_current_function_hook earlier.
29115 2013-01-21 Jakub Jelinek <jakub@redhat.com>
29117 * reload1.c (init_reload): Only initialize reload_obstack
29118 during the first call.
29120 2013-01-21 Marek Polacek <polacek@redhat.com>
29122 * cfgloop.c (verify_loop_structure): Fix up grammar.
29124 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
29126 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
29127 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
29129 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
29132 * config/arm/marvell-pj4.md: Update copyright year.
29133 Fix up use of alu to alu_reg and simple_alu_imm.
29135 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
29137 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
29139 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
29142 * lra-constraints.c (curr_insn_transform): Don't reuse original
29143 insn for secondary memory move when memory mode should be different.
29145 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
29147 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
29148 atomic_storedi_1): New patterns.
29150 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
29152 btver2 pipeline descriptions.
29153 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
29155 * config/i386/i386.md (btver2_decode): New type attributes.
29156 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
29158 * config/i386/btver2.md: New file describing btver2 pipelines.
29160 2013-01-19 Andrew Pinski <apinski@cavium.com>
29162 PR tree-optimization/52631
29163 * tree-ssa-sccvn (visit_use): Before looking up the original
29164 statement, try looking up the simplified expression.
29166 2013-01-19 Anthony Green <green@moxielogic.com>
29168 * config/moxie/moxie.c (moxie_expand_prologue): Set
29169 current_function_static_stack_size.
29171 2013-01-18 Jakub Jelinek <jakub@redhat.com>
29173 PR tree-optimization/56029
29174 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
29175 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
29177 2013-01-18 Sharad Singhai <singhai@google.com>
29179 PR tree-optimization/55995
29180 * dumpfile.c (dump_loc): Print location only if available.
29181 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
29183 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
29186 * lra-constraints.c (curr_insn_transform): Reuse original insn for
29187 secondary memory move.
29188 (inherit_reload_reg): Use rclass instead of cl for
29189 check_secondary_memory_needed_p.
29191 2013-01-18 Jakub Jelinek <jakub@redhat.com>
29193 PR middle-end/56015
29194 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
29195 the case where writing real complex part of target modifies op1.
29197 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
29199 * config/aarch64/aarch64-simd.md
29200 (aarch64_vcond_internal<mode>): Handle unordered cases.
29201 * config/aarch64/iterators.md (v_cmp_result): New.
29203 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
29204 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
29206 * config/arm/marvell-pj4.md: New file.
29207 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
29208 * config/arm/arm.md (generic_sched): Add marvell_pj4.
29209 (generic_vfp): Likewise.
29210 * config/arm/arm-cores.def: Add marvell-pj4.
29211 * config/arm/arm-tune.md: Regenerate.
29212 * config/arm/arm-tables.opt: Regenerate.
29213 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
29214 * doc/invoke.texi: Document marvell-pj4.
29216 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
29218 * config/aarch64/arm_neon.h: Map scalar types to standard types.
29220 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
29225 * var-tracking.c (negative_power_of_two_p): New.
29226 (global_get_addr_cache, local_get_addr_cache): New.
29227 (get_addr_from_global_cache, get_addr_from_local_cache): New.
29228 (vt_canonicalize_addr): Rewrite using the above. Adjust the
29230 (vt_stack_offset_p): Remove.
29231 (vt_canon_true_dep): Always canonicalize loc's address.
29232 (clobber_overlapping_mems): Make sure we have a MEM.
29233 (local_get_addr_clear_given_value): New.
29234 (val_reset): Clear local cached entries.
29235 (compute_bb_dataflow): Create and release the local cache.
29236 Disable duplicate MEMs clobbering.
29237 (emit_notes_in_bb): Clobber MEMs likewise.
29238 (vt_emit_notes): Create and release the local cache.
29239 (vt_initialize, vt_finalize): Create and release the global
29240 cache, respectively.
29241 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
29243 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
29245 PR libmudflap/53359
29246 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
29247 not found in the symtab.
29249 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
29252 PR rtl-optimization/55547
29253 PR rtl-optimization/53827
29256 * alias.c (offset_overlap_p): New, factored out of...
29257 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
29258 the conservative special case for symbolic constants. Don't
29259 adjust zero sizes on alignment.
29261 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
29263 PR rtl-optimization/52573
29264 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
29265 REG_UNUSED for the same register.
29267 2013-01-17 Richard Biener <rguenther@suse.de>
29268 Marek Polacek <polacek@redhat.com>
29270 PR rtl-optimization/55833
29271 * loop-unswitch.c (unswitch_loops): Move loop verification...
29272 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
29273 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
29274 Set it to true when we're removing a loop from hierarchy tree in
29275 an irreducible region.
29276 (fix_bb_placements): Adjust caller.
29277 (fix_loop_placements): Likewise.
29279 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
29281 * config/avr/builtins.def (DEF_BUILTIN): Factor out
29282 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
29283 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
29284 Remove ID. Adjust comments.
29285 * config/avr/avr-c.c (avr_builtin_name): Remove.
29286 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
29287 * config/avr/avr.c (avr_tolower): New static function.
29288 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
29289 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
29290 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
29293 2013-01-17 Jan Hubicka <jh@suse.cz>
29295 PR tree-optimization/55273
29296 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
29298 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
29301 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
29302 store through atomic_store<mode>_1.
29303 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
29305 2013-01-17 Martin Jambor <mjambor@suse.cz>
29307 PR tree-optimizations/55264
29308 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
29309 for virtual methods.
29310 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
29311 virtual methods before inlining is over.
29312 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
29314 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
29317 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
29319 PR rtl-optimization/56005
29320 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
29321 pending reads for prefetch.
29323 2013-01-16 Ian Bolton <ian.bolton@arm.com>
29325 * config/aarch64/aarch64.md
29326 (*cstoresi_neg_uxtw): New pattern.
29327 (*cmovsi_insn_uxtw): New pattern.
29328 (*<optab>si3_uxtw): New pattern.
29329 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
29330 (*<optab>si3_insn_uxtw): New pattern.
29331 (*bswapsi2_uxtw): New pattern.
29333 2013-01-16 Richard Biener <rguenther@suse.de>
29335 * tree-inline.c (tree_function_versioning): Remove set but
29336 never used variable.
29338 2013-01-16 Richard Biener <rguenther@suse.de>
29340 PR tree-optimization/55964
29341 * tree-flow.h (rename_variables_in_loop): Remove.
29342 (rename_variables_in_bb): Likewise.
29343 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
29344 (copy_loop_before): Adjust and delete update-ssa status.
29345 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
29346 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
29347 (rename_variables_in_loop): Remove.
29348 (slpeel_update_phis_for_duplicate_loop): Likewise.
29349 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
29350 use available cfg machinery instead of duplicating it.
29351 Update PHI nodes and perform poor-mans SSA update here.
29352 (slpeel_tree_peel_loop_to_edge): Adjust.
29354 2013-01-16 Richard Biener <rguenther@suse.de>
29356 PR tree-optimization/54767
29357 PR tree-optimization/53465
29358 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
29359 (vrp_visit_phi_node): For PHI arguments coming via backedges
29360 drop all symbolical range information.
29361 (execute_vrp): Compute backedges.
29363 2013-01-16 Richard Biener <rguenther@suse.de>
29365 * doc/install.texi: Update CLooG and ISL requirements to
29368 2013-01-16 Christian Bruel <christian.bruel@st.com>
29371 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
29372 (broken_move): Handle UNSPECV_SP_SWITCH_B.
29373 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
29375 2013-01-16 DJ Delorie <dj@redhat.com>
29377 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
29378 (UNSPECV_SP_SWITCH_E): New.
29379 (sp_switch_1): Change to an unspec.
29380 (sp_switch_2): Change to an unspec. Don't use post-inc when we
29383 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
29385 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
29386 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
29387 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
29388 (expand_mem_thread_fence): Ditto.
29389 (expand_mem_signal_fence): Ditto.
29390 (expand_atomic_load): Ditto.
29391 (expand_atomic_store): Ditto.
29393 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
29395 PR rtl-optimization/55547
29396 PR rtl-optimization/53827
29399 * alias.c (memrefs_conflict_p): Set sizes to negative after
29402 2013-01-15 Jakub Jelinek <jakub@redhat.com>
29405 * function.c (thread_prologue_and_epilogue_insns): Always
29406 add crtl->drap_reg to set_up_by_prologue.set, even if
29407 stack_realign_drap is false.
29409 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
29411 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
29412 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
29413 *call): Fix indention.
29415 2013-01-15 Tom de Vries <tom@codesourcery.com>
29418 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
29421 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
29423 PR rtl-optimization/55153
29424 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
29426 2013-01-15 Martin Jambor <mjambor@suse.cz>
29428 PR tree-optimization/55920
29429 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
29430 accesses as grp_to_be_debug_replaced.
29432 2013-01-15 Jakub Jelinek <jakub@redhat.com>
29434 PR tree-optimization/55920
29435 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
29436 there is non-useless type conversion needed from debug rhs to lhs,
29437 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
29439 2013-01-15 Joseph Myers <joseph@codesourcery.com>
29440 Mikael Pettersson <mikpe@it.uu.se>
29443 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
29445 (ASM_OUTPUT_CASE_LABEL): Remove.
29446 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
29447 * final.c (shorten_branches): Update alignment of labels before
29448 jump tables if CASE_VECTOR_SHORTEN_MODE.
29450 2013-01-15 Richard Biener <rguenther@suse.de>
29453 * system.h: Do not include gmp.h for building host tools.
29455 2013-01-15 Richard Biener <rguenther@suse.de>
29457 PR middle-end/55882
29458 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
29459 account for bitpos when computing alignment.
29461 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
29463 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
29464 (ix86_target_macros_internal): Likewise.
29466 * config/i386/i386.c (m_CORE2I7): Removed.
29467 (m_CORE_HASWELL): New macro.
29468 (m_CORE_ALL): Likewise.
29469 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
29470 (initial_ix86_arch_features): Likewise.
29471 (processor_target_table): Initializations for Core avx2.
29472 (cpu_names): New names "core-avx2".
29473 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
29474 PROCESSOR_CORE_HASWELL.
29475 (ix86_issue_rate): New case.
29476 (ia32_multipass_dfa_lookahead): Likewise.
29477 (ix86_sched_init_global): Likewise.
29479 * config/i386/i386.h (TARGET_HASWELL): New macro.
29480 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
29481 (processor_type): New PROCESSOR_HASWELL.
29483 2013-01-15 Jakub Jelinek <jakub@redhat.com>
29485 PR tree-optimization/55955
29486 * tree-vect-loop.c (vectorizable_reduction): Give up early on
29487 *SHIFT_EXPR and *ROTATE_EXPR codes.
29489 PR tree-optimization/48766
29490 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
29491 -ftrapv disable -fwrapv.
29493 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
29496 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
29497 etc. to 1 and not to __flash.
29498 Use LL suffix for __INT24_MAX__ with -mint8.
29499 Use ULL suffix for __UINT24_MAX__ with -mint8.
29501 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
29503 * config/avr/avr-arch.h
29504 (struct base_arch_s): Use typedef avr_arch_t instead.
29505 (struct arch_info_s): Use typedef avr_arch_info_t instead.
29506 (struct mcu_type_s): Use typedef avr_mcu_t instead.
29507 * config/avr/avr.c: Same.
29508 * config/avr/avr-devices.c: Same.
29509 * config/avr/driver-avr.c: Same.
29510 * config/avr/gen-avr-mmcu-texi.c: Same.
29511 * config/avr/avr-mcus.def: Adjust comment.
29513 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
29515 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
29516 * config/aarch64/iterators.md (VALLDI): New.
29518 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
29519 Andi Kleen <ak@linux.intel.com>
29522 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
29523 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
29526 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
29528 * config/avr/avr-stdint.h: Remove trailing blanks.
29529 * config/avr/avr-log.h: Same.
29530 * config/avr/avr-arch.h: Same.
29531 * config/avr/avr-devices.c: Same.
29532 * config/avr/avr-dimode.md: Same.
29533 * config/avr/predicates.md: Same.
29534 * config/avr/avr-c.c: Same. And fix typo.
29536 * config/avr/avr-protos.h: Same. And:
29537 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
29538 (init_cumulative_args): Rename to avr_init_cumulative_args.
29539 (expand_prologue): Rename to avr_expand_prologue.
29540 (expand_epilogue): Rename to avr_expand_epilogue.
29541 (adjust_insn_length): Rename to avr_adjust_insn_length.
29542 (notice_update_cc): Rename to avr_notice_update_cc.
29543 (final_prescan_insn): Rename to avr_final_prescan_insn.
29544 * config/avr/avr.c: Same.
29545 * config/avr/avr.h: Same.
29546 * config/avr/avr.md: Remove trailing blanks.
29547 (prologue): Use avr_expand_prologue.
29548 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
29550 2013-01-14 Richard Biener <rguenther@suse.de>
29552 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
29553 verify_location, collect_subblocks): New functions.
29554 (verify_gimple_in_cfg): Verify that locations only reference
29555 BLOCKs in the functions BLOCK tree.
29557 2013-01-14 Richard Biener <rguenther@suse.de>
29559 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
29561 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
29563 (insert_out_of_ssa_copy_on_edge): Likewise.
29564 (rewrite_close_phi_out_of_ssa): Likewise.
29565 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
29567 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
29568 propagated constants.
29569 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
29572 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
29574 * config/avr/avr-modes.def: Add GPL copyright notice.
29576 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
29578 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
29579 MEMMODEL_MASK to determine memory model.
29580 (atomic_store<mode>): Ditto from operands[2].
29581 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
29583 2013-01-13 Jakub Jelinek <jakub@redhat.com>
29586 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
29587 (fold_gimple_assign): Don't call unshare_expr here.
29588 (fold_ctor_reference): Call unshare_expr.
29590 2013-01-13 Terry Guo <terry.guo@arm.com>
29592 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
29593 * doc/fragments.texi: Document MULTILIB_REUSE.
29594 * gcc.c (multilib_reuse): New internal spec.
29595 (set_multilib_dir): Also search multilib from multilib_reuse.
29596 * genmultilib (tmpmultilib3): Refactor code.
29597 (tmpmultilib4): Ditto.
29598 (multilib_reuse): New multilib argument.
29600 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
29602 * Makefile.in: Update copyright.
29604 2013-01-12 Tom de Vries <tom@codesourcery.com>
29606 PR middle-end/55890
29607 * calls.c (expand_call): Check if arg_nr is valid.
29609 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
29611 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
29612 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
29613 documentation. Add missing '__' in front of
29614 __builtin_ia32_packssdw256.
29616 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
29619 * config/s390/s390.c (s390_preferred_reload_class): Do not return
29620 NO_REGS for larl operands.
29621 (s390_reload_larl_operand): Use s390_load_address instead of
29624 2013-01-11 Richard Biener <rguenther@suse.de>
29626 * tree-cfg.c (verify_node_sharing_1): Split out from ...
29627 (verify_node_sharing): ... here.
29628 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
29630 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
29632 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
29633 Substitute TREECHECKING.
29634 * configure: Regenerate.
29635 * Makefile.in (TREECHECKING): New.
29637 2013-01-11 Richard Guenther <rguenther@suse.de>
29639 PR tree-optimization/44061
29640 * tree-vrp.c (extract_range_basic): Compute zero as
29641 value-range for __builtin_constant_p of function parameters.
29643 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
29645 Update copyright years.
29647 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
29649 PR rtl-optimization/55672
29650 * lra-eliminations.c (mark_not_eliminable): Permit addition with
29651 const to be eliminable.
29653 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
29655 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
29656 * configure: Regenerate.
29658 2013-01-10 Richard Biener <rguenther@suse.de>
29660 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
29662 2013-01-10 Richard Biener <rguenther@suse.de>
29665 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
29666 locations for virtual PHI arguments.
29667 (rewrite_update_phi_arguments): Likewise.
29669 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
29671 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
29674 2013-01-10 Jakub Jelinek <jakub@redhat.com>
29676 PR tree-optimization/55921
29677 * tree-complex.c (expand_complex_asm): New function.
29678 (expand_complex_operations_1): Call it for GIMPLE_ASM.
29680 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
29683 * config/s390/s390.c (s390_symref_operand_p)
29684 (s390_loadrelative_operand_p): Merge the two functions.
29685 (s390_check_qrst_address, print_operand_address): Add parameters
29686 to s390_loadrelative_operand_p invokation.
29687 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
29688 (s390_reload_larl_operand, s390_secondary_reload): Use
29689 s390_loadrelative_operand_p instead of s390_symref_operand_p.
29690 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
29692 2013-01-09 Mike Stump <mikestump@comcast.net>
29694 * dse.c (record_store): Remove unnecessary assert.
29696 2013-01-09 Jan Hubicka <jh@suse.cz>
29698 PR tree-optimization/55569
29699 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
29700 * cfgloop.h (scale_loop_profile): Likewise.
29702 2013-01-09 Jan Hubicka <jh@suse.cz>
29705 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
29707 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
29709 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
29711 PR middle-end/55114
29712 * expr.h (maybe_emit_group_store): Declare.
29713 * expr.c (maybe_emit_group_store): New function.
29714 * builtins.c (expand_builtin_int_roundingfn): Call it.
29715 (expand_builtin_int_roundingfn_2): Likewise.
29717 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
29719 PR rtl-optimization/55829
29720 * lra-constraints.c (match_reload): Add code for absent output.
29721 (curr_insn_transform): Add code for reloads of matched inputs
29724 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
29726 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
29727 attribute of movddup insn to DF.
29728 (*vec_interleave_lowv2df): Ditto.
29729 (vec_dupv2df): Ditto.
29731 2013-01-09 Jan Hubicka <jh@suse.cz>
29733 PR tree-optimiation/55875
29734 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
29735 EVERY_ITERATION parameter.
29736 (number_of_iterations_exit): Check if exit is executed every iteration.
29737 (idx_infer_loop_bounds): Similarly here.
29738 (n_of_executions_at_most): Simplify
29739 to only test for cases where statement is dominated by the
29740 particular bound; handle correctly the "postdominance" test.
29741 (scev_probably_wraps_p): Use max loop iterations info
29742 as a global bound first.
29744 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
29745 Nick Clifton <nickc@redhat.com>
29747 * config/v850/v850.md (cbranchsf4): New pattern.
29748 (cstoresf4): New pattern.
29749 (cbranchdf4): New pattern.
29750 (cstoredf4): New pattern.
29751 (movsicc): Disallow floating point comparisons.
29752 (cmpsf_le_insn): Fix order of operators.
29753 (cmpsf_lt_insn): Likewise.
29754 (cmpsf_eq_insn): Likewise.
29755 (cmpdf_le_insn): Likewise.
29756 (cmpdf_lt_insn): Likewise.
29757 (cmpdf_eq_insn): Likewise.
29758 (cmpsf_ge_insn): Use LE comparison.
29759 (cmpdf_ge_insn): Likewise.
29760 (cmpsf_gt_insn): Use LT comparison.
29761 (cmpdf_gt_insn): Likewise.
29762 (cmpsf_ne_insn): Delete pattern.
29763 (cmpdf_ne_insn): Delete pattern.
29764 * config/v850/v850.c (v850_gen_float_compare): Use
29765 gen_cmpdf_eq_insn for NE comparison.
29766 (v850_float_z_comparison_operator)
29767 (v850_float_nz_comparison_operator): Move from here ...
29768 * config/v850/predicates.md: ... to here. Move GT and GE
29769 comparisons into v850_float_z_comparison_operator.
29770 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
29772 (v850_float_nz_comparison_operator): Likewise.
29774 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
29776 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
29777 with calls to gen_insvsi/gen_insvdi.
29779 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
29781 * config/i386/i386.c (initial_ix86_tune_features): Set up
29782 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
29784 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
29785 Jakub Jelinek <jakub@redhat.com>
29787 PR tree-optimization/48189
29788 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
29789 If nitercst is 0, don't predict the exit edge.
29791 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
29793 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
29794 in asm_fprintf with reg_names.
29795 (aarch64_print_operand_address): Likewise.
29796 (aarch64_return_addr): Likewise.
29797 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
29799 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
29801 * config/pa/pa.h (VAL_U6_BITS_P): Define.
29802 (INT_U6_BITS): Likewise.
29803 * config/pa/predicates.md (uint6_operand): New predicate.
29804 (shift5_operand, shift6_operand): Likewise.
29805 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
29807 (lshrdi3): Use shift6_operand.
29808 (shrpsi4, shrpdi4): New insn patterns.
29809 (extzv): Delete expander.
29810 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
29811 predicates in unamed zero extract patterns. Tighten common constraint.
29812 (extv): Delete expander.
29813 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
29814 predicates in unamed sign extract patterns. Tighten common constraint.
29815 (insv): Delete expander.
29816 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
29817 predicates in unamed insert patterns. Tighten common constraint.
29818 Change uint32_operand predicate to uint6_operand predicate in unamed
29819 DImode pattern to insert constant values of type 1...1xxxx.
29821 2013-01-04 Jan Hubicka <jh@suse.cz>
29823 PR tree-optimization/55823
29824 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
29827 2013-01-08 Jakub Jelinek <jakub@redhat.com>
29828 Uros Bizjak <ubizjak@gmail.com>
29830 PR rtl-optimization/55845
29831 * df-problems.c (can_move_insns_across): Stop scanning at
29832 volatile_insn_p source instruction or give up if
29833 across_from .. across_to range contains any volatile_insn_p
29836 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
29838 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
29839 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
29841 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
29842 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
29844 2013-01-08 Jakub Jelinek <jakub@redhat.com>
29847 * asan.c (asan_clear_shadow): New function.
29848 (asan_emit_stack_protection): Use it.
29850 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
29852 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
29853 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
29854 with tab instead of space.
29856 2013-01-08 Nick Clifton <nickc@redhat.com>
29858 * config/rl78/rl78.c (rl78_expand_prologue): Always select
29859 register bank 0 at the start of an interrupt handler.
29860 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
29863 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
29865 * config/aarch64/aarch64-simd.md
29866 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
29867 (aarch64_simd_bsl): Likewise.
29868 (aarch64_vcond_internal<mode>): Likewise.
29869 (vcond<mode><mode>): Likewise.
29870 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
29871 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
29873 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
29875 * config/aarch64/aarch64-builtins.c
29876 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
29878 2013-01-08 Martin Jambor <mjambor@suse.cz>
29881 * tree-sra.c (analyze_access_subtree): Return true also after
29882 potentially creating a debug-only replacement.
29884 2013-01-08 Jakub Jelinek <jakub@redhat.com>
29886 PR middle-end/55890
29887 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
29889 PR tree-optimization/54120
29890 * tree-vrp.c (range_fits_type_p): Don't allow
29891 src_precision < precision from signed vr to unsigned_p
29892 if vr->min or vr->max is negative.
29893 (simplify_float_conversion_using_ranges): Test can_float_p
29894 against CODE_FOR_nothing.
29896 2013-01-08 Jakub Jelinek <jakub@redhat.com>
29897 Richard Biener <rguenther@suse.de>
29899 PR middle-end/55851
29900 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
29901 types instead of just INTEGER_TYPE types.
29903 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
29905 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
29908 2013-01-07 Steve Ellcey <sellcey@mips.com>
29911 * config/mips/mips.opt: Change mad to mmad to match documentation.
29913 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
29916 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
29917 .progmemx.data now.
29919 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
29922 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
29923 (avr_addrspace_t): Add .section_name field.
29924 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
29926 (avr_addrspace): Same. Initialize .section_name. Remove last
29927 NULL entry. Put __memx into .progmemx.data.
29928 (progmem_section_prefix): Remove.
29929 (avr_asm_init_sections): No need to initialize progmem_section.
29930 (avr_asm_named_section): Use avr_addrspace[].section_name to get
29931 section name prefix.
29932 (avr_asm_select_section): Ditto. And use get_unnamed_section to
29933 retrieve the progmem section.
29934 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
29935 boundary to run over avr_addrspace[].
29936 (avr_register_target_pragmas): Ditto.
29938 2013-01-06 Jakub Jelinek <jakub@redhat.com>
29940 * varasm.c (output_constant_def_contents): For asan_protect_global
29941 protected strings, adjust DECL_ALIGN if needed, before testing for
29943 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
29944 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
29946 (output_object_block): For asan protected decls, emit asan padding
29947 after their contents.
29948 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
29949 (asan_finish_file): Test it here instead.
29951 2013-01-07 Nick Clifton <nickc@redhat.com>
29952 Matthias Klose <doko@debian.org>
29953 Doug Kwan <dougkwan@google.com>
29954 H.J. Lu <hongjiu.lu@intel.com>
29957 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
29959 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
29961 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
29963 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
29965 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
29967 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
29970 * doc/install.texi (Cross-Compiler-Specific Options): Document
29973 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
29975 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
29976 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
29977 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
29978 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
29979 vqmovun_high_s64): Fix source operand number and update copyright.
29981 2013-01-07 Richard Biener <rguenther@suse.de>
29983 PR middle-end/55890
29984 * gimple.h (gimple_call_builtin_p): New overload.
29985 * gimple.c (validate_call): New function.
29986 (gimple_call_builtin_p): Likewise.
29987 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
29988 Use gimple_call_builtin_p.
29989 (find_func_clobbers): Likewise.
29990 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
29991 (strlen_optimize_stmt): Likewise.
29993 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
29995 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
29996 (vld1q_dup_*): Likewise.
29997 (vld1_*): Likewise.
29998 (vld1q_*): Likewise.
29999 (vld1_lane_*): Likewise.
30000 (vld1q_lane_*): Likewise.
30002 2013-01-07 Richard Biener <rguenther@suse.de>
30004 * lto-streamer.h (LTO_minor_version): Bump to 2.
30006 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
30008 * config/aarch64/aarch64-protos.h
30009 (aarch64_const_double_zero_rtx_p): Rename to...
30010 (aarch64_float_const_zero_rtx_p): ...this.
30011 (aarch64_float_const_representable_p): New.
30012 (aarch64_output_simd_mov_immediate): Likewise.
30013 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
30014 move immediate case.
30015 * config/aarch64/aarch64.c
30016 (aarch64_const_double_zero_rtx_p): Rename to...
30017 (aarch64_float_const_zero_rtx_p): ...this.
30018 (aarch64_print_operand): Allow printing of new constants.
30019 (aarch64_valid_floating_const): New.
30020 (aarch64_legitimate_constant_p): Check for valid floating-point
30022 (aarch64_simd_valid_immediate): Likewise.
30023 (aarch64_vect_float_const_representable_p): New.
30024 (aarch64_float_const_representable_p): Likewise.
30025 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
30026 (aarch64_output_simd_mov_immediate): New.
30027 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
30028 (*movdf_aarch64): Likewise.
30029 * config/aarch64/constraints.md (Ufc): New.
30030 (Y): call aarch64_float_const_zero_rtx.
30031 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
30033 2013-01-07 Richard Biener <rguenther@suse.de>
30035 PR tree-optimization/55888
30036 PR tree-optimization/55862
30037 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
30038 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
30039 not if it is contained therein.
30041 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
30043 * config/avr/t-avr: Typo.
30045 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
30048 * config/avr/t-avr: Don't automatically rebuild
30049 $(srcdir)/config/avr/t-multilib
30050 $(srcdir)/config/avr/avr-tables.opt
30051 $(srcdir)/doc/avr-mmcu.texi
30052 (avr-mcus): New phony target to build them on request.
30053 (s-avr-mlib, s-avr-mmcu-texi): Remove.
30054 * avr/avr-mcus.def: Adjust comments.
30056 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
30058 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
30060 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
30062 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
30064 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
30066 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
30068 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
30070 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
30071 to generate profiling.
30072 * config/rs6000/aix64.h (LIB_SPEC): Same.
30074 2013-01-04 Andrew Pinski <apinski@cavium.com>
30076 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
30078 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
30080 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
30082 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
30084 (ix86_expand_move): Ditto.
30085 (ix86_zero_extend_to_Pmode): Ditto.
30086 (ix86_expand_call): Ditto.
30087 (ix86_expand_special_args_builtin): Ditto.
30088 (ix86_expand_builtin): Ditto.
30090 2013-01-04 Richard Biener <rguenther@suse.de>
30092 PR tree-optimization/55862
30093 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
30094 translating them through PHI nodes.
30096 2013-01-04 Martin Jambor <mjambor@suse.cz>
30098 PR tree-optimization/55755
30099 * tree-sra.c (sra_modify_assign): Do not check that an access has no
30100 children when trying to avoid producing a VIEW_CONVERT_EXPR.
30102 2013-01-04 Marek Polacek <polacek@redhat.com>
30104 PR middle-end/55859
30105 * opts.c (default_options_optimization): Clarify error message.
30107 2013-01-04 Richard Biener <rguenther@suse.de>
30109 PR middle-end/55863
30110 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
30113 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
30116 * config/pa/pa.md (movsi): Revert previous change.
30117 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
30120 2013-01-03 Richard Henderson <rth@redhat.com>
30122 * config/i386/i386.c (ix86_expand_move): Always assign to op1
30123 after eliminating TLS symbols.
30125 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
30128 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
30129 * graphite-poly.c (debug_gmp_value): Likewise.
30131 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
30134 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
30135 selected code model, define __code_mode_small__, __code_model_medium__,
30136 __code_model_large__, __code_model_32__ or __code_model_kernel__.
30137 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
30138 xchg temporary register with %k. Declare temporary register as
30140 [__x86_64__]: For medium and large code models, preserve %rbx register.
30142 2013-01-03 Richard Biener <rguenther@suse.de>
30144 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
30145 (dump_subscript): Adjust.
30146 (finalize_ddr_dependent): Do not dump redundant info.
30147 (analyze_siv_subscript): Adjust.
30148 (subscript_dependence_tester): Likewise.
30149 (compute_affine_dependence): Likewise.
30151 2013-01-03 Richard Biener <rguenther@suse.de>
30154 2013-01-03 Richard Biener <rguenther@suse.de>
30156 PR tree-optimization/55857
30157 * tree-vect-stmts.c (vectorizable_load): Do not setup
30158 re-alignment for invariant loads.
30160 2013-01-02 Richard Biener <rguenther@suse.de>
30162 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
30163 invariant load do not generate a vector load from the scalar location.
30165 2013-01-03 Richard Biener <rguenther@suse.de>
30167 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
30168 for not vectorizing.
30169 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
30170 not build INDIRECT_REFs, call get_name once only.
30171 (vect_create_data_ref_ptr): Likewise. Dump base object kind
30172 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
30174 2013-01-03 Richard Biener <rguenther@suse.de>
30176 PR tree-optimization/55857
30177 * tree-vect-stmts.c (vectorizable_load): Do not setup
30178 re-alignment for invariant loads.
30180 2013-01-03 Richard Biener <rguenther@suse.de>
30183 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
30184 prefer a built-in decl.
30186 2013-01-03 Jakub Jelinek <jakub@redhat.com>
30188 * gcc.c (process_command): Update copyright notice dates.
30189 * gcov.c (print_version): Likewise.
30190 * gcov-dump.c (print_version): Likewise.
30192 PR rtl-optimization/55838
30193 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
30194 iv0.step, iv1.step and step.
30196 2013-01-03 Jakub Jelinek <jakub@redhat.com>
30197 Marc Glisse <marc.glisse@inria.fr>
30199 PR tree-optimization/55832
30200 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
30201 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
30202 integer_{one,zero}_node.
30204 2013-01-03 Jakub Jelinek <jakub@redhat.com>
30207 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
30208 * var-tracking.c (reverse_op): Don't add reverse ops to
30209 VALUEs that have already
30210 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
30212 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
30214 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
30216 2013-01-02 Teresa Johnson <tejohnson@google.com>
30218 * dumpfile.c (dump_loc): Print filename with location.
30219 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
30220 new location_t parameter to emit complete unroll message with
30221 new dump framework.
30222 (canonicalize_loop_induction_variables): Compute loops location
30223 and pass to try_unroll_loop_completely.
30224 * loop-unroll.c (report_unroll_peel): New function.
30225 (peel_loops_completely): Use new dump format with location
30226 for main dumpfile message, and invoke report_unroll_peel on success.
30227 (decide_unrolling_and_peeling): Ditto.
30228 (decide_peel_once_rolling): Remove old dumpfile message subsumed
30229 by report_unroll_peel.
30230 (decide_peel_completely): Ditto.
30231 (decide_unroll_constant_iterations): Ditto.
30232 (decide_unroll_runtime_iterations): Ditto.
30233 (decide_peel_simple): Ditto.
30234 (decide_unroll_stupid): Ditto.
30235 * cfgloop.c (get_loop_location): New function.
30236 * cfgloop.h (get_loop_location): Declare.
30238 2013-01-02 Sriraman Tallam <tmsriram@google.com>
30240 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
30243 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
30245 PR middle-end/55198
30246 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
30247 BLKmode objects when EXPAND_MEMORY is specified.
30249 2013-01-02 Sriraman Tallam <tmsriram@google.com>
30251 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
30253 (fold_builtin_cpu): Do not share cpu model decls across statements.
30255 2013-01-02 Jason Merrill <jason@redhat.com>
30258 * tree.c (build_array_type_1): Revert earlier change.
30260 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
30262 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
30264 * config/aarch64/aarch64-tune.md: Re-generate.
30266 2013-01-02 Richard Biener <rguenther@suse.de>
30268 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
30269 invariant load do not generate a vector load from the scalar location.
30271 2013-01-02 Richard Biener <rguenther@suse.de>
30274 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
30275 * configure: Regenerate.
30277 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
30279 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
30280 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
30281 (expand_builtin_int_roundingfn_2): Keep the original target around
30282 for the fallback case.
30284 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
30286 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
30287 to be clear for sign changes.
30289 2013-01-01 Jan Hubicka <jh@suse.cz>
30291 * ipa-inline-analysis.c: Fix formatting.
30293 2013-01-01 Jakub Jelinek <jakub@redhat.com>
30295 PR tree-optimization/55831
30296 * tree-vect-loop.c (get_initial_def_for_induction): Use
30297 gsi_after_labels instead of gsi_start_bb.
30299 Copyright (C) 2013 Free Software Foundation, Inc.
30301 Copying and distribution of this file, with or without modification,
30302 are permitted in any medium without royalty provided the copyright
30303 notice and this notice are preserved.